Worming Your Way Into Bioavailability: Modelling the Uptake of Organic Chemicals in Earthworms 903933384X, 9789039333846

Citation preview

Worming your way into bioavailability Modelling the uptake of organic chemicals in earthworms

Tjalling Jager

Worming your way into bioavailability Modelling the uptake of organic chemicals in earthworms

Wurm je een weg door biobeschikbaarheid Het modelleren van de opname van organische stoffen in regenwormen (met een samenvatting in het Nederlands)

Proefschrift ter verkrijging van de graad van doctor aan de Universiteit Utrecht op gezag van de Rector Magnificus, Prof. Dr. W.H. Gispen, ingevolge het besluit van het College voor Promoties in het openbaar te verdedigen op woensdag 11 juni 2003 des middags te 14.30 uur door

Dirk Tjalling Jager geboren op 12 augustus 1969, te Purmerend

Promotor:

Prof. Dr. C.J. van Leeuwen (Institute for Risk Assessment Sciences, Utrecht University)

Co-promotor: Dr. J.L.M. Hermens (Institute for Risk Assessment Sciences, Utrecht University)

Worming your way into bioavailability. Modelling the uptake of organic chemicals in earthworms / Tjalling Jager ISBN 90-393-3384-X

The research described in this thesis was partly carried out at the National Institute for Public Health and the Environment (RIVM, Bilthoven, The Netherlands), Laboratory for Ecotoxicology, in cooperation with the Institute for Risk Assessment Sciences (IRAS, Utrecht, The Netherlands). Part of this research was commissioned by the Ministry of Housing, Spatial Planning and the Environment (VROM, The Hague, The Netherlands).

Cover drawing: Inside:

Livingstone ©BIODIDAC http://biodidac.bio.uottawa.ca/ Chlorine-free, recycled paper (100% post-consumer waste)

“To sum up, as chance does not determine the manner in which objects are drawn into the burrows, and as the existence of specialized instincts for each particular case cannot be admitted, the first and most natural supposition is that worms try all methods until they at last succeed; but many appearances are opposed to such a supposition. One alternative alone is left, namely that worms, although standing low in the scale of organization, possess some degree of intelligence. This will strike every one as very improbable; but it may be doubted whether we know enough about the nervous system of the lower animals to justify our natural distrust of such a conclusion. With respect to the small size of the cerebral ganglia, we should remember what a mass of inherited knowledge, with some power of adapting means to an end, is crowded into the minute brain of a worker-ant.” Charles Darwin (1881) The formation of vegetable mould, through the action of worms, with observations on their habits.

Contents 1.

General Introduction ......................................................................................................... 1

2. 3.

Using Compartment Models for Bioaccumulation ..................................................... 23 Mechanistic Approach for Estimating Bioconcentration of Organic Chemicals in Earthworms (Oligochaeta) ....................................................... 59

4.

Toxicokinetics of Polycyclic Aromatic Hydrocarbons in Eisenia andrei (Oligochaeta) using Spiked Soil .......................................................................... 79 Availability of Polycyclic Aromatic Hydrocarbons to Earthworms (Eisenia andrei, Oligochaeta) in Field-Polluted Soils and SoilSediment Mixtures ........................................................................................................... 95 Assessing Bioavailability of Organic Chemicals in Contaminated Soils, Evaluation of Bioassays with Earthworms ...................................................... 111 Solid Phase MicroExtraction to Predict Bioavailability and Accumulation in a Field-Contaminated Soil .............................................................. 125

Section A: Theoretical

Section B: Case Studies 5.

6. 7.

Section C: Gut Uptake 8. 9. 10.

11.

Modelling Ingestion as an Exposure Route for Organic Chemicals in Earthworms (Oligochaeta) ....................................................................................... 141 Feeding Activity of the Earthworm Eisenia andrei in Artificial Soil ........................ 155 Elucidating the Routes of Exposure for Organic Chemicals in the Earthworm, Eisenia andrei (Oligochaeta) ..................................................................... 169

Summary and General Discussion .............................................................................. 187

Samenvatting in het Nederlands ............................................................................................ 203 Curriculum Vitae ........................................................................................................................ 210 List of Publications ................................................................................................................... 211 Dankwoord ............................................................................................................................... 213 Explanation of Symbol Use ..................................................................................................... 216

General introduction

1 General Introduction

ABSTRACT  Soil pollution is a substantial problem in many countries. Total concentrations in soil are poor measures for uptake and toxicity, and therefore, to what degree the contaminants are bioavailable requires study. Earthworms are abundant in many soils and are appropriate model organisms because they live in close contact with the soil, have a thin, permeable cuticle, and consume large amounts of soil. Furthermore, these organisms contribute to soil fertility and are on the diet of many birds, mammals, and invertebrate predators. In this thesis, it is investigated how organic chemicals are taken up in earthworms, and to what extent. Furthermore, models are developed to predict accumulation in earthworms through various routes of exposure. This chapter introduces the problems and origins of soil pollution, and also the animal group that is the focus of this thesis: the earthworms. A brief exposé on earthworm ecology is provided, including the classification in different ecological groups and basic physiology. This knowledge is helpful in interpreting the results from bioaccumulation studies as presented in this thesis. Further, the leading theory on uptake of organic chemicals (equilibrium partitioning) is introduced, although a more thorough discussion is postponed to later chapters. This chapter concludes with an outline of this thesis and a definition of the most important terminology.

1

Chapter 1

ORIGINS OF SOIL POLLUTION Soil pollution is, strictly speaking, as old as the soil itself. However, before human interference, the pollution used to be restricted to specific areas, such as locations with superficial metal ores, sites of volcanic activity, and tar pits. Human activities have changed this pattern over the last few thousand years. Even though heavy metals are naturally occurring compounds, metal pollution has become more prominent, particularly since roman times, due to mining and smelting activities. Several organic soil pollutants, like polycyclic aromatic hydrocarbons (PAHs), are also naturally occurring chemicals, originating from incomplete combustion (e.g. as a result of forest fires). However, with the start of the twentieth century, we have witnessed a dramatic increase in soil levels, coinciding with an extensive increase in the use of fossil fuels [36]. Several of these PAHs are carcinogenic, which makes them particularly interesting from a human health perspective. In the beginning of the twentieth century, the industrial production of organic chemicals started for various applications. As an example, polychlorinated biphenyls (PCBs) were produced for electrical transformers and capacitors, but these compounds turned out to be not only toxic, but also extremely persistent in the environment. These compounds travel all the way around the globe, and considerable levels of PCBs and other organochlorine chemicals are, for example, detected in the fat of polar bears [60]. Currently, the emissions of PCBs are declining, but large numbers of new industrial chemicals are produced each year: between 300 and 350 new chemicals are notified in Europe annually1. To name just a few categories: surfactants (e.g. for use in washing powders), photochemicals, intermediates (chemicals used in production of other chemicals), paints and dyes, polymers (and chemicals involved in polymerisation and polymer processing), and flame retardants. With the declining emissions of PCBs, the problems with persistent and toxic chemicals have not become a thing of the past; also among the currently used chemicals, there are suspect groups. For instance, perfluorinated surfactants have recently created a lot of interest. These compounds are used in lubricants, paints and fire-fighting foams, and are highly persistent in the environment, are widely distributed across the globe, and accumulate in the food chain [28,37,56]. These industrial chemicals can reach the soil through various routes: directly (dumping, leakage and spilling), atmospheric emission followed by wet and dry deposition, and via sewage sludge (see Fig. 1). In many countries, emissions to surface water are treated in sewage treatment works. Hydrophobic chemicals that are poorly degraded end up in the sewage sludge and in several countries, this sludge is applied as fertiliser on agricultural soil; a process that can lead to substantial contamination of agricultural soils [21]. In the Netherlands, sewage sludge from municipal plants is no longer applied in agriculture since 1995. However, from the industrial treatment plants, a part of the sludge is used in agriculture (13%), including a small fraction of sludge that is derived from chemical industries [11]. In contrast, nearly 50% of the sludge that is produced in German municipal plants is used in agriculture, usually after a treatment step [78]. This application of sewage sludge has led to contamination of the soil with heavy metals, PCBs, dioxins and furans. In recent years, the levels of some of these compounds are declining, owing to effective legislation, but a new class of chemicals, the polybrominated diphenyl ethers (PBDE), are being increasingly dispersed in the soil environment through this route [54]. Other chemical inputs into the soil result from the dredging of regional waters to ensure

1

2

See http://ecb.jrc.it/new-chemicals

General introduction

sufficient water depth for drainage and navigation. These sediments are placed on soil where they may contribute to the pollution, because sediments act like a sink for airtransported chemicals like PAHs [27]. The risks of dredge materials should however not be exaggerated, as placing sediments on land may also enhance biodegradation of these compounds [26]. The chemicals that are used industrially or in consumer products do not deliberately end up in the soil. Another important group of toxic pollutants are, however, directly and intentionally emitted: the chemicals used in agriculture to deal with unwanted weeds and pest organisms (herbicides, pesticides, fungicides, also euphemistically called “plantprotection products”). The well-known insecticide DDT was developed in 19391, and widely used after the second world war until the 1960’s. It was not only used in agriculture, but also for vector control against malaria, and as contact poison against head louse. However, the long-lasting effectiveness that made this chemical so popular, also led to serious problems. DDT (and its metabolites) are highly persistent and accumulate in food chains, leading to severe effects on bird populations, especially birds of prey [34,81]. In response to these problems, DDT was banned in most western countries, although it is still used in third-world countries to combat malaria (and its production thus continues). However, 20 years after the last application, levels in orchard soils are still high enough to pose a threat to birds feeding on earthworms [33]. In the Netherlands, banning took place in 1973, but in a large number of soil samples taken between 1993 and 1995, the levels still exceeded the quality objective [32]. For agricultural locations, all of the top soil samples exceeded the quality objective, by up to a factor of 63. Similar conclusions were drawn for other notorious pesticides like dieldrin (banned in 1980). Clearly, we are still seeing the results of the agricultural applications of these chemicals in the soil, decades after their use has seized.

Figure 1. Major sources of soil pollution and removal processes.

The Swiss chemist Paul Muller who invented DDT, received for this feat the Nobel price for medicine in 1948. 1

3

Chapter 1

Even though most of the persistent pesticides are no longer in use today (at least, in the western world), this does not mean that the remaining agricultural chemicals are harmless. In the Netherlands, we are still facing an annual pesticide application on agricultural soil of nearly 12 million kg per year (mainly herbicides and fungicides) [46]. Given approximately 2 million ha of agricultural soil, this makes for an average emission of 6 kg active ingredient per ha per year. However, a much higher use per hectare is normal in the cultivation of flower bulbs (47 kg/ha/yr), and flowers in greenhouses (28 kg/ha/yr)1. These chemicals are intended to kill living organisms, and place a heavy burden on the soil community. The impact of pesticides is not restricted to the area where they are applied. Even the pesticides that are currently used are regularly found in rain and air samples, and are transported in the atmosphere to remote areas, hundreds of kilometres away [86]. The volume of agricultural chemicals that is used has decreased over the past years (in 1990, it was 20 million kg), but this is partly due to a higher efficacy (read: toxicity) of the chemicals. This is therefore not necessarily a positive trend for soil organisms. On the other hand, a more positive development is that new pesticides are generally rapidly degraded. This implies that future generations are not burdened by these chemicals when the land is used for a different purpose, and that they are unlikely to be transported across the globe. On grasslands, the pesticide load is relatively low, but a serious threat to this soil is the increasing use of pharmaceuticals in farm animals. These chemicals are used, amongst others, in the treatment of ectoparasites (e.g. in dips to treat sheep for blowflies), endoparasites (e.g. anthelmintics against parasitic worms), and bacterial diseases (antibiotics, which are also used as growth promoters). Several of these compounds are used externally, or are excreted with the faeces, and may thus end up on the grassland soil (or on arable land when the manure is used as fertiliser). As example, avermectins are used in the treatment of parasitic worms, and these chemicals are to a large extent excreted with the faeces, where they pose a threat for dung-eating animals [76]. The problem with the risk assessment for these compounds is that the main focus lies on the residues in milk and meat for human consumption; the risk assessment for soil organisms is only recently developing [55]. Careless use and disposal of chemicals has led to large numbers of polluted sites throughout the world. Some of the polluted sites in the Netherlands go back to the middle ages (e.g. heavy metal pollution in the heart of old cities, related to the production of leather, glass and paint), but most of the seriously polluted sites originate from practices between 1960 and 1980, including dumping of chemical waste, sites of former gas works, and deposits for dredge spoil from harbours. For example, dredge spoil from the Rotterdam harbour has led to the serious soil pollution in polder De Esch, with particularly high concentrations of the persistent pesticides dieldrin and telodrin (see Chapters 6 and 7). The largest and probably best known polluted site in the Netherlands is De Volgermeerpolder, north of Amsterdam. This site has been used as a refuse dump from 1927 until its closing in 1981. However, apart from household waste, the site was used for illegal dumping of industrial chemical waste in the 1960’s (in particular, waste resulting from pesticide production). Currently, the site is heavily polluted with metals, PAHs, benzene, chlorobenzenes, dioxins, lindane and DDT (and its metabolites), and plans are drawn up for its cleanup2. “Cleanup” may not be the best choice of words as the authorities decided to isolate the pollution by applying a clean layer of soil on top of the contamination. Total costs for remediation: 100 million euro. Another well-known site is Het Griftpark in Utrecht; the former location of a coal gas facility, and a waste-handling 1 2

4

See http://www.cbs.nl See http://www.milieudienst.amsterdam.nl/bodem/volgermeerpolder/ (in Dutch)

General introduction

site for the city. The gasworks have left a thick layer of tar with aromatic hydrocarbons in the soil, while at the waste-handling site, all kinds of chemicals (like toluene and sulphuric acid) were carelessly spilled1. The choice was made to isolate the pollution, building a dam wall of 1200 metres long, and a maximum depth of 64 metres. The water from this box is pumped out and subsequently treated. On top of the pollution, a clean soil layer of 1–1.6 metres is placed, and the site is turned into a park. Total costs for remediation: 110 million euro. However, not all polluted sites are this extensive; in 2001, over 200 petrol stations in the Netherlands were cleaned up [69]. To give an overview of the seriousness of the situation in the Netherlands, the total number of polluted sites is estimated at around 175,000, of which probably 35% is urgent with regard to soil cleanup2. In the period 1980–2000, 900 soil cleanups were performed by the competent authorities (provincial and the four largest cities), and 4,300 by the private sector. The current rate of cleanup is over 1000 sites per year, with an area of more than 100 ha, and 1.8 million tonnes of soil released [69]. Soil cleanup is an economic problem; the total costs for soil cleanup in the Netherlands are estimated at 40 billion euro, and 360 million euro was spent in 2001 (of which less than half is paid by the private sector) [69]. In the United States, the situation is different because the parties responsible for the pollution are always the ones that have to pay for cleaning them up. The government only has to prove that the polluter is actually responsible for the contamination, even when his/her actions were within the laws of the time. This very strong legal ground results from the CERCLA: Comprehensive Environmental Response, Compensation, and Liability Act of 1980, better known as Superfund. The name Superfund comes from the trust fund that this law authorised, containing 1.6 billion dollar, and in 1986 increased to 8.5 billion dollar. This fund allows the Environmental Protection Agency to conduct the cleanup of a site itself (and recover the costs from the responsible parties later), without waiting for the court to determine who was responsible. Between 1980 and 2000, cleanup has been completed for 757 Superfund sites, and private parties settlements for over 18 billion dollar were achieved [83]. However, it must be noted that a remediation is not just an expensive process, and a drain for the economy; cleanup is also an ecological problem. Many remediation techniques are quite destructive for the soil community, and it may take decades before such a site has developed a “representative” flora and fauna. Soil pollution differs in several respects from air and water pollution. Most organic chemicals are degraded to some extent in air by reaction with OH-radicals (e.g. the lifetime of PCBs in air is in the order of months [93]). Both air and water are (at least part of the time) turbulently mixed, implying that air and water pollution can be rapidly diluted into a larger volume. The importance of dilution is vividly illustrated by its absence: e.g. during smog episodes in big cities, when there is little wind. In soil, on the other hand, little transport of the pollutants is possible. Water-soluble chemicals can be transported to the ground water and travel with ground water plumes, but hydrophobic compounds tend to be immobile and persistent. This is mainly caused by their absorption to soil particles. Soil sorption is correlated to hydrophobicity [71], and chemicals sorbed to soil particles are not available for degradation by micro-organisms [95] or transfer to the ground water. This particular behaviour of soil pollution has two consequences: firstly, the pollution tends to be localised in areas where application, spilling or dumping has occurred, and tends to stay there. On the other hand, concentrations tend to be high (due to limited mixing with clean soil), and stay high for long periods of time. As an 1 2

See http://ublad.warande.uu.nl/~ublad/ubladen/29/21/1213Grift.html (in Dutch) See http://arch.rivm.nl/environmentaldata/ 5

Chapter 1

illustration, the sites where mining and smelting activities were located in roman times are still contaminated today, and in the Netherlands, activities in the seventies have led to contaminated sites that still continue to worry the authorities today. Clearly, soil pollution is a serious problem in many countries and it is important to understand how the chemicals affect organisms.

THE IMPORTANCE OF EARTHWORMS Earthworms form a large part of the invertebrate biomass in most temperate soil systems. Exceptions are arctic areas, extremely sandy soils (like dunes and deserts) and acidic soils. Numbers and biomass in several soils are shown in Table 1. They reach especially large numbers in grasslands, whereas arable land usually contains fewer worms (probably due to a lack of food) [24]. In the 19th century, it was gradually established in the scientific community that earthworms are beneficial and essential for soil fertility. The most influential contribution was probably made by Darwin [16], recognising the importance of earthworms in the formation of the humus layer, and the burial of stones and ruins. Earthworms improve fertility of the soil through their contribution to decomposition and nutrient cycling, soil aeration and drainage [24,41]. Without worms, soils often develop a thick layer of undecomposed litter at the surface [24]. Table 1. Numbers, biomass, and diversity of earthworms in various habitats [62]. Environment Cultivated fields Orchards Pastures and grasslands Meadows and alfalfas Deciduous forests and woodlands Coniferous forests and woodlands Tropical and subtropical forests

Numbers (per m2) 67–109 197–227 248–367 148–252 82–185 36–67 64–88

Biomass (gwwt /m2) 25–43 85–90 92–124 46–83 25–56 16–29 12–16

Number of species 1–10 6–15 4–11 2–8 3–11 1–11 7–8

Earthworms not only contribute to soil fertility, they also form an important part of the diet for many birds and mammals (reviewed by [51]). Well-known birds that feed on worms are the blackbird (Turdus merula), song thrush (T. philomelos) and starling (Sturnus vulgaris). Also wading birds like the lapwing (Vanellus vanellus) and oystercatcher (Haematopus ostralegus), as well as gulls (Larus sp.), regularly include worms in their diet when foraging on grasslands. Perhaps less well known is that typical predatory birds like the tawny owl (Strix aluco) and the kestrel (Falco tinnunculus) occasionally catch earthworms on the surface. The most specialised mammalian predator on earthworms is without doubt the mole (Talpa europaea), whose diet consist at least half of earthworms (and on some sites up to 100%). Other species that rely heavily on earthworms are the hedgehog (Erinaceous europaeus), badger (Meles meles), red fox (Vulpes vulpes) and shrews (Sorex sp.). These species mostly feed on earthworms on the soil surface, and the surface foraging species Lumbricus terrestris is therefore their main prey. Several amphibians and reptiles also include worms in their diet; earthworms form a (small) part of the diet for toads (Bufo bufo) and frogs (Rana temporaria). Among the invertebrate predators on earthworms, there are ants, centipedes and several carabid and staphylinid beetle species (larvae and adults) [44]. A relatively recent contributor in the United Kingdom are carnivorous flatworms from Australia and New Zealand, introduced in the 1960’s [75].

6

General introduction

The broad range of animals that regularly feed on earthworms is not surprising, given the fact that the worms have a good nutritional value, and are rich in proteins [44]. And, although few people in the western world would be volunteering, they also constitute a good potential food source for humans (see for example the recipe in [41]). In several parts of the world, consumption of earthworms in various forms is more common. For the indigenous people of the Amazonian rain forest in Venezuela, consumption of smoked earthworms is an important part of their tradition, and a vital contribution to their diet [53]. Furthermore, earthworms can also be used commercially as food for rearing rainbow trout [77].

SPECIES AND CLASSIFICATION Earthworms belong to the phylum Annelida, class of Oligochaeta. To the casual observer, it may appear that there is only one species of earthworm. However, there are world-wide some 3000 species [75], although in the Netherlands only some fifteen are common (personal observations). Most of the European species belong to the family Lumbricidae, and these species have been introduced by the human colonisers in all parts of the world. Even within this family, the species are quite different in their ecology, which means that scientists need to be able to classify them. This seems to be a sensitive point. For example, Darwin [16] never specified the earthworms he was studying, and in some field studies all species are pooled into one sample [30,57]. Although earthworm taxonomy is quite well studied (at least in Europe), identification can be severely hampered by the fact that several species are quite flexible morphologically (see [75] on Aporrectodea caliginosa, Allolobophora chlorotica and Dendrodrilus rubidus), also because polyploid subspecies are quite common in the Lumbricidae [92]. This makes it difficult to judge whether we are dealing with a race or an entirely new species (and leads to problems regarding nomenclature). For instance, modern electrophoresis techniques indicated that the turgida and tuberculata morphs of A. caliginosa are in fact different species (see [68]). Table 2. Classification of earthworms according to Bouché, adapted from [75].

Habitat

Litter dwellers (epigeic species) Usually outside the mineral substrata, in litter, compost etc.

Deep burrowers (anecic species) Deep vertical burrows (several metres), feeding on the surface at night

Colouration

Usually red or rosy

Brown to brown black (especially anterior)

Mobility and respiration rate

Highly mobile and a relatively high respiration rate

Diapause

Unknown, only quiescence Lumbricus rubellus, Dendrodrilus rubidus

Usually less active than litter dwellers, but capable of rapid retreat into the burrow and surface migration Common, often obligatory Lumbricus terrestris, Aporrectodea longa

Typical WestEuropean species

Horizontal burrowers (endogeic species) Horizontal burrows in the mineral soil, ingesting organic materials with soil Pigmentation absent, often blue-grey or flesh coloured Least active of all groups

Common Aporrectodea caliginosa, Allolobophora chlorotica

Although any classification is to some degree artificial, the one proposed by Bouché is probably best known and widely used (see [75]). Bouché distinguished the three groups

7

Chapter 1

shown in Table 2, according to behaviour and morphology. Another classification stems from Perel [65], dividing the earthworms into two major groups based on their feeding habits: the humus formers (feeding on slightly decomposed plant matter), and the humus feeders (feeding on well-decomposed matter, dispersed in the soil). This difference in feeding habits is reflected in the folded typhlosole (a longitudinal fold in the lumen of the intestine, providing additional surface area for absorption, see Fig. 2) of the humus feeders, whereas the humus formers have simpler, unfolded, typhlosoles. Perel divided these groups even further, based upon their vertical distribution in the soil. Such classification may be useful in understanding any inter-species differences in toxicity and accumulation, but some care should be taken. Analysis of stable carbon and nitrogen isotopes indicated that classifications may be site specific (although the general distinction in humus formers and humus feeders was confirmed) [58].

O

C

G I T

P C

L

N

Figure 2. Lateral anterior (left) and transverse (right) section through a lumbricid earthworm. The anterior view showing the pharynx (P), oesophagus (O), crop (C), gizzard (G), and the start of the intestine (I). The transverse view showing the coelom (C), typhlosole (T), gut lumen (L), and a nephridium with nephridiopore (N). Reproduced with permission, J. Soucie  BIODIDAC1.

PHYSIOLOGY: FEEDING AND WATER RELATIONS Generally, earthworms can be considered detrivores. The geophageous (soil-eating) species (the “humus feeders” sensu Perel) consume soil and digest organic constituents. Although these species consume large amounts of soil, they do not feed indiscriminately, but are able to select specific fractions from the soil that are rich in organic matter [8]. Their crop contents (Fig. 2) are rich in well-decomposed detritus [66], and it seems that living fungi and bacteria, ingested with the soil, also play a nutritional role [44]. Furthermore, several typical geophages, like A. caliginosa, are also known to feed on dung at the soil surface [3,44]. Some of the information of food preference and diets is contradictory, and it is likely that diet in the same species differs between sites [44]. The litter feeders (or “humus formers”), feed mostly on slightly decomposed matter at the soil surface, like dung and leaf litter. Their crop contents contain relatively undecomposed plant remains [66], but mineral soil also seems to be an essential part of the diet [20]. Furthermore, some species (presumably L. terrestris) are also keen on fat, meat, and even dead worms [16]. Several typical species from this group live outside the soil under the 1

8

See http://biodidac.bio.uottawa.ca/

General introduction

bark of dead trees, in the litter layer, or in man-made accumulations of organic debris (manure and compost heaps). Most striking examples of this group are the popular test species Eisenia andrei and E. fetida1 that are seldom, if ever, found in the soil2. The earthworm’s digestive system is relatively simple. as the alimentary canal passes like a straight tube from mouth to anus. The first part is shown in Figure 2. In the first four body segments, it is differentiated into a pharynx (used as a suction pump in feeding). The pharynx is followed by a oesophagus, a thin-walled crop (used to store food, before it is passed to the subsequent gut parts), and a muscular gizzard (used for grinding up food, aided by mineral particles). The gizzard opens up into the intestine, which forms the largest part of the alimentary canal, and where most of the digestion and absorption takes place. The surface area in the intestine is enlarged due to many small folds and a large longitudinal dorsal downfold; the typhlosole. The size and shape of this fold differs between species, and between ecological groups (the classification by Perel [65] uses the shape of the typhlosole as characteristic). The pH along the entire stretch from mouth to anus is quite constant between 6 and 7. Digestion in the gut therefore cannot be acidic, but is enzymatic [24], and the enzyme activity includes protease, cellulase and chitinase [43]. These enzymes are indeed produced by the worms themselves, and not only by microorganisms in the gut [63]. Moisture is a very important precondition for earthworm populations, limiting their dispersal and activity. In fact, earthworms more closely resemble aquatic than terrestrial animals, given their respiratory and excretory physiology [44]. Respiration occurs through the body surface, without any special structures, and relies on a thin layer of moisture at the interface [24,43]. In soil, this mechanism leads to large water losses. Furthermore, earthworms excrete nitrogen as toxic ammonia and urea that have to be diluted to urine [43]. The water losses in earthworms in soil are estimated to amount 10–20% of the body weight per day [44]. Exposed to air, earthworms rapidly dry out and die. At 70–80% relative humidity, L. terrestris lost more than 50% of its body water in 24 hours (and two out of 10 died) [44]. This water loss must be replenished across the skin and from the gut contents. Several species of earthworm can survive dry periods (i.e. summer and winter) by entering a period of diapause or quiescence (the difference in the two lies mainly in the cues used to enter or stop the resting period) [24]. They curl into a tight ball and surround themselves with mucus secretions [43] or moist faeces [75], which protects them from desiccation. The epigeic species usually do not survive periods of drought, but the cocoons are much more tolerant to desiccation, and hatch when conditions are more favourable.

EARTHWORMS IN ECOTOXICOLOGY Given their importance for soil fertility and the terrestrial food chain, it is not surprising that earthworms play a prominent role in soil ecotoxicology. Furthermore, earthworms can be considered as model organisms. The body residues in earthworms reflect the local pollution status because of their limited mobility and intimate relationship to the soil (owing to their thin, permeable cuticle, and the fact that most species consume soil in large quantities). In many cases, earthworms also seem to accumulate more chemicals The invalid name E. foetida is still often used. For this reason, Paoletti [62] places these species in a separate category (apart from the ones mentioned in Table 2): the coprophagic (dung-eating) species.

1 2

9

Chapter 1

than other soil invertebrates [17,64,84], and can therefore be considered as a worst case for toxicant availability. It is therefore not surprising that earthworms have been proposed as bioindicators for soil pollution and sustainable land use [25,62]. A more pragmatic reason for their popularity is their size and ease of culturing and handling, making them appropriate test organisms. However, soil ecotoxicology lagged behind aquatic ecotoxicology for a long time, and probably still does. This interest in the aquatic may be prompted by the obvious visible effects of extreme toxic insult (massive fish kills in rivers). Soil-dwelling animals are not so visible, and are not directly interesting from a commercial or recreational perspective. Although not as spectacular as floating fish, effects on the soil community can severely decrease the soil’s ability to support a healthy above-ground ecosystem or agricultural crops. Furthermore, as earthworms are on the diet of so many interesting wildlife species, their body burden of pollutants should be carefully considered. Around 1960, it became clear that earthworms take up organochlorine residues from the soil, and that they experience toxic effects (reviewed by [25]). In the second half of the sixties, many papers appeared, reporting levels of organochlorine insecticides (especially DDT and drins) in soils and earthworms (a.o. [17,29,94]). In the Netherlands, Van Rhee, reviewing ten years of earthworm ecotoxicology studies, warned in 1977 about the toxic effects that chemicals are having on earthworm populations, and the potential consequences for soil functions and predating animals [91]. Most of these studies focussed on the potential toxic effects on wildlife of agricultural practices, also because most of the organochlorine pesticides are not very toxic to earthworms [25]. Cases of wildlife poisoning after feeding on earthworms have indeed been observed in relation to pesticide use [14]. These hazards do not only result from the use of persistent compounds like DDT [33], but also for more degradable compounds like diazinon [12]. When the use of DDT became severely restricted in the 1970’s, it was replaced by other insecticides that have much more profound effects on earthworms, such as carbaryl [25,41]. Earthworms appear to be particularly sensitive to the majority of fumigants, the majority of carbamate insecticides and fungicides (e.g. benomyl, carbaryl, aldicarb), and some organophosphate insecticides (especially phorate) [25,41]. Most of the soil-ecotoxicological work from the 1960’s and 1970’s was not performed very mechanistically, as most studies were in fact observations in response to the problems related to agricultural practices. One of the first mechanistic studies on toxicokinetics in earthworms was the experimental work by Davis in 1971 [18], establishing the importance of feeding behaviour and organic matter content in accumulation. More detailed studies appeared in the 1980’s, with laboratory work investigating the mechanisms of uptake and metabolism (e.g. [10,49,79]), and more structural work on toxicity and testing (especially, the work presented in the PhD thesis by Van Gestel [89]). However, it was not until the early 1990’s that the study of the accumulation kinetics in earthworm reached the level of detail as common in aquatic ecotoxicology (with as most prominent example the PhD thesis by Belfroid [4]). Clearly, toxicity testing with earthworms can learn from experience gained in aquatic studies; e.g. Lanno & McCarty [42] proposed to adopt several concepts from aquatic toxicology in soil testing, like the use of critical body burden and incipient lethal levels. The head start of aquatic ecotoxicology is also obvious in chemical risk assessment. For new industrial chemicals, three acute tests with aquatic species (algae, Daphnia, and fish) are required, whereas the soil assessment must initially rely on extrapolations from the aquatic data [23]. Nevertheless, earthworms are the only soil animal that became a regular part of risk assessment, and an OECD test guideline for acute toxicity testing with earthworms was published in 1984 [61], and an ISO test for reproduction in 1998 [35]. The 10

General introduction

acute test is regularly used for new pesticides, but risk assessors for new industrial chemicals can only request this test when a chemical is produced at more than 10 tonnes per year (see [47]). The most popular test species is without doubt Eisenia andrei and its close relative E. fetida. Their popularity is largely prompted by their rapid growth and large fecundity. Although the natural habitat of this species is limited to accumulations of organic debris (litter and manure heaps), its performance in (artificial) soil is very good. However, it is not only the worms themselves that are at risk, but also the organisms that feed on them. For this reason, Romijn et al. [70] proposed the inclusion of the pathway soil-earthworm-bird/mammal in the derivation of environmental quality criteria, to account for secondary poisoning. This route is also included in the European risk assessment guidelines [23], and is quantified using estimated body residues in earthworms.

UPTAKE ROUTES AND MECHANISMS As was argued in the previous sections, earthworms tend to accumulate large amounts of chemicals from the soil due to their behaviour and physiology. A thorough discussion on the mechanisms of uptake is postponed to the following chapters, but here, a general outline will be provided. The leading theory on uptake of chemicals by soil- and sediment-dwelling organisms is the equilibrium partitioning (EP) theory, formulated and broadly adopted around 1990 [19,73,85,90]. Basically, this approach states that organisms do not take up chemicals from soils or sediments directly, but only from the freelydissolved phase in the pore water. A chemical will tend to distribute itself between the soil, water and organism phases until it is in thermodynamic equilibrium. This implies that the chemical residues in organisms can be predicted when we know the sorption coefficient of the chemical (partitioning between solids and water) and the bioconcentration factor (partitioning between water and organism). Furthermore, we have to assume that the system is in equilibrium. As the pore water concentration in soil or sediment uptake/elimination represents the bioavailable phase, it is possible to derive quality standards for sediment from the available data for aquatic organisms. Because aquatic toxicity data are available for most chemicals (and sediment data are not), all that is needed for sediment quality standards is a good estimate for the sorption partition sorption/desorption coefficient (assuming that sediment organisms are equally sensitive as aquatic organisms).

worm worm

water water

soil soil

This approach is not limited to sediment. Long before EP became established in ecotoxicology, it was already known that for earthworms, uptake of organic chemicals is a passive process, because uptake is linearly related to soil concentrations (in the same soil) [18,49], and that pore water is the important bioavailable phase. The relationship with pore water was first indicated because of the correlation between uptake and organic matter (the main sorption site for organic chemicals) [18]. Van Gestel & Ma [87,88] took this finding one step further by directly linking accumulation and toxicity to concentrations in the soil solution, confirming the general picture of EP for earthworms in soil. Later, the usefulness of EP for earthworms has been demonstrated for a broader range of chemicals [13]. Equilibrium partitioning has thus become an integral part of chemical risk

11

Chapter 1

assessment for soil and sediment (see e.g. [23]), to predict toxicity (from aquatic data) as well as body residues (from total concentrations) in soil- or sediment-dwelling organisms. Currently, the term EP is often used in a broader sense, relaxing the precondition of equilibrium, and denoting the fact that (time-varying) concentrations in organisms can be predicted from the (time-varying) concentrations in pore water. Despite its popularity, also limitations of EP have been observed (extensively reviewed by [7]). The most striking deviations are discussed below. Sequestration or “ageing” is the process by uptake/elimination which chemicals tend to become less available in time (for uptake by organisms as well as by “soft” chemical extraction techniques) [2]. The most likely mechanism for this behaviour is that the chemical is moving deeper into the organic matrix with increasing contact time [67]. Sequestration has been presented as a sorption/desorption deviation from EP [7], but in fact it is a strong support. Granted, the use of sequestration equations where sorption is estimated from hydrophobicity will fail to predict the effects of sequestration, but EP (in the broad sense) appears to be quite robust as long as good estimates or measurements of pore-water concentrations are available for the specific situation of interest [38]. Desorption from sediments was found to be tri-phasic, although in some media, only two phases are observed [82]. The rapidly desorbing fraction appears to be the one that is linearly partitioning with the pore water [15], and is best related to body residues in organisms [39].

worm worm

water water

soil soil soil soil soil soil

Another deviation from EP that is extensively discussed is feeding. Chemicals uptake/elimination are not only taken up by worms from (pore) water through the skin [49], but also from the gut (e.g. [5]). It is a generally held view feeding that the existence of multiple routes of entry into an organism leads to deviations from EP predictions (see e.g. [7,40,50,73,74]), especially for very hydrophobic chemicals sorption/desorption that have a low solubility in water. It was predicted that feeding becomes an important uptake route for earthworms when log Kow exceeds 5 [6]. For sediment organisms, there is evidence that feeding is important for very hydrophobic chemicals [45], and may lead to deviations from EP up to a factor of 5 (data collected in [48]). However, there are few studies that succeed in experimentally separating both uptake routes, and often conclusions on uptake routes are drawn without confirming that equilibrium was established, and without knowing the actual pore-water concentrations. Furthermore, it is unlikely that chemicals are transferred directly from a solid phase to an organism without intervention of a solution phase. Probably the most important mechanistic work on gut uptake was presented by Gobas and co-workers. These authors demonstrated that, in fish, uptake of organic chemicals from the gut contents also obeys passive diffusion from a solution phase [31]. Digestion of lipids promotes the uptake of chemicals into the body as it reduces the number of sorption sites for hydrophobic

worm worm

water water

soil soil

12

General introduction

compounds, thereby increasing the dissolved concentrations. However, for earthworms, this type of modelling has not yet been attempted. Biotransformation in earthworms may also lead biotransformation to deviations from EP, but this process is not well studied. Biotransformation usually results uptake/elimination in more water-soluble metabolites that are easily excreted. Especially when the exchange with the pore water is slow, even low levels of transformation may affect toxicokinetics. Earthworms do posses the P450 enzyme system, although its activity is low compared to that in fish, mammals and birds [22,80], and is sorption/desorption apparently not induced by for example PAHs [1]. There are some claims of biotransformation in earthworms, judging indirectly from deviations observed in accumulation experiments [9,40], although some cases may also be explained by changes in availability (see Chapter 2). Nevertheless, there is also more direct evidence for pesticides: epoxidation of heptachlor [30] and aldrin [59], and conversion of aldicarb [10] and oxamyl [79].

worm worm

water water

soil soil

Despite the limitations, EP is still the reference theory for discussing the accumulation of organic chemicals in soil and sediment organisms. This implies that body residues observed in earthworms first have to be related to pore water concentrations before alternative theories can be explored. This is also the strategy, followed in this thesis.

OUTLINE OF THIS THESIS Given the numerous sources of pollution that threaten the soil ecosystem, we need to understand how chemicals are accumulated, and to what extent. Understanding the uptake into earthworms relates to the protection of species (earthworms, as well as their predators), and the functioning of soil processes. Furthermore, earthworms can act as a model organism for other species. The purpose of this thesis is to understand and quantify the uptake of organic compounds into the earthworm. Models on a mechanistic basis are developed and applied to describe the uptake kinetics and the steady-state body residues, and hopefully to allow predictions for risk assessment in field situations. This work entails a rigorous evaluation of EP theory, in the broadest sense; implying both the prediction of body residues from (estimated) partition coefficients, as well as the underlying assumption of pore water as the bioavailable phase. A large part of this thesis is dedicated to uptake from the gut and the contribution of the different uptake routes. Although many researchers attribute deviations from the expected bioaccumulation pattern to uptake from food, the number of quantitative studies into this route is very small. Finally, I will conclude this thesis by an overview of the understanding achieved so far on bioavailability for earthworms: which processes play a role, how can we measure bioavailability, and how should we include these issues in risk assessment. This thesis is divided into three large sections: A. Theoretical. This sections contains two chapters. The first is a treatment of the onecompartment model for bioaccumulation, providing a diffusion-based interpretation of the rate constants. Further, the limitations of this model approach are discussed, and the

13

Chapter 1

resulting accumulation pattern when the basic assumptions are violated (e.g. when the exposure concentration changes in time). The second chapter studies the relationship between chemical hydrophobicity and bioaccumulation in earthworms, using literature data, and providing estimation routines for BCF. B. Case studies. This section contains four experimental studies where bioavailability to earthworms is studied and related to soil concentrations. The first two deal with PAHs (in artificial and field-collected soils), and the specific problems with the bioavailability of these compounds. The third and fourth chapter describe an experimental study with a site polluted with pesticides (drins) and PCBs. This study employed an extended set-up for laboratory bioassays, and compares these results to field-collected worms and a biomimetic approach (SPME fibres). The experimental results are discussed in relation to EP, so this section can act as a verification of EP in practice. C. Gut uptake. The last three chapters deal with uptake from the gut, with the central aims to quantify the contribution of both exposure routes and to assess the deviation from EP. The first of these chapters presents a mechanistic model and attempts to validate it using a study from the literature. In the second, the relevant physiological parameters of the feeding process are experimentally determined, which are applied in the third chapter to validate the model on an experimental dataset for three chemicals. Several chapters have appeared, or will appear, in scientific journals. In this thesis, these papers are reproduced without major changes. However, I have applied a more uniform use of symbols for parameters throughout this work (see last pages of this thesis). The symbols may thus differ from those in the corresponding journal version.

DEFINITION OF TERMINOLOGY Equilibrium partitioning (EP) is the theory that toxic effects and body residues in soiland sediment-dwelling organisms can be predicted from the dissolved concentration in pore water. Strictly speaking, the theory assumes equilibrium between pore water and soil solids. So when a chemical is degraded, we do not have equilibrium and therefore EP is not valid in this situation. I will use the term in a broader sense, indicating the process that internal concentrations result from a fugacity-driven partitioning between pore water and tissues (even when the pore water concentrations are not constant in time). However, an EP-estimate will usually represent the system in equilibrium. Several authors define an EP-estimate with sorption being estimated from Kow. While this is certainly defensible, it must be stressed that any deviation from this estimate does not automatically invalidate EP (it may represent an invalidation of the sorption estimate). The words bioaccumulation, bioconcentration and biomagnification are often used in the literature to describe distinctly different aspects of chemical uptake in organisms. I believe this to be mainly a semantic issue. Furthermore, it is unfortunate that these three terms all seem to imply that the concentration in the organism will be higher than that in the environment (the validity of using a certain term for the uptake process would then depend on the outcome of that process). A further complication is how to define “higher than” when we look at two concentrations in distinctly different compartments. Are we comparing concentrations on a volume basis or dry weight or wet weight or fugacity? I will not dwell on this issue anymore and use the term “bioaccumulation” to describe the

14

General introduction

process of chemicals entering an organism’s tissues, independent of the route of uptake and independent of the extent to which this happens. I will reserve “bioconcentration” when internal levels are related to (pore) water concentrations. The following quantitative expressions are used. The Bioconcentration factor (BCF) is the steady-state concentration in the organism, divided by an exposure concentration in water (which can be pore water or water only). Note that this definition says nothing about the actual route of uptake (which could also be uptake from the gut contents), but only that internal concentrations are related to a water concentration in a quotient. Unit is L water/kg worm (wwt/dwt/lipid). Note that the BCF is never dimensionless (a kg worm is different from a kg water, and we need to specify whether the kilos of worm are wet, dry or lipid). The Biota-soil accumulation factor (BSAF) is the steady-state concentration in the organism, divided by the soil concentration. For neutral organic compounds, the organism concentration may be normalised to lipid content, and the soil concentration may be normalised to organic carbon (OC) or organic matter (OM). Unit is kg soil (dwt/wwt/OM/OC)/kg worm (wwt/dwt/lip). Lipid and OC normalisation is usually preferable as it facilitates comparison between different soils and organisms (assuming that the EP theory is valid). For other compounds than neutral organics, normalisation may not be valid (discussed in more detail in Chapter 11). Hydrophobicity is the degree to which a chemical dislikes being surrounded by water molecules (literally: the fear of water). Hydrophobic chemicals prefer to be surrounded by non-polar solvents like octanol, acetone or lipids. Therefore, hydrophobicity is the main driving force of bioconcentration: a hydrophobic chemical will prefer the organism’s lipid phases above the water environment. On the other hand, a hydrophilic chemicals will prefer the water phase. A standard measure of hydrophobicity is the octanol-water partition coefficient (Kow), i.e. the concentration in octanol divided by the concentration in water in an equilibrium system. As this parameter can range over many orders of magnitude, it is common practice to report the 10log of the Kow. Even though I will not model anything explicitly on fugacity basis, the term fugacity is central to all chemical transport modelling. Fugacity is the “urge” of a chemical to flee from a compartment. Chemical diffusion will transport the chemical from a location with a high fugacity to a location with a lower one, until the fugacity is equal in all locations. Fugacities are related to concentrations by the fugacity capacity, a measure of how much the chemical “likes” the environment it is in. This is explained in more detail in Chapter 2. Even though the term fugacity is already used a hundred years in chemical engineering [72], in environmental chemistry, modelling the chemical transport between media through fugacity was pioneered by Mackay (see e.g. [52]). An important distinction is that between equilibrium and steady state. Even though both terms indicate a situation in which the concentration in a compartment is not changing in time, the reasons for the stable situation differ. In equilibrium, there is no net diffusion between two compartments because there is no fugacity difference. Steady state, in contrast, is a stable situation where there is still diffusive transport between the compartments and a fugacity gradient is maintained. As an example, when the organism is growing exponentially, growth dilution can lead to a steady-state situation. The BCF is in such a situation is lower than that in an organism that does not grow. Bioavailability is that fraction of the chemical in a medium that can be taken up by an organism, which can be species specific and time dependent. This definition is rather vague and not very quantitative, but is used to denote a process, recognising that some 15

Chapter 1

chemical fractions are not taken up. Time must be included in the definition in one way or another. E.g. the chemical in pore water can be considered the bioavailable phase for earthworms, but when the earthworm depletes the pore water, the organic matter pool can be considered as indirectly bioavailable. Most often the term bioavailability is used to denote the fraction that can be instantaneously taken up. The term is also used qualitatively: “The chemical in soil A is less bioavailable than in soil B because the animals accumulate less at the same total soil concentration.” Also note that bioavailability needs an organism (hence, the “bio”). No chemical extraction method can be used to measure bioavailability unless it is shown that it correlates with an (or a particular species of) organism. The experimental BSAF for a species can be used as an operational measure of bioavailability for a chemical in soil or sediment. A compartment is, in this context, a spatially confined part of the environment or an organism that is considered well-mixed with respect to the chemical concentration. Furthermore, this part of the environment is only a compartment when it is treated as a state variable. To illustrate: the pore water is only a compartment when it is linked to the organism and its concentration can change due to uptake and elimination by the organism (otherwise it is a constant or a forcing function).

REFERENCES [1]

[2] [3] [4] [5]

[6]

[7] [8] [9]

[10] [11] [12]

[13] [14]

16

Achazi RK, C Flenner, DR Livingstone, LD Peters, K Schaub and E Scheiwe (1998). Cytochrome P450 and dependent activities in unexposed and PAH-exposed terrestrial annelids. Comp. Biochem. Physiol. C 121:339-350. Alexander M (2000). Aging, bioavailability, and overestimation of risk from environmental pollutants. Environ. Sci. Technol. 34:4259-4265. Barley KP (1958). The influence of earthworms on soil fertility II. Consumption of soil and organic matter by the earthworm Allolobophora caliginosa (Savigny). Aust. J. Agric. Res. 10:179-185. Belfroid A (1994). Toxicokinetics of hydrophobic chemicals in earthworms. Validation of the equilibrium partitioning theory. PhD thesis, University of Utrecht, Utrecht, The Netherlands. Belfroid A, J Meiling, D Sijm, J Hermens, W Seinen and K Van Gestel (1994). Uptake of hydrophobic halogenated aromatic hydrocarbons from food by earthworms (Eisenia andrei). Arch. Environ. Contam. Toxicol. 27:260-265. Belfroid AC, W Seinen, KCAM Van Gestel, JLM Hermens and KJ Van Leeuwen (1995). Modelling the accumulation of hydrophobic organic chemicals in earthworms - Application of the equilibrium partitioning theory. Environ. Sci. Poll. Res. 2:5-15. Belfroid AC, DTHM Sijm and CAM Van Gestel (1996). Bioavailability and toxicokinetics of hydrophobic aromatic compounds in benthic and terrestrial invertebrates. Environ. Rev. 4:276-299. Bolton PJ and J Phillipson (1976). Burrowing, feeding, egestion and energy budgets of Allolobophora rosea (Savigny) (Lumbricidae). Oecologia 23:225-245. Branquart E, R Deleu, A Copin and C Gaspar (1995). Bio-accumulation et metabolisation comparees de l'isoproturon, du linuron et du lindane par Lumbricus terrestris L. Mededelingen Faculteit Landbouwkundige en Toegepaste Biologische Wetenschappen Universiteit Gent 609:511-519. Briggs GG and KA Lord (1983). The distribution of aldicarb and its metabolites between Lumbricus terrestris, water and soil. Pestic. Sci. 14:412-416. CBS (1999). Waterkwaliteitsbeheer. Deel B: Zuivering van afvalwater, 1997. CBS report C-18-B/1997. Centraal Bureau voor de Statistiek, The Netherlands (in Dutch). Cobb GP, R Mellott, LW Brewer, CM Bens and RJ Kendall (2000). Diazinon dissipation from vegetation, occurrence in earthowrms, and presence in avian gastrointestinal tracts collected from apple orchards following D-Z-N 50W application. Environ. Toxicol. Chem. 19:1360-1367. Connell DW and RD Markwell (1990). Bioaccumulation in the soil to earthworm system. Chemosphere 20:91-100. Cooke AS, PW Greig-Smith and SA Jones (1992). Consequences for vertebrate wildlife of toxic residues in earthworm prey. In: Ecotoxicology of earthworms. PW Greig-Smith, H Becker, PJ Edwards, F Heimbach (eds.). Atheneum Press, Newcastle upon Tyne, UK. pp. 139-155.

General introduction [15] Cornelissen G, H Rigterink, PCM Van Noort and HAJ Govers (2000). Slowly and very slowly desorbing organic compounds in sediments exhibit Langmuir-type sorption. Environ. Toxicol. Chem. 19:1532-1539. [16] Darwin C (1881). The formation of vegetable mould through the action of worms, with observations on their habits. John Murray, London, UK. [17] Davis BNK (1968). The soil macrofauna and organochlorine insecticide residues at twelve agricultural sites near Huntingdon. Ann. Appl. Biol. 61:29-45. [18] Davis BNK (1971). Laboratory studies on the uptake of dieldrin and DDT by earthworms. Soil Biol. Biochem. 3:221-233. [19] Di Toro DM, CS Zarba, DJ Hansen, WJ Berry, RC Swartz, CE Cowan, SP Pavlou, HE Allen, NA Thomas and PR Paquin (1991). Technical basis for establishing sediment quality criteria for nonionic organic chemicals using equilibrium partitioning. Environ. Toxicol. Chem. 10:1541-1583. [20] Doube BM, O Schmidt, K Killham and R Correll (1997). Influence of mineral soil on the palatability of organic matter for lumbricid earthworms: a simple food preference study. Soil Biol. Biochem. 29:569-575. [21] Düring RA and S Gäth (2002). Utilization of municipal organic wastes in agriculture: where do we stand, where will we go? J. Plant Nutr. Soil Sci. 165:544-556. [22] Eason CT, LH Booth, S Brennan and J Ataria (1998). Cytochrome P450 activity in 3 earthworm species. In: Advances in earthworm ecotoxicology. SC Sheppard, JD Bembridge, M Holmstrup, L Posthuma (eds.). SETAC press, Pensacola, FL USA. pp. 191-198. [23] EC (1996). Technical Guidance Documents in support of Directive 93/67/EEC on risk assessment of new notified substances and Regulation (EC) No. 1488/94 on risk assessment of existing substances (Parts I, II, III and IV). EC catalogue numbers CR-48-96-001, 002, 003, 004-EN-C. Office for Official Publications of the European Community, 2 rue Mercier, L-2965 Luxembourg, Luxembourg. [24] Edwards CA and JR Lofty (1972). Biology of earthworms. Chapman and Hall, London, UK. [25] Edwards CA and PJ Bohlen (1992). The effects of toxic chemicals on earthworms. Rev. Environ. Contam. Toxicol. 125:23-99. [26] Eijsackers H, CAM Van Gestel, S De Jonge, B Muijs and D Slijkerman (2001). Polycyclic aromatic hydrocarbon-polluted dredged peat sediments and earthworms: A mutual interference. Ecotoxicology 10:35-50. [27] Gevao B, J Hamilton-Taylor and KC Jones (2000). Towards a complete mass balance and model for PCBs and PAHs in a a small rural lake, Cumbria U.K. Limnol. Oceanogr. 45:881-894. [28] Giesy JP and K Kannan (2001). Global distribution of perfluorooctane sulfonate in wildlife. Environ. Sci. Technol. 35:1339-1342. [29] Gish CD (1970). Organochlorine insecticide residues in soils and soil invertebrates from agricultural lands. Pestic. Monitor. J. 3:241-253. [30] Gish CD and DL Hughes (1982). Residues of DDT, dieldrin, and heptachlor in earthworms during two years following application. Special Scientific Report - Wildlife No. 241. U.S. Fish and Wildlife Service, Washington, DC, USA. [31] Gobas FAPC, X Zhang and R Wells (1993). Gastrointestinal magnification: The mechanisms of biomagnification and food chain accumulation of organic chemicals. Environ. Sci. Technol. 27:2855-2863. [32] Groot MSM, JJB Bronswijk, WJ Willems, T De Haan and P Del Castilho (1998). Landelijk meetnet bodemkwaliteit, resultaten 1995. Report no. 714801 024. National Institute for Public Health and the Environment (RIVM), Bilthoven, The Netherlands. [33] Harris ML, LK Wilson, JE Elliott, CA Bishop, AD Tomlin and KV Henning (2000). Transfer of DDT and metabolites from fruit orchard soils to American robins (Turdus migratorius) twenty years after agricultural use of DDT in Canada. Arch. Environ. Contam. Toxicol. 39:205-220. [34] Hickey JJ and DW Anderson (1968). Chlorinated hydrocarbons and eggshell changes in raptorial and fish-eating birds. Science 162:271-273. [35] ISO (1998). Soil Quality - Effects of pollutants on earthworms (Eisenia fetida). Part 2: Determination of effects on reproduction. No. 11268-2. International Organization for Standardization (ISO), Geneva, Switzerland. [36] Jones KC, JA Stratford, KS Waterhouse, ET Furlong, W Giger, RA Hites, C Schaffner and AE Johnston (1989). Increases in the polynuclear aromatic hydrocarbon content of an agricultural soil over the last century. Environ. Sci. Technol. 23:95-101. [37] Kannan K, J Koistinen, K Beckmen, T Evans, JF Gorzelany, KJ Hansen, PD Jones, E Helle, M Nyman and JP Giesy (2001). Accumulation of perfluorooctane sulfonate in marine mammals. Environ. Sci. Technol. 35:1593-1598. [38] Kraaij R, P Mayer, FJM Busser, M Van het Bolscher, W Seinen, J Tolls and AC Belfroid (2003). Measured pore-water concentrations make equilibrium partitioning work - a data analysis. Environ. Sci. Technol. 37:268-274. [39] Kraaij RH, W Seinen, J Tolls, G Cornelissen and AC Belfroid (2002). Direct evidence of sequestration in sediments affecting the bioavailability of hydrophobic organic chemicals to benthic deposit-feeders. Environ. Sci. Technol. 36:3525-3529.

17

Chapter 1 [40] Krauss M and W Wilcke (2001). Biomimetic extraction of PAHs and PCBs from soil with octadecylmodified silica disks to predict their availability to earthworms. Environ. Sci. Technol. 35:3931-3935. [41] Laird JM and M Kroger (1981). Earthworms. CRC Crit. Rev. in Environ. Control May 1981:189-218. [42] Lanno RP and LS McCarty (1997). Earthworm bioassays: adopting techniques from aquatic toxicity testing. Soil Biol. Biochem. 29:693-697. [43] Laverack MS (1963). The physiology of earthworms. Pergamon Press, Oxford, UK. [44] Lee KE (1985). Earthworms. Their ecology and relationships with soils and land use. Academic Press, Sydney, Australia. [45] Leppänen MT and JVK Kukkonen (1998). Relative importance of ingested sediment and pore water as bioaccumulation routes for pyrene to oligochaete (Lumbriculus variegatus, Müller). Environ. Sci. Technol. 32:1503-1508. [46] LNV (2000). Feiten en cijfers 2000. Ministerie van Landbouw, Natuurbeheer en Visserij, Directie Voorlichting, Den Haag, The Netherlands (in Dutch). [47] Løkke H and CAM Van Gestel (1998). Handbook of soil invertebrate toxicity tests. John Wiley & Sons, Chichester, UK. [48] Loonen H, DCG Muir, JR Parsons and HAJ Govers (1997). Bioaccumulation of polychlorinated dibenzo-p-dioxins in sediment by oligochaetes: influence of exposure pathway and contact time. Environ. Toxicol. Chem. 16:1518-1525. [49] Lord KA, GG Briggs, MC Neale and R Manlove (1980). Uptake of pesticides from water and soil by earthworms. Pestic. Sci. 11:401-408. [50] Ma WC, A Van Kleunen, J Immerzeel and PGJ De Maagd (1998). Bioaccumulation of polycyclic aromatic hydrocarbons by earthworms: assessment of equilibrium partitioning theory in in situ studies and water experiments. Environ. Toxicol. Chem. 17:1730-1737. [51] MacDonald DW (1983). Predation on earthworms by terrestrial vertebrates. In: Earthworm Ecology. From Darwin to vermiculture. JE Satchell (ed.). Chapman and Hall, London, UK. pp. 393-414. [52] Mackay D (1991). Multimedia environmental models. Lewis Publishers, Chelsea, MI, USA. [53] Marconi S, P Manzi, L Pizzoferrato, E Buscardo, H Cerda, DL Hernandez and MG Paoletti (2002). Nutritional evaluation of terrestrial invertebrates as traditional food in Amazonia. Biotropica 34:273-280. [54] Matscheko N, M Tysklind, C De Wit, S Bergek, R Andersson and U Sellström (2002). Application of sewage sludge to arable land-soil concentrations of polybrominated diphenyl ethers and polychlorinated dibenzo-p-dioxins, dibenzofurans, and biphenyls, and their accumulation in earthworms. Environ. Toxicol. Chem. 21:2515-2525. [55] Montforts MHMM, DF Kalf, PLA Van Vlaardingen and JBHJ Linders (1999). The exposure assessment for veterinary medicinal products. Sci. Total Environ. 225:119-133. [56] Moody CA, JW Martin, WC Kwan, DCG Muir and SA Mabury (2002). Monitoring perfluorinated surfactants in biota and surface water samples following an accidental release of fire-fighting foam into Etobicoke Creek. Environ. Sci. Technol. 36:545-551. [57] Nair A and MKK Pillai (1992). Trends in ambient levels of DDT and HCH residues in humans and the environment of Delhi, India. Sci. Total Environ. 121:145-157. [58] Neilson R, B Boag and M Smith (2000). Earthworm δ13C δ15N analyses suggest that putative functional classifications of earthworms are site-specific and may also indicate habitat diversity. Soil Biol. Biochem. 32:1053-1061. [59] Nelson PA, RR Stewart, MA Morelli and T Nakatsugawa (1976). Aldrin epoxidation in the earthworm, Lumbricus terrestris L. Pestic. Biochem. Physiol. 6:243-253. [60] Norstrom RJ, M Simon, DCG Muir and RE Schweinsburg (1988). Organochlorine contaminants in arctic marine food chains: identification, geographical distribution, and temporal trends in polar bears. Environ. Sci. Technol. 22:1063-1071. [61] OECD (1984). Guideline for testing of chemicals no. 207. Earthworm, acute toxicity tests. Organization for Economic Cooperation and Development, Paris, France. [62] Paoletti MG (1999). The role of earthworms for assessment of sustainability and as bioindicators. Agric. Ecosyst. Environ. 74:137-155. [63] Parle JN (1963). Micro-organisms in the intestines of earthworms. J. gen. Microbiol. 31:1-11. [64] Pathirana S, DW Connell and PD Vowles (1994). Distribution of polycyclic aromatic hydrocarbons (PAHs) in an urban roadway system. Ecotox. Environ. Saf. 28:256-269. [65] Perel TS (1977). Differences in lumbricid organization connected with ecological properties. Ecol. Bull. 25:56-63. [66] Piearce TG (1978). Gut contents of some lumbricid earthworms. Pedobiologia 18:153-157. [67] Pignatello JJ and B Xing (1996). Mechanisms of slow sorption of organic chemicals to natural particles. Environ. Sci. Technol. 30:1-11. [68] Reynolds JW (1998). The status of earthworm biogeography, diversity and taxonomy in North America revisited with glimpses into the future. In: Earthworm ecology. CA Edwards (ed.). St. Lucie Press, Boca Raton, FL, USA. pp. 15-34.

