Continuous Tracer Infusion to Measure in Vivo Metabolite Turnover Rates in Trout

Reliable methods to measure metabolite turnover rates in vivo are widely available (Hetenyi et al. 1983), and they have been used extensively in mammalian metabolic studies, especially for humans (Wolfe, 1992). In comparison, surprisingly little work has been carried out on fish (Garin et al. 1987; Weber and Zwingelstein, 1995), mainly because it has been difficult to adapt tracer techniques to aquatic animals of smaller relative body size. The bolus injection method has been used almost exclusively in fish studies (Weber and Haman, 1995) because it only requires a single catheter and can consequently be performed with a simple dorsal aortic cannula. This method has serious limitations, however, and it has been almost totally abandoned for mammals because it can only be used under steady-state conditions and each experiment only allows the calculation of a single turnover rate after analysis of a minimum of 6–10 blood samples. In mammals, metabolite turnover rates are quantified using the more versatile continuous infusion technique because it allows measurements under steady-state as well as non-steady-state conditions; also, much more information can be obtained from a single experiment because instantaneous flux can be calculated from each blood sample separately. Unfortunately, this method requires the surgical placement of two catheters – one to infuse the metabolic tracer, the other for blood sampling – and it has never been adapted to investigate the metabolism of trout, the most commonly used teleost model.

The now classic dorsal aorta cannulation technique of Smith and Bell (1964), later modified by Soivio et al. (1975), revolutionized the study of fish metabolism by allowing repeated blood sampling and intravascular injections in undisturbed, non-anaesthetized fish. Over the last 30 years, dorsal aorta cannulation has been used routinely to investigate fundamental aspects of fish biochemistry and physiology. This surgical approach has provided invaluable information on the effects of various stresses including exercise, changes in water temperature, pH, salinity, oxygen content and toxicant concentration. In studies of fish metabolism, the blood concentration of key metabolites has been monitored to examine the biochemical consequences of these stresses. In this context, turnover rate is a much more informative parameter than concentration, and the temptation to use concentration changes to draw conclusions about fluxes has not always been resisted. Such extrapolation is clearly unwarranted, however, because flux and concentration do not necessarily change in parallel (Wolfe, 1992); a very large change in turnover rate can be associated with an increase, no change or even a decrease in concentration, depending on the prevailing difference between rates of appearance and disappearance. Here, therefore, our goals were to develop an easily performed double catheterization for rainbow trout and to use the continuous tracer infusion technique to quantify resting glucose turnover as a practical example.

Rainbow trout Oncorhynchus mykiss (Walbaum) of both sexes (831–1265 g) were purchased from Linwood Acres Trout Farm (Campbellcroft, Ontario, Canada) and held in 550 l flow-through tanks at 15 °C. They were kept in dechloraminated, well-oxygenated Ottawa tap water under a 12 h:12 h light:dark photoperiod. The animals were acclimated to these conditions for at least 1 month before experiments and they were fed Purina trout chow three times a week until satiation.

24 h before isotope infusion, each animal was anaesthetized in 0.1 g l⁻¹ ethyl-N-aminobenzoate sulphonic acid (MS-222) buffered with 0.2 g l⁻¹ bicarbonate. During surgery, the anaesthetic solution was recirculated and aerated to perfuse the gills. The cannulation technique of Soivio et al. (1975) was modified to allow the placement of two separate PE-50 catheters (Intramedic, Clay-Adams) in the dorsal aorta. The first catheter was used for isotope infusion and it was inserted at the third gill arch before being fed 8–10 cm caudally into the vessel. The second catheter was used for blood sampling. It was inserted into the artery at the first gill arch and was fed caudally no more than 2 cm (see Fig. 1). These infusion and sampling sites provide an accurate measurement of glucose turnover because the determination of glucose kinetics does not depend on exact catheter placement (Katz, 1992; Norwich, 1992; Wolfe, 1992). Each catheter was filled with heparinized (10 i.u. ml⁻¹) Cortland saline without glucose (Wolf, 1963) and held securely in place by two stitches to the roof of the mouth. After recovery from anaesthesia, each animal was placed in an opaque Plexiglas chamber (60 cm×16 cm×18 cm) supplied with the same quality water as the holding tanks at a rate of 5–6 l min⁻¹. The chamber was sealed with a lid tightly held in place with screws, and both catheters were allowed to exit the chamber through a small opening in the lid. All experiments were carried out between 24 and 48 h after surgery.

