Possible involvement of somatolactin in the regulation of plasma bicarbonate for the compensation of acidosis in rainbow trout

Somatolactin is a putative pituitary hormone, whose presence has been confirmed only in fish. The name was devised because the protein is structurally related to both growth hormone (somatotropin) and prolactin.

Although a definitive physiological function for somatolactin has yet to be determined, changes in plasma levels of somatolactin and the activity of somatolactin cells during various biological processes have been described in previous studies (see Kaneko and Hirano, 1993; Kaneko, 1996). In salmonids, an involvement of somatolactin in the regulation of gonadal function was suggested by an elevation of plasma somatolactin levels (Rand-Weaver et al. 1992; Rand-Weaver and Swanson, 1993) and by activation of immunoreactive somatolactin cells (Olivereau and Rand-Weaver, 1994a, b) during the spawning season. Plasma somatolactin levels in rainbow trout (Oncorhynchus mykiss) increased in response to stress (Rand-Weaver et al. 1993; Kakizawa et al. 1995). Activation of somatolactin cells was observed in rainbow trout transferred from calcium-rich to calcium-poor water (Kakizawa et al. 1993). Furthermore, an involvement of somatolactin in lipid metabolism has also been suggested in a cobalt variant of rainbow trout, which lacks somatolactin-producing cells and accumulates large amounts of abdominal fat (Kaneko et al. 1993). In addition, an elevation of plasma somatolactin level during adaptation to a dark background was observed in red drum (Sciaenops ocellatus) (Zhu and Thomas, 1995).

We recently found that plasma somatolactin concentration was increased during acidosis induced by water acidification or exhaustive exercise (Kakizawa et al. 1996). Water acidification directly raises plasma [H⁺] (lowers pH) by affecting transepithelial ion transfer (Wood, 1989), whereas exhaustive exercise entails acidosis caused by dissociation of the lactic acid formed anaerobically. Plasma lactate levels were markedly increased after exercise (Kakizawa et al. 1996). Water acidification can also accompany a reduction of blood oxygen levels (Packer, 1979). Thus, it is possible that somatolactin was secreted in response to hypoxaemia.

In the present study, we studied the effects of environmental hypercapnia, two levels of hypoxia, acid infusion and exhaustive exercise on plasma somatolactin levels and acid–base status using chronically cannulated rainbow trout. The partial pressure of oxygen and oxygen content were also monitored in the hypoxia experiments to evaluate the importance of blood oxygen level for somatolactin secretion. Furthermore, plasma somatolactin levels were artificially depressed by immunoneutralization, and the effect of acid infusion on acid–base status in these fish was compared with that in control trout.

In all experiments except for environmental hypoxia, immature rainbow trout Oncorhynchus mykiss (Walbaum), weighing 800–1200 g, were obtained from Shizuoka Prefectural Fisheries Experimental Station, Japan, and reared in a stock tank with recirculating fresh water (pH 7.5; 0.6 mmol l⁻¹ Ca²⁺; 1.0 mmol l⁻¹ Na⁺; 1.0 mmol l⁻¹ Cl⁻) at 12 °C for more than 3 weeks before use. Fish were fed commercial trout pellets (Oriental no. 6, Chiba, Japan) at a ration equivalent to 1 % of body mass per day until 2 days before experiments. The hypoxia experiment was conducted at the National Research Institute of Aquaculture, Nikko Branch, where mature rainbow trout of both sexes weighing 700–1400 g were reared in flowing fresh water at 9 °C.

Each fish was anaesthetized by immersion in a solution of 0.01 % tricaine methane sulphonate (MS-222, Sigma, St Louis, MO, USA) in tap water buffered with NaHCO₃. During surgery, the gills were irrigated to maintain anaesthesia with a solution of 0.005 % MS-222, aerated with pure O₂.

The dorsal aorta was cannulated using a catheter (PE50, Clay Adams, Parsippany, NJ, USA) by application of a modified Seldinger technique, similar to the procedure of Soivio et al. (1975). After surgery, the fish was allowed to recover for at least 20 h in a recirculating water system consisting of fish chambers and a 200 l reservoir tank, or in a flow-through water system consisting of fish chambers and a 150 l tank (hypoxia experiment). The water in the tank was vigorously aerated. The fish chambers were covered to prevent visual disturbance of the fish.

