INT .
J.
RADIAT . BIOL .,
1992,
VOL .
62,
NO .
2, 239-248
The effect of chronic radiation on the humoral immune response of rainbow trout (Onchorhynchus mykiss Walbaum) J. F. KNOWLESt
Int J Radiat Biol Downloaded from informahealthcare.com by University of Melbourne on 11/23/14 For personal use only.
(Received 5 August 1991 ; revision received 16 January 1992; accepted 27 February 1992)
Abstract. Two separate experiments have examined the effect of exposing rainbow trout to chronic y-radiation, commencing immediately after fertilization . In experiment 1 the period of exposure extended for 20 days with groups receiving mean dose rates of 1 . 87, 3 . 73 and 9 .03 mGy h -1, and mean total accumulated doses of 0 . 83, 1 . 66 and 4 . 01 Gy respectively . At 5 months of age fish were tested for specific antibody response to dinitrophenol coupled to keyhole limpet haemocyanin (DNP-KLH) and there was no significant difference in titre between irradiated groups and unirradiated controls . In experiment 2 the exposure period was extended to 246 days from fertilization. Mean dose-rates to the three groups used were the same as in the first experiment until hatching at 21 days and then lower with rates of 0 . 99, 1-9, and 4 . 66 mGy h -1 to the freeswimming fish . The mean total accumulated doses over the whole irradiation period were 5 . 43, 10 . 53 and 25 .43 Gy respectively . The antibody response to DNP-KLH was significantly lower in trout receiving the highest dose-rate when compared with those of unirradiated controls or the lowest dose-rate group . The significance of these results is discussed in relation to radiation levels in areas of radioactive waste disposal, and results from a similar study published previously .
1 . Introduction The possibility that increased exposure to radioactive and chemical environmental contaminants might cause a rise in the incidence of fish diseases is of considerable interest, not only to the fishing and aquaculture industries, but also to the public in general . The immune system plays a major role in disease resistance and the effects of toxic agents on the fish immune system have been the subject of considerable research effort (Zeeman and Brindley 1981 ; Anderson et al. 1984) . In their major review of the subject Zeeman and Brindley showed that, in contrast to the effects of chemical toxicants, radiation effects have been little investigated . Where such investigations have taken place large, single doses of radiation have been given rather than the chronic low dose-rate radiation exposures, which are pertinent to radioactive contamination in the environtMinistry of Agriculture, Fisheries and Food, Directorate of Fisheries Research, Fisheries Laboratory, Lowestoft, Suffolk NR33 OHT, UK .
ment. The important exceptions to this are the studies of Strand and his co-workers (Strand et al. 1973, 1977, 1982) . They exposed developing rainbow trout to low-level fl-radiation by keeping them in tritiated water from fertilization through 20 days of embryogenesis . At 5 months post-hatching the fish were challenged with killed bacteria and there was a significant reduction in antibody titre even in those which had received total doses of only 40 mGy given at a rate of 2 mGy per day . This remarkable result indicated a very high radiosensitivity for the developing immune system in these fishes . In view of the importance of this finding it is surprising that until now no repeat experiments or further investigations have been undertaken . We have carried out similar studies exposing developing trout to chronic y-radiation at mean dose-rates of between 1 and 9 mGy h -1 and this paper reports two experiments : in one the exposure was from fertilization through 20 days of embryogenesis, the same period as in the experiments of Strand et al . (1977) ; in the other irradiation continued from fertilization for 35 weeks . These experiments were carried out separately and at different times using some different methods . For these reasons they are considered separately and comparisons are only made between experimental groups in the same experiment . However, both are addressing the same general problem and many of the methods used were similar .
2. Materials and methods Two separate experiments were carried out but many features of the two were identical. Eggs and sperm were obtained from rainbow trout, Onchorhynchus mykiss, at the MAFF fish farm at Sacrewell, Northants and transferred to Lowestoft in cooled flasks . For each experiment eggs stripped from three females were fertilized with sperm from a similar number of males . The fertilized eggs were divided into four approximately equal groups, each of which
240
J. F. Knowles
kept in well-aerated, dechlorinated running water . Eggs in one tank were controls and those in the other three received radiation . Trays and tanks were identical and mains water passed through the same dechlorinator (Elgamat AC 1, Elga, High Wycombe, UK) to all tanks, irradiated and unirradiated . Water temperatures were similar in all tanks (see below) .
Int J Radiat Biol Downloaded from informahealthcare.com by University of Melbourne on 11/23/14 For personal use only.
2 .1 .
