INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

SEVERAL CONCEPTS have been proposed regarding the endocrine regulation of vertebrate puberty (reviewed in Ref. 11). The "missing link" concept is the most general one. It covers a number of more specified concepts and assumes that one or more components of the brain-pituitary-gonad (BPG) axis are not yet functional. Missing links can be any component at any level of the BPG axis, e.g., hormone synthesis and/or release mechanisms, receptor expression and/or functioning, or intracellular signal transduction. The "gonadostat" concept, on the other hand, relates to the indirect negative feedback effect that sex steroids exert on gonadotropin-releasing hormone (GnRH)-producing neurons in juveniles (e.g., 6, 25, 32, 48). When juveniles approach puberty, the gonadostat becomes less effective, allowing the GnRH neurons to increase their activity. The exact details of the mechanism(s) underlying this change in the gonadostat have yet to be elucidated but the mechanism results in increased gonadotropin secretion, which in turn stimulates steroidogenesis and gametogenesis.

In juvenile fish, treatment with sex steroids initiated and/or accelerated the development of the BPG axis. Testosterone (T) treatment, for example, increased the GnRH content in the brain of male rainbow trout Oncorhynchus mykiss (12); estradiol-17 (E₂) had a similar effect in immature female eel Anguilla anguilla (23). The pituitary content in the luteinizing hormone (LH)-like gonadotropic hormone¹ is also strongly increased in juvenile fish in response to T or E₂ (e.g., Refs. 2, 8, 10). A direct stimulatory effect on spermatogenesis by 11-ketotestosterone (KT), the main circulating androgen in male fish (1), was observed by Miura et al. (21): incubations of testicular explants from immature Japanese eel (Anguilla japonica) with KT induced complete spermatogenesis. Cochran (7), on the other hand, reported that T, but not KT or 11-hydroxytestosterone, stimulated spermatogenesis to some degree in testicular explants of the mummychog Fundulus heteroclitus. Taken together, it appears that steroids produced by the juvenile testes stimulate spermatogenesis in fish, either by direct action on the testis or by activation of the GnRH system and/or the pituitary gonadotrophs. Such stimulatory effects on all levels of the BPG axis in juvenile fish are difficult to reconcile with the gonadostat concept, which is based rather on inhibitory effects of sex steroids. Thus the fish model appears to be suited to study the initiation of puberty in the context of the missing link concept, attributing the role of a candidate missing link to sex steroids.

As a working definition for the African catfish, the model species used in the present study, puberty is considered the period that starts with spermatogonial multiplication (at ~3 mo of age) and ends when the first wave of spermatogenesis is completed by the differentiation of the first spermatozoa (at ~6 mo of age). To obtain information on the steroid environment during puberty, the synthetic capacity of the testis and interrenal tissues was monitored (3). KT was the predominating androgen in the plasma, whereas 11-hydroxyandrostenedione (OHA) was the main testicular product. Moreover, it was shown that OHA is converted into KT by the liver (4). In an earlier study, it has already been shown that KT had no effect and OHA and androstenetrione (OA) had only slightly inhibitory effects on circulating LH levels (5). Hence the present study focused on the effects of sex steroids on the development of the dual function of the testes: spermatogenesis and androgen production. Moreover, effects on the development of secondary sexual characteristics have been monitored.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animals, hormone treatments, and sampling. African catfish (Clarias gariepinus) were bred and raised as described previously (9), except that African catfish pituitary extract was used instead of human chorionic gonadotropin to induce ovulation. The fish were kept in a copper-free recirculation system at a water temperature of 25 ± 2°C, exposed to a 14:10-h light-dark photoperiod, and fed with Trouvit pellets (Trouw, Putten, The Netherlands). Fish of three different broods were used for three subsequent experiments (Table 1). To start experimentation before the natural onset of spermatogenesis (at 11-13 wk of age; Ref. 38), the gonadosomatic index (GSI = gonad wt × 100/body wt) was recorded weekly from some males of the three broods starting at 8 wk of age; the GSI increases >0.01 when meiosis has started (3). Hence steroids were administered at 10 or 11 wk of age (Table 1) with GSI starting values of 0.005-0.01 (Fig. 3). At this stage of development, only spermatogonia are present in the testes.

