Retinoic acid signaling and the cycle of the seminiferous epithelium

Retinoic Acid Signaling

Testicular RA is primarily made in situ [2] from retinol (ROL), which is derived from retinyl ester stores within Sertoli cells or taken up from the blood where it circulates bound to retinol binding protein (RBP) [3]. Conversion of ROL to RA requires two consecutive oxidation reactions. First, ROL is converted to retinaldehyde in a rapid, reversible reaction after which retinaldehyde is irreversibly converted to RA. The first step is mediated by the retinaldehyde dehydrogenases (RDH), of which there are multiple forms. In the testis, RDH10 appears to be the most highly expressed form. The conversion of retinaldehyde to RA requires the actions of retinaldehyde dehydrogenases (RALDH) and RALDH1 and RALDH2 (encoded by Aldh1a1 and Aldh1a2, respectively) are expressed in the testis and are likely responsible for the bulk of testicular RA production [4–6]. Both chemical modulators and genetic modifications have been used to disrupt RA signaling through manipulation of these enzymes and these studies have revealed much about RA signaling in the testis (see below).

The actions of RA are carried out through two families of ligand-activated nuclear receptors, the retinoic acid receptor (RAR) and the retinoid X receptor (RXR). Both RARs and RXRs have three receptor types (RARα, β and δ and RXRα, β and δ) and multiple isoforms [7]. Additionally, several forms of RA exist and while RARs bind all-trans RA and 9-cis RA, RXRs only bind 9-cis RA. To exert transcriptional control, RARs and RXRs form homodimers or heterodimers that bind to retinoic acid response elements (RAREs). RAREs are located in the promoter regions of some genes and are made up of two repeating core motifs. Once bound to a RARE, the receptor dimers recruit corepressors and the target gene is transcriptionally silenced. Upon binding of RA, RARs and RXRs undergo conformational changes that lead to release of corepressors, recruitment of coactivators, and subsequent gene expression [7–9]. The RA signaling network is multi-layered and this complexity is indicative of the critical nature of RA signaling. Indeed, RA signaling plays essential roles in embryonic development, homeostasis and differentiation of many stem cells and its disruption has detrimental or even lethal effects. The complexity of RA signaling results in a redundancy that can, to some level, protect against the harmful effects of mutation of a single retinoid receptor or metabolic enzyme. While this complexity is biologically advantageous for survival, it has also challenged investigators as they parse out the details of RA signaling.

Spermatogenesis and the Spermatogenic Wave

Spermatogenesis can be divided into three main phases. The first phase is highly proliferative and involves undifferentiated spermatogonia undergoing multiple rounds of mitotic divisions. This phase culminates in the transition of undifferentiated spermatogonia to differentiated spermatogonia. Next, differentiated spermatogonia progress through several rounds of mitotic division and ultimately become primary spermatocytes, which then enter meiosis to form haploid spermatids. Finally, the process of spermiogenesis involves extensive restructuring of the spermatids to form spermatozoa, which are released into the lumen of the seminiferous tubule. These events progress as the germ cell moves from the basement membrane of the seminiferous tubule towards the lumen. Therefore, a cross section of a seminiferous tubule reveals a heterogeneous germ cell population, which, although complex, is not random. Based on their observations in the rat testis, Leblond and Clermont [10] first identified distinct cellular associations that repeat in consecutive order with each round of spermatogenesis. They termed these cellular associations the stages of the seminiferous epithelium (Fig1A). Stages of the seminiferous epithelium have been characterized in many different species and while 14 distinct stages were identified in rats, mice and humans have 12 defined stages [11,12]. Because these stages occur repeatedly in numerical order along the length of a seminiferous tubule, the process of spermatogenesis is asynchronous and often referred to as a wave (Fig1B). Wave-like progression of spermatogenesis is essential for continuous sperm production and has been defined in many mammalian species.

Fig 1.

Cell associations of the stages of the seminiferous epithelium in the mouse. Red box surrounds stages marked by peak RA concentration. A, undifferentiated spermatogonia; A1-3, A1-3 differentiated spermatogonia; Inm, intermediate spermatogonia; B, B spermatogonia; Pl, preleptotene spermatocytes; L, leptotene spermatocytes; Z, zygotene spermatocytes; P, pachytene spermatocytes; D, diplotene spermatocytes; m2Om, meiotically dividing spermatocytes; 1-16, stage 1-16 spermatids (A) Stages of the seminiferous epithelium repeat in numerical order along the length of a seminiferous tubule, creating a wave-like progression of spermatogenesis (B) A cross section of an unsynchronized testis reveals tubules in multiple stages of the cycle of the seminiferous epithelium while the synchronized testis is highly enriched for only a few stages (C).