Irradiation
The radiation facility at Lowestoft consists of a shielded tank room within which 137 Cs y-ray sources are arranged in such a way that three sets of tanks may be exposed, each to a different low dose-rate (Rackham et al . 1991) . The sources are exposed by remote control and the precise period of exposure recorded automatically. This was usually about 22 h per day, allowing 2 h for fish feeding, tank cleaning and other maintenance . Dosimetry was carried out using LiF powder, thermoluminescent dosimeters (TLDs) which were read on a Harshaw 3000A TLD reader. Dose-rates to developing eggs in the egg trays were measured by placing TLDs at nine locations in the tray for each dose group . Dose-rates to experimental tanks were obtained by placing 27 TLDs (a 3 x 3 x 3 matrix) in each (Rackham et al . 1991) . Mean dose-rates for egg trays, whole tanks and bottom layer of tanks are given in Table 1 . The attenuation of radiation by water in the tanks is indicated by a comparison of dose-rates in the nearsurface egg trays and the bottom layers of the tanks, and also by the relatively large standard deviation of the mean whole tank dose-rate . 2.2 . Blood sampling and antigen injection After hatching, the fish were fed pelleted food (BP Nutrition, Northwich, UK) of the appropriate size and amount . For antigen injection and blood sampling fish were anaesthetized using a solution of phenoxyethanol in water (1 to 2500) . In both experiments the fish which provided pre-injection
Table 1 . Mean dose-rates to the experimental fish tanks Mean dose rate±standard deviation (mGyh -1 ) to Dose-rate group Low Mid High
Egg trays
Bottom layer
Whole tank
1 . 87±0 . 17 3 . 73±0 . 36 9. 03±0 . 97
0 . 44±0. 08 0 . 89±0. 10 2 . 07±0. 33
0. 99±0 . 52 1 . 90±0 . 99 4. 60±2 . 38
blood samples were then killed . Those in experiment 1 were 10-12 g at this time . After subsequent blood samples fish were allowed to survive . The fish in experiment 1 were about 20 g by the first postantigen injection sample and 0 . 1 ml of blood was taken . Rainbow trout of this size had 0 . 15-0 . 2 ml of blood sampled without affecting survival or activity (Lied et al . 1975) . As the fish grew older and increased in size, volumes of approximately 0 . 2 ml were taken . Blood samples were taken from the caudal vein using a 26-gauge needle . After each sampling the blood was placed in 1 . 5 ml disposable plastic centrifuge tubes (Boehringer, Lewes, UK) and left to clot for 1-2 h at room temperature, then overnight at 4°C . It was then centrifuged at 3000g for 15 min and the serum removed and stored at -20°C . The antigen used in both experiments was dinitrophenol coupled to keyhole limpet haemocyanin (DNP-KLH ; 494 DNP groups per protein molecule) obtained from Calbiochem-Behring (La Jolla, USA) . It was made up at a concentration of 5 mg/ml in Freund's complete adjuvant (FCA ; Sigma, Poole, UK) and injected intraperitoneally at approximately 1 mg per 100 g body weight . FCA-only controls received injections of similar volumes of FCA alone . 2.3 . Experiment 1 Developing trout eggs were exposed to radiation for 20 days (443-8h actual exposure time) from fertilization to the `eyed' stage . At this time the trays containing them were transferred to unirradiated tanks . The mean dose-rates to eggs were 1 . 87, 3 . 73 and 9 . 03 mGy h -1 and total doses received were 0 . 83, 1 . 66 and 4 . 01 Gy respectively . There were, in addition, two unirradiated groups of eggs . After hatching the fry remained in the same tanks and were reared normally until 5 months of age . At this time they were all anaesthetized and each group marked by tattooing with Alcian blue using a 'Panject' needleless injector (Wright Dental Group, Dundee, UK) or by removal of the adipose fin . Pre antigen-injection blood samples were taken from 10 fish in each group at this time . One week later the fish were all anaesthetized again and injected intraperitoneally with antigen in FCA, except for one of the unirradiated groups which received FCA only . The groups which had received antigen were named by their radiation dose rate : 0, 1 . 87, 3 . 75 or 9 . 03 mGy h -1 groups . The other group was known as FCA-only. After injection fish from all groups were kept together in a single large tank (1500 litres) . Blood samples were subsequently taken
Int J Radiat Biol Downloaded from informahealthcare.com by University of Melbourne on 11/23/14 For personal use only.