All steroids were purchased from Sigma (Zwijndrecht, The Netherlands). OHA was administered as it is the main product of testicular steroidogenesis during puberty as well as in adult African catfish (3, 39, 50). KT, androstenedione (A), and T were also studied, as they are prominent plasma steroids (51). DHT and E₂ have been included to be able to identify treatment effects that might result from metabolization of T or A. The African catfish LH preparation used for the stimulation of testicular androgen secretion in vitro was described previously (43).

Steroid hormones were incorporated in solid Silastic pellets (elastomer; 26, 49) that were 2 mm in diameter. The length of the pellets was adjusted to the body weight such that a dose of 30 µg/g body wt was implanted in the body cavity through a 2-mm-long incision (see Table 1 for treatment groups).

In each experiment, a control group received steroid-free pellets. In addition, on the day of implantation, a start control group was sampled for body weight, testis weight, and histological analysis of spermatogenesis. Blood was collected by puncturing the caudal vasculature using a 1-ml syringe (needle: 26 gauge × 1/2 in.) 2 wk after implantation and stored at 20°C for steroid RIA (see Steroid RIA). Blood sampling was followed by decapitation of the animals. The presence of seminal vesicles and the development of the urogenital papilla was recorded. The latter shows a clear sex dimorphism (Fig. 1, A and B). The seminal vesicles (Fig. 1C) display seasonal growth and regression (24) and function as accessory sex glands. Testes were removed and weighed to calculate the GSI. Testis tissue was then prepared for in vitro incubation with African catfish LH, after which the tissue was processed for light microscopical analysis of spermatogenesis.

View larger version (127K): [in this window] [in a new window]   Fig. 1.   Ventral views of the urogenital papilla of adult male (A) and female (B) African catfish (bars = 1 cm). C: reproductive tract of an adult male African catfish. T, testis; SV, seminal vesicles; UP, urogenital papilla; S, stomach; I, intestine; L, liver lobes (bar = 3 cm).

Steroid secretion in vitro. Testicular tissue was prepared for in vitro incubation as described previously (41). Briefly, the left and right testis of each male were weighed separately and placed in separate wells of 24-well plastic plates (Costar, Badhoevedorp, The Netherlands) containing 0.5 ml of Earle's balanced salt solution (M199 EBSS; Life Technologies-GIBCO, Breda, The Netherlands) supplemented with HEPES (0.02 M; adjusted to pH 7.18 with NaOH) and antibiotics (100 U/ml penicillin G and 100 ng/ml streptomycin sulfate; Life Technologies-GIBCO). Testes were cut into fragments of ~2 mm³ and rinsed once with medium. The incubation was started by replacing the medium with 0.5 ml fresh medium for the left testis of each animal (basal androgen secretion). For the right testis, 0.5 ml medium was added containing African catfish LH at a dose of 100 ng/ml. This LH concentration induces a maximal stimulation of androgen secretion (39). The incubation took place in air atmosphere at 25°C for 18 h. Then, the medium was removed and heated for 1 h at 80°C, centrifuged at 16,000 g for 30 min at room temperature, and supernatants were stored at 20°C until quantification of OHA.

Testicular histology. Testicular fragments of control incubations were used for histological analysis of spermatogenesis. They were fixed for 1 h in 0.1 M sodium cacodylate buffer (pH 7.2) containing 2% glutaraldehyde and 1% paraformaldehyde, postfixed with 1% OsO₄ (1 h) in the same buffer, and then dehydrated and embedded in epoxy resin. One micrometer sections were cut on a Reichert-Jung ultramicrotome (Vienna, Austria), collected on gelatin-coated slides, and stained with 1% methylene blue in 1% borax.

In the African catfish, spermatogenesis has been subdivided into four histological stages (3) that are characterized by the presence of the following germ cell types (Fig. 2): stage I, spermatogonia only; stage II, spermatogonia and spermatocytes; stage III, spermatogonia, spermatocytes, and spermatids; and finally, stage IV, all germ cells including spermatozoa. Animals were assigned to a certain testicular stage according to the most advanced spermatogenetic cell type. The results of the histological analysis are expressed as percentages of males found in the different testicular stages (% of testicular stages = number of animals in a certain stage × 100/total number of animals).

