Necrotizing Soft Tissue Infections: Clinical and Pathogenic Aspects [1st ed.] 9783030576158, 9783030576165

Citation preview

Advances in Experimental Medicine and Biology 1294

Anna Norrby-Teglund Mattias Svensson Steinar Skrede  Editors

Necrotizing Soft Tissue Infections Clinical and Pathogenic Aspects

Advances in Experimental Medicine and Biology Volume 1294 Series Editors Wim E. Crusio, Institut de Neurosciences Cognitives et Intégratives d’Aquitaine, CNRS and University of Bordeaux, Pessac Cedex, France Haidong Dong, Departments of Urology and Immunology, Mayo Clinic, Rochester, Minnesota, USA Heinfried H. Radeke, Institute of Pharmacology & Toxicology, Clinic of the Goethe University Frankfurt Main, Frankfurt am Main, Hessen, Germany Nima Rezaei, Research Center for Immunodeficiencies, Children’s Medical Center, Tehran University of Medical Sciences, Tehran, Iran Junjie Xiao, Cardiac Regeneration and Ageing Lab, Institute of Cardiovascular Science, School of Life Science, Shanghai University, Shanghai, China

Advances in Experimental Medicine and Biology provides a platform for scientific contributions in the main disciplines of the biomedicine and the life sciences. This series publishes thematic volumes on contemporary research in the areas of microbiology, immunology, neurosciences, biochemistry, biomedical engineering, genetics, physiology, and cancer research. Covering emerging topics and techniques in basic and clinical science, it brings together clinicians and researchers from various fields. Advances in Experimental Medicine and Biology has been publishing exceptional works in the field for over 40 years, and is indexed in SCOPUS, Medline (PubMed), Journal Citation Reports/Science Edition, Science Citation Index Expanded (SciSearch, Web of Science), EMBASE, BIOSIS, Reaxys, EMBiology, the Chemical Abstracts Service (CAS), and Pathway Studio. 2019 Impact Factor: 2.450 5 Year Impact Factor: 2.324

More information about this series at http://www.springer.com/series/5584

Anna Norrby-Teglund • Mattias Svensson • Steinar Skrede Editors

Necrotizing Soft Tissue Infections Clinical and Pathogenic Aspects

Editors Anna Norrby-Teglund Center for Infectious Medicine, Department of Medicine Karolinska Institutet Huddinge, Sweden

Mattias Svensson Center for Infectious Medicine, Department of Medicine Karolinska Institutet Huddinge, Sweden

Steinar Skrede Department of Clinical Science University of Bergen Bergen, Norway Department of Medicine Haukeland University Hospital Bergen, Norway

ISSN 0065-2598 ISSN 2214-8019 (electronic) Advances in Experimental Medicine and Biology ISBN 978-3-030-57615-8 ISBN 978-3-030-57616-5 (eBook) https://doi.org/10.1007/978-3-030-57616-5 # Springer Nature Switzerland AG 2020 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Preface

During the last four decades, there has been a worldwide upsurge of severe skin and soft tissue infections. These infections are frequently associated with severe complications such as septic shock and multi-organ failure, resulting in need for intensive care and challenging rehabilitation. Unfortunately, in too many cases the outcome is loss of life or limbs. This book is devoted entirely to these severe and rapidly spreading necrotizing soft tissue infections. It includes 12 chapters ranging from case reports written by relatives or survivors providing the patient perspective, as well as clinical case reports written by clinicians caring for the patients. In addition, the book contains chapters describing the current understanding of therapeutic strategies, pathogenic mechanisms, and systems medicine approaches allowing for new concepts for patient stratification. The work has its foundation in a recently completed European Union funded project called INFECT, which during the period 2013–2018 conducted comprehensive studies of these heterogenous and complex life-threatening infections. The project involved the establishment of the largest prospective patient cohort to date, a collection of biobank samples that were further analyzed to delineate key events underlying disease outcome and exploring their value for improved diagnostics and therapeutics. Advanced systems biology analysis and computational modeling were employed, and findings were validated in clinically relevant experimental in vitro and in vivo tissue infection models. To achieve its goals, the INFECT consortium gathered experts within the fields of intensive care medicine, infectious diseases, microbiology, bacterial pathogenesis, systems biology, and diagnostics, as well as the patient organization Lee Spark NF Foundation. A success factor of the INFECT project was the integrated work requiring the close interaction between partners from different disciplines and full utilization of all available expertise and resources provided by the consortium partners (Fig. 1). In this book, different aspects of necrotizing soft tissue infections are summarized by INFECT partners, providing contemporary state-of-the-art presentations, highlighting both advances made within INFECT and elsewhere. The different chapters of this volume collectively demonstrate the great heterogeneity of these infections and demonstrate the necessity for tools to stratify patients for novel, personalized tailored therapies. Hence, the book underlines the need for further actions toward the implementation v

vi

Preface

Fig. 1 INFCT consortium meeting in Stockholm 2017 attended by INFECT partners, i.e. Karolinska Institutet (coordinator, Sweden), Rigshospitalet (Denmark), Karolinska University Hospital (Sweden), Blekinge County Hospital (Sweden), Sahlgrenska University Hospital (Sweden), University of Bergen (Norway), Helmholtz Centre for Infection Research (Germany), Wageningen University and Research (The Netherlands), University of Lyon (France), LifeGlimmer GmbH (Germany), Tel Aviv University (Israel), Lee Spark NF Foundation (UK), University of North Dakota (US), and Cube Dx (Austria)

of personalized medicine for therapeutic advances for patients with necrotizing soft tissue infections. The INFECT consortium was supported by two scientific advisors: Dr. Marina Morgan, consultant clinical microbiologist at Royal Devon and Exeter hospital in the UK, and Prof. Matthias Reuss from Stuttgart Research Center Systems Biology, who provided valuable scientific advice throughout the entire project period. We also want to acknowledge the financial support from the European Commission, and we are grateful to the invaluable guidance we received from the Scientific Officer, Ms. Christina Kyriakopoulou. Last but not least, we would like to express our sincere gratitude to patients and relatives for their participation in the INFECT project. Huddinge, Sweden Huddinge, Sweden Bergen, Norway

Anna Norrby-Teglund Mattias Svensson Steinar Skrede

Contents

1

2

3

4

5

6

7

The INFECT-Project: An International and Multidisciplinary Project on Necrotizing Soft Tissue Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mattias Svensson and Anna Norrby-Teglund Necrotizing Soft Tissue Infections: Case Reports from the Patients Prospective . . . . . . . . . . . . . . . . . . . . . . . . . Doreen Cartledge, Lucy Dove, Emma Richardson, and Robert Wilkie

1

7

Necrotizing Soft Tissue Infections: Case Reports, from the Clinician’s Perspectives . . . . . . . . . . . . . . . . . . . . . . Torbjørn Nedrebø and Steinar Skrede

21

Necrotizing Soft-Tissue Infections: Clinical Features and Diagnostic Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Martin Bruun Madsen, Per Arnell, and Ole Hyldegaard

39

Microbiological Etiology of Necrotizing Soft Tissue Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Steinar Skrede, Trond Bruun, Eivind Rath, and Oddvar Oppegaard Beta-Hemolytic Streptococci and Necrotizing Soft Tissue Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Trond Bruun, Eivind Rath, Oddvar Oppegaard, and Steinar Skrede Treatment of Necrotizing Soft Tissue Infections: Antibiotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Oddvar Oppegaard and Eivind Rath

53

73

87

8

Treatment of Necrotizing Soft Tissue Infections: IVIG . . . . . . 105 Martin Bruun Madsen, Helena Bergsten, and Anna Norrby-Teglund

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Nikolai Siemens, Johanna Snäll, Mattias Svensson, and Anna Norrby-Teglund vii

viii

Contents

10

Systems Genetics Approaches in Mouse Models of Group A Streptococcal Necrotizing Soft-Tissue Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151 Suba Nookala, Karthickeyan Chella Krishnan, Santhosh Mukundan, and Malak Kotb

11

Systems Biology and Biomarkers in Necrotizing Soft Tissue Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167 Edoardo Saccenti and Mattias Svensson

12

Systems and Precision Medicine in Necrotizing Soft Tissue Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187 Vitor A. P. Martins dos Santos, Christopher Hardt, Steinar Skrede, and Edoardo Saccenti

1

The INFECT-Project: An International and Multidisciplinary Project on Necrotizing Soft Tissue Infections Mattias Svensson and Anna Norrby-Teglund

Contents 1.1

Necrotizing Soft Tissue Infections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2

1.2 1.2.1 1.2.2 1.2.3

The INFECT-Project . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Initiative and Consortium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Objectives and Work Plan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Achievements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1.3

Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2 2 2 3

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

Abstract

Keywords

This book describes clinical and microbiologic aspects, pathogenesis, and diagnostics, related to the severe and rapidly spreading necrotizing soft tissue infections. The work has its foundation in a recently completed European Union funded FP7-project called INFECT, which during the period 2013–2018 focused on utilizing a systems medicine approach to increase our understanding of these heterogenous and complex life-threatening infections. In this chapter, the aim and scope as well as key achievements of the INFECT-project are described.

Necrotizing soft tissue infections · Systems medicine · Pathogenesis

M. Svensson (*) · A. Norrby-Teglund Center for Infectious Medicine, Department of Medicine, Karolinska Institutet, Huddinge, Sweden e-mail: [email protected]

Highlights The work accomplished in the INFECT-project is summarized in book Chaps. 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, and 12, and highlights from these are: • Advanced understanding of necrotizing soft tissue infections and associated pathogenesis • Clinical and experimental data of importance for improved diagnostic and therapeutics • Demonstrated the value of a systems medicine approach in infectious diseases

# Springer Nature Switzerland AG 2020 A. Norrby-Teglund et al. (eds.), Necrotizing Soft Tissue Infections, Advances in Experimental Medicine and Biology 1294, https://doi.org/10.1007/978-3-030-57616-5_1

1

2

M. Svensson and A. Norrby-Teglund

• Scientific foundation for personalized medicine in necrotizing soft tissue infections

1.1

Necrotizing Soft Tissue Infections

Necrotizing soft tissue infections (NSTI) are rapidly spreading infections that may cause extensive soft tissue or limb loss, multiorgan failure, and are associated with a considerable fatality rate. These life-threatening infections that are caused by a variety of pathogens are among the few infectious diseases that remain associated with substantial mortality and amputation even in young individuals with no underlying condition. It is undisputed that rapid diagnosis and prompted intervention is directly related to survival, but diagnosis and management are difficult due to heterogeneity in clinical presentation, co-morbidities, and microbiological etiology. To tackle these challenges an international and multidisciplinary consortium, INFECT (Improving Outcome of Necrotizing Fasciitis: Elucidation of Complex Host and Pathogen Signatures that Dictate Severity of Tissue Infection), was formed https://permedinfect.com/projects/infect/.

1.2 1.2.1

The INFECT-Project Initiative and Consortium

The initiative to INFECT was taken in 2012 when a consortium of 14 partners applied for EU funding within the Seventh Framework Programme Health and Innovation, under the programme “Systems medicine: Applying systems biology approaches for understanding multifactorial human diseases.” INFECT was granted EU funding and was sponsored from January 2013 to June 2018. The overall goal of INFECT was to reveal pathophysiological mechanisms and disease signatures, as well as identifying prognostic and diagnostic markers of the multifactorial highly lethal NSTI. To enable this multifaceted work the consortium was designed to include industrial, clinical, and preclinical partners with expertise in different

disciplines including infectious diseases, intensive care, microbiology, cell biology, bacterial pathogenesis, systems biology, computational modelling and diagnostics. INFECT partners are listed in Table 1.1.

1.2.2

Objectives and Work Plan

The specific objectives of INFECT were to: 1. Establish a patient cohort and collect samples for a biobank collection 2. Analyze patient samples and clinical isolates to pin-point key host and pathogen factors involved in the onset and development of infection 3. Identify and quantify disease signatures and underlying networks that contribute to disease outcome 4. Exploit identified disease traits for the innovation of optimized diagnostic tools 5. Translate the advanced knowledge generated into improved classification and management, as well as novel therapeutics To achieve the objectives, a comprehensive and integrated knowledge of diagnostic features, causative microbial agent, treatment strategies, and pathogenic mechanisms (host and bacterial disease traits and their underlying interaction network) were required. Therefore, the work of INFECT was structured around an integrated systems biology approach in patients and different clinically relevant experimental models. The INFECT-project was set up to be multidisciplinary and included clinical partners, academic experimentalists and mathematical modelers, one patient organization, as well as two smalland medium-sized enterprises focusing on data integration in systems and synthetic biology, and development of diagnostic tools. The partners contributed with unique expertise, infectious disease and intensive care clinicians, technical platforms and/or model systems that together provided the means to successfully conduct the multifaceted research proposed (Table 1.1.). The work included seven highly integrated scientific work packages, one work package on

1

The INFECT-Project: An International and Multidisciplinary Project on. . .

3

Table 1.1 INFECT partners Partner organization name Karolinska Institutet

Country Sweden

Rigshospitalet

Denmark

Karolinska University Hospital Blekinge County Hospital Sahlgrenska University Hospital University of Bergen

Sweden

University of North Dakota Helmholtz Center for Infection Research University of Lyon

USA

Wageningen University and Research LifeGlimmer GmbH

The Netherlands Germany

Tel Aviv University

Israel

Cube Dx

Austria

The Lee Spark NF Foundation

UK

Sweden Sweden Norway

Germany France

Team leader Anna NorrbyTeglund Mattias Svensson Ole Hyldegaard Erik Jansen Michael Nekludov Folke Lind Ylva Karlsson Per Arnell Anders Rosén Steinar Skrede Trond Bruun Malak Kotb Suba Nookala Dietmar Pieper Singh Chhatwal Francois Vandenesch Vitor Martins dos Santos Vitor Martins dos Santos Edoardo Saccenti Eytan Ruppin Raphy Zarecki Christoph Reschreiter Doreen Cartledge

dissemination and exploitation of the knowledge obtained, and one management work package. A central part of the INFECT-project was the prospective enrollment of patients with NSTI and creation of a clinical data registry, the collection of biobank samples and the causative microbes, which all formed the basis for the experimental studies (Madsen et al. 2019). Further, INFECT was designed to use an integrated systems biology approach in patients and different clinically relevant experimental infection models, including both murine NSTI models (Chella Krishnan et al. 2016; Nookala et al. 2018) and a human organotypic skin model (Mairpady Shambat et al. 2016; Siemens et al. 2015, 2016). The work-flow also included a comprehensive set of analyses followed by integration of results in advanced computational platforms, which enabled generation of pathophysiological models

Main role Coordinator, bacterial pathogenesis, experimental human tissue model systems Patient cohort, biobank, therapeutics Patient cohort, biobank, therapeutics Patient cohort, biobank, therapeutics Patient cohort, biobank, therapeutics Patient cohort, biobank, therapeutics, microbiology NSTI murine model, forward genetics Systems biology Microbiology, systems biology Systems biology, integration modeling Systems biology, integration modeling, SME

Systems biology, integration modeling Diagnostic tools, SME Patient organization

of the disease and advanced understanding of the underlying mechanisms and host–pathogen interactions. The results were translated into novel diagnostic tests and improved patient management (Fig. 1.1).

1.2.3

Achievements

The INFECT-project has generated comprehensive knowledge of diagnostic features, causative microbial agents, treatment strategies, and pathogenic mechanisms (host and bacterial disease traits and their underlying interaction network) through analyses of the INFECT patient cohort and associated biobank. Also, INFECT has proven the value of systems medicine approaches in acute infectious diseases to achieve improved diagnostics and therapeutics to improve patient

4

M. Svensson and A. Norrby-Teglund

Fig. 1.1 An overview of the INFECT consortium and activities, in which clinical challenges, identification of pathophysiological mechanisms, and development of diagnostic tools were in focus

1

The INFECT-Project: An International and Multidisciplinary Project on. . .

disease outcome. In this book, we have summarized the scope and outcome of the INFECT-project. The results have been compiled into 12 chapters, each of which has a unique focus ranging from patients’ and doctors’ experiences of NSTI, to different treatment alternatives and the use of systems medicine in NSTI. In short, Chaps. 2 and 3 portray affected patients’ and treating physicians’ experiences; Chaps. 4, 5, and 6 communicate clinical characteristics, microbial etiology, and the role of beta-hemolytic streptococci; Chaps. 7 and 8 describe treatment options with particular attention to antibiotics and polyspecific intravenous immunoglobulin (IVIG); Chaps. 9 and 10 depict pathogenic mechanisms revealed from in vitro and in vivo models of NSTI; and finally, Chaps. 11 and 12 illustrate the merits of system biology for identifying NSTI biomarkers, and the use of systems medicine as a tool for stratifying patients with NSTI. Below we summarize some key findings and achievements of INFECT: • Establishment of the world’s largest NSTI patient cohort with extended clinical registry and associated biobank. • Four hundred and nine NSTI patients with completed clinical registries (>2000 variables) were enrolled and associated biobank samples (> 6000) and clinical bacterial isolates collected. • In Chaps. 3, 4, 5, and 6 advanced insight into the clinical aspects of NSTIs providing the basis for evidence-based guidelines for patient management and care is given (Bruun et al. 2020; Madsen et al. 2019). • In Chaps. 9, 10, 11, and 12 it is shown how the systems medicine approach within INFECT substantially advanced our understanding of these life-threatening infections, including the identification of novel pathogenic mechanisms and specific host and bacterial disease traits associated with disease outcome (Afzal et al. 2020; Chella Krishnan et al. 2016; Siemens et al. 2015, 2016; Thanert et al. 2019). • In Chap. 10 it is demonstrated how the pathophysiology of NSTI is influenced both by the

5

causative microbe and by host factors (Chella Krishnan et al. 2016; Siemens et al. 2016; Thanert et al. 2019). This underscores the need for patient stratification and implementation of tailored therapy/personalized medicine in these infections, presented in Chap. 13. • Development and testing in the clinical setting of multiplex diagnostic tools for rapid pathogen identification and monitoring of disease associated biomarkers. • The novel understanding of the disease mechanisms of these infections has resulted in changed clinical practice related to antibiotic usage as well as use of immunomodulatory treatments (Bergsten et al. 2020; Madsen et al. 2017) (Chaps. 7, 8, and 9). • Fostering the new generation of clinical and preclinical scientists within the field of systems medicine in infectious diseases.

1.3

Future Perspectives

We foresee that the achievements of INFECT will form a basis for future advances in the area of implementing personalized medicine in infectious diseases. This book volume is designed to present the contemporary understanding of clinical features, treatment, and pathogenic aspects in NSTI, as well as serve as an overview of the contributions to this field achieved through the INFECT study. Readers will also be guided to other references that should be consulted for further information on NSTI pathogenesis, causative agents, treatments as well as novel diagnostic and therapeutic strategies in order to improve outcome of NSTIs. The foundation that the results of INFECT has laid will be further explored and implemented in the field of personal medicine in infectious diseases in the current multinational projects PerMIT and PerAID, www. permedinfect.com. Acknowledgement Financial support: The work was supported by the European Union Seventh Framework Programme: (FP7/2007-2013) under the grant agreement 305340.

6

References Afzal M, Saccenti E, Madsen MB, Hansen MB, Hyldegaard O, Skrede S, Martins Dos Santos VAP, Norrby-Teglund A, Svensson M (2020) Integrated univariate, multivariate, and correlation-based network analyses reveal metabolite-specific effects on bacterial growth and biofilm formation in necrotizing soft tissue infections. J Proteome Res 19:688–698 Bergsten H, Madsen MB, Bergey F, Hyldegaard O, Skrede S, Arnell P, Oppegaard O, Itzek A, Perner A, Svensson M, Norrby-Teglund A, INFECT Study Group (2020) Correlation between immunoglobulin dose administered and plasma neutralization of streptococcal superantigens in patients with necrotizing soft tissue infections. Clin Infect Dis. https://doi.org/10. 1093/cid/ciaa022 Bruun T, Rath E, Bruun Madsen M, Oppegaard O, Nekludov M, Arnell P, Karlsson Y, Babbar A, Bergey F, Itzek A, Hyldegaard O, Norrby-Teglund A, Skrede S, INFECT Study Group (2020) Risk factors and predictors of mortality in streptococcal necrotizing soft-tissue infections: a multicenter prospective study. Clin Infect Dis. https://doi.org/10.1093/cid/ciaa027 Chella Krishnan K, Mukundan S, Alagarsamy J, Hur J, Nookala S, Siemens N, Svensson M, Hyldegaard O, Norrby-Teglund A, Kotb M (2016) Genetic architecture of group a streptococcal necrotizing soft tissue infections in the mouse. PLoS Pathog 12:e1005732 Madsen MB, Hjortrup PB, Hansen MB, Lange T, NorrbyTeglund A, Hyldegaard O, Perner A (2017) Immunoglobulin G for patients with necrotising soft tissue infection (INSTINCT): a randomised, blinded, placebo-controlled trial. Intensive Care Med 43:1585–1593 Madsen MB, Skrede S, Perner A, Arnell P, Nekludov M, Bruun T, Karlsson Y, Hansen MB, Polzik P, Hedetoft M, Rosen A, Saccenti E, Bergey F, Martins

M. Svensson and A. Norrby-Teglund Dos Santos VAP, INFECT Study Group, NorrbyTeglund A, Hyldegaard O (2019) Patient’s characteristics and outcomes in necrotising soft-tissue infections: results from a Scandinavian, multicentre, prospective cohort study. Intensive Care Med 45:1241–1251 Mairpady Shambat S, Siemens N, Monk IR, Mohan DB, Mukundan S, Krishnan KC, Prabhakara S, Snall J, Kearns A, Vandenesch F, Svensson M, Kotb M, Gopal B, Arakere G, Norrby-Teglund A (2016) A point mutation in AgrC determines cytotoxic or colonizing properties associated with phenotypic variants of ST22 MRSA strains. Sci Rep 6:31360 Nookala S, Mukundan S, Fife A, Alagarsamy J, Kotb M (2018) Heterogeneity in FoxP3- and GARP/LAPexpressing T regulatory cells in an HLA class II transgenic murine model of necrotizing soft tissue infections by group A streptococcus. Infect Immun 86(12):e00432-18 Siemens N, Kittang BR, Chakrakodi B, Oppegaard O, Johansson L, Bruun T, Mylvaganam H, INFECT Study Group, Svensson M, Skrede S, Norrby-Teglund A (2015) Increased cytotoxicity and streptolysin O activity in group G streptococcal strains causing invasive tissue infections. Sci Rep 5:16945 Siemens N, Chakrakodi B, Shambat SM, Morgan M, Bergsten H, Hyldegaard O, Skrede S, Arnell P, Madsen MB, Johansson L, INFECT Study Group, Juarez J, Bosnjak L, Morgelin M, Svensson M, Norrby-Teglund A (2016) Biofilm in group a streptococcal necrotizing soft tissue infections. JCI Insight 1: e87882 Thanert R, Itzek A, Hossmann J, Hamisch D, Madsen MB, Hyldegaard O, Skrede S, Bruun T, Norrby-Teglund A, INFECT Study Group, Medina E, Pieper DH (2019) Molecular profiling of tissue biopsies reveals unique signatures associated with streptococcal necrotizing soft tissue infections. Nat Commun 10:3846

2

Necrotizing Soft Tissue Infections: Case Reports from the Patients Prospective Doreen Cartledge, Lucy Dove, Emma Richardson, and Robert Wilkie

Contents 2.1

Stephen Lee Spark’s Story as Told by His Mum Doreen Cartledge . . . . . . . . .

7

2.2

Robert’s Experience . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

9

2.3

Frankie’s Story Written by His Mother Lucy Dove . . . . . . . . . . . . . . . . . . . . . . . . . . . 11

2.4

Zara’s Story Written by Her Mum Emma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

Abstract

All who have contributed in writing this chapter have been patients and parents that have experienced an horrific life event. The horrific disease named necrotising fasciitis has affected our lives for ever. All four stories have explained how easily an everyday infection can develop incredibly quickly into a lifethreatening experience. Three stories are expressed from the worn hearts of being a mother, fighting for their child every step of the way. Knowing our children and how they react through pain and illness is felt in each word, sentence, paragraph and even between the lines. Dedicating our unmarkable love and devotion for the child we carried for 9 months. To see them suffer in illness is heart wrenching, but to experience this disease necrotising fasciitis is something else. We must live through every day watching them grow with their scars of debridement, and to D. Cartledge (*) · L. Dove · E. Richardson · R. Wilkie Lancashire, UK

support them through further operations, let alone mental scars. Parents show a strength of support like no other and we hope that their lives can be enhanced through the battle they have individually won let alone their family. Robert’s story from a patient’s perspective is quite different and you will read his courage throughout. We continue to raise awareness through education. Keywords

Necrotizing soft tissue infections · Patient experience

2.1

Stephen Lee Spark’s Story as Told by His Mum Doreen Cartledge

We lost Stephen Lee Spark, Son, Brother, Nephew, Cousin and Grandson to a horrific disease named necrotising fasciitis. The time is important when being bereaved it was 9.45 am

# Springer Nature Switzerland AG 2020 A. Norrby-Teglund et al. (eds.), Necrotizing Soft Tissue Infections, Advances in Experimental Medicine and Biology 1294, https://doi.org/10.1007/978-3-030-57616-5_2

7

8

on the 6th October 1999. My son had been taken away from me in 5 days. Lee was a healthy 23-year-old, with a lovely quite naive personality. Martyn, my younger Son, Lee’s brother lived in Sheffield Yorkshire. Lee had an opportunity for work in Sheffield, so he travelled there in readiness of the interview. Approximately 3 days into staying over there Lee felt cold and had symptoms of flu. I informed Martyn to advise Lee to take two paracetamol tablets every 4 h and to drink plenty of water. I could not speak to Lee as he was not on the phone. Lee went for his interview feeling very poorly but managed to get through it and waited for a few days in the hope that he would be called upon to attend a second interview. Five days later Emma, Martyn’s partner, phoned me to say Lee had something more serious than flu and Lee was seriously ill and as a result they had phoned for an ambulance. I remember the daunting tones of Martyn yelling in the background, he sounded like a wild animal and was hysterical. I lived over a 100 miles away and I was in shock and disbelief, so I knew I was unable to drive all the way to Sheffield even at a time like this. The emergency crew arrived and tried their best to apply emergency care to Lee but he was falling into unconsciousness. Martyn looked on in disbelief, why was Lee so very poorly. Martyn explained later that Lee’s eyes were rolling also. What a terrible experience for Martyn to witness. I just wish I could have been there for them both. The emergency team tried their very best to keep Lee alive. Placed appropriate drains in him but Lee was not reacting to any treatment. The ambulance crew decided to take Lee to the nearest hospital and kindly passed the telephone number of the hospital. Martyn travelled in the ambulance with Lee and the crew. I am sure the stress of this for such a young man was truly horrific. The ambulance crew also said to ask mum to allow 15 min for them to arrive at the hospital before telephoning. I phoned the hospital in disbelief how was it possible for flu like symptoms to develop into Lee being so terribly poorly. What was the problem? I phoned Hallamshire Hospital and spoke to the Staff Nurse in A & E. I explained the circumstances and she replied “that there was no one with them by the name of Stephen Lee

D. Cartledge et al.

Spark”. At that moment I really thought I was in a terrible bad dream. But then the staff nurse replied “Oh wait a minute he is in resuscitation” I replied “No, I do not think this is possible”. The Staff Nurse asked me to phone back 15 min later, and suggested I had a sweet cup of tea and something to eat. By this time, I was in shock and was fainting every time I stood on my feet. I desperately tried to eat a slice of toast, it was difficult and akin to chewing on a piece of leather. I had no saliva in my mouth to swallow. Twenty minutes later I phoned the hospital and spoke to the Staff Nurse she replied “I am so sorry to have to tell you that Lee has passed away. Would you like to speak with Martyn?” we could only cry together. I eventually said “Go back home. I will see you there”. All this time I was on my own and I now needed my husband (not Lee’s father) to be with me. I explained the whole horrific story. It was 2 h before he arrived home and I had packed a few things for the long journey to Sheffield. I took a bottle of water, a bucket to for me to vomit in as I was in shock and an old dressing gown that my mum gave me when I was younger to give me comfort. We eventually arrived in Sheffield. During the journey I tried my best to make the decision of whether to see Lee’s body or go straight to Martyn I decided that Martyn needed support he had been through a terrible trauma. After all it was better for me to remember Lee the last time I saw him full of health and happiness. We asked Martyn a few gentle questions as he was crying all the time and spent a long time just holding one another for comfort. We will never, never forget that morning as long as we live. We then awaited the post-mortem results and we were told Lee had died of a Strep milleri infection— from an abscess in his gum and that it had developed into Necrotising Fasciitis. His internal organs had been eaten away. Martyn informed us that when the ambulance crew were applying resuscitation to Lee all his rotting flesh was coming down his nose and out of his mouth. Still to this day he remembers every finer detail and especially the smell of the rotten flesh. Martyn also referred to a dark purple like rash around Lee’s neck. What on earth was this terrible disease all about, I phoned Public Health England several days later wanting to find out more about

2

Necrotizing Soft Tissue Infections: Case Reports from the Patients Prospective

necrotising fasciitis and spoke with the Strep A and Respiratory Department, they were truly compassionate and informative. Over the next few days I spoke with many Consultants and medical staff all informing a little more about this terrible disease and all apologised that they could not provide a telephone number of a support group and all mentioned that there was a need for a support group to offer that listening ear and an understanding of compassion. Therefore, I launched my Charity The Lee Spark NF Foundation in January 2000. My first survivor contacted me in May 2000. What a thrill it was to talk to her. I desperately hung on to every word she said. That way I had a better understanding of what my dear son had been going through. Future Perspectives The early diagnosis and treatment of necrotising fasciitis must be taught at core level in order to save lives and devastation within families. The financial drain on the medical system is also, a major factor, to the patient if they survive a life of surgery, medicine and psychological problems. Martyn strives everyday living through the tragic circumstances of Lee’s death, my dear sons. What has happened to our family? His father never got over the shock of losing Lee and passed away a few years later (Fig. 2.1).

2.2

Robert’s Experience

The best way to describe living with the effects of necrotising fasciitis is to say it is a journey which continuous long after the infection has gone. This journey can often feel like a roller coaster ride with its highs and lows. Looking back, I would now say the first part of my journey was easy as I knew little about what was going on. I certainly had no idea of the seriousness of my condition for the powerful drugs I was given pretty much kept me in a deep sleep for almost a fortnight. I was admitted to my local hospital with expected appendicitis on New Year’s Eve 31st December 1981 and operated on the following day 1st January 1982, but it was not until almost

9

2 days later that the necrotising fasciitis appeared making it more likely that the infection was postoperative. In those days after treatment it was felt that I would heal better at home, so I was released on 27 January 1982; the day my eldest child was born. On this day I remember being pushed through the streets of the hospital complex in a wheelchair by a nurse from the main building to the maternity department to meet my new-born son. Later that day as my wife was to remain in the hospital for 7 days, I was released into the care of my mother. My wife remembers a few days before my release having a conversation with the consultant about my future prospects, he informed her that there was no reason why I could not live a normal life, “but” I would possibly require some corrective surgery in the future. Never was a truer word spoken. After being released from the hospital I took the consultant at his word and virtually just got on with my life. Taking my new parental responsibilities seriously as I set about being a father for the first time. With hindsight I now realise that carrying a pram up three flights of stairs after defeating necrotising fasciitis was not the brightest idea I have ever had. Now as I reflect, I also realise that perhaps I should have taken more time off work and gradually eased myself back into the workplace. It really was not the greatest idea I have ever had to go straight back into such a physical job. After all, a huge area of abdominal muscle had been lost forever leaving vital organs unsupported and unprotected. As a result of this I now have a loopy bowel and internal bruising to the stomach plus I have also suffered no fewer than 17 abdominal hernias, two of which are irreparable and one is a ruptured repair for which I am required to wear a support belt. Wearing a hernia (stoma) belt has its advantages for it eases the pain considerably but it also has disadvantages for it restricts my movement and can be almost intolerable in hot conditions. After my first hernia operation I discovered that I suffered from acute acid reflux and was given a chalky medication to drink, this did a great job of easing the reflux but also caused constipation due to a build-up of chalk in my bowel. After several endoscopy exams and a

10

D. Cartledge et al.

Fig. 2.1 Martyn Spark (left) Stephen Lee Spark (right) deceased

failed colonoscopy, I was sent for a CT colonoscopy which revealed I had loops in my bowel on my damaged side hence the reason for the intestine pain caused by the constipation. It may sound foolish, but this has resulted in me having to re learn how to and what to eat. For example, I cannot eat anything from the onion family except the green part of a spring onion or the green part of a leek, I cannot eat pineapples, lemons, limes, oranges I cannot take vinegar or eat pickles and many other everyday foods. I also must think about the kind of food I need to eat to help prevent any recurrence of acute constipation, alas this means eating out can be difficult if not in some cases impossible. As I mentioned earlier, I also suffer from internal bruising to my stomach, to this day I have never been given a reason for this, but I believe it is caused by my bowel pressing on my stomach when it is full. Recently I got rather fed up having to keep pulling at my waist band and tightening my trouser belt so I changed to using braces, the easing of the discomfort around my stomach was almost instant; the pain is still there but it is a much

more manageable pain. I feel incredibly lucky to have almost accidently discovered the difference wearing braces instead of a belt can make. Pain management is without a doubt the most difficult thing I have had to learn on my journey, for the pain around my stomach causes by internal bruising is relentless. The only help I receive for this are Codeine and Paracetamol, but I find the use of heat also helps, sadly the aforementioned hernias make exercising the area around my wound sight difficult, so exercise does not help. Three years ago, during a gastroscopy procedure it was discovered that I had an ampullary adenoma (pre-cancerous lesion) in the damaged area of my abdomen. This remains benign, it is almost as if this has taken up residence in the space where I lost the muscle due to Necrotising Fasciitis, therefore I see this as a continuation of my NF journey. Journeying with the effects of NF is a continual battle in trying to adapt to the demands placed upon my damaged body. For years I have struggled with pain causing me to overcompensate putting more pressure on my undamaged side which now has osteoarthritis. A

2

Necrotizing Soft Tissue Infections: Case Reports from the Patients Prospective

11

Fig. 2.2 Robert Wilkie survivor of necrotising fasciitis

coincidence perhaps? This brings me to what I find to be the most difficult thing to live with post Necrotizing Fasciitis, dealing with medical people, for there seems to be a willingness to keep NF quiet, why? Surely if I had been better informed, I could have been better prepared and more able to reduce the risks of further damage to an already battle-scarred body. Future Perspective With necrotising fasciitis listening to a survivor I believe is absolutely crucial yet to often I feel like no one is listening and many look at me and think here he comes again. Surely listening to any survivor is crucial to after care and pain management (Fig. 2.2).

2.3

Frankie’s Story Written by His Mother Lucy Dove

Frankie became unwell on Monday 8th April 2013 and he was only a 1 year of age. He had flu like symptoms. I describe his early symptoms as flu like symptoms because Frankie had a runny nose and a temperature of 39  C. His older

brother, Kayne was unwell the day before with pustules on his face, flu like symptoms and a pin prick rash on his back and chest. An out of hours GP referred Kayne to A and E, where I was told that his blood results were consistent with a viral infection and he was prescribed flucloxacillin. The links between Kayne and Frankie’s symptoms are very important. In the late morning of Tuesday 9th April, Frankie appeared to become very unwell with similar pustules on his face and a red pinprick rash all over his arms, chest and back. However, unlike his brother, Frankie was screaming when I moved him and seemed to be in a lot of pain. At our local walk in centre I stripped Frankie’s clothes off and I could see that a boil type lump had developed on his left thigh. I had not been there when I got him ready and he did not like me touching the area, so it was obviously very painful to him. He began hitting me and did not want to be cuddled. Frankie was a very cuddly baby, so this sent alarm bells ringing in my head! I explained this to the nurse practitioner, and they told me his temperature and other observations were fine. (I told then he had been given medication for a high temperature). They were going to

12

D. Cartledge et al.

Fig. 2.3 Frankie’s back the morning after being admitted. Once can clearly see the swelling on his left side and changes to the skin colour on the surface

discharge him home as they said they believed he must have a viral infection. I informed them of Frankie’s history including how his brother had been unwell. I stressed that I was not happy with how much pain Frankie was in, the paracetamol and ibuprofen were not working to ease his pain and so I really was not happy to go home with him. They then told me I could take him to A and E. I took Frankie straight to Children’s A and E. He was a very tough little boy, and nothing ever fazed him and so I knew that something had been missed because Frankie was in absolute agony. He could not be pacified and seemed to be getting worse before my eyes. He would not stand up and kept screaming. When seen by a consultant I pointed out the lump under the skin on his thigh. Back in the waiting room Frankie kept deteriorating and now he was howling in pain as if he had been really bad injured. At this point I was becoming increasingly frustrated. I felt as though my concerns were not being understood. The only way I could get him comfortable was to lay him on my chest facing me. This was when I began examining him myself. I am not medically trained but nor was I paranoid first-time mum. I knew something was wrong and I decided to check everywhere on his body. I refused to accept this was just a viral infection. I began checking his

back. Om his left side approximately halfway up, I pressed Frankie cried out in pain. I worked out there was an area of approximately 2 in. by 2 in. that was very painful for Frankie when I touched it. I then noticed that it could in fact be seen from the side as a swelling under his skin. When another doctor examined Frankie, she concluded that she believed he had a tissue infection. At roughly 9 pm a doctor took a blood sample from Frankie’s foot. An Orthopaedic doctor examined Frankie and concluded that Frankie had a possible tissue infection. In the early hours a doctor came into the room and told me that a “marker” in Frankie’s bloods was very high, but they decided not to give antibiotics and await more tests in the morning. I trusted that the doctors knew what was medically right for my son. During the Frankie kept waking and projectile vomiting. He kept crying on and off for hours. He was given codeine earlier in the evening which had made him drowsy, but he was still in a great deal of pain. I had a tough night with him as I could not comfort him, and he was in so much pain. In the morning Frankie looked totally delirious. He could barely move. I noticed that the skin on Frankie’s back was now red in parts (Fig. 2.3). It looked like the swelling had spread up his back and as you all can see his back appears to be swollen now (but more obvious on his left side).

2

Necrotizing Soft Tissue Infections: Case Reports from the Patients Prospective

The left side now had red marks appearing on the surface. The boil on his left thigh seemed to be moving outwards as if spreading. Were the boil had been was now looking dark blue like a bruise. Unfortunately, I could not get a photo of his leg. He would not let me touch anywhere on the left side of his back or indeed his body. I remember touching his left arm and him crying out in pain. A group of doctors came into the room I told them that the swelling on Frankie’s back looked like it had spread, and I showed the doctors Frankie’s left thigh. I was told Frankie was to start antibiotics, that he needed an emergency MRI scan. Frankie was given oral sedation and I went down with him for his MRI scan which lasted over 45 min. After the consultant told me Frankie needed to be sent by ambulance to another hospital and that he would need to go to theatre to have the pus removed from under his skin. I asked him what it was, and he replied; “Necrotising fasciitis”. I had never heard of this before and before I had had time to look it up on google two doctors came into the room and began trying to put cannulas into Frankie’s hands. It was now just before 3 pm. The doctors started putting what she said were antibiotics into Frankie’s cannula and pushing large syringes of saline into him one after the other. This was when I became worried something was really, seriously wrong. Frankie was transferred to the other hospital by ambulance. When we arrived, we were taken to the resuscitation area where doctors, consultants and different specialist surgeons began introducing themselves. As they were about to put Frankie onto a ventilator, we kissed him goodbye never knowing if we would ever see him alive again. A consultant plastic surgeon came into the room and told us that Frankie was very seriously ill and that his condition was life threatening. She told us that until they got him into theatre and tried to remove all the dead layers of skin and tissue, they could not know how bad it was. Frankie was in surgery for almost 10 h and when he came back the lead plastic surgeon and PICI Consultants took us into a private room. We were told that Frankie’s body had been extensively damaged by the infection. The surgeon

13

had removed all the layers of skin and fascia from almost his entire back, left side and quite a lot of his left thigh and muscle had been removed. We were told that they had not been able to remove all the infection because Frankie was too weak for any further surgery and that he was been given strong antibiotics intravenously. The consultant told us that they had done everything they could at the point and that it was now up to Frankie’s body to fight the infection with the help of some broad-spectrum antibiotics. We were asked to get everyone we needed up to the hospital because they did not believe our baby was going to survive the night. Not only did he have NF but the deadly toxins in his body causes sepsis and put his internal organs at great risk. As parents we could not believe we were being told that our son might not survive never mind the fact that he had no skin covering his back and sides. They allowed us in to see Frankie and it was unbearable to see him the way he was. Frankie was fully ventilated, in an induced coma and swollen to about five times his normal size. As parents we could hardly recognise our own son. The only thing that made him Frankie was his beautiful white curly hair. We cannot fault anything the second hospital for the care they gave to our son. Frankie by some miracle made it to Friday 12th April when doctors took him back down to surgery. We were told that if the infection had spread then Frankie would be made comfortable and allowed to pass away peacefully. He was down in the theatre for five and a half hours the surgeon had managed to remove the dead tissue from deep behind his shoulder blade and had been forced to cut deeper down into the muscle in his thigh. They were confident that they had removed everything and that everything now looked free from infection. We were informed that the bacteria that has caused the NF to develop was called strep A, which knowing this meant that they could adjust the antibiotics he was on. After Frankie’s second operation we felt more positive but were consistently told that Frankie still had a very long way to go because of the extensive damage the NF had causes to his body. Over the weekend Frankie remained in an

14

D. Cartledge et al.

Fig. 2.4 Frankie’s first dressing change on the ward with me

induced coma in PICU and then on Monday 15th April 2013 he underwent a further 5 h of surgery were expert plastic surgeons who specialise in burns patients, grafted healthy skin from his tummy, chest, right calf, left shin and back of his thigh onto his back and left leg. They were amazed with how well everything went and following two dressing changes in theatre on Friday 19th April 2013 Frankie was stable enough to be taken of his ventilator. He remained on high doses of morphine and clindamycin as well as other drugs, but he became well enough to be transferred onto the burn’s unit. Fast forward 7 years and Frankie is now 8 years old. We stayed in hospital for 3 months as he underwent painful dressing changes and procedures. He wore tight pressure suits and suffered break downs in his scars for 2 years after.

He had recurrent infections and 1 year later had to have his central back re-grafted. He went to theatre every 6 weeks for steroid injections. His grafts became raised and sore. He had to take gabapentin for a year which really affected his behaviour. He takes regular pain killers the strongest being oral morphine. He has low self-esteem and struggles with his confidence, but thankfully he is very active, and we encourage him to not let his skin grafts hold him back. We cream him four times a day. I have included picture of Frankie’s skin grafts from 2013 and now in 2020 (Figs. 2.4, 2.5, 2.6, and 2.7). Frankie is an inspiration. As he grows, he will need more skin grafts and procedures. We are incredibly lucky that he survived but our only wish is that it was diagnosed and treated sooner.

2

Necrotizing Soft Tissue Infections: Case Reports from the Patients Prospective

15

Fig. 2.5 Tissue loss on Frankie’s leg with skin graft. May 2013

Fig. 2.6 Frankie May 2013

I think there are four vital lessons that could be learnt from Frankie’s case. Future Perspective Doctors must pay close attention to any bacterial infections that a person may have been

predisposed to when taking history as this could help with a faster diagnosis. I would argue that this is very important with young children or those with a weaker immune system. The links between Kayne and Frankie’s symptoms are very important because Impetigo can be caused by the

16

D. Cartledge et al.

Fig. 2.7 Frankie January 2020

bacteria Streptococcus (most common) or Staphylococcus (staph). Methicillin-resistant Staph aureus (MRSA) is also becoming a common cause. This could then help with early treatment and prevent complications. Although Frankie had similar symptoms to his brother, Frankie was clearly suffering from a great deal of pain. Doctors need to be aware of this. His symptoms should not have caused the level of pain he was in and this is an important indicator that something more serious is developing. Any break in the skin or boil in conjunction with the symptoms and pain score should be taken seriously and treated with IV antibiotics immediately without waiting for blood test results.

Listen to parents! They see their child every day and they can be the best judge of their child’s condition. If they are worried, it is for a reason. This is vital to helping get an early diagnosis.

2.4

Zara’s Story Written by Her Mum Emma

I have four children, two boys and two girls. My daughter Anastazja was 4 years old when she was diagnosed with Scarlet Fever, to be honest I thought it was an old fashioned disease that had been eradicated. Two weeks later my other daughter Zara who was then 3 years old became unwell with what seemed to be a cold, her nose was running but snotty, her cheeks were red, her

2

Necrotizing Soft Tissue Infections: Case Reports from the Patients Prospective

eyes were watery; to me they were classic symptoms of a cold. Forty-eight hours later Zara came out with chickenpox. Over the next 24 h, more pox came out but Zara suddenly become really unwell which I thought was very unusual as children are usually ill before the “pox” appears. Zara started vomiting and developed a temperature of 38  C which could be controlled by paracetamol. The next day her temperature was increasing and had now reached 39  C but was no longer responding to paracetamol. The chickenpox was still coming through but they were now coming through as green/black in colour and were appearing more like “scabs” as opposed to blistered “pox” texture. Zara had now become very ill, she was not eating or drinking and was starting to cry whenever she was touched. There were two significant “pox” that started to look very nasty, one was on Zaras’ abdomen, the other on her sternum. The “pox” were now appearing as big green and black scabs. The pox on her abdomen had now developed a red ring around it, this is what lead me to take her to our local A&E department in Salford Manchester UK. They put us in an isolation room in the children’s department. Zara was now starting to develop a red rash around the “pox” on her sternum. We found out that 2 weeks prior to our visit a 3 year old with similar symptoms had been sent home and had sadly died of sepsis, I think that that tragic event helped my daughter because that meant the staff was extremely vigilante with her. The doctors were really concerned as they did not know why Zara was ill or what was causing it. Zara was now screaming if anyone touched her. Her body was now in so much pain it was heart breaking to watch. The doctors then decided to transfer Zara to the local children’s hospital. We were transferred into an isolation bay as Zara was obviously still infectious and no one had a clue what was happening to her. The ring around the “pox” on her abdomen was now getting larger. Another concern was that redness had now appeared at the “pox” on her sternum. The redness that was around Zara’s neck appeared to

17

look like a necklace. As the day wore on Zara screamed whenever anybody entered her room because she knew that they were going to touch her. Her screams were ear piercing because she was in so much pain. The redness started spreading to what can only be described as a “capped sleeve cardigan”. As the hours went by, redness has now appeared near the “pox” on her abdomen, it now looked like she was wearing a red “belt”. As time passed, “capped sleeve cardigan” had now turned into a red “bolero jacket” and the red “belt” was now a red “pair of hot pants/boxer shorts”. During the night it now looked as if she was wearing a long sleeved red cardigan. The weirdest thing was that whilst the redness was overtaking her body, she had a “band” of normal skin around her midriff. Her “boxer shorts” were now a pair of “jammers”, the redness had now spread to her knees. Within 30 min the redness was now a red pair of tights. Zara was now “vacant”, just starting at nothing. You could have thrown her around the room and she would not have cared, her eyes were glazed, she was not even responding to me her Mummy. In the early hours the hospital room was filled with the Heads of Departments, Medical Consultants, Immunology consultants, Plastic Surgery Consultants, Anaesthetists and other Doctors and Nurses. A decision was made that Zara now needed emergency surgery. Over 4 h later myself and my partner were finally told that she was out of surgery, we made our way up to Intensive Care were we were ushered into a side room where a Surgeon introduced himself and began to explain the he was head of Plastics and the Burns Unit. He told us what had taken place in surgery. He explained that our daughter had contracted Necrotising Fasciitis, we had never heard of it. He told us that it was a flesh eating bug, he explained that Necrotising Fasciitis had eaten her chest wall so they had to debride the whole of her chest wall and that the next 24–48 h were critical as the child mortality rate was 95%. He also informed us that he had operated on the “pox” that was on her abdomen as it had turned into a large abscess, so he had cut the “pox” out and had packed the wound. We were then led into

18

D. Cartledge et al.

Fig. 2.8 Zara had given up, her eyes were empty, she was no longer screaming whenever she was touched. This is the moment they decided she needed emergency surgery

the Intensive Care isolation room where we saw our little girl for the first time, she was unconscious as she was intubated and looking like an American Football player. She was heavily bandaged and sedated; it was such a shock. Five days later, the surgeon needed to see if the Necrotising Fasciitis had spread so operated on Zara whilst being in her intensive Care room where I observed him removing silver foil from her chest as he had left the wound open. He observed that it had not spread any further so decided to close the wound. It looked so surreal. He said he had only seen a handful of cases but had never witnessed Necrotising Fasciitis in the chest wall nor seen in at its earliest detection. Zara was transferred to the Burns Unit after a week in Intensive Care where she remained for 3 weeks. Zara was eventually discharged home after

1 month. We attended various clinics and Zara also wore a pressure vest for a period of time. Zara is now 12 years old and is in her first year of high school. Her scar spans from shoulder to shoulder and is the shape of a “Y”. We see the consultant regularly who has been amazed at her healing process. Zara’s scar has not blanched but has stretched over time as she has grown. Her Consultant has said that she can have corrective surgery to neaten it if she wishes when she is fully grown. He also thinks that she contracted it from Anastazja having Scarlet Fever as both are caused by the Strep A bacteria. He thinks that Zara must have had the bacteria on her skin when she developed chickenpox and that it entered her body when the “pox” erupted. He also thinks that the nastiest “pox” on her sternum was the entry point (Figs. 2.8, 2.9, 2.10, and 2.11).

2

Necrotizing Soft Tissue Infections: Case Reports from the Patients Prospective

19

Fig. 2.9 This was my first image of my 3 year old after surgery. No words could have prepared me. Unconscious and intubated. I could not comprehend or believe this was my little girl

Fig. 2.10 Seventy-two hours after Zara’s surgery they needed to perform more surgery to ensure that necrotising fasciitis had not spread and to remove the silver mesh they had placed into her

20 Fig. 2.11 We finally had our little girl back, smiling and laughing. No one would believe the trauma Zara had been through and survived. Our little miracle

D. Cartledge et al.

3

Necrotizing Soft Tissue Infections: Case Reports, from the Clinician’s Perspectives Torbjørn Nedrebø and Steinar Skrede

Contents 3.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

3.2 3.2.1 3.2.2 3.2.3

Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Case 1—Polymicrobial Infection/Anaerobic Infection Originating in the Pelvic Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Case 2—GAS of the Upper Extremity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Case 3—GAS in an Immunocompromised Patient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.3

Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34

22 22 27 31

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35

Abstract

Necrotizing soft tissue infections (NSTI) are rapidly spreading and life-threatening infections of skin and soft tissue. Essentially there are two types of NSTI, based on the invasive microorganisms. The speed of development and associated clinical features differ markedly depending on the bacterial etiology. Early recognition, extensive surgical debridement, and appropriate antimicrobials are pivotal for successful management. In this

chapter, we present three cases from the INFECT-study population. This study was an international, multicenter, prospective cohort study of adult patients with NSTI. We describe the clinical presentations, pre-, peri-, and postoperative clinical findings, microbiology, and treatment in cases of monobacillary Streptococcus pyogenes necrotizing soft tissue infections NSTI, polymicrobial infection, and an unusual presentation of pelvic monobacillary S. pyogenes infection in an immunocompromised patient. Keywords

T. Nedrebø (*) Department of Anaesthesia, Haraldsplass Deaconess Hospital, Bergen, Norway e-mail: [email protected] S. Skrede Department of Clinical Science, University of Bergen, Bergen, Norway Department of Medicine, Haukeland University Hospital, Bergen, Norway

NSTI · Streptococcus pyogenes · Polymicrobial infections · Immunosuppression · Toxic shock syndrome Highlights Necrotizing soft tissue infections located to the extremities are most commonly associated with

# Springer Nature Switzerland AG 2020 A. Norrby-Teglund et al. (eds.), Necrotizing Soft Tissue Infections, Advances in Experimental Medicine and Biology 1294, https://doi.org/10.1007/978-3-030-57616-5_3

21

22

T. Nedrebø and S. Skrede

monomicrobial infections caused predominantly by Streptococcus pyogenes (group A streptococcus; GAS). GAS infection of non-extremity location is associated with immunosuppression. Extensive surgical debridement is essential for survival. Anaerobic necrotizing soft tissue infections may benefit from hyperbaric oxygen therapy.

3.1

Introduction

Necrotizing soft tissue infections (NSTIs) are lifethreatening infections typically characterized by rapid progression of disease and significant local tissue destruction. However, the speed of development and associated clinical features differ markedly depending on the bacterial etiology (Hakkarainen et al. 2014; Calandra et al. 2005; Anaya and Dellinger 2007). The tissue destruction involves any of the layers of the soft tissue compartment, extending from the epidermis to the deep muscle fascia. The disease frequently progresses below the outer surface, and in the absence of overt skin changes the diagnosis can be difficult to establish in the early phases. However, signs of systemic toxicity and pain often develop that are disproportionate to the findings of the skin examination (Calandra et al. 2005). The diagnosis is surgical with the findings of friability of the superficial fascia, dishwater-gray exudate, and absence of pus (Stevens and Bryant 2017). Several descriptions and classifications of NSTI have been proposed, including stratification based on (1) anatomical localization (i.e., Fournier’s gangrene), (2) depth of infection (necrotizing fasciitis, necrotizing myositis), (3) severity (Sequential Organ Failure Assessment Score (SOFA), simplified acute physiology score II (SAPS II), or (4) microbiological findings. Classifications due to microbiological findings and their clinical characteristics are delineated in Chap. 5. Shortly summarized, type 1 infections are polymicrobial, type 2 are monomicrobial, type 3 is a monomicrobial infection with marine pathogens including Vibrio vulnificus and Aeromonas hydrophila, whereas in type 4 the etiology is fungal (Stevens and Bryant 2017;

Sartelli et al. 2018; Morgan 2010). Type 3 NSTI clinically presents more like type 2 NSTI (Hakkarainen et al. 2014), and the use of unique clinical types 3 and 4 is yet to be justified. However, differentiating type 1 and 2 would serve practical purposes, in which tailored diagnostics and therapy may be warranted. Type 1 polymicrobial infections most often affect older patients with comorbidities like diabetes, and the predominant anatomical localizations are perineum and the abdomen. Type 2 monomicrobial infections are predominantly caused by S. pyogenes (group A streptococcus; GAS) (Bruun et al. 2020). The patients presented in the following three case reports were all included in the INFECT study, an international, multicenter, prospective cohort study of adult patients (Madsen et al. 2019). Four hundred and nine patients undergoing at least one operation with confirmed NSTI were included, of which 402 were admitted to an Intensive Care Unit (ICU). The most apparent initial symptoms and signs were skin bruising (202 patients, 51%) and opioid-requiring pain (172 patients, 41%) and systemic toxicity was frequent; 50% had septic shock and 20% acute kidney injury (Madsen et al. 2019). Microbiological findings in these patients revealed polymicrobial infection in 50% (202) and monomicrobial infection in 44%, of which GAS was found in 126 cases (70%) among these patients. In 26 of the patients (6%), no microbes were identified. The median number of surgical interventions was 4, and 13% had amputation of either an extremity or penis. Most patients received betalactam (96%) and clindamycin (98%) as antibiotic treatment.

3.2 3.2.1

Cases Case 1—Polymicrobial Infection/Anaerobic Infection Originating in the Pelvic Region

This patient was a 48-year-old man with a 30 years lasting history of substance abuse, mostly heroin administered intravenously by

3

Necrotizing Soft Tissue Infections: Case Reports, from the Clinician’s Perspectives

way of the femoral veins. He had HCV RNA PCR positive chronic hepatitis C genotype 3a and had previously been hospitalized several times due to opioid overdosing. Moreover, he had prior hospital admissions for a traumatic fracture of the left femur and for surgical treatment of the right ankle. Known medication at presentation was alprazolam and oxazepam. Six days prior to admission and during a period of drug-abuse, he fell on his left knee resulting in a blunt minor trauma. In the following days he developed local swelling, pain, and redness of the knee. At admission he was dehydrated and opioid intoxicated, with a blood pressure (BP) of 91/57 mmHg, a pulse rate of 96/min, and a body temperature of 36,3
C. Clinical biochemistry and hematologic tests revealed a C-reactive protein (CRP) of 339 mg/L, white blood cell count (WBC) of 26,2  109/L, creatinine of 233 μM, and a Laboratory Risk Indicator for Necrotizing Fasciitis (LRINEC) score of 8, according to Wong et al. (2004). The medial side of the left knee was warm, had a swelling and erythema (Fig. 3.1), and joint flexion of the knee was impaired due to pain. Septic arthritis was suspected, and antibiotic therapy was initiated with a combination of intravenous ceftriaxone and clindamycin. Ultrasound showed fluid in the knee joint and between the superficial muscle layers in the lower parts of the thigh. X-ray revealed gas along the entire femoral shaft all the way from patella to trochanter major, and the patient was suspected to have infection with gas-producing microbes (Fig. 3.2). Under suspicion of septic arthritis and

Fig. 3.1 Swelling and erythema of the knee at admission

23

possibly NSTI, he was taken to surgery. First, the plastic surgeons made incisions on the medial and lateral part of the thigh from patella in proximal direction. The incisions were extended until they reached the level of the deep fascia, but only edema was present. The orthopedic surgeons first did an arthroscopy of the knee, draining purulent fluid (Fig. 3.3), samples of which were collected for microbiologic diagnostics. Subsequently, they made a midline incision over the prepatellar bursa. Gas had been seen along femur on the X-ray, and when they opened the deep fascia lots of necrotic tissue was detected beneath. The incision was prolonged to previous incisions made by the plastic surgeons, revealing direct communication between the prepatellar bursa and the intermediary vastus muscle. Thereafter the vastus lateralis muscle was split to give access to deeper structures. Pockets of pus were found all along the fascia up to the anterior inferior iliac spine, and parts of the quadriceps-muscle had to be removed due to necrosis (Fig. 3.4). The wound was covered with surgical swabs and left open pending a revision the following day, in accordance with contemporary recommendations (Rogers 2020). On the first revision, necrotic parts of the vastus intermedius muscle had to be removed, but there was no more evident pus. Microbiologic cultures from the wound taken during surgery demonstrated Prevotella disiens, Fusobacterium gonidiaformans, Actinomyces sp., and Streptococcus dysgalactiae (SD), a betahemolytic streptococcus often found in co-culture with intestinal flora. In blood culture, only SD was found. Samples from the fluid in the

24

T. Nedrebø and S. Skrede

Fig. 3.2 Gas seen in the soft tissue along femur

knee joint were culture negative. At first, he was treated with ceftriaxone and clindamycin, but later this was changed to a combination of benzylpenicillin and clindamycin following the susceptibility tests, where the microbes were

Fig. 3.3 Purulent fluid during knee-arthroscopy

found to be sensitive to penicillin and/or clindamycin. There were serial wound revisions, before the wounds were surgically closed, starting at day 4 and finished by day 6. Two days after closing

3

Necrotizing Soft Tissue Infections: Case Reports, from the Clinician’s Perspectives

25

Fig. 3.4 Necrotic muscle (parts of quadriceps)

of the wound he developed more pain, and a rise in CRP and WBC occurred. The sutures were removed, and lots of pus were found and the wound was washed with saline. From day 0 he received 4 daily treatments with hyperbaric oxygen therapy (HBOT), 90 min of 2.4 ATA (14/90), as he was suspected to have a myositis due to anaerobic bacteria, which was confirmed during the initial phase of his hospital stay. He had several more revisions before the wound was closed, partly with a skin autograft. Discussion Polymicrobial NSTI is referred to as Type 1 NSTI, and often associated with gas in the tissue (Stevens and Bryant 2017), although gas is also a possible clinical presentation in beta-hemolytic streptococcal (BHS) cases. In the INFECT study, 5 of 114 mono-GAS NSTI had palpable gas (crepitus) in the tissue and 10 of 114 had gas visualized on radiology preoperatively (Trond Bruun, personal communication). In type 1, strictly anaerobic microbes predominate (Madsen et al. 2019; Thanert et al. 2019), often accompanied by facultative anaerobic gram-negative bacilli and gram-positive cocci. A typical patient with NSTI type 1 is likely to have comorbidities like diabetes mellitus, peripheral vascular disease, or intravenous drug use (IVDU) (Sarani et al. 2009; Chen et al. 2001), thereby in a state of qualitative or quantitative immunosuppression. This particular patient had a minor trauma to his knee 1 week prior to admission. Aspiration of pus from the

knee joint during arthroscopy supported a clinical suspicion of septic arthritis. NSTI was actively sought for, but at initial surgery there were normal findings down to the fascia both laterally and medially on the extremity. However, the X-ray had shown gas along the femur, so the idea of a deeply situated primary NSTI caused by IVDU inoculation, from the femoral vein downwards along the femoral shaft to the knee structures, was not abandoned. Notwithstanding, the local trauma of his knee may have created an environment for growth and subsequent spread of the bacteria. Mechanistic studies addressing the pathogenic strategies and complex dynamics of bacterial communities in polymicrobial NSTIs are lacking (Thanert et al. 2019). However, many of the bacteria found in polymicrobial NSTI are the same as in monomicrobial NSTI (e.g., GAS), and thereby elaborate many of the same exotoxins and tissue-degrading enzymes (Shiroff et al. 2014; Snall et al. 2016). SD is an emerging cause of invasive disease in general (Oppegaard et al. 2016) and specifically an increasing etiology of NSTI in our patient population, and it may cause monobacterial type 2 NSTI (Bruun et al. 2013) or equally frequent, it is identified as part of polymicrobial etiology of NSTI (Bruun et al. 2020). In the INFECT study, blood cultures were more commonly positive in NSTI caused by GAS (56%) than in SD (30%). IVDU is a risk factor also for SD bacteremia, as shown elsewhere (Ruppen et al. 2017).

