A P H Y S I O L O G I C A L M O D E L T O P R E D I C T X E N O B I O T I C

C O N C E N T R A T I O N S IN FISHES

By

R O N G Y A N G

B. S c., Nanj ing University, 1988

M . S c., Nanj ing University, 1991

A THESIS S U B M I T T E D IN P A R T I A L F U L F I L M E N T OF

T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F

D O C T O R OF P H I L O S O P H Y

in

T H E F A C U L T Y OF G R A D U A T E STUDIES

D E P A R T M E N T O F Z O O L O G Y

We accept tWsJhe^g^Qrjnforming^to the required standard

T H E U N I V E R S I T Y OF BRITISH C O L U M B I A

F E B R U A R Y 1997

© R o n g Yang, 1997

In presenting this thesis in partial fulfilment of the requirements for an advanced

degree at the University of British Columbia, I agree that the Library shall make it

freely available for reference and study. I further agree that permission for extensive

copying of this thesis for scholarly purposes may be granted by the head of my

department or by his or her representatives. It is understood that copying or

publication of this thesis for financial gain shall not be allowed without my written

permission.

The University of British Columbi! Vancouver, Canada

Department

DE-6 (2/88)

A B S T R A C T

A physiological model was developed to estimate fish body toxicant load based on

information regarding the chemical exposure regime, fish body weight, lipid content and

oxygen uptake. Three organic compounds of different hydrophobicity, 1,2,4,5-

tetrachlorobenzene (TeCB), 3,4,5,6-tetrachloroguaiacol (TeCG) and 4,6-dichlorobenzenediol

(DBD), were chosen as the test chemicals to carry out a series of investigations before an

overall model was assembled. The primary focus of the model was to incorporate

physiological components into traditional compartmental models in order to avoid the

difficulties associated with complex conventional physiological models. The approach taken

was to predict rate constants based on fish oxygen consumption, a parameter speculated, and

subsequently shown, to be closely correlated to toxicant transfer in fish.

A significant correlation was found between the toxicant uptake process, as

characterized by the uptake rate constant (ki), and fish oxygen consumption, regardless of fish

size and species. Moreover, the correlation was improved when fish toxicant body load was

expressed on a percent body lipid basis. Similarly, fish toxicant depuration tests showed that

there also existed a significant relationship between the toxicant depuration rate constant (k2)

and fish oxygen uptake regardless of the differences in the chemical octanol/water partition

coefficients ( K o w ) .

The finding that the chosen test compounds did not interfere with fish oxygen

consumption after prolonged sublethal exposure justified the use of oxygen uptake as an

indicator for fish toxicant transfer and, as equally important, the utilization of a large fish

i i

oxygen consumption database ( O X Y R E F ) in the proposed chemical modeling. A series of

feeding experiments were also carried out and it was concluded that fish toxicant transfer

across the gills plays a dominant role in the toxicant accumulation and depuration of non-

metabolized chemicals in fish. Uptake of these toxicants in the food was negligible in

determining body burden.

In view of the above findings a general model was tested in which O X Y R E F was used

to predict fish toxicant body burden. Based on the quantitative analysis, it was shown that the

model was reliable and accurate in estimating fish body burden of a number of non-

metabolized aquatic toxicants. Values calculated using this model agreed with most

determinations reported in the literature. Despite the restrictions and preconditions associated

with this physiological model, its main advantage over other compartmental or physiological

models lies in the fact that the prediction is based on the actual physiological processes, and

fish oxygen consumption rate is far easier and accurate to measure than other physiological

parameters even in the absence of the O X Y R E F . This modified model possesses some

functional reality which enables more realistic predictions, making it useful for aquatic

environmental risk assessment.

iii

T A B L E OF CONTENTS

Page ABSTRACT ii T A B L E OF CONTENTS iv LIST OF TABLES vi LIST OF APPENDIX TABLES vii LIST OF FIGURES viii LIST OF APPENDIX FIGURES xi LIST OF ABBREVIATIONS xii A C K N O W L E D G E M E N T S xiii

CHAPTER I General Introduction 1 Introduction 2 Xenobiotic Uptake and Depuration in Fish 2 Kinetic Models of Bioaccumulation 7

Compartmental models 7 Physiological models 9 Non-compartmental models 10

Development of a Simplified Physiological Model 11 Chemical exchange in gills and K o w 12 Rate constants and fish oxygen consumption rate 14

Chemical Selection and Model Application 16 Test compounds 16 Prerequisites for the model application 17

Effects of toxicant exposure on fish oxygen uptake 17 Relative importance of toxicant uptake through feeding

and breathing 18 Model testing 18

CHAPTER II Correlation between Oxygen Uptake and Chemical Uptake 19 Introduction 20 Materials and Methods 22

Experimental animals 22 Fish respirometer 22 Experimental procedure 25 Sample preparation and extraction 27

Water 27 Fish 28

GC chromatography analysis 30 Lipid analysis 30 Calculations and statistical analysis 31

Results 33 Discussion 43

iv

CHAPTER III Correlation between Fish Oxygen Uptake and Chemical Depuration 47 Introduction 48 Materials and Methods 50

Fish acquisition and acclimation 50 Depuration test protocols 50

Initial toxicant loading 50 Depuration experiment 51

Chemical analysis and calculation 52 Results 53 Discussion 76

CHAPTER IV Prerequisites for the Model Development and Application 79 Introduction 80 Materials and Methods 82

Effects of toxicant exposure on fish oxygen uptake 82 Experimental fish 82 Toxicants 82 Experimental system and procedures 83 Calculation and statistical analysis 87

TeCB and TeCG depuration tests with initial toxicant loading via food ingestion 87

Test fish 87 Exposure regime for trout fry 87 Juvenile trout depuration experiment 88

Results 89 Discussion 100

Effects of TeCB and TeCG sublethal exposure on fish oxygen uptake 100 Relative importance of toxicant uptake through feeding 102

CHAPTER V Model Application and General Discussion 106

APPENDIX Sublethal Toxicity of the Test Compounds 120 Introduction 121 Materials and Methods 122

Exposure regime 122 Sampling procedure 123 Analytical techniques 124 Statistical analysis 124

Results 126 Haematology 126 Plasma sodium and muscle moisture 133 Na7K +-ATPase 133

Discussion 141

LITERATURE CITED 146

v

LIST OF T A B L E S

Table 1. Oxygen consumption by the OxyGuard® probe during 20 min at 7.07 °C.

Page 90

Table 2. TeCB and TeCG depuration profile in juvenile rainbow trout {Oncorhynchus mykiss) after toxicant uptake through feeding. Fish toxicant body burden was expressed as fig toxicant •g fish wet weight"1.

98

vi

LIST OF A P P E N D I X T A B L E S

Page Table 3. Effects of TeCB and TeCG on the haematocnt 127

(Hct) (%), haemoglobin content (Hb) (g-dl"1), mean cell haemoglobin concentration (MCHC) (100-Hb-Hcf !), plasma sodium level (mmoH"1) and gill, kidney Na + /K + -ATPase activity (|imol P,-mg protein-h"1) of adult trout (Oncorhynchus mykiss) during 200 (ig-1"1 TeCB or TeCG exposure and depuration in freshwater.

Table 4. Effects of TeCG on the haematocnt (Hct) (%), 129 haemoglobin content (Hb) (g-dl"1), mean cell haemoglobin concentration (MCHC) (100-Hb-Hct"1), plasma sodium level (mmolT 1), muscle moisture (%) and gill, kidney Na /K + -ATPase activity (|0.mol P, • mg protein-h" ) of adult coho salmon (Oncorhynchus kisutch) during 100 jag-T1 TeCG exposure and depuration in freshwater or continuous exposure for 11 days in freshwater.

vn

LIST OF FIGURES

Page Figure 1. The general framework of this study. 3

Figure 2. Diagram of the fish respirometer and ancillary equipment 23 (from Gehrke et al, 1990).

Figure 3. Standard curves for TeCB, TeCG and D B D G C analysis. 35

Figure 4. Relationship between toxicant and oxygen absorption in 37 fishes of different species and sizes. L=large, M=medium, S=small; RBT=rainbow trout, CS=coho salmon, LSS=largescale sucker.

Figure 5. Relationship between toxicant uptake rate constant (ki) 39 and fish oxygen consumption. L=large, M=medium, S=smali; RBT=rainbow trout, LSS=largescale sucker. Rate constant ki was calculated based on fish toxicant body burden expressed as (ig toxicant-g body weight" .

Figure 6. Relationship between toxicant uptake rate constant (ki) 41 and fish oxygen consumption. L=large, M=medium, S=small; RBT=rainbow trout, CS=coho salmon, LSS=largescale sucker. Rate constant ki was calculated based on fish toxicant body burden expressed as pig toxicant-g lipid"1.

Figure 7. Depuration profile of TeCB for juvenile rainbow trout 54 (Oncorhynchus mykiss) swimming at 2.0 BL-s" 1, 11.6 °C (n=6).

Figure 8. Depuration profile of TeCB for juvenile rainbow trout 56 {Oncorhynchus mykiss) swimming at 2.0 BL-s" 1, 4.6 °C (n=6).

Figure 9. Depuration profile of TeCB for juvenile rainbow trout 58 (Oncorhynchus mykiss) swimming at 2.9 BL-s" 1, 14.3°C (n=6).

Figure 10. Depuration profile of TeCB for juvenile coho salmon 60 (Oncorhynchus kisutch) swimming at 2.0 BL-s" 1, 6.5°C (n=6).

Figure 11. Depuration profile of TeCB for juvenile coho salmon 62 (Oncorhynchus kisutch) swimming at 2.9 BL-s" 1, 9.6 °C (n=6).

viii

Page Figure 12. Depuration of TeCG for juvenile rainbow trout 64

(Oncorhynchus mykiss) swimming at 2.0 BL/s" 1 , 10.6 °C (n=6).

Figure 13. Depuration profile of TeCG for juvenile coho salmon 66 (Oncorhynchus kisutch) swimming at 2.9 BL-s" 1, 11.3 °C (n=6).

Figure 14. Depuration profile of D B D for juvenile rainbow trout 68 (Oncorhynchus mykiss) swimming at 2.0 BL-s" 1, 7.3 °C (n=6).

Figure 15. Depuration profile of D B D for juvenile coho salmon 70 (Oncorhynchus kisutch) swimming at 2.9 BL-s" 1, 9.2 °C (n=6).

Figure 16. Correlation between TeCB, TeCG and D B D depuration 74 rate constant (k2) and oxygen consumption rate in small rainbow trout (Oncorhynchus mykiss) and coho salmon (Oncorhynchus kisutch) (n=6).

Figure 17. A schematic diagram of the six-vessel, intermittent flow- 84 through, computerized fish respirometer.

Figure 18. Oxygen consumption profile of juvenile rainbow trout 93 (Oncorhynchus mykiss) during the 48 h pre-exposure and 48 h TeCB exposure (100, 200 lig-1"1) in a flow-through respirometer (n=12).

Figure 19. Oxygen consumption profile of juvenile rainbow trout 95 (Oncorhynchus mykiss) during the 48 h pre-exposure and 48 h TeCG exposure (100, 200 figT 1) in a flow-through respirometer (n=12).

Figure 20. Comparison of the initial chemical body burdens, ' 103 acquired respectively from 2 h exposure in water and by feeding on toxicant loaded trout fry, in medium size rainbow trout (Oncorhynchus mykiss) (n=6) and the relative importance of the toxicant uptake from water and food.

Figure 21. Comparison between experimental and predicted values 111 for chemicals of different K o w . Measured values for dichlorodiphenyltrichloroethane (DDT) (log K o W = 3.99), sodium dodecylbenzenesulfonate (LAS) (log Kow = 2.24), ethylenediaminetetraacetic acid (EDTA) (log Kow = 2.56) and tetradecylheptaethoxylate (AE) (log K o w = 3.78) are from Bishop

ix

and Maki,1980; Measured values for Di-2-ethylhexyl phthalate (DEHP) (log K o w = 4.35) are from Tarr, et al, 1990; and those for TeCB (log K o w = 4.97) and TeCG (log Kow = 4.41) are from Yang and Randall, 1995.

Figure 22. Quantitative analysis of the predictive power of the 114 simplified model. Each single data point corresponds to the relevant observed and model-predicted values which constitute the dashed and solid lines, respectively, in the graphs of Figure 21.

x

LIST OF A P P E N D I X FIGURES

Page Figure 23. Effects of TeCG on coho salmon haematological factors. 131

Day 0-3 is freshwater exposure (100 (XgT1) and day 3-8 is clean seawater depuration.

Figure 24. Effects of 3-day TeCB and TeCG exposure (200 |xg-l"l) in 134 freshwater on plasma sodium level and gill Na + /K + -ATPase activity of rainbow trout during 24 hr seawater challenge. Asterisk indicates a significant difference (p<0.05).

Figure 25. Effects of TeCG on coho salmon plasma sodium level 136 and muscle moisture content. Day 0-3 is freshwater exposure (100 (J-gT1) and day 3-8 is clean seawater depuration. Asterisk indicates a significant difference (p<0.05).

Figure 26. Effects of TeCG on coho salmon Na + /K + -ATPase activity 139 in gill and kidney. Day 0-3 is freshwater exposure (100 p.g-1"1) and day 3-8 is clean seawater depuration. Asterisk indicates a significant difference (p<0.05).

xi

LIST OF A B B R E V E V I A T I O N S

1CFOK one-compartment, first-order kinetic model

A E tetradecylheptaethoxylate

D B D 4,6-dichlorobenzenediol

D D T dichlorodiphenyltrichloroethane

DEHP di-2-ethylhexyl phthalate

E D T A ethylenediaminetetraacetic acid

L A S sodium dodecylbenzenesulfonate

LC50 median lethal concentration

O X Y R E F oxygen data bank

TeCB 1,2,4,5-tectrachlorobenzene

TeCG 3,4,5,6-tetrachloroguaiacol

xii

A C K N O W L E D G E M E N T S

First of all, I would like to express my sincere thanks to my supervisor, Dr. David J.

Randall, for his guidance, encouragement and support during this study. I am indebted to my

research committee, Dr. George Iwama, Dr. Tony Farrell, Dr. A l Lewis and Dr. Don McPhail,

for their comments on this manuscript and many constructive discussions throughout the

study. My appreciation also extends to all my colleague students who have left or are still in

Dr. Randall's lab. In particular, I would like to thank Dr. Colin Brauner, Dr. Hong Lin, Dr.

Mark Shrimpton, Mr. Nick Bernier, Ms. Joelle Harris, Mr. Jonathan Wilson and Ms. Ingrid

Burgetz, for all their help and pleasant company. I feel grateful for the assistance from Dr.

Tony Farrell and Dr. Chris Kennedy's labs at Simon Fraser University in providing the

equipment and fish required for the study described in chapter IV, and the help from Dr. R. V.

Thurston's lab at Montana State University in doing some of the chemical analyses.

Last but not least, I would like to thank my wife Y i , my son Jason, and my parents for

their incredible inspiration and support.

xiii

C H A P T E R I

General Introduction

INTRODUCTION

The release of synthesized organic compounds, known as xenobiotics, into the aquatic

environment is a great threat to fish and other aquatic life. Chemicals at very low

concentrations in water can accumulate in fishes and other aquatic animals and have a toxic

action. The deleterious toxic effects can occur at different biological levels depending on the

exposure regime and the specific organisms involved. These biological changes, whether at

the cellular, organ, individual or population level, are usually associated with the

accumulation of toxicants within the body. Toxicant body burden, therefore, is of particular

interest and importance in understanding the behavior and impact of toxicants on fish. In

addition, because of the hazardous effects on humans, i f the toxicants are transferred along

the food chain, the prediction of fish body burden is of importance in the regulation of

fisheries in contaminated waters.

The main objective of this study was to develop a general, accurate and easy-to-apply

model to estimate fish toxicant load at any given time, based on information regarding water

quality, chemical exposure and physiological parameters of exposed fish. The general

framework of the study is described in Figure 1.

X E N O B I O T I C U P T A K E A N D D E P U R A T I O N IN FISH

Xenobiotics are produced by various types of industrial processes, for example those

associated with pulp and paper production, food processing, and the petrochemical industry.

2

Hypothesis I: A simplified physiological model can be

developed to predict xenobiotic concentration in fishes

Chemical selection and model development

Hypothesis II: Fish oxygen uptake is indicative of toxicant uptake across the gill and toxicant uptake rate constant (ki) is

correlated with oxygen uptake

Hypothesis IV: The relationship between ki, k 2 and oxygen uptake is applicable to a group

of chemicals with different octanolAvater

partition coefficient (K o w )

Hypothesis III: Fish oxygen uptake is indicative of toxicant depuration across the

gill and toxicant depuration rate constant (k2) is

correlated with oxygen uptake

Hypothesis V: Toxicant sublethal exposure will not interfere

with fish oxygen uptake

1 Feasibility of using an oxygen data base in the

chemical modeling

1 Hypothesis VI:

Toxicant uptake via food is much less important compared with direct uptake through breathing.

Thus, fish gill is the major site of toxicant transfer

Model Application

Figure 1. The general framework of the study

These chemicals are often released into the aquatic environment and toxicity varies markedly.

Those compounds that have a high lipid solubility readily diffuse across the surface

membranes of animals and accumulate in high concentrations in adipose tissue. Many of these

compounds also resist breakdown and persist in both the environment and the animal. It is

these special physico-chemical properties, e.g. hydrophobicity and tendency to resist chemical

and bio-transformation (Connell, 1990) that make these chemicals hazardous. Uptake of

xenobiotic compounds in fish may involve direct transfer from water to the body and/or

absorption from contaminated food. Many investigations of DDT and related chemicals

have shown an apparent increase in toxicant body burden concentration with trophic level,

which is evidence of food chain biomagnification (Rudd and Genelly, 1956; Zitko et al,

1972; Robinson et al, 1967; Woodwell et al, 1967), especially in carnivorous fish. In

addition, feeding affects not only chemical uptake, but also elimination (Jimenez et al,

1987), causing the clearance of toxicants with the faeces, in addition to elimination through

gills and kidneys as would occur in unfed fish. No clear evidence, however, has been reported

of an increase in the concentration of xenobiotics with trophic level (Shaw and Connell, 1982;

Robinson et al, 1967).

Most organic chemicals must enter the body before they can exert their toxic effects.

Uptake of these chemicals during bioconcentration is, in most cases, by diffusion. The

principal surface across which the toxicant diffuses depends on the size of the animal. In

juvenile and adult fish, the gill constitutes the majority of the animal's surface area (Murphy

and Murphy, 1971; Rombough and Moroz, 1990) and is the major site of entry of toxicants.

Although cutaneous uptake of xenobiotics may occur (Saarikoski et al, 1986) and could play

4

a more important role in small fishes, most of the work done so far tends to ignore the

significance of this pathway to the development of the toxicant body burden (Murphy and

Murphy, 1971).

A large volume of water flows across the gill epithelium, which is a thin barrier with

a large surface area, allowing the efficient exchange of O2 and C O 2 between the water and

blood. In general, conditions at the gills in fish are adjusted to meet the metabolic demand for

oxygen (Randall, 1990). Oxygen and many xenobiotics are lipid soluble molecules which

diffuse across the gill lamellae transcellularly. It is likely that variables such as ventilatory

volume, gill epithelial thickness and perfusion, which influence the rate of oxygen transport

across the epithelial membrane, wil l similarly influence toxicant movement. That is, the

conditions for oxygen transfer are likely indicative of that for toxicant transfer. In fish,

therefore, the main route for toxicant entry is postulated to be across the large surface area of

the gills (Holden, 1962; Murphy and Murphy, 1971; Randall and Brauner, 1990; Erickson and

M c K i m , 1990) and may vary with the rate of oxygen uptake.

The hypothesis that xenobiotic uptake may be correlated with oxygen uptake,

originated from a study by Murphy and Murphy (1971). Subsequently, Black and McCarthy

(1990) found that oxygen and P C B uptake efficiencies were directly correlated under a

number of conditions. In other studies with trout exposed to benzo[a]pyrene, naphthalene and

2,2',5,5'-TeCB during an acute decrease in temperature, a correlation was established between

oxygen and contaminant uptake (Black et al, 1991). Brauner et al. (1992, 1994) and Thurston

et al. (1992) reported a significant relationship between 1,2,4,5-tetrachlorobenzene (TeCB)

uptake and oxygen consumption rate during initial exposure to TeCB under normoxic

5

conditions in several freshwater teleost fish of different sizes. Randall et al, (1993) obtained

similar results using large and small rainbow trout and suckers exposed to the toxicant TeCG

(tetrachloroguaiacol). The results of all these experiments indicate that factors which alter gill

membrane diffusing capacity and/or fish gill ventilation may have a similar effect on the

uptake of both oxygen and hydrophobic xenobiotics.

