Androgen metabolism in skin and skeletal muscle of the rainbow trout (Salmo gairdnerii) and in accessory sexual organs of the spur dogfish (Squalus acanthias).

The in vitro metabolism of testosterone, 4-androstene-3,17-dione (androstenedione) and dehydroepiandrosterone by skin and muscle from the rainbow trout (Salmo gairdnerii), and by skin and accessory sexual tissues from the spur dogfish (Squalus acanthias) was studied. In trout skin, testosterone was transformed mainly into 5α-dihydrotestosterone together with smaller amounts of 5α-androstane-3α, 17β-diol, androstenedione, 5α-androstane-3,17-dione and androsterone. Androstenedione was transformed mainly into 5α-androstane-3,17-dione with smaller amounts of testosterone, 5α-dihydrotestosterone, androsterone and 5α-androstane-3α,17β-diol. Dehydroepiandrosterone was transformed to 5-androstene-3β,17β-diol with trace quantities of androstenedione and 5α-androstane-3,17-dione. Unidentified polar nonconjugated metabolites and traces of steroid glucuronides were formed from the three substrates. The patterns of steroid metabolism were similar in dorsal and ventral skin, and in dorsal skin from male and female, adult and immature fish. Most of the 5α-reductase activity in the skin was in the dermis, only a small fraction of the total activity being in the epidermis. The trout muscle converted testosterone into 5α-dihydrotestosterone but in much lower yields than did skin.The skin, clasper, sperm sac and vas deferens of an adult male spur dogfish converted testosterone to 5α-dihydrotestosterone and androstenedione, though in much lower yields than did trout skin. Androstenedione was converted into testosterone, 5α-androstane-3, 17-dione and androsterone, while dehydroepiandrosterone was converted into 5-androstene-3β,17β-diol. No metabolism of testosterone was detected in the skeletal muscle of the dogfish.     

Gonadal steroids as modulators of the function of the pineal gland

Sex steroids affect pineal function. Castration or the administration of estradiol, testosterone, or progesterone bring about changes in pineal melatonin synthesis, as well as in other aspects of pineal activity in rats. This paper deals with some of the mechanisms underlying steroid action on pineal metabolism. Estradiol is avidly taken up in vivo and in vitro by the pineal gland. A cytosol binding protein having a K_d of, (0.27 ± 0.08) × 10^?9M for estradiol and requiring SH groups for full expression of binding activity was detected in pineal homogenates. Estradiol uptake by the pineal and the uterus of mature rats attained their minima on the day of proestrus. A priming dose of estradiol increased high affinity binding sites for estradiol in pineal and uterine cytosol by about 150%. Testosterone is also readily taken up and retained in vivo and in vitro by the pineal of orchiectomized rats. Saturation kinetics analysis of androgen binding to bovine pineal cytosol indicated a K_d of 1.57 × 10^?9M for testosterone and 0.36 × 10^?9M for 5α-dihydrotestosterone. The cytoplasmic binder was shown to be a protein and to require SH groups for binding activity. Testosterone was metabolized by the pinealocytes into 5α-dihydrotestosterone, 5α-androstanediol, androstenedione and 5α-androstanedione; 5α-dihydrotestosterone was the major metabolite in the nuclei of pineal cells. Superior cervical ganglionectomy or exposure of rats to light increased androstenedione synthesis, whereas ganglionectomy or exposure to darkness decrease 5α-androstanedione synthesis. Progesterone is taken up by the pinealocytes in vitro up to a tissue concentration 18-fold that of the media. [³H]Progesterone is converted in the pinealocytes into 5α-pregnanedione and 3α-hydroxy-5α-pregnan-20-one. These data suggest that the early steps of sex steroid action on pineal cells resemble those of steroid target tissues.     

In vitro inhibition of aldosterone biosynthesis by 17-hydroxylated steroids in rat adrenal

The in vitro effect of the addition of non-17-hydroxylated and 17-hydroxylated steroids on the production of aldosterone has been studied in the rat. The biosynthesis of deoxycorticosterone, corticosterone, and aldosterone was increased when progesterone or pregnenolone was added to the incubation media, but neither cortisol nor 11-deoxycortisol was produced. The addition of 17α-hydroxyprogesterone or 17α-hydroxypregnenolone to the incubation media resulted in the biosynthesis of cortisol and 11-deoxycortisol by the rat adrenal; concomitantly, a decrease in aldosterone production occurred. However, no decrease in aldosterone production resulted when cortisol and 11-deoxycortisol were added directly to the incubation media. These studies show that neither cortisol nor 11-deoxycortisol inhibits the in vitro biosynthesis of aldosterone and suggest that other mechanisms are involved.     

