In this study, we set out to examine the role of the somatotropic axis in the ion-regulation process in rainbow trout. Specifically, our objective was to examine whether plasma insulin-like growth factor-binding proteins (IGFBPs) are modulated by gradual salinity exposure. To this end, freshwater (FW)-adapted rainbow trout were subjected to gradual salinity increases, up to 66% seawater, over a period of 5 days. During this acclimation process, minimal elevations in plasma Ca2+ and Cl− were seen in the salinity-acclimated groups compared with FW controls. There were no changes in plasma Na+ levels, and only a minor transient change in plasma cortisol levels was seen with salinity exposure. The salinity challenged animals responded with elevations in plasma growth hormone (GH) and IGF-I levels and gill Na+-K+-ATPase activity. We identified IGFBPs of 21, 32, 42, and 50 kDa in size in the plasma of these animals, and they were consistently higher with salinity. Despite the overall increase in IGFBPs with salinity, transient changes in individual BPs over the 5-day period were noted in the FW and salinity-exposed fish. Specifically, the transient changes in plasma levels of the 21-, 42-, and 50-kDa IGFBPs were different between the FW and salinity groups, while the 32-kDa IGFBP showed a similar trend (increases with sampling time) in both groups. Considered together, the elevated plasma IGFBPs suggest a key role for these binding proteins in the regulation of IGF-I during salinity acclimation in salmonids.
- insulin-like growth factor binding-proteins
- insulin-like growth factor-I
- growth hormone
- rainbow trout
euryhaline teleosts, such as killifish, tilapia, and eels, survive abrupt full-strength seawater (SW) transfer without prior physiological preparation (14, 42, 92), whereas salmonids undergo a preacclimation process that is essential to SW survival (46). The preacclimation process involves physiological and morphological changes that improve salinity tolerance of developing/migrating salmonids (16, 31, 48). These physiological changes, such as salinity tolerance, are driven by changes in circulating hormones such as growth hormone (GH), prolactin (PRL), somatolactin (SL), insulin-like growth factor-I (IGF-I), and cortisol (8, 24, 45).
Efforts to separate the endocrine and physiological response to salinity exposure, from the acquisition of salinity tolerance by migrating salmonids, have been made difficult by species differences (48) and aspects of the preacclimation process such as developmental and sex-specific differences (8, 10, 30, 48, 50, 55), changes in temperature and photoperiod (5, 8, 9, 27), diet and rearing conditions (64, 97), growth, genetic lineage and domestication (24, 27, 57), and the method and duration used for salinity challenge (9, 72, 73, 75, 93).
Aside from the aforementioned factors that influence salinity tolerance and hormone levels in salmonids, there is good evidence to show that the preacclimation process, SW tolerance, and enhanced growth of SW-acclimated salmonids are controlled by the hormones of the somatotropic axis (GH and IGF-I) (4–6, 24, 64, 97). Consequently, this area of study has drawn a great deal of interest (for review, see 8, 17, 19, 45, 77). In addition to their growth-promoting actions, it is well established that GH and IGF-I stimulate SW tolerance in a number of euryhaline teleosts (45). However, the relationship between the growth-promoting and osmoregulatory actions of GH and IGF-I are not well understood. In this respect, while several studies in salmonids have shown good relationships between circulating GH and IGF-I and gill Na+-K+-ATPase activity (and growth) during either smoltification, or abrupt SW challenge (5, 8, 9, 45, 75, 94, 97), others failed to show such clear relationships following abrupt SW challenge or SW entry in salmonids (23, 57, 93, 96), including rainbow trout (72, 75).
To gain a clearer picture of the underlying endocrine and physiological adjustments inherent with salinity acclimation in a salmonid, we have used the nonmigratory rainbow trout (Oncorhynchus mykiss) wherein preparative changes that could influence or alter the acclimation process to higher salinity should not occur. Despite this advantage, differences in the endocrine response of rainbow trout, to abrupt salinity challenge, have been reported (72, 75). Specifically, Sakamoto and co-workers (75) and Sakamoto and Hirano (72) showed only transient increases in plasma GH at 2 and 1 days posttransfer to 75% and 80% SW, respectively. While plasma IGF-I levels were not measured in these studies on rainbow trout, many of the aforementioned studies have examined the relationship between plasma GH and IGF-I during the preacclimation process in Atlantic and Pacific salmon, as well as the effects of salinity challenge on plasma GH and liver IGF-I mRNA levels. However, given that IGF-I also exhibits changes during the preacclimation process in salmonids, changes in plasma IGF-I levels (and tissue mRNA) following salinity challenge, or SW rearing, are few and the results are difficult to interpret (9, 24, 50). As for reported increases in liver IGF-1 mRNA in teleosts, following salinity challenge or endocrine manipulation, evidence from recent studies show that hepatic IGF-I mRNA levels may not accurately reflect plasma IGF-I levels in euryhaline teleosts such as the tilapia (33, 69, 70) and rainbow trout (71). Consequently, the relationship between plasma GH, plasma IGF-I levels, and tissue IGF-I mRNA levels, following salinity challenge and/or endocrine manipulation, are not clearly established.
