_ *, .* — _ — _ _ — _ — — — — = “ "'"""\"':, ." z: ' ' 't I "u 1 ’ .'." .’ fttio‘. - : J -1 3 3 a, ‘ -' .: ‘ . . ' .4 - . . .. .1 ..a\!.\- \32 .7.....‘- -5 ..'. i" 145 "\_3" ”V "~ ‘. ‘|.‘"” "f". " ‘~. '2‘“ ‘ ”—39% "" l ' _ ' v \ . ‘:' . , E _ . -‘.' ' -- .' '. 3- \ ‘ , .. 04’ '\:\r..;:. \Sx 5....1: \"..\. . x..-.\: xxx? .o. v '3: .2": :32: 2.2.: r.'. a; 1 x-z'sr'g '& '31-‘\ ‘t: "'-. 133‘s. . ‘ O f 2' L: g. . .3 I 432:3" ’3'.’ 5's. .. 5 far a! . “kl 1:3. ;; :2 f 2' .13“... ca 59 -14-! h" H ‘3"’3”\7" zfi'xn "we ‘ .. .1 -' .J' '. 0’ 4 ~ ' .I‘ .P': ‘ ‘, . Q a . 3 \‘ “ ...L.“¢ . .OV“ - '.,. ' u 0‘. a - arr. r. 'r-- ‘v‘fl ”"f" f‘" o u 1‘ .; v - on . ..A F .. (‘0..3-‘f‘ . no-.. '”"‘°'r"’ O u. g . ' o . ’ " ’1’- .‘.. 0:, '1 . ' 1. f \‘ ;' .t’l‘ \¢0.-. .1. .. u Jox.-t-. if, I I I. Q ‘ fl ‘ 0 g; - - Q A V a" . .‘.. fix/rt .. I 'i '9; .. ’ .'~ -"'. \' ‘ z 0 I 7'22-17X ‘ In ‘ I 3 i o 0" I Isl If n I 1+ 7 ‘_‘ _. ‘ .‘l ;' ~ . ‘u I” .~'- .I- ‘J K0. :4:- v ,’ . ‘ J 'A' 0 . - ' 3‘ ‘9‘“? V a. . ‘ ,_._ . .3 - “‘3’.“ ‘ . V U. "\ A q 0") ¢_.J-.} U . o a u ' 1' " ‘ ‘ v . .W x C ~‘ "" 1 3-: I ~ .‘o‘- L \ 'a' P . ‘ C a. F. -. \ I I I 0 1'9- 0". BINDING IV III)“ 8: SOIS’ 800K BINDEBY Ll? RARY muons .. ~ "‘91. mum; n mum; “1211mm!“ m1 1111 (Ill 11w" nu HM fl ll ABSTRACT EFFECTS ON LINEAR ALKYLATE SULPHONATE 0N TRITIATED WATER UPTAKE BY ISOLATED PERFUSED GILLS 0F RAINBOW TROUT By William F. Jackson Tritiated water (3H0H) uptake by isolated perfused gill arches of rainbow trout (Salmo gairdneri) was observed after exposure to step inputs (l, 5, lo, 20, and l00 mg l']) of the anionic detergent linear alkylate sulphonate (LAS). Perfusate flow (0.6 ml min-1) through the arches and the 3HOH and osmotic gradients were held constant. During the 65 min experimental period gills treated with 5, lO, 20, and 100 T 3 LAS responded with an exponential increase in HOH uptake with gml' time described by the equation: mm = %o(m)(1-e"‘t) where %D(t) is the percent deviation of sample activities from initial conditions (before LAS treatment), %D(m) is the equilibrium value at infinite time, and k is the rate constant. Isolated gills were also perfused with solutions containing either epinephrine (lo‘sM) or acetyl— choline (lOT8M). When these arches were treated with a lo mg l"1 step input of LAS similar responses were obtained. Epinephrine treated gills showed a greater response to the detergent exposure than gills perfused with saline alone. The results indicate that LAS has a rapid William F. Jackson dose—related effect on the 3HOH uptake of the isolated gill prepara— tion. It is suggested that the responses seen are due to changes in the permeability characteristics of gill mucus and/or epithelium, rather than to alterations in the pattern of perfusion flow through the gill. A compartmental model of the isolated perfused gill preparation is proposed. EFFECTS OF LINEAR ALKYLATE SULPHONATE ON TRITIATED HATER UPTAKE BY ISOLATED PERFUSED GILLS OF RAINBOW TROUT By William F. Jackson A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Physiology 1976 “My God, something happened!" Dr. L. F. Wolterink ii ACKNOULEDGMENTS I would like to extend special thanks to the members of my com- mittee: Dr. P. 0. Fromm for his support and guidance during this study, Dr. J. R. Hoffert for many helpful suggestions and photographic work, and Dr. L. F. Holterink for valuable ideas and help with the data analysis. I also thank Esther Brenke for her invaluable technical assis- tance and for typing the rough draft of this thesis. The people of the fish lab are to be thanked for valuable suggestions and moral support, especially Paul Sorenson for many hours of discussion. I am indebted to Dr. H. L. Bergman for the guidance and training he offered me during the initial months of my Masters program. Financial support for this study came from the NIH Training Grant No. HL05873 for which I am more than grateful. iii TABLE OF CONTENTS Page LIST OF TABLES .................................................. v LIST OF FIGURES ................................................. vi INTRODUCTION .................................................... l MATERIALS AND METHODS ........................................... 7 Experimental Animals ...................................... 7 Perfusion Apparatus ....................................... 7 Gill Dissection and Cannulation ........................... ll Perfusion Solutions ....................................... l3 LAS Solutions ............................................. l3 Experiment Protocols ...................................... l4 Data Treatment and Statistics ............................. l5 RESULTS ...................................................... . 20 DISCUSSION .................................................... . 37 CONCLUSIONS ..................................................... 5l RECOMMENDATIONS ................................................. 52 LIST OF REFERENCES .............................................. 53 APPENDIX ........................................................ 55 iv LIST OF TABLES TABLE —I . Parameters and statistics for fitted exponential curves.... . Parameters and statistics for hyperbolic curve fit ......... . Parameters and statistics for exponential curve fit; the effects of vasoactive agents and 10 mg l"1 LAS ............. . The net influx of water calculated from %D(w) values ....... Page 23 24 32 41 LIST OF FIGURES FIGURE Page 1. Schematic diagram of perfusion apparatus .................. 9 2. Results of a representative detergent experiment .......... l8 3. The response of isolated gills exposed to various concen- trations of detergent ..................................... 22 4. Comparison of %D(65) values for gills exposed to various concentrations of detergent ............................... 26 5. The effects of a step input of detergent on perfusion pressure ............... . ...... . ........................... 29 l 6. The effects of a lo mg l' step input of LAS on 3HOH uptake .................................................... 31 7. Comparison of %D(65) for arches treated with vasoactive agents .................................................... 34 8. Hypothetical two compartment model of the isolated per— fused gill ................................................ 44 vi INTRODUCTION The gills of teleost fish are in intimate contact with the animal's external environment. Thus, any toxic substance present in this environment will contact the gills. Synthetic detergents are an example of such a toxicant. Synthetic detergents are a diverse group of compounds, part of a larger family known as surface active agents, or simply, surfactants. All compounds of this type have one common feature; they are amphipathic, that is, they contain both hydrophilic and hydrophobic groups. This polarity endows surfactants with three general properties: 1) The tendency to congregate at surfaces or interfaces. 2) The reduction of surface tension in solution. 