PDF hosted at the Radboud Repository of the Radboud University Nijmegen

Transcription

1 PDF hosted at the Radboud Repository of the Radboud University Nijmegen The following full text is a publisher's version. For additional information about this publication click this link. Please be advised that this information was generated on and may be subject to change.

2 .ENERAL AND COMPARATIVE ENDOCRINOLOGY 96, (1994) A d re n o c o rtic o tro p ic H o rm o n e in R elatio n to In terren al F u n c tio n d u rin g S tre s s in T ilapia (O r e o c h r o m is m o s s a m b ic u s ) P. H. M. B a l m, 1 P. P e p e l s, S. H e l f r i c h, M. L. M. H o v e n s, a n d S. E. W e n d e l a a r B o n g a Department o f Animal Physiology, Faculty o f Science, University of Nijmegen, Toernooiveld ED Nijmegen, The Netherlands Accepted July 15, 1994 This study examines ACTH-like immunoreactivity in the pituitary pars distalis and pars intermedia of the freshwater teleost Oreochromis mossambicus (tilapia). Two products (tactha and tacthb) were present in both lobes. These two products also accounted for the majority of the ACTH i.r. when in vitro pars distalis incubation medium was analyzed by HPLC. In a homologous bioassay the two tilapia ACTH-like molecules and human ACTH,,39 possessed similar corticotropic potency. The peptides were quantified using a newly validated radioimmunoassay, which was also used to measure ACTH in plasma of unstressed and stressed fish. Short-term (<12 min) stress rapidly increased plasma cortisol, reaching levels of around 3 ng/ml in 5 min. Surprisingly, this initial elevation was not accompanied by a rise in plasma ACTH levels. A more prolonged (3 hr) confinement in pairs resulted in high plasma cortisol and ACTH levels in one fish of every pair. The second fish had control ACTH levels and only marginally elevated cortisol levels. Therefore, in this species social interactions seem to influence the reaction of the pituitary-interrenal axis to stress. The short-term cortisol response to disturbance could be abolished completely by pretreating fish in vivo with cortisol for 48 hr. This treatment did not alter circulating ACTH levels. It is concluded that tilapia did not rely on circulating ACTH for a rapid elevation of plasma cortisol levels. Both neuronal mechanisms and cortisol feedback may regulate the pituitary-interrenal axis at the level of the interrenal Academic Press, inc. As in other vertebrates, in teleosts a tressor initially stimulates adrenal cortical.ctivity (Donaldson, 1981). Consequently, levated plasma corticosteroid levels are >ften considered an operational definition >f stress. Cortisol, the main teleost corticoteroid produced by the interrenal cells loated in the headkidney (Sangalang et al., 972; Balm et al., 1989, for tilapia), exerts a ariety of functions, including regulation of nergy metabolism, ionic homeostasis, and mmune competence (Donaldson, 1981; vlaule et al., 1987; Barton and Iwama, 991). Although there is no comprehensive nodel integrating these functions, most auhors favor an adaptive function of cortisol luring stress, in particular during the initial 1To whom correspondence should be addressed. phase (Munck and Naray-Fejes-Toth, 1992). Prolonged elevations of plasma cortisol, however, will be one of the causative factors leading to the deleterious effects of chronic stress (Barton and Iwama, 1991). This illustrates the careful balance fish have to maintain in regulating the activity of interrenal cortisol release. The stress response in fish has been studied in a variety of species, and recently strains have been selected on the basis of the cortisol response (Pottinger et al., 1992; Fevolden and Red, 1993). Interestingly, some of these strains also differed in their immune characteristics (Fevolden and Red, 1993). It has been reported that turkeys selected for low plasma corticosterone response to stress exhibited lower natural mortality and higher growth rate and egg production than /94 $5. Copyright 1994 by Academic Press, Inc. All rights of reproduction in any form reserved.