26

In our hospital we have noted an emerging association of severe SD infections and genotype stG62647, which has predominated in invasive SD infection in our population since 2013 (Oppegaard et al. 2017). This sequence type was also dominating in SD NSTI in the Scandinavian patient cohort of the INFECT study. In this patient, SD was identified along with a number of obligate anaerobic species that are rarely culture positive in blood culture compared to rates of identification in deep tissue cultures (Brook and Frazier 1995), essentially confirmed by the INFECT consortium (Madsen et al. 2019) and even more so, in a study using nucleic acid based detection of microbes in tissue (Thanert et al. 2019). These results are in support of the historic opinion that the importance of obligate anaerobic species may be underestimated in NSTI (Brook and Frazier 1995). In this patient case, time from presentation at hospital to surgery was 18 h, mostly explained by the patient’s non-compliance and refusal to receive help, delaying the time to surgery. This puts the patient at risk of poor outcome. Since the upsurge of NSTI in the mid 1980s, it was soon evident that there was an association of time to surgery and risk for poor outcome (Boyer et al. 2009; Elliott et al. 1996; Kobayashi et al. 2011; McHenry et al. 1995; Wong et al. 2003). Two recent meta-analyses reported a significantly lower mortality if the initial surgery was performed within 12 h of presentation (Gelbard et al. 2018; Nawijn et al. 2020). Gas in the tissue is one of the clinical signs more often associated with polymicrobial NSTI (Stevens and Bryant 2017). In this patient gas had been seen on the initial X-ray and although no evident sign of deeply situated NSTI was detected in the early phases of surgery, this radiologic finding spurred the surgeons to continue, leading them to explore muscle and other soft tissue all the way into the femoral shaft, a strategy leading to drainage of deeply situated necrotic tissue and pus, most likely improving the outcomes for the patient substantially. That a septic arthritis caused by SD appears secondary to a NSTI of an extremity is an observation made on several previous occasions in our hospital.

T. Nedrebø and S. Skrede

Undertreatment with antibiotics in early phases of obscure NSTI is common (Marwick et al. 2012). Our patient received empirical therapy with broad (1) gram-negative and (2) grampositive coverage that included effect on (3) Staphylococcus aureus that in our community is methicillin susceptible in >99% of invasive cases, (4) anaerobic species other than Bacteroides (that frequently expresses clindamycin resistance), and on (5) toxin producing species that might be involved in a NSTI that most likely was of polymicrobial etiology (Peetermans et al. 2020). When results on susceptibility were available, penicillin sensitivity in SD was confirmed, whereas all anaerobic species were either susceptible to both clindamycin and benzylpenicillin or one of these agents. In line with treatment guidelines (Stevens et al. 2014), therapy was narrowed accordingly. Temporary worsening seen in our patient was evaluated to be caused by too early closure of wounds. In NSTI the role of some of the therapeutic measures is controversial. Firstly, in our hospital HBOT is offered sparingly to patients with severe NSTI, organ failure, and evidence of obligate anaerobic microbial etiology. According to local guidelines for treatment of NSTI in our hospital, this patient did receive HBOT, as 1/3 of our patients in the INFECT cohort did (unpublished result). Furthermore, this patient was not offered intravenous polyspecific human immunoglobulin G (IVIG), which has been restrictedly used in our hospital. Until the INFECT study, orders for use were evidence of invasive GAS infection, shock, and unstable patient as described by Anaya and Dellinger (Anaya and Dellinger 2007). Only two patients in our local INFECT cohort fulfilled these criteria (unpublished data). However, there is increasing evidence that patients with monobacillary GAS NSTI or septic shock benefit on IVIG treatment with improved outcome (Parks et al. 2018), recently supported also by data for the streptococcal cohort of the INFECT study (Bergsten et al. 2020; Bruun et al. 2020). In this case, SD and obligate anaerobic microbes were identified. There is to date only one single completed randomized, blinded, placebocontrolled trial of immunoglobulin G treatment

3

Necrotizing Soft Tissue Infections: Case Reports, from the Clinician’s Perspectives

for patients with NSTI (Madsen et al. 2017). This study failed to show any effect on primary and secondary outcomes. Most patients included had polymicrobial infection. In an analysis of association between outcome and anatomic site of infection, there was a tendency in favor of placebo in infections of the anogenital and abdominal areas (Madsen et al. 2017), in which obligate anaerobes are frequently involved, different to BHS, that are rarely encountered (Madsen et al. 2019). This would support that IVIG may have little effect in the polymicrobial, non-BHS infections. A case report on beneficial use of IVIG in NSTI caused by SD has been presented, the effect attributed to blocking of streptokinase-mediated fibrinolysis, with subsequent reduced bacterial spread (Andreoni et al. 2018). The first case presented herein is a type 1 NSTI, a patient with a polymicrobial infection caused by co-occurrence of obligate anaerobic microbes and SD, a beta-hemolytic streptococcus. The disease course developed slowly, but there was a need for very resource demanding handling including HBOT. This case illustrates host risk factors (IVDU), as well as major clinical learning point; the anticipated primary focus of infection was an intra-articular knee joint infection as it was here infection first became visible. Different to this, the primary disease was judged to be advanced NSTI originating in the groin, descending along the femoral shaft.

3.2.2

27

Case 2—GAS of the Upper Extremity

This patient was a 48-year-old male with a history of psoriatic disease, but he was independent of medication. One week prior to hospitalization the patient got a minor laceration, as he cut a finger of the left hand on a paper edge. He treated the finger with a topical antibiotic due to a possible infection, and the finger was fine for the next few days. However, while cross-country skiing his left elbow rapidly became erythematous and painful within a few hours, and his temperature rose to 39.1
C. He was sent to the local community hospital on suspicion of severe skin and soft tissue infection. On admission his left elbow was painful, and there was an area of localized, well defined erythema, and swelling (Fig. 3.5). There were no signs of injury or breach of skin barriers. Vital signs included a BP of 134/82 mmHg, pulse 91, temperature 39
C, respiratory rate (RR) 12. CRP was 65 mm Hg had been initiated. The erythema and swelling had spread in distal direction from the MCP joints to the middle upper arm (Fig. 3.6). The swelling over the olecranon bursa was punctured, and fluid was aspirated and sent to microscopy and culture and gram-positive microbes with streptococcal morphology were detected. Cloxacillin was discontinued, and treatment was changed to clindamycin, benzylpenicillin, and gentamycin. He went to immediate surgery under general anesthesia. In the surgical theatre a large dorsal incision from olecranon to the lower forearm was made. Extensive amounts of dishwater-like fluid was found together with necrotic subcutaneous tissue. An area of 7  20 cm of the skin over the elbow, equaling approximately 1.5% of skin surface, had to be removed due to necrosis (Fig. 3.7). The muscle fascia and the muscle tissue looked vital with no necrosis. The wound was left open, covered with surgical swabs. Subsequently, he was postoperatively transferred to the intensive care unit (ICU), where SAPS III score 52, and SOFA score 8 were noted at admission. The next morning a second look was performed. The muscle fascia on the lower forearm was cut open, and the muscle tissue below appeared to be vital. Then the surgeons explored proximally towards the elbow. Still the fascia looked and felt vital, but extensive amounts of pus-like fluid emanated from deeper musclecompartments. Immediately after surgery the patient’s condition got worse with septic shock and respiratory failure. On examination the

margins of the infection with redness and edema had spread up to the shoulder. Initial bacteriological tissue and fluid cultures grew GAS, whereas blood cultures from the first admission were culture negative. A third revision uncovered some purulent fluid just below the shoulder, but the fascia and muscle tissue were vital with no signs of progression (Fig. 3.8). His condition improved, and 2 days later he was extubated and in no need of circulatory support. The patient went through several further revisions, but no more necrotic tissue or signs of infection were found. On day 4 they started to close the wound, but he had a skin defect which had to be transplanted at a later stage. On day 5 he was discharged from ICU, with a SOFA score 5. Discussion Case 2 is a representative patient of type 2 NSTI: GAS infection in a previously healthy young individual, affecting an extremity (Madsen et al. 2019). GAS is a gram-positive coccus capable of causing a broad array of infections, most commonly pharyngitis and non-necrotizing soft tissue infections (Stevens 2020). Invasive GAS infections include NSTI, which has a significant potential for local spread and aggressive clinical course, as experienced by this young individual. This patient started out with swelling and edema of his elbow, which only 12 h later spread distally to his MCP joint and proximally up to his shoulder. One week prior he had a small cut in his finger. Most likely the cut in his finger caused an entry of the bacteria, that may have been dormant or in a very isolated, minute focus of infection not giving rise to any attention. This was followed by a NSTI development, first in the elbow due to a

3

Necrotizing Soft Tissue Infections: Case Reports, from the Clinician’s Perspectives

29

Fig. 3.7 First surgery over the elbow and forearm. 7 * 20 cm of the skin removed due to necrosis

possible minor tissue trauma caused by the crosscountry skiing. NSTI has historically been associated with penetrating trauma, but recently blunt trauma has also been verified as an independent risk factor of GAS NSTI development (Bruun et al. 2020).

In retrospective studies of NSTIs, upper extremity infections are perceived as uncommon and lethal, as reviewed elsewhere (Uehara et al. 2014). Differently, we demonstrate that upper extremity is as common focus as head-neck NSTI, and even half that of anogenital-abdominal

Fig. 3.8 Surgery after second revision had to go below the shoulder

30

or lower extremity focus of NSTI, and they are not more lethal (Madsen et al. 2019). In extremity NSTI GAS predominates vastly (Bruun et al. 2020; Madsen et al. 2019). In this case blood cultures from the first admission were culture negative, as seen in 44% of GAS cases (Bruun et al. 2020), underlining the importance of collecting bacteriological tissue and fluid for routine culture. In the entire INFECT cohort 93% of the cases were tissue culture positive (Madsen et al. 2019). Our patient went to surgery 24 h after presentation of symptoms, 18 h after admission to hospital, and by then he was clinically affected with hypotension in need of vasopressor. This is somewhat later than the average GAS patient in our cohort. We showed that GAS cases receive surgery after a median of 16 h compared to non-GAS cases, whom received it 2 h later (Bruun et al. 2020). Invasive GAS infection can cause toxic shock syndrome (TSS), in our study reported in 65% (Bruun et al. 2020). Clinical criteria for TSS include hypotension (systolic blood pressure >90 mmHg) and multi-organ involvement (Breiman et al. 1993). This patient met the clinical criteria of TSS, and the diagnosis was confirmed with isolation of GAS from tissue biopsies and bursa fluid. This massive invasion and degradation of the soft tissue caused by GAS are due to both the bacteria’s production of exotoxins and the features of M-proteins, the latter being cell surface molecules on the bacteria. They both result in a massive pro-inflammatory cytokine release and thereby causing septic shock. A study has also shown that patients with GAS NSTI exhibited a lack of specific antibodies directed against the causative S. pyogenes strains and the majority of their exotoxins during initial stages of infection (Babbar et al. 2018; NorrbyTeglund et al. 1994; Basma et al. 1999), antibodies meant to prevent the spread of infection. Several in vitro and in vivo studies have indicated that the protein synthesis inhibitor clindamycin abates the toxin production in GAS, and improves survival rates in NSTIs and streptococcal TSS (Andreoni et al. 2017;

T. Nedrebø and S. Skrede

Carapetis et al. 2014; Linner et al. 2014; Mascini et al. 2001). Upon confirmation of streptococcal etiology our patient received a combination of penicillin and clindamycin, as advocated in contemporary guidelines for the treatment of GAS NSTIs (Stevens and Bryant 2017). However, he initially received inadequate antibiotic therapy, as evaluated retrospectively. In a prospectively enrolled cohort of skin and soft tissue infection cases, Marwick et al. reported inadequate empiric therapy in 10/11 (91%) cases with NSTI, highlighting the diagnostic challenges of distinguishing NSTIs from cellulitis in the early phase (Marwick et al. 2012). The LRINEC score has been proposed as a clinical tool for identifying patients in need of surgical treatment. However, at the primary admission our patient had a LRINEC score of 0/13, demonstrating the shortcomings of this scoring system, reviewed elsewhere (Fernando et al. 2019). In our hospital its use has been discouraged following our studies (Bruun et al. 2013; Bruun et al. 2020; Madsen et al. 2019). The diagnosis of NSTI is established by surgery, but frequently diagnosis is delayed (Goh et al. 2014). In general, it may therefore be advisable to encourage surgical exploration earlier when facing suspected NSTI patients, which means a certain ratio of negative explorative procedures must be accepted. It is our experience that we explore suspected NSTI in the range of 1,5:1 vs. confirmed non-NSTI (unpublished results), resembling published results of 1,2:1 that has been presented by others (Howell et al. 2019). Extensive surgery is main therapy in NSTI. In this case the fascia looked vital, but below the fascia extensive amounts of necrotic tissue were detected and removed. Intraoperative findings of muscle tissue involvement showed that GAS spread across tissue borders, and not only along the tissue planes. Surgically it is important to consider exploring below the fascia, even in cases where they appear to be vital. The extent of the excisions should be down to healthy bleeding tissue at all margins. Often the spread in the subcutaneous layer is more extensive than the skin changes suggest. This was the case here

3

Necrotizing Soft Tissue Infections: Case Reports, from the Clinician’s Perspectives

too, as the patient only few hours after second surgery developed TSS and had to undergo a second excision, in proximal direction to reach just below the shoulder, confirming challenges in infectious source control of this case. This case demonstrates the severe systemic manifestations including streptococcal TSS in a previously healthy man, the significance of blunt versus penetrating trauma, initial antibiotic undertreatment, and impact of too early surgical closure.

3.2.3

Case 3—GAS in an Immunocompromised Patient

A 59-year-old woman was admitted to her local hospital after 2 days of increasing pain in her right buttocks and groin. In addition, foul-smelling dark green vaginal discharge was noted. She previously had one cesarean section and more recently, surgery for urinary incontinence (tension-free vaginal tape (TVT)) a few years before enrollment in the INFECT study. Due to an active psoriatic arthritis, she received regular treatment with methotrexate per oral 10 mg once weekly and the tumor necrosis factor-alpha inhibitor, etanercept subcutaneously 50 mg once weekly. On admission to the hospital BP was

31

115/51 mmHg, pulse rate 107, RR 20, WBC 4.6  109/L, and temperature 38,2
C. CRP was 275 mg/L, while a LRINEC score was calculated retrospectively showing a low score of 4/13. Upon examination, she had no erythema or signs of infection on her buttocks or thigh. A gynecologic examination did not reveal any sign of infection, but a vaginal ultrasound gave suspicion of retroperitoneal edema against the pelvic floor. A CT scan showed either solid fluid or a phlegmon along the pelvic muscles from upper part of the right thigh to 5 cm above the pubic tubercle. Edema was present in gluteal, pelvic, and adductor muscles. A pigtail catheter was inserted prevesically, in an attempt to drain some of the fluid. Culture from blood, and this fluid later grew GAS. Subsequently her condition deteriorated only a few hours later, and soon she was in need of vasopressors (norepinephrine and dobutamine). She was rapidly sent by helicopter to the regional hospital where she arrived 15 h after first admittance to hospital. At arrival she was awake, selfbreathing, BP 89/35 mmHg (on vasopressor), pulse rate 127, and temperature 39,5
C. On examination, a swelling and discolored area on the back of her thigh was found (Fig. 3.9) and a CRP value of 225 mg/L was noted.

Fig. 3.9 Swelling and discolored area of the thigh 15 h after first admission

32

T. Nedrebø and S. Skrede

Fig. 3.10 Fasciotomy of the medial side of the thigh at first surgery

She went immediately to surgery under general anesthesia. An explorative laparotomy did not reveal any intra-abdominal pathology, except from localized retroperitoneal edema in the inguinal opening on the right side. Then a fasciotomy both laterally and medially on the thigh was performed (Fig. 3.10). There was edema of the subcutaneous fat and dishwater-like fluid under the fascia (Fig. 3.11). Both the fascia and the muscle were found viable. The antibiotic regimen was changed to meropenem and clindamycin, and she was transferred to the ICU. Her condition worsened the next few hours, and a swelling close to the right labium majus appeared. She was once again taken to immediate surgery. An incision was made between the labium minus and labium majus on the right side, to uncover devitalized subcutaneous tissue. There was no bleeding, due to widespread thrombosis of minor vessels. A digital dissection down to the fascia of the inferior part of urogenital diaphragm was made. Necrotic subcutaneous tissue was removed, but the fasciae were vital. The same procedure was done on the

left side, and an external vulvectomy was indicated. The wound was left open with wet surgical swabs. A second look at the fasciotomy of the right thigh revealed some necrotic skin and a small necrotic area of adductor magnus muscle. During the surgical procedure the patient received IVIG according to the contemporary treatment protocol, as there was evidence of GAS etiology and she was circulatory unstable and in a critical state. Upon return to the ICU after the second surgery, her condition was judged critical and unstable, supported by maximal SOFA score of 17, SAPS II score of 66, and SAPS III score of 86. Arterial blood gas lactate exceeded 8 mM, two vasopressors (norepinephrine and epinephrine) were administered as a Cardiac Index of 32-fold ratio of minimum bactericidal concentration (MBC) to minimum inhibitory concentration (MIC), has been reported as a phenomenon for several decades, particularly in various streptococcal species (Rolston et al. 1984). For S. pyogenes, the presence of penicillin tolerance has been reported for up to 29% clinical isolates derived from streptococcal tonsillitis (Conley et al. 2003). However, the lack of standardized laboratory methodology and reproducibility has been criticized, and clinical implication for this observation has been difficult to ascertain. A related growth dependent phenomenon was discovered by Eagle in 1952 in a murine S. pyogenes pyomyositis model (Eagle 1952). He reported that the efficacy of penicillin correlated to the size of the bacterial inoculum, and the beta-lactam antibiotic lost its potency when the bacterial burden increased beyond a certain point. It has become known as the Eagle effect, and is presumed to reflect the reduced activity of penicillin on S. pyogenes reaching a stationary plateau-phase in their growth curve. Later, Stevens et al. reproduced the results, but also demonstrated that clindamycin was not affected by growth phase or inoculum size, and significantly outperformed penicillin in the treatment of experimental pyomyositis (Stevens et al. 1988). A high inoculum of bacteria is a hallmark of NSTIs as compared to less severe SSTIs, and the Eagle effect is presumed to be of significance in the treatment of NSTIs. The superiority of clindamycin in bacteria in stationary growth phase, combined with its ability to abrogate toxin production in S. pyogenes, forms the experimental basis for the recommendations of adjunctive clindamycin treatment in NSTIs caused by this pathogen. Notably, the Eagle effect as well as cytotoxin production has also been verified in S. dysgalactiae-infections, but the potency of clindamycin in this setting has not been examined (Lam and Bayer 1983; Siemens et al. 2015).

7.4.5

95

Treatment of Biofilm

The organization of microbial communities into a biofilm has long been recognized as a phenotypic trait in several clinical pathogens, including S. aureus (Gebreyohannes et al. 2019). Recently, biofilm was revealed to also be implicated in NSTIs caused by S. pyogenes (Siemens et al. 2016). The encapsulation of the microbes by rich extracellular matrix and the distinct layering of bacteria in different metabolic states is the hallmark biofilm formation, and has been shown to significantly alter the efficacy of several antibiotics (Fiedler et al. 2015; Gebreyohannes et al. 2019; Conley et al. 2003). The beta-lactam antibiotics inhibit bacterial growth though disruption of the cell wall synthesis. However, in the biofilm community some bacteria enter an inactive stationary phase, rendering the beta-lactams ineffective. Conversely, antibacterial agents targeting DNA (fluoroquinolones), RNA (rifampicin) or protein synthesis (clindamycin and linezolid) appear to retain their activity also in biofilm, and have been suggested as potential therapeutic agents in biofilm-associated infections (Zimmerli and Sendi 2017; Chai et al. 2016). In S. pyogenes, biofilm formation has previously been linked to penicillin treatment failure in tonsillitis (Conley et al. 2003). The presence of S. pyogenes biofilm has been demonstrated in crypts in excised tonsillar tissue, and biofilm capacity and reduced susceptibility to beta-lactam antibiotics has been verified in S. pyogenes isolates obtained from recurrent tonsillitis cases (Conley et al. 2003; Roberts et al. 2012). Treatment with clindamycin appears to more effective in eradication of difficult to treat S. pyogenes tonsillitis cases, though at the expense of more adverse effect (Brook and Hirokawa 1985). How discovery of S. pyogenes biofilm in NSTIs translates into clinical practice has yet to be elucidated, but it likely strengthens the position of adjunctive clindamycin in the treatment of these infections.

96

7.5 7.5.1

O. Oppegaard and E. Rath

Antimicrobial Treatment of NSTIs Treatment Strategy

In the absence of clinical trials pertaining to the antimicrobial treatment of NSTIs, general recommendations for therapeutic management are predominantly based on expert opinion. Knowledge extrapolated from experimental studies on the pharmacodynamics and pharmacokinetics of relevant antibiotics as delineated above, combined with data from clinical studies on non-necrotizing skin and soft tissue infections, have often been used as guidance in the decision process. Importantly, the antimicrobial treatment needs to cover the etiologic spectrum of pathogens associated with NSTI. In general, two different strategies exist in the empiric treatment of infectious diseases. In some clinical conditions, including uncomplicated skins and soft tissue infections, the initiation of narrow spectrum antibiotics is advised, and treatment escalation can be considered in the event of therapeutic failure. Inversely, in more severe disease manifestations it is often recommended to administer broad spectrum antibiotics, and

subsequently de-escalate treatment based on culture and susceptibility data. Given the substantial mortality and morbidity associated with NSTIs and the importance of swift infection control, a de-escalation strategy is advocated by most guidelines, and empiric coverage of all pathogens frequently associated with this condition is essential (Stevens et al. 2014; Sunderkotter et al. 2019). However, the major geographic differences in antibiotic susceptibility described previously, highlights the need to adapt antimicrobial treatment to the local epidemiology of bacterial resistance. Moreover, regional differences in etiological distribution merit considerations, especially in areas where water exposure associated microbes (i.e. Vibrio and Aeromonas spp.) are reported to be more frequent (Hsiao et al. 2020). Hence, a globally valid guideline for treatment of NSTIs is perhaps not feasible.

7.5.2

Empirical Treatment

National guidelines for the treatment of NSTIs have been developed in several countries (Table 7.2), though many simply refer to the recommendations from the Infectious Diseases

Table 7.2 Overview of contemporary guideline recommendations for empirical treatment of NSTIs IDSA 2014 (Stevens et al. 2014) Standard Pip/tazo or Meropenem or 3. gen ceph and metro Plus Vancomycin MRSA or coverage Linezolid Plus NA toxin inhibit

WSES/SIS-E 2018 (Sartelli et al. 2018) Pip/tazo or Meropenem

Germany 2018 (Sunderkotter et al. 2019) Pip/tazo or Meropenem

France 2019 (Urbina et al. 2019) Pip/tazoa

Linezolid or Daptomycin Clindamycin

NA

NA

Norway 2013 (Helsedirektoratet 2013) Penicillinb and Gentamicin or 3. gen ceph and metro NA

Clindamycin

NA

Clindamycin

South Korea 2017 (Kwak et al. 2017) Pip/tazo or Meropenem or 3. gen ceph and metro Vancomycin or Linezolid NA

NSTIs necrotizing soft tissue infections, IDSA Infectious Diseases Society of America, WSES/SIS-E World Society of Emergency Surgery/ Surgical Infection Society—Europe, pip/tazo, piperacillin/tazobactam, 3. gen ceph, third generation cephalosporin (e.g. cefotaxime or ceftriaxone), metro metronidazole, MRSA methicillin-resistant Staphylococcus aureus, toxin inhibit toxin inhibitor, NA not advised as empirical treatment a The French guideline recommends to use meropenem or vancomycin in the presence of risk factors for ESBLceph or MRSA, respectively b The Norwegian guideline recommends using third generation cephalosporin and metronidazole, or even meropenem, in NSTIs emanating from the abdominal or genitourinary region

7

Treatment of Necrotizing Soft Tissue Infections: Antibiotics

Society of America (IDSA) or World Society of Emergency Surgery/Surgical Infections Society Europe (WSES/SIS-E) (Stevens et al. 2014; Sartelli et al. 2018). The various guidelines are largely overlapping, but demonstrate some important differences, predominantly reflecting the variations in antimicrobial resistance. One of the major ambiguities is the incorporation of ani-MRSA coverage in the empiric treatment regime. Understandably, guidelines originating from US (IDSA) and from Asia (South Korean guideline), where the proportion of methicillin resistance in S. aureus approximates 50%, advocate to initiate empiric treatment with an MRSA active antibiotic in all NSTI-patients (Stevens et al. 2014; Kwak et al. 2017). The inverse holds true in recommendations from regions of low MRSA endemicity, including central and northern Europe (Sunderkotter et al. 2019; Urbina et al. 2019; Helsedirektoratet 2013). Another important consideration is the selection of empiric treatment for Enterobacteriaceae. Most guidelines suggest administration of either piperacillin/tazobactam or meropenem, and leave it to the treating physician to decide the appropriate choice based on local resistance epidemiology (Stevens et al. 2014; Kwak et al. 2017; Sartelli et al. 2018; Sunderkotter et al. 2019). Both these agents have excellent broad spectrum activity against streptococci, staphylococci (excluding methicillin-resistant strains), anaerobes, and Enterobacteriaceae, but carbapenems are usually necessary to treat gram-negative bacteria harboring ESBLceph. In general, piperacillin/ tazobactam should be regarded as the first line of defense in areas with low prevalence of ESBLceph, but meropenem is warranted in geographic regions where resistance to piperacillin/ tazobactam is high. A third major strategic decision pertains to the administration of a protein synthesis inhibitor as part of the empiric treatment regime. The German guidelines and the WSES/SIS-E recommendations incorporate clindamycin in their therapeutic algorithm for all patients (Sunderkotter et al. 2019; Sartelli et al. 2018). However, the prevalence of S. pyogenes in

97

NSTIs emanating from the abdominal or genitourinary region is very low (Madsen et al. 2019). Accordingly, reserving the toxin inhibitors for NSTIs confined to the extremities or head/neck area is suggested in France (Urbina et al. 2019). The IDSA guidelines have also omitted clindamycin from their empiric management, but recommend administering a protein synthesis inhibitor to all cases of verified S. pyogenes and Clostridial NSTIs (Stevens et al. 2014). In Norway, a country with very low prevalence of both MRSA and ESBLceph, a narrow spectrum approach has been recommended in the national guidelines for empiric treatment of NSTIs located to the extremities; a combination of penicillin, gentamicin, and clindamycin (Helsedirektoratet 2013). Inversely, the local antimicrobial treatment recommendations in the neighboring country Denmark, advise meropenem and clindamycin as the treatment of choice for all cases of NSTIs (Rigshospitalet 2020). The rates of antimicrobial resistance in these two countries are comparable, but two diametrically opposing strategies have been chosen. Interestingly, in the prospective INFECT study including patients from both these countries, no significant difference in mortality was observed (Madsen et al. 2019). However, randomized clinical trials are needed to address the optimal treatment strategy in NSTIs, and to examine the feasibility of tailoring the antimicrobial therapy to the anatomical location.

7.5.3

Pathogen Specific Treatment

If the causative pathogen is identified, the antimicrobial treatment can often be de-escalated. The treatment recommendations for monomicrobial NSTIs are fairly similar across continents (Table 7.3). The combination of penicillin and clindamycin is uniformly recommended for the treatment of S. pyogenes NSTIs, to counteract bacterial toxin production, biofilm formation and the Eagle effect (Stevens et al. 2014; Helsedirektoratet 2013). However, therapeutic consideration in the event of clindamycin resistance are not discussed

98

O. Oppegaard and E. Rath

Table 7.3 Overview of contemporary guideline recommendations for treatment of NSTI with known etiology

S. pyogenes

S. aureus MSSA S. aureus MRSA Polymicrobial infections

V. vulnificus

A. hydrophilia

C. perfringens

IDSA 2014 (Stevens et al. 2014) Penicillin and Clindamycin Oxacillin Vancomycin or Linezolid Pip/tazo or Meropenem or 3. gen ceph and metro Doxycycline and Ceftriaxone Doxycycline and Ceftriaxone Penicillin and Clindamycin

WSES/SIS-E 2018 (Sartelli et al. 2018) NA

Germany 2018 (Sunderkotter et al. 2019) NA

France 2019 (Urbina et al. 2019) NA

NA

NA

NA

Norway 2013 (Helsedirektoratet 2013) Penicillin and Clindamycin NA

Linezolid or Daptomycin Pip/tazo or Meropenem

Linezolid or Daptomycin Pip/tazo or Meropenem

Vancomycin

NA

Pip/tazoa

3. gen ceph and metro or Meropenem

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

NA

Penicillin and Clindamycin

NSTIs necrotizing soft tissue infections, IDSA Infectious Diseases Society of America, WSES/SIS-E World Society of Emergency Surgery/Surgical Infection Society—Europe, pip/tazo, piperacillin/tazobactam, 3. gen ceph, third generation cephalosporin (e.g. cefotaxime or ceftriaxone), metro metronidazole, MSSA methicillin susceptible Staphylococcus aureus, MRSA methicillin-resistant Staphylococcus aureus, NA not addressed a The French guideline recommends to use meropenem or vancomycin in the presence of risk factors for ESBLceph or MRSA, respectively

in any guidelines. Although Andreoni et al. found evidence for a clindamycin mediated attenuation of toxin production in S. pyogenes even in the presence of clindamycin resistance, the size of the bacterial inoculum remained unchanged (Andreoni et al. 2017). Thus, the clinical implications have yet to be elucidated, and until further studies have been performed, it is advisable to switch to linezolid in clindamycinresistant cases. S. dysgalactiae is an emerging cause of NSTIs (Bruun et al. 2020; Oppegaard et al. 2015). Experimental data has demonstrated its capability of cytotoxin production, biofilm formation as well occurrence of the Eagle effect when exposed to penicillin (Siemens et al. 2015, 2016; Lam and Bayer 1983). There are no studies pertaining to

the adjunctive treatment of clindamycin in this pathogen, but the addition of a toxin inhibitor could be considered in monomicrobial S. dysgalactiae NSTIs, especially in the event of therapeutic failure. Different weight has been given to recent studies indicating a possible superiority for linezolid over vancomycin in SSTIs, especially for MRSA associated cases (Yue et al. 2016). Both IDSA guidelines and French recommendations for NSTI, as well as IDSA guidelines for MRSA infections, maintain vancomycin as the treatment of choice in MRSA infections (Liu et al. 2011; Stevens et al. 2014; Urbina et al. 2019). Differently, WSES/SIS-E and the German recommendations prefer linezolid or even daptomycin (Sartelli et al. 2018; Sunderkotter

7

Treatment of Necrotizing Soft Tissue Infections: Antibiotics

et al. 2019). The quality of the evidence favoring linezolid is very low, and both alternatives are likely to be equally successful. However, avoiding vancomycin in overt renal failure, and possibly in the presence of diabetic microangiopathy is advisable. The role of novel 5th generation cephalosporins and tigecycline in the treatment of MRSA NSTIs has not been firmly established, but should be reserved for cases were the first line options cannot be used. Vibrio spp. and Aeromonas spp. are still rare causes of NSTI in most parts of the world, and few guidelines address specific treatment recommendations for these pathogens. A combination of doxycycline and ceftriaxone is advocated in the South Korean and IDSA guidelines, although doxycycline plus ciprofloxacin might be preferable for Aeromonas infections (Stevens et al. 2014; Kwak et al. 2017). In South East Asia Vibrio spp. has become the dominant cause of monomicrobial NSTIs, implicated in 20% of the cases in a recent study from Taiwan (Hsiao et al. 2020). Accordingly, the South Korean guidelines advice empiric coverage for Vibrio spp. in patients with chronic liver conditions and recent sea water exposure (Kwak et al. 2017).

7.5.4

Treatment Duration

Studies investigating the optimal treatment duration in NSTI are lacking, and considerable variations in clinical practice has been observed. In a retrospective study from the US comparing the management and outcomes in NSTIs between three different hospitals, the treatment duration ranged from 7 to 28 days (Faraklas et al. 2016). Significant variations in treatment length were detected between the three sites, but no differences in clinical outcome was evident. However, the study was not designed to clarify the appropriate treatment duration. Several guidelines offer a pragmatic approach, and suggest continuation of treatment until surgery is no longer indicated, clinical improvement has occurred, and fever has been absent the past 72 h (Stevens et al. 2014; Sartelli et al. 2018).

7.6

99

Conclusions

Necrotizing soft tissue infection is a lifethreatening clinical condition, and rapid administration of potent antibiotics is essential. Empiric antimicrobial treatment should cover the likely culprit, and the therapeutic regimen should be adapted to local resistance epidemiology. The decision to broadly administer vancomycin or linezolid depends on the endemicity of MRSA in the geographic region. Similarly, in areas with low prevalence of ESBLceph, piperacillin/ tazobactam should be regarded as first line of defense, but meropenem might be warranted in regions where resistance to piperacillin/ tazobactam is high. Coverage of Vibrio spp. and Aeromonas spp. should not be forgotten in areas where these pathogens are endemic, particularly in patients with pre-existing chronic liver conditions and recent sea water exposure. The use of a protein synthesis inhibitor could be omitted in the empiric treatment regime in NSTIs emanating from the abdominal and genital areas, but is definitely indicated in cases localized to the extremities, and in verified S. pyogenes and Clostridial NSTIs. Furthermore, it should be strongly considered in monomicrobial cases caused by S. dysgalactiae and S. aureus. Randomized clinical trials investigating the optimal antimicrobial treatment of NSTIs are needed. Acknowledgment This work was supported by the European Union Seventh Framework Programme: (FP7/2007-2013) under the grant agreement 305340 (INFECT project), NordForsk (Project no. 90456, PerAID), the Norwegian Research Council under the frame of ERA PerMed (Project 2018-151, PerMIT) and The Swedish Research Council Framework Grant in infections and antibiotics (2014-8609-117204-47).

References Andreoni F, Zurcher C, Tarnutzer A, Schilcher K, Neff A, Keller N, Marques Maggio E, Poyart C, Schuepbach RA, Zinkernagel AS (2017) Clindamycin affects group A streptococcus virulence factors and improves clinical outcome. J Infect Dis 215:269–277 Arnaud FCS, Liborio AB (2020) Attributable nephrotoxicity of vancomycin in critically ill patients: a marginal

100 structural model study. J Antimicrob Chemother 75:1031–1037 Barbour A, Scaglione F, Derendorf H (2010) Classdependent relevance of tissue distribution in the interpretation of anti-infective pharmacokinetic/pharmacodynamic indices. Int J Antimicrob Agents 35:431–438 Barcia-Macay M, Seral C, Mingeot-Leclercq MP, Tulkens PM, van Bambeke F (2006) Pharmacodynamic evaluation of the intracellular activities of antibiotics against Staphylococcus aureus in a model of THP-1 macrophages. Antimicrob Agents Chemother 50:841–851 Baude J, Bastien S, Gillet Y, Leblanc P, Itzek A, Tristan A, Bes M, Duguez S, Moreau K, Diep BA, NorrbyTeglund A, Henry T, Vandenesch F, Group IS (2019) Necrotizing soft tissue infection Staphylococcus aureus but not S. pyogenes isolates display high rates of internalization and cytotoxicity toward human myoblasts. J Infect Dis 220:710–719 Brindle R, Williams OM, Davies P, HARRIS T, Jarman H, Hay AD, Featherstone P (2017) Adjunctive clindamycin for cellulitis: a clinical trial comparing flucloxacillin with or without clindamycin for the treatment of limb cellulitis. BMJ Open 7:e013260 Brindle R, Williams OM, Barton E, Featherstone P (2019) Assessment of antibiotic treatment of cellulitis and erysipelas: a systematic review and meta-analysis. JAMA Dermatol 155(9):1033–1040 Brook I, Hirokawa R (1985) Treatment of patients with a history of recurrent tonsillitis due to group A betahemolytic streptococci. A prospective randomized study comparing penicillin, erythromycin, and clindamycin. Clin Pediatr 24:331–336 Brook I, Wexler HM, Goldstein EJ (2013) Antianaerobic antimicrobials: spectrum and susceptibility testing. Clin Microbiol Rev 26:526–546 Bruun T, Rath E, Bruun Madsen M, Oppegaard O, Nekludov M, Arnell P, Karlsson Y, Babbar A, Bergey F, Itzek A, Hyldegaard O, Norrby-Teglund A, Skrede S; Group IS (2020) Risk factors and predictors of mortality in streptococcal necrotizing soft-tissue infections: a multicenter prospective study. Clin Infect Dis. https://doi.org/10.1093/cid/ciaa027 Burnham JP, Burnham CA, Warren DK, Kollef MH (2016) Impact of time to appropriate therapy on mortality in patients with vancomycin-intermediate Staphylococcus aureus infection. Antimicrob Agents Chemother 60:5546–5553 Carapetis JR, Jacoby P, Carville K, Ang SJ, Curtis N, Andrews R (2014) Effectiveness of clindamycin and intravenous immunoglobulin, and risk of disease in contacts, in invasive group a streptococcal infections. Clin Infect Dis 59:358–365 Castanheira M, Deshpande LM, Mendes RE, Canton R, Sader HS, Jones RN (2019) Variations in the occurrence of resistance phenotypes and carbapenemase genes among enterobacteriaceae isolates in 20 years of the SENTRY antimicrobial surveillance program. Open Forum Infect Dis 6:S23–S33

O. Oppegaard and E. Rath CDCgov (2010) Prevention of perinatal group B streptococcal disease, revised guidelines from CDC. MMWR Morb Mortal Wkly Rep 2010:1–32 Chai D, Liu X, Wang R, Bai Y, Cai Y (2016) Efficacy of linezolid and fosfomycin in catheter-related biofilm infection caused by methicillin-resistant Staphylococcus aureus. Biomed Res Int 2016:6413982 Conley J, Olson ME, Cook LS, Ceri H, Phan V, Davies HD (2003) Biofilm formation by group a streptococci: is there a relationship with treatment failure? J Clin Microbiol 41:4043–4048 Coyle EA, Society of Infectious Diseases, P (2003) Targeting bacterial virulence: the role of protein synthesis inhibitors in severe infections. Insights from the Society of Infectious Diseases Pharmacists. Pharmacotherapy 23:638–642 Coyle EA, Cha R, Rybak MJ (2003) Influences of linezolid, penicillin, and clindamycin, alone and in combination, on streptococcal pyrogenic exotoxin a release. Antimicrob Agents Chemother 47:1752–1755 Dahesh S, Hensler ME, van Sorge NM, Gertz RE, Schrag S, Nizet V, Beall BW (2008) Point mutation in the group B streptococcal pbp2x gene conferring decreased susceptibility to beta-lactam antibiotics. Antimicrob Agents Chemother 52:2915–2918 Das DK, Baker MG, Venugopal K (2012) Risk factors, microbiological findings and outcomes of necrotizing fasciitis in New Zealand: a retrospective chart review. BMC Infect Dis 12:348 Dechet AM, Yu PA, Koram N, Painter J (2008) Nonfoodborne Vibrio infections: an important cause of morbidity and mortality in the United States, 19972006. Clin Infect Dis 46:970–976 Diekema DJ, Hsueh PR, Mendes RE, Pfaller MA, Rolston KV, Sader HS, Jones RN (2019a) The microbiology of bloodstream infection: 20-year trends from the SENTRY antimicrobial surveillance program. Antimicrob Agents Chemother 63(7):e00355 Diekema DJ, Pfaller MA, Shortridge D, Zervos M, Jones RN (2019b) Twenty-year trends in antimicrobial susceptibilities among Staphylococcus aureus from the SENTRY antimicrobial surveillance program. Open Forum Infect Dis 6:S47–S53 Duncan LR, Smith CJ, Flamm RK, Mendes RE (2019) Regional analysis of telavancin and comparator antimicrobial activity against multidrug-resistant Staphylococcus aureus collected in the USA 2014-2016. J Glob Antimicrob Resist 20:118–123 Eagle H (1952) Experimental approach to the problem of treatment failure with penicillin. I. Group A streptococcal infection in mice. Am J Med 13:389–399 eCDC (2019) Surveillance of antimicrobial resistance in Europe 2018. European Centre for Disease Prevention and Control. https://www.ecdc.europa.eu/en/ publications-data/surveillance-antimicrobial-resis tance-europe-2018. Accessed 26 March 2020 Faraklas I, Yang D, Eggerstedt M, Zhai Y, Liebel P, Graves G, Dissanaike S, Mosier M, Cochran A (2016) A multi-center review of care patterns and

7

Treatment of Necrotizing Soft Tissue Infections: Antibiotics

outcomes in necrotizing soft tissue infections. Surg Infect 17:773–778 Farrell DJ, Flamm RK, Sader HS, Jones RN (2013) Spectrum and potency of ceftaroline tested against leading pathogens causing skin and soft-tissue infections in Europe. Int J Antimicrob Agents 41:337–342 Ferrer R, Martin-Loeches I, Phillips G, Osborn TM, Townsend S, Dellinger RP, Artigas A, Schorr C, Levy MM (2014) Empiric antibiotic treatment reduces mortality in severe sepsis and septic shock from the first hour: results from a guideline-based performance improvement program. Crit Care Med 42:1749–1755 Fiedler T, Koller T, Kreikemeyer B (2015) Streptococcus pyogenes biofilms-formation, biology, and clinical relevance. Front Cell Infect Microbiol 5:15 Fuursted K, Stegger M, Hoffmann S, Lambertsen L, Andersen PS, Deleuran M, Thomsen MK (2016) Description and characterization of a penicillinresistant Streptococcus dysgalactiae subsp. equisimilis clone isolated from blood in three epidemiologically linked patients. J Antimicrob Chemother 71:3376–3380 Gajdacs M, Spengler G, Urban E (2017) Identification and antimicrobial susceptibility testing of anaerobic bacteria: Rubik’s cube of clinical microbiology? Antibiotics 6(4):25 Gebreyohannes G, Nyerere A, Bii C, Sbhatu DB (2019) Challenges of intervention, treatment, and antibiotic resistance of biofilm-forming microorganisms. Heliyon 5:e02192 Helsedirektoratet (2013) Norwegian national guidelines for antimicrobial treatment of necrotizing soft tissue infections. https://www.helsedirektoratet.no/ retningslinjer/antibiotika-i-sykehus/hud-ogblotdelsinfeksjoner/nekrotiserendeblotdelsinfeksjoner?malgruppe¼undefined. Accessed 26 March 2020 Hertzen E, Johansson L, Kansal R, Hecht A, Dahesh S, Janos M, Nizet V, Kotb M, Norrby-Teglund A (2012) Intracellular Streptococcus pyogenes in human macrophages display an altered gene expression profile. PLoS One 7:e35218 Hsiao CT, Chang CP, Huang TY, Chen YC, Fann WC (2020) Prospective validation of the laboratory risk indicator for necrotizing fasciitis (LRINEC) score for necrotizing fasciitis of the extremities. PLoS One 15: e0227748 JANIS (2018) Annual surveillance report on antimicrobial resistance in Japan: 2018. https://janis.mhlw.go.jp/ english/report/index.html. Accessed 26 March 2020 Kaplan EL, Chhatwal GS, Rohde M (2006) Reduced ability of penicillin to eradicate ingested group A streptococci from epithelial cells: clinical and pathogenetic implications. Clin Infect Dis 43:1398–1406 Kiang TK, Hafeli UO, Ensom MH (2014) A comprehensive review on the pharmacokinetics of antibiotics in interstitial fluid spaces in humans: implications on dosing and clinical pharmacokinetic monitoring. Clin Pharmacokinet 53:695–730

101

Kim T, Park SY, Kwak YG, Jung J, Kim MC, Choi SH, Yu SN, Hong HL, Kim YK, Park SY, Song EH, Park KH, Cho OH, Choi SH, Korean SSG (2019) Etiology, characteristics, and outcomes of community-onset necrotizing fasciitis in Korea: a multicenter study. PLoS One 14:e0218668 King M, Rose L, Fraimow H, Nagori M, Danish M, Doktor K (2019) Vibrio vulnificus infections from a previously nonendemic area. Ann Intern Med 171 (7):520–521 Kwak YG, Choi SH, Kim T, Park SY, Seo SH, Kim MB, Choi SH (2017) Clinical guidelines for the antibiotic treatment for community-acquired skin and soft tissue infection. Infect Chemother 49:301–325 Lam K, Bayer AS (1983) Serious infections due to group G streptoccocci. Report of 15 cases with in vitro-in vivo correlations. Am J Med 75:561–570 Lee TC, Carrick MM, Scott BG, Hodges JC, Pham HQ (2007) Incidence and clinical characteristics of methicillin-resistant Staphylococcus aureus necrotizing fasciitis in a large urban hospital. Am J Surg 194:809–812 Lee AS, de Lencastre H, Garau J, Kluytmans J, MalhotraKumar S, Peschel A, Harbarth S (2018) Methicillinresistant Staphylococcus aureus. Nat Rev Dis Primers 4:18033 Lewis JS, Lepak AJ, Thompson GR, Craig WA, Andes DR, Sabol-Dzintars KE, Jorgensen JH (2014) Failure of clindamycin to eradicate infection with betahemolytic streptococci inducibly resistant to clindamycin in an animal model and in human infections. Antimicrob Agents Chemother 58:1327–1331 Linner A, Darenberg J, Sjolin J, Henriques-Normark B, Norrby-Teglund A (2014) Clinical efficacy of polyspecific intravenous immunoglobulin therapy in patients with streptococcal toxic shock syndrome: a comparative observational study. Clin Infect Dis 59:851–857 Liu C, Bayer A, Cosgrove SE, Daum RS, Fridkin SK, Gorwitz RJ, Kaplan SL, Karchmer AW, Levine DP, Murray BE, Talan DA, Chambers HF, Infectious Diseases Society of, A (2011) Clinical practice guidelines by the infectious diseases society of america for the treatment of methicillin-resistant Staphylococcus aureus infections in adults and children. Clin Infect Dis 52:e18–e55 Louis A, Savage S, Utter GH, Li SW, Crandall M (2019) NSTI organisms and regions: a multicenter study from the American Association for the Surgery of Trauma. J Surg Res 243:108–113 Lu B, Fang Y, Fan Y, Chen X, Wang J, Zeng J, Li Y, Zhang Z, Huang L, Li H, Li D, Zhu F, Cui Y, Wang D (2017) High prevalence of macrolide-resistance and molecular characterization of Streptococcus pyogenes isolates circulating in China from 2009 to 2016. Front Microbiol 8:1052 Madsen MB, Skrede S, Perner A, Arnell P, Nekludov M, Bruun T, Karlsson Y, Hansen MB, Polzik P,

102 Hedetoft M, Rosen A, Saccenti E, Bergey F, Martins Dos Santos VAP, Norrby-Teglund A, Hyldegaard O, Group IS (2019) Patient’s characteristics and outcomes in necrotising soft-tissue infections: results from a Scandinavian, multicentre, prospective cohort study. Intensive Care Med 45:1241–1251 Mascini EM, Jansze M, Schouls LM, Verhoef J, van Dijk H (2001) Penicillin and clindamycin differentially inhibit the production of pyrogenic exotoxins A and B by group A streptococci. Int J Antimicrob Agents 18:395–398 Melzer M, Petersen I (2007) Mortality following bacteraemic infection caused by extended spectrum beta-lactamase (ESBL) producing E. coli compared to non-ESBL producing E. coli. J Infect 55:254–259 Miller LG, Perdreau-Remington F, Rieg G, Mehdi S, Perlroth J, Bayer AS, Tang AW, Phung TO, Spellberg B (2005) Necrotizing fasciitis caused by communityassociated methicillin-resistant Staphylococcus aureus in Los Angeles. N Engl J Med 352:1445–1453 Minichmayr IK, Schaeftlein A, Kuti JL, Zeitlinger M, Kloft C (2017) Clinical determinants of target non-attainment of linezolid in plasma and interstitial space fluid: a pooled population pharmacokinetic analysis with focus on critically ill patients. Clin Pharmacokinet 56:617–633 NORM/NORM-VET (2018) Usage og antimicrobial agents and occurrence of antimicrobial resistance in Norway. https://unn.no/fag-og-forskning/norm-norskovervakingssystem-for-antibiotikaresistens-hosmikrober. Accessed 26 Mar 2020 Oppegaard O, Mylvaganam H, Kittang BR (2015) Betahaemolytic group A, C and G streptococcal infections in Western Norway: a 15-year retrospective survey. Clin Microbiol Infect 21:171–178 Pea F (2016) Practical concept of pharmacokinetics/pharmacodynamics in the management of skin and soft tissue infections. Curr Opin Infect Dis 29:153–159 Peetermans M, de Prost N, Eckmann C, Norrby-TeglundA, Skrede S, de Waele JJ (2020) Necrotizing skin and soft-tissue infections in the intensive care unit. Clin Microbiol Infect 26:8–17 Perianez-Parraga L, Martinez-Lopez I, Ventayol-Bosch P, Puigventos-Latorre F, Delgado-Sanchez O (2012) Drug dosage recommendations in patients with chronic liver disease. Rev Esp Enferm Dig 104:165–184 Public Health England (2018) Laboratory surveillance of pyogenic and non-pyogenic streptococcal bacteraemia in England, Wales and Northern Ireland: 2017. Health Protection Report, 41 Rigshospitalet (2020) Danish guidelines for antimicrobial treatment of necrotizing soft tissue infections. https:// www.rigshospitalet.dk/afdelinger-og-klinikker/ diagnostisk/klinisk-mikrobiologisk-afdeling/for_ fagfolk/Sider/retningslinjer-for-antibiotika.aspx. Accessed 26 March 2020 Roberts AL, Connolly KL, Kirse DJ, Evans AK, Poehling KA, Peters TR, Reid SD (2012) Detection of group A Streptococcus in tonsils from pediatric patients reveals

O. Oppegaard and E. Rath high rate of asymptomatic streptococcal carriage. BMC Pediatr 12:3 Rolston KV, Chandrasekar PH, Lefrock JL (1984) Antimicrobial tolerance in group C and group G streptococci. J Antimicrob Chemother 13:389–392 Sartelli M, Guirao X, Hardcastle TC, Kluger Y, Boermeester MA, Rasa K, Ansaloni L, Coccolini F, Montravers P, Abu-Zidan FM, Bartoletti M, Bassetti M, Ben-Ishay O, Biffl WL, Chiara O, Chiarugi M, Coimbra R, de Rosa FG, de Simone B, di Saverio S, Giannella M, Gkiokas G, Khokha V, Labricciosa FM, Leppaniemi A, Litvin A, Moore EE, Negoi I, Pagani L, Peghin M, Picetti E, Pintar T, Pupelis G, Rubio-Perez I, Sakakushev B, SegoviaLohse H, Sganga G, Shelat V, Sugrue M, Tarasconi A, Trana C, Ulrych J, Viale P, Catena F (2018) 2018 WSES/SIS-E consensus conference: recommendations for the management of skin and soft-tissue infections. World J Emerg Surg 13:58 Shumba P, Mairpady Shambat S, Siemens N (2019) The Role of Streptococcal and Staphylococcal exotoxins and proteases in human necrotizing soft tissue infections. Toxins 11:6 Siemens N, Kittang BR, Chakrakodi B, Oppegaard O, Johansson L, Bruun T, Mylvaganam H, Svensson M, Skrede S, Norrby-Teglund A, Group IS (2015) Increased cytotoxicity and streptolysin O activity in group G streptococcal strains causing invasive tissue infections. Sci Rep 5:16945 Siemens N, Chakrakodi B, Shambat SM, Morgan M, Bergsten H, Hyldegaard O, Skrede S, Arnell P, Madsen MB, Johansson L, Juarez J, Bosnjak L, Morgelin M, Svensson M, Norrby-Teglund A, Group IS (2016) Biofilm in group A streptococcal Necrotizing soft tissue infections. JCI Insight 1:e87882 Skhirtladze K, Hutschala D, Fleck T, Thalhammer F, Ehrlich M, Vukovich T, Muller M, Tschernko EM (2006) Impaired target site penetration of vancomycin in diabetic patients following cardiac surgery. Antimicrob Agents Chemother 50:1372–1375 Spellberg B, Talbot GH, Boucher HW, Bradley JS, Gilbert D, Scheld WM, Edwards J, Bartlett JG, Antimicrobial Availability Task Force of the Infectious Diseases Society of A (2009) Antimicrobial agents for complicated skin and skin-structure infections: justification of noninferiority margins in the absence of placebo-controlled trials. Clin Infect Dis 49:383–391 Stevens DL, Bryant AE (2017) Necrotizing soft-tissue infections. N Engl J Med 377:2253–2265 Stevens DL, Gibbons AE, Bergstrom R, Winn V (1988) The Eagle effect revisited: efficacy of clindamycin, erythromycin, and penicillin in the treatment of streptococcal myositis. J Infect Dis 158:23–28 Stevens DL, Aldape MJ, Bryant AE (2012) Lifethreatening clostridial infections. Anaerobe 18:254–259 Stevens DL, Bisno AL, Chambers HF, Dellinger EP, Goldstein EJ, Gorbach SL, Hirschmann JV, Kaplan SL, Montoya JG, Wade JC, Infectious Diseases

7

Treatment of Necrotizing Soft Tissue Infections: Antibiotics

Society of A (2014) Practice guidelines for the diagnosis and management of skin and soft tissue infections: 2014 update by the Infectious Diseases Society of America. Clin Infect Dis 59:e10–e52 Sunderkotter C, Becker K, Eckmann C, Graninger W, Kujath P, Schofer H (2019) S2k guidelines for skin and soft tissue infections Excerpts from the S2k guidelines for “calculated initial parenteral treatment of bacterial infections in adults - update 2018”. J Dtsch Dermatol Ges 17:345–369 Thanert R, Itzek A, Hossmann J, Hamisch D, Madsen MB, Hyldegaard O, Skrede S, Bruun T, Norrby-Teglund A, Medina E, Pieper DH, Group IS (2019) Molecular profiling of tissue biopsies reveals unique signatures associated with streptococcal necrotizing soft tissue infections. Nat Commun 10:3846 Tirupathi R, Areti S, Salim SA, Palabindala V, Jonnalagadda N (2019) Acute bacterial skin and soft tissue infections: new drugs in ID armamentarium. J Commun Hosp Intern Med Perspect 9:310–313 Udy AA, Roberts JA, Lipman J (2013) Clinical implications of antibiotic pharmacokinetic principles in the critically ill. Intensive Care Med 39:2070–2082 Urbina T, Hua C, Sbidian E, Bosc R, Tomberli F, Lepeule R, Decousser JW, Mekontso Dessap A, Chosidow O, de Prost N, Henri Mondor Hospital Necrotizing Fasciitis, G (2019) Impact of a multidisciplinary care bundle for necrotizing skin and soft tissue infections: a retrospective cohort study. Ann Intensive Care 9:123

103

van Hal SJ, Jensen SO, Vaska VL, Espedido BA, Paterson DL, Gosbell IB (2012) Predictors of mortality in Staphylococcus aureus bacteremia. Clin Microbiol Rev 25:362–386 Vannice K, Ricaldi J, Nanduri S, Fang FC, Lynch J, Bryson-Cahn C, Wright T, Duchin J, Kay M, Chochua S, van Beneden C, Beall B (2019) Streptococcus pyogenes pbp2x mutation confers reduced susceptibility to beta-lactam antibiotics. Clin Infect Dis 71 (1):201–204 Wong CH, Chang HC, Pasupathy S, Khin LW, Tan JL, Low CO (2003) Necrotizing fasciitis: clinical presentation, microbiology, and determinants of mortality. J Bone Joint Surg Am 85:1454–1460 Wu VC, Wang YT, Wang CY, Tsai IJ, Wu KD, Hwang JJ, Hsueh PR (2006) High frequency of linezolidassociated thrombocytopenia and anemia among patients with end-stage renal disease. Clin Infect Dis 42:66–72 Yue J, Dong BR, Yang M, Chen X, Wu T, Liu GJ (2016) Linezolid versus vancomycin for skin and soft tissue infections. Cochrane Database Syst Rev 7:CD008056 Zhao-Fleming HH, Wilkinson JE, Larumbe E, Dissanaike S, Rumbaugh K (2019) Obligate anaerobes are abundant in human necrotizing soft tissue infection samples - a metagenomics analysis. APMIS 127:577–587 Zimmerli W, Sendi P (2017) Orthopaedic biofilm infections. APMIS 125:353–364

8

Treatment of Necrotizing Soft Tissue Infections: IVIG Martin Bruun Madsen, Helena Bergsten, and Anna NorrbyTeglund

Contents 8.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106

8.2

Immunoglobulins: Structure and Function of Different Ig Isotypes . . . . . . 106

8.3 8.3.1 8.3.2 8.3.3 8.3.4

IVIG Preparations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Manufacturing and Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanisms of Action . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Therapeutic Indications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

108 108 108 111 112

8.4 8.4.1 8.4.2 8.4.3 8.4.4

Rationale for the Use of IVIG for NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Laboratory Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Clinical Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . IVIG in NSTI, Unresolved Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

116 116 117 119 119

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120

Abstract

Immunoglobulins are key effector molecules in the humoral immune response. Intravenous polyspecific immunoglobulin (IVIG) is a preparation of polyclonal serum immunoglobulins, typically IgG, from thousands of donors. It has been used as adjunctive therapy in critically ill patients with severe infections, i.e. sepsis, M. B. Madsen Department of Intensive Care, Copenhagen University Hospital, Rigshospitalet, Copenhagen, Denmark H. Bergsten · A. Norrby-Teglund (*) Center for Infectious Medicine, Department of Medicine Huddinge, Karolinska Institutet, Karolinska University Hospital, Huddinge, Sweden e-mail: [email protected]

septic shock, and necrotizing soft tissue infections. IVIG has been used for patients with severe invasive group A streptococcal infection since the early nineties and off-label use of IVIG for necrotizing soft tissue infections is common. It is also used for a variety of autoimmune, inflammatory, and immunodeficiency diseases. A meta-analysis of the clinical studies available for IVIG use in group A streptococcal toxic shock syndrome indicates a survival benefit. A blinded, placebo-controlled clinical trial (INSTINCT) assessed the effect of IVIG in 100 intensive care unit patients with necrotizing soft tissue infections, including all bacterial etiologies. The study did not demonstrate any effect on

# Springer Nature Switzerland AG 2020 A. Norrby-Teglund et al. (eds.), Necrotizing Soft Tissue Infections, Advances in Experimental Medicine and Biology 1294, https://doi.org/10.1007/978-3-030-57616-5_8

105

106

M. B. Madsen et al.

self-reported physical functioning at 6 months. In this chapter, we review the mechanisms of action of IVIG and the clinical studies that are available for necrotizing soft tissue infections as well as severe group A streptococcal infections. Keywords

Intravenous immunoglobulin · Streptococcus pyogenes Group A streptococcus · Superantigen neutralization · Necrotizing soft tissue infections Highlights • IVIG can neutralize virulence factors expressed by group A streptococcus and Staphylococcus aureus. • Epidemiologic studies of IVIG use in streptococcal toxic shock syndrome demonstrate a survival benefit in clindamycin treated patients. • Only one RCT of IVIG versus placebo has been conducted in patients with necrotizing soft tissue infection, and no difference in physical quality of life between the two groups was demonstrated.