3) The formation of micelles when present in solution above a Certain critical concentration. Certain implications stem from this surface activity. First, because surfactants gather at interfaces, they will tend to congregate at biological surfaces (i.e., cell membranes). Also, because of their amphipathic nature detergents have been shown to interact with proteins and lipids, the structural components of cell membranes (Helenius and Simons, 1975). Finally, due to the tendency to form aggregates when present in high concentrations, detergents can solubilize membranes. These effects of detergents are well known to biochemists, as surfac- tants have long been used as protein denaturants and solubilizing agents (Putnam, l948). Three classes of synthetic detergents can be identified as anionic, cationic, or non—ionic, based on the charge of their hydro- philic group. The anionic detergents have been well studied, as they are found in a majority of industrial and household detergents and hence represent the greatest potential pollution problem. The most common type of anionic detergents are the alkyl aryl sulphonates, such as the alkyl benzene sulphonates (ABS). If the alkyl group is un- branched, the detergent is designated linear alkyl benzene sulphonate (LAS). At the present time LAS is the most widely used anionic deter- gent, as the linear alkyl group lends itself more readily to bio- degradation than do the branched chains. It must be noted that these unbranched detergents are not immediately degraded, and thus will pose a threat to aquatic fauna and flora (Abel, l974). Synthetic anionic detergents have been shown by many investiga- tors to be toxic to fish (see reviews by Marchetti, 1965 and Abel, l974). Little is known, however, about the physiological effects of exposure to detergents, as most studies have been directed toward determinations of median lethal doses (L050). The 24 to 96 hr L050 values for LAS range from 0.5 to 6 mg l"1 (Abel, l974). The toxicity of anionic detergents varies depending on several factors. LAS has been shown to increase in toxicity as the length of the alkyl group increases from C8 to C18 with the CM-C16 lengths being the most toxic (Abel, l974). Calamari and Marchetti (1973) reported that water hardness increased the toxicity of LAS and they speculated that the increase in toxicity was due to a complex formed between two detergent molecules and one di-valent cation molecule. Low dissolved oxygen concentrations have been shown to cause an increase in the toxicity of not only detergents, but a wide variety of poisons (Abel, 1974). The increase in toxicity could be due to the combined effect of internal hypoxia, along with the specific toxic action of a pollutant, since gill damage is often seen in fish exposed to various pollutants (Skidmore, 1970). Tovell, Howe and Newsome (1975) demonstrated that radioulabeled sodium lauryl sulphate (SLS), an anionic detergent similar to LAS, is absorbed by goldfish primarily through the gills, and is distributed throughout the body, metabolized, and excreted by the kidney. The highest concentration of labeled metabolite of SLS was found in the gallbladder. However, in other animals, anionic detergents have been shown to be only slightly toxic as internal poisons (Gloxhuber, 1974), thus the absorption of surfactants by fish is thought to be of minor importance when considering their toxicity. The major effect of detergents on fish appears to be related to the damage they cause to the gills. The changes seen consist of a detachment and thickening of the epithelium of the secondary 1ame1lae, and adhesion of adjacent secondary lamellae (Schmid and Mann, 1961; Cairns and Scheier, 1963; Abel, 1974). Skidmore (1970) and Skidmore and Tovell (1972), in reporting the effects of zinc on the gills of rainbow trout, state that many toxic pollutants evoke similar changes and that these responses represent a non-specific inflammatory response. This statement is supported by the observations of Abel and Skidmore (1975) on the changes induced in the gills of rainbow trout (Salm9_ gairdneri Richardson) by the anionic detergent sodium lauryl sulphate (SLS) at a concentration of 100 mg 1-]. SLS was shown to cause lifting of the gill epithelium, invasion of the subepithelial spaces by lympho- cytes and granulocytes, and a sloughing off of a large number of dead epithelial cells. These responses are nearly identical to those re- ported for gills damaged by zinc (Skidmore and Tovell, 1972). The only differences noted were related to death of pillar cells and supportive tissue which is not seen in gills of zinc treated animals. Since the gills of fish are involved in respiration, excretion and osmoregulation, any damage or changes of this organ could affect one or more of these functions. Abel (1974) reported that in most stud- ies of detergent toxicity exposed fish appear to expire due to asphyxia, as all the signs of respiratory distress are present. These symptoms include gulping of air, coughing, and increased ventilatory movements. On the cellular level surfactants have been shown to bind to cell membranes, and at low concentrations (lO'SM) alter the permeability characteristics of these membranes to ions and water (Helenius and Simons, 1975). At higher concentrations detergents have a lytic effect on cells. The gill, however, represents a slightly more complex system as compared to a single cell. There are not only several cell mem— branes between the environment and the milieu interieur but also a variable layer of mucus. Thus any surfactant present in the environment will encounter the mucus barrier before reaching the gill epithelium. What effect this may have on the transfer characteristics of the gill is not known. Still, surface active agents have been shown to increase the uptake of substances into fish, via the gills, presumably by alter— ing the permeability. Levy and Anello (1968) demonstrated that Polysor- bate 80, a non-ionic detergent, significantly increases the absorption rate of secobarbitol by goldfish. This detergent has also been shown to significantly increase the absorption and exsorption of 4 animoanti- pyrine in goldfish (Anello and Levy, 1969). There are relatively few reports concerning gill damage by pollu- tants and its effect on osmotic exchange. In their experiments with trout, Schmid and Mann (1961) speculated that osmoregulatory processes were probably hindered by the action of the anionic detergent sodium dodecyl sulphate but presented no supportive evidence. Cairns and Scheier (1966) treated the pumpkinseed sunfish (Lepomis gibbosus) with 18 mg l"1 ABS for 42 days and studied the effects of this exposure on blood chloride levels. This quantity of detergent was sufficient to cause severe gill damage, and yet no significant change in blood chloride levels could be demonstrated, as compared with control, even when treated fish were challenged by high external chloride concen- trations. These results do not exclude loss of osmotic integrity by the gills, as renal compensation could be sufficient to maintain homeo— stasis. Skidmore (1970) reported that trout exposed to zinc showed an increase in oxygen consumption, increased ventilation volume and fre« quency, and decreased oxygen utilization but no change was seen in blood ion parameters. He concluded that the damage produced by zinc decreased respiratory efficiency without affecting the osmoregulatory function of the gill. The decreased oxygen utilization was attributed to an increase in the water-blood diffusion distance caused by exposure to zinc. Thus, in spite of the above considerations, it remains uncertain as to whether detergents affect gill osmotic exchange. The isolated perfused gill preparation, described by Bergman (1973) and Bergman, Olson and Fromm (1974) appears to be an ideal experimental system to study this problem. By exposing the gills of rainbow trout to various concentrations of the anionic detergent LAS, and measuring the uptake of tritiated water (3HOH), and vascular resistance across the gill, insight may be gained into this phenomenon. Along with the detergent, gills were also treated with epinephrine or acetylcholine to determine if these vasoactive substances modify the effects of LAS on the iso- lated perfused gills of rainbow trout (Salmo gajrdneri). MATERIALS AND METHODS Experimental Animals Experimental animals were 150 to 300 g rainbow trout (Salmg_ gairdneri) obtained from a local hatchery (Midwest Fish Farm Enter- prises, Inc., Harrison, Michigan). They were held at Michigan State University in large fiberglass tanks located in a refrigerated room (12 :_1°C) and exposed to a 16 hr light, 8 hr dark photoperiod. The tanks were supplied with flowing tap water from which excessive chlorine had been removed. Trout were fed every other day on Silver Cup trout pellets (Murray Elevators, Murray, Utah), and starved one week prior to experi— mental use. Perfusion Apparatus A schematic diagram of the apparatus used to perfuse the gills is shown in Figure l and was essentially the same as that described by Bergman (1973). Perfusion fluids were pumped from reservoirs through small diameter Tygon tubing (0.8 mm i.d., 2.4 mm o.d.) by a multi- channel peristaltic pump (Brinkman Instruments, Inc., Nestbury, New York). The perfusate then moved through 10 cm of silicon rubber tubing (to simulate the “Nindkessel” effect of the conus arteriosus Figure 1. Schematic diagram of perfusion apparatus. GP = FCD DC _.