Inhibition of Spermatogonial Differentiation by Testosterone

In this review we describe a variety of pathological conditions in rodents that result in seminiferous tubule atrophy, and which are characterized by the absence of all germ cells except for type A spermatogonia. In many such cases, suppression of gonadotropins and testosterone with gonadotropin-releasing hormone (GnRH) analogues restores spermatogonial differentiation and spermatogenic progression. In some cases, spermatogenesis is maintained after the cessation of hormonal treatment and fertility is restored. We and others have shown that the hormones normally responsible for the maintenance of spermatogenesis—testosterone, and in some cases, follicle-stimulating hormone (FSH)—actually inhibit spermatogonial differentiation in these conditions. This inhibitory function is a completely new role for androgens in the testis. It has long been known that systemic administration of low levels of testosterone can inhibit the completion of spermatogenesis (Steinberger, 1971) as a result of decreasing gonadotropin levels, thereby reducing testosterone production by the Leydig cells and actually reducing intratesticular testosterone (ITT) concentrations. In the pathological conditions we and others have described, however, the ITT concentration is not reduced and it is responsible for the inhibition of spermatogonial differentiation. As will be described later, it is not clear whether the direct action of testosterone is to block an actual step of differentiation of the spermatogonia or to cause the apoptosis of the spermatogonia prior the step at which they would differentiate. However, throughout this review we will use the concept of "inhibition of spermatogonial differentiation" to encompass both possibilities. The stem spermatogonia, designated As, maintain their numbers by self-renewal, and some differentiate to form by sequential divisions Apr (A-paired), and Aal−4 and Aal−8 (A-aligned) spermatogonia, which go on to produce A1 spermatogonia. This differentiation may be blocked in 3 ways. In one way, which is the focus of this review, undifferentiated spermatogonia proliferate but their numbers remain relatively constant because of apoptosis (Figure 1) (Allard and Boekelheide, 1996; Shuttlesworth et al, 2000). We will call this the proliferation-apoptosis (PAp) block to distinguish it from the other 2 ways. The second type of block in spermatogonial differentiation, which is caused by vitamin A deprivation, is characterized by spermatogonial differentiation to the Aal stage, but then proliferation ceases and spermatogonia can remain at this stage for a period of only several weeks (van Pelt and de Rooij, 1990) and is designated Ar (arrest). In the third kind of block, which is observed in several types of transgenic mice, including bax-deficient, or bcl-2-overexpressing, or glial cell line-derived neurotropic factor (GDNF)-overexpressing mice (Knudson et al, 1995; Furuchi et al, 1996; Meng et al, 2000), type A spermatogonia proliferate and accumulate but produce few differentiated cells, and is designated proliferation-accumulation (PAc). . Outline of stem cell kinetics in (A) normal rodents and (B) rodents with a proliferation-apoptosis block in spermatogonial differentiation as described for toxicant-treated rats, cryptorchid mice, and some mutant mice. In normal rodents, no apoptosis is observed at these stages, and the Aal and some Apr spermatogonia are induced to undergo differentiation into A1 spermatogonia at stage VII—VIII of the cycle of the seminiferous epithelium. In the rodents with the PAp block, spermatogonia of all clonal sizes undergo apoptosis, with the probability of undergoing an apoptotic event, as opposed to a mitotic division, increasing with chain length. The precise relationship between these 3 blocks in rodents and the clinical phenotype of spermatogonial arrest in humans, which is often the result of hypogonadotropism (Johnsen, 1970), is not known. However, the types of spermatogonia present in humans and their proliferative status have not been studied. A variety of testicular toxicants produce similar testicular histology in rats consistent with the PAp type of block. These agents include hexanedione (Boekelheide and Hall, 1991), boric acid (Ku et al, 1993), radiation (Kangasniemi et al, 1996), procarbazine (Meistrich, 1999), dibromochloropropane (DBCP; Meistrich, unpublished results), and indenopyridines (Hild et al, 2001). The type A spermatogonia proliferate in atrophic tubules but they do not accumulate because they continue to be lost by apoptosis many months after the original acute or subchronic