Abstract Blood samples are used in many organisms as an indication of health and physiological state. In yellowfin tuna, blood analysis is complicated by stress responses to handling. These responses include changes in serum osmolarity, blood metabolites, and release of red blood cells from the spleen leading to increased hematocrit. Samples from stressed fish therefore do not accurately represent the resting health and physiology of an individual. This study presents a new method developed to minimize stress levels during blood sampling. Blood is drawn from the bulbus arteriosis of captive yellowfin tuna while the fish are still in the tank. Twenty-seven samples were taken from 18 individuals over the course of seven weeks. Hematocrit, hemoglobin, red blood cell counts, ions, and metabolites were measured and compared to values obtained from fish sampled using other common techniques. Hematocrit and hemoglobin values were similar to those measured in chronically cannulated tuna, and lower than for fish sampled using a common practice of capture in a net (hematocrit 31.6% + 0.8 vs. 44.1% + 8.5 ; hemoglobin 11.1 g/dl +0.3 vs. 17.0 g/dl +2.0). Creatine phosphokinase, alkaline phosphatase, potassium, and creatinine (all related to stress and/or increased muscle activity) significantly increased in net-sampled fish. Therefore, fish sampled using a bulbus puncture appear to be less stressed. Using this sampling technique, an experiment was run in which diet in one tank was changed from a normal 1% fat diet to a high 8% fat diet. Blood samples were taken before and after the 8% diet treatment. Serum free fatty acid concentrations doubled (0.25 mmol/L +0.03 vs. 0.524 mmol/L + 0.03) in the 8% fat fed fish that were sampled. The results of these experiments demonstrate an ability to measure changes in blood chemistry using the new sampling techniques. This study will help determine what is physiologically "normal" in yellowfin tuna blood. Introduction Blood chemistry and hematology are commonly used as an indicator of health and physiological status. However, the accurate interpretation of blood parameters relies on the assumption that sampling procedure does not affect resting values (Houston, 1990). At the Tuna Research and Conservation Center (TRCC), we are interested in determining baseline values for captive yellowfin tuna (Thunnus albacares) blood hematology and chemistry without injuring or stressing the fish. The effects of handling stress and muscle activity on fish blood chemistry and hematology are well documented. In rainbow trout, handling stress causes increased hematocrit and disruption of ionic balance leading to increased potassium and decreased chloride levels in the blood (Railo, E. et al). The stress response involves, in particular, an adrenaline mediated release of red blood cells from the spleen into circulation (McDonald and Milligan, 1992). Yamamoto et al. (1980) showed an increase in red blood cells in circulation as measured by increased hematocrit and decreased spleen volume in capture- stressed yellowtail. In addition, Wells et al (1984) showed increased hematocrit and hemoglobin and decreased pH in Antarctic cod subjected to net-capture stress. Likewise, Devries (1981) demonstrated large increases in blood potassium levels and altered erythrocyte ion exchange abilities in fish that have undergone net-capture stress. A variety of methods are used to control a fish before blood sampling. Past samples have been obtained at the TRCC by capturing a fish with a net and removing it from the water before drawing blood. Observation has shown that this method involves a considerable amount of struggling by the fish for a prolonged period of time before the sample can be taken. Use of cannulation to obtain repeat samples has been helpful in determining resting blood measurements in fish (Houston, 1990). Hematocrit values are generally lower in cannulated fish, which implies less-stress and therefore more realistic measurements (Sovio, et al, 1975). Cannulation of yellowfin tuna has been successful and has shown relatively low hematocrit and hemoglobin values (Korsmeyer, unpublished results) The use of anesthesia in blood sampling is another common method. A blood sample can be taken without physically restraining the fish. However, anesthesia is known to induce many changes in blood chemistry, including activation of adrenaline stress response as well as increased hematocrit and hemoglobin measures (Houston, 1971). Fish must also be handled first in order to anesthetize, and anesthesia has been shown to cause a high level of mortality in tuna (Heidi Dewar, personal comment) The dynamics