3 348 BALM ET AL material in the pituitary [neurointermediate lobe (NIL) and pars distalis (PD)] and in incubation medium by HPLC and subsequent bioassay. To investigate the role of ACTH in stress-related corticosteroidogenesis in this species, two stress protocols were tested: capture by netting (minutes) and capture + subsequent confinement in pairs (3 hr). Under these conditions, plasma cortisol, ACTH, and glucose were measured. Finally, aspects of the regulation of the corticotropes, including cortisol feedback, were investigated in vivo and in vitro. high responding counterparts (Brown and Nestor, 1973). These results indicate the fundamental impact of the stress-related rise in corticosteroid output. Although evidence is accumulating that the regulation of the interrenal cells in fish is more complex than previously assumed (Donaldson, 1981), ACTH is still considered the major corticotropic factor, basically because of its potency (Ranee and Baker, 1981). ACTH is the main splicing product of the prohormone POMC in the pituitary corticotropes, both in fish (Rodrigues and Sumpter, 1983) and in higher vertebrates (Jones and Gillham, 1988). Despite the many studies on the effects of stress on cortisol output in teleosts, relatively little is known about the effects of stress on the pituitary corticotropes; this reflects a lack of sufficiently specific and/or sensitive assays for fish ACTH, in particular for nonsalmonids. Nevertheless, information on ACTH, in addition to cortisol, would further increase insight into the regulation of the stress response in fish. The cortisol response to stressors is finely tuned by many mechanisms operating at the levels of the hypothalamus, the pituitary, and interrenal. To date, evidence on these mechanisms in fish largely stems from in vitro data (Fryer et al., 1984; Bradford et al.t 1992). The simultaneous measurement of plasma ACTH and plasma cortisol in experimental animals will yield valuable information on the relevance of these regulations in vivo. Furthermore, in salmonids (Sumpter et al., 1986) a rise in ACTH output precedes the stress-related cortisol response. Therefore, in some cases ACTH might be a more sensitive or even a more precise index for stress than cortisol. An ACTH RIA for tilapia (Oreochromis mossambicus) has been developed and validated using an antiserum previously demonstrated to detect ACTH in a variety of vertebrates (Dores et al., 1993; R. M. Dores, personal communication). This assay was used to characterize ACTH-related MATERIALS AND METHODS Mature male tilapia (O. mossambicus) from our laboratory stock were kept in groups of nine in 12-liter all-glass aquaria filled with artificial freshwater under a 12L/12D light regime. The fish, weighing 18 ± 2 g (n = 25), were fed tropical fish food (Tetramin). The response to a short-term handling was investigated by netting fish one by one from aquaria at regular intervals. This procedure took 12 min, during which seven fish were sampled. Nine aquaria were sampled this way, which yielded nine fish captured as firsts, nine fish captured as seconds, etc. Immediately upon removal from the aquaria, 2 p.1 blood was collected from the caudal vessels in EDTA/aprotinin; 1.5 mg Na:EDTA/3 KIU aprotinin (Serva GmbH, Heidelberg, Germany). No anesthesia was applied; the fish were killed by spinal transection. Plasma was obtained by centrifugation (3 min, 1,#), aliquoted, and stored at -2. In each aquarium, the two remaining fish were confined in a small net for 3 hr and then sampled as described. To study the effect of steroid feedback on the capture response and in vitro cortisol release, fish were pretreated for 48 hr with cortisol delivered via the food. Briefly, Tetramin flakes were sprayed with cortisol dissolved in ethanol, which was evaporated. Control food was treated with ethanol only. Eight groups of fish in = 7 per aquarium) were either given cortisol-containing meals (four groups) or given control meals (four groups) at / = -4 8 hr and t = -2 4 hr prior to sampling. Each meal was \% of total body weight; total dose per 48 hr was 2 mg cortisol/kg body wt. Sampling took 9 min per aquarium. Assays Cortisol was measured by radioimmunoassay using a commercial antiserum (Steranti Res. Ltd., UK). The procedure of the RIA using this antiserum (validated previously for the measurement of cortisol in fish plasma; Pickering et al., 1987) was slightly modified.

4 ACTH AND STRESS IN TILAPIA 349 briefly, 5 [l\ standards or unknowns were incubated >vernight with 1 fxl tritiated cortisol (4 dpm; Am- rsham Nederland BV, s Hertogenbosch, The Neth- rlands) and 5 ^1 antiserum (diluted to yield one-third >f the titer recommended by the suppliers). All continents were in phosphate buffer- (.1 M, Sigma -4417; Sigma Chemie, Bornem, Belgium) containing.1% (w/v) BSA (Fluka 5475, Fluka Chemie AG, iuchs, Switzerland). Free and bound hormone were jparated by adding 1 ml of ice-cold dextran-coated harcoal to each tube containing 2 mg Norit A (Fisher Scientific Company, Fair Lawn, New Jersey) and.2 ng dextran (7 kda, Fluka 3139, Fluka Chemie AG). The tubes were centrifuged at 4 (16#, 1 min in a ouan CR 412), 8 fxl supernatant was mixed with 4 nl Aqualuma scintillation fluid, and counted in a Phar- nacia Wallac 141 liquid scintillation counter. The ensitivity of this assay was 1.6 pg/tube, and 25 pg ortisol half maximally displaced the labeled hormone rom the antibody. The intra- and interassay coeffi- ients of variation were 7.9 and 8.5%, respectively, he steroid was measured in unextracted samples; outinely, tilapia plasma samples were diluted 1 times n assay buffer. Tilapia ACTH was measured by RIA using an tntiserum raised against ACTH,.24 kindly donated by rofessor R. M. Dores (see Dores et al., 1993). Rou- inely, 25 \i.\ standards (NIBSS, London, UK; iacth,.39) or unknowns were incubated at 4 with 5 il antiserum (1:7) and 1 jjli assay buffer (63 mm Ma2H P 4, 13 mm Na2EDTA,.2% (w/v) NaN3) con- aining 25, KIU aprotinin (Bayer), 1 fxl Triton <-1 (Sigma), and 25 mg BSA ORH D 2/21 Hoechst) per 1 ml. After 72 hr, 45 cpm iodinated iacth,_39 (2 Ci/mmol; Amersham Nederland BV) n 1 fxl assay buffer was added to the tubes, which vere incubated for another 24 hr at 4. Second antibody (1 fxl; SAR diluted 1 times in a.1% (w/v) abbit IgG Sigma I 56 solution) was then added o the tubes, which were incubated 2 min at 21. The econd antibody was donated by Professor Benraad, )epartment of Clinical Endocrinology, University of Nijmegen, The Netherlands. Finally, 1 ml of an ice- old PEG solution (7.5% w/v PEG 6; Merck, Darm- tadt, Germany) in assay buffer (from which aprotinin nd BSA had been omitted) was added, and the tubes vere centrifuged (1 min, 15#, Jouan CR 412). The upernatants were decanted, and the pellets were ounted in a LKB Wallac Clinigamma. Pituitary and PD homogenates were prepared in all- :lass homogenizers in.1 N HC1, centrifuged (3 nin, 1,#), and the supernatants were diluted in ssay buffer. For the displacement curves only, RIA ample volume was 5 jjli. a-msh immunoreactivity was measured by RIA fol- owing van Zoest et al. (1989). Plasma glucose was neasured in deproteinized samples using a commer- ial kit (Boehringer-Mannheim, Germany). HPLC Reverse-phase high-performance liquid chromatography was performed as described by Martens et al. (1982). To the fractions collected, 1 xg BSA was added, and the samples were dried in a Speedvac (Savant). Peptides were dissolved in.1 N HCl/MeOH (1:1), and ACTH and a-msh immunoreactivities were measured by RIA. Superfusion Superfusions of headkidney or pituitary rostral pars distalis tissues were performed as described previously (Balm, 1986) with the exception of the flow rate during the headkidney incubations (75 fil/min). Headkidney tissue was pulsed with dried (Speedvac) HPLC fractions or hacth,_39 (Peninsula Laboratories, England) dissolved in superfusion medium. Pars distalis tissue was challenged for 1 min with a hypothalamic extract (HE; 2 hypothalami prepared in.1 N HC1, and subsequently diluted 1 times in superfusion medium), ovine CRH (1 min,.5 nm; Bachem, Buben- dorf, Switzerland), or a high K + pulse (5 min, 5 mm). The maximally stimulated release rates were tested against the effect of a saline pulse (to overcome the increase in spontaneous ACTH release; see Results) and were expressed as percentage of prepulse values. Prepulse release rates were 3.6 ±.5 pg/min/pd (HE experiment, n = 1), 3.4 ±.4 pg/min/pd (CRH experiment, n = 12), and 4.9 ±.6 pg/min/pd (K ' experiment, n = 3). The saline values from the three experiments were pooled, yielding n = 26. The pulses were given at t = 22 min. Statistics All data are presented as average ± SEM. To test differences between experimental groups, data were analyzed by the Mann-Whitney U test; P <.5 was accepted as a significant level of difference. RESULTS Serial dilutions of plasma from stressed tilapia and pituitary homogenates displaced radiolabeled ACTH from the antibody in parallel with dilutions of the standard used (Fig. 1). The assay procedure produced a very reproducible standard curve and could measure ACTH levels in plasma samples (25 jjil) from unstressed animals: 47 ± 4 pg/ ml (n = 25). The detection limit and the ED5 were.32 and 4.3 pg per tube, respectively; intra- and interassay coefficients of variation were 6.2 and 7.2%, respectively.