While recent studies have demonstrated the growth-promoting and osmoregulatory actions of IGF-I (45, 49), the biological activity of IGF-I is regulated by high-affinity plasma-binding proteins, termed insulin-like growth factor-binding proteins (IGFBPs) (35, 37). The IGFBPs, which possess IGF-independent biological actions in teleosts (2, 20), are themselves regulated by stress, nutritional, endocrine, and developmental status in salmonids (e.g., 6, 38, 39, 78, 79, 81). With the exception of a single report on the effects of brackish water exposure on plasma IGFBPs in the stenohaline channel catfish (32), there have been no reports on the effects of salinity challenge on plasma GH, IGF-I, and IGFBPs in an euryhaline teleost. This deficiency in our understanding of the roles of the hormones of the somatotropic axis (GH and IGF-I) in growth and osmoregulation, along with the IGFBPs, underscores the need for further studies in this area.
The goal of this study was to investigate the role of the salmonid somatotropic axis during the process of acclimation to higher environmental salinities. To this end, we gradually exposed freshwater-adapted (nonanadromous) rainbow trout to incremental increases in environmental salinity over a 5-day period to minimize the compounding effect of stress associated with abrupt salinity transfer. Our specific objective was to examine IGFBP profiles in conjunction with changes in plasma hormone (GH, IGF-I, and cortisol) levels and ionoregulatory disturbances following gradual exposures to higher salinities in rainbow trout.
MATERIALS AND METHODS
General experimental protocol.
Rainbow trout (Oncorhynchus mykiss; body mass ∼50 g) were purchased from Rainbow Springs Hatchery in Thamesford, Ontario, Canada. Fish were housed at the Wilfrid Laurier University aquatic facility according to guidelines set by the Canadian Council of Animal Care and were initially held for at least 1 wk in aerated well water before the start of the experiment. Trout were fed ∼5% of the groups total body weight every other day with 3.0 pt trout chow (Corey Aquaculture Trout and Charr grower; Corey Feed Mills, Fredericton, New Brunswick). Trout were acclimated for 1 wk to soft water, produced by a reverse osmosis system (Culligan, Series E Reverse Osmosis system), before SW exposure. All experiments, sample collection, and analyses were carried out in accordance with the approved animal care guidelines at the National Center for Cool and Cold Water Aquaculture/ARS/USDA and the Universities of Kentucky (Lexington, KY, USA) and Waterloo (Ontario, CA).
Groups of 50 fish each were randomly divided into two 90-liter polyethylene containers, and both groups were unfed throughout the acclimation period. Both containers were aerated and prepared as flow-through systems with flow rates of 150 ml/min each. In one of the two containers a freshwater control was prepared with the salinity kept at ∼1 mM Cl− by using a full-strength salt water drip (Kent Marine Salts, Acworth, GA). In the second container, a salt water stock solution (Kent Marine Salts) was delivered by peristaltic pump to gradually increase salinity to the desired level over 10 h. Trout were acclimated to 33% strength SW and sampled at day 1 [10 parts per thousand (ppt), 160 mM Cl−], after which the salinity was increased gradually to 50% SW and sampled at day 3 (16 ppt, 260 mM Cl−), followed by 66% SW (22 ppt, 360 mM Cl−) and sampled at day 5. Twelve fish (6 freshwater control and 6 SW acclimating fish) were sampled at 1, 3 and 5 days after the start of salinity acclimation.
Trout were randomly sampled one at a time with minimal disturbance (confirmed later by lack of correlation between removal time and plasma cortisol levels) and killed by an overdose of buffered MS-222 (84g/l methanesulfonate salt; Sigma), pH-adjusted using KOH. Blood (500–1,000 μl) was drawn by caudal puncture using an ice-cold, heparinized 1-ml luer lock syringe (Hamilton) and dispensed into a microcentrifuge tube and spun at 12,700 g (IEC Micro-MB Centrifuge) for 2 min. Plasma was removed and stored at −80°C. A portion of gill tissue was excised from the second gill arch from the right side of each fish, rinsed in ultrapure water, frozen in liquid nitrogen, and stored at −80°C until analysis.
Water and plasma ions.
Water temperature (14 ± 1.0°C) and pH (7.7 ± 0.3) were recorded at each sampling time throughout the experiment. At each sampling time water from each container was dispensed into two 7-ml scintillation vials. Acidified water and plasma samples were analyzed for Na+ and Ca2+ content by flame AAS (Varian Spectra AA-880). Plasma and nonacidified water samples were analyzed for Cl− content using a mercuric-thiocyanate method (Sigma-Aldrich kit 461-M) and read at 460 nm on a Spectronic 301 spectrophotometer (Milton Roy, Rochester, NY).
Determination of Blood/Plasma Parameters and Gill Na+-K+-ATPase activity.
For determination of blood hematocrit, ∼100 μl of sample was dispensed into two heparinized hematocrit tubes and spun for 5 min in a READACRIT centrifuge (Clay Adams, NJ). Hematocrit was measured by dividing the packed cell length by the total sample length in the tube and is expressed as a percentage. Hemoglobin was measured colorimetrically by a cyanmethemoglobin method with 20-μl aliquots of whole blood (Sigma kit 525-A). Plasma glucose was measured using the Trinder method (Sigma Chemical, St. Louis, MO; kit 16–20). Plasma protein was determined according to the method of Alexander and Ingram (1). Gill Na+-K+-ATPase activity was measured according to the method of McCormick (47).