8.1

Introduction

Immunoglobulins, also known as antibodies, are glycoproteins secreted by plasma cells, and are an important part of the immune system where they help neutralize pathogens. Immunoglobulins for medical use are made from human serum and are used for a variety of indications. Already in 1890, von Behring and Kitasato reported of an agent in the blood that could neutralize diphtheria toxin— “Antikörper,” and proposed that this agent in serum could react against foreign antigens (Schroeder and Cavacini 2010). In the 1930s, clinical studies by McKhann showed that placental extracts could prevent measles (Robinson and McKhann 1935), and by 1939, electrophoresis and later cold fractioning made it possible for serum to be separated into albumin and globulin fractions, and immunoglobulins to be separated

into different groups (heavy (IgM), regular (IgA, IgE, IgD, and IgG), and light (light chain dimers) (Gordon 1987). The product was used to prevent measles and hepatitis, and for the treatment of sepsis due to pneumococci in pediatric patients with agammaglobulinemia (Bruton 1952). Immunoglobulins have also been used for a variety of autoimmune and inflammatory diseases. The first reports of its use in patients with toxic shock syndrome caused by group A streptococcus (GAS) were published in 1992 (Barry et al. 1992; Stegmayr et al. 1992). In this chapter we will review the structure and function of immunoglobulins and discuss the indications of intravenous immunoglobulins with emphasis on necrotizing soft tissue infections.

8.2

Immunoglobulins: Structure and Function of Different Ig Isotypes

Immunoglobulins (Ig) are heterodimeric Y-shaped glycoproteins, highly abundant in plasma with concentrations of 7–12 g/L. The Ig molecules are composed of two identical heavy polypeptide chains and two identical light chains bridged together by disulphide bonds (Schroeder and Cavacini 2010). Each chain contains an NH2terminal variable region and a COOH-terminal constant (Fc) region (Fig. 8.1). Functionally, the molecules are separated by a hinge region into a variable, antigen binding, Fab part, and a constant Fc-part that mediate immunomodulatory effector function through complement interaction and binding to cellular Fc receptors (Schroeder and Cavacini 2010) (Fig. 8.1). The Fc region of the heavy chain defines the five classes (isotypes) of Ig: IgA, IgD, IgE, IgG, and IgM (Black 1997). IgG exists in four subclasses IgG1-IgG4, and IgA in two subclasses IgA1 and IgA2. The isotypes and Ig subclasses differ with respect to abundance, location, and effector function. IgG is the most abundant Ig isotype in serum and provides the majority of pathogen-specific antibodies involved in humoral immunity (Vidarsson et al. 2014). The four IgG subclasses vary greatly in abundance with IgG1 being most

8

Treatment of Necrotizing Soft Tissue Infections: IVIG

107

Fig. 8.1 Schematic model of the immunoglobulin structure. Fab fragment antigen binding, Fc fragment constant, VL variable region of the light chain, CL constant region of the light chain, VH variable region of the heavy chain, CH constant region of the heavy chain, SS disulfide

abundant (60%) followed by IgG2 and considerably less of IG3 and IgG4. In addition, the subclasses vary in function due to differences in the Fc region. IgG1 and IgG3 have a stronger binding to C1q and to Fc receptors (FcR), and are common in response to protein antigens (Vidarsson et al. 2014). In contrast, IgG2 dominate responses to bacterial polysaccharides (Ferrante et al. 1990; Siber et al. 1980), while IgG4 to allergens (Vidarsson et al. 2014). IgA is the second most abundant isotype in serum, and the principal Ig isotype in all mucosal secretions such as saliva, tears, and secretions covering the respiratory, gastrointestinal, and urogenital tract (Woof and Kerr 2006). IgA therefore represents an important first line of defense towards inhaled or ingested pathogens, neutralizing toxins and inhibiting colonization of epithelial surfaces, while at the same time tolerating commensals and innocuous antigens. In mucosal secretions, IgA exists predominantly in a dimeric form where the two IgA molecules are linked together by a small polypeptide, called J chain. The dimeric IgA also includes the

secretory component; a residual polypeptide of the extracellular proteolytic fragment of the polymeric Ig receptor required for transport of dimeric IgA into the secretions. The complex of dimeric IgA, J chain and secretory component is referred to as secretory IgA (sIgA). In serum, monomeric IgA is more common. Human IgA exists in two subclasses, IgA1 and IgA2, and IgA1 is the dominant subclass in serum as well as in many mucosal secretions. However in colon, IgA2 is most frequent (Brandtzaeg and Johansen 2005). IgM is the first isotype to be produced in a humoral immune response and it offers an important early defense against infections. IgM exists in a large pentameric form, displaying 10 antigenbinding sites and consequently high avidity interaction with antigens. The pentameric structure also supports an efficient complement activation (Schroeder and Cavacini 2010). Both IgD and IgE are found at very low levels in serum. Despite this, IgE is highly potent and predominates in the immune response to allergens and helminths parasites.

108

8.3 8.3.1

M. B. Madsen et al.

IVIG Preparations Manufacturing and Composition

IVIG is prepared from pooled plasma from several thousands of blood donors and the resulting polyvalent Ig preparation is supplied either in a liquid or a lyophilized form. As a human bloodderived product, manufacturing is tightly regulated starting with the selection and testing of individual donors to ensure quality and safety control (Medicines Agency E 2018; WHO 2020). Important quality and safety aspects include retained structural and functional characteristics of the Ig molecules, no infectious agents, no procoagulant activity and absence of agglutination of anti-A and anti-B hemagglutinins (Radosevich and Burnouf 2010). To this end, the manufacturing of IVIG include several procedures to remove such factors that might be present in the plasma pool, including not the least multiple steps (e.g. solvent/detergent treatment, pasteurization, and nanofiltration) to reduce the risk of disease transmission including viruses and prions (for further detail see Sect. 8.3.4). There are many IVIG preparations available on the market, most of which are composed of IgG, but there are also preparations enriched in IgM and IgA (Table 8.1). The IgG subclass distribution in the IVIG preparation is comparable to that found in serum with IgG1 being most abundant followed by IgG2, and only small amounts of IgG3 and IgG4. The Ig molecules are polyvalent and offers a broad spectrum of antibodies specific for a variety of foreign antigens (e.g. viral, bacterial, parasitic, etc.), but also antibodies to selfantigens, such as towards cytokines (natural

autoantibodies) and antibodies (anti-idiotypic antibodies). As a consequence of the plasma pool originating from different donors as well as variations in manufacturing procedures, IVIG preparations will vary in their final Ig composition and their functionality. This is evident not only between different IVIG preparations but also between batches of the same IVIG preparation (Dhainaut et al. 2013; Norrby-Teglund et al. 1998).

8.3.2

Mechanisms of Action

IVIG has several mechanisms of actions attributed to the dual functionality of Ig molecules: (1) as cell-surface receptors allowing for cellular signaling and activation, and (2) as soluble effector molecules binding and neutralizing antigens. This results in a direct targeting of disease-causing antigens or agents, and numerous immunomodulatory functions that have been implicated in the mechanistic action of IVIG in infections, as well as in autoimmune and acute inflammatory diseases (Gelfand 2012; Schwab and Nimmerjahn 2013). In the subsection below, we focus on the role of IVIG in acute infectious diseases, such as sepsis, toxic shock syndrome and necrotizing soft tissue infections (NSTI) with special attention on how IVIG can neutralize specific virulence factors. However, also the potent immunomodulatory functions of IVIG are of key importance as these conditions are characterized by dysregulated host responses (Johansson and Norrby-Teglund 2012; NorrbyTeglund et al. 2003; Steinhagen et al. 2020). The immunomodulatory effects of IVIG are numerous and include both Fc- and

Table 8.1 Different types of IVIG preparations used in randomized placebo-controlled trials of IVIG as adjunctive therapy for severe acute bacterial infections Commercial name Privigen Endobulin Polyglobin Trimodulin Pentaglobin

Ig content >98% IgG >98% IgG >97% IgG 56% IgG; 23% IgM; 21% IgA 76% IgG; 12% IgM; 12% IgA

Study subjects NSTI STSS Sepsis Severe CAP Sepsis, septic shock

Reference Madsen et al. (2017) Darenberg et al. (2003) Werdan et al. (2007) Welte et al. (2018) Rodríguez et al. (2005)

NSTI necrotizing soft tissue infections, STSS Streptococcal toxic shock syndrome, CAP community acquired pneumonia

8

Treatment of Necrotizing Soft Tissue Infections: IVIG

109

Fab-mediated effects (Schwab and Nimmerjahn 2013). Fab-dependent effects comprise (1) killing of target cells through antibody-dependent cytotoxicity, (2) blockade of cellular receptors and consequently cell–cell interactions, (3) neutralization of autoantigens (e.g. cytokines, FAS or FAsL, sialic acid-binding Ig-like lectin (SIGLECs) expressed on neutrophils and eosinophils) and autoantibodies by anti-idiotypic antibodies, and (4) scavenging of the anaphylatoxins C3a and C5a. The Fc-dependent effects comprise (1) blockade of activating Fc-gamma-receptors, (2) saturation of neonatal FcR (FcRn) resulting in an accelerated IgG catabolism, (3) expansion of regulatory T cell population, (4) modulation of FcγR expression on innate immune cells and B cells, and modulation of dendritic cells. FcγRIIIB has been shown in several studies to be involved in the immunomodulatory activities of IVIG. This inhibitory FcR is either upregulated on myeloid cells or the number of myeloid cells expressing this receptor is increased following IVIG administration (Bruhns et al. 2003; Kaneko et al. 2006; Samuelsson et al. 2001). Many of these IVIG-mediated effects result in immunosuppressive and antiinflammatory responses, which are likely to have impact on hyperinflammatory conditions like NSTI and toxic shock syndrome (NorrbyTeglund et al. 2000). However, it should be noted that the vast majority of studies investigating IVIG-mediated immunomodulation have been performed in autoimmune or inflammatory diseases. With regard to IVIG mediating neutralization of virulence factors through specific antibodies, more data are available. In sepsis, studies have reported that low Ig-levels are associated with increased mortality (Bermejo-Martín et al. 2014; Martin-Loeches et al. 2017). A similar association between lack of protective humoral immunity and severity of infection has been reported in GAS infections, where patients with streptococcal toxic shock syndrome or NSTI have lower levels of protective antibodies to streptococcal virulence factors than patients with less severe GAS infections (Babbar et al. 2018; Eriksson et al. 1999; Norrby-Teglund et al. 1994). In light of

this, IVIG was explored as a source of such protective antibodies towards streptococcal virulence factors. Studies revealed that IVIG contains broad spectrum antibodies that efficiently neutralize bacterial virulence factors shown to be key contributors to severe GAS infection (Table 8.2). This is not surprising, considering the abundance of asymptomatic carriage (especially in children) and frequency of mild streptococcal diseases in the population (Carapetis et al. 2005). The antibodies target both surface attached virulence factors, including the classical virulence factor M protein, resulting in increased opsonophagocytic killing (Table 8.2 and Fig. 8.2). In a recent study applying immunoprecipitation of GAS cell wall extracts revealed a broad range of surface proteins, e.g. M protein, C5a peptidase, S. pyogenes cell envelope protease (SpyCEP), phosphoglycerate kinase, that were recognized by IVIG (Reglinski et al. 2015). Also, secreted factors including superantigens, proteases, DNases, and cytolytic toxins are inhibited by IVIG, which leads to reduced pro-inflammatory responses (Table 8.2). Similarly, IVIG has been reported to contain antibodies that neutralize the activity of S. aureus virulence factors involved in toxic shock syndrome and NSTI, including superantigens and cytolytic toxins such as Panton Valentine leukocidin (PVL) and α-hemolysin (Table 8.2). IVIG was tested in an in vivo model of streptococcal toxic shock syndrome using HLA class II transgenic mice with increased susceptibility to GAS superantigens (Sriskandan et al. 2006). The results revealed that IVIG treatment resulted in improved bacterial clearance, neutralization of circulating superantigen, reduced pro-inflammatory responses, and reduce illness severity. Similarly, in an experimental murine model of necrotizing fasciitis caused by GAS, IVIG administration resulted in smaller lesions and reduced levels of several important GAS virulence factors, including the SLO, Sda1, and SpyCEP (Tarnutzer et al. 2019). However, in this study, no effect on bacterial load in treated and non-treated mice, indicating that the clinical treatment effect noted in the infected mice was linked to neutralization of virulence factors.

110

M. B. Madsen et al.

Table 8.2 Neutralizing antibodies in IVIG that target streptococcal and staphylococcal virulence factors Virulence factors IVIG effect Group A streptococcus M proteins Opsonophagocytic killing Serine protease Inhibition of IL8 cleaving SpyCEP activity Superantigens Neutralization of immune stimulation Streptolysin O Inhibition of hemolytic activity (SLO) The DNase Sda1 Inhibition of DNase activity SDSE Streptokinase Inhibition of Ska-mediated (Ska) fibrinolysis Staphylococcus aureus Cell wall proteins Opsonophagocytic killing Superantigens Neutralization of immune stimulation LukAB (LukGH) Inhibition of cytotoxic activity PVL Inhibition of cytotoxic activity α-hemolysin

Inhibition of cytotoxic activity

References Reglinski et al. (2015), Basma et al. (1998) Tarnutzer et al. (2019) Sriskandan et al. (2006), Schrage et al. (2006), NorrbyTeglund et al. (1996a) Tarnutzer et al. (2019) Tarnutzer et al. (2019) Andreoni et al. (2018)

Reglinski and Sriskandan (2019) Darenberg et al. (2004), Darville et al. (1997), Takei et al. (1993) Wood et al. (2017) Diep et al. (2016), Mairpady Shambat et al. (2015), Gauduchon et al. (2004) Diep et al. (2016), Mairpady Shambat et al. (2015), Farag et al. (2013)

SDSE Streptococcus dysgalactiae subspecies equisimilis; Sda Streptodornase 1, SLO Streptolysin O, SpyCEP S. pyogenes cell envelope protease, Luk Leukocidin, PVL Panton Valentine Leukocidin

IVIG was also shown to be protective against necrotizing pneumonia caused by MRSA strains in an in vivo rabbit model and in particular antibodies towards the pore-forming toxins α-hemolysin and PVL were implicated in the protective effect (Diep et al. 2016). Also Mairpady Shambat et al. reported that IVIG neutralized the effect of α-hemolysin and PVL thereby attenuating tissue pathology in a human organotypic model of lung (Mairpady Shambat et al. 2015). One important aspect is whether all IVIG preparations are equally efficient. Studies have investigated this and reported that variation do exist between different preparation in their efficacy to neutralize different virulence factors (Norrby-Teglund et al. 1998; Schrage et al. 2006; Wood et al. 2017). When different IVIG preparations were tested for neutralization of broad panel of streptococcal superantigens, significant differences were demonstrated between preparation but also towards different superantigens (Schrage et al. 2006). The

superantigens SSA, Spe-A, Spe-C, and Spe-K/L were more efficiently neutralized by IVIG as compared to Spe-J, Spe-H, and SmeZ-2. Notably, within the concentration range tested none of the preparations could completely block the activity of SPE-J. This has implication for the dosage of IVIG used and clinical efficacy. However, in studies where supernatants from GAS strains have been used, even from Spe-J-positive strains, the superantigenic activity has always been completely blocked by IVIG (Norrby-Teglund et al. 1998, 1996a, b; Emgård et al. 2019). Most of the studies described above have utilized IgG preparations. However, it should be mentioned that in studies comparing the IgG versus IgMand IgA-enriched IVIG preparations differences have been noted. IgM-enriched IVIG has been reported to contain higher anti-LPS antibody titers in the IgM fraction (Trautmann et al. 1998), as well as increased opsonic activity towards the Gram-negative bacteria Pseudomonas aeruginosa, E coli, and Klebsiella pneumoniae, as compared to IVIG containing

8

Treatment of Necrotizing Soft Tissue Infections: IVIG

111

Fig. 8.2 Mechanisms of action of IVIG in severe invasive group A streptococcal infections. SAg superantigen, SLO streptolysin, SdaI streptodornase I, APC antigen presenting cell

only IgG (Garbett et al. 1989). A similar difference was also reported in regard to neutralization of mitogenic and cytokine-inducing activity of one of the group A streptococcal superantigens, i.e. Spe-A (Norrby-Teglund et al. 1998). Whether these noted differences would translate into a varying clinical efficacy has not yet been explored.

8.3.3

Therapeutic Indications

IVIG is mostly used as substitution treatment or for immunomodulatory purposes. According to the U.S. Food and Drug Administration (FDA) and the European Medicines Agency, IVIG is approved for replacement therapy in primary immunodeficiencies with impaired antibody production and secondary immunodeficiencies with

112

M. B. Madsen et al.

recurrent infections or low IgG levels, as well as for immunomodulatory effect in primary immune thrombocytopenia, acute inflammatory demyelinating polyneuropathy (Guillain–Barré syndrome), Kawasaki disease, and Chronic inflammatory demyelinating polyradiculoneuropathy (Table 8.3 and Fig. 8.3) (European Medicines Agency 2020). IVIG can also be used as prophylaxis against hepatitis A and measles. In outbreaks of measles, IVIG is often recommended for healthy, non-immune persons (Arciuolo et al. 2017; Matysiak-Klose et al. 2018; Tunis et al. 2018). There is also widely accepted off-label use of IVIG in several additional conditions, supported by positive results but in small-scale trials. In short, there is some level of evidence for the use of IVIG against rheumatoid and neurological autoimmune/inflammatory diseases (Ballow 2011; Viard et al. 1998; Li et al. 2005; Ephrem et al. 2008; Perlmutter et al. 1999), viral infections (Elsterova et al. 2017; Vanderven and Kent 2020), bacterial infections (Kakoullis et al. 2018; Ohlsson and Lacy 2020), hematological conditions (Chang et al. 2008; Sultan et al. 1984), obstetric (Egerup et al. 2015), and pulmonological conditions (Wade and Chang 2015). How common off-label usage of IVIG is for these indications is currently unknown by the authors of this work.

8.3.4

Safety

Immunoglobulin use was generally considered safe until the early 1990s, where patients who received immunoglobulins were infected with hepatitis C virus (Chapel 1999). In this section we will review the safety profile of immunoglobulins for medical use, with emphasis on use for patients with NSTI.

8.3.4.1

Donor Selection and Manufacturing Process Plasma donation varies slightly from whole blood donation, as a different technique is used in order to obtain more plasma and reinfuse the remaining blood products in the patient. It is also possible to extract plasma from a “whole blood donation,”

but this technique renders less plasma. Donation practices vary between countries; in some places, donation is based on a voluntary basis, and some have systems where the donor is compensated economically (Farrugia et al. 2015). Plasma derived from the national blood services is sold to the medicine industry. Almost all plasma derived products are manufactured by the industry. Transmission of viral disease was previously a concern to patients receiving IVIG (Siegel 2006), but few cases of viral transmission have been reported through time. Cold ethanol fractioning was used as part of the manufacturing process, and this removed most viruses. No reports of HIV transmission exist, but in the nineties two series of incidents of hepatitis C virus transmission were published (Farrugia and Quinti 2014; Power et al. 1994; Bresee et al. 1996). Consequently, the FDA demanded viral removal and deactivation to be included in the preparation process. Today, most donor blood is screened for syphilis, hepatitis B, C, and E and HIV and human T-lymphotropic virus (NHS 2019). A variety of different methods are used to prevent transmission of virus. In the quest to manufacture Ig preparation, Cohn developed cold ethanol fractioning, which separated plasma proteins (Cohn et al. 1940). Addition of pepsin or incubation at low pH also help eliminate anticomplementary activity (Gordon 1987). This process itself eliminates HIV, but not HCV (Siegel 2006), and was the only elimination process for many years. The main methods for removing or inactivating virus are described in Table 8.4

8.3.4.2 Supply Immunoglobulin is a scarce blood product. Immunoglobulin use is increasing (Sundhedstyrelsen 2016; Blood 2019), and the indications for its use are expanding. Data from Australia suggest that the majority of immunoglobulin use is for acknowledged indications, and the increase is partly due to variations in administration frequency, dosage, and population, and not as a consequence of off-label use (Blood 2019). However, a report from Canada suggests

8

Treatment of Necrotizing Soft Tissue Infections: IVIG

113

Table 8.3 Indications for treatment with IVIG Disease Substitution Primary humoral immunodeficiency

Secondary humoral immunodeficiency

Hematopoietic stem cell transplantation, allogenic

Immune modulation Immune thrombocytopenia/ idiopathic thrombocytopenic purpura

Indication

Dosage

Main mechanism

Congenital agammaglobulinemia and hypogammaglobulinemia, common variable immunodeficiency, severe combined immunodeficiency, Wiskott–Aldrich syndrome Blood or bone marrow disorders especially chronic lymphocytic leukemia (CLL) or multiple myeloma (MM) with hypogammaglobulinemia and recurrent bacterial infections, congenital acquired immunodeficiency syndrome (AIDS) with recurrent infections, in rare cases: patients treated with immunosuppressant/'biological' drugs or undergoing chemotherapy, transcobalamin deficiency, gut lymphangiectasia Hypogammaglobulinemia, earlier also prophylaxis at transplant surgery and treatment of graft versus host disease

0.4–0.8 g/kg initially, then 0.2–0.8 g/kg every 3–4 weeks

Substitution of Ab (Ammann et al. 1982; Oates et al. 1991)

0.2–0.4 g/kg every 3–4 weeks

Substitution of Ab (Raanani et al. 2009)

2 g/kg during 4 days, then 0.2–0.4 g/kg every 3–4 weeks

Substitution of Ab, infection prophylaxis, decrease donor specific Ab (NIH 1990; Jamil and Mineishi 2015)

Life threatening bleeding due to ITP, prophylaxis during surgery

0.8–1 g/kg one/twice during 3 days, or 0.4 g/kg every day during 2–5 days 2 g/kg once

Fc-rec blockade, decrease auto-Ab (Ballow 2011; Imbach et al. 1981)

Kawasaki disease

First line of treatment with acetylsalicylic acid

Acute inflammatory demyelinating polyneuropathy/Guillain– Barré syndrome, Chronic inflammatory demyelinating polyneuropathy, Multifocal motor neuropathy Passive immunity Hepatitis A prophylaxis

First line of treatment

0.4 g/kg daily during 5 days for AIDP, 2 g/kg during 2–5 days, then 1 g/kg every third week for CIDP and MMN

Pre/post-exposure for patients with immunodeficiency disorders

0,0032 g/kg intramuscularly once

Passive immunity (FASS 2020; Askling and Herzog 2019)

0.4 g/kg within 72 h

Passive immunity (Arciuolo et al. 2017; Matysiak-Klose et al. 2018; Tunis et al. 2018) (continued)

Common off-label clinical practice Measles Post-exposure for non-immune persons

Decrease pro-inflammatory cytokines, neutralize bacterial toxins (Ballow 2011; Eleftheriou et al. 2014) Decrease auto-Ab (Ballow 2011; Finsterer 2005)

114

M. B. Madsen et al.

Table 8.3 (continued) Disease Myasthenia gravis

Indication Second line of treatment—severe symptoms, myasthenic crisis

Dosage 2 g/kg during 5 days, then 0.4 g/kg every month

Streptococcal toxic shock syndrome

Adjunctive treatment

0.5–2 g/kg initially, followed by lower doses

Necrotizing soft tissue infections /fasciitis

Streptococcal or staphylococcal etiology, adjunctive treatment

0.5–2 g/kg initially, followed by lower doses

Main mechanism Decrease auto-Ab (Ballow 2011; GarcíaCarrasco et al. 2007; Zinman et al. 2007; Barth et al. 2011) Neutralize bacterial toxins, opsonization of bacteria (Parks et al. 2018; Linnér et al. 2014) Neutralize bacterial toxins, opsonization of bacteria (Madsen et al. 2017; Bergsten et al. 2020)

The dosages are given intravenously unless otherwise specified. Ab indicate antibodies. Subcutaneous alternatives are available for primary immunodeficiencies and chronic inflammatory demyelinating polyneuropathy

Fig. 8.3 FDA-registered compounds with human normal polyclonal immunoglobulin and approved indications. AIDP acute inflammatory demyelinating polyneuropathy, AIDS acquired immune deficiency syndrome, CIDP chronic inflammatory demyelinating polyneuropathy, CLL chronic lymphocytic leukemia, GBS Guillain–Barré

syndrome, HSCT hematopoietic stem cell transplantation, allogenic, ITP idiopathic thrombocytopenic purpura, IVIg intravenous immunoglobulin, MM multiple myeloma, MMN multifocal motor neuropathy, PID primary humoral immunodeficiency, SCIg subcutaneous immunoglobulin, SID secondary humoral immunodeficiency

8

Treatment of Necrotizing Soft Tissue Infections: IVIG

115

Table 8.4 Methods for removing or inactivating of virus Cold ethanol fractioning (Cohn process) Octanoic/caprylic acid fractioning Filtration (Ultra-, dia- and nanofiltration) Anion exchange chromatography Pasteurization Low pH incubation

Involves increasing ethanol fraction, decreasing temperature and decreasing pH Addition of caprylic acid will make non-immunoglobulin proteins precipitate A membrane separation process, where components are removed based on their solvated size and structure Separation of molecules based on their net surface charge Heating in solution

the opposite (Foster et al. 2010). Plasma-self-sufficiency is not possible in a number of countries—even in countries with high levels of voluntary blood donation—and they are reliable on plasma bought commercially.

8.3.4.3 Batch and Preparation Variation Different manufacturers use different compositions of immunoglobulin (Siegel 2015). Furthermore, as IVIG is extracted from human plasma, the content and thus the neutralizing properties are subject to variation, dependent on the exposure of the donors (Shankar-Hari et al. 2012). 8.3.4.4

Adverse Reactions of IVIG Infusion It is difficult to get an overview of adverse reactions as use is indicated for a wide variety of diseases. Furthermore, dosage, number of infusions, infusion rate and composition of the product vary. Consequently, different studies have focused on different aspects of safety profiles. According to the summary of product characteristics, common adverse reactions include chills, headache, dizziness, fever, vomiting, allergic reactions, nausea, arthralgia, low blood pressure, and moderate lower back pain. Severe adverse reactions include hypersensitivity, hemolytic anemia, aseptic meningitis syndrome, thromboembolism, acute renal failure, transfusion-related acute lung injury, and transmissible agents (FDA 2019; Pierce and Jain 2003).

Adverse effects can be divided into immediate, delayed, and late. As for all medications, infusion of intravenous immunoglobulin may result in an inflammatory response resulting in headache, flushing, tight chestedness, dyspnea, and circulatory collapse (anaphylactic reaction). Headache is the most frequent adverse reaction, and although symptoms such as headache, malaise, and myalgia occur in about 5% of patients, these often abate with a reduction in infusion rate. Rash is seen in approximately 6% of patients. As immunoglobulin preparations contain varying amounts of IgA, patients with total IgA deficiency may develop anti-IgA antibodies, ultimately resulting in anaphylactic reaction (Nydegger and Sturzenegger 1999). Acute renal failure has been described in patients with known impairment of renal function (Cantú et al. 1995). The proposed mechanism might be the presence of sucrose, which is used as a stabilizing agent, but this is merely speculative (Nydegger and Sturzenegger 1999). Serum viscosity may increase as a result of high IgG concentrations, immune complex formation, and immunoglobulin aggregates, and may cause symptoms as headache and visual blurring, and could result in impaired circulation in patients with atherosclerosis or other conditions that result in poor vascular tone. Aseptic meningitis is a condition mimicking bacterial meningitis with affection of the cerebrovascular fluid, but the mechanism is not clear. In a study on patients with neuromuscular diseases receiving IVIG, 6/54 (11%) developed severe symptoms with acute, severe headache, stiffneckedness, lethargy, fever, and photophobia (Sekul et al. 1994).

116

The definitions of adverse events, and the collection and reporting of safety data are not standardized. Overall, the trials assessing effects of IVIG in general did not have the statistical power to detect differences in adverse events and serious adverse reactions, and the clinical trials performed to support marketing applications for IVIG had insufficient power to detect adverse events (Pierce and Jain 2003). Regarding IVIG use in patients with NSTI, only two trials have systematically collected data on adverse reactions of IVIG (Darenberg et al. 2003; Madsen et al. 2017); for a review of these trials, please see the section “Rationale for the use of IVIG for NSTI” below. In trials of IVIG for sepsis and septic shock, the reporting of adverse events is scarce (Pildal and Gotzsche 2004). In one of the larger trials (n ¼ 653), a total of 19 adverse events were reported in 17 patients; 13 events in 11 patients in the IVIG group, and 6 events in 6 patients in the placebo group (Werdan et al. 2007). The adverse events in the IVIG group included exanthema, thrombosis, cardiac failure and arrest, anaphylactic reaction, intracerebral hemorrhage, overhydration, and acute kidney injury. In an observational study with the specific aim of studying the safety profile of IVIG use in a general ICU population, acute kidney injury (AKI) (defined as a rise in creatinine 20% within 28 days of IVIG admission) occurred in 117 of 145 (81%) patients, and 30 (21%) required renal replacement therapy (Foster et al. 2010). Although most patients had other factors that may have contributed to renal insufficiency, 10% had no obvious concurrent renal insult other than IVIG. The number of patients with thromboembolism was 4 (3%). The risk of AKI may be associated with pre-existing conditions such as sepsis and hypovolemia, but also with the specific product used (Siegel 2006). Lastly, it is worth considering the fact that critically ill patients receive many different therapies. In generally, little is known about drug interactions, and the addition of IVIG is likewise poorly described.

M. B. Madsen et al.

8.4 8.4.1

Rationale for the Use of IVIG for NSTI Laboratory Data

Patients with severe GAS infections including NSTI have been reported to have lower levels of protective antibodies towards several important GAS virulence factors including Spe-B and superantigens, as compared to uncomplicated cases such as GAS tonsillitis or erysipelas cases (Babbar et al. 2018; Eriksson et al. 1999; NorrbyTeglund et al. 1994). IVIG has been shown to contain protective antibodies targeting GAS virulence factors, which are efficiently transferred to the patients upon IVIG administration as evident by an increased neutralizing activity in plasma collected post-IVIG therapy (Tarnutzer et al. 2019; Norrby-Teglund et al. 1996a, b; Basma et al. 1998). Important points for discussion are what dosages of IVIG that are efficacious, and concentration-levels achieved at the tissue site of infection. There are, as of yet, no clinical IVIG dosage studies performed on NSTI or toxic shock syndrome patients. However, insight was gained through analyses of plasma collected from patients included in the INSTINCT and the INFECT studies (Bergsten et al. 2020). The plasma analyses revealed that a dose of 25 g of IVIG was sufficient to yield neutralizing activity towards GAS superantigens. However, the level of inhibition varied between strains, and towards some GAS strains the inhibition was only borderline protective. Taken together with the report by Bruun et al. (Bruun et al. 2020) on GAS NSTI patients in the observational INFECT study, which showed an association between patients who had not received IVIG and 90-day mortality, Bergsten et al. (Bergsten et al. 2020) proposed a dosage regimen of: 0.5 g/kg bodyweight (or a minimum of 25 g) on day 1, followed by fixed doses of 25 g daily for 1–2 additional days. The higher dose on day 1 of 25 g offered only borderline protective towards some strains in combination with often high levels of toxins and bacterial

8

Treatment of Necrotizing Soft Tissue Infections: IVIG

117

load at this time point and at the tissue site. However, as acknowledged by the authors this treatment regimen, i.e. dosages and treatment days, should be evaluated in future trials. Bioavailability of IVIG in the deep necrotic tissue has been questioned, and the possibility of local applications of antibodies, as used for postexposure prophylaxis in rabies (CDC 2020), has been raised. In an in vivo model of GAS NSTI, intraperitoneal and subcutaneous administration of IVIG was compared (Tarnutzer et al. 2019). This study showed that the subcutaneous administration yielded 1.6 times higher IgG concentration in the skin, but the serum concentration was half as high compared to intraperitoneal administration. Subcutaneous injection also resulted in reduced clinical severity (i.e. smaller lesion size) as well as reduced SLO activity.

albumin. After inclusion of 21 patients, 10 receiving IVIG (of whom 6 had NSTI) and 11 receiving placebo (of whom 7 had NSTI), the trial was prematurely terminated due to slow inclusion, after enrolling only 21 of the intended 120 patients. The primary endpoint was 28-day mortality and there was no difference between the two groups; 1 of 10 patients in the IVIG group had died, and 4 of 11 patients in the placebo group ( p ¼ 0.30). After administration of IVIG, patient serum contained sufficient neutralizing antibodies to suppress proliferation of T-lymphocytes to below pretreatment levels, whereas in the placebo group, no change in neutralizing activity was seen. The mean time to resolution of shock was 100 h for the IVIG group and 122 h for the placebo group. The mean time to no further progression of necrotizing fasciitis or cellulitis was 69 h for the IVIG group, compared with 36 h for the placebo group. Besides death, a total of 12 adverse events were observed, none of which were found related to the study drug. None of these secondary endpoints were statistically significant. The INSTINCT trial was a randomized, clinical trial including 100 patients with NSTI (Madsen et al. 2017). Patients with NSTI, irrespective of etiology, were included after admission to the intensive care. Patients were randomized to receive IVIG 25 g/day for three consecutive days or placebo. The primary outcome was physical quality of life assessed by the physical component summary (PCS) score of the Short Form-36 questionnaire (Ware et al. 1992). No difference in the primary outcome (mean PCS score 29 versus 28, mean difference 1 (95% CI 7 to 10)) or in secondary outcomes, including mortality, serious adverse reactions (SARs), organ failure, organ support, bleeding, resolution of shock and days in hospital, was observed. SARs included acute kidney injury, allergic reactions, aseptic meningitis syndrome, hemolytic anemia, thrombi, and transmittable agents. In the preplanned subgroup analysis on patients with presumed GAS infection, no difference was seen. Importantly, the distribution of patients with GAS or S. aureus was uneven between the two intervention groups. Some

8.4.2

Clinical Studies

The proportion of patients receiving IVIG as part of the treatment for NSTI, as well as approaches to IVIG use varies. Typically, IVIG is considered in NSTI patients where the bacterial etiology is highly suspicious of GAS (Anaya and Dellinger 2007). In this setting, 25% of all intensive care units participating in a survey reported to always use IVIG (de Prost et al. 2015), and in an Australian observational study, 27% of patients received IVIG (Carapetis et al. 2014). Overall use independent of bacterial etiology has been described, with 58% of patients receiving IVIG (Madsen et al. 2019).

8.4.2.1 Randomized Clinical Trials Despite the promising effects from in vitro studies, only two trials have compared the effects of IVIG with placebo in patients with NSTI. In 1999, Darenberg and colleagues aimed to evaluate the efficacy and safety of IVIG as adjunctive therapy in patients with streptococcal toxic shock syndrome (STSS) in a randomized, double blind, placebo-controlled trial (Darenberg et al. 2003). Patients were randomly assigned to either IVIG for three consecutive days (1 g/kg day one, 0.5 g/ kg day 2 and 3) or an equal volume of 1%

118

limitations of the trial deserve mentioning; the trial had reduced power, as patients who died where given a PCS score of 0. Furthermore, 20 (40%) patients in the placebo group had received one dose of IVIG prior to enrolment. This might have diluted a potential effect. Lastly, the trial was not powered to detect differences in secondary outcomes, making it difficult to conclude whether IVIG use is safe in NSTI patients.

8.4.2.2

Systematic Reviews and Meta-Analyses A systematic review from 2018 included only one RCT on IVIG for NSTI, and thus could not add additional data (Hua et al. 2018). In a metaanalysis from 2018 (Parks et al. 2018), evaluating the effect of IVIG in patients with STSS, data from four observational studies (Carapetis et al. 2014; Kaul et al. 1999; Adalat et al. 2014; Linnér et al. 2014) and one RCT (Darenberg et al. 2003) were included. Studies were included if patients were identified prospectively, met the consensus criteria of STSS and whose antibiotic treatment included clindamycin. A total of 165 patients were included in the analysis; 70 received IVIG and 95 did not. IVIG use was associated with a reduction in overall mortality rate from 33.7% to 15.7% (RR, 0.46; 95% confidence interval, 0.26–0.83; p ¼ 0.01). The analysis has limitations, the main being that data from 147 of the 165 patients included were from observational studies, and thus no causative conclusions can be drawn. The studies included used different comparators, and for some of the studies, no information on dosing or timing was given. 8.4.2.3 Observational Studies In a retrospective observational study by Kadri and colleagues, 4127 patients with necrotizing fasciitis and shock from 121 centers were identified based on diagnosis code (Kadri et al. 2017). Only 164 of these patients had received IVIG. These were propensity-matched and riskadjusted with those who had not received IVIG. The groups were similar regarding age, diagnoses, comorbidities, day of IVIG dose and destination of discharge of survivors. The primary outcome was in-hospital mortality; the crude

M. B. Madsen et al.

mortality rate was 27.3% in the IVIG group and 23.6% in the non-IVIG group (adjusted OR 1.00, 95% CI: 0.55–1.83). In a publication focusing on the subgroup of patients with GAS NSTI (Bruun et al. 2020) from the Scandinavian INFECT observational study (Madsen et al. 2019), the administration of IVIG was associated with decreased risk of mortality by Lasso Regression analysis, but not by ordinary logistic regression analysis. Due to the observational design, these data may support the hypothesis that IVIG can have a beneficial effect in this subgroup of patients, but no causal conclusions can be drawn. In an Australian surveillance programme of invasive GAS infections from 2002 to 2004, the mortality rate in patients treated with IVIG and clindamycin was lower (1/14 (7%)) than in those who were treated with clindamycin alone (7/39 (18%)), but this was not statistically significant (Carapetis et al. 2014). In a similar Swedish surveillance programme of patients with STSS, also from 2002 to 2004, 67 patients were identified of whom 23 received IVIG and 44 did not (Linnér et al. 2014). At day 28, 3/23 (13%) in the IVIG group, and 22/44 (50%) in the non-IVIG group, had died (OR for death 0.18 (95% CI: 0.04–0.83), adjusted for SAPS II and use of clindamycin). The proportion of patients with necrotizing fasciitis was larger in the IVIG group; in a subgroup analysis of patients with necrotizing fasciitis (n ¼ 19), the OR for death was 0.17 (95% CI: 0.01–2.5). A third surveillance programme for invasive GAS infections in Ontario identified 62 patients from 1992 to 2002; 35 (56%) patients received IVIG (Mehta et al. 2006). No difference in ICU/hospital-mortality by univariate analysis was seen. Lastly, a prospective observational study enrolled patients with STSS in Canada in 1992–1994 (Kaul et al. 1999). A total of 53 patients with STSS were included; OR for death for the 21 patients receiving IVIG was 0.18 (95% CI: 0.03–1.02).

8.4.2.4 IVIG for Sepsis Several trials on IVIG for patients with sepsis/ septic shock have been conducted, and systematic reviews have shown potential beneficial effects

8

Treatment of Necrotizing Soft Tissue Infections: IVIG

119

on mortality (Soares et al. 2012; Turgeon et al. 2007), although results vary, depending on the inclusion criteria (Alejandria et al. 2013). Besides general concerns regarding IVIG use (ShankarHari et al. 2018) (see section below), there are several other concerns that need attention, in particular the quality of the included trials. Inclusion criteria varied, as several definitions of sepsis were used. Sepsis is a heterogenous disease, and the patients response to infection may differ according to the site infected (Shankar-Hari et al. 2016; Burnham et al. 2017). Most of the trials were of a relatively small sample size, ranging from 24 to 682 patients. Whereas IVIG in NSTI patients is believed to function through a neutralizing effect of exotoxins and other bacterial virulence factors, the mechanisms in sepsis is more poorly understood, and this patient group is even more heterogenous than NSTI patients. Lastly, the trials were conducted more than 10 years ago—some of them more than 20 years ago. In conclusion, there is no evidence to support the use of immunoglobulin for treating patients with sepsis or septic shock (Rhodes et al. 2017).

treatment. IVIG has been used as adjuvant therapy for NSTI (de Prost et al. 2015), but not necessarily accounting for bacterial etiology, and the neutralizing abilities of IVIG have primarily been shown in patients affected by NSTI caused by GAS or S. aureus. Questions on dosage, timing of administration, and duration of treatment are still debated. Traditionally, dosing has ranged from 0.5 g/kg to 2 g/ kg (Table 8.5). As discussed previously, one dose of 25 g seems to be sufficient to suppress toxin activity in NSTI patients infected by GAS, and after two doses, no toxin activity is seen (Bergsten et al. 2020). However, this has not been confirmed in a clinical setting, and optimally, a randomized clinical trial on GAS NSTI should be performed to assess whether this treatment has a beneficial effect and no unacceptable serious adverse reactions. The optimal composition of immunoglobulins is not known. IgG has many potential mechanisms of action (Di Rosa et al. 2014), but whether an additional effect of adding IgM is achieved is not known. Furthermore, the properties of the preparations are dependent on the donors and thus immunomodulating properties may vary. Only few prospective studies have evaluated the safety profile of IVIG in patients with NSTI. Of concern is the severity of the disease and polypharmacy in these patients, with the possibility of drug interactions (Kane-Gill et al. 2012).

8.4.3

IVIG in NSTI, Unresolved Issues

The use of IVIG for patients with NSTI is a complex intervention, with multiple aspects that are unknown, or have not been addressed adequately. NSTI is not a well-defined disease, and NSTI patients represent a heterogenic population. The term encompasses many different variations of severe soft tissue infections (Madsen et al. 2019). The varying nomenclature and classification schemes also reflect this diversity; Necrotizing Fasciitis, describing an infection along the fascia in the subcutis; Fournier’s gangrene, describing a progressing soft tissue infection originating from the anogentital area.; Type 1 and 2 necrotizing fasciitis, depending on a polyor monomicrobial etiology. It is likely, that the pathophysiology varies depending on the affected body-part, mechanism of bacterial entry and bacterial etiology (Madsen et al. 2019). Especially the latter is of importance regarding IVIG

8.4.4

Future Perspectives

From in vitro and in vivo studies it has been demonstrated that IVIG has the potential to alleviate the course of NSTIs where GAS or S. aureus is the causative agent. However, we do not have data from randomized clinical trials to show that this treatment is beneficial to patients and, just as important, is safe to use in patients that are critically ill. Furthermore, plasma is a scarce product. A large, multi-center RCT evaluating the effect of IVIG on mortality in patients with GAS NSTI is therefore warranted.

120

M. B. Madsen et al.

Table 8.5 Doses of IVIG used in patients with severe invasive infections, including necrotizing soft tissue infections and toxic shock syndrome

Study Kadri et al. (2017)

Patient cohort Debrided NF with shock Madsen et al. NSTI admitted to the (2017) ICU Adalat et al. (2014) Streptococcal and staphylococcal TSS Carapetis et al. Severe invasive GAS (2014) infections Linnér et al. (2014) STSS Mehta et al. (2006) Invasive GAS infections Norrby-Teglund Severe GAS soft tissue et al. (2005) infections Arnholm et al. Invasive GAS (2004) infections Darenberg et al. STSS (2003) Haywood et al. Streptococcal NF (1999) Kaul et al. (1999) STSS

Number of included patients (number receiving IVIG) 4127 (164)

Repeat dosages n/a

Number of treatment days 1b

100 (50)

Initial IVIG dosage 1 g/kg (0.66–1)a 25 g

25 g

2–3

49 (10)

n/a

n/a

n/a

84 (14)

n/a

n/a

n/a

67 (23) 62(45)

0.5 g/kg 1.61  0.72 g/kgc 2 g/kg

0.5 g/kg 0.75  0.52 g/kgc n/a

1–6 n/a 1–3

50 g (n ¼ 7) 80 g (n ¼ 1) 1 g/kg

n/a

2-3

7 (7) 41 (11) 21 (10) 20 (16) 53 (21)

0.5 mg/kg day 2 and 3 0.001–0.002 g/ n/a kg 0.2–2 g/kgd n/a

3 n/a Up to 5

Case-reports not included. NSTI necrotizing soft tissue infection, NF necrotizing fasciitis, ICU intensive care unit, TSS toxic shock syndrome, GAS group A streptococcal infections, STSS streptococcal toxic shock syndrome a Median dose (IQR) b Median number of treatment days c Mean dosage d The median cumulative dose was 2 g/kg (range 0.2–3.6 g/kg) Acknowledgement Financial support: The work was supported by the European Union Seventh Framework Programme: (FP7/2007–2013) under the grant agreement 305340 (INFECT project); the Swedish Governmental Agency for Innovation Systems (VINNOVA) under the frame of NordForsk (Project no. 90456, PerAID), and the Swedish Research Council under the frame of ERA PerMed (Project 2018-151, PerMIT).

References Adalat S, Dawson T, Hackett SJ et al (2014) Toxic shock syndrome surveillance in UK children. Arch Dis Child 99:1078–1082 Alejandria MM, Lansang MAD, Dans LF, Mantaring Iii JB (2013) Intravenous immunoglobulin for treating sepsis, severe sepsis and septic shock. Cochrane Database Syst Rev 9:CD001090 Ammann AJ, Ashman RF, Buckley RH et al (1982) Use of intravenous gamma-globulin in antibody immunodeficiency: results of a multicenter controlled trial. Clin Immunol Immunopathol 22:60–67

Anaya DA, Dellinger EP (2007) Necrotizing soft-tissue infection: diagnosis and management. Clin Infect Dis 44:705–710 Andreoni F, Ugolini F, Keller N et al (2018) Immunoglobulin attenuates streptokinase-mediated virulence in Streptococcus dysgalactiae subspecies equisimilis necrotizing fasciitis. J Infect Dis 217:270–279 Arciuolo RJ, Jablonski RR, Zucker JR, Rosen JB (2017) Effectiveness of measles vaccination and immune globulin post-exposure prophylaxis in an outbreak setting-New York City, 2013. Clin Infect Dis 65:1843–1847 Arnholm B, Lundqvist A, Strömberg A, Stromberg A (2004) High-dose immunoglobulin--life-saving in invasive group A streptococcal infection. Report of eleven cases with only one fatality. Lakartidningen 101:2642–2644 Askling HH, Herzog C (2019) Hepatitis A vaccination in immunocompromised patients – the need for individualized vaccination strategies and correct methodology. Travel Med Infect Dis 32:101526 Babbar A, Bruun T, Hyldegaard O et al (2018) Pivotal role of preexisting pathogen-specific antibodies in the development of necrotizing soft-tissue infections. J Infect Dis 218:44–52

8

Treatment of Necrotizing Soft Tissue Infections: IVIG

121

Ballow M (2011) The IgG molecule as a biological immune response modifier: mechanisms of action of intravenous immune serum globulin in autoimmune and inflammatory disorders. J Allergy Clin Immunol 127:315–323 Barry W, Hudgins L, Donta ST, Pesanti EL (1992) Intravenous immunoglobulin therapy for toxic shock syndrome. JAMA 267:3315–3316 Barth D, Nabavi Nouri M, Ng E et al (2011) Comparison of IVIg and PLEX in patients with myasthenia gravis. Neurology 76:2017–2023 Basma H, Norrby-Teglund A, McGeer A et al (1998) Opsonic antibodies to the surface M protein of group A streptococci in pooled normal immunoglobulins (IVIG): potential impact on the clinical efficacy of IVIG therapy for severe invasive group A streptococcal infections. Infect Immun 66:2279–2283 Bergsten H, Madsen MB, Bergey F et al (2020) Correlation between immunoglobulin dose administered and plasma neutralization of streptococcal superantigens in patients with necrotizing soft tissue infections. Clin Infect Dis. https://doi.org/10.1093/cid/ciaa022 Bermejo-Martín JF, Rodriguez-Fernandez A, HerránMonge R et al (2014) Immunoglobulins IgG1, IgM and IgA: a synergistic team influencing survival in sepsis. J Intern Med 276:404–412 Black CA (1997) A brief history of the discovery of the immunoglobulins and the origin of the modern immunoglobulin nomenclature. Immunol Cell Biol 75:65–68 Blood (2019) Criteria for the clinical use of intravenous immunoglobulin in Australia, 2nd ed. https://www. blood.gov.au/pubs/ivig/intravenous-immunoglobulinsupply-and-demand.html. Accessed 23 Jan 2019 Brandtzaeg P, Johansen F-E (2005) Mucosal B cells: phenotypic characteristics, transcriptional regulation, and homing properties. Immunol Rev 206:32–63 Bresee JS, Mast EE, Coleman PJ et al (1996) Hepatitis C virus infection associated with administration of intravenous immune globulin. A cohort study. JAMA 276:1563–1567 Bruhns P, Samuelsson A, Pollard JW, Ravetch JV (2003) Colony-stimulating factor-1-dependent macrophages are responsible for IVIG protection in antibodyinduced autoimmune disease. Immunity 18:573–581 Bruton OC (1952) Agammaglobulinemia. Pediatrics 9:722–728 Bruun T, Rath E, Bruun Madsen M, et al (2020) Risk factors and predictors of mortality in streptococcal necrotizing soft-tissue Infections: a multicenter prospective study. Clin Infect Dis. https://doi.org/10. 1093/cid/ciaa027 Burnham KL, Davenport EE, Radhakrishnan J et al (2017) Shared and distinct aspects of the sepsis transcriptomic response to fecal peritonitis and pneumonia. Am J Respir Crit Care Med 196:328–339 Cantú TG, Hoehn-Saric EW, Burgess KM et al (1995) Acute renal failure associated with immunoglobulin therapy. Am J Kidney Dis 25:228–234

Carapetis JR, Steer AC, Mulholland EK, Weber M (2005) The global burden of group A streptococcal diseases. Lancet Infect Dis 5:685–694 Carapetis JR, Jacoby P, Carville K et al (2014) Effectiveness of clindamycin and intravenous immunoglobulin, and risk of disease in contacts, in invasive group a streptococcal infections. Clin Infect Dis 59:358–365 CDC (2020) Rabies postexposure prophylaxis (PEP)|Medical Care|Rabies|CDC. https://www.cdc.gov/rabies/ medical_care/index.html. Accessed 5 Feb 2020 Chang J, Shi PA, Chiang EY, Frenette PS (2008) Intravenous immunoglobulins reverse acute vaso-occlusive crises in sickle cell mice through rapid inhibition of neutrophil adhesion. Blood 111:915–923 Chapel HM (1999) Safety and availability of immunoglobulin replacement therapy in relation to potentially transmissable agents. IUIS Committee on Primary Immunodeficiency Disease. Clin Exp Immunol 118 (Suppl):29–34 Cohn EJ, Luetscher JA, Oncley JL et al (1940) Preparation and properties of serum and plasma proteins. III. Size and charge of proteins separating upon equilibration across membranes with ethanol—water mixtures of controlled ph, ionic strength and temperature. J Am Chem Soc 62:3396–3400 Darenberg J, Ihendyane N, Sjölin J et al (2003) Intravenous immunoglobulin G therapy in streptococcal toxic shock syndrome: a European randomized, double-blind, placebo-controlled trial. Clin Infect Dis 37:333–340 Darenberg J, Soderquist B, Normark BH, Norrby-Teglund A (2004) Differences in potency of intravenous polyspecific immunoglobulin G against streptococcal and staphylococcal superantigens: implications for therapy of toxic shock syndrome. Clin Infect Dis 38:836–842 Darville T, Milligan LB, Laffoon KK (1997) Intravenous immunoglobulin inhibits staphylococcal toxin-induced human mononuclear phagocyte tumor necrosis factor alpha production. Infect Immun 65:366–372 de Prost N, Sbidian E, Chosidow O et al (2015) Management of necrotizing soft tissue infections in the intensive care unit: results of an international survey. Intensive Care Med 41:1506–1508 Dhainaut F, Guillaumat P-O, Dib H et al (2013) In vitro and in vivo properties differ among liquid intravenous immunoglobulin preparations. Vox Sang 104:115–126 Di Rosa R, Pietrosanti M, Luzi G, Salemi S (2014) Polyclonal intravenous immunoglobulin: an important additional strategy in sepsis? https://doi.org/10.1016/j. ejim.2014.05.002 Diep BA, Le VTM, Badiou C et al (2016) IVIG-mediated protection against necrotizing pneumonia caused by MRSA. Sci Transl Med 8:357 Egerup P, Lindschou J, Gluud C, Christiansen OB (2015) The effects of intravenous immunoglobulins in women with recurrent miscarriages: A systematic review of randomised trials with meta-analyses and trial sequential analyses including individual patient data. PLoS One 10:e0141588

122 Eleftheriou D, Levin M, Shingadia D et al (2014) Management of Kawasaki disease. Arch Dis Child 99:74–83 Elsterova J, Palus M, Sirmarova J et al (2017) Tick-borne encephalitis virus neutralization by high dose intravenous immunoglobulin. Ticks Tick Borne Dis 8:253–258 Emgård J, Bergsten H, McCormick JK et al (2019) MAIT cells are major contributors to the cytokine response in group A streptococcal toxic shock syndrome. Proc Natl Acad Sci U S A 116:25923–25931 Ephrem A, Chamat S, Miquel C et al (2008) Expansion of CD4+CD25+ regulatory T cells by intravenous immunoglobulin: a critical factor in controlling experimental autoimmune encephalomyelitis. Blood 111:715–722 Eriksson BK, Andersson J, Holm SE, Norgren M (1999) Invasive group A streptococcal infections: T1M1 isolates expressing pyrogenic exotoxins A and B in combination with selective lack of toxin-neutralizing antibodies are associated with increased risk of streptococcal toxic shock syndrome. J Infect Dis 180:410–418 European Medicines Agency (2020) Clinical investigation of human normal immunoglobulin for intravenous administration (IVIg). https://www.ema.europa.eu/en/ clinical-investigation-human-normal-immunoglobu lin-intravenous-administration-ivig#current-effectiveversion-section. Accessed 19 Feb 2020 Farag N, Mahran L, Abou-Aisha K, El-Azizi M (2013) Assessment of the efficacy of polyclonal intravenous immunoglobulin G (IVIG) against the infectivity of clinical isolates of methicillin-resistant Staphylococcus aureus (MRSA) in vitro and in vivo. Eur J Clin Microbiol Infect Dis 32:1149–1160 Farrugia A, Quinti I (2014) Manufacture of immunoglobulin products for patients with primary antibody deficiencies - the effect of processing conditions on product safety and efficacy. Front Immunol 5:665 Farrugia A, Penrod J, Bult JM (2015) The ethics of paid plasma donation: a plea for patient centeredness. HEC Forum 27:417–429 FASS (2020) FASS Allmänhet - Startsida. https://www. fass.se/LIF/startpage. Accessed 19 Feb 2020 FDA (2019) Fractionated plasma products - immune globulins. https://www.fda.gov/ BiologicsBloodVaccines/BloodBloodProducts/ ApprovedProducts/LicensedProductsBLAs/ FractionatedPlasmaProducts/ucm127589.htm. Accessed 24 Jan 2019 Ferrante A, Beard LJ, Feldman RG (1990) IgG subclass distribution of antibodies to bacteria and viral antigens. Pediatr Infect Dis J 9:S16–S24 Finsterer J (2005) Treatment of immune-mediated, dysimmune neuropathies. Acta Neurol Scand 112:115–125 Foster R, Suri A, Filate W et al (2010) Use of intravenous immune globulin in the ICU: a retrospective review of prescribing practices and patient outcomes. Transfus Med 20:403–408