exposure. Atrophic tubules with actively dividing stem type A spermatogonia were also observed in testis cross-sections from 27-month-old Brown-Norway rats (Schoenfeld et al, 2001). The failure of these cells to differentiate is in part responsible for the decline in spermatogenesis with age in these rats. In contrast to that of rats, brief exposures to such toxicants do not induce such a block in spermatogonial differentiation in mice. Whereas 3.5 Gy of irradiation was sufficient to induce this block in LBNF1 rats, mouse spermatogonia maintain their ability to differentiate even after doses of 12 Gy (Meistrich et al, 1978). But a block in spermatogonial differentiation can be induced by continuous elevation of temperature. In cryptorchid C57BL/6 mouse testes, spermatogenesis fails to progress past the Aal spermatogonial stage (Haneji et al, 1983); these cells actively proliferate but die by apoptosis (de Rooij et al, 1999). Similar blocks in the differentiation of A spermatogonia were also observed in jsd (juvenile spermatogonial depletion) mice (Beamer et al, 1988) and in Sl17H mice, which have an altered form of stem cell factor in Sertoli cells (Brannan et al, 1992). In these mice, an initial wave of spermatogenesis is not maintained, so that the adult testis tubules contain only Sertoli cells and type A spermatogonia; the latter proliferate but die by apoptosis (de Rooij et al, 1999). In addition, certain other mouse mutants, including XO-Sxrb (Sutcliffe and Burgoyne, 1989) and Dazl (Schrans-Stassen et al, 1999), have PAp blocks in spermatogonial differentiation but they differ from the above 2 models in that this condition is apparent by postnatal day 10 and there is no initial wave of spermatogenesis. It was indeed surprising that such a wide variety of toxicant exposures, conditions, and genetic mutations produced such a similar phenotype. For example, some of these toxicants, such as irradiation, are believed to act directly on germ cells (Lee et al, 1999), whereas other toxicants, such as hexanedione, are believed to act on Sertoli cells. Furthermore, at least in some instances, the block to spermatogonial differentiation does not begin to develop for almost 6 weeks after the insult. Both these observations imply that the block to spermatogonial differentiation is not a direct consequence of the initial event, but that different initiating events produce a common outcome, which in turn, leads to the block. Although the PAp blocks in spermatogonial differentiation caused by different agents have much in common (Figure 1), they show some quantitative differences in terms of the stage to which spermatogonia differentiate before undergoing apoptosis. The type A spermatogonia in atrophic testes were first identified following exposure of either Sprague-Dawley or Fischer F344 rats to hexanedione (Boekelheide and Hall, 1991). Stem cells (isolated type A spermatogonia), although reduced in number from controls, still constituted a substantial proportion of these remaining A spermatogonia (Allard et al, 1995). The A spermatogonia were in active proliferation, but their numbers remained constant because they underwent apoptosis (Allard and Boekelheide, 1996). Calculations based on numbers of stem cells and total spermatogonia indicated that the cells were progressing to the A2 or A3 spermatogonial stage (Allard and Boekelheide, 1996). In contrast, direct, whole-mounted tubule analysis of mitotic clones of A spermatogonia in irradiated LBNF1 rats revealed that most of the clones were isolated or paired A spermatogonia and few had a clone size greater than 4, indicating that they were early progeny from the stem cells (Shuttlesworth et al, 2000). Very few clones progressed to become Aal−8 and Aal−16, which are the clone sizes that most often undergo differentiation in normal rats (Figure 1A) because the probability of apoptosis increased as clone size increased (Figure 1B). Thus, failure of spermatogonia to differentiate appeared to be a consequence of their undergoing apoptosis first. In jsd, Sl17H, and cryptorchid mice, the clones of A spermatogonia in whole-mounted tubules were arranged as 1 to 16 cells (de Rooij et al, 1999). There were appreciable and similar numbers of clones of Aal−8 and Aal−16 in all 3 models. These undifferentiated A spermatogonia were proliferating, but they did not accumulate, and the larger clones in particular underwent apoptosis. Because the clone sizes indicate that spermatogonia develop to the point at which the Aal cells should differentiate into A1 spermatogonia, the failure to do so indicates the lack of a signaling system rather than prior