The uptake of xenobiotics directly from water by fish is determined by numerous

factors, the most dominant of which are the physico-chemical properties of the compound and

the physiological process which occurs at the gills. The physico-chemical characteristics

which may influence the flux of xenobiotics across biological membranes include water and

lipid solubility, molecular weight and volume, fugacity, tendency to ionize, and susceptibility

to metabolic transformation. Among all the rate-limiting physico-chemical factors of an

organic chemical, the most important ones in predicting toxicant movement across gill

epithelium membranes are K o w and the active concentration of the compound, which includes

both the concentration of the aqueously dissolved chemical and the level of the non-

dissociated form if the toxicant is a weak acid.

The term depuration has been associated with a specific set of experimental condition,

i.e., loss of chemical from the animal after being placed in water without the chemical (Barron

et al, 1990). Overall, the polluted fish can have toxicant eliminated via either the gills, or the

faecal matter, or through metabolic transformation (Clark et al, 1990). As mentioned above,

the gill is the main route of toxicant uptake directly from water by fish. Heath (1987) argued

that the fish gill is also an important mode of exit for small compounds that require little

biotransformation before excretion, and that this is presumably a passive process. For

6

example, the gills turned out to be an especially important organ for elimination of aromatic

hydrocarbons and the rate of excretion of individual hydrocarbons by the gills was shown to

be inversely proportional to the size of the molecule (Thomas and Rice, 1981). There is a

potential, therefore, that depuration rate could be related to oxygen consumption, as is uptake

rate, although the possibility has never been investigated.

KINETIC M O D E L S OF BIO A C C U M U L A T I O N

The mathematical characterization of chemical absorption, distribution and elimination

processes permits a quantitative prediction of the amounts and concentrations of a chemical in

the body of an animal as a function of certain parameters. A number of predictive models

have been developed for both mammals and aquatic organisms in the past few decades. A l l

the models already established fall into three categories: Compartmental; Physiologically-

based; and Non-compartmental models.

Compartmental model:

A compartmental model is a simplified mathematical description of a chemical's

behaviour in an animal, where the body is represented as a system of compartments. This type

of model has been predominantly applied in aquatic toxicology because, despite their

limitations, they allow an essential description of the kinetics of xenobiotics in aquatic

animals. The simplest and the most commonly used is a one-compartment, first-order kinetics

(1CFOK) model which has been argued by some researchers to be the standard for

7

pharmacokinetic analysis in aquatic toxicology (Spacie and Hamelink, 1982). The 1CFOK

model is actually a rate constant-based model with the basic relationships expressed as

follows:

for uptake: Cf=C w(ki/k 2)(l-e" k 2 t); (1)

and for elimination: C^CVe" 1 ^ (2)

where Cf: concentration of the chemical in the fish at time t;

C w : water concentration;

C t 0 : concentration in the fish at time zero;

k i , k 2: uptake and elimination rate constants, respectively.

The uptake of chemicals from the diet can be described with a similar model.

Bruggeman et al. (1981) used the following equation to describe bioaccumulation after dietary

exposure.

Cf(t)=Ef/k2C f d(l-exp(-k 2t)) (3)

where Cfd = concentration of chemical in food;

f = feeding rate(food weight - fish weight"1 • time"1);

E = absorption efficiency for ingested chemical;

Constant dietary exposure results in increasing fish concentrations until a plateau level

is reached when the clearance rate equals the uptake rate. The ratio between the concentration

in fish and in the food at steady state is given by the biomagnification factor K m :

km=Cf(t)/C f d=Ef/k 2 (4)

More recently, Gobas (1993) has suggested a comprehensive model which combines all the

possible factors that may affect the chemical concentration within a fish body in an overall

flux equation, describing the net flux of chemical into the fish as the sum of all of the uptake

and loss fluxes:

dCf/dt = k fCw + k D C f d - (k2 + k E + k M + k G ) C f (5)

where k i , k 2 = uptake and elimination rate constants via gills;

k D = uptake rate constant from food;

k£ = elimination rate constant by faecal egestion;

k M = metabolic transformation rate constant of the chemical.

The application of these relationships requires three assumptions: 1) The system

operates by first-order kinetics, indicative of passive diffusion of the toxicant into the

organism; 2) A steady-state can be reached; and 3) The body can be treated as one well-mixed

compartment, with the rate of distribution of the toxicant within the organism exceeding the

rate of exchange with the surroundings.

Although the model should be applied with some appropriate restrictions, the 1 C F O K

model approach can provide a useful simple approximation. In order to estimate steady-state

chemical concentrations in fish, values are required for all the rate constants (k's) for different

chemicals and fish species, etc.

Physiological model:

A physiologically-based pharmacokinetic (PBPK) model incorporates the underlying

physiological processes and important tissues involved in chemical deposition. Hayton and

Barron (1990) have proposed a physiological model to describe xenobiotic uptake across the

fish gill:

9

Uptake rate = (C w -C f ) [(d(Dm ') A Km)+(h(D a- 1)A)+V b(K b- 1+Vw ' 1)]- 1 (6)

C w and Cf = concentration of the chemical in the external water and the plasma

water, respectively;

d = aqueous stagnant layer thickness;

A = gill surface area;

Vb = effective gill blood flow;

V w = effective water flow;

D m = diffusion coefficient in epithelium;

D a = diffusion coefficient in water;

k m = blood/water distribution coefficient.

This model consists of three components: 1) The concentration gradient between the

environment and the fish; 2) Physiological and anatomical characteristics of the gills; and 3)

Physical constants specific to the compound in question. It is apparent that the physiological

model requires detailed information not only about the properties of the chemical but also

about the animal, and data collection of the latter is both technically difficult and time-

consuming. A n advantage of the P B P K model is that once the dominant transport

mechanisms are characterized, it could provide a rational basis from which to extrapolate the

model to other conditions and species.

Non-compartmental model:

Non-compartmental pharmacokinetic analysis avoids many of the assumptions of

compartmental modelling, while characterizing capacities for elimination, storage and

10

persistence. Non-compartmental analysis has received only limited application in aquatic

toxicology, apparently due both to the method and the requirement for blood or plasma

concentration-time data.

D E V E L O P M E N T OF A SIMPLIFIED P H Y S I O L O G I C A L M O D E L

There are a large number of models predicting organic toxicant bioaccumulation in

fish. These models, however, are very complicated and not easily applied to practical

biomonitoring. A simplified physiological model needs to be developed to predict toxicant

amount in a fish body. Fish oxygen consumption has the potential to be used as an indicator of

toxicant transfer.

The overall objective of this research is to develop a simple model in which some

easily measured fish physiological parameters and certain chemical properties are

incorporated to predict the amount of xenobiotics in a fish body with acceptable

approximation. The presumption of this study is that the fish gill dominates as the xenobiotic

exchange site between the fish body and the ambient environment during both uptake and

depuration periods, as discussed earlier.

The most commonly used one-compartment, first order kinetic (1CFOK) model was

chosen as the basis of this work. As this is a rate-constant based model, effort was directed

toward the prediction of uptake and depuration rate constants according to some easily

measured physiological parameter so as to avoid the complicated measurements necessary for

the physiological models developed so far.

11

Chemical exchange in gills and octanol/water partition coefficient (Kow):

Chemicals are delivered to the gill surface by an unidirectional water flow over the

gills; they diffuse across the gills and are then distributed to the tissues by the blood flow.

Potentially water flow, blood flow and/or diffusion across the gill epithelium could limit

uptake of the chemical by the fish (Randall and Brauner, 1990). The ability to deliver a

chemical to the gills wi l l depend on gill ventilation and the water solubility of the chemical,

whereas diffusion across the gill epithelium wil l depend on its lipid solubility. Lipid content in

fish blood is approximately 5%, so during initial exposure to a xenobiotic with log K o w above

2, almost all of the xenobiotic in the blood wil l be bound to lipids and proteins (Schmieder

and Henry, 1988), and the ability of the blood to remove the toxicant wil l far exceed the

ability of water to deliver it. Thus, transfer of these xenobiotics into the fish is likely to be

limited by either the volume flow of water over the gills (ventilation limited) or by the rate of

diffusion across the gills (diffusion limited).

The rate at which chemicals are absorbed by fish via the gills is expressed by the gill

uptake rate constant ki (1-kg-day'1). The kinetic elimination process is called the depuration

rate constant k 2 (day"1). Some authors have reported that the uptake and clearance rate

constants of some lipophilic compounds by various fish species have a fixed relationship to

the Kow over a certain range (Hawker et al, 1985; Saarikoski et al, 1986). Studies of the

relationship between ki and K o w in fish have shown that 1) ki increases with K o w i f logKoW is

low (<4); 2) ki is constant i f l o g K o w is large (between 4 to 6); and 3) ki drops with increasing

K0w for chemicals with extremely high K o w (log K o w above 6) (McKim et al, 1985; Gobas et

al, 1986; Gobas and Mackay, 1987). Hawker et a/.(1985) reported a very good linear

12

correlation between k 2 and logK o w , within the range from 2.5 to 6.0, as depuration rate

constant decreases with increased K o w .

A n elevation in K o w could result from either an increase in lipid solubility or a decrease

in water solubility. Dobbs and Williams (1983) showed that there was a linear relationship

between water and fat solubility, but a steep inverse relationship between water solubility and

logKow. That is, high l o g K o w values were associated with very low water solubilities. So, the

xenobiotics with high K o w should be termed as extremely hydrophobic rather than lipophilic,

and transfer of these compounds into fish, therefore, is probably ventilation limited. For

chemicals with low log K o w (<4), the capacity of the fish to deliver the chemicals to the gill

surface is large because of the high water solubility, but diffusion across the gill epithelial

membrane plays a critical role in the uptake process. In other words, toxicant uptake tends to

be diffusion limited, hence the uptake rate constant of chemicals in this category increases

with K o w . For chemicals with log K o w greater than about 4, the importance in the ability of the

fish to deliver the chemical to the gills, due to the much reduced water solubility of the

chemicals, wi l l overshadow the diffusion capacity across the gills in the uptake process. It is

not surprising, then, to see the independence of uptake rate constant on K o w because uptake of

these chemicals is ventilation limited. The hydrophobicity of compounds with l o g K o w above

about 6, however, results in an increased importance of binding to any organic matter present

in the inhalant water (Black and McCarthy, 1988) or to micelle formation. The observed drop

in uptake in ki for extremely high K o w substances, therefore, may be due to reduced

bioavailability and/or experimental error associated with the difficulty in measuring the

concentration of the aqueously dissolved chemical (Gobas and Mackay, 1987).

13

During depuration, the toxicants already stored in the lipid within the tissues will be in

equilibrium with the blood stream which serves as a transporting compartment, for

compounds that eventually wil l be removed from the fish body across the gills. It has been

shown that depuration rate constant k2 remains more or less the same for chemicals with a

low log K o w (< 4) and then gradually decreases along with increasing K o w (Gobas et al,

1986). In the case of chemicals with log K o w lower than about 4, the increased lipid solubility

and decreased water solubility of a specific chemical wil l have opposite impact on the overall

depuration rate, and it is likely that the combined effect of these two gives rise to a fairly

constant k 2 for toxicants with low K o w . For chemicals with log K o w above about 4, the

solubility of the chemical in the aqueous phase may dominate the depuration process such that

transport is ventilation limited rather than diffusion limited. Aqueous solubility decreases with

increasing K o w and results in a decreased depuration rate.

Rate constants and fish oxygen consumption rate:

Oxygen consumption rate can easily and accurately be measured and there is some

evidence already, both theoretical (Randall and Brauner, 1990) and experimental (Black and

McCarthy, 1990; Black et al, 1991; Brauner, et al, 1992; 1994; Thurston et al, 1992) that

oxygen uptake may be an indicator of toxicant transfer across fish gills. The notion of using

oxygen uptake together with other chemical properties to predict the amount of xenobiotics in

fish body is theoretically sound, provided that a correlation exists between fish metabolic rate

and toxicant uptake/depuration kinetic processes.

14

Randall and Brauner (1990) related oxygen and toxicant uptake rates, and derived a

toxicant transfer coefficient (k) for the chemical, TeCB, during an initial exposure period.

This coefficient was expressed in mg 02 -kg-fish' 1 ,h" 1 exposure per mgT 1 chemical gradient

(water to blood); i.e., the uptake of toxicant per unit toxicant gradient per unit oxygen uptake

(mg-kg"!,h"1). Thus, the actual unit of this toxicant transfer coefficient is hour"1.

In rate constant based models, the proportionality constant that relates the rate of

change of either the amount or concentration of chemical within a compartment to the driving

force is a first-order rate constant (kj, units of time"1) (Barron, 1990). The driving force may

be expressed as a gradient of either concentration or amount, but the choice must be

consistent with the rate expression. A rate constant, therefore, is simply a fraction rate of

removal of chemical per unit time.

From a toxicological point of view, the toxicant transfer coefficient (X) derived by

Randall and Brauner (1990) is, as a matter of fact, an uptake rate constant (ki). Although the

linear relationship in this study between oxygen consumption and toxicant uptake and is

generated during initial exposure, it does provide some basis for the hypothesis that fish

metabolic rate is indicative of the entire toxicant uptake and depuration kinetic processes

occurring at fish gills. That is to say, the first order rate constants may be closely correlated

with oxygen uptake during either toxicant uptake or elimination, because of the similar

physiological process taking place in both cases.

As mentioned above, for chemicals with l o g K o w less than about 4, the uptake rate

constant (ki) increases with K o w but depuration rate constant (k2) remains fairly constant, and

for chemicals with l o g K o w above 4, it is the opposite; i.e., ki does not change while k 2 drops

15

along with increased K o w . Thus, the relationship between oxygen consumption and ki derived

from a specific type of chemical may only be applied to a certain log K o w range of the

hydrophobic xenobiotics. This correlation, therefore, needs to be further substantiated not

only with fishes of different sizes and species, but also with chemicals of different physico-

chemical properties ( K o w ) . If fish oxygen uptake can be used to predict the uptake and

depuration rate constants for a variety of xenobiotics with different K o w values, this would

make it possible, with the availability of a large number of oxygen uptake measurements in

the literature, to predict the amount of xenobiotics in a fish body by incorporating fish

metabolic rate, octanol-water partition coefficient and possibly other parameters into a first

order one compartment kinetic model.

C H E M I C A L S E L E C T I O N A N D M O D E L A P P L I C A T I O N

Test compounds:

To study the correlation between toxicant transfer and fish metabolic rate, some test

chemicals were chosen as representative of a group of xenobiotic compounds. The test

chemicals selected were based on the following criteria: a) They were not metabolized; b)

They covered a range of octanol/water partition coefficients ( K o w ) ; and c) They were

significant in industrial production. Three chemicals, 1,2,4,5-tetrachlorobenzene (TeCB),

3,4,5,6-tetrachloroguaiacol (TeCG) and 4,6-dichlorobenzenediol (DBD), were chosen as the

test compounds. TeCB and TeCG are stable, non-volatile, not easily metabolised and have

relatively high hydrophobicity, their octanol-water partition coefficient ( K o w ) being 4.97 and

16

4.41, respectively, at 25°C. 1,2,4,5-tetrachlorobenzene (TeCB) and tetrachloroguaiacol

(TeCG) belong to a group of chlorinated compounds produced by the pulp and paper industry

in British Columbia (Garrett 1980). The presence of chlorinated compounds in kraft pulp mill

bleach effluent in northwestern Canada was reported as early as 1975 (Leach and Thakore,

1975). These two chemicals, although considered lipophilic compounds, are sufficiently water

soluble that they are bioavailable for fish (U.S. Environmental Protection Agency, 1994). It

was reported, as discussed earlier, that the uptake and clearance rates of some lipophilic

compounds by various fish species are associated with their K o w (Hawker et al, 1985;

Saarikoski et al, 1986). D B D has a much lower lipid solubility (log K o w = 2.35) and was

selected to represent less hydrophobic chemicals.

Prerequisites for the model application:

Effects of toxicant exposure on fish O^uptake:

The pre-requisite of using fish oxygen uptake as an indicator of toxicant transfer, as

mentioned earlier, is that it should not be affected by chemical exposure. Only under this

situation can the oxygen uptake values in the literature, derived in the absence of any chemical

exposure, be applied in the proposed model. Therefore, the effects of the chosen chemicals on

the oxygen consumption of fish during prolonged exposure must be investigated to determine

if toxicant exposure is a complicating factor when using fish oxygen uptake as the indicator of

toxicant uptake and depuration.

17

Relative importance of toxicant uptake through feeding; and breathing:

The enormous amount of substance exchange continuously occurring across fish gills

tends to support the hypothesis that direct uptake from water is dominant in fish toxicant

accumulation. Toxicant intake via food, however, may also contribute to the total toxicant

load, especially at high feeding rates. In order to evaluate the relative importance of toxicant

uptake from food and to test the hypothesis that the impact of toxicant uptake via food on the

total body burden is negligible, feeding experiments were conducted using adult rainbow trout

and small salmon fry.

Model Testing:

In this rate constant (ki and k 2) based model, it becomes a matter of predicting the

uptake and depuration rate constants according to fish oxygen consumption rate. A model test

was conducted to look at the feasibility of predicting chemical concentration in fish by

incorporating the relationships established between fish oxygen uptake and kinetic rate

constants (ki and k 2 ) into the one compartment, first order kinetic model. The main advantage

of this modified 1CFOK model over the other compartmental models is its basis on actual

physiological processes. Also, fish oxygen consumption rate is well documented in the

literature, and is far easier and accurate to measure than some of the parameters required by

other physiological models, such as gill surface area and blood flow, epithelial thickness and

diffusion coefficients. Based on the outcome of the modeling trial, the implication and

limitations of the simplified model are discussed at the end of this thesis.

18

C H A P T E R II

Correlation between Oxygen Uptake and Chemical Uptake

INTRODUCTION

Uptake of xenobiotic compounds during bioconcentration, defined as the direct entry

of chemicals across the respiratory surface (Connell, 1990), is most often by diffusion. In

juvenile and adult fish, the respiratory epithelium forms an extremely large and thin barrier

between the water and blood to permit the efficient gas exchange and comprises the majority

of the body surface area (Murphy and Murphy, 1971; Rombough and Moroz, 1990). Due to

the much lower oxygen content per unit volume in water than in blood (Cameron and Davis,

1970), fish gills are hyperventilated with water relative to blood flow to ensure that oxygen

delivery to the gills by water is matched to the oxygen transport capacity of the blood. Thus

the gills are most likely the main site of xenobiotic transfer.

Oxygen is a lipid-soluble molecule which diffuses across the gill lamellae

transcellularly as do most hydrophobic xenobiotics. It is likely that variables such as

ventilatory volume, gill epithelial thickness and perfusion, which influence the rate of oxygen

transport across the epithelial membrane, wil l similarly influence toxicant movement. Not

surprisingly, it has been postulated that the conditions for oxygen transfer could be indicative

of those for toxicant transfer (Holden, 1962; Murphy and Murphy, 1971). Black and

McCarthy (1988) found that, in rainbow trout, oxygen and P C B uptake efficiencies from

water flowing over the gills were directly correlated over a wide range of uptake efficiencies,

ranging from 20 to 80%. In other studies with trout exposed to benzo[a]pyrene, naphthalene

and 2,2',5,5'-TeCB during an acute decrease in temperature, a correlation between oxygen and

contaminant uptake was established (Black et al., 1991). Brauner et al. (1994) reported a

20

significant relationship between 1,2,4,5-tetrachlorobenzene (TeCB) uptake and oxygen

consumption rate during initial exposure to TeCB under normoxic conditions. These findings

imply that factors which alter gill membrane diffusion properties and/or gill ventilation may

have a similar effect on the uptake by the fish of both oxygen and hydrophobic xenobiotics.