5 35 BALM ET AL. dilution lilapia pituitary/pars cjislalis dilution plasma 1:15 1:14 1:13 ^ B :8 1:4 1:2 neal trout plasma tilapia plasma pituitary pars distalis. 2 h ACTH pg ACTH pg u-msh F ig. I. ACTH R1A standard and displacement curves. The standard curves presented is constructed from three separate curves. Plasma was obtained from stressed (2 hr confinement) tilapia and trout: 5 p.l of undiluted ( neat") or diluted (up to 16 times with assay buffer) plasma was used. For the pituitary and the pars distalis homogenates, three and two lobes respectively were homogenized and diluted up to 1, times (1:1s) in assay buffer; 5 (jli of these diluted samples was assayed. cx-msh was monoacetylated a-msh from Bachem. The cross-reactivity with monoacetylated a- MSH was negligible. Three other antibodies were tested: one donated by the NIDDK (previously demonstrated to crossreact with goldfish ACTH-like molecules; Tran et al., 1989), and an antibody directed against the N-terminal portion of the ACTH molecule (un-term," donated by Professor Sumpter; Sumpter, 1986), and an antibody from the IgG Corporation (Nashville, TN), which is highly sensitive and has been used to measure salmonid ACTH by RIA (Sumpter and Donaldson, 1986; Balm and Pottinger, 1993). These antibodies did not satisfactorily detect tilapia ACTH-like material in our RIA system, assessed from the lower sensitivity (NIDDK, N-term), or poor displacement curves with serial dilutions of tilapia pars distales homogenates (NIDDK, IgG; not shown). Conversely, the present antibody could not measure rainbow trout ACTH in plasma as judged by the poor displacement of radiolabeled ACTH by dilutions of plasma from stressed troui (Fig. 1). HPLC analysis of the ACTH immunoreactivity in tilapia pituitary homogenates revealed the presence of two products in both the PD and the NIL, which eluted from the column prior to the hacth,.39 standard (Fig. 2). The ratio between the peak areas of the two ACTH-IR peaks was similar in both lobes, but the pars distalis contained approximately eight times more ACTH immunoreactivity than the NIL. The more apolar ACTH product ( A ) coeluted with monoacetylated a-msh. Although the majority of the a-msh immunoreactivity was recovered from the NIL homogenate, significant amounts of desacetylated-a-msh in particular were present in the PD. The relative amounts of the three a-msh prod-

6 ACTH AND STRESS IN TILAPIA 351 PD N IL hacth standard ACTH i.r. O f\o 2-1 (pg.fraction' ) 5 B f\? \ n O A I B u-msh standard MSH i.r. (pg.fraction'1) o, o 4 mono- 2 il 3 o b 2 des di- 2 l\ des- 1. mono- ; *V v V HPLC elution time (min) F ig. 2. HPLC immunograms of tilapia pars distalis and neurointermediate lobe homogenates ( ). The open circles illustrate the elution profiles of ACTH and a-msh standards. Please note the different Y axes in the two a-msh panels. PD and NIL lobes from two Fish were homogenized and subjected to HPLC. Between 1- and 25-min.3-min fractions were collected;.5-min fractions were collected between 25 and 45 min. The major a-msh immunoreactive peaks in this system have been demonstrated previously to represent tilapia des-, mono-, and diacetylated a-msh (Lamers et al., 1991). cts was distinctly different in the two jbes. When tested in a homologous bioassay, oth HPLC purified immunoreactive vcth products were corticotropic (Fig. 3). apparently, at the concentration tested 25 pg ACTH-IR/ml), both tilapia prodicts evoked similar responses as the human tandard. During in vitro incubation, tilapia corticotropes released both ACTH products (Fig. 4a; A and B). The ratio between the amounts of both ACTH forms was similar to that in the lobe. In addition, a minor fraction of total medium ACTH immunoreactivity was recovered in several smaller peaks. Initially, the NIL and the PD released equivalent amounts of ACTH (Fig. 4b), but with time the melanotropes did not maintain this release rate, and from 8 min