Effect of aminoglutethimide on extraglandular metabolism of exogenous testosterone

Aminoglutethimide (AG), an inhibitor of steroid biosynthesis, seems to have an extraglandular site of action on steroid catabolism. To study this effect, five males with peripheral hypogonadism were first given testosterone propionate and then the same dose was repeated combined with AG and urinary testosterone, and its metabolites were measured. AG was shown to have a very evident effect on the peripheral degradation of exogenous testosterone. This may be responsible for a few signs of virilization and fetal masculinization in women taking AG.     

Androgen-binding proteins in rat epididymis: properties of a cytoplasmic receptor for androgen similar to the androgen receptor in ventral prostate and different from androgen-binding protein (ABP)

The cytoplasmic receptor (CR) in rat epididymal 105,000 $\text{g}$ supernatant was separated from the androgen-binding protein (ABP) by gel electrophoresis following labeling with [1,2,6,7-³H]-testosterone in vivo. ABP disappeared from epididymal supernatants after castration 01 hypophysectomy, while CR remained unchanged. CR was evenly distributed between caput and cauda, while much more ABP was present in caput. Properties of CR in epididymis and prostate were similar and distinctly different from ABP. Binding to CR was destroyed by charcoal treatment (1 mg/mg protein) of supernatant at 0 °C for 6 h, heating at 50 °C for 30 min, or exposure to the sulfhydryl blocking reagent, $\text{p}$-chloromercuriphenylsulfonate (1 mM) at 25 °C for 30 min, while binding to ABP was unaffected. The isoelectric pH of CR (5.8) was higher than that of ABP (4.6). Dissociation of radioactive 5α-dihydrotestosterone (DHT) from CR and nuclear receptors was extremely slow (half-time at 0 °C >2 days), while dissociation from ABP was rapid (half-time at 0 °C ~ 6 min). Cyproterone acetate (250 mg/100g body weight) inhibited binding to CR both in epididymis and ventral prostate but did not affect binding to ABP. Nuclear uptake was inhibited by cyproterone to the same extent as binding to CR, indicating that nuclear uptake and binding are dependent on CR and independent of ABP. The time-course of uptake and binding in epididymal supernatant and nuclear fractions was essentially the same 1 day after bilateral castration when both CR and ABP were present or 8 days after castration when CR alone was present. It is concluded that the cytoplasmic receptor for androgen in rat epididymis has properties very similar to the androgen receptor in ventral prostate but different from ABP.     

Small doses of 5α-dihydrotestosterone mimic the effects of 5α-androstane-3β,17β-diol on acid phosphatase activity in the adult rat prostate gland

These studies were designed to further investigate whether 5α-androstane-3β,17β-diol was exerting unique effects on rat prostate acid phosphatase activity or could possibly be exerting its actions by a small peripheral conversion to 5α-dihydrotestosterone. Intraperitoneal administration of 5α-dihydrotestosterone in doses of 1 mg, 100 μg or 50 μg per day starting 7 days after castration led to the restoration of normal characteristics of acid phosphatase activity. However, when 5α-dihydrotestosterone was given in a dose of only 25 μg per day starting 7 days after castration, the changes in acid phosphatase activity were indistinguishable from those found when 5α-androstane-3β,17β-diol was administered in a dose of 2 mg per day. This suggests that the effects of 5α-androstane-3β,17β-diol can be explained by its conversion to small amounts of 5α-dihydrotestosterone.     

Androgen biosynthesis in vitro by testes from amphibia.

Testes from three species of anuran Amphibia, Bufo marinus, Rana catesbeiana, and Rana esculenta, were incubated with radioactive pregnenolone, progesterone, and testosterone. In all incubations the major metabolite was dihydrotestosterone, which accounted for 30–47% of the initial radioactivity after a 3-hr incubation. In addition, 5α-androstanedione and 5α-androstane-3β,17β-diol were formed from all three substrates with testes of Bufo marinus and Rana catesbeiana. Testes of Rana esculenta however converted only testosterone into 5α-androstanedione and 5α-androstanediol, pregnenolone and progesterone being transformed to 5α-pregnane-3β,17α,20ξ-triol. Since in both R. catesbeiana and B. marinus significantly higher yields of 5α-androgens were obtained from pregnenolone and progesterone than from testosterone, it is possible that at least some of these compounds arise from a biosynthetic pathway not involving testosterone.     

