Retinoic acid metabolism links the periodical differentiation of germ cells with the cycle of Sertoli cells in mouse seminiferous epithelium

MECHANISMS OF DEVELOPMENT

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

Available at www.sciencedirect.com

journal homepage: www.elsevier.com/locate/modo

Retinoic acid metabolism links the periodical differentiation of germ cells with the cycle of Sertoli cells in mouse seminiferous epithelium Ryo Sugimoto

a,b,c

, Yo-ichi Nabeshima b, Shosei Yoshida

a,b,d,*

a

Division of Germ Cell Biology, National Institute for Basic Biology, 5-1 Higashiyama, Myodaiji, Okazaki, Aichi 444-8787, Japan Department of Pathology and Tumor Biology, Graduate School of Medicine, Kyoto University, Yoshida-Konoe, Sakyo, Kyoto 606-8501, Japan c Laboratory of Functional Biology, Graduate School of Biostudies, Kyoto University, Yoshida-Konoe, Sakyo, Kyoto 606-8501, Japan d Department of Basic Biology, Sokendai, 5-1 Higashiyama, Myodaiji, Okazaki, Aichi 444-8787, Japan b

A R T I C L E I N F O

A B S T R A C T

Article history:

Homeostasis of tissues relies on the regulated differentiation of stem cells. In the epithe-

Received 11 November 2011

lium of mouse seminiferous tubules, the differentiation process from undifferentiated

Received in revised form

spermatogonia (Aundiff), which harbor the stem cell functions, to sperm occurs in a period-

10 December 2011

ical manner, known as the ‘‘seminiferous epithelial cycle’’. To identify the mechanism

Accepted 12 December 2011

underlying this periodic differentiation, we investigated the roles of Sertoli cells (the

Available online 20 December 2011

somatic supporting cells) and retinoic acid (RA) in the seminiferous epithelial cycle. Sertoli cells cyclically change their functions in a coordinated manner with germ cell differentia-

Keywords: Spermatogonia Retinoic acid Seminiferous epithelial cycle Sertoli cells

tion and support the entire process of spermatogenesis. RA is known to play essential roles in this periodic differentiation, but its precise mode of action and its regulation remains largely obscure. We showed that an experimental increase in RA signaling was capable of both inducing Aundiff differentiation and resetting the Sertoli cell cycle to the appropriate stage. However, these actions of exogenous RA signaling on Aundiff and Sertoli cells were strongly interfered by the differentiating germ cells of intimate location. Based on the expression of RA metabolism-related genes among multiple cell types – including germ and Sertoli cells – and their regulation by RA signaling, we propose here that differentiating germ cells play a primary role in modulating the local RA metabolism, which results in the timed differentiation of Aundiff and the appropriate cycling of Sertoli cells. Similar regulation by differentiating progeny through the modulation of local environment could also be involved in other stem cell systems. Ó 2011 Elsevier Ireland Ltd. All rights reserved.

1.

Introduction

Tissue homeostasis is ensured by stem cells, which maintain their pool and persistently give rise to differentiating progeny. To maintain normal tissue organization and function, differentiation of stem/progenitor cells needs to be

appropriately regulated – both temporally and spatially. In mammalian spermatogenesis, undifferentiated population of spermatogonia that are located on the peripheral basement membrane of seminiferous tubules differentiate at regular intervals. The differentiation accompanies their movement toward the lumen, resulting in the stratified organization of

* Corresponding author at: Division of Germ Cell Biology, National Institute for Basic Biology, 5-1 Higashiyama, Myodaiji, Okazaki, Aichi 444-8787, Japan. Tel.: +81 564 59 5865; fax: +81 564 59 5868. E-mail address: [email protected] (S. Yoshida). 0925-4773/$ - see front matter Ó 2011 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.mod.2011.12.003