View larger version (191K): [in this window] [in a new window]   Fig. 2.   Sections (1 µm) of African catfish testes at different stages of spermatogenesis. A: stage I, lobules contain large, isolated spermatogonial stem cells (SSC) surrounded by Sertoli cells (Se), smaller spermatogonia (SG), and groups of spermatogonia also surrounded by Sertoli cells, thus forming intralobular cysts. Large clusters of Leydig cells (Le) are located in the interstitium. B: stage II, spermatocytes (SC) in different stages of meiosis are found next to SG. Leydig cells become dispersed. C: stage III, the lobules are larger. In addition to spermatogonia and spermatocytes, cysts with spermatids (ST) are present. D: stage IV, cells at all stages of spermatogenesis, including spermatozoa (SZ). Bars in A and B = 20 µm; bars in C and D = 10 µm.

Steroid RIA. The plasma levels of KT, OHA, T, and E₂ and the levels of OHA in incubation media were determined by specific RIAs as described previously (3, 39). The detection limit of all RIAs was 8 pg/tube; rates of crossreaction of the antisera were reported earlier (36, 37).

Statistics. Results are expressed as means ± SE. Steroid levels are given as nanograms per milliliter plasma or as nanograms secreted per milligram of testis tissue incubated. For statistical analysis, data on GSI and OHA secretion in vitro were log₁₀-transformed and then processed by one-way ANOVA, followed by the Fisher's least-significant difference test. For the analysis of the stimulatory effect of LH, we used a one-tailed paired t-test. Secondary sexual characteristics and histological data were analyzed by Fisher's exact test.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Plasma levels of sex steroids. Males carrying steroid-containing implants showed significantly elevated circulating hormone levels at the end of the experimental period (Table 2). The plasma levels of KT were increased after treatment with KT, OHA, and OA but not after treatment with T, A, and E₂. OHA or E₂ plasma levels were only elevated in the OHA- or E₂-treated groups, respectively. T levels were elevated in the T- and A-treated groups. Although E₂ is undetectable in males irrespective of the stage of development, the plasma levels of KT, OHA, and T in the respective treatment groups are in the physiological range of adolescent male African catfish (3).

GSI, testicular histology, and secondary sexual characteristics. In all three experiments, an increase in GSI (Fig. 3, left) and a higher percentage of males with spermatocytes (Fig. 3, right) were observed when comparing start and end control groups. This reflects the normal testicular development during the experimental period of 2 wk. As expected for the early stages of puberty, development of the secondary sexual characteristics did not occur, except for experiment 3, in which the fish were in a more advanced stage of development already at the beginning of the experiment, as indicated by their higher GSI values and incidental presence of spermatocytes in the start control group.

View larger version (41K): [in this window] [in a new window]   Fig. 3.   Gonadosomatic index (GSI; means ± SE), %testicular stages (no. of animals at a certain stage × 100/total no. of animals per treatment group), and presence of SV and UP calculated as the number of animals with SV or UP × 100/total no. of animals per treatment group in experiments 1 (n = 20-35), 2 (n = 19-28), and 3 (n = 15-25) in African catfish 2 wk after implantation of different steroids: androstenetrione (OA), 11-hydroxyandrostenedione (OHA), androstenedione (A), testosterone (T), estradiol-17 (E2), 11-ketotestosterone (KT), and 5-dihydrotestosterone (DHT). Groups sharing the same letter are not significantly different (P < 0.05, ANOVA, followed by the Fisher's least-significance difference test for GSI and Fisher's exact test P < 0.05 for %testicular stages, SV, and UP). See Fig. 2 for histological appearance of stages I-III.

In experiment 1, treatment with OHA and OA increased the GSI above the end control level (Fig. 3, left) and promoted spermatogenesis, as indicated by a higher proportion of spermatocytes and/or the presence of spermatids (Fig. 3, right). Moreover, the development of secondary sexual characteristics was stimulated in the OHA- and OA-treated groups. Treatment with A, T, and E₂ did not increase the GSI levels, whereas T slightly stimulated the development of the seminal vesicles. Spermatogenesis did not differ from the end control group in T- and A-treated fish, whereas spermatogenesis was inhibited in the E₂-treated group.