Radiation effects on fish immune response
at intervals as shown in Figure 1 . Usually 10 fish were sampled from each group except for 3. 73 mGy h -1 and FCA-only where the samples were four to six fish at each time up to 12 weeks and only one and two fish respectively at 19 weeks . Throughout the experiment all tanks received water from the same mains supply. In the tanks holding fish prior to antigen injection the water was dechlorinated, but the supply to the single large tank used later was not . However, chlorine levels were checked regularly and remained below 0 .06 mg 1 -1 . Water temperatures were 8 .5-10°C during pre-hatching and then rose steadily 18°C at injection before falling to 12°C by 9 weeks post-injection. Sera from blood samples were assayed for antibody using a latex particle agglutination assay modified from that described by Hudson and Hay (1989) : (1) A suspension of latex particles 400 µl; (Difco, East Molesey, UK) were washed in 15 ml of 0.05m glycine buffered saline at pH 8 . 7 (GBS), centrifuged at 12 500g for 15 min and the particles resuspended in a further 5 ml of 0.05 M GBS . (2) Particles were coated with antigen by adding 40 tl of DNP-KLH solution in 0 .05 M GBS (10 mg/ml) to the suspension, mixing and leaving for 3-4h at room temperature-this step was omitted for uncoated, latex-only suspensions . (3) The particles were washed twice as before, finally suspended in 5 ml of 0 . 25 M GBS containing 0 .5% bovine serum albumin (BSA) and stored at 4°C until use . (4) The trout sera were thawed and serial doubling dilutions made with 0 . 25 M GBS containing 0 .5% BSA . The dilutions were made in microtitre plates (LIP, Shipley, England) with 25 yl of diluted serum per well, and a duplicate series was made for each test serum . (5) DNP-KLH-treated latex suspension (25 µl) was added to each well for one series, while latex-only suspension was added for the other . (6) The microtitre plates were covered and left for 24-48 h at room temperature before the titres were read and recorded .
2.4. Experiment 2 Developing trout were irradiated continuously from fertilization for a total of 246 days . After hatching at a mean of 21 days the fry passed through holes in the egg trays and fell to the tank floors where they remained for a mean of 32 days before becom-
24 1
ing free-swimming throughout the tank for the remaining 193 days of exposure . The dose-rates thus varied during development and are given for the various positions in Table 1 . For simplicity the groups were named by the mean dose-rates received for the majority of their exposure period, i.e. 0, 0 .99, 1 .90 and 4 .60 mGy h -1 groups . It should, however, be remembered that mean dose-rates other than these were received during the first 53 days postfertilization(Table 1) . The integrated total accumulated doses received by the 0.99, 1 .9 and 4.60 mGy h -1 groups were 5 . 43, 10 . 53 and 25 .43 Gy respectively . At the termination of irradiation blood samples were taken from 10 fish in each irradiated and unirradiated group . The trout were then transferred to eight new tanks (450 litres each), two replicate tanks for each irradiated group and the 0 mGy h -1 group . Fifty fish were placed in each tank, giving a total of a 100 for each group . Fifty unirradiated fish which formed the FCA-only group were left in their original tank . After a week of acclimatization all fish were injected with antigen in FCA, except for the FCA-only which received adjuvant alone . Blood samples were then taken at intervals until 21 weeks post-injection . To ensure against oversampling of the same fish, samples of 10 antigeninjected fish per group were taken from one tank at each sampling and the two replicate tanks were sampled alternately. Five fish per sample were taken from the FCA-only group, which was maintained in a single tank . The water supply to all tanks was from the same mains but that to the FCA-only group remained dechlorinated throughout while the supply to the eight new tanks was not. However, chlorine levels were checked regularly, as in experiment 1, and remained below 0 .06mg1 -1 . Water temperature was 12-14°C during pre-hatching, fell to 9-11°C during the following 4 months, rose to 16°C at antigen injection (8 months) and this was maintained for the next 3 months before falling to 11-13°C at the termination of the experiment . Sera from blood samples were assayed for specific antibody using an enzyme-linked immunosorbent assay (ELISA) which was a modification of that used by P. Dixon and H . de Groot (personal communication) . ELISA methods are more elegant and sensitive than agglutination methods, and it was possible to develop one as an improvement for this second experiment mainly thanks to the gift of monoclonal mouse anti-trout y-globulin from Dr C . Secombes, University of Aberdeen, who had obtained the clonal line, 1-14, from Professor G . Warr, University of S . Carolina . The protocol used was arrived at by extensive chequer board and other trials, and may be summarized as follows :
Int J Radiat Biol Downloaded from informahealthcare.com by University of Melbourne on 11/23/14 For personal use only.
242
J . F. Knowles
(1) Microtitre plates (Costar, Cambridge, USA) coated by adding 200 µl per well of 0 . 5 µg/ml DNP-KLH in carbonate buffer (pH 9 . 6) and incubating overnight at 20 ° C . (2) Plates washed (all washes are three times) in phosphate buffered saline (PBS, pH 7 . 4) . (3) Blocking carried out by adding 300µl per well of 1 % bovine serum albumin (BSA) in carbonate buffer and incubating for 1 h at 37°C . (4) Wash in PBS with 0. 05% Tween-20 (PBST) . (5) Plates dried overnight, sealed and stored at -20°C until required . (6) Test sera were diluted 1 :1000 in PBST, 200 µl added per coated well and incubated for 90 min at 20-22°C . (7) Wash in PBST . (8) 200 p1 of ascites fluid containing monoclonal mouse anti-trout y-globulin diluted 1 :2000 in PBST with 1 % BSA were added per well and incubated for 90 min at 4°C . Wash in PBST. 200 µl of goat anti-mouse IgG conjugated with alkaline phosphatase (Sigma, St Louis, USA) diluted 1 :1000 in PBST with 1 % BSA were added per well and incubated for 90 min at 4 ° C . (11) Wash in PBST . (12) 200µl of a substrate solution containing 1 mg/ml p-nitrophenyl phosphate in diethanolamine buffer pH 9 . 8 were added per well and incubated for approximately 30 min (see later) at room temperature . (13) The reaction was stopped by adding of 50 yl of 3 M NaOH . (14) Optical density (ODs) were read at 405 nm .