of stress-mediated changes in blood chemistry and hematological measurements, combined with our interest in determining average tuna blood characteristics, led us to ask the question: can we develop a repeatable blood sampling method that involves minimal stress and damage to the fish? The success of a sampling technique can be assessed by comparing results to those obtained using techniques known to cause stress-induced or other chemical changes in the blood. A consistent, accurate sampling technique would prove valuable both in determining base-line values for captive vellowfin tuna blood and for monitoring the health of the fish. Methods: Tanks and Maintenance All yellowfin tuna sampled were held at the TRCC in large holding tanks 30 ft. in diameter and 6 ft. deep. Fish from two tanks were used for blood sampling (T2 and T3). Tank temperatures were held constant at 22°C. Oxygen and pH levels in the water were monitored daily. Blood Sampling We recognized two possible sampling sites, the caudal vessels in the tail region and the bulbus arteriosis, which leads from the ventricle to the ventral aorta (Figure 1). Contact with the tail region is observed to cause the fish to struggle, so sampling from the caudal region would require forced physical restraint of the entire body. Also, because the caudal vein and artery are close to each other, it would not be known if samples were of arterial or venous blood. Sampling from the bulbus is risky because of its close proximity to the ventricle (Figure 1). However, the bulbus is fairly elastic, which could prevent bleeding after sampling. Careful dissection showed that a needle entering anterior to the bone covering the bulbus would not hit the ventricle. An angle of about 45 ° toward posterior was required to avoid hitting the structurally weaker ventral aorta. The bulbus puncture was used to obtain all samples. Blood samples were taken inside the tank. Water levels were lowered to about 1 meter and an average of 7 workers were in the tank throughout the sampling process. Fish were constrained in a small area and caught with water filled vinyl slings. The first three samples were carried out using a forced restraint method. A fish was caught in an open-ended vinyl sling, turned upside down, and placed on a modified v¬ board with as little handling as possible. A vinyl covered foam flap (25 cm wide) was then placed over the body and held firmly. The fish remained calm when a strip of blanket soaked in PolyAqua fish Protectant was placed over the eyes. The sample was then taken from the bulbus and the fish was released. Fish sampled using forced restraint showed significant signs of handling trauma. This led to the development of the sling technique, in which handling of the fish was minimized. A fish was caught in a closed-ended vinyl sling, which was held closed under-water to allow the fish to calm down. In this way, the fish was protected by an envelope of water. The sling was opened and the fish was rolled over. Fish that fought or would not calm down within a few seconds were immediately released to prevent injury and to insure low- stress samples. The fish was held lightly in place and a blood sample was taken. Sampling took less than 30 seconds. The relative stress levels of each fish in terms of struggle and physical response to handling were recorded. Approximately 3ml of blood was drawn into a 5 ml heparinized syringe (needle size 18- 20 ga 1.5 in.). A small amount was put into a tube containing EDTA and immediately placed on ice. The remainder was put into a wax bottom serum separator with clot activator. Blood from the EDTA tube was used to measure hematocrit, hemoglobin (Sigma Diagnostics total hemoglobin kit), and red blood cell counts (Becton-Dickinson unopette microcollection system). Blood in the serum separator was allowed to clot for one hour, then centrifuged for 7 minutes at 3500 rpm. Serum was frozen at -80 C for analysis of free fatty acid concentrations (Wako NEFA kit, Biochemical Diagnostics) and chemistry analysis (Stanford University, Department of Comparative Medicine, Diagnostic Laboratory). See table 1 for blood chemistry parameters measured. Over the course of 7 weeks, 27 samples were obtained from 18 different fish using the sling technique, and blood analysis results were compared to values obtained from both net sampling methods (TRCC data) and chronically cannulated fish (Korsmeyer, unpublished) In three individuals, repeat samples were taken 6 days apart, and hematology values were compared in each fish between sampling events. Feeding Manipulation In order to test the ability of the sling technique in measuring