| 11 PP PSR SM Grass polygraph Fraction collector driver Drop counter Bath Pressure transducer Peristaltic pump Perfusion solution reservoir Signal marker GP ism ;_:_._:.--——— F. CHI IIHIHHIIIHIHHHI. hfifi3£muuummmmfl CH 4 PM Figure 1 10 and the ventral aorta), a “T“-connector, 20 cm of polyethylene tubing (PE 60), and finally the afferent cannula. This cannula was made from a cut off, blunted 20 gauge needle. The efferent cannula was similarly constructed. Forty cm of PE 60 tubing connected the efferent cannula to one of two identical drop counting assemblies mounted over a fraction col- 1ector. Output from these photoelectric drop counters was fed into a Grass model SD polygraph (Grass Instrument Co., Quincy, Mass.), step- wise amplified, and the signal displayed by the oscillograph such that each drop was represented by a deflection of the recording pen. The fraction collector (ISCO mode 563, Instrumentation Special- ties Co., Lincoln, Neb.) was driven by a timer at the rate of 1 sample per 0.5 min. Each time the fraction collector advanced, a signal was relayed to the Grass polygraph and a mark was made by the signal marker pen. Perfusion pressures were monitored using Statham 523AC pressure transducers (Statham Transducers, Inc., Hato Rey, Puerto Rico) con- nected to the "T“-connectors mentioned above. The signa1 from these transducers was then fed into two separate channels of the polygraph and recorded. The pressures recorded were analogous to ventral aortic pressures in the intact animal. After cannulation the gill arches were suspended in glass stain- ing dishes containing 300 ml of 1% non-nutrient Ringer solution 3 (Appendix) spiked with tritiated water (3H0H) (Biological grade HOH, Sigma Chemical Co., St. Louis, Mo.). The specific activity of the 11 baths was approximately 4.4 x 105 disintegrations per minute per ml, and the osmolarity less than 5 mOsm. The staining dishes were placed on magnetic stirring motors, and Teflon coated stirring bars were placed in the baths to insure adequate mixing and aeration. A 2.5 cm thick piece of styrofoam was placed between the dish and the motor to insulate the baths from the heat of the stirring motor. All experiments were performed at 11 :_1°C. Gill Dissection and Cannulation Gill arches were isolated, cannulated and perfused as described by Bergman (1973), with minor modifications. Fish were netted, then rather than delivering a blow to the cranium to stun the animals, they were wrapped in a damp towel and decapitated just posterior to the opercula. The heads were allowed to fall into a beaker containing 600 ml of heparinized non-nutrient Ringer solution (2 USP units Na heparin/ml; Abbott Laboratories, North Chicago, 111.). Since the hearts continue to beat this solution is pumped through the gills clearing them of blood. The blow to the head was eliminated as it tended to cause cardiac arrest in the group of rainbow trout used for these experiments. The heads were allowed to clear themselves for 15 to 20 min, at which time they were removed from the beaker, and the ventricle of the heart cut so that air would not be pumped into the gill vasculature. The opercula were then removed and the branchial basket isolated. 12 Next, the first and second pair of gill arches were dissected free from the branchial basket, and the first pair of gills discarded. The second pair of arches were then separated by cutting longitudinally through the basibranchial bone. Only the second pair of arches were used in these experiments. After separation, gill arches were placed on watch glasses and covered with non-nutrient Ringer to prevent desiccation. With the per— fusion pump delivering a flow of 0.25 ml min‘1 , the afferent cannula was inserted into the afferent branchial artery, and a single ligature tied around the arch with number 0 nylon suture material to hold the cannula in place. During this procedure perfusion pressure was monitor- ed to assure correct placement of the cannula. Use of a stereoscopic dissecting microscope facilitated insertion of the efferent cannula. After a cannula was placed in the efferent branchial artery, flow in the efferent tube was checked. If flow had been achieved the efferent cannula was tied in place as above. The preparation was then suspended in a bath and the PE 60 tubing from the efferent cannula connected to the appropriate drop counter. The drop counters were placed such that the efferent tube was raised about 20 cm above the level of the gill arches. This represented a pressure of 15 mm Hg on the efferent side of the gill, and mimicked systemic resistance in the intact animal. The flow rate from the pump was then increased to 0.6 ml min”1 and the arches allowed to equilibrate for one hour. When flow was in- creased, perfusion pressure initially increased from 40-50 mm Hg to 13 70-80 mm Hg, and then began to fall. A successful preparation was achieved when the pressures returned to 60-70 mm Hg and remained con- stant for the entire equilibration period. If the perfusion pressures did not attain steady state, the experiment was discarded. Perfusion Solutions Perfusion solutions consisted of Cortland saline (Wolf, 1963) with 3 g Dextran (6.0 x 104 1111) added per 100 ml of solution to mimic the colloid osmotic pressure of plasma proteins. The solution was then filtered through a 0.22 pm Millipore filter, and shaken to assure atmospheric equilibration. The final solution had an osmilarity of 270-280 mOSm, and a pH of 7.5-7.6. When vasoactive substances (10-5M epinephrine (Epi) or 10'8M acetylcholine (Ach); Appendix) were used, they were added after filtra— tion of the perfusion solution. This concentration of Epi was chosen 14 as it has been shown to cause maximal uptake 0f C-urea in isolated perfused gill arches (Bergman ££.214’ 1974). They also demonstrated that Ach reduces the uptake of 14 8 C-urea. Preliminary experiments with Ach resulted in choosing 10- M because at this concentration perfusion pressures remained relatively constant (i_10 mm Hg) for the duration of the experimental periods. LAS Solutions Linear alkulate sulphonate (LAS) reference acids were obtained from the U. S. Environmental Protection Agency, Appendix). l4 Concentrations of LAS used in experiments were 1, 5, 10, 20 and 100 mg 1‘1 (3.14 x 10-6M, 1.57 x 10‘5M, 3.14 x 10‘5M, 6.29 x 10‘5M and 3.14 x 10-4M respectively). Dilute LAS solutions were freshly prepared for each experiment. The appropriate amount of LAS solution was weighed out into a 10 m1 volumetric flask and distilled water added to volume such that one m1 of this solution, when added to a 300 ml bath gave the desired 1, 5, 10 or 20 mg 1“ 100 mg 1“ concentrations. In experiments where LAS was used, the surfactant was weighed into a 1 ml tubercu- line syringe and added directly to the bath as this amount of the detergent occupied approximately one ml. Experiment Protocols Usually two gills were perfused simultaneously with treatments randomly assigned. After allowing the preparation to equilibrate fer 60 min, an initial period lasting 25 min was defined. During this time six samples of perfusate were collected for each preparation. At the end of this interval (defined as time zero, t = 0) a step input of LAS was added to the bath of experimental arches. Fractions of perfusate effluent were then collected at the following intervals (in minutes) after addition of detergent: 1, 2, 3, 4, 5, 10, 15, 25, 35, 45, 55, and 65. At the beginning and end of the 65 min experimental period 0.1 ml aliquots of the bath were taken and counted to assure that the specific activity of the bath remained constant during an experiment. Perfusate was collected directly into vials containing a dioxane based scintillation cocktail (Omni-fluor, New England Nuclear, Boston. 