All samples taken at any one time, a total of 45, were assayed together and two identical replicate microtitre plates were prepared . Outer wells were not used, the remaining 60 accommodating the 45 samples and various standards, all of which were placed in a set pattern . Reference and external positive standard sera were obtained from trout which had been injected with DNP-KLH in FCA 8 weeks (reference) or 12 weeks (external) previously . Sera from completely untreated trout of similar ages to experimental fish were used as a negative standard . In each case pooled sera were frozen in 20 p1 aliquots until use, when they were diluted as for test samples . Each microtitre plate had two replicate wells of each standard . In order to standardize substrate incubation time the ODs in the reference positive wells were
always allowed to develop to approximately 1 .2 before stopping the reaction . Slight under- or overshooting was compensated for by obtaining the mean of the two reference positive values and dividing this into 1 . 2 to obtain a correction factor, which was used for all ODs on the microtitre plate . The external standard was used as a quality control measure to monitor for errors (e .g. dilution) arising from the reference positive standard . The ELISA was developed as an improvement for this second experiment, as already noted . In order to check the correlation between results obtained by ELISA and latex agglutination, samples of sera taken at 8 and 16 weeks were assayed by both methods .
2 .5 . Statistical analysis Results from both experiments have been analysed in a similar way. From a preliminary analysis of the data it was clear that for both experiments the response within each treatment group at each time was not normally distributed, and that no simple transformation of the data would result in normality . The data for FCA-only groups, which received no DNP-KLH, showed them to have levels of specific antibody to this antigen which were clearly lower than in any of the antigen-injected groups, a result which was to be expected . These data were excluded from the analyses described below, as their inclusion would have confounded the results by introducing much variation, attributable to the low antibody responses, when the purpose of the analyses was to determine whether or not exposure to radiation was affecting antibody response to antigen . To compare the treatment groups non-parametric, one-way analyses of variance using the Kruskal-Wallis test on the ranked data (Conover 1980) were made for subsets of the data . These subsets were restricted to data collected from week 8 post-injection until the end of sampling (i .e . after achievement of full response as determined by eye from Figures 1 and 3) . To remove any remaining time effects the median titre (over all groups) at each sample time was subtracted from each of the titre values at that time . In order to display differences between treatment groups, mean ranks for each were plotted with a significance band based on Fisher's least significant difference (Figures 2 and 4) . When these bands do not overlap they are significantly different at the 95% level (Nicholson et al . 1991) . The mean titre values corresponding to each mean rank are also given in the figures .
Radiation effects on fish immune response
3. Results
Int J Radiat Biol Downloaded from informahealthcare.com by University of Melbourne on 11/23/14 For personal use only.
3 .1 . Experiment I The titres obtained at each sample time are shown in Figure 1 . All antigen-injected groups show a rise in titre until 9 weeks and then a plateau . There is some variation in response with occasional samples having low titres, most notably in the 3 . 73 mGy h - 1 group . As expected the FCA-only group had significantly lower titres than the antigen-injected groups . However, even here a rise was seen, though considerably lower and less sustained. There was no significant difference between the antigen-injected groups . The plot of mean rank with significance bands (Figure 2) illustrates this . Bands from irradiated groups all overlap with the 0 mGy h -1 group and with each other .
3 .2. Experiment 2 The ODs obtained following antigen injection are shown for each treatment group in Figure 3 . Those for the FCA-only group are not shown as most were less than 0 . 1 and the highest only 0 .25 . In all groups shown there is an increase in OD starting at 4 weeks post-injection and plateauing at 7-8 weeks . At most sampling times there are also some fish in all groups which have low ODs . There was a significant difference between antigen-injected groups . The plot of mean rank with significance bands (Figure 4) shows, formally, that the band for the 4 .6 mGy h -1 group does not overlap the 0 or 0 .99 mGy h -1 groups, indicating significant difference from these . Informally, the plot also shows a progressive decrease in rank (i .e. in OD, the mean OD corresponding to each mean rank is also given in Figure 4) with increasing dose-rate . The samples taken at 8 and 16 weeks and assayed by latex agglutination and ELISA showed a high correlation (r=0-87, p
J.