changes in blood chemistry over changes in conditions, fish were sampled from an experimental tank (T2) in which the diet was manipulated. For two weeks the fish were maintained on the standard TRCC low fat diet (28 Kcal/Kg/day, 1% fat) of squid and smelt. The diet was then switched to a high fat diet (28 Kcal/Kg/day, 8% fat) of sardines for four weeks. Blood samples were taken from T2 fish at the end of each of the feeding periods. T3 fish were held to a constant diet over all sampling periods (low fat, 40 Kcal/Kg/day). T3 fish were sampled on three occasions at least one week apart. Results: Repeat Sampling Of the twenty-seven samples using the sling technique there was one mortality, giving a success rate of 95%. The mortality was associated with the forced restraint method. Hematocrit (Figure 2), hemoglobin (Figure 3), and red blood cell count (Figure 4) values are shown for the three repeat samples. There was no significant differences in any of the measurements between sampling dates (paired Students t-test, pr.05). There was no significant difference in hematocrit (p=. 11) or hemoglobin (p=. 6) over dietary treatment (Figure 5, one-way ANOVA, p=05). There were no significant changes in any other blood parameter between diets or tank, therefore values from 27 sling-samples were pooled to determine measured ranges. A summary of blood analysis from all samples in this study is compared to results from net sampling techniques (Table 1). Stress Indices Hemoglobin values of net-sampled fish were significantly higher than sling-sampled fish (Student's t-test, p.05). Significant increases in net-samples were also seen in alkaline phosphatase, total bilirubin, cholesterol, creatinine, phosphorous, potassium, SGOT, and SGPT (Student's t-test, pS.05). Although not statistically significant, an increase was seen in creatine phosphokinase (2000 %). A further comparison between sling-samples, net-samples, and chronically cannulated samples showed significant increases in both hematocrit (Figure 6) and hemoglobin (Figure 7) in net-sampled fish (one way ANOVA, Tukey's HSD, pS.05). There was no significant difference between values obtained from cannulation (hemoglobin -11.3 +7. hematocrit =30.7 + 1.3) and sling-samples (Tukey's HSD, p«05). Diet Manipulation Figure 8 shows the results of the T2 diet manipulation experiment. Free fatty acid concentrations in blood serum doubled (0.25 mmol/L + O.03 to 0.524 mmol/L +0.03) in the high fat diet (Student's t-test, p5.0005). This was the only significant change recorded in the blood over change in diet. Discussion Results from measures of blood chemistry and hematology in this study indicate that bulbus puncture blood sampling can be successfully repeated and results in a lower stress response than in other techniques. We are interested in determining if this technique can provide accurate, repeatable data on yellowfin blood. The hematology results from the three repeat samples show that we can be fairly confident about generating consistent data. This is necessary if we wish to use the information either to set base-line blood parameters or use the results to evaluate the physiological status of an individual. In other words, if we see a change in blood chemistry from one week to the next, we want to be sure this change is due to actual alteration of physiological processes, and not a result of the sampling process itself Pooled data in this study includes samples from fish in different tanks being fed different diets. All hematology and chemistry measures were compared between tanks and diet with no differences being found. Thus, the pooled data represent a good estimate for what the average captive tuna blood should look like. Only free fatty acid concentrations show an increase over the two diet treatments. This is consistent with previous experiments. Reynolds et al (1994) showed an increase in serum free fatty acid concentration in dogs on a high fat diet compared to a high carbohydrate diet. Because the free fatty acid measurements were done under variable conditions, our results tell us very little about the normal range of free fatty acid in tuna blood. However, it does add confidence to the repeatability of the sampling method. The way in which the blood was sampled does not significantly alter the ability to accurately measure changes in blood chemistry. A number of lines of evidence show that samples obtained using bulbus puncture techniques represent a relatively low level of stress. The hematocrit and hemoglobin values obtained in this study are nearly identical to those from chronically cannulated fish. Because cannulated fish are allowed to