Action des hormones androgenes sur l'epididyme d'un reptile lacertilien, Lacerta vivipara jacquin. Effets de la testosterone et de ses principaux mètabolites en culture organotypique

The lizard epididymis is controlled by the testis, as is demonstrated by castration. It selectively retains testosterone and is capable of converting this hormone into 5α-dihydrotestosterone and 5α-androstanediols. Furthermore, it shows marked modifications of structure, particularly of the epithelium, during the annual cycle and is characterized by a copious secretion when active. In organotypic culture of regressed cells from castrated animals we have obtained secretory cells of normal size after treatment with a variety of androgens. Different degrees of stimulation were observed: 5α-DHT and 3β-androstanediol are more potent than 3α-androstanediol and the latter is more potent than testosterone.Electron-microscopic examination of treated tissue shows that 5αDHT produces an increase in nuclear and nucleolar size and conspicuously stimulates the production of rough endoplasmic reticulum.It is suggested that this organ might be used as a model to study the mechanism of androgen action.     

Cortisol metabolism in lymphocytes from cancer-bearing patients

A study was conducted to determine the pattern of cortisol metabolism by lymphocytes obtained from four groups of subjects: 27 male and female patients suffering from various types of malignancy other than malignancy of lymphatic tissues; and 26 healthy male and female controls. Known concentrations of cells were incubated with 1,2-³H-cortisol and the products were isolated by thin-layer and paper chromatography. Three metabolites were found to be produced by lymphocytes from both normal and cancer-bearing patients: 20α-hydroxycortisol, 20β-hydroxycortisol, and tetrahydrocortisol. Cells from the female control group were found to be more active than those from the male controls, while cells from cancer-bearing patients were markedly more active than the normal cells, regardless of sex. It is suggested that this finding of increased metabolism of cortisol by lymphocytes from patients with different types of malignancy other than lymphoma may provide the basis for a new diagnostic aid.     

Yolk androgen deposition in rockhopper penguins, a species with reversed hatching asynchrony

To maximize fitness, females should invest optimally in the siblings within a litter or brood and adapt this investment to environmental conditions. Chick mass and yolk androgens have been shown to influence the outcome of sibling competition. In birds, asynchronous hatching plays a major role in this process and often leads to brood reduction. We studied maternal deposition of yolk androgens in eggs of southern rockhopper penguins (Eudyptes chrysocome chrysocome). Contrary to other avian models, laying and hatching sequences do not coincide in this species, which exhibits reversed hatching asynchrony. This provides a unique model to test whether the first egg to hatch (B-egg), which is the most likely to survive, differs in composition from the second egg to hatch (A-egg). We found that B-eggs had higher egg masses, yolk masses, yolk androgen concentrations and total yolk androgen amounts than A-eggs. This was observed consistently for the three androgens analyzed (testosterone, androstenedione and 5α-dihydrotestosterone). Laying date affected androgen deposition into A- and B-eggs differently. Interestingly, late clutches had proportionally higher androgen levels in the B-egg compared to the A-egg than early clutches. We discuss these results in relation to the chronology of egg formation and the potential effect of the observed differences on embryo development and brood reduction.     

SERUM PREGNENOLONE, PROGESTERONE, 17α-HYDROXYPROGESTERONE, ANDROSTENEDIONE, TESTOSTERONE, 5α-DIHYDROTESTOSTERONE AND ANDROSTERONE DURING PUBERTY IN BOYS