RADIAT . BIOL .,
1992,
VOL .
62,
NO .
2, 239-248
The effect of chronic radiation on the humoral immune response of rainbow trout (Onchorhynchus mykiss Walbaum) J. F. KNOWLESt
Int J Radiat Biol Downloaded from informahealthcare.com by University of Melbourne on 11/23/14 For personal use only.
(Received 5 August 1991 ; revision received 16 January 1992; accepted 27 February 1992)
Abstract. Two separate experiments have examined the effect of exposing rainbow trout to chronic y-radiation, commencing immediately after fertilization . In experiment 1 the period of exposure extended for 20 days with groups receiving mean dose rates of 1 . 87, 3 . 73 and 9 .03 mGy h -1, and mean total accumulated doses of 0 . 83, 1 . 66 and 4 . 01 Gy respectively . At 5 months of age fish were tested for specific antibody response to dinitrophenol coupled to keyhole limpet haemocyanin (DNP-KLH) and there was no significant difference in titre between irradiated groups and unirradiated controls . In experiment 2 the exposure period was extended to 246 days from fertilization. Mean dose-rates to the three groups used were the same as in the first experiment until hatching at 21 days and then lower with rates of 0 . 99, 1-9, and 4 . 66 mGy h -1 to the freeswimming fish . The mean total accumulated doses over the whole irradiation period were 5 . 43, 10 . 53 and 25 .43 Gy respectively . The antibody response to DNP-KLH was significantly lower in trout receiving the highest dose-rate when compared with those of unirradiated controls or the lowest dose-rate group . The significance of these results is discussed in relation to radiation levels in areas of radioactive waste disposal, and results from a similar study published previously .
1 . Introduction The possibility that increased exposure to radioactive and chemical environmental contaminants might cause a rise in the incidence of fish diseases is of considerable interest, not only to the fishing and aquaculture industries, but also to the public in general . The immune system plays a major role in disease resistance and the effects of toxic agents on the fish immune system have been the subject of considerable research effort (Zeeman and Brindley 1981 ; Anderson et al. 1984) . In their major review of the subject Zeeman and Brindley showed that, in contrast to the effects of chemical toxicants, radiation effects have been little investigated . Where such investigations have taken place large, single doses of radiation have been given rather than the chronic low dose-rate radiation exposures, which are pertinent to radioactive contamination in the environtMinistry of Agriculture, Fisheries and Food, Directorate of Fisheries Research, Fisheries Laboratory, Lowestoft, Suffolk NR33 OHT, UK .
ment. The important exceptions to this are the studies of Strand and his co-workers (Strand et al. 1973, 1977, 1982) . They exposed developing rainbow trout to low-level fl-radiation by keeping them in tritiated water from fertilization through 20 days of embryogenesis . At 5 months post-hatching the fish were challenged with killed bacteria and there was a significant reduction in antibody titre even in those which had received total doses of only 40 mGy given at a rate of 2 mGy per day . This remarkable result indicated a very high radiosensitivity for the developing immune system in these fishes . In view of the importance of this finding it is surprising that until now no repeat experiments or further investigations have been undertaken . We have carried out similar studies exposing developing trout to chronic y-radiation at mean dose-rates of between 1 and 9 mGy h -1 and this paper reports two experiments : in one the exposure was from fertilization through 20 days of embryogenesis, the same period as in the experiments of Strand et al . (1977) ; in the other irradiation continued from fertilization for 35 weeks . These experiments were carried out separately and at different times using some different methods . For these reasons they are considered separately and comparisons are only made between experimental groups in the same experiment . However, both are addressing the same general problem and many of the methods used were similar .
2. Materials and methods Two separate experiments were carried out but many features of the two were identical. Eggs and sperm were obtained from rainbow trout, Onchorhynchus mykiss, at the MAFF fish farm at Sacrewell, Northants and transferred to Lowestoft in cooled flasks . For each experiment eggs stripped from three females were fertilized with sperm from a similar number of males . The fertilized eggs were divided into four approximately equal groups, each of which
240
J. F. Knowles
kept in well-aerated, dechlorinated running water . Eggs in one tank were controls and those in the other three received radiation . Trays and tanks were identical and mains water passed through the same dechlorinator (Elgamat AC 1, Elga, High Wycombe, UK) to all tanks, irradiated and unirradiated . Water temperatures were similar in all tanks (see below) .
Int J Radiat Biol Downloaded from informahealthcare.com by University of Melbourne on 11/23/14 For personal use only.
2 .1 .