It has been recommended that fish lipid content be taken into account in modeling

bioaccumulation (Dobbs & Williams, 1983; Connell, 1990), since most of the xenobiotics

have relatively high lipid solubility and, as such, wil l tend to be highly concentrated in the

adipose tissues relative to the aqueous phases of the fish body. Thus toxicant uptake may be

modulated by lipid content of the fish.

The objectives of the study described in this chapter were 1) to investigate whether the

rate of toxicant uptake was correlated with the oxygen consumption of various species of fish

of different sizes and lipid content; and 2) to test the hypothesis that toxicant uptake kinetic

process, characterized by toxicant uptake rate constant (ki), is dictated by fish oxygen

consumption.

21

M A T E R I A L S A N D M E T H O D S

Experimental animals:

Fish used in this study were small (5.8 ± 0.32 g) rainbow trout (Oncorhynchus

mykiss), medium size (25.35 ± 0.65 g) rainbow trout and coho salmon (Oncorhynchus

kisutch) and large rainbow trout (425 ± 22.6 g) and were obtained from fish farms in British

Columbia. Large fish were held in 10,000 L , and smaller fish in 200 L fibreglass tanks

containing dechlorinated Vancouver city tap water, circulating at approximately 1 body length

per second (BL-s"1), for at least two weeks prior to experimentation. Water temperature varied

seasonally from 5 - 10 °C. A l l experiments were performed at the temperature to which the

animals had been acclimated. Following arrival, fish were fed commercial pellets bi-weekly

ad libitum, but were starved at least 72 h prior to introduction to the experimental conditions

to ensure they were in a post-absorptive state. In order to reduce handling stress, visual

estimates of length and weight were made before the tests and actual parameters determined

upon completion of the tests.

Fish respirometer:

A 130 L modified Brett type respirometer (Fig. 2) was used to force fish to swim at

predetermined velocities and is described in detail by Gehrke et al. (1990). The swim chamber

is a plexiglass tunnel, 1000 mm in length and 200 mm in diameter, with stainless steel mesh

at each end. A n opaque plastic covering was placed around the middle of the swim tube to

entice the animals to swim in this region.

22

Figure 2. Diagram of the fish respirometer and ancillary equipment (from Gehrke et al., 1990)

23

24

The large volume of this respirometer permits a fairly constant water toxicant

concentration over time; however, accurate oxygen consumption measurement require small

volume. Small fish, therefore, were swum in groups of 20 -25 to ensure the combined body

mass exceeded 100 g, the lower limit determined for accurate oxygen uptake measurement.

Before the introduction of the fish to the respirometer, water temperature was adjusted

to that of the holding tank, the PO2 was calibrated and the PO2 of the respirometer was

increased to 100% air saturation levels. A l l air bubbles were removed from the swim tunnel,

the fish were placed in the swim chamber and forced to swim at 18 cnrs"1 (the minimum flow

velocity of the system) for the introductory period of two hours for all the tests unless

otherwise specified. During this period there was a slow continuous overflow of

dechlorinated city water to prevent the accumulation of metabolic wastes.

Experimental procedure:

At the time of the test, individual large fish or groups of smaller fish were introduced

into the respirometer and acclimated for at least two hours at low swimming speeds prior to

chemical dosing. Following this introductory phase, a pre-mixed solution of 1,2,4,5-

tetrachlorobenzene (TeCB), or 3,4,5,6-tetrachloroguaiacol (TeCG), or 4,6-

dichlorobenzenediol (DBD), dissolved in approximately 67 ml of acetone was slowly added

downstream of the test animals into the swim tube to make the nominal toxicant level of 200

ug-1"1. The water velocity was then gradually increased over the next 5 min to the final

swimming speed (ranging from 1.3 to 3.9 BL-s" 1 depending on the size and status of the test

fish) required for each particular test. Duplicate water samples were taken in 40 ml I-Chem

25

vials with teflon lined lids 5 min after the addition of the chemical. The system was

subsequently sealed for the rest of the 2 h test duration when fish oxygen consumption rate

was measured.

The 2 h uptake test period was chosen for several reasons, primarily due to the fact that

this was long enough to measure the differences in the dissolved oxygen levels in the

exposure water as long as the total body weight of the test fish exceeded 100 g. Also, during

the initial stage of the uptake, e.g., the first two hours, it was unlikely that the rate of the

toxicant uptake by the test fish would change to a measurable extent, making the calculation

of the uptake rate constant easy to achieve. At the end of each 2 h test period, fish oxygen

consumption was calculated according to the following equation: Oxygen uptake = Decrease

of dissolved oxygen in the water during the test period (mg) / total body weight of fish

swimming in the tunnel (kg) / time (h).

Final water samples were taken in duplicate at the end of each test, and the fish were

removed from the respirometer and terminated with a sharp blow to the head. Then the fish

were rinsed quickly in freshwater and lightly blotted dry with paper towel. Each fish was

weighed and measured for total and fork length, and then placed in a pre-labeled plastic bag.

A l l water and fish samples were stored under -80°C until later analysis for toxicant and/or

lipid content.

26

Sample Preparation and Extraction:

a) Water

TeCB

Aliquots of the water samples were added to 100 ml square cross section volumetric

flasks containing 50 ml of laboratory water and a teflon coated magnetic stirrer. Hexane (10

ml) was added and TeCB was extracted by stirring the mixture for at least 30 min. After a

settling period, the hexane layer was forced into the neck of the flask by the addition of more

water. A 1.0 ml aliquot of this hexane extract was transferred to a 10 ml volumetric flask and

exactly 0.5 ml of internal standard solution (Pentachlorobenzene) (50 ugT 1 ) was then added.

This solution was made to volume with hexane and TeCB concentrations were determined by

GC.

TeCG and D B D

A different method was required for TeCG or D B D sample extraction. According to

the exposure concentration, aliquots of water samples were placed in 100 ml square-shaped

volumetric flasks containing 0.5 ug surrogate, ca. 50 ml double distilled water, 2 ml 18N

sulfuric acid, 10 ml 1:1 methyW-butyl-ether/hexane, and a magnetic stir bar. Extraction

proceeded by mixing the flask contents on a magnetic stirrer for 30 min. Extraction solvent

was then forced up into the neck of the flask by addition of more water, and 1.0 ml of this

solvent was transferred to a 10-ml volumetric flask. Diazomethane was removed, internal

standard was added and the contents diluted to volume with hexane. Method blanks and

TeCG or D B D (0.5 ug) fortified blanks were analyzed along with water samples by GC.

27

b) Fish

TeCB

Small whole fish (10 g), or sub-samples of homogenates of larger fish ground up in a

conventional meat chopper, were blended to a fine powder using a Polytron™ tissue

homogenizer (Model #PT-10/35, Brinkmann Instruments Inc., Westbury, New York) with the

aid of dry ice and anhydrous Na2SC>4 while still frozen. At least two separate samples were

prepared from homogenates of medium and large fish. Only single samples were prepared

from small fish because of limitations in sample quantity. TeCB was extracted (>8 h) using a

soxhlet extraction apparatus with Friedrichs condenser including 24/40 250 ml capacity flask

( V W R Scientific) from each sample using hexane. Surrogate (1,2,3,4-TeCB) in amounts

comparable to the amount anticipated for TeCB results (usually 100 ug) was added at the

initiation of the extraction step. Lipids were removed from portions of this extract using

florisil column chromatography; TeCB and surrogate were eluted with 5% methyW-butyl

efher/hexane. Pentachlorobenzene (PCB) internal standard was added to the purified extract

and component concentrations were determined by electron capture gas chromatography

(ECD-GC).

TeCG and D B D

Small fish were weighed and then diced into small pieces while still frozen. These

small pieces were placed into a tared square-shaped extraction bottle and weight recorded.

Medium and large fish were partially thawed, cut into several pieces and then ground using a

conventional kitchen meat grinder. The ground sample was further mixed to achieve as

28

homogenenous a tissue sample as possible. TeCG or D B D analyses were performed, in

duplicate, on portions of the homogenate. At the time of extraction, weighed aliquots (ca. 5 g)

of homogenates were placed in square wide mouth extraction bottles. Each sample, either

homogenate aliquot or diced small fish, was acidified by the addition of 2.0 ml 18N sulfuric

acid followed by thorough mixing. Surrogate (pentachlorophenol) was then added to each

sample followed by blending this mixture for 30 seconds using a Polytron™ tissue

homogenizer. The extract was decanted through anhydrous sodium sulfate into a 250 ml

volumetric flask. The residue was extracted again, both extracts were combined, and then the

volume adjusted to 250 ml with additional extraction solvent.

Aliquots (5 ml) of extract were taken to near dryness with the aid of a gentle stream of

dry nitrogen and low heat. The residue was immediately methylated in a diazomethane

solution generator (>30 min) using approximately 3 ml ethereal diazomethane solution.

Excess diazomethane was removed by passing a gentle stream of nitrogen over the sample

and then the entire contents were quantitatively transferred to a 11x300 mm glass column

filled with 5 g florisil (60-80 mesh) topped with 1-2 cm of anhydrous sodium sulfate. These

prepared columns were stored at 130 °C for at least 24 h prior to use. Just before sample

addition, each column was rinsed with several milliliters of 5% methyW-butyl ether/hexane

solution (elution solvent). Methylated chemicals and surrogate were eluted with 40 ml elution

solvent and collected in a 100 ml volumetric flask. Hexachlorobenzene internal standard

solution (1.0 ml 880 ug-1"1) was added and the contents diluted to volume. Required dilutions

were made with internal standard solution at 8.8 ug-1"1. Controls and matrix spike samples

were treated similarly.

29

GC Chromatography Analysis:

A l l G C analyses were performed using a Varian model 6000 gas chromatograph

equipped with N i 6 3 electron capture detector, auto sampler and data system. The column used

was 180 cm X 2 mm i.d. glass column packed with 3% Carbowax-20M on 100-120 mesh

Supelcoport™. This column was operated at 130 °C for TeCB and 160 °C for TeCG or D B D

to accomplish complete resolution of internal standard, target compounds and surrogates.

Argon-methane (10%) was used as the carrier gas at 30 ml-min"1. The detector and inlet

temperatures were set at 300 and 200°C, respectively. Regression equations were established

from a seven point calibration curve at concentrations ranging from 0.5 to 10 ug-f1 for TeCB

and TeCG and 20 to 200 pgT 1 for D B D . Required dilutions were made using an internal

standard solution at 2.5 ugT 1 . Method blanks were taken through the entire procedure to

check for gas chromatographic interference at the retention times of the target compounds,

surrogates and internal standards. A l l standards and samples were injected in duplicate using a

Varian model 8000 autosampler.

Lipid analysis:

Total lipids were determined (Folch, 1957) on six small fish or in duplicate on each

medium and large fish tissue homogenate. This procedure involved blending 5 g aliquots of

frozen tissue homogenate or diced small whole fish (previously weighed) to a powder in a

small stainless steel blender, to which dry ice and anhydrous sodium sulfate were also added.

Lipids were extracted by blending this powder with 100 ml 1:1 chloroform:methanol. Two

extractions were used. The combined extracts were purified by washing with aqueous

30

potassium chloride solution (0.88 % w/v). Layers were allowed to separate overnight and the

aqueous layer discarded. The organic layer containing the lipids was vacuum filtered through

a 0.2 u Nylon-66 membrane and made to a 250 ml volume. A 25 ml aliquot was added to a

tared beaker and solvent was removed using a gentle stream of nitrogen and low heat. The

beaker and lipid residue was vacuum dried, weighed and percent lipid calculated.

Calculations and statistical analysis:

Gas chromatographic peak areas were measured with a Varian Vista Data System. The

general formula used to calculate toxicant and surrogate sample concentrations was described

as follows:

X = ( A ) ( B ) ( C ) / ( D ) ( E ) ( F ) where

X = the concentration of target compound and surrogate in the sample (ug -g _ 1 for

tissue and u.g-1"1 for water)

A = ug^l"1 concentration of toxicant and surrogate in the final extract determined using

calibration curve linear regression equations

B = volume of the final extract (1)

C = dilution factor

D = initial weight or volume of the sample taken for analysis (g or 1)

E = aliquot of initial extract taken for cleanup (i.e., 10/250 = 0.04)

F = aliquot of initial extract taken for concentration (i.e., 10/100 = 0.10)

Since the toxicant uptake rate was thought to be fast and linear initially, the uptake rate

constant (ki) was calculated as toxicants accumulated in the test fish during the 2 h exposure

31

period per water/blood toxicant gradient which could also be considered as the measured

mean water toxicant concentration during the first 2 h of exposure. A n F-test performed upon

the Pearson correlation coefficient was used to test for statistical significance of all

regressions and correlations. Differences between the means of various exposure groups was

tested by a one-way A N O V A following data transformation, i f required, and a Tukey test. In

all tests, a probability of 0.05 was chosen as the level of statistical significance. A l l values are

expressed, unless otherwise specified, as the means ± s.e..

32

R E S U L T S

Fish toxicant body burden was calculated based on the G C standard curves for TeCB,

TeCG and D B D (Fig. 3) and the peak area of the measured samples on each corresponding

chromatogram. Fish oxygen consumption rate was determined based on the raw data recorded

during the 2 h test period in a print file in an Olivetti M24 personal computer.

A significant correlation was established between fish oxygen consumption and the

amount of TeCB, TeCG and D B D absorbed by rainbow trout and coho salmon of different

sizes under normoxic conditions (Fig. 4, r2=0.94). The data on largescale sucker collected by

Brauner et al. (unpublished data) fitted into the general regression generated in this study,

which indicated that the amount of toxicant absorbed was correlated with the total amount of

oxygen consumed by the animal during the initial uptake stage regardless of the relative lipid

solubility of the target compound, the sizes and species of the fish being used.

The toxicant uptake rate constant (ki), which was the main concern for the toxicant

uptake model, was calculated for each test and was also found to be closely correlated with

fish oxygen uptake (Fig. 5). In this study, lipid concentrations in large fish ranged from 6.04 -

10.28 % (mean 8.16 %) and small rainbow trout and coho salmon lipid values ranged from

3.35 - 7.61%, averaging 5.64%. If the toxicant body load was calculated based on the lipid

content of the fish, i.e., ug toxicant per gram fish lipid, then the significance of the linear

relationship between fish toxicant uptake rate (expressed as ug toxicant per gram fish body

weight per unit chemical gradient) and fish oxygen consumption (mg 0 2 per kg fish body

weight per hour) is improved (Fig. 6). Fish metabolic rate, therefore, seemed to dictate the

33

chemical uptake kinetics and the relationship was more significant when data were expressed

on a lipid basis.

34

Figure 3. Standard curves for TeCB, TeCG and D B D G C analysis

Figure 4. Relationship between toxicant and oxygen absorption in fishes of different species and sizes. L=large, M=medium, S=small; RBT=rainbow trout, CS=coho salmon, LSS=largescale sucker.

37

10000

Y = 1.41 X °- 7 5 7

R 2 = 0.942

1000

100

V MCS, DBD O MRBT, TeCB • MRBT, DBD • LRBT, DBD o LSS,TeCB T MRBT, TeCB A SCS,TeCG • SRBT, DBD

10 10 100

O X Y G E N U P T A K E

(mg/h)

38

Figure 5. Relationship between toxicant uptake rate constant (ki) and fish oxygen consumption. L=large, M=medium, S=small; RBT=rainbow trout, LSS=largescale sucker. Rate constant ki was calculated based on fish toxicant body burden expressed as \lg toxicant • g body weight"1.

39

1 / • S R B T , T C G

/ -

• L S S , T C G • g

• M R B T , T C G • M R B T , T C B • 9»° o

V

S R B T , T C B L R B T , T C G

* M °

A L R B T , T C B A / V A A W v •

a A tfv • A / A B •

• / •

A W v • a A tfv •

A / A B • • / •

• rv m

/ • • • /

/ •

•

T

7 Y=1.012X-2.946

/ R2=0.734

i

102 10 3

Oxygen Consumption Rate

(mg/kg/h)

Figure 6. Relationship between toxicant uptake rate constant (ki) and fish oxygen consumption. L=large, M=medium, S=small; RBT=rainbow trout, CS=coho salmon, LSS=largescale sucker. Rate constant ki was calculated based on fish toxicant body burden expressed as |i,g toxicant • g lipid"1.

41

• S R B T , T e C G

• L S S . T e C G

• M R B T , T e C G • M R B T , T e C B

0 S R B T , T e C B

V L R B T , T e C G A L R B T , T e C B

• M C S , D B D

• S C S , D B D

0 S R T , D B D

M R B T , D B D

Y=1.054X-2.923 R2=0.873

102

Oxygen Consumption Rate (mg/kg/h)

103

DISCUSSION

These studies support the contention of Randall and Brauner (1990) that the rate of

transfer of toxicants of different hydrophobicity is correlated with the oxygen consumption of

various fish species of various sizes. Moreover, the fact that toxicant uptake rate constant (ki)

is significantly related to fish metabolic rate (Fig. 5) indicates that the fish toxicant uptake

kinetic process can be characterized by fish oxygen consumption, and this finding is of great

importance to the development of a simplified physiological model.

Oxygen molecules transferred through the gill epithelium wil l be bound to hemoglobin

and then carried away in the circulatory system. Similarly, almost all hydrophobic toxicants

diffusing into the blood stream wil l be bound by plasma proteins or blood lipids and this

bound portion of the toxicant increases drastically with an increase in K o w as reported by

Schmieder and Henry (1988). During the initial stage of toxicant uptake the ability of the

blood to remove the toxicant wil l far exceed the ability of water to deliver it. As a result, the

initial uptake rate wi l l be dependent on the gradient for diffusion of toxicant across the fish

gills, which should be very close to the concentration of the toxicant in the ambient water.

Hence, the toxicant transfer rate should be fairly constant for at least the first 2 h, which was

used as the exposure duration in all the uptake experiments. Consequently, the uptake rate

constant (ki) was calculated as toxicants accumulated in the test fish during the 2 h exposure

period per water/blood toxicant gradient, which is also the mean water toxicant concentration

during the 2 h uptake experiment. This uptake rate constant calculation can be easily made

under these circumstances and should be indicative of the whole toxicant uptake profile. In

43

other words, it is valid to use the initial toxicant exposure regime as a vehicle to investigate

the correlation between fish toxicant uptake kinetics and fish oxygen consumption. This

approach, therefore, is theoretically sound and technically more cost-efficient compared with

deriving ki from either the traditional "plateau" method where toxicant exposure is continued

until tissue concentrations reach a steady state or the short-term "kinetic" experiments which

normally last from 48 h up to 5 days (Bishop and Maki, 1980).

Some authors have reported, as mentioned earlier, that the uptake rate constants of

some lipophilic compounds by various fish species have a fixed relationship to the K o w over a

certain range (Hawker et al., 1985; Saarikoski et al., 1986). In this study the selected target

compounds included relatively highly hydrophobic chemicals such as TeCB, TeCG ( K o w =

4.97 and 4.41, respectively) and the less lipid soluble chemical D B D ( K 0 w = 2 . 3 5 ) . The

correlation between fish toxicant uptake rate constant (ki) and oxygen uptake (Y=1.054x-

2.923, r =0.87) was derived based on the data on all three chemicals, and should be

applicable to basically all xenobiotics, except those which are dissociated and/or easily

metabolized. Theoretically, ki should vary among the three target compounds based on their

different K o w - It has been indicated by Hawker and Connell (1985) that a general relationship

exists with the ki of a variety of lipophilic compounds and their K o w :

Log ki = 0.337 Log K o W - 0 . 3 7 3 (1)

Therefore, as K o w increases, the uptake kinetic process should also be faster (higher

ki). According to the result in this study, ki is also markedly affected by fish metabolic rate

which could be a complicating factor in this relationship. In fact, the degree of elevation in ki

induced by the difference in log K o w between TeCB (4.97) and D B D (2.35) could also be

44

easily achieved through the increase in fish oxygen consumption rate at routine levels by 5-

10%. Consequently, we use these three chemicals to cover a whole range of xenobiotics for

the development of the relationship and make sure that it is generally applicable. Oxygen

uptake was found to be the primary factor dictating the toxicant uptake kinetic processes.