We have determined the concentrations of pregnenolone, progesterone, 17α-hydroxyprogesterone, androstenedione, testosterone, 5α-dihydrotestosler-one and androsterone in serum samples collected from a total of seventy-nine boys who were 8–18 years of age at the time the first blood samples were drawn. Additional samples were drawn from sixty-six and forty-four of these boys after 1 and 2 year intervals, respectively. The first increases in serum steroid concentrations were those of androstenedione and androsterone, thus supporting the hypothesis that an early activation of the adrenal cortex is the first hormonal change in puberty. This increase in serum androstenedione occurred 2 years earlier than the first significant increase in serum testosterone, which took place by 13 5 years, after the first signs of external genital development, but in concert with the onset of pubic hair growth. The serum concentrations of 5α-dihydrotestosterone were closely related to those of testosterone, but the relative increases were smaller, the consequence of which is a decrease in the ratio of 5α-dihydrotestosterone: testosterone throughout puberty. The main increase in serum androsterone tends to take place after that of 5α-dihydrotestosterone and testosterone. In comparison with the four androgens, the serum concentrations of their C₂₁ precursors correlated rather poorly with the physical signs of advancing puberty. As in the case of androstenedione, the concentrations of pregnenolone and 17α-hydroxyprogesterone were relatively high in the youngest age groups, which probably also reflects an early adrenal activation. In contrast to testosterone, the major increases in precursor steroids occur at a relatively late stage of puberty. It seems likely therefore that a major qualitative shift in the testicular secretion of steroids occurs during puberty in boys.     

High-affinity binding of the antiestrogen [3H]tamoxifen to the 8S estradiol receptor

The interaction of tamoxifen (ICI 46,474), a synthetic antiestrogen, with uterine cytosol proteins of immature calf and rat has been studied directly using the tritiated compound labeled with a high specific activity. The binding complexes were measured by the dextrancoated charcoal, protamine sulfate and hydroxyapatite assays. Scatchard plots revealed a single class of high-affinity (KD congruent to 1.7 nM) binding sites, with a binding capacity similar to that of estradiol. Competitive experiments showed the same binding specificity for estrogens and antiestrogens. Sucrose gradient analysis revealed an 8S binding protein which could be partially proteolysed by trypsin into a 4S binding protein. Kinetic studies showed that the association rate of tamoxifen was 5 times lower than that of estradiol and reacted according to a second order kinetics. The first-order kinetics of dissociation was considerably higher than that of estradiol, giving a half-dissociation time of 20--40 min at 0--2 degrees C. In some cases tamoxifen displayed two slopes of dissociation, but the proportion of the slow-dissociating complex was always inferior to that found with estradiol. In contrast to estradiol, the kinetic constants ratio (k-/k+) gave a calculated dissociation constant, similar to that determined in equilibrium conditions (KD), agreeing with a simple reactional scheme. We conclude that the antiestrogen tamoxifen binds directly to the 8S cytosol receptor for estrogens and not to another receptor for the antagonists. In contrast to estradiol, the antagonist is rapidly dissociated from the receptor sites and is unable to protect them against thermal inactivation. The affinity of tamoxifen for its receptor sites as determined directly is surprisingly high when compared to its affinity evaluated indirectly by competitive experiments. It is then suggested that the two ligands either bind on two different sites of the same protein or induce a different conformational change of the same binding site.     

Photoperiodic control of testosterone metabolism, plasma gonadotrophins, cloacal gland growth, and reproductive behavior in the Japanese quail.

The changes in plasma gonadotrophins and testosterone, in cloacal gland area, and in reproductive behavior were observed in male Japanese quails after transfer to long days or testosterone implantation and were correlated with the testosterone metabolism in the hyperstriatum, hypothalamus, pituitary, and cloacal gland. Long days stimulate the growth of the cloacal gland and at the same time enhance its production of androstenedione from testosterone. This increased 17β-hydroxysteroid dehydrogenase activity is correlated at the individual level with cloacal gland area but not with plasma testosterone. Similarly the changes observed in some reproductive activities (aggressive behavior and struts) are correlated with individual differences in the brain metabolism of testosterone (hyperstriatal or hypothalamic production of androstenedione), but not with plasma testosterone. The plasma luteinizing hormone is also related to testosterone metabolism rather than to the circulating level of the hormone (negative correlation with the production of 5α-dihydrotestosterone in the pituitary). The testosterone metabolism in target organs thus appears of critical importance in the control of reproduction and seems largely responsible for the occurrence of individual differences. After exposure to 12 long days the mean testosterone metabolism in the pituitary is also strongly affected (increased 5β-reduction) though not in the same way as after 2 months of photostimulation. It is suggested that these metabolic changes at the pituitary level may play some role in the induction of the increased responsiveness to luteinizing hormone-releasing hormone which can be observed in quails after exposure to 7 long days.     