M. B. Madsen et al. Garbett ND, Munro CS, Cole PJ (1989) Opsonic activity of a new intravenous immunoglobulin preparation: Pentaglobin compared with sandoglobulin. Clin Exp Immunol 76:8–12 García-Carrasco M, Escárcega RO, Fuentes-Alexandro S et al (2007) Therapeutic options in autoimmune myasthenia gravis. Autoimmun Rev 6:373–378 Gauduchon V, Cozon G, Vandenesch F et al (2004) Neutralization of Staphylococcus aureus panton valentine leukocidin by intravenous immunoglobulin in vitro. J Infect Dis 189:346–353 Gelfand EW (2012) Intravenous immune globulin in autoimmune and inflammatory diseases. N Engl J Med 367:2015–2025 Gordon DS (1987) Intravenous immunoglobulin: historical perspective. Am J Med 83:1–3 Haywood CT, McGeer A, Low DE et al (1999) Clinical experience with 20 cases of group A streptococcus necrotizing fasciitis and myonecrosis: 1995 to 1997. Plast Reconstr Surg 103:1567–1573 Hua C, Bosc R, Sbidian E et al (2018) Interventions for necrotizing soft tissue infections in adults. Cochrane Database Syst Rev 5:CD011680 Imbach P, Barandun S, d’Apuzzo V et al (1981) High-dose intravenous gammaglobulin for idiopathic thrombocytopenic purpura in childhood. Lancet 1:1228–1231 Jamil MO, Mineishi S (2015) State-of-the-art acute and chronic GVHD treatment. Int J Hematol 101:452–466 Johansson L, Norrby-Teglund A (2012) Immunopathogenesis of streptococcal deep tissue infections. In: Current topics in microbiology and immunology. pp 173–188 Kadri SS, Swihart BJ, Bonne SL et al (2017) Impact of intravenous immunoglobulin on survival in necrotizing fasciitis with vasopressor-dependent shock: a propensity score-matched analysis from 130 US hospitals. Clin Infect Dis 64:877–885 Kakoullis L, Pantzaris N-D, Platanaki C et al (2018) The use of IgM-enriched immunoglobulin in adult patients with sepsis. J Crit Care 47:30–35 Kane-Gill SL, Kirisci L, Verrico MM, Rothschild JM (2012) Analysis of risk factors for adverse drug events in critically ill patients. Crit Care Med 40:823–828 Kaneko Y, Nimmerjahn F, Madaio MP, Ravetch JV (2006) Pathology and protection in nephrotoxic nephritis is determined by selective engagement of specific Fc receptors. J Exp Med 203:789–797 Kaul R, McGeer A, Norrby-Teglund A et al (1999) Intravenous immunoglobulin therapy for streptococcal toxic shock syndrome--a comparative observational study. Clin Infect Dis 28:800–807 Li N, Zhao M, Hilario-Vargas J et al (2005) Complete FcRn dependence for intravenous Ig therapy in autoimmune skin blistering diseases. J Clin Invest 115:3440–3450 Linnér A, Darenberg J, Sjölin J et al (2014) Clinical efficacy of polyspecific intravenous immunoglobulin therapy in patients with streptococcal toxic shock

8

Treatment of Necrotizing Soft Tissue Infections: IVIG

123

syndrome: a comparative observational study. Clin Infect Dis 59:851–857 Madsen MB, Hjortrup PB, Hansen MB et al (2017) Immunoglobulin G for patients with necrotising soft tissue infection (INSTINCT): a randomised, blinded, placebo-controlled trial. Intensive Care Med 43:1585–1593 Madsen MB, Skrede S, Perner A et al (2019) Patient’s characteristics and outcomes in necrotising soft-tissue infections: results from a Scandinavian, multicentre, prospective cohort study. Intensive Care Med 45:1241–1251 Mairpady Shambat S, Chen P, Nguyen Hoang AT et al (2015) Modelling staphylococcal pneumonia in a human 3D lung tissue model system delineates toxinmediated pathology. Dis Model Mech 8:1413–1425 Martin-Loeches I, Muriel-Bombín A, Ferrer R et al (2017) The protective association of endogenous immunoglobulins against sepsis mortality is restricted to patients with moderate organ failure. Ann Intensive Care 7:44 Matysiak-Klose D, Santibanez S, Schwerdtfeger C et al (2018) Post-exposure prophylaxis for measles with immunoglobulins revised recommendations of the standing committee on vaccination in Germany. Vaccine 36:7916–7922 Medicines Agency E (2018) Committee for Medicinal Products for Human Use (CHMP) Guideline on the clinical investigation of human normal immunoglobulin for intravenous administration (IVIg) Mehta S, McGeer A, Low DE et al (2006) Morbidity and mortality of patients with invasive group A streptococcal infections admitted to the ICU. Chest 130:1679–1686 NHS (2019) Tests we carry out - NHS Blood Donation. https://www.blood.co.uk/the-donation-process/fur ther-information/tests-we-carry-out/. Accessed 11 Nov 2019 NIH (1990) NIH consensus conference. Intravenous immunoglobulin. Prevention and treatment of disease. JAMA 264:3189–3193 Norrby-Teglund A, Pauksens K, Holm SE, Norgren M (1994) Relation between low capacity of human sera to inhibit streptococcal mitogens and serious manifestation of disease. J Infect Dis 170:585–591 Norrby-Teglund A, Kaul R, Low DE et al (1996a) Evidence for the presence of streptococcal-superantigenneutralizing antibodies in normal polyspecific immunoglobulin G. Infect Immun 64:5395–5398 Norrby-Teglund A, Kaul R, Low DE et al (1996b) Plasma from patients with severe invasive group A streptococcal infections treated with normal polyspecific IgG inhibits streptococcal superantigen-induced T cell proliferation and cytokine production. J Immunol 156:3057–3064 Norrby-Teglund A, Basma H, Andersson J et al (1998) Varying titers of neutralizing antibodies to streptococcal superantigens in different preparations of normal

polyspecific immunoglobulin G: implications for therapeutic efficacy. Clin Infect Dis 26:631–638 Norrby-Teglund A, Chatellier S, Low DE et al (2000) Host variation in cytokine responses to superantigens determine the severity of invasive group A streptococcal infection. Eur J Immunol 30:3247–3255 Norrby-Teglund A, Ihendyane N, Darenberg J (2003) Intravenous immunoglobulin adjunctive therapy in sepsis, with special emphasis on severe invasive group A streptococcal infections. Scand J Infect Dis 35:683–689 Norrby-Teglund A, Muller MP, Mcgeer A et al (2005) Successful management of severe group A streptococcal soft tissue infections using an aggressive medical regimen including intravenous polyspecific immunoglobulin together with a conservative surgical approach. Scand J Infect Dis 37:166–172 Nydegger UE, Sturzenegger M (1999) Adverse effects of intravenous immunoglobulin therapy. Drug Saf 21:171–185 Oates JA, Wood AJ, Buckley RH, Schiff RI (1991) The use of intravenous immune globulin in immunodeficiency diseases. N Engl J Med 325:110–117 Ohlsson A, Lacy JB (2020) Intravenous immunoglobulin for preventing infection in preterm and/or low birth weight infants. Cochrane Database Syst Rev 1: CD000361 Parks T, Wilson C, Curtis N et al (2018) Polyspecific intravenous immunoglobulin in clindamycin-treated patients with streptococcal toxic shock syndrome: a systematic review and meta-analysis. Clin Infect Dis 67:1434–1436 Perlmutter SJ, Leitman SF, Garvey MA et al (1999) Therapeutic plasma exchange and intravenous immunoglobulin for obsessive-compulsive disorder and tic disorders in childhood. Lancet 354:1153–1158 Pierce LR, Jain N (2003) Risks associated with the use of intravenous immunoglobulin. Transfus Med Rev 17:241–251 Pildal J, Gotzsche PC (2004) Polyclonal immunoglobulin for treatment of bacterial sepsis: a systematic review. Clin Infect Dis 39:38–46 Power JP, Lawlor E, Davidson F et al (1994) Hepatitis C viraemia in recipients of Irish intravenous anti-D immunoglobulin. Lancet 344:1166–1167 Raanani P, Gafter-Gvili A, Paul M et al (2009) Immunoglobulin prophylaxis in chronic lymphocytic leukemia and multiple myeloma: systematic review and metaanalysis. Leuk Lymphoma 50:764–772 Radosevich M, Burnouf T (2010) Intravenous immunoglobulin G: trends in production methods, quality control and quality assurance. Vox Sang 98:12–28 Reglinski M, Sriskandan S (2019) Treatment potential of pathogen-reactive antibodies sequentially purified from pooled human immunoglobulin. BMC Res Notes 12:228 Reglinski M, Gierula M, Lynskey NN et al (2015) Identification of the Streptococcus pyogenes surface antigens

124 recognised by pooled human immunoglobulin. Sci Rep 5:15825 Rhodes A, Evans LE, Alhazzani W et al (2017) Surviving sepsis campaign: international guidelines for management of sepsis and septic shock: 2016. Intensive Care Med 43:304–377 Robinson ES, McKhann CF (1935) Immunological application of placental extracts. Am J Public Health Nations Health 25:1353–1358 Rodríguez A, Rello J, Neira J et al (2005) Effects of highdose of intravenous immunoglobulin and antibiotics on survival for severe sepsis undergoing surgery. Shock 23:298–304 Samuelsson A, Towers TL, Ravetch JV (2001) Antiinflammatory activity of IVIG mediated through the inhibitory Fc receptor. Science 291:484–486 Schrage B, Duan G, Yang LP et al (2006) Different preparations of intravenous immunoglobulin vary in their efficacy to neutralize streptococcal superantigens: implications for treatment of streptococcal toxic shock syndrome. Clin Infect Dis 43:743–746 Schroeder HW, Cavacini L (2010) Structure and function of immunoglobulins. J Allergy Clin Immunol 125: S41–S52 Schwab I, Nimmerjahn F (2013) Intravenous immunoglobulin therapy: how does IgG modulate the immune system? Nat Rev Immunol 13:176–189 Sekul EA, Cupler EJ, Dalakas MC (1994) Aseptic meningitis associated with high-dose intravenous immunoglobulin therapy: frequency and risk factors. Ann Intern Med 121:259–262 Shankar-Hari M, Spencer J, Sewell WA et al (2012) Bench-to-bedside review: immunoglobulin therapy for sepsis - biological plausibility from a critical care perspective. Crit Care 16:206 Shankar-Hari M, Harrison DA, Rowan KM (2016) Differences in impact of definitional elements on mortality precludes international comparisons of sepsis epidemiology—a cohort study illustrating the need for standardized reporting. Crit Care Med 44:2223–2230 Shankar-Hari M, Madsen MB, Turgeon AF (2018) Immunoglobulins and sepsis. Intensive Care Med 44:1923–1925 Siber GR, Schur PH, Aisenberg AC et al (1980) Correlation between serum IgG-2 concentrations and the antibody response to bacterial polysaccharide antigens. N Engl J Med 303:178–182 Siegel J (2006) Safety considerations in IGIV utilization. Int Immunopharmacol 6:523–527 Siegel J (2015) Immune globulins: therapeutic, pharmaceutical, cost, and administration considerations Soares M, Welton N, Harrison D et al (2012) An evaluation of the feasibility, cost and value of information of a multicentre randomised controlled trial of intravenous immunoglobulin for sepsis (severe sepsis and septic shock): incorporating a systematic review, meta-analysis and value of informati. Health Technol Assess 16:1–186

M. B. Madsen et al. Sriskandan S, Ferguson M, Elliot V et al (2006) Human intravenous immunoglobulin for experimental streptococcal toxic shock: bacterial clearance and modulation of inflammation. J Antimicrob Chemother 58:117–124 Stegmayr B, Björck S, Holm S et al (1992) Septic shock induced by group A streptococcal infection: clinical and therapeutic aspects. Scand J Infect Dis 24:589–597 Steinhagen F, Schmidt SV, Schewe J-C et al (2020) Immunotherapy in sepsis - brake or accelerate? Pharmacol Ther 107:476 Sultan Y, Kazatchkine MD, Maisonneuve P, Nydegger UE (1984) Anti-idiotypic suppression of autoantibodies to factor VIII (antihaemophilic factor) by high-dose intravenous gammaglobulin. Lancet 2:765–768 Sundhedstyrelsen (2016) Redegørelse for blodproduktområdet 2015 Takei S, Arora YK, Walker SM (1993) Intravenous immunoglobulin contains specific antibodies inhibitory to activation of T cells by staphylococcal toxin superantigens. J Clin Invest 91:602–607 Tarnutzer A, Andreoni F, Keller N et al (2019) Human polyspecific immunoglobulin attenuates group A streptococcal virulence factor activity and reduces disease severity in a murine necrotizing fasciitis model. Clin Microbiol Infect 25:512 Trautmann M, Held TK, Susa M et al (1998) Bacterial lipopolysaccharide (LPS)-specific antibodies in commercial human immunoglobulin preparations: superior antibody content of an IgM-enriched product. Clin Exp Immunol 111:81–90 Tunis MC, Salvadori MI, Dubey V et al (2018) Updated NACI recommendations for measles post-exposure prophylaxis. Can Commun Dis Rep 44:226–230 Turgeon AF, Hutton B, Fergusson DA et al (2007) Metaanalysis: intravenous immunoglobulin in critically ill adult patients with sepsis. Ann Intern Med 146:193–203 Vanderven HA, Kent SJ (2020) The protective potential of Fc-mediated antibody functions against influenza virus and other viral pathogens. Immunol Cell Biol. https:// doi.org/10.1111/imcb.12312 Viard I, Wehrli P, Bullani R et al (1998) Inhibition of toxic epidermal necrolysis by blockade of CD95 with human intravenous immunoglobulin. Science 282:490–493 Vidarsson G, Dekkers G, Rispens T (2014) IgG subclasses and allotypes: from structure to effector functions. Front Immunol 5:520 Wade A, Chang C (2015) Evaluation and treatment of critical asthma syndrome in children. Clin Rev Allergy Immunol 48:66–83 Ware JE, Sherbourne CD, Jew J (1992) The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 30:473–483 Welte T, Dellinger RP, Ebelt H et al (2018) Efficacy and safety of trimodulin, a novel polyclonal antibody preparation, in patients with severe community-acquired pneumonia: a randomized, placebo-controlled, double-blind, multicenter, phase II trial (CIGMA study). Intensive Care Med 44:438–448

8

Treatment of Necrotizing Soft Tissue Infections: IVIG

125

Werdan K, Pilz G, Bujdoso O et al (2007) Score-based immunoglobulin G therapy of patients with sepsis: the SBITS study. Crit Care Med 35:2693–2701 WHO (2020) Quality assurance and safety: blood products and related biologicals. https://www.who.int/ bloodproducts/en/. Accessed 5 Feb 2020 Wood JB, Jones LS, Soper NR et al (2017) Commercial intravenous immunoglobulin preparations contain

functional neutralizing antibodies against the Staphylococcus aureus leukocidin LukAB (LukGH). Antimicrob Agents Chemother 61:e00968 Woof JM, Kerr MA (2006) The function of immunoglobulin A in immunity. J Pathol 208:270–282 Zinman L, Ng E, Bril V (2007) IV immunoglobulin in patients with myasthenia gravis: a randomized controlled trial. Neurology 68:837–841

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections Nikolai Siemens, Johanna Snäll, Mattias Svensson, and Anna Norrby-Teglund

Contents 9.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

9.2

Initial Steps of Tissue Infections: Bacterial Adhesins and Colonizing Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129

9.3 9.3.1 9.3.2 9.3.3

Secreted Molecules Contributing to Host Evasion and Tissue Pathology Streptolysins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Proteases and DNases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Streptococcal Inhibitor of Complement (SIC) . . . . . . . . . . . . . . . . . . . . . . . . . . . .

130 130 132 135

9.4 9.4.1 9.4.2 9.4.3

Local and Systemic Inflammation in NSTIs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Neutrophil Responses in NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Macrophage Responses in NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Superantigens . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

135 135 137 138

9.5

Biofilm and Complex Modelling of GAS NSTIs in the Human Tissue Setting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 139

9.6

Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

Abstract

Necrotizing skin and soft tissue infections (NSTIs) are severe life-threatening and rapidly progressing infections. Beta-hemolytic streptococci, particularly S. pyogenes (group A streptococci (GAS)) but also

N. Siemens (*) Department of Molecular Genetics and Infection Biology, University of Greifswald, Greifswald, Germany e-mail: [email protected] J. Snäll · M. Svensson · A. Norrby-Teglund Department of Medicine, Center for Infectious Medicine, Karolinska Institutet, Huddinge, Sweden

S. dysgalactiae subsp. equisimilis (SDSE, most group G and C streptococcus), are the main causative agents of monomicrobial NSTIs and certain types, such as emm1 and emm3, are over-represented in NSTI cases. An arsenal of bacterial virulence factors contribute to disease pathogenesis, which is a complex and multifactorial process. In this chapter, we summarize data that have provided mechanistic and immuno-pathologic insight into hostpathogens interactions that contribute to tissue pathology in streptococcal NSTIs. The role of streptococcal surface associated and secreted factors contributing to the hyper-inflammatory

# Springer Nature Switzerland AG 2020 A. Norrby-Teglund et al. (eds.), Necrotizing Soft Tissue Infections, Advances in Experimental Medicine and Biology 1294, https://doi.org/10.1007/978-3-030-57616-5_9

127

128

N. Siemens et al.

state and immune evasion, bacterial load in the tissue and persistence strategies, including intracellular survival and biofilm formation, as well as strategies to mimic NSTIs in vitro are discussed. Keywords

Streptococcus · Necrotizing soft tissue infections · Pathogenesis · Neutrophils · Macrophages Highlights Necrotizing soft tissue infections (NSTIs) located to the extremities are associated with monomicrobial infections caused predominantly by group A streptococcus (GAS). Current data indicate that the pathogenesis of GAS NSTIs is associated with a variety of bacterial virulence factors, such as superantigens, cytolysins, and Sda1. IL-1β, CXCL9, CXCL10, and CXCL11 have been implicated as potential biomarkers for GAS NSTIs. Within the tissue, GAS have evolved different strategies to promote bacterial persistence and antibiotic resistance strategies. These include (1) residing within macrophages and (2) biofilm formation. 3D-organotypic skin tissue models recapitulate key anatomical and functional features of the skin and their use as infection model systems has provided novel insights into the pathogenesis of GAS NSTIs.

9.1

Introduction

Necrotizing soft tissue infections (NSTIs) are rare, severe, and rapidly progressing infections of the skin and soft tissue that account for significant morbidity and mortality (Anaya et al. 2005; Madsen et al. 2019; Morgan 2010). NSTIs can occur after traumatic injuries, including minor breaches of the skin or mucosa and non-penetrating injuries of the soft tissue (Stevens and Bryant 2017). Based on the invading microorganisms, NSTIs are commonly classified into two types. Type 1 NSTIs are of polymicrobial nature and are affecting around 70–80% of patients. Type 2 NSTIs, affecting around 20–30% of

patients, are monomicrobial infections, mostly due to Gram-positive microorganisms (Harbrecht and Nash 2016; Morgan 2010). Among these, Streptococcus pyogenes (group A streptococcus; GAS) remains the most common pathogen (Bruun et al. 2020; Madsen et al. 2019). Compared to type 1 NSTI, type 2 infections are more likely to occur in younger patients often without underlying illnesses (Stevens and Bryant 2017). It is estimated that more than 18 million people worldwide develop invasive streptococcal diseases annually (Sims Sanyahumbi et al. 2016). Several prospective population-based studies have reported skin and soft tissue as the dominating foci of invasive streptococcal infections (Davies et al. 1996; Lamagni et al. 2008; Moses et al. 2002). Depending on the geographic region and varying between years, the annual incidence rate is ranging from 1.4 to 4.7 severe cases per 100,000 populations (Kaul et al. 1997; Lamagni et al. 2008; Naseer et al. 2016; Nelson et al. 2016; Stevens et al. 1989). A Canadian prospective study on invasive GAS infections linked 48% of all cases to soft tissue infections and 24% met the criteria for NSTIs. Among these patients, 47% also presented with streptococcal toxic shock syndrome (STSS). NSTI patients who also developed STSS had a mortality of 67%, as compared to only 4.9% of patients who did not (Kaul et al. 1997). Recently an international multicenter, prospective cohort study, called INFECT, with the aim of studying the clinical profile of NSTI patients was performed in Scandinavia (Madsen et al. 2018). The INFECT study is the largest prospective cohort study in patients with NSTIs to date. In total, 409 patients were included. 179 (44%) had a monomicrobial infection, and GAS was predominant bacterial species (126 [31%]). Among GAS NSTI patients, 65% presented with STSS. All-cause mortality was 14% at day 30 and 18% at day 90. However, GAS NSTIs were associated with lower mortality rates (10%). Interestingly, NSTIs located to the extremities were associated with monomicrobial infections caused predominantly by GAS (Madsen et al. 2019). Such anatomical preference by GAS was also seen in a study of a selected cohort of 148 patients from

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections

the INFECT study by Thänert and colleagues (Thanert et al. 2019). 16S rRNA gene sequencing revealed that the diversity of the bacterial communities was highly dependent on anatomical location of NSTIs. Upper and lower extremities showed a Gini—Simpson diversity index lower than 0.25 with significant detection of streptococcal species. However, 16S rRNA sequencing of tissue biopsies also revealed the abundance of obligate anaerobes in NSTI cases (Thanert et al. 2019; Zhao-Fleming et al. 2019).

9.2

Initial Steps of Tissue Infections: Bacterial Adhesins and Colonizing Factors

NSTIs can present with and without a defined portal of entry (Stevens and Bryant 2017). In approximately half of the cases the bacteria gain entry to the deep tissue through superficial lesions, penetrating trauma or breaches of the skin (Morgan 2010). In the other 50% of cases, NSTIs initiate deep in the soft tissue, sometimes at sites of a non-penetrating trauma (Kaul et al. 1997; Stevens et al. 1989). It has been suggested that transient bacteremia originating from the nasopharynx, a potential reservoir for streptococci, is the source of such infections (Johansson et al. 2010; Stevens and Bryant 2017). However, reports providing such evidence are lacking. The ability of GAS to establish an infection can be primarily attributed to the surface anchored and associated adhesins. Among these, the streptococcal M protein is the most abundant and probably best characterized protein, which is also used for emm-typing (Beall et al. 1996). Since its discovery more than 200 different emm serotypes have been described and nearly all GAS isolates harbor an emm gene (Bessen et al. 2015). In addition, many GAS strains encode emm-like genes, mrp and enn (Bessen et al. 2015). Based on the chromosomal arrangement of these three genes, five distinct patterns, A to E, are described. Emm pattern A-C isolates show a significant association with nasopharynx, while pattern D strains are often isolated from skin infections (Bessen

129

2009). In contrast, emm pattern E isolates account for almost equal fractions at both tissue sites (Bessen et al. 2015). Epidemiological studies revealed that severe disease manifestations, including NSTIs and STSS, are predominantly associated with emm1 and emm3 serotypes (Bruun et al. 2020; Darenberg et al. 2007; LucaHarari et al. 2009; Stevens 1999). The adhesion by GAS to epithelial surfaces is a two-step process. The first step is a relatively weak and unspecific and overcomes electrostatic repulsion. It was suggested that lipoteichoic acid (LTA) is a mediator of the first step of adhesion. This interaction is reversible and not sufficient to provide for tissue tropism (Courtney et al. 2002). The second step of adhesion may then involve classical and non-classical surface proteins of GAS. These adhesins either bind directly to the cell surface receptors or use extracellular host matrix proteins or plasma proteins as a prime target for specific interactions (Kreikemeyer et al. 2004). In skin infections, M proteins can mediate cell adhesion to keratinocytes directly via CD46 binding (Okada et al. 1995) or indirectly by binding human fibronectin as a target on host epithelial cells (Courtney et al. 1986). In addition to proteinaceous ligands, M proteins bind to glycosaminoglycans and that interaction promotes GAS attachment to human skin fibroblasts (Frick et al. 2003b). There are other classical and non-classical adhesins described for GAS, including hyaluronic acid capsule that binds to CD44 on keratinocytes, fibronectinbinding proteins FbaA, F2, serum opacity factor (SOF), Fbp54, ScpA, Lbp, and FbaB, collagenbinding protein Cpa, vitronectin-binding protein, galactose-binding protein, and plasminogenbinding proteins Epf, GAPDH, enolase, and phosphoglycerate kinase (PGK) (Rohde and Cleary 2016). Hence, GAS express a variety of surface factors that employ different mechanisms contributing to bacterial adhesion. However, (1) not all of them occur in all serotypes, (2) they are expressed at different bacterial growth cycles, and (3) many of them are host cell type specific (Courtney et al. 2002). For example, M proteins mediate adhesion to human keratinocytes, but not to buccal cells (Okada et al. 1994).

130

On another note, the secreted exotoxins, namely superantigens (SAgs), were recently shown to be critical for colonization. The reports showed that the SAgs elicited an inflammatory response that promoted bacterial colonization and establishment of infection (Kasper et al. 2014; Zeppa et al. 2016, 2017). This provided new insight into the function of this family of bacterial exotoxins, most well known for their pro-inflammatory activities and association with the cytokine storm characterizing toxic shock syndrome (detailed in separate subsections below). Once the infection is established GAS can translocate to deeper tissue or initiate internalization into the cells. Although GAS have long been considered to be an extracellular pathogen, now it is appreciated that GAS can internalize into non-phagocytic cells even to an equal level as classical intracellular pathogens (Lapenta et al. 1994). For example, M protein forms complexes with fibronectin and interacts with α5β1 integrins. Extensive integrin clustering triggers caveolae aggregation to form invaginations that ingest the bacteria without detectable actin polymerization. In that way intracellular bacteria bypasses fusion with lysosome and intracellular streptococci reside within caveolin-1-contating endosomes, the so-called caveosomes. Alternatively, M protein-fibronectin interaction leads to the cytoskeletal rearrangement with actin accumulation around the streptococci. This results in trafficking of the bacteria inside the cells through early and late endosomes, which fuse with lysosomes to form phagolysosomes (Rohde and Cleary 2016). The first mechanism directs GAS to the safer caveosomal compartment, while the second results in most of the bacteria being killed by the host cells (Dombek et al. 1999). In line with this, Thänert and colleagues have shown that transcripts of genes encoding fibronectin-binding proteins, which mediate adherence to the extracellular matrix and invasion into human host cells, are highly abundant in NSTI patient biopsies (Thanert et al. 2019). Thus, the results underline the importance of streptococcal surface proteins in NSTIs and suggest that there are additional functions to be explored in future studies.

N. Siemens et al.

9.3

9.3.1

Secreted Molecules Contributing to Host Evasion and Tissue Pathology Streptolysins

To breach dermal barriers and enable the bacteria to disseminate through the tissue, GAS secrete several pore-forming exotoxins, proteases, and DNases (summarized in Table 9.1). Studies have shown that secreted factors are highly up-regulated during fulminant tissue infections and play a crucial role in NSTIs (Johansson et al. 2010; Olsen and Musser 2010; Thanert et al. 2019). Pore-forming toxins are a class of bacterial virulence factors that cause cytolysis, disrupt membranes, and have immunemodulatory functions. Nearly all GAS isolates secrete two potent hemolysins, namely Streptolysin S (SLS) and Streptolysin O (SLO) (Barnett et al. 2015; Yoshino et al. 2010). SLO is a 57 kDa oxygen sensitive, cholesterol-dependent cytolysin. It disrupts cytoplasmic membrane integrity of epithelial and endothelial cells, monocytes, macrophages, and neutrophils, thereby inducing cell death through different mechanisms, including apoptosis, necrosis, and/or pyroptosis (Chandrasekaran and Caparon 2016; Keyel et al. 2013; Timmer et al. 2009). For example, SLO blocks the clathrin-dependent pathway for GAS internalization and induces keratinocytes apoptosis through dysregulation of the calcium signaling (Cywes Bentley et al. 2005). At sub-lytic concentrations, SLO suppresses crucial neutrophil functions, including chemotaxis and formation of neutrophil extracellular taps (NETs) (Uchiyama et al. 2015). Furthermore, SLO impairs intracellular GAS clearance by neutrophils and professional phagocytes. It inhibits dendritic cell (DC) maturation by suppressing the expression of major histocompatibility complex (MHC) class II and co-stimulatory molecules CD80 and CD83 (Cortes and Wessels 2009) and prevents phagolysosome acidification in macrophages by damaging the phagolysosome membrane (Bastiat-Sempe et al. 2014). In murine models of

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections

131

Table 9.1 GAS secreted virulence factors GAS factor SLO

Main immunological functions Cytolysis, suppression of neutrophil chemotaxis

SLS

Cytolysis, activation of pain producing sensory nociceptor neurons

SpeB

Hydrolysis of a broad spectrum of host and bacterial substrates leading to an impaired recognition and killing of GAS Systemic spread through plasminogen activation and fibrinolysis

Ska SpyCEP EndoS/ IdeS Sda1 SIC SAgs

Chemokine degradation resulting in an impaired chemotaxis Immunoglobulin degradation, impaired opsonization DNA degradation leading to an impaired bacterial recognition and killing by phagocytes Interferes with complement functions and blocks lytic activities of antimicrobial peptides Cytokine storm through excessive T cell and APC activation, colonization

soft tissue infections, it was shown that GAS mutants lacking slo gene were less virulent as compared to the parental strain (Zhu et al. 2017). In contrast to SLO, SLS is a small non-immunogenic peptide of 2.7 kDa, which is encoded by the SLS-associated gene (sag) locus consisting of nine genes (sagABCDEFGHI). The sagA gene encodes the premature form of SLS, while other genes encode proteins for posttranslational processing and export (Nizet et al. 2000). SLS targets mainly human red blood cells, platelets, and leukocytes. It accumulates on cell membranes of eukaryotic cells and induces pore formation by a yet not fully characterized mechanism (Barnett et al. 2015). Higashi and colleagues suggested that SLS interacts with the major erythrocytes anion exchange protein band 3, thereby inducing osmotic change, rapid Cl influx, and erythrocyte lysis (Higashi et al. 2016). Whether similar mechanisms apply to other human host cells needs to be proven. However, experimental NSTI models have shown that SLS facilitates translocation of GAS across epithelial barriers through direct cleavage of junctional proteins E-cadherin and occludin (Sumitomo et al. 2011). Furthermore, SLS-dependent cytotoxicity to human keratinocytes is mediated through activation of pro-inflammatory mediators

Key references Cywes Bentley et al. (2005), Uchiyama et al. (2015) Nizet et al. (2000), Pinho-Ribeiro et al. (2018) Nelson et al. (2011), Larock et al. (2016) Boxrud et al. (2004), Nitzsche et al. (2016) Zinkernagel et al. (2008) Collin et al. (2002), Naegeli et al. (2019) Keller et al. (2019), Uchiyama et al. (2012) Westman et al. (2018) Kasper et al. (2014), Llewelyn and Cohen (2002), Levy et al. (2016)

p38 MAPK and NFκB. Subcutaneous infections of mice result in IL-1β production in a SLS-dependent manner (Flaherty et al. 2018). A recent study by Pinho-Ribeiro and colleagues has shown that SLS is also responsible for activation of pain producing sensory nociceptor neurons (Fig. 9.1) (Pinho-Ribeiro et al. 2018). This is of great interest as an out-of-proportion pain at early stages of infections is one of the critical hallmark symptoms of NSTIs. SLS-dependent activation of nociceptor neurons results in a release of calcitonin gene-related peptide (CGRP) into infected tissue, which in turn suppresses critical neutrophil functions, including recruitment to the site of infection, opsonophagocytosis, and killing of the bacteria (Pinho-Ribeiro et al. 2018). These actions allow the bacteria to spread in the tissue. Once deeper skin layers are reached, direct cytotoxicity of SLS towards different cell compartments provokes neutrophil influx (Humar et al. 2002). In turn, SLS actively destroys neutrophils which are recruited to the site of infection (Humar et al. 2002). These events potentially negatively contribute to the outcome in patients: (1) early CGRP-mediated suppression of influx resulting in a lack or reduced number of neutrophils is an unfavorable prognostic sign in GAS NSTIs (Bakleh et al. 2005) and (2) hyper-

132

N. Siemens et al.

Fig. 9.1 Immune evasion strategies by GAS. GAS secrete a variety of virulence factors responsible for bacterial spread in the tissue. Streptokinase (Ska) forms complexes with plasminogen (Plg). These Ska-Plg complexes degrade fibrin clots, LL-37, and all classes of human histones. Streptococcal pyrogenic exotoxin B (SpeB) is a cysteine protease with a broad spectrum of human substrates. SpeB degrades all classes of immunoglobulins, complement components (e.g., C3), antimicrobial peptides (e.g., LL-37), extracellular matrix proteins (ECM), and CXCL chemokines including

CXCL1-7. SpyCEP, a subtilisin-like protease, degrades CXCL8 which results in impaired neutrophil recruitment to the site of infection. SLS-dependent activation of nociceptor neurons (N) results in a release of calcitonin generelated peptide (CGRP), which suppresses neutrophil recruitment. Furthermore, GAS have developed strategies to avoid phagocytic killing. They reside within macrophages. However, certain environmental signals trigger the egress out of the cells and bacterial spread within the tissue. In addition, GAS form biofilms within the tissue

activation and SLS-dependent damage of neutrophils at later time points result in augmented tissue pathology through neutrophilderived factors (Johansson et al. 2009, 2010; Johansson and Norrby-Teglund 2013; Nuwayhid et al. 2007).

9.3.2

Proteases and DNases

9.3.2.1 The Cysteine Protease SpeB In addition to pore-forming toxins, streptococcal proteases damage tissue barriers, inhibit immune cell transmigration to the site of infection, and

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections

impair directly or indirectly immune cell functions. One of the first proteases identified in GAS is the streptococcal pyrogenic exotoxin B (SpeB) (Elliott 1945). SpeB gene is highly conserved in GAS (Yu and Ferretti 1991) and it encodes a zymogen of 40 kDa which is autocatalytically cleaved into a mature 28 kDa protease (Liu and Elliott 1965). SpeB has a broad spectrum of host and bacterial substrates. It removes GAS surface anchored proteins, including M protein, fibronectin-binding proteins, and C5a-peptidase, and hydrolyzes secreted proteins, such as streptokinase (Ska), SLO, EndoS, and superantigens (Nelson et al. 2011). At the host site, SpeB degrades immunoglobulins and components of the complement system (Fig. 9.1) (Kuo et al. 2008). These events result in impaired opsonophagocytosis and subsequent failure in phagocytic clearance of the bacteria (Collin et al. 2002). Moreover, SpeB degrades antimicrobial peptides produced by the host, including LL-37, which was shown to provide protection against murine soft tissue infections (Nizet et al. 2001). Neutrophil-derived LL-37 is highly abundant in human biopsy specimens and its expression is positively correlated to the bacterial load (Johansson et al. 2008). However, such a positive correlation suggests that LL-37 is not efficient in bacterial killing. Nyberg and colleagues have shown that SpeB gets entrapped by α2-macroglobulin-GRAB complex on bacterial surface. As a consequence, SpeB is retained at the bacterial surface and protects bacteria against killing by LL-37 (Nyberg et al. 2004). Furthermore, SpeB degrades a wide range of chemokines, including CXCL chemokines CXCL1-7, CXCL10-14, CXCL16, and CCL20, XCL1, and CX3CL1 (Egesten et al. 2009), and cleaves pro-IL-1β into mature form (Kapur et al. 1993). LaRock and colleagues suggested that IL-1β acts as an intracellular sensor of bacterial proteolytic activity (Larock et al. 2016). GAS have adapted different strategies, including mutations abrogating SpeB expression to evade detection by IL-1β and subsequent killing of the pathogen (Sumby et al. 2006). Genes encoding for CovR/S two component system (TCS) and transcriptional stand-alone regulator RopB,

133

which directly or indirectly control SpeB expression, are such mutational hot-spots (Hollands et al. 2008; Sumby et al. 2006). Although SpeB shows such a broad spectrum of substrates, its role in invasive diseases is still under debate. The speB gene can be found in all types of GAS isolates (Darenberg et al. 2007; Luca-Harari et al. 2009). It was shown that SpeB is readily detectable in sera and tissues of NSTI patients (Gubba et al. 1998; Siemens et al. 2016). Some studies demonstrate that GAS isolates from non-invasive infections express higher amounts of SpeB and subsequently show greater activity as compared to invasive strains (Kansal et al. 2000). However, other studies demonstrate the opposite. GAS isolates from invasive infections produce high amounts of SpeB and low anti-SpeB antibody titers in patients were associated with fatal outcome (Holm et al. 1992). Furthermore, the controversy of SpeB contribution in NSTIs continues in murine soft tissue infection models. Some authors report that SpeB highly contributes to disease severity, tissue damage, bacterial dissemination, and mortality (Kuo et al. 1998; Lukomski et al. 1997, 1998, 1999). Others demonstrated that speB-deficient mutant strains are as virulent as their parental wild type strains (Ashbaugh et al. 1998; Ashbaugh and Wessels 2001). However, human tissue biopsies from NSTI patients are strongly positive for SpeB and a mixture of SpeB-positive and negative clones is usually seen in patients (Johansson et al. 2008; Siemens et al. 2016).

9.3.2.2 Streptokinase (Ska) and SpyCEP In contrast to eliminated SpeB expression in natural GAS covR/S mutants, several other secreted factors are up-regulated in the same genetic background. This includes SLO, SpyCEP, Sda1, and Streptokinase (Ska) (Cole et al. 2011; Siemens et al. 2015; Sumby et al. 2006; Walker et al. 2007). Ska is a non-enzymatic plasminogen (Plg) activator. GAS cover their surface with Plg via different surface proteins, which in turn results in acquisition of endogenously produced and secreted Ska (Chandrahas et al. 2015; Siemens et al. 2011). The Ska-Plg complex exposes the active site and proteolytically

134

converts surface-bound Plg to plasmin (Boxrud et al. 2004). Plasmin is a broad-spectrum host protease that is involved in fibrinolysis. In the absence of SpeB, active plasmin accumulates on bacterial surface, resulting in protease activity that enables the bacteria to degrade tissue barriers and facilitate invasive spread (Cole et al. 2006; Sanderson-Smith et al. 2008). Furthermore the Ska-Plg complex degrades LL-37 and all classes of human histones and abrogates their antibactericidal effects (Fig. 9.1) (Hollands et al. 2012; Nitzsche et al. 2016). Due to human hostspecificity of Ska, no differences between wild type and ska-deficient GAS strains were seen in murine models (Khil et al. 2003). In contrast, by using humanized mice expressing human plasminogen it was shown that activation of host Plg by Ska leads to accelerated clearance of host fibrin, which plays a central mechanism in GAS invasion and spread in tissues (Sun et al. 2004). It is becoming more evident that antimicrobial peptides including LL-37 act as immunomodulatory and chemotactic agents (Beisswenger and Bals 2005). Furthermore, LL-37 influences bacterial virulence properties. It induces an invasive phenotype of GAS by up-regulating gene expression of SpyCEP, among others (Gryllos et al. 2008). SpyCEP is a surface localized subtilisinlike protease of 180 kDa, which is highly expressed in vivo (Zinkernagel et al. 2008). It cleaves human CXCL chemokines, including CXCL1, CXCL2, CXCL3, CXCL5, CXCL6, and CXCL8, and murine chemokines CXCL1 and CXCL2 (Fig. 9.1) (Hidalgo-Grass et al. 2004, 2006; Sumby et al. 2008; Zinkernagel et al. 2008). Considering the role of CXCL8 in human neutrophil chemotaxis to the infected tissue, it is tempting to assume that lack of neutrophils due to CXCL8 degradation will result in failed bacterial clearance and severe tissue pathology caused by bacterial toxins. Indeed, several murine studies of soft tissue infections have demonstrated a paucity of neutrophil influx resulting from chemokine degradation by SpyCEP (Edwards et al. 2005; Hidalgo-Grass et al. 2004, 2006; Sumby et al. 2008; Zinkernagel et al. 2008). However, infiltration of neutrophils in superficial fascia and dermis is one of the

N. Siemens et al.

pathological criteria for diagnosis of NSTIs (Stamenkovic and Lew 1984). Several studies utilizing patient’s biopsy material have shown that neutrophils represent one of the major immune cell infiltrating populations in NSTIs (Johansson et al. 2009, 2014; Siemens et al. 2016; Thulin et al. 2006). The infiltration of neutrophils positively correlated with bacterial load in the tissue (Thulin et al. 2006). These facts suggest that the murine models are limited in mimicking the complexity of human infection setting, in which a plethora of chemotactic signals exists.

9.3.2.3

The Immunoglobulin Degrading Enzymes IdeS and EndoS GAS have also developed strategies to evade adaptive immunity at the site of infection. The bacteria secrete two enzymes, namely immunoglobulin degrading enzyme (IdeS) and endoglycosidase (EndoS). IdeS hydrolyzes four subclasses of IgG (Von Pawel-Rammingen et al. 2002). However, its role in NSTIs is not yet clear. EndoS, which transcription is highly up-regulated during NSTIs (Thanert et al. 2019), specifically cleaves the conserved N-glycan of IgG antibodies. A recent study demonstrated that EndoS specifically alters IgG Fc glycosylation in vivo. At the site of infection, IgG glycan hydrolysis positively correlates with bacterial load and EndoS secretion (Naegeli et al. 2019). These actions potentially result in failed opsonophagocytosis and pathogen clearance. The authors showed that ndoS-mutant is more susceptible to phagocytic killing as compared to its parental wild type strain. In contrast to tissue findings, glycan hydrolysis in circulation was only detected in patients with septic shock (Naegeli et al. 2019). 9.3.2.4 Streptodornase 1(Sda1) Sda1 is a secreted enzyme which is predominantly associated with emm1 GAS NSTI isolates. It displays potent sequence-nonspecific nuclease activity on DNA substrates (Aziz et al. 2004a). It occurs during phase-shifts to hyper-virulent invasive infection and is believed to play a role in evasion of the host’s innate immune responses

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections

(Aziz et al. 2004b; Walker et al. 2007). Sda1 is involved in degradation of the DNA component of chromatin-rich NETs (Walker et al. 2007), a mechanism which provides a protective effect against bacterial killing by neutrophils. Recently it was also shown that Sda1-mediated degradation of streptococcal DNA alters recognition of GAS by Toll-like receptor (TLR)9 (Uchiyama et al. 2012), which preferentially senses bacterial and viral DNA. DNA degradation by Sda1 prevented GAS induced IFNα and TNFα secretion from murine macrophages and contributed to bacterial survival. Furthermore, IFNα and TNFα levels were significantly decreased in wild type mice infected with Sda1-positive GAS strain as compared to TLR9-deficient mice (Uchiyama et al. 2012). In addition, Sda1 impaired plasmacytoid (p)DC recruitment to the site of infection. Keller and colleagues suggested that Sda1 interferes with stabilization of the DNA in a HMGB1dependent manner, which may result in decreased IFN-1 levels and subsequently lead to reduced numbers of pDCs (Keller et al. 2019).

9.3.3

The Streptococcal Inhibitor of Complement (SIC)

In addition to proteases and DNases, a limited number of GAS M serotypes (emm1, emm12, emm55, emm57) secrete a protein called streptococcal inhibitor of complement (SIC). SIC is a highly polymorphic 31 kDa negatively charged protein which interferes with complement functions (Akesson et al. 1996; Fernie-King et al. 2001). Furthermore, SIC blocks the lytic activity of antimicrobial peptides such as LL-37, defensins, H-kininogens, and lysozyme (FernieKing et al. 2002, 2004; Frick et al. 2003a, 2011). Recently it was shown that SIC also binds to extracellular histones, a group of host-associated damage-associated molecular patterns (DAMPs) released during necrotizing tissue infections, and neutralizes their antimicrobial activity. Such SIC-histone aggregates are readily detectable in biopsies of NSTI patients (Westman et al. 2018). Although SIC is considered to be an antiinflammatory agent (Akesson et al. 2010), SIC

135

binding to histones boosted the release of a broad range of pro-inflammatory cytokines and chemokines (Westman et al. 2018). Such actions can potentially further contribute to hyperinflammation in tissue environment.

9.4 9.4.1

Local and Systemic Inflammation in NSTIs Neutrophil Responses in NSTI

Despite the fact that GAS factors contribute to tissue pathology, it is widely accepted that immune cells and host-derived factors play a crucial role in hyper-inflammatory and tissue destructive processes (Herwald et al. 2004; Johansson et al. 2009; Kahn et al. 2008). Several immune cell populations are recruited to the site of infection, including neutrophils, macrophages, T cells, and dendritic cells (Norrby-Teglund et al. 2001). Neutrophil migration to the site of infection is a crucial step in host defense. Once neutrophils have reached the infection site, three distinct mechanisms to fight an infection can occur: phagocytosis, degranulation, and formation of NETs (Kolaczkowska and Kubes 2013). Critical tissue damaging components of all these processes are granule effector molecules, including proteolytic enzymes and antimicrobial peptides (Borregaard et al. 2007). GAS have been demonstrated to have a significant impact on neutrophils, by inducing an apoptotic program and triggering degranulation (Kobayashi et al. 2003; Snall et al. 2016; Soehnlein et al. 2008) and both, surface associated as well as secreted factors are potent inducers of neutrophil degranulation (Snall et al. 2016; Uhlmann et al. 2016b). Strong neutrophil influx in the tissue is a typical characteristic for GAS NSTIs and there is a significant correlation between bacterial load, neutrophil influx, presence of degranulation markers heparin-binding protein (HBP) and resistin, and tissue pathology and inflammation (Fig. 9.2) (Johansson et al. 2008, 2009, 2014; Snall et al. 2016). Streptococcal M protein is one of the major triggers of neutrophil activation. Soluble M1 protein forms complexes with fibrinogen

136

N. Siemens et al.

Fig. 9.2 Immune cell mediated hyper-inflammatory response in the tissue triggered by secreted streptococcal virulence factors. Several GAS factors contribute to hyperinflammation of the tissue. For example, soluble M1 protein (sM1) forms complexes with fibrinogen, which in turn activates neutrophils. Furthermore, PGK and streptolysins S and O (SLS/SLO) trigger neutrophil activation. This results in a release of granule effector molecules, including resistin, heparin-binding protein (HBP), LL-37, and myeloperoxidase, which augments tissue inflammation.

SLS and SLO induce through different mechanisms host cell lysis resulting in a release of pro-inflammatory cytokines and chemokines. In addition, SpeB cleaves proIL-1β into mature IL-1β leading to pyroptosis of macrophages. GAS also secrete superantigens (SAgs). SAgs bind without cellular processing to α- and β-chains of MHC class II molecules on antigen-presenting cells (APC) and variable β-chains of T cell receptor. These interactions result in a massive cytokine storm

and binds β2-integrins on neutrophil surface. This event results in a release of massive amounts of the granule proteins, including HBP, which is

solely responsible for vascular leakage (Gautam et al. 2001; Herwald et al. 2004). NSTI patient biopsies are highly positive for such M protein/

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections

fibrinogen complexes (Herwald et al. 2004; Kahn et al. 2008). Furthermore, streptococcal secreted factors which are susceptible to SpeB degradation, e.g., PGK and SLO, are potent inducers of neutrophil degranulation (Nilsson et al. 2006; Uhlmann et al. 2016b). Nilsson and colleagues observed that GAS supernatants containing high levels of SLO activated neutrophils by perforating these cells. As a consequence, neutrophils secreted HBP, LL-37, α-defensins, and neutrophil elastase (Fig. 9.2). Bacterial supernatants with negligible levels of SLO failed to induce degranulation (Nilsson et al. 2006). Most likely, these effects were observed due to the presence or absence of SpeB in bacterial supernatants. Uhlmann et al. demonstrated that high degranulation response was triggered exclusively by supernatants from SpeB-negative strains (Uhlmann et al. 2016b). Global proteome analyses identified phosphoglycerate kinase (PGK) as a stimulatory factor for neutrophils. Further experiments confirmed that PGK is susceptible to SpeB degradation (Uhlmann et al. 2016b). The neutrophil response and subsequent tissue hyper-inflammation can be further augmented through the streptococcal vesicle-like structures. It was shown that GAS respond to sub-inhibitory concentrations of LL-37 by releasing vesicles containing several virulence factors. These vesicles are of pro-inflammatory nature and elicit resistin and myeloperoxidase release from neutrophils (Uhlmann et al. 2016a).

9.4.2

Macrophage Responses in NSTI

Several studies have shown that in addition to neutrophils, macrophages are also predominantly infiltrating highly infected tissue of NSTI patients (Johansson and Norrby-Teglund 2013; Johansson et al. 2010, 2014; Siemens et al. 2016; Thulin et al. 2006). Macrophages are highly effective professional phagocytes. They ingest bacteria by recognition of GAS via surface pattern recognition receptors (e.g., TLRs), which in turn induces remodeling of the actin cytoskeleton to form a phagosome (Valderrama and Nizet 2018). The phagosome interacts with endosomes and

137

lysosomes, leading to phagolysosome formation, which would usually kill the pathogen. In vitro studies have shown that GAS activate NFκB and MAPK in a MyD88-dependent manner leading to production of pro-inflammatory cytokines, including TNF and IL-6 (Gratz et al. 2008). In addition, an up-regulation of the maturation markers such as CD80 and CD86 was noted. Experimental mice models of GAS skin infection demonstrated that impaired MyD88 signaling results in diminished pro-inflammatory response, including IL-12, IFNγ, and TNF production as well as chemoattractants such as CXCL1 and CCL2. This in turn leads to the impaired recruitment of macrophages and neutrophils (Loof et al. 2010). In addition, TNF-deficient mice is highly susceptible to GAS soft tissue infections due to the defect in macrophage recruitment to the site of infection (Mishalian et al. 2011). However, mice studies are highly controversial. MAPK and NFκB activation and MyD88-dependent cytokine production are TLR-independent processes in mice as compared to humans (Gratz et al. 2008). Analyses of human NSTI biopsies revealed that even if the biopsies were highly infiltrated by neutrophils and macrophages, high amounts of viable bacteria were readily detectable, even in those patients with prolonged antibiotic therapy. Viable bacteria were mainly residing within macrophages and such biopsies were characterized by lower inflammation level and lower bacterial load (Thulin et al. 2006). As previously mentioned, bacteria are normally killed by phagolysosomal degradation. However, GAS have acquired specific mechanisms to avoid such killing. These processes involve M1-proteindependent trafficking and impaired fusion of phagosome with lysosome leading to a persistent infection (Hertzen et al. 2010), adaptation to acidic environment, and escape from the phagolysosome (Fig. 9.1) (Bastiat-Sempe et al. 2014). GAS pore-forming toxin SLO and its co-toxin NAD-glycohydrolase (NADase) mediate GAS intracellular survival in macrophages. The two toxins prevent phagolysosomal acidification. SLO perforates the phagolysosomal membrane and enables translocation of NADase into the macrophage cytosol. Subsequently, NADase

138

augments SLO-mediated cytotoxicity by depleting cellular energy stores (Bastiat-Sempe et al. 2014). However, the bacteria do not only persist, but also replicate within macrophages and egress (Hertzen et al. 2010). Ihk/Irr TCS was identified as a key regulatory system at early stages of bacterial persistence within macrophages. Furthermore, Kachroo and colleagues have shown that Ihk/Irr TCS is highly up-regulated in a necrotizing myositis model in primates (Kachroo et al. 2020). In contrast, up-regulation of CovR/S and down-regulation of Ihk/Irr TCS were seen at late stages when the bacteria were exiting the cells (Hertzen et al. 2012). It was also suggested that SpeB, which is regulated by the CovR/S TCS, contributes to intracellular survival and bacterial egress out of the cell (Thulin et al. 2006). LaRock and colleagues demonstrated that SpeB cleaves IL-1β, which is usually seen as a marker of canonical and caspase-1 dependent inflammasome activation and subsequent pyroptosis of the cells (Larock et al. 2016). However, the direct cleavage by SpeB induces a non-canonical IL-1β activation in macrophages (Fig. 9.2). Analyses of patient’s tissue biopsies revealed that once IL-1β is released it forms immuno-stimulatory complexes with High Mobility Group Box 1 (HMGB1) protein (Johansson et al. 2014). HMGB1 is a ubiquitously expressed nuclear protein that is released through active secretion by innate immune cells or through passive release by injured cells (Skinner 2010). Extracellular HMGB1 has been reported to act as an alarmin with ability to activate the immune system and elevated levels of HMGB1 were found in patients diagnosed with sepsis and/or septic shock (Sunden-Cullberg et al. 2005). In GAS tissue infections, HMGB1 expression correlated with disease severity and macrophages were identified as a predominant source of HMGB1 release (Johansson et al. 2014). A close proximity between HMGB1 and neutrophils suggested also a chemotactic role of this molecule in GAS NSTIs which was further confirmed by in vitro experiments (Johansson et al. 2014). Despite the fact that macrophages contribute to the highly pro-inflammatory state in the tissue, there is evidence rising that GAS induce a

N. Siemens et al.

mixed activation of murine macrophages (Goldmann et al. 2007). Transcriptomic analyses revealed that GAS infections induce both, Th1 (classical) signatures characterized by induction of IL-1 and IL-6 cytokines, as well as Th2 (alternative) responses, e.g., induction of IL-1 decoy receptor and IL-10 (Goldmann et al. 2007). However, this phenomenon was observed in mice.

9.4.3

Superantigens

NSTIs are often complicated by STSS (Johansson et al. 2010). Sepsis-3 criteria define sepsis as a life-threatening organ dysfunction caused by a dysregulated host response to an infection. STSS is a subset of sepsis in which particularly profound circulatory, cellular, and metabolic abnormalities are associated with a greater risk of mortality (Singer et al. 2016). STSS is a result of an invasive GAS infection (Low 2013) and superantigens (SAgs) are key mediators of a systemic excessive inflammatory response (Llewelyn and Cohen 2002). To date, 13 GAS SAgs were identified, which include streptococcal pyrogenic exotoxins (Spe) A, C, G-M, streptococcal superantigen (SSA), and streptococcal mitogenic exotoxin Z (SmeZ) (Commons et al. 2014) and the two recently identified superantigens SpeQ and SpeR (Reglinski et al. 2019). SAgs bind intact to α- and/or β-chains of the major histocompatibility complex (MHC) class II molecules on antigen-presenting cells (APCs) and to the variable β-chains of the T cell receptor. In addition, SAgs can also bind a co-stimulatory molecule CD28 and its ligand CD86 (B7-2) (Arad et al. 2011; Levy et al. 2016). Once the MHC-peptide specificity of T cells is bypassed, these interactions result in a massive cytokine storm, which includes production of TNF, IFN-γ, IL-1, IL-2, IL-6, CXCL8, CCL2, and CCL3 (Fig. 9.2) (Chatila and Geha 1993). However, such responses may vary due to specific HLA class II haplotypes which may influence the severity of GAS infections. Kotb and colleagues have shown that certain haplotypes (e.g., DRB1*1501/DQB1*0602) are of a protective nature resulting in attenuated

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections

inflammatory responses to the streptococcal SAgs. In contrast, other haplotypes (e.g., DRB1*14/DQB1*0503 and DRB1*07/ * DQB1 0201) are associated with a risk to develop severe systemic responses (Kotb et al. 2002). A recent report identified that Mucosa-associated invariant T (MAIT) cells are strongly activated by SAgs and represent the major T cell source of IFNγ and TNF in the early cytokine response associated with STSS (Emgard et al. 2019). However, the majority of the SAg studies are confined to the systemic effects and only a limited number of studies investigated SAg-driven events at the deep tissue site (Johansson et al. 2010; NorrbyTeglund et al. 2001). Norrby-Teglund and colleagues showed a massive in vivo production of SAgs in biopsies of GAS NSTI patients (Norrby-Teglund et al. 2001). The infected tissue was characterized by an enormous infiltration with monocytes, macrophages, and T cells, which correlated with a typical SAg-driven excessive inflammatory response and the severity of infection (Norrby-Teglund et al. 2001). In addition, up-regulation of the homing receptors CCR5, CD44, and cutaneous lymphocyteassociated antigen correlated with the expression of Th1 cytokines at the tissue site (NorrbyTeglund et al. 2001). T cell derived Th1 cytokines and predominantly IFN-γ induce production of IL-1β by monocytes and macrophages (Chizzolini et al. 1997). In line with this, it was shown that IL-1β plays a crucial role in GAS NSTIs. Mice lacking IL-1R are more susceptible to GAS infections (Hsu et al. 2011). Furthermore, it was reported that patients who received the IL-1R antagonist Anakinra experience more frequently invasive GAS diseases including NSTIs, which results in a greater risk of mortality (Larock et al. 2016). In addition, by utilizing BXD mice Chella Krishnan and colleagues have shown that IL-1β participates in modulating GAS NSTIs (Chella Krishnan et al. 2016). The authors demonstrated that IL-1β gene expression was significantly up-regulated in tissue biopsies of GAS NSTI susceptible mice strains as compared to non-susceptible strains. This finding was further verified in plasma samples and tissue biopsies from GAS NSTI patients, as well as samples

139

from an in vitro experimental model of human skin. High IL-1β mRNA levels were detected in infected tissues. In addition, analyses of systemic IL-1β revealed high levels in patient plasma collected during an acute phase of GAS NSTIs. These levels declined during the treatment (Chella Krishnan et al. 2016). A study by Hansen and colleagues underlined the importance of IL-1β in NSTIs. IL-1β was determined as an independent predictor for 30-day mortality in NSTIs (Hansen et al. 2017). In addition, significantly higher IL-6 and TNF plasma levels in streptococcal infected patients were detected as compared to patients infected with other microbiological agents, which is consistent with septic shock being more prevalent in streptococcal NSTIs (Hansen et al. 2017). Comparative RNAseq analyses of tissue specimens from streptococcal and polymicrobial NSTIs identified a core of common up-regulated host genes encoding pro-inflammatory mediators, including IL-6 and CXCL8, complement components, alarmins, and proteolytic enzymes (Thanert et al. 2019). However, GAS NSTI biopsies were characterized by a higher expression of genes encoding interferon-inducible mediators, which includes CXCL9, CXCL10, CXCL11, MX1, and MX2. This finding was further verified in patient plasma samples. CXCL9, CXCL10, and CXCL11 chemokine plasma levels were higher in GAS NSTI patients as compared to patients with polymicrobial NSTIs (Thanert et al. 2019). Thus, these chemokines might serve as potential biomarkers for GAS NSTIs.

9.5

Biofilm and Complex Modelling of GAS NSTIs in the Human Tissue Setting

Biofilm formation is recognized as a virulence trait of many chronic tissue infections, including cystic fibrosis, burn wound infections, and infected diabetic foot ulcers (Bjarnsholt et al. 2018). Although GAS biofilm studies were performed in the past, until recently, GAS have not been considered a major biofilm forming species due to the lack of clinically relevant reports.