Lipid normalized data improved the relationship between fish oxygen consumption

rate and chemical ki (Fig. 6). So, it can be concluded that fish metabolic rate influences the

chemical uptake kinetics and the relationship is more significant when data were expressed on

a lipid basis. Lipid normalized experimental values were first used for the standardization and

easy comparison of bioaccumulation among different groups in aquatic animals. It seems that

the total amount of lipids in an animal, where all the chemicals' residues wil l eventually be

found, should affect the final concentration of the toxicant being accumulated rather than the

uptake kinetic process depicted by ki . So, it is not clear why the significance of the

relationship between oxygen consumption and ki was enhanced by incorporating the lipid

content factor. One possible explanation is that, the fatter an animal becomes the longer it

tends to take for them to get rid of the toxicant being accumulated. And the instant the uptake

starts so does the chemical depuration from the body. Due to the established fixed relationship

among k i , k 2 and K o W (Hawker and Connell, 1985): ki/k 2 = 0.048 K o w , it is likely that the

decrease in k 2 wi l l result in corresponding changes in ki as well.

The possible routes for toxicant entry in fish include direct uptake across the gills,

uptake by the skin, and through ingestion of contaminated food. Biomagnification, which

refers to the accumulation of a chemical from feeding, is generally thought to be of primary

importance in air breathing animals (Connell, 1990). The relative contribution of chemical

45

uptake from food in fish wil l be discussed in detail in Chapter IV. For non-air breathing

animals, such as fish, direct uptake from the environment is normally considered to be the

main mode of toxicant uptake (Shaw and Connell, 1982; Reinert, 1972). Uptake by the skin

should vary with size, because of changing surface to volume ratios. The relationship between

ki and oxygen uptake, however, was not influenced by size, indicating that uptake by the skin

was not sufficient to significantly affect this relationship. Thus the relationship between ki

and oxygen uptake holds for a range of xenobiotics and fish sizes, and several species of fish.

46

CHAPTER III

Correlation between Fish Oxygen Uptake and Chemical Depuration

INTRODUCTION

Bioaccumulation is considered one of the most important environmental transport and

partitioning processes (Kathryn, et al, 1990), and the toxicant residue being accumulated in

the fish body is dependent not only on the chemical inflow (uptake) but also outflow

(depuration). Overall, a polluted fish can have a toxicant eliminated via the gills to the water,

via faecal matter or through metabolic transformation (Clark et al, 1990). Fish are capable of

metabolizing organic chemicals to enhance the elimination process, thereby decreasing the

equilibrium level of the chemical accumulated. Heath (1987) argued that the fish gill is an

important mode of exit for small compounds that require little bio-transformation before

excretion, and that this is presumably a passive process. For example, the gills turned out to

be an especially important organ for elimination of aromatic hydrocarbons and the rate of

excretion of individual hydrocarbons by the gills was shown to be inversely proportional to

the size of the molecule (Thomas and Rice, 1981). For non-metabolized organic compounds,

the fish gill is postulated to be the main route of toxicant excretion, as well as entry, of

hydrophobic toxicants (Yang and Randall, 1995).

Due to the fact that fish gills constitute the majority of the body surface and represent a

very thin barrier between blood and ambient water, there is a potential that toxicant

depuration rate could be related to oxygen consumption, as is toxicant uptake rate, although

the possibility has never been investigated. In this chapter, I report on 21-day depuration tests

designed for small rainbow trout and coho salmon, conducted at different water temperatures

under low and high swimming velocities, to determine if there was a correlation between

48

oxygen uptake and depuration rate of toxicants. Fish were exposed to TeCB, TeCG, or DBD

alone, or simultaneously to all three chemicals to see if the depuration dynamics of more than

one chemical can be determined without interaction among chemicals. Correlation between

toxicant depuration and fish oxygen consumption was explored to investigate the possibility

of using oxygen uptake as an indicator for toxicant depuration rate in fish.

The main theme of this thesis is the derivation of a simplified physiological model to

predict fish xenobiotic body burden which is essentially controlled by two opposite dynamic

processes, uptake and depuration, occurring at the same time. Since toxicant uptake was

demonstrated to be dependent upon fish oxygen uptake (see chapter II), whether or not the

rate of depuration of non-metabolized chemicals is also related to oxygen uptake, as

hypothesized, becomes another crucial issue for the development of the general model.

49

M A T E R I A L S A N D M E T H O D S

Fish acquisition and acclimation:

Fish used in this study were small (5.4 ± 0.52 g) rainbow trout (Oncorhynchus mykiss)

and coho salmon (Oncorhynchus kisutch) (4.96 ± 0.32 g), purchased respectively from

Westbrook Trout Farm in Aldergrove and Capilano Hatchery in North Vancouver, British

Columbia. The fish were held in 200 L fibreglass tanks containing dechlorinated Vancouver

city tap water, circulating at approximately 1 body length per second (BL-s"1), for at least two

weeks prior to experimentation. Water temperature varied seasonally from 5 - 10 °C. A l l

experiments were performed at the temperature to which the animals had been acclimated.

Following arrival, fish were fed commercial pellets once a week but were starved at least 48 h

prior to introduction to the experimental conditions to ensure a post-absorptive state.

Depuration test protocols:

Initial Toxicant loading.

Prior to the depuration tests the experimental fish were loaded with target compounds

1,2,3,4-tetrachlorobenzene (TeCB), 3,4,5,6-tetrachloroguaiacol (TeCG) or 4,6-

dichlorobenzenediol (DBD), through gill diffusion. Water exposure was achieved through a

double 12 h dosing regime with either single chemicals or all three chemicals simultaneously

at predetermined concentrations. TeCB, TeCG or D B D were added to a fully automated

Brett-type respirometer (Gehrke et al., 1990) giving the water a calculated toxicant

concentration of 270 - 300 g-1"1. After a 5-min mixing time, water samples were taken and the

50

fish were transferred to the respirometer. The system was then sealed and water flow

resumed. In order to achieve adequate aeration, oxygen was introduced into the system

through a tube penetrating the top lid of the swim tunnel at a flow of approximately 10 ml-

min"1. Water oxygen levels were routinely monitored to ensure normoxic conditions within

the respirometer. After 12 h, fish were temporarily removed from the respirometer and

duplicate water samples were taken for chemical analysis. Then the swim tunnel was flushed

and the same start-up procedure was repeated for the second 12 h exposure period followed

by the fish being reintroduced back into the system. At the end of the second 12 h dosing

period, duplicate water samples were taken again and all fish were removed from the

respirometer. At this time, 6-8 fish were killed, briefly rinsed with clean fresh water and dried

with paper towel before weight and body length were measured. Then, these fish were stored

in plastic bags under -80 °C for subsequent toxicant analysis and were treated as the day-0

sample into the depuration tests.

Depuration experiment

Depuration tests were conducted after the fish had accumulated toxicants via gill

diffusion. The double 12 h dosed fish were temporary held in a bucket while the respirometer

was cleaned, and then were put back into the swim tunnel and the clean freshwater flow was

resumed. During the 21 day depuration periods, fish were swum against water flow rate of 18

or 27 cm-s"1, but at different temperatures ranging from 4.6 to 14.3 °C for different depuration

tests. The respirometer was operated with the artificial lung being incorporated in the flow

path and the swim tunnel partially opened such that water could overflow the chamber as

51

freshwater was added. This operating mode ensured that the depurated toxicant was removed

from the receiving water in the respirometer. Normally, 6-8 fish were taken from the chamber

at 1, 1.5, 4, 7, 14, 21 day intervals during the depuration phase, fish length and body weight

were recorded prior to being stored in plastic bags and frozen for subsequent whole body

toxicant burden analysis.

Oxygen consumption rate was monitored twice a day as long as the total body weight

of all remaining fish was not less than 100 g which is the lower limit for accurate oxygen

uptake measurement in this 130 L respirometer. A combined oxygen uptake was determined

for all fish in the respirometer at that time based on the rate of oxygen depletion in the tunnel

over that period of time.

Chemical analysis and calculation:

GC analysis and calculations were similar to those described in Chapter II. Soxhlet

extractors were used for TeCB samples and methylation protocol required for TeCG or DBD

samples. All data including calculations and regression analysis were done using QUATTRO

PRO™ spreadsheet software. All values are expressed, unless otherwise specified, as the

means ± s.e..

52

RESULTS

In these depuration experiments, juvenile rainbow trout and coho salmon depurated

TeCB, TeCG or DBD while swimming at different velocities under various water

temperatures. The oxygen consumption rate of the fish remaining in the respirometer after

each sampling interval was measured and corrected after all the fish body weights were

known at the end of each test. Changes in body weight were small (<5%) over the 21 day

depuration even though the fish were not fed, and these changes were ignored in the

calculation of oxygen uptake. The toxicant body burden declined over the 21 day depuration

period as shown in each depuration data profile (Fig. 7-15). In fact, the 21 day depuration

tests resulted in insignificant changes in fish body weight and the close-to-complete toxicant

elimination towards the end of the tests. Considering Fig. 7 as an example, juvenile rainbow

trout were loaded with a mean level of 75.6 g TeCB • g body weight"1 initially and the whole

body TeCB concentration decreased to 1.5 g TeCB • g body weight"1 after being forced to

swim at 2.0 BL1"1 at 11.6 °C for 21 days.

Comparison of TeCB concentrations in rainbow trout collected on a given sampling

day with corresponding depuration time provided a best curve-fitting routine in the

exponential relationship: Ct = 69.20 e "0171 ' (r2 = 0.996), where C ( represents TeCB

concentration in rainbow trout at any given time. Other depuration tests, conducted using

either juvenile rainbow trout or coho salmon swimming at 2.0 or 2.9 BL's"1 but under different

Figure 7. Depuration profile of TeCB for juvenile rainbow trout (Oncorhynchus mykiss) swimming at 2.0 BL-s"1, 11.6 °C (n=6).

54

0 5 10 15 20 25

Time (day)

Figure 8. Depuration profile of TeCB for juvenile rainbow trout (Oncorhynchus mykiss) swimming at 2.0 BL-s"1, 4.6 °C (n=6).

56

Time (day)

Figure 9. Depuration profile of TeCB for juvenile rainbow trout (Oncorhynchus mykiss) swimming at 2.9 BL-s"1, 14.3°C (n=6).

58

'3 > -7

£ bi)

C

o

O c o O

3 O

o g

O

10 15 20 25

Time (day)

Figure 10. Depuration profile of TeCB for juvenile coho salmon (Oncorhynchus kisutch) swimming at 2.0 BL-s"1, 6.5°C (n=6).

60

100

0 5 10 15 20 25

Time (day)

Figure 11. Depuration profile of TeCB for juvenile coho salmon {Oncorhynchus kisutch) swimming at 2.9 BL-s"1, 9.6 °C (n=6).

62

Time (day)

Figure 12. Depuration of TeCG for juvenile rainbow trout (Oncorhynchus mykiss) swimming at 2.0 BL-s"1, 10.6°C(n=6).

64

0 5 10 15 20 25

Time (day)

Figure 13. Depuration profile of T e C G for juvenile coho salmon (Oncorhynchus kisutch) swimming at 2.9 BL-s" 1 , 11.3 °C (n=6).

66

e o

'§ •+-» c o c o O O O CD

bD bo

C o

o o

o

10 15 20 25

Time (day)

Figure 14. Depuration profile of DBD for juvenile rainbow trout (Oncorhynchus mykiss) swimming at 2.0 BL-s"1, 7.3 °C (n=6).

68

- fefl

o <3 +-> c o o O Q PQ Q

=3 O

E—1

o g

10 15 20 25

Time (day)

Figure 15. Depuration profile of DBD for juvenile coho salmon (Oncorhynchus kisutch) swimming at 2.9 BL-s"1, 9.2 °C (n=6).

70

temperatures ranging from 4.6 to 14.3 °C, showed similar depuration profiles and all the data

fitted an exponential relationship.

A general trend shown in the depuration experiments was that fish swimming at either

faster speeds or higher temperatures, both of which were associated with higher metabolic

rates, would lose the toxicant more quickly as indicated by the steepness of the depuration

curves. For example, Test 2 (rainbow trout TeCB depuration, Fig. 7) was conducted under the

same conditions as Test 1 (also rainbow trout TeCB depuration, Fig. 8), except that the

respirometer water temperature was set at 4.6°C while the temperature in Test 1 was

maintained at 11.6 °C. As a result, the depuration rate observed was found to be slower than

that of Test 1 which displayed a depuration rate constant of 0.119 g-g -day"1 (Fig. 8). The

mean 21 day TeCB concentrations in fish swum at this low temperature were found to be 3.5

times higher than those swum at temperatures 7 °C higher. Test 3 (rainbow trout TeCB

depuration, Fig. 9) was carried out under the same conditions as Test 1 and 2 except that the

temperature was at 14.3 °C and, most importantly, the water velocity was increased such that

the fish were swimming at 2.9 instead of 2.0 BL-s"1. In this case, the whole body TeCB

concentration in the rainbow trout declined at a significantly faster rate (Fig. 9) than those in

the above-mentioned two depuration tests. It is not surprising to see the depuration rate

constant calculated for this test (2.9 BL-s"1, 14.3 °C) to be -1.135, 9.5 and 33.4 times higher

than that in Test 1 (2.0 BL-s"1, 11.6 °C) and Test 2 (2.0 BL-s"1, 4.6 °C), respectively.

To further quantify the correlation between fish metabolic rate and toxicant depuration

kinetics, the depuration rate constant (k2) was calculated using the BIOFAC computer

72

program for each depuration test and plotted against the mean value of fish oxygen uptake

rate in a specific depuration period (Fig. 16). An increase in k2 was observed when fish

metabolic rate was elevated and the linear regression between the two was significant ( k2 =

0.0099 M0 2 - 2.2975, r2 = 0.87).

7 3

Figure 16. Correlation between TeCB, TeCG and DBD depuration rate constant (k2) and oxygen consumption rate in small rainbow trout (Oncorhynchus mykiss) and coho salmon (Oncorhynchus kisutch) (n=6).

74

R2=0.87

250 260 270 280 290 300 310 320

Oxygen Consumption Rate (mg/kg/h)

DISCUSSION

The pathway of toxicant elimination by fish is either branchial, biliary, urinary, and/or

cutaneous, depending on the physico-chemical property of the specific chemical and the

animal's physiological status. Cutaneous excretion was found to be less than 2% of the total

rotenone loss from rainbow trout (Gingerich, 1986) and it is generally accepted that cutaneous

excretion plays a minor role in the depuration of most xenobiotic compounds. If the chemical

is subject to high oxidative metabolism, the elimination through kidney or bile is expected to

increase proportionally (Gingerich and Rach, 1985).

Given the fact that fish gills represent the major interface between the water and the

body of a fish and high rates of 0 2 and C 0 2 transfer occur across the gill lamellae all the time

as the fish respires, the branchial route should be considered dominant in not only toxicant

uptake (Yang and Randall, 1995), but also elimination provided the toxicant involved is not

metabolised. Xenobiotics of this nature, therefore, are thought to be eliminated essentially

unchanged across the gills and the dynamic process is dominated by passive diffusion. The

other important factor that could impose great impact on the depuration kinetic is fish oxygen

consumption and this is based on the same argument discussed earlier (Chapter II).

The hypothesis that oxygen uptake is indicative of the toxicant depuration kinetic

process was tested in the depuration experiments using three undissociated chemicals resistant

to biological transformation. The depuration tests for chemically dosed rainbow trout or coho

salmon indicated that different metabolic rates result in different rates of chemical depuration

which was reflected in the slope of the depuration curve (Fig. 7-9). As elimination of

76

toxicants from the whole fish body, and major tissues depots, conforms to simple first-order

kinetics (Gingerich, 1986), the depuration rate constant (k2) was calculated for each of the

depuration experiments (Fig. 16) and an increase in k 2 was observed when fish metabolic rate

was elevated. It was also found that the depuration data profile was similar for fish exposed

only to one chemical and those exposed to three chemicals simultaneously, indicating there

was no interaction among these chemicals during the depuration process. The increase in

chemical elimination rate associated with higher fish oxygen uptake rates, indicates that

transfer across the fish gill is the rate limiting step in the depuration of most non-metabolized

chemicals.

According to the relationship between uptake/depuration efficiency and the octanol-

water partition coefficient ( K o w ) , characterised by M c K i m et al. (1985), it appears that the rate

limiting process controlling both uptake and depuration varies with logK o w . Considering K o w

as an important determinant of the depuration of xenobiotics in fish, target compounds with

log K o w ranging from 4.97 to 2.53 were chosen for these depuration studies, aimed at

developing a general relationship characterised by the influence of fish "oxygen consumption

on toxicant elimination dynamics. As expected, k 2 derived from TeCB (log K o w=4.97), TeCG

(log K o w

= 4 . 4 1 ) and D D B (log K o w=2.35) depuration experiments, conducted under different

conditions associated with various fish metabolic rates, were significantly correlated with fish

oxygen consumption, indicating that fish metabolic rate is an important indicator of

xenobiotic depuration kinetics for several species of fish. It was reported by Hawker and

Connell (1985) that a general relationship exists with the k 2 of a variety of lipophilic

compounds and their K o w :

77

-Log k 2 = 0.663 Log K o w -0 .947 (1)

The depuration rate constant (k2) should decrease with the increase in K o w - However, it

has been shown in this study that fish metabolic rate has a pronounced effect on k 2 which

could be a complicating factor in this relationship. With fish oxygen uptake being the primary

concern in these tests, it was not surprising to see little effect of K o w on k 2 ; which could

probably be due to the great extent of k 2 being influenced by a small amount of changes in fish

metabolic rate. We use these three chemicals to cover a wide range of xenobiotics with

different hydrophobicity for the purpose of developing a generally applicable relationship.

Geyer et al. (1995) demonstrated that the depuration half lives (ti/2), which are

inversely correlated with k 2 of T C D D , in mussel and fish at different developmental stages,

increased with their lipid content. The lipid contents of the small rainbow trout or coho

salmon were all close to 5%; as a result, although k 2 was also calculated using lipid

normalized data in these depuration experiments, it was found to be affected by fish metabolic

rate in a similar manner as i f k 2 was derived from the data bearing the unit of g toxicant • g

wet body weight"1.

In conclusion, fish toxicant depuration is enhanced by elevated metabolic rate. The fact

that there is a significant relationship between toxicant depuration rate constant (k2) and fish

oxygen uptake, regardless of chemical hydrophobicity and fish species, is another important

component for the proposed simple physiological model which enables the prediction of

toxicant load in fish body at any given time.

CHAPTER IV

Prerequisites for the Model Development and Application

INTRODUCTION

In the present investigation, TeCB, TeCG and DBD were chosen as the target

compounds for deriving a physiological model based on oxygen consumption to predict

xenobiotic concentrations in fishes. It has been shown that both uptake and depuration

dynamics of these compounds in fish are significantly affected by oxygen consumption (see

Chapters II and III). Consequently, a rate constant-based model can be developed based on

oxygen uptake. A major advantage of the proposed physiological model is that it uses fish

oxygen uptake as an indicator for toxicant transfer, thereby avoiding the time-consuming and

complicated measurements normally required for most other physiological models. Oxygen

uptake is not only easy to measure, but there also exists a fish oxygen consumption database

(Thurston and Gehrke, 1993). This database, called OXYREF, contains data from over 6800

individual laboratory tests in which oxygen consumption was measured. This data bank

includes information on fish species, fish weight, test water conditions, and the level of

activity of the fish, i.e., "standard"(resting), "routine"(moving about), or "active"(measured

swimming speed).

The prerequisite, however, of applying OXYREF in this general physiological model

is that toxicant exposure does not interfere with fish oxygen consumption since the

measurements in the database were taken in the absence of toxicant exposure. Brauner (1991)

reported no direct effect of 1 or 4 h TeCB exposure on the oxygen consumption rates of

rainbow trout at rest or during exercise. These findings are in agreement with McKim et al.

(1985) who exposed rainbow trout to 14 organic chemicals and found no effect on oxygen

80

consumption rate, ventilation rate, or ventilation volume.

One of the objectives of the study presented in this chapter was to look at the effect of

prolonged constant sublethal exposure of TeCB and TeCG on the oxygen uptake of juvenile

rainbow trout (Oncorhynchus mykiss) using a newly modified, multi-vessel, computerized

flow-through fish respirometer (Duval et al. 1981; Johansen and Geen 1990). The hypothesis

is that toxicant sublethal exposure will not add any complicating factors to the relationship

already established between fish oxygen consumption and toxicant transfer.