7 352 BALM ET AL cortisol release (pg.mirf1 g '1) 8 tacth, tacth D hacth t (mm) F ig. 3. Corticotropic activity of tilapia ACTH immunoreactive products. Peak HPLC fractions from a pars distalis homogenate and a ACTH standard run were dried and dissolved in superfusion medium to yield a concentration of 25 pg/ml. Headkidneys from control fish were challenged with these solutions for 2 min. Results presented as increase in cortisol release over baseline release. Baseline cortisol release was 1.6 ±.4 pg/min/g (tactha; n = 12), pg/min/g (tacthb; n = 12), and 1.1 ±.4 pg/min/g (hacth,_39; n = 8). in vitro onwards the PD released significantly more ACTH than the NIL (P =.14). At t = 34 min, the release ratio between the PD and the NIL was 11.7 ± 3.4 (Fig. 4; n = 4). In vitro a-msh release by the two lobes (not shown) gave the opposite picture, with the PD releasing undetectable amounts of a-msh. The amount of ACTH released by the NIL during the initial 8 min approximated the ACTH content of a fresh NIL (Fig. 1). The in vitro release of ACTH by the PD could be stimulated by a hypothalamic extract, CRH, and K * depolarization (Fig. 4c). Tilapia appeared very sensitive to sequential sampling from groups judged by the extremely fast cortisol response (Fig. 5). At none of the sample points was the rise in plasma cortisol values (P <.1 compared with the animals caught first) accompanied by significant rises in either ACTH or glucose in plasma. The discrepancy between ACTH and cortisol levels was most notable in the fish caught second. From each of the nine groups sampled, the last two remaining fish were confined in their home aquaria for 3 hr. After this period, a marked dichotomy in cortisol levels was observed (Fig. 5). In every pair, one fish had control or moderately elevated cortisol levels, whereas the other fish exhibited levels of several hundred ng/ml. This difference was not related to the sequence in which the fish were sampled from each pair (3 sec in between the fish from each pair). When averaged separately, the cortisol levels in low responders (L) were lower than those in the fish captured just prior to the confinement period, but were twice those of fish captured as firsts. The high responders (H), however, maintained extremely high cortisol levels, which were significantly higher than those in the fish captured just prior to the confinement period (P =.27). The high significant difference in cortisol levels between the high

8 ACTH AND STRESS IN TILAPIA 353 % lolal ACTH i.r. released 12 a B ACTH release (pg.m m '1 lobe' 1) 12 b ACTH release ( o b a s a l) c p.o 2 PD 4 P ' O? B p< I 2 I m M İ I \ NIL H P L C elution lime (mm) lime (mm) pre-pulse sal H E CRH K F ig. 4. in vitro ACTH release by tilapia PD lobes. For the ACTH HPLC immunogram (a), two PDs were incubated for 1.5 hr in 1 ml of incubation medium, which was subsequently subjected to HPLC. To assess the unstimulated in vitro ACTH release rates (b), four PDs and four NILs were superfused individually for 34 min. The effect of a hypothalamic extract (HE), CRH. and K depolarization was investigated using PDs in superfusion (c). nd low responders also corresponds to dif- 3rences in the ACTH levels of the same idividuals. In this case, only the ACTH :vels in the high responders were signifiantly higher than those of the unstressed sh ( firsts'1). In contrast to the divergence i the cortisol and ACTH responses to concernent, both groups were equally hyperlycemic. The cortisol response to the capture proedure could be abolished completely by in ivo cortisol treatment (Fig. 6a): none of ie cortisol values of the cortisol-treated sh were different from the value in the fish aught first. In the control groups, there ;as a significant capture sequence effect 'hich was evident from the fish caught secnd onwards {P =.14). Cortisol levels in ortisol-treated animals caught first were ignificantly higher than their control counîrparts (P =.29). Despite the different Drtisol response to the capture protocol in le experimental groups, their plasma CTH levels were similar: 55 ± 9 versus pg/ml for controls (n = 27) and corti- )l-treated fish (n = 28), respectively. The ilues from the two groups were pooled because there was no fish number effect on plasma ACTH in either group. Two other parameters were affected by cortisol treatment: plasma glucose [68.9 ± 5.2 versus ± 8.2 mg/1 ml for the controls (n = 28) and for cortisol-treated fish (n = 28), respectively; P <.1] and in vitro cortisol release (Fig. 6b). To limit the possible influence of the capture sequence on this parameter, only the headkidneys from the first two fish from each groups were superfused in vitro. The spontaneous cortisol release in the cortisol-treated group was lower than in controls (P =.14 during the first fraction collected). This difference persisted during the entire incubation, although both curves seemed to have similar baselines. DISCUSSION The present study demonstrates the presence of two major ACTH immunoreactive products in the corticotropes and the melanotropes of tilapia (O. mossambicus). Both forms are released in vitro into the medium, in particular by the corticotropes.