Plasma hormone levels.
Plasma cortisol concentration was determined using commercially available radioimmunoassay (RIA) kit (ICN Biomedicals, CA) that has been validated for trout (90). Plasma IGF-I levels were determined according to the method of Moriyama (56) as modified by Shimizu et al. (81), using commercially available recombinant trout (rt) IGF-I and the anti-barramundi IGF-I antibody (GroPep, South Australia, Australia). Inter- and intra-assay coefficients of variation were 5% and 4%, respectively.
Assay measurement of plasma GH levels.
Plasma GH levels were measured using a modified (low volume), and further validated (described below), radioimmunoassay (RIA) as previously described by Peterson and co-workers (63). The RIA utilized commercially available recombinant salmon/trout GH (rtGH; lot DAB-GHB1) and anti-salmon/trout GH primary antibody (lot AJI-PAN1) (GroPep). Sources for hormones used to determine antibody specificity are as previously described by Drennon et al. (18).
The RIA was performed overnight using a double-antibody method (PanSorbin cells were used in lieu of a secondary antibody for pelleting) under disequilibrium conditions. Hormone standards (rtGH) and plasma, heterologous hormones, and pituitary homogenates (for assay validation) were serially diluted with assay buffer (PBS-BSA: 20 mM PBS pH 7.4, 0.05% Tween-20, 0.2 g/l protamine sulfate, 0.5% BSA). Standards were run in triplicate and plasma and pituitary samples (unknowns and plasma pools) in duplicate. Standards consisted of serially diluted hormone (30–0.12 ng/ml) in PBS-BSA, and plasma (15–25 μl) was diluted with PBS-BSA to a final volume of 50 μl. On day 1, 50 μl of primary antibody, diluted to 1:30,000 in EDTA-PBS-NRS (20 mM PBS, pH 7.4, 0.15 M NaCl, 0.05% Tween-20, 0.05% sodium azide, 0.2 g/L protamine sulfate, 10 mM EDTA, and 0.5% normal rabbit serum), was added to the standard, pools, and unknown tubes and vortexed. Immediately after vortexing, 50 μl of 30,0000 CPM of 125I-labeled rtGH (iodinated according to the manufacturer's protocol), diluted in PBS-BSA, was added to each tube, vortexed, and allowed to incubate overnight at room temperature. On the following day, antibody-bound rsGH was precipitated with the addition of 100 μl of PanSorbin cells (CalBiochem, San Diego, CA), diluted to 0.35% in EDTA-PBS, to each tube. Assay tubes were incubated for 5–6 h, with occasional vortexing (gentle). Following incubation, 100 μl of PBS-BSA was added to each tube, vortexed, and centrifuged for 30 min at 3,000 g and 4°C. Following centrifugation, the supernatant was aspirated, and tubes were counted on a gamma counter (Packard Instruments, Autogamma 5500, Meriden, CT).
The antibody specificity was confirmed using heterologous plasma, pituitary homogenate, and purified homologous and heterologous hormones. For the pituitary and plasma dilutions, tissues were used from taxonomically diverse teleosts as previously described (18). Multiple dose-response inhibition curves for recombinant salmon GH, using rtGH and other purified hormones, revealed no cross-reactivity with other hormones, particularly the forms of prolactin (tilapia, catfish, and ovine) tested (data not shown). Serial dilutions of rainbow trout pituitary homogenates (1:1 × 10−1 to 1:1 × 10−8) revealed a displacement curve that paralleled the rtGH standard curve, whereas pituitary homogenate from taxonomically primitive teleosts, such as carp, but not channel catfish, exhibited some parallelism to the rtGH standard curve, but only at the lower doses tested (data not shown). By contrast, pituitary homogenates from neoteleosts (hybrid striped bass and yellow perch) exhibited no parallelism to the rtGH standard curve (data not shown). Serially diluted rainbow trout plasma (1:1, 1:2, 1:4, 1:8, and 1:16) displayed an inhibition curve that was parallel to the rtGH standard curve (data not shown).
The intra-assay coefficient of variation (CV) was determined from multiple plasma pools from a single assay, whereas the interassay CV was determined from the ED50 value from standard curves and plasma pool values obtained from separate assays. For assessment of intra-assay CV, the mean pool (n = 5) value from five separate assays of plasma pool samples, within the same assay, gave a mean value of 5.7 ± 0.2 ng/ml and a CV of 7.0%. For assessment of the inter-assay CV, mean ED50 values from separate assay standard curves (n = 26) were determined to be 3.4 ± 0.3 ng/ml with a CV of 7.6%. Plasma pool values obtained from separate assays (n = 9) were also used to determine the inter-assay CV and showed a mean pool value of 5.7 ± 0.5 ng/ml and a CV of 8.4%. The detection limit of the assay, defined as the 90% binding, was 0.45 ± 0.14 (n = 26 assays).
Western ligand blotting procedure for plasma IGFBPs.