Another concern in developing this model based on uptake of toxicant from the water

across the gills is the necessary assumption that uptake via food was negligible. To test this

assumption, feeding experiments using adult rainbow trout and small trout fry were conducted

to compare the contribution to the total toxicant body burden from uptake across the gills and

food ingestion. Because of the continuous and massive exchange of materials across fish gills,

it was postulated that, for fish, toxicant uptake via feeding plays a minor role in determining

the toxicant body burden in fish.

81

MATERIALS AND METHODS

Effects of toxicant exposure on fish oxygen uptake:

Experimental fish

Juvenile rainbow trout, Oncorhynchus mykiss (43.8 ± 3.6 g; fork length 15.2 ± 0.3

cm), were purchased from West Creek Trout Farm, Aldergrove, British Columbia. Fish were

held in a 2.5 m diameter fiberglass tank under natural photoperiod in dechlorinated

Vancouver city water (temperature 5-10 °C, pH 6.1-6.3, hardness 5.2 to 6.0 mgT1 as CaC03

and O2 saturation approx. 95%) at Simon Fraser University, Burnaby, British Columbia. Fish

were fed commercial trout pellets, Biodiet Grover (purchased from Bioproducts, Warrenton,

Oregon), weekly ad libitum. Fish were starved at least 96 h prior to the trials to assure a post-

absorptive state (Goss and Wood 1988; Beamish 1978) and were allowed a 24 h acclimation

period within the glass vessel before the initiation of each test.

Toxicants

1,2,4,5-tetrachlorobenzene (TeCB) and tetrachloroguaiacol (TeCG) are chlorinated

compounds formed in the bleaching processes of pulp and paper industry. Both TeCB and

TeCG are relatively hydrophobic with the log Kow being 4.97 and 4.41, respectively

(USEPA 1994). Although their water solubilities were sufficient to make them bioavailable in

the water, acetone was used to facilitate their dissolution in the test medium. Acetone, used as

a chemical carrier, has been proved to be non-toxic to fish at the concentrations administered

in this study (Brauner, unpublished data). TeCB and TeCG were acquired from Pfalz and

82

Bauer Inc., Stamford, Connecticut, and Chem Services, Westchester, Pennsylvania.

Experimental system and procedures

All experimental trials were conducted in a six-glass-vessel computer-controlled,

intermittent-flow fish respirometer (Fig. 17) originally designed by Duval et al. (1981) and

modified subsequently by Johansen and Geen (1990). The volume of each glass vessel was

about 8 L and was covered with opaque polythylene during the whole test period to minimize

possible disturbance. The oxygen readings were given by the OxyGuard® DO probe, mounted

through the lid into each glass vessel, instead of the modified YSI Model 53 oxygen monitors

used by Johansen and Geen (1990). A transduction 386 compatible PC and a custom made

A/D-D/A interface were used for system control and data acquisition. Preliminary trials were

conducted to find out the optimal total fish body weight each vessel can accommodate in

order to avoid any possible continuous oxygen deficit or interactive stress within each vessel,

and to assure DO measurement accuracy. Four fish (~ 200 g in total body weight) were placed

in each vessel and acclimated for 24 h under continuous flow conditions before each

experiment. All vessels were immersed in a 175 L water bath to minimize the water

temperature fluctuation (7.11 ± 0.14 °C). The fluorescent lights produced a simulated 14 : 10

hr light: dark photo-period. Parameters such as pre-exposure time, exposure time, cycle time,

measurement frequency, purge time/rate, bypass time, toxin flow rate, toxin stock solution

concentration, and toxin pump time, etc., were manipulated in order

Figure 17. A schematic diagram of the six-vessel, intermittent flow-through, computerized fish respirometer.

84

Toxin Stock Solution

I Computer Control Unit

Clean Dechlorinated Water Reservoir

Mixing Connector

Outflow Rate Control Valve

Bypass

85

to achieve a pre-determined toxicant exposure regime.

Two trials, one for TeCB and one for TeCG, were carried out under identical

conditions. Each trial lasted for 96 h, with both the pre-exposure and exposure period being

48 h. Each vessel was flushed with freshwater every 15 min with the purge rate at 1.0 1-min"1

for 2.0 min. The oxygen probe inserted into each vessel was set to record the oxygen level in

each vessel every 5 min. The bypass time between each purge was 10 sec, when the

connecting tubes were flushed to prevent any possible toxicant residue accumulation. The

toxicant flow rate was 2 ml-min"1 and the concentration of the toxicant stock solution was 100

mgT1 TeCB or TeCG. The toxicant was pumped into each vessel, where required, during a

purge. The pump time was adjusted, respectively, at 0 min for two control group vessels, 1.0

min for two 100 figl"1 treatment groups and 2.0 min for 200 figT1 treatment groups. Eight

control fish at the beginning and all fish in the vessels at the end of each test were terminated.

Body weight and fork length were measured.

At five-minute intervals, the dissolved oxygen in each vessel was measured by the

inserted OxyGuard® Probe and each vessel was partially flushed with freshwater every fifteen

minutes. Completion of the sequential oxygen measurements and freshwater flushing in six

vessels constituted one cycle. Oxygen consumption rate was based on the pooled value for

four fish in each vessel, and total fish body weight in that vessel was measured at the end of

the 48 h exposure when the fish were killed.

86

Calculation and statistical analysis

The average of oxygen consumption in 30 min was calculated and the oxygen

consumption data profile analysed between control and treatment groups, or within treatment

groups but between 48 h pre-exposure and 48 h exposure periods. One way ANOVA

followed by a Dunnett's test or student t-test were applied to see if there is any effect of the

two chemicals on the oxygen consumption pattern of these fish under the condition in this

study.

TeCB and TeCG depuration tests with initial toxicant loading via food ingestion:

Test fish

Juvenile rainbow trout {Oncorhynchus mykiss) (58.3 ± 5.9 g) or rainbow trout fry

(0.23 ± 0.03 g) were obtained from West Creek Trout Farm, Aldergrove, British Columbia.

Fish were held in a 2.5 m diameter fiberglass tank or 80 L glass aquarium under natural

photoperiod in dechlorinated Vancouver city tap water. Fish were fed with appropriately-sized

commercial trout pellets once a week and were starved at least 96 hr prior to the trials to

ensure a post-absorptive state.

Exposure regime for trout fry

Trout fry were exposed in a 80 L aquarium for 6 days, with the exposure solution

partially changed (20 1-day"1) during the first 4 days. The initial calculated TCB (or TeCG)

concentration in the water before exposure of the fry was 200 fig-l"1. Test water in the

aquarium was sampled each day of the 6-day exposure duration. At the end of the TCB (or

87

TeCG) dosing procedure, 6 toxicant loaded fry were sampled immediately after the exposure

and, in addition, 6 contaminated fry that had spent 1.5 h in clean water were sampled.

Juvenile trout depuration experiment

In the TeCB test, adult trout were maintained for the feeding and depuration period in

separated compartments within a large tank with fresh water overflowing continuously. Water

temperature was 8.1 - 14.1 °C. Ten toxicant-saturated fry. were fed to each adult trout and the

fry were all eaten in 1-1.5 h. After trout were fed, 3 animals were collected at day 0, 1, 3, 8,

12, 17 and 29. Six unfed control adult trout were also sampled. The TeCG depuration test was

carried out in a similar fashion except that water temperature during the whole depuration test

was 6.0 - 9.6 °C and three fish were collected at day 0, 1, 3, 8, 14, 21 and 26. All water and

fish samples were immediately stored in the freezer (-80 °C) for subsequent chemical analysis.

The GC analytical procedures conformed with those described in the relevant section of

Chapter II.

88

RESULTS

Effects of TeCB or TeCG prolonged exposure on rainbow trout oxygen uptake:

A major modification in this flow-through respirometer compared with that described

by Johansen and Geen (1990) is the usage of the OxyGuard® DO probe, which has a built-in

temperature compensation, a large anode and electrolyte volume. Probe air calibration was

performed according to the manual (Point Four Systems Inc.) prior to each experiment. The

probe uses very little oxygen for its measurement and, therefore, it functions correctly with

liquid movement as low as 2 cm/sec (measured at 7 ppm and 13 °C). The total amount of

oxygen the probe consumed by itself was minimal (Tab. 1) during 20 min and each cycle,

which lasted for 15 min, was always initiated by partial flushing of fresh solution in the glass

vessel, indicating the feasibility of using the oxygen probe for accurate oxygen consumption

measurement in this study.

Marked respiratory changes occur in fish under hypoxic conditions (Hughes 1973;

Randall 1982; Bushnell et al. 1984). Therefore, the maximum amount of fish each vessel can

accommodate is another parameter that, if inappropriate, could potentially affect the survival

of the test fish, let alone the accuracy of oxygen uptake measurement. In pre-experimental

trials, four or six fish were put into a glass vessel and oxygen depletion after 20 min was

measured. A 20.7% drop of dissolved oxygen was detected in the vessel of four fish and

35.7% in the vessel of six fish. After 5 min freshwater flushing, the DO level in the vessel

containing four fish went back to 102.8% of the original level, while the DO was only 84.1%

Table 1. Oxygen consumption by the OxyGuard probe during 20 min at 7.07 °C.

90

PROBE 1 2 3 4 Dissolved O2 (ppm) 8.09 8.51 10.29 10.41

7.88 8.33 10.24 10.35 7.77 8.20 10.21 10.31 7.72 8.13 10.19 10.29 7.70 8.10 10.18 10.28

% Decrease in 20 min 4.8 4.8 1.1 1.2

of the initial concentration if six fish were present in the chamber. As a result, only four fish

were included in each 8 L vessel during the actual experiments.

A circadian rhythm in fish oxygen consumption rate was detected in this study (Fig. 18

and 19), reflecting the normal daily activity pattern (Brett and Zala, 1975) of the fish being

used. The mean fish oxygen consumption was similar to those found by other investigations

using the same apparatus (MacKinnon and Farrell, 1992; Janz et al. 1991; Johansen and

Geen, 1990), indicating that the system provided reliable oxygen recordings with the newly

installed OxyGuard® DO probes.

After being acclimated in the vessel for 48 h, rainbow trout exposed to nominal

concentrations of 100, 200 (ig-1"1 TeCB (Fig. 18) or TeCG (Fig. 19) did not show any

significant difference in their oxygen consumption throughout the 48 h exposure period.

This conclusion was based on the comparison between the fish oxygen uptake profile before

and after exposure in the same group, and those in the control and treatment groups.

Depuration profile ofjuvenile rainbow trout via toxicant uptake from food:

Concentrations in the TeCB and TeCG exposed trout fry, sacrificed immediately after

dosing, ranged from 2.44-7.83 Ltg-g"1 and 15.9-28.2 Ltg-g"1, respectively. TeCB and TeCG

levels in the fry maintained in freshwater for 1.5 h after dosing were 2.23-11.3 Ltg-g"1 and

10.9-21.1 Ltg-g"1, respectively. There was no significant decrease of either TeCB or TeCG

level in toxicant loaded fry after being allowed to swim 1.5 h in freshwater, indicating that

toxicant loss from fry before they were consumed by juvenile trout was minimal in these

92

Figure 18. Oxygen consumption profile of juvenile rainbow trout {Oncorhynchus mykiss) during the 48 h pre-exposure and 48 h TeCB exposure (100, 200 Hg r1) in a flow-through respirometer (n=12).

93

CD

E

0) CO

CL

E </) c o O

CD

200

150

100

Control

200

150

100

50 ]

40 60

100 ug/l TeCB

100

Pre-exposure i Exposure

• •• • V . . '

200

150

100

50

20 40 60 80 100

200 ug/l TeCB

Pre-exposure Exposure

• • • • * • •

20 40 60 80 100

Time (hr)

94

Figure 19. Oxygen consumption profile of juvenile rainbow trout (Oncorhynchus mykiss) during the 48 h pre-exposure and 48 h TeCG exposure (100, 200 HgT1) in a flow-through respirometer (n=12).

95

CD

E

CD •*-> CO

E tn a o O CD CO >> X O

200 Control

150

100

50

Pre-exposure 1 Exposure

1 • • • • • L • • • • . •

250

200

150

100

50

20 40 60 80 100

100ug/l TeCG

Pre-exposure Exposure

• • - • i ••. •

200

150

100

50

20 40 60

200 ug/l TeCG

80 100

Pre-exposure • Exposure

20 40 60 80 100

Time (hr)

96

experiments.

The mean value of TeCB concentration in juvenile trout sampled at test time zero was

0.170 Aig-g"1, ranging from 0.118 to 0.207 //g-g"1. This value was, not surprisingly, close to a

theoretical value of 0.173 Mg*g 1 calculated from the total TeCB consumed in the food.

Concentrations of TeCB in the other test fish during the depuration period declined over time

reaching non-detectable levels in 12 days (Tab. 2). TeCG concentrations in fish sampled

immediately after feeding averaged 0.783 ^cg'g'\ ranging from 0.609 to 1.02 /ig'g"1. This

mean level at day-0 was higher than the calculated value (0.491 //g-g"1). Measured TeCG

concentrations in fish did not decline until after the first day of depuration and TeCG was still

detectable, at the established detection limit of 0.0125 Mg'g"1, in some fish even at the end of

the depuration (Tab. 2).

97

Table 2. TeCB and TeCG depuration profile in juvenile rainbow trout (Oncorhynchus mykiss) after toxicant uptake through feeding. Fish toxicant body burden was expressed as |ig toxicant -g fish wet weight"1.

98

Depuration Day TeCB (//g-g1) TeCG Og-g"1) 0 0.1702 ± 0.0219 0.7832 ± 0.0991 1 0.1070 ± 0.0211 0.9013 ± 0.1153 3 0.0616 ± 0.0465 0.5414 ± 0.1061 8 0.0375 ± 0.0281 0.2794 ± 0.0353 12 0.0012 ± 0.0002 14 0.1860 ± 0.0760 18 Not Detectable 21 0.1615 ± 0.0862 26 0.1145 ± 0.0669 29 Not Detectable

DISCUSSION

Effects of TeCB and TeCG sublethal exposure on fish oxygen uptake:

Exposure to TeCB or TeCG did not affect oxygen consumption in trout. This is

perhaps surprising because Yang and Randall (1996b) demonstrated that 72 h partially static

exposure to TeCB or TeCG (at similar concentrations) could exert sublethal effects on the

muscle moisture, plasma sodium concentration and gill/kidney Na+/K+-ATPase activity in

rainbow trout of a similar size. In this study, gill Na+/K+-ATPase activity disturbances were

also measured and found to change in the same way as reported earlier. It seems that fish

stress may not be reflected in their oxygen consumption even though many other changes

have been observed (Yang and Randall 1996a; 1996b; MacKinnon and Farrell 1992; Heath

1987).

MacKinnon and Farrell (1992) have shown that a concentration-dependent response in

oxygen consumption of juvenile coho salmon {Oncorhynchus kisutch) was not observed when

sublethally exposed to 2-(thiocyanomethylthio) benzothiazole (TCMTB). It has been shown

that 1-4 h sublethal exposure of TeCB, even at concentrations approaching LC50, did not

exert any effect on the oxygen uptake of adult rainbow trout at rest or during exercise

(Brauner et al, 1994). It was therefore suggested that the correlation studies, usually carried

out at much lower concentrations, would not be complicated by any toxicant effects on

oxygen consumption rate. This result is in agreement with studies using similar exposure

regimes with a variety of chemicals and fish species, such as rainbow trout (McKim et al,

1985), mosquito fish (Murphy and Murphy, 1971) and coho salmon (Janz et al, 1991). As

100

most environmental exposures to fish residing in a polluted area are occurring not only under

sublethal concentrations but also for prolonged periods, it is of special interest to look at the

effects of chemicals on fish oxygen uptake after a prolonged sustainable exposure, as in this

study. No effect on oxygen consumption was observed in fish exposed to these chemicals for

prolonged periods.

The pre-requisite for the use of OXYREF in the model is that fish oxygen

consumption is not influenced by toxicant exposure. Because toxicant exposure does not

affect oxygen consumption, the proposed model can utilize an existing oxygen data base

OXYREF (Thurston and Gehrke, 1993) compiled from the fish oxygen consumption rate

values from the literature, measured in the absence of toxicant exposure. Oxygen uptake can

be used as an indicator of toxicant uptake under normoxic conditions. Altered fish ventilation

rate or diffusion capacity without changes in whole animal metabolic rate, which may occur

during hypoxia or hyperoxia (Randall, 1990), will change the rate of toxicant movement

across fish gills without changes in oxygen uptake (McKim et al, 1985). In other words,

toxicant absorption may be greatly enhanced under these conditions without there being any

significant changes in fish oxygen uptake. Thus, the use of OXYREF in the model is limited

to normoxic conditions. The use of the database under hypoxic conditions will tend to

underestimate toxicant load whereas using it in hyperoxia will tend to overestimate the

toxicant load in fish.

101

Relative importance of toxicant uptake through feeding:

Toxicant intake via food may contribute to the total toxicant loading, especially during

high feeding seasons. The rationale behind the feeding experiments, therefore, was to test this

hypothesis and evaluate the relative importance of toxicant uptake from food by making

comparison with those depuration experiments where the fish were loaded with toxicants via

gill diffusion.

In the previous TeCB and TeCG depuration tests, the initial fish toxicant body

burdens, achieved through the double 12 h aqueous exposure in which the maximum toxicant

loading was supposedly ensured, were 11.06 fxg TeCB-g"1 and 6.15 /xg TeCG-g"1 when fish

were swimming at 2.0 BL-s"1, and 15.00 fxg TeCB-g"1 and 8.94 fxg TeCG-g"1 when fish

swimming speed was 2.9 BL-s". In comparison, the initial TeCB and TeCG concentrations in

the juvenile rainbow trout, trained to feed on trout fry and starved for two weeks before the

feeding experiment, were 0.17 fig TeCB-g"1 and 0.78 fxg TeCG-g"1, respectively. Even when

fish were swimming at low velocities, 88.7% of the total body burden of TeCB and 98.5% of

TeCG resulted from uptake across the gills (Fig. 20). Although there exist some uncertainties

in this comparison, it provides some evidence based on laboratory data to support the

argument that, for aquatic animals such as fish, uptake of certain hydrophobic, non-

dissociated toxicant residues through food ingestion is probably much less important

compared with that through direct water exposure.

The subsequent one-month depuration phase in these feeding tests showed that the

elimination profiles of toxicants were similar to those described in previous depuration tests

Figure 20. Comparison of the initial chemical body burdens, acquired respectively from 2 h exposure in water and by feeding on toxicant-loaded trout fry, in medium size rainbow trout (Oncorhynchus mykiss) (n=6) and the relative importance of toxicant uptake from water and food.

103

2.0 BL's"1 2.9 BL's"1

Uptake from water Uptake from food by at low and high consuming toxicant

swimming speeds loaded trout fry

104

where the initial toxicant concentration was achieved through uptake from water. The

majority of the toxicant load was eliminated in the first 3-4 days, especially when fish were

swum at high speeds. The actual digestion of food in the fish gut normally takes about that

much time as well. As a result, it seems that the potential interference of feeding with fish

toxicant depuration is minimal.

The main conclusion in this chapter is that the chosen target compounds will not

interfere with fish oxygen consumption after prolonged sublethal exposure, which has

justifies the use of Oxygen Database (OXYREF) in the proposed physiological modeling.

Also, the notion that fish toxicant transfer across the gills tends to play a dominant role in the

accumulation and depuration of non-metabolised compounds has been further supported by

the outcome of the feeding tests. Thus the effects of feeding are validly ignored and not

incorporated into the model.

105

CHAPTER V

Model Application and General Discussion

Interest in the physiologically-based toxicokinetic (PBTK) models has increased in

recent years (Nichols, et al, 1993) because of a growing need to predict the time-course of

chemical concentration in specific tissues. The major advantage of any PBTK model is that it

defines an organism, be it terrestrial or aquatic, in terms of its anatomy, physiology and

biochemistry. Consequently, PBTK models can be parameterized independently of exposure

information and provide a basis for the comparison of kinetic data between species. This type

of model, which leads to improved understanding of the uptake and disposition of chemicals

in different animal tissues (Andersen, et al, 1987; Reitz et al, 1990), is also associated with

certain restrictions and disadvantages.