9 354 BALM ET AL. capture confinement 3h plasma cortisol ng.ml -1 p< 1 4 p< 1 T H p< 25 piasma ACTH pg.m - 1 p= 27 5 i t T p< 25 P<1 plasma glucose mg.dl & 9 fish number time (min) Fig. 5. The effects of two handling treatments (capture or capture followed by confinement) on plasma cortisol, ACTH, and glucose in tilapia (n = 9 for each data point). The effect of capture + confinement in both L and H" fish (for explanation, see text) was compared to values of the fish captured first at t = min (significant differences indicated over the bars). Differences between the L (n = 9) and H (// = 9) responses are also given. In contrast to the situation in higher vertebrates and in salmonids (Sumpter et cil., 1986), the cortisol response to handling was not preceded by a rise in plasma ACTH. A prolonged stress treatment, however, (confinement in pairs) resulted in both cortisol and ACTH responses. However, this was evident in only one fish of every pair, pos-

10 ACTH AND STRESS IN TILAPIA 355 plasma cortisol (ng.ml*1) in vitro cortisol release (pg.miri'1.g'1) X I fish number t (min) Fig. 6. The effects of cortisol administration on the reaction of the pituitary-interrenal axis to a capture protocol (a), (b) The in vitro cortisol release by the headkidneys of the experimental fish (/; = 8 per group; see text for details). Closed and open circles refer to control and cortisol-treated fish, respectively. ibly indicating a social interaction in the egulation of the pituitary-interrenal axis uring confinement stress in this species. There are several reasons why the two Timunoreactive products identified by IPLC are concluded to represent tilapia vcth. First, the antiserum used has been emonstrated to detect ACTH molecules in variety of (nonmammalian) vertebrate pecies (Dores et cil., 1993; R. M. Dores, ersonal communication). Second, both roducts were bioactive in a homologous ioassay. The finding that the bioactivity /as comparable with that of the hacth,_39 upports our identification since none of íe other POMC-derived peptides identied to date in other vertebrates possess Drticotropic activity comparable to ACTH ^owry, 1984; Ranee and Baker, 1981). For iis reason the small contamination of the VCTHA HPLC fraction with monoacetyted a-msh (5% maximally) may be assumed not to interfere in the bioassay. Prolactin, the only major non-pomc factor in the tilapia rostral pars distalis, elutes at 46 and 5 min (tprl,77 and tprl188, respectively) in the HPLC system used (Balm, unpublished). Third, the different results obtained with tilapia PD and NIL tissue confirm the characteristic picture of POMC processing in these two pituitary lobes. Finally, the release of ACTH immunoreactivity could be markedly stimulated with CRH. The ACTH radioimmunoassay described combined sensitivity with specificity for tilapia ACTH-like molecules in particular, and could be used to measure ACTH immunoreactivity in tilapia, but not in trout plasma. It appears that the species specificity between fish ACTH molecules is considerable because for three species studied (goldfish, Tran et cil., 1989; salmonids, Sumpter, 1986; tilapia, this study) three dif

11 356 BALM ET AL ferent antibodies have been used which were demonstrated not to be interchangeable (Sumpter, 1986; this study). The ACTH antibody used possesses 1% crossreactivity with CLIP (Dores et al., 1993). Therefore, on the basis of the amount of a-msh in the NIL homogenate, a major (third) peak in the NIL ACTH immunogram was anticipated. Its absence indicates that the antibody does not detect tilapia CLIP. Comparison of the present data with those obtained with other species confirms the species specificity: two major immunoreactive products in tilapia (see Tran et al., 1989, for goldfish), which appear identical in PD and NIL (see Sumpter, 1986, for Chinook salmon), and which both possess a ratio between immunoreactivity and bioactivity of around 1 (see Scott and Baker, 1975, for rainbow trout). Therefore, the tilapia data in particular comply reasonably well w'ith the characteristic vertebrate picture (Jones and Gillham, 1988). It seems unlikely that these discrepancies between fish can be explained solely by the different methods (separation, assays) employed, and therefore indicate species-specific corticotrope functions. At present, the nature of the different forms in tilapia is unclear. The fact that both tilapia forms are released into the medium might indicate differential processing of POMC in the corticotropes of this species, possibly leading to phosphorylated and nonphosphorylated forms (Bennett et al., 1983) or to molecules of varying size (e.g., by carboxypeptidase activity; Friedman et al., 1993). The possibility that one of the forms represents POMC can be excluded, since the prohormone elutes at 41 min in the HPLC system used (Martens et al., 1982). Different degrees of glycosylation appear unlikely on the basis of the data by Iturriza and Estivariz (1986). The a-msh immunoreactivity present in the tilapia corticotropes indicates processing of ACTH in PD POMC cells of this species. The presence of a-msh did not result from contamination of the PD with NIL tissue or of residual NIL a-msh on the HPLC column, since the ratio between the three a-msh forms was different between the two tissues. The processing of ACTH into desacetylated a-msh in particular has also been demonstrated in the corticotropes of other vertebrates (Tanaka and Kurosomi, 1986; Dores et al., 1993). A function of this process is unclear, but it has been observed to be most prominent during fetal/ neonatal developmental stages (Dores et al., 1993) and to be under regulation by CRH (Iturriza and Ferese-Spinelli, 1991). A surprising finding was the increase in unstimulated ACTH release by tilapia PD tissue with time in vitro since these data are in conflict with in vitro data of Tran et al. (1989) on goldfish. However, Ball et al. (1972), studying the corticotropes of several fish species during in vitro culture, also reported an apparent activation of eel and rainbow trout ACTH cells under these conditions. This was taken to indicate a species-specific regulation of corticotrope function. The present data may indicate an overriding inhibitory control of basal secretory activity of tilapia corticotropes in vivo. At present the additional peaks in the HPLC immunogram of the incubation medium are unknown. They might either result from release-associated modifications of one or both of the major products or arise in the medium (e.g., sulfoxide forms of ACTH). However, only a small fraction o the total immunoreactivity was recovered in these peaks. These products are concluded to be of minor importance because the high steroidogenic potency of the two major ACTH-like products (tactha and tacthb), probably precludes, a major biological role for other products. In addition to the differences in both ACTH and a-msh release between the NIL and the PD POMC-producing cells, it also appeared that the pattern of unstimulated ACTH release by the melanotropes differs significantly from that of unstimulated a-msh release by these cells (Lamers et al., 1991).