Some properties of androgen-binding activity in rat testis.

High-affinity (Ka approximately equal to 5 X 10(8) M-1 for testosterone) androgen-binding activity in rat testis was shown to have a rapid dissociation rate constant (t1/2 = 3 min, 0 degrees C, 30% glycerol buffer) using dextran-coated charcoal to separate bound from free hormone. Because of this fact, exchange of endogenous and labeled hormone was complete in the assay incubation time (16 h, 0 degrees C) and Scatchard plots of the high-affinity binding data were shown to measure total as contrasted to available sites. The binding was highly specific for androgens. Polyacrylamide gel electrophoresis separated high-affinity androgen-binding protein (Rf 0.54) from albumin (Rf 0.62). Binding site estimates under saturating conditions or by Scatchard analysis of electrophoresis data utilizing [3H]dihydrotestosterone agreed reasonably well with estimates made by the charcoal technique using [3H]testosterone.     

Progestins inhibit fsh-stimulated steroidogenesis in cultured rat granulosa cells

It has been hypothesized that progesterone (P) exerts a direct inhibitory effect on ovarian follicular development, an effect which could be mediated by P receptors located in granulosa cells. We tested this hypothesis by examining the effect of several progestins on FSH-stimulated estrogen (E), P, and 20α-dihydroprogesterone (DHP) production by cultured rat granulosa cells, and correlated the results with the ability of the progestins to bind to the granulosa cell P receptor. Granulosa cells from immature hypophysectomized DES-treated rats produced 9 ng/ml E, 21 ng/ml P and 29 ng/ml DHP during a 2-day incubation in McCoy's 5a medium containing 10^?7 M androstenedione and 10 ng/ml oFSH. The FSH-induced increase in E production was inhibited by 50 and 95% following concomitant treatment with 3 × 10^?6 and 10^?5 M resp. of R5020, a potent synthetic progestin. Added R5020 at these concentrations also significantly inhibited P and DHP production. R5020 had no effect on granulosa cell viability or plating efficiency, and the inhibitory action of R5020 on E production was reversible. In studies of the specificity of the progestin inhibitory action, the relative abilities of various progestins to inhibit E production were: R5020 > P > DHP > 17α-hydroxyprogesterone (170HP). The relative abilities of these progestins to bind to the ovary P receptor were also: R5020 > P > DHP > 170HP. These results indicate that exogenous progestins directly inhibit the FSH-stimulation of granulosa cell steroidogenesis in vitro and suggest that the progestin effect may be mediated by the P receptor. Such results offer a possible mechanism whereby progesterone could exert a direct but reversible inhibitory action on ovarian follicular development.     

Competitive binding studies with glucocorticoid receptors from rat-thymus cells: differential temperature-dependence of steroid binding.

Competitive steroid-binding studies were performed with intact rat thymus cells and with cytosol preparations at different temperatures using [1,2-3H]dexamethasone as the labelled ligand. Steroids lacking a 17 alpha-hydroxyl group, such as corticosterone, were better able to compete with [1,2-3H]dexamethasone for binding to glucocorticoid receptors at 0 degrees C than compounds containing a 17 alpha-hydroxyl substituent, such as cortisol. At 37 degrees C the reverse was true. This temperature-dependent change in relative affinities appeared to be unrelated to steroid metabolism or receptor activation, and to depend only on the thermodynamic parameters of the steroid--receptor interaction. Relative biological activities for different steroids agree more closely with the relative affinities determined at 37 degrees C than with those determined at lower temperatures.     

Effect of androgens on the biosynthesis of estradiol-17beta by isolated periovulatory rat follicles.