18

General introduction [69] RIVM/VROM (2002). Jaarverslag bodemsanering over 2001, de monitoringsrapportage. RIVM/VROM, Den Haag, Bilthoven, The Netherlands (in Dutch). [70] Romijn CAFM, R Luttik and JH Canton (1994). Presentation of a general algorithm to include effect assessment on secondary poisoning in the derivation of environmental quality criteria. 2. Terrestrial food chains. Ecotox. Environ. Saf. 27:107-127. [71] Sabljić A, H Güsten, H Verhaar and J Hermens (1995). QSAR modelling of soil sorption. Improvements and systematics of log Koc vs. log Kow correlations. Chemosphere 31:4489-4514. [72] Schwarzenbach RP, PM Gschwend and DM Imboden (1993). Environmental organic chemistry. John Wiley & Sons, New York, NY, USA. [73] Shea D (1988). Developing national sediment quality criteria. Environ. Sci. Technol. 22:1256-1261. [74] Sijm D, R Kraaij and A Belfroid (2000). Bioavailability in soil or sediment: exposure of different organisms and approaches to study it. Environ. Poll. 108:113-119. [75] Sims RW and BM Gerard (1985). Earthworms. Synopsis of the British fauna No. 31. E.J. Brill Publishing Company, Leiden, The Netherlands. [76] Spratt DM (1997). Endoparasitic control strategies: implications for biodiversity of native fauna. Intern. J. Parasitol. 27:173-180. [77] Stafford EA and AGJ Tacon (1988). The use of earthworms as a food for rainbow trout Salmo gairdneri. In: Earthworms in waste and environmental management. CA Edwards, EF Neuhauser (eds.). SPB Academic Publishing, The Hague, The Netherlands. pp. 193-208. [78] Statistisches Bundesamt (1995). Öffentliche Wasserversorgung und Abwasserbeseitigung. Umwelt, Fachserie 19, Reihe 2.1/2.2. Statistisches Bundesamt, Germany (in German). [79] Stenersen J and N Øien (1980). Action of pesticides on earthworms. Part IV: uptake and elimination of oxamyl compared with carbofuran. Pestic. Sci. 11:396-400. [80] Stenersen J (1992). Uptake and metabolism of xenobiotics by earthworms. In: Ecotoxicology of earthworms. PW Greig-Smith, H Becker, PJ Edwards, F Heimbach (eds.). Athenaeum Press, Newcastle upon Tyne, UK. pp. 129-138. [81] Stickel LF (1973). Pesticide residues in birds and mammals. In: Environmental pollution by pesticides. CA Edwards (ed.). Plenum Press, London, U.K. pp. 254-305. [82] Ten Hulscher TEM, BA Vrind, H Van den Heuvel, LE Van der Velde, PCM Van Noort, JEM Beurskens and HAJ Govers (1999). Triphasic desorption of highly resistant chlorobenzenes, polychlorinated biphenyls, and polycyclic aromatic hydrocarbons in field contaminated sediment. Environ. Sci. Technol. 33:126-132. [83] US-EPA (2000). Superfund: 20 years of protecting human health and the environment. United States Environmental Protection Agency, EPA 540-R-00-007, see www.epa.gov/superfund. [84] Van Brummelen TC, RA Verweij, SA Wedzinga and CAM Van Gestel (1996). Polycyclic aromatic hydrocarbons in earthworms and isopods from contaminated forest soils. Chemosphere 32:315-341. [85] Van der Kooy LA, D Van de Meent, CJ Van Leeuwen and WA Bruggeman (1991). Deriving quality criteria for water and sediment from the results of aquatic toxicity tests and product standards: Application of the equilibrium partitioning theory. Wat. Res. 25:697-705. [86] Van Dijk HFG and R Guicherit (1999). Atmospheric dispersion of current-use pesticides: a review of the evidence from monitoring studies. Wat. Air Soil Pollut. 115:21-70. [87] Van Gestel CAM and WC Ma (1988). Toxicity and bioaccumulation of chlorophenols in earthworms, in relation to bioavailability in soil. Ecotox. Environ. Saf. 15:289-297. [88] Van Gestel CAM and WC Ma (1990). An approach to quantitative structure-activity relationships (QSARs) in earthworm toxicity studies. Chemosphere 21:1023-1033. [89] Van Gestel CAM (1991). Earthworms in ecotoxicology. PhD thesis, University of Utrecht, Utrecht, The Netherlands. [90] Van Leeuwen CJ, PTJ Van der Zandt, T Aldenberg, H Verhaar and JLM Hermens (1992). Application of QSARs, extrapolation and equilibrium partitioning in aquatic effects assessment. I. Narcotic industrial pollutants. Environ. Toxicol. Chem. 11:267-282. [91] Van Rhee JA (1977). Effects of soil pollution on earthworms. Pedobiologia 17:201-208. [92] Viktorov AG (1997). Diversity of polyploid races in the family Lumbricidae. Soil Biol. Biochem. 29:217221. [93] Wania F and GL Daly (2002). Estimating the contribution of degradation in air and deposition to the deep sea to the global loss of PCBs. Atm. Environ. 36:5581-5593. [94] Wheatley GA and JA Hardman (1968). Organochlorine insecticide residues in earthworms from arable soils. J. Sci. Fd. Agric. 19:219-225. [95] Wodzinski RS and JE Coyle (1974). Physical state of phenanthrene for utilization by bacteria. Appl. Microbiol. 27:1081-1084.

19

Section A

Theoretical worm worm

uptake/elimination

water water soil soil

sorption/desorption

“Good theories consist of just one problem-solving strategy, or a small family of problem-solving strategies, that can be applied to a wide range of problems. The theory succeeds as it is able to encompass more and more problem areas. Failure looms when the basic problem-solving strategy (or strategies) can resolve almost none of the problems without the “aid” of untestable auxiliary hypotheses.” P. Kitcher (1982) Abusing Science. The case against creationism.

Theory of compartment models

2 Using Compartment Models for Bioaccumulation1

ABSTRACT  Compartment models are the dominant modelling approach in ecotoxicology, but their limitations and preconditions are not always recognised. As an example, the standard one-compartment accumulation model assumes that the parameters, including the exposure concentration, are constant. Furthermore, the organism’s volume is not allowed to change during exposure. This chapter attempts to explain the basic nature of the one-compartment model and its rate constants, and explores what happens when we violate its assumptions. How is the accumulation pattern affected when the exposure concentration is not constant? Especially in a soil situation, constant exposure is not easily tested. This chapter thus provides background for readers who are not experts on transport models. There are no radical new insights or models in this chapter; I tried to compile existing knowledge and follow it to its logical consequences. The chapter is written in a text-book style instead of a journal style to increase its palatability, and to allow a more extensive and coherent discussion.

1 This chapter started as a lecture for PhD students at the Institute for Risk Assessment Sciences (IRAS) in Utrecht. It also includes elements from the discussions I had with Prof. Bas Kooijman (Vrije Universiteit, Amsterdam) on the rate constants, and their relation to hydrophobicity.

23

Chapter 2

INTRODUCTION Probably the most important advantage of building a model is to help us understand a problem and to test mechanistic theories. Is our theory (when translated into mathematics) sufficient to describe phenomena we see in the real world? The trade-off between model complexity and model realism is central for any modelling exercise. More complex models can be more realistic, but because there are more (unknown) parameters, the model may become impossible to (in)validate. Such a model will provide little understanding into the process studied, and will not help to increase our confidence in the theory. So the dilemma in modelling is to describe only the most dominant processes, without sacrificing too much realism. It is my opinion that, whenever possible, models should be developed on a mechanistic basis for scientific reasons (testing theories), and to allow for better predictions (extrapolation to different situations). However, there are also useful equations that are purely descriptive, like regressions. Although it can be argued whether regressions can be called “models”, they prove to be useful in many fields. In environmental chemistry and ecotoxicology, quantitative structure-activity relationships (QSARs) are widely used to predict an expensive property of a chemical from a cheap one. Popular QSARs relate hydrophobicity (expressed as Kow) to molecular structure [79], and bioconcentration factors (BCF) to Kow [60]. Some QSARs are based on mechanistic considerations while others are purely descriptive. An extreme example of the latter are polynomial regressions of BCF against log Kow [18]: hydrophobicity can be mechanistically linked to bioconcentration, but there is no mechanism leading to a polynomial relationship. Multi-variate regression falls into the same category; for example, the description of internal concentrations of metals in organisms from a variety of environmental properties [71]. Even though the included properties are all known (or suspected) to influence metal bioavailability, it is unlikely that all these properties are working additively after log-transformation. Other QSARs follow a much more mechanistic path, like the estimation of Kow from specific structural units of a molecule (see e.g. [79]). These estimations are trusted to such an extent that calculated Kow values for extremely hydrophobic chemicals are often preferred above measured ones (even though there is no way to validate these estimates when the measurements cannot be trusted). Many problems in biology or ecotoxicology can be addressed by applying compartment modelling. In essence, the analysis starts with identifying the “compartments”, which can be any entity that we are willing to treat as “well-mixed” and homogeneous with regard to some other entity that is transported to or from the compartment. To be more explicit, I will focus on ecotoxicological problems, where the compartments can be parts of the environment (e.g. a lake or an experimental container), or an organism (e.g. a fish or an earthworm), or even parts of organisms (the gut or an internal organ). The entity that is transported is in this case the chemical. Our choice of compartment will not only depend on the system that we are describing, but also on the scale of the problem of interest. As an example, for describing the transfer of chemicals through the food chain, earthworms can be modelled as a single compartment. On the other hand, when we are interested in the kinetics of a chemical in a specific organ, we may need to add more compartments to describe the system. Compartment models are probably as old as differential equations themselves. The earliest application that I could find of this model, applied to chemical accumulation in organisms, is from 1924 [102]. The one-compartment model became especially popular to describe bioaccumulation kinetics in fish (see [85] for an overview and discussion, as well as some alternative formulations). The standard one-compartment model is usually written in the form of this differential equation:

24

Theory of compartment models

dC b = k uC w − k eC b dt

Eq. 1

The concentration in the organism (Cb) changes with the net result of the relative uptake flux1 from the water phase (ku×Cw) and elimination from the organism (ke×Cb). When the water volume is large enough, its concentration (Cw) may be assumed constant, in which case we only have to deal with one compartment (the organism). So the organism is viewed as a well-mixed and homogeneous compartment. This is of course a gross simplification. Nevertheless, this assumption may be warranted, especially when the internal redistribution in the organism is rapid compared to exchange with the water phase. A further thing to note is that the relative chemical fluxes (ki×Cj) are directly proportional to the concentration in the source compartment, as long as the rate constants are indeed constant (“donor control”). This implies that we assume a passive transport process. In case the uptake or elimination would be actively controlled by the organism, we would see a flux that depends non-linearly on the concentration (e.g. saturation at high concentrations in the case of Michaelis-Menten kinetics, see dedicated section in this chapter). When all parameters apart from Cb are constant in time, the general solution of this differential equation is:

b

C b = C b 0 e − k et +

(

ku C w 1 − e − ket ke

)

w

Eq. 2

As shown in the schematic picture (below Eq. 1), this solution implies a one-compartment system (as the water concentration needs to be constant for this solution, it does not qualify as a compartment). The quotient ku/ke is also known as the bioconcentration factor (BCF). Multiplied with the concentration in water, this BCF gives the steady-state concentration in the organism (Cb∞). If the initial concentration in the organism is not zero, it can be included2 as Cb0. The value of this conceptual model was demonstrated as early as 1975, by Branson et al. [13]. Achieving a steady-state level in fish may take a long time for hydrophobic chemicals, and BCF tests are thus very expensive. These authors proposed to replace steady-state tests with shorter accumulation and eliminations tests, to estimate the rate constants ku and ke (which are than used to estimate BCF). Although Equation 2 looks deceptively simple, it hinges on several (implicit) assumptions: 1. The compartment is assumed homogeneous with regard to the chemical concentration and mixing is instantaneous. 2. Uptake and elimination are the result of passive processes (i.e. fluxes are proportional to the concentration in the donating compartment). 3. The exposure concentration is constant and not influenced by chemical uptake into the organism (i.e. the water volume is infinitely large, and there are no losses due to degradation, volatilisation, etc.). 4. The compartments are of constant volume (e.g. the organism does not grow or shrink). 5. The rate constants are indeed constants (in time as well as in concentration).

Fluxes are defined with the dimension # t-1 (molecules per time). Here, ku×Cw has dimension # L-3 , which can be called a relative flux. The product ku×Cw×Vb is a true flux. 2 The way C is included in Equation 2 assumes that the molecules initially present are no different b0 than the ones that are taken up (they are excreted in a similar way, evidenced from the exponent in the first term). In some formulations (e.g. [72]), the choice is made to view the initial body residue as a passive pool that is not eliminated (thus without the exponent). 1

t-1

25

Chapter 2

In this chapter, several problems regarding these assumptions will be dealt with. What is the nature of the rate constants, how do they change with growth, and what happens when the assumptions are violated? This chapter will focus mainly on neutral organic contaminants, although many of the examples will also have relevance for other compounds. I have added references where model formulations have been applied, but do not claim these to be complete or the first mention in the literature. Rather, they provide representative examples. The discussion in this chapter is not limited to a particular organism, or group of organisms. These concepts are equally relevant for aquatic, as well as sediment and terrestrial bioaccumulation studies, and even for uptake from the air. A thorough understanding of the concepts and problems in this chapter will help to fully appreciate some of the concepts in the following chapters on earthworms.

NATURE OF THE RATE CONSTANTS From the form of Equation 1, in which the process of bioconcentration is written down, it appears that uptake and elimination are two separate processes that have nothing to do with each other. The BCF is just the coincidental result of these processes. This concept of bioconcentration may be appropriate for metals, for which uptake and elimination are largely mediated by carrier molecules in the membrane [58]. For these compounds, it is not clear beforehand whether transport across the membrane is governed by the same rate constant in both directions. For most organic chemicals, there appear to be no specific channels, and bioconcentration is seen as passive diffusion of the chemical (e.g. for fish [30], for earthworms [56], for plants from air [76]). This means that the same processes govern uptake and elimination as they take place across the same membrane, and are subject to the same resistances in both directions. I therefore believe it is more appropriate to look at bioconcentration as a diffusion process. THE CHEMICAL NATURE OF DIFFUSION AND THE CONCEPT OF FUGACITY It is well-established in environmental chemistry (see e.g. [79]) that the passive transport processes of a chemical in a system can be described in terms of the energy status that a molecule of a compound carries when it is mixed in the environment of another chemical (the solvent phase). The chemical potential (µi) of compound i is the amount of Gibbs-free energy added to the system by adding a mol of i. The compound flows from a high chemical potential to a lower one to minimise the energy content of the system (or maximise the entropy), i.e. until thermodynamic equilibrium is reached. Chemical potentials are related to the more pragmatic fugacities, the “urge” of a chemical to flee from a certain system. The chemical potential of compound i in liquid solution is the chemical potential of i in the ideal state (the pure liquid) plus a term dealing with the nonideality of the solvent environment. The main term in this deviation from ideality is the “activity” of the compound; a measure of how active the compound is in a given solution, compared to the ideal state. The activity is related to the concentration by the activity coefficient γ which is a measure of the additional free energy that the compound is carrying in a non-ideal mixture, compared to the reference state (the pure liquid). The γ can be interpreted as a measure of how much the compound “likes” the solvent environment (a large γ means a great dislike). The octanol-water partition coefficient can therefore be written as the ratio of the activity coefficients of the chemical in water and octanol: Kow = γw/ γo. For non-polar hydrophobic substances, γo is generally between 1 and 10, indicating that octanol comes close to being an ideal solvent. The γw varies much more, and it is therefore the dislike of water that drives the partition coefficient (therefore, the

26

Theory of compartment models

designation “hydrophobic” is more accurate than “lipophilic”). The activity coefficients are inversely related to the solubilities of the chemical in a solvent. In summary, the difference in chemical potential or fugacity determines the direction and the magnitude of the passive chemical transport between two points. When considering diffusion within a single solvent, fugacities are directly proportional to concentrations. In those cases, we may therefore just as well use a concentration difference as our driving force for diffusion. However, when dealing with diffusion from one solvent to the other (as we have to when describing bioconcentration), we have to consider the difference in activity coefficients of the chemical in both solvents. Concentrations alone do not suffice; in equilibrium, the chemical fugacity will be equal in both compartments, but not their concentrations. We therefore have to recalculate all concentrations to fugacities, as pioneered by Mackay (see e.g. [61,62]), or correct the concentrations for the differences in solubility of the chemical in both media. Fugacities (f) and concentrations (C) are related by the fugacity capacity (Z):

f =

C Z

Eq. 3

The fugacity capacity is specific to the chemical and the medium in which it is dissolved. It is a measure of how much the compound “likes” the solvent, just like the activity coefficient (but inversely: the higher Z, the more the chemical likes the solvent). In this chapter, I will not attempt to write everything in fugacities. Even though the use of fugacities leads to a more consistent model description, it leads to a less obvious interpretation of the parameters. Therefore, I will stick to the use of concentrations here. DIFFUSION FROM WATER TO ORGANISM According to Fick’s first law (see e.g. [79]), diffusion of a chemical in a single solvent is driven by the concentration difference between two locations (∆C). The diffusion mass flux (dN/dt) depends on the area through which diffusion takes place (A) and the distance the chemical needs to travel (∆x):

dN D = A∆C dt ∆x

Eq. 4

The proportionality constant (D) is called the diffusion coefficient, which differs between chemicals (although not spectacularly; it depends mainly on molecular size) and between solvents (more important). The diffusion coefficient is very high in gasses, lower in liquid solution, and very low in solids. In most practical applications, diffusion coefficients and transport distances are not known and their quotient is used as a single unknown parameter; the mass transfer coefficient (abbreviated MTC, with symbol m):

m=

D ∆x

Eq. 5

This coefficient has the dimensions of a speed (l t-1) and is therefore sometimes referred to as a “piston velocity” (as an imaginary piston, pushing the chemical through the medium). When we consider chemical diffusion from one solvent (e.g. water) to a different solvent (e.g. an organism), we cannot simply apply the concentration difference between the two media. Concentration in organisms and concentration in water are not comparable quantities (their units even differ). However, we know that when we wait 27

Chapter 2

long enough, the organism-water system will evolve to an equilibrium situation. In equilibrium, the chemical is where it “wants” to be (the fugacities are equal) and there is no net transport of chemical. We may therefore use the deviation from equilibrium instead of the concentration difference (∆C), leading to a mass flux into the organism of:

dN b = mb+ Abw (C b∞ − C b ) dt

Eq. 6

in which Abw is the exchange area between organism and water, and mb+ is the total MTC, seen from the perspective of the organism (and thus with dimension L t-1). These total transfer coefficients will be dealt with later. More conveniently, the steady-state concentration (Cb∞) may be expressed in terms of a partition coefficient between water and organism, the bioconcentration factor (Kbw):

C b∞ = K bwC w

Eq. 7

The mass flux dN/dt can be translated to the concentration change in the organism by dividing both sides of Equation 6 by the volume of the organism (Vb):

dC b A = mb+ bw (K bwC w − C b ) dt Vb

Eq. 8

Note that I have expressed the driving force for diffusion as a concentration difference in which one of the concentrations in corrected for the differences in solubility of the chemical in the two “solvents”. As this equation describes the same organism system as Equation 1, new definitions of ku and ke can be derived:

k u = mb +

Abw K bw Vb

Eq. 9

k e = mb +

Abw Vb

Eq. 10

Thus, ku and ke are similar and deviate only by a factor of Kbw (which is obvious, given the fact that the bioconcentration factor is also given by ku/ke). Parameter ke has the dimensions of a rate constant (t-1), whereas ku strictly does not (l3 L-3 t-1) because it takes the water concentration and calculates its effect on the organism concentration. Both rate constants depend in the same manner on the area-volume ratio of the organism. Note that the MTC does not depend on areas or volumes, but only on the intrinsic properties of the interface, and (to some extent) on chemical properties. I prefer the description of the onecompartment model with ke and Kbw (Eq. 8–10) above the one with the two rate constants ku and ke (Eq. 1–2). The choice is rather arbitrary and leads to the same numerical results. However, this option has my preference because the description of Equation 8 is closer to the diffusion mechanism that, by assumption, underlies the process. Secondly, the partition coefficient acts as a model parameter. The partition coefficient plays a central role in environmental chemistry applications (as in the Kow and Henry’s law constant), and for neutral organic chemicals, Kbw can be estimated from Kow (e.g. [60]). This estimation is possible as octanol and organism lipids are to a large degree comparable solvents. So instead of having two unknowns, we have one truly unknown parameter (ke), and one where we can make a first estimation on the basis of the chemical’s Kow and the organisms lipid content. The Kbw is therefore not an accidental result of the uptake and elimination rates; it is a parameter with physical meaning, and deserves to be treated as a model parameter. Furthermore, as we saw earlier, ku does not have the strict dimensions of a rate

28

Theory of compartment models

constant. And finally, fitting Kbw as model parameter has the practical advantage that a confidence interval of the bioconcentration factor can be estimated. This is not so straightforward from the standard errors of ku and ke as these tend to be (heavily) correlated. TOTAL MASS-TRANSFER COEFFICIENTS In the previous section, I have introduced the total MTC for the transfer between organism and water but postponed the explanation of its nature to this section. The coefficient mb+ was said to be “referenced” to the perspective of the organism. This is a necessary specification as this coefficient acts on a concentration difference that I expressed in organism concentrations: Cb∞–Cb. However, although the choice for organisms is an obvious one, the choice of perspective is in fact arbitrary. We could just as well express the driving force of diffusion in water concentrations:

 dN b C  = mw+ Abw  C w − b  dt K bw  

Eq. 11

This equation describes the exact same process, and must therefore lead to exactly the same numerical results. The concentration difference in units organism is however numerically different from that expressed as water (unless the partition coefficient Kbw is exactly one). It thus follows that the MTC referenced to water (mw+) does not have the same value as the one referenced to organism (mb+). In fact, comparison of Equation 6–7 and 11 reveals that the difference between them is a factor of Kbw. Note that this analysis is a very general one. No mechanistic explanation is provided for the value of the total MTC, nor how it depends on chemical and organism properties. The only assumption is that the process is driven by passive diffusion. The speculations start when we want to infer about the contributors to the total MTCs. As the chemical may diffuse from fish to water and vice versa, the elements making up mb+ and mw+ must be the same. Traditionally, the transfer process has been characterised by the “two-resistance mass-transfer coefficient approach”, published by Whitman in 1923 (see [61,79], and Fig. 1). It is assumed that the total MTC is made up of several “resistances” that the chemical encounters when travelling from one medium to another. For instance, in air-water exchange, the resistance is thought to be localised in the boundary layers at both sides of the air-water interface. In the bulk of the medium, chemical transport is rapid, owing to turbulent eddy diffusion. However, at the interface, the eddies are dampened and transport slows down. The exact mechanism of chemical movement in the boundary layers is not fully understood, but there are several theories providing adequate descriptions under different circumstances. In the “stagnant two-film model”, the chemical moves through the boundary layers by (slow) chemical diffusion, whereas the “surface renewal model” deals with the refreshment of air and water parcels at the interface. The technicalities of these models do not concern us here (the interested reader is referred to textbooks on environmental chemistry [61,79]), only that they lead to a similar formulation of the total MTC (Eq. 12). In a strict sense, these mechanism are not compatible with one-compartment models for the media, because the media are not homogenous and well mixed anymore (there is a concentration gradient near the interface), thus violating assumption 1 in the introduction of this chapter. However, as long as the bulk of the medium can be considered to be well mixed, the one-compartment model can be used as valid approximation of this mechanism (the other assumptions in the list are not conflicting with the two-resistance model).

29

Chapter 2

turbulent air eddy mixing ~1 mm “stagnant” air boundary layer

Ra

molecular diffusion

phase transfer

air-water interface

Rw molecular diffusion

“stagnant” water boundary layer

resistances

~0.1 mm

turbulent water

eddy mixing

Figure 1. Schematic representation of the physical processes responsible for the transport of chemicals between water and air, according to the two-phase resistance model (adapted from [61,79]). Ra and Rw are the air and water resistances, respectively.

For explaining the two-resistance approach, I will focus first on air-water exchange, as the application to this problem is more intuitive than to living organisms. The inverse of the MTC can be thought of as a resistance to diffusion, so the total coefficient is made up of the resistances in the air and the water boundary layer. These resistances are placed in series so their contributions may be added (just as electrical resistances). Please note that I switch from a mechanistic description to a more phenomenological one here. The resistance of the interface itself appears to be small (in the air-water case) and may be ignored [61]. The total transfer resistance is thus given by:

1 1 K aw = + m a + ma mw

Eq. 12

log rate constant

where ma and mw are the “partial” MTCs at each side of the interface and the air-water partition coefficient (Kaw) enters the equation to correct the MTC in water for the fact that water has a different capacity for the chemical than air (the imaginary ke (air) “piston” has to move at a ku (air) different velocity in water than Ra = Rw in air to move the same amount of chemical). The partial MTCs depend on the diffusivity of the chemical in the medium and the log air-water partition coeff. boundary thickness, or the mean contact time of the parcel Figure 2. Expected relationship between the air-water at the interface (depending on partition coefficient (Kaw) and the rate constants for uptake the theory that is followed). in (ku), and elimination from (ke), the air compartment. Nevertheless, these coefficients The vertical line shows where both transfer resistances are depend only little on chemical equal (Ra = Rw).

30

Theory of compartment models

properties (mainly on molecular weight), so we can predict how the total MTC (and with it the rate constants) depends on Kaw (which varies over many orders of magnitude). From Equation 12, it follows that the air-phase resistance dominates when Kaw is very small; the water-phase resistance dominates when Kaw is very high. Furthermore, it follows that at low values of Kaw, the total resistance (as seen from air) does not depend on Kaw as the partition coefficient is not influencing the main resistance 1/ma. At high values of Kaw, however, the water-phase dominates, which implies that the total resistance increases linearly with Kaw. The rate constant for “elimination” from the air phase will thus decrease linearly with Kaw at high values of Kaw. This dependence is illustrated in Figure 2. As stated earlier, the point of reference is arbitrary. Seen from the water phase, the total resistance is given by: 1 1 1 = + mw + K aw ma mw

Eq. 13

In this case, the air resistance needs to be corrected in a similar way. At high values of Kaw, the water phase dominates again, but the total resistance seen from the water phase is now independent of Kaw. At first glance, this appears to be inconsistent; at high values of Kaw, the total resistance is influenced in a different way by Kaw, depending from which side we look. Here we touch at the fundamental asymmetry of the problem when we deal with two different phases. However, the internal consistency becomes clear by expressing the partition coefficient from the viewpoint of water: K 1 1 = wa + m w+ ma mw

Eq. 14

Now we can see that the water-phase total MTC is related in the same manner to the water-air partition coefficient as the air-phase MTC is related to the air-water coefficient. MASS-TRANSFER COEFFICIENTS IN BIOLOGICAL SYSTEMS The description of diffusion between different media, outlined above, is well-established in environmental chemistry for chemical exchange between liquid and gaseous environmental media (e.g. [54] for ocean-air exchange). Furthermore, the expected pattern for the rate constants against Kow was also beautifully demonstrated for SPME fibres in water for a broad range of chemicals [94]. For soil-air exchange, the situation is much more complicated as soil cannot usually be considered as an instantaneously mixed compartment (e.g. [61]). The application of this theory to organisms also implies that we are willing to consider the organism to be a well-mixed volume of water and lipids. This extrapolation from established chemical theory to ecotoxicology is however not so straightforward. In principle, one could look at the fish-water system in a similar way as the air-water system, although the nature of the interface is entirely different. Instead of two mobile “bulk” phases, we now have a bulk water phase and a complexly organised “organism phase”, separated by a biological membrane. Several authors have attempted to reconcile the bioconcentration processes in a phenomenological way, with a number of diffusion resistances in series or parallel (e.g. plants from air [76], fish from water [30]). One could look at the organism as a bulk phase, leading to a similar formulation as for the air-water system. Others described a different system for fish, envisaging the process as diffusion across an interface consisting of two water layers (inside and outside of the organism), separated by a lipid membrane [30,84]. Inside the organism is a lipid storage compartment that rapidly exchanges with the interior water compartments. As a

31

Chapter 2

consequence of this model, the elimination rate of the fish is expected to follow a similar pattern as the elimination from air, depicted in Figure 1. The elimination rate of the fish (ke) is independent of the Kbw at low Kbw values, but linearly decreases with Kbw at higher values. The change in the relationship reflects the transition from a membrane-dominated resistance to a water-dominated resistance. This transition has been verified experimentally in an artificial system, where two water phases were separated by a hydrophobic membrane [28], and by using in vivo fish preparations [66]. The uptake rate constant (ku) is related in a similar way to the inverse of the bioconcentration factor (the water-fish partition coefficient), which has also been observed in fish [23]. In practice, Kbw increases with hydrophobicity, and so does molecular weight. Because diffusion coefficients depend on molecular weight, ku will decrease again slightly, at higher Kow values [84]. Although these attempts provide a reasonable description of the bioconcentration process [30,84], it must be stressed that they are simplifications. Nevertheless, this line of thinking appears to be the right way to proceed, providing an insight into the bioconcentration process on the basis of well-established chemical theories. It must be noted that many researchers plot ke against Kow and infer about their relation, whereas in fact ke should be plotted against Kbw (Kbw drives the chemical exchange and not Kow). Of course, Kbw and Kow will be related, but in many cases this relationship is not directly proportional: slopes of log Kbw against log Kow tend to be somewhat less than unity (for an overview, see [93]). Above a log Kow of around 6, many Kow-Kbw relations tend to level off. This has urged several authors to change the linear model to a polynomial relationship [18]. However, the data for such extremely hydrophobic chemicals need be carefully considered. Values for log Kow above 6 are difficult to determine reliably [60], although the introduction of the “slow-stirring” method in 1989 by De Bruijn et al. [22] has seriously improved the quality of the Kow data for very hydrophobic chemicals like PCBs. Further, bioconcentration experiments are hampered by poor water solubility and sorption to glass walls etcetera. There are, however, also valid reasons to believe in reduced accumulation of these compounds. Firstly, molecular size tends to increases with Kow, and the size may hamper membrane passage [68]. Secondly, elimination rates decrease with Kow. At high Kow, elimination may become so slow that growth dilution or biotransformation can become the main “elimination” process, thus leading to non-equilibrium body residues (at lower values than predicted from hydrophobicity). For hydrophobic compounds, plots of log ke against log Kow tend to show a decrease of ke as expected, but seldom with a slope of –1, as predicted (see Fig. 2). In the log Kow range of 3–6, slopes for fish vary from approximately –0.4 [40], to –0.6 [18], and –0.66 [34]. For molluscs and daphnids, values around –0.5 are reported [35], but in midges, a slope was found of –0.99 [53]. Values of ku are relatively constant in this range: between 1 [40], or 1.5 [18] log units, although a slope with log Kow of 0.34 was also found [34]. Legierse et al. [51] showed that the expected relationship of the rate constants with Kow was confirmed for chlorobenzenes in guppies (Poecilia reticulata), although the slope of ke with Kow was slightly less than unity on log scale (–0.84). For the pond snail (Lymnaea stagnalis), the authors observed a stronger deviation from predicted patterns, as ku showed a clear increase with Kow (a slope of 0.76 on log scale). The reasons for this deviation are not clear, but it shows that the two-resistance diffusion model for biota may not be sufficient to cover all situations. For now, this theory serves as a reference model. The general pattern, predicted by the model has been verified, but several questions remain. Deviations may be caused by the differences between octanol and animal lipids (hence the preference to plot ke against Kbw), and differences between compounds (e.g. steric hindrance and 32

Theory of compartment models

metabolism). However, the deviations may also point at more structural problems, as the two-resistance model may be too simplistic to encompass all the resistances between water and tissue. For example, the water around the gill surface may not be well represented by a stagnant water layer in which molecular diffusion is rate limiting, but the rate of refreshment (the ventilation rate) may be limiting (see e.g. [3,15]). Sijm et al. [83] calculated that the water flow will limit the uptake of hydrophobic compounds for larger fish (more than 5 g), and that also the blood flow limits uptake to some extent. ACCUMULATION IN SOIL SYSTEMS The previous sections have focussed on organism exposed to a water environment (like fish in an experimental aquarium). However, the situation is different when we look at soil and sediment organisms. Although it is likely that diffusive uptake takes place from the water phase [56,89] (diffusion in solids is extremely slow), the pore water is not likely to be well-mixed. The volume of the pore water that is in contact with the organism is relatively small, as well as its capacity for hydrophobic chemicals, which may lead to problems with replenishment [8]. Micro-organisms may degrade the dissolved chemical and thus alter the apparent sorption [25]. These situations will be discussed in the section on variable exposure concentrations. For earthworms, the focus of this thesis, it is difficult to judge toxicokinetics solely on the basis of passive diffusion from the pore water. Earthworms will also feed on soil particles so we will always have two uptake and elimination routes simultaneously. Furthermore, as feeding behaviour will depend on soil properties and species, different datasets cannot be compared, and a single relationship between Kow and ke is not to be expected (see the dedicated section at the end of this chapter). However, the basic processes are the same in soil and water, and the relationship between Kbw and Kow has been confirmed [4,19] (further discussed in Chapter 3). Furthermore, ke decreases with Kow with a slope of –0.66 on log scale for chlorobenzenes and PCBs [7], and for PAHs and PCBs around –0.3 [63]. Experiments with chlorobenzenes in a water-only environment have shown the log ke to decrease with log Kow with a slope of –1.3 [4], with estimates of ku constant within a factor of three.

NON-LINEARITIES IN THE ACCUMULATION MODEL RELATIONSHIP BETWEEN BODY SIZE AND KINETICS Equation 9 and 10 give the mechanistic nature of the rate constants, from a diffusion perspective. A not so obvious consequence of these equations is the fact that the same area-volume ratio is present in both rate constants. When the organism changes in volume (e.g. growth), the area will change in a different way than the volume. As a result, the rate constants ku and ke will change with the size of the organism, but in the same way, leaving their quotient (Kbw) unaffected. This is to be expected as Kbw will only depend on the solubility of the chemical in the organism compared to that in water, which does not necessarily change with area/volume ratio of the organism (Kbw will change only when the lipid content of the organism changes). The fact that the area/volume ratio determines the rate constants implies that we cannot directly compare rate constants between different species (or even between juveniles and adults of the same species), although this can be partially circumvented by using allometric relations [37]. The contact area for diffusion is in most species an unknown quantity, as it involves the surface area of the gills or the internal surface of the gut wall. Nevertheless, we may assume that the contact area is proportional to the total surface area. When we assume that the organism does not change it’s shape dramatically during growth, we may take the surface area proportional

33

Chapter 2

to body volume to the power 2/3 (see also [42]). We may assume that the MTC changes little during growth (it only depends on intrinsic properties of the exchange interface), which leads to an expression for the relation between ke and the volume of organism (Vb): k e = k e−ref

Vb1−/ref3

Eq. 15

Vb1 /3

The ke at a certain size (ke at Vb) is derived from the ke at an arbitrary reference size (ke-ref at Vb-ref), analogous to [42,84]. Thus ku and ke are both affected by the growth of an organism in the same way: they depend on an area/volume, or decrease with a measure of length. Please note that volumes can be transferred into weights any time by applying the density of the organism (with dimension m L-3), which may be assumed constant. A change of the volume of the organism has another effect on the body concentration. Since the amount is divided by the different volume, we see that the chemical is “diluted” when the organism grows and “concentrated” when the organism shrinks. For PCBs in fish, growth dilution may be the most important elimination route [82]. This can be included in the differential equation as a term affecting the concentration in fish by the relative growth rate. In its most general form [42]: Vb1−/ref3 dC b dVb 1 = k e−ref 1 / 3 (K bwC w − C b ) − C b dt dt Vb Vb

Eq. 16

Note that this formulation makes no assumptions on how the animal grows, only that Vb is a function of time. In case the organism grows (or shrinks) exponentially, the relative growth rate is a constant factor (the exponential growth rate)1, which has the same dimensions and the same effect as an elimination rate constant. Often, a total “elimination” rate is mentioned in ecotoxicological studies, which is made up of these two terms. The fact is however ignored that the rate constants themselves change when the volume changes. Admittedly, the change in rate constant is not so dramatic as the effect of growth dilution, as the first one changes with the third root of the volume. In short-term experiments, growth can safely be taken as exponential, but the situation becomes more complicated when growth slows down as the organism reaches maturity. This is a general phenomenon, as organisms they do not grow indefinitely. The Von Bertalanffy curve follows from energetic considerations, and provides an excellent description for the individual growth of many species, as long as the food density is constant [42]:

b

(

dVb1 /3 = k B Vb1∞/3 − Vb1 /3 dt

)

w

Eq. 17

This equations yields a skewed S-shaped curve for body size (see Fig. 3), leading to a final volume Vb∞. The growth rate kB is known as the Von Bertalanffy growth rate. Another popular growth curve is the logistic equation, although this form is more often used for 1 Exponential growth is defined as dV /dt = k ×V , where k b exp b exp is the exponential growth rate. This implies that the last term in Eq. 16 can be replaced by kexp×Cb thus showing that exponential growth can be treated as a simple elimination term.

34

Theory of compartment models

concentrations (% of eq.)

100

100

80

80

60

60

40

40

20

20

elimination rate (1/d)

weight of organism (grams)

3 populations than for individual growth (e.g. in case of growing duckweed in a 2.5 pond, or yeast in a batch culture). To give an example of the effect of growth 2 on accumulation for earthworms; [42] mentions for the earthworm Eisenia 1.5 veneta1 a Von Bertalanffy growth rate of 12.04 per year and a final volumetric 1 length of 14.5 mm. In more familiar units, this converts to a kB of 0.033 d-1 0.5 and a final weight of 3.0 grams (see Fig. 3). For a hydrophobic chemical like 0 0 50 100 150 benzo[a]pyrene, an elimination rate time (days) constant of 0.27 may be taken at a reference weight of 0.2 grams [38] Figure 3. Von Bertalanffy growth curve for the (although this was derived for a earthworm Eisenia veneta. different species: E. andrei). Another simulation run was performed with a ke of 0.02, the value found for hexachlorobenzene for worms in a water-only experiment [4]. Using these assumptions, the accumulation curves as shown in Figure 4 are obtained.

0.8

0

0

50

100

150

0

50

100

150

0.2 elimination rel. growth

0.6

0.1

0.2

0.05

0

50 100 time (days)

elimination rel. growth

0.15

0.4

0

0

150

0

0

50 100 time (days)

150

Figure 4. Top: accumulation curves for organisms of different age, assuming a Von Bertalanffy growth curve (Fig. 3), for two chemicals with different elimination rates: ke = 0.27 (left) and ke = 0.02 (right). Bottom: corresponding elimination rates and relative growth rates as function of time.

Clearly, the effect of the size of the worm on the uptake kinetics is visible. With the slow elimination rate, we can clearly see that the curve is deviating from a one-compartment pattern. For benzo[a]pyrene, growth dilution is less important than elimination (lower left panel of Fig. 4), whereas for the smaller rate constant, the growth dilution dominates as 1

This species used to be called Dendrobaena veneta. 35

Chapter 2

long as the worm is less than 40 days old (lower right panel of Fig. 4). When working with growing animals, it is important that we are aware of these effects, and incorporate them into our models, if needed. SATURATING UPTAKE KINETICS Earlier, we have assumed that the uptake flux into the organism was directly proportional to the external concentration (see Eq. 1). This seems to be a valid assumption for neutral organic compounds for which passive diffusion is the transport mechanism. For metals, however, it is usually observed that uptake rates, as well as BCFs, decrease with increasing external concentration [36,77,97], while the elimination rate is independent of external concentrations [27]. Metal ions have difficulties passing lipid membranes, so metal uptake is probably largely mediated by carrier molecules [58]. As the number of carriers is limited, it is conceivable that the uptake is saturating at high external concentrations. One way to describe this saturation is by Michaelis-Menten kinetics, that is usually applied to enzyme reactions. Michaelis-Menten kinetics is often presented in this form:

V (S ) =

Vmax S Km + S

Eq. 18

(The symbols in Eq. 18 are the ones usually applied in enzyme kinetics, and differ from those in this thesis). The rate of a certain enzymatic process (V) is a function of the substrate concentration (S). The maximum rate is given by Vmax, and when the substrate concentration equals the half-saturation level (Km), V is half the maximum value. The uptake flux into an organism can be represented by the uptake rate constant (ku) times the external concentration (Cw). However, in this situation, ku is not a constant but a function of the external concentration. The Michaelis-Menten equation can be rewritten to:

k u (C w ) =

k u−max C 1+ w C w 50

Eq. 19

At low concentrations, we see the maximum uptake rate ku-max, but as the external concentration increases, ku decreases. When the external concentration equals the halfsaturation concentration (Cw50), ku is half the maximum value. As the external concentration goes to infinity, so ku decreases to zero. Combining Equation 19 with Equation 1 gives: dC b k = u−max C w − k eC b C dt 1+ w C w 50

Eq. 20

This description has been successfully applied to uptake of cadmium in carp [90]. The effect of this form of saturation is shown in Figure 5. At a given external concentration, the curve is a simple one-compartment curve. However, as the external concentration increases, the steady-state BCF decreases. When the external concentration is low in relation to the saturation concentration (Cw50), it may not be obvious that the system can saturate (see e.g. the difference between cadmium and thallium in [46]). A form of saturation is seen in Chapter 4 for pyrene in earthworms, although the mechanism in that

36

case is different. At high concentrations in soil, the maximum water solubility is reached, and added pyrene will remain in the crystalline form, unavailable to earthworms.

internal conc./external conc. (% of max.)

Theory of compartment models

100

0.01 x Cw50

90

0.1 x Cw50

80 70 60 50

1 x Cw50

40

BIOTRANSFORMATION 30 For an overview of 20 biotransformation, textbooks 10 x Cw50 10 like [93] provide a good 0 starting point. This process is 0 2 4 6 8 10 12 14 16 18 20 time (days) more often included in models dealing with human health (as Figure 5. Internal concentration (relative to external in the physiologically-based concentration) with a saturating uptake, shown at different pharmaco-kinetic, or PB-PK, external concentrations (given as fraction of the halfmodels) than in ecotoxicology saturation concentration, Cw50). (this is probably because transformation rates in mammals tend to be higher than in fish and, especially, invertebrates). Nevertheless, for the sake of completeness, a brief discussion is included in this chapter. In contrast with fish, there is little evidence for biotransformation in earthworms. In some cases, a peak-shaped accumulation curve has been interpreted as indication for transformation [12,75]. However, these patterns likely represent problems with sequestration or degradation (see next section). Although earthworms are able to transform pyrene, the amount of metabolites formed is very low [38], and the P450 system is apparently not induced by exposure to PAHs [1]. Nevertheless, there is evidence of transformation of the relatively hydrophilic pesticides aldicarb [14] and oxamyl [86]. Because metabolites are usually more hydrophilic than the parent compound, biotransformation can act as an additional elimination route for chemicals. As passive elimination rates tend to decrease with Kow, even low biotransformation rates can influence the toxicokinetics [26]. The result of biotransformation will thus be a higher elimination rate and a lower body residue than predicted on the basis of hydrophobicity, as for example observed for several organophosphorous pesticides in fish [23]. As long as the biotransformation rate is constant, we can add the rate constant to the passive elimination rate (ke) to obtain an overall elimination rate. However, transformation is an active process, implying that we may need to apply Michaelis-Menten kinetics. In this case, it needs to be applied to an additional elimination flux, instead of an uptake flux (see e.g. [42,85]):

k dC b = k uC w − k eC b − m−max C b C dt 1+ b C b 50

Eq. 21

Here, km-max is the maximum rate constant for biotransformation; Cb50 is the half-saturation concentration in the organism. This formulation suggests that rate of the removal flux due to transformation saturates with increasing external concentration. Furthermore, this equation assumes that the (passive) elimination rate of the metabolites is fast compared to the (saturating) transformation of the parent compound. If this is not the case, a morecompartment approach can be followed, in which the kinetics of the parent compound

37

Chapter 2

and the metabolites are modelled separately [78]. However, the biggest problem in modelling biotransformation will be the induction of the enzyme production. Induction has the consequence that also the km-max changes in time (probably as a reaction to internal concentration on a specific target). Furthermore, the presence and activity of enzymes may depend on the development stage of the organism, which further complicates a general picture of this process. A practical approach to provide indicative rate constants for biotransformation was presented by Van der Linde et al. [88]. Using a model based on the two-phase resistance mechanism, these authors take the biotransformation rate constant as the difference between expected and measured elimination rates. This analysis for example indicated that in Annelida, PAHs are the only group of chemicals for which the total elimination was influenced by transformation. ELIMINATION BY REPRODUCTION The production of offspring by organisms may form an additional elimination route for chemicals. The females will lose a fraction of their chemical burden through this process, and, depending on the reproductive effort of the species, this may be significant (for guppy, this was indicated by [82]). At the same time, however, the body weight of the female decreases through reproduction. Assuming that the eggs or the young have the same body composition (read: lipid content) and chemical concentration as the parents, the net result is zero. In this way, reproduction can be compared to growth dilution: it is production of new biomass although it is not attached to the parent anymore. However, when the eggs or offspring contain more fat than the parent, or in the case of mammals that feed their young milk, this route can seriously affect the toxicokinetics. When chemicals are not transferred to the offspring, reproduction will still affect body residues as the weight of the female drops (thereby concentrating the remaining chemical). A problem with modelling this process is that, in many species, reproduction is not continuous. Therefore, this process can lead to a kind of “saw tooth” accumulation curve [82]. In any case, reproduction is a time-dependent process, which depends on the body size (which varies in time) and feeding conditions. Therefore, the effect of reproduction on body concentrations is best addressed in conjunction with a model for the growth and development of an organism. This is beyond the scope of this chapter, but good examples are available [41,42]. A rather special toxicokinetic problem is presented by fish larvae. The egg membrane seems to slow down the intrusion of chemicals into the developing embryo [92]. After hatching, many species have what is known as a sac-fry stage (in which the yolk is attached to the larvae). Hydrophobic chemicals seem to preferentially accumulate in the yolk, and the resorption of the yolk by the growing larvae leads to high body burdens in the larvae themselves, as well as toxicity, at the transition to the swimming fry stage [80,92].

VARIABLE EXPOSURE CONCENTRATIONS The solution of the differential equation for bioconcentration (Eq. 2) is only valid when the exposure concentration remains constant over time. If this is not the case, morecompartment approaches are required. Ecotoxicologists are often hesitant to move towards more compartments, partly because they feel uneasy about working with differential equations, and because analytical solutions are not always possible (thus requiring numerical simulation).

38

Theory of compartment models

EXPOSURE CONCENTRATIONS INFLUENCED BY THE ORGANISM: DEPLETION In the previous discussions, the water phase was assumed to be infinitely large. As a consequence, the uptake of chemicals by the organisms has a negligible effect on the water concentration itself. As long as the water concentration remains constant, the bioconcentration process can be described with a one-compartment model. However, in many experimental as well as environmental situations, this is not the case. When testing very hydrophobic chemicals, the capacity of the fish for the chemical is orders of magnitude higher than that of the water, and one risks depletion of the chemical pool in the water. This is why many tests are performed under “flow-through” conditions (see [93] for a discussion of exposure systems for aquatic studies). For example, consider an accumulation experiment with fish and a chemical with a log Kow of 6. We may estimate Kbw from the lipid fraction of the fish (say 5%) and the Kow, resulting in a Kbw of 50,000. Suppose we tolerate some 10% depletion of the water, so we take the water’s capacity for the chemical at 10 times that of the fish. This implies we need to take the water volume half a million times as large as that of the fish. In other words, when we have 10 grams of fish in the aquarium, we require some 5,000 litres of water. It is clear that depletion can be a serious concern.

Soil organisms may also encounter the effects of depletion in the laboratory as well as in the field. However, in contrast to the aquatic situation, it is usually less clear that depletion has occurred. Generally for hydrophobic chemicals, the soil’s capacity for a chemical is large enough to prevent depletion of the total concentration. However, chemicals are not directly taken up from the solid phase but from the aqueous phase (the chemical dissolved in pore water). Hydrophobic chemicals tend to sorb heavily to organic matter in soil and will thus have very low concentrations in pore water. It is therefore possible that the small dissolved pool is depleted by the organism when replenishment is slow. Replenishment can come from the sorbed phases (by desorption) or from another part of the soil by mass transfer (e.g. diffusion). Desorption of very hydrophobic chemicals like PAHs and PCBs is a slow process, especially for older (“aged”) contamination [21]. The organism may thus deplete the bioavailable phase even when the total soil concentration remains constant, and the desorption rate could govern the uptake kinetics [48]. A further problem may be the heterogeneity of contaminants in the soil. Soil is not a “well-mixed compartment” and diffusion of chemicals in soil is slow. As a result, the chemical concentration in the soil may show a heterogeneous pattern. The soil-dwelling organism is therefore not exposed to the pore water of the entire world, but rather to a limited personal “influence sphere”. In this personal sphere, the organism may deplete the local dissolved phase even when this is not obvious from the total soil or total dissolved concentrations. A solution to this problem is not so easy, as soil tests cannot be performed under continuous-stirring or flow-through conditions. However, the experiments of Lord et al. [56] clearly demonstrated this phenomenon: uptake of dieldrin into earthworms was higher when the soil was regularly stirred. Clearly, one needs to be aware of this problem in soil tests, especially when the organism:soil ratio is high in a test or when the organisms are not moving through soil (plant roots, earthworms in diapause, and organisms experiencing narcotic toxic effects). For situations where the organism depletes its environment, a two-compartment approach is needed. It is essential that we start to define the differential equations on the bases of masses and not as concentrations. Chemical masses are preserved when the chemical moves from one compartment to another, but concentrations are not (unless the volumes of both compartment are the same):

39

Chapter 2

dN b = mb+ Abw (K bwC w − C b ) dt

Eq. 22

dN w = mb+ Abw (C b − K bwC w ) dt

Eq. 23

Note that this system is completely closed, with no losses due to volatilisation or degradation, and sorption to other phases (like particles or glass walls). The differential equations for the mass fluxes (dN/dt) of the two compartments differ only in sign; the sum of both fluxes is zero because no chemical leaves or enters the two-compartment system. This is called a “closed mass balance”:

b

dN b dN w + =0 dt dt

w

Eq. 24

The differential equations can now be written in the form of concentrations, through division by the volume of the compartment. This is usually more convenient as concentrations are the properties that are measured: A dC b = mb+ bw (K bwC w − C b ) dt Vb

Eq. 25

A dC w = mb+ bw (C b − K bwC w ) dt Vw

Eq. 26

If we now like to express this system in terms of rate constants instead of mass-transfer coefficients, we face a dilemma. A rate constant is, by its very nature, related to the volume of the compartment it acts upon. It is also determined by the contact area, but this is the same on the organism side and the water side. Usually, the choice is made to reference the elimination rate constant to the organism: dC b = k e (K bwC w − C b ) dt

Eq. 27

But this has the logical consequence that this rate constant has to be corrected in the equation of the water for the ratio of both volumes: dC w V = k e b (C b − K bwC w ) dt Vw

Eq. 28

If this is not done, the principle of mass balance is violated and we have an inconsistent system. Furthermore, we have to be aware of the point of perspective when using rate constants from the literature (e.g. desorption rate constants are referenced to the solid phase). This notation (Eq. 27) is convenient for most bioconcentration purposes but it must be noted that the choice of perspective is an arbitrary one. We could also have expressed the concentration difference in terms of the water phase, which leads to different rate constants:

40

Theory of compartment models

dC b A = mw+ bw dt Vb

 C   C w − b  K bw  

Eq. 29

By comparing Equation 25 and 29, it is clear that mb+ and mw+ differ only in a factor Kbw. TIME TO EQUILIBRIUM AND Kbw UNDER DEPLETION The fact that the organism depletes the water phase also affects the time needed for the system to reach equilibrium. As uptake is proceeding, the organism is approaching equilibrium, but the water concentration with which it wants to equilibrate is decreasing. This implies that there is less time needed to achieve equilibrium in such a system. In equilibrium, the Kbw will still describe the ratio of the concentrations in organism and water. However, when only the initial concentration in water is known, Kbw cannot be directly derived from the animal’s concentration. However, in this simple system, both the time needed for equilibrium and the apparent Kbw can be derived analytically. The total amount of chemical in the two-compartment system (N+) can be written as:

N + = C bVb + C wVw

Eq. 30

Rearranging this equation provides an alternative expression for the concentration in water: Cw =

N + − C bVb Vw

Eq. 31

which in turn can be inserted in Equation 22, as derived previously:   dN b N − C bVb = mb+ Abw  K bw + − C b  dt Vw  

Eq. 32

Which can be rewritten as:   V dN b N = mb + Abw K bw + −  mb + Abw K bw b + mb + Abw  C b Vw  Vw dt 

Eq. 33

Several observations can be made, based on this equation. Firstly, the form of the equation is similar to Equation 1. This implies that this two-compartment system will lead to a simple one-compartment accumulation pattern in the organism. Even though the water phase is depleted, this will not be obvious from the accumulation pattern in the organism, which will show a smooth increase to steady state. However, the height of the steady state and the speed with which it is attained now depend on the volume of the water phase and Kbw. From Equation 33 we can derive that the final body residue in the organism is given by: C b∞ = C w 0

1 Vb 1 + Vw K bw

Eq. 34

Or, when written as the apparent Kbw:

41

Chapter 2

K bw−app

V  C = b∞ = K bw  b K bw + 1  C w0  Vw 

−1

Eq. 35

When Vw is very high compared to Vb, Equation 34 reduces to the simple accumulation equation Cb∞ = Cw0×Kbw. However, if Kbw is 100 and Vw is 100× larger than Vb, the final concentration in the organism is half of what we can expect when the water volume is infinitely large. Note that in equilibrium, it is always true that Cb∞ = Cw∞×Kbw. So if we measure the concentration in water at the end of the experiment, we still arrive at a valid measure of Kbw. Depletion of the water phase not only influences the final body residue in the organism, but also the speed with which equilibrium is reached in the system. By dividing Equation 33 by Vb and comparing the result to Equation 1, it follows that the apparent rate constant can be written as: k e − app = m b +

Abw A K bw + m b + bw Vw Vb

Eq. 36

or, when looking at Equation 10: k e−app = k e

V  Vb K bw + k e = k e  b K bw + 1  Vw  Vw 

Eq. 37

As long as Vb/Vw is much smaller than Kbw, we will observe the “true” rate constant when we observe the organism. Problems occur when Kbw is high and/or Vb is high compared to Vw. For example, if Kbw is 100 and Vw is 100× larger than Vb, we will see equilibrium reached at twice the speed of what we observe when the water volume is infinitely large. The factor with which the rate constant appears to increase is the same as the factor with which the final body residue decreases. This behaviour is shown in Figure 6. As the presented analysis shows, the telltale marks of depletion are the rapid attainment of steady state at a lower level than expected. Furthermore, a depleting system can be recognised when an elimination experiment is performed by placing the loaded organism in clean medium: the elimination rate constant from this experiment will be lower than the one calculated from the accumulation experiment. The fact that depletion occurs in a system does not invalidate that experiment. The “true” bioconcentration factor and rate constants can still be derived from the experiment as long as the proper model is used (Eq. 33). An example is provided in the work of Banerjee et al. [2], who proposed to use static tests with small volume to estimate bioconcentration factors and rate constants, by only measuring the decrease in the water concentration. This method, of course, rests on the assumption that the chemical is lost from the water only due to uptake into the fish, precluding adsorption, degradation and volatilisation. A comparable effect of small exposure volume occurs in elimination experiments. A loaded animal is placed in clean medium and allowed to eliminate the chemical from its tissue. However, as the chemical leaves the organism’s body, it pollutes the medium. In time, an equilibrium will be achieved with the organism having a concentration that is non-zero. An added difficulty is that tissue concentrations of eliminating organisms are usually plotted on log scale, making it hard to appreciate the amount of chemical that the organism has introduced into the experimental container. A simulation of this effect is shown in Figure 7.