RA Signaling and Spermatogenesis

RA plays an essential role in each of the three phases of spermatogenesis. Spermatogonial differentiation, meiotic entry, spermatid elongation and release of spermatozoa into the lumen of the seminiferous tubule all require RA signaling to proceed normally [•13–16]. Additionally, RA is essential for dynamic reorganization of the blood-testis barrier tight junctions [17]. Initial evidence for RA’s role in spermatogonial differentiation came from studies in VAD animals. When rodents were made VAD at or shortly after birth, their testes were found to contain only Sertoli cells, undifferentiated spermatogonia and rare spermatocytes, even into adulthood [18–20]. Later it was discovered that this effect could be recapitulated by administering the bis-dichloroacetyl-diamine (BDAD) WIN 18,446 to rodents and rabbits and that this effect was likely mediated through inhibition of the Aldh1a enzymes [21,22]. Upon administration of exogenous RA to VAD animals, the spermatogenic block was released and undifferentiated spermatogonia progressed through differentiation [23–25]. Furthermore, arrested spermatocytes entered meiosis, indicating that RA is also required for meiotic initiation [25]. A recent study by Endo et al. [•13] highlighted the importance of RA signaling in spermatid elongation and spermiation through manipulation of RA levels. RA depletion in mice was accomplished through injection of WIN 18,446 and a buildup of round spermatids was observed in their testes, indicating that elongation did not occur efficiently. An increased number of aligned spermatozoa that were not released into the lumen of the seminiferous epithelium was also observed, which agrees with previous reports that RA depletion prevents efficient spermiation [15]. Conversely, when animals received an injection of exogenous RA, there was a significant increase in the number of elongated spermatids and a decrease in the number of aligned spermatozoa at the luminal edge of the seminiferous epithelium [•13].

Interestingly, the RA-dependent processes of spermatogonial differentiation, meiotic entry, tight junction reorganization and spermiation, occur nearly simultaneously within stages VII-IX. The timing of these events led many researchers to hypothesize that RA production in the testis is pulsatile and driven by stage-specific production of RA metabolic and degradative enzymes. This was supported by gene expression analyses and circumstantial evidence [5,26,27] and by studies in VAD animals. Upon administration of retinoids to animals made VAD either by dietary restriction or chemical inhibition, spermatogenesis would recommence. However, spermatogenesis progressed synchronously throughout the testis, creating a system in which a cohort of sperm were released every 8.6 days, rather than continuously [28]. Detailed histological analyses of cross sections from synchronized testes revealed that the tubules were all restricted to one or two stages of spermatogenesis at any given time point (Fig1C). Thus, withdrawal of RA from the neonatal testis, followed by administration of a single dose of RA disrupts the wave-like progression of spermatogenesis and results in a testis enriched for only one or two spermatogenic stages, rather than all twelve. Using this synchronization technique, Hogarth et al. directly measured the levels of RA at specific stages, providing definitive evidence for pulsatile production of RA [29]. Their findings suggest that a pulse of RA is produced during stages VII-IX, which further supports the notion that RA signaling plays a role in the maintenance of asynchronous spermatogenesis.

The hypothesis that RA signaling is responsible for establishing asynchronous spermatogenesis is further galvanized by two studies in which Snyder et al. assessed RA signaling in neonates using the RARE-hspLacZ mouse [30,31]. The RARE-hspLacZ mouse model contains a LacZ transgene under the control of a RARE promoter so that any cells responding to RA stain blue upon treatment with B-galactosidase [32]. Using this model, they found that the unperturbed RARE-LacZ neonatal testis commences RA signaling in patches throughout the testis and that both germ cells and Sertoli cells respond to RA. Furthermore, exposure of 2 day post-partum pups, but not adults, to exogenous RA ablates this patchy initiation and instead, the entire testis commences RA signaling simultaneously [30]. This suggests that patchy initiation of spermatogenesis in the neonate is driven by availability of RA rather than through ability of certain germ cells to respond to ubiquitous RA. Indeed, a recent study by Lord et al. demonstrated that preservation of the spermatogonial stem cell (SSC) population through protection from RA-induced differentiation was not intrinsic to SSCs but was dependent on the architecture of the seminiferous epithelium [••33]. It is also possible that RA degrading enzymes, such as the cytochrome P450 enzyme CYP26, could be responsible for eliminating RA in localized regions to initiate localized spermatogenesis. A large dose of exogenous RA could simply overrun these enzymes, resulting in the loss of patchy spermatogenesis initiation. In fact, degradation of RA by CYP26b1 during embryogenesis is responsible for preventing precocious meiosis in fetal XY germ cells [34–••36]. Thus it is plausible that these degradative enzymes could play a role in establishing the asynchronous spermatogenic wave in neonates.