Arterial blood (1.0 ml or 1.5 ml in the hypoxia experiment) was sampled through a catheter into a 1 ml glass syringe whose dead space was filled with heparinized (lithium salt, Sigma, 50 i.u. ml⁻¹) saline. 0.2 ml (or 0.5 ml in the hypoxia experiment) was used for pH (and ⁠) determinations. The vremaining blood was centrifuged at 10 000 g for 5 min at 4 °C to obtain plasma. The plasma was used for measurements of total ⁠, Na⁺, K⁺, Ca²⁺, Cl⁻ and somatolactin levels. In the hypoxia experiment, the O₂ content of whole blood was also determined. Plasma samples for measurement of ion and somatolactin levels were stored at −80 °C until analysis. Blood used for and pH measurements was reinfused through the cannula after analysis. The red blood cells obtained after centrifugation were suspended in heparinized (sodium salt, Sigma, 50 i.u. ml⁻¹) saline containing 1.2 mmol l⁻¹ Ca²⁺ (Ca²⁺-saline) and also reinfused into the fish.

After two control samples had been taken 2 h apart, the system water was equilibrated with 2 % CO₂, and arterial blood samples were taken subsequently 1, 3, 6, 24 and 48 h after the onset of hypercapnia and 1, 3 and 24 h post-hypercapnia.

The protocol was essentially the same as described by Tang et al. (1988). After two control samples had been taken 2 h apart, the fish was infused with 0.25 ml per 100 g body mass of 0.1 mmol l⁻¹ HCl in Ca²⁺-saline, followed by a Ca²⁺-saline wash (0.1 ml per 100 g body mass). The infusion rate was 0.5 ml min⁻¹ (infusion took 4–6 min). The time when the acid infusion finished was defined as time zero. Arterial blood samples were taken 5, 30 and 120 min post-infusion.

The protocol was essentially the same as described by Kakizawa et al. (1995). After two control samples had been taken, the fish was forced to swim by chasing it for approximately 20 min in a 1000 l water tank until it failed to respond to further stimulation. An arterial blood sample was taken immediately after the exercise. All fish recovered 5–10 min after chasing.

After two control samples had been taken (2 h apart), water in the flow-through system was bubbled with N₂ to achieve a final water of 9.3 kPa (first experiment) or 6.1 kPa (second experiment). Subsequently, arterial blood samples were taken 1 and 2 h after the onset of hypoxia, and 1 and 15 h post-hypoxia.

To decrease plasma somatolactin levels, the fish was infused with 0.1 ml per 100 g body mass of specific rabbit anti-salmon somatolactin serum, diluted with Ca²⁺-saline (1:10), after two control samples had been taken. The infusion rate was 0.5 ml min⁻¹ (infusion took 100–140 s). The first sample (1.0 ml) was taken 3 min after the infusion ended. 17 min after the first sample had been taken (20 min after the antiserum infusion), the fish was infused with acid, and blood samples were taken as described above for the acid infusion.

Arterial blood pH and were determined using pH (E301/E351, Cameron Instruments, Port Aransas, TX, USA) and (E101, Cameron Instruments) electrodes, adjusted to the water temperature. was determined using a Capni-con V (Cameron Instruments). Plasma was calculated from the measured pH and values, using the Henderson–Hasselbalch equation with appropriate constants (Boutilier et al. 1985). Plasma [HCO₃⁻] was calculated by subtracting the fraction of molecular CO₂ from ⁠. Plasma O₂ content was monitored using an Oxycon (Cameron Instruments).

Plasma levels of Na⁺, K⁺ and Ca²⁺ were determined using an electrolyte analyzer (AVL 984-S, Graz, Austria) and Cl⁻ levels using a chloridometer (Buchler 4-2500, USA).

Plasma concentrations of somatolactin were measured by the specific radioimmunoassay described by Kakizawa et al. (1993). The immunoglobulin G (IgG) in the plasma obtained from the immunoneutralized fish was completely removed using Protein A (IgG-sorb, The Enzyme Center, Malden, MA, USA), which did not affect the plasma somatolactin assay. The minimal detected level was 0.5 ng ml⁻¹, the ED₅₀ was 11.7 ng ml⁻¹, the intra-assay coefficient of variation was 2.5 % (N=10) and the inter-assay coefficient of variation was 3.9 % (N=10).

Dunnett’s test was used to determine statistically significant differences between the mean control values and other values in the same experimental series. Student’s t-tests or paired t-tests (effects of exhaustive exercise) were applied after Bartlett’s test for variance to detect statistically significance differences between control and experimental groups at the same time point in the experiments for acid infusion and immunoneutralization. Data are presented as means ± S.E.M. A level of P<0.05 was considered significant.