apoptosis. The difference in numbers and stage of development of spermatogonia between the irradiated rat and the mouse models appears to be real because the same methodology was employed. It is not known whether these differences are due to how mice and rats respond to blocks at the spermatogonial level or whether differences in the cause of the blocks. The difference in stage at which the block was reported to occur in irradiated vs hexanedione-treated rats could be a result of the different analytical methods employed, in addition to the possible contributions of rat strain or the nature of the original toxic insult. As is typical in cases of testicular tubular atrophy, FSH and luteinizing hormone (LH) levels rise in most cases in which only type A spermatogonia remain in the tubules. FSH levels were elevated 1.5-fold to 2-fold and LH levels were elevated 2-fold to 4-fold after treatment of rats with hexanedione (Boekelheide and Hall, 1991), gamma radiation (Kangasniemi et al, 1996), procarbazine (Meistrich et al, 1999), indenopyridine (Hodel and Suter, 1978), boric acid (Ku et al, 1993), and DBCP (Meistrich, unpublished results) and in jsd mice (Shetty et al, 2001). In all the cases studied, serum testosterone remained unchanged. It has been shown that when the germ cells in the testes are lost, testicular mass and, consequently, blood flow decline (Wang et al, 1983). The maintenance of serum testosterone levels is a result of the hypothalamic-pituitary axis acting to keep serum testosterone constant when there is a decline in testicular blood flow by adjusting LH levels accordingly. This results in a 2.5-fold to 3-fold increase in ITT concentrations, which was confirmed in irradiated, procarbazine-treated, and DBCP-treated rats and jsd mice. The greater proportion of Leydig cells (their numbers are not decreased) in the testis, the decreased clearance rate of newly produced testosterone from the testis, and the elevated LH levels are all responsible for the increase in ITT concentrations. There were 2 exceptions to this pattern of hormone changes. First, in cryptorchid mice, FSH was elevated 1.5-fold, but LH was unchanged (Mendis-Handagama et al, 1990). Second, in aged rats, both serum and testicular interstitial fluid testosterone levels were depressed (Schoenfeld et al, 2001). This depression in testosterone levels may be a combined result of the general depression with aging in LH and Leydig cell function, which can no longer respond by increasing testicular testosterone production. Nevertheless, these results show that above normal levels of ITT are not necessarily required for inhibition of spermatogonial differentiation, which will be discussed later. We first demonstrated the stimulation of recovery of spermatogenesis in rats using hormone treatment given after irradiation (Meistrich and Kangasniemi, 1997). All previous studies had focused on the possible protective effect of giving the suppressive hormones before the toxicant exposure (Ward et al, 1990). However, we ruled out many possible mechanisms (Meistrich et al, 1997) by which the hormone treatment could have protected the survival of the spermatogonia and concluded that the only explanation that fit the data was that the hormonal treatment given before the toxic insult helped somatic cells to support sustained recovery of spermatogenesis from surviving stem cells after the toxicant exposure (Meistrich et al, 2000). In all subsequent work we have focused on giving the GnRH analogue treatment after toxicant exposure, although others have given the hormones before and after the toxicant. In our initial study (Meistrich and Kangasniemi, 1997), the tubule differentiation index (TDI; the percentage of tubule cross-sections containing differentiated cells) was only 37% at 10 weeks after 3.5 Gy irradiation in the absence of hormone treatment. When GnRH agonist treatment was started immediately after irradiation, the TDI at 10 weeks was dramatically increased to 91%. However, because GnRH analogue treatments suppress testosterone, which is required for spermatid differentiation, there is histological recovery to the round spermatid stage, but no sperm are produced. The production of sperm after cessation of a transient GnRH analogue block will be discussed below. We also showed that systemic exogenous administration of testosterone, which suppresses ITT concentrations, also maintains spermatogonial differentiation after irradiation. In other cases involving a toxicant-induced PAp block to spermatogonial differentiation (Table 1) maintenance or recovery of spermatogenesis was enhanced by giving GnRH analogues after the toxicant