42

concentrations (% of max.)

100

concentrations (% of max.)

Theory of compartment models

100

100

80

80

60

60

40

40

20

20

100

80

80

60

60

40

40

20

20

0

0

0

0

5

5

10

15

10 15 time (days)

20

20

0

0

0

5

0

5

10

15

20

10 15 time (days)

20

concentrations (% of max.)

concentrations (% of max.)

Figure 6. Simulations of a depleting system. Solid line is the concentration in the organism; dotted line is the water concentration. The sequence left to right, top to bottom shows the effect of a decreasing water volume. 100

100

80

80

60

60

40

40

20

20

0

0

5

10

15

20

0

100

100

80

80

60

60

40

40

20

20

0

0

5

10 15 time (days)

20

0

0

5

0

5

10

15

20

10 15 time (days)

20

Figure 7. Simulations of elimination from an organism in a small experimental container. Solid line is the concentration in the organism; dotted line is the water concentrations. The sequence left to right, top to bottom shows the effect of a decreasing water volume.

43

Chapter 2

Clearly, when the water volume decreases, the elimination pattern changes from an exponential decay to a curve decreasing to a new equilibrium level. A very clear example of this pattern was shown for PCBs in earthworms [50]. I believe that this process may well explain some of the claims that are made for bi-phasic elimination in organisms. Biphasic elimination has been observed in some cases in earthworms [6,9] and fish [39,40]. Although I do not claim all these results to be artefacts, it is important to first ensure that the observed pattern is not caused by re-uptake from the medium, before turning to morecompartment models for the organism. There are two conditions to be met before adding compartments to the organism: can we give a plausible mechanism for the compartments, and is the second compartment really contributing to the toxicokinetics? To clarify the latter, when the second compartment only contains a small fraction of the total body residue, it can safely be ignored for most applications. Again, plotting the elimination phase on logarithmic scale tends to overstress the importance of the second phase. INFLUENCE OF DESORPTION The analysis described above is valid for aquatic situations, or soil situations where the water phase is not replenished. However, when a soil contamination is left for sufficiently long, an equilibrium will be established between the dissolved phase and the sorbed phase (mainly organic matter for non-polar organic compounds). When the water concentration decreases, this will initiate a mass flux from the sorbed phase to the pore water. This system is slightly more complicated than the previous one because of the addition of a third compartment (soil) and two more constants (the soil-water partition coefficient Ksw and the desorption rate constant ks). The mass-balance equations of this system are:

b

w s

dN b = mb+ Abw (K bwC w − C b ) dt

Eq. 38

dN w = mb + Abw (C b − K bwC w ) + ms + Asw (C s − K swC w ) dt

Eq. 39

dN s = ms+ Asw (K swC w − C s ) dt

Eq. 40

And in terms of concentrations and rate constants: dC b = k e (K bwC w − C b ) dt

Eq. 41

V V dC w = k e b (C b − K bwC w ) + ks s (C s − K swC w ) dt Vw Vw

Eq. 42

dC s = ks (K swC w − C s ) dt

Eq. 43

Note that here the desorption rate constant (ks) is referenced to the soil phase, and therefore needs to be corrected in the equation for the water phase. In many cases, this system can be simplified by taking the soil concentration constant. Although it is not true that the soil concentration indeed remains constant, its decrease may often be too little to

44

Theory of compartment models

concentrations (% of max.)

concentrations (% of max.)

notice. This system is more difficult to examine analytically, and therefore, only simulations are given in Figure 8. Three scenarios are calculated: 1) desorption in infinitely slow (this is the same as in the previous section), 2) desorption is infinitely rapid (this is the same as in Eq. 1), 3) desorption is of intermediate speed. In most cases, the system will tend to one of the two extremes. Only for specific parameter values, the intermediate pattern will occur. There are indeed indications from the literature that desorption or transport rates may limit biodegradation [16,74], and uptake by sediment organisms [43,48], fish [69], and earthworms (Chapter 4, 5). 100

100

80

80

60

60

40

40

20

20

0

0

5

10

15

20

0

100

100

80

80

60

60

40

40

20

20

0

0

5

10 15 time (days)

20

0

0

5

0

5

10

15

20

10 15 time (days)

20

Figure 8. Simulation of a depleting system. Solid line is the concentration in the organism; dotted line is the water concentration. The sequence left to right, top to bottom shows an increasing rate of desorption from zero to very rapid.

EXPOSURE CONCENTRATIONS INFLUENCED BY EXTERNAL FACTORS It is not only possible that the organism itself depletes the water phase, other processes can also bring about a decrease in the water concentration. Examples include (bio)degradation and volatilisation. Although at first sight, this looks like a similar problem as the depletion from the previous section, it leads to a very different behaviour. The equation for the organism is the same as Equation 27 and 41, but the one for the water phase is different:

b

dC w = − k dC w dt

w

Eq. 44

The water phase, in this case, does not depend on the organism’s concentration at all but decreases exponentially in time (taking Vw as very large, to preclude depletion). Because the water concentration will eventually go to zero, the organism concentration will also go

45

Chapter 2

concentrations (% of max.)

concentrations (% of max.)

to zero (see Fig. 9). This is different from depletion by the organism itself were the decrease in the water concentration will seize when equilibrium is obtained in the system. 100

100

80

80

60

60

40

40

20

20

0

0

5

10

15

20

0

100

100

80

80

60

60

40

40

20

20

0

0

5

10 15 time (days)

20

0

0

5

0

5

10

15

20

10 15 time (days)

20

Figure 9. Simulations of a system in which the chemical is degraded in the water phase. Solid line is the concentration in the organism; dotted line is the water concentration. The sequence left to right, top to bottom shows an increasing rate of degradation from zero to rapid.

In a soil system, degradation will not always have such a straightforward effect. Just as uptake by the organism, biodegradation and volatilisation will primarily occur from the water phase. However, as the degradation of the water phase proceeds, a mass flux will initiate, attempting to replenish the dissolved phase from the sorbed phases. The net effect of these processes on the accumulation pattern in the organism is hard to predict, as it will depend on the size of the rate constants and partition coefficients. Furthermore, in case of biodegradation, the effect may depend on non-linear kinetics of a growing bacterial population. These peak-shaped uptake patterns have been observed mainly for PAHs, in earthworms [59,63] (see also Chapter 4, 5), sediment amphipods [43,48], and polychaetes [44,73,91], although it cannot be proven that biodegradation is the cause; e.g. a rapid firstorder sequestration or induced biotransformation will produce similar results. In a specific situation, even a constant rate of biotransformation can produce the peak-shaped accumulation curves as shown in Figure 9. As shown by De Maagd et al. [24], peak-shaped curves will result when a compound is significantly biotransformed in a static test system (i.e. a depleting environment). When the water concentration remains constant (i.e. no depletion), biotransformation will lead to a standard accumulation curve (but to a lower level than expected without transformation). However, in the depleting system, the chemicals that are transformed cannot be replenished, as chemicals are lost from the

b

46

w

Theory of compartment models

system. Therefore, this situation is quite comparable to degradation in the water phase, although the loss term is now from the organism compartment. INFLUENCE OF DISSOLVED ORGANIC PHASES For hydrophobic contaminants, the concentration dissolved in water is in many cases not fully bioavailable. Organic molecules dissolved in water may sorb appreciable amounts of chemical, thus rendering them unavailable for uptake by the organism. These molecules can be humic and/or fulvic acids, and are collectively designated by the term dissolved organic matter (DOM). Evidently, DOM-bound chemicals are not directly available for uptake by organisms [10,47], because chemicals bound to DOM do not influence the chemical’s fugacity in the water. The effect of DOM on accumulation for aquatic organisms has been reviewed by [33]. The following equations result if desorption from DOM is assumed to be rapid, in comparison to the uptake kinetics of the chemical in the organism:

b

(

dC b = k e K bwC w F free − C b dt

(

)

dC w V = k e b C b − K bwC w F free dt Vw

w dom Eq. 45

)

Eq. 46

The fraction that is freely dissolved in the pore water is calculated from the concentration of DOM and the DOM-water partitioning (Kdom) [17]: F free =

1

1 + K dom [DOM ]

Eq. 47

This model system is simulated in Figure 10. It is clear that the uptake decreases when DOM is added to the system, when you relate uptake to the total water concentration. However, the uptake pattern related to the dissolved phase remains unchanged. In pore water from soils and sediments, DOM can occur in high concentrations (see Chapter 5). However, in soil we also have a large pool of chemicals bound to soil organic matter (OM). When all the soil’s phases are in equilibrium, the pore water will also be in equilibrium with soil OM. This implies that sorption (Koc) is not influenced by DOM (as long as Koc is related to the freely-dissolved concentration), and neither is the Kbw (also related to free concentrations). In other words: adding DOM to a soil system will not influence the body residues of the organisms. There are some preconditions: the amount of DOM added must not be so high as to decrease the amount of chemical sorbed to soil OM. Furthermore, DOM may be facilitating uptake in case the organism threatens to deplete the pore water by providing a rapidly desorbing buffer phase. If sorption to DOM is a significant process in a system, pore-water measurements, based on total pore-water concentrations cannot be directly related to uptake (see Chapter 5). It is necessary to have a good estimate of the freely-dissolved pool, or estimate the dissolved concentration from QSAR-derived Koc values.

47

concentrations (% of max.)

concentrations (% of max.)

Chapter 2

100

100

80

80

60

60

40

40

20

20

0

0

5

10

15

20

0

100

100

80

80

60

60

40

40

20

20

0

0

5

10 15 time (days)

20

0

0

5

0

5

10

15

20

10 15 time (days)

20

Figure 10. Simulation of a system without depletion, but with dissolved organic matter (DOM). Solid line is the concentration in the organism relative to the total water concentration; broken line is the concentration in the organism, related to the freely-dissolved phase. The sequence left to right, top to bottom shows the effect of an increasing concentration of DOM.

UPTAKE FROM FOOD It has long been recognised that chemicals are not only taken up through the skin or through respiratory surfaces. Hydrophobic chemicals in the food are also taken up into the tissues by fish [31,55] and earthworms [5,96]. Feeding may constitute a significant uptake route for PAHs [52,101], metals [81] and hexachlorobenzene [11] in sediment organisms. It is furthermore conceivable that feeding is the most important route for soil organisms with a hard outer surface, like isopods [87]. However, care must be taken in interpreting these studies, as it is almost impossible to study both uptake routes in isolation under the exact same conditions. The experimental set-ups do not usually allow quantification of the contribution from water and food [58]. There is however good evidence that uptake from the gut contents obeys the same rules as uptake through other surfaces. Note that a chemical in the gut contents is technically not taken up yet; the gut contents are a piece of outside world inside the organism. Most organic chemicals are not actively taken up from the gut into the tissue, so we are still dealing with an essentially diffusion-driven process [31,67,99]. Another possible route could be co-transport with the active uptake of small particles or lipid droplets across the membrane (pinocytosis). However, this process is highly inefficient for chemical transport [67], and also lipids are primarily taken up by diffusion [29]. Although the dominant mechanism of chemical uptake is by passive diffusion, the situation is different from the other uptake routes in the following respects:

48

Theory of compartment models

1) Refreshment of the exposure medium (the gut contents) is relatively slow, and the volume of this medium is limited. This means that we quickly have to deal with depletion of this exposure phase. 2) The animal is digesting sorption sites (lipids and OM), and may add all kinds of secretions to the gut contents (e.g. saliva, digestive enzymes, acids) that may change the bioavailability. SIMPLE MODELLING The proper mechanistic manner to address the process of gut uptake is by adding a compartment to the model. Uptake from the gut contents is a passive diffusion process from the dissolved phase in the gut contents, and the concentration in the gut contents is determined by advective processes (ingestion and egestion) and digestive processes (digestion of sorption sites and compaction of its contents). However, this modelling strategy requires quantitative knowledge on the feeding behaviour and physiology of the animal (see Section C of this thesis for more details on the modelling). It is probably for this reason that many researchers apply a more simple model, treating uptake from food as a process that can be added to uptake from water by diffusion:

dC b = k uC w + k f C f − k eC b dt

b

w

g

f ege

Eq. 48

With the uptake rate from food (kf) given by the feeding rate times the assimilation efficiency of the chemical from the food (see for applications in aquatic studies e.g. [55,98], and for earthworms [5,96]). The elimination rate ke represents, in this situation, elimination through all possible routes (including with the faeces). For some strange reason, this type of model is sometimes referred to as “bioenergetics-based” (see e.g. [49]). This formulation has been extended to estimate the rate constants for uptake from food and elimination to the faeces [37,88]. These extensions are based on the two-resistances model (see Eq. 12), and include a flow delay (the gut transfer time may limit accumulation), as well as allometric scaling. A problem with the formulation of Equation 48 is that it suggests that the steady-state body residue will always increase due to feeding. When there is only uptake from water, Cb∞ equals ku/ke×Cw, but when the organism is also feeding, the body residue is always higher:

C b∞ =

kf ku Cw + Cf ke ke

Eq. 49

This is however not true. The confusion is generated by the fact that ke is also influenced by feeding. The rate constant for elimination will be higher when an organism is feeding, as elimination with the faeces becomes an important excretion route. So it is better to introduce an additional elimination rate with the gut contents (kg) in Equation 48, which means that Equation 49 becomes:

C b∞ =

kf ku Cw + Cf ke + k g ke + k g

Eq. 50

49

Chapter 2

Now it is less clear how the final body residue will differ from ku/ke×Cw, because it depends on the concentrations in food and water, and on four rate constants. This formulation makes clear that uptake follows two routes, but elimination does too. It is often valid to use an overall elimination rate to describe accumulation from food (see e.g. [96]), but one has to be careful in using this formulation to describe uptake from both routes simultaneously. One example where this equation is incorrectly used is [7]. These authors want to model accumulation in earthworms, and include the effects of feeding on soil particles, as well as uptake from pore water. A value for kf can be derived from experiments where worms are fed contaminated food, but an estimate for ku is more problematic. This parameter cannot be derived from soil experiments as the gut route will always contribute, and the relevance of water-only exposure is questionable. These authors rewrite Equation 50 to remove ku by assuming that ku can be derived from the bioconcentration factor in water-only exposure (Kbw) and the total elimination rate in soil:

K bw =

(

ku → ku = K bw k e + k g ke + k g

)

Eq. 51

However, this is not true; the uptake rate from water cannot be derived from Kbw and the total elimination rate, but only when the elimination to pore water alone (ke) is known (see also [98]). The result of this conceptual flaw is a model in which feeding on soil particles always leads to deviations from equilibrium partitioning (EP) (although the deviation can be very small when the contribution from feeding is small). Another limitation of this model solution is that results cannot be extrapolated to different situations because the uptake and elimination rate from food (kf and kg) not only depend on the feeding rate, but also on digestibility of the food [32] and the gut loading. Extrapolation to soils with a different OM contents [7] is therefore invalid, as increasing OM will also decrease the chemical assimilation efficiency from the food (because the sorption in the gut will increase proportionally). In summary, the model of Equation 50 may be used to describe uptake from food if a description of an experiment suffices. However, care must be taken that the elimination rate consists of two rates (to water and with the faeces). Furthermore, this model form does not allow useful extrapolations to other situations, and provides little insight in the accumulation process. In the more complex mechanistic model (see Chapters 8 and 10), more data are needed, especially on the feeding activity (e.g. gut load, digestion efficiency of OM, and gut retention time). EFFECT OF SECRETIONS Many organisms will add secretions to the gut contents to aid the digestion process. Biosurfactants may increase the solubility of a chemical in the gut contents [64], largely due to the surfactant micelles [95]. The effect of secretions on chemical solubility has mainly been studied in marine species [65]. The fact that secretions are involved in uptake from the gut does not mean that the mechanism is not by simple passive diffusion. Even though in mammals the uptake through the gut wall is largely mediated by micelles, the available evidence supports a diffusion-driven process [67] (e.g. because the process is bidirectional). Surfactants may speed up the rate of desorption and mass transport of the chemical, they do not change the fugacity gradient between chemical dissolved in the gut fluids and in the animals tissues. Even though the dissolved concentration in the gut fluid has increased, the surfactants have also changed the solvent properties of the gut fluid. In fact, the gut fluid is now a less hostile environment for the chemical, compared to water, and the chemical has less “urge” to flee from it. The gut fluid is not water anymore, but a solvent with a higher affinity for the chemical. In other words, the fugacity capacity is

50

Theory of compartment models

increased by the same amount as the dissolved concentration is increased (see Eq. 3). If we take a compartment of water and a compartment of gut fluid (with concentration Cg), we can observe a partition coefficient as:

K gw =

Cg Cw

Eq. 52

This partition coefficient will be larger than one if solubility is increased in the gut fluid. The partition coefficient between water and organism tissue is given by:

K bw =

Cb Cw

Eq. 53

So the partition coefficient between organism and gut fluid is now given by:

K bg =

C b K bw = C g K gw

Eq. 54

This implies that increasing the Kgw decreases the driving force for uptake (determined by Kbg) by the same degree. Thus, steady-state body residues in organisms cannot be increased by increasing the solubility of the chemical in its environment (either external or in the gut contents). The same point was raised by Lu et al. [57]. As an illustration, consider the following system: a closed vessel with an air and water phase. A hydrophobic chemical is introduced into the system and allowed to travel freely between the phases. After some time, the system is in equilibrium and the air-water partition coefficient is given by:

K aw =

Ca Cw

Eq. 55

As we have seen earlier, the fugacities of the chemical in air and water are equal in equilibrium:

fa =

Ca C = fw = w Za Zw

Eq. 56

What happens when we add some ethanol to the system? Ethanol is miscible with water and will decrease the polarity of the aqueous phase. The solution can now contain more of the chemical, so the fugacity capacity of the water+ethanol solution is higher than that of the pure water. The chemical has less urge to flee from the solution. When we look at Equation 56, we see that an increase in Zw will decrease the fugacity fw. As fa and fw are not equal anymore, this drives a chemical transport from air to water until both fugacities are equal again. This new equilibrium will have more chemical in solution, so the Kaw will be lower. Now we change focus to another system: a closed vessel with an air, water and soil phase. Again when we add the hydrophobic chemical, we will see that an equilibrium between all three phases will be established so that the fugacities are equal:

air

water

51

Chapter 2

C a C w Cs = = Za Zw Zs

Eq. 57

Again we add some ethanol to the water compartment and see what happens. First we have to assume that the ethanol does not affect the properties of the phases (solvents tend to be detrimental to the organic matter in soil). Addition of ethanol will increase the fugacity capacity of the water phase (the water becomes a less hostile environment for the chemical, so more of the chemical will be dissolved). This will lead to a desorption flux from the soil phase to the water. Now assume that the chemical is highly sorbed to the soil phase and that desorption does not appreciably decrease the soil concentration. In the new equilibrium, the soil’s fugacity has changed little (Zs as well as Cs are unchanged). As a consequence, the fugacities in air and water cannot change either. However, the fugacity capacity of the water phase has increased due to the ethanol addition. The change in fugacity capacity is thus counteracted by an increase in Cw (coming from the soil compartment, which as we assumed, decreased little in concentration). In this case, even though the concentration in solution is increased, the concentrations in soil and air have not changed by adding the ethanol, so neither has the soil-air partition coefficient. If the soil concentration will decrease significantly by adding ethanol to the water phase, we will also see a decrease in the air concentration. In other words, the air concentration cannot be increased by adding ethanol to the water phase.

air

water soil

The effect of an addition of DOM to the water falls in the same category, as it increases the apparent water solubility but decreases the fraction of the water concentration that is bioavailable (see Eq. 45–47 and Fig. 10). Experimental work with excised mussel gills confirms that surfactants decrease bioavailability (as in Fig. 10), when there is no replenishment from sorbed phases [70]. Thus, measuring the total concentration in the gut fluid does not help to predict body residues because the Kgw is unknown, and cannot be estimated from Kbw. Similarly, measuring total pore-water concentrations does not help to estimate body residues in earthworms (see Chapter 5). It is better to estimate or measure the freely-dissolved concentration. What a surfactant may achieve, however, is that equilibrium is rapidly achieved, and may possibly free more heavily sorbed contaminants from soil or sediment [20]. Furthermore, we saw that DOM will counteract the effects of depletion in a system. Therefore, dissolution in the gut fluid may be related to body residues [100], although not through deviating from EP, as suggested [95]. MAGNITUDE OF FEEDING EFFECT FOR SOIL AND SEDIMENT FEEDERS The model with gut compartment is discussed in more detail in the Chapter 8 of this thesis. However, we can provide the following general observations for animals feeding on the medium they live in (e.g. earthworms). Digestion is the main process driving a deviation of the body residues from EP with the soil. Digestion destroys sorption sites for organic chemicals, and potentially decreases the volume of the gut contents (compaction). However, compaction also increases the concentration of the remaining sorption sites (Fom), and is therefore expected to contribute little to the deviation from EP. Roughly speaking, when the organism digests 50% of the OM in the gut contents (which would be very high for earthworms), this could double the fugacity of the chemical in the gut contents. In other words, the potential maximum deviation from EP is a factor of two. Whether this deviation is indeed reached depends on several factors:

52

Theory of compartment models

1. The values of the rate constants for uptake across the gut wall and the skin. When exchange across the skin is very slow, the animal can reach body residues in equilibrium with the gut (thus, higher than EP with the soil). 2. When the retention time is fast and the gut volume large, there is sufficient refreshment of the gut contents. This could prevent a depletion of the chemical pool in the gut contents. In short, digestion provides the maximum possible deviation from EP, but the actual deviation will depend on the kinetics of the various fluxes. The animal can end up anywhere between equilibrium with the soil and equilibrium with the gut contents. Selective feeding is an issue in this process, but its contribution is not straightforward. When the animal is selecting a diet from the medium that is twice as organic, it will ingest twice as much of the chemical. However, it will also ingest twice as many sorption sites, which makes the net result close to zero. There are however possible exceptions when there is substantial heterogeneity in the OM. When the organism is not indiscriminately selecting for OM, but for a specific form of OM (e.g. by selecting a specific particle size [45]), the effect on chemical concentration and sorption may be less predictable.

CONCLUSIONS Bioaccumulation of neutral organic compounds can be viewed as passive diffusion from a liquid phase (usually water) or a gas phase (usually air) to the organism. The driving force is the fugacity difference between the two phases, but can also be expressed in terms of concentrations and partition coefficients. Although the resistances theory underlying the rate constants is well-established and confirmed in environmental chemistry, its extrapolation to organisms is not beyond discussion. Most of the work has been done on fish, where there is sufficient evidence supporting this theory to grant it the status of a “reference model” (experimental data are discussed in relation to their deviations from, or consistency with, this model). The assumptions for applying the standard onecompartment model may be violated, especially for soil organisms. These violations may lead to typical deviations from the expected patterns. When these situations occur, morecompartment modelling is needed for a proper description, and examples are given in this chapter. Furthermore, additional measurements may be needed in such situations (e.g. of the pore-water concentrations, or desorption rates). Feeding can be an important uptake route for animals. However, it must be stressed that feeding does not imply uptake from a solid phase (an intermediate solution phase is always required). Preferably, this process should be handled with the gut as a separate model compartment, obeying the same principles as uptake across the skin or the gills.

ACKNOWLEDGEMENTS I would like to thank the following persons for their critical comments on earlier versions of this chapter: Bas Kooijman, Dick Sijm, Kees van Leeuwen and Joop Hermens.

53

Chapter 2

REFERENCES [1]

[2] [3] [4]

[5]

[6] [7]

[8] [9]

[10]

[11]

[12]

[13] [14] [15]

[16] [17] [18] [19] [20] [21]

[22]

[23]

54

Achazi RK, C Flenner, DR Livingstone, LD Peters, K Schaub and E Scheiwe (1998). Cytochrome P450 and dependent activities in unexposed and PAH-exposed terrestrial annelids. Comp. Biochem. Physiol. C 121:339-350. Banerjee S, RH Sugatt and DP O'Grady (1984). A simple method for determining bioconcentration parameters of hydrophobic compounds. Environ. Sci. Technol. 18:79-81. Barron MG (1990). Bioconcentration. Will water-borne organic chemicals accumulate in aquatic animals? Environ. Sci. Technol. 24:1612-1618. Belfroid A, A Van Wezel, M Sikkenk, K Van Gestel, W Seinen and J Hermens (1993). The toxicokinetic behavior of chlorobenzenes in earthworms (Eisenia andrei): experiments in water. Ecotox. Environ. Saf. 25:154-165. Belfroid A, J Meiling, D Sijm, J Hermens, W Seinen and K Van Gestel (1994). Uptake of hydrophobic halogenated aromatic hydrocarbons from food by earthworms (Eisenia andrei). Arch. Environ. Contam. Toxicol. 27:260-265. Belfroid A, M Sikkenk, W Seinen, K Van Gestel and J Hermens (1994). The toxicokinetic behavior of chlorobenzenes in earthworm (Eisenia andrei) experiments in soil. Environ. Toxicol. Chem. 13:93-99. Belfroid AC, W Seinen, KCAM Van Gestel, JLM Hermens and KJ Van Leeuwen (1995). Modelling the accumulation of hydrophobic organic chemicals in earthworms - Application of the equilibrium partitioning theory. Environ. Sci. Poll. Res. 2:5-15. Belfroid AC, DTHM Sijm and CAM Van Gestel (1996). Bioavailability and toxicokinetics of hydrophobic aromatic compounds in benthic and terrestrial invertebrates. Environ. Rev. 4:276-299. Belfroid AC and DTHM Sijm (1998). Influence of soil organic matter content on elimination rates of hydrophobic compounds in the earthworm: possible causes and consequences. Chemosphere 37:12211234. Black MC and JF McCarthy (1988). Dissolved organic macromolecules reduce the uptake of hydrophobic organic contaminants by the gills of rainbow trout (Salmo gairdneri). Environ. Toxicol. Chem. 7:593-600. Boese BL, H Lee and DT Specht (1990). Comparison of aqueous and solid-phase uptake for hexachlorobenzene in the tellinid clam Macoma nasuta (Conrad): a mass balance approach. Environ. Toxicol. Chem. 9:221-231. Branquart E, R Deleu, A Copin and C Gaspar (1995). Bio-accumulation et metabolisation comparees de l'isoproturon, du linuron et du lindane par Lumbricus terrestris L. Mededelingen Faculteit Landbouwkundige en Toegepaste Biologische Wetenschappen Universiteit Gent 609:511-519. Branson DR, GE Blau, HC Alexander and WB Neely (1975). Bioconcentration of 2,2',4,4'tetrachlorobiphenyl in rainbow trout as measured by an accelerated test. Trans. Am. Fish. Soc. 4:785-792. Briggs GG and KA Lord (1983). The distribution of aldicarb and its metabolites between Lumbricus terrestris, water and soil. Pestic. Sci. 14:412-416. Bruggeman WA, LBJM Martron, D Kooiman and O Hutzinger (1981). Accumulation and elimination kinetics of di-, tri- and tetra chlorobiphenyls by goldfish after dietary and aqueous exposure. Chemosphere 10:811-832. Carmichael LM, RF Christman and FK Pfaender (1997). Desorption and mineralization kinetics of phenanthrene and chrysene in contaminated soils. Environ. Sci. Technol. 31:126-132. Chiou CT, RL Malcolm, TI Brinton and DE Kile (1986). Water solubility enhancement of some organic pollutants and pesticides by dissolved humic and fulvic acids. Environ. Sci. Technol. 20:502-508. Connell DW and DW Hawker (1988). Use of poynomial expressions to describe the bioconcentration of hydrophobic chemicals by fish. Ecotox. Environ. Saf. 16:242-257. Connell DW and RD Markwell (1990). Bioaccumulation in the soil to earthworm system. Chemosphere 20:91-100. Conrad AU, SD Comber and K Simkiss (2002). Pyrene bioavailability; effect of sediment-chemical contact time on routes of uptake in an oligochaete worm. Chemosphere 49:447-454. Cornelissen G, PCM Van Noort and HAJ Govers (1997). Desorption kinetics of chlorobenzenes, polycyclic aromatic hydrocarbons, and polychlorinated biphenyls: sediment extraction with Tenax® and effects of contact time and solute hydrophobicity. Environ. Toxicol. Chem. 16:1351-1357. De Bruijn J, F Busser, W Seinen and J Hermens (1989). Determination of octanol/water partition coefficients for hydrophobic organic chemicals with the "slow-stirring" method. Environ. Toxicol. Chem. 8:499-512. De Bruijn J and J Hermens (1991). Uptake and elimination kinetics of organophosphorus pesticides in the guppy (Poecilia reticulata): correlations with the octanol/water partition coefficient. Environ. Toxicol. Chem. 10:791-804.

Theory of compartment models [24] De Maagd PGJ, A Van Kleunen, M Overdijk, J De Poorte, H De Graaf, A Opperhuizen and DTHM Sijm (1996). Biotransformation of polycyclic aromatic hydrocarbons in fathead minnow (Pimephales promelas): the use of a static exposure system and the biotransformation inhibitor piperonyl butoxide. In: Polycyclic aromatic hydrocarbons: fate and effects in the aquatic environment. PGJ De Maagd. PhD thesis, University of Utrecht, Utrecht, The Netherlands. pp. 69-93. [25] De Maagd PGJ, TL Sinnige, SM Schrap, A Opperhuizen and DTHM Sijm (1998). Sorption coefficients of polycyclic aromatic hydrocarbons for two lake sediments: influence of the bactericide sodium azide. Environ. Toxicol. Chem. 17:1899-1907. [26] De Wolf W, JHM De Bruijn, W Seinen and JLM Hermens (1992). Influence of biotransformation on the relationship between bioconcentration factors and octanol-water partition coefficients. Environ. Sci. Technol. 26:1197-1201. [27] Evans RD, GC Balch, HE Evans and PM Welbourn (2002). Simultaneous measurement of uptake and elimination of cadmium by caddisfly (Trichoptera: Hydropsychidae) larvae using stable isotope tracers. Environ. Toxicol. Chem. 21:1032-1039. [28] Flynn GL and SH Yalkowsky (1972). Correlations and prediction of mass transport across membranes I: Influence of alkyl chain length on flux-determining properties of barrier and diffusant. J. Pharm. Sci. 61:838-852. [29] Friedman HI and B Nylund (1980). Intestinal fat digestion, absorption, and transport. A review. Am. J. Clin. Nutr. 33:1108-1139. [30] Gobas FAPC, A Opperhuizen and O Hutzinger (1986). Bioconcentration of hydrophobic chemicals in fish: relationship with membrane permeation. Environ. Toxicol. Chem. 5:637-646. [31] Gobas FAPC, JR McCorquodale and GD Haffner (1993). Intestinal absorption and biomagnification of organochlorines. Environ. Toxicol. Chem. 12:567-576. [32] Gobas FAPC, JB Wilcockson, RW Russel and GD Haffner (1999). Mechanism of biomagnification in fish under laboratory and field conditions. Environ. Sci. Technol. 33:133-141. [33] Haitzer M, S Höss, W Traunspurger and C Steinberg (1998). Effects of dissolved organic matter (DOM) on the bioconcentration of organic chemicals in aquatic organisms - a review -. Chemosphere 37:1335-1362. [34] Hawker DW and DW Connell (1985). Relationships between partition coefficient, uptake rate constant, clearance rate constant and time to equilibrium for bioaccumulation. Chemosphere 14:1205-1219. [35] Hawker DW and DW Connell (1986). Bioconcentration of lipophilic compounds by some aquatic organisms. Ecotox. Environ. Saf. 11:184-197. [36] Hendriks AJ and A Heikens (2001). The power of size. 2. Rate constants and equilibrium ratios for accumulation of inorganic substances related to species weight. Environ. Toxicol. Chem. 20:1421-1437. [37] Hendriks AJ, A Van der Linde, G Cornelissen and DTHM Sijm (2001). The power of size 1. Rate constants and equilibrium ratios for accumulation of organic substances related to octanol-water partition ratio and species weight. Environ. Toxicol. Chem. 20:1399-1420. [38] Jager T, FA Antón Sánchez, B Muijs, EG Van der Velde and L Posthuma (2000). Toxicokinetics of polycyclic aromatic hydrocarbons in Eisenia andrei (Oligochaeta) using spiked soil. Environ. Toxicol. Chem. 19:953-961. (Chapter 4 of this thesis) [39] Kalsch W, R Nagel and K Urich (1991). Uptake, elimination, and bioconcentration of ten anilines in zebrafish (Brachydanio rerio). Chemosphere 22:351-363. [40] Könemann H and K Van Leeuwen (1980). Toxicokinetics in fish: accumulation and elimination of six chlorobenzenes by guppies. Chemosphere 9:3-19. [41] Kooijman SALM and RJF Van Haren (1990). Animal energy budgets affect the kinetics of xenobiotics. Chemosphere 21:681-693. [42] Kooijman SALM (2000). Dynamic energy and mass budgets in biological systems. Cambridge University Press, Cambridge, UK. [43] Kraaij RH, S Ciarelli, J Tolls, BJ Kater and A Belfroid (2001). Bioavailability of lab-contaminated and native polycyclic aromatic hydrocarbons to the amphipod Corophium volutator relates to chemical desorption. Environ. Toxicol. Chem. 20:1716-1724. [44] Kukkonen J and PF Landrum (1994). Toxicokinetics and toxicity of sediment-associated pyrene to Lumbriculus variegatus (Oligochaeta). Environ. Toxicol. Chem. 13:1457-1468. [45] Kukkonen J and PF Landrum (1995). Measuring assimilation efficiencies for sediment-bound PAH and PCB congeners by benthic organisms. Aquat. Toxicol. 32:75-92. [46] Kwan KHM and S Smith (1991). Some aspects of the kinetics of cadmium and thallium uptake by fronds of Lemna minor L. New Phytol. 117:91-102. [47] Landrum PF, SR Nihart, BJ Eadie and LR Herche (1987). Reduction in bioavailability of organic contaminants to the ampipod Pontoporeia hoyi by dissolved organic matter of sediment interstitial waters. Environ. Toxicol. Chem. 6:11-20. [48] Landrum PF (1989). Bioavailability and toxicokinetics of polycyclic aromatic hydrocarbons sorbed to sediments for the amphipod Pontoporeia hoyi. Environ. Sci. Technol. 23:588-595.

55

Chapter 2 [49] Landrum PF, H Lee and MJ Lydy (1992). Toxicokinetics in aquatic systems: model comparison and use in hazard assessment. Environ. Toxicol. Chem. 11:1709-1725. [50] Larsen B, F Pelusio, H Skejø and A Paya-Perez (1992). Bioavailability of polychlorinated biphenyl congeners in the soil to earthworm (L. rubellus) system. Intern. J. Environ. Anal. Chem. 46:149-162. [51] Legierse KCHM, DTHM Sijm, CJ Van Leeuwen, W Seinen and JLM Hermens (1998). Bioconcentration kinetics of chlorobenzenes and the organophosporus pesticide chlorthion in the pond snail Lymnea stagnalis - a comparison with the guppy Poecilia reticulata. Aquat. Toxicol. 41:301-323. [52] Leppänen MT and JVK Kukkonen (1998). Relative importance of ingested sediment and pore water as bioaccumulation routes for pyrene to oligochaete (Lumbriculus variegatus, Müller). Environ. Sci. Technol. 32:1503-1508. [53] Leslie HA, TL Ter Laak, FJM Busser, MHS Kraak and JLM Hermens (2002). Bioconcentration of organic chemicals: is a solid-phase microextraction fiber a good surrogate for biota? Environ. Sci. Technol. 36:5399-5404. [54] Liss PS and PG Slater (1974). Flux of gases across the air-sea interface. Nature 247:181-184. [55] Loonen H, JR Parsons and HAJ Govers (1991). Dietary accumulation of PCDDs and PCDFs in guppies. Chemosphere 23:1349-1357. [56] Lord KA, GG Briggs, MC Neale and R Manlove (1980). Uptake of pesticides from water and soil by earthworms. Pestic. Sci. 11:401-408. [57] Lu X, DD Reible, JW Fleeger and Y Chai (2003). Bioavailability of desorption-resistant phenanthrene to the oligochaete Ilyodrilus templetoni. Environ. Toxicol. Chem. 22:153-160. [58] Luoma SN (1983). Bioavailability of trace metals to aquatic organisms - a review. Sci. Total Environ. 28:122. [59] Ma WC, J Immerzeel and J Bodt (1995). Earthworm and food interactions on bioaccumulation and disappearance in soil of polycyclic aromatic hydrocarbons: Studies on phenanthrene and fluoranthene. Ecotox. Environ. Saf. 32:226-232. [60] Mackay D (1982). Correlation of bioconcentration factors. Environ. Sci. Technol. 16:274-278. [61] Mackay D (1991). Multimedia environmental models. Lewis Publishers, Chelsea, MI, USA. [62] Mackay D and A Fraser (2000). Bioaccumulation of persistent organic chemicals: mechanisms and models. Environ. Poll. 110:375-391. [63] Matscheko N, S Lundstedt, L Svensson, M Harju and M Tysklind (2002). Accumulation and elimination of 16 polycyclic aromatic compounds in the earthworm (Eisenia fetida). Environ. Toxicol. Chem. 21:1724-1729. [64] Mayer LM, Z Chen, RH Findlay, J Fang, S Sampson, RFL Self, PA Jumars, C Quetel and OFX Donard (1996). Bioavailability of sedimentary contaminants subject to deposit-feeder digestion. Environ. Sci. Technol. 30:2641-2645. [65] Mayer LM, DP Weston and MJ Bock (2001). Benzo[a]pyrene and zinc solubilization by digestive fluids of benthic invertebrates - a cross-phyletic study. Environ. Toxicol. Chem. 20:1890-1900. [66] McKim J, P Schmieder and G Veith (1985). Absorption dynamics of organic chemical transport across trout gills as related to octanol-water partition coefficient. Toxicol. Appl. Pharmacol. 77:1-10. [67] Moser GA and MS McLachlan (2002). Modeling digestive tract absorption and desorption of lipophilic organic contaminants in humans. Environ. Sci. Technol. 36:3318-3325. [68] Opperhuizen A, EW Van der Velde, FAPC Gobas, DAK Liem, JMD Van der Steen and O Hutzinger (1985). Relationship between bioconcentration in fish and steric factors of hydrophobic chemicals. Chemosphere 14:1871-1896. [69] Opperhuizen A and RCAM Stokkel (1988). Influence of contaminated particles on the bioaccumulation of hydrophobic organic micropollutants in fish. Environ. Poll. 51:165-177. [70] Park SS, JW Park, C Uchrin and MA Cheney (2002). A micelle inhibition model for the bioavailbility of polycyclic aromatic hydrocarbons in aquatic systems. Environ. Toxicol. Chem. 21:2737-2741. [71] Peijnenburg WJGM, R Baerselman, AC De Groot, T Jager, L Posthuma and RPM Van Veen (1999). Relating environmental availability to bioavailability: soil-type-dependent metal accumulation in the oligochaete Eisenia andrei. Ecotox. Environ. Saf. 44:294-310. [72] Peijnenburg WJGM, L Posthuma, PGPC Zweers, R Baerselman, AC De Groot, RPM Van Veen and T Jager (1999). Prediction of metal bioavailability in Dutch field soils for the oligochaete Enchytraeus crypticus. Ecotox. Environ. Saf. 43:170-186. [73] Penry DL and DP Weston (1998). Digestive determinants of benzo[a]pyrene and phenanthrene bioaccumulation by a deposit-feeding polychaete. Environ. Toxicol. Chem. 17:2254-2265. [74] Pignatello JJ and B Xing (1996). Mechanisms of slow sorption of organic chemicals to natural particles. Environ. Sci. Technol. 30:1-11. [75] Reinecke AJ and RG Nash (1984). Toxicity of 2,3,7,8-TCDD and short-term bioaccumulation by earthworms (Oligochaeta). Soil Biol. Biochem. 16:45-49.

56

Theory of compartment models [76] Riederer M (1995). Partitioning and transport of organic chemicals between the atmospheric environment and leaves. In: Plant contamination. Modelling and simulation of organic chemical processes. S Trapp, JC McFarlane (eds.). Lewis Publishers, Boca Raton, FL, USA. pp. 153-190. [77] Sample BE, GW Suter, JJ Beauchamp and RA Efroymson (1999). Literature-derived bioaccumulation models for earthworms: development and validation. Environ. Toxicol. Chem. 18:2110-2120. [78] Schuler LJ, M Wheeler, AJ Bailer and MJ Lydy (2003). Toxicokinetics of sediment-sorbed benzo[a]pyrene and hexachlorobiphenyl using the freshwater invertebrates Hyalella azteca, Chironomus tentans, and Lumbriculus variegatus. Environ. Toxicol. Chem. 22:439-449. [79] Schwarzenbach RP, PM Gschwend and DM Imboden (1993). Environmental organic chemistry. John Wiley & Sons, New York, NY, USA. [80] Seinen W, T Helder, H Vernij, A Penninks and P Leeuwangh (1981). Short term toxicity of tri-nbutylchloride in rainbow trout (Salmo gairdneri Richardson) yolk sac fry. Sci. Total Environ. 19:155-166. [81] Selck H, VE Forbes and TL Forbes (1998). Toxicity and toxicokinetics of cadmium in Capitella sp. I: relative importance of water and sediment as routes of cadmium uptake. Mar. Ecol. Progr. series 164:167178. [82] Sijm DTHM, W Seinen and A Opperhuizen (1992). Life-cycle biomagnification study in fish. Environ. Sci. Technol. 26:2162-2174. [83] Sijm DTHM, ME Verberne, P Part and A Opperhuizen (1994). Experimentally determined blood and water-flow limitations for uptake of hydrophobic compounds using perfused gills of rainbow trout (Oncorhynchus mykiss) - Allometric applications. Aquat. Toxicol. 30:325-341. [84] Sijm DTHM and A Van der Linde (1995). Size-dependent bioconcentration kinetics of hydrophobic organic chemicals in fish based on diffusive mass transfer and allometric relationships. Environ. Sci. Technol. 29:2769-2777. [85] Spacie A and JL Hamelink (1982). Alternative models for describing the bioconcentration of organics in fish. Environ. Toxicol. Chem. 1:309-320. [86] Stenersen J and N Øien (1980). Action of pesticides on earthworms. Part IV: uptake and elimination of oxamyl compared with carbofuran. Pestic. Sci. 11:396-400. [87] Van Brummelen TC and NM Van Straalen (1996). Uptake and elimination of benzo[a]pyrene in the terrestrial isopod Porcellio scaber. Arch. Environ. Contam. Toxicol. 31:277-285. [88] Van der Linde A, AJ Hendriks and DTHM Sijm (2001). Estimating biotransformation rate constants of organic chemicals from modeled and measured elimination rates. Chemosphere 44:423-435. [89] Van Gestel CAM and WC Ma (1988). Toxicity and bioaccumulation of chlorophenols in earthworms, in relation to bioavailability in soil. Ecotox. Environ. Saf. 15:289-297. [90] Van Ginneken L, L Bervoets and R Blust (2001). Bioavailability of Cd to the common carp, Cyprinus carpio, in the presence of humic acid. Aquat. Toxicol. 52:13-27. [91] Van Hoof PL, JVK Kukkonen and PF Landrum (2001). Impact of sediment manipulations on the bioaccumulation of polycyclic aromatic hydrocarbons from field-contaminated and laboratory-dosed sediments by an oligochaete. Environ. Toxicol. Chem. 20:1752-1761. [92] Van Leeuwen CJ, PS Griffioen, WHA Vergouw and JL Maas-Diepeveen (1985). Differences in susceptibility of early life stages of rainbow trout (Salmo gairdneri) to environmental pollutants. Aquat. Toxicol. 7:59-78. [93] Van Leeuwen CJ and JLM Hermens (1995). Risk assessment of chemicals: an introduction. Kluwer Academic Publishers, Dordrecht, The Netherlands. [94] Verbruggen EMJ, WHJ Vaes, TF Parkerton and JLM Hermens (2000). Polyacrylate-coated SPME fibers as a tool to simulate body residues and target concentrations of complex organic mixtures for estimation of baseline toxicity. Environ. Sci. Technol. 34:324-331. [95] Voparil IM and LM Mayer (2000). Dissolution of sedimentary polycyclic aromatic hydrocarbons into the lugworm's (Arenicola marina) digestive fluids. Environ. Sci. Technol. 34:1221-1228. [96] Wågman N, B Strandberg and M Tysklind (2001). Dietary uptake and elimination of selected polychlorinated biphenyl congeners and hexachlorobenzene in earthworms. Environ. Toxicol. Chem. 20:1778-1784. [97] Wang WX and RCH Dei (1999). Kinetics measurements of metal accumulation in two marine macroalgae. Mar. Biol. 135:11-23. [98] Wang WX and NS Fisher (1999). Assimilation efficiencies of chemical contaminants in aquatic invertebrates: a synthesis. Environ. Toxicol. Chem. 18:2034-2045. [99] Weber LP and RP Lanno (2001). Effect of bile salts, lipid, and humic acids on absorption of benzo[a]pyrene by isolated channel catfish (Ictalurus punctatus) intestine segments. Environ. Toxicol. Chem. 20:1117-1124. [100] Weston DP and LM Mayer (1998). Comparison of in vitro digestive fluid extraction and traditional in vivo approaches as measures of polycyclic aromatic hydrocarbon bioavailability from sediments. Environ. Toxicol. Chem. 17:830-840.

57

Chapter 2 [101] Weston DP, DL Penry and LK Gulmann (2000). The role of ingestion as route of contaminant bioaccumulation in a deposit-feeding polychaete. Arch. Environ. Contam. Toxicol. 38:446-454. [102] Widmark E and J Tandberg (1924). Über die Bedingungen für die Akkumulation indifferenter Narkotika. Theoretische Berechnungen. Biochemische Zeitschrift 147:358-369 (in German).

58

Estimating bioconcentration

3 Mechanistic Approach for Estimating Bioconcentration of Organic Chemicals in Earthworms (Oligochaeta)1

Tjalling Jager Appeared in 1998: Environmental Toxicology and Chemistry 17(10):2080-2090

ABSTRACT  Earthworms (Oligochaeta) represent an important food source for many vertebrates and as a result, predators may encounter toxic effects via the food chain (secondary poisoning) from consumption of contaminated worms. Therefore, including an assessment of secondary poisoning in risk assessment procedures is advisable. In this study, a mechanistic model is presented for estimating bioconcentration of organic chemicals in earthworms. It is assumed that bioconcentration can be described by a thermodynamic partitioning between soil solids, soil water, and the resident organism’s tissues. For most chemicals, the lipid phase is the dominant site for sorption in the earthworm, but for more hydrophilic compounds, the water phase may also play a role. Model predictions are compared to literature data that were derived from experiments with earthworms in water, laboratory experiments with various soils, and from field experiments. Without calibration, the model was able to accurately predict bioconcentration factors (BCFs) from experiments in water, indicating the applicability of this theoretical approach. However, BCFs in soil were consistently overestimated by the model (on average a factor of 5.6), which may be due to the absence of true equilibrium conditions in the soil-pore water-earthworm system. The collected experimental data reveal no net influence of uptake via soil ingestion, growth dilution, or sorption to dissolved organic carbon (DOC). Field data were more variable, but were generally consistent with the model. Nevertheless, before field data can be accurately predicted, the influences of chemical sorption, sorption kinetics, and earthworm behaviour must be quantified under field conditions.

1 Direct motivation for this paper was a report written on behalf of the Ministry of Housing, Spatial Planning and the Environment as part of the RIVM project 679102 (Jager, DT and T Hamers, 1997, RIVM report no.: 679102 013).

59

Chapter 3

INTRODUCTION Risk assessment of chemical substances is traditionally concerned primarily with direct effects of chemicals on organisms: the concentration in an environmental compartment is compared to a relevant effect or no-effect level for organisms living in, or in close contact with, that compartment. Chemicals can, however, be passed on in the food chain and predators may be largely exposed via their food. This accumulation in the food source is in itself no reason for concern when levels remain below toxic thresholds, but may eventually lead to adverse effects in the organisms preying on them (secondary poisoning). Earthworms (Oligochaeta) play a central role in terrestrial food chains because of their abundance, their relatively large size compared to other soil invertebrates, and the fact that they comprise a large part of the diet of many vertebrate species (e.g. moles, badgers and thrushes). For certain pesticides, incidents of wildlife poisoning could be directly attributed to feeding on earthworms from treated agricultural areas [16]. The importance of earthworms in secondary poisoning is reflected in the implementation of this specific pathway in the derivation of environmental quality criteria [40] and European risk-assessment guidances [18]. For these purposes, a short food chain is modelled to indicate a chemical's potential to cause secondary poisoning: soil→worm→predating bird or mammal. In these procedures, it is assumed that this food chain also protects more intricate food webs. Table 1. Regressions of earthworm BCFs to the octanol-water partition coefficient (Kow). The BCFs are expressed on solution basis (L/kg), either on wet or dry weight basis (wwt and dwt, respectively). The earthworm species is abbreviated as: Ea = Eisenia andrei, Lr = Lumbricus rubellus, Lt = L. terrestris. log BCF =

1.06 log Kow – 2.36 (wwt) 0.398 log Kow + 0.724 (dwt) 0.547 log Kow – 0.405 (dwt) 0.476 log Kow + 1.04 (dwt) 1 log Kow – 0.6 (dwt) a Several species combined.

Sp.

Compounds

Log Kow range

Steady state

Ea

Chlorobenzenes

4.2–5.7

yes

Lr

Chlorophenols

2.5–5.0

no

Ea

Chlorophenols

2.5–5.0

no

Lt

Pesticides

1–7.5

unk.

a

Mainly pesticides

1–6.5

unk./ no

Exp. design

Whole worms in water Worms in soil in laboratory Worms in soil in laboratory Macerated worms in water Data of [49] and [30], and field data

Ref.

[1] [49] [49] [30] [15]

The assessment of secondary poisoning requires so-called bioconcentration or bioaccumulation factors (BCFs or BAFs), defined as the steady-state ratio between the concentration of a pollutant in the organism and the concentration in its environment. Although the difference between bioconcentration and bioaccumulation is more or less semantic, the first term is generally used to describe uptake from a water phase whereas the second describes the net result of all routes of exposure. In contrast with aquatic organisms, measured data and estimation routines for BCFs or BAFs are scarce for soil organisms. For the soil compartment, a clear need for descriptive and transparent estimation routines exists. Estimation routines in the form of empirical regressions with the octanol-water partition coefficient (Kow) are useful for obtaining a first impression of the accumulating behaviour of a chemical. Several authors have reported log-linear regressions for earthworm BCFs on soil-solution basis (Table 1). Large discrepancies exist

60

Estimating bioconcentration

among the various regressions (both in slope and intercept), although these studies are not comparable because of radically different experimental designs. The slope and intercept depend heavily on the selected data (the training set), and therefore on the type of chemicals used and the experimental conditions. Extrapolating regressions beyond their domain can lead to serious errors and, clearly, care must be taken when applying these estimation routines in risk assessment procedures. Furthermore, regressions provide little insight into the mechanism of bioaccumulation and the role of organism and soil properties in this process. A more mechanistic approach is therefore preferred, as long as it is sufficiently descriptive for a broad range of chemicals and earthworm species. For organic chemicals, the main route SOIL of exposure for earthworms is uptake from the soil solution through the outer SOLIDS SOLUTION [6], and the pore-water skin K sw concentration should therefore be an important factor determining soil ingestion “bioavailability”. This is supported by BCF studies where bioaccumulation and uptake from food toxicity for earthworms were found to EARTHWORM metabolism be related to organic matter in the soil reproduction (the main sorption site for organic growth chemicals) [17,49]. Bioavailability is a rather vague term but is defined here Figure 1. Processes affecting the concentration of as the fraction of the bulk amount of xenobiotics in earthworms. Thick lines represent chemical in soil that can potentially be the equilibrium partitioning theory, thin lines taken up into the organism’s tissues represent processes that may influence the during its lifespan [6]. A mechanistic validity of this theory. approach should distinguish between sorption processes in soil and uptake by the earthworm (Fig. 1). When the solid phase, pore water and organism are in thermodynamic equilibrium, the concentration in the organism is determined by the concentration in the solid phase and the steady-state partition coefficients (Ksw and BCF in Fig. 1). This “equilibrium partitioning” (EP) concept is widely applied in soil and sediment risk-assessment procedures but has several shortcomings (extensively reviewed by Belfroid et al. [6]); some processes that may reduce the applicability of the EP approach are also shown in Figure 1. The first process in this three-phase equilibrium is the partitioning between soil solids and soil water. Sorption of neutral organic chemicals in soil is dominated by hydrophobic interactions with soil organic matter, and therefore, good correlations are observed between the partition coefficient normalised to organic carbon (Koc) and the octanol-water partition coefficient (Kow) [41]. The second process is the actual bioconcentration process from soil water to earthworm. For aquatic organisms, lipid tissue is the main dissolving medium for organic chemicals [14,32], and relationships between BCF and Kow provide satisfactory descriptions. In its simplest form, a mechanistic model for BCF depends on the lipid fraction of the organism and the Kow of the chemical [32]. The same approach was proposed by Connell & Markwell [15] for the soil solution-earthworm system but, unfortunately, their empirical regression is not consistent with the lipid fraction in worms (their data suggest a lipid phase of 25% of the dry weight whereas 4–6% is more realistic [27]). The purpose of this study is to evaluate the mechanistic approach for the accumulation of organic chemicals in earthworms. The role of other phases in the worm (water and protein) will be investigated, as well as the possible influence of sorption to 61

Chapter 3

dissolved organic carbon (DOC) in soil. Subsequently, this mechanistic approach will be compared to evaluated experimental data from the literature, and the applicability for risk assessment will be discussed.