140

Bacteria can exist in two states, planktonic or biofilm and it has become clear that it is a dynamic process allowing the bacteria to frequently flux between these states (Mcdougald et al. 2011). The process of biofilm formation can be summarized in four steps: (1) reversible attachment to abiotic or biotic surfaces driven by environmental signals (e.g., pH, temperature, and nutrients availability), (2) irreversible attachment with initial microcolony formation and development of an extracellular matrix, (3) biofilm maturation, and (4) dispersal (Kostakioti et al. 2013). The first report providing evidence of GAS biofilm in vivo was published by Akiyama and colleagues (Akiyama et al. 2003). Confocal laser scanning microscopy (CLSM) of impetigo skin biopsies revealed concanavalin A positive GAS microcolonies consistent with biofilm formation (Akiyama et al. 2003). The second study performed on tonsils from children with recurrent GAS pharyngitis identified three dimensional communities of GAS within reticulated crypts (Roberts et al. 2012). However, both studies were of descriptive nature and have not addressed potential pathogenic effects of GAS biofilms within the host. A more recent clinical study on GAS NSTIs presented a patient with a persistent infection over a period of 24 days. The surgeons reported a thick layer of biofilm covering fasciotomy wounds (Siemens et al. 2016). Analyses of biopsies collected from different anatomical locations of 31 GAS NSTI patients revealed that in 32% of cases, aggregations consistent with biofilm communities were evident. CLSM and scanning electron microscopy (SEM) analyses confirmed the presence of amorphous or fibrous biofilm structures consisting of bacteria, polysaccharides, lipids, and DNA. Biofilmpositive biopsies were characterized by a higher bacterial load. In addition, higher neutrophil influx and elevated pro-inflammatory responses were noted in biofilm-positive tissue (Siemens et al. 2016). Classical mechanistic studies on biofilm formation are confined to coated and uncoated plastic or glass surfaces. These studies showed that (1) not all clinically relevant GAS serotypes form biofilms (Lembke et al. 2006), (2) biofilm

N. Siemens et al.

forming abilities within and between the serotypes may vary (Conley et al. 2003), and (3) GAS biofilms are up to ten times more resistant to antibiotics as compared to planktonic cells (Ogawa et al. 2011). In addition, several GAS virulence factors have been implicated in biofilm development. Among these, surface localized factors, including M proteins, pili, and LTA, were associated with initial attachment of bacteria to different surfaces. Lack or diminished abundance of such factors resulted in reduced biofilm formation (Fiedler et al. 2015). Furthermore, capsule and SpeB expression were implicated in biofilm formation. However, reports on these two factors are contradicting. While some authors report a positive correlation, others report a negative impact (Marks et al. 2014; Roberts et al. 2010). Therefore, a recent study underlines the importance of studying biofilm in a physiologically relevant setting (Siemens et al. 2016). Although not all tested GAS serotypes were able to form biofilms on abiotic surfaces, all of them readily formed biofilms in a tissue setting. Furthermore, Nra regulator, which is encoded within the FCT-region (fibronectin-binding, collagenbinding, T-pilus) and controls a variety of surface localized factors including pili was implicated in this process. A nra-negative mutant lost its ability to form biofilm. In contrast to previous findings, no associations between SpeB, capsule, CovR/S, M protein, and Mga regulator and biofilm formation were found (Siemens et al. 2016). In support of the importance to study biofilms in in vivo settings, Vajjala and colleagues have demonstrated biofilm formation in a mice NSTI model (Vajjala et al. 2019). Both streptolysins, SLS and SLO, which were previously excluded as biofilm-contributing factors, were implicated in this process. While GAS wild type strain formed biofilms across the entire fascia already 12 h post infection, GAS mutant lacking both streptolysins did not. The authors demonstrated that streptolysins induce significant up-regulation of endoplasmic reticulum (ER) stress which in turn activates the unfolded protein response causing a release of host factors that support biofilm formation. Treatment of mice with pharmacological inducer of ER stress restored the ability of

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections

streptolysin-null mutant to form biofilm (Vajjala et al. 2019). These two studies underline the importance of studying NSTIs in a complex setting since simplified experimental in vitro approaches did not revealed biofilms as clinical feature of NSTIs. Such findings have direct implications for antibiotic efficacy in the tissue setting and call for reevaluation of treatment protocols in which biofilm is considered. Although in vitro assays in monolayer cultures and the use of mice as an in vivo model of choice contribute greatly to our understanding of infectious diseases, human tissue biopsies and human 3D-organotypic tissue models are the most valuable tools to study host-pathogen interactions. As described above, human NSTI tissue biopsies provided valuable insight in host-pathogen interactions at the tissue site. However, NSTIs are rare acute and devastating conditions, making it difficult to obtain patients consent for, e.g., sampling infected tissue or blood. Therefore, 3D-organotypic models might serve as a useful tool to study these infections in a more complex setting. Tissue engineering approaches originated in the field of regenerative medicine (Langer and Vacanti 1993). In contrast to standard monolayer cell cultures, tissue models much more closely resemble the architecture, cellular composition, and matrix complexity of the respective organ (Fig. 9.3a). In recent years 3D-organotypic models were successfully employed in a number of studies of human diseases involving the tissue setting (Svensson and Chen 2018). These studies include, but are not limited to, GAS, group G streptococcus or staphylococcal skin/soft tissue infections in human skin tissue model setting (Mairpady Shambat et al. 2016; Siemens et al. 2015, 2016). The use of such organotypic tissue models suggests a great potential of this approach in infectious diseases. Considering the fact that GAS are exclusively human pathogens, it becomes convincing to use a complex humanized model system. The engineered tissue recapitulates key anatomical and functional features, including a dermal layer consisting of fibroblasts and stratified epidermis with stratum corneum (Fig. 9.3a). The tissue produces important structural frame work proteins as well as the tight

141

junction and adherence junction proteins (Fig. 9.3a). It was shown that GAS are able to initiate tissue infection in this model system (Fig. 9.3b). While at 8 h post infection most of the bacteria were predominantly found with the stratum corneum, at later time points, bacteria disseminated throughout the entire tissue and formed biofilms (Fig. 9.3c) (Siemens et al. 2016). 3D tissue models are also suitable for implanting and studying immune cells, including dendritic cells, monocytes, macrophages, and even peripheral blood mononuclear cells (Mairpady Shambat et al. 2015; Nguyen Hoang et al. 2012). Such studies highlight a significant progress in the field of infectious diseases and contribute to the three R principle (replacement, reduction, refinement) of animal experimentation (Russell 1995).

9.6

Future Perspectives

The pathogenesis of NSTIs is a multifactorial process which involves both, bacterial and hostderived factors. These infections are hyperinflammatory conditions and are characterized by a massive immune cell infiltration and high bacterial load. GAS have evolved various immune evasion mechanisms, including intracellular survival and replication within phagocytes, egress out of the cells, and biofilm formation in the tissue. All these processes contribute greatly to the bacterial persistence and as a result, the antibiotic therapy fails to clear the infection. Future studies should aim to improve therapeutic strategies targeting intracellular bacterial reservoir and biofilm formation. Furthermore, early diagnosis of GAS NSTIs might improve the outcome for the patients. Therefore, the results from the INFECT study, including anatomical location and IL-1β, CXCL9, CXCL10, and CXCL11 as potential promising biomarker candidates, should be validated in a bigger cohort which encompass clinical samples of patients with different infections and other critical illnesses. Early diagnosis and subsequent early treatment might reduce local and systemic hyper-inflammation and will most likely further improve the outcome.

Fig. 9.3 GAS infections of the 3D-organotypic skin tissue models. (a) Comparative microscopic analyses of the human skin (left) and 3D-organotypic skin tissue model (right). Hematoxylin and eosin staining of human skin and tissue model (upper panel) and images from immuno-fluorescence analyses of key anatomical proteins (lower

panels) are shown. (b) CLSM analyses of GAS within different epidermal and dermal compartments of the 3D-organotypic model after 24 h of infection. (c) Comparative CLSM analyses of GAS biofilms on glass surfaces (right panel) and within the skin tissue model (left panel)

142 N. Siemens et al.

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections

Acknowledgments Financial support: The work was supported by the European Union Seventh Framework Programme (FP7/2007–2013) under the grant agreement 305340 (INFECT project), the Swedish Governmental Agency for Innovation Systems (VINNOVA) under the frame of NordForsk (Project no. 90456, PerAID), the Swedish Research Council under the frame of ERA PerMed (Project 2018-151, PerMIT), the German Research Foundation (DFG; grant no. 407176682), and the Federal Excellence Initiative of Mecklenburg Western Pomerania and European Social Fund (ESF) grant KoInfekt (ESF_14-BM-A55-0001_16).

References Akesson P, Sjoholm AG, Bjorck L (1996) Protein SIC, a novel extracellular protein of Streptococcus pyogenes interfering with complement function. J Biol Chem 271:1081–1088 Akesson P, Herwald H, Rasmussen M, Hakansson K, Abrahamson M, Hasan AA, Schmaier AH, MullerEsterl W, Bjorck L (2010) Streptococcal inhibitor of complement-mediated lysis (SIC): an antiinflammatory virulence determinant. Microbiology 156:3660–3668 Akiyama H, Morizane S, Yamasaki O, Oono T, Iwatsuki K (2003) Assessment of Streptococcus pyogenes microcolony formation in infected skin by confocal laser scanning microscopy. J Dermatol Sci 32:193–199 Anaya DA, Mcmahon K, Nathens AB, Sullivan SR, Foy H, Bulger E (2005) Predictors of mortality and limb loss in necrotizing soft tissue infections. Arch Surg 140:151–157. discussion 158 Arad G, Levy R, Nasie I, Hillman D, Rotfogel Z, Barash U, Supper E, Shpilka T, Minis A, Kaempfer R (2011) Binding of superantigen toxins into the CD28 homodimer interface is essential for induction of cytokine genes that mediate lethal shock. PLoS Biol 9: e1001149 Ashbaugh CD, Wessels MR (2001) Absence of a cysteine protease effect on bacterial virulence in two murine models of human invasive group A streptococcal infection. Infect Immun 69:6683–6688 Ashbaugh CD, Warren HB, Carey VJ, Wessels MR (1998) Molecular analysis of the role of the group A streptococcal cysteine protease, hyaluronic acid capsule, and M protein in a murine model of human invasive softtissue infection. J Clin Invest 102:550–560 Aziz RK, Ismail SA, Park HW, Kotb M (2004a) Postproteomic identification of a novel phage-encoded streptodornase, Sda1, in invasive M1T1 Streptococcus pyogenes. Mol Microbiol 54:184–197 Aziz RK, Pabst MJ, Jeng A, Kansal R, Low DE, Nizet V, Kotb M (2004b) Invasive M1T1 group A Streptococcus undergoes a phase-shift in vivo to prevent proteolytic degradation of multiple virulence factors by SpeB. Mol Microbiol 51:123–134

143

Bakleh M, Wold LE, Mandrekar JN, Harmsen WS, Dimashkieh HH, Baddour LM (2005) Correlation of histopathologic findings with clinical outcome in necrotizing fasciitis. Clin Infect Dis 40:410–414 Barnett TC, Cole JN, Rivera-Hernandez T, Henningham A, Paton JC, Nizet V, Walker MJ (2015) Streptococcal toxins: role in pathogenesis and disease. Cell Microbiol 17:1721–1741 Bastiat-Sempe B, Love JF, Lomayesva N, Wessels MR (2014) Streptolysin O and NAD-glycohydrolase prevent phagolysosome acidification and promote group A Streptococcus survival in macrophages. MBio 5: e01690–e01614 Beall B, Facklam R, Thompson T (1996) Sequencing emm-specific PCR products for routine and accurate typing of group A streptococci. J Clin Microbiol 34:953–958 Beisswenger C, Bals R (2005) Functions of antimicrobial peptides in host defense and immunity. Curr Protein Pept Sci 6:255–264 Bessen DE (2009) Population biology of the human restricted pathogen, Streptococcus pyogenes. Infect Genet Evol 9:581–593 Bessen DE, Mcshan WM, Nguyen SV, Shetty A, Agrawal S, Tettelin H (2015) Molecular epidemiology and genomics of group A Streptococcus. Infect Genet Evol 33:393–418 Bjarnsholt T, Buhlin K, Dufrene YF, Gomelsky M, Moroni A, Ramstedt M, Rumbaugh KP, Schulte T, Sun L, Akerlund B, Romling U (2018) Biofilm formation - what we can learn from recent developments. J Intern Med 284:332–345 Borregaard N, Sorensen OE, Theilgaard-Monch K (2007) Neutrophil granules: a library of innate immunity proteins. Trends Immunol 28:340–345 Boxrud PD, Verhamme IM, Bock PE (2004) Resolution of conformational activation in the kinetic mechanism of plasminogen activation by streptokinase. J Biol Chem 279:36633–36641 Bruun T, Rath E, Bruun Madsen M, Oppegaard O, Nekludov M, Arnell P, Karlsson Y, Babbar A, Bergey F, Itzek A, Hyldegaard O, Norrby-Teglund A, Skrede S, INFECT Study Group (2020) Risk factors and predictors of mortality in streptococcal necrotizing soft-tissue infections: a multicenter prospective study. Clin Infect Dis. https://doi.org/10.1093/cid/ciaa027 Chandrahas V, Glinton K, Liang Z, Donahue DL, Ploplis VA, Castellino FJ (2015) Direct host plasminogen binding to bacterial surface M-protein in pattern D strains of Streptococcus pyogenes is required for activation by its natural coinherited SK2b protein. J Biol Chem 290:18833–18842 Chandrasekaran S, Caparon MG (2016) The NADasenegative variant of the Streptococcus pyogenes toxin NAD(+) glycohydrolase induces JNK1-mediated programmed cellular necrosis. MBio 7:e02215–e02215 Chatila T, Geha RS (1993) Signal transduction by microbial superantigens via MHC class II molecules. Immunol Rev 131:43–59

144 Chella Krishnan K, Mukundan S, Alagarsamy J, Hur J, Nookala S, Siemens N, Svensson M, Hyldegaard O, Norrby-Teglund A, Kotb M (2016) Genetic architecture of group A streptococcal necrotizing soft tissue infections in the mouse. PLoS Pathog 12:e1005732 Chizzolini C, Chicheportiche R, Burger D, Dayer JM (1997) Human Th1 cells preferentially induce interleukin (IL)-1beta while Th2 cells induce IL-1 receptor antagonist production upon cell/cell contact with monocytes. Eur J Immunol 27:171–177 Cole JN, Mcarthur JD, Mckay FC, Sanderson-Smith ML, Cork AJ, Ranson M, Rohde M, Itzek A, Sun H, Ginsburg D, Kotb M, Nizet V, Chhatwal GS, Walker MJ (2006) Trigger for group A streptococcal M1T1 invasive disease. FASEB J 20:1745–1747 Cole JN, Barnett TC, Nizet V, Walker MJ (2011) Molecular insight into invasive group A streptococcal disease. Nat Rev Microbiol 9:724–736 Collin M, Svensson MD, Sjoholm AG, Jensenius JC, Sjobring U, Olsen A (2002) EndoS and SpeB from Streptococcus pyogenes inhibit immunoglobulinmediated opsonophagocytosis. Infect Immun 70:6646–6651 Commons RJ, Smeesters PR, Proft T, Fraser JD, RobinsBrowne R, Curtis N (2014) Streptococcal superantigens: categorization and clinical associations. Trends Mol Med 20:48–62 Conley J, Olson ME, Cook LS, Ceri H, Phan V, Davies HD (2003) Biofilm formation by group a streptococci: is there a relationship with treatment failure? J Clin Microbiol 41:4043–4048 Cortes G, Wessels MR (2009) Inhibition of dendritic cell maturation by group A Streptococcus. J Infect Dis 200:1152–1161 Courtney HS, Ofek I, Simpson WA, Hasty DL, Beachey EH (1986) Binding of Streptococcus pyogenes to soluble and insoluble fibronectin. Infect Immun 53:454–459 Courtney HS, Hasty DL, Dale JB (2002) Molecular mechanisms of adhesion, colonization, and invasion of group A streptococci. Ann Med 34:77–87 Cywes Bentley C, Hakansson A, Christianson J, Wessels MR (2005) Extracellular group A Streptococcus induces keratinocyte apoptosis by dysregulating calcium signalling. Cell Microbiol 7:945–955 Darenberg J, Luca-Harari B, Jasir A, Sandgren A, Pettersson H, Schalen C, Norgren M, Romanus V, Norrby-Teglund A, Normark BH (2007) Molecular and clinical characteristics of invasive group A streptococcal infection in Sweden. Clin Infect Dis 45:450–458 Davies HD, Mcgeer A, Schwartz B, Green K, Cann D, Simor AE, Low DE, Fletcher A, Kaul R, Scriver S, Willey B, Demers B, Gold W, Lovgren M, Talbot J, Naus M (1996) Invasive group a streptococcal infections in Ontario, Canada. N Engl J Med 335:547–554 Dombek PE, Cue D, Sedgewick J, Lam H, Ruschkowski S, Finlay BB, Cleary PP (1999) High-

N. Siemens et al. frequency intracellular invasion of epithelial cells by serotype M1 group A streptococci: M1 proteinmediated invasion and cytoskeletal rearrangements. Mol Microbiol 31:859–870 Edwards RJ, Taylor GW, Ferguson M, Murray S, Rendell N, Wrigley A, Bai Z, Boyle J, Finney SJ, Jones A, Russell HH, Turner C, Cohen J, Faulkner L, Sriskandan S (2005) Specific C-terminal cleavage and inactivation of interleukin-8 by invasive disease isolates of Streptococcus pyogenes. J Infect Dis 192:783–790 Egesten A, Olin AI, Linge HM, Yadav M, Morgelin M, Karlsson A, Collin M (2009) SpeB of Streptococcus pyogenes differentially modulates antibacterial and receptor activating properties of human chemokines. PLoS One 4:e4769 Elliott SD (1945) A proteolytic enzyme produced by group A Streptococci with special reference to its effect on the type-specific m antigen. J Exp Med 81:573–592 Emgard J, Bergsten H, Mccormick JK, Barrantes I, Skrede S, Sandberg JK, Norrby-Teglund A (2019) MAIT cells are major contributors to the cytokine response in group A Streptococcal toxic shock syndrome. Proc Natl Acad Sci U S A 116:25923–25931 Fernie-King BA, Seilly DJ, Willers C, Wurzner R, Davies A, Lachmann PJ (2001) Streptococcal inhibitor of complement (SIC) inhibits the membrane attack complex by preventing uptake of C567 onto cell membranes. Immunology 103:390–398 Fernie-King BA, Seilly DJ, Davies A, Lachmann PJ (2002) Streptococcal inhibitor of complement inhibits two additional components of the mucosal innate immune system: secretory leukocyte proteinase inhibitor and lysozyme. Infect Immun 70:4908–4916 Fernie-King BA, Seilly DJ, Lachmann PJ (2004) The interaction of streptococcal inhibitor of complement (SIC) and its proteolytic fragments with the human beta defensins. Immunology 111:444–452 Fiedler T, Koller T, Kreikemeyer B (2015) Streptococcus pyogenes biofilms-formation, biology, and clinical relevance. Front Cell Infect Microbiol 5:15 Flaherty RA, Donahue DL, Carothers KE, Ross JN, Ploplis VA, Castellino FJ, Lee SW (2018) Neutralization of streptolysin S-dependent and independent inflammatory cytokine IL-1beta activity reduces pathology during early GROUP A Streptococcal skin infection. Front Cell Infect Microbiol 8:211 Frick IM, Akesson P, Rasmussen M, Schmidtchen A, Bjorck L (2003a) SIC, a secreted protein of Streptococcus pyogenes that inactivates antibacterial peptides. J Biol Chem 278:16561–16566 Frick IM, Schmidtchen A, Sjobring U (2003b) Interactions between M proteins of Streptococcus pyogenes and glycosaminoglycans promote bacterial adhesion to host cells. Eur J Biochem 270:2303–2311 Frick IM, Shannon O, Akesson P, Morgelin M, Collin M, Schmidtchen A, Bjorck L (2011) Antibacterial activity of the contact and complement systems is blocked by

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections

SIC, a protein secreted by Streptococcus pyogenes. J Biol Chem 286:1331–1340 Gautam N, Olofsson AM, Herwald H, Iversen LF, Lundgren-Akerlund E, Hedqvist P, Arfors KE, Flodgaard H, Lindbom L (2001) Heparin-binding protein (HBP/CAP37): a missing link in neutrophilevoked alteration of vascular permeability. Nat Med 7:1123–1127 Goldmann O, Von Kockritz-Blickwede M, Holtje C, Chhatwal GS, Geffers R, Medina E (2007) Transcriptome analysis of murine macrophages in response to infection with Streptococcus pyogenes reveals an unusual activation program. Infect Immun 75:4148–4157 Gratz N, Siller M, Schaljo B, Pirzada ZA, Gattermeier I, Vojtek I, Kirschning CJ, Wagner H, Akira S, Charpentier E, Kovarik P (2008) Group A streptococcus activates type I interferon production and MyD88dependent signaling without involvement of TLR2, TLR4, and TLR9. J Biol Chem 283:19879–19887 Gryllos I, Tran-Winkler HJ, Cheng MF, Chung H, Bolcome R 3rd, Lu W, Lehrer RI, Wessels MR (2008) Induction of group A Streptococcus virulence by a human antimicrobial peptide. Proc Natl Acad Sci U S A 105:16755–16760 Gubba S, Low DE, Musser JM (1998) Expression and characterization of group A Streptococcus extracellular cysteine protease recombinant mutant proteins and documentation of seroconversion during human invasive disease episodes. Infect Immun 66:765–770 Hansen MB, Rasmussen LS, Svensson M, Chakrakodi B, Bruun T, Madsen MB, Perner A, Garred P, Hyldegaard O, Norrby-Teglund A, INFECT Study Group (2017) Association between cytokine response, the LRINEC score and outcome in patients with necrotising soft tissue infection: a multicentre, prospective study. Sci Rep 7:42179 Harbrecht BG, Nash NA (2016) Necrotizing soft tissue infections: a review. Surg Infect 17:503–509 Hertzen E, Johansson L, Wallin R, Schmidt H, Kroll M, Rehn AP, Kotb M, Morgelin M, Norrby-Teglund A (2010) M1 protein-dependent intracellular trafficking promotes persistence and replication of Streptococcus pyogenes in macrophages. J Innate Immun 2:534–545 Hertzen E, Johansson L, Kansal R, Hecht A, Dahesh S, Janos M, Nizet V, Kotb M, Norrby-Teglund A (2012) Intracellular streptococcus pyogenes in human macrophages display an altered gene expression profile. PLoS One 7(4):e35218 Herwald H, Cramer H, Morgelin M, Russell W, Sollenberg U, Norrby-Teglund A, Flodgaard H, Lindbom L, Bjorck L (2004) M protein, a classical bacterial virulence determinant, forms complexes with fibrinogen that induce vascular leakage. Cell 116:367–379 Hidalgo-Grass C, Dan-Goor M, Maly A, Eran Y, Kwinn LA, Nizet V, Ravins M, Jaffe J, Peyser A, Moses AE, Hanski E (2004) Effect of a bacterial pheromone peptide on host chemokine degradation in group A

145

streptococcal necrotising soft-tissue infections. Lancet 363:696–703 Hidalgo-Grass C, Mishalian I, Dan-Goor M, Belotserkovsky I, Eran Y, Nizet V, Peled A, Hanski E (2006) A streptococcal protease that degrades CXC chemokines and impairs bacterial clearance from infected tissues. EMBO J 25:4628–4637 Higashi DL, Biais N, Donahue DL, Mayfield JA, Tessier CR, Rodriguez K, Ashfeld BL, Luchetti J, Ploplis VA, Castellino FJ, Lee SW (2016) Activation of band 3 mediates group A Streptococcus streptolysin S-based beta-haemolysis. Nat Microbiol 1:15004 Hollands A, Aziz RK, Kansal R, Kotb M, Nizet V, Walker MJ (2008) A naturally occurring mutation in ropB suppresses SpeB expression and reduces M1T1 group A streptococcal systemic virulence. PLoS One 3:e4102 Hollands A, Gonzalez D, Leire E, Donald C, Gallo RL, Sanderson-Smith M, Dorrestein PC, Nizet V (2012) A bacterial pathogen co-opts host plasmin to resist killing by cathelicidin antimicrobial peptides. J Biol Chem 287:40891–40897 Holm SE, Norrby A, Bergholm AM, Norgren M (1992) Aspects of pathogenesis of serious group-A Streptococcal infections in Sweden, 1988-1989. J Infect Dis 166:31–37 Hsu LC, Enzler T, Seita J, Timmer AM, Lee CY, Lai TY, Yu GY, Lai LC, Temkin V, Sinzig U, Aung T, Nizet V, Weissman IL, Karin M (2011) IL-1 beta-driven neutrophilia preserves antibacterial defense in the absence of the kinase IKK beta. Nat Immunol 12:144–U54 Humar D, Datta V, Bast DJ, Beall B, De Azavedo JC, Nizet V (2002) Streptolysin S and necrotising infections produced by group G streptococcus. Lancet 359:124–129 Johansson L, Norrby-Teglund A (2013) Immunopathogenesis of streptococcal deep tissue infections. Curr Top Microbiol Immunol 368:173–188 Johansson L, Thulin P, Sendi P, Hertzen E, Linder A, Akesson P, Low DE, Agerberth B, Norrby-Teglund A (2008) Cathelicidin LL-37 in severe Streptococcus pyogenes soft tissue infections in humans. Infect Immun 76:3399–3404 Johansson L, Linner A, Sunden-Cullberg J, Haggar A, Herwald H, Lore K, Treutiger CJ, Norrby-Teglund A (2009) Neutrophil-derived hyperresistinemia in severe acute streptococcal infections. J Immunol 183:4047–4054 Johansson L, Thulin P, Low DE, Norrby-Teglund A (2010) Getting under the skin: the immunopathogenesis of Streptococcus pyogenes deep tissue infections. Clin Infect Dis 51:58–65 Johansson L, Snall J, Sendi P, Linner A, Thulin P, Linder A, Treutiger CJ, Norrby-Teglund A (2014) HMGB1 in severe soft tissue infections caused by Streptococcus pyogenes. Front Cell Infect Microbiol 4:4 Kachroo P, Eraso JM, Olsen RJ, Zhu L, Kubiak SL, Pruitt L, Yerramilli P, Cantu CC, Ojeda Saavedra M,

146 Pensar J, Corander J, Jenkins L, Kao L, Granillo A, Porter AR, Deleo FR, Musser JM (2020) New pathogenesis mechanisms and translational leads identified by multidimensional analysis of necrotizing myositis in primates. MBio 11(1):e03363-19 Kahn F, Morgelin M, Shannon O, Norrby-Teglund A, Herwald H, Olin AI, Bjorck L (2008) Antibodies against a surface protein of Streptococcus pyogenes promote a pathological inflammatory response. PLoS Pathog 4:e1000149 Kansal RG, Mcgeer A, Low DE, Norrby-Teglund A, Kotb M (2000) Inverse relation between disease severity and expression of the streptococcal cysteine protease, SpeB, among clonal M1T1 isolates recovered from invasive group A streptococcal infection cases. Infect Immun 68:6362–6369 Kapur V, Majesky MW, Li LL, Black RA, Musser JM (1993) Cleavage of interleukin 1 beta (IL-1 beta) precursor to produce active IL-1 beta by a conserved extracellular cysteine protease from Streptococcus pyogenes. Proc Natl Acad Sci U S A 90:7676–7680 Kasper KJ, Zeppa JJ, Wakabayashi AT, Xu SX, Mazzuca DM, Welch I, Baroja ML, Kotb M, Cairns E, Cleary PP, Haeryfar SM, Mccormick JK (2014) Bacterial superantigens promote acute nasopharyngeal infection by Streptococcus pyogenes in a human MHC class II-dependent manner. PLoS Pathog 10:e1004155 Kaul R, Mcgeer A, Low DE, Green K, Schwartz B (1997) Population-based surveillance for group A streptococcal necrotizing fasciitis: clinical features, prognostic indicators, and microbiologic analysis of seventyseven cases. Ontario Group A Streptococcal Study. Am J Med 103:18–24 Keller N, Woytschak J, Heeb LEM, Marques Maggio E, Mairpady Shambat S, Snall J, Hyldegaard O, Boyman O, Norrby-Teglund A, Zinkernagel AS (2019) Group A streptococcal DNase Sda1 impairs plasmacytoid dendritic cells’ type 1 interferon response. J Invest Dermatol 139:1284–1293 Keyel PA, Roth R, Yokoyama WM, Heuser JE, Salter RD (2013) Reduction of streptolysin O (SLO) poreforming activity enhances inflammasome activation. Toxins (Basel) 5:1105–1118 Khil J, Im M, Heath A, Ringdahl U, Mundada L, Cary Engleberg N, Fay WP (2003) Plasminogen enhances virulence of group A streptococci by streptokinasedependent and streptokinase-independent mechanisms. J Infect Dis 188:497–505 Kobayashi SD, Braughton KR, Whitney AR, Voyich JM, Schwan TG, Musser JM, Deleo FR (2003) Bacterial pathogens modulate an apoptosis differentiation program in human neutrophils. Proc Natl Acad Sci U S A 100:10948–10953 Kolaczkowska E, Kubes P (2013) Neutrophil recruitment and function in health and inflammation. Nat Rev Immunol 13:159–175 Kostakioti M, Hadjifrangiskou M, Hultgren SJ (2013) Bacterial biofilms: development, dispersal, and

N. Siemens et al. therapeutic strategies in the dawn of the postantibiotic era. Cold Spring Harb Perspect Med 3:a010306 Kotb M, Norrby-Teglund A, Mcgeer A, El-Sherbini H, Dorak MT, Khurshid A, Green K, Peeples J, Wade J, Thomson G, Schwartz B, Low DE (2002) An immunogenetic and molecular basis for differences in outcomes of invasive group A streptococcal infections. Nat Med 8:1398–1404 Kreikemeyer B, Klenk M, Podbielski A (2004) The intracellular status of Streptococcus pyogenes: role of extracellular matrix-binding proteins and their regulation. Int J Med Microbiol 294:177–188 Kuo CF, Wu JJ, Lin KY, Tsai PJ, Lee SC, Jin YT, Lei HY, Lin YS (1998) Role of streptococcal pyrogenic exotoxin B in the mouse model of group A streptococcal infection. Infect Immun 66:3931–3935 Kuo CF, Lin YS, Chuang WJ, Wu JJ, Tsao N (2008) Degradation of complement 3 by streptococcal pyrogenic exotoxin B inhibits complement activation and neutrophil opsonophagocytosis. Infect Immun 76:1163–1169 Lamagni TL, Darenberg J, Luca-Harari B, Siljander T, Efstratiou A, Henriques-Normark B, Vuopio-Varkila J, Bouvet A, Creti R, Ekelund K, Koliou M, Reinert RR, Stathi A, Strakova L, Ungureanu V, Schalen C, Jasir A, Grp S-ES (2008) Epidemiology of severe Streptococcus pyogenes disease in Europe. J Clin Microbiol 46:2359–2367 Langer R, Vacanti JP (1993) Tissue engineering. Science 260:920–926 Lapenta D, Rubens C, Chi E, Cleary PP (1994) Group A streptococci efficiently invade human respiratory epithelial cells. Proc Natl Acad Sci U S A 91:12115–12119 Larock CN, Todd J, Larock DL, Olson J, O’Donoghue AJ, Robertson AAB, Cooper MA, Hoffman HM, Nizet V (2016) IL-1beta is an innate immune sensor of microbial proteolysis. Sci Immunol 1:eaah3539 Lembke C, Podbielski A, Hidalgo-Grass C, Jonas L, Hanski E, Kreikemeyer B (2006) Characterization of biofilm formation by clinically relevant serotypes of group A streptococci. Appl Environ Microbiol 72:2864–2875 Levy R, Rotfogel Z, Hillman D, Popugailo A, Arad G, Supper E, Osman F, Kaempfer R (2016) Superantigens hyperinduce inflammatory cytokines by enhancing the B7-2/CD28 costimulatory receptor interaction. Proc Natl Acad Sci U S A 113:E6437–E6446 Liu TY, Elliott SD (1965) Streptococcal proteinase: the zymogen to enzyme transfromation. J Biol Chem 240:1138–1142 Llewelyn M, Cohen J (2002) Superantigens: microbial agents that corrupt immunity. Lancet Infect Dis 2:156–162 Loof TG, Goldmann O, Gessner A, Herwald H, Medina E (2010) Aberrant inflammatory response to Streptococcus pyogenes in mice lacking myeloid differentiation factor 88. Am J Pathol 176:754–763

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections

Low DE (2013) Toxic shock syndrome: major advances in pathogenesis, but not treatment. Crit Care Clin 29:651–675 Luca-Harari B, Darenberg J, Neal S, Siljander T, Strakova L, Tanna A, Creti R, Ekelund K, Koliou M, Tassios PT, Van Der Linden M, Straut M, VuopioVarkila J, Bouvet A, Efstratiou A, Schalen C, Henriques-Normark B, Strep ESG, Jasir A (2009) Clinical and microbiological characteristics of severe Streptococcus pyogenes disease in Europe. J Clin Microbiol 47:1155–1165 Lukomski S, Sreevatsan S, Amberg C, Reichardt W, Woischnik M, Podbielski A, Musser JM (1997) Inactivation of Streptococcus pyogenes extracellular cysteine protease significantly decreases mouse lethality of serotype M3 and M49 strains. J Clin Invest 99:2574–2580 Lukomski S, Burns EH Jr, Wyde PR, Podbielski A, Rurangirwa J, Moore-Poveda DK, Musser JM (1998) Genetic inactivation of an extracellular cysteine protease (SpeB) expressed by Streptococcus pyogenes decreases resistance to phagocytosis and dissemination to organs. Infect Immun 66:771–776 Lukomski S, Montgomery CA, Rurangirwa J, Geske RS, Barrish JP, Adams GJ, Musser JM (1999) Extracellular cysteine protease produced by Streptococcus pyogenes participates in the pathogenesis of invasive skin infection and dissemination in mice. Infect Immun 67:1779–1788 Madsen MB, Skrede S, Bruun T, Arnell P, Rosen A, Nekludov M, Karlsson Y, Bergey F, Saccenti E, Martins Dos Santos VAP, Perner A, Norrby-TeglundA, Hyldegaard O (2018) Necrotizing soft tissue infections - a multicentre, prospective observational study (INFECT): protocol and statistical analysis plan. Acta Anaesthesiol Scand 62:272–279 Madsen MB, Skrede S, Perner A, Arnell P, Nekludov M, Bruun T, Karlsson Y, Hansen MB, Polzik P, Hedetoft M, Rosen A, Saccenti E, Bergey F, Martins Dos Santos VAP, INFECT Study Group, NorrbyTeglund A, Hyldegaard O (2019) Patient’s characteristics and outcomes in necrotising soft-tissue infections: results from a Scandinavian, multicentre, prospective cohort study. Intensive Care Med 45:1241–1251 Mairpady Shambat S, Chen P, Nguyen Hoang AT, Bergsten H, Vandenesch F, Siemens N, Lina G, Monk IR, Foster TJ, Arakere G, Svensson M, Norrby-Teglund A (2015) Modelling staphylococcal pneumonia in a human 3D lung tissue model system delineates toxin-mediated pathology. Dis Model Mech 8:1413–1425 Mairpady Shambat S, Siemens N, Monk IR, Mohan DB, Mukundan S, Krishnan KC, Prabhakara S, Snall J, Kearns A, Vandenesch F, Svensson M, Kotb M, Gopal B, Arakere G, Norrby-Teglund A (2016) A point mutation in AgrC determines cytotoxic or colonizing properties associated with phenotypic variants of ST22 MRSA strains. Sci Rep 6:31360

147

Marks LR, Mashburn-Warren L, Federle MJ, Hakansson AP (2014) Streptococcus pyogenes biofilm growth in vitro and in vivo and its role in colonization, virulence, and genetic exchange. J Infect Dis 210:25–34 Mcdougald D, Rice SA, Barraud N, Steinberg PD, Kjelleberg S (2011) Should we stay or should we go: mechanisms and ecological consequences for biofilm dispersal. Nat Rev Microbiol 10:39–50 Mishalian I, Ordan M, Peled A, Maly A, Eichenbaum MB, Ravins M, Aychek T, Jung S, Hanski E (2011) Recruited macrophages control dissemination of group A Streptococcus from infected soft tissues. J Immunol 187:6022–6031 Morgan MS (2010) Diagnosis and management of necrotising fasciitis: a multiparametric approach. J Hosp Infect 75:249–257 Moses AE, Goldberg S, Korenman Z, Ravins M, Hanski E, Shapiro M, Grp I (2002) Invasive group A streptococcal infections, Israel. Emerg Infect Dis 8:421–426 Naegeli A, Bratanis E, Karlsson C, Shannon O, Kalluru R, Linder A, Malmstrom J, Collin M (2019) Streptococcus pyogenes evades adaptive immunity through specific IgG glycan hydrolysis. J Exp Med 216:1615–1629 Naseer U, Steinbakk M, Blystad H, Caugant DA (2016) Epidemiology of invasive group A streptococcal infections in Norway 2010-2014: a retrospective cohort study. Eur J Clin Microbiol Infect Dis 35:1639–1648 Nelson DC, Garbe J, Collin M (2011) Cysteine proteinase SpeB from Streptococcus pyogenes - a potent modifier of immunologically important host and bacterial proteins. Biol Chem 392:1077–1088 Nelson GE, Pondo T, Toews KA, Farley MM, Lindegren ML, Lynfield R, Aragon D, Zansky SM, Watt JP, Cieslak PR, Angeles K, Harrison LH, Petit S, Beall B, Van Beneden CA (2016) Epidemiology of invasive group A streptococcal infections in the United States, 2005-2012. Clin Infect Dis 63:478–486 Nguyen Hoang AT, Chen P, Juarez J, Sachamitr P, Billing B, Bosnjak L, Dahlen B, Coles M, Svensson M (2012) Dendritic cell functional properties in a three-dimensional tissue model of human lung mucosa. Am J Physiol Lung Cell Mol Physiol 302:L226–L237 Nilsson M, Sorensen OE, Morgelin M, Weineisen M, Sjobring U, Herwald H (2006) Activation of human polymorphonuclear neutrophils by streptolysin O from Streptococcus pyogenes leads to the release of proinflammatory mediators. Thromb Haemost 95:982–990 Nitzsche R, Kohler J, Kreikemeyer B, Oehmcke-Hecht S (2016) Streptococcus pyogenes escapes killing from extracellular histones through plasminogen binding and activation by streptokinase. J Innate Immun 8:589–600 Nizet V, Beall B, Bast DJ, Datta V, Kilburn L, Low DE, De Azavedo JC (2000) Genetic locus for streptolysin S

148 production by group A streptococcus. Infect Immun 68:4245–4254 Nizet V, Ohtake T, Lauth X, Trowbridge J, Rudisill J, Dorschner RA, Pestonjamasp V, Piraino J, Huttner K, Gallo RL (2001) Innate antimicrobial peptide protects the skin from invasive bacterial infection. Nature 414:454–457 Norrby-Teglund A, Thulin P, Gan BS, Kotb M, Mcgeer A, Andersson J, Low DE (2001) Evidence for superantigen involvement in severe group a streptococcal tissue infections. J Infect Dis 184:853–860 Nuwayhid ZB, Aronoff DM, Mulla ZD (2007) Blunt trauma as a risk factor for group A streptococcal necrotizing fasciitis. Ann Epidemiol 17:878–881 Nyberg P, Rasmussen M, Bjorck L (2004) alpha2-macroglobulin-proteinase complexes protect Streptococcus pyogenes from killing by the antimicrobial peptide LL-37. J Biol Chem 279:52820–52823 Ogawa T, Terao Y, Okuni H, Ninomiya K, Sakata H, Ikebe K, Maeda Y, Kawabata S (2011) Biofilm formation or internalization into epithelial cells enable Streptococcus pyogenes to evade antibiotic eradication in patients with pharyngitis. Microb Pathog 51:58–68 Okada N, Pentland AP, Falk P, Caparon MG (1994) M protein and protein F act as important determinants of cell-specific tropism of Streptococcus pyogenes in skin tissue. J Clin Invest 94:965–977 Okada N, Liszewski MK, Atkinson JP, Caparon M (1995) Membrane cofactor protein (CD46) is a keratinocyte receptor for the M protein of the group A streptococcus. Proc Natl Acad Sci U S A 92:2489–2493 Olsen RJ, Musser JM (2010) Molecular pathogenesis of necrotizing fasciitis. Annu Rev Pathol 5:1–31 Pinho-Ribeiro FA, Baddal B, Haarsma R, O’Seaghdha M, Yang NJ, Blake KJ, Portley M, Verri WA, Dale JB, Wessels MR, Chiu IM (2018) Blocking neuronal signaling to immune cells treats streptococcal invasive infection. Cell 173:1083 Reglinski M, Sriskandan S, Turner CE (2019) Identification of two new core chromosome-encoded superantigens in Streptococcus pyogenes; speQ and speR. J Infect 78:358–363 Roberts AL, Connolly KL, Doern CD, Holder RC, Reid SD (2010) Loss of the group A Streptococcus regulator Srv decreases biofilm formation in vivo in an otitis media model of infection. Infect Immun 78:4800–4808 Roberts AL, Connolly KL, Kirse DJ, Evans AK, Poehling KA, Peters TR, Reid SD (2012) Detection of group A Streptococcus in tonsils from pediatric patients reveals high rate of asymptomatic streptococcal carriage. BMC Pediatr 12:3 Rohde M, Cleary PP (2016) Adhesion and invasion of Streptococcus pyogenes into host cells and clinical relevance of intracellular streptococci. In: Ferretti JJ, Stevens DL, Fischetti VA (eds) Streptococcus pyogenes: basic biology to clinical manifestations. University of Oklahoma Health Sciences Center, Oklahoma City, OK

N. Siemens et al. Russell WM (1995) The development of the three Rs concept. Altern Lab Anim 23:298–304 Sanderson-Smith ML, Dinkla K, Cole JN, Cork AJ, Maamary PG, Mcarthur JD, Chhatwal GS, Walker MJ (2008) M protein-mediated plasminogen binding is essential for the virulence of an invasive Streptococcus pyogenes isolate. FASEB J 22:2715–2722 Siemens N, Patenge N, Otto J, Fiedler T, Kreikemeyer B (2011) Streptococcus pyogenes M49 plasminogen/ plasmin binding facilitates keratinocyte invasion via integrin-integrin-linked kinase (ILK) pathways and protects from macrophage killing. J Biol Chem 286:21612–21622 Siemens N, Kittang BR, Chakrakodi B, Oppegaard O, Johansson L, Bruun T, Mylvaganam H, INFECT Study Group, Svensson M, Skrede S, Norrby-Teglund A (2015) Increased cytotoxicity and streptolysin O activity in group G streptococcal strains causing invasive tissue infections. Sci Rep 5:16945 Siemens N, Chakrakodi B, Shambat SM, Morgan M, Bergsten H, Hyldegaard O, Skrede S, Arnell P, Madsen MB, Johansson L, INFECT Study Group, Juarez J, Bosnjak L, Morgelin M, Svensson M, Norrby-Teglund A (2016) Biofilm in group A streptococcal necrotizing soft tissue infections. JCI Insight 1: e87882 Sims Sanyahumbi A, Colquhoun S, Wyber R, Carapetis JR (2016) Global disease burden of group a streptococcus. In: Ferretti JJ, Stevens DL, Fischetti VA (eds) Streptococcus pyogenes: basic biology to clinical manifestations. University of Oklahoma Health Sciences Center, Oklahoma City, OK Singer M, Deutschman CS, Seymour CW, ShankarHari M, Annane D, Bauer M, Bellomo R, Bernard GR, Chiche JD, Coopersmith CM, Hotchkiss RS, Levy MM, Marshall JC, Martin GS, Opal SM, Rubenfeld GD, Van Der Poll T, Vincent JL, Angus DC (2016) The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA 315:801–810 Skinner M (2010) Autophagy: in the hands of HMGB1. Nat Rev Mol Cell Biol 11:756–757 Snall J, Linner A, Uhlmann J, Siemens N, Ibold H, Janos M, Linder A, Kreikemeyer B, Herwald H, Johansson L, Norrby-Teglund A (2016) Differential neutrophil responses to bacterial stimuli: streptococcal strains are potent inducers of heparin-binding protein and resistin-release. Sci Rep 6:21288 Soehnlein O, Oehmcke S, Ma X, Rothfuchs AG, Frithiof R, Van Rooijen N, Morgelin M, Herwald H, Lindbom L (2008) Neutrophil degranulation mediates severe lung damage triggered by streptococcal M1 protein. Eur Respir J 32:405–412 Stamenkovic I, Lew PD (1984) Early recognition of potentially fatal necrotizing fasciitis. The use of frozensection biopsy. N Engl J Med 310:1689–1693 Stevens DL (1999) The flesh-eating bacterium: what’s next? J Infect Dis 179(Suppl 2):S366–S374

9

Pathogenic Mechanisms of Streptococcal Necrotizing Soft Tissue Infections

Stevens DL, Bryant AE (2017) Necrotizing soft-tissue infections. N Engl J Med 377:2253–2265 Stevens DL, Tanner MH, Winship J, Swarts R, Ries KM, Schlievert PM, Kaplan E (1989) Severe group A streptococcal infections associated with a toxic shock-like syndrome and scarlet fever toxin Aa. N Engl J Med 321:1–7 Sumby P, Whitney AR, Graviss EA, Deleo FR, Musser JM (2006) Genome-wide analysis of group a streptococci reveals a mutation that modulates global phenotype and disease specificity. PLoS Pathog 2:e5 Sumby P, Zhang S, Whitney AR, Falugi F, Grandi G, Graviss EA, Deleo FR, Musser JM (2008) A chemokine-degrading extracellular protease made by group A Streptococcus alters pathogenesis by enhancing evasion of the innate immune response. Infect Immun 76:978–985 Sumitomo T, Nakata M, Higashino M, Jin Y, Terao Y, Fujinaga Y, Kawabata S (2011) Streptolysin S contributes to group A streptococcal translocation across an epithelial barrier. J Biol Chem 286:2750–2761 Sun H, Ringdahl U, Homeister JW, Fay WP, Engleberg NC, Yang AY, Rozek LS, Wang X, Sjobring U, Ginsburg D (2004) Plasminogen is a critical host pathogenicity factor for group A streptococcal infection. Science 305:1283–1286 Sunden-Cullberg J, Norrby-Teglund A, Rouhiainen A, Rauvala H, Herman G, Tracey KJ, Lee ML, Andersson J, Tokics L, Treutiger CJ (2005) Persistent elevation of high mobility group box-1 protein (HMGB1) in patients with severe sepsis and septic shock. Crit Care Med 33:564–573 Svensson M, Chen P (2018) Novel models to study stromal cell-leukocyte interactions in health and disease. Adv Exp Med Biol 1060:131–146 Thanert R, Itzek A, Hossmann J, Hamisch D, Madsen MB, Hyldegaard O, Skrede S, Bruun T, Norrby-Teglund A, INFECT Study Group, Medina E, Pieper DH (2019) Molecular profiling of tissue biopsies reveals unique signatures associated with streptococcal necrotizing soft tissue infections. Nat Commun 10:3846 Thulin P, Johansson L, Low DE, Gan BS, Kotb M, Mcgeer A, Norrby-Teglund A (2006) Viable group A streptococci in macrophages during acute soft tissue infection. PLoS Med 3:e53 Timmer AM, Timmer JC, Pence MA, Hsu LC, Ghochani M, Frey TG, Karin M, Salvesen GS, Nizet V (2009) Streptolysin O promotes group A Streptococcus immune evasion by accelerated macrophage apoptosis. J Biol Chem 284:862–871 Uchiyama S, Andreoni F, Schuepbach RA, Nizet V, Zinkernagel AS (2012) DNase Sda1 allows invasive M1T1 group A Streptococcus to prevent TLR9dependent recognition. PLoS Pathog 8:e1002736 Uchiyama S, Dohrmann S, Timmer AM, Dixit N, Ghochani M, Bhandari T, Timmer JC, Sprague K, Bubeck-Wardenburg J, Simon SI, Nizet V (2015) Streptolysin O rapidly impairs neutrophil oxidative

149

burst and antibacterial responses to group A Streptococcus. Front Immunol 6:581 Uhlmann J, Rohde M, Siemens N, Kreikemeyer B, Bergman P, Johansson L, Norrby-Teglund A (2016a) LL-37 triggers formation of Streptococcus pyogenes extracellular vesicle-like structures with immune stimulatory properties. J Innate Immun 8:243–257 Uhlmann J, Siemens N, Kai-Larsen Y, Fiedler T, Bergman P, Johansson L, Norrby-Teglund A (2016b) Phosphoglycerate kinase-a novel streptococcal factor involved in neutrophil activation and degranulation. J Infect Dis 214:1876–1883 Vajjala A, Biswas D, Tay WH, Hanski E, Kline KA (2019) Streptolysin-induced endoplasmic reticulum stress promotes group A Streptococcal host-associated biofilm formation and necrotising fasciitis. Cell Microbiol 21:e12956 Valderrama JA, Nizet V (2018) Group A Streptococcus encounters with host macrophages. Future Microbiol 13:119–134 Von Pawel-Rammingen U, Johansson BP, Bjorck L (2002) IdeS, a novel streptococcal cysteine proteinase with unique specificity for immunoglobulin G. EMBO J 21:1607–1615 Walker MJ, Hollands A, Sanderson-Smith ML, Cole JN, Kirk JK, Henningham A, Mcarthur JD, Dinkla K, Aziz RK, Kansal RG, Simpson AJ, Buchanan JT, Chhatwal GS, Kotb M, Nizet V (2007) DNase Sda1 provides selection pressure for a switch to invasive group A streptococcal infection. Nat Med 13:981–985 Westman J, Chakrakodi B, Snall J, Morgelin M, Bruun Madsen M, Hyldegaard O, Neumann A, Frick IM, Norrby-Teglund A, Bjorck L, Herwald H (2018) Protein SIC secreted from Streptococcus pyogenes forms complexes with extracellular histones that boost cytokine production. Front Immunol 9:236 Yoshino M, Murayama SY, Sunaoshi K, Wajima T, Takahashi M, Masaki J, Kurokawa I, Ubukata K (2010) Nonhemolytic Streptococcus pyogenes isolates that lack large regions of the sag operon mediating streptolysin S production. J Clin Microbiol 48:635–638 Yu CE, Ferretti JJ (1991) Frequency of the erythrogenic toxin-B and toxin-C genes (Speb and Spec) among clinical isolates of group-A Streptococci. Infect Immun 59:211–215 Zeppa JJ, Wakabayashi AT, Kasper KJ, Xu SX, Haeryfar SMM, Mccormick JK (2016) Nasopharyngeal infection of mice with Streptococcus pyogenes and in vivo detection of superantigen activity. Methods Mol Biol 1396:95–107 Zeppa JJ, Kasper KJ, Mohorovic I, Mazzuca DM, Haeryfar SMM, Mccormick JK (2017) Nasopharyngeal infection by Streptococcus pyogenes requires superantigen-responsive Vbeta-specific T cells. Proc Natl Acad Sci U S A 114:10226–10231 Zhao-Fleming HH, Wilkinson JE, Larumbe E, Dissanaike S, Rumbaugh K (2019) Obligate anaerobes are abundant in human necrotizing soft tissue infection

150 samples - a metagenomics analysis. APMIS 127:577–587 Zhu L, Olsen RJ, Lee JD, Porter AR, Deleo FR, Musser JM (2017) Contribution of secreted NADase and streptolysin O to the pathogenesis of epidemic serotype M1 Streptococcus pyogenes infections. Am J Pathol 187:605–613

N. Siemens et al. Zinkernagel AS, Timmer AM, Pence MA, Locke JB, Buchanan JT, Turner CE, Mishalian I, Sriskandan S, Hanski E, Nizet V (2008) The IL-8 protease SpyCEP/ScpC of group A Streptococcus promotes resistance to neutrophil killing. Cell Host Microbe 4:170–178

Systems Genetics Approaches in Mouse Models of Group A Streptococcal Necrotizing Soft-Tissue Infections

10

Suba Nookala, Karthickeyan Chella Krishnan, Santhosh Mukundan, and Malak Kotb

Contents 10.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152

10.2 10.2.1 10.2.2 10.2.3

Host Genetics and Group A Streptococcus Infections . . . . . . . . . . . . . . . . . . . . Systems Genetics Approaches Using BXD Recombinant Inbred Mouse Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ARI-BXD Mouse Models of GAS NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Assessing Heterogeneity of Treatment Effects in BXD Models of GAS NSTI

10.3

HLA-II Mouse Models of Invasive GAS Infections . . . . . . . . . . . . . . . . . . . . . . . 158

10.4

Future Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

153 154 155 158

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162

Abstract

Mouse models are invaluable resources for studying the pathogenesis and preclinical evaluation of therapeutics and vaccines against many human pathogens. Infections caused by group A streptococcus (GAS, Streptococcus pyogenes) are heterogeneous ranging from mild pharyngitis to severe invasive necrotizing fasciitis, a subgroup of necrotizing soft-tissue infections (NSTIs). While several strains of mice including BALB/c, C3H/HeN, CBA/J, S. Nookala (*) · S. Mukundan · M. Kotb Department of Biomedical Sciences, School of Medicine and Health Sciences, University of North Dakota, Grand Forks, ND, USA e-mail: [email protected] K. C. Krishnan Department of Medicine/Division of Cardiology, David Geffen School of Medicine, University of California, Los Angeles, CA, USA

and C57BL/10 offered significant insights, the human specificity and the interindividual variations on susceptibility or resistance to GAS infections limit their ability to mirror responses as seen in humans. In this chapter, we discuss the advanced recombinant inbred (ARI) BXD mouse model that mimics the genetic diversity as seen in humans and underpins the feasibility to map multiple genes (genetic loci) modulating GAS NSTI. GAS produces a myriad of virulence factors, including superantigens (SAg). Superantigens are potent immune toxins that activate T cells by cross-linking T cell receptors with human leukocyte antigen class-II (HLA-II) molecules expressed on antigen-presenting cells. This leads to a pro-inflammatory cytokine storm and the subsequent multiple organ damage and shock. Inbred mice are innately refractive to SAg-mediated responses. In this chapter, we

# Springer Nature Switzerland AG 2020 A. Norrby-Teglund et al. (eds.), Necrotizing Soft Tissue Infections, Advances in Experimental Medicine and Biology 1294, https://doi.org/10.1007/978-3-030-57616-5_10

151

152

discuss the versatility of the HLA-II transgenic mouse model that allowed the biological validation of known genetic associations to GAS NSTI. The combined utility of ARI-BXD and HLA-II mice as complementary approaches that offer clinically translatable insights into pathomechanisms driven by complex traits and host genetic context and novel means to evaluate the in vivo efficiency of therapies to improve outcomes of GAS NSTI are also discussed. Keywords

Streptococcus pyogenes · NSTI · ARI-BXD · IL-1β, HLA-II Highlights A systems biology approach using advanced intercross-derived recombinant inbred BXD (ARI-BXD) mouse models and an HLA-II transgenic mouse model to validate known genetic associations and superantigen-mediated responses reveals: • A significant quantitative trait locus (QTL) on Chromosome 2 with pathways involving IL1β as key upstream regulator mediating severity and survival in NSTI susceptible ARI-BXD mice. • Decreased survival was associated with a significant loss in the frequency of myeloidderived suppressor cells (MDSCs) in NSTI susceptible ARI-BXD strains, which was reversed with clindamycin treatment. • Induction of FoxP3-expressing T regulatory phenotype by the HLA-II mice expressing the protective allele in contrast to a Th1 phenotype with concomitant IFN-γ and IL-2 by mice expressing the HLA-II allele associated with neutral risk during GAS NSTI. • Transcriptome profile of GAS-infected skin from HLA-II mice reveals unique signaling pathways that are amenable to treatment with existing therapeutic approaches.

S. Nookala et al.

10.1

Introduction

Host genetic variations have been implicated as significant modulators of clinical presentation and disease severity in almost all infectious diseases and have been reported on extensively. The complex polymorphic and nonpolymorphic traits impact host immune responses and confer susceptibility to a specific infection while at the same time resistance to another type of infection. The host genetic background significantly influences not only the onset of the disease but also treatment outcomes (Skamene 1983; Kotb 2004; Frodsham and Hill 2004; Sadikot et al. 2005; Howard 2013; Asner et al. 2014; Patarčić et al. 2015; Çalişkan et al. 2015). Polymorphisms in HLA class II haplotypes, cytokine genes, and innate pathogen recognition receptors including TLR4 are among the key genetic determinants that affect host immune responses to invading pathogens. However, the complete spectrum of genetic determinants that influence susceptibility or resistance to infections remains unknown. The identification of these factors can be impactful in understanding the pathogenesis and inform management strategies. Furthermore, there is a growing concern and risk to human health with the increase in the emerging and reemerging infectious diseases, as well as natural or accidental biological threats (Fauci 2005). Therefore, recognizing emerging and reemerging infections and preparedness to biothreats through thorough knowledge and understanding of the mechanisms that govern the pathogenesis and associated host immune responses are key for substantially controlling disease outbreaks and the development of treatment possibilities. Despite significant variations in the microbial factors that are mostly enough to cause morbidity and mortality, host genetic factors have a major impact on disease susceptibility and outcomes (Hill 1999). Mouse models are cornerstones for studying human diseases and are indispensable for the prediction of outcomes of infections and diseases including genetic abnormalities, neurodegeneration, and cancer (Bedell et al. 1997). The knowledge gained from mouse

10

Systems Genetics Approaches in Mouse Models of Group A Streptococcal. . .

models as surrogates of human infections progresses to preclinical investigations that can be assessed in the same mouse models. Further, the successful use of mouse models to study human infections and test potential interventions is because of the genetic homology between mouse and human (Peltonen and McKusick 2001).