While the fish was resting quietly in the Plexiglas chamber, a continuous infusion of 6-[³H]glucose (New England Nuclear, 1.6 TBq mmol⁻¹; Amersham, 1.11 TBq mmol⁻¹) was started with a calibrated syringe pump at 1 ml h⁻¹ (Harvard Apparatus, South Natick, MA, USA). Glucose tritiated in position 6 has been shown to yield the best estimate of glucose turnover because detritiation due to hepatic cycling is minimal (Katz, 1979). A priming dose equivalent to 90 min of infusion was injected as a bolus before starting the pump to decrease the time necessary to reach isotopic steady state in the relatively large glucose pool (Wolfe, 1992). The infusate was prepared daily by drying a sample of the stock solution under N₂ and redissolving it in Cortland saline without heparin. Exact infusion rate was determined for each experiment by counting a sample of the infusate; infusion rates ranged between 30 059 and 108 333 disints min⁻¹ kg⁻¹ min⁻¹. In the Results section, glucose specific activities have been divided by infusion rate for each experiment to provide standardized values allowing meaningful comparisons. Blood samples (300 μl each) were drawn after 40, 50 and 60 min of infusion. Each sample was diluted in 500 μl of perchloric acid (8 %) and centrifuged. After separation, the perchloric acid extracts were used to measure glucose and lactate concentrations at 340 nm using standard enzymatic methods (Bergmeyer, 1985) on a Milton Roy spectrophotometer (Spectronic 1001). Glucose specific activity was measured by counting 100 μl of perchloric acid extract in 10 ml of ACSII scintillation fluid (Amersham, Oakville, Ontario, Canada). This procedure was used because total tritium activity has been shown to be restricted to glucose and water (Katz et al. 1974). Therefore, this simple technique to determine specific activity is used routinely for glucose turnover measurements in dogs (Moates et al. 1988), humans (Molina et al. 1990; Levy et al. 1989), birds (Marsh and Dawson, 1982) and fish (West et al. 1994a,b). Here, the drying step was eliminated because negligible amounts of ³H₂O were found in our experiments owing to the very high water rate turnover rate found in fish. Preliminary experiments showed that ³H₂O activity represented less than 3 % of total activity, even after 3 h of 6-[³H]glucose infusion. Counting was performed on a Tri-Carb 2500 scintillation counter (Packard, Canada) with external quench correction. Glucose turnover rate (R_t) was calculated with the steady-state equation of Steele (1959) and glucose clearance rate was determined by dividing turnover rate by glucose concentration.

Oxygen consumption was measured by stopping the external water supply for three 10 min periods during each isotope infusion, and recycling the chamber water within a 15 l closed system. Particular care was taken to eliminate air bubbles within the system and to avoid exchange between the recirculated water and atmospheric air. Water flow rate through the chamber remained at 5–6 l min⁻¹ during the metabolic rate measurements. During each 10 min interval, O₂ concentration was recorded every minute with a calibrated oxygen electrode (Oxyguard, Handy MK III, Valox Ltd). The total decrease in never exceeded 10 % of saturation but it was large enough to measure oxygen consumption accurately. The total volume of recirculated water (15 l) was selected to provide an optimal ratio of fish size to respirometer volume (Steffensen, 1989). The rate of oxygen consumption was calculated using the closed-system equation of Steffensen (1989). After each metabolic rate measurement, return to saturation was accomplished by flushing the system for 5 min. All results are given as means ± S.E.M.

The double cannulation technique proposed in this paper is summarized in Fig. 1 where exact catheter placements are indicated. Complete surgery was achieved in 25–40 min depending on the individual fish (where surgery time is the time elapsed between starting the anaesthesia and complete recovery with the animal having regained its normal balance).