In experiment 2, the effects of T and A on spermatogenesis and secondary sexual characteristics were similar to those observed in the first experiment, i.e., there was no difference to the end control group in regard to spermatogenesis, and T, but not A, stimulated the development of seminal vesicles. An inhibitory effect of A and E₂ was noted on testicular growth, but E₂ did not inhibit spermatogenesis in this experiment. The DHT treatment had no effect and thus differed from the T-treated group in regard to the development of the secondary sexual characteristics. Treatment with KT had stimulatory effects on testicular growth, spermatogenesis, and secondary sexual characteristics.

Experiment 3 was carried out to reconfirm and to directly compare the effects of the three 11-oxygenated androgens. They all significantly stimulated testicular growth and advanced spermatogenesis and the development of secondary sexual characteristics to a similar degree.

Basal and LH-stimulated OHA secretion in vitro. Compared with testis tissue of the end control group, basal OHA secretion was significantly reduced after treatment with all steroids, E₂ being least effective (Fig. 4, experiment 1). A significant increase of OHA secretion above basal levels was induced by LH (100 ng/ml) in all cases. However, this increase was only two- to threefold after OHA treatment, whereas an at least sixfold increase was observed in all other groups, including the end control group (Table 3). As regards the absolute amounts of OHA secreted in response to LH, OHA treatment led to the strongest reduction followed by OA, T, and A; E₂ treatment had no effect.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   In vitro secretion (ng/mg testis, means ± SE, n = 15-20) of OHA by African catfish testes 2 wk after implantation of different steroids. C, end control. Groups sharing the same underscore are not significantly different (P < 0.05, ANOVA, followed by the Fisher's least-significance difference test). Irrespective of the pretreatment in vivo, OHA secretion in the presence of LH (100 ng/ml; filled bars) was significantly higher (P < 0.05, 1-tailed paired t-test) than in the absence of LH (basal; hatched bars).

In experiment 2, basal OHA secretion was reduced after treatment with KT, T, and A but not after treatment with DHT or E₂ (Fig. 4). Again, 100 ng LH/ml always induced significant increases in OHA secretion above basal levels. The secretion of OHA in response to LH was strongly reduced after treatment with KT and T, whereas A, E₂, and DHT were progressively less effective.

In experiment 3, treatment with the three 11-oxygenated androgens resulted in a significant reduction of both basal and LH-stimulated OHA secretion to a similar extent. As in experiment 1, OHA treatment resulted in a reduction of the response to LH (Table 3).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This study aimed to assess the effects of a number of steroid hormones on testis development in juvenile African catfish. The treatment effects can be summarized as follows: the 11-oxygenated androgens KT, OHA, and OA stimulated spermatogenesis and the development of secondary sexual characteristics but reduced the capacity of the testis to secrete OHA. On the contrary, the androgens T, A, and DHT failed to stimulate spermatogenesis and had no or limited (T) effects on the development of secondary sexual characteristics. However, T and, to a lesser extent, A also reduced both basal and LH-stimulated OHA production. Although E₂ and DHT also had similar effects on OHA production, they were much less effective than the other steroids tested.

In previous experiments the 11-oxygenated androgens had no (KT) or weak (OHA and OA) inhibitory effects on circulating LH levels (5). This suggests that both the stimulation of spermatogenesis and the impairment of steroidogenesis in the present experiments reflect direct effects of the 11-oxygenated androgens on the testis rather than inhibitory effects on LH release from the pituitary.

The similarity of the effects after treatment with OHA, OA, or KT may be explained on the basis of the assumption that these androgens exert their effects via a common mechanism of action or that KT is the biologically active androgen after bioconversion of OHA or OA (4). We favor the latter possibility, as previous studies have shown that treatment of adult male African catfish with OA (40) or OHA (4) led to increased plasma levels of KT. The conversion of OHA and OA to KT was shown to take place in the liver (4), and the present results indicate (Table 2) that this conversion is also very effective in juvenile males.