10.2

Host Genetics and Group A Streptococcus Infections

Infections caused by group A streptococcus (GAS, Streptococcus pyogenes) are an ideal model to explore the effects of interindividual variations on susceptibility or resistance. The pathophysiology of infections caused by GAS is heterogeneous. Some individuals experience only mild skin infections or sore throat, while in others, GAS infections can lead to severe, invasive, life-threatening complications such as streptococcal toxic shock syndrome (STSS) and necrotizing fasciitis (NF). In about 50% of the cases, NF, a subgroup of necrotizing soft-tissue infection (NSTI) can be complicated by STSS. (Stevens 2000; Carapetis et al. 2005; Ralph and Carapetis 2012). Despite early surgical debridement, an appropriate combination of penicillin and clindamycin, and adjunct therapies, an overarching pro-inflammatory cytokine storm and aggressive tissue destruction lead to significant morbidity and lethal outcomes with mortality as high as 25% (Reglinski and Sriskandan 2014; Barnett et al. 2019). The global burden of invasive GAS infections remains alarming with over 600,000 new cases each year and an estimated 160,000 deaths annually (Ermert et al. 2015; Sims Sanyahumbi et al. 2016). Antibiotics remain effective against GAS infections; however, antibiotic-resistant strains have been observed and present a potential threat to public health (Gilmer et al. 2013; Andreoni et al. 2017). Due to the extensive genomic diversity in GAS strains, there has been limited progress in GAS vaccine development. However, vaccine discovery efforts have increased more recently, and several candidates including peptide antigens have been

153

proposed for undergoing preclinical studies (Azuar et al. 2019; Vekemans et al. 2019). GAS produces a spectrum of potent virulence factors that are either secreted or surface associated that aid in their establishment, colonization, multiplication and dissemination, and to avoid recognition by the host immune system, antimicrobials, and antibiotics (Thulin et al. 2006; Siemens et al. 2016; Andreoni et al. 2017; Keller et al. 2018). Potential post-streptococcal autoimmune complications that arise from GAS mimics of host proteins include rheumatic heart disease, rheumatic heart fever, arthritis, glomerulonephritis, and Sydenham’s chorea (Dubost et al. 2004; Kim et al. 2004; Chandnani et al. 2015; Sims Sanyahumbi et al. 2016). Superantigens (SAgs) are the most important among the virulence factors secreted by GAS bacteria. SAgs, which are potent immune toxins, simultaneously cross-link T-cell receptor (TCR)bearing Vβ regions and conserved sequences of HLA class II expressed on antigen-presenting cells (APCs), thereby inducing the activation of both types of cells (Kotb 1998; Proft and Fraser 2003). The net result is the fulminant activation of the immune system and the massive release of pro-inflammatory mediators leading to septic shock and multiorgan failure with high mortality rates (Norrby-Teglund et al. 2000; Kotb et al. 2003). The Kotb research group, through their large epidemiological studies, demonstrated a significant association of certain HLA class II haplotypes with susceptibility or resistance to invasive GAS infections (Kotb et al. 2002, 2003). The haplotype DRB1*15/DQB1*06 (DR15/DQ6) conferred strong protection from severe systemic disease (SSD), the haplotype DRB1*14/DQB1*05 (DR14/DQ5) was associated with predisposition to SSD, and other haplotypes, e.g., DRB1*04/DQB1*0302 (DR4/DQ8), were neutral with respect to SSD (Kotb et al. 2002). Kotb et al. validated their findings in in vitro human cell cultures and demonstrated that the inflammatory responses to the M1T1 SAgs were much attenuated in the protective allele compared to high-risk or alleles associated with neutral outcomes of invasive GAS infections (Kotb et al. 2002).

154

10.2.1

S. Nookala et al.

Systems Genetics Approaches Using BXD Recombinant Inbred Mouse Models

It is only logical that HLA-II molecules are key determinants of pathological outcomes of GAS STSS and combined NF/STSS because GAS SAgs, pivotal mediators of systemic shock, simultaneously cross-link HLA-II molecules expressed on antigen-presenting cells (APCs) and TCRVβ elements and elicit pro-inflammatory cytokine storm leading to multiple organ damage and shock (Norrby-Teglund et al. 2001; Proft et al. 2003). More recently Giesbrecht et al. demonstrated that cross-linking of the SAg streptococcal pyrogenic exotoxin A (SpeA) to HLA-II alone can cause activation of the APCs independent of the responding T cells (Giesbrecht et al. 2019). However, in addition to SAgs, GAS produces a plethora of other virulence factors that might modulate the host immune responses besides involving the highly polymorphic HLA class II alleles. It is likely that polymorphisms in cytokine genes and innate pathogen recognition receptors including TLR4 affect host innate immune responses and might contribute to the severity and susceptibility to GAS infections. However, the complete spectrum of genetic determinants that influence susceptibility remains unknown. Identification of the key host factors will significantly impact the understanding of mechanisms underlying pathogenesis due to invasive infections caused by GAS, improve therapeutics, and management strategies. One way to accomplish is to perform a genomewide search for susceptibility loci in humans that requires the involvement of affected family members; while this may be possible, it is quite impractical and might be confounded by other variables. The influence of genetic background on susceptibility or resistance to invasive GAS infections has been reported using inbred mouse lines, including BALB/c, C3H/HeN, CBA/J, and C57BL/10 (Medina et al. 2001). While the use of these inbred lines offered insights, they lack the power to mirror the host responses as seen in

humans that are associated with complex polygenic traits due to their uniform background and limited genetic pool. To address this limitation, recombinant inbred (RI) mice, and particularly, the BXD family of RI mice, a murine reference population with high genetic diversity, was created using a breeding design originally constructed by Benjamin A. Taylor by crossing a female C57BL/6J (B6 or B) and a male DBA/2J (D2 or D) (BXD) and followed by at least 20 consecutive generations of sib-matings (Taylor et al. 1973) (Fig. 10.1). These BXD RI strains have been used extensively to study variation in pathogenicity and uncover genes modulating lethality (Benjamin et al. 1986; Melvold et al. 1990; Watters et al. 2001; Chesler et al. 2003). While there is extensive variation between each strain, there is virtually no genetic variation between individuals of one strain (Peirce et al. 2004). Thus, each BXD strain resembles a large human population of monozygotic twins. The most compelling feature of the BXD RI panel is that the genomes of the parental inbred strains (C57BL/6J and DBA/2J) have been fully sequenced, and they are a replicable, replenishable, and recapitulatable resource to map genetic loci of phenotypic traits as well as to study causal and mechanistic links of pathogenesis, treatment, prevention, and assessment of the countermeasures of global biothreats (Waterston et al. 2002). To further improve the power and precision, the breeding strategy of BXD lines was modified, and accordingly, the production of strains BXD43 through BXD102, started in the late 1990s, was derived from an advanced intercross (BXD-ARI) progeny that have been bred following a schema that incorporated roughly twice as many recombinations between the parental genomes B6 and D2 compared to the F2-derived BXDs (BXD-RI) and the ability to explore one or more quantitative trait loci (QTL) associated with a given disease or complex trait (Darvasi 1998; Williams et al. 2001; Chesler et al. 2003; Peirce et al. 2004; Parker et al. 2012; Pandey and Williams 2014).

10

Systems Genetics Approaches in Mouse Models of Group A Streptococcal. . .

155

Fig. 10.1 BXD strains are generated by crossing two fully inbred strains (C57BL/6J [female] and DBA/2J [male]), abbreviated as B6 and D2 or simply B and D, respectively. The resulting F1 generation (B6D2F1 or F1) is in turn crossed to produce F2 progeny. Systemic inbreeding is then performed beginning with the F2 generation for over 20 generations (recombinant inbred, RI) or after random mating for 8–14 generations of F2 before

inbreeding (advanced intercross-derived recombinant inbred, ARI) to allow random recombination of the chromosomes and generate progressively inbred BXD strains whose genomes exhibit unique B to D or D to B recombinations that are stably fixed. ARI mice increase the recombinations and diversity of the genetic pool (Peirce et al. 2004; Abdeltawab et al. 2008)

10.2.2

strain were more resistant than the male DBA/2J mice, suggesting the propensity of sex in addition to host genetics in driving susceptibility to GAS NSTIs (Krishnan et al. 2016). In order to identify the candidate genes involved in mediating differential susceptibility to GAS NSTIs, the Kotb research group capitalized on the BXD-ARI mice and established a mouse model of GAS NSTI (subcutaneous infections) with the hypervirulent M1T1 clinical isolate, GAS 5448 (originally isolated from a patient with STSS) (Chatellier et al. 2000). Using a large panel of 33 BXD strains along with the parental strains C57BL/6J and DBA/2J, and their F1 strains (B6D2F1), the GAS NSTI associated QTLs were identified through genome-wide linkage

ARI-BXD Mouse Models of GAS NSTI

In order to mimic the genetic diversity as seen in humans, the Kotb research group undertook a systems genetics approach to study the effect of multiple genes in modulating disease phenotypes to invasive GAS infections. Prior to establishing the BXD ARI mouse models of GAS NSTIs, the distinct role of host genetics and sex differences in driving susceptibility to GAS NSTIs were demonstrated in subcutaneous infections of conventional mouse strains C57BL/6J, DBA/2J, AJ, and CD-1 (Krishnan et al. 2016). Krishnan et al. reported that DBA/2J strain was more susceptible than C57BL/6J, and female mice of DBA/2J

156

S. Nookala et al.

Fig. 10.2 Swivel graph showing survival against NSTIs caused by GAS 5448 within the 33 BXD (black bars), B6 (green bar), D2 (red bar), and their F1 (blue bar) strains is expressed as mean values of corrected relative survival indices (cRSIs). Data is rank-ordered with positive indices

indicating increased survival and negative indices indicating decreased survival. Error bars indicate SEM. P-values were calculated by Generalized Linear Model (GLM) using ordinary least squares (OLS) ANOVA. Data from Chella Krishnan et al. (2016)

scans by mapping quantitative phenotypic traits to the BXD mouse genotype available from the gene network (Chella Krishnan et al. 2016). There were significant differences in the differential susceptibility to GAS NSTI among different BXDs, their parental, and F1 strains with BXD40, BXD101, and BXD64 at the susceptible end while BXD87, BXD85, and BXD73 strains were extremely resistant (Fig. 10.2). Several genomic loci proved significantly associated with survival, body weight change, and maximum lesion size due to GAS NSTI. The strongest and highly significant QTL modulating mouse survival against GAS NSTI mapped to mouse Chromosome (Chr) 2 between 24.5 and 35 Mb. The QTLs for body weight change kinetics

mapped to Chr 7 between 125 and 131 Mb, while the QTL modulating maximum lesion area was mapped to Chr 6 (131.6–141.8 Mb) and Chr 18 (49.5–56.3 Mb) (Fig. 10.3). Aziz et al. established the GAS sepsis model (intravenous infections) using the M1T1 clinical isolate, GAS5448, using B6 and D2 parents and a panel of 20 BXD strains, and reported the influence of genetics, sex, age, body weight, and inoculum titer on susceptibility and survival to GAS sepsis (Aziz et al. 2007). Their study revealed that bacteremia and dissemination to organs were major predictors of severity and mortality due to GAS sepsis, and preliminary mapping results suggested two significant QTLs on Chr 2 that needed to be verified and further narrowed down

10

Systems Genetics Approaches in Mouse Models of Group A Streptococcal. . .

157

Fig. 10.3 Genome-wide interval mapping for survival (a), body weight change (b), and maximum lesion size (c) revealed suggestive QTLs on mouse Chr 2, 7, and

6 and 18, respectively, suggesting a likely role for genes under these QTLs in modulating the respective responses during GAS NSTI (data from Chella Krishnan et al. 2016)

using additional BXD strains. By analyzing disease phenotypes in the context of BXD mice genotypes in a larger panel of BXD strains (32 strains including the parental strains, B6 and D2), Abdeltawab et al. not only confirmed the findings by Aziz et al. (2007) of a highly significant QTL to mouse Chr 2 modulating BXD

mouse GAS sepsis survival, but fine-mapped genes between 22 and 34 Mb and verified the differential expression of transcripts of key genes in uninfected or GAS-infected blood and spleens from highly resistant or susceptible BXD strains, at selected times post-infection (Abdeltawab et al. 2008). Further, Abdeltawab

158

S. Nookala et al.

et al. reported a second less significant QTL that was also mapped on the same chromosome between 125 and 150 Mb, and a third QTL on Chr X. The QTLs for bacteremia and bacterial dissemination to spleen overlapped with those for survival, with a slight difference in significance, with interleukin-1 alpha and prostaglandin E synthases pathways as key networks involved in modulating GAS sepsis (Abdeltawab et al. 2008). GAS sepsis and NSTI were independent studies with different routes of infection performed on BXD-ARI mice yet pointed out that the genetic susceptibility loci for survival in both cases was mapped to QTL on Chr 2. About 125 genes were significantly differentially expressed in the GAS NSTI susceptible strains (BXD 40 and 64) compared to uninfected controls. BXD strains that were resistant to GAS NSTIs were identified to be BXD 73 and BXD87. Parsing of the 125 differentially expressed genes identified two highscoring networks that likely mediate susceptibility to GAS NSTIs in susceptible hosts. Particularly, IL-1β was identified as the top upstream regulator likely mediating pro-inflammatory effects driving severity seen in susceptible hosts. The differential expression level of IL-1β transcripts was verified in GAS-infected skin tissue from susceptible mice and validated in tissue biopsies and plasma from GAS-infected patients (GAS 2006, INFECT Consortium) and in GAS-infected 3D reconstructed skin models (Chella Krishnan et al. 2016).

derived from studies on these mouse models are comparable to real-world data from diverse human populations. These models help tease out the mechanistic insights into disease and to assess the efficacy of potential therapeutics. For example, preliminary findings by Nookala et al. (manuscript in preparation) show that the survival and the frequency of CD45+, CD11b+ monocytes, and LY6G+LY6C+ myeloid-derived suppressor cells (MDSCs) are significantly lower in BXD40 (NSTI susceptible) compared to BXD87 (NSTI resistant) during subcutaneous GAS infections. However, this condition was reversed with clindamycin treatment (unpublished findings, Nookala et al., Fig. 10.4a, b). The accumulation of MDSCs is crucial for the host control of infections. MDSCs are a heterogeneous collection of monocytes and polymorphonuclear innate immune cells and function as initial effector cells to restrict infection burden and suppress the responses of activated effector T and NK-cells (Ost et al. 2016; Dorhoi and Plessis 2018). The understanding of the induction and frequency of MDSCs in GAS infections is just beginning to evolve. It is an attractive possibility that the loss or lack of induction of MDSCs could be one of the mechanisms against survival in GAS NSTI susceptible mice, which needs to be further investigated.

10.2.3

Invasive infections caused by GAS is one of the most complex human diseases affecting multiple organs driven by complex interactions that necessarily include HLA-II haplotype, immunological background, GAS burden, and the variations in the secreted and surface-bound virulence factors of the GAS pathogen. SAgs are among the most potent toxins secreted by GAS. While the BXD mice models described above are invaluable tools to study the effect of complex traits, conventional inbred mice are naturally resistant to SAg-mediated outcomes of GAS infections (Miethke et al. 1992; Blank et al. 1997; Sriskandan et al. 2001). These setbacks were

Assessing Heterogeneity of Treatment Effects in BXD Models of GAS NSTI

Invasive GAS infections have complex etiologies that vary depending on the host genetic background, environment, gender, and other as yet unidentified complex traits. The heterogeneity in the outcomes and the cause, therefore, have to be investigated to determine the best course of treatment and prevention. The BXD mice offer a promising model system to provide different information and insights into pathomechanisms underlying GAS infections. Importantly, data

10.3

HLA-II Mouse Models of Invasive GAS Infections

10

Systems Genetics Approaches in Mouse Models of Group A Streptococcal. . .

159

Fig. 10.4 Immunophenotype of myeloid-derived suppressor cells assessed by flow cytometry in blood from 10 days post GAS-infected BXD mice either susceptible or resistant to GAS NSTI. Data is represented as a t-distributed stochastic neighbor embedding (t-SNE) plot to aid the visualization of distribution and clustering of indicated subsets in untreated or clindamycin-treated

groups. t-SNE plot was run on ~5000 live cells discriminated by viability dye on singlets. Manually gated populations were overlaid onto t-SNE plots (a). Kaplan–Meier survival curves for the susceptible or resistant BXD mice 10 days post GAS infection that were either untreated or treated with clindamycin (b)

overcome through the use of transgenic mice expressing human plasminogen (Sun et al. 2004; Cole et al. 2006) and HLA-II alleles associated with differential susceptibility to invasive GAS infections (Nooh et al. 2007). Several studies have reported the use of HLA-II transgenic mice models to mimic responses to streptococcal and staphylococcal SAgs as seen in humans (Unnikrishnan et al. 2002; Welcher et al. 2002; DaSilva et al. 2002; Roy et al. 2005; Rajagopalan et al. 2006; Mangalam et al. 2008). The first evidence of the dominant and overriding effect of HLA-II molecules to GAS infection in vivo and SAg stimulations in vitro was reported by Nooh et al. (2007). Nooh et al. (2007) demonstrated that HLA-II mice expressing the neutral risk haplotype, DRB1*04/ DQB1*0302 (DR4/DQ8), mounted significantly

higher levels of the pro-inflammatory cytokines, TNF-α and IFN-γ, in vivo and in vitro compared to mice expressing DQ6, the DQ allele that is in linkage disequilibrium with the DR15 allele (DRB1*15/DQB1*06 (DR15/DQ6)). While studies by Nooh et al. (2007) used models of invasive GAS sepsis, Nookala et al. recently reported the use of HLA-II transgenic mice expressing either DRB1*1501 (DR15) or DRB1*04/DQB1*0302 (DR4/DQ8) as mouse models of GAS NSTI (Nookala et al. 2018). They demonstrated that HLA-II transgenic mice expressing the protective DR15 allele significantly induced surface GARP/ LAP expressing FoxP3+ T regulatory cells compared to mice expressing HLA-II DR4/DQ8, a haplotype associated with neutral risk for combined NF/STSS, which skewed the responses toward a Th1 type during in vivo infections with

160

S. Nookala et al.

Fig. 10.5 Proposed pathomechanisms underlying severe or nonsevere outcomes of GAS revealed from HLA-II transgenic mouse models of GAS NF/STSS. GAS superantigens simultaneously crosslink HLA-II and distinct variable beta regions in the TCR (TCR-Vβ), thereby eliciting robust levels of pro-inflammatory cytokines. Cell activation driven by the pro-inflammatory cytokine–receptor interactions induces phosphorylation of STAT molecules mediating the translocation of phosphorylated

STAT molecules into the nucleus resulting in the activation of transcription factors. The net effect is the polarization of naïve T cells into distinct subsets (Th1/Th17 in case of severe and Tregs in case of nonsevere). Data presented is a summary of analysis from transcriptomics data from in vitro and in vivo studies in HLA-II mouse models of GAS NF/STSS. Red arrows indicate increased expression of transcripts

a GAS 2006 (clinical isolate from a patient with combined NF/STSS, INFECT Consortium) and in vitro stimulations to SmeZ with the concomitant expression of IFN-γ and IL-2 (Nookala et al. 2018). The summary of various cytokines and transcription factors that are significantly modulated during SAg stimulations in vitro in splenocytes from HLA-II mice expressing risk (severe) or protective (nonsevere) alleles and

in vivo during GAS NSTI is shown in Fig. 10.5. Transcriptome analysis of GAS2006-infected skin from HLA-II DR4/DQ8 (a haplotype associated with neutral risk for combined NF/STSS) mice offered a plethora of information, specifically, clues for the existence of a combined adipogenic, metabolic, and inflammatory axis in GAS NSTI (Nookala et al. 2018). The top five most significantly downregulated pathways

10

Systems Genetics Approaches in Mouse Models of Group A Streptococcal. . .

161

Fig. 10.6 Heat map showing top 10 hits for canonical pathways with significant activated (red) and inhibited (green) z-score (A) associated with the differentially expressed genes in GAS-infected skin lesions from HLA-II mice expressing DR4/DQ8, a haplotype associated with neutral risk for combined GAS NF/STSS compared to PBS. A negative z-score indicates a predicted

inactivation, while a positive z-score indicates a predicted activation of the enriched pathway. These pathways predicted by the Ingenuity Pathway Analysis based on the gene expression changes seen in our dataset may be responsible for the pathophysiology underlying GAS NF/STSS

during GAS NSTI were oxidative phosphorylation, PPAR, TCA cycle II, PPARα/RXRα activation, and fatty acid β-oxidation 1, while the top five most significantly upregulated canonical signaling pathways were neuroinflammation, TREM1, IL-6, acute phase response, p38MAPK, and leukocyte extravasation (Nookala et al. manuscript in preparation). The top ten canonical pathways that are significantly enriched in HLA-II mice expressing the DR4/DQ8 haplotype are represented as a heat map in Fig. 10.6. Collectively, we provided strong evidence through a combination of systems genetics using a large panel of BXD mice and humanized mouse models of invasive GAS infections that host genetic variations contribute significantly through distinct molecular events involved in mediating susceptibility and clinical outcomes of invasive GAS infections. The systems genetics approach

using the ARI-BXD model provided the first evidence of the architecture of complex traits that were crucial in the understanding and the identification of genes, networks, and pathways underlying invasive GAS infections. The reductionist (HLA-II) approach in murine models of invasive GAS infections has contributed to unraveling the important role of HLA class II allelic polymorphisms leading to selective polarization into distinct Th1/Th17 or T-regulatory subsets during GAS NSTI, and the resulting pro-inflammatory or anti-inflammatory pathogenic processes they drive through their complex cross talk with the immune system. Although polymorphisms in HLA-II alleles are critical determinants of severity of GAS infections, systems genetics is becoming an important approach to understanding both the underlying biology and the distinct genetic loci, possible genes that modulate the particular traits, and the

162

S. Nookala et al.

mechanisms driven by complex traits in the genome.

10.4

Future Perspectives

There is an urgent need to reevaluate host responses to identify host-directed strategies for interference with various signaling pathways combined with standard antibiotic care to attenuate severity and improve outcomes of GAS NSTI. The identification of the interactions driven by the HLA-II haplotype as well as the complex traits of the genome is critical because therapeutic effectiveness is influenced by pathogen and hostrelated genetic and nongenetic factors. Over the last few years, the concept of “theranostics,” a convergence of therapy and diagnostics catered to patients based on their individual genetic characteristics (Degrauwe et al. 2019) is becoming increasingly popular. Given the disease heterogeneity, it is imperative to test whether or not a patient may benefit from therapy or vaccination. Due to the key role of HLA-II allelic polymorphisms in self and nonself-recognition, it is only logical that disease outcomes and therapy benefits show heterogeneity due to interindividual variation. Both BXD and HLA-II transgenic mouse models are invaluable tools to help uncover biomarkers, therapeutic targets, and vaccine candidates and test their efficacy for safe and personalized therapeutic and vaccination strategies. Most importantly, data from these studies will translate directly to humans, will enable the genetic evaluation of outcomes of infections, and also appears as a promising forthcoming revolutionary approach to patient-tailored bedside theranostics. Particularly important, yet poorly understood, is the host genetics-dependent onset and progression of metabolic dysfunctions and post-streptococcal sequelae, including rheumatic heart disease, rheumatic fever, arthritis, glomerulonephritis, and psychiatric conditions. Through carefully designed studies, the established murine models will facilitate the dissection of consequences due to long-term and recurrent GAS infections and complex host–pathogen interactions. Treatment or vaccination

strategies can then be personalized depending on the host genetic context that confer susceptibility or protection from particular disease in as much as customized interventions can produce more effective results than generalized therapeutic approaches. Acknowledgements The work presented was supported by the grants from the European Union (FP7/2012–2017) under the grant agreement 305340 (MK), grants from the Swedish Research Council under grant number 20150338 (MK), UND CoBRE Host–Pathogen Interactions Pilot Award (SN), and UND Genomics Core funded through grant support from the National Institutes of Health grants P20GM104360 (SN) and by the National Institute of General Medical Sciences of the National Institutes of Health under grant numbers P20GM103442, U54GM128729, and P20GM113123. We thank John Lee, Graphic Design Specialist, at Information Resources of the University of North Dakota School of Medicine & Health Sciences, for his assistance with the figure illustrations.

References Abdeltawab NF, Aziz RK, Kansal R, Rowe SL, Su Y, Gardner L, Brannen C, Nooh MM, Attia RR, Abdelsamed HA, Taylor WL, Lu L, Williams RW, Kotb M (2008) An unbiased systems genetics approach to mapping genetic loci modulating susceptibility to severe streptococcal sepsis. PLoS Pathog 4: e1000042. https://doi.org/10.1371/journal.ppat. 1000042 Andreoni F, Zörcher C, Tarnutzer A, Schilcher K, Neff A, Keller N, Maggio EM, Poyart C, Schuepbach RA, Zinkernagel AS, Zürcher C, Tarnutzer A, Schilcher K, Neff A, Keller N, Marques Maggio E, Poyart C, Schuepbach RA, Zinkernagel AS (2017) Clindamycin affects group a streptococcus virulence factors and improves clinical outcome. J Infect Dis 215:269–277. https://doi.org/10.1093/infdis/jiw229 Asner SA, Morré SA, Bochud PY, Greub G (2014) Host factors and genetic susceptibility to infections due to intracellular bacteria and fastidious organisms. Clin Microbiol Infect 20:1246–1253. https://doi.org/10. 1111/1469-0691.12806 Aziz RK, Kansal R, Abdeltawab NF, Rowe SL, Su Y, Carrigan D, Nooh MM, Attia RR, Brannen C, Gardner LA, Lu L, Williams RW, Kotb M (2007) Susceptibility to severe streptococcal sepsis: use of a large set of isogenic mouse lines to study genetic and environmental factors. Genes Immun 8:404–415. https://doi.org/ 10.1038/sj.gene.6364402 Azuar A, Jin W, Mukaida S, Hussein WM, Toth I, Skwarczynski M (2019) Recent advances in the development of peptide vaccines and their delivery systems

10

Systems Genetics Approaches in Mouse Models of Group A Streptococcal. . .

against group a streptococcus. Vaccine 7:58. https:// doi.org/10.3390/vaccines7030058 Barnett TC, Bowen AC, Carapetis JR (2019) The fall and rise of Group A Streptococcus diseases. Epidemiol Infect 147:e4. https://doi.org/10.1017/ S0950268818002285 Bedell MA, Jenkins NA, Copeland NG (1997) Mouse models of human disease. Part I: techniques and resources for genetic analysis in mice. Genes Dev 11:1–10. https://doi.org/10.1101/gad.11.1.1 Benjamin WH, Turnbough CL, Posey BS, Briles DE (1986) Salmonella typhimurium virulence genes necessary to exploit the Ity(s/s) genotype of the mouse. Infect Immun 51:872–878 Blank C, Luz A, Bendigs S, Erdmann A, Wagner H, Heeg K (1997) Superantigen and endotoxin synergize in the induction of lethal shock. Eur J Immunol 27:825–833. https://doi.org/10.1002/eji.1830270405 Çalişkan M, Baker SW, Gilad Y, Ober C (2015) Host Genetic Variation Influences Gene Expression Response to Rhinovirus Infection. PLoS Genet 11: e1005111. https://doi.org/10.1371/journal.pgen. 1005111 Carapetis JR, Steer AC, Mulholland EK, Weber M (2005) The global burden of group A streptococcal diseases. Lancet Infect Dis 5:685–694. https://doi.org/10.1016/ S1473-3099(05)70267-X Chandnani HK, Jain R, Patamasucon P (2015) Group C streptococcus causing rheumatic heart disease in a child. J Emerg Med 49:12–14. https://doi.org/10. 1016/j.jemermed.2014.12.057 Chatellier S, Ihendyane N, Kansal RG, Khambaty F, Basma H, Norrby-Teglund A, Low DE, Mcgeer A, Kotb M (2000) Genetic relatedness and superantigen expression in group A streptococcus serotype M1 isolates from patients with severe and nonsevere invasive diseases. Infect Immun 68:3523–3534. https://doi. org/10.1128/IAI.68.6.3523-3534.2000 Chella Krishnan K, Mukundan S, Alagarsamy J, Hur J, Nookala S, Siemens N, Svensson M, Hyldegaard O, Norrby-Teglund A, Kotb M (2016) Genetic architecture of Group A streptococcal necrotizing soft tissue infections in the mouse. PLoS Pathog 12:1–27. https:// doi.org/10.1371/journal.ppat.1005732 Chesler EJ, Wang J, Lu L, Qu Y, Manly KF, Williams RW (2003) Genetic correlates of gene expression in recombinant inbred strains: a relational model system to explore neurobehavioral phenotypes. Neuroinformatics 1:343–357. https://doi.org/10.1385/ NI:1:4:343 Cole JN, Mcarthur JD, Mckay FC, Sanderson-Smith ML, Cork AJ, Ranson M, Rohde M, Itzek A, Sun H, Ginsburg D, Kotb M, Nizet V, Chhatwal GS, Walker MJ (2006) Trigger for group A streptococcal M1T1 invasive disease. FASEB J 20:1745–1747. https://doi. org/10.1096/fj.06-5804fje Darvasi A (1998) Experimental strategies for the genetic dissection of complex traits in animal models. Nat Genet 18:19–24. https://doi.org/10.1038/ng0198-19

163

Dasilva L, Welcher BC, Ulrich RG, Aman MJ, David CS, Bavari S (2002) Humanlike immune response of human leukocyte antigen–DR3 transgenic mice to staphylococcal enterotoxins: a novel model for superantigen vaccines. J Infect Dis 185:1754–1760. https://doi.org/10.1086/340828 Degrauwe N, Hocquelet A, Digklia A, Schaefer N, Denys A, Duran R (2019) Theranostics in interventional oncology: versatile carriers for diagnosis and targeted image-guided minimally invasive procedures. Front Pharmacol 10:450. https://doi.org/10.3389/ fphar.2019.00450 Dorhoi A, Plessis ND (2018) Monocytic myeloid-derived suppressor cells in chronic infections. Front Immunol 8:1895. https://doi.org/10.3389/fimmu.2017.01895 Dubost JJ, Soubrier M, De Champs C, Ristori JM, Sauvezie B (2004) Streptococcal septic arthritis in adults. A study of 55 cases with a literature review. Joint Bone Spine 71(4):303–311. https://doi.org/10. 1016/S1297-319X(03)00122-2 Ermert D, Shaughnessy J, Joeris T, Kaplan J, Pang CJ, Kurt-Jones EA, Rice PA, Ram S, Blom AM (2015) Virulence of Group A Streptococci is enhanced by human complement inhibitors. PLoS Pathog 11: e1005043. https://doi.org/10.1371/journal.ppat. 1005043 Fauci AS (2005) Emerging and reemerging infectious diseases: the perpetual challenge. Acad Med 80:1079–1085. https://doi.org/10.1097/00001888200512000-00002 Frodsham AJ, Hill AVS (2004) Genetics of infectious diseases. Hum Mol Genet 13:R187–R194. https://doi. org/10.1093/hmg/ddh225 Giesbrecht K, Förmer S, Sähr A, Heeg K, Hildebrand D (2019) Streptococcal pyrogenic exotoxin A-stimulated monocytes mediate regulatory T-cell accumulation through PD-L1 and kynurenine. Int J Mol Sci 20 (16):3933. https://doi.org/10.3390/ijms20163933 Gilmer DB, Schmitz JE, Euler CW, Fischetti VA (2013) Novel bacteriophage lysin with broad lytic activity protects against mixed infection by streptococcus pyogenes and methicillin-resistant staphylococcus aureus. Antimicrob Agents Chemother 57:2743–2750. https://doi.org/10.1128/AAC. 02526-12 Hill AVS (1999) Genetics and genomics of infectious disease susceptibility. Br Med Bull 55:401–413. https://doi.org/10.1258/0007142991902457 Howard ST (2013) Recent progress towards understanding genetic variation in the Mycobacterium abscessus complex. Tuberculosis 93:S15–S20. https://doi.org/ 10.1016/S1472-9792(13)70005-2 Keller N, Andreoni F, Reiber C, Luethi-Schaller H, Schuepbach RA, Moch H, Marques Maggio E, Zinkernagel AS (2018) Human Streptococcal necrotizing fasciitis histopathology mirrored in a murine model. Am J Pathol 188:1517–1523. https://doi.org/ 10.1016/j.ajpath.2018.03.009

164 Kim SW, Grant JE, Kim SI, Swanson TA, Bernstein GA, Jaszcz WB, Williams KA, Schlievert PM (2004) A possible association of recurrent streptococcal infections and acute onset of obsessive-compulsive disorder. J Neuropsychiatry Clin Neurosci 16:252–260. https://doi.org/10.1176/jnp.16.3.252 Kotb M (1998) Superantigens of gram-positive bacteria: structure-function analyses and their implications for biological activity. Curr Opin Microbiol 1:56–65. https://doi.org/10.1016/S1369-5274(98)80143-4 Kotb M (2004) Genetics of susceptibility to infectious diseases. ASM News 70:457–463 Kotb M, Norrby-Teglund A, Mcgeer A, El-Sherbini H, Dorak MT, Khurshid A, Green K, Peeples J, Wade J, Thomson G, Schwartz B, Low DE (2002) An immunogenetic and molecular basis for differences in outcomes of invasive group A streptococcal infections. Nat Med 8:1398–1404. https://doi.org/10.1038/nm800 Kotb M, Norrby-Teglund A, Mcgeer A, Green K, Low DE (2003) Association of human leukocyte antigen with outcomes of infectious diseases: the streptococcal experience. Scand J Infect Dis 35:665–669. https:// doi.org/10.1080/00365540310015962 Krishnan KC, Mukundan S, Alagarsamy J, Laturnus D, Kotb M (2016) Host genetic variations and sex differences potentiate predisposition, severity, and outcomes of group A Streptococcus-mediated necrotizing soft tissue infections. Infect Immun 84:416–424. https://doi.org/10.1128/IAI.01191-15 Mangalam AK, Rajagopalan G, Taneja V, David CS (2008) HLA Class II transgenic mice mimic human inflammatory diseases. Adv Immunol 97:65–147. https://doi.org/10.1016/S0065-2776(08)00002-3 Medina E, Goldmann O, Rohde M, Lengeling A, Chhatwals GS (2001) Genetic control of susceptibility to group A streptococcal infection in mice. J Infect Dis 184:846–852. https://doi.org/10.1086/323292 Melvold RW, Jokinen DM, Miller SD, Dal Canto MC, Lipton HL (1990) Identification of a locus on mouse chromosome 3 involved in differential susceptibility to Theiler’s murine encephalomyelitis virus-induced demyelinating disease. J Virol 64:686–690 Miethke T, Wahl C, Heeg K, Echtenacher B, Krammer PH, Wagner H (1992) T cell-mediated lethal shock triggered in mice by the super-antigen staphylococcal enterotoxin B: critical role of tumor necrosis factor. J Exp Med 175:91–98. https://doi.org/10.1084/jem.175. 1.91 Nooh MM, El-Gengehi N, Kansal R, David CS, Kotb M (2007) HLA transgenic mice provide evidence for a direct and dominant Role of HLA class II variation in modulating the severity of streptococcal sepsis. J Immunol 178:3076–3083. https://doi.org/10.4049/ jimmunol.178.5.3076 Nookala S, Mukundan S, Fife A, Alagarsamy J, Kotb M (2018) Heterogeneity in FoxP3- and GARP/LAPexpressing T regulatory cells in an HLA class II transgenic murine model of necrotizing soft tissue

S. Nookala et al. infections by Group A Streptococcus. Infect Immun 86:12. https://doi.org/10.1128/IAI.00432-18 Norrby-Teglund A, Chatellier S, Low DE, Mcgeer A, Green K, Kotb M (2000) Host variation in cytokine responses to superantigens determine the severity of invasive group A streptococcal infection. Eur J Immunol 30:3247–3255. https://doi.org/10.1002/ 1521-4141(200011)30:113.0.CO;2-D Norrby-Teglund A, Thulin P, Gan BSS, Kotb M, Mcgeer A, Andersson J, Low DEE, Norrby-Teglund A, Thulin P, Gan BSS, Kotb M, Mcgeer A, Andersson J, Low DEE (2001) Evidence for superantigen involvement in severe group A streptococcal tissue infections. J Infect Dis 184:853–860. https://doi.org/10.1086/ 323443 Ost M, Singh A, Peschel A, Mehling R, Rieber N, Hartl D (2016) Myeloid-derived suppressor cells in bacterial infections. Front Cell Infect Microbiol 6:37. https:// doi.org/10.3389/fcimb.2016.00037 Pandey AK, Williams RW (2014) Genetics of gene expression in CNS. Int Rev Neurobiol 116:195–231. https://doi.org/10.1016/B978-0-12-801105-8.00008-4 Parker CC, Sokoloff G, Cheng R, Palmer AA (2012) Genome-wide association for fear conditioning in an advanced intercross mouse line. Behav Genet 42:437–448. https://doi.org/10.1007/s10519-0119524-8 Patarčić I, Gelemanović A, Kirin M, Kolčić I, Theodoratou E, Baillie KJ, De Jong MD, Rudan I, Campbell H, Polašek O (2015) The role of host genetic factors in respiratory tract infectious diseases: systematic review, meta-analyses and field synopsis. Sci Rep 5:116119. https://doi.org/10.1038/srep16119 Peirce JL, Lu L, Gu J, Silver LM, Williams RW (2004) A new set of BXD recombinant inbred lines from advanced intercross populations in mice. BMC Genet 179:1069–1078. https://doi.org/10.1186/1471-2156-57 Peltonen L, Mckusick VA (2001) Genomics and medicine: dissecting human disease in the postgenomic era. Science 291:1224–1229. https://doi.org/10.1126/science. 291.5507.1224 Proft T, Fraser JD (2003) Bacterial superantigens. Clin Exp Immunol 133:299–306. https://doi.org/10.1046/j. 1365-2249.2003.02203.x Proft T, Sriskandan S, Yang L, Fraser JD (2003) Superantigens and streptococcal toxic shock syndrome. Emerg Infect Dis 9:1211–1218. https://doi. org/10.3201/eid0910.030042 Rajagopalan G, Iijima K, Singh M, Kita H, Patel R, David CS (2006) Intranasal exposure to bacterial superantigens induces airway inflammation in HLA class II transgenic mice. Infect Immun 74:1284–1296. https://doi.org/10.1128/IAI.74.2.1284-1296.2006 Ralph AP, Carapetis JR (2012) Group A streptococcal diseases and their global burden. Curr Top Microbiol Immunol 368:1–27. https://doi.org/10.1007/82_2012_ 280

10

Systems Genetics Approaches in Mouse Models of Group A Streptococcal. . .

Reglinski M, Sriskandan S (2014) The contribution of group A streptococcal virulence determinants to the pathogenesis of sepsis. Virulence 5:1. https://doi.org/ 10.4161/viru.26400 Roy CJ, Warfield KL, Welcher BC, Gonzales RF, Larsen T, Hanson J, David CS, Krakauer T, Bavari S (2005) Human leukocyte antigen-DQ8 transgenic mice: a model to examine the toxicity of aerosolized staphylococcal enterotoxin B. Infect Immun 73:2452–2460. https://doi.org/10.1128/IAI.73.4.24522460.2005 Sadikot RT, Blackwell TS, Christman JW, Prince AS (2005) Pathogen-host interactions in pseudomonas aeruginosa pneumonia. Am J Respir Crit Care Med 171:1209–1223. https://doi.org/10.1164/rccm. 200408-1044SO Siemens N, Chakrakodi B, Shambat SM, Morgan M, Bergsten H, Hyldegaard O, Skrede S, Arnell P, Madsen MB, Johansson L, Juarez J, Bosnjak L, Mörgelin M, Svensson M, Norrby-Teglund A (2016) Biofilm in group A streptococcal necrotizing soft tissue infections. JCI Insight 1:1–13. https://doi.org/10.1172/ jci.insight.87882 Sims Sanyahumbi A, Colquhoun S, Wyber R, Carapetis JR (2016) Global disease burden of group A streptococcus. Streptococcus pyogenes : Basic Biol Clin Manifest 5:685–694. https://doi.org/10.1016/S14733099(05)70267-X Skamene E (1983) Genetic regulation of host resistance to bacterial infection. Rev Infect Dis 5:S823–S832. https://doi.org/10.1093/clinids/5.supplement_4.s823 Sriskandan S, Unnikrishnan M, Krausz T, Dewchand H, Van Noorden S, Cohen J, Altmann DM (2001) Enhanced susceptibility to superantigen-associated streptococcal sepsis in human Leukocyte antigen–DQ transgenic mice. J Infect Dis 184:166–173. https://doi. org/10.1086/322018 Stevens DL (2000) Streptococcal toxic shock syndrome associated with necrotizing fasciitis. Annu Rev Med 51:271–288. https://doi.org/10.1146/annurev.med.51. 1.271 Sun H, Ringdahl U, Momeister JW, Fay WP, Engleberg NC, Yang AY, Rozek LS, Wang X, Sjöbring U, Ginsburg D (2004) Plasminogen is a critical host pathogenicity factor for group A streptococcal infection. Science 305:1283–1286. https://doi.org/10.1126/sci ence.1101245 Taylor BA, Heiniger HJ, Meier H (1973) Genetic analysis of resistance to cadmium-induced testicular damage in mice. Proc Soc Exp Biol Med 143(3):629–633. https:// doi.org/10.3181/00379727-143-37380 Thulin P, Johansson L, Low DE, Gan BS, Kotb M, Mcgeer A, Norrby-Teglund A (2006) Viable group A streptococci in macrophages during acute soft tissue infection. PLoS Med 3:371–379. https://doi.org/10. 1371/JOURNAL.PMED.0030053 Unnikrishnan M, Altmann DM, Proft T, Wahid F, Cohen J, Fraser JD, Sriskandan S (2002) The bacterial superantigen streptococcal mitogenic exotoxin Z is the

165

major immunoactive agent of Streptococcus pyogenes. J Immunol 169:2561–2569. https://doi.org/10.4049/ jimmunol.169.5.2561 Vekemans J, Gouvea-Reis F, Kim JH, Excler JL, Smeesters PR, O’Brien KL, Van Beneden CA, Steer AC, Carapetis JR, Kaslow DC (2019) The path to group A Streptococcus vaccines: World Health Organization research and development technology roadmap and preferred product characteristics. Clin Infect Dis 69:877. https://doi.org/10.1093/cid/ciy1143 Waterston RH, Lindblad-Toh K, Birney E, Rogers J, Abril JF, Agarwal P, Agarwala R, Ainscough R, Alexandersson M, An P, Antonarakis SE, Attwood J, Baertsch R, Bailey J, Barlow K, Beck S, Berry E, Birren B, Bloom T, Bork P, Botcherby M, Bray N, Brent MR, Brown DG, Brown SD, Bult C, Burton J, Butler J, Campbell RD, Carninci P, Cawley S, Chiaromonte F, Chinwalla AT, Church DM, Clamp M, Clee C, Collins FS, Cook LL, Copley RR, Coulson A, Couronne O, Cuff J, Curwen V, Cutts T, Daly M, David R, Davies J, Delehaunty KD, Deri J, Dermitzakis ET, Dewey C, Dickens NJ, Diekhans M, Dodge S, Dubchak I, Dunn DM, Eddy SR, Elnitski L, Emes RD, Eswara P, Eyras E, Felsenfeld A, Fewell GA, Flicek P, Foley K, Frankel WN, Fulton LA, Fulton RS, Furey TS, Gage D, Gibbs RA, Glusman G, Gnerre S, Goldman N, Goodstadt L, Grafham D, Graves TA, Green ED, Gregory S, Guigó R, Guyer M, Hardison RC, Haussler D, Hayashizaki Y, Lahillier DW, Hinrichs A, Hlavina W, Holzer T, Hsu F, Hua A, Hubbard T, Hunt A, Jackson I, Jaffe DB, Johnson LS, Jones M, Jones TA, Joy A, Kamal M, Karlsson EK, Karolchik D, Kasprzyk A, Kawai J, Keibler E, Kells C, Kent WJ, Kirby A, Kolbe DL, Korf I, Kucherlapati RS, Kulbokas EJ, Kulp D, Landers T, Leger JP, Leonard S, Letunic I, Levine R, Li J, Li M, Lloyd C, Lucas S, Ma B, Maglott DR, Mardis ER, Matthews L, Mauceli E, Mayer JH, McCarthy M, McCombie WR, Mclaren S, Mclay K, McPherson JD, Meldrim J, Meredith B, Mesirov JP, Miller W, Miner TL, Mongin E, Montgomery KT, Morgan M, Mott R, Mullikin JC, Muzny DM, Nash WE, Nelson JO, Nhan MN, Nicol R, Ning Z, Nusbaum C, O’Connor MJ, Okazaki Y, Oliver K, Overton-Larty E, Pachter L, Parra G, Pepin KH, Peterson J, Pevzner P, Plumb R, Pohl CS, Poliakov A, Ponce TC, Ponting CP, Potter S, Quail M, Reymond A, Roe BA, Roskin KM, Rubin EM, Rust AG, Santos R, Sapojnikov V, Schultz B, Schultz J, Schwartz MS, Schwartz S, Scott C, Seaman S, Searle S, Sharpe T, Sheridan A, Shownkeen R, Sims S, Singer JB, Slater G, Smit A, Smith DR, Spencer B, Stabenau A, Stange-ThomannN, Sugnet C, Suyama M, Tesler G, Thompson J, Torrents D, Trevaskis E, Tromp J, Ucla C, UretaVidal A, Vinson JP, Von Niederhausern AC, Wade CM, Wall M, Weber RJ, Weiss RB, Wendl MC, West AP, Wetterstrand K, Wheeler R, Whelan S, Wierzbowski J, Willey D, Williams S, Wilson RK,

166 Winter E, Worley KC, Wyman D, Yang S, Yang SP, Zdobnov EM, Zody MC, Lander ES (2002) Initial sequencing and comparative analysis of the mouse genome. Nature 420:520–562. https://doi.org/10. 1038/nature01262 Watters JW, Dewar K, Lehoczky J, Boyartchuk V, Dietrich WF (2001) Kif1C, a kinesin-like motor protein, mediates mouse macrophage resistance to anthrax lethal factor. Curr Biol 11:1503–1511. https://doi.org/ 10.1016/S0960-9822(01)00476-6 Welcher BCC, Carra JHH, Dasilva L, Hanson J, David CSS, Aman MJJ, Bavari S (2002) Lethal shock

S. Nookala et al. induced by streptococcal pyrogenic exotoxin A in mice transgenic for human leukocyte antigen–DQ8 and human CD4 receptors: implications for development of vaccines and therapeutics. J Infect Dis 186:501–510. https://doi.org/10.1086/341828 Williams RW, Gu J, Qi S, Lu L (2001) The genetic structure of recombinant inbred mice: high-resolution consensus maps for complex trait analysis. Genome Biol 2:0046. https://doi.org/10.1186/gb-2001-2-11research0046

Systems Biology and Biomarkers in Necrotizing Soft Tissue Infections

11

Edoardo Saccenti and Mattias Svensson

Contents 11.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168

11.2

State of the Art: Existing Biomarkers for NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . 169

11.3 11.3.1 11.3.2 11.3.3 11.3.4

Biomarker Type and Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Diagnostic Biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Prognostic Biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Predictive Biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Predisposing Biomarkers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

169 170 170 171 171

11.4 11.4.1 11.4.2 11.4.3 11.4.4 11.4.5 11.4.6 11.4.7

Biomarker Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Genomic Biomarkers in NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transcriptomic Biomarkers in NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Epigenomic Biomarkers in NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Proteomic Biomarkers in NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Metabolomic Biomarkers in NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Microbiomic Biomarkers in NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . From One-Category-Analyses Towards Systems Biological Approaches . . .

172 172 172 173 173 174 174 174

11.5

Systems Biology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 175

11.6

Systems Biology Approaches to Biomarker Discovery . . . . . . . . . . . . . . . . . . . 176

11.7

Networks and Network Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177

11.8

Multivariate Biomarker and Omics Signatures . . . . . . . . . . . . . . . . . . . . . . . . . . . 180

11.9

Future Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184

E. Saccenti (*) Laboratory of Systems and Synthetic Biology, Wageningen University and Research, Wageningen, The Netherlands e-mail: [email protected] M. Svensson Center for Infectious Medicine, Department of Medicine, ANA Futura, Karolinska Institutet, Karolinska University Hospital, Huddinge, Sweden

Abstract

In necrotizing soft tissue infection (NSTI) there is a need to identify biomarker sets that can be used for diagnosis and disease management. The INFECT study was designed to obtain such insights through the integration of patient data and results from different

# Springer Nature Switzerland AG 2020 A. Norrby-Teglund et al. (eds.), Necrotizing Soft Tissue Infections, Advances in Experimental Medicine and Biology 1294, https://doi.org/10.1007/978-3-030-57616-5_11

167

168

E. Saccenti and M. Svensson

clinically relevant experimental models by use of systems biology approaches. This chapter describes the current state of biomarkers in NSTI and how biomarkers are categorized. We introduce the fundamentals of top-down systems biology approaches including analysis tools and we review the use of current methods and systems biology approaches to biomarker discover. Further, we discuss how different “omics” signatures (gene expression, protein, and metabolites) from NSTI patient samples can be used to identify key host and pathogen factors involved in the onset and development of infection, as well as exploring associations to disease outcomes. Keywords

Epigenomics · Genomics · Network inference and analysis · Mathematical modelling · Metabolomics · Microbiome · Multivariate statistics · Proteomics · Top-down approaches Highlights • Comprehensive systems biology analyses and interpretation of omics data are emerging as valuable tools for biomarker discovery in NSTI. • The inflammatory proteins IL1β, IL-6, CXCL9, CXCL10, and CXCL11, were identified as potential biomarkers to discriminate S. pyogenes NSTI from NSTI caused by other pathogens. • Metabolite–metabolite association networks analysis has proven able to define sets of metabolites associated with biofilm formation in NSTI. • Metabolites in combination with multivariate and machine learning modelling have the potential to be utilized as predictive markers in NSTI.

11.1

Introduction

In search for biomarkers in NSTI, analyses have usually focused specifically on measuring certain selected factors, such as cytokines. This approach is, to some extent, distorted because it is based on already existing knowledge that somehow

indicates that these factors must be associated with the disease. In this context, a systems biological approach, used correctly, has the greater opportunity to identify hitherto unknown factors that may be associated with disease and that may be used as biomarkers. Systems biology is a holistic approach to deciphering the complexity of biological systems that starts from the understanding that the networks that form the whole of living organisms are more than the sum of their parts. The systems biology approach to understand disease-related biology has the potential to revolutionize our understanding of the cellular pathways and gene networks underlying the onset of disease, and the mechanisms of treatment strategies that ameliorate disease phenotypes. In the post-genomic era, a wealth of novel approaches for generating and analyzing large, high-dimensional genomic, transcriptomic, proteomic, and metabolomic data sets from cohorts of normal and diseased individuals can be used in combination with systems biology approaches (Rosato et al. 2018). In research related to infectious diseases, such as NSTI, the data sets include information from both pathogen- and host-oriented analyses that can be used at a systems biology level. The emerging field of systems biology attempts to harness complex, multi-parameter systems by computationally integrating gene-level data with molecular pathways and networks to extract new biological insight. Systems biology may combine and augment current strategies to biomarker discovery, generating novel, experimentally testable candidates. A biomarker is a measurable indicator of some biological state or condition, and in recent years systems biology approaches have been applied to screen and identify diagnostic, prognostic, predictive, and predisposing biomarkers as well as targets for prevention and treatment of disease (Feala et al. 2013; Vafaee et al. 2018; Anvar et al. 2018). There are several successful examples of molecular biomarkers that are currently the clinical standard for diagnostic screening in several diseases, for example, in myocardial infarction (Hajar 2016) and certain cancers (Mehta et al. 2010; Goossens et al. 2015). The search for novel molecular biomarkers

11

Systems Biology and Biomarkers in Necrotizing Soft Tissue Infections

continues to be a major research drive in many biomedical fields. A PubMed (https://pubmed. ncbi.nlm.nih.gov/) query for NSTI and biomarker (Biomarker AND (NSTI OR “necrotising soft tissue infection” OR “necrotising fasciitis”) in March 2020 returned only 39 hits (44 using the spelling “necrotizing”) demonstrating how little is known and how limited the research activity on this topic is, despite an incidence ranging from 0.3 to 15.5 case per 100.000 population (Stevens and Bryant 2017) and all-cause 90-day mortality of 18% (Madsen et al. 2019). In NSTI biomarkers for early detection and guidance of treatment may improve outcome and help to reduce the length of hospital stay and mortality. There are good opportunities that systems biology approaches in NSTI can identify biomarkers that could be used as novel tools for rapid and early identification and stratification of patients with NSTI upon admission. In this book chapter, we review biomarker types and detection methods, as well as a number of systems biology approaches potentially useful to identify candidate biomarkers in NSTI.

11.2

State of the Art: Existing Biomarkers for NSTI

There are currently no approved biomarkers for the diagnosis or prognosis in NSTI, and the molecular mechanisms of host responses during NSTI remain poorly understood. This lack of understanding reflects the complex, multifactorial nature of NSTI, which is believed to involve a network of interweaving host and pathogen molecular pathways that dictate disease onset and progression. Available information on NSTI biomarkers are restricted to a few research studies that focused on a limited number of selected factors or information of factors included as components of relatively inaccurate scoring systems, such as the LRINEC (Laboratory Risk Indicator for Necrotizing Fasciitis). LRINEC is generated from six routinely performed laboratory tests including the analyses of patients´ C-reactive protein, white blood cell count (
10,000/μL), hemoglobin (g/dL), sodium (mEq/L), creatinine, and glucose.

169

Although LRINEC scoring is not optimal, it has been used to distinguish NSTI from less severe soft tissue infections (Wong et al. 2004). Recently, LRINEC score has also been analyzed in relation to pro-inflammatory cytokines, and this revealed no significant association between the LRINEC score and analyzed specific cytokine levels on admission (Hansen et al. 2017a). However, the analyses of cytokines revealed IL-6 to be associated with disease severity, and that IL-1β was strongly associated with 30-day mortality (Hansen et al. 2017b). Another factor suggested to have biomarker potential in NSTI is Pentraxin-3 (PTX3) a glycoprotein which belongs to the pentraxin protein superfamily, including acute phase proteins, C-reactive protein (CRP), and serum amyloid protein (SAP). PTX3 is produced by vascular cells or inflammatory cells and is released into the circulation, possibly reflecting systemic inflammation. The PTX3 level in patients with NSTI at time of admission was associated with septic shock, amputation, and risk of death in patients, but was not an independent predictor of 180-day mortality in this patient group (Hansen et al. 2016a). Levels of s-lactate >6 mM and creatinine >144 mM are considered prognostic markers. In addition, microbiological cultures of Clostridium or Vibrio are negatively associated, while a microbiological culture of GAS is positively associated with survival in NSTI. Nevertheless, there is a great need for robust early measurable biomarkers to improve identification and management of NSTI.

11.3

Biomarker Type and Detection

Molecular biomarkers generally consist of molecules measured in body fluids or in samples from the affected tissue. Challenges facing biomarker research include the lack of standardized methods for fast reliable detection. Biomarkers can be divided into four categories: diagnostic, prognostic, predictive, and predisposing/susceptibility (Quezada et al. 2017; Simon 2011, 2014; Drucker and Krapfenbauer 2013). The typical pattern of biomarker discovery is shown in Fig. 11.1.

170

E. Saccenti and M. Svensson

Fig. 11.1 Relationship between the classical biomarker discovery pipeline and the systems biology cycle. Systems biology can play a pivotal role in both discovery and analytical validations steps. Figure adapted from Del Campo et al. (2015) and Rosato et al. (2018)

11.3.1

Diagnostic Biomarkers

Diagnostic biomarkers are used, e.g. to determine the health status of a patient to ensure appropriate management. In infectious diseases such as NSTI and sepsis, immediate treatment is required, necessitating quick, early, and accurate diagnosis which can help decision making. A singular ideal biomarker may not be identified, but rather an alternative approach focusing to determine the diagnostic relevancy of multiple biomarkers when used in concert may be useful. The ongoing efforts in the development of a multiplex pointof-care testing kit, enabling quick and reliable detection of serum biomarkers, may have great potential for early diagnosis of acute bacterial infections. Thänert et al. (2019) provided insight into the pathophysiology of mono- and polymicrobial NSTIs and protein multiplex analysis highlighted three potential biomarkers, CXCL9, CXCL10, CXCL11, as these three

chemokines displayed statistically significant concentration differences between S. pyogenes and polymicrobial NSTIs. In addition, this study performed RNA-seq analyses on tissue biopsies and this showed that genes with significantly greater expression in polymicrobial compared to monomicrobial NSTIs were those encoding extra cellular matrix (ECM) components like collagen, fibronectin, lumican, and connective tissue growth factor. In addition, gene sequencing data showed, in line with protein data, that genes encoding interferon-inducible mediators such as CXCL9, CXCL10, CXCL11 were expressed at higher levels in monomicrobial NSTIs, in particular those caused by S. pyogenes.

11.3.2

Prognostic Biomarkers

A prognostic marker identifies outcome in patients. Usually, these markers provide insight

11

Systems Biology and Biomarkers in Necrotizing Soft Tissue Infections

into unfavorable outcome, such as mortality. The presence or absence of a prognostic marker can be useful for the selection of patients for a specific treatment but does not directly predict the response to a treatment. In NSTI plasma biomarkers representing the early inflammatory response are thought to be useful as prognostic markers of disease severity and mortality. To determine the association of admission biochemical markers, such as lactate, to limb loss and mortality both univariate and multivariable analyses have been used. For example, arterial lactate was found to be associated with both mortality (odds ratio [OR], 1.5; 95% confidence interval [CI], 1.1 to 2.0; Pvalue ¼ 0.009) and limb loss (OR, 1.3; 95% CI, 1.0 to 1.7; P-value ¼ 0.02), and lactate was suggested to be used early on to guide aggressive therapeutic interventions (Schwartz et al. 2013). In addition, attempts have been made to associate markers with ICU (Intensive Care Unit) scoring systems including simplified acute physiology score (SAPS) II (Le Gall et al. 1993) and sequential organ failure assessment (SOFA) scores (Vincent et al. 1996), as well as the LRINEC score (Wong et al. 2004), presence of septic shock, microbial etiology, renal replacement therapy, and amputation. In a study by Hansen et al. it was demonstrated that plasma levels of mannose-binding lectin and ficolins, which can activate the complement pathway, were significantly lower in patients with NSTI than in controls (Hansen et al. 2016b). Furthermore, this study revealed that a high baseline ficolin-2 level indicated a 94% chance of surviving the first 28 days after admission. Another study by Hansen et al. reported on the association between cytokine (IL-1β, IL-6, IL-10, and TNF-α) levels and the LRINEC score, disease severity and mortality in NSTI patients (Hansen et al. 2017b). Although no significant association between the LRINEC score and cytokine levels on admission was found, the authors found that IL-6 was consistently associated with disease severity and that high IL-1β (OR 3.86 [95% CI, 1.43–10.40], P-value ¼ 0.008) and IL-10 (4.80 [1.67–13.78], P-value ¼ 0.004) were associated with 30-day mortality. Also,

171

PTX3 may be an appropriate marker of severity and prognosis of NSTI (Hansen et al. 2016a). Although PTX3 was reported to perform even better than the “classical” inflammatory markers CRP and procalcitonin (PCT), its usefulness has been a matter of debate (Honore and Spapen 2016).