15 Mass.) and was counted within 48 hours on a Mark I liquid scintillation counter (Nuclear-Chicago Corp., Des Plaines, 111.). Counts per minute (cpm) were converted to disintegrations per minute (dpm) of 3H using an internal standard channel ratio technique to determine loss of counting efficiency due to quenching. Background activity was sub- tracted from all experimental activities. The effects of Epi and Ach on the response of gills to 10 mg l-1 LAS was studied as preliminary experiments using this concentration gave fairly consistent results in gills not treated with vasoactive agents. Data Treatment and Statistics The quantity of 3 HOH taken up by the isolated gill depended not only on the effects of treatment but also on such variable factors as the size of the gill arch and the amount of tissue actually perfused. To reduce this inter—gill variability, all sample activities were con- verted to percentages of the mean 3H dpms from the initial period for that arch. The dpms of individual samples from the initial period were averaged and this mean defined as 0% deviation of initial sample activity. The activities of fractions taken during the experimental period were then converted to a percent deviation of this mean initial activity as follows: _ dpm§§p(t) - afifitnit x 100 %D(t) - 355' 16 Where: %D(t) = percent deviation of experimental sample dpms from the mean initial dpm (diagni (t) = activity of experimental sample at time t. %D(t) was t) at time t after addition of LAS to the bath and dpmexp plotted against time for each experiment: a representative plot is given in Figure 2. Provided that perfusate flow into and out of the gill is con- stant for any single arch, %D(t) values are directly proportional to the specific activity of samples (dpm ml']) and hence this transforma- tion corrects for variations in the volume of samples collected between experiments. Percent deviation normalized data was fitted to several mathema- tical functions based on the general shape of the plotted %D(d) data array. Functions tried included a rising exponential with general fOrm y(t) = y(w)(l - e'kt) and a rectangular hyperbola. Using a generalized non-linear least squares technique (Pitha and Jones, 1966; Simon, 1972) and the MSU CDC 6500 computer, functions were fitted to experimental data. This method involves minimizing the sum of squared differences of calculated and experimental points (sum of the squared residuals) by simultaneously varying any parameters contained in the desired function from some initial estimates of these parameters. Unlike the usual linear least squares technique the non— linear case requires several iterations of the above step until the sum of the squared residuals is minimized, i.e., until small changes in the parameters no longer reduce the sum. At this point the iterative non—linear least squares procedure is said to have converged. Figure 2. 17 Results of a representative detergent experiment. At time zero LAS was added to the bath solution. Closed circles represent percent deviation (%D(t)) values obtained from a single experiment. 18 90- 80- . . . 70- +- 60- . =50_' 40- 304. 20- lo. 0% l I I I I 1 O 5 I5 25 35 45 55 65 TIME (min) Figure 2 19 Approximate statistics (variance-covariance matrix, partial and multiple correlation coefficients) were generated for the parameters of the desired function. A discussion of the validity of these linear estimates is given by Hamilton (1964). Because the inter-experiment variances were found to be signifi- cantly heterogeneous, the statistics obtained from the least squares analysis could not be used in hypothesis testing without serious restric- tions. To analyze the steady state response of isolated gills to various concentrations of surfactant, %D(t) values for t = 10, 35, and 65 min were compared within each experiment. To assess treatment effects the percent deviation values of samples obtained during the 65th min (%D(65)) were compared between experiments. Percent deviation data was normalized via log transformation, and single classification analyses of variance (one way ANOVA) performed. Student—Newman-Keuls procedure was used for multiple comparison of means (Sokal and Rohlf, 1969). Log transformed means and means :_one standard deviation were retrans- formed for display purposes. RESULTS Isolated perfused gills of rainbow trout exposed to a step input of the anionic detergent linear alkylate sulphonate responded with an exponential increase in the uptake of 3HOH, and showed a dose related response at detergent concentrations greater than or equal to 5 mg 1-]. Percent deviation data for gills treated with 5, 10, 20, and 100 -1 mg l LAS fit a single rising exponential function of the form: mm = :00») (1 - e‘kt) where 100») represents the percent deviation value at infinite time (t = m), and k is the rate constant (min’I). Fitted curves are pre— sented in Figure 3 along with the mean + or - standard deviation of %D(t) values for t = 5-65 min. Values for the parameters %D(m) and k along with approximate statistics are given in Table 1. Rate con- stants were similar except in the case of 100 mg 1"1 experiments. The Specific activity of perfusate effluent at equilibrium ranged between 5700 and 17,200 dpm ml"1 for experimental arches as compared to approximately 4800 dpm ml"1 for controls. The data mentioned above also converged in the case of a rectangu- lar hyperbole given by the equation: 20 Figure 3. 21 The response of isolated gills exposed to various concen- trations of detergent. Points represent mean %D(t) + or - standard deviation. Although values between 0 and 5 min were used in the curve fitting procedure they were not plotted for clarity. Solid lines are the fitted exponential function of the form: %D(t) = %D(w) (l — e‘kt). (:) = Control (untreated) 1 D = Treated with 1 mg l- LAS at t = 0 A = Treated with 5 mg 1" LAS at t = o O = Treated with 10 mg 1"] LAS at t = o - = Treated with 20 mg 1'] LAS at t = 0 ‘ = Treated with 100 mg l.1 LAS at 1; = o '65 45' . 3.5. 2'5 1'5 -' ES . 5'5 TIME (min) Figure 3 23 Table 1. Parameters and statistics for fitted exponential curves. LAS (mg 1‘1) Parameter _+_ sal ego k “150 k Md 5 7100») 17. 37 _+_1. 07 0.031 0.55 99 k 0. 245 1 0. 053 10 7100») 53. 54 :1. 69 0.019 0.54 66 k 0.189 1. 0. 022 20 200») 96. 89 i 3. 76 0.052 0.54 78 k 0.180 t 0. 026 100 7:00») 255. 75 i 26. 4 0.32 0.84 60 k 0. 052 i 0. 015 gLinear estimate of the standard deviation cCovariance of the parameters dPartial correlation coefficient Number of data points used in the curve fitting procedure where t0.5 is the time, in min, required to attain one half the maximum percent deviation value (%D(w)). Since the parameter covariance and partial correlation coefficients were greater (thus indicating a greater interdependence of the parameters) for the hyperbolic fit than those seen in the exponential case (Table 2), the single exponential function was considered to be the better fit to the data. Data for untreated gills (controls) and gills treated with 1 mg 1-1 LAS did not converge on either function (Figure 3), and were not significantly different than zero %0 throughout the entire range. Analysis of 10(65) values, which should reflect the %D(m) para- meter obtained from the curve fitting procedure, gave results similar to those discussed above. Exposure of gills to 10, 20, and 100 mg 1"1 24 Table 2. Parameters and statistics for hyperbolic curve fit. -l a b c d LAS (mg 1 Parameter + S o R N " 7"”05 7“3:105 5 %D(w) 19.44 :_1.53 1.70 0.75 99 t 3 403 + 1.070 0.5 - 10 %D(w) 60.32 :_2.48 1.40 0.75 66 t 4.504 + 0.756 0.5 - 20 %D(w) 109.24 :_5.60 4.00 0.76 78 t0 5 4.639 10.970 100 %D(w) 304.74 :_38.40 190.0 0.90 60 t0 5 15.827 3*. 