METHODS THEORETICAL MODEL Bioconcentration in aquatic organisms can be regarded as the result of a chemical’s distribution between water and the phases inside the organism [11,32]. With this assumption, the organism is reduced to an inanimate container that is seeking a thermodynamic equilibrium with its medium. To reach equilibrium, the organism must be in close contact with the bioavailable phase in its environment to allow for sufficient diffusive exchange of chemicals. For organisms such as fish, this precondition is easily satisfied, as large amounts of water are pumped over the large surface of the gills. For earthworms, this assumption needs further consideration, although the earthworm’s physiology provides clues to its appropriateness. Earthworms likely lose 10–20% of their body weight in moisture each day due to their respiratory system, which requires the maintenance of a moist outer surface, and because of their nitrogen excretion mechanism, which requires water to dilute ammonia and urea to more hypotonic urine [29]. This water loss can only be replenished by ingestion or direct contact with free water. To estimate the equilibrium BCF, the chemical’s affinity to each of the phases inside the organism must be defined relative to the surrounding medium. With the lipid pool as the main sorbing medium for organic chemicals, a partition coefficient between the body of the earthworm and water in equilibrium (Kbw) can be defined as a function of Kow and the fraction lipids in the organism (Flip), as proposed by Connell & Markwell [15]:

K bw =

[worm] (mg/L) = Flip K ow [soil solution] (mg/L)

(Lwater/Lworm)

Eq. 1

For the affinity of the lipid phase, the octanol-water partition coefficient is used as a surrogate parameter. Even though octanol and natural lipids are structurally dissimilar, octanol was found to be an appropriate model for sorption of neutral compounds to biomembranes consisting of phospholipids [20]. It should be noted that Kow is not truly dimensionless but is defined as the concentration in the octanol phase (mol/Loctanol) divided by the concentration in the water phase (mol/Lwater), i.e. a volume ratio. This implies that the fraction of the lipid phase must be a volume fraction also. This fact is easily overlooked [15,42,46,47], which stresses the relevance of conscientiously specifying units of an equation. A BCF with the general unit of L/kg is derived by dividing the partition coefficient Kbw by the bulk density of the organism’s body (ρb in kgwwt/L):

BCF =

K bw

ρb

(Lwater/kgwwt )

Eq. 2

Because the bulk density is defined here on a wet-weight basis, the BCF is now also expressed on a worm wet-weight basis. This facilitates comparison of concentrations in worms with results of bird or mammalian toxicity tests [40]. In a situation of thermodynamic equilibrium, the water phase inside the organism can be expected to reach the same concentration as the external water (i.e. a partition coefficient of 1). When this phase is added (Fw), Equation 1 is extended to: 62

Estimating bioconcentration

K bw = Fw + Flip K ow

(Lwater/Lworm)

Eq. 3

Quantitatively, the internal water phase will be less important than the lipid phase for most hydrophobic chemicals. Equations of this form were successfully applied to describe bioconcentration in fish [42] and uptake of chemicals from air by plants (in that case also including an air phase) [39]. Similar equations are also used in environmental chemistry to describe chemical partitioning between the different phases in abiotic media (soil, sediment, water). For aquatic organisms, lipid is thought to be the dominant sorbing medium for hydrophobic organic chemicals, and other phases are usually ignored. Earthworms have a low lipid and high protein content compared to aquatic organisms (approximately 10 times more protein than lipid [27]), and the protein phase may therefore also play a role in bioconcentration. However, data on the affinity of proteins for organic chemicals are scarce. Neely et al. [35] observed a relationship between bioconcentration in trout muscle (a tissue rich in protein) and Kow. Nevertheless, compared to the combined lipid and water phases, the affinity of these proteins is insufficient to exert influence on bioconcentration. Based on these considerations, the protein phase is ignored as a significant binding site in the model. PARAMETERS OF THE GENERAL MODEL Earthworm properties seem to be quite variable among various studies or various species (e.g. [27,28]). Furthermore, the water and lipid content of the animals may also vary as a result of climatic conditions and nutritional status. In particular, the water content may vary because earthworms do not have efficient mechanisms for water conservation and can survive the loss of up to 70–80% of their water content [28,29]. Table 2. Selected standard properties for earthworms. Parameter Bulk density worm Bulk density lipids Dry/wet weight ratio

Values (range) 1 kgwwt/L 0.83 kg/L 0.16 (0.14–0.20) kgdwt/kgwwt

Sources assumed equal to water assumed equal to octanol [28,31,40,48]

Fraction water Fraction lipids

Wet-weight basis 0.84 (0.80–0.86) 0.01 (0.006–0.02)

[28,31,40,48] [1,22,27,31,48]

Volume basis 0.84 0.012

For the model definition, parameter values for a “standard worm” were selected from literature sources (Table 2). Typical ranges are included for water and lipid contents. Volume fractions are required in the theoretical model, as discussed previously, and weight fractions of a phase were recalculated to a volume fraction by using the densities of the phase (i stands for water or lipid) and the organism: Fi (volume) = Fi (weight)

ρb ρi

Eq. 4

DATA COLLECTION AND EVALUATION Experimental bioaccumulation data were collected from the literature by on-line search in recent literature and tracing through references. A distinction is made between different experimental designs: earthworms exposed in aqueous medium under laboratory conditions, exposure in soil under laboratory conditions (either spiked or fieldcontaminated soil), and exposure in soil under field conditions. These exposure situations all provide valuable information but are very different and difficult to compare. Going

63

Chapter 3

from water to laboratory soil to the field not only increases the relevance for risk assessment but also the number of variables, thereby diminishing the possibilities for validation of the process mechanisms. The data from these categories are therefore presented separately. The following criteria were applied to evaluate the literature studies: 1. Only data were used with exposure via soil or aqueous medium (thereby discarding food-only or filter paper exposure). In the selected field studies, however, uptake from contaminated food and/or soil ingestion may also have occurred. 2. Organic matter content of the soil must be specified. At very high or very low organic matter contents, sorption is no longer linearly related to organic matter and therefore a range in organic matter content of 1 to 30% is taken as acceptable. 3. Preferably, a steady-state situation between soil and earthworm must be achieved in the experiment. In several studies, it was not determined whether steady state was actually achieved. These studies were still included when the exposure duration was at least 10 d (this period seems sufficient for most chemicals to reach steady state [3,4]). In other studies, chemicals disappeared from the exposure medium, thereby excluding steady-state situations. Although less valuable for validating the mechanistic model, these studies still provide relevant information for its applicability for risk assessment purposes, as the same non-equilibrium conditions will also occur under field conditions. The exposure basis used to calculate BCF is also listed as additional information in the appendix. 4. Data for dissociating substances were included as long as the fraction in the neutral form (calculated from pKa of the chemical and pH of the soil) was at least 5%. This percentage was taken as it is around this value that the ionic species starts to influence uptake in biomembranes (calculated from [20]). Because the anionic species is more polar than the neutral form, it is likely that sorption and uptake are dominated by the latter. Sorption is therefore calculated from Kow of the neutral form and experimental BCFs are also related to the Kow of the neutral species. 5. In several studies, evidence of biotransformation was found. This will usually result in more polar metabolites that are more readily excreted. These studies were still included in the data set to test if metabolism will seriously affect the steady-state BCFs. The BCFs, as well as important characteristics from all selected studies, are given in the appendix. All BCF data were recalculated to wet weight of worms (using the standard worm defined in Table 2) and expressed on soil-solution basis. The soil solids-water distribution in soil (Ksw) is calculated from the partition coefficient normalised to organiccarbon (Koc) and the reported fraction organic carbon (Foc) in the study (a fixed ratio of 1.7 between organic matter and organic carbon is assumed [40]): K sw =

[soil dwt] = Foc K oc [soil solution]

(L/kgdwt)

Eq. 5

The Koc values were estimated using Quantitative Structure-Activity Relationships (QSARs) as advised for risk assessment of new and existing chemicals [18]. The QSAR for the group of “predominantly hydrophobic” chemicals was used for most chemicals in this study [41] (for several polychlorinated biphenyls -PCBs-, this QSAR was applied outside its log Kow domain of 1–7.5). For chlorophenols and pesticides containing nitrogen groups, QSARs were used as derived for the chemical groups “phenols and benzonitriles” and “agricultural chemicals”, respectively [41]. The Kow values were obtained from the MedChem database [34]. The advised measured value (lopP*) was preferred, otherwise a

64

Estimating bioconcentration

calculated value was used (ClogP version 2.10, as given by MedChem). All curve fits and statistical analyses were performed with the software package GraphPad Prism™ (version 2.00, San Diego, CA, USA). DISSOLVED ORGANIC CARBON Dissolved humic and fulvic acids can enhance the apparent water solubility of hydrophobic organic chemicals [13]. Chemicals associated with dissolved organic carbon (DOC) are generally assumed to be unavailable for diffusive uptake by organisms. This was experimentally confirmed in fish [9], and it is therefore reasonable to assume that DOC-bound chemicals are also not taken up by earthworms through their skin. The association with DOC can be described as a partition-like process and the truly dissolved fraction (Ffree) can be calculated from the DOC concentration (in kg/L) and the normalised partition coefficient with DOC (Kdoc in L/kg) [13]:

F free =

solubility 1 = apparent solubility 1 + [DOC] K doc

Eq. 6

The Kdoc for soil-derived humic acids was roughly a factor of two lower than Koc (for fulvic acids, this factor was 6–8) [13]. The difference between bulk and dissolved organic matter was attributed to the size of the DOC, its polarity, and molecular configuration. The DOC concentrations in several typical soils were 20–150 mg/L [26]. When only the truly dissolved fraction is bioavailable to the worm, the apparent BCF will be the true BCF multiplied with Ffree (Eq. 6).

RESULTS AND DISCUSSION 7

log BCF

GENERAL MODEL BEHAVIOUR Figure 2 shows the general behaviour of the model and the effect of adding the water phase in the earthworm as described by Equation 3. BCF is shown as log-transformed values on wetweight basis using Equation 2. For most of the Kow range, the lipid fraction dominates the BCF. The water phase affects the modelled BCF for hydrophilic compounds only (log Kow < 2), resulting in a minimum BCF equal to the fraction of water in the worm. Figure 2 also shows the predicted minimum and maximum effect of DOC on apparent BCFs (Eq. 6). When chemicals associated with DOC are not available for uptake by earthworms, slopes less than unity at log Kow greater than 5 or 6 will result. The predictions of the theoretical model (as defined by Eq. 2 and 3) are compared to experimental data from the literature

lipids

6

lipids + water

5

DOC min. influence DOC max. influence

4 3 2 1 0 -1 0

1

2

3

4

5

6

7

8

log Kow

Figure 2. Behaviour of the theoretical model (BCF in L/kgwwt). The potential influence of DOC on the apparent BCF, when DOC-associated chemicals are not available to the organism. The maximum effect is obtained assuming Kdoc = 0.5 Koc and [DOC] = 150 mg/L, the minimum effect assumes Kdoc = 0.125 Koc and [DOC] = 20 mg/L.

65

Chapter 3

(see appendix). The different exposure situations (water, soil in the laboratory, and field sampling) are treated separately.

log BCF

EXPOSURE IN WATER The theoretical model describes the 5 data for earthworms in water very well without calibration (Fig. 3). As shown 4 in Table 3, the fit is good (r2 = 0.90) and improves when only experiments are 3 included in which steady state was 2 attained (r2 = 0.97). This supports the assumption of earthworms as 1 inanimate containers of water and lipids in the model. For chemicals with aldicarb 0 isoproturon a log Kow > 2, the slope of the oxamyl experimental data is very close to unity -1 (Table 3), which means that Kow is an -1 0 1 2 3 4 5 6 7 appropriate descriptor for the affinity log Kow of these chemicals to worm lipids (at least up to a log Kow of 6). When Figure 3. Comparison of the theoretical model isoproturon is removed from the data with experimental BCF data from experiments in set, the correspondence with linearity water (BCF expressed in L/kgwwt). The dotted line of the model is almost perfect. This represents the behaviour of the lipid phase only. chemical can be considered an outlier as it was intensively degraded in the solution and in the worm tissue [10], so that steadystate conditions were not achieved. Table 3. Model fitting (Eq. 2 and 3) and log-linear regressions of the experimental data against Kow. The modifying factor M represents a constant factor by which the model BCF must be divided to provide the best fit to the data. Data sets Water exposure a only steady-state data b Laboratory soil data neutral compounds Chlorophenols PCBs in field

Water exposure c excluding isoproturon Laboratory soil data neutral compounds Chlorophenols PCBs in field a No curve fitting performed. b See appendix. c Only data for log K > 2 included. ow

M 1 1 5.6 7.1 3.5 38

Slope 1.09 1.00 0.87 0.94 0.74 0.93

Fitting the theoretical model n 95% conf. r2 — 0.90 11 — 0.97 8 3.9–7.2 0.83 69 4.7–9.6 0.86 45 1.5–5.5 0.54 24 32–44 0.81 55 Linear regressions (log Kow - log BCF) n 95% conf. Intercept r2 0.86–1.32 0.95 9 –2.49 0.86–1.14 0.98 8 –2.03 0.78–0.96 0.84 69 –2.00 0.82–1.05 0.86 45 –2.43 0.46–1.01 0.58 24 –1.42 0.81–1.05 0.82 55 –3.01

The data for aldicarb and oxamyl are above the predicted BCF for the lipid phase only (dotted line in Fig. 3), which supports the inclusion of the earthworm’s water phase in the 66

Estimating bioconcentration

model for hydrophilic compounds. The experimental BCF for oxamyl is nevertheless lower than the predictions including the water phase. This BCF is likely to be erroneously low because it was based on the initial water concentration (in a similar experiment with carbofuran in the same paper, the concentration in water decreased by 32% during the experiment), and degradation products of oxamyl were detected in the worm [44]. A lack of steady state is a more probable cause for deviation from the model than metabolism because metabolites are unlikely to have very different BCFs (oxamyl lies in the Kow range where hydrophobicity is no longer expected to affect BCF). Furthermore, aldicarb was also intensively metabolised [12] but fits the model quite well. However, for pesticides containing nitrogen groups (e.g. isoproturon and oxamyl) Kow is possibly not as good a predictor of bioconcentration as for more simple substances like chlorobenzenes.

log BCF

EXPOSURE TO SOIL IN THE LABORATORY The laboratory data for earthworms exposed via soil show that residues in earthworms are consistently lower than expected from the model (Fig. 4). On average, this difference is a factor of 5.6 (Table 3). Many factors can mediate the uptake of chemicals from soil or sediment matrices, as reviewed by Belfroid et al. [6]. These factors can be abiotic (e.g. the composition of the soil matrix, sorption kinetics) as well as biotic (e.g. feeding strategies, avoidance, burrowing activity), or a combination of both. Earthworms may deplete the chemical pool in the pore water in their immediate surroundings (especially for very hydrophobic chemicals, which 6 will have low concentrations Model in the water phase). Neutral compounds 5 Desorption from soil solids or Chlorophenols diffusion through soil can 4 subsequently become the ratelimiting processes [6]. 3 Evidence for this depletion is found in studies where 2 stirring of the soil increased concentrations in worms [30], 1 and where wet soil resulted in higher uptake rates compared 0 to soil with a lower moisture content [12]. Furthermore, in -1 1 2 3 4 5 6 7 8 several studies, the residues in earthworms reached a log Kow maximum in time, after which they declined [10,31,38] (even Figure 3. Comparison of the theoretical model with experimental BCF data from laboratory experiments in soil. when the bulk concentrations All BCFs expressed on soil-solution basis (L/kgwwt). in soil remained relatively constant).

The earthworm’s behaviour may also affect bioavailability. With regard to feeding habits, earthworms can be roughly divided into two groups: the “humus formers” or litter feeders, eating slightly decomposed plant materials (e.g. Lumbricus terrestris and L. rubellus), and “humus feeders” or soil feeders (geophageous species), which consume organic materials dispersed in the soil (e.g. Aporrectodea caliginosa) [37]. The popular test species Eisenia fetida and E. andrei can be considered litter feeders although they are not typical soil-dwelling species. They prefer accumulations of organic matter, like rotten vegetation and compost and manure heaps [43]. Within these groups, several sub-groups 67

Chapter 3

of species can be identified according to their vertical distribution. Litter feeders will only ingest large amounts of soil when burrowing through compact soil [29] or possibly when driven by hunger stress [31]. In two studies, strong indications were found that behaviour influenced BCFs. Beyer [8] found much lower BCFs for chlorobenzenes than other authors [3] (both in artificial soil). Furthermore, he observed a cyclic trend in the residues of hexachlorobenzene in earthworms, coinciding with transferring the earthworms to fresh soils. The experiments were performed with L. terrestris, a species that constructs permanent burrows, so the construction of new burrows in the fresh soil could have temporarily increased exposure to HCB. Ma et al. [31] found higher BCFs for L. rubellus when kept without a food source, probably because the worms were ingesting more soil when driven by hunger stress. This is supported by the experiments of Martin [33], who found that lack of food increased burrowing and soil ingestion. The present data set reveals no significant differences between different species or between natural soil and artificial medium. Even the data for field-contaminated soil are not particularly deviating from laboratory-spiked soils. This does not imply that these factors are unimportant, as the effect may be largely disguised by the combination of these factors and differences in experimental design. For example, when only studies with chlorobenzenes are examined, a clear difference between two species in artificial soil can be seen (E. andrei [3] has higher BCFs than L. terrestris [8], p < 0.0001) as well as a slight difference between artificial [3] and field-contaminated soil [4] for E. andrei (p < 0.1). In a direct comparison, a difference between two species was found for chlorophenols [49]: BCFs for L. rubellus were generally higher than those for E. andrei. This difference may be explained by their difference in size, differences in behaviour, or by the different incubation temperatures. However, determining the cause of the difference is impossible because steady state was not achieved (the chemicals disappeared from the soil), and the worms were only measured at the end of the experiment. The slope of the laboratory BCFs is very close to unity as expected from the model, and, when chlorophenols are excluded, not significantly different from 1 (Table 3). For chlorophenols [49], the low slope in Table 1 may originate from the fact that BCFs were expressed by the original author on a total soil-solution basis, thereby including the neutral as well as the anionic species. The anionic form is more polar and the apparent hydrophobicity of the chemical will therefore decrease with increasing pH. The Kow in the regression should therefore be corrected for the degree of dissociation, leading to steeper curves. It is also possible that interactions other than hydrophobic ones play a role in the bioconcentration process for these compounds. The BCFs for chlorophenols are quite variable and show little relation with Kow (Table 3). Sorption of chlorophenols also shows less dependence on Kow than for neutral compounds [41]. Furthermore, these BCFs are also on average closer to the model than those of neutral compounds (p < 0.05). However, drawing firm conclusions from these data is not possible because steady-state conditions were not achieved (see appendix). Slopes slightly less than unity are commonly found for aquatic bioconcentration data [36] and plant tissues [46], and are usually explained by structural differences between octanol and natural lipids, or hindered transport over membranes. Interestingly, very low regression slopes (0.4–0.55) were reported for earthworms by several other authors (Table 1). For the pesticide QSAR [30] the low slope cannot be entirely explained, although the experimental design is in this case disputable. These data were excluded from this study because macerated worms were used, and only the concentration in water was measured. Steady state was unlikely to have been achieved, and for most chemicals, there were indications of substantial chemical loss from the containers (by volatilisation, metabolism, and/or degradation). For one chemical (aldicarb), BCFs were a factor of 10 lower in a comparable study where whole worms 68

Estimating bioconcentration

were used [12]. No loss of linearity of the Kow-BCF relationship could be discerned for very hydrophobic compounds, as is commonly observed for aquatic organisms [36]. This implies that the collected experimental data do not support a net effect of soil ingestion or sorption to DOC on BCF. Increased BCFs due to soil ingestion at log Kow > 5 were predicted from an uptake model by Belfroid et al. [5]. However, it should be noted that the expected effect was at maximum only a factor of two. This kind of influence cannot easily be discerned on a log scale, as applied in this study. Worms are able to take up organic chemicals through feeding [2] but, nevertheless, this route does not seem to lead to increased body burdens. Apparently, the equilibrium-partitioning hypothesis still holds. It is furthermore unlikely that DOC plays a role in the bioconcentration process, as a profound loss of linearity is expected when DOC-bound chemicals are not taken up (Fig. 2). This could mean that DOC-bound chemicals are remobilised during gut passage of soil when organic materials are digested, or that the sorption estimate (Eq. 5) already reflects the freely-dissolved concentration only. FIELD-COLLECTED EARTHWORMS In the field, the soil environment is much more heterogeneous than in the laboratory and further variability can be expected due to the earthworm’s ecology. Earthworms in a field situation can express more of their natural behaviour than in a laboratory jar, which may affect exposure to chemicals (e.g. due to feeding habits, burrowing, diapause, surface activity) [19]. Furthermore, additional variation can be expected as chemical sorption may increase in time (ageing of the contamination). Four different field studies were selected (described in the appendix) and are shown in Figure 5A–D. The BCFs found for PCBs are much lower than expected from the model (Fig. 5A); the difference is on average a factor of 38 (Table 3), which is also lower than observed for these chemicals in a fieldcontaminated soil, tested in the laboratory [4]. It is therefore unlikely that ageing is solely responsible for this deviation. Species differences may have contributed to a different exposure (L. rubellus sampled from the polluted site, E. andrei in the laboratory study), although both species are litter feeders. Another difference between the field and the laboratory study is that, in the latter, the soil is homogenised before earthworms are introduced, thereby enhancing bioavailability. Furthermore, laboratory conditions diminish the possibilities for the earthworms to avoid the contamination (actively or unintentionally) [45]. The soil concentration in the field study was reported as an average of the top 20 cm, but the worms may confine their activity to only a small part of this range (depending on the soil moisture content), and thereby may modify their exposure. Despite the deviations from the model, the slope of the data was very close to 1 (Table 3).

Figure 5B shows data for some chlorobenzenes, dioxins and furans in earthworms from a refuse dump. In contrast with the PCBs in Figure 5A, the data are quite close to the model and some are even somewhat higher than predicted. Although the contamination was at least 15 years old, no ageing effects were visible and bioavailability was quite high. The soil samples were taken from the top 10 cm and the earthworms sampled were all shallow-living species, including the soil feeder Aporrectodea rosea. These factors may have led to a closer correspondence with the model predictions than in the study with PCBs (Fig. 5A), were the litter feeding L. rubellus and the top 20 cm of soil were sampled. Figures 5C and 5D show data for surface-applied pesticides. The short-term average BCFs in Figure 5C are quite close to the model (data points indicate the geometric average over 0.5–4 months after application). The large variability in time indicated by the error bars is, however, striking. Generally, the BCFs decreased in the two years following application. Surface application will inevitably lead to higher concentrations in the upper soil layers, thereby increasing the exposure of surface-feeding or surface-active species in the initial 69

Chapter 3

period after treatment. In time, this effect will disappear as the chemical distributes more evenly through the soil layers and, furthermore, ageing may result in lower apparent BCFs. The variability is especially large for heptachlor and its metabolite heptachlor epoxide. This transformation process was faster in the worms than in soil, leading to high initial concentrations of heptachlor epoxide in worms compared to soil (Fig. 5C). The pesticide BCFs in Figure 5D resulted from surface application, followed up to 11 years after treatment. These BCFs are also quite comparable to the model. The BCFs of heptachlor epoxide and dieldrin declined only a factor of 1.5 and 4 over this period, respectively. The BCFs for total DDT residues remained remarkably stable in time and no ageing effect could be discerned. 6

A

PCBs

5

log BCF

6

Model

B

Refuse dump

5

others

4

4

3

3

2

2

1

1

0

0 0

1

6

2

3

4

5

6

7

8

9

0

1

6

Model

C

Pesticides

5

log BCF

Model

3

4

Model

5

6

7

8

9

6

7

8

9

D

Pesticides

5

4

2

4

3

3

heptachlor epoxide

2

2

heptachlor 1

1

0

0 0

1

2

3

4

5

log Kow

6

7

8

9

0

1

2

3

4

5

log Kow

Figure 4. BCFs (L/kgwwt) for field-collected earthworms: A) Sampling from polluted floodplains. B) Sampling from a refuse dump. C) Agricultural soil with surface-applied pesticides. Points represent geometric averages over 0.5 to 4 months post application; the range represents minimum and maximum over the entire two-year period. D) Agricultural soil with surface applied pesticides. Points represent samples taken at 0, 5.5 and 11 years after treatment (BCFs decreasing in time).

In summary, field data are reasonably consistent with the model, although they seem to be more variable than the laboratory data. This can be expected as the exact exposure concentration that the earthworms encounter in their micro-habitat within the heterogeneous soil environment is not known. The interpretation of the field BCFs therefore strongly depends on the species sampled, and the method and depth of soil sampling. Nevertheless, the slope of the field data remains close to 1, as predicted by the model, and shows no net result of uptake through food or sorption to DOC.

70

Estimating bioconcentration

CONCLUSIONS The collected experimental data are generally consistent with the theory that bioconcentration in earthworms is governed by thermodynamic partitioning of the chemical between surrounding water and the water and lipid phases inside the organism. Data from worms exposed to hydrophobic compounds in aqueous media are especially well predicted by this model. Generally, soil exposure leads to lower BCFs than expected; probably caused by non-equilibrium conditions because of slow desorption from soil solids, slow diffusion in soil, and/or earthworm behaviour. Differences in experimental design and lack of direct comparisons preclude conclusions on differences between species and soil types. Uptake from ingested soil, dilution by growth, and sorption to DOC do not seem to exert a net influence on the body residues. It seems possible to treat dissociating substances like neutral compounds by ignoring the ionic form in the calculation of sorption and bioconcentration (at least when the degree of dissociation is below 95%). Nevertheless, indications exist that the mechanism of bioconcentration differs from neutral compounds. The theoretical model also seems to work for field situations, although field BCFs can only be reproduced when bioavailability is adequately quantified in terms of sorption kinetics and earthworm behaviour. For risk assessment purposes, Equations 2 and 3 can be used to predict BCFs for earthworms in the log Kow range 0–8. This estimate should be regarded as a maximum BCF that is not always reached in soil situations. The theoretical model seems sufficiently protective to cover most field situations, but pesticide spraying requires special care. The heterogeneous vertical distribution of the chemical in soil or the specific contamination of food sources may result in high exposure for specific species. As an example, L. terrestris, a litter feeder that constructs semi-permanent burrows, is more susceptible to chemicals present in litter and granular pesticide formulations, but less susceptible to chemicals incorporated in the top soil layer, compared to shallow-living soil feeders like A. caliginosa [17,22,45]. The ecology of individual species can thus be a dominant factor influencing body residues in the field. Uptake and sorption to soil solids are inversely related to hydrophobicity of the compound. On a total soil basis, bioaccumulation factors will therefore show little dependence on Kow. Instead, concentration factors will be low and depend mainly on organism properties and organic matter content of the soil. Low concentration factors, however, do not imply that no danger exists to vertebrates feeding on earthworms, because the risk will also depend on the absolute concentrations in soil and the sensitivity of the predator. Furthermore, despite the relative ease of modelling uptake by earthworms from pore water, this approach may not apply to other soil organisms that live in less direct contact with the soil solution (e.g. [48]).

ACKNOWLEDGEMENTS I would like to thank Jari Haimi (University of Jyväskylä, Finland) for providing me with the original data of his experiments. Furthermore, for their critical comments on earlier versions of this manuscript, I would like to thank: Angélique Belfroid, Kees van Gestel, Wim Ma, Willie Peijnenburg, Lennart Weltje, Leo Posthuma, and Herman Eijsackers.

71

Chapter 3

REFERENCES [1]

[2]

[3] [4]

[5]

[6] [7] [8] [9]

[10]

[11]

[12] [13] [14] [15] [16]

[17] [18]

[19]

[20] [21]

[22] [23]

72

Belfroid A, A Van Wezel, M Sikkenk, K Van Gestel, W Seinen and J Hermens (1993). The toxicokinetic behavior of chlorobenzenes in earthworms (Eisenia andrei): experiments in water. Ecotox. Environ. Saf. 25:154-165. Belfroid A, J Meiling, D Sijm, J Hermens, W Seinen and K Van Gestel (1994). Uptake of hydrophobic halogenated aromatic hydrocarbons from food by earthworms (Eisenia andrei). Arch. Environ. Contam. Toxicol. 27:260-265. Belfroid A, M Sikkenk, W Seinen, K Van Gestel and J Hermens (1994). The toxicokinetic behavior of chlorobenzenes in earthworm (Eisenia andrei) experiments in soil. Environ. Toxicol. Chem. 13:93-99. Belfroid A, M Van den Berg, W Seinen, J Hermens and K Van Gestel (1995). Uptake, bioavailability and elimination of hydrophobic compounds in earthworms (Eisenia andrei) in field-contaminated soil. Environ. Toxicol. Chem. 14:605-612. Belfroid AC, W Seinen, KCAM Van Gestel, JLM Hermens and KJ Van Leeuwen (1995). Modelling the accumulation of hydrophobic organic chemicals in earthworms - Application of the equilibrium partitioning theory. Environ. Sci. Poll. Res. 2:5-15. Belfroid AC, DTHM Sijm and CAM Van Gestel (1996). Bioavailability and toxicokinetics of hydrophobic aromatic compounds in benthic and terrestrial invertebrates. Environ. Rev. 4:276-299. Beyer WN and CD Gish (1980). Persistence in earthworms and potential hazards to birds of soil applied DDT, dieldrin and heptachlor. J. Appl. Ecol. 17:295-307. Beyer WN (1996). Accumulation of chlorinated benzenes in earthworms. Bull. Environ. Contam. Toxicol. 57:729-736. Black MC and JF McCarthy (1988). Dissolved organic macromolecules reduce the uptake of hydrophobic organic contaminants by the gills of rainbow trout (Salmo gairdneri). Environ. Toxicol. Chem. 7:593-600. Branquart E, R Deleu, A Copin and C Gaspar (1995). Bio-accumulation et metabolisation comparees de l'isoproturon, du linuron et du lindane par Lumbricus terrestris L. Mededelingen Faculteit Landbouwkundige en Toegepaste Biologische Wetenschappen Universiteit Gent 609:511-519. Briggs GG (1981). Theoretical and experimental relationships between soil adsorption, octanol-water partition coefficients, water solubilities, bioconcentration factors, and the parachor. J. Agric. Food Chem. 29:1050-1059. Briggs GG and KA Lord (1983). The distribution of aldicarb and its metabolites between Lumbricus terrestris, water and soil. Pestic. Sci. 14:412-416. Chiou CT, RL Malcolm, TI Brinton and DE Kile (1986). Water solubility enhancement of some organic pollutants and pesticides by dissolved humic and fulvic acids. Environ. Sci. Technol. 20:502-508. Clayton JR, SP Pavlou and NF Breitner (1977). Polychlorinated Biphenyls in coastal marine zooplankton: Bioaccumulation by equilibrium partitioning. Environ. Sci. Technol. 11:676-682. Connell DW and RD Markwell (1990). Bioaccumulation in the soil to earthworm system. Chemosphere 20:91-100. Cooke AS, PW Greig-Smith and SA Jones (1992). Consequences for vertebrate wildlife of toxic residues in earthworm prey. In: Ecotoxicology of earthworms. PW Greig-Smith, H Becker, PJ Edwards, F Heimbach (eds.). Atheneum Press, Newcastle upon Tyne, UK. pp. 139-155. Davis BNK (1971). Laboratory studies on the uptake of dieldrin and DDT by earthworms. Soil Biol. Biochem. 3:221-233. EC (1996). Technical Guidance Documents in support of Directive 93/67/EEC on risk assessment of new notified substances and Regulation (EC) No. 1488/94 on risk assessment of existing substances (Parts I, II, III and IV). EC catalogue numbers CR-48-96-001, 002, 003, 004-EN-C. Office for Official Publications of the European Community, 2 rue Mercier, L-2965 Luxembourg, Luxembourg. Edwards CA (1992). Testing the effects of chemicals on earthworms: the advantages and limitations of field tests. In: Ecotoxicology of earthworms. PW Greig-Smith, H Becker, PJ Edwards, F Heimbach (eds.). Atheneum Press, Newcastle upon Tyne, UK. pp. 75-84. Escher BI and RP Schwarzenbach (1996). Partitioning of substituted phenols in liposome-water, biomembrane-water, and octanol-water systems. Environ. Sci. Technol. 30:260-270. Gish CD and DL Hughes (1982). Residues of DDT, dieldrin, and heptachlor in earthworms during two years following application. Special Scientific Report - Wildlife No. 241. U.S. Fish and Wildlife Service, Washington, DC, USA. Haimi J, J Salminen, V Huhta, K J. and H Palm (1992). Bioaccumulation of organochlorine compounds in earthworms. Soil Biol. Biochem. 24:1699-1703. Haque A (1984). Behaviour of some fungicides in soil and their uptake by earthworms and plants. In: Proceedings, Institut National de la Recherche Agronomique (INRA) 31: Comportement et Effets Secondaires des Pesticides dans le Sol., Versailles, France (June, 4-8, 1984). pp. 279-289.

Estimating bioconcentration [24] Heida H, K Olie and E Prins (1986). Selective accumulation of chlorobenzenes, polychlorinated dibenzofurans and 2,3,7,8-TCDD in wildlife in the Volgermeerpolder, Amsterdam, Holland. Chemosphere 15:1995-2000. [25] Hendriks AJ, WC Ma, JJ Brouns, EM De Ruiter-Dijkman and R Gast (1995). Modelling and monitoring organochlorine and heavy metal accumulation in soils, earthworms, and shrews in Rhine-delta floodplains. Arch. Environ. Contam. Toxicol. 29:115-127. [26] Janssen RPT, WJGM Peijnenburg, L Posthuma and MAGT Van den Hoop (1997). Equilibrium partitioning of heavy metals in Dutch field soils. I. Relationships between metal partition coefficients and soil characteristics. Environ. Toxicol. Chem. 16:2470-2478. [27] Laird JM and M Kroger (1981). Earthworms. CRC Crit. Rev. in Environ. Control May 1981:189-218. [28] Laverack MS (1963). The physiology of earthworms. Pergamon Press, Oxford, UK. [29] Lee KE (1985). Earthworms. Their ecology and relationships with soils and land use. Academic Press, Sydney, Australia. [30] Lord KA, GG Briggs, MC Neale and R Manlove (1980). Uptake of pesticides from water and soil by earthworms. Pestic. Sci. 11:401-408. [31] Ma WC, J Immerzeel and J Bodt (1995). Earthworm and food interactions on bioaccumulation and disappearance in soil of polycyclic aromatic hydrocarbons: Studies on phenanthrene and fluoranthene. Ecotox. Environ. Saf. 32:226-232. [32] Mackay D (1982). Correlation of bioconcentration factors. Environ. Sci. Technol. 16:274-278. [33] Martin NA (1982). The interaction between organic matter in soil and the burrowing activity of three species of earthworms (Oligochaeta: Lumbricidae). Pedobiologia 24:185-190. [34] MedChem Project. MedChem Database, Ver. 3.55. Pomona College, Claremont, CA, USA. [35] Neely WB, DR Branson and GE Blau (1974). Partition coefficient to measure bioconcentration potential of organic chemicals in fish. Environ. Sci. Technol. 8:1113-1115. [36] Nendza M (1991). QSARs of bioconcentration: validity assessment of log Pow/log BCF correlations. In: Bioaccumulation in aquatic systems. Contributions to the assessment. R Nagel, R Loskill (eds.). VCH Verlagsgesellschaft mbH, Weinheim, Germany. pp. 43-66. [37] Perel TS (1977). Differences in lumbricid organization connected with ecological properties. Ecol. Bull. 25:56-63. [38] Reinecke AJ and RG Nash (1984). Toxicity of 2,3,7,8-TCDD and short-term bioaccumulation by earthworms (Oligochaeta). Soil Biol. Biochem. 16:45-49. [39] Riederer M (1990). Estimating partitioning and transport of organic chemicals in the foliage/atmosphere system: discussion of a fugacity-based model. Environ. Sci. Technol. 24:829-837. [40] Romijn CAFM, R Luttik and JH Canton (1994). Presentation of a general algorithm to include effect assessment on secondary poisoning in the derivation of environmental quality criteria. 2. Terrestrial food chains. Ecotox. Environ. Saf. 27:107-127. [41] Sabljić A, H Güsten, H Verhaar and J Hermens (1995). QSAR modelling of soil sorption. Improvements and systematics of log Koc vs. log Kow correlations. Chemosphere 31:4489-4514. [42] Sijm DTHM and A Van der Linde (1995). Size-dependent bioconcentration kinetics of hydrophobic organic chemicals in fish based on diffusive mass transfer and allometric relationships. Environ. Sci. Technol. 29:2769-2777. [43] Sims RW and BM Gerard (1985). Earthworms. Synopsis of the British fauna No. 31. E.J. Brill Publishing Company, Leiden, The Netherlands. [44] Stenersen J and N Øien (1980). Action of pesticides on earthworms. Part IV: uptake and elimination of oxamyl compared with carbofuran. Pestic. Sci. 11:396-400. [45] Tomlin AD (1992). Behaviour as a source of earthworm susceptibility to ecotoxicants. In: Ecotoxicology of earthworms. PW Greig-Smith, H Becker, PJ Edwards, F Heimbach (eds.). Athenaeum Press, Newcastle upon Tyne, UK. pp. 116-125. [46] Trapp S and M Matthies (1995). Generic one-compartment model for uptake of organic chemicals by foliar vegetation. Environ. Sci. Technol. 29:2333-2338. [47] Trapp S and M Matthies (1996). Generic one compartment model for uptake of organic chemicals by foliar vegetation (correction to vol. 29, pg 2333, 1995). Environ. Sci. Technol. 30:360. [48] Van Brummelen TC, RA Verweij, SA Wedzinga and CAM Van Gestel (1996). Polycyclic aromatic hydrocarbons in earthworms and isopods from contaminated forest soils. Chemosphere 32:315-341. [49] Van Gestel CAM and WC Ma (1988). Toxicity and bioaccumulation of chlorophenols in earthworms, in relation to bioavailability in soil. Ecotox. Environ. Saf. 15:289-297.

73

Chapter 3

APPENDIX: COLLECTED BIOCONCENTRATION DATA Bioconcentration factors for earthworms on (soil-)solution basis (L/kgwwt). For each study the tested species is shown and whether steady state is achieved (column St-st). Species names abbreviated in column Sp.: Ea = Eisenia andrei, Ef = E. fetida, Lt = Lumbricus terrestris, Lr = L. rubellus, Ac = Aporrectodea caliginosa. Also shown is the exposure measure on which BCF is based. For chlorophenols, Kow and BCFs are given for the neutral species only. Ref. Chemical in water [1]

[10]

[12] [44]

1,3,5-trichlorobenzene 1,2,3-trichlorobenzene 1,2,3,4-tetrachlorobenzene pentachlorobenzene hexachlorobenzene isoproturon linuron lindane aldicarb carbofuran oxamyl

Ref. Chemical in soil [3]

[4]

[8]

[10] [17]

74

1,2,3,4-tetrachlorobenzene pentachlorobenzene hexachlorobenzene 1,2,3,4-tetrachlorobenzene pentachlorobenzene hexachlorobenzene PCB 101 PCB 118 PCB 138 PCB 153 PCB 156 PCB 167 PCB 180 1,3,5-trichlorobenzene 1,2,4-trichlorobenzene 1,2,3-trichlorobenzene 1,2,3,5-tetrachlorobenzene 1,2,4,5-tetrachlorobenzene 1,2,3,4-tetrachlorobenzene pentachlorobenzene hexachlorobenzene linuron lindane dieldrin

log Kow 4.19 4.14 4.64 5.18 5.73 2.87 3.20 3.72 1.13 2.32 –0.47 log Kow 4.64 5.18 5.73 4.64 5.18 5.73 6.50 7.12 7.25 7.16 7.57 7.50 8.04 4.19 4.02 4.14 4.66 4.60 4.64 5.18 5.73 3.20 3.72 5.20

log BCF 1.86 2.12 2.75 3.24 3.61 0.05 1.24 1.85 –0.16 0.22 –1.06

Sp.

St-st

Exposure basis

Ea Ea Ea Ea Ea Lt Lt Lt Lt Ef Ef

Yes Yes Yes Yes Yes No Yes Yes Yes No Unk.

Constant Constant Constant Constant Constant Solution conc. at 24 h Solution conc. at 24 h Solution conc. at 24 h Solution conc. at 6 h Solution conc. at 8 h Initial solution conc.

log BCF 2.64 3.25 3.84 1.50 1.71 3.52 3.86 4.51 4.49 4.51 4.83 4.69 5.10 1.21 1.08 1.26 1.50 1.50 1.52 2.30 2.94 0.27 2.12 2.43 2.36 2.44 2.74 2.67

Sp.

St-st

Exposure basis

Ea Ea Ea Ea Ea Ea Ea Ea Ea Ea Ea Ea Ea Lt Lt Lt Lt Lt Lt Lt Lt Lt Lt Ac Ac Ac Ac Ac

Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes No No Unk. Unk. Unk. Unk. Unk.

Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Constant Average soil conc. Average soil conc. Final soil conc. Final soil conc. Final soil conc. Final soil conc. Final soil conc.

Estimating bioconcentration Ref. Chemical in soil

log Kow

[17]

DDT

6.91

[23]

imazalil triadimenol

3.82 3.08

[38]

2,3,7,8-TCDD

6.42

[31]

phenanthrene

4.46

fluoranthene

5.16

3-chlorophenol

2.50

3,4-chlorophenol

3.33

2,4,5-chlorophenol

3.72

2,3,4,5-chlorophenol

4.21

pentachlorophenol

5.12

2,3,4,6-tetrachlorophenol

4.45

Pentachlorophenol

5.12

[49]

[22]

Ref. Chemical in field [25] PCBs and pesticides

[21]

Pesticides

[24]

Chlorobenzenes, furans Pesticides

[7]

log BCF 3.41 3.61 3.58 3.66 3.71 0.57 1.05 1.25 4.76 3.81 3.72 3.43 1.60 1.23 0.94 2.49 1.88 0.09 0.41 1.26 1.27 0.78 0.64 0.63 0.88 0.92 0.56 1.14 1.88 0.72 0.84 1.63 1.60 2.26 2.29 2.14 2.66 2.36 2.82 2.79 2.67

Sp.

St-st

Exposure basis

Ac Ac Ac Ac Ac Ac Ac Lt Ac Ac Ac Ac Lr Lr Lr Lr Lr Ea Ea Lr Lr Ea Ea Lr Lr Ea Ea Lr Lr Ea Ea Lr Lr Ea Ea Lr Lr Ac Lr Ac Ac

Unk. Unk. Unk. Unk. Unk. Unk. Unk. Unk. No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No No

Initial soil conc. Initial soil conc. Initial soil conc. Initial soil conc. Initial soil conc. Initial soil conc. Initial soil conc. Initial soil conc. Maximum BCF Maximum BCF Maximum BCF Maximum BCF Maximum BCF Maximum BCF Maximum BCF Maximum BCF Maximum BCF Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc. Average soil conc.

Sp. Lr

Description of study Polluted floodplains in the Rhine delta (the Netherlands) at two locations (5 and 9% OM). DDT residues summed. Several Treated study plots followed for 2 years (3.2–4.6% OM). DDT residues summed. dioxins, Several Earthworms sampled at contaminated refuse dump (21–24% OM). Several Study site in a hayfield followed for 11 years (3–5% OM). DDT residues summed.

75

Section B

Case Studies

“The wealth of experience of animals among horse and dog trainers, veterinarians and pet owners is generally dismissed as anecdotal. This happens so often that I looked up the origin of this word to find out what it means. It comes from the Greek roots an + ekdotos, meaning ‘not published’. An anecdote is an unpublished story. Some fields of research, for example medicine, rely heavily on anecdotes, but when they are published they literally cease to be anecdotes; they are promoted to the rank of case histories.” R. Sheldrake (1999) Dogs that know when their owners are coming home. And other unexplained powers of animals.

PAHs in artificial soil

4 Toxicokinetics of Polycyclic Aromatic Hydrocarbons in Eisenia andrei (Oligochaeta) using Spiked Soil1

Tjalling Jager, Paco Antón Sánchez, Barry Muijs, Els van der Velde, Leo Posthuma Appeared in 2000: Environmental Toxicology and Chemistry 19(4):953-961 (and erratum in Environmental Toxicology and Chemistry 19(6):1702)

ABSTRACT  The accumulation of four polycyclic aromatic hydrocarbons (PAHs; phenanthrene, pyrene, fluoranthene and benzo[a]pyrene) was tested in the earthworm Eisenia andrei in a spiked artificial soil medium. A typical peak in the body residues was observed for all PAHs around day 7 which could not be explained from changes in the total soil concentration. It is argued that the most likely cause of this peak is a decrease in the concentration in pore water, the main bioavailable phase for earthworms. This decrease is caused by biodegradation while the low rate of mass transfer from the solid state precludes replenishment. To describe the data, bioavailability was assumed to decline exponentially in time, but the shape of the accumulation curves suggests a more abrupt change. Estimates of the uptake rate (ku) are similar for all PAHs when expressed on soil-solution basis (approximately 2000 L/kg/d); the elimination rate (ke) shows a decrease with Kow as expected, but the values tend to be slightly lower than literature data. The dynamic bioconcentration factors (ku/ke) agree well with an equilibrium partitioning between soil water and the phases inside the organism.

The experiments described in this chapter were performed by Paco Antón Sánchez, during his stay at the RIVM as post-doctoral fellow, supported by the Spanish ministry of Education and Science. Tragically, in 1996, Paco was involved in a fatal motorcycle accident. Several years later, I finished the data analysis on his accumulation studies and published the results.

1

79

Chapter 4

INTRODUCTION Polycyclic aromatic hydrocarbons (PAHs) are widespread contaminants in soil. They are emitted as a result of combustion processes (anthropogenic as well as natural) and processing of fossil fuels, and are transported over long distances in association with soot particles. Furthermore, PAHs are deposited on soils, where they prove to be very persistent (especially under anaerobic conditions), and thus constitute a potential threat to soil organisms. In ecotoxicity studies, earthworms are often selected as representatives of the soil fauna as they live in close contact with the soil (and influence its structure), and because they represent a major part of the diet of many vertebrate species. Accumulation of PAHs in earthworms is therefore not only of concern for the health of the earthworms themselves but also for the terrestrial food chain. Bioconcentration or bioaccumulation factors are required in the risk assessment of chemicals for higher organisms that feed on earthworms [27]. However, few bioconcentration data are available for earthworms, compared with aquatic organisms. In the absence of data, quantitative structure-activity relationships (QSARs) may be used to provide an initial estimate. A recent evaluation of the available literature data resulted in a mechanistic approach [11], confirming that the main route of uptake is from, or related to, pore water, as proposed earlier [7,32] . Bioconcentration in the earthworm may be seen as a partitioning process between the chemical in the pore water and the internal phases in the organism (lipid and water). The exact contribution of the different routes of uptake (water and food) for oligochaetes remains a point of discussion. For compounds of intermediate hydrophobicity (log Kow < 5), the consensus is that pore water dominates exposure, but for more hydrophobic chemicals assimilation from ingested solids may play a role [7]. For sediment oligochaetes, ingestion of very hydrophobic compounds is indeed responsible for higher steady-state body residues [15,16], although the quantitative impact is relatively small (a factor of 1.4 and 2.5 in the respective studies), and may thus be easily masked by other factors. Earthworms are also able to take up chemicals from ingested materials [3], and the kinetics in the presence of soil differ from those in water-only exposure [8]. Nevertheless, there is at present no evidence that the equilibriumpartitioning theory consistently underestimates uptake of organic chemicals in earthworms by ignoring other uptake routes [11]. Although bioconcentration factors for earthworms are sometimes available, the kinetics of uptake and elimination are not routinely investigated. Interesting about the toxico-kinetics of PAHs is that, after initial accumulation, there seems to be a decrease in the concentration in the organism, even when the concentration in soil remains constant [17]. The same peak pattern was observed for 2,3,7,8-tetrachlorodibenzo-p-dioxin in earthworms [26], for PAHs in sediment amphipods [12] and sediment polychaetes [24], but not for chlorobenzenes in earthworms [4]. The reasons for this behaviour are unclear but several authors proposed different mechanisms, including induced biotransformation or active excretion [24,26], movement to a biologically unavailable pool [12], and increased soil sorption [17]. Another option may be a change in earthworm activity over time, as behaviour is known to affect exposure to chemicals [31]. The purpose of this work was to obtain data on the behaviour of different PAHs and to evaluate the existing QSAR for bioconcentration in earthworms. The working hypothesis for the experiments was that earthworms accumulate contaminants from the soil solution only. As substrate, artificial soil [22] was spiked with a series of PAH concentrations. The experiments described in this paper were set up as a combined accumulation and toxicity

80

PAHs in artificial soil

study; this paper focuses on the kinetics in relation to the behaviour in soil, the toxic effects will be described in a separate paper (Muijs et al., unpublished data).

MATERIALS AND METHODS ANIMALS AND CHEMICALS Specimens (Eisenia andrei) were obtained from mass cultures kept in climatised conditions (temperature 20 ± 2°C ) for many generations on a substrate of sphagnum peat, potting soil, and horse dung. For each experiment, the animals were obtained from a single culture container in which cocoons were produced 18–21 or 21–24 wk before the start of the experiment. All animals were adults with a well-developed clitellum. The OECD test protocol proposes the use of E. andrei or E. fetida for testing. Although these species are not typical soil dwellers (they are usually confined to accumulations of organic matter; especially manure), their ease of culturing and the amount of research with these species make E. andrei the preferred species for these experiments. The following chemicals were used in the experiments: phenanthrene (PHE, Riedel de Haën, Seelze, Germany, purity 98%), fluoranthene (FLU, Fluka, Buchs, Switzerland, purity 98%), pyrene (PYR, Janssen Chimica, Geel, Belgium, purity > 98%), and benzo[a]pyrene (BaP, Janssen Chimica, Geel, Belgium, purity > 98%). Some relevant physico-chemical properties are summarised in Table 1. Except when stated otherwise, chemicals were kept in the dark to avoid photolysis. For each compound six concentrations were tested: a solvent control and five concentrations in a geometric series (dose range in Table 1). Table 1. Physico-chemical properties of the selected PAHs and the tested dose range.

Phenanthrene (PHE) Pyrene (PYR) Fluoranthene (FLU) Benzo[a]pyrene (BaP)

Mol wt. (g/mol) 178.2 202.3 202.3 252.3

Log Kow [20] 4.46 4.88 5.16 5.97

Solubility (mg/L) [19] 1.29 0.135 0.26 0.0038

Dose range (mg/kgdwt) 50–800 83.5–1800 80–1800 48.7–1043

EXPOSURE MEDIUM All experiments were performed in an artificial medium [22], often referred to as OECD artificial soil. The medium is made of (expressed in percentage dry weight) 10% peat (Sphagnum, dried, ground and sieved, particle size < 1 mm), 20% kaoline clay, and 70% fine quartz sand (50% particles 0.05–0.2 mm). Before mixing these constituents, the sand fraction was coated with the appropriate amount of PAH dissolved in acetone, or with pure acetone (control treatments). Acetone was evaporated in a gentle air stream in a fume hood. Sand treated in this way was mixed with the appropriate amounts of clay and peat. Water was added until the water content was 35% (v/w). The pH (KCl) of the mixture was set to 5.5 ± 0.5 (optimal for cocoon production of E. andrei [34]) by adding calcium carbonate (Merck, Amsterdam, The Netherlands, purity > 98.5%). Three to seven days elapsed between mixing of sand with other components and the start of the exposure phase (for BaP treatments, this was 25 d for logistic reasons). To avoid personal exposure, all experimental handling involving PAHs was executed in a cabinet designed for carcinogenicity research. EXPERIMENTAL SORPTION ISOTHERMS The sorption characteristics of the selected PAHs in the medium were determined experimentally and expressed with Freundlich isotherms. Separate concentration series in 81

Chapter 4

wet artificial soil were made for the four PAHs, the concentration ranges overlapped with the exposure concentrations of the accumulation experiments (five concentrations, ~30– 1200 mg/kgdwt). The pH (KCl) of the medium was 5.3. After 1 wk of equilibration (20°C), soil was shaken overnight (200 rpm) with 0.01 M CaCl2-solution (in a 1:4 ratio of gwwt soil/ml salt solution). Centrifugation (3500 rpm) was used to separate the solid and liquid phase, and the supernatant was collected and analysed for PAHs (see below). The solid phase was not analysed as it was assumed that the amount of PAH in the liquid phase is negligible compared to the solid phase. The concentration of the chemical in soil water was related to the concentration sorbed to the soil solid phase as: C s = K F C w1 / n

Eq. 1

where Cs is the concentration in soil (mg/kgdwt), Cw the concentration in the liquid phase (mg/L), and KF and n are the Freundlich sorption parameters. The sorption isotherms were calculated from log-transformed liquid and solid-phase concentrations with linear regression. EXPERIMENTAL DESIGN The design of the assays was modified from a combination of proposed standard assays [1,33]. Each experiment consisted of three consecutive phases, including an acclimation phase (outside the cabinet, 1 wk, non-polluted medium), an exposure phase (inside the cabinet, 3 wk), and a depuration phase (inside the cabinet, 1 wk, clean medium). To start the acclimation phase, animals were randomly assigned to groups of ten adults: 12 groups were used for studying the changes of soil and body concentrations, 24 groups were used for toxicity assessment and subsequently for elimination kinetics. The test containers contained ~750 gwwt medium, and were placed inside a climate room (24 h light and 20 ± 1°C). Food was added once a week as ground cow dung on the soil surface (7.75 g, 55% water v/w). After 1 wk of acclimation to the medium, animals were recaptured. The experiments with different PAHs were executed consecutively due to space limitations in the carcinogenicity cabinet. To start exposure, recaptured animals were transferred to the cabinet, and put into plastic containers (1 L) with the (contaminated) medium. In the cabinet, a day/night cycle of 12/12 h was used and the temperature was ~20°C (minimum and maximum were monitored). The containers had a transparent lid, with small holes for gas exchange. Each container was filled with ~770 gwwt medium and food was added weekly. During the exposure phase, water evaporation was balanced weekly by adding the appropriate volume of water, as assessed from weight loss. To quantify the change of soil and body concentrations, a sample of soil and four worms were taken after various time intervals (1, 3, 7, 14, 21/22 d). The amount of soil removed was chosen to maintain a constant ratio of earthworms per soil weight. Sampled soil was immediately frozen; recaptured animals voided their gut overnight on moist filter paper, and were kept frozen until analysis. After three weeks of exposure, the animals recaptured from the 24 containers were transferred to clean soil to determine the elimination kinetics. These animals were captured after various periods in the clean medium (3 or 4 points in 7 d, four worms per sample). Similarly, they voided their guts overnight, were frozen, and were analysed for PAHs. PAH ANALYSES For PAH analysis, 5 gwwt soil was extracted in an extraction tube with 25 ml of acetone with 6-methylchrysene as internal standard during 20 min on a shaking machine (for the samples with higher levels of PAH, 75 ml of acetone was added and no internal standard

82

PAHs in artificial soil

was used, because samples were diluted for analysis). The tubes were centrifuged (2500 rpm, 3 min) and 5 ml of extract (or a dilution) was diluted with 7.5 ml of water for HPLCanalysis. Pooled samples of four animals (0.2–1 gwwt) were homogenised in a dismembrator (2000 rpm, 3 min). A weighed amount was transferred to an extraction tube and 5 ml of acetone with 6-methylchrysene as internal standard was added. For the samples with higher levels of PAH, 50 ml of acetone was added and no internal standard was used, because samples were diluted for analysis. Samples were extracted using a shaking machine during 20 min, and 1.0 ml of extract (or a dilution) was filtered and subsequently diluted with 1.5 ml of water. For quality control, blank medium and unexposed animals were analysed using the same procedures. Recoveries were determined by adding PAHs at different levels to control animal or soil samples. Extracts of 200 µL (animal or soil) were injected into an on-line SPE-HPLC system. The extract was on-line cleaned on a C8-cartridge with methanol/water and subsequently analysed on a Bakerbond PAH 16 HPLC-column using gradient elution from acetonitrile/water (60/40) to acetonitrile (100%). The analytes were detected with programmable fluorescence detection at specific wavelengths. All samples were diluted to concentrations within the linear range of the detector. The PAHs were quantified on external standards of PAHs, corrected for internal standard (for losses during the extraction procedure), procedural blanks (phenanthrene; fluoranthene) and recoveries. Recoveries for animals at lower exposure levels were 70–80% and for higher levels 100–110%; for soil, recoveries varied between 95 and 110%. Limits of determination were 1 ng/gwwt for worm tissue and soil. A separate sample of soil was used to assess the water content (all soil concentrations were recalculated to dry weights). SOIL AND EARTHWORM DATA ANALYSES All linear and non-linear curve fitting was performed with the software package GraphPad Prism™ (version 2.00, San Diego, CA, USA). Measured soil concentrations (Cs) were fitted with a log-linear regression assuming exponential decay (with or without a lag-phase), resulting in a first-order degradation rate (kd in d-1).

C s (t ) = C s 0 e − k d t

(mg/kgdwt)

Eq. 2

The elimination rate of the earthworms in clean medium (ke in d-1) was estimated first, fitting a first-order exponential model through the body residues (Cb) as measured during the depuration phase: C b (t ) = C b 0 e − k e t

(mg/kgwwt)

Eq. 3

Subsequently, a modified one-compartment model was used to describe the observed peak in the body residues (see introduction). Assuming that the peak reflects a change in the available PAH pool, the exposure concentration is modified with an exponential decay, as proposed by Landrum [12]: dC b (t ) = k uC s 0 e − kdist − k eC b (t ) dt

Eq. 4

The same model formulation was applied by Widianarko & Van Straalen [37], although in their case, the factor kdis did not reflect a change in bioavailability but a change in total concentration of a pesticide in soil. The non-linear regression resulted in estimates of the rate of uptake (ku in kg/kg/d) and an empirical “disappearance” rate (kdis in d-1). The

83

Chapter 4

elimination rate constant (ke) was fixed to the value found in the independent elimination experiments to avoid identification problems; when ke was also allowed as free parameter in Equation 5, the optimisation routine was often not able to distinguish between ke and kdis. When the initial concentration in the earthworm is negligible (as is the case for PAHs) this leads to the general solution: C b (t ) =

(

k uC s 0 − kdist e − e − ket k e − k dis

)

(mg/kgwwt)

Eq. 5

Using calculated pore-water concentrations (Eq. 1), the uptake rate constant ku was also expressed on soil-solution basis (L/kg/d). Biota-soil accumulation factors (BSAF) were calculated by dividing the lipid-normalised concentrations in earthworms (Flip taken as 1% [11]) by the initial concentration in the medium, normalised to organic carbon (Foc in OECD medium is 5.9%):

BSAF(t ) =

C b (t ) Foc C s 0 Flip

(kgoc/kglip)

Eq. 6

It should be noted that in this definition, BSAF does not represent a steady state but is a function of time. A bioconcentration factor (BCF) is calculated on a soil solution basis (assuming this is the dominant route of uptake). Because of the dynamics of the body residues, BCFs were calculated as the quotient of uptake and elimination rate constants (ku on solution basis and ke), thus reflecting a (theoretical) situation of equilibrium:

BCF = k u / k e

(L/kgwwt)

Eq. 7

RESULTS AND DISCUSSION Since the experiments were set up as a combined accumulation and toxicity experiment, a series of five soil concentrations and a control were tested. For assessing toxico-kinetics, the results from the different exposure concentrations were combined as much as possible to obtain a larger data set to fit the accumulation models. Because accumulation kinetics are preferably studied at non-toxic levels, only the doses were used in which no overt toxic damage on growth, mortality or reproduction was observed (B. Muijs, pers. comm.). Moreover, there was also direct evidence of a different accumulation pattern in the earthworms at the higher doses (discussed in the section on accumulation). Therefore, for PYR, FLU and BaP, the lowest three doses are used; for PHE, only the lowest two doses are used. FREUNDLICH SORPTION ISOTHERMS The sorption isotherms of Equation 1 provided a good description of the data (r2 > 0.90); the sorption parameters are summarised in Table 2. The non-linearity constant (n) is clearly less than 1 for PYR and BaP, which probably reflects solubility limitations (in view of the high test concentrations, see Table 1). Solids-water partition coefficients were calculated in the selected dose range and normalised to the amount of organic carbon in the medium (Koc in Table 2). No correction for any degradation in the 7 d equilibration were made because of the lag phase observed and because degradation is lower in soil without earthworms (see section on degradation). The resulting Koc values are similar to 84

PAHs in artificial soil

values found in lake sediments [10], which were determined in the presence of a bactericide (and therefore not influenced by biodegradation). The relevance of this sorption experiment for bioavailability to earthworms is, however, limited. The procedure of shaking overnight with an excess amount of CaCl2 solution may not represent what the organisms experience in their immediate surrounding. Table 2. Results of model regressions on sorption, degradation and earthworm kinetics (s.e. in parenthesis, where appropriate). All data are expressed on the basis of soil dry weight and worm wet weight. PHE

PYR

FLU

BaP

r2 log KF n log Koc (L/kg)

0.94 2.91 1.05 4.20

Sorption parameters 0.96 0.92 4.00 3.57 0.826 0.992 4.93 4.79

0.94 5.40 0.680 5.54

r2 kd (1/d) lag phase (d)

0.95 a 0.17 (0.011) 5 (0.84)

r2 ke (1/d)

0.87 0.16 (0.020)

Elimination phase 0.96 0.97 0.76 (0.048) 0.66 (0.034)

r2 ku (kg/kg/d) ku (L/kg/d × 103) kdis (1/d) ke (1/d)

0.89 1.4 (0.19) 1.3 (0.18) 0.71 (0.12) 2.1 (0.33)

Accumulation phase 0.78 0.54 0.51 (0.047) 0.81 (0.15) 2.5 (0.23) 2.9 (0.55) 0.033 (0.0099) 0.020 (0.018) 0.98 (0.34)

Soil degradation 0.81 n.s. 0.0076 (0.00093) 0b

0.54 0.0042 (0.0015) 5.2 (4.2)

0.96 c 0.27 (0.017)

0.72 0.14 (0.019) 2.9 (0.39) 0.088 (0.021)

Bioconcentration and bioaccumulation max. BSAF (kgoc/kglip) 7.3 (0.33) 3.9 (0.58) 8.2 (2.3) 2.4 (0.21) 3.9 3.5 3.6 4.0 log BCF (L/kg) 2.8 3.4 log BCF (ke from accum.) a All doses pooled. b Linear regression used. c Doses 2 to 4 used because the lowest dose elimination pattern would not fit.