Irradiation
The radiation facility at Lowestoft consists of a shielded tank room within which 137 Cs y-ray sources are arranged in such a way that three sets of tanks may be exposed, each to a different low dose-rate (Rackham et al . 1991) . The sources are exposed by remote control and the precise period of exposure recorded automatically. This was usually about 22 h per day, allowing 2 h for fish feeding, tank cleaning and other maintenance . Dosimetry was carried out using LiF powder, thermoluminescent dosimeters (TLDs) which were read on a Harshaw 3000A TLD reader. Dose-rates to developing eggs in the egg trays were measured by placing TLDs at nine locations in the tray for each dose group . Dose-rates to experimental tanks were obtained by placing 27 TLDs (a 3 x 3 x 3 matrix) in each (Rackham et al . 1991) . Mean dose-rates for egg trays, whole tanks and bottom layer of tanks are given in Table 1 . The attenuation of radiation by water in the tanks is indicated by a comparison of dose-rates in the nearsurface egg trays and the bottom layers of the tanks, and also by the relatively large standard deviation of the mean whole tank dose-rate . 2.2 . Blood sampling and antigen injection After hatching, the fish were fed pelleted food (BP Nutrition, Northwich, UK) of the appropriate size and amount . For antigen injection and blood sampling fish were anaesthetized using a solution of phenoxyethanol in water (1 to 2500) . In both experiments the fish which provided pre-injection
Table 1 . Mean dose-rates to the experimental fish tanks Mean dose rate±standard deviation (mGyh -1 ) to Dose-rate group Low Mid High
Egg trays
Bottom layer
Whole tank
1 . 87±0 . 17 3 . 73±0 . 36 9. 03±0 . 97
0 . 44±0. 08 0 . 89±0. 10 2 . 07±0. 33
0. 99±0 . 52 1 . 90±0 . 99 4. 60±2 . 38
blood samples were then killed . Those in experiment 1 were 10-12 g at this time . After subsequent blood samples fish were allowed to survive . The fish in experiment 1 were about 20 g by the first postantigen injection sample and 0 . 1 ml of blood was taken . Rainbow trout of this size had 0 . 15-0 . 2 ml of blood sampled without affecting survival or activity (Lied et al . 1975) . As the fish grew older and increased in size, volumes of approximately 0 . 2 ml were taken . Blood samples were taken from the caudal vein using a 26-gauge needle . After each sampling the blood was placed in 1 . 5 ml disposable plastic centrifuge tubes (Boehringer, Lewes, UK) and left to clot for 1-2 h at room temperature, then overnight at 4°C . It was then centrifuged at 3000g for 15 min and the serum removed and stored at -20°C . The antigen used in both experiments was dinitrophenol coupled to keyhole limpet haemocyanin (DNP-KLH ; 494 DNP groups per protein molecule) obtained from Calbiochem-Behring (La Jolla, USA) . It was made up at a concentration of 5 mg/ml in Freund's complete adjuvant (FCA ; Sigma, Poole, UK) and injected intraperitoneally at approximately 1 mg per 100 g body weight . FCA-only controls received injections of similar volumes of FCA alone . 2.3 . Experiment 1 Developing trout eggs were exposed to radiation for 20 days (443-8h actual exposure time) from fertilization to the `eyed' stage . At this time the trays containing them were transferred to unirradiated tanks . The mean dose-rates to eggs were 1 . 87, 3 . 73 and 9 . 03 mGy h -1 and total doses received were 0 . 83, 1 . 66 and 4 . 01 Gy respectively . There were, in addition, two unirradiated groups of eggs . After hatching the fry remained in the same tanks and were reared normally until 5 months of age . At this time they were all anaesthetized and each group marked by tattooing with Alcian blue using a 'Panject' needleless injector (Wright Dental Group, Dundee, UK) or by removal of the adipose fin . Pre antigen-injection blood samples were taken from 10 fish in each group at this time . One week later the fish were all anaesthetized again and injected intraperitoneally with antigen in FCA, except for one of the unirradiated groups which received FCA only . The groups which had received antigen were named by their radiation dose rate : 0, 1 . 87, 3 . 75 or 9 . 03 mGy h -1 groups . The other group was known as FCA-only. After injection fish from all groups were kept together in a single large tank (1500 litres) . Blood samples were subsequently taken
Int J Radiat Biol Downloaded from informahealthcare.com by University of Melbourne on 11/23/14 For personal use only.