5.560 aLinear estimate of the standard deviation cCovariance of the parameters dPartial correlation coefficient Number of data points used in the curve fitting procedure LAS resulted in a significant increase in 3HOH uptake as seen by the increased %D(65) in each case. The percent deviation of samples taken during the 65th experimental min for gills treated with l and 5 mg 1'1 detergent were not significantly different than control, although, as mentioned above, the data obtained from 5 mg 1"1 experiments did fit an exponential function. These results are summarized in Figure 4. When %D(t) values for t = 20, 35, and 65 min were compared within each experimental group no significant differences were found except 1 in the case of gills exposed to 100 mg l- LAS. However, from the results of the exponential curve fit this was to be expected as the 1 gills treated with 100 mg l- LAS had not reached the proposed equi— librium state (%D(w)) during the interval analyzed. 25 Figure 4. Comparison of %D(65) values for gills exposed to various concentrations of detergent. Mean 1 standard deviation and N value are shown for 0, 1, 5, 10, 20, and 100 mg 1" LAS experiments. Means underscored by the same (solid) line are not significantly different (a = 0.05): LAS (mg 1") 0 1 5 10 20 100 -0.66 -9.66 18.00 56.51 93.26 265.85 26 400-4 350—- 300-1 250—4 200—4 ‘36 0(65) 155C)-—+ 100-3 (5) {— (6) L___ 1 1 F O l 5 I 10 1 20 LAS (mg 1") Figure 4 1 100 27 No pressure changes were recorded when concentrations of deter- 1 gent other than 100 mg 1' were used. When the bath LAS concentration was increased to 100 mg 1"1 the perfusion pressure of isolated arches increased from an average control level of 70 mm Hg to 90 mm Hg, and then decreased to approximately 75 mm Hg 10 min after addition of LAS to the bath (Figure 5). Associated with this increase in perfusion pressure was a 5 to 40 percent decrease in the flow out of the gill. Flow returned to control levels as pressures did. The percent devia- tion transformation reflects the specific activity of samples only when flow is constant. Thus activities of samples collected when flow was less than 90% of the initial value were corrected by multiplying the dpm per sample by the ratio of experimental to initial drops per sample. With this correction calculated %0 values were directly pro- portional to the specific activities of the samples. The response of isolated gill arches perfused with solutions containing 10'5M Epi or 10'8M Ach and treated with a 10 mg 1"1 step input of LAS is shown in Figure 6 along with data from similar experi- ments where gills were perfused with saline only. Percent deviation data collected from arches perfused with vaso- active agents converged on the proposed exponential function (Table 3) showing an increase and leveling off of 3 HOH uptake with time. %D(t) values for gills perfused with acetylcholine did not fit the function as well as other data presented thus far as can be seen in Figure 6C, and as indicated by the large covariance of %D(w) and k (Table 4). Rate constants for the Ach and control LAS step input experiments were 28 _-F as oo_ eeep hoop m<4 eo “see? eoom Am C u 3 Se mwpep=mmmeame seem mmcwueoome mezmmmea szpo< .mezmmmea cowmzeema co “somewumu yo page? ampm a mo muompmm one .m mezmwm 29 m oesa_a 8:52. m3 5:5 2...: as: n. O. DVMN.O [Irr-rblrlulrrrrlllll on m NT... 26 m on cow 1 2.. 00° 0: Figure 6. 30 The effects of a 10 mg 1"1 step input of LAS on 3HOH uptake. A) Arches perfused with saline only 8) Arches perfused with solutions containing 10"5 M Epi C) Arches perfused with solutions containing 10"8 M Ach Solid lines represent fitted curves for the function: %D(t) = %D(oo) (1 - e'kt) Means :_standard deviation are indicated. . = Mean %D(t) for arches treated with detergent III = Mean %D(t) for arches not treated with detergent (controls) 31 CONTROL E131 r 45 1 35 TIME (min) T 25 Figure 6 32 Table 3. Parameters and statistics for exponential curve fit; the effects of vasoactive agents and 10 mg l"1 LAS. a 5* c d e Perfusate Parameter :_S 0%D,k R%D,k N Control %D(w) 53.54 :_1.69 0.019 0.52 66 k 0.189 :_0.022 Epi %D(w) 33.30 1.0.84 0.015 0.51 60 k 0.342 :_0.035 Ach %D(w) 33.23 :_3.15 0.16 0.51 62 k 0.276 :_0.102 2Control = Cortland saline + Dextran Epi = Control + 10'5M epinephrine aAch = Control + 10'3M acetylcholine cLinear estimate of standard deviation dCovariance of the parameters Partial correlation coefficient Number of data points used in curve fit similar, while k for Epi pre-treated gills was greater (standard devia- tions did not overlap). The mean %D(65) was similar to %D(w) for arches perfused with 5 10' M epinephrine, while %D(65) appeared to be less than %D(w) for the Ach experiments. Figure 7A shows the comparison of %D(65) values for control, Epi, and Ach preparations exposed to a 10 mg 1"1 step input of LAS. As can be seen there was no significant difference between %D(65) for control and Epi, or Epi and Ach experiments, but Ach %D(65) was significantly less than control data. Values of mean %D(65) for gills perfused but not exposed to detergent (control, Epi-control, Ach-control) are shown in Figure 7B. Figure 7. 33 Comparison of %D(65) for arches treated with vasoactive agents. Means :_standard deviation and N value are shown for each group. A) %D(65)values for control (saline only), Epi, and Ach 1 LAS at t = 0. perfused gill arches exposed to 10 mg 1’ B) %D(65) values for control, Epi-control, and Ach- control (not treated with detergent) experiments. Means underscored by the same line are not significantly different (a = 0.05) Perfusate Treatment Control Epi Ach LAS 56.51 34.46 16.39 None -0.66 7.10 -17.16 34 (m 60- (- _L (5) ”‘ 5 _ 5340- if- F E _L #220- o-___J ______ J ______ 1______] ‘ .1- '10 1 1 1 CONTROL EPI Ach PERFUSATE 20- B (5) £5) 1- _ 1' "5 1 g o_——.__ fin __. “(4.)"- 0 _ 33 -IO- _1_ 1' ~20- _L. ‘10 1 l 1 CONTROL EPI Ach PERFUSATE Figure 7 35 The Ach-control %D(65) was significantly less than the control (not treated with Epi, Ach or LAS) and epinephrine control %D(65) values. There was no significant difference between the percent deviation data collected at t = 10, 35, 65 min within an experiment for untreated preparations or gills exposed to detergent. Activity of samples collected prior to LAS exposure were similar for control and Ach experiments (4800 vs 4700 dpm ml") while sample activity from gills perfused with 10'5M Epi were about 29,900 dpm ml']. Thus, %D(m) values for control, Ach, and Epi experiments represent specific activities of approximately 7400, 6300, and 39,800 dpm ml-1 respectively. No transient perfusion pressure changes were observed when LAS was added to the bath in any experiment with vasoactive agents. Systolic pressure for arches perfused with epinephrine averaged 40 mm Hg, while control and acetylcholine gills showed pressures of 60 and 80 mm Hg respectively. Arches perfused with solutions containing 10‘8M Ach showed a gradual increase in perfusion pressure during the 65 min experimental period. At t = 65 min peak pressures were 5-10 mm Hg greater than initial. No observable changes were seen in the volume flow out of the gill in spite of the increase in perfusion pressure for these arches. Comparison of the equilibrium percent deviation values (%D(w)) obtained from the curve fitting procedure and %D(65), the percent devia- tion values during the 65th experimental min, indicate that there was a consistent similarity between the two