DEGRADATION IN THE SOIL SYSTEM The total loss of chemicals from the system was attributed to biodegradation due to the observation of a distinct lag phase for PHE and BaP of approximately 5 d. This probably reflects adaptation of micro-organisms, although the actual microbial activity was not investigated. Estimated degradation rates, lag time and the goodness of fit of Equation 2 are shown in Table 2. As was also found by Ma et al. [17], the degradation rates of PHE and PYR were slightly lower in soils without worms (p < 0.05; data from pilot experiments, not shown), indicating that the presence of earthworms (or the manure added as feed) enhances degradation. For FLU there was no significant decrease in soil concentrations observed. Although the degradation rates of PYR and BaP differ significantly from zero, the rate of PHE degradation is more than an order of magnitude

85

Chapter 4

higher, which is consistent with its relatively low sorption to solids and its linear structure (making it less recalcitrant) [9]. 6

400

177 1824

300

BSAF

body residues

500

836

200

81 379 2

379 100

4

836

177

1824

81

0 0

5

10

15

time (days)

20

25

0 0

5

10

15

20

25

time (days)

Figure 1. Individual accumulation curves for pyrene, absolute values (left, mg/kgwwt), and normalised as BSAFs (right, kgoc/kglip). The boxed numbers in the graphs are the measured initial soil concentrations (mg/kgdwt).

ACCUMULATION IN EARTHWORMS In the accumulation phase, the concentration in the earthworm showed a peak around 3 to 7 d for all PAHs. This effect is too large to result from the decrease in total soil concentration. Figure 1 shows the results for PYR for the individual doses as illustration. The two highest doses reveal a different pattern of accumulation with slower uptake rates than the lower three doses. This may relate to toxic effects, although it is likely that poor water solubility leads to lower availability at the higher doses than expected (the estimated pore water concentration reaches the solubility at a soil concentration of 900 mg/kg). Based on these findings, the lowest three doses were selected and pooled as BSAFs (Eq. 6). The model of Equation 5 fitted to the pooled data provided a reasonable description (Table 2), although a typical misfit is observed (Fig. 2). The BSAFs at t = 7 are underestimated whereas the values at t = 14 are overestimated. For PHE, this misfit is at least partly caused by the problems of toxicity affecting ke, as discussed in the next section. This misfit implies that the pattern of availability decrease differs from a smooth exponential decay, as postulated in Equation 5, but is more abrupt.

The uptake rate constants expressed on total soil concentrations show a clear decrease with Kow (Fig. 3). This relationship disappears when the data are expressed on soilsolution basis (using the experimental Freundlich isotherms). Differences in sorption can fully explain the differences in uptake rate between the PAHs, thus supporting the hypothesis that the main route of uptake is via the water phase in soil (or a related route). These findings are consistent with data on fish, where ku in this Kow range is independent of hydrophobicity and is around 1000 L/kg [23,28]. For chlorobenzenes in artificial soil, similar results are found (ku = 1500–3300 L/kg as calculated from [4]). Lower values of ku are found in water-only exposure: ~200 L/kg for chlorobenzenes [2], and ~400 L/kg for PAHs in Lumbricus rubellus [18]. Although BCFs in water are essentially similar to soilexposed animals [4,11,18], the kinetics are very different [8]. This is also to be expected, because earthworms in soil have more contact with their environment as solid material is passed through their gut. Furthermore, earthworms have poorly developed mechanisms for water conservation [14] and therefore require higher water uptake in a soil environment.

86

PAHs in artificial soil

8

5

PHE

6

PYR

4 3

4 2

BSAF

2

1

0

0 0

5

10

15

20

25

0

5

10

15

20

25

3

12 10

BaP

FLU

8

2

6 1

4 2 0 0

5

10

15

20

25

0

0

5

10

15

20

25

time (days) Figure 2. Pooled data of the BSAFs (kgoc/kglip) for the different PAHs. The line represents the fit of Equation 5, bars represent s.e.

10

uptake rate (L/kg/d)

uptake rate (kg/kg/d)

The empirical rates of disappearance (kdis) found in the regressions are larger than the decrease in the total soil concentration (kd) (Table 2). The decreasing soil concentration itself therefore cannot fully explain the peak shape in the accumulation pattern. The estimated disappearance rate shows no clear relation with Kow (Table 2). This differs from findings of Landrum for sediment amphipods [12], who found a clear decrease in kdis with increasing hydrophobicity. The absolute values of kdis in this study must, however, be interpreted with care as ke and kdis have a similar effect on the curve shape. To avoid identification problems in fitting Equation 5, ke was fixed to the value estimated in the depuration phase. This identification problem does not seriously influence the estimate of the uptake rate, as ku is predominantly governed by the first (“linear”) phase of the accumulation curve.

1

0.1 4

5

6

log K ow

7

104

103

102 4

5

6

7

log K ow

Figure 3. Uptake rate constants shown against log Kow (symbols with s.e.). Figure left shows ku referenced to the total soil concentration, right on estimated pore-water basis.

87

Chapter 4

ELIMINATION FROM EARTHWORMS The estimated elimination rates (ke) seem to have little relation with Kow (Fig. 4), which is contrary to the decrease with increasing Kow as found in studies with other chemicals in earthworms [6] and for aquatic organisms [23,28]. Our data for PHE in particular are much lower than data for chlorobenzenes with comparable Kow [4]. The rate constant observed in the depuration phase of this compound corresponds poorly with the observed behaviour in the accumulation phase. When the initial part of the accumulation phase is considered to follow normal one-compartment kinetics (which is acceptable, given the lag-phase observed in the degradation), an additional estimate for ke is obtained. Especially for PHE, this results in a higher value then found in the depuration phase (Table 2), which compares better to the literature data and reveals a relationship between ke and Kow (Fig. 4). The data sets for FLU and BaP did not allow an accurate estimation of this additional ke. The reason for this discrepancy between the accumulation and elimination phase is not known, but it is conceivable that the tested concentrations of PHE were sufficiently high to exert a narcotic effect. Narcosis will reduce the animal’s activity and feeding rate, thereby decreasing contact with soil and possibly the elimination of compounds from the body. Table 2 contains the pooled results of the lowest doses only, but when the individual elimination curves are examined, a dose-related decrease in ke is observed for PYR and FLU, thus providing additional support for this hypothesis.

elimination rate (1/d)

The literature data underlying the 10 regression lines shown in Figure 4 are taken from experiments with earthworms in water [2], artificial soil [3-5] and field-contaminated soil [6] 1 with other chemicals (mainly chlorobenzenes and PCBs). The elimination in the presence of soil or sediment is in principle bi-phasic, with the onset of 0.1 the second (slower) phase delayed with increasing Kow [8]. In our experiments, only one phase was observed; the slower second phase 0.01 may have been missed as it would 4 5 6 7 8 have occurred outside the time log K window of the experiment. The slow ow phase, when detected, corresponds well to the elimination rates found in Figure 4. Elimination rates (ke) from the present water-only exposure. The elimination study (symbols with s.e.). Open symbols represent the results from the depuration experiment, closed rates from this study show a symbols the estimate from the accumulation reasonable correspondence to the phase. Literature data are summarised as literature data, although they are regression lines. Broken line represents the slow consistently lower than the regression phase and water exposure (log ke = 5.4 – 1.2 log line for the rapid phase. In another Kow), solid line the rapid phase (for log Kow < 7: log study with earthworms and PAHs [18], ke = 4.0 – 0.79 log Kow). elimination rates found for water-only exposure showed little correlation with Kow (around 0.2 to 0.4 d-1). Especially the high value found for BaP (comparable to our data) does not correspond with the expectation that elimination is larger in the presence of soil particles. It should, however, be noted that the elimination rates were not determined in independent depuration experiments (the

88

PAHs in artificial soil

values were estimated from the changes in the concentrations in water and worms) and that a different species of earthworm was used (L. rubellus).

log BCF

BIOCONCENTRATION FACTORS Because the concentration in the 5 organisms did not reach a steady state in the experiment, a static accumulation factor (concentration in worm divided 4 by concentration in soil) cannot be obtained. The maximum BSAFs (Eq. 6) range from 2.4–8.2 kgoc/kglip (Table 2) 3 for the different PAHs. Therefore, bioconcentration factors (BCF) are best 2 expressed dynamically as ratio of the uptake and elimination rate constants (Eq. 7). Because we hypothesised that 1 pore water is the dominant exposure route, BCF needs to be expressed on 0 solution basis to allow comparison with 1 2 3 4 5 6 7 data from other soils [11]. The results are given in Table 2 and can be seen as log Kow an equilibrium concentration factor that will be achieved when the exposure Figure 5. Estimated bioconcentration factors concentration remains constant for a (ku/ke in L/kgwwt) for the different PAHs on a soil sufficiently long period of time. Figure solution basis compared to a theoretical estimate based on equilibrium partitioning (solid line). ! 5 compares these dynamic BCFs to a = dynamic BCF; " = dynamic BCF using ke from mechanistic estimate (based on log Kow the accumulation phase. and the lipid and water phases in the organism [11]). Our BCFs are slightly higher than expected from the model which may be a feature of PAHs in particular but is more likely caused by experimental errors (e.g. a slightly overestimated sorption). The BCFs based on the ke from the accumulation phase agree much better with the expected values, thereby indicating that the ke from the depuration experiment (especially for PHE) is erroneously low. EXPLANATION OF THE PEAK Several processes may explain the observed peak in the BSAF of PAHs: biotransformation, ageing or a change in bioavailability other than ageing. Another possibility is a decline in lipid content of the earthworms during the experiment. As lipid contents were not determined, this could have influenced the accumulation pattern. However, the animals were fed during the experiments and, especially for PHE, the change in body residue is too large to result from a lipid decrease alone. Furthermore, the lipid content will be most effected at higher exposure levels but Figure 2 shows the reverse trend: i.e. a less distinct peak at higher doses. It is not possible that the peak represents depletion of the pore water by the organism. Such a mechanism may explain lower than expected uptake and BCFs of very hydrophobic chemicals when desorption or diffusion become rate limiting for the uptake [7], but it cannot explain the peak in the body residues. As long as the organism follows linear one-compartment behaviour, the concentration in pore water will decrease as the concentration in the earthworm increases. The organism-water system will approach a steady state smoothly, albeit at a lower body residue than expected on the basis of equilibrium. A peak can only follow from biotransformation or a decreasing concentration in the water phase. 89

Chapter 4

A slowly induced biotransformation is not a likely explanation. Biotransformation of xenobiotics in earthworms has received little study so far, compared to mammals, fish and insects. Although cytochrome P450 is present in earthworms, the ability to oxidise planar molecules like PAHs seems limited [29]. This was recently confirmed in an analysis of pyrene metabolites in animals, left over from our experiments, that had been exposed for 21 d to pyrene (G. Stroomberg, unpublished data; method adapted from [30]). Chromatograms of the worms clearly showed the presence of pyrene, 1-hydroxypyrene and three unknown peaks, not present in the unexposed animals. The absorption spectra of the three unknown peaks were comparable to that of pyrene, indicating that they are pyrene conjugates. Samples hydrolysed with glucuronidase/arylsulfatase showed the removal of two conjugate-peaks and an increased amount of 1-hydroxypyrene. The total amount of metabolised pyrene comprised 0.1–1% of the amount of parent pyrene in the earthworms. The un-hydrolysed conjugate-peak was not quantified, but was estimated to be within 30% of the total amount of conjugated 1-hydroxypyrene. This indicates that E. andrei has some capacity for metabolising PAHs, such as pyrene, but that it is unlikely to represent a significant removal process. The hypothesis of biotransformation is also not compatible with the observed elimination rates. The most distinct peak is observed for PHE but for this compound, ke in the elimination phase is lower than estimated from the first part of the accumulation phase (see Table 2). Induction of metabolism dictates that the highest elimination rate should occur after exposure, in the depuration phase, when the enzymes are optimally induced. The peak most likely represents a change in the available pool of PAH in the soil medium, as proposed by Landrum [12], who observed the same phenomenon with amphipods in a spiked natural sediment. In his experiments, the apparent decline in bioavailability was up to an order of magnitude faster than the decrease in chemical extractability with time (ageing). Although ageing will affect the uptake of PAHs by earthworms [36], this process is slow compared to the time frame of this study. For PHE and PYR in sediments, a twofold increase in sorption was observed over the course of half a year [13] and, for PHE in earthworms, a decrease in uptake of a factor two to five in the same period [36]. In contrast, this study reveals a decrease in body residues of PHE of nearly two orders of magnitude in two weeks. A more likely explanation is that biodegradation causes this effect. Like earthworms, micro-organisms primarily take up chemicals from the water phase [38], and it is known that the rate of degradation can be sufficiently fast to cause mass transfer (desorption or dissolution) to be the limiting factor [21,25,35] . This implies that micro-organisms are able to effectively keep the dissolved concentration low and hence increase apparent sorption [10]. Because PAHs are strongly sorbed to the solid phases, and because the rates of desorption and dissolution are low, the effect of rapid biodegradation in pore water need not be directly apparent on a total soil basis. In this experiment, the medium was spiked by coating the sand fraction with PAH. After spiking, the chemical will slowly migrate from the crystalline phase, via the pore water, to the organic matrix. Because sorption is relatively rapid, and dissolution relatively slow, we expect that the pore water will be in equilibrium with the organic matter (resulting, after 1 wk equilibration, in Koc values comparable to natural sediments [10]). At the start of the exposure phase, most of the PAH will still be present in crystalline form, coated to the sand, due to the low mobility of the chemicals (H. Mulder, pers. comm.). Adding the earthworms stimulates the degradation and, possibly after an adaptation phase (as deduced from the change in total concentration), degradation starts but is rapidly limited by the transport from the crystalline form to the pore water. The pore water concentration can thus be expected to follow three stages: a relatively high concentration in the adaptation phase (equilibrium between water and organic matter), followed by 90

PAHs in artificial soil

exponential decrease (micro-organisms start to grow), and finally, low or near-zero concentrations (degradation and transport rates are balanced). In practice, the exponential phase will usually be very short and difficult to determine (H. Mulder, pers. comm.). The earthworms will respond to these changes in availability: after a high uptake in the initial phase, the earthworms must compete with the growing micro-organisms for the PAHs, and the rate of dissolution from the crystalline coating will limit both degradation and bioconcentration. This relatively rapid change in availability is consistent with the misfit in Figure 2. It is therefore likely that the observed peak in body residues represents a combination of biodegradation in the soil water and limited mass transport. For PHE and BaP, the ratio between these factors is such that the micro-organisms deplete the pore water more effectively than for PYR and FLU. Further support for this hypothesis is found in experiments with SPME fibres (a surrogate for organism lipids) in spiked sediments (T. Parkerton, pers. comm.). The same concentration peak was observed in the fibres as shown for earthworms and other organisms, but the peak disappeared when the system was poisoned with mercuric chloride (effectively stopping biodegradation). This strongly suggests that biodegradation in pore water is indeed responsible for the peak.

CONCLUSIONS In this paper, the kinetics of uptake and elimination of four PAHs are presented in the earthworm E. andrei. A typical peak in the body residues was observed for all PAHs, most likely caused by a change in bioavailability, in turn caused by biodegradation in pore water combined with a slow rate of mass transfer from the solid phase. Admittedly, the evidence for this hypothesis is circumstantial and deserves a more mechanistic experimental study. To describe the data, bioavailability was assumed to decline exponentially (Eq. 5), although the shape of the accumulation curves (Fig. 2) and the microbiological literature [35] suggest a more abrupt change. Estimates of the uptake rate (ku) are similar for all PAHs when expressed on soil solution basis (Fig. 3, ~2000 L/kg/d), thereby supporting the hypothesis that pore water is the bioavailable phase for earthworms. The elimination rate shows a decrease with Kow as expected but the values tend to be slightly lower than data from the literature (Fig. 4). Nevertheless, the dynamic BCFs (ku/ ke) compare well to the mechanistic partition approach [11] (Fig. 5). Because of the intricate dynamics of the available pool of PAHs, accumulation and toxicity experiments with spiked substrate are difficult to compare to field situations on the basis of external exposure. In the field, one would not expect to see such a rapid decline in availability because a steady state will be established between the rate of dissolution and desorption on the source side, and degradation and uptake on the demand side. Furthermore, at the same organic matter content, bioavailability will be lower in field situations because availability tends to decrease with increasing age of the contamination [13,25,36]. Nevertheless, when pore water is indeed the main route of exposure, it may be expected that the kinetic rate constants ku and ke also apply to field situations when they are expressed on a reliable soil-solution basis. It should, however, be noted that uptake and elimination rates are not only a property of the chemical and the size of the organism [28], but also depend on the substrate and the general activity of the organisms. This is evidenced from the influence of organic matter content on elimination [8], the fact that water exposure leads to different kinetics (although BCFs are similar), and experiments in which hunger stress leads to higher uptake rates [17].

91

Chapter 4

From the observations in this study, it can be concluded that accumulation experiments should be done dynamically to ensure that steady-state is actually achieved and to capture the kinetics. In the case of PAHs, a standard accumulation experiment (where the worms are only analysed after several weeks exposure) would have produced misleading results. The optimal method for spiking soils and equilibration times need further study because, with the current OECD procedure, an exposure system is created with little relevance to most field situations, with consequences for both risk assessment of soil organisms (estimation of toxicity) and for secondary poisoning (estimation of body residues).

ACKNOWLEDGEMENTS This paper is dedicated to the late Paco Antón Sánchez, who died in 1996. He executed the experiments reported in this paper, and did part of the data analysis during his stay at RIVM. Gerard Stroomberg (IVM, Amsterdam) performed the analysis of the metabolites, Kees van Gestel, Herman Eijsackers, Kees van Leeuwen, and Dick Sijm provided valuable comments. Finally we would like to thank the following persons at the RIVM laboratory for organic-analytical chemistry: G.S. Stil, C.J. Berkhoff, W.C. Hijman, I. de Lange, and S.H.M.A. Linders.

REFERENCES [1] [2]

[3]

[4] [5] [6]

[7] [8]

[9] [10]

[11] [12] [13]

92

BBA (1994). Richtlinien für die Zulassung von Pflanzenschutzmitteln im Prüfungsverfahren, Teil VI, 2-2, Auswirkungen von Pflanzenschutzmitteln auf die Reproduktion und das Wachstum von Eisenia fetida/Eisenia andrei. Saphir-Verlag, Ribbesbüttel, Germany. Belfroid A, A Van Wezel, M Sikkenk, K Van Gestel, W Seinen and J Hermens (1993). The toxicokinetic behavior of chlorobenzenes in earthworms (Eisenia andrei): experiments in water. Ecotox. Environ. Saf. 25:154-165. Belfroid A, J Meiling, D Sijm, J Hermens, W Seinen and K Van Gestel (1994). Uptake of hydrophobic halogenated aromatic hydrocarbons from food by earthworms (Eisenia andrei). Arch. Environ. Contam. Toxicol. 27:260-265. Belfroid A, M Sikkenk, W Seinen, K Van Gestel and J Hermens (1994). The toxicokinetic behavior of chlorobenzenes in earthworm (Eisenia andrei) experiments in soil. Environ. Toxicol. Chem. 13:93-99. Belfroid A, J Meiling, HJ Drenth, J Hermens, W Seinen and K Van Gestel (1995). Dietary uptake of superlipophilic compounds by earthworms (Eisenia andrei). Ecotox. Environ. Saf. 31:185-191. Belfroid A, M Van den Berg, W Seinen, J Hermens and K Van Gestel (1995). Uptake, bioavailability and elimination of hydrophobic compounds in earthworms (Eisenia andrei) in field-contaminated soil. Environ. Toxicol. Chem. 14:605-612. Belfroid AC, DTHM Sijm and CAM Van Gestel (1996). Bioavailability and toxicokinetics of hydrophobic aromatic compounds in benthic and terrestrial invertebrates. Environ. Rev. 4:276-299. Belfroid AC and DTHM Sijm (1998). Influence of soil organic matter content on elimination rates of hydrophobic compounds in the earthworm: possible causes and consequences. Chemosphere 37:12211234. Cerniglia CE (1992). Biodegradation of polycyclic aromatic hydrocarbons. Biodegradation 3:351-368. De Maagd PGJ, TL Sinnige, SM Schrap, A Opperhuizen and DTHM Sijm (1998). Sorption coefficients of polycyclic aromatic hydrocarbons for two lake sediments: influence of the bactericide sodium azide. Environ. Toxicol. Chem. 17:1899-1907. Jager T (1998). Mechanistic approach for estimating bioconcentration of organic chemicals in earthworms (Oligochaeta). Environ. Toxicol. Chem. 17:2080-2090. (Chapter 3 of this thesis) Landrum PF (1989). Bioavailability and toxicokinetics of polycyclic aromatic hydrocarbons sorbed to sediments for the amphipod Pontoporeia hoyi. Environ. Sci. Technol. 23:588-595. Landrum PF, BJ Eadie and WR Faust (1992). Variation in the bioavailability of polycyclic aromatic hydrocarbons to the amphipod Diporeia (spp.) with sediment aging. Environ. Toxicol. Chem. 11:1197-1208.

PAHs in artificial soil [14] Lee KE (1985). Earthworms. Their ecology and relationships with soils and land use. Academic Press, Sydney, Australia. [15] Leppänen MT and JVK Kukkonen (1998). Relative importance of ingested sediment and pore water as bioaccumulation routes for pyrene to oligochaete (Lumbriculus variegatus, Müller). Environ. Sci. Technol. 32:1503-1508. [16] Loonen H, DCG Muir, JR Parsons and HAJ Govers (1997). Bioaccumulation of polychlorinated dibenzo-p-dioxins in sediment by oligochaetes: influence of exposure pathway and contact time. Environ. Toxicol. Chem. 16:1518-1525. [17] Ma WC, J Immerzeel and J Bodt (1995). Earthworm and food interactions on bioaccumulation and disappearance in soil of polycyclic aromatic hydrocarbons: Studies on phenanthrene and fluoranthene. Ecotox. Environ. Saf. 32:226-232. [18] Ma WC, A Van Kleunen, J Immerzeel and PGJ De Maagd (1998). Bioaccumulation of polycyclic aromatic hydrocarbons by earthworms: assessment of equilibrium partitioning theory in in situ studies and water experiments. Environ. Toxicol. Chem. 17:1730-1737. [19] Mackay D and WY Shiu (1977). Aqueous solubility of polunuclear aromatic hydrocarbons. J. Chem. Engin. Data 22:399-402. [20] MedChem Project. MedChem Database, Ver. 3.55. Pomona College, Claremont, CA, USA. [21] Mulder H, AM Breure and JG Van Andel (1997). Physico-chemical processes affecting the bioavailability of PAHs. In: In Situ and On-Site Bioremediation: Volume 5. BC Alleman, A Leeson (eds.). Battelle Press, Columbus, Ohio, USA. pp. 643-648. [22] OECD (1984). Guideline for testing of chemicals no. 207. Earthworm, acute toxicity tests. Organization for Economic Cooperation and Development, Paris, France. [23] Opperhuizen A (1991). Bioconcentration and biomagnification: is a distinction necessary? In: Bioaccumulation in aquatic systems. Contributions to the assessment. R Nagel, R Loskill (eds.). VCH Verlagsgesellschaft mbH, Weinheim, Germany. pp. 67-80. [24] Penry DL and DP Weston (1998). Digestive determinants of benzo[a]pyrene and phenanthrene bioaccumulation by a deposit-feeding polychaete. Environ. Toxicol. Chem. 17:2254-2265. [25] Pignatello JJ and B Xing (1996). Mechanisms of slow sorption of organic chemicals to natural particles. Environ. Sci. Technol. 30:1-11. [26] Reinecke AJ and RG Nash (1984). Toxicity of 2,3,7,8-TCDD and short-term bioaccumulation by earthworms (Oligochaeta). Soil Biol. Biochem. 16:45-49. [27] Romijn CAFM, R Luttik and JH Canton (1994). Presentation of a general algorithm to include effect assessment on secondary poisoning in the derivation of environmental quality criteria. 2. Terrestrial food chains. Ecotox. Environ. Saf. 27:107-127. [28] Sijm DTHM and A Van der Linde (1995). Size-dependent bioconcentration kinetics of hydrophobic organic chemicals in fish based on diffusive mass transfer and allometric relationships. Environ. Sci. Technol. 29:2769-2777. [29] Stenersen J (1992). Uptake and metabolism of xenobiotics by earthworms. In: Ecotoxicology of earthworms. PW Greig-Smith, H Becker, PJ Edwards, F Heimbach (eds.). Athenaeum Press, Newcastle upon Tyne, UK. pp. 129-138. [30] Stroomberg GJ, JA De Knecht, F Ariese, CAM Van Gestel and NH Velthorst (1999). Pyrene metabolites in the hepatopancreas and gut of the isopod Porcellio scaber, a new biomarker for polycyclic aromatic hydrocarbon exposure in terrestrial ecosystems. Environ. Toxicol. Chem. 18:2217-2224. [31] Tomlin AD (1992). Behaviour as a source of earthworm susceptibility to ecotoxicants. In: Ecotoxicology of earthworms. PW Greig-Smith, H Becker, PJ Edwards, F Heimbach (eds.). Athenaeum Press, Newcastle upon Tyne, UK. pp. 116-125. [32] Van Gestel CAM and WC Ma (1988). Toxicity and bioaccumulation of chlorophenols in earthworms, in relation to bioavailability in soil. Ecotox. Environ. Saf. 15:289-297. [33] Van Gestel CAM, WA Van Dis, EM Van Breemen and PM Sparenburg (1989). Development of a standardized reproduction toxicity test with the earthworm species Eisenia fetida andrei using copper, pentachlorophenol, and 2,4-dichloroaniline. Ecotox. Environ. Saf. 18:305-312. [34] Van Gestel CAM, EM Dirven-Van Breemen and R Baerselman (1992). Influence of environmental conditions on the growth and reproduction of the earthworm Eisenia andrei in an artificial soil substrate. Pedobiologia 36:109-120. [35] Volkering F, AM Breure, A Sterkenburg and JG Van Andel (1992). Microbial degradation of polycyclic aromatic hydrocarbons: effect of substrate availability on bacterial growth kinetics. Appl. Microbiol. Biotechnol. 36:548-552. [36] White JC, JW Kelsey, PB Hatzinger and M Alexander (1997). Factors affecting sequestration and bioavailability of phenanthrene in soils. Environ. Toxicol. Chem. 16:2040-2045. [37] Widianarko B and N Van Straalen (1996). Toxicokinetics-based survival analysis in bioassays using nonpersistent chemicals. Environ. Toxicol. Chem. 15:402-406.

93

Chapter 4 [38] Wodzinski RS and JE Coyle (1974). Physical state of phenanthrene for utilization by bacteria. Appl. Microbiol. 27:1081-1084.

94

PAHs in field-polluted soils

5 Availability of Polycyclic Aromatic Hydrocarbons to Earthworms (Eisenia andrei, Oligochaeta) in Field-Polluted Soils and Soil-Sediment Mixtures1

Tjalling Jager, Rob Baerselman, Ellen Dijkman, Arthur de Groot, Elbert Hogendoorn, Ad de Jong, Jantien Kruitbosch, Willie Peijnenburg Appeared in 2003: Environmental Toxicology and Chemistry 22(4):767-775

ABSTRACT  The bioavailability of polycyclic aromatic hydrocarbons (PAHs) for earthworms (Eisenia andrei) was experimentally determined in seven field-polluted soils and 15 soil-sediment mixtures. The pore-water concentration of most PAHs was higher than predicted. However, most of the compound was associated with dissolved organic carbon (DOC) and not directly available for uptake by earthworms. The apparent sorption could be reasonably predicted on the basis of interactions with DOC. However, the biota-soil accumulation factors (BSAF) for earthworms were up to two orders of magnitude lower than predicted by equilibrium partitioning. The large variability between sites was not fully explained by differences in sorption. Experimental results indicate that the pool of freely-dissolved PAHs in the pore water became partially depleted because of uptake by the earthworms, and that bioaccumulation is thus also influenced by the kinetics of PAH desorption and mass transport. A pilot study with Lumbricus rubellus showed that steady-state body residues were well correlated to E. andrei. Current results show that depositing dredge spoil on land may lead to increased bioavailability of the lower molecular weight PAHs. However, risk assessment can conservatively rely on equilibrium partitioning, but accurate prediction requires quantification of the kinetics of bioavailability.

This research was performed on behalf of the Ministry of Housing, Spatial Planning and the Environment, within the RIVM project 733007. Part of the work was presented in the RIVM report of Peijnenburg et al. in 2000 (no.: 733007 007).

1

95

Chapter 5

INTRODUCTION Polycyclic aromatic hydrocarbons (PAHs) are common pollutants in soils, and typically tend to be persistent in soils because of their relatively low mobility and high resistance to degradation. Over the past 100 years, levels of PAHs in soils have been steadily increasing, primarily as a result of combustion of fossil fuels, followed by atmospheric deposition [15]. An additional input to soils in the Netherlands follows from dredging practices. Sediments in regional waters are dredged on a regular basis to ensure sufficient water depth for navigation and water discharge. The main part of these dredged sediments is placed on soils, where they can contribute to the local PAH contamination [8]. To prevent unacceptable levels in soil, environmental quality criteria for sediment are in place, based on the total level of PAHs in the sediment. However, quality standards and risk assessment should preferably be based on actual bioavailable concentrations in soil, which appears to be the chemical freely dissolved in pore water [3]. Earthworms are common representatives of the soil macrofauna, that live in close contact with the soil. Uptake of organic chemicals in earthworms is assumed to occur through passive diffusion, driven by the fugacity difference between pore water and the organism’s tissues. This equilibrium partitioning (EP) approach generally provides a satisfactory description for earthworms [12], although it seems likely that in many situations limited diffusion or other transport processes prevent establishment of true equilibrium [4,21]. Bioavailability and uptake of PAHs is also quite predictable based on pore-water concentrations, at least under laboratory conditions when using spiked artificial soil [13]. However, under field conditions, bioavailability of PAHs is generally much more difficult to predict as a multitude of factors play a role, including sequestration or “ageing” [16], mass-transport limitations [25], strong binding to soot [10], and variability in the polarity of organic matter [17]. The purpose of this study was to examine the current practice of depositing dredge spoil on soils. This was done by experimentally determining the bioavailability of PAHs in typical field-polluted soils and soil-sediment mixtures. The measure of availability that is used here is the dynamic accumulation pattern in the compost worm (Eisenia andrei), expressed by the steady-state BSAF and the elimination rate constant. Bioavailability may differ between species, and although the compost worm is not a typical soil dweller, it is used as standard test organism for soil [26]. The BSAFs are compared to EP estimates to determine the predictive power of the EP approach for bioavailability of PAHs in typical field-polluted media. Furthermore, we attempt to explain the observed bioavailability from the dynamic sorption and uptake processes that determine it.

MATERIALS AND METHODS SAMPLING OF SOILS AND SEDIMENTS Sediment samples were taken on 45 contaminated sites in the Netherlands. From these, 15 sites were selected for further use on the basis of elevated levels of either metals (not discussed in this paper) or PAHs, and representing sediments with different predominant constituents (sand, clay and peat). Furthermore, seven soils with the same predominant constituents were sampled. After removal of the vegetation, soil was taken from the upper soil horizon (0–20 cm). Soil and sediments were stored in closed containers at 5°C, and within two weeks the samples were sieved (4 mm) and homogenised in a baker’s mill. For several samples, the high clay content prohibited this procedure. For these soils,

96

PAHs in field-polluted soils

large objects (e.g. stones and roots) were removed, after which the samples were homogenised manually. Mixtures of soil and sediment were created by adding a sediment to one of the seven soils in a fixed proportion: sand and clay 1:4, peat 1:2 (sediment:soil on dwt basis). This mimics realistic practices in the Netherlands. In total 22 media (seven soils and 15 mixtures) were used for subsequent testing (see Table 1). Table 1. Soil properties for the soils (S) and soil-sediment mixtures (M) used in this study, including fraction organic carbon (Foc), clay content, pH, dissolved organic carbon (DOC), and total content of PAHs (ΣPAH). Soil/mixture code S1 S2 S3 S4 S5 S6 S7 M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 M12 M13 M14 M15

mixed with

S5 S4 S6 S4 S2 S3 S6 S3 S2 S1 S5 S4 S7 S1 S7

Character peat sand Clay peat Clay sand clay clay peat sand peat sand clay sand clay sand peat clay peat clay peat clay Minimum Maximum

Foc (%) 10% 2.4% 4.5% 14% 3.7% 1.3% 6.4% 4.4% 11% 1.9% 9.8% 2.4% 4.6% 2.2% 3.3% 3.6% 10% 4.7% 14% 7.0% 9.8% 6.3% 1.3% 14%

Clay (%) 14% 6.1% 12% 8.8% 21% 0.69% 35% 20% 14% 0.61% 11% 4.7% 11% 2.4% 9.8% 7.9% 48% 23% 14% 24% 16% 39% 0.61% 48%

pH (CaCl2) 7.0 4.7 6.8 7.2 7.4 7.1 6.1 7.5 6.8 7.1 7.0 5.5 7.1 7.3 6.8 6.6 7.3 7.3 7.3 6.2 7.4 6.7 4.7 7.5

DOC (mg/L) 76 315 57 77 47 91 108 54 95 152 121 311 116 156 78 298 76 62 102 113 33 84 33 315

ΣPAH (mg/kgdwt) 10 0.25 2.8 25 1.8 17 1.2 2.3 13 16 11 4.6 3.8 19 3.9 2.6 12 1.3 23 1.1 14 1.8 0.25 25

PHYSICAL AND CHEMICAL ANALYSES In the 22 exposure media, the following parameters were determined: pH (CaCl2), organic carbon content (Elemental Analyser by Fisons Instruments, Model EA 1108, Rodana, Italy), clay content (≤ 2 µm, gravimetric analysis), water holding capacity, and DOC content of pore water (Dohrmann Division DC-190, Santa Clara, CA, USA). The media were analysed for selected PAHs (Table 2) as described earlier [13], after cryogenic grinding. For the determination of PAHs in pore water, soils were brought to 85% of the water holding capacity using 2 mmolar Ca(NO3)2. After at least two weeks of equilibration, pore water was collected using a centrifugation method in which the water is pushed upwards and can be collected with a Pasteur pipette. This procedure resulted in a clear supernatant, and filtering was judged unnecessary (pilot experiments showed that a 0.45 µm filter also removed large amounts of dissolved PAHs). Total PAH contents in pore water were determined after exhaustive extraction, employing acetone and petroleum ether, drying over sodium sulphate, and evaporation on a Kuderna Danish apparatus (custom made at our institute). Residues were dissolved in 1 ml acetone/water (2:3), and 100 µL was injected into a liquid chromatography (LC) system with fluorescence detection.

97

Chapter 5 Table 2. Summary of the PAHs included in this study, their log Kow values [24], and the percentage of the particular PAH in the total concentration (mean and range). The last three columns give the number of steady-state body residues (Cb∞), elimination rates (ke) and sorption coefficients (Koc-app) that could be determined. Naphthalene Acenaphthene Fluorene Phenanthrene Anthracene Fluoranthene Pyrene Benzo[a]anthracene Chrysene Benzo[b]fluoranthene Benzo[k]fluoranthene Benzo[a]pyrene Dibenzo[ah]anthracene Benzo[ghi]perylene

log Kow 3.37 3.92 4.18 4.57 4.54 5.22 5.18 5.91 5.86 5.80 6.00 6.04 6.75 6.50

Percentage of ΣPAH 1.8 (0.46–5.5) 0.75 (0–1.4) 1.0 (0.50–1.5) 8.6 (4.1–14) 2.0 (0.58–3.2) (10–23) 18 13 (8–16) 9.0 (6.6–10) (6.9–10) 9.0 12 (8.7–18) 1.1 (0.85–1.4) 4.7 (3.4–5.7) 1.4 (0.81–2.6) 3.7 (2.1–5.7)

No. Cb∞ 0 0 0 4 7 12 12 18 19 18 17 17 4 17

No. ke 0 0 0 2 4 7 10 18 18 18 17 15 0 6

No. Koc-app 16 0 14 21 5 15 20 15 16 17 19 19 16 16

Freely-dissolved concentrations of benzo[a]pyrene were determined by leading 600 µL pore water through a capillary chemical-bonded siloxane (CBSIL) gas chromatography (GC) column (55 cm × 0.32 mm ID, film thickness 1.2 µm) at 0.3 mL/min. The capillary was rinsed with 600 µL high-performance liquid chromatography water with d12-labeled PAH isotopes as internal standards. Nitrogen was led through the capillary, after which it was connected to a GC with high-resolution mass spectrometry detection. Using a cold trap, the analytes were concentrated at the top of the analytical column and subsequently separated and determined. PAHs were extracted from earthworm tissues by means of an Ultra-Turrax extraction with petroleum ether. After drying over sodium sulphate, the samples were concentrated to 1 mL and transferred to pre-weighed autosampler bottles. Part of the sample was injected into a Gel-Permeation-Chromatography system (Model 305 LC pump, Model 321 autosampler, fraction collector Model FC 204 from Gislon, Villier-le-Bel, France) for cleanup (lipid separation). The fraction with analytes was evaporated, dissolved in 1 ml acetone-water (2:3), and 100 µL was injected into the LC system. BIOASSAYS Earthworms (Eisenia andrei) were obtained from mass cultures at our institute, kept under climatised conditions (temperature 20 ± 2°C ). Juvenile worms were selected to ensure that no loss of accumulated chemicals would occur due to reproduction. The worms were weighed, placed in plastic containers (10 worms per container, total ~3 gwwt) with 600–900 gwwt of medium, and incubated at 20°C in full light (to prevent escape). Instead of maximising replication, we chose to use a large number of exposure periods: 0.25, 1, 2, 3, 4, 7, 10, 12, 14 and 21 d. Worms were recaptured, and depurated (48 h on moist filter paper, which was changed after 24 h). Subsequently, the worms were divided into two groups (five for PAH analysis and five for metal analysis; metals not discussed in this paper) and frozen before analysis. Four groups of five worms, taken directly from the culture, were used for t = 0, after depuration. To test the validity of E. andrei as a model for other species, a similar, but limited set of assays was performed with Lumbricus rubellus, a species that occurs in a range of Dutch field soils. Adult or sub-adult specimens were obtained from a low-intensity culture, and six worms (total ~3 gwwt) were placed in containers with 750–900 gwwt of medium. Only two soil-sediment mixtures were tested,

98

PAHs in field-polluted soils

with five exposure periods (1, 3, 7, 14, 21 d). Three worms from each container were frozen for PAH analysis, and four groups of three worms were used for t = 0, after depuration. DATA ANALYSIS Apparent sorption coefficients (Koc-app) were calculated from PAH concentrations in the solid phase (Cs) and the total PAH concentration in pore water (Cw-tot), and normalised to the fraction of organic carbon (Foc) in the solid test medium:

K oc − app =

Cs

(L/kgoc)

C w −tot Foc

Eq. 1

In case of “significant accumulation” (body residues at least 2 standard deviations above the level in the culture, and above the detection limit), the measured body residues in earthworms (Cb) were fitted with a standard one-compartment model:

(

C b (t ) = C b 0 e − k e t + C b ∞ 1 − e − k e t

)

(µg/kgwwt)

Eq. 2

resulting in estimates of the steady-state concentration (Cb∞) and the overall elimination rate constant (ke in d-1). Note that the calculation of the elimination rate from an accumulation experiment is valid only when the bioavailable exposure concentration is constant during the bioassay. The Cb0 is the initial concentration in the organisms from the culture, which was close to the detection limit for all PAHs. The initial concentration was fitted on the data unless negative or unrealistically high values were obtained (in that case, Cb0 was fixed to the initial value measured in the worms). Regression analyses were performed with the software package Graphpad Prism™ 2.01 (San Diego, CA, USA). The BSAFs were calculated from the steady-state levels in earthworm, normalised to lipids (Flip, taken as 1% on wet-weight basis [12]), and the total soil concentrations, normalised to the organic carbon fraction in the soil (Foc): BSAF =

C b ∞ Foc C s Flip

(kgoc/kglip)

Eq. 3

Steady-state body residues were also converted to bioconcentration factors (BCF) on soilsolution basis by using both measured and estimated pore-water concentrations (Cw): BCF =

Cb∞ Cw

(Lwater/kgwwt)

Eq. 4

CORRELATIONS AND DIFFERENCES BETWEEN SITES AND PAHS Pearson correlation coefficients were calculated between the estimates of Koc-app, BSAF and ke. Furthermore, these parameters were correlated to Kow and soil properties (after logtransformation, using Graphpad Prism™ 2.01). Only significant correlations (p < 0.05) are reported. Not all parameters could be determined for all PAHs in all soils. The missing data hinder a proper comparison between the various PAHs, and between mixtures and soils from different sites. To obtain a stronger data set, five PAHs were selected for which most parameters could be determined at most sites (see Table 2): benzo[a]anthracene, chrysene, benzo[b]fluoranthene, benzo[k]fluoranthene and benzo[a]pyrene. For these PAHs, completely filled matrices of PAH × site could be obtained for Koc-app, BSAF and ke.

99

Chapter 5

These data matrices were log transformed and analysed with analysis of variance (ANOVA, one-way, repeated measures) using Graphpad Prism™ 2.01, to test for differences between sites and between PAHs.

RESULTS AND DISCUSSION The ranges of the properties of the soils and soil-sediment mixtures are shown in Table 1. Total PAH levels were hardly influenced by the addition of sediment; the ΣPAH in the mixtures was between a factor of 0.4 and 1.4 of the level in the original soil. Only in the least contaminated soil (S2), was a substantial increase in PAH levels observed after adding sediment (a factor of 10–18 on ΣPAH). The individual PAHs included in this study are shown in Table 2, along with their range of concentrations in soil. Even though the total level of PAHs varies between soils by two orders of magnitude (Table 1), the proportional contribution of the individual PAHs is quite similar in all soils (Table 2). This indicates that the contamination of these soils is likely due to non-point source contamination, without substantial impact from specific local sources. Table 2 also shows the number of steady-state concentrations and elimination rates determined from the earthworm bioassays, and the number of Koc-app values derived.

log K oc

SORPTION The apparent sorption coefficient 6 (Koc-app, Eq. 1) was calculated after normalisation with respect to the organic carbon (OC) content in 5 each soil. No difference was observed in Koc-app values that could be determined in soil-sediment 4 mixtures and soils alone. However, the low-molecular-weight PAHs (naphthalene and fluorene) were 3 not detected in pore water from most terrestrial soils. As the total 3 4 5 6 7 levels of PAHs did not differ log K ow between soils and mixtures, it is Figure 1. Measured apparent sorption data (Koc-app in conceivable that Koc-app for these L/kgoc, geometric mean and s.d.) in relation to several compounds in soils was actually model predictions. Furthermore, Koc based on higher than those in the mixtures. measured freely-dissolved concentrations of The Koc-app values show appreciable benzo[a]pyrene is shown. ○ apparent sorption; ■ BaP variation between sites and only freely dissolved; ····· predicted [27]; - - - predicted [9]; little relationship with Kow (Fig. 1). ―― predicted [9] + DOC. This is contrary to the expectations of a linearly increasing relationship with hydrophobicity [27]. As can be seen from Figure 1, the PAHs with a low molecular weight have higher experimental values of Koc-app than predicted on the basis of the quantitative structure-activity relationship (QSAR) of Sabljić et al. [27]. In contrast, the PAHs of higher molecular weight are below this estimate. Two factors may explain this pattern. First, PAHs are known to sorb more strongly to OC than expected on the basis of their polarity [6,7]. For this reason, this QSAR (derived from a wide range of hydrophobic chemicals) may not be appropriate. Additionally, Koc values

100

PAHs in field-polluted soils

were calculated from an empirical relationship found by Gerstl [9], specific for PAHs in soils (Fig. 1): log K oc = 0.762 log K ow + 1.051

(n = 20)

Eq. 5

A second explanation concerns the procedure used in this study to determine Koc. Generally, Koc is derived by shaking the soil with excess water (or an aqueous solution), but in our experiments, the soils were centrifuged to obtain a better estimate of the pore water in situ. A result of this procedure is that the solution thus obtained contains a relatively high concentration of DOC (we contribute all the organic carbon in our pore water samples to DOC although some particles may be included also). PAHs interact with DOC, thereby increasing the apparent water solubility of the compounds and thus lowering the apparent Koc. The “true” sorption coefficient (Koc-free) and the fraction of the chemical that is freely dissolved in the soil solution determine the measured “apparent” sorption coefficient (Koc-app). The dissolved fraction can be calculated from the concentration of DOC ([DOC] in kg/L) and the partition coefficient with DOC (Kdoc in L/kg), leading to [6]:

K oc − app =

K oc − free 1 + [DOC ] K doc

(L/kgoc)

Eq. 6

Only for benzo[a]pyrene (BaP), the freely-dissolved concentration was directly measured using an experimental method (see materials and methods). The method was not extensively validated yet, so results must be interpreted with care. The measured Koc-free for this compound is also shown in Figure 1. The average value corresponds quite nicely to the QSAR of Gerstl (Eq. 5), although the variation is still considerable. Using Equation 6, Kdoc can be calculated for BaP. The resulting Kdoc was on average a factor of three lower than Koc-free, which is well in line with the factor of 2 found for humic acid [6]. The total deviation was almost a factor of ten around this value. This deviation is caused by differences in organic matter composition, but may also reflect technical difficulties in determining freely-dissolved fractions [5]. It is expected that Kdoc is related to the hydrophobicity of the compound [5], which allows for a prediction of Koc-app for PAHs, based on the following assumptions. Firstly, the QSAR of Equation 5 is a valid estimation of Koc-free. This assumption is supported by the correspondence of the QSAR with Koc-app for the less hydrophobic PAHs, and the measured Koc-free for BaP. Secondly, Kdoc is a factor of 3 lower than Koc-free for all PAHs, and, finally, we assume an average DOC concentration of 100 mg/L in all soils (see Table 1). With these assumptions, the trend in the data is well described (solid line in Fig. 1), indicating that sorption to DOC can indeed explain the loss of linearity in Figure 1. The observed variation in Kdoc for BaP, and the DOC content between soils, is largely sufficient to cover the observed variation in this figure. However, several of the individual Koc-app and Koc-free values exceed the line of the QSAR by Gerstl (indicated by the standard deviations exceeding the broken line in Fig. 1). These values must represent higher values of Koc-free than predicted, indicating that not all variation between soils is explained by the action of DOC. As stated in the introduction, several processes are known to influence the observed sorption and partitioning of PAHs in soils. BIOTA-SOIL ACCUMULATION FACTORS The one-compartment model of Equation 2 agreed reasonably well with most of the data although the variation in goodness of fit was considerable. Most r2-values ranged from 0.80–0.98, but several fits were poor (r2 ≈ 0.50). Several typical examples of accumulation

101

Chapter 5

curves are shown in Figure 2. Most data showed a smooth increase in time, but several PAH-soil combinations showed a very rapid steady state (within 1 d), precluding a firm estimation of ke. This pattern was most clear for dibenzo[ah]anthracene and benzo[ghi]perylene, the most hydrophobic PAHs examined. It is unlikely that true equilibrium is reached so rapidly for these PAHs, and this pattern may therefore be related to an experimental artefact. This could be depletion of the bioavailable phase, although the BSAFs for these PAHs are not particularly low when a rapid steady state is evident. A few curves showed the distinct pattern of an apparent maximum in the body residues. This behaviour was only found in three soil-sediment mixtures (M4, M5, M7), and was limited to one or more of the lighter PAHs (anthracene, fluoranthene and pyrene). In the few cases where this pattern or a rapid steady state was observed, only the estimated BSAF derived from the one-compartment model was used. The same pattern of an apparent maximum and subsequent decrease in the accumulation was previously observed in studies of PAHs spiked into soils [13,22]. A likely cause for this pattern is a substantial decrease in bioavailability during the bioassay, possibly mediated by biodegradation in pore water [13], or resulting from a very rapid aging or sequestration of PAHs in soil [16] during the course of exposure, thus decreasing the concentration of PAHs in the soil solution. Induced biotransformation could also lead to such a peak. Although earthworms are able to transform pyrene, the amount of metabolites formed is very low [13], and the P450 system is apparently not induced by exposure to PAHs [1]. Furthermore, if biotransformation in the earthworm is the cause of this result, the pattern is expected to be evident in more than just a few soils. 60

body residues

8 40

6 4

20 2

A

0

body residues

0

5

10

15

20

0

75

150

50

100

25

50

C

0 0

5

10

time (days)

15

B

0 5

10

15

D

0 0

5

10

15

20

time (days)

Figure 2. Examples of typical accumulation curves for earthworms in soil-sediment mixtures (body residues in µg/kgwwt): near perfect agreement to the model (A), and a poor fit with rapid achievement of steady state (B); distinct peak (C), and less distinct peak with new steady state after 7 d (D).

Overall, soil-sediment mixtures yielded similar BSAFs as soils alone, but on careful observation a difference can be observed. For the low-molecular-weight PAHs (up to pyrene in Table 2), few BSAFs could be calculated, and most of them are for soil-sediment mixtures. In fact, only one soil yielded significant body residues in the earthworms (S6, which has the highest level of PAHs and the lowest Foc, see Table 1). The BSAFs for soil 102

PAHs in field-polluted soils

only and soil-sediment mixtures are shown in Figure 3A and 3B, respectively. As little difference was observed in total levels between soils and mixtures, it appears that the low-molecular-weight PAHs are hardly available for uptake in the soils, but are more bioavailable in the sediments. The calculated BSAFs are quite variable and show little relationship with Kow (Fig. 3A/B). The BSAFs are on average 0.23 kgoc/kglip, but the variation is large. The lack of influence of hydrophobicity on BSAF is to be expected as both sorption and accumulation from pore water increase with increasing Kow. The EP estimate of BSAF in Figure 3 is taken as the bioconcentration factor (BCF) from an equilibrium partitioning estimate [12], divided by the QSAR from Equation 5. soil only

0

-1

A

-2 4

5

6

0

-1

B

7

4

5

6

7

log K ow

estimated pore water

5

BaP freely dissolved EP estimate

4 log BCF

EP estimate

-2

log Kow

3 2 1 0

soil-sediment mixtures

1

EP estimate

log BSAF

log BSAF

1

C 4

5

6

7

Figure 3. Shown against log Kow are biota-soil accumulation factors (BSAF kgoc/kglip) in soils (A), and soil-sediment mixtures (B), and bioconcentration factors (BCF L/kgwwt, C; based on estimated freely-dissolved concentrations). Also shown is the estimate based on equilibrium partitioning (EP), and BCFs of benzo[a]pyrene, based on measured freely-dissolved concentrations. The broken line is a log-linear regression with 95% confidence intervals.

log K ow

The observed BSAFs are much lower than the EP estimate (up to two orders of magnitude), and also lower than the maximum observed in spiked artificial soil medium [13] (values of 2–8 kgoc/kglip), although body residues in artificial soil decreased after reaching this maximum. The large variation in BSAFs may partly reflect sequestration in these field-polluted media, however, in the previous section we argued that Equation 5 was likely to be a reasonable estimate of the average Koc-free in our soils. These consistently lower BSAFs must therefore also have other explanations. Generally, the experimental BSAFs are quite comparable to similar bioassays with L. terrestris [19] (BSAFs on average between 0.13 and 0.41 for PAHs, with large variation between sites). Ma et al. [23] reported BSAFs for PAHs in L. rubellus, sampled from floodplain sites. Their BSAFs also show a substantial variation between sites, but their values are on average a factor of four lower than the present results. However, the actual concentration that these field-collected earthworms had been exposed to is not easily reconstructed (the soil concentration was taken as a bulk sample from the top 20 cm).

103

Chapter 5

BIOCONCENTRATION FACTORS The BCFs (Eq. 4) calculated on the basis of the measured total concentrations in pore water were low (~50 L/kgwwt), and showed little variation between the different PAHs. This effect was predicted earlier [12] in case BCF is estimated on the basis of the total concentration in pore water (thus including DOC). This supports the findings by other authors who claim that only the freely-dissolved concentration in pore water is available for uptake by organisms [11,20]. The BCFs on the basis of estimated freely-dissolved concentrations (using Eq. 5 as estimate of Koc-free) show a linear increase with Kow as expected from the theoretical relationship (Fig. 3C). The data are generally lower than the theoretical estimate (on average a factor of 11), which is consistent with the low BSAFs, discussed in the previous section. This seems to be a general trend, as BCFs for earthworms tested in a soil medium are on average a factor of six lower than expected on the basis of EP theory [12]. The BCF of BaP can also be expressed on the basis of the measured freely-dissolved concentration in pore water. It is interesting to see that this BCF is similar to the data based on the QSAR for Koc-free, and that they have a similar variation. Although these measurements must be interpreted carefully, it appears that free concentrations do not fully determine BCFs. It is possible that the free pool is partially depleted by the earthworms and that differences between soils (e.g. the rate of desorption) determine the final body residues. A simple mass-balance calculation for BaP shows that the amount of chemical accumulated in earthworms during the bioassay exceeds the amount freely dissolved in pore water by an order of magnitude, on average. This implies that the freely-dissolved pool of PAHs in soil solution is concurrently being partially depleted and replenished before a steady state is achieved. The total amount of BaP in pore water (including DOC) is in the same order of magnitude as the amount accumulated by the worms. Therefore, it is likely that the kinetics of depletion of pore water PAHs, and the kinetics of their replenishment, have influenced the accumulation patterns and the steady-state body residues.