Surgical success rate was higher than 90 % with the large rainbow trout used here (see Table 1). Additional experiments (results not shown) revealed that it is possible to cannulate smaller animals (down to 600 g) with the same success rate; however, performing this double catheterization in animals smaller than 500 g became increasingly difficult because of the lack of working space within the mouth. The surgical procedure occasioned minimal blood loss, as indicated by the maintenance of high haematocrit values (Table 1). After removing catheterized fish from the anaesthesia solution, recovery occurred in 5–10 min.

Both catheters were kept patent between surgery and isotope infusion without problems. The opaque Plexiglas chamber, low infusion rate (1 ml h⁻¹) and good sampling catheters allowed isotope infusion and blood sampling to occur without disturbing the fish, as indicated by the normal resting metabolic rates measured during the experiments (⁠ ranged between 33.5 and 63.3 μmol O₂ kg⁻¹ min⁻¹, and it averaged 42.2±3.7 μmol O₂ kg⁻¹ min⁻¹, see Table 1). In addition, no thrashing or unusual movements were observed as a result of isotope infusion or blood sampling; low blood lactate and normal blood glucose concentrations were maintained throughout the infusion procedure (Table 1).

The infusion rates, priming dose and blood sampling schedule selected here were adequate to achieve concentration and isotopic steady-state conditions for glucose in 40 min of infusion. Glucose concentration, specific activity and turnover rate did not change significantly over time after 40 min of infusion (Fig. 2, P>0.7, ANOVA). Results for individual infusion experiments are presented in Table 1. The resting glucose turnover rate R_t of rainbow trout ranged between 6.2 and 11.8 μmol kg⁻¹ min⁻¹ and it averaged 9.0±0.7 μmol kg⁻¹ min⁻¹. Turnover rate and concentration were not significantly correlated (r=0.49, P=0.22). Glucose clearance rate ranged between 0.9 and 2.7 ml kg⁻¹ min⁻¹ and averaged 1.7±0.2 ml kg⁻¹ min⁻¹.

This paper describes a double dorsal aorta catheterization technique allowing the easy measurement of substrate turnover rates by continuous infusion of metabolic tracers in rainbow trout. The time necessary to carry out this new surgical procedure is only slightly longer than for a single aortic cannulation (Soivio et al. 1975) and it can be performed with minimal training. Moreover, the animals recover rapidly from a simple operation causing negligible blood loss and no more stress than the commonly used single cannulation. Other, less satisfactory techniques were considered before selecting the method described in this paper; the most promising alternative was to place the infusion catheter in the coeliac artery and to sample blood from the dorsal aorta. However, this approach was eventually abandoned because coeliac artery cannulation is far more invasive, and therefore much more stressful (i.e. longer surgery time, more blood loss and a large external wound to heal). In addition, we found it impossible to reach isotopic steady state when labelled glucose was administered through the coeliac artery, even after prolonged infusion in the healthiest animals tested.

The technique proposed in this paper makes it possible to measure metabolic substrate fluxes by continuous infusion in rainbow trout, instead of relying on the commonly used, but much less versatile, bolus injection method (Weber and Haman, 1995; Weber and Zwingelstein, 1995). There are three main reasons why using continuous infusion is advantageous for fish metabolism studies: (1) this technique has been shown to yield much more accurate estimates of substrate flux than bolus injection under steady-state conditions (when concentration and flux of the metabolite investigated are constant throughout the experiment) (Allsop et al. 1978, 1979); (2) bolus injection cannot be used to monitor rates of appearance and disappearance independently under non-steady-state conditions (Hetenyi et al. 1983); and (3) a single value of flux can be measured with much less blood (only one or two blood samples are necessary using continuous infusion, whereas a complete specific activity decay curve must be generated using bolus injection) (Wolfe, 1992).