The pH decreased by approximately 0.5 units from the control level of 7.96±0.09 to 7.47±0.05 at 1 h, and then gradually increased to 7.70±0.05 at 48 h, although the latter value was still significantly lower than the control value. The pH compensation resulted from an increase in plasma [HCO₃⁻] from 14.3±1.3 mmol l⁻¹ at 1 h to 33.1±4.4 at 48 h. Blood remained at approximately 1.7 kPa throughout the hypercapnic period. Upon recovery, blood declined rapidly, while plasma [HCO₃⁻] slowly returned to the control level. The discrepancy in the time course of changes in and [HCO₃⁻] resulted in a transient alkalosis at 3 h post-hypercapnia (Fig. 1).

There was no significant change in the plasma concentration of somatolactin, which remained within a narrow range of values (4.2–6.7 ng ml⁻¹) throughout the experiment. Plasma [Na⁺] increased transiently at 1 h, by 3 h, and decreased at 72 h. Plasma [Cl⁻] was significantly reduced (115±6 mmol l⁻¹) below the control level (127±5 mmol l⁻¹) at 24 h, remained low until hypercapnia ended, and returned to the control level at 72 h (24 h after hypercapnia). Plasma [Ca²⁺] showed a similar time course of change. Plasma [K⁺] showed a transient but significant decrease at 1 h and increase at 49 h (1 h after hypercapnia) (Fig. 2).

Blood pH in acid-infused fish decreased significantly from the control level of 7.97±0.03 to 7.75±0.04 at 5 min, but then recovered quickly. The changes in were insignificant. Plasma [HCO₃⁻] was significantly lower than the initial level at 5 and 30 min. Plasma somatolactin level increased sharply in response to the acid infusion and became significantly higher than in saline-injected fish at 5 min, but the difference from the initial level was significant only at 120 min. Sham infusion with Ca²⁺-saline caused small, but significant, changes in plasma pH, [HCO₃⁻] and ⁠, but not in somatolactin level (Fig. 3).

Plasma [K⁺] was significantly lower and plasma [Ca²⁺] significantly higher at 5 min, but neither was significantly different from the control level at 30 min (data not shown).

Fig. 4 shows effects of exhaustive exercise on blood pH and and plasma levels of [HCO₃⁻] and somatolactin. The exhaustive exercise resulted in a pronounced and significant decrease in blood pH (from 7.98±0.04 to 7.31±0.12), whereas tended to increase (from 0.33±0.06 to 0.85+0.19 kPa), although the increase was not significant (P=0.088). Plasma [HCO₃⁻] decreased by more than 50 % (from 10.6±2.2 to 5.2±1.2 mmol l⁻¹). Plasma somatolactin concentration tended to increase from 3.9±1.3 ng ml⁻¹ before exercise to 13.4±3.3 ng ml⁻¹ immediately after exercise, although the change was not significant (P=0.068). Plasma [Ca²⁺] also increased from 1.13±0.04 to 1.54±0.16 mmol l⁻¹ (P=0.05) (data not shown). In contrast, plasma levels of Na⁺, Cl⁻ and K⁺ did not change significantly in response to exhaustive exercise (data not shown).

The two different levels of environmental hypoxia resulted in different changes in blood gas status and plasma factors.

As is shown in Fig. 5, an environmental hypoxia of 9.3 kPa resulted in significant decreases in blood ⁠, O₂ content and Blood and returned to the initial level by 1 h post-hypoxia, while blood O₂ content was still significantly lower than the control value at this time. Blood pH and plasma [HCO₃⁻] and somatolactin levels were not affected by the treatment. Plasma levels of Na⁺, K⁺, Ca²⁺ and Cl⁻ were also unaffected (data not shown).

Fig. 6 shows the effects of a more severe environmental hypoxia of 6.1 kPa on blood and plasma factors. Blood decreased by 90 % and plasma O₂ content by 60–80 % during the hypoxia. Blood pH became significantly lower than the initial level at 1 h after the onset of the hypoxia (a reduction from 7.84±0.03 to 7.56±0.05). Plasma [HCO₃⁻] was also lower than the initial level at 2 h (8.7±0.7 mmol l⁻¹ initially, 5.1±0.3 mmol l⁻¹ at 2 h). There was no change in plasma ⁠. Plasma somatolactin levels showed a tendency to increase after the onset of the more severe hypoxia and reached a peak at 2 h, although the change was not significant. A significant decrease in plasma [K⁺] occurred at 2 h post-hypoxia, but no significant change was seen in plasma Na⁺, Cl⁻ and Ca²⁺ levels throughout the experiment (data not shown).