treatment (Table 2). These include hexanedione, procarbazine, or DBCP. In some other cases, GnRH analogue treatment has also proved beneficial to the maintenance or recovery of spermatogenesis after exposure to a toxicant for which the blocks in spermatogonial differentiation were not well characterized. Treatment with GnRH agonist for about 12 weeks after exposure to the anticancer agent busulfan significantly increased the TDI at week 18 (Udagawa et al, 2001). However, a 4-week hormone treatment prior to busulfan injection was ineffective. The irreversible loss of spermatogenic function that occurred after a single dose of heat to rat testes was likewise counteracted by GnRH analogue posttreatment (Setchell et al, 2001), and treatment with GnRH agonist before heating was also effective (Setchell et al, 2002). Finally, prevention of the indenopyridine-induced block to spermatogonial differentiation was achieved when GnRH analogues were given both before and after drug treatment (Hild et al, 2001). However, in a subsequent study using a GnRH antagonist, only prior, but not subsequent, treatment with the GnRH analogue was effective at restoring recovery of spermatogenesis following indenopyridine treatment (S.A. Hild, personal communication). GnRH analogue treatment also enhanced the stimulation of recovery of spermatogenesis from stem cells following spermatogonial transplantation. When mouse testicular cells were transplanted into busulfan-treated mouse recipients, the efficiency of differentiated germ cell production from transplanted stem cells in the recipient tubules was enhanced with GnRH analogue treatment (Ogawa et al, 1998; Dobrinski et al, 2001). However, a significant benefit was derived only from pretreatment with GnRH analogue, indicating that the hormone treatment may be important for the stem cells to attach in their proper niche in the seminiferous tubules, but not for the initiation of differentiation. The importance of suppressing ITT levels with either GnRH agonist or exogenous testosterone treatment was also demonstrated in studies in which rat or mouse spermatogonia were transplanted into busulfan-treated rat hosts (Ogawa et al, 1999). When GnRH treatment is given relative to the toxic exposure is important. Data from irradiated and hexanedione-treated rats showed that treating immediately after exposure to a toxicant was more effective than delayed treatments in the restoration of spermatogonial differentiation (Meistrich et al, 1999). However, there has not been a strict comparison between the effects of pretreatments and posttreatments in any of the models in which both treatments are effective. Fertility can be restored in these pathological situations by GnRH analogue treatment. When a 10-week GnRH agonist or GnRH antagonist treatment was started immediately after 3.7-Gy irradiation, fertility was maintained at week 20 in the GnRH agonist and GnRH antagonist treated rats at normal and nearly normal levels, respectively, whereas none of the irradiated-only rats were fertile (Meistrich et al, 2001b). When treatment was initiated 10 weeks after 5 Gy irradiation, at which point spermatogenesis had completely declined, fertility was restored at week 30 to subnormal levels in 83% of GnRH agonist and 50% of GnRH antagonist treated rats. Thus we conclude that normal fertility can be restored by GnRH treatment after irradiation, although that may depend on initiation of the GnRH analogue treatment soon after a toxicant exposure that is not too severe. We have also demonstrated that GnRH analogue posttreatment significantly increases recovery of fertility in rats after procarbazine treatment (Meistrich et al, 1999). In contrast in the jsd mice, a transient increase in spermatogonial and spermatocyte differentiation was produced by the GnRH antagonist treatment; testicular sperm extraction and intracytoplasmic sperm injection (ICSI) were both required to produce offspring (Tohda et al, 2002). Although the TDI in rats receiving 3.5 Gy of radiation and GnRH agonist for 10 weeks was 91%, testicular sperm head counts were only 0.1% of controls because the hormone treatment suppressed spermiogenesis. However, when additional time without further GnRH treatment was allowed before the rats were killed, the TDI recovered to 100%, and sperm counts reached about 50% of normal control levels at 6.5 weeks after stopping treatment and were maintained at this level for at least another 3.5 weeks. The maintenance of spermatogenesis in irradiated rats after GnRH analogue treatment is stopped depends on the toxicant dose and time of initiation and duration of the hormone treatment. For