As a practical example, we measured the glucose kinetics of resting rainbow trout to establish the necessary experimental conditions for future use of the proposed technique in this species. If the glucose pool is initially primed with the equivalent of a 90 min infusion, isotopic steady state is achieved in less than 40 min with infusion rates ranging between 30 ×10³ and 110 ×10³ disints min⁻¹ kg⁻¹ min⁻¹. For the measurement of glucose turnover rate, injecting a priming dose is necessary to reach steady state rapidly enough, because the total glucose pool of vertebrates is known to be large (reaching steady state without priming the pool would take several hours). In all our experiments, glucose concentration and specific activity remained constant after 40 min (Fig. 2), and these results show that, beyond this time, the proposed experimental set-up can be used to quantify rates of glucose production and disposal under non-steady-state conditions if required (e.g. during swimming, when water quality is altered or when hormonal manipulations are imposed on the animal).

Resting glucose turnover rates of rainbow trout averaged 9.0±0.7 μmol kg⁻¹ min⁻¹ (Table 1) and variability among individuals was low [R_t ranging between 6.2 and 11.8 μmol kg⁻¹ min⁻¹ (N=8)]. In addition, all our fish were in a true resting state during the infusions because they showed normal resting metabolic rates for this species and their blood lactate concentration remained low throughout the experiments. We were able to find three comparable published values of resting glucose R_t in rainbow trout, all of them measured by bolus injection at 13–15 °C (Andersen et al. 1991; Cowey et al. 1977; Washburn et al. 1992). These values were 4–9 times lower (ranging from 1 to 2.4 μmol kg⁻¹ min⁻¹) than we measured here using continuous infusion. It is difficult to compare these values with ours because the bolus injection method is much more prone to experimental errors. In the studies by Andersen et al. (1991) and Cowey et al. (1977), the first blood sample was only taken 60 min after isotope injection, and the most crucial portion of the decay curve was therefore totally missed. An accurate estimate of the surface area under the specific activity decay curve would only have been possible if the early part of the curve had been defined adequately (i.e. many samples must be taken during the first 60 min after injection to generate a usable decay curve). Furthermore, their animals were not truly at rest; they were taken repeatedly out of water for isotope injection and to sample blood from the caudal vein (Cowey et al. 1977), they showed hyperglycaemia (Andersen et al. 1991) or they were anaesthetized with MS-222 shortly before isotope injection (Washburn et al. 1992). Contrary to these three studies, West et al. (1994b) recently found resting glucose turnover rates twice as high as were measured here, but these high fluxes were probably caused by an unusually high plasma glucose concentration (averaging 20 mmol l⁻¹). In the present study, glucose concentration and turnover rate were not correlated. In contrast, West et al. (1994a,b) found a significant correlation in both hyperglycaemic trout (6–38 mmol l⁻¹ plasma glucose) and carp (3–17 mmol l⁻¹). These results suggest that flux and concentration only show a significant correlation when the animals become hyperglycaemic.

Resting glucose turnover rates have also been measured in several fish species other than trout, and Fig. 3 summarizes published results. Except for a single study where continuous infusions were carried out in American eels Anguilla rostrata (Cornish and Moon, 1985), glucose turnover rates were always measured using the bolus injection technique. Interestingly, the bolus injection experiments all provided lower estimates of turnover rate than we observed here for trout, whereas the American eel values previously obtained by continuous infusion were six times higher. Critical problems with the use of the bolus injection technique have been mentioned above or discussed previously (Weber and Haman, 1995). Because continuous infusion bypasses many of these problems (Wolfe, 1992), and because it is known to be more accurate than the bolus injection technique (Allsop et al. 1978, 1979), previous estimates of glucose turnover measured by bolus injection probably underestimate true rates.

In conclusion, the simple surgical technique described here makes the accurate measurement of metabolite turnover rates possible in live trout and opens the door to the dynamic study of substrate kinetics under a variety of experimental conditions. The method can be adapted to the investigation of most metabolic substrates, including fatty acids, glycerol, amino acids and lactate, in addition to glucose. In particular, infusion/sampling sites may have to be modified to provide the best possible estimates of flux for metabolites entering the circulation at multiple locations. Future use of the continuous infusion technique under steady-state and non-steady-state conditions will add a new dimension to our understanding of fish metabolism.