In stage I of spermatogenesis, African catfish testes show numerous clusters of highly active Leydig cells, supposedly generating high intratesticular levels of OHA, as reflected in the highest basal production of OHA per milligram testis tissue that was recorded throughout puberty (39). These high intratesticular OHA levels render it unlikely that OHA is the hormonal signal pushing spermatogenesis toward more advanced stages, again favoring the concept that KT is the biologically active androgen. However, it is possible that high OHA levels are required as a permissive factor for spermatogonial multiplication and/or their differentiation into primary spermatocytes. In any case, the high Leydig cell activity appears to be, at least in part, independent of LH, because circulating (5) and pituitary (44) LH levels are very low in stage I males.

During normal development, the initially low testicular mass (1-4 mg in stages I and II) and hence a low total testicular OHA output, apparently does not allow a quantitatively important hepatic KT production; moreover, the testicular production of KT is very low (3), ~1/100th of the OHA production (39). Accordingly, KT and OHA plasma levels were low during stages I and II (3). Starting in stage II, Leydig cells become dispersed in the rapidly growing testes, and basal OHA production per milligram tissue in stages II, III, and IV was 3-, 7-, and 13-fold lower than in stage I (39), so that intratesticular OHA levels will probably have decreased. The rapid testicular growth, on the other hand, more than compensated for the decreased OHA production per milligram of tissue, resulting in a strong (44-fold) increase in the total basal testicular OHA production comparing stages I and IV (39). Together with an increased hepatic capacity to convert OHA to KT in stages III and IV (15), these traits may underlie the observed increases in circulating OHA and KT levels (3). The present study showed that spermatogenesis was advanced by increasing the availability of precursors for hepatic KT production by treatment with OHA or OA or by direct treatment with KT. This is in line with the concept that KT has a direct stimulatory effect on the testis and drives spermatogenesis toward and beyond the first meiotic division, as has been shown in the Japanese eel (21). Finally, it is possible that the progressive decrease of intratesticular OHA concentrations, which was inferred from the decreasing basal OHA production in stages II-IV (see above), plays a permissive role in KT-driven spermatogenesis. Pituitary LH levels increase dramatically shortly after the appearance of spermatocytes in African catfish (44), which, however, is not reflected in circulating LH levels, being rather low (<1 ng/ml) throughout puberty (42). This indicates that LH secretion is under inhibitory control, possibly involving dopamine (9) and steroid-mediated feedback (40). Nevertheless, it cannot be excluded that changes in testicular LH sensitivity (39, 41) also contribute to the puberty-associated increases in circulating steroid levels.

The molecular mechanism of action used by KT to stimulate testicular development is not clear. A direct action of KT on the testis was demonstrated in the Japanese eel, where KT induced all stages of spermatogenesis in testicular explants in vitro (21), possibly involving KT-induced expression of activin B by Sertoli cells (20) and depending on a direct contact between germ cells and Sertoli cells (19). Surprisingly, androgen receptors in fish show only a weak affinity for KT (see below). However, a cell membrane-associated KT-binding protein has been partially characterized in the stickleback kidney (13), which may indicate the existence of a nonclassical, nonnuclear KT receptor in fish. Clearly, more work is needed in this context. Nevertheless, our results show that KT has important biological functions in vivo to stimulate testicular development.

Androstenedione is a major product of pregnenolone metabolism in both immature and mature African catfish testes (3, 35), and this androgen is readily converted to OHA, whereas T is produced in much smaller amounts. However, analysis of the plasma androgen levels (Table 2) showed that no significant increase occurred in OHA or KT levels after implantation of A or T, and neither A nor T stimulated spermatogenesis. We therefore assume that the low testicular mass of immature males did not allow a quantitatively significant conversion of A or T to KT or its precursors, resulting in a failure to promote spermatogenesis. E₂ and DHT tended to inhibit testicular growth or spermatogenesis, whereas they had no effects on secondary sexual characteristics and only weak effects on OHA production. These effects are distinct from those of T and A, so that their further metabolism (5-reduction or aromatization) is unlikely to be responsible for the effects observed after treatment with T or A. We therefore assume that these two androgens have specific effects, possibly related to the presence of a classic, nuclear androgen receptor. Considering that the effects of T on basal and LH-stimulated OHA secretion in vitro were stronger than those of A, the African catfish androgen receptor may show a higher affinity for T. Although experimental evidence is not available for the African catfish, the androgen receptor-like ligand binding characteristics of other fish are distinct from those of mammalian androgen receptors in that DHT is a 5- to 10-fold weaker competitor than T (27, 30). The latter may explain the lack of effect of DHT in the present experiments or the lack of effect of DHT on basal and GnRH-stimulated LH release in goldfish (49). Finally, it is important to note that KT is a poor competitor for the binding of T to classic androgen receptors in several species (13, 27, 30, 47). Taken together, we assume that the effects of 11-oxygenated androgens on spermatogenesis are mediated by a mechanism that involves specific receptors for the 11-oxygenated androgens that have yet to be characterized. A nuclear androgen receptor, on the other hand, probably showing a preference for T, is unlikely to mediate stimulatory effects on spermatogenesis, a situation clearly different from the one in mammals (17, 18, 46). The fact that the 11-oxygenated androgens and T were able to stimulate the development of the seminal vesicles suggests that the latter may be subject to a dual regulation by these two different classes of androgenic steroids.