example, when a 7-week GnRH analogue treatment was initiated at week 15 after 6 Gy of irradiation, the TDI was elevated from 0% in irradiated-only rats to 95% at week 24 (2 weeks after stopping the GnRH treatment), but then declined to 50% at week 36 (14 weeks after stopping GnRH; G.A. Shuttlesworth and M.L. Meistrich, unpublished data). Thus permanent progression and maintenance of spermatogenesis is not assured by this technique. Although no time course studies were done, extensive recovery of spermatogenesis in tubules after hexanedione treatment was observed 9 weeks after the end of a 10-week GnRH agonist treatment, and the degree of recovery was inversely correlated with the dose of hexanedione (Blanchard et al, 1998). In contrast to the toxicant-treated rat models, spermatogenesis degenerated rapidly in jsd mice after withdrawal of the GnRH antagonist. Whereas a 6-week GnRH antagonist treatment increased the TDI from 11% in non-hormone treated mice to 95%, 5 weeks after cessation of the treatment the TDI progressively declined to 78% and to 8% after 13 weeks (Shetty et al, 2001). Although one wave of late spermatids was produced from the differentiating spermatogonia and spermatocytes that developed during the GnRH antagonist treatment, the maximum percentage of tubules that contained elongated spermatids was only 20% at week 4 after the hormone treatment was stopped (Tohda et al, 2002). However, these elongated spermatids were used in ICSI to effect a pregnancy. The difference between the maintenance of spermatogenesis in the irradiated rat model and jsd mice is that the former likely involves an epigenetic change, whereas the latter is a genetic alteration. The epigenetic change caused by irradiation to render spermatogonial differentiation sensitive to inhibition by testosterone can be largely reversed by hormonal treatment. But the underlying defect in a genetic disorder manifests itself again as soon as the hormone treatment is stopped. Because the GnRH analogues that were used to stimulate or maintain spermatogonial differentiation in the various cases described above generally suppress LH, FSH, and testosterone, these hormones were implicated in the inhibition of spermatogonial differentiation. Using irradiated rat and jsd mouse models, we and others investigated the roles of these hormones in the regulation of spermatogonial differentiation. One study involved the administration of exogenous LH to GnRH antagonist—treated jsd mice (Tohda et al, 2001). Whereas the GnRH antagonist restored spermatogonial differentiation, the addition of exogenous LH inhibited it. However, other experiments with jsd mice (Shetty et al, 2001; Tohda et al, 2001) and with irradiated rats indicated that it was the testosterone production stimulated by the LH, and not the LH itself, that inhibited spermatogonial differentiation. For example, GnRH agonist treatment of LBNF1 rats did not suppress LH levels, but it did suppress ITT, serum testosterone, and FSH levels and stimulated spermatogonial differentiation (Meistrich and Kangasniemi, 1997; Meistrich et al, 1999). In another study, when irradiated rats treated with GnRH agonist were given exogenous testosterone, spermatogonial differentiation was inhibited despite a suppression of LH levels (Shetty et al, 2001). This led us to further investigate the precise roles of testosterone and FSH in the inhibition of spermatogonial differentiation after irradiation. Various studies have indicated that testosterone had an inhibitory effect. Because there is a major increase in the ITT concentration in mice between 30 and 40 days of age (Jean-Faucher et al, 1978), the large decline in the numbers of B spermatogonia in jsd testes, which occurs between 6 and 7 weeks of age (Kojima et al, 1997), could very well be a consequence of the increase in ITT. In addition in these mice, the stimulation of spermatogonial differentiation by suppression of testosterone with GnRH antagonist was reversed by exogenous testosterone (Shetty et al, 2001). Furthermore, that inhibition by testosterone was reversed by treatment with the androgen-receptor antagonist flutamide. In irradiated rats, we have shown that testosterone dose-dependently reduced the GnRH antagonist-stimulated spermatogonial differentiation. (Shetty et al, 2000, 2002). Further, the stimulatory action of low-dose testosterone alone, which reduces ITT concentrations, was also reduced with increasing doses of testosterone that increased both ITT and serum testosterone concentrations. The TDIs and the serum and ITT levels were similar for each given dose of testosterone, with or without the GnRH antagonist, showing