Western ligand blotting was accomplished using digoxigenin (DIG) labeled-recombinant barramundi (rb)IGF-I (GroPep) according to the method of Shimizu et al. (81) as modified in our laboratory by Johnson et al. (32). To confirm specific binding of the DIG-rbIGF-I probe, competitive reactions were carried out using unlabeled rbIGF-I as previously described (32).
Data were analyzed using two-way analysis of variance (ANOVA) with treatment (FW and salinity) and time (1, 3, and 5 days) as independent variables. For all other dependent variables (where indicated in the figure legend), FW values were pooled when there were no significant differences between FW time points (1, 3, and 5 days) and compared with days in salt water with one-way ANOVA. If the FW samples were not significant, they were left separate (see results). Where necessary, the data were log- or arcsine-transformed to meet assumptions of homogeneity of variance within treatment group although nontransformed values are shown in the tables and figures. Where significant differences occurred (P < 0.05) following ANOVA, comparisons between groups were made using Fisher's protected least significant difference (FPLSD) for pair-wise comparisons (86). Statistical analyses were completed using SPSS computer software (SPSS, Chicago, IL). Data are presented as means ± SE calculated within each treatment group unless indicated otherwise.
Gradual salinity exposure to ∼66% SW over 5 days did not significantly affect rainbow trout plasma Na+ concentrations (Table 1; one-way ANOVA, P > 0.05). Plasma Cl− concentrations were significantly higher at 3 and 5 days, and plasma Ca2+ concentrations were significantly higher at 1 and 5 days (Table 1; one-way ANOVA; P < 0.05). There were no fish mortalities in any of the treatments during the 5-day experimental period. At day 6, however, air flow into the SW tank had stopped for a period of 6 h; therefore, data from subsequent samplings (8 and 10 days in full-strength SW) are not included.
Plasma cortisol and other blood parameters.
There was no significant effect of either treatment (salinity exposure) or time (day postexposure) on plasma cortisol concentrations, but there was a significant (P < 0.05) interaction of treatment × time (Table 2). Specifically, cortisol levels in the salinity exposure groups were not different at day 1 but were significantly lower than FW controls at day 3; however, by day 5, cortisol levels were significantly (P < 0.05) higher in the 66% SW group compared with levels seen in the FW controls at this time (Table 2; two-way ANOVA, P < 0.05). Plasma glucose concentrations were slightly, but significantly (P < 0.05), lower in the salinity-exposed fish compared with the FW group regardless of sampling time (Table 2). Plasma glucose concentration also showed a significant time effect, declining at 3 and 5 days compared with 1-day sampling period. Blood hematocrit and hemoglobin levels were unaffected by salinity exposure or sampling time (Table 2). There was a significant effect of sampling time on plasma protein levels (Table 2: two-way ANOVA, P < 0.05), with elevated levels occurring in the FW 1-day group, compared with all other groups; however, there was no significant effect of salinity on plasma protein levels (Table 2).
Gill Na+-K+-ATPase activity.
Gill Na+-K+-ATPase activity was significantly affected by salinity (FW vs. saline water) and sampling time (increasing salinity/time) (Fig. 1A; two-way ANOVA, P < 0.001). There was no significant (P > 0.05, two-way ANOVA) salinity (FW vs. saline water) × sampling time (increasing salinity/time) interaction. Salinity exposure resulted in significantly (P < 0.001, FPLSD) higher gill Na+-K+-ATPase activity compared with the FW group (Fig. 1A). There was also a significant time effect, with gill Na+-K+-ATPase activity being significantly (P < 0.05, FPLSD) lower at 5 days compared with the 1- and 3-day time points, regardless of treatment.
Plasma GH and IGF-I levels.
Plasma GH levels were significantly (P = 0.03, two-way ANOVA) altered by environmental salinity, with the highest levels occurring in the SW groups (Fig. 1B). There were no significant effects of sampling time or a significant interaction of salinity (FW or saline water) × sampling time (increasing salinity/time) on plasma GH levels (Fig. 1B). Plasma GH levels were significantly (P < 0.05, FPLSD) elevated in the salinity-challenged groups at days 1 (33% SW) and 5 (66% SW) compared with the respective FW control groups. Plasma GH levels, within the FW controls, did not significantly differ according to sampling time; however, in the SW groups, plasma GH levels were significantly (P < 0.05, FPLSD) higher at day 1 (33% SW) compared with day 3 (50% SW) and day 5 (66% SW).
Plasma IGF-I levels were significantly (P < 0.001, two-way ANOVA) affected by salinity, with higher levels occurring in the salinity-challenged groups (Fig. 1C). There were no significant effects of sampling time or a salinity (FW or saline water) × sampling time (increasing salinity/time) interaction on plasma IGF-1 levels (Fig. 1C). Plasma IGF-I levels were significantly (P < 0.05, FPLSD) elevated in salinity-challenged groups sampled at days 3 (50% SW) and 5 (66% SW) compared with their respective FW control groups. Plasma IGF-I levels, within the FW controls, were not significantly different over the sampling time course. By contrast, mean levels of plasma IGF-I progressively increased over the sampling time course within the SW groups, but these increases were not significant (P > 0.05, FPLSD).