Infusion of anti-somatolactin serum markedly depressed plasma somatolactin level (Fig. 7). The somatolactin level 20 min after the antiserum infusion (time 0, 2.0±0.6 ng ml⁻¹) was significantly lower than the initial (time 0) level in the untreated fish (fish not subjected to antiserum infusion) (4.2±0.8 ng ml⁻¹). Until 30 min after the acid infusion, somatolactin levels in antiserum-infused fish remained significantly lower (1.8±0.9 ng ml⁻¹ at 30 min) than in the untreated fish. Somatolactin levels in antiserum-infused fish did not increase significantly from the initial level even 120 min after the acid infusion. Blood pH in both groups was significantly lower than the initial levels at 5 min and then recovered by 30 min. There was no significant difference in either blood pH or between the two groups throughout the experiment. Plasma [HCO₃⁻] in the immunoneutralized fish was significantly lower than in the untreated fish at 30 and 120 min.

Except for the higher Ca²⁺ level in the immunoneutralized fish at 120 min, there was no difference in plasma levels of Na⁺, Cl⁻, Ca²⁺ and K⁺ between the two groups of fish throughout the experiment. A transient increase in plasma [Cl⁻] in immunoneutralized fish and a decrease in plasma [K⁺] in both groups occurred at 5 min (data not shown).

In the results described above, plasma somatolactin level was found to be elevated along with the decrease in plasma [HCO₃⁻] induced by transient stimulations, namely acid infusion and exhaustive exercise. Thus, the relationship between plasma [HCO₃⁻] and somatolactin level at 5 min after saline or acid infusion and just after exhaustive exercise was examined. There was a significant negative correlation (N=24, r=−0.497; P<0.05) between plasma [HCO₃⁻] and somatolactin level (each value is given as a percentage of the initial level), as shown in Fig. 8.

Elevation of plasma somatolactin level during acidosis in rainbow trout was first observed in our previous study (Kakizawa et al. 1996), in which acidosis was induced by water acidification or exhaustive exercise. In the present study, acidosis was also induced by environmental hypercapnia, severe hypoxia, acid infusion or exhaustive exercise. However, plasma somatolactin levels were increased only when metabolic acidosis was induced by acid infusion, exhaustive exercise or severe hypoxia, but not when respiratory acidosis resulted from environmental hypercapnia.

Environmental hypercapnia induced a rise in blood ⁠. Elevated CO₂ levels resulted in increases in [H+TION O] and [HCO₃⁻], and thereby depressed blood pH with a concomitant rise in [HCO₃⁻] (respiratory acidosis). Plasma somatolactin levels were not affected under these conditions. The sustained increase in [HCO₃⁻] during the hypercapnic period seems to be due to uptake of HCO₃⁻ from or proton extrusion into the ambient water, mainly through the branchial epithelium (Heisler, 1993a). In contrast, acid infusion, exhaustive exercise or severe hypoxia induced an increase in [H⁺] directly or indirectly by dissociation of non-volatile acids. Part of the excess [H⁺] is buffered by bicarbonate in the body fluids, thereby depressing [HCO₃⁻] (metabolic acidosis). Plasma somatolactin level tended to increase in response to these treatments. The moderate hypoxia did not affect the acid–base status or the plasma somatolactin level. Although the changes in plasma somatolactin levels were not significant after exhaustive exercise or during severe hypoxia, the lack of significance is likely to be due to the small number of fish used. Large individual variations in plasma somatolactin levels resulted in the large standard deviations shown in Figs 4 and 6. However, if the data are examined closely from fish to fish, plasma somatolactin levels increased in all fish during acidosis induced by both exhaustive exercise (150–920 % of the initial levels) and severe hypoxia (140–690 %). Thus, plasma somatolactin levels were found to increase only when a depression of pH was due to metabolic causes where [HCO₃⁻] was reduced, but not in response to respiratory causes where [HCO₃⁻] was raised. These results suggest that somatolactin release was triggered by the reduction of [HCO₃⁻] but not by the fall in pH per se.