Radiation effects on fish immune response
at intervals as shown in Figure 1 . Usually 10 fish were sampled from each group except for 3. 73 mGy h -1 and FCA-only where the samples were four to six fish at each time up to 12 weeks and only one and two fish respectively at 19 weeks . Throughout the experiment all tanks received water from the same mains supply. In the tanks holding fish prior to antigen injection the water was dechlorinated, but the supply to the single large tank used later was not . However, chlorine levels were checked regularly and remained below 0 .06 mg 1 -1 . Water temperatures were 8 .5-10°C during pre-hatching and then rose steadily 18°C at injection before falling to 12°C by 9 weeks post-injection. Sera from blood samples were assayed for antibody using a latex particle agglutination assay modified from that described by Hudson and Hay (1989) : (1) A suspension of latex particles 400 µl; (Difco, East Molesey, UK) were washed in 15 ml of 0.05m glycine buffered saline at pH 8 . 7 (GBS), centrifuged at 12 500g for 15 min and the particles resuspended in a further 5 ml of 0.05 M GBS . (2) Particles were coated with antigen by adding 40 tl of DNP-KLH solution in 0 .05 M GBS (10 mg/ml) to the suspension, mixing and leaving for 3-4h at room temperature-this step was omitted for uncoated, latex-only suspensions . (3) The particles were washed twice as before, finally suspended in 5 ml of 0 . 25 M GBS containing 0 .5% bovine serum albumin (BSA) and stored at 4°C until use . (4) The trout sera were thawed and serial doubling dilutions made with 0 . 25 M GBS containing 0 .5% BSA . The dilutions were made in microtitre plates (LIP, Shipley, England) with 25 yl of diluted serum per well, and a duplicate series was made for each test serum . (5) DNP-KLH-treated latex suspension (25 µl) was added to each well for one series, while latex-only suspension was added for the other . (6) The microtitre plates were covered and left for 24-48 h at room temperature before the titres were read and recorded .
2.4. Experiment 2 Developing trout were irradiated continuously from fertilization for a total of 246 days . After hatching at a mean of 21 days the fry passed through holes in the egg trays and fell to the tank floors where they remained for a mean of 32 days before becom-
24 1
ing free-swimming throughout the tank for the remaining 193 days of exposure . The dose-rates thus varied during development and are given for the various positions in Table 1 . For simplicity the groups were named by the mean dose-rates received for the majority of their exposure period, i.e. 0, 0 .99, 1 .90 and 4 .60 mGy h -1 groups . It should, however, be remembered that mean dose-rates other than these were received during the first 53 days postfertilization(Table 1) . The integrated total accumulated doses received by the 0.99, 1 .9 and 4.60 mGy h -1 groups were 5 . 43, 10 . 53 and 25 .43 Gy respectively . At the termination of irradiation blood samples were taken from 10 fish in each irradiated and unirradiated group . The trout were then transferred to eight new tanks (450 litres each), two replicate tanks for each irradiated group and the 0 mGy h -1 group . Fifty fish were placed in each tank, giving a total of a 100 for each group . Fifty unirradiated fish which formed the FCA-only group were left in their original tank . After a week of acclimatization all fish were injected with antigen in FCA, except for the FCA-only which received adjuvant alone . Blood samples were then taken at intervals until 21 weeks post-injection . To ensure against oversampling of the same fish, samples of 10 antigeninjected fish per group were taken from one tank at each sampling and the two replicate tanks were sampled alternately. Five fish per sample were taken from the FCA-only group, which was maintained in a single tank . The water supply to all tanks was from the same mains but that to the FCA-only group remained dechlorinated throughout while the supply to the eight new tanks was not. However, chlorine levels were checked regularly, as in experiment 1, and remained below 0 .06mg1 -1 . Water temperature was 12-14°C during pre-hatching, fell to 9-11°C during the following 4 months, rose to 16°C at antigen injection (8 months) and this was maintained for the next 3 months before falling to 11-13°C at the termination of the experiment . Sera from blood samples were assayed for specific antibody using an enzyme-linked immunosorbent assay (ELISA) which was a modification of that used by P. Dixon and H . de Groot (personal communication) . ELISA methods are more elegant and sensitive than agglutination methods, and it was possible to develop one as an improvement for this second experiment mainly thanks to the gift of monoclonal mouse anti-trout y-globulin from Dr C . Secombes, University of Aberdeen, who had obtained the clonal line, 1-14, from Professor G . Warr, University of S . Carolina . The protocol used was arrived at by extensive chequer board and other trials, and may be summarized as follows :
Int J Radiat Biol Downloaded from informahealthcare.com by University of Melbourne on 11/23/14 For personal use only.
242
J . F. Knowles
(1) Microtitre plates (Costar, Cambridge, USA) coated by adding 200 µl per well of 0 . 5 µg/ml DNP-KLH in carbonate buffer (pH 9 . 6) and incubating overnight at 20 ° C . (2) Plates washed (all washes are three times) in phosphate buffered saline (PBS, pH 7 . 4) . (3) Blocking carried out by adding 300µl per well of 1 % bovine serum albumin (BSA) in carbonate buffer and incubating for 1 h at 37°C . (4) Wash in PBS with 0. 05% Tween-20 (PBST) . (5) Plates dried overnight, sealed and stored at -20°C until required . (6) Test sera were diluted 1 :1000 in PBST, 200 µl added per coated well and incubated for 90 min at 20-22°C . (7) Wash in PBST . (8) 200 p1 of ascites fluid containing monoclonal mouse anti-trout y-globulin diluted 1 :2000 in PBST with 1 % BSA were added per well and incubated for 90 min at 4°C . Wash in PBST. 200 µl of goat anti-mouse IgG conjugated with alkaline phosphatase (Sigma, St Louis, USA) diluted 1 :1000 in PBST with 1 % BSA were added per well and incubated for 90 min at 4 ° C . (11) Wash in PBST . (12) 200µl of a substrate solution containing 1 mg/ml p-nitrophenyl phosphate in diethanolamine buffer pH 9 . 8 were added per well and incubated for approximately 30 min (see later) at room temperature . (13) The reaction was stopped by adding of 50 yl of 3 M NaOH . (14) Optical density (ODs) were read at 405 nm .