for all experiments except those 36 involving acetylcholine perfused gill arches. Further, it has been demonstrated that there was no significant difference between %0 data collected at t = 10, 35, and 65 min within an experimental group, except for 100 mg 1'1 step input experiments. This indicates that the preparations had indeed reached a new steady state after detergent exposure. From the above findings, and the fact that the nonlinear least squares analysis takes into account the entire data array from t = 0 to 65 min, it is felt that the exponential function: %D(t) = 110(5) (1 - e‘kt) provides a better description of the data presented than the single interval analysis. Therefore all further discussion of the detergent data will be in terms of the parameters of the fitted curves. DISCUSSION There are two possible explanations for the observed increase in 3HOH (water) uptake by isolated perfused gills exposed to the anionic detergent LAS: 1) The surfactant altered the permeability character- istics of the gill to water, 2) the detergent caused a change in the internal perfusion pathway thus altering the functional surface area of the gill available for water exchange. Alteration of perfusion pathway can be all but ruled out as a cause for the increased water influx into gill, as such an effect would most likely be accompanied by some change in perfusion pressure. 1 LAS no concomitant varia- 1 Except when gills were exposed to 100 mg 1' tion in perfusion pressure was seen. In the 100 mg 1' 3 LAS experiments pressure increased as did the uptake of HOH. If the change in pres- sure is indicative of the transfer state of the gill, as is the case when dealing with vasoactive substances (Bergman 33.21,, 1974), one 3 would have expected to see a decrease in HOH uptake when perfusion pressures increased. Thus, the most likely explanation for the increased 3 HOH uptake - by the gill is that the surfactant in Some way increased the permeabil- ity characteristics of the gill to water (3HOH). An increased perme- ability could be due to an interaction of the detergent with the mucus coating of the gill, or by a direct action of the detergent on the gill epithelium, or both. 37 38 Gill mucus has been ascribed the function of physically and biologically protecting this delicate respiratory organ. It probably serves as a diffusion barrier and it may significantly influence the rheological properties of the gill. Thus, alteration of the mucus layer could significantly affect the transfer characteristics of the gill to water. LAS could modify the mucus covering per_§e_or it could promote “washing“ of variable amounts of this coating off the gill surface. In so doing the detergent could change the permeability and/or thick- ness of the mucus layer, and hence alter the mobility of water through this barrier. Detergents have been shown to alter the permeability of lipid bilayers and biological membranes of single cells to both ions and water (Helenius and Simon, 1975). In more complex systems such as the gastric mucosa of dogs (Keller et.gl,, 1976) and whole goldfish (Anello and Levy, 1969) detergents have been shown to increase efflux and influx of drugs, respectively, presumably by increasing the permeabil- ity of the membranes in question. Thus, it would be expected that linear alkylate sulphonate could, through a direct action, increase the permeability of the epithelial cell membranes of the gill. Calcium ions, which are known to affect the permeability of pills (Potts and Fleming, 1970) could be removed by interacting with the detergent. Also, by binding to protein and lipid components LAS could alter the structure of (gill) membranes (Helenius and Simon, 1975). Such changes could conceivably increase the size 39 and/or number of the hypothetical membrane "pores". Finally, because detergents bind in high quantities to cell membranes, such agents could, feasibly, increase the hydrophilic character of the external membrane and hence allow water to approach and subsequently move through the membrane more readily. Although critical experiments have not been designed Or performed, it seems likely that the increased flux of 3 HOH into gills after deter- gent exposure is probably related to effects of LAS on both the gill mucus and the epithelial surface of the gills. The increased perfusion pressure seen when gills were exposed to 100 mg l"1 LAS could be due to l) a pharmacologic effect of the sur- factant on the gill vasculature. 2) an osmotic effect on the vessels causing constriction, and/or 3) a response to cellular damage caused by this very high level of surfactant. Although gills perfused with saline containing epinephrine appeared to be less responsive to a 10 mg 1"1 step input of LAS, than arches perfused with saline alone, this difference is probably an arti- fact of the %D normalization. Comparison of the initial activities and those at equilibrium for control and epinephrine treated gill arches show that the difference is actually greater in the epinephrine treated arches than in controls (9900 vs. 2600 dpm ml'], respectively). If Epi increases the functional surface area of the gill as proposed by Bergman et_al, (1974), and the detergent increases the permeability of the gill, it would be expected that for a given amount of change in 3 the permeability of the gill, the increase in HOH uptake seen would be 40 greater for Epi perfused arches than arches perfused with saline only. The activity of samples collected from Ach control experiments decreased with time during the experimental period. The decrease in %D(t) observed after 25 min exposure to 10 mg 1"1 LAS is most likely due to the additive effect of the decrease seen in control preparations. Also, although flow out of the gill remained constant, the increased perfusion pressures seen with time could have resulted from a change in 3HOH into the the perfusion geometry which would affect the movement of gill. The reported 96 hr median lethal dose (L050) for linear alkylate sulphonate ranges between 0.5 and 6 mg l"1 (Abel, 1974). Thus, the concentrations of detergent used in isolated gill experiments were all within or greater than this range. An increased uptake of water by fish, similar to that seen in isolated gill arches exposed to LAS may not be acutely lethal. However, this increased permeability would impose a considerable burden on homeostatic mechanisms such as the kidney, which could prove to be lethal. As shown in Table 4, the influx of water (54.4 vs. 6.6 u1 min']) was 8-fold greater when arches were exposed to 10 mg 1'1 LAS and perfused with solutions containing 10'5M Epi than for untreated control preparations. It should be noted that gills in these experiments were treated with detergent for only 65 min, longer exposures may cause more drastic changes which could prove to be fatal jn_ijg, Also, this study has not indicated if or how LAS might affect the transfer of other substances such as oxygen across the gill. Because the movement of water into the gill is osmotic and not diffusive in 41 Table 4. The net influx of water calculated from %D(w) values. ___ No vasoactive agents With vasoactive agents Treatment Net flux Treatment Net flux (u1 min' ) (ul m1n' ) Control 6.6 Epi-control 40.8 5 mg 1"1 LAS 7.7 Epi-10 mg 1‘1 LAS 54.4 10 mg 1“1 LAS 10.1 Ach-control 6.6 20 mg 1“ LAS 12.9 Ach-10 mg 1‘1 LAS 8.6 100 mg 1"1 LAS 23.4 nature, the detergent could alter the diffusion of oxygen differently than it affects water movement. Subtle changes in oxygen transfer through the gill by the surfactant could result in the toxicity seen in more chronic studies. This hypothesis is supported by indirect evidence that fish appear to be in respiratory distress when placed in solutions containing detergent. Although the increased water movement into an animal caused by detergent might not be acutely lethal, if other toxicants