The hypoxia experiment also allows us to examine whether reductions in blood and/or O₂ content can influence plasma somatolactin levels. The major differences between mild and severe hypoxia are the extent of the reductions of blood and O₂ content and the development of metabolic acidosis. Therefore, it is possible that there is a threshold of blood and/or O₂ content below which a release of somatolactin is enhanced; the drop in blood oxygen level during mild hypoxia may not have been large enough to evoke somatolactin secretion. However, we consider this unlikely for the following reasons. Holeton et al. (1983) reported increases in both and O₂ content in rainbow trout forced to undergo vigorous activity by electrical shock. The fish showed a plasma pH decrease of 0.5 units, with blood lactate levels being considerably elevated until 8 h after exercise. Burst swimming until exhaustion reduced blood by approximately 4 kPa but maintained O₂ content owing to a 40 % rise in haematocrit. Blood pH dropped by approximately 0.3 units, with a sharp rise in lactate level (Primmett et al. 1986). In Atlantic salmon exhausted by manual chasing (the same protocol used in the present study), blood and O₂ content were decreased transiently by only 1.6 kPa for and 2.3 vol% (or by 20 % from the control value) for O₂ content. Blood oxygen levels were restored by 0.5 h after exercise, whereas plasma lactate levels remained elevated even at 8 h (Tufts et al. 1991). Overall, exhaustive exercise appears to influence blood oxygen levels only slightly. Our finding that plasma somatolactin level was increased by exercise thus appears to contradict the idea that blood oxygen level may be important for the release of somatolactin. The above studies as well as our previous study (Kakizawa et al. 1996) have demonstrated consistent increases in blood lactate levels in response to exhaustive exercise. Therefore, we cannot dismiss the possibility that anaerobiosis, not necessarily accompanied by reductions in blood oxygen level, constitutes a stimulus for the release of somatolactin.

The hypothesis that somatolactin release is triggered by the decrease in plasma [HCO₃⁻] led us to investigate the effects of somatolactin deficiency on acid–base balance, in particular on plasma [HCO₃⁻]. Plasma somatolactin level was decreased 3 min after infusion of anti-somatolactin serum and remained low until 120 min after acid infusion, thus confirming successful immunoneutralization of somatolactin in the antiserum-infused fish. The effect of somatolactin deficiency on acid–base regulation was obvious; plasma [HCO₃⁻] in the immunoneutralized fish was significantly lower than in the fish without immunoneutralization 30 and 120 min after acid infusion, and no significant recovery in plasma [HCO₃⁻] was seen in the immunoneutralized fish. The results also suggest the involvement of somatolactin in the regulation of plasma [HCO₃⁻] during acid–base regulation.

The mechanisms of acid–base compensation in fish are still controversial (for recent reviews, see Goss et al. 1995; Lin and Randall, 1995). What is evident is that fish manipulate rates of transepithelial ion transfer with the ambient water in order to correct acid–base disturbances (Heisler, 1986, 1993b). The gills are the main site responsible for the process, although the types and localization of ion exchangers in the branchial epithelium remain unknown. The determination of the molecular basis for the ionic exchange systems in the gills and the involvement of somatolactin in the regulation of these systems will require further investigation.

A hypercalcaemic action of somatolactin has been indicated in previous studies: somatolactin cells are activated by low environmental Ca²⁺ levels and plasma somatolactin levels are elevated in association with the increase in plasma Ca²⁺ levels in stressed and exercised fish (Kakizawa et al. 1993, 1995, 1996). In the present study, however, plasma Ca²⁺ levels in the immunoneutralized fish showed a tendency to increase in parallel with Ca²⁺ levels in the fish without immunoneutralization 5 min after acid infusion, when severe acidosis was observed. The results suggest that the elevation of plasma [Ca²⁺] during acidosis induced by acid infusion and exhaustive exercise in the present study and in response to stress and exhaustive exercise in previous studies (Kakizawa et al. 1995, 1996) may be the result of a dissociation of calcium phosphate in bone and/or scales caused by acidosis (Herrmann-Erlee and Flik, 1989) rather than of the action of somatolactin.

Decreases in plasma [HCO₃⁻] resulting from elevated plasma [H⁺] induced by acid infusion, exhaustive exercise or severe hypoxia caused an elevation of plasma somatolactin levels. Furthermore, the recovery from the decreased [HCO₃⁻] during acidosis was inhibited by somatolactin deficiency induced by immunoneutralization. Thus, somatolactin seems to be involved in the retentive regulation of plasma [HCO₃⁻], especially during acidosis, in rainbow trout.

We wish to thank Dr Mariann Rand-Weaver, Department of Biology, Brunel University, USA, for providing us with the anti-coho-salmon somatolactin serum. We are grateful to Professor Hiroshi Kawauchi, School of Fisheries, Kitasato University, Japan, for providing the purified chum salmon somatolactin. We are also grateful to Professor Howard A. Bern, University of California at Berkeley, USA, for his critical reading of the manuscript. We wish to thank Dr Teruo Azuma and other members of the National Research Institute of Aquaculture, Nikko Branch, for helping in the hypoxia experiment. We also thank the staff of Shizuoka Prefectural Fisheries Experimental Station for providing us with rainbow trout. This study was supported by a fellowship from the Japan Society for the Promotion of Science to S.K., and also by grants-in-aid from the Ministry of Education, Science and Culture and from the Fisheries Agency, Japan to H.T. and T.K.