Firstly, the schematic representation of a PBTK model is normally fairly complicated,

requiring a large number of physiological inputs which may not be easily accessible or even

available. This has caused problems, for example, in the development of a PBTK model for

channel catfish (Nichols, et al, 1993), since only very limited physiological information is

available for this fish species in the literature (Burggren and Cameron, 1980; Cameron, 1980),

resulting in the absence of estimates for several important parameters such as the cardiac

output and respiratory volume. Thus, this model can only be applied following the collection

of physiological information for this fish, vital for accurate estimation of toxicant loading

using this model.

Secondly, the kind of physiological measurements needed, if not available in the

literature, are often extremely difficult and time consuming to measure. Considering the

channel catfish PBTK model as an example, some of the physiological parameters required

are gill surface area, gill epithelial thickness, effective respiratory volume, blood/water

107

partition coefficient, etc. (Nichols, et al, 1993). Such data are available in the literature for a

relatively small number of fish species. In addition, measurement of these physiological

parameters, acquired with great effort, can be quite variable due to the differences in the

experimental set-up, temperature and animal preparation, making the data difficult to compare

and extrapolate and, subsequently, to standardize. Moreover, these models are usually

developed using a single species which is easy to culture and handle and assumed to be of

importance as a model system for kinetic studies. The same model parameters, however, when

used for a new species, may not provide an accurate prediction due to the lack of an

interspecies extrapolative validity among fish species.

Thirdly, the dependence of toxicant concentration prediction on anatomical

measurements could lead to errors because fish, at different activity levels, may have different

toxicant accumulation and elimination behaviours, not reflected in the model

parameterization. For example, environmental or physiological factors that influence

ventilation volume also influence the uptake of chemicals. According to Spacie and Hamelink

(1982), ventilation volume (Rv) increases with metabolic rate (Q), which is a function of fish

body weight (F): Q = F* . As the weight exponent is approximately 0.8, the metabolic rate

per gram (Q/F) decreases with increase of fish body weight. Consequently, a 10-g fish

consumes about 60% as much oxygen per gram as does a 1-g fish. A similar weight

relationship should exist for ki as well. Murphy and Murphy (1971) measured the values of

0.75 and 0.77 for the uptake of DDT by mosquito fish at 5 and 20 °C, respectively. These

values were similar to the corresponding exponents for oxygen consumption (0.71 and 0.68),

but lower than the exponent for gill surface area per gram (0.89), indicating that uptake is

more a function of the fish gill's physiological rather than anatomical properties.

Thus, results obtained using these PBTK models involving the measurement of several

anatomical and physiological parameters should be accepted with caution and, in general, are

only applicable after verification, which negates the value of the model as a predictive tool.

The question is posed, therefore, as to whether there exists any suitable physiological

parameter that is not only easy to obtain and but is also functionally realistic. All of the above-

mentioned physiological parameters such as gill surface area, gill epithelial thickness, gill

blood flow, diffusion distance and coefficient, etc., presumably affecting toxicant transfer, are

all related to gas exchange across fish gills. As discussed in the general introduction, fish

oxygen consumption is the end-result of the optimization of all of the above parameters. In

other words, fish oxygen uptake is facilitated in a similar manner as is toxicant transfer across

fish gills. The results of this study, presented in Chapters II and III, indicate that oxygen

uptake can be used as an indicator for toxicant transfer regardless of fish species and the

physico-chemical property of the non-metabolised chemicals in question. In addition,

OXYREF has been shown to be applicable to the proposed model since fish oxygen

consumption is not altered by toxicant exposure (Chapter IV), meaning that the physiological

component critical for this physiological model can be obtained easily.

Given the argument that fish gills are the main site for toxicant transfer (Chapter IV)

and the toxicant uptake/depuration kinetic processes are characterized by fish metabolic rate

(Chapters II & III), it becomes a matter of predicting the rate constants such that the overall

109

toxicant body load can then be estimated at any time using the appropriate modified rate

constant model.

As mentioned earlier in the general introduction, the most commonly used one-

compartment, first order kinetic (1CFOK) model (Spacie and Hamelink, 1982) was chosen as

the basis of this work. The basic relationships for toxicant uptake and depuration were

expressed, respectively, in equation (1) and (2) of Chapter I. The uptake and depuration rate

constants can be calculated according to fish oxygen consumption rate by using the

correlations determined in this study (Chapters II & III). A model test was conducted to look

at the feasibility and accuracy of predicting chemical concentration in fish by incorporating

the relationships established between fish 0 2 uptake and kinetic rate constants (ki and k2) into

the one compartment, first order kinetic model. In order to test the validity of this modified

model, experiments were chosen from the literature and the experimental data for toxicant

load were compared with those derived through the modeling based on relevant information,

e.g., size of the fish being used, experimental conditions including toxicant exposure

concentration, temperature, etc., reported in that study and used to derive oxygen uptake and,

therefore, uptake and elimination rate constants. The following chemicals were selected for

the model test: dichlorodiphenyltrichloroethane (DDT) (log Kow - 3.99), sodium

dodecylbenzenesulfonate (LAS) (log Kow = 2.24), ethylenediamine- tetraacetic acid

(EDTA) (log Kow = 2.56) and tetradecylheptaethoxylate (AE) (log Kow = 3.78) (Bishop and

Maki,1980), di-2-ethylhexyl phthalate (DEHP) (log Kow = 4.35) (Tarr, et al, 1990) and TeCB

(log K o w = 4.97) and TeCG (log K o w = 4.41) (Yang and Randall, 1995). The result of the

model application is shown in Fig. 21.

110

Figure 21. Comparison between experimental and predicted values for chemicals of different Kow. Measured values for dichlorodiphenyltrichloroethane (DDT) (log K o w = 3.99), sodium dodecylbenzenesulfonate (LAS) (log K o w = 2.24), ethylenediaminetetraacetic acid (EDTA) (log K o w = 2.56) and tetradecylheptaethoxylate (AE) (log Kow - 3.78) are from Bishop and Maki,1980; Measured values for di-2-ethylhexyl phthalate (DEHP) (log K o w = 4.35) are from Tarr, et al., 1990; and those for TeCB (log K o w = 4.97) and TeCG (log Kow = 4.41) are from Yang and Randall, 1995.

I l l

400 DEHP TeCG

1 2 4 8 16 32 48 64 96 Time (h)

0.5 1 2 4 8 16 21 Time (day)

TeCB TeCB

0.5 1 8 16 21 Time (day) Time (day)

LAS 120 100

LAS

24 48 96 Time (h)

144 3 4 5 Time (day)

DDT DDT

1 6 12 24 48 120 Time (h)

1 6 12 24 48 120 Time (h)

EDTA

3 8 24 48 120 Time (h)

12 24 Time (h)

48 120

- Experimental Data - Model Data

112

The predicted uptake data of DDT at both test concentrations agreed with the

measured ones quite well up to exposure time 48 h. The experimental data of EDTA,

however, seemed to level off after 24 h while the predicted values kept increasing. As for the

depuration data, model values of LAS and DEHP concentrations fitted extremely well with

the test ones at two water concentrations., and so did the data for high speed TeCG

depuration.

Assuming the model is perfect, the predicted values will be exactly the same as the

observed ones. As a result, the linear regression derived between predicted and measured data

should have a slope of 1.0. However, when the values predicted using the present simplified

model are plotted against those obtained from the experiments, the regression line has a slope

of 1.288 (Fig. 22), meaning there exist some descrepancies. Consequently, the predictive

power of the simplified model can be judged by the degree to which the perfectly matched

regression line overlaps with the 95% confidence limits of the real regression. It has turned

out that the perfect regression line falls within the 95% confidence intervals of 92% of all the

predictions performed (67 out of 73 data points). The six points which do not overlap with the

perfect match line are all from one single test with the same chemical, i.e. DEHP

(logKow=4.35), using rainbow trout at temperature 12 0 C. It is with these 6 points that the

model tends to overestimate, or maybe the measured values are lower than what they should

be. Discrepancies between predicted and measured values may be due to inaccuracies in

either the estimation of oxygen uptake and/or bioavailability of the toxicant. One

possibility is that, in these uptake experiments, the chemical exposure concentration may have

declined over time due to the initial fast uptake rate by the fish, and as a result measured

113

Figure 22. Quantitative analysis of the predictive power of the simplified model. Each single data point corresponds to the relevant observed and model-predicted values which constitute the dashed and solid lines, respectively, in the graphs of Figure 21.

114

values are underestimated. The relationship between metabolic rate and temperature varies

between species and may account for some of the discrepancies. Veith et al. (1979) found

that, between 5 to 15 °C, the bioconcentration of Aroclor® increased much more for the

fathead minnows (Pimephales promelas) and green sunfish (Lepomis cyanellus), which are

warmwater fish, than it did for rainbow trout (Oncorhynchus mykiss), a coldwater species. It

was also reported that increases in oxygen uptake per gram were greater for carp (Cyprinus

carpio) than for brook trout (Salvelinus fontinalis) (Beamish, 1964). Since the model

development was based on data derived from experiments using salmonids, it is probably

more appropriate and safe to conclude that the established relationships may be applicable to

these groups of fish in particular and precautions should be taken when the model is used for

fish markedly different in terms of life history, habitat requirements and geographical

distribution.

In any model of the uptake and elimination of xenobiotics in fish, one of the most

important components is the concentration of the chemical that can be absorbed from the

water via the gills of fish. Environmental factors that affect the concentration of a chemical in

true solution (Cw) will affect both the initial uptake and steady-state concentration in fish

tissues (Spacie and Hamelink, 1982). In particular for very hydrophobic chemicals, the

concentration of absorbable or bioavailable chemical is often only a fraction of the total

chemical concentration in the water. This fraction is normally referred to as the Bioavailable

Solute Fraction (BSF), or simply the bioavailability (Landrum et al. 1985; McCarthy and

Jimenez, 1985). The major factors affecting bioavailability of organics in natural waters is the

presence of adsorbents (Gobas, 1993), namely sediments, suspended solids and dissolved

116

polyelectrolytes such as humic acids. For groups of pesticides (Wershaw and Goldberg, 1972;

Kenaga and Goring, 1980), polynuclear aromatic hydrocarbons (PAH) (Means et al, 1980)

and industrial organics (Lopez-Avila V. and R.A. Hites. 1980), sorption to the organic

fraction of sediments generally increases as the octanol-water partition coefficient (K™)

increases. Similarly, dissolved organic matter, including humic substances, binds chlorinated

hydrocarbons, phthalates and PAH (Josephson, 1982; Sullivan et al, 1982), reducing their

bioavailability. The presence of humic substances retards the uptake of benzo(a)pyrene (log

K o w = 6.06) from water by bluegills, but does not significantly affect the accumulation of

anthracene (log K o w = 4.45) (Spacie et al, 1983). All of these effects should not alter the

uptake rate constant ki since the kinetic process is independent of the real toxicant

concentration in the ambient water. However, they do sometimes reduce the "true" chemical

concentration (Cw) which is one of the important components of the model presented here

(Equation 1, Chapter I). Moreover, the actual value of C w may be unknown since it is difficult,

analytically, to distinguish between the free and adsorbed fractions in natural water samples,

meaning the validity of this model is, to a large extent, dependent on the accurate

determination of toxicant bioavailability in the first place. The traditional PBTK models, on

the other hand, can be parameterized independently of exposure information and the

prediction is accomplished partially by using a number of in vitro equilibrium partitioning

coefficients between water/blood and/or tissue/blood of different organic compounds (Nichols

et al, 1993). These equilibrium coefficients, however, are also influenced by bioavailability in

the water, so the problem is obscured rather than absent in the traditional PBTK models.

117

There are limitations to the use of the model developed in this thesis. Any of these

limitations are also applicable to other PBTK models. Whether the purpose is to predict

environmental residues or to acquire insight into toxic mechanisms, bioconcentration must be

viewed as a dynamic process with competing rates of uptake and depuration. Factors affecting

either one of these two kinetic processes will effect changes to the overall chemical

accumulation within the fish body. The relationships between fish toxicant uptake/depuration

and fish oxygen consumption were developed using target compounds resistant to bio­

transformation, and assuming fish gills are the sites where toxicant transfer is limited.

Toxicant loss from the fish can be achieved through routes other than breathing for

metabolized chemicals (Southworth et al, 1980; Southworth et al, 1981), as discussed

earlier. As a result, the model reported here is not adequate for predicting tissue residues of

substances that are biotransformed once they have entered the fish. From an environmental

standpoint, however, persistent chemicals (i.e., chemicals that are not biotransformed) are

often more toxic and generally have a high priority in bio-residue monitoring. These

chemicals that resist biotransformation can have a persistent and adverse impact on the

aquatic ecosystem, so the prediction of their levels in fish is important. The model reported

here is the means to achieve this goal.

Despite all the aforementioned restrictions and preconditions associated with this

newly developed physiological model, its main advantage over the other compartmental or

physiological models lies in the fact that the prediction is based on actual physiological

processes and fish oxygen consumption rate is easier to measure with acceptable accuracy

than many physiological parameters used in other models (Hayton and Barron, 1990).

118

Furthermore, oxygen consumption rate data are available through OXYREF (Thurston and

Gehrke, 1993) provided the fish body weight is known. The rationale behind this study was to

combine the advantages of both compartmental and physiological models by incorporating

physiological processes into a compartmental model, using an easily accessible physiological

parameter correlated with the rate of toxicant transfer, i.e., fish oxygen consumption rate, to

predict chemical concentration in fish with acceptable accuracy. The model developed in this

thesis possesses some functional reality which enables more realistic predictions, and is

convenient and practical for application in biomonitoring and toxicant risk assessment.

119

APPENDIX

Sublethal Toxicity of the Test Compounds

INTRODUCTION

A series of experiments were carried out to determine the effects of TeCB and TeCG

on fish. The effect of these toxicants on the ability of trout and coho salmon to withstand a

seawater challenge was investigated because pulp mills are often situated close to estuaries

and salmonids are euryhaline. Na+/K+-ATPase, Plasma [Na+] hematocrit and muscle moisture

content were measured in a search for possible biomarkers, normally viewed as sensitive

indicators of exposure, response, and/or sensitivity (Weeks, 1995), to indicate the effect of

these toxicants on osmoregulatory ability in these fish. Since environmental exposures are,

more often than not, at low levels, the uptake and depuration tests were conducted at

concentrations 5-10 times lower than the LC50 values for the specific chemicals. All sublethal

tests were focused on TeCB and TeCG, whose toxicity is much higher than DBD due to the

extent of chlonnation.

121

MATERIAL AND METHODS

Exposure regime:

Fish used in this study were adult rainbow trout (Oncorhynchus mykiss) (56.39 ± 1.49g)

and coho salmon (Oncorhynchus kisutch) (25.35 ± 0.65 g), obtained from Westbrook Trout

Farm in Aldergrove and Capilano Hatchery in North Vancouver, British Columbia,

respectively. The fish were held in dechlorinated freshwater at 12°C in the Zoology

Department, The University of British Columbia, before being transported to Bamfield

Marine Station, Vancouver Island. All fish were left in a freshwater tank for one week after

transportation in order for them to recover from transportation stress. Once the experiment

started, the fish being tested were held in 70 L plastic garbage cans sitting in three large

square tanks full of continuously running water to minimise temperature fluctuations. Control

fish were also held in these cans to preclude the effect of plasticizers.

There were three toxicant treatment groups and two control groups. The freshwater tests

were as follows: (1) Rainbow trout were exposed to 200 mg TeCBT1, 200 mg TeCGT1,

separately, and coho salmon were dosed at 200 mg TeCGl"1 for 3 days with the exposure

solution partially changed (20 L-d"1) to ensure adequate toxicant loading. The dosed fish were

then kept in clean freshwater during the 8-day depuration period; (2) Coho salmon were

exposed to 100 jj.g-1"1 TeCG for 11 days with the exposure solution partially changed (20 L-d"

'); (3) Rainbow trout and coho salmon control groups were maintained in freshwater with no

toxicant for 11 days.

The seawater challenge tests were as follows: the fish followed the same exposure

122

process as in the freshwater group, i.e., 3 day toxicant exposure in freshwater, but were

transferred directly to 307oo seawater after toxicant exposure termination. Fish in the control

group were simply held in toxicant free freshwater, with 20 L water changed every day in the

first three days of the experiment, and then were transferred to 307oo saltwater for the

seawater challenge test, which lasted for 8 days. Fish in all groups were put in aerated plastic

garbage cans with either freshwater (except in the continuous exposure group) or seawater

overflowing at a turnover rate of 2 h during depuration or seawater challenge periods. The

experiment lasted for three weeks and the mean water temperature was 13.1 ± 1.6 °C.

Sampling procedure:

Before exposure started, six control fish were sampled and on each successive sampling

day six fish were also sampled both in the control and the toxicant treated groups. The

sampling days were as follows: exposure-day-3 (d3) in all groups; depuration-day-1, 3, 8

(ddl, dd3, dd8) in seawater challenge groups and all freshwater groups with the exception of

the continuous exposure group where fish were sampled at exposure-day-4, 6 and 11 (d4, d6,

dll). Fish sampling was achieved by terminal anaesthetisation with MS 222 (200 mgT1)

(Aqualife TMS, Syndel Laboratories). Anaesthetisation prior to sampling has been

demonstrated to have no effect on plasma ion concentrations (Blackburn and Clarke, 1987).

Blood was sampled through the caudal ventral vein using heparinized 100 pi syringes

and 60 pi microhematocrit tubes. At the same time duplicate 20 pi blood samples were also

taken into small test tubes containing 10 ml Drabkin's solution for later hemoglobin

measurement. The microhematocrit tubes were then centrifuged at 11,500 rpm for 3 min in a

123

Damon IEC MB microhematocrit centrifuge and hematocrit (Hct) was measured in

quadruplicate. Fish were then weighed and dissected. Approximately 0.1 g gill and 0.3 g

kidney tissues were taken from each fish and maintained in Eppendorf tubes with 500-ul SEI

solution (0.3 M sucrose, 0.02 M EDTA, 0.1 M Imidazole) for the Na+/K+-ATPase assay

(Zaugg, 1982). About 1 g white muscle was removed from the same place in each coho

salmon and weighed before and after being dried to constant weight in the oven. The fish

carcass, tissue samples and plasma were stored at -80°C for later analysis.

Analytical techniques:

Blood hemoglobin concentrations were determined by analysing the hemolyzed blood

and Drabkin's solution mixture at 540 nm using a Shimadzu UV-160 visible recording

spectrophotometer. Plasma sodium concentrations were measured with a Perkin-Elmer model

2380 atomic absorption (aa) spectrophotometer. The plasma held in microhematocrit tubes

under -80°C prior to ion level determination was thawed and diluted to within the linear range

of the machine's detection. Gill and kidney Na+/K+ ATPase activity assays were conducted

according to procedures described by Zaugg (1982).

Statistical analysis:

All data are expressed as mean ± standard error. Statistical differences between

treatment and control groups were determined using parametric one way ANOVA followed

by a Dunnett's test for freshwater groups, and student 7-test for seawater groups. Under

124

conditions where the equal variance test failed, the non-parametric equivalent was used as

appropriate. A probability level of 0.05 was chosen as the limit of statistical significance.

125

RESULTS

Rainbow trout did not tolerate the transfer from freshwater directly to full strength

seawater. No trout, including controls, survived more than three days in saltwater. No

mortality was recorded in the rainbow trout control group, however, 24 h after seawater

transfer while 5.5% of the trout died in the groups pre-exposed to TeCB and TeCG. On day-3

in seawater, the mortality increased drastically in all groups, reaching 50, 83 and 100% in the

control, TeCB, and TeCG groups, respectively. Although the test was intended to be

sublethal, these results indicate that TeCB and TeCG may impair the ability of rainbow trout

to withstand any subsequent seawater challenge. Coho salmon, on the other hand, were much

better at dealing with hyperosmotic stress. It was not surprising to see all the TeCG-exposed

coho survived the transfer well into the end of the 8-day depuration period in clean seawater.

Haematology:

Rainbow trout showed a gradual, but not significant, decline in haematocrit and

haemoglobin content during depuration in freshwater (Tab. 3) after a three-day sublethal

exposure to 200 mg-1"1 TeCB or TeCG. The same pattern was observed with coho salmon

throughout the 3-day and 11-day 100 M-g'l"1 TeCG exposure (Tab. 4) and the 8-day post­

exposure period in clean seawater (Fig. 23).