recover from handling prior to sampling, it is assumed that they are only minimally stressed. These results indicate that the samples we get using the sling technique are representative of fairly non-stressed tuna. Blood chemistry differences were found between sling-sampled and net-sampled fish. Increases in potassium, creatine phosphokinase (CPK), alkaline phosphatase, and creatinine are all likely to be related to sampling condition. CPK is a muscle enzyme that increases rapidly with motor activity (McDonald and Milligan, 1992). Average measures for CPK were 2000% higher in net-sampled fish. This result shows CPK as a possible indicator of stress in captive yellowfin tuna. In one instance, an individual that struggled for up to 30 seconds in the sling before sampling had a CPK level that was 900% higher than the average sling sample (personal observation). Alkaline phosphatase and creatinine are also enzymes used in muscle activity (McDonald and Milligan, 1992). It is apparent that one of the first effects of stress response in these fish is to increase muscle activity. This is probably related to the adrenaline mediated changes in which red blood cells are mobilized to prepare for increased oxygen demand in the body. Serum potassium has been shown to increase during stress due to leakage from muscle cells and lysing of erythrocytes (Railo, 1985). A combination of stress and cell lysis probably occurred in this study, as evidenced by visible red coloring of serum from some net sampled fish (personal observation). The increases we observed in potassium, CPK, alkaline phosphatase, and creatinine in net-sampled fish therefore correlate well with what we know about the probable stress state at the time of sampling. Increases in SGOT and SGPT are often indicative of liver dysfunction (McDonald and Milligan, 1992). We have to take into account the health state of the fish chosen for net- sampling. These fish were usually sick or lethargic, and were chosen for sampling before being euthenized. It is likely that other physiological factors contributed to the results we see. However, many of the changes are probably a result of extended stress and muscle activity that occurs when netting a fish. The observed increases in phosphorous and cholesterol in net-sampled fish do not seem to be directly attributable to a stress response. This increase could either be an artifact of variable handling or environmental conditions that were not controlled, or a secondary effect of changes in blood chemistry and osmolarity. To explain this effect, there will need to be more samples taken using both methods. Likewise, to better quantify all the changes we have observed in this study, more research needs to be done on the comparative aspects of sampling techniques. Conclusion We developed a new method of blood sampling in which blood is taken from the bulbus arteriosis of captive yellowfin tuna. This method has proven to be consistently repeatable. Furthermore, based on hematology and blood chemistry comparisons between this method and other common sampling methods, it appears that samples analyzed in the current study are representative of minimally stressed fish. Hemoglobin and hematocrit values closely resemble those obtained using cannulation methods Therefore, we are confident that this technique can be used to determine accurate parameters for healthy tuna blood, as well as aid in the comparison between various physiological states. Literature Cited Devries, A.L., and Ellory, T.C. 1981. The effect of stress on ion transport in fish erythrocytes. Journal of Physiology. 324:51pp Houston, A.H., Madden, J.A., Woods, R.J., and Miles, H.M. 1971. Some physiological effect of handling and tricain methanesulphonite anesthesiation upon brook trout, Salvelinus fontinalus. Journal of Fisheries Research Board of Canada. 28:625-633. Houston, A.H. (1990) Blood and circulation. In Methods for Fish Biology (ed. C.B. Schreck and P.B. Moyle). pp. 273-334. Maryland: American Fisheries Society. Korsmeyer, 1996. Unpublished results. McDonald, D.G. and Milligan, C.L. (1992). Chemical properties of the blood. In Fish Physiology Vol. 12B (ed. W.S. Hoar, D.J. Randall, and A.P. Farrell). pp. 55-133. San Diego: Academic Press. Railo, E., Nikinmaa, M., and Soivio, A. 1985. Effects of sampling on blood parameters in rainbow trout, Salmo gairdner: Richardson. Journal of Fish Biology. 26: 725- 732. Reynolds, A.J., Fuhrer, L., Dunlap, H., Finke, M.D., and Kallfelz, F.A. 1994. Lipid metabolite responses to diet and training in sled dogs. Journal of Nutrition. 124:27545-2759s. Sovio, A., Nyholm, K., and Westman, K. 1975. A technique for repeated sampling of the blood of individual resting fish. Journal of Experimental Biology. 