All samples taken at any one time, a total of 45, were assayed together and two identical replicate microtitre plates were prepared . Outer wells were not used, the remaining 60 accommodating the 45 samples and various standards, all of which were placed in a set pattern . Reference and external positive standard sera were obtained from trout which had been injected with DNP-KLH in FCA 8 weeks (reference) or 12 weeks (external) previously . Sera from completely untreated trout of similar ages to experimental fish were used as a negative standard . In each case pooled sera were frozen in 20 p1 aliquots until use, when they were diluted as for test samples . Each microtitre plate had two replicate wells of each standard . In order to standardize substrate incubation time the ODs in the reference positive wells were
always allowed to develop to approximately 1 .2 before stopping the reaction . Slight under- or overshooting was compensated for by obtaining the mean of the two reference positive values and dividing this into 1 . 2 to obtain a correction factor, which was used for all ODs on the microtitre plate . The external standard was used as a quality control measure to monitor for errors (e .g. dilution) arising from the reference positive standard . The ELISA was developed as an improvement for this second experiment, as already noted . In order to check the correlation between results obtained by ELISA and latex agglutination, samples of sera taken at 8 and 16 weeks were assayed by both methods .
2 .5 . Statistical analysis Results from both experiments have been analysed in a similar way. From a preliminary analysis of the data it was clear that for both experiments the response within each treatment group at each time was not normally distributed, and that no simple transformation of the data would result in normality . The data for FCA-only groups, which received no DNP-KLH, showed them to have levels of specific antibody to this antigen which were clearly lower than in any of the antigen-injected groups, a result which was to be expected . These data were excluded from the analyses described below, as their inclusion would have confounded the results by introducing much variation, attributable to the low antibody responses, when the purpose of the analyses was to determine whether or not exposure to radiation was affecting antibody response to antigen . To compare the treatment groups non-parametric, one-way analyses of variance using the Kruskal-Wallis test on the ranked data (Conover 1980) were made for subsets of the data . These subsets were restricted to data collected from week 8 post-injection until the end of sampling (i .e . after achievement of full response as determined by eye from Figures 1 and 3) . To remove any remaining time effects the median titre (over all groups) at each sample time was subtracted from each of the titre values at that time . In order to display differences between treatment groups, mean ranks for each were plotted with a significance band based on Fisher's least significant difference (Figures 2 and 4) . When these bands do not overlap they are significantly different at the 95% level (Nicholson et al . 1991) . The mean titre values corresponding to each mean rank are also given in the figures .
Radiation effects on fish immune response
3. Results
Int J Radiat Biol Downloaded from informahealthcare.com by University of Melbourne on 11/23/14 For personal use only.
3 .1 . Experiment I The titres obtained at each sample time are shown in Figure 1 . All antigen-injected groups show a rise in titre until 9 weeks and then a plateau . There is some variation in response with occasional samples having low titres, most notably in the 3 . 73 mGy h - 1 group . As expected the FCA-only group had significantly lower titres than the antigen-injected groups . However, even here a rise was seen, though considerably lower and less sustained. There was no significant difference between the antigen-injected groups . The plot of mean rank with significance bands (Figure 2) illustrates this . Bands from irradiated groups all overlap with the 0 mGy h -1 group and with each other .
3 .2. Experiment 2 The ODs obtained following antigen injection are shown for each treatment group in Figure 3 . Those for the FCA-only group are not shown as most were less than 0 . 1 and the highest only 0 .25 . In all groups shown there is an increase in OD starting at 4 weeks post-injection and plateauing at 7-8 weeks . At most sampling times there are also some fish in all groups which have low ODs . There was a significant difference between antigen-injected groups . The plot of mean rank with significance bands (Figure 4) shows, formally, that the band for the 4 .6 mGy h -1 group does not overlap the 0 or 0 .99 mGy h -1 groups, indicating significant difference from these . Informally, the plot also shows a progressive decrease in rank (i .e. in OD, the mean OD corresponding to each mean rank is also given in Figure 4) with increasing dose-rate . The samples taken at 8 and 16 weeks and assayed by latex agglutination and ELISA showed a high correlation (r=0-87, p