12 ACTH AND STRESS IN TILAPIA 357 1he differences in PD and NIL ACTH connt and release suggest that the PD is the i ajor source of circulating ACTH in tila-? a. A similar conclusion was drawn for out by Rodrigues and Sumpter (1983). The almost immediate cortisol response > handling in O. mossambicus corrobo- tes recent results of Foo and Lam (1993) 31 the same species. Although these auors also sampled their fish by netting iem one by one, they did not observe the ' sh number effect reported here, possibly "}cause of the additional anesthetization, ep used by these authors. The prompti:ss of the response in tilapia is illustrated 3St by comparison with that in brown trout a similar handling protocol (Pickering et 1982). The data of Sumpter et al. (1986) n rainbow trout also indicate that plasma >rtisol levels did not react to handling and mfinement within 8 min. Also, in humans >rtisol responses become manifest after 1 in (Hermus et al., 1984). In addition to the pid onset of the cortisol response to disirbance, a second notable feature of the sponse in tilapia is the ability of (a subset f) the animals to rapidly lower their iasma cortisol levels in the continuous *esence of the stressor. An efficient corti- >1 clearance mechanism probably plays an iportant role in this process. Foo and am (1993) demonstrated that during the rst hour after a stress treatment plasma >rtisol in O. mossambicus had dropped om 144 ng/ml to control levels (5 ng/ml). *evious results on the effects of cortisol lministration on plasma cortisol levels»aim, 1986) corroborate the short half-life cortisol in tilapia plasma suggested by e data of Foo and Lam. The ability of apia to quickly downregulate the activity the pituitary interrenal axis will be of obdus advantage to the fish, since in this \y they are able to benefit from the alarm nction of interrenal activation, but can oid the deleterious consequences of proiged elevations in plasma cortisol (Bari and Iwama, 1991). A unique feature of the initial interrenal response in tilapia subjected to handling is its apparent ACTH independence. In some of the groups studied, individual fish had elevated plasma ACTH levels. However, there was no correlation between plasma cortisol and plasma ACTH (r =.12, P >.5, n = 63), which further supports the conclusion that during the first minutes ACTH and cortisol are independently regulated. The response to confinement in particular in the high responders, however, indicates that a more prolonged activation of interrenal cortisol output might rely on elevated ACTH levels. The divergence in the interrenal response to confinement ties in with reports on eel (Peters et al., 1981) and rainbow trout (Pottinger and Pickering, 1992), in which fish were also exposed to confinement in pairs. The present data extend this difference to the pituitary level. The different response in pairs usually is discussed in terms of hierarchy, the low responders being dominant. Tilapia species are hierarchal (de Oliveira Fernendez and Volpato, 1993), and therefore we suggest that in our fish differences in hierarchal status (established prior to or during the confinement period) underly the different stress responses. As suggested for other species, this phenomenon might selectively predispose high responders to infectious diseases (Maule et al., 1987), as well as reduce their development and growth rate. Other relevant features include the facts that subordinates generally suffer most from additional stress conditions (Schreck, 1981) and social rank in some fish is associated with interrenal mass (Erickson, 1967) and interrenal cell activity (Noakes and Leatherland, 1977; Schreck, 1981). Whether differences in HPA axis activity associated with social status are also expressed anatomically at levels other than the interrenal is unknown. Social regulation of the brain-pituitary-gonadal axis in Haplochromis burtoni (an African cichlid fish) occurs through reversible modulation of