that the testosterone levels in the testis or the serum, or both, limit the ability of spermatogonia to differentiate. The inhibition of spermatogonial differentiation by testosterone was further confirmed by showing that flutamide reversed the inhibition induced by exogenous testosterone in GnRH antagonist—treated, irradiated rats (Shetty et al, 2000). Further support for our hypothesis that it is indeed testosterone acting through the androgen and not a of testosterone that spermatogonial differentiation was by showing that various including androgen but one that can be and also suppressed spermatogonial differentiation in GnRH antagonist—treated irradiated rats (Shetty et al, 2002). In the same study, we showed that was not When testicular testosterone levels in irradiated rats treated with various GnRH analogues and testosterone were with the an was observed (Figure with only 1 point significantly from each of the (Figure B and Although a general was also for serum testosterone vs there was a very significant (Figure A and and in which the irradiated rats with a of serum testosterone showed no differentiation. However, irradiated rats treated with GnRH analogues and testosterone (Figure A and and showed serum testosterone, but a significantly percentage of the tubules contained differentiating cells. This led us to conclude that ITT is the major as the irradiated-only rats had much ITT than also treated with GnRH analogues and testosterone (Figure B and and However, there were some but significant in the between ITT and For example, GnRH irradiated rats (Figure showed a TDI but also ITT than a similar of rats that also testosterone Because the former had much serum testosterone levels (Figure we that although the ITT was the major factor spermatogenic serum testosterone to have a inhibitory The point that from the in was a result of treatment of irradiated rats with GnRH antagonist and of testosterone which may result in levels of testosterone throughout the course of treatment. . between serum testosterone and ITT during hormone treatment and the levels of recovery of spermatogenesis at the end of Data from of testosterone with GnRH given during weeks after 6 Gy irradiation. TDI analysis was on testicular histological on week Data from of testosterone with GnRH antagonist, given during weeks after 5 Gy irradiation. TDI analysis was on testicular histological on week in through are from the same were to the data with the of the in and indicate from Data from 2 (Shetty et al, 2000, were In all these ITT in the normal to inhibit the differentiation of spermatogonia. that even ITT of of testis inhibited spermatogonial differentiation. Further, the observed block in the spermatogonial differentiation in aged rats that had ITT normal and spermatogonial differentiation was stimulated by further suppression of ITT with a GnRH agonist show that above normal levels of ITT are not necessarily required for the inhibition of spermatogonial differentiation. in these spermatogonial differentiation sensitive to levels of on the concept that testosterone inhibited spermatogenesis in toxicant-treated rats, rats were treated with which Leydig cells, by GnRH which Leydig cell et al, 2002). reduced testosterone levels to levels, the treatment inhibited the recovery of spermatogonial differentiation that the GnRH agonist would normally Although the results of this study to the hypothesis that testosterone spermatogonial differentiation, that hypothesis could still be a Leydig cell factor is required for the stimulation of spermatogenic recovery in the atrophic testis and this factor were by but not by GnRH analogue treatment. The elevated FSH levels in these pathological models of testicular could to the inhibition of spermatogonial differentiation. Although as shown testosterone appears to be an inhibitory it is to whether FSH also has a The possible of serum testosterone to spermatogonial differentiation that testosterone may act at an One such likely is the it could act by gonadotropin We have ruled out LH as a significant to the inhibition of spermatogonial differentiation, so we focused on a possible role for However, testosterone has a action on production of When testosterone is given to rats or mice that have normal GnRH production and it suppresses FSH levels by a combined action on the and However, when testosterone is given to GnRH antagonist—treated rats, but not mice (Shetty et al, 2001), it the GnRH of FSH levels in these rats by direct of in the et al, The levels of FSH in the of exogenous testosterone to be of whether or not a GnRH antagonist is also given (Shetty et al, 2000). There was