11.3.3

Predictive Biomarkers

Predictive biomarkers aim to objectively evaluate the outcome of a specific clinical intervention, or the differential outcomes of two or more interventions, including toxicity. Predictive and prognostic markers often get confused because many markers can be both predictive and prognostic. A marker can be an unfavorable prognostic factor but could predict favorably for response to therapy, or vice versa. Furthermore, it is possible that markers can be prognostically unfavorable, but predict favorably for one therapy and predict unfavorably for another therapy. When searching PubMed with the joint medical subject headings “NSTI” and “predictive markers,” most, if not all, publications indexed are studies that have identified prognostic rather than predictive markers. In a recent study Afzal and coworkers (Afzal et al. 2020) performed metabolomics analysis combined with advanced bioinformatics analysis of metabolite–metabolite association networks, indicating potential of metabolites as predictive markers in NSTI (described in more detail in Sect. 11.4.5). In conclusion, there is very little information on biomarkers that can be used to predict patient who benefits from different clinical interventions. Thus, this could potentially be an accelerating research field which would have a significant clinical impact.

11.3.4

Predisposing Biomarkers

The fourth category of biomarkers is used to identify individuals with a predisposition or susceptibility to develop disease. In NSTI, studies have revealed that allelic variations in human

172

E. Saccenti and M. Svensson

leukocyte antigen (HLA) class II haplotypes result in striking differences in severity and outcomes of invasive GAS infections via the differential presentations of GAS superantigens (SAgs) by class II HLAs to host T-cell receptor (TCR) Vβ elements (Kotb et al. 2002, 2003; Norrby-Teglund et al. 2000; Norrby-Teglund and Kotb 2000). In particular, specific HLA class II haplotypes have been found to confer strong protection from severe systemic disease, whereas others increased the risk of severe disease; patients with the DRB1*1501/DQB1*0602 haplotype mounted significantly reduced responses and were less likely to develop severe systemic disease. It is likely that in NSTI a complex interaction of multiple host and microbial factors that do not lend themselves to reduction into a simple formula exist, and therefore more holistic approaches need to be implemented to identify biomarkers for diagnosis, prognosis, prediction, or predisposition/susceptibility.

11.4

made possible by technological developments that have expanded the throughput and reduced the cost of single nucleotide polymorphisms (SNP)-based research and allowed SNP-based techniques to identify DNA copy number variation. Nevertheless, to identify SNPs remains a significant challenge, due to the difficulty of separating these. One of the limitations of GWA studies is that it captures only common SNPs, which by themselves may only contribute a small extent of risk to develop disease. NSTI patients infected with the same bacterial strain can develop very different manifestations and a role of immunogenetics of the host in shaping the outcome of invasive streptococcal infections can be identified (Kotb et al. 2002). However, to identify rare SNPs that contribute to develop NSTI for a small percentage of patients, more cost-effective technologies will be required. Despite its disadvantages DNA sequencing may become more important for patient care over the next 5 years as the costs are reduced and capabilities extended.

Biomarker Use 11.4.2

11.4.1

Genomic Biomarkers in NSTI

To identify genomic biomarkers DNA sequencing (host and pathogen) technologies, such as whole-genome sequencing can be used. While cancer research has reached the farthest using whole-genome sequencing (Staaf et al. 2019), it is difficult to find examples where DNA sequencing has been used to identify biomarkers in NSTI. It is likely that NSTI, with its acute course, poses other challenges than cancer and in NSTI perhaps gene sequencing may be best used to identify predisposing conditions. However, the relatively high costs likely prevent whole-genome sequencing analyses from becoming routine on a large scale. Instead, “exome” sequencing (targeted sequencing of protein-coding regions) may become a cost-effective way to provide access to the entire transcribed genome of individuals. In addition, genome-wide association (GWA) studies identifying risk loci are feeding into the clinical use of DNA variants. This research has been

Transcriptomic Biomarkers in NSTI

Mapping genotypes to phenotypes is one of the long-standing challenges in biology and medicine, and a powerful strategy for tackling this problem is performing transcriptome analysis. Transcriptomics entails the measurement and analysis of mRNA using microarray or RNA sequencing technologies. At the beginning microarrays offered tremendous opportunities by capturing parallel information about many more mRNAs than quantitative PCR techniques (RT-PCR), which mostly focus on single mRNA analyses. However, like all molecular techniques there are limitations, such as dependence on template quality (Bustin and Nolan 2004). Although microarrays have been highly productive research tools and used in several current prognostic and predictive tests, other emerging technologies have replaced the microarrays. Next-generation DNA sequencing technology known as “RNA-seq” allows simultaneous

11

Systems Biology and Biomarkers in Necrotizing Soft Tissue Infections

analysis of all RNA molecules, including alternative splice variants, mRNAs, non-coding RNAs (ncRNAs), and miRNAs. RNA-seq has revolutionized the analysis of RNAs. In a recent study Thänert et al. took advantage of the nextgeneration sequencing tools to enable microbial profiling using 16S rRNA sequencing with transcriptional analysis of host and microbe using dual RNA sequencing (RNA-seq) in tissue biopsies from NSTI patients (Thänert et al. 2019). The results from this study demonstrated that, despite the similar clinical presentation of NSTIs, the pathophysiology of mono- and polymicrobial etiology differs significantly and that these differences can potentially be exploited for diagnostic purposes. The transcriptional analysis of infected tissue indicated that the gene expression profile differed significantly between monomicrobial streptococcal infections and polymicrobial NSTI. Among the genes with significantly greater expression in polymicrobial compared to monomicrobial NSTIs were those encoding ECM components like collagen, fibronectin or lumican, as well as connective tissue growth factor, proteins. On the other hand, a set of genes encoding interferoninducible mediators such as CXCL9, CXCL10, CXCL11, MX1, and MX2 as well as the guanylate-binding GTP1 and GTP2 were most prominently higher expressed in monomicrobial NSTIs, in particular those caused by S. pyogenes. Because of the many diverse cell types in our body each of which express a unique transcriptome, conventional bulk population sequencing can provide only the average expression signal for an ensemble of cells. More recently there has been an explosion of interest in obtaining high-resolution views of single-cell transcriptomics (scRNA-seq) heterogeneity. Assessing the differences in gene expression between individual cells has the potential to identify rare populations that cannot be detected from an analysis of pooled cells. Due to issues with cost and analysis challenges, the implementation of scRNA-seq as a tool for biomarker identification in NSTI may still be some years away and is unlikely to give the complete picture.

11.4.3

173

Epigenomic Biomarkers in NSTI

Epigenetics can be defined as the field of inheritable changes in gene expression that are not caused by alterations in DNA sequences. Epigenetic mechanisms include DNA methylation, histone modifications, and non-coding RNAs, and epigenetic alterations have an increasingly clear role in modulating inflammatory and other immunological processes. Identification of epigenetic alterations may provide important prognostic, predictive markers. This is an exciting field that is likely to grow in clinical importance. There are no studies on epigenetics in NSTI, while the often-associated condition sepsis has been subjected to epigenetic analyses (Cross et al. 2019). To date, mostly the epigenetic modification associated with various stages of sepsis has been assessed. A focus has been to identify mechanisms involved in endothelial dysfunction during the hyperinflammatory response and those underpinning aspects of immunosuppression.

11.4.4

Proteomic Biomarkers in NSTI

Large scale analysis of specific proteomes, also known as proteomics, defines protein diversity and understands its biological consequences. Analysis of proteins as inflammatory markers in diseases could provide important information about disease severity and guide decision making. In true proteomics analysis all proteins from a sample of interest are usually extracted and digested with one or several proteases (typically trypsin alone or in combination with Lys-C (Wiśniewski and Mann 2012) to generate a defined set of peptides. The peptides obtained are subsequently analyzed by liquid chromatography coupled to mass spectrometry (LC-MS). The two most common approaches are either designed to achieve a deep coverage of the proteome (shotgun MS) or to collect as much quantitative information as possible for a defined set of proteins/peptides (targeted MS) (Picotti and Aebersold 2012). Protein content information can also be obtained through targeted approaches,

174

E. Saccenti and M. Svensson

including multiplex analyses made to identify a set of predetermined proteins. Since the test results are needed within hours in NSTI, approaches will most realistically include analyses of samples, such as plasma, that can be obtained quickly. However, rapid assessment based on proteomics and/or multiplex analyses of NSTI patients to identify protein-based biomarkers is challenging, as these are time consuming analyses. Nevertheless, multiplex approaches have been used to analyze NSTI plasma samples and has provided important insight to potential diagnostic biomarkers (Thänert et al. 2019; Hansen et al. 2017a).

biofilm formation assays revealed metabolite-specific effects on both bacterial growth and biofilm formation. A biofilm can be considered as a potential complicating microbiological feature of NSTI and, consequently, emphasizes reconsideration of antibiotic treatment protocols. This study identifies for the first time an NSTI-specific metabolic signature with implications for optimized diagnostics and therapies. Thus, the finding of metabolites in NSTI associated with biofilm formation emphasizes the potential of metabolites as predictive markers in NSTI.

11.4.6 11.4.5

Metabolomic Biomarkers in NSTI

Metabolomics takes a special position among the omics disciplines in the system top-down approach as the metabolome of biological processes, carrying imprints of genetic, epigenetic, and environmental factors, provides the link between the genotype and phenotype (Karakitsou et al. 2019; Rosato et al. 2018; Dunn and Ellis 2005; Kell 2004). Since metabolites represent the endpoint of many molecular pathways, metabolomics which quantifies the metabolite content of cells or tissues is in addition to have diagnostic and prognostic value also potentially useful for predicting treatment response. Metabolomics has been identified as a novel tool to discover targets for sepsis diagnosis and prognosis, as well as to gain insight into pathogenic disease mechanisms (Eckerle et al. 2017). Recently, we used untargeted metabolomics analyses of plasma from NSTI patients and healthy controls to identify the metabolic signatures and connectivity patterns among metabolites associated with NSTI. Out of 97 metabolites detected, the abundance of 33 was significantly altered in NSTI patients. Analysis of metabolitemetabolite association networks identified 20 metabolites differently connected between NSTI and controls. Testing of a set of differently connected metabolites (ornithine, ribose, urea, and glucuronic acid) in in vitro

Microbiomic Biomarkers in NSTI

Analyzing the microbiome at the system biology level for comparison in diseases to that of healthy subjects has become increasingly popular. Rapid progress in the development of next-generation sequencing (NGS) technologies in recent years has provided many valuable insights into complex biological systems, including diverse microbial communities. As far as NSTI is concerned, this field of research is in its infancy, but it may prove helpful in providing biomarkers which will give insights into the pathological process behind disease evolution and progression by determining specific etiological factors.

11.4.7

From One-Category-Analyses Towards Systems Biological Approaches

The classical approach to biomarker discovery usually entails the comparison of molecules or group of molecules among two or more groups, usually in a case/control scenario. In NSTI typical comparison is monomicrobial vs polymicrobial infection or patient outcomes (septic shock/noseptic shock (Hansen et al. 2018), amputation/ no-amputation, or mortality (Hansen et al. 2018)) or some clinical score (low/high LRINEC) (Hansen et al. 2017a). Statistical approaches are usually simple: candidate biomarkers are mostly

11

Systems Biology and Biomarkers in Necrotizing Soft Tissue Infections

tested one at a time in a univariate fashion using procedure like ANOVA or linear mixed models, logistic regression, or Kaplan–Meier curves when association with survival is of interest or by correlating biomarker concentration with some clinical outcome. Most new biomarkers proposed in the literature never reach the clinic, often because of a lack of reproducibility. In a meta-analysis of highly cited articles announcing new biomarker candidates for a variety of diseases, it was shown that follow-up experiments with greater statistical power generally fail to reproduce the same effect size as the original studies (Ioannidis and Panagiotou 2011; Feala et al. 2013). Moreover, in addition to the need for validation of such candidates, integration of the information provided by each biomarker is also needed for a comprehensive representation of the disease process. Here we argue that biomarker discovery would be greatly enhanced by applying systems biology principles such as the multi-scale integration of information and the analysis of dynamic patterns with the help of computational tools.

11.5

Systems Biology

Systems biology is an interdisciplinary field in which experimental and computational methods concur with mathematical models to the analysis and the understanding of biological phenomena. In systems biology different types of molecular knowledge are integrated and exploited thanks to the synergistic use and combination of statistical and mathematical model together with experimental data (Bruggeman and Westerhoff 2007). Systems biology embraces a system prospective as opposed to the reductionistic vision which is common in the mainstream biology: it aims to tackle biological problems by considering the interaction among the different parts of the systems rather than subdividing it into independent sub-problems. In biological systems the interactions between components dominate the component themselves in shaping the systemwide behavior and for this reason the reductionist approach is less effective to understand the

175

biomolecular mechanisms underlying complex biological processes (Ahn et al. 2006) Systems biology relies on the use of a mathematical model to integrate and understand data that are acquired using a vast array of experimental approaches ranging from molecular techniques to comprehensive measurement often encompassing multiple layers of the omics cascade such as transcriptomics, proteomics, and metabolomics. Different types of models can be used depending on the data available, the research questions, and whether a deductive or an inductive approach is followed. The deductive approach, usually referred to as Top-down approach (Oltvai and Barabási 2002), aims to gain insights into a given phenomenon starting from the system-wide data acquired using omics technologies from which information is extracted, usually using statistical and machine learning methods in combination with network inference and network analysis (Rosato et al. 2018; Ideker and Krogan 2012). The ultimate goal is to describe the interactions among the molecular constituents of the system (genes, proteins, metabolites), possibly across different conditions (Rosato et al. 2018, Ideker and Krogan 2012) to understand how these parts interact to shape the system-wide behavior of the system. On the other side, the inductive approach, referred to as Bottom-up approach (Oltvai and Barabási 2002), starts from a detailed molecular and biochemical knowledge of certain biological mechanisms and aims to create mathematical models that can reproduce experimental data. The overarching goal is to predict the systemwide behavior by building a complete genomescale model to provide an integrative view of the biological interactions occurring inside living systems (Shahzad and Loor 2012). The iterative cycle of systems biology is illustrated in Fig. 11.1 in relationship to the pipeline for biomarker discovery.

176

11.6

E. Saccenti and M. Svensson

Systems Biology Approaches to Biomarker Discovery

Determining whether a given biomarker is causal or reactive to NSTI (or, in general, to any disease) is likely to provide information on the underlying pathogenic process and ultimately impact its general applicability (Villoslada and Baranzini 2012). However, standard approaches that are based on just univariate analysis can fail, by their own nature, to account for the interrelatedness existing among genes, proteins, and metabolites that behave in an orchestrated way for what concerns regulation, transcription, and translation. For instance, Edwards and co-workers (Edwards et al. 2018) investigated S. pyogenes infections including NSTI among others, using a proteomic approach and identified several human proteins associated with NSTI. Yet, data was analyzed in a univariate fashion, missing the potential of highlighting association patterns among proteins that could have been uniquely associated with one or more type of infections. Systems biology approaches rely on the concept of mutual interaction among the constituents of a biological system and thus poses a great emphasis on measuring, describing, analyzing, and interpreting the relationships among molecular features (genes, proteins, metabolites...) rather than considering them one at a time. In this light, top-down systems biology approaches can be considered multivariate in nature, that is, the analysis is performed taking into consideration the relationships among the variables and not only their mean levels (i.e. concentration and abundances) (Saccenti et al. 2014a) as in the standard analysis for single molecule biomarkers. Relationships among genes, proteins, and metabolites are quantified using association measures which describe the similarity of the concentration and abundance profiles of these molecular features measured on a set of biological replicate samples. Correlations (either the Pearson’s or Spearman’s indexes) are commonly used and biological and biochemical information can be derived from both the strength and the sign

of correlations; for instance, if analyzing metabolites, a strong positive correlation can indicate an equilibrium condition or enzyme dominance, while strong negative correlation can indicate the presence of a conserved moiety (Camacho et al. 2005). It is fundamental to understand that the associations among variables can carry information also when concentrations (and abundances) do not carry relevant information, for instance, to discriminate between the case and control groups. This situation, so common in the biomarker discovery setting, is better illustrated with a simulated example. With reference to Fig. 11.2a, consider two metabolites (or protein or any other biomolecular quantity) A and B measured from case patients and controls: under this simulation there is no difference between the concentration levels of A and B in the two groups. However, if the case and control group are compared taking ratio of the concentrations of A and B a clear difference between case and controls is evident, as shown in Fig. 11.2b (see figure caption for more details) and it is clear that the information discriminating between cases and controls relies on both metabolites and their mutual relationship This phenomenon is a consequence of metabolites A and B being correlated as shown in Fig. 11.2c: from biological point of view what is observed here is that given a certain concentration of metabolite A, the concentration of metabolite B changes with class (Saccenti et al. 2014a). This indicates that when the relation or interaction between pairs of metabolites or other biomolecular features is thought to be important and to carry information about the problem studied, the ratios of metabolite concentration (or abundances) could be considered instead of single metabolite concentrations. Examples of this exist in the clinical setting where the deviation from the normal range of the ratios of metabolites or protein is used as diagnostic biomarkers (Petersen et al. 2012): ornithine-δ-aminotransferase deficiency in young children is diagnosed using the proline to citrulline ratio (De Sain-Van Der Velden et al. 2012), while the ratio between blood phenylalanine and tyrosine concentrations is used to

11

Systems Biology and Biomarkers in Necrotizing Soft Tissue Infections

177

Fig. 11.2 Idealized experiment aiming to find biomarkers discrimination between the case and control group based on the measurement and analysis of two metabolites A and B measure on n1 ¼ 75 cases and n2 ¼ 75 controls. (a) Comparison of the concentrations of metabolites A and B tested with 1-way ANOVA: there is no statically significant difference at the α ¼ 0.05 level (P-value 0.32 and

0.98, respectively, for metabolite A and B. (b) Comparison of the ratios of the concentrations of metabolites A and B: a discrimination between the case and control groups is evident (1-way ANOVA P-value ¼ 2.7
1018). (c) Scatter plot of the concentration profiles of metabolites A and B

identify heterozygous carriers of phenylketonuria risk alleles (Hsia 1958). Metabolite and protein ratios and variation thereof have often direct biological relevance and can provide mechanistic insights. In fact, if two metabolites are connected within (or by) a biochemical pathway, metabolite ratios approximate (under idealized steady state assumptions) the reaction rates (Petersen et al. 2012). For instance, considering the elementary biochemical reaction where metabolite A is converted to B and then degraded

½A k 2 ¼ ½B k 1

k1

k2

A !B! described by the differential equations 8 > >
> : d ½B ¼ þk1 ½A  k2 ½B dt where [A] and [B] are the concentrations of A and B at time t. At steady state the ratio of the concentration of A and B gives an approximation of the ratios of the rate constants for the two reactions

and can be used as a proxy to characterize the system. Since reaction constants are tightly constrained under normal physiological conditions, increased (or decreased) metabolite ratios can be an indication of alteration in the activity of an enzyme that catalyzes the reaction or a change in flux distribution. For instance, variation of the ratios between sphingolipids that differ by two carbon moieties has been linked to a modified beta-oxidation, while variation of the ratios between different classes of phospholipids has been proposed to describe modified activity of enzymes in the phospholipid pathways (Altmaier et al. 2008).

11.7

Networks and Network Analysis

When many metabolites or other biochemical and biomolecular features are measured, their mutual relationships can be conveniently represented using networks. A (biological) network is a graphical representation of objects (called nodes) and their relationships described by links

178

E. Saccenti and M. Svensson

or edges connecting different nodes. The exact meaning of the nodes and edges, i.e. of the kind of relationships they represent, depends on the context. In gene regulatory networks the nodes represent genes and the edge represents the existence of regulatory mechanism where the protein produced by a given gene regulates the expression of the target gene (Emmert-Streib et al. 2014). In networks describing protein–protein interaction the edge represents the existence of a physical interaction between two proteins (Jordán et al. 2012). In the context of biomarker discovery, the most relevant networks are probably the so-called metabolite–metabolite association networks, or correlation networks, where the nodes are metabolites and the edges represent the existence of associations between two metabolites, quantified through the correlation of their concentration profiles measured on replicated samples. The simplest approach to build metabolite– metabolite association networks is first taking the correlation rij among pairs of metabolites i and j and then imposing a threshold on either the magnitude of the correlation or on the statistical significant (or both), setting to 0 the correlation that do not satisfy such conditions. For instance  r ij ¼

r ij if jr ij j > θ and Pval < α 0 otherwise

ð11:1Þ

where θ and α (statistical significance) are parameters to be determined; a commonly accepted value for θ is 0.6 which is usually taken as a threshold to discriminate between low and medium-high correlation (Camacho et al. 2005). Other methods, like the Probabilistic Context Likelihood of Relatedness for Correlations (PCLRC) (Saccenti et al. 2014b), dispose of the necessity of a pre-defined threshold on the correlation but are context dependent, and the significance of the correlations is established with respect to the correlation background observed in the data.

If the value of the correlation (or any other association measure) is retained, the network is said to be weighted as every edge connecting two nodes describes the weight of the association. In other cases, only the presence/absence of an association is retained, and in this case the network is said to be unweighted  r ij ¼

1 if jr ij j > θ and Pval < α 0 otherwise

ð11:2Þ

The nodes in a network can be characterized using functions that can be derived from the patterns of its association: a common measure is the node degree or connectivity. For unweighted networks this is simply the number of its connection. For weighted networks a weighted connectivity is defined as the sum of the absolute values of the edges: χi ¼

X

j r ij j

ð11:3Þ

i>j

In the framework of biomarker discovery, the patterns of correlations between metabolites and their connectivity can be exploited by comparing them across conditions, for instance, cases and controls, to identify associations that are disrupted or altered by pathophysiological conditions. This approach is called differential network analysis and has been applied in NSTI research (Afzal et al. 2020) as well in other fields like cardiovascular disease (Vignoli et al. 2020; Saccenti et al. 2014b) and aging (Vignoli et al. 2018). Afzal et al. (2020) combined classical univariate and multivariate approaches together with differential analysis of metabolite–metabolite correlation networks which were built from metabolite concentrations of blood plasma metabolites measured using GC-MS from NSTI patients (n ¼ 34) and healthy (noninfected) controls (n ¼ 24) enrolled in the INFECT project (Madsen et al. 2019). In this study Spearman’s rank correlation was used as an index of metabolite association and implemented a conservative selection procedure involving resampling to obtain robust

11

Systems Biology and Biomarkers in Necrotizing Soft Tissue Infections

179

Fig. 11.3 (a) Absolute differences in metabolite connectivity in networks from NSTI and healthy control subject (see Eqs. (11.3) and (11.4)) against its statistical significance (Benjamini–Hochberg corrected -log10(P-value)). The horizontal lines represent the significance thresholds at α ¼ 0.05 and 0.01 level. (b) Scatter plot of the

metabolite connectivity observed in the metabolite–metabolite correlation networks built from NSTI and noninfected surgical controls. The colors are proportional to – log10(P-value) and the marker size is proportional to the difference in connectivity

estimates of the correlation because of the small sample size available. Metabolite association networks were constructed for both NSTI and healthy and for each of the 97 measured metabolites they calculated the differential connectivity Δχ i defined as

NSTI network, becoming hub node. In network biology a hub is a node that is densely connected, i.e. it is associated with many other nodes. Because of this, hub nodes are assumed to play crucial biological roles. The concept of hub was first introduced within the analysis of yeast protein–protein interaction networks and it was shown that highly connected nodes are more likely to be essential for survival (Jeong et al. 2001; Carter et al. 2004). Analysis of hub nodes has been extended to metabolite– metabolite association networks: comparing low and high cardiovascular risk patients. Saccenti et al. (2014b) identified conserved hub metabolite, and differentially conserved hubs and linked them to alteration of mitochondrial activity. Given this, Afzal et al. speculated that metabolites showing differential connectivity could play a role in disease pathogenesis either by affecting the host response or the bacterial infection. A set of significantly altered metabolites was selected and tested in vitro to

 χ Control Δχ i ¼ χ NSTI i i

ð11:4Þ

and χ Control are the connectivity for where χ NSTI i i the metabolite i in the association network specific for NSTI and healthy control, respectively. The significance of the observed difference in connectivity was assessed using a permutation test. Figure 11.3a shows the results of differential network analysis and Fig. 11.3b shows the metabolite connectivity in the NSTI specific network against connectivity in the uninfected surgery controls. The authors noted that several metabolites (urea, ribose, ribitol, pseudo uridine penta, malic acid) that were not connected in the control network were highly connected in the

180

E. Saccenti and M. Svensson

investigate potential influence on NSTI group A streptococcal strain growth and biofilm formation. The selected metabolites (ornithine, ribose, urea, and glucuronic acid) were supplemented to chemically defined media and revealed metabolite-specific effects on both bacterial growth and biofilm formation.

11.8

Multivariate Biomarker and Omics Signatures

A generalized class of biomarkers are the so-called multivariate biomarkers which are a generalization of the bivariate biomarker described in Sect. 11.6. This kind of biomarkers are sometimes denoted with the term signatures, especially in the context on omics research where gene, protein and metabolite expression, abundances and/or concentration levels are measured in a comprehensive fashion on biological samples using high-throughput analytical techniques like RNA sequencing, mass spectrometry, or nuclear magnetic resonance spectrometry. These experimental techniques are multivariate in nature since hundreds to thousands of molecular features are measured simultaneously and such molecular signatures are exploited through multivariate and machine learning approaches. Sung et al. (2012) defined operatively molecular signature as a set of biomolecular features together with a pre-defined computational procedure that applies those features to predict a phenotype of clinical interest on a previously unseen patient sample, stressing the computational/statistical nature of such signatures. Signatures are defined using statistical approaches in combination with machine learning techniques. Principal component analysis (PCA) (Hotelling 1931; Pearson 1901; Jolliffe 2002) is a common starting point when looking for biomarker signatures and it is used to explore data and to highlight if such a signature may exist. PCA is data reduction technique that allows for visualization and exploration of high-dimensional data by reducing its dimensionality. This is

accomplished by taking linear combinations of the original variables that explain as much as data variability as possible which allows to visualize the data in a lower dimensional space. Figure 11.4 shows the results of a PCA analysis on the data from Afzal and colleagues (Afzal et al. 2020). A clear separation between the metabolite profiles of NSTI patient and healthy controls is evident, indicating the existence of a metabolomics (in this case) signature as shown in the PCA score plot in Fig. 11.4. Figure 11.5 shows the relative importance (i.e. the PCA loadings) corresponding to the PCA score plot. The loadings are measure of the contribution of each variable to explain the variability observed in the data, in this case the separation between NSTI and controls. As it can be seen all variables (only the first 40 are shown for simplicity) contribute to the model, showing the multivariate nature of this metabolomics signature differentiating NSTI subjects from the controls. It is interesting to note that among the most important metabolites there are urea, ribose, ribitol, pseudo uridine penta, malic acid, the same metabolites that showed significant differential connectivity in the network analysis (see Fig. 11.3). This is because PCA, like all multivariate approaches, takes into account the correlation among variables which are also reflected in the network structure underlying the data, making network analysis and multivariate statistics two complementary approaches. Omics signatures consisting of many biomolecular features have been defined and routinely applied in the clinical setting. For instance, the MammaPrint test (Van ’T Veer et al. 2002) used to determine the most appropriate chemotherapy treatment in lymph node negative breast cancer patients with either positive or negative ER, consists of 70 gene expression signatures which were found to correlate with clinical phenotype (metastatic vs. non-metastatic) and validated on independent patient cohorts (Van de Vijver et al. 2002). Given the particular nature of NSTI, which is a fast spreading infection requiring extremely fast diagnosis and treatment, this type of highdimensional genomics signatures may not of immediate applicability in the clinical setting

11

Systems Biology and Biomarkers in Necrotizing Soft Tissue Infections

181

Fig. 11.4 Score plot of principal component analysis of concentration profiles of 97 metabolites measured on using GC-MS from NSTI patients (n ¼ 34) and healthy (surgery, noninfected) controls (n ¼ 24) enrolled in the INFECT project (Madsen et al. 2019). Data is from (Afzal et al. 2020). A clear separation between NSTI patient and controls is evident, indicating the existence of a multivariate metabolomics signature specific to NSTI. See Fig. YY for the corresponding loading plot

although some applications are emerging (See also Chap. 13, Sect. 11.3.2). However, omics signature can be explored to single out key players and provide mechanistic information about disease onset and progression and can generate hypotheses that can be further

test in vitro and or in vivo in animal models, closing the iterative cycle typical of systems biology approaches. In this respect the main limitation of PCA based signatures is that (1) all variables contribute to the model and (2) PCA base signatures are exploratory in nature and do

Fig. 11.5 Loading plot of principal component analysis of the concentration profiles of 97 metabolites measured on using GC-MS from NSTI patients (n ¼ 34) and healthy (surgery, noninfected) controls (n ¼ 24) enrolled in the INFECT project (Madsen et al. 2019). Data is from (Afzal

et al. 2020). Only the first 40, most contributing metabolites are shown. Urea, ribose, ribitol, pseudo uridine penta, and malic acid are also found to be relevant from network analysis, See Fig. 11.4 for the corresponding score plot

182

not allow, per se, the prediction of the clinical phenotype on previously unseen patient samples. Signatures that are easier to interpret can be obtained using advanced explorative techniques like sparse approaches (Camacho et al. 2017; Saccenti et al. 2018; Camacho et al. 2020). The existence of a predictive power of such signatures can be investigated using chemometrics or machine learning approaches like partial least squares discriminant analysis (PLS-DA) (Wold and Eriksson 2001) or some of its extensions (Lê Cao et al. 2008; Bylesjö et al. 2006; Camacho and Saccenti 2018), Random Forests (Breiman 2001), Support vector machines (Cortes and Vapnik 1995), or other machine learning tools. When properly implemented, these methods can allow unbiased prediction of the clinical phenotype on previously unseen patient samples: this is accomplished by implementing crossvalidation, which is a computational strategy that implies removing a set of samples from the data set, building a predictive model, and then testing in an unbiased way the ability of the model to predict the left out samples, mimicking, in this way, validation on an external cohort of samples. This procedure can be particularly relevant in NSTI where validation cohorts are difficult, to obtain. Random Forest was used to define a multivariate biomarker consisting of 11 blood analytes attaining specificity and sensitivity around 90% for the early detection of ovarian cancer (Bertenshaw et al. 2008), while Support vector machine has been used to derive a panel of 5 markers able to attaining a prediction sensitivity of 95.3% and a specificity of 99.4% (Mor et al. 2005). Predictive omics signatures have great potential in NSTI. As an example, we show here a Random Forest classification model fit on the data from (Afzal et al. 2020) with default parameters and implemented by splitting the data into a validation set (20% of the data) and a training set (80%) and repeating the crossvalidation procedure 100 times to take into account sampling variability results. On the training set the accuracy for the prediction of NSTI and control samples is 0.981 (0.978, 0.982 95%

E. Saccenti and M. Svensson

CI), while on the validation set one get average accuracy equal to 0.980 (0.975–0.995, 95% CI), average sensitivity 0.954 (0.938–0.969, 95% CI), and an average specificity of 1. The classification results on the validation set are similar to those on the training set and this indicates that, in principle, these results may be generalizable to samples acquired on external validation cohorts. Many classification tools also produce measures of importance of the predictors used in the model and this allows to single out which variables are most contributing to the discrimination among the groups to gain understanding of the underlying biological processes or identify potential biomarkers. Figure 11.6 shows the variable importance plot of the Random Forest discriminant model previously described to discriminate between NSTI and healthy controls. It is interesting to note that the group of variables selected as important is somehow different from the one from the PCA analysis (see Fig. 11.5): only ribose is present in the top five most important variables, while urea, which is important in the PCA model, is not important in the Random Forest model. This should not be surprising, since PCA and Random Forest have different aims and exploit different characteristics of the data: PCA is an unsupervised method, while Random Forest is a supervised method that uses information about sample class (i.e. whether they belong to NSTI patient or to controls) to build a model able to predict to which class an unknown sample may belong.

11.9

Future Perspective

Technology advances and critical analysis of molecular pathways have opened new horizons for management of diseases, exploring therapeutic solutions to each individual patient beyond the one-size fits all practice. As the amount of information available grows, there comes a stage when there is too much information to manually integrate reliably. This leads to the need for formal mathematical or computational models of disease that incorporate all this information. For this reason, clinical, microbiological, and experimental

11

Systems Biology and Biomarkers in Necrotizing Soft Tissue Infections

183

Fig. 11.6 Variable importance plot of a Random Forest discriminant model fit to the concentration profiles of 97 metabolites measured on using GC-MS from NSTI patients (n ¼ 34) and healthy (surgery, noninfected) controls (n ¼ 24) enrolled in the INFECT project (Madsen

et al. 2019). Data is from Afzal et al. (2020). Only the first 40, most contributing metabolites are shown. The model is used to investigate the predictive capability of the metabolomics signature discriminating NSTI patient from controls

data can be integrated using various types of mathematical algorithms to provide decision support tools for clinicians. This can pave the way for application of personalized medicine in infectious diseases, such as NSTI, with the aims to achieve the right diagnosis and right treatment for the right patient at the right time at the right cost. Multivariate statistics, machine learning, and reverse engineering approaches will allow pinpointing and mapping key nodes that can reveal possible biomarker sets from heterogeneous (meta-)data both from the tested pathogens and the hosts (including the various “omics,” pathotyping, patient stratification data). These top-down analyses will be iteratively combined with bottom-up modelling of specific sub-networks/biomarkers. By thoroughly applying an iterative “dry-wet” cycle centered around the cross-linking of bottom-up, hypothesis-driven modelling of selected pathways and/or signatures, with top-down, unbiased systems biology approaches biomarkers that may contribute to disease onset and outcome can be identified. In particular, biomarker discovery in NSTI will benefit from the integration of heterogenous data, such as patient data and metadata, data from the tested pathogens and the murine models

and the wealth of omics data that can be collected from them. Even if the vast heterogeneity and/or quality of the data will not allow detailed mechanistic modelling of specific pathways, these top-down, unbiased approaches will at least enable to identify basic structures in the topology of the various networks underlying the disease. This in itself is a substantial contribution to the state of the art in biomarker identification. Finally, it will be important to validate identified signatures/biomarkers for patient classification through the use of large patient cohorts. Indeed, validation remains one of the major bottlenecks for biomarker discovery. To become a clinically approved test, a potential biomarker should be confirmed and validated using large cohorts and should be reproducible, specific, and sensitive (Drucker and Krapfenbauer 2013). Although not uncommon, NSTI is still a rare condition whose characteristics (difficult diagnosis, fast spread, high mortality, among others) make the creation of validation cohorts difficult. However, the INFECT cohort and the methodologies and strategies for data collection and analysis developed within the INFECT project (Madsen et al. 2018, 2019; Afzal et al. 2020) are a first and

184

promising step towards the definition of clinically relevant biomarker in NSTI. Acknowledgement Financial support: The work was supported by the European Union Seventh Framework Programme: (FP7/2007-2013) under the grant agreement 305340 (INFECT project); the Swedish Governmental Agency for Innovation Systems (VINNOVA) under the frame of NordForsk (Project no. 90456, PerAID), and the Swedish Research Council and The Netherlands Organization for Health Research and Development (ZonMv) under the frame of ERA PerMed (Project 2018-151, PerMIT).

References Afzal M, Saccenti E, Madsen MB, Hansen MB, Hyldegaard O, Skrede S, Martins Dos Santos VAP, Norrby-Teglund A, Svensson M (2020) Integrated univariate, multivariate, and correlation-based network analyses reveal metabolite-specific effects on bacterial growth and biofilm formation in necrotizing soft tissue infections. J Proteome Res 19:688–698. https://doi. org/10.1021/acs.jproteome.9b00565 Ahn AC, Tewari M, Poon C-S, Phillips RS (2006) The clinical applications of a systems approach. PLoS Med 3:e209 Altmaier E, Ramsay SL, Graber A, Mewes H-W, Weinberger KM, Suhre K (2008) Bioinformatics analysis of targeted metabolomics—uncovering old and new tales of diabetic mice under medication. Endocrinology 149:3478–3489 Anvar MS, Minuchehr Z, Shahlaei M, Kheitan S (2018) Gastric cancer biomarkers; a systems biology approach. Biochem Biophys Rep 13:141–146 Bertenshaw GP, Yip P, Seshaiah P, Zhao J, Chen T-H, Wiggins WS, Mapes JP, Mansfield BC (2008) Multianalyte profiling of serum antigens and autoimmune and infectious disease molecules to identify biomarkers dysregulated in epithelial ovarian cancer. Cancer Epidemiol Biomarkers Prev 17:2872–2881 Breiman L (2001) Random forests. Mach Learn 45:5–32 Bruggeman FJ, Westerhoff HV (2007) The nature of systems biology. Trends Microbiol 15:45–50 Bustin SA, Nolan T (2004) Pitfalls of quantitative realtime reverse-transcription polymerase chain reaction. Jf Biomol Tech 15:155 Bylesjö M, Rantalainen M, Cloarec O, Nicholson JK, Holmes E, Trygg J (2006) OPLS discriminant analysis: combining the strengths of PLS-DA and SIMCA classification. J Chemom 20:341–351 Camacho J, Saccenti E (2018) Group-wise partial least square regression. J Chemometr 32:e2964 Camacho D, de la Fuente A, Mendes P (2005) The origin of correlations in metabolomics data. Metabolomics 1:53–63

E. Saccenti and M. Svensson Camacho J, Rodríguez-GÓMEZ RA, Saccenti E (2017) Group-wise principal component analysis for exploratory data analysis. J Comput Graph Stat 26:501–512 Camacho J, Smilde AK, Saccenti E, Westerhuis JA (2020) All sparse PCA models are wrong, but some are useful. Part I: computation of scores, residuals and explained variance. Chemom Intel Lab Syst 196:103907 Carter S, Brechbuhler C, Griffin M, Bond A (2004) Gene co-expression network topology provides a framework for molecular characterization of cellular state. Bioinformatics 20:2242–2250 Cortes C, Vapnik V (1995) Support-vector networks. Mach Learn 20:273–297 Cross D, Drury RE, Hill JL, Pollard AJ (2019) Epigenetics in sepsis: understanding its role in endothelial dysfunction, immunosuppression and potential therapeutics. Front Immunol 10:1363 De Sain-Van Der Velden MGM, Rinaldo P, Elvers B, Henderson M, Walter JH, Prinsen BHCMT, Verhoeven-Duif NM, DE Koning TJ, van Hasselt P (2012) The Proline/Citrulline ratio as a biomarker for OAT deficiency in early infancy. JIMD Rep 6:95–99 Del Campo M, Jongbloed W, Twaalfhoven HA, Veerhuis R, Blankenstein MA, Teunissen CE (2015) Facilitating the validation of novel protein biomarkers for dementia: an optimal workflow for the development of sandwich immunoassays. Front Neurol 6:202 Drucker E, Krapfenbauer K (2013) Pitfalls and limitations in translation from biomarker discovery to clinical utility in predictive and personalised medicine. EPMA J 4:7 Dunn WB, Ellis DI (2005) Metabolomics: current analytical platforms and methodologies. Trends Anal Chem 24:285–294 Eckerle M, Ambroggio L, Puskarich MA, Winston B, Jones AE, Standiford TJ, Stringer KA (2017) Metabolomics as a driver in advancing precision medicine in sepsis. Pharmacotherapy 37:1023–1032 Edwards RJ, Pyzio M, Gierula M, Turner CE, AbdulSalam VB, Sriskandan S (2018) Proteomic analysis at the sites of clinical infection with invasive Streptococcus pyogenes. Sci Rep 8:5950 Emmert-Streib F, Dehmer M, Haibe-Kains B (2014) Gene regulatory networks and their applications: understanding biological and medical problems in terms of networks. Front Cell Dev Biol 2:38 Feala JD, Abdulhameed MDM, Yu C, Dutta B, Yu X, Schmid K, Dave J, Tortella F, Reifman J (2013) Systems biology approaches for discovering biomarkers for traumatic brain injury. J Neurotrauma 30:1101–1116 Goossens N, Nakagawa S, Sun X, Hoshida Y (2015) Cancer biomarker discovery and validation. Transl Cancer Res 4(3):256–269 Hajar R (2016) Evolution of myocardial infarction and its biomarkers: a historical perspective. Heart Views 17:167 Hansen MB, Rasmussen LS, Garred P, Bidstrup D, Madsen MB, Hyldegaard O (2016a) Pentraxin-3 as a

11

Systems Biology and Biomarkers in Necrotizing Soft Tissue Infections

marker of disease severity and risk of death in patients with necrotizing soft tissue infections: a nationwide, prospective, observational study. Crit Care 20:40 Hansen MB, Rasmussen LS, Pilely K, Hellemann D, Hein E, Madsen MB, Hyldegaard O, Garred P (2016b) The lectin complement pathway in patients with necrotizing soft tissue infection. J Innate Immun 8:507–516 Hansen MB, Rasmussen LS, Svensson M, Chakrakodi B, Bruun T, Madsen MB, Perner A, Garred P, Hyldegaard O, Norrby-Teglund A, INFECT Study Group (2017a) Association between cytokine response, the LRINEC score and outcome in patients with necrotising soft tissue infection: a multicentre, prospective study. Sci Rep 7:42179–42179 Hansen MB, Rasmussen LS, Svensson M, Chakrakodi B, Bruun T, Madsen MB, Perner A, Garred P, Hyldegaard O, Norrby-Teglund A, INFECT Study GROUP, Nekludov M, Arnell P, Rosén A, Oscarsson N, Karlsson Y, Oppegaard O, Skrede S, Itzek A, Wahl AM, Hedetoft M, Bærnthsen NF, Müller R, Nedrebø T (2017b) Association between cytokine response, the LRINEC score and outcome in patients with necrotising soft tissue infection: a multicentre, prospective study. Sci Rep 7:42179 Hansen MB, Rasmussen LS, Garred P, Pilely K, Wahl AM, Perner A, Madsen MB, Hedegaard ER, Simonsen U, Hyldegaard O (2018) Associations of plasma nitrite, L-arginine and asymmetric Dimethylarginine with morbidity and mortality in patients with necrotizing soft tissue infections. Shock 49:667–674 Honore PM, Spapen HD (2016) Pentraxin-3 to better delineate necrotizing soft tissue infection: not really! Crit Care 20:173 Hotelling H (1931) Analysis of a complex of statistical variables into principal components. J Educ Psychol 24:417441 Hsia DYY (1958) Phenylketonuria: the phenylalaninetyrosine ratio in the detection of the heterozygous carrier. J Ment Defic Res 2:8–16 Ideker T, Krogan NJ (2012) Differential network biology. Mol Syst Biol 8:565 Ioannidis JP, Panagiotou OA (2011) Comparison of effect sizes associated with biomarkers reported in highly cited individual articles and in subsequent metaanalyses. JAMA 305:2200–2210 Jeong H, Mason S, Barabasi A, Oltvai Z (2001) Lethality and centrality in protein networks. Nature 411:41–42 Jolliffe IT (2002) Principal component analysis. Wiley, London Jordán F, Nguyen T-P, Liu W-C (2012) Studying protein– protein interaction networks: a systems view on diseases. Brief Funct Genomics 11:497–504 Karakitsou E, Foguet C, de Atauri P, Kultima K, Khoonsari PE, Martins Dos Santos VAP, Saccenti E, Rosato A, Cascante M (2019) Metabolomics in systems medicine: an overview of methods and applications. Curr Opin Syst Biol 15:91–99

185

Kell DB (2004) Metabolomics and systems biology: making sense of the soup. Curr Opin Microbiol 7:296–307 Kotb M, Norrby-Teglund A, Mcgeer A, El-Sherbini H, Dorak MT, Khurshid A, Green K, Peeples J, Wade J, Thomson G (2002) An immunogenetic and molecular basis for differences in outcomes of invasive group A streptococcal infections. Nat Med 8:1398 Kotb M, Norrby-Teglund A, Mcgeer A, Green K, Low D (2003) Association of human leukocyte antigen with outcomes of infectious diseases: the streptococcal experience. Scand J Infect Dis 35:665–669 Lê Cao K-A, Rossouw D, Robert-Granié C, Besse P (2008) A sparse PLS for variable selection when integrating omics data. Stat Appl Genet Mol Biol 7:1 Le Gall J-R, Lemeshow S, Saulnier F (1993) A new simplified acute physiology score (SAPS II) based on a European/north American multicenter study. JAMA 270:2957–2963 Madsen MB, Skrede S, Bruun T, Arnell P, Rosén A, Nekludov M, Karlsson Y, Bergey F, Saccenti E, Martins Dos Santos VAP, Perner A, Norrby-TeglundA, Hyldegaard O (2018) Necrotizing soft tissue infections—a multicentre, prospective observational study (INFECT): protocol and statistical analysis plan. Acta Anaesthesiol Scand 62:272–279 Madsen MB, Skrede S, Perner A, Arnell P, Nekludov M, Bruun T, Karlsson Y, Hansen MB, Polzik P, Hedetoft M (2019) Patient’s characteristics and outcomes in necrotising soft-tissue infections: results from a Scandinavian, multicentre, prospective cohort study. Intensive Care Med 45:1241–1251 Mehta S, Shelling A, Muthukaruppan A, Lasham A, Blenkiron C, Laking G, Print C (2010) Predictive and prognostic molecular markers for cancer medicine. Therapeut Adv Med Oncol 2:125–148 Mor G, Visintin I, Lai Y, Zhao H, Schwartz P, Rutherford T, Yue L, Bray-Ward P, Ward DC (2005) Serum protein markers for early detection of ovarian cancer. Proc Natl Acad Sci U S A 102:7677–7682 Norrby-Teglund A, Kotb M (2000) Host-microbe interactions in the pathogenesis of invasive group A streptococcal infections. J Med Microbiol 49:849 Norrby-Teglund A, Chatellier S, Low DE, Mcgeer A, Green K, Kotb M (2000) Host variation in cytokine responses to superantigens determine the severity of invasive group A streptococcal infection. Eur J Immunol 30:3247–3255 Oltvai ZN, Barabási A-L (2002) Life’s complexity pyramid. Science 298:763–764 Pearson K (1901) On lines and planes of closest fit to systems of points in space. London, Edinburgh, Dublin Philos Mag J Sci 2:559–572 Petersen A-K, Krumsiek J, Wägele B, Theis FJ, Wichmann H-E, Gieger C, Suhre K (2012) On the hypothesis-free testing of metabolite ratios in genome-wide and metabolome-wide association studies. BMC Bioinfo 13:120

186 Picotti P, Aebersold R (2012) Selected reaction monitoring–based proteomics: workflows, potential, pitfalls and future directions. Nat Method 9:555 Quezada H, Guzmán-Ortiz AL, Díaz-Sánchez H, ValleRios R, Aguirre-Hernández J (2017) Omics-based biomarkers: current status and potential use in the clinic. Bol Med Hosp Infant Mex 74:219–226 Rosato A, Tenori L, Cascante M, de Atauri Carulla PR, Martins Dos Santos VAP, Saccenti E (2018) From correlation to causation: analysis of metabolomics data using systems biology approaches. Metabolomics 14:37 Saccenti E, Hoefsloot HC, Smilde AK, Westerhuis JA, Hendriks MM (2014a) Reflections on univariate and multivariate analysis of metabolomics data. Metabolomics 10:361–374 Saccenti E, Suarez-Diez M, Luchinat C, Santucci C, Tenori L (2014b) Probabilistic networks of blood metabolites in healthy subjects as indicators of latent cardiovascular risk. J Proteome Res 14:1101–1111 Saccenti E, Smilde AK, Camacho J (2018) Group-wise ANOVA simultaneous component analysis for designed omics experiments. Metabolomics 14:73 Schwartz S, Kightlinger E, de Virgilio C, de Virgilio M, Kaji A, Neville A, Bennion R (2013) Predictors of mortality and limb loss in necrotizing soft tissue infections. Am Surg 79:1102–1105 Shahzad K, Loor JJ (2012) Application of top-down and bottom-up systems approaches in ruminant physiology and metabolism. Curr Genomics 13:379–394 Simon R (2011) Genomic biomarkers in predictive medicine. An interim analysis. EMBO Mol Med 3:429–435 Simon R (2014) Biomarker based clinical trial design. Chin Clin Oncol 3:3 Staaf J, Glodzik D, Bosch A et al (2019) Whole-genome sequencing of triple-negative breast cancers in a population-based clinical study. Nat Med 25:1526–1533. https://doi.org/10.1038/s41591-0190582-4 Stevens DL, Bryant AE (2017) Necrotizing soft-tissue infections. N Engl J Med 377:2253–2265 Sung J, Wang Y, Chandrasekaran S, Witten DM, Price ND (2012) Molecular signatures from omics data: from chaos to consensus. Biotechnol J 7:946–957 Thänert R, Itzek A, Hoßmann J, Hamisch D, Madsen MB, Hyldegaard O, Skrede S, Bruun T, Norrby-Teglund A, Medina E (2019) Molecular profiling of tissue biopsies reveals unique signatures associated with streptococcal necrotizing soft tissue infections. Nat Commun 10:1–15

E. Saccenti and M. Svensson Vafaee F, Diakos C, Kirschner MB, Reid G, Michael MZ, Horvath LG, Alinejad-Rokny H, Cheng ZJ, Kuncic Z, clarke S (2018) A data-driven, knowledge-based approach to biomarker discovery: application to circulating microRNA markers of colorectal cancer prognosis. NPJ Syst Biol Appl 4:1–12 Van ’T Veer LJ, Dai H, Van de Vijver MJ, He YD, Hart AAM, Mao M, Peterse HL, Van Der Kooy K, Marton MJ, Witteveen AT, Schreiber GJ, Kerkhoven RM, Roberts C, Linsley PS, Bernards R, Friend SH (2002) Gene expression profiling predicts clinical outcome of breast cancer. Nature 415:530–536 Van de Vijver MJ, He YD, Van ’T Veer LJ, Dai H, Hart AA, Voskuil DW, Schreiber GJ, Peterse JL, Roberts C, Marton MJ (2002) A gene-expression signature as a predictor of survival in breast cancer. N Engl J Med 347:1999–2009 Vignoli A, Tenori L, Luchinat C, Saccenti E (2018) Age and sex effects on plasma metabolite association networks in healthy subjects. J Proteome Res 17:97–107 Vignoli A, Tenori L, Giusti B, Valente S, Carrabba N, Balzi D, Barchielli A, Marchionni N, Gensini GF, Marcucci R, Gori AM, Luchinat C, Saccenti E (2020) Differential network analysis reveals metabolic determinants associated with mortality in acute myocardial infarction patients and suggests potential mechanisms underlying different clinical scores used to predict death. J Proteome Res 19:949–961 Villoslada P, Baranzini S (2012) Data integration and systems biology approaches for biomarker discovery: challenges and opportunities for multiple sclerosis. J Neuroimmunol 248:58–65 Vincent JL, Moreno R, Takala J, Willatts S, de Mendonça A, Bruining H, Reinhart CK, Suter PM, Thijs LG (1996) The SOFA (sepsis-related organ failure assessment) score to describe organ dysfunction/ failure. On behalf of the working group on sepsisrelated problems of the European Society of Intensive Care Medicine. Intensive Care Med 22:707–710 Wiśniewski JR, Mann M (2012) Consecutive proteolytic digestion in an enzyme reactor increases depth of proteomic and phosphoproteomic analysis. Anal Chem 84:2631–2637 Wold S, Eriksson L (2001) PLS-regression: a basic tool of chemometrics. Chemom Intell Lab Syst 58:109–130 Wong C-H, Khin L-W, Heng K-S, Tan K-C, Low C-O (2004) The LRINEC (laboratory risk indicator for necrotizing fasciitis) score: a tool for distinguishing necrotizing fasciitis from other soft tissue infections. Crit Care Med 32:1535–1541

Systems and Precision Medicine in Necrotizing Soft Tissue Infections

12

Vitor A. P. Martins dos Santos, Christopher Hardt, Steinar Skrede, and Edoardo Saccenti

Contents 12.1

Introduction: The Case for a Systems and Personalized Approach to NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188

12.2 12.2.1 12.2.2 12.2.3

State of the Art: Systems, Precision and Personalized Medicine in NSTI Systems Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Challenges in Systems Medicine and How to Tackle Them . . . . . . . . . . . . . . . . . Precision and Personalized Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

12.3 12.3.1 12.3.2

Big Data, Machine Learning and Deep Learning in Systems Medicine 193 Big Data Definitions and Characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 193 AI in Systems Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194

12.4 12.4.1 12.4.2 12.4.3 12.4.4

Information Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Personalized and Precision Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Case of NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Integrating Heterogeneous Data with FAIR Principles . . . . . . . . . . . . . . . . . . . . . . Laying the Basis for Computer-Assisted Decision Support . . . . . . . . . . . . . . . . . .

189 190 190 192

196 196 197 197 198

V. A. P. Martins dos Santos (*) Laboratory of Systems and Synthetic Biology, Wageningen University and Research, Wageningen, The Netherlands Lifeglimmer GmbH, Berlin, Germany e-mail: [email protected] C. Hardt Lifeglimmer GmbH, Berlin, Germany S. Skrede Department of Clinical Science, University of Bergen, Bergen, Norway Department of Medicine, Haukeland University Hospital, Bergen, Norway E. Saccenti Laboratory of Systems and Synthetic Biology, Wageningen University and Research, Wageningen, The Netherlands # Springer Nature Switzerland AG 2020 A. Norrby-Teglund et al. (eds.), Necrotizing Soft Tissue Infections, Advances in Experimental Medicine and Biology 1294, https://doi.org/10.1007/978-3-030-57616-5_12

187

188

V. A. P. Martins dos Santos et al. 12.5 12.5.1 12.5.2 12.5.3 12.5.4

Clinical Decision Support Systems for Soft Tissue Infections . . . . . . . . . . . . . .199 The Need to Enhance Medical Decisions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .199 What Are CDSS What Is Their Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .199 CDSS in NSTI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .200 Current and Future Developments in Relation to Dedicated CDSS for NSTI . 201

12.6

Conclusion and Perspectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .204

References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .204

Abstract

Necrotizing soft tissue infections (NSTI) are multifactorial and characterized by dysfunctional, time dependent, highly varying hyperto hypo-inflammatory host responses contributing to disease severity. Furthermore, host-pathogen interactions are diverse and difficult to identify and characterize, due to the many different disease endotypes. There is a need for both refined bedside diagnostics as well as novel targeted treatment options to improve outcome in NSTI. In order to achieve clinically relevant results and to guide preclinical and clinical research the vast amount of fragmented clinical and experimental datasets, which often include omics data at different levels (transcriptomics, proteomics, metabolomics, etc.), need to be organized, harmonized, integrated, and analyzed taking into account the Big Data nature of these datasets. In this chapter, we address these matters from a systems perspective and yet personalized approach. The chapter provides an overview on the increasingly more frequent use of Big Data and Artificial Intelligence (AI) to aggregate and generate knowledge from burgeoning clinical and biochemical information, addresses the challenges to manage this information, and summarizes current efforts to develop robust computer-aided clinical decision support systems so to tackle the serious challenges in NSTI diagnosis, stratification, and optimized tailored therapy. Keywords

Artificial intelligence · Big data · Clinical decision support systems · Deep learning ·

Information management · Personalized medicine · Semantic technologies Highlights • Systems, precision and personalized medicine approaches are methods to facilitate and improve contemporary diagnosis, accurate stratification, and optimized tailored therapy in NSTI. • Big Data analytics and artificial intelligence are increasingly allowing to uncover disease mechanisms, giving a basis for patient stratification. • We have established a prototype of an advanced platform for personalized medicine in NSTI according to FAIR principles. • We have developed the basis for a clinical decision support system in NSTI.

12.1

Introduction: The Case for a Systems and Personalized Approach to NSTI

NSTI are complex multifactorial diseases that can be caused by a variety of microbes. They are frequently complicated by septic shock and multi-organ failure. Despite modern medicine, the mortality is high, often exceeding 25%, and amputation is required in up to 15% of the cases (Peetermans et al. 2020). Most patients affected are individuals with co-morbidities, e.g. cardiovascular diseases and diabetes mellitus, but some patients are young immunocompetent individuals. The fulminant, often rapid course of these invasive infections (Kittang et al. 2010), demands early diagnosis and immediate intervention. However, misdiagnosis and subsequent doctor’s delay is frequent (Goh et al. 2014), due

12

Systems and Precision Medicine in Necrotizing Soft Tissue Infections

to discrete and unspecific initial symptoms, scarce clinical findings and lack of biomarkers in early stages of many NSTI. The cornerstones in the treatment of NSTI are early aggressive surgical debridement and appropriate antimicrobial therapy. Frequently there is a need for advanced supportive measures. There is clearly an urgent need for refined bedside diagnostics and novel, targeted treatment options to improve outcome in these patients. Because of the multifactorial nature of NSTI and sepsis, it is becoming increasingly clear that individualized approaches are required to improve outcome of these patients. These infections are characterized by dysfunctional, highly varying hyper- to hypo-inflammatory, host responses contributing to disease severity (Goldstein et al. 2007; Sarani et al. 2009; Peetermans et al. 2020). Thus, individualized therapeutic strategies targeting both pathogen as well as the host response have great potential, but require patient stratification based on disease signatures. Data are becoming available showing the power of omics in identification of disease signatures (Davenport et al. 2016), as well as in providing the basis for patient stratification through large-scale analysis of clinical data (Hardt et al., in preparation), but fail to acknowledge the complexity of the host-pathogen interactions and the patient’s individual responses to infection and its progression. In this chapter, we address these problems from a systems medicine perspective, provide an overview on the increasingly more frequent use of Big Data and Artificial Intelligence (AI) to aggregate, and generate knowledge from burgeoning clinical and biochemical information directing towards personalized approaches. Moreover, we address the challenges to manage this information and summarize current efforts to translate this information into actionable knowledge through the development of robust computer-aided clinical decision support systems (CDSS) to tackle the serious challenges in NSTI diagnosis, stratification, and optimized tailored therapy.

12.2

189

State of the Art: Systems, Precision and Personalized Medicine in NSTI

Current medical science is largely conducted under a reductionist paradigm, which involves the notion that complex phenomena like mechanisms underlying onset of infectious diseases, interactions between host and pathogens and disease progression may be better understood by breaking them down into smaller, simpler components (Ahn et al. 2006). However, given the complexity of the conditions under study and the myriad of underlying factors, this often leads to biased focus and oversimplification (i.e. by focusing only on a handful of major factors with the biggest effect, while the sum of minor factors may be considerable) and generalization (i.e. assuming that a common causeeffect relationship applies equally in all cases). Such simplifications and generalizations often limit our ability to understand how multiple variables interact with one another to create emergent effects and hamper not only our understanding of the disease, but more importantly, our capability of delivering better treatments. There is clearly a need to address health and disease from a systems perspective, that is, one that accounts for all factors and interactions. For instance, Chap. 12 illustrated the application of such a systems (biology) approach for the identification of potential biomarker sets in NSTI. This entailed the uncovering of metabolite–metabolite association networks and analysis and deployment of machine learning methods (Afzal et al. 2020). These findings would not have been possible with a traditional, reductionist approach of trying to identify potential biomarkers by standard approaches that are based, e.g. on simple univariate analysis of datasets, as these fail to account for the interrelatedness existing among genes, proteins, and metabolites that behave in an orchestrated way for what concerns regulation, transcription, and translation (Rosato et al. 2018).