13 358 BALM ET AL. soma size changes in a population of GnRH neurons in the ventral forebrain, territorial males having larger cell bodies (Francis et al., 1993). The results of the cortisol administration experiment confirm the apparent ACTH independence of the initial cortisol response since cortisol treatment abolished the cortisol response to capture but did not affect plasma ACTH levels. The data further suggest that tilapia corticotropes are relatively insensitive to direct corticosteroid feedback, which may seem surprising in view of previous in vivo and in vitro results in other fish (Pickering et a l 1987, Olivereau and Olivereau, 1989). However, Fryer and Peter (1977) demonstrated that implanting cortisol-releasing pellets adjacent to the NLT or NPO of the goldfish hypothalamus, but not in the pituitary, inhibits the cortisol They concluded that in their species, cortisol mainly acted at the level of the hypothalamus to inhibit the reaction of the corticotropes to handling. One conclusion that can be drawn from the present data is that the plasma cortisol level is not a reliable index for corticotrope activity (Fryer and Peter, 1977) because the results demonstrate that plasma cortisol levels may be elevated in the absence of corticotrope activation. Two mechanisms might explain this cortisol-induced inhibition of the interrenal stress response in tilapia, which apparently bypassed the ACTH cells. First, direct cortisol feedback at the level of the interrenal cells: our in vitro data might support this notion. Assuming that the initial in vitro cortisol release reflects in vivo secretory rates (Balm, 1986; Balm and Pottinger, 1993), the data demonstrate that interrenal cells of cortisol-treated tilapia release less cortisol than in controls, despite similar plasma ACTH levels. This would imply that ACTH is of minor importance for the maintenance of basal in vivo cortisol output. Cortisol has recently been suggested on the basis of in vitro experimental evidence to act at the level of the interrenal cells to render them less sensitive to ACTH (Bradford et al., 1992). However, this mechanism cannot explain the suppression of the interrenal response to capture by exogenous cortisol, since that initial response does not rely on ACTH. Alternatively, the results would be consistent with the idea that the initial phase of the interrenal response is regulated by neuronal mechanisms, which are sensitive to cortisol, either at the level of the CNS or the interrenal cells. The kinetics of the cortisol response seem to favor this type of regulation. Furthermore, corticosteroidogenesis is modulated by a number of neurotransmitters delivered from nerve endings (Charlton, 199) or released from nonsteroidal adrenal cells, regulating steroidogenesis in a paracrine fashion. response to handling (saline injection). ACKNOWLEDGMENT The authors gratefully acknowledge Professor R. M. Dores (University of Denver. CO) for the generous supply of the ACTH antibody. REFERENCES Ball, J. N., Baker, B. I., Olivereau, M., and Peter, P. E. (1972). Investigations on hypothalamic control of adenohypophysial functions in teleost fishes. Gen. Comp. Endocrinol. Suppl. 3, Balm, P. H. M. (1986). Osmoregulation in teleosts by cortisol and prolactin. Adaptation to low ph environments. Thesis, University of Nijmegen, The Netherlands. Balm. P. H. M., Lambert, J. D. G., and Wendelaar Bonga, S. E. (1989). Corticosteroid biosynthesis in the interrenal cells of the teleost fish Oreochromis mossambicus. Gen. Comp. Endocrinol. 76, Balm, P. H. M., and Pottinger, T. G. (1993). Acclimation of rainbow trout (Oncorhynchus my kiss) to low environmental ph does not involve an activation of the pituitary-interrenal axis, but evokes adjustments to branchial ultrastructure. Can. J. Fish. Aquat. Sci. 5, Barton, B. A., and Iwama, G. K. (1991). Physiological changes in fish from stress in aquaculture with emphasis on the response and effects of corticosteroids. Annu. Rev. Fish. Dis. 1, Bennett, H. P. J., Brubaker, P. L., Seger, M. A., and

14 ACTH AND STRESS IN TILAPIA 359 Solomon, S. (1983). Human phosphoserine 31 corticotropin,,39. Isolation and characterization. J. Biol. Chem. 258, radford. C. S., Fitzpatrick. M. S., and Schreck, C. B. (1992). Evidence for ultra-short-loop feedback in ACTH-induced interrenal steroidogenesis in coho salmon: Acute self-suppression of cortisol secretion in vitro. Gen. Comp. Endocrinol. 87, >rown, K. I., and Nestor, K. E. (1973). Some physiological responses of turkeys selected for high and low adrenal response to cold stress. Poultry' Sci. 52, harlton, B. G. (199). Adrenal cortical innervation and glucocorticoid secretion. J. Endocrinol. 126, 5-8. e Oliveira Fernandes, M., and Volpato, G. L. (1993). Heterogeneous growth in the Nile tilapia: Social stress and carbohydrate metabolism. Physiol. Behavior 54, )onaldson, E. M. (1981). The pituitary-interrenal axis as an indicator of stress in fish. In Stress and Fish (A. D. Pickering, Ed.), pp Academic Press, London. )ores, R. M., Sandoval. F. L., and McDonald, L. K. (1993). Proteolytic cleavage of ACTH in corticotropes of sexually mature axolotls (Ambystoma mexicanium). Peptides 14, Irickson, J. G. (1967). Social hierarchy, territoriality, and stress reactions in sunfish. Physiol. Zool. 4, ;evolden, S. E., and Red, K. H. (1993). Cortisol and immune characteristics in rainbow trout (Oncorhynchus mykiss) selected for high or low tolerance to stress. J. Fish Biol. 43, ', J. T. W., and Lam, T. J. (1993). Serum cortisol response to handling stress and the effect of cortisol implantation on testosterone levels in the tilapia, Oreochromis mossambicus. Aquaculture 115, :rancis, R. C., Soma, K., and Fernald, R. D. (1993). Social regulation of the brain-pituitary-gonadal axis. Proc. Natl. Acad. Sci. USA 9, 'riedman, T. C., Chen, H.-C., and Peng Loh, Y. (1993). Generation of 1-37 and 1-38 forms of adrenocorticotropin by mono- and dipeptidyl serine carboxypeptidase activities in bovine pituitary secretory vesicles. Endocrinology 133, ryer, J. N., and Peter, R. E. (1977) Hypothalamic control of ACTH secretion in goldfish III. Hypothalamic cortisol implant studies. Gen. Comp. Endocrinol. 33, ^ryer, J. N., Lederis, K., and Rivier, J. (1984). Cortisol inhibits the ACTH-releasing activity of Uro- tensin I, CRF and sauvagine observed with superfused goldfish pituitary cells. Peptides 5, Hermus, A. R. M. M., Pieters, G. F. F. M., Smals, A. G. H., Benraad, Th. J., and Kloppenborg, P. W. C. (1984). Plasma adrenocorticotropin, cortisol, and aldosterone responses to corticotropin- releasing factor: modulatory effect of basal cortisol levels. J. Clin. Endocrinol. Metab. 58, Iturriza, F. C., and Estivariz, F. E. (1986). Lack of glycosylation of pro-opiomelanocortin might account for the periodic acid-schiff-negative reaction in ACTH cells of teleost fishes. Gen. Comp. Endocrinol. 61, Iturriza, F. C., and Ferese-Spinelli, C. G. (1991). Transplantation of the pituitary pars distalis induces the corticotrophs to store melanocyte- stimulating hormone, an effect reversed by the administration of corticotropin-releasing factor. Neuroendocrinology 53, Jones, M. T., and Gillham. B. (1988). Factors involved in the regulation of adrenocorticotropic hormone/(3-lipotropic hormone. Physiol. Rev. 68, Lamers, A. E., Balm, P. H. M., Haenen, H. E. M. G., and Wendelaar Bonga, S. E. (1991). Regulation of differential release of a-melanocyte- stimulating hormone forms from the pituitary of a teleost fish, Oreochromis mossambicus. J. Endocrinol. 129, Lowry, P. J. (1984). Pro-opiocortin: The multiple adrenal hormone precursor. Biosci. Rep. 4, Martens, G. J. M., Jenks, B. G., and van Overbeeke, A. P. (1982). Biosynthesis of pairs of peptides related to melanotropin, corticotropin and endorphin in the pars intermedia of the amphibian pituitary gland. Eur. J. Biochem. 122, 1-1. Maule, A. G., Schreck, C. B., and Kaattari, S. L. (1987). Changes in the immune system of coho salmon (Oncorhynchus kisutch) during parr-to- smolt transformation and after implantation of cortisol. Can. J. Fish. Aquat. Sci. 44, Munck, A., and Naray-Fejes-Toth, A. (1992). The ups and downs of glucocorticoid physiology. Permissive and suppressive effects revisited. Mol. Cell. Endocrinol. 9, C1-C4. Noakes, D. L. G., and Leatherland, J. F. (1977). Social dominance and interrenal cell activity in rainbow trout, Salmo gairdneri (Pisces, Salmonidae). Environ. Biol. Fish 2, Olivereau, M., and Olivereau, J. (1989). Quantitative changes of CRF-like immunoreactivity in eels treated with reserpine and cortisol. Peptides 9, Peters, G., Klinger, H., and Delventhal, H. (1981). Stress diagnosis of eels. In Stress and Fish (A. D. Pickering, Ed.), pp Academic Press, London.