Four specific IGFBPs (21, 32, 42 and 50 kDa) were identified in the plasma of rainbow trout from this study (Figs. 2 and 3). In this study, the higher molecular weight (MW) IGFBPs (42 and 50 kDa) routinely appeared as a doublet (Fig. 2). To determine whether DIG-labeled rbIGF-I binding was specific, identical membranes were probed with varying amounts (10- to 50-fold excess) of unlabeled (competitor) rbIGF-I. Addition of unlabeled rbIGF-I significantly (P < 0.05, one-way ANOVA followed by FPLSD) reduced or eliminated the appearance of the all of the IGFBP bands. When using a 10-fold excess of unlabeled rbIGF-I, binding of the DIG-rbIGF-I was significantly reduced for all of the IGFBPs except the IGFBP-42K which displayed only a 1.3-fold reduction. When using a 50-fold excess of unlabeled rbIGF-I, binding of the DIG-rbIGF-I was significantly (P < 0.05) reduced for all of the IGFBPs (data not shown). When no ligand (labeled or unlabeled IGF-I) was present, band intensities were absent for the IGFBP-21 and −32kDa bands and significantly (P < 0.05, one-way ANOVA followed by FPLSD) reduced for the IGFBP-42K and IGFBP-50K proteins.
Plasma levels of the 21-kDa IGFBP (IGFBP-21K) were significantly altered by salinity (FW vs. saline water) (P < 0.001, two-way ANOVA) and sampling time (increasing salinity/time) (P < 0.001, two-way ANOVA) (Fig. 3A). There was also a significant (P < 0.001, two-way ANOVA) interaction of salinity × sampling time on plasma IGFBP-21K levels. Plasma IGFBP-21K levels were significantly (P < 0.001, two-way ANOVA) elevated in the salinity exposed groups compared with the FW controls (Fig. 3A). Levels of the IGFBP-21K differed significantly (P < 0.001, two-way ANOVA) according to sampling time, with mean levels decreasing over the sampling period, with exception of an increase in IGFBP-21K levels in the day 5 SW (66% SW) group (FPLSD). With regard to the salinity × sampling time interaction, plasma levels of the IGFBP-21K in the FW groups were significantly (P < 0.05, FPLSD) higher at day 1 with mean levels decreasing through day 5 (Fig. 3A). For the SW groups, levels of the IGFBP-21K were significantly (P < 0.05, FPLSD) higher at days 1 and 5 compared with day 3. Furthermore, levels of the IGFBP-21K, in the 66% SW group (day 5), were significantly (P < 0.05, FPLSD) elevated over the respective FW control group sampled at day 5 (Fig. 3A).
Plasma levels of the 32-kDa IGFBP (IGFBP-32K) were significantly altered by salinity (P < 0.01) and sampling time (P < 0.01), with higher levels occurring in the SW groups (Fig. 3B; two-way ANOVA). There was no significant (P > 0.05, two-way ANOVA) salinity × sampling time interaction on plasma IGFBP-32K levels. For the FW groups, mean plasma levels of the IGFBP-32K increased over time with significantly (P < 0.05, FPLSD) higher levels at day 5 vs. days 1 and 3 (Fig. 3B). For the SW groups, plasma IGFBP-32K levels were significantly (P < 0.05, FPLSD) elevated at days 3 (50% SW) and 5 (66% SW) compared with day 1 (33% SW).
Levels of the 42-kDa IGFPB (IGFBP-42K) were significantly (P < 0.001, two-way ANOVA) altered by salinity, with the highest levels occurring in the SW groups at all sampling times, compared with the respective FW control groups (Fig. 3C). There was no significant effect of sampling time nor was there a significant salinity × sampling time interaction on plasma IGFBP-42K levels. For the FW groups, mean levels decreased over time, with the highest mean levels occurring in the groups sampled at days 1 and 3; however, this decrease was only significant (P < 0.05, FPLSD) for the groups sampled at day 1 and day 5. For the SW groups, plasma IGFBP-42K levels were similar across all sampling times; however, all the SW groups were significantly (P < 0.05, FPLSD) higher than the FW controls (Fig. 3C).
Levels of the 50-kDa IGFBP (IGFBP-50K) were significantly (P < 0.001, two-way ANOVA) altered by salinity, with the highest levels occurring in the SW groups at all sampling times, compared with the respective FW control groups (Fig. 3D). There was no significant (P > 0.05, two-way ANOVA) salinity × sampling time interaction on plasma IGFBP-50K levels. For the FW groups, plasma IGFBP-50K levels were not significantly (P > 0.05, FPLSD) different between the sampling times. For the SW groups, plasma IGFBP-50K levels were significantly (P < 0.05, FPLSD) elevated at day 3 compared with those animals sampled at days 1 and 5 (Fig. 3D).
We demonstrate for the first time that plasma IGFBPs are modulated by gradual salinity exposure in rainbow trout. The involvement of GH and IGF-I in salmonid smoltification, and SW tolerance, appear to be unequivocal (8, 11, 22, 44, 45, 51, 55, 56, 71, 74); however, very little is known about the relationships between IGF-I and the IGFBPs in these processes and others. While some studies have begun to address this problem (2, 18, 78, 80), no studies have examined the relationships between salinity exposure and the attendant changes in plasma GH, IGF-I, and the IGFBPs in a teleost. Herein, we report that gradual increases in environmental salinity provoke an activation of the hormones (GH and IGF-I) of the somatotropic axis, with a corresponding increase in plasma IGFBPs (important regulators of IGF-I action and part of the somatotropic axis). The increase in plasma IGFBPs suggests a key role for these binding proteins in the function of IGF-I during salinity acclimation in salmonids.