The calculation of a BCF implies that pore water is considered to be the primary bioavailable phase for earthworms. The earthworms were feeding on soil too, and a large part of the body burden may actually be derived from the gut contents. Nevertheless, we believe the focus on exposure via pore water to be appropriate as body residues were in virtually all cases lower than expected on the basis of EP. Furthermore, uptake from the gut contents is not fundamentally different from uptake through the external skin, as both are likely mediated through a dissolved phase and driven by passive diffusion [14]. However, processes in the gut may release residual PAH fractions that are otherwise not participating in EP. Thus, EP can be considered to estimate the maximum amount that can be taken up, but the total variation in body residues and uptake kinetics may be driven by differences in assimilation efficiencies between soils, as well as differences in desorption kinetics of PAHs from soils. ELIMINATION RATE CONSTANTS An estimate of the elimination rate constant (ke) follows from fitting Equation 2 to the accumulation data. The rate constant was generally poorly identified by the data, and these results must therefore be interpreted with care. Even though the relationship between log ke and log Kow is significant (p < 0.05), the slope of the regression is only –0.16 (Fig. 4). These findings are in contrast with those of Belfroid et al. [2] who derived a slope of –0.66 for PCBs and chlorobenzenes. Our ke values are in fact quite similar for all PAHs (95% are within 0.72 d-1 ± factor 3). The rate constants for fluoranthene and pyrene are in the same range as values reported from earthworms in OECD medium [13], but the values for benzo[a]pyrene are generally larger in the present study. It is however very well 104

PAHs in field-polluted soils

elimination rate (1/d)

possible that these apparent 1 elimination rate constants are artefacts. As discussed in the previous section, desorption of PAHs from organic carbon phases 0 is necessary to establish the observed body residues in the worms. If these desorption rates are relatively slow, Equation 2 is not a -1 valid description of the accumulation process anymore as the exposure concentration (the 4 5 6 7 dissolved pool) is not constant. log Kow Any ke estimate will then be a Figure 4. The rate constant for elimination (ke) as compound parameter, including the estimated from the accumulation bioassays, shown kinetics of depletion and replenishment of the freely- against log Kow. Broken line is a log-linear regression dissolved pool. Furthermore, the with 95% confidence limits. estimate of ke will include uptake from the gut as well as across the skin. CORRELATIONS AND DIFFERENCES BETWEEN PAHS AND BETWEEN SITES Correlations for the complete data set are listed in Table 3. Because there are so many significant correlations between the estimated parameters and with the soil properties (which are also correlated amongst themselves), it is difficult to make inferences about causal relationships. BSAF and ke are negatively correlated, which implies that when a rapid steady state is achieved, the BSAF is lower than expected on the basis of EP. This finding is consistent with the hypothesis of partial depletion and replenishment. BSAF correlates negatively to Koc-app, indicating low total pore-water levels when BSAF is low. This correlation supports the assumption that uptake is driven by the pore-water concentration, but care must be taken as Koc-app is not necessarily representative for freelydissolved concentrations. Table 3. Significant correlations (p < 0.05) for the estimated sorption (Koc-app), BSAF, and elimination rates (ke), after log transformation (n.s. is not significant). Also correlations between these parameters and hydrophobicity (Kow), Foc, clay content, pH, DOC, concentration in soil of individual compounds (Cs), and the total level of PAHs (ΣPAH). Koc-app 1 –0.53 0.30

BSAF

ke

Koc-app BSAF ke

1 –0.48

1

Foc clay pH DOC Cs ΣPAHs

Koc-app n.s. –0.32 0.31 n.s. 0.63 0.46

BSAF 0.33 0.58 –0.28 n.s. –0.65 –0.61

ke n.s. –0.23 n.s. 0.38 0.24 n.s.

Kow 0.31 0.33 –0.24 Other correlations Foc and clay correlated 0.70 pH and DOC correlated –0.69 correlated to pH 0.62

The correlations with Kow are significant, but not very large. The correlations with the soil properties are difficult to interpret, although a few conclusions seem justified. The large differences in Koc-app between sites (Fig. 1) may be related to differences in the quality of

105

Chapter 5

the soil organic matter between sites (and thus differences in Koc-free), but they may also reflect differences in DOC. The correlation with the quantity of DOC is not significant, but the apparent sorption may be determined largely by DOC composition (which is also indicated by the correlation with pH) [5,11,20]. This in turn would help explain the correlation between apparent sorption and the PAH concentration in soil. DOC-bound chemicals are potentially available for leaching, and it is conceivable that a low value of Koc-app at a site may contribute to leaching of PAHs from the top-soil layer. The highest BSAFs are found in soils with a high organic matter and clay content, a low pH, and low PAH levels. Although these correlations may be practical for predicting potential problems in the field, it is likely that the causal relationship is through the action of these properties on PAH sorption. The differences between PAHs and between sites are calculated using data for five PAHs only, as explained in materials and methods. Even though the selected PAHs have very similar log Kow values (5.80–6.04), significant (ANOVA, p < 0.0001) differences exist between them in Koc-app, BSAF and ke. This shows that the behaviour of the PAHs is determined not only by their hydrophobicity. The mean difference between the PAHs is, however, not very large (less than a factor of 7 for Koc-app and BSAF, and less than a factor of 2 for ke). Significant (ANOVA, p < 0.0001) differences also exist in Koc-app, BSAF and ke between the sites for the five selected PAHs. The mean differences are much larger than between PAHs (a factor of ~30 for Koc-app and BSAF, a factor of 9 for ke), indicating that Foc is not the only soil property governing sorption and bioaccumulation. No clear difference was observed between soil-sediment mixtures and soils alone, and no new relations with soil properties (the correlations are similar to those shown in Table 3).

Lumbricus rubellus

PILOT STUDY WITH LUMBRICUS RUBELLUS A pilot study was performed with measured 2 two soil-sediment mixtures (M3 and 1:1 line M15) and a different species of earthworm: L. rubellus. In contrast to E. andrei, this species is common in 1 Dutch field soils. As shown in Figure 5, the steady-state body residues are clearly correlated between both species (r = 0.91), but body residues 0 of PAHs in L. rubellus are on average a factor of 2 lower than in E. andrei. It is unclear what has caused the 1 2 difference between these species, but Eisenia andrei differences in lipid content (not Figure 5. Relation between the 10log of the steadymeasured) or in body size may state body residues (µg/kgwwt) of the two species of contribute. The weight ratio of earthworm in two soil-sediment mixtures. The worm:soil was similar for both broken line is a log-linear regression with 95% species but the individuals of L. confidence limits. rubellus were much larger, causing differences in uptake kinetics (through the ratio of the surface area to body volume). Especially deviating is a value for benzo[ghi]perylene where levels in L. rubellus were 17 times lower. For this compound, a rapid steady-state was in many cases observed for E. andrei (also in the same soil), so accumulation of this compound may be especially influenced by kinetic constraints.

106

PAHs in field-polluted soils

CONCLUSIONS In this study, bioassays were performed with the earthworm Eisenia andrei in seven soils and 15 soil-sediment mixtures. The soil-sediment mixtures tested do not apparently differ in sorption or accumulation of the high-molecular-weight PAHs from the terrestrial soils tested. However, the low-molecular-weight PAHs (up to pyrene) have a very low bioavailability in soil, but are more readily taken up from the soil-sediment mixtures. In several of the mixtures, bioavailability was already declining during the course of the bioaccumulation experiment. Risk assessment for dredged materials thus has to be aware that the low-molecular-weight PAHs can be more bioavailable than in the terrestrial soil found on-site, directly after depositing dredge spoil. Total levels of PAHs in pore water were higher than predicted, but most of the dissolved compound is associated with DOC. The current data set supports earlier assumptions that DOC-bound chemicals are not directly available for uptake by earthworms. Prediction of Koc-free can be done on the basis of Equation 5 and is probably accurate within an order of magnitude. In most soils, the earthworms do not reach the body residue expected by EP (on average, one order of magnitude lower). Furthermore, the variability is very high for BSAF in these fieldcontaminated soils, pointing at large differences in bioavailability that are not fully explained by differences in sorption. Even though E. andrei is not a typical soil species, a pilot study with L. rubellus indicates that the two species are well correlated with regard to steady-state body residues. The different PAHs are quite similar in their behaviour, although between PAHs with similar Kow, consistent differences exist up to a factor of seven in BSAF and Koc-app. The differences between sites are larger and are consistent for all PAHs. Soils with a high content of clay and organic carbon, and a low pH and PAH level, lead to higher BSAFs (i.e. closer to EP predictions). Indications exist that the freely-dissolved pool of PAHs in soil pore water is partially depleted, and that bioaccumulation is influenced by the kinetics of PAH desorption and mass transport in soil, and equilibrium partitioning is not achieved. Similar conclusions were drawn for biodegradation of PAHs in soils [25] and accumulation in sediment amphipods [18]. It is therefore likely that the apparent elimination rate constant also reflects these processes, and that Equation 2 is not a valid representation of bioaccumulation. Furthermore, differences in feeding rates and assimilation efficiencies may have contributed to the total variation between soils. Nevertheless, risk assessment for PAHs in soils and soil-sediment mixtures may rely upon QSARs for Koc [9] and BCF [12]. Combined with the Foc of the site and the total content of PAHs, this generally seems to provide a worst-case estimate of body residues in earthworms. Nevertheless, the applicability of this approach is limited as body residues in bioassays may be overestimated by two orders of magnitude (although this does not necessarily reflect levels in earthworms under field conditions). For a better explanation of the toxicokinetics, methods need to be applied that measure or estimate the freely-dissolved concentration in pore water with greater precision, but this alone is not sufficient to predict bioavailability of PAHs. Bioavailability of PAHs appears to be a highly dynamic problem, that additionally requires quantification of the kinetics of desorption and mass transfer, as well as the influence of uptake from the gut contents.

107

Chapter 5

ACKNOWLEDGEMENTS We would like to thank the following persons for constructive comments on drafts of this manuscript: Kees van Leeuwen, Joop Hermens, Leo Posthuma, Gerard van den Berg, and two anonymous reviewers.

REFERENCES [1]

[2]

[3] [4] [5] [6] [7] [8]

[9] [10]

[11]

[12] [13]

[14] [15]

[16] [17] [18]

[19] [20]

[21]

108

Achazi RK, C Flenner, DR Livingstone, LD Peters, K Schaub and E Scheiwe (1998). Cytochrome P450 and dependent activities in unexposed and PAH-exposed terrestrial annelids. Comp. Biochem. Physiol. C 121:339-350. Belfroid AC, W Seinen, KCAM Van Gestel, JLM Hermens and KJ Van Leeuwen (1995). Modelling the accumulation of hydrophobic organic chemicals in earthworms - Application of the equilibrium partitioning theory. Environ. Sci. Poll. Res. 2:5-15. Belfroid AC, DTHM Sijm and CAM Van Gestel (1996). Bioavailability and toxicokinetics of hydrophobic aromatic compounds in benthic and terrestrial invertebrates. Environ. Rev. 4:276-299. Briggs GG and KA Lord (1983). The distribution of aldicarb and its metabolites between Lumbricus terrestris, water and soil. Pestic. Sci. 14:412-416. Burkhard LP (2000). Estimating dissolved organic carbon partition coefficients for nonionic organic chemicals. Environ. Sci. Technol. 34:4663-4667. Chiou CT, RL Malcolm, TI Brinton and DE Kile (1986). Water solubility enhancement of some organic pollutants and pesticides by dissolved humic and fulvic acids. Environ. Sci. Technol. 20:502-508. Chiou CT, SE McGroddy and DE Kile (1998). Partition characteristics of polycyclic aromatic hydrocarbons on soils and sediments. Environ. Sci. Technol. 32:264-269. Eijsackers H, CAM Van Gestel, S De Jonge, B Muijs and D Slijkerman (2001). Polycyclic aromatic hydrocarbon-polluted dredged peat sediments and earthworms: A mutual interference. Ecotoxicology 10:35-50. Gerstl Z (1990). Estimation of organic chemical sorption by soils. J. Contam. Hydrol. 6:357-375. Gustafsson Ö, F Haghseta, C Chan, J MacFarlane and PM Gschwend (1997). Quantification of the dilute sedimentary soot phase: implications for PAH speciation and bioavailability. Environ. Sci. Technol. 31:203-209. Haitzer M, BK Burnison, S Höss, W Traunspurger and CEW Steinberg (1999). Effects of quantity, quality, and contact time of dissolved organic matter on bioconcentration of benzo[a]pyrene in the nematode Caenorhabditis elegans. Environ. Toxicol. Chem. 18:459-465. Jager T (1998). Mechanistic approach for estimating bioconcentration of organic chemicals in earthworms (Oligochaeta). Environ. Toxicol. Chem. 17:2080-2090. (Chapter 3 of this thesis) Jager T, FA Antón Sánchez, B Muijs, EG Van der Velde and L Posthuma (2000). Toxicokinetics of polycyclic aromatic hydrocarbons in Eisenia andrei (Oligochaeta) using spiked soil. Environ. Toxicol. Chem. 19:953-961. (Chapter 4 of this thesis) Jager T (2003). Modelling ingestion as an exposure route for organic chemicals in earthworms (Oligochaeta). Accepted for publication in Ecotoxicology. (Chapter 8 of this thesis) Jones KC, JA Stratford, KS Waterhouse, ET Furlong, W Giger, RA Hites, C Schaffner and AE Johnston (1989). Increases in the polynuclear aromatic hydrocarbon content of an agricultural soil over the last century. Environ. Sci. Technol. 23:95-101. Kelsey JW and M Alexander (1997). Declining bioavailability and inappropriate estimation of risk of persistent compounds. Environ. Toxicol. Chem. 16:582-585. Kile DE, RL Wershaw and CT Chiou (1999). Correlation of soil and sediment organic matter polarity to aqueous sorption of nonionic compounds. Environ. Sci. Technol. 33:2053-2056. Kraaij RH, S Ciarelli, J Tolls, BJ Kater and A Belfroid (2001). Bioavailability of lab-contaminated and native polycyclic aromatic hydrocarbons to the amphipod Corophium volutator relates to chemical desorption. Environ. Toxicol. Chem. 20:1716-1724. Krauss M, W Wilcke and W Zech (2000). Availability of polycylic aromatic hydrocarbons (PAHs) and polychlorinated biphenyls (PCBs) to earthworms in urban soils. Environ. Sci. Technol. 34:4335-4340. Landrum PF, SR Nihart, BJ Eadie and LR Herche (1987). Reduction in bioavailability of organic contaminants to the ampipod Pontoporeia hoyi by dissolved organic matter of sediment interstitial waters. Environ. Toxicol. Chem. 6:11-20. Lord KA, GG Briggs, MC Neale and R Manlove (1980). Uptake of pesticides from water and soil by earthworms. Pestic. Sci. 11:401-408.

PAHs in field-polluted soils [22] Ma WC, J Immerzeel and J Bodt (1995). Earthworm and food interactions on bioaccumulation and disappearance in soil of polycyclic aromatic hydrocarbons: Studies on phenanthrene and fluoranthene. Ecotox. Environ. Saf. 32:226-232. [23] Ma WC, A Van Kleunen, J Immerzeel and PGJ De Maagd (1998). Bioaccumulation of polycyclic aromatic hydrocarbons by earthworms: assessment of equilibrium partitioning theory in in situ studies and water experiments. Environ. Toxicol. Chem. 17:1730-1737. [24] Mackay D, WY Shiu and KC Ma (1992). Illustrated handbook of physical-chemical properties and environmental fate for organic chemicals. Lewis Publishers, Boca Raton, Florida, US. [25] Mulder H, AM Breure and WH Rulkens (2001). Prediction of complete bioremediation periods for PAH soil pollutants in different physical states by mechanistic models. Chemosphere 43:1085-1094. [26] OECD (1984). Guideline for testing of chemicals no. 207. Earthworm, acute toxicity tests. Organization for Economic Cooperation and Development, Paris, France. [27] Sabljić A, H Güsten, H Verhaar and J Hermens (1995). QSAR modelling of soil sorption. Improvements and systematics of log Koc vs. log Kow correlations. Chemosphere 31:4489-4514.

109

Chapter 5

110

Evaluation of bioassays

6 Assessing Bioavailability of Organic Chemicals in Contaminated Soils, Evaluation of Bioassays with Earthworms1

Tjalling Jager, Leon van der Wal, Roel Fleuren, Arjan Barendregt, Joop Hermens Manuscript in preparation

ABSTRACT  Earthworms live in close contact with the soil and can thus be considered representative for the bioavailability of chemicals at contaminated sites. Bioavailability can either be assessed by analysing earthworms from the contaminated location, or by exposing laboratory-reared specimens to soil samples from the field (bioassays). In this study, we investigate the relevance of bioassays by using an extended bioassay design (to identify signs of depletion), and by using two species of earthworm (the standard test species Eisenia andrei and the field-relevant Aporrectodea caliginosa). Furthermore, the bioassay results are compared to body residues of worms collected from the field site: a polder in the city of Rotterdam, heavily polluted with dredge spoil from the harbour. We focussed on the following chemicals: telodrin, dieldrin, hexachlorobenzene and eight PCBs. With our bioassay design, it was shown that depletion of the bioavailable phase was unlikely, although more subtle effects could have occurred (e.g. changes in sorption during the experiments). E. andrei is a good choice for bioassays because its body residues correlate well to those in A. caliginosa, as well as to those in the field-collected worms. Nevertheless, E. andrei accumulates systematically more than the other species, and appeared to be more sensitive to the poor conditions in the soil from one of our sites.

The bioassays in this study were performed within the RIVM project 711701, on behalf of the Ministry of Housing, Spatial Planning and the Environment. Chemical analyses were performed at the IRAS, supported by the Technology Foundation STW, applied science division of NWO and the technology programme of the Ministry of Economic Affairs (project number UBI-4552), and partly by the European FP5 project Liberation under contract number: (EVK1-CT-2001-00105). 1

111

Chapter 6

INTRODUCTION Most of the seriously polluted soils in the Netherlands originate from practices between 1960 and 1980, including the dumping of chemical wastes and the disposal of dredge materials from harbours. Risk assessment for these contaminated sites is seriously hampered by a lack of quantitative knowledge about bioavailability of the pollutants. It is generally accepted that the total concentration is a poor measure for predicting accumulation and toxic effects, and bioavailability tends to decrease with increasing contact time between chemical and soil (sequestration) [1]. Earthworms are appropriate model organisms for bioavailability as they live in close contact with the soil, have a thin and permeable cuticle, and also consume large amounts of soil. However, bioassays with earthworms have several shortcomings. Firstly, bioavailability may depend on the behaviour of the organism and may thus differ between species [30], and secondly, assays are generally performed with homogenised and sieved soil samples whereas exposure in the field is more heterogeneous. The capacity of the pore-water pool for hydrophobic chemicals is very small, so desorption from the solid phases is needed to establish the observed body residues. When this desorption is slow, deviating accumulation patterns may result. To illustrate the possible effects of depletion, three extreme cases are shown in Figure 1. When desorption does not occur at all, the worms may deplete the pore-water phase. This will show up as a rapid equilibration in the first accumulation stage, and substantially less uptake in the second accumulation stage (when re-using the soil from the first stage). For example, there are indications that uptake of dieldrin was limited by chemical transport in the soil [17], and that the bioavailable phase for PAHs can be depleted [13,19]. Furthermore, the rate constant in the elimination stage will be smaller than the apparent rate constant from the accumulation stages. When desorption is slow and rate-limiting, we should observe a first rapid increase in body residues, followed by a slower rate of increase (governed by desorption). depletion

rate-limiting desorption

! "

#

!

degradation/sequestration

! " #

" #

Figure 1. Model predictions for a system where the organism is depleting the bioavailable phase (left), uptake is limited by slow desorption (middle), and with rapid degradation or sequestration in the course of the experiment (right). Three stages are shown: 1) accumulation, 2) elimination and 3) accumulation re-using the soil from stage 1 with fresh worms.

Several authors have reported peak-shaped accumulation curves [12,18,19,23], most often for PAHs. The exact cause of this pattern is unknown, but suggested explanations include the induction of active excretion by the worm [23], an increase of sorption in soil [18], or biodegradation coupled to slow desorption from organic matter [12]. When there is rapid degradation or sequestration, the bioavailable concentration will decrease during the experiment, leading to peak-shaped uptake curves. When this shape is caused by induced biotransformation, one can expect that the second accumulation stage will be identical to the first stage (as bioavailability is unaffected).

112

Evaluation of bioassays

In an attempt to deal with these shortcomings, we used an extended set-up for our bioassays. As location, we selected a polder, “De Esch”, within the city of Rotterdam in the Netherlands. This polder has served as a depot for heavily contaminated dredge spoil from the harbour of Rotterdam in the seventies. To address potential problems with depletion in laboratory experiments, we follow the accumulation in time, as well as elimination on a reference soil. Furthermore, we re-use the soil from the accumulation assays to assess possible changes in availability or depletion of the bioavailable phase (stage 3 in Fig. 1). The resulting patterns can be compared to the predicted curves of Figure 1. Besides depletion, we turn to the more practical problem of whether the standard assays are representative for the field situation and for different species. The standard test species is Eisenia andrei [21], even though its natural habitat is limited to accumulations of organic matter (like compost heaps and manure). Additionally, we therefore use a typical soil-dwelling species (Aporrectodea caliginosa). Furthermore, we collected earthworms from the field site to compare their body residues to those of the species in the laboratory experiments. In the field, exposure is different from the laboratory as the worms may avoid the most polluted spots, and problems associated with laboratory experiments do not occur (homogenisation of the soil, or depletion in the test container). In the paper by Van der Wal et al. (Chapter 7), solid-phase micro-extraction (SPME) is used as a chemical measure of availability, which is compared to the levels observed in the earthworms. Table 1. Soil characteristics of the three locations sampled, vegetation type and earthworm species. N.a. is not analysed. Genus names are abbreviated: L. = Lumbricus, Ap. = Aporrectodea, All. = Allolobophora. Soil property

soil 3

Organic matter (Fom) pH (KCl) WHC50 (g water/gdwt soil) Clay content ≤ 2 µm

11% 7.7 0.50 n.a.

soil 4 Soil properties 15% 7.5 0.53 24.3

soil 1 7.4% 7.7 0.30 n.a.

Vegetation and resident earthworm population Thistle, stinging Grassland, stinging Grassland, blackberry nettle nettle L. castaneus All. chlorotica Ap. caliginosa Earthworms (dominant) Ap. caliginosa L. castaneus L. rubellus L. rubellus Ap. longa Earthworms (few) L. rubellus Ap. caliginosa a L terrestris a The three specimens of this species showed a healthy behaviour, although their colour was remarkably pale. Vegetation type

MATERIALS AND METHODS SAMPLING The polder “De Esch” is located in the city of Rotterdam. Within this area, three sites were selected, based on previous studies by the local authorities. Site 1 is the reference site; it is located within the polder, but clean soil has been put on top. Site 3 and 4 are within the most polluted part of the polder. These sites are comparable in their metal and PAH content, but site 3 has much higher levels of PCBs and drins (especially dieldrin). Soil 113

Chapter 6

parameters and concentrations are shown in Table 1 and 2. After removing the superficial vegetation, we sampled the top soil (~20 cm) on each location. All earthworms found were collected and transferred back to the laboratory in soil from their site. Field-collected worms were allowed to evacuate their gut contents for 48 h at 10°C (comparable to soil temperatures in the field) on moist filter paper (filter paper was changed after 24 h). After depuration, the worms were frozen per two or four at –20°C in aluminium foil. The soil samples were stored in plastic containers at 4°C. The soil was sieved (4 Table 2. Soil concentrations of the three locations sampled (µg/kgdwt). Values for Kow are mm) and homogenised manually, experimental and calculated data from [6]. N.a. is although the high clay content posed not analysed, d.l. is detection limit. problems, as clay beads were remaining soil 4 soil 1 on the sieve. These clay beads were air Chemical log Kow soil 3 3 Telodrin 5.2 64 3.6 4.1⋅10 dried and ground by hand over the 4 3 Dieldrin 5.4 550 39 38⋅10 mm sieve. The water holding capacity HexaCB 5.73 360 32 0.91 (WHC) was determined for each soil. PCB 95 6.69 n.a. n.a. n.a. The soils from site 3 and 4 were already PCB 110 6.84 53 58 0.26 very close to 50% of the maximum PCB 149 7.3 120 110 0.39 WHC, so only for soil from site 1, some PCB 153 7.53 54 64 0.52 tap water was added to reach the same PCB 179 7.68 14 15 0.65 relative moisture level. The fraction PCB 138 7.45 89 92 0.52 7.83 3.3

0.88 after log-transformation), although the variation in the data from worms that were fed manure is clearly higher than in soil only. The lanthanides appear to function well as inert markers, judging from the fact that the concentration in the worms does not increase further after the retention time. After 48 h exposure and 48 h starvation on filter paper, the concentration of Lu in the worm samples is 4–11% of the concentration in worms with gut contents, and for Tm this was even lower (1–3%). The model parameters are quite accurately identified in the Bayesian fitting procedure; the results are shown in Table 3 as the highest-probability estimate with 90% probability intervals. The earthworms are clearly selecting an Lu-enriched diet in OECD medium, independent of the test temperature (Fsel is 1.6 and 1.7, somewhat less than for OM). A clear lag time can be observed before the worms start feeding on the surface, which is larger at 10°C (2.3 vs. 1.2 hr). The selectivity factor (Fsel) from soil when manure is present is lower than in the OECD-only experiment. The worms are thus selecting less Lu-enriched soil fractions when a preferred food source is available. Despite the preference for manure, this matrix makes up less than half of the total diet: 0.22 at 20°C and 0.34 at 10°C (probability intervals overlap). The retention time of the gut contents (Tret) is approximately two times higher at 10 than at 20°C (~5.5 vs. 2.9 h). The posterior distribution of retention time for the OECD-only experiments was used as prior information for the experiments with manure (see Table 2). This was helpful, as feeding activity was more variable when a food source was present in the system (the data from the manure-fed worms did not provide clear information on the retention time). As a result, the resulting posterior distribution

163

Chapter 9

for retention time from the manure data was very similar to those of the OECD-only experiments (Table 3).

OECD 10°C

concentration Lu

12 10 8

10

6 4

5

2 0

concentration Lu and Tm

OECD 20°C

15

0

10

20

30

40

50

Manure 10°C

6

0

0

10

30

40

50

Manure 20°C

10

5

20

8

4

6

3 4

2

2

1 0

0

10

20

30

40

50

0

0

time (hours)

10

20

30

40

50

time (hours)

Figure 2. Concentration of lanthanides in the earthworms (mg/kgdwt including gut contents) against time. Filled symbols are Lu, open symbols Tm from manure, the solid line is the highest probability fit. Triangles are the concentrations in worms after 48 h starvation.

DISCUSSION SELECTION, DIGESTION AND COMPACTION The dissection of Eisenia andrei turned out to be rather difficult because of the small size of the tested sub-adults. Furthermore, not all animals had material in the last part of their gut, and some of the crop contents could not be properly removed. An additional difficulty was the loss of samples at 10°C and the apparent outliers. For this reason, the discussion will focus on the results at 20°C. In general, the variation in the data is large, precluding conclusions on differences between the treatments. This variation partly reflects inter-individual variation in feeding pattern and physiology, but is also caused by the form of Equation 3. An increase in digestion does not lead to a proportional decrease of Fom in egesta because digestion also increases compaction (Eq. 2), especially when Fom of the ingesta is high. Small measurement errors can therefore be magnified in the estimates of digestion and compaction. The observed digestion efficiencies around 40% are much higher than values predicted earlier for E. andrei [15]. They are also much higher than

164

Feeding activity in artificial soil

values reported for Aporrectodea rosea (2% [4]), and are more in line with values given for A. longa and Lumbricus terrestris (30–40% [21]). These species are larger and have a much higher gut retention time (20 h for L. terrestris, [23]). Compaction is often ignored in earthworms feeding on soil [24], but was demonstrated to be up to a factor of four when worms were kept on a litter-only diet [9]. In the present study, compaction was less extreme, though still considerable (9% in OECD medium, 18% with manure). One should note that the estimates of digestion efficiency and compaction are not independent, as discussed above. The measured carbon content in the crop suggests a selectivity for OM of a factor of 2.1 on OECD medium. However, it is not clear whether the carbon measured in the crop and in the gut is derived from the diet only; it may include secretions from the organism itself (from the calciferous glands, mucus and digestive enzymes), or contamination of the samples with coelomic fluid. The exact origins of the carbon in the gut requires further elaboration, but for now we will assume that the main contribution is from the diet. GUT LOAD Weighing the faecal production proved to be very simple to perform on a routine basis although it is likely to underestimate the faecal output, as it is difficult to collect all of the faeces from the filter paper. Here, the faeces produced over 48 h are taken as the total output. Even then, the worm may not be completely clean. In most bioaccumulation studies, 24 h is taken to evacuate the gut contents. However, the data in Table 3 show that at least 6–21% of the gut contents remains after this period, and worse, that this fraction may depend on the treatment. For this reason, several authors propose methods with a longer duration [8,26], which decreases the bias from remaining gut contents but may also lead to substantial elimination of the chemical from the tissues. Stafford & McGrath [28] proposed to correct for remaining soil in the worm by measuring the acid-insoluble residue in the worm and comparing it to the soil. However, the validity of this approach is questionable, given the fact that the earthworm is selecting a particular fraction from the total soil (which may differ in insoluble residues and in chemical content from the bulk soil). The relative egested weight (Fege) is quite constant around 10% (dwt/wwt worm, Table 3). This fraction seems to be lower when food is present, but this difference may also be caused by an increased compaction in the gut (at least at 20°C, see Table 3). Hartenstein et al. [12] found a much higher effect of temperature and addition of manure on E. fetida which, given the close relation of this species with E. andrei, is remarkable. The reasons remain unclear, but may be related to differences between the studies in the quality of the soil and manure. The addition of manure allows the worms to grow considerably, but OECD soil still contains sufficient nutrition to sustain them throughout the experiment. The total amount of OM digested when fed manure can be calculated from the parameters, and is 1.4 times larger than on OECD soil only (at 20°C). However, the quality of the OM may also have led to the differences in growth. INERT MARKERS The model fits in Figure 2 clearly indicate the usefulness of the lanthanides Lu and Tm as markers of the feeding activity, and supports the assumption of plug-flow conditions in the gut. Nevertheless, some Lu and, to a lesser degree, Tm is remaining in the worm after 48 h of starvation. It is possible that this represents soil particles still left in the gut, or sorption to the outer skin of the worm, but some assimilation cannot be ruled out. Compared to simpler methods, the use of markers like Lu and Tm has the advantage that the markers are unlikely to irritate the worm (and thus influence retention), provide a dynamic picture of feeding (and are thereby able to test the plug-flow assumption, and show lag times), and allow quantification of feeding from multiple sources (here from soil 165

Chapter 9

and from manure). The Bayesian framework allows accurate parameter estimations, but the computations cannot be done by standard statistical packages, and are quite intensive, numerically. However, this framework lends itself readily to the use of existing (uncertain) information in the analysis of new data [11]. The selection of Lu from soil is less than estimated on the basis of the carbon measured in the crop contents (Fsel, Table 3). Even though the variation in both values is large enough to ignore this difference, it is conceivable that Lu is also partly sorbed to clay minerals and is therefore not entirely representative for organic matter selection (although the affinity for kaolinite clay is not very high at this pH [6]). However, if not all of the carbon in the crop is actually derived from the soil, these two figures may be more consistent. The use of a marker that is more specifically bound to organic matter may resolve this matter [20]. The retention times (Tret) observed, are quite comparable to the estimates made by Hartenstein et al. [12], also showing a clear effect of temperature. As the gut load hardly depends on temperature, the feeding rate (Eq. 6) of the worms will be nearly halved at 10°C, compared to 20°C. The manure data do not seriously affect the estimate of the retention time, although there is a slight shift to lower values. Care must be taken in applying the estimates for retention time directly to other studies. It seems likely that retention is directly related to the length of the gut. A large differences between adults and juveniles was indeed found in a polychaete [1], although Hartenstein et al. [12] showed no influence of size in E. fetida. The data collected for adult A. rosea show slightly higher retention times than juveniles, and suggest that temperature influences gut retention for juvenile worms more than for adults [4]. The data for the manure-fed worms are clearly more variable than for the worms kept in OECD medium only (Fig. 2). It is likely that the worms do not feed continuously on a mix of soil and manure but alternate between feeding on soil and manure (as indicated by the co-varying data of Lu and Tm in Fig. 2). The worms also use the soil in a different way when manure is present, as the selection factor is reduced to around unity. Possibly, the worms now feed on soil indiscriminately as a better source of nutrition is available, and the soil probably only serves to add mineral particles to the ingesta [27]. Even though the worms were acclimatised to a situation with manure on the surface, it still took several hours for them to start feeding after transfer to the spiked media (but ingestion of soil started immediately). This lag time (Tlag) was nearly twice as long at 10°C. Despite the high variation, the fraction of manure in the diet (Ff) is fixed quite accurately at 0.22 at 20°C and 0.34 at 10°C. These figures are, however, not consistent with the measured Fom in the crop contents (data not shown). If we can assume that no selection of OM from soil occurs, the Fom data in the crop indicate a fraction of manure in the diet of approximately 0.5. This discrepancy could point at the presence of carbon in the crop, derived from various secretions by the worm. On the other hand, the carbon content in the crop and in the faeces were determined after one-week exposure to manure, whereas the Tm-spiked manure was only a few days old and included a small amount of chloride (the counter-ion in the metal salt). Possibly, the chloride and the age of the manure play a role in its quality for the worm and may thus influence the diet composition. Variation in gut load and retention time with dung age were observed for L. festivus [14]. These observations stress the sensitivity of the feeding process to all kinds of factors. Care should thus be taken to perform the estimation of digestion and gut load, and the marker experiments under the exact same conditions.

166

Feeding activity in artificial soil

CONCLUSIONS In this study, the feeding activity of the compost worm Eisenia andrei is examined, and methods are provided to estimate the physiological parameters. Gut load was derived from weighing the faecal output; selection, digestion and compaction from measuring carbon in the gastro-intestinal tract. Lanthanides (Lu and Tm) are successfully used as inert markers of the feeding process and provide estimates of the gut retention time and the fraction of manure in the diet. The methods and models applied in this study may also be used for experiments with more field-relevant worm-soil combinations in ecological studies. This study clearly shows that the compost worm does not feed on soil indiscriminately but is able to select an OM-enriched diet from apparently homogeneous OECD medium. When manure is present, a large part of the diet still consists of soil particles. The gut load is not significantly affected by the presence of manure or temperature, but the gut retention time nearly doubles by a 10-degree temperature decrease. The present data may be used to aid the interpretation of routine studies with E. andrei in OECD medium (e.g. how chemical exposure via ingestion is affected by temperature and providing manure). Especially, these data can be used to parameterise bioaccumulation models that include the feeding process. However, care must be taken in using these data, as feeding activity may be influenced by subtle differences in experimental set-up (e.g. the age of the manure used as feed).

ACKNOWLEDGEMENTS The authors would like to thank the department of inorganic analytical chemistry at the RIVM (especially Rob Ritsema and Carlo Strien) for the measurements of the lanthanides and organic carbon. Furthermore, we would like to thank Lennart Weltje, Willie Peijnenburg, Kees van Leeuwen, Joop Hermens and two anonymous reviewers for valuable comments on drafts of this paper.

REFERENCES [1]

Ahrens MJ, J Hertz, EM Lamoureux, GR Lopez, AE McElroy and BJ Brownawell (2001). The effect of body size on digestive chemistry and absorption efficiencies of food and sediment-bound organic contaminants in Nereis succinea (Polychaeta). J. Exp. Mar. Biol. Ecol. 263:185-209. [2] Austreng E, T Storebakken, MS Thomassen, S Refstie and Y Thomassen (2000). Evaluation of selected trivalent metal oxides as inert markers used to estimate apparent digestibility in salmonids. Aquaculture 188:65-78. [3] Barley KP (1958). The influence of earthworms on soil fertility II. Consumption of soil and organic matter by the earthworm Allolobophora caliginosa (Savigny). Aust. J. Agric. Res. 10:179-185. [4] Bolton PJ and J Phillipson (1976). Burrowing, feeding, egestion and energy budgets of Allolobophora rosea (Savigny) (Lumbricidae). Oecologia 23:225-245. [5] Box GEP and GC Tiao (1992). Bayesian inference in statistical analysis. Wiley-Interscience, New York, US. [6] Coppin F, G Berger, A Bauer, S Castet and M Loubet (2002). Sorption of lanthanides on smectite and kaolinite. Chem. Geol. 182:57-68. [7] Curry JP and T Bolger (1984). Growth, reproduction and litter and soil consumption by Lumbricus terrestris L. in reclaimed peat. Soil Biol. Biochem. 16:253-257. [8] Denneman WD (1994). Extended water starvation; an improved method for sample preparation of lumbricidae in ecotoxicological studies. Fresenius J. Anal. Chem. 348:684-687. [9] Dickschen F and W Topp (1987). Feeding activities and assimilation efficiencies of Lumbricus rubellus (Lumbricidae) on a plant-only diet. Pedobiologia 30:31-37. [10] Ellis WC (1968). Dysprosium as an indigestible marker and its determination by radioactivation analysis. J. Agric. Food Chem. 16:220-224.

167

Chapter 9 [11] Ellison AM (1996). An introduction to Bayesian inference for ecological research and environmental decision-making. Ecological Applications 6:1036-1046. [12] Hartenstein F, E Hartenstein and R Hartenstein (1981). Gut load and transit time in the earthworm Eisenia foetida. Pedobiologia 22:5-20. [13] Helmke PA, WP Robarge, RL Korotev and PJ Schomberg (1979). Effects of soil-applied sewage sludge on concentrations of elements in earthworms. J. Environ. Qual. 8:322-327. [14] Hendriksen NB (1991). Gut load and food-retention time in the earthworms Lumbricus festivus and L. castaneus: A field study. Biol. Fertil. Soils 11:170-173. [15] Jager T (2003). Modelling ingestion as an exposure route for organic chemicals in earthworms (Oligochaeta). Accepted for publication in Ecotoxicology. (Chapter 8 of this thesis) [16] Janssen RPT, WJGM Peijnenburg, L Posthuma and MAGT Van den Hoop (1997). Equilibrium partitioning of heavy metals in Dutch field soils. I. Relationships between metal partition coefficients and soil characteristics. Environ. Toxicol. Chem. 16:2470-2478. [17] Janssen RPT, L Posthuma, R Baerselman, HA Den Hollander, RPM Van Veen and WJGM Peijnenburg (1997). Equilibrium partitioning of heavy metals in Dutch field soils. II. Prediction of metal accumulation in earthworms. Environ. Toxicol. Chem. 16:2479-2488. [18] Jones DL (1997). Trivalent metal (Cr, Y, Rh, La, Pr, Gd) sorption in two acid soils and its consequences for bioremediation. Eur. J. Soil Sci. 48:697-702. [19] Klepper O and JJM Bedaux (1997). Nonlinear parameter estimation for toxicological threshold models. Ecol. Mod. 102:315-324. [20] Kukkonen J and PF Landrum (1995). Measuring assimilation efficiencies for sediment-bound PAH and PCB congeners by benthic organisms. Aquat. Toxicol. 32:75-92. [21] Morgan JE and AJ Morgan (1992). Heavy metal concentrations in the tissues, ingesta and faeces of ecophysiologically different earthworm species. Soil Biol. Biochem. 24:1691-1697. [22] OECD (1984). Guideline for testing of chemicals no. 207. Earthworm, acute toxicity tests. Organization for Economic Cooperation and Development, Paris, France. [23] Parle JN (1963). Micro-organisms in the intestines of earthworms. J. gen. Microbiol. 31:1-11. [24] Piearce TG (1972). The calcium relations of selected Lumbricidae. J. Anim. Ecol. 41:167-188. [25] Piearce TG (1978). Gut contents of some lumbricid earthworms. Pedobiologia 18:153-157. [26] Pokarzhevskii AD, NM Van Straalen and AM Semenov (2000). Agar as a medium for removing soil from earthworm guts. Soil Biol. Biochem. 32:1315-1317. [27] Schulmann OP and AV Tiunov (1999). Leaf litter fragmentation by the earthworm Lumbricus terrestris L. Pedobiologia 43:453-458. [28] Stafford EA and SP McGrath (1986). The use of acid insoluble residue to correct for the presence of soilderived metals in the gut of earthworms used as bio-indicator organisms. Environ. Poll. 42:233-246. [29] Weltje L, H Heidenreich, W Zhu, HT Wolterbeek, S Korhammer, JJM De Goeij and B Markert (2002). Lanthanide concentrations in freshwater plants and molluscs, related to those in surface water, pore water and sediment. A case study in The Netherlands. Sci. Total Environ. 286:191-214. [30] Whalen JK, KH Paustian and RW Parmelee (1999). Simulation of growth and flux of carbon and nitrogen through earthworms. Pedobiologia 43:537-546.

168

Elucidating the exposure routes

10 Elucidating the Routes of Exposure for Organic Chemicals in the Earthworm, Eisenia andrei (Oligochaeta)1

Tjalling Jager, Roel Fleuren, Elbert Hogendoorn, Gert de Korte Manuscript submitted to Environmental Science & Technology

ABSTRACT  Earthworms take up organic compounds through their skin as well as from their food, but the quantitative contribution of each route is unclear. In this contribution, we experimentally validate an accumulation model containing a separate compartment for the gut. Uptake from the gut is modelled as passive diffusion from the dissolved phase in the gut contents. For the experiments, we exposed Eisenia andrei in artificial soil, spiked with tetrachlorobenzene, hexachlorobenzene, and PCB 153. Apart from the standard accumulation and elimination experiments, we ligatured the worm (using tissue adhesive) to prevent feeding. Model fits were good, thus supporting the validity of the model. The contribution of the gut route increased with increasing hydrophobicity of the chemical, and for PCB 153 the gut route clearly dominated. Despite the importance of the gut route, the final steady-state body residues did not exceed equilibriumpartitioning predictions by more than 25%. Rate constants for exchange across the skin and the gut wall could be separately identified. The rate constant across the skin decreases with Kow, but was higher than data derived from water-only exposure. However, the rate constant across the gut wall appeared to be largely independent of hydrophobicity.

The experimental work described in this chapter was performed at RIVM, on behalf of the Ministry of Housing, Spatial Planning and the Environment, within the RIVM project 607220.

1

169

Chapter 10

INTRODUCTION Earthworms are able to take up organic chemicals through their skin [19] as well as from their food [2]. However, the quantitative contribution of each route remains unclear. As earthworms regularly consume soil, it is difficult to study both routes in isolation in a relevant experimental set-up. Earthworms can be exposed in water alone [1], but the relevance of this design for pore-water uptake from soil is not obvious. More work has been done in this area for sediment organisms, showing that ingestion is an important pathway for very hydrophobic chemicals like pyrene and dioxins [17,18]. For earthworms, Belfroid et al. [3] predicted, based on model extrapolations, that food uptake becomes an important exposure route for very hydrophobic chemicals (log Kow > 5). It seems to be a generally held opinion that feeding on soil can lead to the invalidation of equilibrium partitioning (EP), which is why additional safety factors were prompted in European risk assessment guidelines [7]. In most case, uptake from food is modelled by simply adding uptake routes [15], but in a previous contribution [13], we proposed a more mechanistic accumulation model. Based on the work of Gobas et al. [10], the model includes a separate compartment for the gut contents, and a closed mass-balance. The mechanism for uptake from the gut is likely to be the same as for uptake across the skin, i.e. passive diffusion [10]. The validity of this assumption for earthworms was indicated, although a proper validation was impossible, because the physiological data regarding the feeding process were lacking. Most routine studies with earthworms are carried out with the compost worm (Eisenia andrei/fetida) in an artificial soil medium [25]. It is for this system that the essential feeding parameters have recently been identified [14], including gut loading, gut retention time and digestion efficiency. In this study, we set out to validate the accumulation model with three organic compounds: tetrachlorobenzene (TeCB), hexachlorobenzene (HeCB), and PCB 153, in artificial soil. A series of experiments was performed with these chemicals, starting with a straightforward accumulation and elimination phase. Subsequently, soil from these experiments was re-used to see whether the bioavailable phase had been altered, as indications of depletion have been observed [19]. Finally, an accumulation experiment was performed with worms, sealed with a tissue adhesive, thus preventing feeding. This procedure was pioneered by Vijver et al. [29], to demonstrate uptake routes for heavy metals. Other forms of ligaturing have been applied to demonstrate that pesticides are mainly taken up through the skin in water-only exposure [19], and that calcium is mainly taken up from the diet [26]. However, this study is, to our knowledge, the first to separate exposure routes for organic chemicals in a soil situation. The data from all these experiments are used together to fit the accumulation model and to identify the rate constants for uptake through the skin and from the gut. Furthermore, the model can shed light on the central questions: which exposure route dominates, and does feeding lead to body residues exceeding the predictions made by equilibrium partitioning.

EXPERIMENTAL SECTION EXPOSURE MEDIA AND SPIKING PROCEDURE Artificial soil was used for the experiments [25]. The water content was brought to 40% (weight basis, water/dry medium) with a lutetium solution (Lu, hydrated chloride salt, purity 99.9%, Alfa Aesar, Karlsruhe, Germany), to obtain a nominal concentration of 15 mg/kgdwt. The Lu is used as a non-assimilated tracer, to study the feeding activity and

170

Elucidating the exposure routes

estimate the retention time of materials in the earthworm’s gut [14]. After wetting the soil, it was thoroughly mixed and stored in closed plastic containers at 5°C for 1 wk, prior to spiking with organic chemicals. After storage, the pH (KCl) of the soil was 5.0. 1,2,3,4tetrachlorobenzene and hexachlorobenzene were obtained from Riedel de Haën, Seelze, Germany (99% purity, Pestanal®); PCB 153 was synthesised at the IRAS, Utrecht, the Netherlands (99% purity). The spiking procedure for these compounds was adapted from Northcott & Jones [24]. Because we needed to spike 4 kgdwt of medium, we had to do the procedure in steps (dilution spike). First, the chemicals needed to achieve a nominal concentration of 10 mg/kgdwt for each chemical were dissolved in 100 ml acetone (pro analysis). Next, 1 kg of wet soil was placed in a kitchen blender and the acetone solution was slowly added while mixing. Mixing continued for several minutes (stopping a few times to crush aggregates with a spatula). The spiked soil was left in a fume cabinet overnight after which the acetone, and also most of the water, had evaporated. Next, 1/5 of the spiked soil was put in the blender with 900 gwwt of uncontaminated medium and the water needed to restore the 40% water content. This was mixed for several minutes, stopping a few times to prevent the medium from overheating, and crushing aggregates. This entire procedure was followed five times until the entire medium was spiked. The moisture content was checked by oven drying at 80°C and was 41%. The fraction organic matter (Fom) in the soil was 10.5% (loss on ignition). To allow equilibration, the medium was stored at 10°C in glass jars for 1 wk before animals were introduced. One day before animals were introduced, soils were transferred to the test temperature of 20°C. TEST ANIMALS AND EXPERIMENTAL DESIGN Sub-adult earthworms (Eisenia andrei), were taken from mass cultures at the RIVM (Bilthoven, the Netherlands). The animals weighed between 200 and 300 mgwwt. First, the animals were allowed to evacuate their gut contents by keeping them on moist filter paper for 24 h at 20°C. Next, the animals were transferred to plastic containers with 175 gwwt of uncontaminated medium (four worms per container), and the containers were placed at 20°C, covered by a black plastic pot to minimise disturbance. The animals were left to acclimatise for one week under these conditions. After this, they were exposed to the chemicals in glass 1 L jars, using 250 gwwt of spiked medium and four animals per jar. For the determination of the feeding activity, one container was sacrificed at 1, 3 and 24 h. The animals were captured, rinsed with tap water, and immediately frozen at –20°C (including their gut contents). Later, the samples were freeze dried and analysed for Lu as described by Jager et al. [14], including three soil samples.

For the accumulation experiment, the worms were recaptured following exposure (0, 1, 2, 3, 5, 7, 14 and 21 d), and placed in a petri dish on moist filter paper for 24 h at 20°C. After this period, worms were packed in aluminium foil and frozen at –20°C. The petri dish and filter paper were dried at 80°C and the excreted material was gathered and weighed. Soil samples were taken and frozen in glass jars at –20°C (four samples at t = 0, and two samples at day 14 and 21). At t = 14, three additional jars were emptied and worms recaptured. The worms were transferred to 250 gwwt of uncontaminated medium and allowed to eliminate for 2, 5 and 11 d. The bioavailability of the chemicals may change during the experiment, which is why the spiked soil from the four jars emptied at t = 14 was re-used. Fresh worms were taken from the culture, placed on wet filter paper for 24 h, and subsequently on uncontaminated medium for another 24 h, and then introduced in the re-used soil. These worms were recaptured after 1, 3, 7 and 11 d. For the ligaturing experiment, worms were taken from the culture, and allowed to empty their gut for 24 h on moist filter paper. Their anterior end was ligatured using a tissue 171

Chapter 10

adhesive (Indermil™, from Loctite Ireland Ltd.) conform the procedure by Vijver et al. [29]. Gluing turned out to be difficult for this species because the worms were irritated by the procedure, excreting coelomic fluid which interfered with the setting of the glue. As exposure time, 0, 1, 2, 3, 5, 7 and 14 d were employed, using five worms per jar with 250 gwwt of soil. After exposure, the worms were recaptured and placed individually in a petri dish with moist filter paper overnight. Only the worms that did not evacuate any solids were used for chemical analysis. For t = 0, 5 worms were taken; on day 1 and 5, three worms had been successfully exposed; on day 2 only one worm (after longer periods, all worms excreted solid materials and were not analysed). ANALYSIS OF ORGANOCHLORINE COMPOUNDS For the sample pre-treatment of soil, 10 gwwt was mechanically shaken during 10 min in a glass tube with 25 ml of acetone. Next, 50 ml of light petroleum were added and the contents were mechanically shaken for 20 min. After centrifugation (5 min at 3,000 rpm), the liquid phase was transferred into a shaking funnel. The remaining part was extracted again following the same procedure and the liquid was transferred to the funnel. After the addition of 500 ml of water, the funnel was manually shaken for 1 min. The aqueous phase was discharged and the upper layer was extracted once more for 1 min with 500 ml water. The light petroleum phase was passed through a funnel with about 10 g of anhydrous sodium sulphate and concentrated to a volume of 10 ml. For the analysis of the worms, a sample of ~1 g was placed into a glass tube and weighed. After addition of 100 µl of the internal standard (~100 ng of 13C12-PCB 153), 9 ml isopropylalcohol and 10 ml cyclohexane were added and the mixture was macerated with an ultra-speed homogeniser for 2 min. Next, 10 ml of water was added and the mixture was macerated again for 1 min. The phases were separated by centrifugation for 10 min at 3,000 rpm. The upper organic layer was transferred through a funnel with sodium sulphate to a Kuderna Danish evaporation apparatus by means of a Pasteur pipette. The remaining part of the sample was macerated again for 1 min with 10 ml of a mixture of iso-propanolcyclohexane (13:87, v/v). After centrifugation, the upper layer was added to the first extract and the sodium sulphate was rinsed with 5 ml of cyclohexane; the organic layer was concentrated to 1 ml. For cleanup (fat destruction), the extract was brought onto a chromatography column filled with 0.5 g of silicagel impregnated with sulphuric acid (100 g of silica heated for 4 h at 200°C and 43.5 ml of concentrated sulphuric acid, mixed by rotating for 12 h). Next, 5 ml hexane was passed through the column and the organic solvent was collected in a calibrated glass tube and brought to a volume of 5 ml. For instrumental analysis with GC/MS operating in the electron impact (EI) mode, 1 ml of extract was transferred into an auto sampler vial and 10 µl of pentachlorobenzene (PeCB, 1 µg) was added as internal standard. Next, 1.5 µl was on-column injected into a fused silica DB-5MS capillary column, coated with 5% cross-linked 5% phenyl methyl siloxane with a length of 30 m × 0.25 mm I.D. and a film thickness of 0.25 µm. Helium was applied as carrier gas at a flow of 1 ml/min. Quantification was done using the internal standard PeCB for calibration and, in case of worm samples, the isotope PCB 153 to correct for losses during sample pre-treatment. The average recoveries performed at levels between 2–10 µg/g, ranged between 83–103% with s.d. below 9% (n = 4 for each analyte-matrix combination). THE MODEL The model is set up as a three-compartment model with a closed mass balance (Fig. 1). Diffusion (the two-way arrows) and advection (one-way arrows) are the basic transport processes. The model has been described earlier [13], and the full model formulation is placed in the appendix of this chapter. Each compartment is assumed to be well mixed 172

Elucidating the exposure routes

and of constant volume. The degradation chemicals will be taken up into the volatilization tissue of the worm from the outside diffusion Worm tissue soil as well as from the gut contents; both processes are driven by passive diffusion from the dissolved phase diffusion Soil [10]. The diffusion gradient between egestion soil or gut contents and worm tissue is defined by the concentrations in Gut contents ingestion both compartments and the partition coefficient between organic matter (OM) and earthworm tissue (Kbs in kgom/kgwwt). The Kbs can be viewed Figure 1. Schematic representation of the as the ratio of the bioconcentration accumulation model. factor (BCF in L/kgwwt) and the organic-matter specific sorption coefficient (Kom in L/kgom). The OM-worm partition coefficient is the same for soil to worm as from gut contents to worm. However, the magnitude of the gradient differs between these two uptake routes because the chemical concentration, as well as the Fom in the gut contents, differs from that in soil (due to selective feeding, compaction, and OM digestion) [10]. In this study, we chose to ignore degradation in the gut and metabolism as removal processes (which is acceptable, given our choice of test chemicals). However, the analysis results prompted us to include a firstorder loss term for the soil (kd). Supported by pilot calculations [13], instantaneous chemical equilibrium between solid and water phases is assumed. Several adaptations to the previous model are made. Firstly, compaction of the gut contents is included, meaning that the gut volume decreases as food is absorbed [14]. Also, a slightly different formulation for the Fom in the gut is taken: as average of ingesta and egesta, instead of egesta only. MODEL FITTING The model equations are implemented in matrix form in MatLab Version 6.1 (The Mathworks, Inc.) and solved with the matrix exponential function. For each chemical, we have five data sets (soil concentrations, two accumulation phases, one elimination phase, and accumulation in glued worms) that must be described by the same model, and with the same parameter values. Therefore, all datasets must be fitted simultaneously; something that is not possible with standard software tools. To this end, we define a likelihood function on the basis of the sums-of-squares (SSQ) from the model fits on each dataset, assuming normally-distributed data [4]:

l(θ |data ) ∝

5

∏ SSQ(θ ; data ) i

− ni / 2

Eq. 1

i =1

where θ is the entire set of parameters, ni is the number of points in data set i. Different likelihood functions may be multiplied, so in this way we end up with one expression for the overall likelihood of the model parameters given all five datasets. For the gluing experiment, the number of worms that were successfully glued was taken as a weight coefficient in the SSQ. The overall log-likelihood function is maximised by a Nelder-Mead simplex search in MatLab, yielding maximum-likelihood (ML) estimates of the parameters. The likelihood function is also used to construct confidence intervals by calculating the profile likelihood [22].

173

Chapter 10 Table 1. Fixed parameters for the feeding process [14]. Symbol Fege Frem Fsel Fdig Fcom Tret Fsol

Description Egested faeces as fraction of body weight Fraction of gut contents remaining after 24 h depuration Selectivity for OM in diet Digestion efficiency of OM Factor by which gut contents are compacted Gut retention time Fraction solids in worm

Unit kgdwt/kgwwt kgdwt/kgdwt — — — h kgdwt/kgwwt

Value 0.12 0.056 2.1 0.35 1.09 2.9 0.14

The worms had been depurated for 24 h, which is insufficient to remove all of the gut contents, but a longer depuration could lead to bias as also chemicals will be lost from the tissues. However, using an estimate of the remaining fraction (Frem, Table 1), we corrected the modelled concentration for the remaining gut contents (see appendix). The MLestimates are used to estimate the observed biota-soil accumulation factor (BSAF), the net chemical assimilation efficiency (FAE), and the deviation from EP. The BSAF (in kgom/kgwwt) is calculated from the modelled body residue in the worm (Cb in mg/kgwwt) at t = 21, the concentration in the soil (Cs in mg/kgdwt) and the Fom in soil:

BSAF =

C b 21 Fom−s C s 21

Eq. 2

The assimilation efficiency can be calculated from the uptake flux from the gut contents into the worm tissues and the chemical flux with feeding [13], and the deviation from EP is made by comparing Cb21 to the EP estimate based on the concentration in soil and the gut (Cg):

C b ( EP , soil ) =

C b (EP , gut ) =

K bs C s 21 Fom−s K bsC g 21 Fom− g

Eq. 3

Eq. 4

The worm can come to equilibrium with the soil or the gut contents, or end up somewhere in between (in a non-equilibrium steady state). Where the body residues will end up between these extremes will especially depend on the kinetics of the different transfer processes, and is not simply the sum of the individual uptake routes. To obtain a confidence interval on these derived results, we applied a random parameter search. Parameters were randomly drawn from log-uniform distributions. When the likelihood of these parameters (Eq. 1) was not significantly lower than the ML estimate, the parameter combination was stored. Random parameters were drawn until 200 acceptable parameter combinations were obtained. For each parameter combination in this set, the BSAF, AE and the deviation from EP were calculated. The maximum and minimum values serve as confidence intervals.

174

Elucidating the exposure routes

RESULTS AND DISCUSSION concentration Lu (mg/kg dwt)

GENERAL OBSERVATIONS We want to apply the parameter values 14 for the feeding activity, as observed 12 previously (Table 1), to our current 10 experiments. For this reason, we first checked whether these values are 8 indeed representative for this study. The 6 Lu concentrations determined in the best fit worms (with gut contents) 4 Lu corresponded to the expected levels 90% probability 2 from the previous study, corrected for the different actual Lu concentration in 0 0 5 10 15 20 25 the medium (Fig. 2). Faeces collected time (hours) after 24 h depuration were dried and weighed. The relative egested weight Figure 2. Lutetium concentration versus time (Fege) from the previous experiments was and the model fit derived earlier (Table 1) with 90% probability intervals. calculated after 48 h depuration. Using the estimate for the remaining fraction after 24 h (Frem, Table 1), we obtain an estimate for this experiment of 0.13 gdwt/gwwt worm, which compares well to that determined previously (Table 1). The worms increased approximately 10% in weight during the 21 d exposure, but the potential effects on accumulation kinetics were ignored. Table 2. Parameter estimates and derived measures, resulting from the model fits. Maximumlikelihood estimates with 95% likelihood-based confidence intervals. Kws is the OM-worm partition coefficient. HeCB

PCB 153

Chemical properties and QSAR estimations 4.64 [6] 5.73 [6] kgom/kgwwt 0.12 0.20

6.92 [11] 0.33

Parameter Log Kow Kbs estimated from QSARs [12,28]

Unit

Kbs (1st accumulation exp.) Kbs (2nd accumulation exp.) Rate constant skin (ke) Rate constant gut wall (kg) Degradation and/or volatilisation rate (kd)

TeCB

Estimated model parameters kgom/kgwwt 0.11 (0.077,0.13) 0.20 (0.19,0.22) kgom/kgwwt —a 0.16 (0.15,0.18) d-1 0.74 (0.42,2.1) 0.30 (0.22,0.43) d-1 0.27 (–∞,+∞) 0.43 (0.18,0.72) 0.0065 0.0055 d-1 (0.0051,0.0077) (0.0051,0.0058)

Derived results BSAF at t = 21 kgom/kgwwt 0.12 (0.072,0.13) Max. assimilation efficiency % 10 (2.6 10-4,50) Deviation from EP with soil % 7.3 (0,18) Deviation from EP with gut % –14 (–21,0) a Assumed the same value as in the first accumulation phase.