126

Table 3. Effects of TeCB and TeCG on the haematocrit (Hct) (%), haemoglobin content (Hb) (g-dl"1), mean cell haemoglobin concentration (MCHC) (100'Hb-Hcf1), plasma sodium level (mmoM'1) and gill, kidney Na+/K+-ATPase activity (pimol P, -mg protein-h"1) of adult trout (Oncorhynchus mykiss) during 200 (J-g-1"1 TeCB or TeCG exposure and depuration in freshwater.

127

•s s

so o

r-H co

ON C N O in NO

r-O •<*

IT) C N

ro NO

SO

r- O so 00 in

.10

+1.

00 +1 I-H ro

00 +1 ro "°.

•<*' +1 1 ro

ro +1 ON

00 +1 o co ?;

C N r~

+i o\ o

+i C N ro

i—I +1 o

00 +1 SO i—i

co +1 co O

+i SO SO

ro' +1

<N O 1-C N in ON

co C N 00 in C N as C N ro 00

CO

^ "* r-1 ^ O N oo r~ oo

? ?; +i ?; ? "1 so C N 00 r-H C N C N oo in ro © oo t— so so

rr1 C N co ON O

O 00 I-H >*

§ +; ?; ?; ?; , -* CN V> CN csi (N "* CN r--

°^ so CN CN * * C O so O N

SO SO so so

3 5 ? 3 +U •n r-H rH ON S C N ON r~- WI CO J O SO •*' m \© £

o e E.

Z S3

s

o C N 00 O N r- -* C N

in C N CO 00 in so sq C N r-H ,—' <n rH +1 +1 +1 +1 +1 +1 +1 +1 in CO o >n C N •* 00 in in 00 SO 00 CO r~ © O N SO CO r-' so C N CO ro ro CO CO CO

l-H rH i—i rH i-H rH i-H I-H

* ON 00 o 00

* so

>n

r-- rH so >n >n 00 +1 +1 +i m so >n 00 >n 00 r-' ON SO CO CO CO

(-H rH I—1

o ON

CN 1 +i +) +i

<n CN CN

m n

u n u

© © © - H

so so <n co in

oo O H H

C N co in oo SO 00 ON 00 n-

O rH O rH

so so in co in

oo o •* ON

SO CN 00 C N O N O

re in

O CN so ' CO

CN 00 Tt" >n co 5 5 5 ? ? 5 5 5 5 5 5 5 5 5 5

00 H CO I

in ir

SO i •* 'i

a

o

o

•a

w

NT

4} a. E

t /5

00 SO

CN ON

O N r-o

ro O 00

so so SO

I-H o o

00 00 00 SO

ro so •*

ro

o ro ON

,94

±0

+1 ON

in ? in o

I-H

+1 O N CN

+1 ro O ? r^

00

i-H +1 r~->n

i—i +1 O N CN

?; O N CO O N

?; SO 00

i—i +1 SO 00

+1 CO

?; CO CN

o CN CN 00

r-H

SO l-H

r~' o CN

00 00 I-H

SO l-H

so' 1—1 O CN in rH

CN rH

1—1

CO H O N

CO o SO O N

in ro O N O o O in

"* O N

CO SO ON

O N

in ro 00

.08

+1 CO

+1 CO 00

+! SO ro

+1 00 00 +i O N 00

+1 00 o

+i in 00 ON

ro +1 SO

I-H +1 1 in

+1 00 o

l-H +1 CN CN

+i ro sq

?i ro

00 in ON

CO r-' ro

CN CO ro CO CN CO CN CO o in ro

r-' CN

o CO

T J

T3

o U

I

o

T ) i c o I

s o -rJ -rJ ^ a U P C H

<o T J

O OH g, CJ

T J

OH O

T J T J

H

rH CO 00 m >. i aj & T J T J T J

T J c B CJ O o ,o rH 'H-J '•*-»

s CS cd cs rH rH rH o P P P OH OH OH OH X CJ CJ t> CJ T J T J T J

•a s o

T J

T J i U rH P o X C J

l-H CO 00 & >N

cd cd T J T J T J

JS c P _o O O •a '43 cd cd cd tH rH rH P P P O , OH OH CJ CJ CJ T J T J T J

CA O

o

o

p CJ cd CJ U P CJ CJ CJ -o CJ CJ rH

!9 T J

cd o

'2 OJO

cd o T J P

CJ

r5

128

Table 4. Effects of TeCG on the haematocrit (Hct) (%), haemoglobin content (Hb) (g'dT1), mean cell haemoglobin concentration (MCHC) (100 Hb-Hcf 'Xplasma sodium level (mmolT1), muscle moisture (%) and gill, kidney Na+/K+-ATPase activity (jlmol P,-mg protein-h"1) of adult coho salmon (Oncorhynchus kisutch) during 100 pigi"1

TeCG exposure and depuration in freshwater or continuous exposure for 11 days in freshwater.

129

T 1 3

S 3 w ft

jO T3 J3 -g s n .3 S S u

3 o ft <->

S o o o O >—1

X *°

£1 at s * •a £

03 no

O -S

a 8 I t 3 3

u -3

o o ca a I

-r o 'a ^ to

s bp

a a 0 ••_ •S o 1 * .

O ID ^

S J2 ^ <D ca .a J= a o a o H

3

3

-a

§ 1

•s 1

0 ft

g g 1

O

CS

i f £

s

3

+ C «

z

« 2

W •

i f o

•a W

n

a,

t/3

NO •«t

ON NO Tl-ON r-_ C N co NO •* ON 00 NO

r-in co i—i NO ON o

00 o

C N O NO O

.5912

+1 +-; o ON co +1

00 O

+1 i—i

+1 Os VI

+i o r-_

C O +1

00 +1 NO in

+i C N

O N <r)

^ ?i o o O m +i NO 00

+1 >n in in C N

ON NO C N

r-C N

00 C N

in C N co co C N co NO «n

C N

C O O i—i C N ON >n

00 >n

C N 00 C N C O

o •<* 00 in

* C N 00 o

NO C O

00 in

* C N 00 in ON

in 0+86

< N +1 C O 00

+1 o NO

+i o ON ? in ON

+i

ON in C N

?; Os

?; in ON +i

ON ? l

o ?; C N O N

C O

in Tt o in C N o in C N C N

NO cn o NO C N

NO C O NO

* in C N

00 O N

O in

NO C O

r~ NO

* 00 in

NO ON C N ©

.87±2

+1 O

+1 C N NO

>n +1 C N C O

in +1 C N 00

+J 00

« +i o o NO

+1 C N • *

•<*' +1 C O r- o

00 TJ-

NO +1

C O +1

O NO

C O

+1 in oo r-(*N

NO m

C N C O

C O C O C O C O

t~ C N

O N O i—1 C O C N C O C N

C N O C N C N in

NO OS cn

C O C O

r-

00 NO NO *

O N NO o

NO ON C N

NO NO *

.13

+1

+1 00 ON +1 Os

7—i +1 C O

?; C N

+1 C O

5 o C O

?; o

+\ NO r~ +i NO o

+i C O r-H o

C O

?; C N

+i O N r-

+1 in in cn

NO o NO

o NO

NO NO C O NO o NO

NO 00 in

NO NO C O NO o NO

in NO

00 NO NO

C N O N cn

00 NO ON o C N * *

00 in o

o C N 00 ON ON NO C O

.97+

2

C N

+1

>n +1 o o

+i NO C O

+1 1—1 1—1

C N +1 1^ ON

+1 +1

NO C O +1 00 00

7—i +1 o >n

+i

ON +1 ?; in 00

+1

m +1 C N

t~ >n <—i NO

o NO

("-" in ©

NO r--

in ON in

o NO

C N in in r-' m ON in

C N

NO ©

NO in in

t~ cn

>n 00

1—1

NO 00 in r~ C O

o ON in

OS •>* O N

r-co o C O ©

m in i

.77±0

?; ?; in

?; C N NO

?; ON

?! r~

+i C O 00

?i 00 o ?; Os

? i 00 00 ?; +1

C O 00 ?; C N O

? l Tf

ON ?; i NO • C O !

00 Os 00 00 oo' 00 i> ON in NO 00 p~' 00 NO i

NO C N

C N 00 o NO

OS I-

o NO C N

o

00 o ON

in in Tf NO C N

o 00

C O r- i

.67±1

+i cn cn

?; in C N

+1 NO

?; o in +1

NO 1—1

+1 C O 00

?; NO

?; +1 C O C O

+i r-NO

+i C O 00

+1 NO

?; NO

+1 i co ; 00 : C N

cn C N cn

C N C O

C N C O C O

C N C O

ON C N C O

o C O

ON C N

C N C O ON

C N

00 C N

ON C N

ON i C N i

9 S 3 >> ft ft T3 D T3

O

u

e o

I PH O

u 3 ft

T 3

I O

u-3 ft

•a T3

u H

a 3 O

X

CS •a

C O 00 CS ca ca T3 V V el 3 3

o o o •& ca ca CS u l-3 3 3 ft ft ft <L) u •a T3 •a

C O ^ > NO £ CS cS CS V •a i kH 3 3 3

© CO • O o o & ft ft CS X X X

T3 CL) ID

T 3

O

X a)

o &

o u a

H

3 O s B

O

u 3 O ft H V

m © © V &

3 °

t cs

3 o '•e CS ft •s 3 o

1 o o

3

I a 3 (L) I D

CJ o 3 CJ

s ca o ia '3 0J3 ta

T 3

a

CL>

O

2

Figure 23. Effects of TeCG on coho salmon haematological factors. Day 0-3 is freshwater exposure (100 ug-1"1) and day 3-8 is clean seawater depuration.

131

Time (day)

Plasma sodium and muscle moisture:

Rainbow trout treated with TeCB or TeCG showed no significant change in plasma

sodium levels until after being held for 3 days in clean freshwater following toxicant dosing,

when a decrease was observed throughout the rest of depuration in both the TeCB and TeCG

groups (Tab. 3). Plasma sodium concentration was fairly stable in the freshwater coho

salmon groups until 4 days after the beginning of toxicant exposure, when a significant

decrease (p<0.05) was seen (Tab. 4). Rainbow trout plasma sodium levels were elevated in

TeCB and/or TeCG predosed fish given a seawater challenge (Fig. 24). Plasma [Na+] in coho

salmon exposed to TeCG was not significantly higher (p<0.05) than that in control group

fish until 8 days after being transferred to toxicant free seawater (depuration-day-8) (Fig.

25). Muscle moisture data were collected from coho salmon and a significant water gain in

muscle was found only at depuration-day-1 and exposure-day-4 and 6 in freshwater (Tab. 4).

There was a non significant trend of increased muscle water content when the coho were

transferred to and then held in seawater (Fig. 25).

Na/K*-ATPase:

A significant decrease in gill Na+/K+-ATPase was observed in freshwater groups at the

end of the 3-day exposure and ATPase activity remained low until depuration-day-1 in TeCB

and depuration-day-3 in TeCG dosed rainbow trout, respectively (Tab. 3). There was a

tendency for recovery of ATPase activity toward the end of freshwater depuration, which was

also apparent in the TeCG exposed coho salmon either post exposure, or during the

continuous exposure period in freshwater (Tab. 3 & 4). In the 24 h seawater transfer test,

133

Figure 24. Effects of 3-day TeCB and TeCG exposure (200 (igT1) in freshwater on plasma sodium level and gill Na+/K+-ATPase activity of rainbow trout during 24 h seawater challenge. Asterisk indicates a significant difference (p<0.05).

134

P O C/5

tZ3

PM

o £ =

250

200

150

100

50

0

1 I SI

H U < es

.0-H

+

d

"53 O

o. WD a

12

10

8

6

4

2

0

A.

I

I

I CONTROL TCB

I TCG Exposure day-0 hNNXI Exposure day-3

24 hours after seawater transfer

135

Figure 25. Effects of TeCG on coho salmon plasma sodium level and muscle moisture content. Day 0-3 is freshwater exposure (100 (igT1) and day 3-8 is clean seawater depuration. Asterisk indicates a significant difference (p<0.05).

136

Time (day)

rainbow trout gill Na /K -ATPase remained inhibited after being held one day in seawater no

matter whether the fish were exposed to TeCB or TeCG (Fig. 24). An upregulation of

this enzyme in coho salmon gill, however, was found at the end of 3-day TeCG exposure

(p<0.05) and no significant difference was detected at other sampling times (Tab. 4, Fig. 26).

The kidney Na+/K+-ATPase activity in rainbow trout was not influenced throughout the

whole test (Tab. 3). TeCG exposed coho salmon, on the other hand, showed an inhibition in

kidney Na+/K+-ATPase at the end of either the 8-day depuration seawater or after being

continuously exposed to TeCG for 11 days in freshwater (Tab. 4, Fig. 26).

138

Figure 26. Effects of TeCG on coho salmon Na /K+-ATPase activity in gill and kidney. Day 0-3 is freshwater exposure (100 (igi"1) and day 3-8 is clean seawater depuration. Asterisk indicates a significant difference (p<0.05).

139

0 2 4 6 8 10 12

Time (day)

DISCUSSION

Both increases and decreases in haematological factors have been observed in response

to environmental pollutants (Heath, 1987). In this study, however, neither rainbow trout nor

coho salmon demonstrated a significant change in either haematocrit or haemoglobin levels.

The calculated mean cell haemoglobin concentration was fairly stable throughout the whole

experiment for both species, indicating no net loss of red blood cells from the circulating

blood. As a result, the oxygen carrying capacity of the dosed fish should not be altered and it

is, therefore, unlikely that the gas exchange in the gill was affected. This is in agreement with

the observation that there is no impact of sublethal exposure to TeCB on fish swimming

performance (Brauner et al, 1994). Although haematological factors are usually easy to

measure compared with other parameters, their varied and reduced response to sublethal

exposure to TeCB or TeCG exclude the possibility of them being used as effective biomarkers

in this case.

In freshwater, the changes in plasma [Na+] in coho salmon, although exposed at half the

concentration of toxicant to which rainbow trout were exposed, were more rapid than that

observed in trout. If the toxicant-exposed fish were transferred to seawater, rainbow trout

showed elevated plasma sodium 24 h after seawater transfer while no change was seen in

coho salmon until they had been held in clean saltwater for 8 days, suggesting coho salmon

appeared to be more capable of maintaining the plasma ion levels when subjected to a

hyperosmotic environment (Fig. 24 & 25). For freshwater tests, therefore, coho salmon

indicates more sensitivity than rainbow trout if plasma ion concentration is used as a

141

biomarker; however, rainbow trout show a greater osmoregulatory impairment if seawater

challenge tests are chosen. Plasma samples are easy to collect and relatively convenient and

less costly to measure compared with other physiological parameters, such as enzyme activity.

A wide range of organic xenobiotics are capable of affecting fish active ion transport

(Evans, 1987). The inhibition of gill Na+/K+-ATPase has been shown in different fish exposed

to a variety of organic chemicals. Davis and Wedemeyer (1971) reported that the

organochlorines, DDT, dicofol, and endosulfan inhibited rainbow trout gill Na+/K+-ATPase

by 60 to 100% in vitro at concentrations between 10"5 and 10"4M In this study, the response

time for both species in freshwater was similar except that coho recovered a little faster than

trout (Tab. 3 & 4), although the toxicant-induced change was in the opposite direction; gill

Na+/K+-ATPase for rainbow trout exposed to a nominal concentration of 200 pigT1 TeCB or

TeCG for 3 days showed an inhibition, while an upregulation was found in coho salmon after

being dosed with 100 (ig'l1 TeCG for the same amount of time. An inhibition of Na+/K+-

ATPase was observed in both TeCB and TeCG treated trout 24 h after the seawater challenge

(Fig. 24). The enzyme activity in coho gill recovered after transfer to seawater (Fig. 26), as in

freshwater (Tab. 4). These findings agree with the aforementioned observation that coho

salmon, as smolts, are less sensitive during freshwater to seawater transition than rainbow

trout. Elevated growth of juvenile coho salmon exposed to kraft pulpmill effluent was

reported and it was found to be dose-dependent (Mason and Davis, 1976). Although there

was no direct evidence, a variety of hypotheses to explain this effect were proposed, including

hormone analogs in the wood extracts that stimulate growth hormone and/or appetite. Another

possibility is that hormesis occurred (Sansone and Martens 1981). The term "hormesis" refers

142

to an overcompensation to some inhibitory challenge. The stimulation seen in the gill Na+/K+-

ATPase activity of coho salmon could be described as a form of hormesis.

The kidney also plays an important role in osmoregulation, particularly in freshwater

fish where glomerular filtration is high and extensive reabsorption takes place. According to

the results of this study, kidney Na+/K+-ATPase in rainbow trout was largely unaffected,

whereas the enzyme was inhibited 3 days into seawater depuration and at the end of continued

freshwater exposure in the TeCG treated coho salmon, although the response was much

delayed in the kidney compared with gill (Tab. 4, Fig. 26). The inhibition of the kidney

Na+/K+-ATPase was similar in freshwater and seawater fish, indicating that the reduction was

a direct effect of the toxicant on the ATPase rather than any secondary response to osmotic

and ionic changes. It would seem that coho kidney Na+/K+-ATPase is more responsive than

that of trout.

The parameters monitored are directly or indirectly related to each other. The inhibition

of gill Na+/K+-ATPase in rainbow trout could have been the cause for the subsequent fall in

plasma ions. For TeCG exposed coho, however, the increase in plasma [Na+] was associated

with no change in gill Na+/K+-ATPase activity and a decrease in kidney Na+/K+-ATPase

activity, which should have resulted in a drop in sodium level, rather than the observed

increase (Fig. 25 & 26). Also, the upregulation of gill Na+/K+-ATPase was not consistent with

the decrease in plasma [Na+] recorded in the freshwater TeCG treated coho (Tab. 4). It would

appear, therefore, that the changes in plasma sodium are not brought about by the changes in

Na+/K+-ATPase in the body but are due probably to increased passive flux of water and/or

sodium, induced by changes in gill membrane permeability. Coho salmon muscle moisture

143

content increased at about the same time as plasma sodium levels decreased in the freshwater

groups (Tab. 4), indicating that in freshwater the reduction in plasma ion concentration

(hemodilution) is associated with an increased water content in muscle. That is, the dilution is

general and water entering the body is distributed to both blood and muscle tissues.

McCarty (1995) has argued that biomarkers, i.e., changes in physiological or

biochemical parameters, can become a useful and scientifically valid tool if a biomarker-

biological indicator relationship is established. In general, plasma sodium level, gill Na+/K+-

ATPase activity, and/or muscle moisture are potential biomarkers. Na+/K+-ATPase in gill is a

better choice as a biomarker than that in kidney and its sensitivity to the toxicant is even

higher than that of plasma [Na+]. The enzyme assay, however, is more time consuming and

costly relative to ion measurements. Nevertheless, the efficiency of Na+/K+-ATPase analysis

has been greatly enhanced by the development of a microassay method (McCormick, 1993),

and could possibly serve as a valuable, easy-to-apply biomarker. Based on this study, muscle

moisture content is indicative of disturbed ion balance in coho salmon to the same degree as

the direct measurement of plasma ion content, and as a result it could be used as a less

expensive biomarker. Muscle moisture, plasma ion level and ion transport enzyme activity,

although functionally linked, do not always co-vary.

In addition, all these parameters show considerable normal variability. Thus, based on

this study, the potential use of enzyme activity, plasma Na+ levels or muscle moisture as

biomarkers is of limited value. In the Fraser River, hundreds of thousands of salmon are

undergoing smoltification each year, and the ability to osmoregulate will certainly determine

fish survival rate after seawater transition. Increased sodium concentration in the plasma has

144

been found to affect the critical swimming speed Ucrit in coho salmon (Brauner et al, 1992).

Assuming fish aerobic swimming performance is impaired due to the failure to maintain ion

balance after going into the ocean, their survival could be at stake not only due to

osmoregulatory failure but also by their being less capable of escaping from potential

predators. The parameters measured in these experiments are valuable, not as potential

biomarkers, but rather because they indicate that both TeCB and TeCG impair

osmoregulation. Such an impairment could have a detrimental effect on the size of the fish

population. The impact of these toxicants might be minimised if the exposed fish are left for

several days to recover in freshwater before they head for the ocean. Consequently, location of

pulp mills to either freshwater or seawater sites, but away from the estuary could potentially

reduce the adverse effects of TeCB and TeCG on the local fish population.