Although the studies by Snyder et al. revealed RA response in germ and Sertoli cells, it was still not clear which cells produce RA in the testis. For decades it had been assumed that, as with organogenesis, RA signaling in the testis is paracrine, with one cell making RA and another responding [37]. In the case of testicular RA signaling, it was thought that Sertoli cells produced RA that would act on neighboring undifferentiated spermatogonia [38,39]. This seemed reasonable, since Sertoli cells express the requisite enzymes for RA production and are in close association with germ cells. Furthermore, data from several studies indicated that germ cells are capable of responding directly to RA [27,40]. However, mounting evidence suggests that RA signaling in the testis is vastly more complicated than this, with both germ and Sertoli cells producing and responding to RA. Much of this evidence comes from expression profiles of RA receptors and metabolic enzymes. Although there is some controversy over which isoforms of the RAR and RXRs are expressed in Sertoli cells and which are expressed in germ cells, it is clear that both cell types express receptors required to respond to RA signaling [41–44,5].

Recent advances in understanding testicular RA signaling have come through functional analyses of the RDH and RALDH enzymes. Mice lacking Rdh10 expression in Sertoli cells and germ cells (Rdh10 sgKO) were infertile and their testes contained only undifferentiated spermatogonia and Sertoli cells. However, a single dose of retinoic acid caused synchronized spermatogenesis to commence and continue indefinitely. Furthermore, Rdh10 sgKO animals greater than 9 weeks old that did not receive a RA injection were found to be normally fertile with normal spermatogenesis [45]. Thus, an RA source independent of RDH10 is present in the adult testis and is sufficient to drive spermatogenesis after puberty.

The notion that separate RA-producing programs are present in neonatal and adult animals is further supported by studies in which production of Aldh1a-derived RA was manipulated. Beedle et al. assessed the effects of treating animals with WIN 18,446, a pan-Alhd1a inhibitor, followed by an RA injection then more WIN 18,446 treatment. They found that the single RA injection rendered subsequent WIN 18,446 treatment ineffectual in blocking spermatogenesis, which indicates that once advanced germ cells are present in the testis spermatogenic progress no longer relies exclusively on Aldh1a-derived RA [•46]. In a more precise approach, Raverdeau et al. explored the effects of eliminating Aldh1a-derived RA in Sertoli cells through Sertoli cell-specific ablation of the Aldh1a1–3 genes (Aldh1a1–3^Ser−/−). Interestingly, although spermatogenesis was arrested prior to differentiation, a single injection of retinoic acid allowed spermatogenesis to progress indefinitely [14]. Teletin et al. built upon these findings by eliminating Aldh1a1–3 expression in germ cells (Aldh1a1–3^Germ−/−) and found that, although a significant decrease in RA concentration was observed in Aldh1a1–3^Germ−/− animals’ testes, spermatogenesis initiated and proceeded normally and animals were fertile. However, animals lacking Aldh1a1–3 expression in both Sertoli and germ cells (Aldh1a1–3^{Germ−/−;Ser−/−}) were infertile, and their testes contained only Sertoli cells and undifferentiated spermatogonia. When given an injection of exogenous RA, Aldh1a1–3^{Germ−/−;Ser−/−} animals underwent a single round of spermatogenesis followed by spermatogenic arrest, suggesting that Aldh1a is solely responsible for converting retinaldehyde to RA in the testis [••47]. This is in contrast to the findings of Beedle et al., which suggested that sources of RA independent of Aldh1a are present and sufficient to drive spermatogenesis following the first wave [•46]. However, if advanced germ cells produce RA and WIN 18,446 cannot effectively penetrate the blood testis barrier to block Aldh1a, this could explain the apparent contradiction between these studies. Thus, although more experimentation is required to determine whether Aldh1a is the sole source of RA in the testis, it is clear that RA production by Sertoli cells through the Aldh1a program is essential for initiation of spermatogenesis. Furthermore, germ cell-derived RA produced through the Aldh1a pathway is sufficient but not essential to drive spermatogenesis in the adult animal.

Remaining Questions

Despite recent advances in our understanding of RA signaling in the testis there are still many unanswered questions. Although it has become apparent that germ cells play a role in RA production, it is still not clear which subtypes have the capacity to synthesize RA and thus drive spermatogenesis in the adult animal. Furthermore, the discovery that RA is produced in pulses at specific stages of the seminiferous epithelium has also led to more questions. Perhaps the most compelling is whether the RA pulse is generated through temporal control of RA synthesis or degradation. Answering these questions will provide a much clearer picture of RA signaling in the testis and provide a more complete understanding of spermatogenesis.