To analyse the mechanism for the earlier reported decline in estrogen synthesis by the periovulatory rat follicle, prepubertal rats injected with 8 IU PMS on day 28 were killed following the endogenous gonadotrophin surge (15.00–18.00) on day 30. Isolated preovulatory follicles were incubated for 2 h in a chemically defined medium. Steroids were measured by specific RIA methods. Follicles exposed in vivo to the gonadotrophin surge and extirpated 19.00–22.00 h on day 30 secreted significantly lower amounts of androstenedione, testosterone and estradiol-17β but significantly higher amounts of progesterone than did follicles extirpated from rats in which the gonadotrophin surge had been prevented by a Nembutal injection. Secretion of estradiol-17β by follicles isolated following the endogenous gonadotrophin surge remained low when LH, FSH or dibutyryl cyclic 3′,5′-AMP was added to the medium. However, the addition of testosterone (0.1–1 μg/ml) or androstenedione (1 μg/ml) to the incubation medium restored estradiol biosynthesis to values similar to those seen prior to gonadotrophin exposure. There was no effect of 5α-dihydrotestosterone or 17α-hydroxyprogesterone on the estradiol-17β synthesis. The results indicate that cleavage of the 17:20 sidechain rather than the aromatase enzyme limits estradiol synthesis in the periovulatory follicle following the gonadotrophin surge. It is suggested that the combined action of LH and FSH of the gonadotrophin surge might explain the lack of inhibitory effect on the aromatase enzyme recently reported by Katz and Armstrong (1976) 6–8 h after the injection of LH.     

Testosterone restores ovarian aromatase activity in rats treated with a 17,20-lyase inhibitor.

Immature female rats were primed with a non-ovulating dose of pregnant mare serum gonadotropin (PMS), at 24 days of age, and later treated by an antisteroidogenic drug — either 17β-acetamidoandrost-4-ene-3-one (AAA), a competitive inhibitor of the steroid 17,20-lyase, or aminoglutethimide phosphate (AGP), an inhibitor of hydroxylase-associated cytochrome P-450 action. Both drugs initiated a fall in the ovarian androgen content, concomitant with decreased ovarian aromatase activity, manifested in vivo and in vitro. Testosterone therapy, administered to the antisteroidogenic drugs-treated rats, elevated the ovarian estradiol-17β production only in the AAA-treated ones. This elevation, detectable both in vivo and in vitro, was to values indistinguishable from controls. No such accomplishment was observed in the AGP-treated rats. Analysis of the AGP-treated rats' ovaries, 9 h after the administration of the drug, with and without a testosterone supplement, revealed residual amounts of the drug sufficient to inhibit the aromatase, both in vivo and in vitro. The data support the theory suggesting that reduced androgen levels, due to the deactivation of the 17,20-lyase, result in cessation of estrogen secretion from the ovary.     

In vitro metabolism of androstenedione and identification of endogenous steroids in Helix aspersa

In vitro metabolism of androstenedione in gonads of juvenile and adult Helix aspersa has been investigated. The conversion of [3H]androstenedione into testosterone, 5 alpha-dihydrotestosterone, androsterone, and estriol was demonstrated. In juvenile animals testosterone (59.8%) is the major metabolite whereas in adult animals androsterone (18.8%) is. The following endogenous steroids have been identified by gas chromatography-mass spectrometry in adult gonads: androsterone, dehydroepiandrosterone, androstenedione, 3 alpha-androstanediol, estrone, estradiol-17 beta, and estriol. The levels of testosterone, 5 alpha-dihydrotestosterone, androstenedione, and progesterone have been measured by RIAs in gonads and hemolymph. Their levels vary with the physiological stage: the gonadal and circulating levels of testosterone decrease with the sexual maturation whereas the 5 alpha-dihydrotestosterone increases. These differences observed in metabolism and in level of steroids between the juvenile and the adult snails allow us to suppose that these steroids have a biological role.     

The effect of temperature on the hepatic catabolism of testosterone in the rainbow trout (Salmo gairdneri) and the goldfish (Carassius auratus)

Liver of the rainbow trout Salmo gairdneri and the goldfish Carassius auratus has been incubated with [³H]testosterone at a range of temperatures from 1 to 46°. With trout liver the main metabolites were identified as androstenedione, testosterone glucuronide, 5β-dihydrotestosterone glucuronide, 5β-androstane-3α, 17β-diol glucuronide, and 6β-hydroxytestosterone glucoronide. With goldfish liver the main products were testosterone glucuronide and androstenedione. With both species the effect of temperature on testosterone glucuronide formation was very similar to that found in the testis, maximum yields being obtained at 31 and 41° for the trout and the goldfish, respectively. An inverse pattern was observed for the recovery of unconjugated testosterone, indicating a more rapid removal of free androgen at higher temperatures. It is suggested that the hepatic glucuronyl transferase may act in conjunction with the testicular enzyme to regulate free plasma androgen levels so that reproductive development takes place at the environmentally most favourable temperature.