145

LITERATURE CITED

Andersen, M.E., H.J. Clewell, M.L. Gargas, FA. Smith and RH. Reitz. 1987. Physiologically based pharmacokinetics and the risk assessment process for methylene chloride. Toxicol. Appl. Pharmacol. 87: 185-205.

Barron, M.G. 1990. Bioconcentration. Will water-borne organic chemicals accumulate in aquatic animals? Environ. Sci. Technol. 24(11), 1612-18.

Barron, M.G., Stehly, G.R. and Hayton, W.L. 1990. Pharmacokinetic modeling in aquatic animals. I. Models and concepts. Aquatic Toxicology 18:61-86.

Beamish, F.W.H. 1978. Swimming capacity. In. W.S. Hoar and D.J. Randall (eds). Fish Physiology, Vol 7. Academic Press, New York, NY, pp. 101-187.

Bertelsen, S.L., AD. Hoffman and CA. Gallinat. 1993. Determination of in vitro chemical partition coefficients in fish tissues: A comparative study of four species. EPA-600/J-93-045.

Bishop WE. and A.W. Maki. 1980. A critical comparison of two bioconcentration test methods. Aquatic Toxicology, ASTM STP 707, J.G. Eaton et al, Eds., pp. 61-77.

Black, M.C., and J.F. 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.

Black, M.C., D.S. Millsap and J.F. McCarthy. 1991. Effects of acute temperature change on respiration and toxicant uptake by rainbow trout, Salmo gairdneri (Richardson). Physiol. Zool., (in review).

Blackburn, J. and W.C. Clark. 1987. Revised procedure for the 24 hour seawater challenge test to measure seawater adaptability of juvenile salmonids. Can. Tech. Rep. Fish. Aquat. Sci. 1515: 35.

Brauner, C. J. 1991. The relationship between oxygen and 1,2,4,5-tetrachlorobenzene uptake, in five species of fishes, M.Sc. Thesis, The University of British Columbia.

Brauner, C.J, J.M. Shrinpton, and DJ. Randall. 1992. Effects of short-duration seawater exposure on plasma ion concentration and swimming performance of coho salmon (Oncorhynchus kistch) parr. Can. J. Fish. Aquat. Sci. Vol. 49. 11, pp.2399-2405.

Brauner, C.J., D.J. Randall, J.F. Neuman and R.V. Thurston. 1994. The effect of exposure to 1,2,4,5-tetrachlorobenzene and the relationship between toxicant and oxygen uptake in rainbow trout (Oncorhynchus mykiss). Environ. Toxicol. Chem., Vol. 13, No. 11, pp. 1813-1820.

Brett, JR. 1964. The respiratory metabolism and swimming performance of young sockeye salmon. J. Fish. Res. Bd. Canada, 21(5): 1183-1226.

147

Brett, J.R., Zala, CV. 1975. Daily pattern of nitrogen excretion and oxygen consumption of sockeye salmon (Oncorhynchus nerka) under controlled conditions. J Fish Res Board Can 32:2479-2486.

Bruggeman, W.A., Matron, L.B.J.M., Kooiman, D. and Hutzinger, O. 1981. Accumulation and elimination kinetics of di-, tri and tetrachlorobiphenyls by goldfish after dietary and aqueous exposure, Chemosphere, 10(8):811-832.

Burggren, W.W. and J.N. Cameron. 1980. Anaerobic metabolism, gas exchange, and acid-based balance during hypoxic exposure in channel catfish, Ictaluruspunctatus. J. Exp. Zool. 213: 405-416.

Bushnell P.G., Steffensen J.F., Johansen K. 1984. Oxygen consumption and swimming performance in hypoxia-acclimated rainbow trout (Salmo gairdneri). J. Exp. Biol. 113: 225.

Cameron, J.N. and Davis, J.C. 1970. Gas exchange in rainbow trout {Salmo gairdneri) with varying blood oxygen capacity. J. Fish. Res. Bd. Canada. 27:1069-1085.

Chiou, C.T., V.H. Freed, D.W. Schmedding and R.L. Kohnert. 1977. Partition coefficient and bioaccumulation of selected organic chemicals. Environ. Sci. Technol. 11: 475-478.

Clark, K.E., F.A.P.C. Gobas and D. Mackay. 1990. Model of organic chemical uptake and clearance by fish from food and water, Environ. Sci. Technol., 24:1203-1213.

Connell, D.W. 1990. Bioaccumulation of xenobiotic compounds, CRC Press, Inc., Boca Raton, Florida. 213pp.

Davis PW. and G.A. Wedemeyer. 1971. Na+, Reactivated ATPase inhibition in rainbow trout: a site for organochlorine pesticide toxicity? Comp. Biochem. Physiol. 40B:823-827.

Dobbs, A.J. and N. Williams. 1983. Fat solubility- a property of environmental relevance? Chemosphere. 12(1):97-104.

Duval W.S., Fink R.P., Geen G.H. 1981. An aquatic respirometer for use in sublethal studies. Water Res. 15:817-822.

Eddy, F.B. 1981. Effects of stress on osmotic and ionic regulation in fish. In: A.D. Pickering (Ed.), Stress in fish: Academic Press, London, 77-102.

Erickson, R.J. and J.M. Mckim. 1990. A model for exchange of organic chemicals at fish gills: Flow and diffusion limitations. Aquat. Toxicol. 18, 175-198.

Evans, D.H. 1987. The fish gill: site of action and model for toxic effects of environmental pollutants. Env. Health Perspect. 71: 54-58.

148

Folch, J., M. Lees and G.H. Stanley. 1957. A simple method for the isolation and purification of total lipids from animal tissues. Journal of Biology and Chemistry 226: 497-509.

Garrett, CL. 1980. Toxic Organic Contaminants, Water Quality, Fraser River Estuary Study. Government of Canada.

Gehrke, P C , L.E. Fidler, D.C. Mense and D.J. Randall. 1990. A respirometer with controlled water quality and computerized data acquisition for experiments with swimming fish, Fish Physiology and Biochemistry, 8(1):61-67.

Geyer, H., I. Scheunert, R. Bruggemann, V. Zitko, C.E.W. Steinberg, R. Pruell, A. Kettrup, J. Olson, P. Schmieder, D.C.G. Muir. 1995. Relationship between elimination half-lives (T1/2) of 2,3,7,8-tetrachlorodibenzo-/>-dioxin (TCDD) in different mussel and fish species and their lipid content or body size, Organohalogen Compounds, Vol.25, 273-278.

Gingerich, W.H. 1986. Tissue distribution and elimination of rotenone in rainbow trout, Aquatic Toxicology, 8: 27-40.

Gingerich, W.H. and J.J. Rach. 1985. Uptake, accumulation and depuration of 14C-retenone in bluegills (Lepomis macrochirus). Aquat. Tox. 6: 170-196.

Gobas, F.A.P.C. 1993. A model for predicting the bioaccumulation of hydrophobic organic chemicals in aquatic food-webs: application to Lake Ontario, Ecological Modelling, 69:1-17.

Gobas, F.A.P.C. and Mackay, D. 1987. Dynamics of hydrophobic organic chemical bioconcentration in fish. Environ. Toxicol. Chem., 6: 495-504.

Gobas, F.A.P.C, Opperhuizen, A. and Hutzinger, O. 1986. Bioconcentration of hydrophobic chemicals in fish: relationship with membrane permeation. Environ. Toxicol. Chem., 5:637-646.

Goss, G.G. and Wood, CM. 1988. The effects of acid/aluminum exposure on circulating plasma Cortisol levels and other blood parameters in the rainbow trout, Salmo gairdneri, J Fish Biol. 32: 63-76.

Hamelink, J.L. and Spacie, A. 1971. A proposal: exchange equilibria control the degree chlorinated hydrocarbons are biologically magnified in lentic environments, Trans. Am. Fish. Soc, 100: 207.

Hawker, D.W. and Connell, D.W. 1985. Relationship between partition coefficient, uptake rate constant, clearance rate constant and time to equilibrium for bioaccumulation, Chemosphere, 14(9): 1205-1219.

Hayton, W.L. and M.G. Barron. 1990. Rate-limiting barriers to xenobiotic uptake by the gill. Environ. Toxicol. Chem. 9:151-157.

149

Heath, A.G. 1987. Water pollution and fish physiology. CRC Press Inc., Boca Raton.

Hoar, W.S. and Randall, D.J., Eds. 1984. Fish Physiology, Vol. 10 (Part B), Gills: Ion and Water Transfer, Academic Press, New York.

Hocutt, CH. and Stauffer, J.R., Eds. 1980. Biological Monitoring of Fish, Lexington Books, Lexington, KY.

Holden, A.V. 1962. A study of the absorption of 14C-labelled DDT from water by fish. Ann. App. Biol. 50, 467-477.

Hughes, G.M. 1973. Respiratory responses to hypoxia in fish, Am. Zool. 13: 475.

Janz, DM., Farrell, A.P., Morgan, J.D., and Vigers, GA. 1991. Acute physiological stress responses of juvenile coho salmon (Oncorhynchus kisutch) to sublethal concentrations of Garlon 3A and Vision herbicides, Environ. Toxicol. Chem. 10: 81-90.

Johansen, J.A. and Geen, G.H. 1990. Sublethal and acute toxicity of the ethylene glycol butyl ether ester formation of triclopyr to juvenile coho salmon (Oncorhynchus mykiss), Bull. Environ. Contam. Toxicol. 19:610-616.

Jones, D.R. and Randall, D.J. 1978. The respiratory and circulatory systems during exercise. In: Fish Physiology. (Hoar, W.S. and D.J. Randall, Eds.). Vol 7. Academic Press, New York. pp. 425-501.

Josephson, J. 1982. Humic substances. Environ. Sci. Technol. 16: 20A-24A.

Kathryn E C , A.P.C. Gobas and D. Mackay. 1990. Model of organic chemical uptake and clearance by fish from food and water, Environ. Sci. Technol., 24: (8) 1203 - 1213.

Kenaga, E.E. and C.A.I. Goring. 1980. Relationship between water solubility, soil sorption, octanol-water partitioning, and concentration of chemicals in biota. In: J.G. Eaton, P.R. Parrish, and A.C. Hendricks, eds., Aquatic Toxicology, ASTM ATP 707. American Society for Testing and Materials, pp. 78-115.

Kiceniuk, J.W. and Jones, D.R. 1977. The oxygen transport system in trout (Salmo gairdneri) during sustained exercise. J. Exp. Biol. 69:247-260.

Landrum, P.F., Reinhold, M.D., Nihart, S.R. and Eadie, B.J. 1985. Predicting the bioavailability of organic xenobiotics to Pontoporeia hoyi in the presence of humic and fulvic materials and natural dissolved organic carbon. Environ. Toxicol. Chem., 4: 459-467.

150

Leach, J.M. and A.N. Thakore. 1975. Isolation and identification of constitutes toxic to juvenile rainbow trout (Salmo gairdneri) in caustic extraction effluents from kraft pulp mill bleach plants. Journal of the Fisheries Research Board of Canada 32: 1249-1257.

Lopez-Avila V. and R.A. Hites. 1980. Organic compounds in an industrial wastewater. Their transport into sediments. Environ. Sci. Technol. 14: 1382-1390.

MacKay, D. 1982. Correlation of bioconcentration factors. Environ. Sci. Technol. 16: 274-278.

MacKinnon, D.L. and Farrell, A.P. 1992. The effect of 2-(thiocyanomethylthio) benzothiazole on the juvenile coho salmon (Oncorhynchus kisutch): sublethal toxicity testing, Environ. Toxicol. Chem. 11:1541-1548.

Mason B.J. and J.C. Davis. 1976. Growth stimulation and biochemical changes in juvenile coho salmon Oncorhynchus kisutch during chronic exposure to sublethal levels of neutralized bleached kraft mill waste. In: Davis J.C. (Ed.) Progress in sublethal effect studies with kraft pulpmill effluent and salmonids. J. Fish. Res. Bd. Can. 33:2031.

Means, J.C, S.G. Wood, and J.J. Hassett and W.L. Beanwart. 1980. Sorption of polynuclear aromatic hydrocarbons by sediments and soils. Environ. Sci. Technol. 14: 1524-1528.

McCarty, L.S. 1995. Environmental biomarkers: fiction, fantasy, or functional? The 2nd SETAC World Conference, Abstract book. p. 39.

McCarthy, J.F. and Jimenez, B.D., 1985. Reduction in bioavailability to bluegills of polycyclic aromatic hydrocarbons bound to dissolved humic materials. Environ. Toxicol. Chem., 4: 511-521.

McCormick, S.D. 1993. Methods for nonlethal gill biopsy: Measurements of Na+/K+-ATpase activity. Can. J. Fish. Aquat. Sci. Vol. 50. pp. 656-658.

McKim, J., Schmieder, P., 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.

Murphy, P.G. and J.V. Murphy, 1971. Correlation between respiration and direct uptake of DDT in the mosquito fish Gambusia affinis. Bull. Environ. Contam. Toxicol. 6, 581-588.

Nichols, J.W., J.M. McKim, G.J. Lien, AD. Hoffman, S.L. Bertelsen and CA. Gallinat. 1993. Physiologically-based toxicokinetic modelling of three waterborne chloroethanes in channel catfish, Ictahuruspunctatus. Aquatic Toxicology, 27: 83-112.

Neely, W.B., D.R. Branson and G.E. Blau. 1974. Partition coefficient to measure bioconcentration potential of organic chemicals in fish. Environ. Sci. Technol. 8: 1113-1115.

151

Randall, D.J. 1982. The control of respiration and circulation in fish during exercise and hypoxia, J Exp Biol 100: 275.

Randall, D.J. and C.J. Brauner. 1990. Toxicant uptake across fish gills. In: Proceedings of the second international symposium on Fish Physiology, Fish Toxicology, and Water Quality Management. Sacramento, California, September 18-20, 1990. United States Environmental Protection Agency, Environmental Research Laboratory, Athens, Georgia, USA.

Randall, D.J. 1990. Control and co-ordination of gas exchange in water breathers. In: Advances in comparative and environmental physiology. (R.G. Boutilier, Ed.). Vol. 6. Springer-Verlag, Berlin, Heidelberg, pp 253-278.

Randall, D.J., C.J. Brauner and R. Yang, 1993. Uptake of chemicals by fish, In: The Aquatic Resource Research Project Technical Report, Edited by A.P. Farrell, Simon Fraser University, pp. 181-194.

Randall, D.J. and C. Daxboeck, 1984. Oxygen and carbon dioxide transfer across fish gills. In: Fish Physiology, Vol. 10A: 263-307. W.S. Hoar and D.J. Randall(Eds). Academic Press Inc., New York.

Reinert, R.E. 1972. Accumulation of Dieldrin in an alga (Scenedesmus obliquus), Daphnia magna, and the guppy (Poecilia reticulata). J. Fish. Res. Bd. Canada. 29:1413-1418.

Reitz, R.H., P.S. McCroskey, C.N. Park, M.E. Andersen and ML. Gargas. 1990. Development of a physiologically based pharmacokinetic model for risk assessment with 1,4-dioxane. Toxicol. Appl. Pharmacol. 105: 37-54.

Rombough, P.J. and B.M. Moroz. 1990. The scaling and potential importance of cutaneous and branchial surfaces in respiratory gas exchange in young chinook salmon Oncorhynchus tshawytscha. J. Exp. Biol. 154, 1-12.

Rudd, R.L. and Genelly, R.E. 1956. Pesticides: their use and toxicity in relation to wildlife, Calif. Fish and Game, 7.

Saarikoski, J., Lindstrom, R., Tyynela, M. and M. Viluksela. 1986. Factors affecting the absorption of phenolics and carboxylic acids in the guppy {Poecilia reticulata). Ecotoxicol. Environ. Safety. 11:158-173.

Sansone, F.J. and C.S. Martens. 1981. Hormesis: A response to low environment concentrations of petroleum hydrocarbons. Sci. Vol. 211. pp.705-709.

Schmieder, PK. and T. A. Henry. 1988. Plasma binding of 1-butanol, phenol, nitrobenzene and pentachlorophenol in the rainbow trout and rat: a comparative study. Comp. Biochem. Physiol. 91C(2):413-418.

152

Shaw, G.R. and Connell, D.W. 1982. Factors influencing concentration of polychlorinated biphenyls in organisms from an estuarine system, Aust. J. Mar. Freshwater Res., 33: 1507.

Southworth, G.R., C.C. Keffer and J.J. Beauchamp. 1980. Potential and realized bioconcentration. A comparison of observed and predicted bioconcentration of amaarenes in the fathead minnow (Pimephales promelas). Environ. Sci. Technol. 14: 1529-1531.

Southworth, G.R., C.C. Keffer and J.J. Beauchamp. 1981. The accumulation and disposition of benz(a)acridine in the fathead minnow, Pimephales promelas. Arch. Environ. Contam. Toxicol. 10: 561-570.

Spacie, A. and J.L. Hamelink, 1982. Alternative models for describing the bioconcentration of organics in fish. Environ. Toxicol. Chem. 1, 309-320.

Spacie, A., P.F. Landrum and G.J. Leversee. 1983. Uptake and elimination of anthracene and benzo(a)pyrene in bluegill sunfish. (in press).

Sullivan, K.F., E.L. Atlas and CS. Giam. 1982. Absorption of phthalic acid esters from seawater. Environ. Sci. Technol. 16: 428-432.

Tarr, B.D., Mace G. Barron and William L. Hayton. 1990. Effect of body size on the uptake and bioconcentration of di-2-ethylhexyl phthalate in rainbow trout. Environ. Toxicol. Chem. Vol. 9, pp. 989-995.

Thomas, RE. and S.D. Rice, 1981. Metabolism and clearance of phenolic and mono-, di-, and polynuclear aromatic hydrocarbons by Dolly Varden char. In: W.B. Vernberg et al., eds. Physiological mechanisms of marine pollutant toxicity. Academic Press, New York., pp. 425-448.

Thurston, R.V. and P.C Gehrke. 1993. Respiratory oxygen requirements of fishes: Description of OXYREF, a datafile based on test results reported in the published literature. In: Proceedings of the second international symposium on Fish Physiology, Fish Toxicology, and Water Quality Management. Sacramento, California, September 18-20, 1990. United States Environmental Protection Agency, Environmental Research Laboratory, Athens, Georgia, USA.

USEPA, 1994. Predicting bioconcentration and residue-based effects of xenobiotics in and on aquatic biota. Cooperative agreement CR995189.

Veith, G.D., D.L. DeFoe and B.V. Bergstedt. 1979. Measuring and estimating the bioconcentration factor of chemicals in fish. J. Fish. Res. Board Can. 36: 1040-1048.

153

Wershaw, R.L. and M.C. Goldberg. 1972. Interaction of organic pesticides with natural organic polyelectrolytes. In: R.F. Gould, ed., Fate of organic pesticides in aquatic environment, advances in chemistry series III, American Chemistry Society, Washington D. C, pp. 149-158.

Yang, R. and Randall, D.J. 1995. Relationships between xenobiotic uptake/depuration and fish oxygen consumption: A physiological model. Proceedings of the 4th International Symposium on Fish Physiology, Toxicology, and Water Quality Management.

Yang, R. and Randall D.J. 1996a. Effects of Tetrachloroguaiacol (TeCG) on the osmoregulation of adult Coho Salmon (Oncorhynchus kisutch). Bull. Environ. Contam. Toxicol. 56: 788-795.

Yang, R. and Randall, D.J. 1996b. Biomarkers for rainbow trout (Oncorhynchus mykiss) and coho salmon (Oncorhynchus kisutch) exposed to 1,2,4,5-tetrachlorobenzene and 3,4,5,6-tetrachloroguaiacol, Chemosphere (In press).

Zaugg, W.S. 1982. A simplified preparation for adenosine triphosphatase determination in gill tissue. Can. J. Fish. Aquat. Sci. 39:215-217.

Zitko, V. and O. Hutzinger. 1976. Uptake of chloro- and bromobiphenyls, hexachloro- and hexabromobenzene by fish, Bull. Environ. Contam. Toxicol. 16: 665-673.

154