were present in the environment, along with the surfactant, an increase in their uptake might prove to be so. Under the conditions that rainbow trout gills were perfused, there should be a net flux of water into the gill due to the difference in bath and perfusate osmolarity (AOsm = 270 mOsm). The rate at which water moves osmotically into the gill should depend on 1) the osmotic permeability of the organ, including mucus, 2) the distance over which permeation must occur, and 3) the physiological surface area of the 42 gill, i.e., the portion of the gill perfused and ventilated. A deter- gent such as LAS could alter the influx of water by modifying any of these factors. As mentioned above, the most likely candidates are the osmotic permeability and functional thickness of the gill. How could 3HOH a change in these parameters result in the exponential increase in uptake seen experimentally? Let us consider a simple two compartment model of the isolated perfused gill shown in Figure 8 with the following properties: 1) Flow into and out of the gill compartment (Vin, V t respec- ou tively) does not vary with time. 2) The specific activity of the perfusate entering the gill is zero. 3) Any movement of water from the gill to the bath compartment (Vback) due to filtration, leakage, and/or diffusion is constant. 4) The membrane separating the two compartments is permeable only to water such that the driving force (AOSm) for osmotic flow into the gill (V ) remains constant. osm 5) The effective volume of the gill (V9) is time invariant. 6) There is instantaneous mixing in the bath compartment. 7) The 3HOH activity of the bath solution does not vary with time, i.e., the bath serves as an infinite source of 3HOH. Given these conditions the rate of change of 3HOH activity of the solution leaving the gill compartment (Cg(t)) can be described by the following differential equation: 43 com: m—onexm “cm:_m$m mammawema ecu mo xum>wpom Owwwumqm eewpspom seen oee ea xu_>wpoe oeewooam _me one to oe=_e> mmmxemp eo\ucm cowumepFPe on mac cums ou acmsuemaeoo —F_m Eoew emuez mo xapm ppwm one one? gene: mo xzpe proswo pamEpemanO Ppwm esp mo pzo zo_$ mpemzwema pewsuequou Ppwm on» oucw zope mummaeemg ea_pwe_eoa xoea Emo Apv -> use o> o> emmemema .zo—mn cm>wm mew .prm ummaeema umumpomw we» mo _meoe acmEuemanu oz» Feuwpmcuoa»: .m mezmwm 44 300 m mesmwu 45 d C (t) Vosm(t) Cb V + V g = V5 _ out vback cg(t) (l) 9 9 where the symbols are as those used in Figure 8. Now, assume that when LAS is added to the bath solution it inter- acts with the membrane separating the two compartments and, in some way, rapidly increases Qosm to some new steady state, V* os which depends on the dose of detergent. This change in Vosm can be m’ the value of expressed as a step function given 0y: .* Vosm u(t) (2) '* where u(t) represents the unit step function (Riggs, 1970) and Vosm is the height of the step. Substituting equation (2) into (1) for V (t) and solving the osm differential equation for Cg(t), the following relationship is found: -* v c c (t) = c (o) e‘kt + .—°§i“—b—.— (1 - 6"“) (3) g g + v out back where Cg(0) is the specific activity of the perfusate effluent before treatment with detergent and k is equal to: 1'1 1 out + back V 9 At infinite time (t = 00), equation (3) reduces to: Cg(m) = . vosm Sb (4) vout + vback Thus equation (3) can be written: Cg(t) = 09(0) e‘kt + cg(e) (1 — e‘kt) (5) Now, defining: such that: egos) = (%+ 1) cg(01 (6) Substituting equation (6) into (5) and simplifying gives: %D(t) = %D(oo) (1 - e‘kt) which is identical to the function fitted to experimental data. As the rate constant, k, of this function is independent of gosm and hence treatment effects, K should be a constant for all LAS concen- trations used, provided that V out’ vback’ and V9 are similar. Experi- mentally this has been demonstrated for gill arches exposed to 5, 10, and 20 mg 1"1 step inputs of LAS. The decreased k value seen in 100 mg 1"1 experiments would also be expected as flow out of the gill (gout) decreased as compared to other experimental groups which remained at the control level. However, since the initial conditions of the model were not satisfied for 100 mg 1'1 LAS experiments (Vout not constant) no real conclusions can be drawn. Epinephrine is supposed to alter the vascular geometry of the gill in such a way that the transfer of substances across the gill is increased (Bergman et_al,, 1974; Steen and Kruysse, 1964). Hence, the 47 1 difference in rate constants seen in 10 mg 1' LAS step input experi- ments for gill arches perfused with solutions containing Epi and those 1 respectively) is probably perfused with saline alone (:0.3 and 0.2 min- a result of changes in the gill, and hence initial conditions, induced by epinephrine. To further test the appropriateness of this hypothetical model, preliminary experiments were performed in which a step input of 3HOH was added to baths of isolated gill arches. At time zero the activity 5 1 and the of bath solutions was raised from 0 to 4.4 x 10 dpm ml- appearance of labeled water in the perfusate effluent monitored. From the differential equation describing the proposed system (equation (1)) it can be shown that the application of such a step input, with all other factors constant, can be described by: v c c (t) = 49-53—13.— (1 — e'kt) (7) or in terms of percent deviation from initial (equations (4) and (6) above): %D(t) = 40(5) (1 - e‘kt) where k is identical to that given in equation (3) above and is equal to: vback + vout V 9 If the response of the gill can be simulated by the proposed model, rate constants from detergent experiments (5, 10, and 20 mg 1-] 48 LAS) should be similar to those obtained from 3 HOH step input experi- ments. The rate constants, k, for 3HOH step input experiments thus far performed was 0.874 1 0.217 min'1 which is quite different than k obtained for detergent experiments (0.205 :_0.063 min-1). If the proposed model system is accurate, how can this difference be accounted for? First, it was assumed that there was only one time- varying factor other than Cg(t); gosm in the detergent case and Cb for 3 the HOH step input. It may be that more than one treatment effect is responsible for the increased 3 HOH uptake observed, especially in the case of detergent experiments. Another possible explanation for the observed difference in rate constants is that the parameters of the model equations are not independent, i.e., there is a nonlinearity in the system such that perturbation of one factor causes another to change concomitantly. Thus, the equations describing the response of Cg(t) to two different inputs may be of similar form, but cannot be compared due to the presence of a non-linear term. Further, it could be that there is some fundamental inaccuracy in the model. This could be due to the presence of undefined factors in the actual system not included in the model, or the isolated gill may be more analogous to multi-compartment (greater than 2) model. If in- deed a more complex model could describe the preparation better than the simple two compartment system described above, one would expect that the data collected could be described by a more complex function, probably of multiple exponential form. 49 To test this hypothesis %D(t) data was fitted to several multiple exponential functions. In each case either the added parameters were essentially zero (less than 10"3 with standard deviations of at least 10-3) such that the resulting equation reduced to the single rising exponential function, or the data would not converge on the function. Thus, adoption of a more complex system than the two compartment model is all but ruled out. The presence of as yet undefined factors operating within the framework of the proposed model cannot be eliminated from consideration. Finally, differences in the value of rate constants for the two types of step input experiments could result from the assumption that the detergent interacts with the gill arch and instantaneously increases the osmotic influx of water. If this effect were to take place over a period of several min (5-10 min), such that the change of gosm could not be approximated by a step function the above derivation would be invalid. Under such conditions the change in 3 HOH uptake with time would be proportional to the change in Qosm which in turn would be a function of the duration of exposure of the gill to detergent. Such a relationship could be described by a hyperbolic function. However, as discussed above, a rectangular hyperbola did not fit the data as well as the exponential function. Also, the reaction of detergents in solution is very rapid, the rate constant for micelle formation, for example, being approximately 6 x 1010 min"1 (Helenius and Simon, 1975). Thus, this consideration as a possible explanation for the observed differences in k seems improbable. 