0.23 (0.21,0.26) 26 (0.19,42) 13 (1.0,23) –7.1 (–21, –1.9)

0.24 (0.23,0.27) 0.16 (0.13,0.21) 0.027 (0.020,0.037) 0.16 (0.13,0.20) 0.0062 (0.0058,0.0066) 0.30 (0.25,0.36) 16 (12,22) 24 (18,27) –3.3 (–6.4, –0.3)

The initial concentration in the soil was 5.8 mg/kgdwt for TeCB, 7.2 for HeCB, and 6.8 for PCB 153. These were lower than the nominal 10 mg/kgdwt. However, the homogeneity of the spiking was acceptable as the standard deviations were 4–5% (n = 4) of the average

175

Chapter 10

value. The soil concentrations appeared to decrease some 10% in the course of the exposure experiment. It is striking that the rate constants for this disappearance (kd, Table 2) are practically identical for all three chemicals, and could perhaps reflect the formation of resistant fractions [27]. The accumulation model can adequately describe the data from the different experiments simultaneously (Fig. 3), with only five free parameters (see Table 2), an average of one parameter per data set. Only the data for TeCB are rather scattered (especially in the first accumulation phase). In most cases, the parameters are accurately identified by the data, evidenced from the tight 95% confidence intervals (Table 2). Only the poor fit for TeCB results in ill-defined estimates. The good fit supports the basic assumption that gut uptake is indeed mediated through passive diffusion from the dissolved phase in the gut [10]. Ligaturing the earthworm allows to isolate exchange across the skin in a soil situation. Even though the worms survive for several days without problems, the treatment is stressful. As a consequence, their behaviour will be affected, which may also bias chemical uptake. However, the difference between glued and intact worms is remarkably small for both chlorobenzenes, providing some reassurance that the earthworms are sufficiently active to take up the chemicals through the skin. EXPOSURE ROUTES AND ASSIMILATION EFFICIENCY The estimated uptake fluxes clearly show an increase of the importance of the gut in the total uptake with increasing hydrophobicity (Fig. 3). Furthermore, the net fluxes decrease in time as the animal is approaching steady state. Figure 3 shows that for TeCB, the skin is probably the most important route, although the large confidence intervals preclude firm conclusions. For HeCB, both routes are approximately equally important, but for PCB 153, the gut route is truly the dominant exposure route. Although we do not agree with the approach taken by Belfroid et al. [3], our conclusion is comparable: the gut begins to become an important route for chemicals with a log Kow above approximately 5, and dominates above 6. This is also reflected in the deviation from EP with soil, which increases with Kow from 7 to 24%. However, the confidence intervals are quite large, and only the deviation for PCB is clearly higher than 0%. Similarly, the deviation from EP with the gut contents (Eq. 4) decreases with increasing Kow, so that for PCB 153 the tissue residues are nearly in equilibrium with the gut contents. The parameter estimates also allow calculation of the net chemical assimilation efficiency (FAE) from the estimated fluxes in the model. As shown previously [13], the net efficiency depends on time, and therefore, only the maximum is given here. No general trend with Kow is observed, although a trend may be obscured by the large confidence intervals. PARTITION COEFFICIENTS The OM-worm partition coefficient (Kbs) was accurately predicted by the QSARs for worm-water [12] and organic carbon-water partitioning [28], assuming a factor of 1.7 between organic carbon and organic matter (Table 2). Note that Kbs in this case is a model parameter; the final body residue in steady state depends on the kinetics, and lies in between the extremes: equilibrium with the soil or the gut contents (Eq. 3 and 4). The actual BSAF is thus a secondary result (Eq. 2), and is slightly higher than the Kbs, indicating that the concentrations in the worms exceeds equilibrium with the soil. The Kbs in the re-used soil was generally somewhat lower than the value in the initial accumulation phase. Only for TeCB, a higher Kbs was estimated in the re-used soil. As we judged this behaviour to be an artefact (because of the scatter in the initial accumulation phase), we chose to use only one Kbs for both accumulation experiments with this chemical.

176

Elucidating the exposure routes

A

7

B

18 16

6

14

concentration in worm

5

12

4

10

3

8 6

2

4 1 0

2 0

5

10

15

20

25

0

0

5

10

15

20

25

C

18 16 14

accumulation 1

12

elimination

10

accumulation 2 glued worms model fit EP prediction

8 6 4 2 0

0

5

10

15

20

25

time (days)

A

10

B

10

8

8

6 6

4

4

2 0

2

uptake flux

-2 0

-4 -6

0

1

2

3

4

5

6

7

8

9

10

-2

0

1

2

3

4

5

6

7

8

9

10

C 3

flux across skin flux across gut wall

2

1

0

0

1

2

3

4

5

6

7

8

9

10

time (days) Figure 3. Model fits for the different accumulation and elimination experiments (top, mg/kgwwt), and the modelled uptake fluxes (µg/d) from soil and gut contents, with 95% confidence intervals (bottom). The EP prediction marks the estimated body residue for a worm in equilibrium with the soil. A = TeCB, B = HeCB, C = PCB 153.

177

Chapter 10

For PCB 153, we saw a clear decrease in Kbs for the re-used soil, showing that bioavailability had declined over two weeks. We first assumed that this difference was caused by depletion of the pore-water pool for PCB, coupled with slow desorption, but our attempts to expand the model to include this process failed to describe the data. This is not so strange as equilibration between water and solids is a matter of hours for many chemicals, at least in sediment [5]. Sequestration remains a possible explanation. Even though the soil was allowed to equilibrate for one week, this may be insufficient to achieve equilibrium between the different soil phases (as also indicated by the apparent loss of the chemicals from the soil). In the model, we assume the same partition coefficient (Kbs) for exchange from soil to worm (via pore water) as from gut contents to worm (via the gut fluid). However, gut fluids differ from pore water in that they include secretions from the worm to aid digestion. Mayer et al. [21] have shown that, in marine deposit feeders, these secretions include surfactants that also act to solubilise organic contaminants above levels expected in sea water. It may appear plausible that the action of gut fluids invalidates the hypothesis of passive diffusion via a water phase. However, we do not believe this to be the case, as there is strong experimental support for a diffusion-driven uptake. Even though micelles play an important role in the transport, the available evidence supports a diffusion-driven process in mammals [23], as well as in isolated gut segments of catfish [30]. Furthermore, secretions with surface active-properties will increase the dissolved chemical concentration, but not the fugacity gradient. Because the gut fluid becomes less polar than water due to these secretions, the chemical has less urge to flee to the earthworm’s tissues. The gut fluid-worm partition coefficient is decreased by the same factor as the solubility is increased, leading to the same net uptake. The same conclusion was reached by Lu et al. [20]. We therefore believe that gut secretions act mainly on the gut rate constant (kg), and not on the OM-worm partition coefficient (Kbs).

rate constant (1/d)

RATE CONSTANTS 1 The rate constants for 10 rate constant skin (ke) exchange across the skin (ke) rate constant gut (kg) and the gut wall (kg) could be regression ke regression data [1] separately identified (Table (elimination in water) 0 2). Only for TeCB, a 10 confidence interval for the gut rate constant cannot be made. There is a value that -1 has the highest likelihood, 10 but very high and very low values are not significantly worse (apparently, exchange -2 across the skin is so rapid 10 4 4.5 5 5.5 6 6.5 7 7.5 that the gut route cannot log Kow compete). The rate constants for chemical exchange are Figure 4. Rate constants for exchange across the skin and shown in Figure 4, which also across the gut wall versus log Kow. Error bars represent 95% gives a fit on data for the likelihood-based confidence intervals. The broken line elimination from worms in a indicates elimination rates for earthworms in water only [1]. water-only situation [1]. The skin rate constant (ke) for TeCB is quite comparable to the water-only data, but our value for HeCB is much higher. This is striking, given the fact that in a water-only situation, the 178

Elucidating the exposure routes

contact between worm and water is likely to be more intensive. Furthermore, rate constants for passive-diffusive exchange are expected to decrease with Kow with a slope close to one (on log scale) in this Kow range [9]. However, for the rate constant across the skin, we find a slope that is less steep than for water-only exposure (–0.63 vs. –1.3). An explanation may be derived from the earthworm’s physiology. In a soil situation, earthworms lose 10–20% of their body weight in moisture each day due to their respiratory system, which requires the maintenance of a moist outer surface [16]. In wateronly exposure, the water loss will likely be much less. These losses need to be replenished, requiring water transport across the skin (also advectively transporting the chemical). This process can explain a higher exchange rate in soil than in water, and the different relationship with Kow. The rate constant across the gut wall is quite constant over the studied Kow range (Fig. 4), although this conclusion is very preliminary as a reliable estimate for TeCB cannot be made. Again, what we expected was a decrease with a slope around unity. The slope of –1 was predicted because for very hydrophobic compounds, the diffusion across a stagnant water layer is rate-limiting [9]. Absorption of lipids in the gut is a passive diffusion process, and the rate of uptake is primarily determined by transport across the unstirred water layer, enhanced by bile salts [8]. It is conceivable that the surfactants in the gut facilitate crossing the boundary layer. CONSEQUENCES FOR EP IN RISK ASSESSMENT To our knowledge, this is the first time that, in earthworms, the uptake of organic chemicals from soil through the skin has been separated from uptake resulting from feeding on soil particles. The importance of the gut route increases with increasing hydrophobicity, and very hydrophobic chemicals (log Kow > 6) will mainly be absorbed from the gut contents. There is some additional uptake as a result of feeding on soil, but the deviation from EP with the soil is less than a factor of 1.3, which is well within the accuracy of risk assessment applications. The general fear that feeding leads to the invalidation of EP is thus unwarranted. The model presented here can adequately describe the experimental data, and the diffusion mechanism for gut uptake [10] is thus supported by this study. However, in view of the small deviations, risk assessment can rely on EP, and specific modelling of the gut compartment is usually not necessary. Nevertheless, this model may be useful for specific cases, especially when the worms are not feeding on soil alone, but on a diet that is specifically contaminated (e.g. manure from farm animals treated with pharmaceuticals, or pesticide residues in leaf litter). In these situations, soil concentrations along with EP are insufficient to predict body residues.

ACKNOWLEDGEMENTS We would like to thank the laboratory for inorganic chemistry at the RIVM for performing the lutetium measurements in soil and worms, and the Institute for Risk Assessment Sciences (IRAS, Utrecht) for kindly providing the organic chemicals. Furthermore, we would like to thank Martina Vijver for demonstrating the ligaturing procedure, Rob Baerselman for support in the experiments, and Willie Peijnenburg and Joop Hermens for reviewing drafts of this manuscript.

179

Chapter 10

REFERENCES [1]

[2]

[3]

[4] [5]

[6]

[7]

[8] [9] [10] [11] [12] [13] [14] [15] [16] [17]

[18]

[19] [20] [21]

[22] [23] [24] [25]

180

Belfroid A, A Van Wezel, M Sikkenk, K Van Gestel, W Seinen and J Hermens (1993). The toxicokinetic behavior of chlorobenzenes in earthworms (Eisenia andrei): experiments in water. Ecotox. Environ. Saf. 25:154-165. Belfroid A, J Meiling, D Sijm, J Hermens, W Seinen and K Van Gestel (1994). Uptake of hydrophobic halogenated aromatic hydrocarbons from food by earthworms (Eisenia andrei). Arch. Environ. Contam. Toxicol. 27:260-265. Belfroid AC, W Seinen, KCAM Van Gestel, JLM Hermens and KJ Van Leeuwen (1995). Modelling the accumulation of hydrophobic organic chemicals in earthworms - Application of the equilibrium partitioning theory. Environ. Sci. Poll. Res. 2:5-15. Box GEP and GC Tiao (1992). Bayesian inference in statistical analysis. Wiley-Interscience, New York, US. Cornelissen G, PCM Van Noort and HAJ Govers (1997). Desorption kinetics of chlorobenzenes, polycyclic aromatic hydrocarbons, and polychlorinated biphenyls: sediment extraction with Tenax® and effects of contact time and solute hydrophobicity. Environ. Toxicol. Chem. 16:1351-1357. De Bruijn J, F Busser, W Seinen and J Hermens (1989). Determination of octanol/water partition coefficients for hydrophobic organic chemicals with the "slow-stirring" method. Environ. Toxicol. Chem. 8:499-512. EC (1996). Technical Guidance Documents in support of Directive 93/67/EEC on risk assessment of new notified substances and Regulation (EC) No. 1488/94 on risk assessment of existing substances (Parts I, II, III and IV). EC catalogue numbers CR-48-96-001, 002, 003, 004-EN-C. Office for Official Publications of the European Community, 2 rue Mercier, L-2965 Luxembourg, Luxembourg. Friedman HI and B Nylund (1980). Intestinal fat digestion, absorption, and transport. A review. Am. J. Clin. Nutr. 33:1108-1139. Gobas FAPC, A Opperhuizen and O Hutzinger (1986). Bioconcentration of hydrophobic chemicals in fish: relationship with membrane permeation. Environ. Toxicol. Chem. 5:637-646. Gobas FAPC, JR McCorquodale and GD Haffner (1993). Intestinal absorption and biomagnification of organochlorines. Environ. Toxicol. Chem. 12:567-576. Hawker DW and DW Connell (1988). Octanol-water partition coefficients of polychlorinated biphenyl congeners. Environ. Sci. Technol. 22:382-387. Jager T (1998). Mechanistic approach for estimating bioconcentration of organic chemicals in earthworms (Oligochaeta). Environ. Toxicol. Chem. 17:2080-2090. (Chapter 3 of this thesis) Jager T (2003). Modelling ingestion as an exposure route for organic chemicals in earthworms (Oligochaeta). Accepted for publication in Ecotoxicology. (Chapter 8 of this thesis) Jager T, RHLJ Fleuren, W Roelofs and AC De Groot (2003). Feeding activity of the earthworm Eisenia andrei in artificial soil. Soil Biol. Biochem. 35:313-322. (Chapter 9 of this thesis) Landrum PF, H Lee and MJ Lydy (1992). Toxicokinetics in aquatic systems: model comparison and use in hazard assessment. Environ. Toxicol. Chem. 11:1709-1725. Lee KE (1985). Earthworms. Their ecology and relationships with soils and land use. Academic Press, Sydney, Australia. Leppänen MT and JVK Kukkonen (1998). Relative importance of ingested sediment and pore water as bioaccumulation routes for pyrene to oligochaete (Lumbriculus variegatus, Müller). Environ. Sci. Technol. 32:1503-1508. Loonen H, DCG Muir, JR Parsons and HAJ Govers (1997). Bioaccumulation of polychlorinated dibenzo-p-dioxins in sediment by oligochaetes: influence of exposure pathway and contact time. Environ. Toxicol. Chem. 16:1518-1525. Lord KA, GG Briggs, MC Neale and R Manlove (1980). Uptake of pesticides from water and soil by earthworms. Pestic. Sci. 11:401-408. Lu X, DD Reible, JW Fleeger and Y Chai (2003). Bioavailability of desorption-resistant phenanthrene to the oligochaete Ilyodrilus templetoni. Environ. Toxicol. Chem. 22:153-160. Mayer LM, Z Chen, RH Findlay, J Fang, S Sampson, RFL Self, PA Jumars, C Quetel and OFX Donard (1996). Bioavailability of sedimentary contaminants subject to deposit-feeder digestion. Environ. Sci. Technol. 30:2641-2645. Meeker WQ and LA Escobar (1995). Teaching about approximate confidence regions based on maximum likelihood estimation. Am. Stat. 49:48-53. Moser GA and MS McLachlan (2002). Modeling digestive tract absorption and desorption of lipophilic organic contaminants in humans. Environ. Sci. Technol. 36:3318-3325. Northcott GL and KC Jones (2000). Developing a standard spiking procedure for the introduction of hydrophobic organic compounds into field-wet soil. Environ. Toxicol. Chem. 19:2409-2417. OECD (1984). Guideline for testing of chemicals no. 207. Earthworm, acute toxicity tests. Organization for Economic Cooperation and Development, Paris, France.

Elucidating the exposure routes [26] Piearce TG (1972). The calcium relations of selected Lumbricidae. J. Anim. Ecol. 41:167-188. [27] Pignatello JJ and B Xing (1996). Mechanisms of slow sorption of organic chemicals to natural particles. Environ. Sci. Technol. 30:1-11. [28] Sabljić A, H Güsten, H Verhaar and J Hermens (1995). QSAR modelling of soil sorption. Improvements and systematics of log Koc vs. log Kow correlations. Chemosphere 31:4489-4514. [29] Vijver MG, JPM Vink, CJH Miermans and CAM Van Gestel (2003). Oral sealing using glue: a new method to distinguish between intestinal and dermal uptake of metals in earthworms. Soil Biol. Biochem. 35:125-132. [30] Weber LP and RP Lanno (2001). Effect of bile salts, lipid, and humic acids on absorption of benzo[a]pyrene by isolated channel catfish (Ictalurus punctatus) intestine segments. Environ. Toxicol. Chem. 20:1117-1124.

181

Chapter 10

APPENDIX: MODEL EQUATIONS DIFFERENTIAL EQUATIONS ON THE BASIS OF NET FLUXES (dN) The basis of the model is formed by the mass-balance equations for the three compartments, written as function of the net fluxes (dN, shorthand for dN/dt).

Chemical in the earthworm’s tissue:

Wb

Chemical in the earthworm’s gut:

Wg

Chemical in the soil:

Ws

dC b (t ) = dN u−s + dN u− g dt dC g (t ) dt

= dN ing − dN u− g − dN ege

dC s (t ) = dN ege − dN ing − dN u−s − dN d dt

NET FLUXES (mg/d) The fluxes for uptake and elimination from the worm’s tissues are described as passive diffusion processes. The concentration gradient is given in parentheses as the difference between the equilibrium tissue concentration and the actual tissue concentration.

uptake/elimination across the skin:

  K dN u−s = k e W b  C s bs − C b   Fom−s 

uptake/elimination from gut contents:

  K dN u− g = k g W b  C g bs − C b    Fom− g  

Note that the equation for the gut is quite comparable to that for skin uptake. The rate constant is a different one, and so is the equilibrium concentration (calculated from the concentration in the gut and the Fom in the gut contents). The ingestion and egestion fluxes are modelled as advective processes, determined principally from the feeding rate (Qs). The ingestion flux is also determined by the selectivity of the worm for OM in its diet (assuming that all of the chemical is sorbed to OM). The egestion rate is smaller than the feeding rate due to compaction of the gut contents (Fcom). chemical flux from ingestion of soil:

dN ing = Qs C s Fsel

Qs Cg Fcom The degradation/volatilisation flux is a straightforward first-order decay.

chemical flux from egestion:

dN ege =

Flux for degradation and/or volatilisation: dN d = k d W s C s SORPTION AND FEEDING RATES The fraction organic matter (OM) in the ingested material is calculated from the Fom in the soil and the selectivity factor of the worm. The fraction OM in the egested material is calculated from the Fom in the ingesta: digestion decreases the fraction OM, but compaction increases this fraction.

182

Elucidating the exposure routes

Fraction OM in ingesta:

Fom−ing = Fom−s Fsel

Fraction OM in egesta:

Fom− ege = Fom−ing Fcom (1 − Fdig )

Compaction is defined as the ratio of the ingestion and egestion fluxes and can be calculated as (assuming that the weight decrease of the gut contents is solely due to digestion of OM): Fraction compaction of gut contents:

Fcom =

1 1 − Fdig Fom−ing

The fraction organic matter in the gut contents is calculated as the average of the values in ingesta and egesta.

Fom− g =

Fraction OM in the gut:

Fom−ing + Fom−ege 2

The weight of the egested materials can be measured when worms are depurating on filter paper. It is expressed as fraction of the empty wet weight of the worm. The ingested fraction is calculated by assuming that the gut contents have decreased in weight due to the digestion of OM (this decrease is defined as compaction). Ingested weight as fraction of worm wt.:

Fing = Fege Fcom

Feeding rates on soil:

Qs =

Weight of the gut (avg. of ingesta and egesta):

W g = Wb

Fing Wb Tret Fing + Fege 2

ADDITIONAL EQUATIONS The concentration in the worm that is actually measured also contains a contribution from the gut contents that remain after 24 h depuration. As the gut concentration is also modelled, the total predicted concentration in the depurated worm becomes:

C b (measured ) =

C b Wb + C g W g Frem Wb + W g Frem ( 1 + Fw )

The net chemical-assimilation efficiency is given by the ratio of the uptake flux from the gut into the tissue and the ingestion flux:

FAE (t ) =

dN u− g dN ing

183

Chapter 10

EXPLANATION OF SYMBOLS AND SUBSCRIPTS Sym. Subscript dN u-s u-g ing ege d C b s/g W b s/g Q s K bs k e/g d F ege com dig ing sel om-s/g om-ing/ege rem w AE T ret

184

Description chemical flux with uptake from soil across the skin (mg/d) chemical flux with uptake from the gut across the gut wall (mg/d) chemical flux with ingestion from soil (mg/d) chemical flux with egestion from the gut (mg/d) chemical flux with degradation from soil (mg/d) concentration of chemical in the earthworm (mg/kgwwt) concentration of chemical in soil and gut contents (mg/kgdwt) total weight of earthworm (kgwwt) weight of soil and gut contents in the system (kgdwt) rate of feeding on soil solids (kgdwt/d) worm-soil OM partition coefficient (kgom/kgwwt) rate constant for exchange across skin and gut wall (d-1) rate constant for desorption and volatilisation in soil (d-1) faecal weight as fraction of worm weight (kgdwt/kgwwt) fraction compaction of gut contents during digestion (—) fraction of OM digested during gut passage (—) ingested weight as fraction of worm weight (kgdwt/kgwwt) enrichment of the diet with OM compared to soil (—) fraction OM in soil solids or gut solids (kgom/kgdwt) fraction OM in ingesta or egesta (kgom/kgdwt) fraction of the gut contents remaining after 24 h starvation (—) fraction water in soil (L/kgdwt) chemical assimilation efficiency (—) retention time of gut contents in the GIT (d)

Elucidating the exposure routes

185

Summary and discussion

11 Summary and General Discussion

ABSTRACT  This chapter provides a summary and general discussion of this thesis. There is no detailed discussion of various aspects in relation to the literature, as this has already been done in the previous chapters. This chapter contains a more general overview of the work, main findings, recommendation and open questions. In conclusion, this thesis shows that equilibrium partitioning is valid, as long as good estimated or measured pore-water concentrations are available. In most cases, EP based on hydrophobicity (Kow) will lead to a worst-case estimate, and can in this way be applied in risk assessment. However, considerable uncertainty remains about its applicability to “difficult” compounds, like surfactants, pharmaceuticals and polar chemicals. Feeding is an important exposure route for very hydrophobic compounds (log Kow > 5 or 6), but the deviations from EP are small. Based on the work in this thesis, body residues are not expected to exceed the EP predictions by more than a factor of two for any species of earthworm or any soil type. This also implies that the additional safety factor of 10, applied in EU risk assessment guidelines, is unsupported. Explicitly accounting for this route is thus only relevant for situations in which the food source of the worm (e.g. dung or litter) is specifically contaminated. Advice is given for proper design of bioassays and dealing with bioavailability issues in risk assessment. Finally, directions for possible further research are identified.

187

Chapter 11

VERIFICATION OF EQUILIBRIUM PARTITIONING Uptake of organic chemicals into earthworms can be described as a hydrophobic partitioning between the (pore) water and the internal phases of the worm (lipids and water). This view of the bioaccumulation process provides an excellent description of water-only exposure (Chapter 3), indicating that octanol is a good substitute for earthworm lipids. When earthworms are exposed in soil, the body residue is a combination of two partitioning processes: sorption (determined by the sorption coefficient) and bioconcentration (determined by the bioconcentration factor; BCF). This is the basic assumption of equilibrium partitioning (EP). When sorption is estimated from hydrophobicity (Kow), the measured body residues tend to be similar to, or lower than, predicted by EP (Chapters 3–7, 10). This shows that EP can be taken as an estimate for the upper limit of the body residues. Because the measured body residues are usually below the EP predictions, there is little room for additional uptake from feeding. This is confirmed by dedicated experiments (discussed in more detail later in this chapter). Therefore, the general applicability of EP for the uptake of neutral organic chemicals in earthworms is supported by this thesis, as long as good estimates of the dissolved concentration are available. Furthermore, the deviations for ionisable chemicals like chlorophenols, are, under certain conditions, not very large (Chapter 3). The key uncertainty in the verification of EP lies most likely with the estimation of the actual sorption (see next section), and not in the estimation of BCF. As good measurements of the actual dissolved concentration in pore water are usually not available, a definite verification uptake/elimination of EP in the broad sense (see Chapter 1) cannot be provided. Any deviation from EP cannot be properly assessed without measured dissolved concentrations. The total concentration in pore water also fails as a good measure of the exposure level, because a large part of the chemical pool may be associated with dissolved sorption/desorption organic matter (Chapter 5), especially for very hydrophobic chemicals.

worm worm

water water

soil soil

It is striking that the relationship between BCF and Kow apparently remains linear (on log scale), even for extremely hydrophobic chemicals (Chapter 3, 5, 7, 10). This is in marked contrast with accumulation studies in fish that reveal a general loss of linearity at log Kow > 6 (e.g. [23]). Various reasons are given for this loss of linearity, including physiological ones (differences between octanol and lipids, membrane permeability), kinetic ones (biotransformation and growth are more influential as passive excretion decreases with Kow), and experimental ones (lack of equilibrium, and adsorption to glass walls and suspended materials). The experience with earthworms in this work indicates that either fish gills are different from earthworm membranes, or that for fish the explanation of the loss of linearity must be found in kinetic and experimental considerations. In earthworms, the kinetic problems are usually less important, because biotransformation does not play a large role, growth is often slow in experiments, and the presence of soil usually provides a sufficiently large buffer for the pore-water concentration. A huge potential problem for BCF-Kow relationship is the fact that most of the work has been performed on simple, neutral, organic hydrophobic chemicals; usually (chlorinated) hydrocarbons. Most chemicals that are newly introduced on the market are highly-specific chemicals, computer-engineered for specific properties, and often containing unusual molecular groups. It is currently largely uncertain to what extent the behaviour of these chemicals 188

Summary and discussion

can be described on the basis of hydrophobic partitioning. For polar chemicals, other measures of hydrophobicity than Kow may be appropriate to model bioconcentration, such as phospholipid-water partitioning [10,32], while for surfactants, the critical micelle concentration was proposed [30]. These “difficult” compounds not only pose problems for BCF estimation, but also for predicting sorption in soil. For example, veterinary pharmaceuticals sorb more heavily to soil than expected from their hydrophobicity, and normalisation to organic carbon is not helpful in reducing the variability between soils [31]. Furthermore, the general use of Kow as the only parameter for estimating sorption has been questioned, as partitioning will always result from various interactions in soil [14]. This also implies that the role of other soil properties (pH, cation-exchange capacity, clay content, etc.) needs to be investigated, and if possible, quantified. This is still a major problem in environmental chemistry.

NON-EQUILIBRIUM CONDITIONS Even though EP can thus be considered as valid for earthworms, estimation of pore-water concentrations remains a bottleneck. Sequestration (or “ageing”) is probably the most important factor in explaining the lower body residues in field-polluted soils (see Chapters 3, 5, 6). As the contact time between the chemicals and the soil increases, so does the sorption. Often, three distinct sorption phases can be distinguished [29], and the most rapid phase seems to be involved in the exchange with the pore water (i.e. EP). However, there appear to be differences in sequestration status between soils (Chapter 5), and between chemicals (Chapter 6), thereby making a general prediction impossible. Using estimation routines for sorption and BCF may lead to overprediction of body residues by a factor of 1–100. Therefore, a more precise prediction of bioavailability will have to rely on direct measurements of dissolved concentrations (Chapter 5), determination of rapidly desorbing fractions (e.g. with Tenax [6]), or biomimetic extractions (SPME, Chapter 7). In Chapter 7, it is clearly shown that EP is valid up to a log Kow of at least 8, as long as measured freely-dissolved concentrations are available. The same conclusion was reached for sediment organisms [15]. Apart from the poor predictability of sorption in the field, problems may occur uptake/elimination because the capacity of the pore water for hydrophobic chemicals is small, compared to the capacity of the soil and water degradation the worm. The pore-water pool is water therefore insufficient to explain the body residues observed in earthworms, and desorption from the solid phase is sorption/desorption required. The small capacity of this pool may lead to problems when the sequestration replenishment is slow compared to the uptake into the worm and possible degradation (see Chapter 2). There are indications in the literature that replenishment of the pore-water pool can be rate limiting for uptake into organisms [19,21], but the direct evidence is limited. In this thesis, we have probably seen the limits of the pore-water pool. There were indications of depletion (Chapter 5), and also several cases where it was likely that the bioavailability changed during the course of the experiment, leading to peak-

worm worm

soil soil soil soil soil soil

189

Chapter 11

shaped accumulation curves (Chapter 4, 5), or decreased uptake in re-used soil (Chapter 6, 10). The advanced bioassay set-up with re-used soil is, in principle, a good approach although it failed to yield definitive answers (see section on bioassays, later in this chapter). The results from spiked artificial soil (Chapter 10) show that the current data sets are not compatible with the assumption of pore-water depletion and slow/no replenishment. However, for field-polluted soils, depletion cannot be ruled out. Firstly, because pore-water concentrations are often very low due to sequestration, and secondly, because field soils may adversely affect the worm in the assay, leading to decreased activity (Chapter 6). The depletion may also be very local (in the direct surroundings of the worm), which is most clear when working with earthworm species that live in permanent burrows (like Lumbricus terrestris). The rapid achievement of steady state for some PAHs may point at (local) depletion (Chapter 5). However, the ratio worm:soil that was used in this work was quite low (1:200–300 gwwt worm:gwwt soil in Chapter 5 and 6). In other studies, depletion may lead to much more severe effects; e.g. in the work of Tang & Alexander [28], eight adult individuals of Eisenia fetida were used on 10 grams of soil. It can also be speculated that perhaps it is not the pore water that is depleted, but the rapidly-desorbing fraction. Desorption rates for this fraction are quite rapid compared to the accumulation kinetics [5], and this fraction may be small enough to be depleted by the accumulation into the worms. Only for PAHs, peak-shaped accumulation curves were observed. In Chapter 2, we concluded from model calculations that depletion alone is not sufficient to cause this pattern. Biotransformation of PAHs is unlikely to have such a severe effect (Chapter 4), and this process would require this accumulation pattern to show up in more than just a few field-polluted soils (Chapter 5). Recent detailed investigations clearly rule out biotransformation (R. Fleuren, unpublished results). This pattern therefore suggests a rather rapid change of availability during the exposure, and the fact that PCBs do not show this pattern could point at biodegradation. Biodegradation takes place from the dissolved phase in soil, and desorption or dissolution of PAHs can easily become rate limiting (see Chapter 4). This implies that biodegradation can keep dissolved concentrations low, without any apparent effect on the total concentration (as was elegantly demonstrated by De Maagd et al. [7]). On the other hand, a rapid sequestration cannot be ruled out. Directly after spiking the soil with PAHs (Chapter 4), or sieving and homogenising field-polluted soil (Chapter 5), bioavailability may be relatively high. A rapidly proceeding sequestration can produce the observed peak shapes, but the question remains why this pattern is apparently not observed for other chemicals. In summary, sequestration is a general phenomenon for persistent chemicals in the field, over a timescale of years. However, its predictability is currently poor and experimental methods are required. On a shorter timescale (days), changes in availability may occur in bioassays. No proper explanation can be given for this behaviour, but biodegradation, depletion and rapid sequestration processes are likely involved. Fortunately, these processes all act in decreasing the body residues in relation to EP, so EP can still be used as a worst-case estimate. PAHs seem to pose more problems in this respect than other chemicals, but the good news is that the different PAHs behave quite similarly (Chapter 5), so that information on few chemicals can be used for the entire group.

190

Summary and discussion

UPTAKE THROUGH FEEDING In this thesis, a convincing case is made for the assumption that uptake from the gut is uptake/elimination also driven by passive diffusion from a dissolved phase (Chapter 8, 10). For very hydrophobic chemicals (PCBs: Chapter 10, PAHs: R. Fleuren, unpublished results), the gut is the dominant uptake route for the earthworm. For less hydrophobic chemicals (chlorobenzenes, up to and possibly including hexachlorobenzene), the skin feeding/egestion route is likely to dominate. The importance of the gut route also implies that a general relationship between the overall elimination rate and Kow is not to be expected. Feeding behaviour differs between species, and depends on soil properties and the availability of food (Chapter 6, 8, 9). Therefore, feeding behaviour will influence the elimination rate (Chapter 6). The importance of the gut route may also explain the fact that equilibrium in bioassays usually takes less than a week, even for very hydrophobic chemicals. In contrast, SPME fibres take longer to equilibrate in a soil slurry than worms in soil, for very hydrophobic chemicals (Chapter 7).

worm worm

gut gut

soil soil

food food

Even though the gut route may be dominant for very hydrophobic chemicals, this does not imply large deviations from EP. Because uptake from the gut is through passive diffusion from a dissolved phase, EP also applies to this route. This is unintentionally demonstrated in Chapter 4, where the uptake of pyrene from soil was shown to saturate at high concentrations. This saturation is caused by the fact that, at the highest concentrations, the water solubility is reached. At this point, any additional pyrene will remain in the soil in crystalline form, which is unavailable to the worms, apparently even through the gut route. Even though EP clearly applies, the concentration and availability of the chemical in the gut may differ from that in the soil. For earthworms consuming soil, the gut contents result from the soil, although they are enriched in organic matter (OM) content (Chapter 8, 9). This selectivity leads to increased chemical intake (as organic chemicals are mainly sorbed to OM), although this is largely balanced by the corresponding increase in sorption. However, it must be realised that selectivity invalidates gut corrections based on acid-insoluble residues [26], as well as empirical models for metal uptake from the gut [25]. More important for the chemical availability, the earthworm will digest part of the OM for its energy requirements. This digestion will thus also remove sorption sites, and possibly compact the gut contents (Chapter 8–10). For E. andrei in artificial soil, this implied that, even for PCB 153, the body residues will be within a factor of 1.3 from the EP value (which also confirmed for PAHs; R. Fleuren, unpublished results). The maximum deviation from EP is largely determined by the digestive efficiency of OM. Roughly speaking, a 50% digestion efficiency (which would be high for earthworms) can lead to a deviation from EP with the soil by a factor of two. However, to achieve this maximum, the retention time needs to be short (to prevent depletion in the gut), and the exchange across the skin must be slow (to isolate the worm from the soil, and keep its concentration above EP with the soil). A difference within a factor of two will easily be masked by other factors, working to decrease the body residue (see previous section). The net result is that also very hydrophobic compounds like PCBs

191

Chapter 11

do not exceed the EP estimate, even when it is based on measured freely-dissolved concentrations (Chapter 7). This means that there is little support for an additional safety factor in risk assessment to cover the effects of feeding (like the current factor 10 in the EU risk assessment guidelines [8]). E. andrei likely represents a worst case for the effects of feeding, as its digestion efficiency is much higher than values reported for typical geophageous earthworms, and because its retention time is much shorter than anecic (deep-burrowing) species with a similar digestion efficiency (Chapter 8, 9). The fact that feeding does not lead to large deviations from EP also implies that there is little market for the extended bioaccumulation model (Chapter 8, 10). There is however an exception; when the food source of the worms is specifically contaminated, EP does not apply because the soil concentration is not relevant for the body residues. This situation may occur for pesticide residues in leaf litter, and for animal pharmaceuticals excreted with manure. To use the model in a predictive manner, physiological data are required for the earthworm species and soil type under consideration (Chapter 9), an estimate of the sorption in the food source (which may differ from the sorption to soil, Chapter 8), and estimates for the rate constants across the gut wall and the skin. Especially for the rate constants, insufficient information is currently available to predict the parameter values, which may also be species and soil-type dependent. Further work in this direction may want to focus on typical surface-feeding species like L. terrestris and L. rubellus. As a temporary worst case, it is possible to assume that the worm reaches equilibrium with the gut contents (requiring only estimates for the digestive efficiency, OM content of the food source, and sorption in the food). This level of detail may be sufficient to provide a worstcase estimate for risk-assessment applications. The methods applied in this thesis for studying the feeding route are promising (Chapter 8–10), and the accumulation model appears to provide a good framework to describe experimental results. The methods described in Chapter 9 are well suited to measure the physiological parameters regarding feeding, although some work is needed to validate the methods for digestion efficiency and selectivity. Retention times and feeding on specific food sources can be examined using lanthanides as passive tracers, although a (radio)tracer more specific for OM would be better suited (e.g. PDMS [18]). Furthermore, ligaturing the worms with tissue adhesive is a good way to study uptake through the skin in isolation (Chapter 10). Given that the tested chlorobenzenes where largely taken up through the skin, this method appears valid (most conceivable artefacts would lead to less uptake in ligatured worms).

A MORE COMPLETE PICTURE OF BIOAVAILABILITY The simple picture of EP at the beginning of this chapter is clearly insufficient to cover all aspects of bioavailability. Figure 1 provides a more complete picture of the dominant processes governing uptake into the earthworm, as discussed in this thesis, and Table 1 shortly reiterates the importance and predictability of these processes. Clearly, bioavailability is a dynamic issue, depending on the rate constants of all these processes, which implies that equilibrium between all phases will only be achieved in specific cases. Nevertheless, the general validity of the EP predictions shows that the deviations from equilibrium are usually small. It can thus be stated that this picture is overly complex, and that it will seldom be necessary to quantify all these processes (especially for riskassessment applications). We probably know all the players contributing to bioavailability, and we can estimate many of the partition coefficients from hydrophobicity, but the rate constants are more elusive. There is evidence that the desorption rate constants from the different phases in sediments are comparable for a 192

Summary and discussion

large group of chemicals [29], but for the rate constants of the worm, similar evidence is still not available. In Chapter 10, we quantified the rate constant across the skin and across the gut wall for three chemicals in one species of earthworm and one soil, which is insufficient to base an estimation routine on. In practice, most studies focus only on specific aspects of this picture for bioavailability, and could therefore be flawed. In my opinion, future toxicokinetic work with soil organisms should at least be aware of the processes governing bioavailability and, if possible, quantify them, or discuss why they can be ignored, before resorting to ad hoc explanations. As an example, Krauss & Wilcke [17] studied PCBs and PAHs and found a difference between a chemical measure of availability (C18 discs) and accumulation in earthworms. They suggested biotransformation of PAHs and additional uptake from the gut, although their deviations are more likely explained by a local depletion with the C18 discs.

biotransformation uptake/elimination

worm worm

food food

degradation

water water

gut gut feeding/egestion

uptake/elimination

soil soil soil soil soil soil

DOC DOC sorption/desorption

sequestration

Figure 1. A more complete picture of bioavailability, containing all of the processes that are discussed in this thesis.

PROPOSALS FOR DESIGNING BIOASSAYS A direct measure of bioavailability can be obtained in bioassays with earthworms, that can be performed with spiked (Chapter 4, 10), as well as field-polluted soils (Chapter 5, 6). The work in this thesis supports the use of the compost worm E. andrei for bioassays. This species is a compost dweller, and is popular only because of its ease of culturing. Nevertheless, it behaves like a true earthworm, feeding on soil and selecting for OMenriched fractions (Chapter 9). This is not only the case for artificial soil, but also for field soils (Chapter 6, and [12]), although this species does not perform well on heavy clay soils (personal observations). Regarding this aspect, it is a good idea to routinely weigh the faecal matter, produced in the depuration phase on filter paper, as a measure of the feeding activity in bioassays (Chapter 6, 9). This is a relatively easy procedure that provides important information for the interpretation of earthworm tests.

193

Chapter 11

Table 1. Summary of the importance and predictability of the processes in Figure 1. Importance Uptake/elimination Important pathway for from water hydrophobic chemicals that are not extremely hydrophobic (probably log Kow < 5 or 6). Uptake takes place mainly via passive diffusion from the pore water. Uptake/elimination Important route for very from gut hydrophobic chemicals (probably log Kow > 5 or 6). Uptake takes place mainly via passive diffusion from the gut contents. Feeding/ingestion Earthworms select an OMenriched fraction from soil. Soil makes up a large part of the diet, even when a specific food source is available. Sorption/desorption Because uptake into the worm is from soil by passive diffusion, it is of great importance to know the sorption behaviour. A slow desorption may influence the accumulation kinetics. Sequestration It is a general phenomenon that availability decreases with contact time. Bioavailability may be a up to a factor of 100 lower than predicted. Sorption/desorption DOC-associated chemicals are not from dissolved directly available for uptake. Total organic carbon pore-water measurements thus (DOC) need to be corrected for DOC. QSARs for Koc appear to describe the freely-dissolved concentration. Biodegradation Many chemicals are degraded to some extent in soil. Degradation coupled to slow desorption may produce peak-shaped accumulation curves. Biotransformation Not studied in this thesis. Nevertheless, this does not seem to be a process of great importance for earthworms, in general.

Predictability The BCF is well-predicted from Kow for neutral hydrophobics. The rate constants are more problematic. It is likely that the elimination rate decreases with Kow. Deviation of BSAF from EP is small (less than a factor of two). There is currently insufficient information to predict rate constants. Feeding behaviour is soil-, food-, and species-specific. It is therefore impossible to provide generally applicable defaults. The Koc can be estimated from Kow although this estimation is only reliable for neutral hydrophobics, and sequestration often hampers estimation. The rate constants are often unknown. Sequestration cannot be properly estimated from soil and chemical properties. Measurements are thus necessary. The partition coefficient with DOC is probably related to Kow, although the quality of DOC plays a large role. The desorption rate from DOC may be rapid. Biodegradation rates are difficult to predict, and are soil-specific. Furthermore, desorption and diffusion may limit degradation. Predictability of transformation rates in earthworms is low. Measurements are needed, and transformation is likely to be species-specific.

The accumulation results of E. andrei correlate well with those of more field-relevant species like L. rubellus (Chapter 5) and Aporrectodea caliginosa (Chapter 6). On average, E. andrei accumulates approximately 1.5 to 2 times more than the other species, although the reason remains unclear. For heavy clay soils, A. caliginosa may be a better choice although this species takes some more time and effort to culture in large numbers. L. terrestris may a good choice for feeding trials but is a poor choice for soil bioassays, even though it is applied in some cases [4,16]. This species lives in permanent burrows and therefore its

194

Summary and discussion

body burden does not fully reflect the levels in soil (it more likely represent the levels in surface litter), and may underestimate body burdens in other species. As medium, OECD artificial soil behaves like a true soil in all respects, for testing of organic chemicals. Earthworms feed readily on this medium and are able to gain nutrition from it (Chapter 9), and the accumulation results are similar to field soils (Chapter 3). However, results for spiked contaminants do not necessarily reflect the hazard of older soil pollution (see section on non-equilibrium, earlier in this chapter). To properly understand the results of bioassays and evaluate EP, it is advisable to measure chemical availability in the same experiment (e.g. by SPME, Chapter 7) or rapidly-desorbing fractions (e.g. by Tenax, [6]). It is best to use a dynamic experimental design (i.e. determine body residues at several points in time) for the bioassays. In this way, problems with non-equilibrium may be spotted, as well as other deviations from the expected pattern (Chapter 2). It is furthermore better to maximise the number of time steps than to maximise replication. The focus on replication for each measurement stems from the use of statistical testing (testing two or more “treatments” for differences in the means), and may be useful to test the analytical procedures or to assess inter-individual variation in the test animals. However, for fitting a model, spreading the measurements over time is more valuable than replicating time steps. Elimination rates can be estimated from the accumulation curve, but if possible, an elimination experiment in comparable uncontaminated soil should be performed. This is especially useful when non-equilibrium conditions are suspected; the elimination rate from the accumulation phase may be confounded by bioavailability problems (see Chapter 6). However, elimination rates may also be influenced by narcotic effects (Chapter 4), and the uncontaminated soil needs to be exactly the same (which is practically impossible for field-polluted sites, see Chapter 6). Furthermore, one may observe apparent two-phase elimination (see e.g. [3]). In Chapter 6 and 10, something similar was observed, although in this case, the pattern could be explained by contamination of the elimination medium: the worm is eliminating the chemical from its tissues but is thereby polluting the soil. This is not to say that two-phase elimination is an experimental artefact in all cases, but care has to be taken before resorting to more-compartment models (of which the nature of the compartments is purely empirical). “Depletion” experiments in which the soil is re-used with fresh worms (Chapter 6 and 10) do not seem to be useful for routine applications, partly because of experimental problems (e.g. the worms need to be exactly the same size and condition as in the first use of the soil). Even when a difference in accumulation between both accumulation stages is observed, the causes cannot be identified (microbial activity, sequestration, sorption to earthworm faeces). Nevertheless, for PAHs this set-up clearly ruled out biotransformation as a cause of the peak-shaped accumulation curves (R. Fleuren, unpublished results).

ADVICE FOR RISK ASSESSMENT OF SOIL POLLUTION Earthworms play a central role in the risk assessment for soil-dwelling organisms. This is partly because these animals are easily cultured and tested, but also because standard test protocols are in place. However, the focus on earthworms as model organism is defensible, because they reflect the pollution status of the soil owing to their physiology and behaviour. Estimation of body residues for earthworms can be done as shown in the decision schemes of Figure 2. The schemes focus on body residues and not directly on toxicity, and are thus concerned with bioavailability per se, as well as the risk of secondary

195

Chapter 11

poisoning for predating species. Nevertheless, toxic effects can be linked to body residues (conform the concept of critical body residues [22], that is also applicable to earthworms [1,11]), which means that such a scheme can easily be expanded to include toxicity (although this is outside the scope of this thesis). In fact, when critical body residues are established, this could facilitate the extrapolation of toxic effects between different soils, and perhaps even between different species [20]. A distinction is made between generic chemical risk assessment (e.g. for new chemicals) and site-specific assessments (e.g. for determining the urgency of soil clean-up). The main difference between these two forms of risk assessment is that for generic risk assessment, we assess a chemical in general, and not in a specific geographic location. There is often no way of knowing whether the tested soil (often artificial soil) is representative (or worst case) for a broad range of soils. We therefore cannot resort to measurements in the field. This turns out to be especially problematic for “difficult” substances like surfactants, (organo-)metals, ionising compounds, etc. For these chemicals, sorption cannot usually be predicted on the basis of organic matter alone, and it is questionable whether Kow is a good descriptor for bioaccumulation. For site-specific assessments we can always collect worms from the field, or perform bioassays with the field soil, to assess the body burdens in the earthworms. If sampling is properly performed, this is than representative for the location that is to be assessed. Site-specific assessments are, however, complicated by the fact that the pollution is usually made up of a cocktail of (unknown) contaminants. INITIAL ESTIMATE For both types of risk assessment, a first estimate of risk can be made by applying estimation routines based on hydrophobicity. Estimate sorption and BCF from Kow (Chapter 3, 5, 6), and calculate BSAF by dividing BCF by the sorption coefficient. Together with the (estimated or measured) concentration and the organic carbon content (Foc) in soil, this yields a body residue that will be protective for most risk-assessment applications. It can under-predict risk when a contaminated food source is available, but will over-predict most of the times because the contaminant may be less available than predicted from Kow. Care must be taken in applying these estimation routines, as they are only reliable for neutral hydrophobic compounds. REFINING SORPTION COEFFICIENTS A second step will often be the refinement of the sorption, or measurement of pore-water levels. For generic risk assessment, this step may be particularly useful for “difficult” substances that fall outside of the categories for which there are estimation routines (see [24]). For neutral organic compounds, the estimation from Kow will usually suffice. For location-specific risk assessment, measuring freely-dissolved concentrations in pore water is possible. When measuring pore water directly (e.g. Chapter 5), it must be realised that DOC-associated chemicals are not available for uptake. A better approach is to use biomimetic approaches like SPME (Chapter 7), although it is not clear to what extent this method will work for difficult compounds. These procedures will in most cases predict body residues accurately, because the influence of sequestration is accounted for. Overpredictions can occur when the pore-water pool is depleted (e.g. by micro-organisms degrading the chemical; their effect will go unnoticed in SPME procedures with bactericides added), under-predictions occur when a food source is specifically contaminated.

196

Summary and discussion

“Difficult” “Difficult”chemicals chemicals

Neutral Neutralhydrophobic hydrophobic chemicals chemicals Measure or estimate soil concentration and F oc

Estimate Koc and BCF from Kow

Site-specific assessment

Determine K oc of food source and run model

yes

Food? no

no

Risk?

don’t worry

yes Food?

Measure pore water or soil sorption

no

yes no

Risk?

don’t worry

yes

Collect earthworms on site and determine levels

“Difficult” “Difficult”chemicals chemicals

Measure sorption in relevant soil type

Perform bioassays with earthworms in the lab

Neutral Neutralhydrophobic hydrophobic chemicals chemicals

Exposure scenario Estimate soil concentration and choose Foc

Food?

yes

Estimate Koc and BCF from Kow

Generic risk assessment

no

Measure BCF in lab test

yes

Food?

likely

Determine K oc of food source and run model

no don’t worry

no

Risk?

metabolism?

yes

Risk?

no

don’t worry

yes

Refine exposure assessment

unlikely

Bioassay with contaminated food yes

Risk?

no

don’t worry

Figure 2. Decision schemes for dealing with bioavailability and secondary poisoning via earthworms in risk assessment (site-specific and generic). The schemes focus on body residues, and not directly on toxicity; it is assumed that evaluated toxicity data for predators are available. Thick arrows indicate a continuation of the scheme to risk management, or case-by-case judgement. The blocks “Food?” ask whether a contaminated food source is available for the earthworms.

197

Chapter 11

CONTAMINATED FOOD SOURCE Special care is warranted in case a contaminated food source is available, although there is currently little attention for this route in risk assessment. Examples are situations where cattle is treated for parasitic worms (part of the medicine will be excreted with the manure), and when a pesticide is sprayed onto a crop (part of the chemical ends up in the leaf litter). The previous section on uptake through feeding shows that the presented model (Chapter 8, 10) is capable of describing these situations, but more work is required before this model can be applied in a predictive manner. Bioassays with contaminated food are also possible, for example, using treated leaf discs [9], or dung from cattle treated with anti-parasitic agents [27]. This is not standard practice, but is sometimes used to test toxicity of these chemicals to earthworms. “DIFFICULT” COMPOUNDS Care must be taken with chemicals that are extensively metabolised (although the biotransformation capacity of earthworms seems limited), and polar compounds (although the EP analysis appears to be robust to some abuse; see Chapter 3). Furthermore, the new chemicals that are currently produced may differ markedly from the chlorinated hydrocarbons on which a large part of ecotoxicology is founded. It is conceivable that the BCF, as well as sorption, for these compounds is poorly predicted by hydrophobicity (as discussed earlier in this chapter). For these chemicals, laboratory measurement of BCF and sorption may be useful, until sufficiently accurate estimation routines become available. However, one needs to be aware that for these compounds, different soil factors than OM alone may determine bioavailability. BIOASSAYS VS. FIELD-COLLECTING For location-specific risk assessment, bioassays with earthworms may be useful to measure bioavailability (see previous section on bioassays). Another option is to collect the earthworms on site and measure the chemicals they have accumulated (also proposed by [13,33]). This integrates all aspects, is not very laborious, and precludes problems with depletion, homogenisation, and choice of sampling depth. The problem with fieldcollected worms is that their individual exposure history in unknown (a sufficient number of worms must be sampled), and that different species must be analysed separately as different behaviour leads to different exposure. These biological methods may, however, be less easy to standardise or more laborious to execute, compared to chemical methods. COMMENTS TO THE NEW DRAFT OF THE TGD In 2003, a revision will appear of the European technical guidance documents (TGDs) for risk assessment of industrial chemicals. In the draft of this document, the theoretical relationship of Chapter 3 is proposed for the initial estimation of earthworm body residues. However, this is expanded with the contribution of soil in the gut contents (based on the experience of Chapter 8 and 9). Predators will consume the whole worm including the gut contents, but the quantitative effect of adding this fraction is small. In most cases, the concentration in the soil is comparable to that in the worm, and the gut contents also adds to the weight of the worms. Therefore, the net increase in concentration will be limited. The assessment of secondary poisoning furthermore assumes that 50% of the predator’s diet is derived from the local environment (close to a point source), and the other half from the regional environment (i.e. the background). Although the quantitative consequences are small (the diet concentration is at maximum a factor of two lower than the local level), this scenario does not account for several species that do have a small home range like the number one earthworm predator, the mole.

198

Summary and discussion

When EP is applied for the effects assessment for soil or sediment organisms, an additional safety factor of ten is included for chemicals with log Kow > 5. This factor is based on the work of Belfroid et al. [2], and is used to account for possible additional uptake from feeding. However, the work in Chapter 8 and 10 shows that this factor is unsupported, and based upon questionable model extrapolations (see Chapter 2). This is at the very least true for earthworms, although the validity of the mechanism indicates that the same conclusion may be drawn for other organisms. Not only is there no direct evidence for deviations of more than a factor of 1.5 (even for extremely hydrophobic compounds), the mechanism of the uptake process from the gut make it unlikely that any soil-eating species will deviate by more than a factor of two. Strangely enough, this additional safety factor is not applied in the secondary poisoning assessment. If feeding could lead to a higher toxicity for the soil-dwelling organisms, it would surely also lead to higher body residues. Nevertheless, a rather vague footnote is added, stating that in some studies, chemicals may be taken up “directly from the fraction of the substance adsorbed onto soil particles.” Apart from the fact that this idea of uptake is quite wrong, the footnote continuous that in other studies, uptake from the pore water seems to dominate. Clearly, there is still a lot of confusion surrounding uptake via feeding. The TGD includes no scenario to deal with a specifically-contaminated food source for the earthworms, although there is no general model available yet to make predictions for this route of exposure. Therefore, these parts are shown in the schemes with dotted lines (Fig. 2).

OPEN QUESTIONS At the end of this thesis, in my opinion, the following open questions remain for earthworm toxicokinetics, divided in fundamental and practical questions. For the fundamental questions with indirect practical interest: 1. What is the role of non-equilibrium conditions? What exactly determines the peakshaped accumulation curves? It would be interesting to perform an experiment with PAHs in artificial soil, using the extended bioassay set-up, and using increasing contact times. Additionally, the sequestration status could be determined using SPME. 2. What are the values for the rate constants across the gut wall and across the skin? Are they truly constant or do they depend on species or soil type? Addressing this question requires more studies as outlined in Chapter 9 and 10. The questions of direct practical interest: 1. How is the EP prediction for “difficult” chemicals like pharmaceuticals, surfactants and modern pesticides? This requires more basic experimental work in estimating sorption coefficients and BCFs, along with QSAR work to find appropriate descriptors. 2. Is SPME broadly applicable to use as a routine tool for risk assessment (i.e. for all chemicals and all soils)? The usefulness of SPME is clearly indicated by Chapter 7, but more evidence is required for a broader range of chemicals and test organisms. 3. Can the extended bioaccumulation model with the gut compartment play a role as a predictive tool for risk assessment (in case food is contaminated)? Application of the model requires quantitative knowledge on the physiological parameters and the rate constants (see point 2 of the previous list). Furthermore, we need to measure or predict sorption in food matrices.