Gradual salinity increases did not appreciably disturb the plasma ion homeostasis in rainbow trout over a 5-day period. Plasma Na+ levels remained unchanged, while the minimal rise in plasma Cl− concentration were well below the levels seen in rainbow trout following transfer to 67–100% SW (11, 66, 72, 75), as well as below the 160 mM threshold associated with ionoregulatory dysfunction in salmonids (88). The efficient ion regulation seen here is similar to that observed in euryhaline species transferred from FW to SW (15, 40, 84, 91) but unlike that reported in rainbow trout, which displayed elevated plasma chloride, sodium, and cortisol for several days following abrupt exposure to SW (67–100%) (11, 66, 71, 72, 75). The differences between our study and the aforementioned studies, where plasma ions were elevated, may be due to the stress (indicated by high cortisol and ion levels) associated with abrupt transfer to higher salinities, whereas the gradual acclimation approach did not result in severe disturbances in plasma cortisol or chloride levels (Tables 1 and 2). This notion of reduced stress in the animals used in this study is further confirmed by a lack of a glucose response in the groups, which generally follows cortisol stimulation in fish (52).
There were no mortalities, with salinity exposure, with the trout used in this study over the 5-day experimental period. The initial exposure to 33% SW, at day 1, may be playing an important role by priming the animal's ion-extrusion mechanisms to allow effective hypoosmoregulation at even higher salinities. This notion finds support from the significantly higher sodium pump activation at 33% SW (day 1) (Fig. 1A), implying an upregulation of the ion extrusion mechanisms even at a salinity that results only in very small ionic gradient between the animal's extracellular fluid and environment. The minimal changes in plasma parameters, including hematocrit, hemoglobin, and protein, indicate that there were minimal disturbances in plasma volume. Altogether, gradual exposure to incremental salinity changes did not appreciably affect plasma ion homeostasis in the present study.
In this study, we observed significant increases in plasma GH in trout that were exposed to gradual elevations in environmental salinity, compared with the FW control groups (Fig. 1B). This finding is in close agreement with findings in salmonid (72, 73, 75) and nonsalmonid teleosts (41). The hypoosmoregulatory capacity, including maintenance of higher gill Na+-K+-ATPase activity in the salinity-exposed trout over the 5-day period (Fig. 1, A–C), coincided with elevated plasma GH and IGF-I levels, further supporting a key role for the somatotropic axis in ion regulation in rainbow trout (41, 45). Even though there was an overall significant effect of salinity on plasma GH levels, there was no progressive increase with the transition to higher salinities. This result is not unexpected as previous studies in salmonids, which have been exposed to gradual increases in salinity (see Ref. 9), or abrupt salinity challenge (30, 61, 75, 93), have also reported discordant observations in plasma GH levels.
Plasma IGF-I levels, on the other hand, showed an increasing trend over time with the transition to higher salinities. The basis for the differences in the patterns of plasma GH and IGF-I, to gradual salinity acclimation, is unknown, but may be related to the sampling time course or to other unknown factors. Specifically, studies that have involved intraperitoneal injection or salinity challenge to elevate plasma GH levels in rainbow trout showed that elevations in circulating GH levels are transient and persist for only 12 to 72 h following provocative testing (54, 71, 72, 75, 76). This may involve increased plasma clearance rate of GH due to salinity-induced increases in the specific binding and occupancy of the rainbow trout liver GH receptor (72, 75). Moreover, elevated levels of plasma IGF-I, or liver mRNA, have been shown to lag peak GH levels (exogenous treatment) by 3–36 h, depending upon dose and route of administration (7, 53, 54, 60, 76). By contrast, Sakamoto and co-workers (71) did not see an increase in liver IGF-I mRNA levels, despite elevated GH levels in rainbow trout 24 h posttransfer to 80% SW. Consequently, it is not surprising that IGF-I levels increased in the 50% and 66% SW groups without a corresponding increase in plasma GH levels. The decreasing plasma IGF-I levels in the FW group may also be due to nutritional status, especially since the animals were food-deprived for 1 (33% SW), 3 (50% SW), and 5 days (66% SW) before sampling. Fasting appears to more rapidly attenuate IGF-I levels (plasma or tissue mRNA) than GH levels in teleosts and mammals (19, 87); however, the opposite has been reported in coho salmon (21). It is interesting, however, that fasting does attenuate the GH response to salinity challenge in tilapia (91) and in salmon (89). Such a phenomenon may also explain the lower mean GH levels seen in animals from this study held in 50% (3-days fasting) and 66% (5 days fasting) SW.