15 36 BALM ET AL Pickering, A. D., Pottinger, T. G., and Christie. P. (1982). Recovery of the brown trout, Salmo trutta L., from acute handling stress: A time-course study. J. Fish Biol. 2, Pickering, A. D., Pottinger, T. G., and Sumpter, J. P. (1987). On the use of dexamethasone to block the pituitary-interrenal axis in the brown trout, Salmo trutta L. Gen. Comp. Endocrinol. 65, Pottinger. T. G., and Pickering, A. D. (1992). The influence of social interaction on the acclimation of rainbow trout, Oncorhynchus mvkiss (Walbaum) to chronic stress. J. Fish. Biol. 41, Pottinger, T. G., Pickering. A. D., and Hurley. M. A. (1992). Consistency in the stress response of individuals of two strains of rainbow trout, Oncorhxnchus mykiss. Aquaculture 13, Ranee, T. A., and Baker, B. 1. (1981). The in vitro response of the trout interrenal to various fragments of ACTH. Gen. Comp. Endocrinol. 45, Rodrigues, K., and Sumpter, J. P. (1983). The distribution of some proopiomelanocortin-related peptides in the pituitary gland of the rainbow trout, Salmo gairdneri. Gen. Comp. Endocrinol. 51, Sangalang, G. B., Truscott, B., and Idler, D. R. (1972). A comparison of steroidogenesis in vivo in two teleosts, the marine herring, Clupea, and the freshwater Atlantic salmon, Salmo. J. Endocrinol. 53, Schreck, C. B. (1981). Stress and compensation in teleostean fishes: Response to social and physical factors. In Stress and Fish (A. D. Pickering. Ed.), pp Academic Press, London. Scott, A. P., and Baker, B. I. (1975). ACTH production by the pars intermedia of the rainbow trout pituitary. Gen. Comp. Endocrinol. 27, Sumpter, J. P. (1986). ACTH-related material in the pituitary gland of the Chinook salmon (Oncorhynchus tschawytscha). Gen. Comp. Endocrinol. 62, Sumpter, J. P., and Donaldson, E. M. (1986). The development and validation of a radioimmunoassay to measure plasma ACTH levels in salmonid fishes. Gen. Comp. Endocrinol. 62, Sumpter, J. P., Dye, H. M., and Benfey, T. J. (1986) The effects of stress on plasma ACTH, MSH, and cortisol levels in salmonid fishes. Gen. Comp. Endocrinol. 62, Tanaka, S., and Kurosomi, K. ( 1986). Differential sub- cellular localization of ACTH and a-msh in corticotropes of the rat anterior pituitary. Cell Tissue Res. 243, Tran, T. N., Fryer, J. N., Bennett, H. P. J., Tonon, M. C., and Vaudry, H. (1989). TRH stimulates the release of POMC-derived peptides from goldfish melanotropes. Peptides 1, van Zoest, I. D., Heijmen, P. S., Cruijsse, P. M. J. M., and Jenks, B. G. (1989). Dynamics of background adaptation in Xenopus laevis: Role of catecholamines and melanophore stimulating hormone. Gen. Comp. Endocrinol. 76,