63:207-217. Wells, RM.G., et al. (1984). Recovery from stress following capture and anesthesia of Antarctic fish: hematology and blood chemistry. Journal of Fish Biology. 25:567- 576. Yamamoto, K.I., Itazawa, Y., and Kobayashi, H. 1980. Supply of erythrocytes into the circulating blood from the spleen of exercised fish. Comparative Biochemistry and Physiology. 65A:5-11. Table 1 Comparison of blood analysis between fish sampled using sling technique and fish that were were netted or dying. Data for sling sampled fish represents baseline values obtained during this study. Sling Technique Netted Fish Hematology Mean t-SE Mean t/-SEn t-test % increase 31.6 +.8 Hematocrit (%) 44.1 + 11.0 5 N.S. (p=.2) 40 11.1 +.3 Hemoglobin (g/dl) 16.2+2.1 8 “ (p=.02) 46 2.25 +.06 RBC (E6/cu mm) Chemistry 7.6 +5 43.5 + 8.6 8 “ (p=.002) 472 Alkaline Phophatase (IU/L) 8.3 +.8 Anion gap Bilirubin - direct (mg/dl) Bilirubin - indirect (mg/dl) .02 +.008 175 + 04 8 * (p=.004) 775 02 +.008 Bilirubin - total (mg/dl) 2.7 +.2 3.4 + 4 8 N.S. (p=.3) BUN (mg/dl) 13.6 + 9 8 N.S. (p=.5) 12.9 +2 Calcium (mg/di) 155.0 + 15,5 8 N.S.(p=.24) 77.1 +-1.3 Chloride (meq.L) 193.6 + 24.8 8 " (p=.025) 119.0 + 3.0 Cholesterol (mg/dl) 1653 + 1018 8 N.S. (p=.16) CPK (IU/) 78.5 + 25.6 2000 49 + 06 8 " (p=.002) 19 +.03 Creatinine (mg/dl) 158 94.8 + 3.8 80.9 + 14.3 8 N.S. (p=.44) Glucose (mg/dl) 5.0 +.3 10.6 + 1.4 8 " (p=.002) Phosphorous (mg/dl) 112 1.96 +.07 Potassium (meg/L) 7.4 + 1.5 8 * (p=.007) 278 25 +.06 RBC (E6/cu mm) 189.5 + 58.4 8 " (p=.01) 2.4 + 5 7795 SGOT (AST) (U/) 11 +.08 16 + 5.9 8 " (p=.03) 144454 SGPT (ALT) (U/) 197.7 + 6 193 + 14.2 8 N.S. (p=.8) Sodium (meg/L) 4.4 +6 8 N.S. (p=.4) 3.9 +.06 Total protein (gm/di) 1.6 +.15 8 N.S. (p=.1) 1.2 +.07 Albumin (gm/di) 2.8 + 5 8 N.S. (p=.6) Globulin (gm/di) 2.6 +.09 Figure Legends Figure 1. Blood sampling locations on the yellowfin tuna. In this study, blood was sampled from the bulbus arteriosis. Figure 2. Repeat hematocrit measurements in three individual fish separated by 6 days There is no significant difference (paired Student's t-test, p-66) Figure 3. Repeat hemoglobin measurements in three individual fish separated by 6 days There is no significant difference (paired Student's t-test, p-.44). Figure 4. Repeat red blood cell counts in three individual fish separated by 6 days. There is no significant difference (paired Student's t-test, p- 88). Figure 5. Hematocrit and hemoglobin values in different tank and diet treatments. There are no significant differences (one-way ANOVA, hematocrit p= 11, hemoglobin p-.6). Figure 6. A comparison of hematocrit values in sling-samples (current study), cannulated samples (Korsmeyer, 1996), and net-samples (TRCC data). Net samples are significantly higher than sling samples and cannulated samples (one-way ANOVA, Tukey's HSD, p5.05), while there was no significant difference between sling samples and cannulated samples. Figure 7. A comparison of hemoglobin values in sling-samples (current study), cannulated samples (Korsmeyer, 1996), and net-samples (TRCC data). Net samples are significantly higher than sling samples and cannulated samples (one-way ANOVA, Tukey's HSD, pS.05), while there was no significant difference between sling samples and cannulated samples. Figure 8. Changes in free fatty acid concentrations ([FFA)) in high fat vs low fat diet Increase in [FFA] in high fat diet is significant (Student's t-test, p5.005). Acknowledgments Thanks to everyone at the TRCC who made this project possible. Especially to my advisor Dr. Barbara Block for her patience, enthusiasm, and motivational skills; Heidi Dewar for making this project fun and always having time; Dr. Tom for teaching me about the finer points and keeping things loose, Chuck Farwell because there is no one better at catching a tuna; Ellen Freund for her wisdom and wit; Doug Fudge for his technical support and tuna knowledge, Denise Imai for all of her help and support; Beth Condie and Victor Tubbesing for their assistance; and everyone else who joined in when extra hands were needed. 40 35 30 15 YFT 1 Figure 2 5/1/97 5/7/97 YFT 2 YET 3 14 12 YET Figure 3 5/1/97 5/7/97 YFT 2 YFT 3 YET YFT3 Figure 4 5/1/97 — 5/7/97 Figure 5 Tank 3 samples Tank 2 before fat diet Tank 2 after fat diet 14 8 60 40 20 Figure 6 Sling Technique YFT Chronically cannulated YFT (Korsmeyer, 1996) Netted or Dying YFT n=5 20 15 10 Figure 7 — Sling Technique YFT Chronically cannulated YFT (Korsmeyer, 1996) Netted or Dying YFT n=9 n=6 n=27 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 Figure 8 T2 before fat diet T2 after fat diet