Of equal interest are the mild changes in plasma cortisol levels, which suggest that gradual acclimation to higher salinities may not elicit a strong stress response that typically follows abrupt SW transfer. Despite methodological differences between these two approaches (gradual vs. abrupt SW transfer), studies (cited herein) using abrupt SW challenges to assess the hypoosmoregulatory capacity of salmonids have either reported transient increases in plasma cortisol and GH, inconsistent changes, or no increases at all. The reason for the differences in the endocrine response of migratory and nonmigratory salmonids is unclear, but one possible factor may be that the majority of salinity challenge studies have been conducted on migratory salmon, during the parr-smolt stage (8), where changes in hormone levels (GH, IGF-I, cortisol, and thyroid hormones, etc.) are already occurring as part of the preparatory (metabolic, morphological, and behavioral) changes needed to complete their life cycle (8, 9, 27, 58, 94, 95). The ability to fully understand how and when the somatotropic axis, and its components, function solely to regulate growth or SW adaptation (or both?) will require further study and intense focus. In any case, it is more evident that salinity challenge is increasingly associated with an upregulation of a number of components (GH, IGF-I, and IGFBPs) of the somatotropic axis, whereas the plasma cortisol response appears inconsistent.
Role of IGFBPs.
Circulating IGFs, and those found in the extracellular fluids, are bound to high-affinity plasma proteins (IGFBPs) (67, 68). We have identified four IGFBPs of 21, 32, 42 and 50 kDa (Figs. 2–3) using a sensitive Western ligand blotting procedure (32, 81). Our findings agree with previous work in rainbow trout where IGFBPs of 23, 34, 41.5, and 55 kDa were detected in plasma using a combination of acid gel filtration chromatography and Western ligand blotting (59). In general, IGFBPs that are <32 kDa have been associated with stress or catabolic states (including cortisol treatment) in teleosts (2, 20, 33, 35, 43, 62, 79, 81, 82), whereas the higher molecular weight IGFBPs (>32 kDa) are generally associated with anabolic processes such as positive nutritional state, GH treatment (34, 37, 38, 78, 79, 81–83) or the season-dependent smoltification (upregulation of the somatotropic axis) process seen in salmonids (6, 56, 78, 81).
Classification of these IGFBPs in teleosts is based upon similar molecular weights seen in mammals and similar responses to endocrine and nutritional manipulation (35) and, more recently, the biomolecular characterization of the mammalian IGFBP-1 (25–30 kDa) and IGFBP-2 (31 kDa) homologues in rainbow trout and zebrafish (2, 20, 43) and an IGFBP-3 (41 kDa) homologue in chinook salmon (80). The IGFBPs not only modulate the IGF (-I & -II) activity but also possess IGF-independent actions (2, 13, 20, 36). Despite recent advances, the biological roles of IGFBPs in teleost fish are largely unknown, and even less is known regarding their involvement in osmoregulation and during salinity exposure (18).
We show, for the first time, that IGFBPs are upregulated by salinity exposure in rainbow trout. The lack of an effect of salinity on plasma total protein concentration rules out hemoconcentration as a possible reason for the higher levels of plasma IGFBPs in the salinity groups. The 21-kDa IGFBP (IGFBP-21K) decreased in intensity over the sampling intervals in the FW controls but increased in the salinity-challenged groups. The reason for the decrease in the FW group is not clear especially since plasma cortisol levels did not show any consistent changes with time. However, IGFBP-21K showed a transient profile in FW similar to that of plasma IGF-I levels, suggesting that the nutritional state of the animal may also be a key modulator of this binding protein. Given that the low molecular weight IGFBPs are upregulated during stress, catabolic states and/or cortisol treatment (35), the decrease in the FW control animals in this study could also be interpreted as a progressive decrease in the initial handling stress experienced by this group; however, this was not reflected in plasma cortisol changes over the sampling time period.
The concurrent increases in levels of plasma IGFBP-21K, with that of GH and IGF-I, are not characteristic with the regulation of the putative zebrafish (26 kDa, predicted from cDNA sequence) (43) and rainbow trout (30 kDa) (2) homologues to mammalian IGFBP-1, or the putative zebrafish homologue (31 kDa) to mammalian IGFBP-2 (20). Despite this dissimilarity, reports of low molecular weight IGFBPs, that were present/increased following GH treatment or fasting, have been made in coho salmon (22 kDa) and striped bass (28–30 kDa) (25, 79). Indeed, salinity also modulates plasma IGFBP-21K in rainbow trout, and the concurrent increases in plasma GH and IGF-I levels point to a role for this protein in the regulation of plasma IGF-I during salinity acclimation; however, much more work is needed before an identity to this lower molecular weight IGFBP can be ascribed.