50 No conclusions are warranted concerning the model of isolated perfused rainbow trout gills as it has not been adequately tested. However, as a first approximation, this simple model does provide a foundation for further study. CONCLUSIONS Isolated perfused gill arches of rainbow trout (Salmo gairdneri) were exposed to step inputs of the anionic detergent linear alkylate sulphonate and the uptake of tritiated water monitored. 1) Step inputs of 5, 10, 20, and 100 mg 1" 3 LAS increased the uptake of HOH from initial levels, the response being dose related. 2) The 3HOH uptake with time after application of a step input of detergent can be described, in terms of percent deviation from initial levels, by a single rising exponential function of the form: kt %D(t) = %D(m) (1 - e‘ ) where %D(w) is the equilibrium percent deviation value at infinite time, and k is the rate constant. 3) The response of gills to surfactant can be explained ih terms of an increase in the permeability characteristics of the gill. This increase could be related to alterations in gill mucus and/or epithalial cell membranes. 4) Epinephrine increased the effects of a 10 mg 1.1 step input of LAS on 3 HOH uptake of isolated gills, probably by altering the vascular geometry of the gill. 4) These results suggest an hypothetical compartmental model of isolated perfused rainbow trout gills. 51 RECOMMENDATIONS To test the hypothetical model of the gill the following experi- ments need to be performed: 1) Quantify aback, the flux of water from gill to bath (with and without LAS present). This could be accomplished by perfusing isolated 3 gill arches with solutions spiked with HOH, and placing the prepara- tion in a bath solution of the same osmolarity as the perfusate. By measuring the difference in the activity of inflow and outflow, Vback could be determined. 2) Vary the flow rate into the gill and hence the flow out of the 3 gill (Vout) and determine how this affects the uptake of HOH. 3) Increase the specific activity of the bath in a stepwise fashion with gills perfused at different flow rates. Determine how this affects the rate constant (k) for 3 3 HOH uptake. 4) Apply a step input of 3 HOH to the bath of an isolated gill and follow the rise of HOH activity of the perfusate effluent (Cg(t)). Add detergent to the bath and allow Cg(t) to reach steady state. Increase bath 3 HOH activity and compare the response with that obtained when no detergent was present. By comparing the results of the above experiments it could be determined if the proposed model is accurate, and if there_are any non- linear factors present in the system. 52 LIST OF REFERENCES LIST OF REFERENCES Abel, P. D. 1974. Toxicity of Synthetic Detergents to Fish and Aquatic Invertebrates. J. Fish Biol., 6:279-298. Abel, P. D. and J. F. Skidmore. 1975. Toxic Effects of an Anionic Detergent on the Gills of Rainbow Trout. Water Res., 9:759-765. Anello, J. A. and G. Levy. 1969. Effect of Complex Formation on Drug Absorption. X: Effect of Polysorbate 80 on the Permeability of Biologic Membranes. J. Pharm. Sci., 58:721-724. Bergman, H. L. 1973. Regulation of Blood Flow through the Isolated- Perfused Gills of Rainbow Trout: Effects of Vasoactive Agents on Functional Surface Area. Ph.D. Thesis, Michigan State University, East Lansing, Michigan. Bergman, H. L., K. R. Olson and P. 0. Fromm. 1974. The Effects of Vasoactive Agents on the Functional Surface Area of Isolated Perfused Gills of Rainbow Trout. J. Comp. Physiol., 94:267-286. Calamari, D. and R. Marchetti. 1973. The Toxicity of Mixtures of Metals and Surfactants to Rainbow Trout (Salmo gairdnerii, Richardson). Water Res., 7:1453-1464. Cairns, J. and A. Scheier. 1963. The Acute and Chronic Effects of Standard Sodium Alkyl Benzene Sulphonate on the Pumpkinseed Sunfish, Lepomis Gibbosus (Linn.) and the Bluegill Sunfish, Lepomis macrodhirus (Rafi). Proc. 17th Ind. Waste Conf., Purdue unit. Engng. Ext. Ser., 112:14-28. Cairns, J. and A. Scheier. 1966. The Influences of Exposure to ABS Detergent on the Blood Chloride Regulation of the Pumpkinseed. Prog. Fish-Cult., 28:128-132. Gloxhuber, C. H. 1974. Toxicological Properties of Surfactants. Arch. Toxicol., 32:245-270. Hamilton, W. C. 1964. Statistics in Physical Science: Estimation, Hypothesis Testing and’Least Sguares. Ronalngress Co., New York. 324pp. Helenius, A. and K. Simons. 1975. Solubilization of Membranes by Detergents. Biochim. Biophys. Acta, 415:29-79. 53 54 Keller, R. W., M. M. Berenson, F. Moody and J. W. Preston. 1976. The Effect of Detergent Treatment of the Gastric Mucosa on Drug Transport. Proc. Soc. Exp. Biol. and Med., 151:730-735. Levy, G. and J. A. Anello. 1968. Effects of Complex Formation on Drug Absorption. V: Studies on the Mechanism of Secobarbitol Absorption - Enhancing Effect of Polysorbate 80 in Goldfish. J. Pharm. Sci., 57:101-104. Marchetti, R. 1965. Critical Review of the Effects of Synthetic Detergents on Aquatic Life. Stud. Rev. gen. Fish. Coun. Medit., 26:1-31. Pitha, J. and R. N. Jones. 1966. A Comparison of Optimization Methods for Fitting Curves to Infrared Band Envelopes. Can. J. Chem., 44:3031-3050. Potts, W. T. W. and W. R. Fleming. 1970. The Effects of Prolactin and Divalent Ions on the Permeability to Water of Fundulus kansae. J. Exp. Biol., 53:317-327. Putnam, F. W. 1948. The Interactions of Proteins and Synthetic Deter- gents. Adv. Proteins Chem., 4:79-122. Riggs, D. S. 1970. Control Theory and Physiolo ical Feedback Mechan- isms. The Williams and Wilkins Co., Ba timore, 599pp. Schmid, 0. J. and H. Mann. 1961. Action of a Detergent (Dodecyl Benzene Sulphonate) on the Gills of Trout. Nature, Lond., 192:675. Simon, W. 1972. Mathematical Techniqges for Physiology and Medicine. Academic Press, Inc.,—NewFYork, 267pp. Skidmore, J. F. 1970. Respiration and Osmoregulation in Rainbow Trout with Gills Damaged by Zinc Sulphate. J. Exp. Biol., 52:481-494. Skidmore, J. F. and P. W. A. Tovell. 1972. Toxic Effects of Zinc Sulphate on the Gills of Rainbow Trout. Water Res., 6:217-230. Sokal, R. R. and F. J. Rohlf. 1969. Biometry. W. H. Freeman and Co., San Francisco. 776pp. Tovell, P. W. A., D. Howe and C. S. Newsome. 1975. Absorption, Metab- olism, and Excretion by Goldfish of the Anionic Detergent Sodium Lauryl Sulphate. Toxicology, 4:17-29. Wolf, K. 1963. Physiological Salines for Freshwater Teleosts. Prog. Fish-Cult., 25:135-140. APPENDIX APPENDIX Non-nutrient Ringer solution NaCl 8.5 9 KCl 0.1 g CaCl2 0.5 g HOH 1000.0 9 Dilute l to 100 for 1% solution. LAS reference samples Lot #07181 8-74 Avg. M.W. 318 Active LAS 5.5% Shake flask test 90% Act. sludge test 95+% Obtained from: National Environmental Research Center U. S. E. P. A. Cincinnati, Ohio 45268 Source of vasoactive agents Epinephrine HCl, Sigma Chemical Co., St. Louis, Mo. Acetylcholine C1, Sigma Chemical Co., St. Louis, Mo. 55 MICHIGAN STATE UNIV. LIBRARIES 31293100251465