All androgens except for DHT reduced the testicular OHA secretory capacity. In view of the regularly high concentration of OHA in immature testes (see above), rendering it unlikely that OHA itself is inhibitory, we again assume that among the 11-oxygenated androgens, KT was the active compound to reduce the androgen secretory capacity. No information is available for the African catfish in regard to the mechanism(s) of action involved, although KT may use a mechanism of action distinct from the one used by T or A, as discussed above. However, our data suggest that androgen treatment may have suppressed the expression or activity of key steroidogenic enzymes, analogous to the situation in the mouse, where T decreased the steady-state mRNA levels and the synthesis of the 17-hydroxylase/C_17-20 lyase enzyme (28). Moreover, ultrastructural data suggest that the reduced androgen secretion is associated with cytological changes compatible with a reduced Leydig cell activity (e.g., reduced cell size; Cavaco, unpublished results).

In regard to the African catfish, it has already been discussed that the effects of the 11-oxygenated androgens on steroidogenesis and spermatogenesis are probably direct effects on the testes, as plasma LH levels were not affected by these androgens, although T, A, or E₂ treatment did significantly decrease plasma LH levels (5). However, whereas T and A also decreased androgen production, E₂ and DHT did not. This indicates that the actions of T and A to reduce androgen production are not dependent on aromatization or reduction to E₂ and DHT, respectively. These data also support our suggestion that at least part of the actions of T and A on testicular function are independent of pituitary LH.

We have no evidence to explain the apparently OHA-specific reduction of the fold-increase of LH-induced OHA secretion over basal (Table 3). However, the treatment with OHA leads to increased plasma levels of both OHA and KT, which may be an endocrine signal distinct from elevating KT only.

In mammals, follicle-stimulating hormone (FSH) and T are key hormones needed for spermatogenesis. FSH promotes, for example, spermatogonial division and meiosis or is involved in controlling germ cell apoptosis (17). There is no information on the role of the FSH-like gonadotropin for spermatogenesis in fish, except that receptor binding for salmon FSH was localized to Sertoli cells (22). In African catfish, an FSH-like gonadotropin has not yet been detected despite repeated trials (38), and we have recently proposed that LH might fulfill all functions that depend on gonadotropic regulation (44). However, what exactly these functions are, next to regulating Leydig cell steroid production, is presently unknown. The intratesticular importance of T is well established in mammals. For example, substantial restoration after complete regression of spermatogenesis can be achieved by administration of T to GnRH-immunized rats (18) or gonadotropin-deficient mice (46). This role for T in mammalian spermatogenesis is likely fulfilled by KT in many teleosts (1, 19-21, present study). Gonadotropin-induced spermatogenesis in immature Japanese eel was accompanied by an activation of Leydig cells and increasing KT blood levels (33). In the African catfish, Leydig cells are already very active before the beginning of spermatogenesis and produce OHA (39); circulating OHA can be converted to KT in the liver (4, 15). It seems that in species as distantly related as the African catfish and the Japanese eel, KT is an important endocrine signal to promote spermatogenesis, although the mechanisms to provide KT can differ between species.

With regard to the concept of the endocrine regulation of puberty, the present data suggest that in the African catfish the production of KT and/or the expression of its cognate receptor(s) are (part of) the missing link for the initiation of puberty.

Perspectives