Additional IGFBPs of 32, 42, and 50 kDa, that appeared to correlate with GH/IGF-I status, were also identified in the rainbow trout from this study. Levels of the 32-kDa IGFBP (IGFBP-32K) increased (coincident with the increase in GH) over the sampling time course in the FW and SW groups, with overall levels being significantly higher in the SW groups. IGFBPs of this size have been associated with physiological stress in teleosts (20, 26, 33, 35, 43, 78). Recently, Bauchat and colleagues (2) isolated and characterized a 30-kDa IGFBP in rainbow trout that is structurally and biologically similar to that of mammalian IGFBP-1. Consequently, it is possible that the IGFBP-32K protein seen in this study is the same protein, based on molecular size alone. However, given that levels of this IGFBP were significantly elevated with an upregulation of the somatotropic axis (GH and IGF-I), in the SW groups, this could suggest that the IGFBP-32K protein is either a truncated form of IGFBP-3 (the 42- and 50-kDa IGFBPs seen in this study) or a nonglycosylated form of IGFBP-3. Recently, Shimizu and co-workers (80) treated purified chinook salmon IGFBP-3 (40- and 43-kDa doublet proteins) with glycopeptidase F and reported an approximate 6.0-kDa shift in the deglycosylated proteins to 34 and 37 kDa, respectively (M. Shimizu, personal communication). If we were to predict that the IGFBP-42 and −50 kDa proteins from trout in this study would be similarly deglycosylated, this change in molecular size would not account for the difference in mass between the 42- and 50-kDa IGFBP doublet (IGFBP-3) and the 32-kDa IGFBP. Consequently, the most likely explanation would seem to be that the IGFBP-32K protein is the IGFBP-1 (30 kDa) characterized in rainbow trout (2) and that there is an, as yet, unidentified regulatory mechanism responsible for the salinity-induced increases seen in this study.
Plasma levels of the 42-kDa (IGFBP-42K) and 50-kDa (IGFBP-50K) binding proteins showed no consistent time-dependent changes over the sampling interval, but IGFBP-42K and -50K levels were significantly elevated in all of the SW groups, compared with the FW control groups. The higher molecular weight IGFBPs (42 and 50 kDa) consistently appeared as a doublet (Fig. 2). The appearance of a higher molecular weight (41 and 43 kDa) doublet has been reported in chinook salmon and is thought to represent the same protein with varying states of N-glycosylation and is the putative salmonid IGFBP-3 homologue(s) to mammalian IGFBP-3 (80). Concurrent increases in plasma IGFBP-42 and -50K, with that of plasma IGF-I (GH-dependent) levels, agrees well with previous work in salmonids (6, 38, 54, 78, 80) and other teleosts (37) where increases in the high molecular weight IGFBPs (40–50 kDa) were associated with GH treatment, higher growth rate, and plasma IGF-I (6). The associated increases in the IGFBP-42K and -50K doublet in rainbow trout, following salinity-induced increases in plasma GH and IGF-I levels, is good (independent) indication that these IGFBPs are homologous to mammalian IGFBP-3 and points to their potential involvement in metabolic and hypoosmoregulatory processes.
In conclusion, we have demonstrated a positive relationship between environmental salinity, increases in plasma GH and IGF-I levels, and the attendant increases in plasma IGFBPs in the rainbow trout. We have identified four IGFBPs that are regulated by salinity, but the physiological significance of the increases in IGFBPs, during salinity acclimation, remains to be elucidated. The efficient ion regulation, coupled with the elevated plasma GH and IGF-I levels, underscores a key role for the IGFBPs in the somatotropic axis modulation of hypoosmoregulatory ability in trout during salinity acclimation.
Osmoregulation is an ancient and ubiquitous physiological process that is essential to life (85), and teleosts are ideal vertebrate models for studies on the endocrine control of osmoregulation (12, 28, 29, 45, 65, 74). Our findings in rainbow trout appear to be the only study to demonstrate a positive relationship between environmental salinity, the attendant increases in GH and IGF-I levels, and the higher molecular weight IGFBPs in a euryhaline teleost; however, recent work conducted in the stenohaline channel catfish (I. punctatus) shows that these higher molecular weight IGFBPs (35, 44 and 47 kDa) are inversely related to GH status (18, 32). The significance and basis for these differences is unclear but could be linked to changes in plasma osmolality (ion content), osmoregulatory performance, or species differences (3). Other considerations would include the prior history of the animal, as well as the cortisol response of the animal elicited by salinity challenge. To our knowledge, these findings are among the first to illustrate taxonomic differences (see Ref. 32) between environmental salinity and IGFBP regulation in a euryhaline teleost (rainbow trout) and a stenohaline teleost (channel catfish). Obviously, as our understanding of teleost endocrine and osmoregulatory physiology develops, these sorts of evolutionary questions and physiological differences can be addressed.
This study was funded by grants from the National Research Initiative Competitive Grants Program/USDA award no. 2002-35206-11629 and 2004-05124, the USDA/ARS, and the support of the U.S. Geological Survey and Kentucky Water Resources Research Institute Grant Agreement No. 01HQGR0133 to B. S. Shepherd, the Natural Sciences and Engineering Research Council of Canada (NSERC) Industrial Oriented Research Program grant with Kodak Canada to R. C. Playle, and by an NSERC Discovery grant and Premier's Research Excellence Award (PREA) to M. M. Vijayan.
We thank Dr. G. Weber for technical assistance. We also gratefully acknowledge materials (ovine PRL and IGF-I) provided by Dr. A. F. Parlow of the NIDDK, National Hormone and Pituitary Program. We thank S. Raptis, University of Waterloo, for excellent technical assistance.
The views contained in this document are those of the authors and should not be interpreted as necessarily representing the official policies, either expressed or implied, of the U.S. Government. This manuscript is submitted for publication with the understanding that the United States Government is authorized to reproduce and distribute reprints for governmental purposes.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 the American Physiological Society