199

Chapter 11

REFERENCES [1] [2]

[3]

[4] [5]

[6]

[7]

[8]

[9] [10] [11]

[12] [13]

[14] [15]

[16] [17] [18]

[19] [20]

[21] [22] [23]

200

Belfroid A, W Seinen, K Van Gestel and J Hermens (1993). The acute toxicity of chlorobenzenes for earthworms (Eisenia andrei) in different exposure systems. Chemosphere 26:2265-2277. Belfroid AC, W Seinen, KCAM Van Gestel, JLM Hermens and KJ Van Leeuwen (1995). Modelling the accumulation of hydrophobic organic chemicals in earthworms - Application of the equilibrium partitioning theory. Environ. Sci. Poll. Res. 2:5-15. Belfroid AC and DTHM Sijm (1998). Influence of soil organic matter content on elimination rates of hydrophobic compounds in the earthworm: possible causes and consequences. Chemosphere 37:12211234. Beyer WN (1996). Accumulation of chlorinated benzenes in earthworms. Bull. Environ. Contam. Toxicol. 57:729-736. Cornelissen G, PCM Van Noort and HAJ Govers (1997). Desorption kinetics of chlorobenzenes, polycyclic aromatic hydrocarbons, and polychlorinated biphenyls: sediment extraction with Tenax® and effects of contact time and solute hydrophobicity. Environ. Toxicol. Chem. 16:1351-1357. Cornelissen G, H Rigterink, DEM Ten Hulscher, BA Vrind and PCM Van Noort (2001). A simple Tenax® extraction method to determine the availability of sediment-sorbed organic compounds. Environ. Toxicol. Chem. 20:706-711. De Maagd PGJ, TL Sinnige, SM Schrap, A Opperhuizen and DTHM Sijm (1998). Sorption coefficients of polycyclic aromatic hydrocarbons for two lake sediments: influence of the bactericide sodium azide. Environ. Toxicol. Chem. 17:1899-1907. EC (1996). Technical Guidance Documents in support of Directive 93/67/EEC on risk assessment of new notified substances and Regulation (EC) No. 1488/94 on risk assessment of existing substances (Parts I, II, III and IV). EC catalogue numbers CR-48-96-001, 002, 003, 004-EN-C. Office for Official Publications of the European Community, 2 rue Mercier, L-2965 Luxembourg, Luxembourg. Edwards CA and PJ Bohlen (1992). The effects of toxic chemicals on earthworms. Rev. Environ. Contam. Toxicol. 125:23-99. Escher BI and RP Schwarzenbach (1996). Partitioning of substituted phenols in liposome-water, biomembrane-water, and octanol-water systems. Environ. Sci. Technol. 30:260-270. Fitzgerald DG, RP Lanno, U Klee, A Farwell and DG Dixon (1997). Critical body residues (CBRs): applications in the assessment of pentachlorophenol toxicity to Eisenia fetida in artificial soil. Soil Biol. Biochem. 29:685-688. Fleuren RHLJ, T Jager, W Roelofs, AC De Groot, R Baerselman and WJGM Peijnenburg (subm.). Feeding behaviour of Eisenia andrei in two different field contaminated soils. Submitted to Pedobiologia. Gish CD and DL Hughes (1982). Residues of DDT, dieldrin, and heptachlor in earthworms during two years following application. Special Scientific Report - Wildlife No. 241. U.S. Fish and Wildlife Service, Washington, DC, USA. Goss KU and RP Schwarzenbach (2001). Linear free energy relationships used to evaluate equilibrium partitioning of organic compounds. Environ. Sci. Technol. 35:1-9. Kraaij R, P Mayer, FJM Busser, M Van het Bolscher, W Seinen, J Tolls and AC Belfroid (2003). Measured pore-water concentrations make equilibrium partitioning work - a data analysis. Environ. Sci. Technol. 37:268-274. Krauss M, W Wilcke and W Zech (2000). Availability of polycylic aromatic hydrocarbons (PAHs) and polychlorinated biphenyls (PCBs) to earthworms in urban soils. Environ. Sci. Technol. 34:4335-4340. Krauss M and W Wilcke (2001). Biomimetic extraction of PAHs and PCBs from soil with octadecylmodified silica disks to predict their availability to earthworms. Environ. Sci. Technol. 35:3931-3935. Kukkonen J and PF Landrum (1995). Effects of sediment-bound polydimethylsiloxane on the bioavailability and distribution of benzo(a)pyrene in lake sediment to Lumbriculus variegatus. Environ. Toxicol. Chem. 14:523-531. Landrum PF (1989). Bioavailability and toxicokinetics of polycyclic aromatic hydrocarbons sorbed to sediments for the amphipod Pontoporeia hoyi. Environ. Sci. Technol. 23:588-595. Lanno RP, SC LeBlanc, BL Knight, R Tymowski and DG Fitzgerald (1998). Application of critical body residues as a tool in the assessment of soil toxicity. In: Advances in earthworm ecotoxicology. SC Sheppard, JD Bembridge, M Holmstrup, L Posthuma (eds.). SETAC press, Pensacola, FL USA. pp. 41-53. Lord KA, GG Briggs, MC Neale and R Manlove (1980). Uptake of pesticides from water and soil by earthworms. Pestic. Sci. 11:401-408. McCarty LS and D Mackay (1993). Enhancing ecotoxicological modeling and assessment. Body residues and modes of toxic action. Environ. Sci. Technol. 27:1719-1728. Nendza M (1991). QSARs of bioconcentration: validity assessment of log Pow/log BCF correlations. In: Bioaccumulation in aquatic systems. Contributions to the assessment. R Nagel, R Loskill (eds.). VCH Verlagsgesellschaft mbH, Weinheim, Germany. pp. 43-66.

Summary and discussion [24] Sabljić A, H Güsten, H Verhaar and J Hermens (1995). QSAR modelling of soil sorption. Improvements and systematics of log Koc vs. log Kow correlations. Chemosphere 31:4489-4514. [25] Saxe JK, CA Impellitteri, WJGM Peijnenburg and HE Allen (2001). Novel model describing trace metal concentrations in the earthworms Eisenia andrei. Environ. Sci. Technol. 35:4511-4529. [26] Stafford EA and SP McGrath (1986). The use of acid insoluble residue to correct for the presence of soilderived metals in the gut of earthworms used as bio-indicator organisms. Environ. Poll. 42:233-246. [27] Svendsen TS, C Sommer, P Holter and J Grønvold (2002). Survival and growth of Lumbricus terrestris (Lumbricidae) fed on dung from cattle given sustained-release boluses of ivermectin or fenbendazole. Eur. J. Soil Biol. 38:319-322. [28] Tang J and M Alexander (1999). Mild extractability and bioavailability of polycyclic aromatic hydrocarbons in soil. Environ. Toxicol. Chem. 18:2711-2714. [29] Ten Hulscher TEM, BA Vrind, H Van den Heuvel, LE Van der Velde, PCM Van Noort, JEM Beurskens and HAJ Govers (1999). Triphasic desorption of highly resistant chlorobenzenes, polychlorinated biphenyls, and polycyclic aromatic hydrocarbons in field contaminated sediment. Environ. Sci. Technol. 33:126-132. [30] Tolls J and DTHM Sijm (1995). A preliminary evaluation of the relationship between bioconcentration and hydrophobicity for surfactants. Environ. Toxicol. Chem. 14:1675-1685. [31] Tolls J (2001). Sorption of veterinary pharmaceuticals in soils: a review. Environ. Sci. Technol. 35:33973406. [32] Vaes WHJ, E Urrestarazu Ramos, HJM Verhaar and JLM Hermens (1998). Acute toxicity of non-polar versus polar narcosis: is there a difference? Environ. Toxicol. Chem. 17:1380-1384. [33] Wheatley GA and JA Hardman (1968). Organochlorine insecticide residues in earthworms from arable soils. J. Sci. Fd. Agric. 19:219-225.

201

Samenvatting in het Nederlands (Voor niet-ingewijden) DE AANLEIDING Met name na de tweede wereldoorlog is de hoeveelheid chemicaliën die we gebruiken sterk toegenomen. We gebruiken deze stoffen voor een breed scala aan activiteiten, bijvoorbeeld: • in de industrie (bijv. als grondstof voor het produceren van andere stoffen, als smeermiddel, of voor het ontwikkelen van foto’s), • in consumentenproducten (bijv. in cosmetica, als brandvertrager in kleding en als wasmiddel), • als bestrijdingsmiddel in de landbouw en door particulieren (bijv. tegen schimmels of insecten op kamerplanten en in de tuin), • als geneesmiddel voor mensen en (landbouw)huisdieren (bijv. als antibioticum, voor anticonceptie, of als ontwormingsmiddel).

Deze stoffen hebben in grote mate bijgedragen aan de mate van luxe die we vandaag de dag in de westerse wereld genieten, maar de keerzijde van de medaille is de vervuiling van ons milieu. Als voorbeeld, gechloreerde bifenylen (PCB's) werden lange tijd toegepast in o.a. transformatoren en condensatoren. Maar deze stoffen bleken niet alleen zeer giftig te zijn, ze zijn ook bijzonder slecht afbreekbaar in het milieu. De PCB’s worden bijvoorbeeld in hoge concentraties aangetroffen in het vet van ijsberen en robben aan noordpool. Inmiddels is de productie van PCB’s aan banden gelegd en nemen de emissies (bijv. door lekkage uit elektrische apparatuur) wereldwijd af, maar daarmee zijn de risico’s van stoffen nog niet verdwenen: de lijst van “bestaande stoffen” telt ruim 100,000 chemicaliën en elk jaar komen er alleen al in Europa 300 tot 350 nieuwe stoffen bij. Er is vrijwel geen toepassing van stoffen denkbaar waarbij geen verliezen naar het milieu optreden. Hierbij kan men denken aan emissies tijdens de productie van de stof, of het verwerken van de stof in een product. Verder kunnen stoffen weglekken uit producten waarin ze verwerkt zijn: bijvoorbeeld de al genoemde PCB’s uit elektrische apparaten en ftalaten (weekmakers) uit kinderspeelgoed en vloerbedekking. Andere stoffen zoals wasmiddelen en geneesmiddelen komen in veel gevallen via het afvalwater in een rioolwaterzuivering terecht, en worden daar deels afgebroken. Met name de slecht afbreekbare stoffen zullen de zuivering overleven en worden vervolgens op het oppervlaktewater geloosd. Andere stoffen zullen aan het zuiveringsslib binden. Dit slib bestaat grotendeels uit dode bacteriën, en door het hoge organisch stofgehalte wordt dit vervolgens in veel landen als meststof in de landbouw gebruikt. Op deze manier kunnen de op het riool geloosde stoffen toch de bodem bereiken. Andere giftige stoffen worden doelbewust in het milieu gebracht om schade aan organismen te berokkenen: bestrijdingsmiddelen. De beruchte gechloreerde pesticiden als DDT en dieldrin zijn weliswaar al dertig jaar verboden, maar de concentraties zijn in vele Nederlandse bodems nog steeds ruim boven de norm. Dit wordt veroorzaakt doordat deze stoffen weinig mobiel zijn en slechts langzaam afbreken in het milieu. De middelen die tegenwoordig gebruikt worden zijn gelukkig meestal goed afbreekbaar zodat eventuele schadelijke effecten relatief snel weer verdwijnen. Desalniettemin zijn de hoeveelheden die worden toegepast aanzienlijk: op dit moment gebruiken we in Nederland jaarlijks 12 miljoen kilo bestrijdingsmiddelen, voornamelijk ter bestrijding van

203

onkruid en schimmels. Met name voor de teelt van bloemen en bloembollen zijn grote hoeveelheden per hectare normaal. Onzorgvuldig omgaan met chemicaliën heeft geleid tot ernstige vormen van bodemvervuiling in Nederland en in veel andere landen. Enkele van de meest in het oog springende voorbeelden zijn zwaar vervuilde locaties als de Volgermeerpolder nabij Broek in Waterland en het Griftpark in Utrecht, en natuurlijk een grote hoeveelheid kleine locaties zoals tankstations en spoorwegterreinen. Het totaal aantal vervuilde locaties in Nederland wordt geschat op 175.000 en de totale kosten om dit schoon te maken op 40 miljard euro. Jaarlijks geven we momenteel zo’n 360 miljoen euro uit om ruim 1000 locaties op te ruimen, met een totaal oppervlak van 100 hectare. Hierbij komt 1.8 miljoen ton vuile grond vrij die weer een bestemming moet krijgen (een deel wordt gereinigd, een ander deel wordt gebruikt voor minder kritische toepassingen zoals voor de aanleg van wegen). Tabel 1. Aantallen, biomassa, en diversiteit van regenwormen in verschillende gebieden. Grondgebruik Cultuurgrond Boomgaarden Weide en grasland Loofbossen Naaldbossen

Aantallen (per m2) 67–109 197–227 248–367 82–185 36–67

Biomassa (g natgew. /m2) 25–43 85–90 92–124 25–56 16–29

Aantal soorten 1–10 6–15 4–11 3–11 1–11

Behalve een economisch probleem is bodemverontreiniging ook een ecologisch probleem. De stoffen die we in de bodem introduceren kunnen diverse negatieve effecten hebben op het bodemleven, variërend van sterfte tot meer subtiele effecten op de populatie (zoals verminderde groei of voortplanting van bodemorganismen). Uiteindelijk kan de vervuiling ook repercussies hebben voor het functioneren van processen in de bodem, zoals de afbraak van strooisel en de kringloop van nutriënten. Regenwormen vormen, met name in gematigde streken, een groot deel van de ondergrondse biomassa (tabel 1). In Nederland zijn zo’n 15 soorten algemeen, maar de meeste hebben geen goede Nederlandse namen (alleen hengelsporters gebruiken verschillende namen, hoewel die naamgeving niet altijd eenduidig is). Deze soorten verschillen niet alleen in uiterlijk, maar ook in gedrag. Zo zijn er wormen die diepe verticale gangen graven en aan het oppervlak strooisel eten, maar ook wormen die dichter aan de bodemoppervlak leven en zich een horizontale weg eten door de grond. Andere soorten tref je alleen aan in strooisellagen en mesthopen, zoals de meest gebruikte soort voor experimenten: de mestpier Eisenia andrei/fetida. Regenwormen dragen bij aan de bodemvruchtbaarheid doordat ze de grond beluchten, strooiselafbraak bevorderen en voedingsstoffen voor planten vrijmaken. Verder staan de regenwormen op het menu bij een groot aantal dieren. De meest gespecialiseerde predator is ongetwijfeld de mol, wiens dieet bijna volledig uit wormen kan bestaan. Maar ook andere zoogdieren zoals dassen, vossen en egels weten grote hoeveelheden te verorberen. Verder zijn ook een groot aantal vogels dol op wormen, zoals merels, lijsters, meeuwen, en scholeksters. Omdat wormen giftige stoffen uit de bodem kunnen accumuleren, kunnen deze stoffen via deze route weer worden doorgegeven aan hun belagers. Bodemvervuiling is dus niet alleen een probleem voor de wormen zelf, maar ook voor de vruchtbaarheid van de bodem en de gezondheid van een grote groep vogels en zoogdieren. Regenwormen zijn ook goede modelorganismen, d.w.z. de gehalten aan stoffen in hun lichaam zijn een goede afspiegeling van de mate van vervuiling in de bodem. Dit komt omdat wormen niet al te mobiel zijn en in nauw contact met de grond

204

leven. Ze hebben namelijk een dunne huid en veel soorten eten grote hoeveelheden grond (ze leven van de organische resten in de bodem). EVENWICHTSPARTITIE, DE BASISTHEORIE Niet elke stof is even problematisch in de bodem. Allereerst verschilt natuurlijk de giftigheid tussen stoffen, maar ook de beschikbaarheid. Algemeen wordt aangenomen dat alleen dat deel van de stof dat is opgelost in het bodemwater kan worden opgenomen door organismen. Dit betekent dus dat stoffen die sterk binden aan bodemdeeltjes weinig beschikbaar zijn voor organismen, en dus in de bodem minder giftig zijn dan de totale concentratie doet vermoeden. Voor organische stoffen (i.h.a. stoffen op basis van koolstof) blijkt één eigenschap de binding aan grond te domineren: de vetoplosbaarheid oftewel de “hydrofobiciteit” (letterlijk de “angst voor water”). Hoe hydrofober de stof, des te slechter lost hij op in water, en des te liever bindt hij aan het organische materiaal in de bodem. Van een erg hydrofobe stof zal dus maar een klein deel van de in de bodem aanwezige stof in opgeloste vorm in het poriewater aanwezig zijn. Dit heeft als consequentie dat deze stof relatief weinig wordt opgenomen door organismen, slecht afbreekbaar is (microorganismen kunnen er ook niet bij), en weinig mobiel is (de stof kan niet vervluchtigen uit de bodem, of weglekken naar het grondwater). Naarmate een vervuiling langer in een bodem zit zal de beschikbaarheid vaak bioconcentratiefactor afnemen. Dat wil zeggen, de stoffen migreren dieper in de bodemdeeltjes en zijn daardoor niet beschikbaar om opgenomen te worden door organismen zoals regenwormen. Dit proces wordt aangeduid als “veroudering”. Het is dus van groot belang om te begrijpen in hoeverre de vervuilende stoffen in de sorptiecoëfficiënt bodem ook daadwerkelijk worden opgenomen door organismen. Een belangrijke basistheorie is de evenwichtspartitie (EP). Deze theorie stelt dat stoffen alleen beschikbaar zijn voor organismen als ze zijn opgelost in water. De stof kan vrijelijk bewegen tussen de drie fasen, maar op een gegeven moment zal de verdeling niet meer veranderen: het systeem is in evenwicht. Het maakt niet uit waar de stof zich initieel in het systeem bevond; in evenwicht is de verhouding tussen de concentraties in de drie fasen steeds hetzelfde. Als men dus weet hoe in de evenwichtsituatie de verhouding is tussen de concentraties in water en bodem (weergegeven door de sorptiecoëfficiënt), en tussen water en organisme (middels de bioconcentratiefactor), kan men de gehalten van de stof in het organisme voorspellen vanuit de totale concentratie in de bodem. Dit staat in formulevorm in tabel 2.

worm worm

water water

bodem bodem

Tabel 2. Evenwichtspartitie kan worden gebruikt om de concentratie in de worm te schatten uit de concentratie in de bodem. Deze vergelijkingen gelden in de evenwichtssituatie.

bioconcentratiefactor = sorptiecoefficient = concentratie in de worm =

concentratie in de worm concentratie in bodemwater

concentratie in de bodem concentratie in bodemwater

concentratie in de bodem × bioconcentratiefactor sorptiecoefficient

205

Zowel de bioconcentratiefactor en de sorptiecoëfficiënt zullen afhangen van eigenschappen van de stof, en verder zal de bioconcentratiefactor afhangen van eigenschappen van de worm, en de sorptiecoëfficiënt van de bodem. Als beide coëfficiënten betrouwbaar geschat kunnen worden, is het mogelijk de concentratie in de worm te voorspellen, zonder dat deze experimenteel bepaald hoeft te worden. Het is belangrijk te onthouden dat EP dus uitgaat van evenwicht tussen de drie fasen. Als er geen evenwicht is, is het belangrijk de beginsituatie te kennen (waar zit de stof en hoeveel), en de transportsnelheden tussen bodem en water (de desorptiesnelheid) en tussen water en de worm (de opnamesnelheid). Hierdoor wordt het schatten van de concentratie in de worm dus beduidend lastiger. DE ONDERZOEKSVRAAG Het doel van mijn onderzoek was om te begrijpen hoe regenwormen stoffen opnemen en hoeveel. Kunnen we de bioconcentratiefactor schatten uit eigenschappen van de stof en hoe goed werkt de EP theorie daadwerkelijk, en wat zijn de randvoorwaarden? Kunnen we er vanuit gaan dat er inderdaad evenwicht optreedt tussen de drie fasen? Behalve hoeveel er wordt opgenomen is het ook de vraag hoe, d.w.z. via welke route. Het is bekend dat regenwormen stoffen kunnen opnemen via de huid, maar ook vanuit hun voedsel, over de darmwand. De kwantitatieve bijdrage van elke route is tot nog toe onbekend omdat de routes in experimenten lastig te scheiden zijn. De kennis van het opnameproces kan vervolgens gebruikt worden voor het ontwikkelen en valideren van wiskundige modellen, waarmee uiteindelijk voorspellingen kunnen worden gedaan (een noodzaak voor risicobeoordeling). Een belangrijk voordeel van het gebruik van modellen is dat daarmee theorie getoetst kan worden. Als de theorie bruikbaar blijkt, kun je met meer vertrouwen uitspraken doen over de opname in regenwormen in situaties die je niet experimenteel getest hebt: bijvoorbeeld voor een andere stof, een andere soort regenworm, of een andere vervuilde locatie. Zonder het gebruik van modellen zul je dus voor elke nieuwe situatie weer experimenten met wormen moeten verrichten. Hoofdstuk 2 van dit proefschrift gaat dieper in op de algemeen gebruikte modellen en hun randvoorwaarden. Hierdoor kun je bijvoorbeeld voorspellen wat voor afwijkingen je zult zien als er geen evenwicht is tussen bodem, water, en worm. HOEVEEL STOF NEMEN REGENWORMEN OP? Evenals de sorptie aan bodemdeeltjes, kan de opname van stoffen in regenwormen beschreven worden op basis van de hydrofobiciteit van de stof. Naarmate een stof hydrofober is (dus minder graag in water zit) wordt deze sterker opgenomen in de worm (omdat deze vet bevat waar de stof bij voorkeur in oplost), mits de stof beschikbaar is voor opname (dus opgelost is in het bodemvocht). De worm kan dus, weinig flatteus, worden gezien als een zakje met water en vet dat ronddrijft in het bodemwater. Hoofdstuk 3 laat zien dat deze manier van modelleren een goede beschrijving geeft voor wormen die worden blootgesteld in water alleen. De bioconcentratiefactor neemt recht evenredig toe met de hydrofobiciteit van de stof, d.w.z. de oplosbaarheid van de stof in vet bepaalt de opname in de worm. Echter, zodra de wormen in grond worden blootgesteld hebben we ook met de sorptie te maken. Omdat de sorptiecoëfficiënt evenals de bioconcentratiefactor toeneemt met de hydrofobiciteit, werken deze processen elkaar tegen (zie tabel 2). Om het netto effect te kunnen schatten maken we gebruik van de EP methode. Deze aanpak geeft echter in het algemeen een overschatting van de gemeten gehalten. Dit geldt zowel voor grond waar stoffen in het laboratorium aan zijn toegevoegd (hoofdstuk 3), als voor vervuilde grond die van een locatie komt (hoofdstuk 5 en 6). Dit kan betekenen dat er geen evenwicht is tussen bodem, water en worm, zodat EP niet opgaat. Een andere mogelijkheid is dat de 206

stof zwaarder bindt aan de bodem dan voorspeld (en dus een hogere sorptiecoëfficiënt heeft). In dit werk zijn diverse aanwijzingen gevonden voor een gebrek aan evenwicht: in hoofdstuk 4 is het aannemelijk dat micro-organismen de stof afbraken in het bodemvocht; in hoofdstuk 5 zagen we afwijkingen van voorspelde patronen die duiden op gebrek aan evenwicht (zie hoofdstuk 2), en in de hoofdstukken 5 en 6 was sprake van veroudering van de vervuiling, waardoor de stoffen minder beschikbaar waren dan verwacht. Echter, hoofdstuk 7 laat zien dat de gehalten in regenwormen prima kunnen worden voorspeld als de daadwerkelijk opgeloste concentraties worden gemeten middels zogenaamde SPME-fibers (dunne vezels met een coating van materiaal waarin stoffen kunnen oplossen). Voor de stoffen die we in dat hoofdstuk bepaald hebben (oude bestrijdingsmiddelen en PCB’s) was gebrek aan evenwicht dus geen groot probleem. De meest lastige te voorspellen stoffen bleken de polycyclische aromatische koolwaterstoffen (PAK’s), een groep van stoffen die met name vrijkomen bij (onvolledige) verbranding. Hieronder zijn ook de kankerverwekkende stoffen die gemaakt worden bij het branden van tabak. Deze stoffen waren het onderwerp in hoofdstuk 4 en 5. Het blijkt dat de PAK’s vaak minder beschikbaar zijn dan verwacht op basis van hun hydrofobiciteit, maar niet altijd. Verder lijkt het er op dat gebrek aan evenwicht een belangrijk probleem is, mogelijk omdat deze stoffen door micro-organismen kunnen worden afgebroken in het bodemwater. Al met al kan ik dus concluderen dat de EP theorie goed werkt voor regenwormen. De bioconcentratiefactor laat zich goed inschatten, maar het grootste probleem ligt in de praktijk bij de schatting van de sorptie van de stof aan bodemdeeltjes. Voor een nauwkeurige voorspelling van de gehalten in wormen is dus een betrouwbare meting van de concentratie in bodemwater onontbeerlijk. In het geval dat de sorptie wordt geschat vanuit stofeigenschappen, en als er geen evenwicht optreedt, zal EP een overschatting geven. Dit betekent dat EP in ieder geval kan worden ingezet als een “worst case” in de risicobeoordeling. HOE NEMEN REGENWORMEN STOFFEN OP? Wormen die in water worden blootgesteld nemen stoffen op via hun huid. Dit is een passief diffusieproces, d.w.z. de moleculen bewegen bioconcentratie willekeurig tussen de worm en het water en verdelen zich al naar gelang de omgeving ze bevalt. De moleculen bewegen langzamer als de omgeving ze bevalt, en een hydrofobe stof zal dus uiteindelijk in een hogere concentratie in de worm te vinden zijn dan in het water. Dit is niet het resultaat van actieve opname door de worm, maar het resultaat van eten en poepen willekeurige bewegingen van de moleculen. Als wormen zich echter in grond bevinden zullen ze hiervan ook eten. Dit proces is echter lastig te bestuderen omdat het leven van de worm zich grotendeels ondergronds afspeelt. In hoofdstuk 9 gebruik ik daarom diverse indirecte methoden, zoals het gebruik van merkstoffen om de eetsnelheid te bepalen, en het opensnijden van de worm en vervolgens doormeten van het materiaal in de darm. De resultaten laten zien dat de worm bij het eten selectief te werk gaat: hij eet niet wat hem voor de mond komt, maar selecteert een fractie die organischer is dan de gehele bodem. Die organische fractie bestaat waarschijnlijk uit de resten van planten en andere organismen, en uit levende schimmels en bacteriën die als voedsel kunnen dienen voor de worm.

worm worm

darm darm

bodem bodem

207

Met het eten van de grond krijgt de worm ook de stof in zijn darm, van waaruit deze kan worden opgenomen in zijn lichaam. Het proces van opname is echter precies hetzelfde als over de huid, namelijk passieve diffusie (hoofdstuk 8 en 10). Er zijn echter enkele verschillen. Ten eerste heeft de darminhoud een andere samenstelling dan de bodem omdat de worm selectief eet. Ten tweede verteert de worm een deel van het organische materiaal voor zijn levensbehoeften. Omdat hydrofobe stoffen graag binden aan organisch materiaal in de bodem betekent dit dat ze door vertering meer beschikbaar komen voor opname in de worm (ze worden als het ware in de waterfase gedwongen doordat het organische materiaal wordt afgebroken). Verder is de uitwisseling tussen de bodem en de darminhoud geen passieve diffusie: de stof diffundeert niet de mond in, maar reist mee met het voedsel dat de worm eet. Het is dus belangrijk om te kwantificeren wat de worm eet, hoeveel hij eet, en hoeveel hij verteert (hoofdstuk 9). Op basis van deze visie is een nieuw wiskundig model bioconcentratie bioconcentratie geformuleerd en getest (hoofdstuk 8 en 10). Het is duidelijk dat de worm te maken heeft met twee opname- en uitscheidingsroutes. De worm kan niet in evenwicht komen met de sorptiecoëfficiënt darminhoud en het water eten en poepen tegelijk; de uiteindelijke concentratie in de worm zal dus ergens tussen beide evenwichten in komen te liggen. Waar precies hangt af van de snelheid van de verschillende processen. Hoofdstuk 10 laat zien dat dit beeld van de opname uit het voedsel inderdaad gebruikt kan worden om meetgegevens te beschrijven. Om beide opnameroutes te kunnen scheiden is het de wormen onmogelijk gemaakt om te eten: hun mond werd dichtgemaakt met een medicinale weefsellijm, zodat alleen opname door de huid mogelijk was. Uit deze experimenten bleek dat het belang van de twee opnameroutes (huid en darm) afhangt van de stof waarnaar men kijkt. Hoe hydrofober de stof is, des te groter het aandeel van de route via de darm (uit het voedsel). Erg hydrofobe verbindingen zoals PAK’s en PCB’s worden voornamelijk via de darm opgenomen. Voor de meest hydrofobe stof die we getest hebben (een PCB), was de opname in de “dichtgelijmde” wormen beduidend langzamer dan in intacte wormen. Meer hydrofiele (“waterlievende”) stoffen gaan voornamelijk door de huid. Voor deze stoffen was er weinig verschil in opname tussen intacte en dichtgelijmde wormen. Waarschijnlijk komt dit doordat voor deze stoffen relatief snel evenwicht ontstaat tussen worm en bodemwater. De opname vanuit de darminhoud wordt dan beperkt door de eetsnelheid van de worm. Het feit dat de darm de belangrijkste opnameroute is voor hydrofobe stoffen betekent nog niet dat het plaatje van EP, zoals eerder geschetst, niet meer opgaat. Als de worm veel organisch materiaal verteert wordt de beschikbaarheid van de stof verhoogd t.o.v. de bodem. Dit zou betekenen dat de worm hogere gehalten van de stof kan bevatten dan voorspeld op basis van EP. Regenwormen zijn echter niet bijzonder efficiënt in het verteren van bodem; in mijn experimenten is de afwijking van EP maximaal een factor 1,3 en het theoretische maximum is ongeveer een factor twee (hoewel het onwaarschijnlijk is dat dit in werkelijkheid gevonden zal worden). Opname vanuit de darm zorgt er wel voor dat de wormen snel in evenwicht komen met de bodem (vaak sneller dan de eerder besproken SPME-fibers). Deze afwijkingen van EP zullen in de praktijk van de risicobeoordeling niet snel tot problemen leiden, allereerst omdat de spreiding in de meetgegevens een verschil van een factor twee vaak zal verbergen, en ten tweede omdat

worm worm

water water

darm darm

bodem bodem

208

andere processen weer juist zorgen voor lager-dan-verwachte concentraties (gebrek aan evenwicht en veroudering van de vervuiling). Oftewel, ook al werkt het uitgebreide model goed, in de praktijk kan worden volstaan met een simpele EP berekening. Er zijn echter situaties denkbaar waar dit model wel extra informatie kan opleveren: namelijk als de worm een voedselbron heeft die specifiek vervuild is. Bijvoorbeeld, als landbouwhuisdieren een ontwormingskuur krijgen zal een groot deel van het geneesmiddel met de mest het dier verlaten en op de bodem komen. Veel wormen zijn dol op mest, en zullen op deze manier aan de stof worden blootgesteld. TEN SLOTTE Op basis van het werk in dit proefschrift worden in hoofdstuk 11 aanbevelingen gedaan voor de opzet van experimenten met regenwormen, en voor de risicobeoordeling van bodemverontreiniging. Het werk uit hoofdstuk 3 heeft al geleid tot een aanpassing van de Europese richtlijnen voor risicobeoordeling, maar verdere verbeteringen zijn mogelijk. Zo wordt bijvoorbeeld in de richtlijn een extra veiligheidsfactor van 10 gebruikt voor erg hydrofobe stoffen, om rekening te houden met additionele opname via de darm. Het werk in dit proefschrift toont aan dat zo’n factor niet onderbouwd wordt door experimenten of modelstudies. Een kanttekening die bij dit werk gemaakt moet worden is dat ik alleen gekeken heb naar stoffen met een redelijk “brave” moleculaire structuur. Veel van de stoffen die tegenwoordig op de markt komen hebben een ingewikkelde structuur en typische eigenschappen. Voor deze stoffen zal additioneel onderzoek moeten uitwijzen of EP geschikt is.

209

Curriculum Vitae Tjalling Jager was born in Purmerend on August 12, 1969. After finishing his preuniversity education (VWO) at the St. Igatiuscollege, Purmerend in 1987, he started a study Biology at the Vrije Universiteit (VU) in Amsterdam. He specialised in ecophysiology of plants and animals, and theoretical biology, with periods of practical training at the department of plant ecology of the VU (the role of water relations and cellwall elasticity in the growth process of plants, supervised by Ir. H. Linders, Ir. G. Lenssen, Drs. W. van Duin), and at the National Institute for Public Health and the Environment (RIVM) in Bilthoven (PB-PK modelling of dioxin in the human population, supervised by Prof. Dr. W. Slob, and Prof. Dr. S.A.L.M. Kooijman). After graduating in 1992, he started working at the National Institute for Public Health and the Environment (RIVM) in Bilthoven with the Laboratory for Ecotoxicology, as a scientific researcher in the area of chemical risk assessment. In this period, he worked intensively on the 1996 update of the European risk assessment guidelines (TGD), and the development and model analysis of risk assessment systems (the EU System for the Evaluation of Substances – EUSES – and its predecessors). Furthermore, his work developed to include bioaccumulation in earthworms (experimental as well as modelling), opening up a possibility to write this PhD thesis. Since 2002, he has been working at the Vrije Universiteit in Amsterdam as a post-doc researcher with the department of Theoretical Biology, on the model analysis of ecotoxicity tests (simultaneous assessment of multiple endpoints), as well as the translation to population effects. This project is done in cooperation with the Wageningen University, department of Nematology.

210

List of Publications IN THIS THESIS (chronological) Jager, T (1998). Mechanistic approach for estimating bioconcentration of organic chemicals in earthworms (Oligochaeta). Environ. Toxicol. Chem. 17: 2080-2090. Jager, T, FA Antón Sánchez, B Muijs, EG van der Velde and L Posthuma (2000). Toxicokinetics of polycyclic aromatic hydrocarbons in Eisenia andrei (Oligochaeta) using spiked soil. Environ. Toxicol. Chem. 19: 953-961. And erratum in Environ. Toxicol. Chem. 19:1702. Jager, T, R Baerselman, E Dijkman, AC de Groot, EA Hogendoorn, A de Jong, JAW Kruitbosch and WJGM Peijnenburg (2003). Availability of polycyclic aromatic hydrocarbons to earthworms (Eisenia andrei, Oligochaeta) in field-polluted soils and soil-sediment mixtures. Environ. Toxicol. Chem. 22(4):767-775. Jager, T, RHLJ Fleuren, W Roelofs and AC de Groot (2003). Feeding activity of the earthworm Eisenia andrei in artificial soil. Soil Biol. Biochem. 35(2):313-322. Jager, T (2003/2004). Modeling ingestion as an exposure route for organic chemicals in earthworms (Oligochaeta). Accepted for publication in Ecotoxicology. Jager, T, RHLJ Fleuren, EA Hogendoorn and G de Korte (subm.). Elucidating the routes of exposure for organic chemicals in the earthworm, Eisenia andrei (Oligochaeta). Submitted to Environ. Sci. Technol. OTHER SCIENTIFIC PAPERS (chronological, published and accepted) Jager, DT, TG Vermeire, W Slooff and H Roelfzema (1994). Uniform System for the Evaluation of Substances II: Effects Assessment. Chemosphere 29: 319-335. Jager, DT, CJM Visser and D van de Meent (1994). Uniform System for the Evaluation of Substances IV: Distribution and Intake. Chemosphere 29: 353-369. Polder, MD, EM Hulzebos and DT Jager (1995). Validation of models on uptake of organic chemicals by plant roots. Environ. Toxicol. Chem. 14: 1615-1623. Vermeire, TG, DT Jager, B Bussian, J Devillers, K den Haan, B Hansen, I Lundberg, H Niessen, S Robertson, H Tyle and PTJ van der Zandt (1997). European Union System for the Evaluation of Substances (EUSES): Principles and structure. Chemosphere 34: 1823-1836. Polder, MD, EM Hulzebos and DT Jager (1998). Bioconcentration of gaseous organic chemicals in plant leaves. Environ. Toxicol. Chem. 17: 962-968. Severinsen, M and T Jager (1998). Modelling the influence of terrestrial vegetation on the environmental fate of xenobiotics. Chemosphere 37: 41-62. Peijnenburg, WJGM, L Posthuma, PGPC Zweers, R Baerselman, AC de Groot, RPM van Veen and T Jager (1999). Prediction of metal bioavailability in Dutch field soils for the oligochaete Enchytraeus crypticus. Ecotox. Environ. Saf. 43: 170-186. Peijnenburg, WJGM, R Baerselman, AC de Groot, T Jager, L Posthuma and RPM van Veen (1999). Relating environmental availability to bioavailability: Soil-type-dependent metal accumulation in the oligochaete Eisenia andrei. Ecotox. Environ. Saf. 44: 294-310. Huijbregts, MAJ, U Thissen, JB Guinée, T Jager, D Kalf, D van de Meent, AMJ Ragas, A Wegener Sleeswijk and L Reijnders (2000). Priority assessment of toxic substances in life cycle assessment I: Calculation of toxicity potentials for 181 substances with the nested multi-media fate, exposure and effects model USES-LCA. Chemosphere 41(4): 541-573. Huijbregts, MAJ, U Thissen, T Jager, D van de Meent and AMJ Ragas (2000). Priority assessment of toxic substances in life cycle assessment II: Assessing parameter uncertainty and human variability in the calculation of toxicity potentials. Chemosphere 41(4): 575-588. Jager, T, TG Vermeire, MGJ Rikken and P van der Poel (2001). Opportunities for a probabilistic risk assessment of chemicals in the European Union. Chemosphere 43: 257-264. Peijnenburg, W, R Baerselman, A de Groot, T Jager, D Leenders, L Posthuma and R van Veen (2001). Quantification of metal bioavailability of lettuce (Lactuca sativa L.) in field soils. Arch. Environ. Toxic. Contam. 39: 420-430.

211

Vijver, M, T Jager, R van Veen, R Baerselman, A de Groot, L Posthuma and W Peijnenburg (2001). The impact of metal pools and soil properties on metal accumulation in Folsomia candida (Collembola). Environ. Toxicol. Chem. 20(4): 712-720. Vermeire, T, T Jager, G Janssen, P Bos and M Pieters (2001). A probabilistic human health risk assessment for environmental exposure to dibutylphthalate. Human Ecol. Risk Assess. 7(6): 16631679. Jager, T, HA den Hollander, P van der Poel, MGJ Rikken and T Vermeire (2001). A probabilistic environmental risk assessment for dibutylphthalate (DBP). Human Ecol. Risk Assess. 7(6): 16811697. Tarazona, J, K Hund, T Jager, M S-Salonen, AMVM Soares, JU Skaare and M Vighi (2002). Standardizing chemical risk assessment, at last. Nature 415: 14. Van Wezel, AP, T Jager (2002). Comparison of two screening level risk assessment approaches for six disinfectants and pharmaceuticals. Chemosphere 47:1113-1128. Artola-Garicano, E, I Borkent, K Damen, T Jager and WHJ Vaes (2003). Sorption kinetics and microbial degradation activity of hydrophobic chemicals in sewage sludge: model and measurements based on free concentrations. Environ. Sci. Technol. 37: 116-122. Vijver, M, T Jager, L Posthuma and W Peijnenburg (2003). Metal uptake from soils and soilsediment mixtures by larvae of Tenebrio molitor (L.) (Coleoptera). Ecotox. Environ. Saf. 54:277-289. Roelofs, W, MAJ Huijbregts, T Jager and AMJ Ragas (2003). Prediction of ecological no-effect concentrations for initial risk assessment: combining substance-specific data and database information. Accepted for publication in Environ. Toxicol. Chem. 22(6). Peijnenburg, WJGM and T Jager (acc.). Monitoring approaches to assess bioaccessibility and bioavailability of metals: matrix issues. Accepted for publication in Ecotox. Environ. Saf. Heugens, EHW, T Jager, R Creyghton, MHS Kraak, AJ Hendriks, NM van Straalen and W Admiraal (acc.). Temperature dependent effects of cadmium on Daphnia magna: accumulation versus sensitivity. Accepted for publication in Environ. Sci. Technol. BOOK CHAPTERS Jager, DT (1995). Human exposure through the environment. Chapter 4.6 in: Van Leeuwen and Hermens (eds.). Risk assessment of chemicals: an introduction. Dordrecht, Kluwer Academic Publishers, ISBN 0-7923-3740-9. Jager, T and J de Bruijn (2001). The EU-TGD for new and existing chemicals: does it predict risk? In: Rainbow, Hopkin and Crane (eds.). Forecasting the environmental fate and effects of chemicals. John Wiley & Sons, Ltd. Sussex, United Kingdom. Breure, AM, DT Jager, D van de Meent, C Mulder, WJGM Peijnenburg, L Posthuma, M Rutgers, AJ Schouten, A Sterkenburg, J Struijs, P van Beelen, M Vonk and D de Zwart (2002) Ecological Risk Assessment of Environmental stress. In: EOLSS, Encyclopedia of Life-Support Systems. Unesco, Paris/World Summit on Sustainable Development (Rio +10), Johannesburg. On line at http://www.eolss.net. RISK ASSESSMENT GUIDELINES AND SYSTEMS (as co-editor and co-author) VROM, WVC, RIVM and RPC (1992). Uniform System for the Evaluation of Substances (USES), Second Prototype December 1992. RIVM report no. 679120 002. RIVM, VROM and WVC (1994). Uniform System for the Evaluation of Substances (USES), version 1.0. The Hague, Ministry of Housing, Spatial Planning and the Environment. Distribution No. 11144/150. EC (1996). Technical Guidance Documents in support of Directive 93/67/EEC on risk assessment of new notified substances and Regulation (EC) No. 1488/94 on risk assessment of existing substances (Parts I, II, III and IV). EC Catalogue Numbers CR-48-96-001, 002, 003, 004-EN-C. Office for Official Publications of the European Community, 2 Rue Mercier, L-2965 Luxembourg. EC (1996). EUSES, the European Union System for the Evaluation of Substances. National Institute of Public Health and the Environment (RIVM), the Netherlands. Available from the European Chemicals Bureau (EC/DGXI), Ispra, Italy. RIVM, VROM and VWS (1998). Uniform System for the Evaluation of Substances 2.0 (USES 2.0.). RIVM report no. 679102 044. 212

Dankwoord Het zit erop, het boekje is af. Ik heb er met veel plezier aan gewerkt, en ook aan de publicaties die er bij horen, maar ik ben ook blij dat het klaar is. Vooral de laatste loodjes van het promotiecircus hebben me heel wat buikpijn opgeleverd; ik denk dat je als man niet dichter in de buurt van een bevalling kan komen als dit. Zo aan het einde van dit epos heb ik toch de behoefte iets over de totstandkoming uit te leggen, omdat dit anders dan gebruikelijk is. De hele reis begon bij het schrijven van een RIVM-rapport in 1996 over de accumulatiemodellen die worden gebruikt in de risicobeoordeling. Bij het schrijven werd me duidelijk dat de schattingsroutine voor opname in regenwormen die tot dan toe gebruikt werd niet alleen verouderd was, maar ook slordige fouten bevatte. Dit leek me dus een makkelijke manier om te scoren (zie hoofdstuk 3). Verschillende wormensoorten hebben verschillend gedrag en dus verschillende blootstelling. Ik heb me daarom verdiept in de taxonomie van de regenworm, kreeg de smaak van wormen (in figuurlijk opzicht) te pakken, en heb een stuk of tien soorten “gekweekt” (zowel thuis – in de woonkamer en de koelkast - als op het lab). Dit heeft uiteindelijk tot een redelijk lopende kweek geleid van caliginosa (gebruikt in hoofdstuk 6) en rubellus (gebruikt in hoofdstuk 5). Toen ik de kans kreeg de data van Paco te analyseren (zie hoofdstuk 4), begon het idee voor een proefschrift vorm te krijgen. Dit leek me vooral een goed idee omdat ik het promoveren als een soort breekijzer kon gebruiken om van de baas steun te krijgen voor leuke (maar niet direct beleidsrelevante) experimentele projecten. Modelleerwerk is wel leuk, maar zonder gericht experimenteel werk blijft het droogzwemmen. Desalniettegenstaande is de dataanalyse en het schrijfwerk voornamelijk ten koste van mijn vrije tijd gegaan (tja, dat loopt wel eens een beetje doorelkaar …). Via mijn promotor en copromotor heb ik kunnen samenwerken met het IRAS, wat behalve goede discussies ook direct tot twee hoofdstukken heeft geleid (hoofdstuk 6 en 7), en indirect aanleiding was voor hoofdstuk 2. Veel mensen hebben op vele manieren geholpen bij de totstandkoming van dit boekje, maar een aantal wil ik er even speciaal uitlichten. ☺Als eerste wil ik mijn promotor en copromotor bedanken. Toen het idee voor een boekje vastere vormen begon aan te nemen heb ik in Kees en Joop goede sparring partners gevonden. Jullie hebben de samenwerking met IRAS vorm gegeven, mijn promotieplannen gestructureerd, mijn schrijfsels gelardeerd met kritische noten, gemotiveerd, schouderklopjes uitgedeeld, en puntjes op i-en gezet. ☺Het grootste deel van dit proefschrift-ei is gelegd op het RIVM. Ik wil dus graag mijn labhoofden (m.n. Herman en Hans C.) en afdelingshoofden (m.n. Carla en Ton B.) bedanken voor hun support en motivatie. Jullie hebben uiteindelijk de randvoorwaarden geschapen om de experimenten voor de leukste hoofdstukken (6, 7 en 9, 10) te doen. Herman, ik ben blij dat je lid wilde zijn van mijn leescommissie. ☺Verder natuurlijk mijn ex-collega’s van ECO voor de plezierige samenwerking en buitenschoolse activiteiten (borrelen, labuitjes, etentjes, etc.). Allemaal erg bedankt voor het onvergetelijke afscheid. Voor hun (in)directe bijdrage aan dit boekje, speciaal bedankt: wormengoeroe Rob voor support, gezelligheid en wijze raad op het lab, Arthur voor je bikkelwerk bij de monstercampagnes (hoofdstuk 5–7), diverse meet- en data-analyses

213

(hoofdstuk 5–7 en 9), je goede humeur en de post-monstering-Big-Mac-campagnes, Willie voor projectleiding van het werk dat leidde tot hoofdstuk 5, 9 en 10, de geanimeerde discussies en je uitgebreide (maar onleesbare) commentaar op menig manuscript, Leo voor de hulp bij het naspeuren van Paco’s gegevens en de discussies daarover (hoofdstuk 4) en je commentaar op diverse stukken, Henri voor het steeds weer draaiende houden van de computers, Dik voor menig kritische laat-op-de-middag discussie, en Olivier voor de modelleertips en de broodnodige relativering (“BMI staat toch niet voor Besluiteloos, Moedeloos en Indolent?”). En natuurlijk de secretariële ondersteuning van de Karins, Ellen, en de M&Ms (Marga en Margreet). ☺“Willies Angels” bedankt voor wat jullie op diverse wijze bijdroegen: Jantien bij de bioassays van hoofdstuk 5, Mirjam voor het halen van een koe van een fout uit hoofdstuk 4 (toen de drukproeven van het artikel al de deur uit waren; ik zal nooit meer 10log en ln door elkaar halen!), en mijn inmiddels-weer-collega Martina voor menig wormendiscussie, congressenlol, en de uitleg van het lijmen (en de speurtocht naar weefsellijm). ☺Speciale dank voor mijn “acolieten” Willem en Roel: jammer dat ik jullie niet wat langer heb kunnen begeleiden, maar ik heb er veel plezier aan beleefd en veel aan gehad. Fijn dat jullie co-auteur bij mijn artikelen zijn, en ik bij die van jullie! Willem, ik heb met verbazing gekeken hoe snel je je in MatLab inwerkte, en hoe je de meest esthetische “mokkogrammen” produceerde. Roel, je hebt na mijn vertrek het wormenproject niet alleen voortgezet en uitgebreid, maar ook ge-projectleid (voorwaar geen sinecure als nieuweling bij het RIVM). ☺Van de andere RIVM-labs, speciaal bedankt Elbert van LOC voor je positieve aan-deslag-mentaliteit en de goede analyses, en (hoewel met wat minder direct contact) de LACers voor de metaalanalyses. Verder de CSR-ers voor de prettige samenwerking: met name Theo V. voor de geweldige projectleiding in bange (E)USES-dagen (waar toch de kiem gelegd is voor dit boekje, middels hoofdstuk 3), en Annemarie voor de kritische leiding van de projecten die uiteindelijk leidden tot hoofdstuk 6–8. Van LWD, dank ik de geweldenaar Tom A. voor mijn inwijding in de Bayesiaanse statistiek, wat goed van pas kwam in hoofdstuk 9. ☺Hoewel ik bij IRAS promoveer ben ik er niet zoveel geweest als ik wel gewild had. Meestal was ik dan bij Heather, Elsa en Leon te vinden, waar het goed toeven was. Elsa, ik vond het leuk je met het modelleren bij te staan, en mee te schrijven aan één van je artikelen. Speciaal wil ik natuurlijk Leon bedanken voor de goede en gezellige samenwerking en de lol op de gezamenlijke congressen. Het heeft (letterlijk en figuurlijk) wat voeten in de aarde gehad, maar onze Rotterdamse-bikkel-hoofdstukken “De Eschfiles” (hoofdstuk 6 en 7) staan in onze proefschriften. Ik ben blij dat we dat logistiek, organisatorisch, RIVM-technisch en wetenschappelijk voor elkaar hebben gekregen. Verder wil ik Willem bedanken voor het deelnemen aan de leescommissie. ☺Mijn huidige collega’s, de VU-theoreten, bedankt voor de goede (werk)sfeer, de koffiediscussies en het Stellingbier. En met name Bas en Jacques (jammer dat je alweer vertrokken bent) bedankt voor het oppeppen van mijn wiskundekennis, wat goed van pas kwam voor de laatste hoofdstukken (m.n. voor de profile likelihoods in hoofdstuk 10). Bas, bedankt voor je vertrouwen, en dat je mijn commissie wilde verrijken.

214

☺Verder wil ik veel dank zwaaien aan de heren van de ex-jonge-honden van de ex-ECOruïne-club voor de na-schoolse activiteiten: Hans B., Timo, Theo T. en Lennart. Hierbij natuurlijk speciale aandacht voor Theo: we hebben jarenlang een kamer gedeeld en ik heb veel van je opgestoken qua MatLab en modelleren, en enorm gelachen natuurlijk: “Pak aan!” Daarom ben ik heel blij dat je mijn paranimf wilt zijn. Nog maar even en dan zal ook jouw proefschrift als een mokerslag door ecotoxicologisch Nederland klinken. Lennart bedankt voor menig middernachtelijk discussie over de Ecotox (“Ik MOET met je praten …”) en zijn Godenwereld, het was leuk jouw paardenymfomaan te zijn en aan het promoveren te snuffelen. “Three down, one to go!” ☺Voor het kritisch becommentariëren van manuscripten ben ik verder dank verschuldigd aan diverse mensen. Ik heb ze bij de desbetreffende hoofdstukken vermeld, maar ik wil Angélique, Kees van G. en Dick S. er speciaal even uitlichten. Bedankt voor de tijd om de verhalen serieus door te worstelen, en Dick en Kees bedankt dat jullie in mijn leescommissie wilden plaatsnemen. ☺Thanks to my foreign friends and colleagues for many vivid discussions and good fun on international meetings, while enjoying beers and cigars. Especially: Maike, Kristine, Henriette and Roman. And hi to all the Fellow Earthworm Freaks who were present at the International Workshops on Earthworm Ecotox in Amsterdam and Århus. ☺Omdat het grootste deel van het werk in mijn vrije tijd gedaan moest worden, heeft dit een niet te verwaarlozen invloed gehad op mijn privé-leven. Van mijn vrienden wil ik speciaal de ex-leden van Zeikspoor, Steven en Patrick, er even uitlichten voor menig avondje concert, stappen en weekjes weg. Steven bedankt dat je zo vaak weer een goed idee had om het kluizenaarschap te doorbreken (“Je MOET toch uitgaan!?”), fijn dat je mijn paranimf wilt zijn. Verder sport-billy Robert, die mij aan de aerobics en tae-bo kickpunch heeft gekregen (hoewel dat niet blijvend bleek). ☺Pa en Ma, bedankt voor het mij op de wereld schoppen en het op jeugdige leeftijd stimuleren van mijn biologie-bewustzijn. Pa, dank je voor al de keren dat we met de hond wandelden (onderwijl over de natuur babbelend en eendeneieren tellend), en gingen vissen. Ma, dank je voor het tolereren van de meest uiteenlopende huisdieren (van de obligate cavia’s tot muizen, vissen, kreeften, slakken, woestijnratten …) en mijn eerste vakliteratuur (het veel te dure “Elseviers gids van planten en dieren in de tuin”). ☺Geen betere manier om de zinnen te verzetten dan vanaf de rug van een paard. Bedankt, mijn 4-benige vrienden Lytse en Aran (en Mirjam natuurlijk), Dakota (en iedereen bij White Socks Stables), en de IJslanders (en mensen) van stal Edda. ☺Als laatste degene die waarschijnlijk het meest heeft geleden … Marina, bedankt voor het doorstaan van menig kluizenaarsweekend, het aanhoren van “slimme modelleertruukjes”, je “praktische” oplossingen voor allerhande problemen, het beste verjaarskado ooit (zie vorige punt), en nog vééél meer dan ik hier kwijt wil.

215

Explanation of Symbol Use In this section, I will briefly explain the symbols that are used throughout this thesis. The notation may differ from that in the published manuscripts, as I tried to be consistent in all chapters. Although I have not been entirely successful (especially with the terms BCF and BSAF), the symbols are specified again in each chapter to limit possible confusion. Dimensions are mainly used in Chapter 2, whereas the rest of this thesis specifies units directly. Note that I make a distinction between length of organism and length of environment, because the sum of these lengths is not a meaningful measure. This explains why BCF is never dimensionless (it has the dimensions l3 L-3 or l3 m-3), and why the uptake rate constant is not a real rate constant (it has dimensions l3 L-3 t-1 or l3 m-3 t-1 instead of t-1). Dimensions t time (units e.g. days, hours) # number of chemicals (moles in Chapter 2, otherwise chemical mass) l length (of environment) L length (of organism) 3 l volume of environment (units e.g. metres3 or litres) L3 volume of an organism (units e.g. metres3 or litres) m mass (weight, units e.g. grams or kg) p pressure

For the symbols, the convention is used to apply one symbol for each parameter. In some cases, the definition is used in a slightly different manner in some chapters, e.g. partition coefficients are sometimes applied with dimensions m m-1 instead of l3 l-3. Units are always supplied in Chapter 3–10, to avoid confusion. Symbols A C D F f k K m N dN/dt Q t T V W x Z

ρ θ

216

dim. l2 or L2 # L-3 or # l-3 l2 t-1 – p t-1 l3 l-3 or l3 L-3 l t-1 or L t-1 # # t-1 m t-1 t t L3 or l3 m l # l-3 p-1 m l-3 or m L-3 various

interpretation surface area concentrations (also on mass basis) diffusion coefficient fraction (dimensionless, unless stated otherwise) fugacity rate constants (only ku has a different dimension) partition coefficients (also on mass basis) mass-transfer coefficient (MTC) moles of chemical in a compartment chemical flux (often shortened to dN) feeding rate time (as variable) a specific period of time (a constant) volume of a compartment weight of a compartment distance (for diffusion) fugacity capacity bulk density used to indicate a set of more than one parameter

For the subscripts, the convention of one symbol is loosened to facilitate interpretation. This leads to some inconsistencies, like e.g. Kow would mean the partition coefficient between octanol and water, but Koc is the coefficient between organic carbon and water. There is no w in Koc, but as Koc is always used to denote this particular coefficient, I do not wish to introduce new terminology in this thesis. In some cases a specification is added to the subscript, as in Foc-g which is the fraction OC in the gut contents. Sub. a AE app b com d dig doc / dom e ege f free g i/j ing lag lip m max oc ow ref rem ret s sel sol SPME u w ∞ + 0

interpretation air assimilation efficiency (in FAE the chemical assimilation efficiency) apparent (usually applied together with other subscripts as in Koc-app) body of an organism (e.g. fish or earthworm) compaction (in Fcom the weight decrease during gut transit) biodegradation (in kd) digestion in the gut (in Fdig) dissolved organic carbon (DOC) and dissolved organic matter (DOM) elimination from organism (in ke) egestion (faecal production) food freely dissolved (usually applied together with other subscripts as in Cw-free) gut used as counters or “wildcards”, and represent other subscripts or a number ingestion (feeding) lag time (in Tlag) lipid (as in Flip, the lipid content of the organism) metabolism or biotransformation (in km) maximum (as in ku-max) organic carbon (as in Foc) or OC-water (as in Koc) octanol-water (only in Kow, the octanol-water partition coefficient) parameter in a reference state (an arbitrary reference body volume in ke-ref) remaining solids in the gut (in Frem) retention (in Tret) soil or sediment solids (the dry part) selection of OM in the diet (in Fsel) solids in the earthworm (used in the dwt/wwt ratio Fsol) solid-phase micro-extraction fibres (in CSPME), or SPME-water (in KSPME) uptake into organism (as in ku or dNu-s) water or pore water after a long time (usually referring to steady-state conditions as in Cb∞) sum or total initial (at t = 0)

The abbreviations BCF and BSAF are sometimes also used as symbols, although they are in most situation the same as Kbw and Kbs, respectively. However, in most cases, BCF and BSAF are only used to indicate a normalisation of measured body residues. In Chapter 3, Kbw is used to specify the partition coefficient in L/L, and BCF for the coefficient in L/kg. In Chapter 10, Kbs is a model parameter: the partition coefficient between OM and worm tissue (the actual coefficient that drives the uptake from the soil as well as from the gut contents). However, as uptake is from two routes, the earthworm ends up somewhere between equilibrium with the soil and the gut contents. BSAF is used in this chapter to indicate a secondary model result: the steady-state ratio between worm and soil. 217

“Although they are indifferent to undulations in the air audible by us, they are extremely sensitive to vibrations in any solid object. When the pots containing two worms which had remained quite indifferent to the sound of the piano, were placed on this instrument, and the note C in the bass clef was struck, both instantly retreated into their burrows. After a time they emerged, and when G above the line in the treble clef was struck they again retreated.” Charles Darwin (1881) The formation of vegetable mould, through the action of worms, with observations on their habits.

218