190

V. A. P. Martins dos Santos et al.

Fig. 12.1 Schematic illustration of the core differences between reductionism and systems science, when analyzing the properties of a system (Tillmann et al. 2015). Licensed under Creative Commons

12.2.1

Systems Medicine

Building upon the uncovering of biological mechanisms as above and going beyond that, “Systems medicine” applies systems biology approaches to medical research and medical practice. Its objective is to integrate a variety of biological/medical data at all relevant levels of cellular organization using the power of computational and mathematical modelling, to enable understanding of the pathophysiological mechanisms, prognosis, diagnosis, and treatment of disease (Auffray et al. 2010). Systems Medicine involves iterative and reciprocal feedback between the clinical practice and the research carried on with computational, statistical, and mathematical multiscale analysis and modelling at both the epidemiological and individual patient level (www.easym.eu). This new paradigm of Systems Science and Medicine complements the traditional reductionist approach (as exemplified in Fig. 12.1) and, ideally, leads to the identification of mechanisms related to disease pathophysiology, selection of novel drug targets and biomarkers, patient stratification, risk assessment, and optimized therapy. As described in Chap. 1, the systems medicine INFECT project (https://permedinfect.com/ projects/infect/) has generated comprehensive knowledge of diagnostic features, causative

microbial agents, treatment strategies, and pathogenic mechanisms (host and bacterial disease traits and their underlying interaction networks) by using the largest thus far patient cohort in the world. Also, INFECT has proven the value of systems medicine approaches in acute infectious diseases to achieve improved diagnostics and therapeutics to improve patient disease outcome. In particular, the novel understanding of the disease mechanisms has resulted in changed clinical practice related to antibiotic usage as well as use of immunomodulatory treatments (Bergsten et al. 2020; Madsen et al. 2018) (see also Chaps. 7, 8, and 9). However, the insights gained also underscore the need for patient stratification and implementation of tailored therapy. Despite considerable progress, there are substantial hurdles to be overcome related to the difficulties in accounting for patient’s individual conditions, or intrinsic to Systems Medicine itself.

12.2.2

Challenges in Systems Medicine and How to Tackle Them

The recent transition to data-rich (i.e. molecular characterization) applications, which followed the omics revolution with the advent of high throughput genomics, transcriptomics, metabolomics, and all other omics disciplines,

12

Systems and Precision Medicine in Necrotizing Soft Tissue Infections

have enabled Systems Medicine approaches (Noble 2008). However, this data deluge is at the same time an opportunity and a challenge. An opportunity because it allows patient phenotyping at different levels in an unprecedented manner, and a challenge because of the heterogeneity observed in both data and patients. There is an inherent data heterogeneity which arises when different data types and different data resources are integrated. This poses challenges for both data handling and analysis which can be addressed by vertical integration, i.e. when different data informs on the same biological level, or by horizontal integration, i.e. when multiple datasets inform on the same biological level (Berthold et al. 2010). Both cases can be handled, using FAIR data approaches (Findable, Accessible, Interoperable, Reusable) and semantics technologies (Berthold et al. 2010) but, as a matter of fact, there are other inherent sources of variability, which poses even greater challenges to Systems medicine approaches. Humans exhibit a great phenotypic diversity, which originates from the complex interplay of genetic, epigenetic, and environmental factors, that affects both disease manifestations and responses to therapy (Assfalg et al. 2008; Bernini et al. 2009). Nonetheless, intra-individual variability is less than half the inter-individual variability, making personalized medicine possible (Hughes et al. 2015; Gruden et al. 2012). Under this light, statistical and computational characterization of individual and inter-individual variability is pivotal to the deployment of systems and personalized medicine approaches, which will allow both higher sensitivity and specificity of personalized assays and substantial new insights into health and disease. Characterization of phenotypic diversity in both health and disease and underlying biomolecular mechanisms should be carried out at different levels. Examples could be developing new statistical and computational methods, to exploit the wealth of information obtainable at all omics levels, or by developing and deploying multiscale approaches to model how processes occurring at widely different scales integrate, results in the phenotypic variability observed in humans.

191

Multiscale modelling is another way to approach the problem of phenotypic variability. This includes bridging molecular and physiological processes, even taking place at very different time and spatial scales (Meier-Schellersheim et al. 2009). This could facilitate a better understanding of the biology of health and disease, allowing us to tailor models to individual patients (e.g. genome-scale metabolic model constrained by patient-specific data). A key challenge is to select measurements and data collected at the small scales and combine them into informative metrics to be transferred to a higher level. This is the realm of metamodelling, which is the statistical approximations or predictions of the relationships between the various model components. Meta-modelling has been proposed as an efficient solution to link models obtained at different scales, to link modelling results and measured data, and to identify the most important metrics determining system functionality at the various levels (Tøndel et al. 2012). Among statistical approaches derived from multivariate analysis, multi-way analysis has been proposed to retain the block-wise structure of temporal data originating from nonlinear dynamic models used to describe the systems at different levels (Tøndel et al. 2012). However, other solutions can be hypothesized that involve other component methods: some examples are principal component regression (Jolliffe 1982) or partial least square regression (Wold and Eriksson 2001) that can be used to model the complex relationships between input parameters and model outputs of nonlinear dynamic models likely with embedded procedure for variable selection to reduce dimensionality and complexity of the parameter space (Tøndel et al. 2012). Furthermore, parameter characteristics, such as distributional properties, covariance, and correlation patterns, could be inferred by stochastics methods designed to deconvolute correlative or noisy patterns: since signals can be separated from background biological variability in datadriven inference framework, this could ultimately lead to a better definition of phenotype characteristics.

192

12.2.3

V. A. P. Martins dos Santos et al.

Precision and Personalized Medicine

Personalized Medicine builds upon Systems Medicine and is an emerging data-driven health care approach that integrates phenotypic, genotypic, epigenetic, lifestyle, and environmental factors unique to an individual. The goal of personalized medicine is to facilitate diagnosis, predict effective therapy, and avoid adverse reactions specific for each patient (Nimmesgern et al. 2017; Union 2015). Precision medicine is the concept of tailoring disease treatment and prevention to account for differences in genetic, environmental, or even lifestyle factors specific to groups of people (Bresnick 2018). Precision medicine takes genetic and biochemical information unique to a group of patients and uses that information to develop more specific and streamlined medications or treatments. The goal is to ensure that each medication or treatment is best suited to treat the individual, resulting in decreased side effects and increased effectiveness (Hulsen et al. 2019). Although the terms precision medicine and personalized medicine are sometimes used interchangeably, generally speaking, “precision medicine” seeks to create treatments that are applicable to groups of individuals who meet certain characteristics, whereas “personalized medicine” implies individualized treatments available for every unique patient (Erikainen and Chan 2019). Precision and personalized medicine are in their infancy in infectious diseases, in particular acute diseases such as sepsis and NSTI (Lazăr et al. 2019). As the patient populations in severe infections are highly heterogeneous due to hostmicrobe interactions governed by different host factors and pathogens driving unique pathogenic mechanisms, personalized and precision medicine approaches may prove crucial. It is recognized that a dysregulated host response to infection is directly linked to severity and outcome of severe infections, such as sepsis and NSTI. Furthermore, the response profiles/disease signatures can be highly variable, ranging from

hyper- to hypo-responses with different mediators involved (Hotchkiss and Karl 2003; Anaya et al. 2005; Huang et al. 2011; Thänert et al. 2019). The identification of disease signatures of value for patient classification studies in welldefined patient cohorts has been undertaken, exploring host responses, pathogen profiles, and their association with disease outcomes (Chella Krishnan et al. 2016; Thänert et al. 2019). However, as for all biomarkers (see Chap. 12), these findings must be validated by use of other patient cohorts, as stated in a landmark position paper by the European Society of Clinical Microbiology and Infectious Diseases (Rello et al. 2018). Accordingly, to advance on precision medicine of NSTI and sepsis, two current multinational projects have been designed and implemented (PerMIT and PerAID, www.permedinfect.com). These projects build upon the unique resources created in the INFECT project, including clinical expertise, patient registry, biobank, multi-omics data, identified candidate disease signatures, data stewardship resources, and basic science experimental model systems. These resources are currently being used to test data-driven working hypotheses through advanced preclinical and clinical studies combined with Big Data integration and information technology solutions, to develop patient stratification schemes allowing for individualized therapy in NSTI. The identified pathogenic mechanisms and biomarkers linked to particular NSTI disease signatures and clinical outcome are being further validated in sepsis patients, a condition that, just like NSTI, is defined by the host’s response to an infection. A major strength of these on-going studies is that they resort on well-defined patient cohorts, allowing for robust conclusions linking host response signatures and pathogenic mechanisms to clinical outcomes encouraging translation of these findings into future patient handling measures. Interestingly as well, these two synergistic projects envision that some stratification and clinical trial designs will be shared for NSTI and sepsis, whereas others will target more strictly defined patient subgroups. For example, sepsis patients with an immunosuppressive response

12

Systems and Precision Medicine in Necrotizing Soft Tissue Infections

profile might be targeted for an immune boosting therapy, whereas NSTI and sepsis patients with a hyperinflammatory profile will benefit from a suppressive agent. Undoubtedly, such insights will allow for effective tailored treatments and development of new tools and concepts for future clinical trials. Crucially, however, for this potential to be fulfilled, some further challenges have to be tackled, in particular with regard to the handling of information and translation to actionable knowledge. Indeed, a key challenge in the personalized/ precision medicine field is the vast amount of fragmented clinical and experimental datasets that need to be organized, harmonized, and integrated in order to achieve clinically relevant results and to guide preclinical and clinical research. This is also essential to enable fruitful Big Data analytics to yield meaningful insights, as well as for development of Clinical Decision Support Systems (CDSS) to assist effectively and efficiently in bedside decision.

12.3

12.3.1

Big Data, Machine Learning and Deep Learning in Systems Medicine Big Data Definitions and Characteristics

Big Data and the development of techniques to handle it have the potential of enhancing our ability to probe and understand which parts of the biological machinery underlying the normal functioning of the organism that may be or become dysfunctional, given the pathophysiology of a condition (Hulsen et al. 2019). Big Data are characterized by the so-called four V’s (Schroeck et al. 2012), which stands for volume, variety, velocity, and variability. 1. Volume refers to the size of data, where size indicates the physical occupancy of data files. In Big Data applications exabytes, zettabytes, and even higher amounts of data need to be handled simultaneously. For what concerns

193

systems medicine applications, the volume of the data is likely to be much smaller. 2. Variety refers to the heterogenous nature and sources of data. In the biomedical field the different types of data that can be collected, mined, and analyzed are virtually endless. Data comes not only in form of health records, clinical data, to which data from omics measurements can be added but also as medical imaging, including X-rays, CT, or MRI, or images recorded on tissue samples, e.g. tissue biopsies. These different data sources need to be handled and integrated to be properly analyzed, taking into account there are structured data (e.g. data stored in excel format) and unstructured data (like doctor’s notes). This is usually one of the most challenging tasks. 3. Velocity refers to rate of data sampling and acquisition. Standard clinical measurements and omics data are usually static or acquired at a (very) low sampling rates. Differently, health monitoring systems, including, but not limited to, smartphones, smart watches, smart bracelets/wristbands, connected sensors, and wearable devices, enable continuous monitoring of patient data by sensing and transmitting measurements such as heart rate, blood pressure, body temperature, respiratory rate, chest sounds, and electrocardiogram (Vitabile et al. 2019) at high frequency, creating a flow of data that often needs to be processed “on-thefly.” 4. Variability concerns the quality of the data acquired and their inconsistency. For biomedical and healthcare applications, data quality is a very critical aspect, because erroneous information can lead to erroneous diagnosis and treatment. The problem of obtaining quality data is complex and cross-disciplinary. Over the years, several organizations have contributed to defining the quality of various products and services and identifying ways of measuring such quality (Brighi 2018). Handling and management of data inconsistency, such those introduced by missing data,

194

V. A. P. Martins dos Santos et al.

is of paramount importance for the exploitation of Big Data. Accordingly, Big Data are high-volume, highvelocity and/or high-variety information assets that demand cost-effective, innovative forms of information processing that enable enhanced insight, decision making, and process automation. This definition (www.gartner.com/it-glos sary/big-data/), albeit stated in the context of informatics, summarizes the challenges and the gain that Big Data present and can offer to Systems Medicine. The ultimate goal of the collection and use of Big Data in Systems Medicine is the possibility of obtaining better description of both health and disease profiles and to use them to build integrative models that can be used to predict disease onset and progression and to tailor better treatment for each patient. Under this light, Big Data approaches are fundamental for the development of Precision medicine, which aim to integrate phenotypic, genomic, epigenetic, and environmental factors unique to an individual to facilitate diagnosis, predict effective therapy, and avoid adverse reactions specific for each patient. Thus, precision medicine needs to operate on different scales to gain other insights into health and disease, utilizing and integrating data from cells, tissues, organs, and ecosystems, e.g. those constituted by microbial communities (Stacy et al. 2016).

12.3.2

AI in Systems Medicine

Artificial intelligence encompasses the use of software and algorithms to emulate the human cognition in the analysis of complex medical data. AI is being successfully employed in several medical fields. Models have been constructed that enable to distinguish high-risk breast lesions (HRL) diagnosed with image-guided needle biopsy that require surgical excision from HRLs that are at low risk for upgrade to cancer at surgery (Bahl et al. 2018) or able to detect pneumonia using chest X-rays with an accuracy level exceeding practicing radiologists (Rajpurkar

et al. 2017). Using images, artificial intelligence approaches have been used to describe the impact of orthognathic treatment on facial attractiveness and age appearance (Patcas et al. 2019). A seminal study (Esteva et al. 2017) tested the performance of AI to distinguish keratinocyte carcinomas versus benign seborrheic keratoses and malignant melanomas versus benign nevi, training it on >15,000 biopsy-proven diagnostic images, against 21 board-certified dermatologists, and found AI to perform on par with all tested experts across both comparisons, and other studies confirmed AI ability in identifying melanoma from dermoscopic images with accuracy similar to that of specialists (Phillips et al. 2019, 2020). There are few applications of Big Data and AI to systems medicine specific to NSTI. The management of NSTI is complex given that clinical presentation is highly variable and range from early sepsis with obvious skin involvement to minimal cutaneous manifestations with a disproportionate systemic response (Bosshardt et al. 1996). Classic signs like fever, diffuse crepitus, and shock are late signs: once large blisters and gangrene develop, the infectious process is already at an advanced stage (Bosshardt et al. 1996). NSTI treatment must be aggressive and rapid and essential elements of the treatment are resuscitation, antimicrobial therapy, surgical debridement, and supportive care (Anaya et al. 2005; Morgan 2010; Hakkarainen et al. 2014; Stevens and Bryant 2017), and constant monitoring is required to achieve fluid, electrolyte, and hemodynamic stability (Bosshardt et al. 1996). Hyperbaric oxygen treatment can also be used as adjunctive therapy when infections involve anaerobic bacteria, specifically the clostridial species (Bakker 2012). For instance, AI could be applied to the analysis of microbiological findings. Identification of the etiological agents can assist infection control measures and antimicrobial therapy decision making, and may offer prognostic information (Anaya et al. 2005; Huang et al. 2011; Madsen et al. 2019). The practical applicability of AI methods for the analysis of microbiology finding to aid diagnostic testing has been postulated and discussed for the image analysis including Gram

12

Systems and Precision Medicine in Necrotizing Soft Tissue Infections

195

Fig. 12.2 Information management model for systems and personalized medicine. Figure from (Ganzinger and Knaup 2017), Licensed under Creative Commons

stains (Smith et al. 2020) which could be applied also to NSTI microbiology (see also Fig. 12.2). The role of radiologic imaging in diagnosis of NSTI is debated (Leichtle et al. 2016; Fernando et al. 2019) and several imaging options are available, such as plain radiographs, ultrasonography, computerized tomography (CT), and magnetic resonance imaging (MRI). Although MRI images were previously not used in Big Data context or analyzed with AI methods, those could be easily applied to support and aid clinical decision. For instance, supervised algorithms could be trained on large datasets containing MRI images of patients with NSTI and related conditions, and these algorithms could be applied instantly, contextually with image acquisition and reconstruction, to provide guided diagnosis that could be integrated with other clinical information. Kim et al. (2011) suggested that magnetic resonance imaging could be used to differentiate between necrotizing and non-necrotizing fasciitis. They compared MRI findings between the two groups and found that patients with necrotizing fasciitis had a significantly greater frequency of,

among others, thick (3 mm) abnormal fascial signal intensity on fat-suppressed T2-weighted images, low signal intensity in the deep fascia on fat-suppressed T2-weighted images. CT scanning has also been proposed (Hietbrink et al. 2016) but MRI scanning proves to have the highest sensitivity and specificity (Hietbrink et al. 2016). Rakus-Andersson and Frey (Rakus-Andersson and Frey 2016) trained a modified neural network, to identify groups of NSTI patient with good prognosis of recovering without HBO compared to patients for which HBO could be beneficial, in such a way as to support clinical decision making. They used data from 13 patients admitted to the Blekinge County City Hospital in Karlskrona between 2006 and 2010. The input data consisted of clinical data (non-disclosed) and the results were satisfactory, with an accuracy of 92%; however, these results were obtained on a very limited sample size and have not been crossvalidated. The use of Big omics data in the clinical setting of NSTI to support diagnosis at the bedside is less immediate since these measurement

196

platforms are usually not available in hospital settings: sample work-up, preparation, acquisition, data cleaning, and processing may require several days, which is a major hindrance, given the severity of NSTI in critically ill patients (Peetermans et al. 2020). A notable exception is the use of nuclear magnetic resonance, which requires minimal sample preparation to measure metabolomics profiles of blood and tissues samples collected in the operating theater and that can provide almost real-time information to the surgeons and to clinicians. Such a setting has been implemented at St. Mary Hospital in London (Nicholson et al. 2012). The advent of new technologies is enabling real-time sequencing of large genomes and it is now possible to perform without delay sequencing and analysis of patient genetic information. Implementations of automated phenotyping and interpretation of genome sequencing by beadbased genome library preparation directly from blood samples and sequencing of paired 100-nt reads obtained 15.5 h and used for fast population-scale, provisional diagnosis of genetic diseases of infants in neonatal and pediatric intensive care units, have been reported (Clark et al. 2019). They reported a prospective 100% precision and a mean time saving of 22 h on diagnosis which subsequently affected treatment. If Big omics data are less applicable for bedside decision support due to technical limitations, they are essential when defining strategies for patient stratifications and individualized therapy in NSTI. Indeed, the lack of stratification strategies (one of the cornerstones of precision medicine) is one of the biggest bottlenecks in NSTI management. Omics data from NSTI subjects (and possibly controls), such as transcriptomics, proteomics, lipidomics, and metabolomics profiles measured on blood and tissue biopsies, could be analyzed in relation to pathogens and clinical parameters using multivariate statistics, machine learning, and reverse engineering approaches to identify subgroups of patients demonstrating a survival benefit or favorable response to given treatments.

V. A. P. Martins dos Santos et al.

12.4 12.4.1

Information Management Personalized and Precision Medicine

A key challenge in the fields of personalized and precision medicine is to organize, harmonize, and integrate the vast amount of fragmented clinical and experimental datasets, in order to achieve clinically relevant results and to guide preclinical and clinical research. Typically, data of different sources such as electronic health record (EHR) systems, clinical research databases, or biomedical knowledge representations like (i.e. ontologies) have to be reviewed and prepared. Furthermore, an often overlooked weakness is the use of patient samples and/or omics data despite lack of linked clinical (meta)data, which greatly reduces the usefulness of the studies. Thus, information management is of paramount importance for systems and personalized medicine in research as well as clinical practice. To tackle these challenges, a variety of approaches have been suggested and implemented in different settings. One such model is a three-layer information technology architecture coupled to a cyclic data management approach, as proposed by Ganzinger and Knaup (2017). The generic high-level architecture of such a three-layer model entails: 1. Data representation, 2. Decision support, and 3. User interface. As for data representation (layer 1), data and knowledge from different sources have to be prepared and made available for use in systems medicine. This includes data harmonization, transformation, and storage. In decision support (layer 2), the data and knowledge from layer 1 are processed by applying decision support approaches or systems biology models (see also Sect. 12.3 on Big Data). Systems medicine applications should be designed to assist and not replace human decisions. Consequently, the user interface for such an application (layer 3) must be

12

Systems and Precision Medicine in Necrotizing Soft Tissue Infections

carefully designed to support well-informed, reproducible, clinical decisions in an appropriate time frame. The core concept of this model is the knowledge base, which contains patient and disease-related data as well as formally represented knowledge including a variety of omics and biomedical data. The Decision Support is suggested as case-based and rule-based components (Ganzinger and Knaup 2017), see also Sect. 12.5 below pertaining Decision Support. This model is shown schematically in Fig. 12.2. The complexity of the data management process depends on the level of heterogeneity prevalent in the data sources. To achieve sufficient case numbers it is often necessary to combine data on the same entity types from different sources. For example, in the multi-center approach hospitals decide to collaborate and share clinical data on a specific disease area to build a joint systems medicine application with a higher number of cases and therefore greater statistical power. In most cases, clinical documentation will not be based on identical specifications. Thus, in a harmonization step data definitions have to be evaluated for each attribute, both on a syntactic and semantic level.

12.4.2

The Case of NSTI

In the context of systems and personalized medicine in NSTI, the most comprehensive endeavor thus far is that being undertaken under the scope of a series of national and international research programs comprising teams from hospitals, medical research, academia, and industry (see Chap. 1 and www.permedinfect.com). These programs have built a platform for personalized medicine in acute infectious diseases with focus on NSTI and sepsis that form the basis for development of tools and concepts for refined diagnosis, patient stratification, and individualized treatment. The various participating hospitals have made substantial standardization efforts, so that clinical partners provide patient cohorts, including both clinical registries and associated biobanks, which were used to populate a common information

197

platform. This required previous ethical approvals and written informed consent from all patients [and to be updated and amended as need arises]. It required as well full compliance on Data & Ethics governance, Good Clinical Practice guidelines, and European General Data Protection Regulations (GDPR) (Regulation 2016). The data platform developed includes systems to handle data in separated data domains and stores to protect data ownership and privacy requirements.

12.4.3

Integrating Heterogeneous Data with FAIR Principles

A key feature and goal of such multinational, multi-center projects like INFECT, PERAID, and PerMIT is to ensure that all data resources (both institutions’ own and public) are properly integrated into a common framework. A distributed data lookup and retrieval service allow users to select relevant datasets for inclusion in analysis based on not only matching metadata but importantly clinical and biological relevance and this adds a vital level of quality control ensuring that only clinically relevant datasets are included. These encompass also demographics, clinical, and treatment aspects. Strict clinical case definition criteria, as well as source of infection and severity scores such as simplified acute physiology score (SAPS) and sequential organ failure assessment (SOFA) are used for precise patient classifications, whereas microbiological results will be documented and allow for stratification according to etiology and virulence properties. A key component is the inclusion of heterogeneous omics and biochemical datasets, which are ultimately essential for establishing the mechanisms underlying clinical conditions. The carefully curated studies/datasets are integrated with a minimal data model for meta-data exposure using a Resource Description Framework (RDF) model (Lassila and Swick 1998), which is further empowered by using distributed search and indexing technologies such as Apache Lucene (Białecki et al. 2012).

198

V. A. P. Martins dos Santos et al.

Data fusion and standardization is ensured throughout the use of specialized data fusion algorithms and well-functioning interactions among participating institutions. The use of a RDF data model to manage data provenance and storage ensures that data comply with the FAIR guidelines for data management. Minimum information checklists are used to facilitate interpretation and reproducibility of results ensuring the inclusion of the relevant meta-data. All data types are (should be) to be represented in ontologies so that they can be integrated and remain interoperable as the types and size of data increase. A systematic treatment of the data is ensured by the use of ontologies devoted to clinical and disease-related terms, such as Systematized Nomenclature of Medicine-Clinical Terms (SNOMED CT) (Cote 1986; Lussier et al. 1998). The structure of the underlying data resources is assessed using tools such as RDF2Graph (van Dam et al. 2015) to ensure that newly integrated datasets readily fulfill the quality standards and indicators. Operation of the www.permedinfect.com platform enables both handling of heterogeneous data and the standardization of operating procedures, data, and models, as well as their storage and stewardship. The structure and workflows comply with those required by the European Union on data management of research (European Commission 2016). The platform itself aligns with international data stewardship infrastructure such as the ESFRI ELIXIR (elixir-europe.org) or Nordic e-Infrastructure Collaboration (NeIC) program (https://neic.no/).

12.4.4

Laying the Basis for Computer-Assisted Decision Support

Any CDSS has to rely on a data platform (see, for instance, the HUNT platform, www.ntnu.edu/

hunt/data which inform the PerAID and PerMIT data projects), allowing different layers with different level of access and different levels of granularity at the data level. At the first level, a close interaction between the clinicians and support personnel exists. The interaction on the electronic devices is optimized using the state-of-the-art technology to monitor and to optimize human-computer interfaces. The second level prepares anonymized databases from level 1 using software that can process various types of data such as patient records, data collected in various forms, and biobank data to a format that can be included in the RDF databases through previous standardization. This software is to be used within the hospitals in a protected environment and only the anonymized data will be transferred to the second level. This data is then to be further digitalized and abstracted to the third level that will represent a statistical view on the original data. The first and second level are unique for a particular clinical center, while data at the third level will be pooled across centers. The threelayered data structure helps to overcome the main limitations of medical support systems, which are usually not scalable and not interconnected. Data are translated to standard English vocabulary while keeping the original granularity of the data and third layer data will be shareable between different clinical sites. This allows the CDSS to be scalable and used in other studied and countries by accounting for local language requirements (see Sect. 12.5). Then the third layer, containing highly processed data, can be made public and used to predict treatment outcomes and contributing to best practice management of the disease. The CDSS will therefore combine the information of all patients in the patient cohorts, thereby allowing for interoperability and scalability of the system.

12

Systems and Precision Medicine in Necrotizing Soft Tissue Infections

12.5

12.5.1

Clinical Decision Support Systems for Soft Tissue Infections The Need to Enhance Medical Decisions

The quantity and quality of clinical data are expanding rapidly, including EHRs, disease registries, patient surveys, and information exchanges. Also, burgeoning amounts of data are becoming available for each patient, as is the increasing body of medical and basic sciences evidence. Hence, clinicians need tools to help them make rational decisions based on all these information (Wasylewicz and Scheepers-Hoeks 2019). Big Data and digitalization, however, does not automatically mean better patient care. Several studies have shown that only implementing an EHR and computerized physician order entry (CPOE) has rapidly decreased the incidence of certain errors, introducing, however, many more (Magrabi et al. 2016). Therefore, high-quality clinical decision support is essential if healthcare organizations are to achieve the full benefits of EHR and CPOE. In the current healthcare setting, healthcare providers often do not know that certain patient data are available in the EHR, do not always know how to access these data, do not have the time to search for them, or are not fully informed on the most current medical insights when facing a decision. It is said the healthcare providers often drown in the midst of plenty (Mamlin and Tierney 2016; Frost and Sullivan 2015; Bresnick 2016). Moreover, decisions by healthcare professionals are often made in conjunction with/as part of direct patient contact, ward rounds, or multidisciplinary meetings. This means that many decisions are made in a matter of seconds or minutes. This way, their quality depends on the healthcare provider having all patient parameters and medical knowledge readily available at the time of the decision. Consequently, current decisions are still strongly confounded by experience and knowledge of the professional.

199

Computer technology and algorithms can assist by generating case-specific advice for clinical decision making. The systems used are usually referred to as CDSS, and are thus intended to improve health care delivery by enhancing medical decisions with targeted clinical knowledge, patient information, and other molecular or health information (Wasylewicz and Scheepers-Hoeks 2019).

12.5.2

What Are CDSS What Is Their Use

A traditional CDSS is comprised of software designed to be a direct aid to clinical decision making, in which the characteristics of an individual patient are matched to a computerized clinical knowledge base. Patient-specific assessments or recommendations are then presented to the clinician for a decision (Sutton et al. 2020). From a historical point of view, medication-related CDSS has been used for a long time and is still the most advanced (Garg et al. 2005). CDSSs today are primarily used at the point-of-care, for the clinician to combine their knowledge with information or suggestions provided by the CDSS. However, there are CDSS being developed with the capability to leverage data and observations otherwise unobtainable or uninterpretable by humans (Wasylewicz and Scheepers-Hoeks 2019). Current CDSS often makes use of web-applications or integration with EHR and CPOE systems. They can be administered not only through desktop, tablet, smartphone but also through other devices such as biometric monitoring and wearable health technology. These devices may or may not produce outputs directly on the device or be linked into EHR databases (Dias and Paulo Silva Cunha 2018). The scope of functions provided by CDSS is vast, including diagnostics, alarm systems, disease management, prescription, drug control, and much more. CDSS ranges from personal digital assistant applications customized by a single clinician to multihospital mainframe-based surveillance systems meant to assure care for

200

V. A. P. Martins dos Santos et al.

Fig. 12.3 Clinical algorithm for suspected fasciitis, as suggested by Hietbrink and co-workers (Hietbrink et al. 2016). The algorithm is used for gate specialties in patients with suspected necrotizing fasciitis. It consists of awareness, early surgical exploration, and early initiation of treatment. When incision biopsy is indicated, the patient

is transported to the operation room for further treatment. Treatment and aftercare are multidisciplinary. Analysis of frozen section, microbiological findings, and biopsy could be supported by AI technologies. Figure and caption from Hietbrink et al. (2016), Licensed under Creative Commons

thousands of patients (Pusic and Ansermino 2004). They can manifest as computerized alerts and reminders, computerized guidelines, order sets, patient data reports, documentation templates, and clinical workflow tools. Regarding the nature of the interaction with the clinician, CDSS may be categorized into those that entail solicited information (e.g. a clinician asking for specific advice for a given condition) or unsolicited information (deliver information or knowledge that beneficially can alter clinical decision making (Pusic and Ansermino 2004)). Such applications can be particular useful if CDSS builds upon robust and flexible data integration and includes a wide variety of data analytics. Such data analytics, in particular those based on AI, Machine Learning, and other

modelling tools, can be essential for rapid diagnosis, stratification, and assistance on decision regarding disease treatment and intervention strategies to be applied. The benefits of CDSS, possible pitfalls, and evidence-based mitigation strategies to overcome have been published recently by Sutton et al. (2020) and discussed as well by Pusic and Ansermino (2004) and (Wasylewicz and Scheepers-Hoeks (2019).

12.5.3

CDSS in NSTI

Algorithmic procedures to handle NSTI patients have been defined (Hietbrink et al. 2016; Peetermans et al. 2020), see Fig. 12.3, but there

12

Systems and Precision Medicine in Necrotizing Soft Tissue Infections

are currently no CDSS or modules thereof dedicated to NSTI. Such modules or built-in tools would possibly be very useful for a range of uses, including those that could be included in CDSSs or applications routinely used by general practitioners, as these are often the first point of entry of potential patients, which are frequently misdiagnosed due to lack of familiarity with the disease (Goh et al. 2014). Indeed, due to the fast progression of the disease, NSTI management requires fast decisions to determine the most appropriate course of action. Such decisions, or part of them, could be performed or supported by trained algorithms, e.g. aiding in the main diagnostic problem, that is to differentiate a lesion requiring surgery from a lesion for which conservative treatment will be sufficient. Furthermore, tasks like selection of antibiotic therapy and adjunctive therapy could be performed or optimized by dedicated algorithms, as suggested in the closely relating field of sepsis management (Komorowski 2020). Simple algorithmic procedures are already routinely applied in the management of NSTI. An example of this is the LRINEC (Laboratory Risk Indicator for Necrotizing Fasciitis) score which is generated from six routinely performed laboratory tests including the analyses of patients´ C-reactive protein, white blood cell count, hemoglobin, sodium, creatinine, glucose. Wong et al. (2004) proposed a tool to distinguishing NSTI from the other severe soft tissue infections. However, recently its performance was evaluated in a prospective cohort study and data were displayed discouraging its use (Hsiao et al. 2020). An older study within the National Surgical Quality Improvement Program (NSQIP, USA) was used to determine data on the incidence, treatment, and outcomes of NSTIs (Mills et al. 2010). Partly on the basis thereof, a 30-day postoperative mortality risk calculator for patients with NSTI was developed and validated using a cohort of 1392 identified NSTI cases, of which 42% were female, median age was 55 years, and median body mass index was 32 kg/m2 (Mills et al. 2010). Thirtyday mortality was 13%. Seven independent variables were identified that correlated with mortality: age older than 60 years, functional status, requiring dialysis, American Society of

201

Anesthesiologists class 4 or higher, emergent surgery, septic shock, and low platelet count. The receiver operating characteristic area was 0.85 (95% CI 0.82-0.87), reflecting a reasonably strong prediction. Using bootstrap validation, the optimism-corrected receiver operating characteristic area was 0.83 (95% CI 0.81–0.86), which was used to develop an interactive risk calculator for future patients. Although not a CDSS, this correlation was nevertheless useful for stratification, according to the authors. Another cohort study was reported (Hua et al. 2015) that included 109 patients with a confirmed diagnosis of NSTI, a median follow-up of 274 days (range 2–6135 days) and of which 31 (28%) died. On multivariate analysis, independent risk factors of mortality were age older than 75 years, multifocal NSTI, severe peripheral vascular disease, hospital-acquired infection, severe sepsis, and septic shock on hospital admission. Although a retrospective cohort, which disallows a precise record of the delay between diagnosis and surgery, these analyses could help building information to develop a true CDSS and help clinicians stratify NSTI severity at clinical course onset. A triple diagnostic procedure has been proposed to manage NSTI (Hietbrink et al. 2016) which combines the analysis of microscopic findings on tissue biopsies together with Gram staining to assess the presence, gram staining, characteristic arrangements, and morphology of microorganisms and analysis of fresh frozen sections to detect necrosis of the superficial fascia with fibrinous thrombi of arteries and veins located in the fascia. The algorithmic procedure is described in Fig. 12.3, in which many steps could be replaced or supported by AI- and Big Data-informed decision.

12.5.4

Current and Future Developments in Relation to Dedicated CDSS for NSTI

Specific efforts to develop and deploy innovative clinical support tools for patient stratification and bedside decisions suitable for the emergency and intensive care setting for NSTI are currently being

202

V. A. P. Martins dos Santos et al.

E e.g xtern . lit al era dat MedTrender tur a, e

Clinical Genetic

INFECT cohort + New data from PerMIT / PerAID

Metabolomic Health records?

Questionnaires?

Consider overlap!

Patient

ML models (+ others?) PerMIT / PerAID Database

Findable Accessible Interoperable Reusable

Other?

Personalised (pre-emptive) treatment

CDSS running on Nordic platform Toolbox Automated scoring Mortality prediction Severity assessment ...

Expert knowledge

Interface (App)

Improved decisionmaking

Fig. 12.4 Simplified representation of a CDSS for NSTI under development in the scope of the projects PerAID and PerMIT (Figure reproduced from permedinfect.com, author’s copyright)

undertaken in a series of national and international research projects comprising teams from hospitals, medical research, academia, and industry (see permedinfect.com). One specific task is the development of Machine Learning and AI models for the prediction of different outcomes in NSTI, such as risk of sepsis, septic shock, or amputation, using a combination of clinical parameters including SAPS and SOFA severity scores and biomarkers that can be measured during the acute stages (e.g. in the emergency room), thereby enabling rapid bedside decisions. The same data approach is being used to identify an optimized scoring of NSTI patients, to overcome some of sensitivity issues identified with the severity score LRINEC indicated above (Hansen et al. 2017). Development of automatic calculations of such severity score is part of the basis for applications offering personalized decision support. Considering that different hospitals have their own sepsis alert system, such automatic calculations linked to the patients charts and

sepsis alarm are considered as a means to achieve a more rapid and optimized identification of patients. This type of systems can help prioritizing severe cases and thereby reduce the clinical burden and efficient use of hospital beds. An effective and efficient CDSS is of particular importance for the stratification of patients and personalized therapy in NSTI. Several dedicated modules are currently under development (www. permedinfect.com) for: (1) the integration of a variety of quantitative or qualitative models (i.e. statistical models, algorithms, etc.) to enable a CDSS to perform data analytics (see previous chapter and section on AI and Big Data above); (2) aggregation of reasoning processes from the domain and inference capabilities (e.g. rule-base and case-based systems) to handle the data/information (i.e. clinical and experimental parameters); and (3) user interfaces to match the need of practitioners in the clinic. The software architecture model for the NSTIdedicated CDSS, the representation of interoperable clinical knowledge, and inference engine are

Systems and Precision Medicine in Necrotizing Soft Tissue Infections

203

C

al

t

cu

la

ea Tr

te

PatientDB

12

Predict

Mine

Fig. 12.5 Representation of a simple mobile app for use by clinicians as a front-end of a CDSS for NSTI (LifeGlimmer GmbH). It enables data mining, querying

databases, calculating specific scores, predicting outcomes and generally supports the clinician for decisions

Example with 2 variables predicting “death after 1 yea ar”

Enter data

Predict

Fig. 12.6 Example of the use of the app as support to beside decision. Based on over 2000 current clinical parameters and by using Machine Learning algorithms (tested in the INFECT patient cohort), it predicts severity

(e.g. 90 days mortality) of NSTI patients with high accuracy, supporting thereby patient stratification for differentiated treatment (LifeGlimmer GmbH)

being designed to form a base for a CDSS framework of wider applicability. The CDSS functionalities are being iteratively developed through requirement-adjustment-development-

validation cycles using enterprise-grade software-engineering methodologies and technologies. The CDSS prepares views for the clinicians and supports personnel with data that

204

V. A. P. Martins dos Santos et al.

allow patient stratification and individualized therapy in NSTI/sepsis. In addition to individual patient data, the system provides information of patient classification into groups with similar disease status and comparable clinical and biological parameters, which are of great value for continued research. A simplified representation of the envisaged CDSS for NSTI is shown in Fig. 12.4. An example of an initial prototype for a frontend app to be directly used by a clinical practitioner is depicted in Figs. 12.5 and 12.6 (www. lifeglimmer.com). These modules are currently under development and testing and it is expected that they will be eventually implemented in the clinical practice, with potentially substantial benefits for the patient and disease management.

12.6

Conclusion and Perspectives

The clinical and biochemical research progress over the last decades have provided a burgeoning body of information on the possible mechanisms underlying NSTI, on the clinical manifestation of this fast-developing disease, and on individualized patient characteristics. The organization and translation of this information into actionable knowledge requires concerted, multidisciplinary efforts and accessible computational systems that assist decision making. Altogether, the efforts, platforms, and variety of modules herein described for systems and personalized medicine in acute infectious diseases form the basis for development of tools and concepts for refined diagnosis, patient stratification, and individualized treatment. Thereby, such approaches hold great promise for accurate and rapid diagnosing and improving outcome in NSTI, as well as potential to increase costefficacy, as it will promote optimized tailored therapy. Acknowledgments The work was supported by the European Union Seventh Framework Programme: (FP7/2007-2013) under the grant agreement 305340 (INFECT project); the Swedish Governmental Agency

for Innovation Systems (VINNOVA), Innovation Fund Denmark and the Research Council of Norway under the frame of NordForsk (Project no. 90456, PerAID), and the Swedish Research Council, Innovation Fund Denmark, the Research Council of Norway, the Netherlands Organisation for Health Research and Development (ZonMW), and DLR Federal Ministry of Education and Research, under the frame of ERA PerMed (Project 2018151, PerMIT).

References Afzal M, Saccenti E, Madsen MB, Hansen MB, Hyldegaard O, Skrede S, Martins Dos Santos VAP, Norrby-Teglund A, Svensson M (2020) Integrated univariate, multivariate, and correlation-based network analyses reveal metabolite-specific effects on bacterial growth and biofilm formation in necrotizing soft tissue infections. J Proteome Res 19:688–698. https://doi. org/10.1021/acs.jproteome.9b00565 Ahn AC, Tewari M, Poon C-S, Phillips RS (2006) The limits of reductionism in medicine: could systems biology offer an alternative? PLoS Med 3:e208 Anaya DA, Mcmahon K, Nathens AB, Sullivan SR, Foy H, Bulger E (2005) Predictors of mortality and limb loss in necrotizing soft tissue infections. Arch Surg 140:151–157 Assfalg M, Bertini I, Colangiuli D, Luchinat C, Schafer H, Schutz B, Spraul M (2008) Evidence of different metabolic phenotypes in humans. Proc Natl Acad Sci USA 105:1420–1424 Auffray C, Charron D, Hood L (2010) Predictive, preventive, personalized and participatory medicine: back to the future. Genome Med 2(8):57. https://doi.org/10. 1186/gm178 Bahl M, Barzilay R, Yedidia AB, Locascio NJ, Yu L, Lehman CD (2018) High-risk breast lesions: a machine learning model to predict pathologic upgrade and reduce unnecessary surgical excision. Radiology 286:810–818 Bakker DJ (2012) Clostridial myonecrosis (gas gangrene). Undersea Hyperb Med 39:731 Bergsten H, Madsen MB, Bergey F, Hyldegaard O, Skrede S, Arnell P, Oppegaard O, Itzek A, Perner A, Svensson M (2020) Correlation between immunoglobulin dose administered and plasma neutralization of streptococcal superantigens in patients with necrotizing soft tissue infections. Clin Infect Dis 9:ciaa022 Bernini P, Bertini I, Luchinat C, Nepi S, Saccenti E, Scha Fer H, Schutz B, Spraul M, Tenori L (2009) Individual human phenotypes in metabolic space and time. J Proteome Res 8:4264–4271 Berthold MR, Borgelt C, Höppner F, Klawonn F (2010) Guide to intelligent data analysis: how to intelligently make sense of real data. Springer, New York Białecki A, Muir R, Ingersoll G (2012) Apache lucene 4. SIGIR 2012 workshop on open source information retrieval, p 17

12

Systems and Precision Medicine in Necrotizing Soft Tissue Infections

Bosshardt TL, Henderson VJ, Organ CH (1996) Necrotizing soft-tissue infections. Arch Surg 131:846–854 Bresnick J (2016) The difference between big data and smart data in healthcare. Health Anal. https:// healthitanalytics.com/features/the-difference-betweenbig-data-and-smart-data-in-healthcare. Accessed May 2020 Bresnick J (2018) What are precision medicine and personalized medicine? Health analytics [online]. https://healthitanalytics.com/features/whatareprecision-medicine-and-personalized-medicine. Accessed May 2020 Brighi R (2018) The quality and veracity of digital data on health: from electronic health records to big data. Revista de Bioética y Derecho 42:163–179 Chella Krishnan K, Mukundan S, Alagarsamy J, Hur J, Nookala S, Siemens N, Svensson M, Hyldegaard O, Norrby-Teglund A, Kotb M (2016) Genetic architecture of group A streptococcal necrotizing soft tissue infections in the mouse. PLoS Pathog 12:e1005732 Clark MM, Hildreth A, Batalov S, Ding Y, Chowdhury S, Watkins K, Ellsworth K, Camp B, Kint CI, Yacoubian C, Farnaes L, Bainbridge MN, Beebe C, Braun JJA, Bray M, Carroll J, Cakici JA, Caylor SA, Clarke C, Creed MP, Friedman J, Frith A, Gain R, Gaughran M, George S, Gilmer S, Gleeson J, Gore J, Grunenwald H, Hovey RL, Janes ML, Lin K, McDonagh PD, McBride K, Mulrooney P, Nahas S, Oh D, Oriol A, Puckett L, Rady Z, Reese MG, Ryu J, Salz L, Sanford E, Stewart L, Sweeney N, Tokita M, Van der Kraan L, White S, Wigby K, Williams B, Wong T, Wright MS, Yamada C, Schols P, Reynders J, Hall K, Dimmock D, Veeraraghavan N, Defay T, Kingsmore SF (2019) Diagnosis of genetic diseases in seriously ill children by rapid wholegenome sequencing and automated phenotyping and interpretation. Sci Transl Med 11:6177 Cote RA (1986) Architecture of SNOMED: its contribution to medical language processing. Proceedings of the annual symposium on computer application in medical care, American Medical Informatics Association, p 74 Davenport EE, Burnham KL, Radhakrishnan J, Humburg P, Hutton P, Mills TC, Rautanen A, Gordon AC, Garrard C, Hill AV (2016) Genomic landscape of the individual host response and outcomes in sepsis: a prospective cohort study. Lancet Respir Med 4:259–271 Dias D, Paulo Silva Cunha J (2018) Wearable health devices—vital sign monitoring, systems and technologies. Sensors 18:2414 Erikainen S, Chan S (2019) Contested futures: envisioning “Personalized,” “Stratified,” and “Precision” medicine. New Genet Soc 38:308–330 Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, Thrun S (2017) Dermatologist-level classification of skin cancer with deep neural networks. Nature 542:115–118

205

European Commission (2016) Guidelines on FAIR data management in horizon 2020. https://www.google.com/ url?sa=t&rct=j&q=&esrc=s&source=web&cd=&ved= 2ahUKEwjOrb3ZgNTrAhWQmIsKHQ2xBhsQFjAAe gQIAhAB&url=https%3A%2F%2Fec.europa.eu% 2Fresearch%2Fparticipants%2Fdata%2Fref% 2Fh2020%2Fgrants_manual%2Fhi%2Foa_pilot% 2Fh2020-hi-oa-datamgt_en.pdf&usg= AOvVaw0RrG7kc_LE3Hp74lceZDA5 Fernando SM, Tran A, Cheng W, Rochwerg B, Kyeremanteng K, Seely AJ, Inaba K, Perry JJ (2019) Necrotizing soft tissue infection: diagnostic accuracy of physical examination, imaging, and LRINEC score: a systematic review and meta-analysis. Ann Surg 269:58–65 Frost & Sullivan (2015) Drowning in big data? Reducing information technology complexities and costs for healthcare organizations, Frost & Sullivan [Online]. https://www.emc.com/collateral/analyst-reports/frostsullivan-reducing-information-technologycomplexities-ar.pdf. Accessed 7 May 2017 Ganzinger M, Knaup P (2017) Information management for enabling systems medicine. Curr Direct Biomed Eng 3:501–504 Garg AX, Adhikari NK, McDonald H, Rosas-Arellano MP, Devereaux PJ, Beyene J, Sam J, Haynes RB (2005) Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA 293:1223–1238 Goh T, Goh L, Ang C, Wong C (2014) Early diagnosis of necrotizing fasciitis. Br J Surg 101:e119–e125 Goldstein EJ, Anaya DA, Dellinger EP (2007) Necrotizing soft-tissue infection: diagnosis and management. Clin Infect Dis 44:705–710 Gruden K, Hren M, Herman A, Blejec A, Albrecht T, Selbig J, Bauer C, Schuchardt J, Or-Guil M, Zupančič K (2012) A “crossomics” study analysing variability of different components in peripheral blood of healthy caucasoid individuals. PLoS One 7:e28761 Hakkarainen TW, Kopari NM, Pham TN, Evans HL (2014) Necrotizing soft tissue infections: review and current concepts in treatment, systems of care, and outcomes. Curr Probl Surg 51:344 Hansen MB, Rasmussen LS, Svensson M, Chakrakodi B, Bruun T, Madsen MB, Perner A, Garred P, Hyldegaard O, Norrby-Teglund A, INFECT Study Group, Nekludov M, Arnell P, Rosén A, Oscarsson N, Karlsson Y, Oppegaard O, Skrede S, Itzek A, Wahl AM, Hedetoft M, Bærnthsen NF, Müller R, Nedrebø T (2017) Association between cytokine response, the LRINEC score and outcome in patients with necrotising soft tissue infection: a multicentre, prospective study. Sci Rep 7:42179 Hietbrink F, Bode LG, Riddez L, Leenen LPH, Van Dijk MR (2016) Triple diagnostics for early detection of ambivalent necrotizing fasciitis. World J Emerg Surg 11:51 Hotchkiss RS, Karl IE (2003) The pathophysiology and treatment of sepsis. N Engl J Med 348:138–150

206 Hsiao C-T, Chang C-P, Huang T-Y, Chen Y-C, Fann W-C (2020) Prospective validation of the laboratory risk indicator for necrotizing fasciitis (LRINEC) score for necrotizing fasciitis of the extremities. PLoS One 15: e0227748 Hua C, Sbidian E, Hemery F, Decousser JW, Bosc R, Amathieu R, Rahmouni A, Wolkenstein P, ValeyrieAllanore L, Brun-Buisson C (2015) Prognostic factors in necrotizing soft-tissue infections (NSTI): a cohort study. J Am Acad Dermatol 73:1006–1012 Huang K-F, Hung M-H, Lin Y-S, Lu C-L, Liu C, Chen C-C, Lee Y-H (2011) Independent predictors of mortality for necrotizing fasciitis: a retrospective analysis in a single institution. J Trauma Acute Care Surg 71:467–473 Hughes DA, Kircher M, He Z, Guo S, Fairbrother GL, Moreno CS, Khaitovich P, Stoneking M (2015) Evaluating intra-and inter-individual variation in the human placental transcriptome. Genome Biol 16:54 Hulsen T, Jamuar SS, Moody A, Karnes JH, Orsolya V, Hedensted S, Spreafico R, Hafler DA, McKinney EF (2019) From big data to precision medicine. Front Med 6:34 Jolliffe I (1982) A note on the use of principal components in regression. J Roy Stat Soc C-App 31:300–303 Kim K-T, Kim YJ, Won Lee J, Kim YJ, Park S-W, Lim MK, Suh CH (2011) Can necrotizing infectious fasciitis be differentiated from nonnecrotizing infectious fasciitis with MR imaging? Radiology 259:816–824 Kittang BR, Langeland N, Skrede S, Mylvaganam H (2010) Two unusual cases of severe soft tissue infection caused by Streptococcus dysgalactiae subsp. equisimilis. J Clin Microbiol 48:1484–1487 Komorowski M (2020) Clinical management of sepsis can be improved by artificial intelligence: yes. Intensive Care Med 46:375–377 Lassila O, Swick RR (1998) Resource description framework (RDF) model and syntax specification. World Wide Web Consortium, Cambridge Lazăr A, Georgescu AM, Vitin A, Azamfirei L (2019) Precision medicine and its role in the treatment of sepsis: a personalised view. J Crit Care Med 5:90–96 Leichtle SW, Tung L, Khan M, Inaba K, Demetriades D (2016) The role of radiologic evaluation in necrotizing soft tissue infections. J Trauma Acute Care Surg 81:921–924 Lussier YA, Rothwell D, Cote R (1998) The SNOMED model: a knowledge source for the controlled terminology of the computerized patient record. Methods Inf Med 37:161–164 Madsen MB, Skrede S, Bruun T, Arnell P, Rosén A, Nekludov M, Karlsson Y, Bergey F, Saccenti E, Martins dos Santos VAP, Perner A, Norrby-Teglund A, Hyldegaard O (2018) Necrotizing soft tissue infections—a multicentre, prospective observational study (INFECT): protocol and statistical analysis plan. Acta Anaesthesiol Scand 62:272–279 Madsen MB, Skrede S, Perner A, Arnell P, Nekludov M, Bruun T, Karlsson Y, Hansen MB, Polzik P, Hedetoft

V. A. P. Martins dos Santos et al. M (2019) Patient’s characteristics and outcomes in necrotising soft-tissue infections: results from a Scandinavian, multicentre, prospective cohort study. Intensive Care Med 45:1241–1251 Magrabi F, Ammenwerth E, Hyppönen H, de Keizer N, Nykänen P, Rigby M, Scott P, Talmon J, Georgiou A (2016) Improving evaluation to address the unintended consequences of health information technology. Yearb Med Inform 25:61–69 Mamlin BW, Tierney WM (2016) The promise of information and communication technology in healthcare: extracting value from the chaos. Am J Med Sci 351:59–68 Meier-Schellersheim M, Fraser ID, Klauschen F (2009) Multiscale modeling for biologists. Wiley Interdiscip Rev Syst Biol Med 1:4–14 Mills MK, Faraklas I, Davis C, Stoddard GJ, Saffle J (2010) Outcomes from treatment of necrotizing softtissue infections: results from the National Surgical Quality Improvement Program database. Am J Surg 200:790–797 Morgan M (2010) Diagnosis and management of necrotising fasciitis: a multiparametric approach. J Hosp Infect 75:249–257 Nicholson JK, Holmes E, Kinross JM, Darzi AW, Takats Z, Lindon JC (2012) Metabolic phenotyping in clinical and surgical environments. Nature 491:384–392 Nimmesgern E, Benediktsson I, Norstedt I (2017) Personalized medicine in Europe. Clin Transl Sci 10:61–63 Noble D (2008) Claude Bernard, the first systems biologist, and the future of physiology. Exp Physiol 93:16–26 Patcas R, Bernini DA, Volokitin A, Agustsson E, Rothe R, Timofte R (2019) Applying artificial intelligence to assess the impact of orthognathic treatment on facial attractiveness and estimated age. Int J Oral Maxillofac Surg 48:77–83 Peetermans M, de Prost N, Eckmann C, Norrby-TeglundA, Skrede S, de Waele JJ (2020) Necrotizing skin and soft-tissue infections in the intensive care unit. Clin Microbiol Infect 26:8–17 Phillips M, Marsden H, Jaffe W, Matin RN, Wali GN, Greenhalgh J, Mcgrath E, James R, Ladoyanni E, Bewley A, Argenziano G, Palamaras I (2019) Assessment of accuracy of an artificial intelligence algorithm to detect melanoma in images of skin lesions. JAMA Netw Open 2:e1913436–e1913436 Phillips M, Greenhalgh J, Marsden H, Palamaras I (2020) Detection of malignant melanoma using artificial intelligence: an observational study of diagnostic accuracy. Dermatology Practical & Conceptual 10:e2020011 Pusic M, Ansermino M (2004) Clinical decision support systems. B C Med J 46:236–239 Rajpurkar P, Irvin J, Zhu K, Yang B, Mehta H, Duan T, Ding D, Bagul A, Langlotz C, Shpanskaya K (2017) Chexnet: radiologist-level pneumonia detection on chest X-rays with deep learning. Preprint arXiv: 1711.05225

12

Systems and Precision Medicine in Necrotizing Soft Tissue Infections

Rakus-Andersson E, Frey J (2016) Fuzzy one-decision making model with fuzzified outcomes in the treatment of necrotizing fasciitis, eTELEMED, Venice, Italy, 2016. International Academy, Research and Industry Association (IARIA), pp 145–152 Regulation GDP (2016) Regulation (EU) 2016/679 of the European Parliament and of the council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing directive 95/46. Official J European Union 59:294 Rello J, Van Engelen T, Alp E, Calandra T, Cattoir V, Kern W, Netea M, Nseir S, Opal S, Van de Veerdonk F (2018) Towards precision medicine in sepsis: a position paper from the European Society of Clinical Microbiology and Infectious Diseases. Clin Microbiol Infect 24:1264–1272 Rosato A, Tenori L, Cascante M, de Atauri Carulla PR, Martins dos Santos VAP, Saccenti E (2018) From correlation to causation: analysis of metabolomics data using systems biology approaches. Metabolomics 14:37 Sarani B, Strong M, Pascual J, Schwab CW (2009) Necrotizing fasciitis: current concepts and review of the literature. J Am Coll Surg 208:279–288 Schroeck M, Shockley R, Smart J, Romero-Morales D, Tufano P (2012) Analytics: the real-world use of big data: how innovative enterprises extract value from uncertain data, Executive Report. IBM Institute for Business Value and Said Business School at the University of Oxford Smith KP, Wang H, Durant TJ, Mathison BA, Sharp SE, Kirby JE, Long SW, Rhoads DD (2020) Applications of artificial intelligence in clinical microbiology diagnostic testing. Clin Microbiol Newsl 42:61–70 Stacy A, McNally L, Darch SE, Brown SP, Whiteley M (2016) The biogeography of polymicrobial infection. Nat Rev Microbiol 14:93 Stevens DL, Bryant AE (2017) Necrotizing soft-tissue infections. N Engl J Med 377:2253–2265

207

Sutton RT, Pincock D, Baumgart DC, Sadowski DC, Fedorak RN, Kroeker KI (2020) An overview of clinical decision support systems: benefits, risks, and strategies for success. NPJ Digital Med 3:1–10 Thänert R, Itzek A, Hoßmann J, Hamisch D, Madsen MB, Hyldegaard O, Skrede S, Bruun T, Norrby-Teglund A, Medina E (2019) Molecular profiling of tissue biopsies reveals unique signatures associated with streptococcal necrotizing soft tissue infections. Nat Commun 10:1–15 Tillmann T, Gibson RA, Scott G, Harrison O, Dominiczak A, Hanlon P (2015) Systems medicine 2.0: potential benefits of combining electronic health care records with systems science models. J Med Internet Res 17:e64 Tøndel K, Indahl UG, Gjuvsland AB, Omholt SW, Martens H (2012) Multi-way metamodelling facilitates insight into the complex input-output maps of nonlinear dynamic models. BMC Syst Biol 6:88 Union C. O. T. E (2015) Council conclusions on personalised medicine for patients. Official J European Union 421:2 van Dam JC, Koehorst JJ, Schaap PJ, Martins dos Santos VA, Suarez-DIEZ M (2015) RDF2Graph a tool to recover, understand and validate the ontology of an RDF resource. J Biomed Semant 6:1–12 Vitabile S, Marks M, Stojanovic D, Pllana S, Molina JM, Krzyszton M, Sikora A, Jarynowski A, Hosseinpour F, Jakobik A (2019) Medical data processing and analysis for remote health and activities monitoring. In: Highperformance modelling and simulation for big data applications. Springer, Cham Wasylewicz A, Scheepers-Hoeks A (2019) Clinical decision support systems. In: Fundamentals of clinical data science. Springer, Cham Wold S, Eriksson L (2001) PLS-regression: a basic tool of chemometrics. Chemom Intel Lab Syst 58:109–130 Wong C-H, Khin L-W, Heng K-S, Tan K-C, Low C-O (2004) The LRINEC (laboratory risk indicator for necrotizing fasciitis) score: a tool for distinguishing necrotizing fasciitis from other soft tissue infections. Crit Care Med 32:1535–1541