MECHANISMS OF DEVELOPMENT

differentiating germ cells (de Rooij and Russell, 2000; Russell et al., 1990a) (Fig. 1). This cyclic program, known as ‘‘seminiferous epithelial cycle’’ (Leblond and Clermont, 1952a; Oakberg, 1956b), ensures the constant production of mature sperm. However, there is currently little understanding of the mechanism that achieves such a periodic differentiation. In mice, stem cell functions mostly reside in a primitive set of spermatogonia, namely, As (singly isolated spermatogonia), Apr (interconnected two-cell cysts), and Aal (cysts of 4, 8, 16 or occasionally 32 cells), which are observed throughout the seminiferous epithelial cycle (Barroca et al., 2009; Nakagawa et al., 2007, 2010; Nishimune et al., 1978; Ohbo et al., 2003; Sada et al., 2009; Shinohara et al., 2000). As, Apr, and Aal spermatogonia (collectively designated as ‘‘undifferentiated’’ spermatogonia or Aundiff) transform into A1 spermatogonia once during every 8.6-day cycle (de Rooij, 1998; Lok et al., 1982). A1 spermatogonia then give rise to five successive stages of spermatogonia (A2, A3, A4, In, and B, designated collectively as ‘‘differentiating’’ spermatogonia) before entering meiosis. All the stages of spermatogonia (both ‘‘undifferentiated’’ and ‘‘differentiating’’) comprise the peripheral germ cell layer in the basal compartment (the space between the tight junctions of Sertoli cells and the basement membrane) (Fig. 1A). At the beginning of the meiotic prophase (preleptotene), spermatocytes vertically translocate across the Sertoli cell junction and subsequently undergo differentiation to sperm through the stages of primary (preleptotene, leptotene, zygotene, pachytene, and diplotene) and secondary spermatocytes, then round and elongating spermatids (Fig. 1B) (Russell et al., 1990a). The differentiation from A1 spermatogonia to spermatozoa takes approximately 35 days, which corresponds to four rounds of the seminiferous epithelial cycle. Thus, the differentiating germ cells form four layers from the basal to the luminal sides. Combinations of germ cell types that appear in a single cross section of the tubules represent the phase of the seminiferous epithelial cycle (Fig. 1D) (Leblond and Clermont, 1952b). In mice, a cycle is divided into stages I to XII based on the combinations of germ cell types (Oakberg, 1956a; Russell et al., 1990a). Differentiation of Aundiff into A1 spermatogonia occurs in stages VII to VIII (de Rooij, 1998). The stratified germ cells are nourished by the epithelium of the Sertoli cells, which are large somatic cells spanning from the basement membrane to the lumen (Fig. 1A). In contrast to the constant turnover of germ cells, Sertoli cells persist over a long period without division (Kluin et al., 1984). On the other hand, Sertoli cells exhibit cyclic changes in their function and gene expression. Particular patterns of Sertoli cell gene expression coincide with particular sets of germ cells, such that Sertoli cells provide the microenvironment appropriate to each step of germ cell differentiation (Fig. 1C and D) (Elftman, 1950). Different segments of the tubules exhibit different stages of a combination of germ cells along with the expression of stage-specific genes, where the temporal order is recapitulated by the spatial arrangement (Supplementary Fig. S1A–D) (Hogarth and Griswold, 2010; Perey et al., 1961). Previous studies have provided a number of suggestive but puzzling findings when assessing how germ cells and Sertoli cells are coordinated in their cycles. On one hand, the Sertoli cells appear to cycle in germ cell-depleted testes as in normal situation (see also Supplementary Fig. S1G–P) (Timmons et al.,

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

611

2002; Yoshida et al., 2006), leading to the idea that Sertoli cells have an intrinsic cyclic program and ‘‘instruct’’ the timing of germ cell differentiation. On the other hand, when rat spermatogenic cells were transplanted into mouse seminiferous tubules, they differentiated over the 13-day rat cycle instead of the 8.6-day mouse cycle, despite being nourished by mouse Sertoli cells (Franca et al., 1998), suggesting that the cyclic differentiation is an intrinsic nature of the germ cell population. Convincing explanations to these seemingly inconsistent findings are warranted. Lines of researches have demonstrated that retinoic acid (RA) plays crucial roles in these cyclic processes. First, a deficiency of vitamin A (VA), the dietary precursor of RA, blocks the Aundiff to A1 differentiation, which is resumed by readministration of VA. Intriguingly, the restored spermatogenesis is synchronized throughout the testis in the seminiferous epithelial cycle over a prolonged period (Morales and Griswold, 1987; van Pelt and de Rooij, 1990). Additionally, blockade of RA signaling by deletion of retinoic acid receptor a (RARa), the prominent RAR member expressed in Sertoli cells, causes spermatogenesis defects and misregulation of cyclic gene expression of Sertoli cells (Chung and Wolgemuth, 2004; Vernet et al., 2006a). Moreover, genes involved in RA metabolism show seminiferous epithelial cycle-related expression, suggesting that RA metabolism is regulated in a cyclic manner (Vernet et al., 2006b). However, the role of RA and its regulation remains obscure and has yet to achieve general agreement. In this study, we have investigated the mechanism that achieves the periodical differentiation of Aundiff and subsequent spermatogenesis, focusing on the role of Sertoli cells and RA signaling. We discovered that RA can both induce differentiation of Aundiff and reset the Sertoli cell cycle, and that the expression of RA-metabolism related genes is modulated by feedback regulation by RA signaling in Sertoli cells but not in germ cells. These findings have been integrated into a model proposing that particular stages of differentiating germ cells determine the timing of differentiation of Aundiff by modulating the local RA metabolism.

2.

Results and discussion

2.1. Impact of altered RA signal strength on Sertoli cell cycling: revisiting the VAD animal model The importance of RA signaling for the proper control of seminiferous epithelial cycle, especially that of Aundiff-to-A1 differentiation, came from classical observations of the testes in VAD animals. In VAD mice and rats, differentiation of Aundiff is affected, while readministration of VA resumes their differentiation synchronously, as revealed by germ cell morphology (Morales and Griswold, 1987; van Pelt and de Rooij, 1990). Compatible with these studies, Aundiff spermatogonia (negative for Kit expression) were accumulated in the seminiferous tubules of VAD mice, while Kit+ differentiating spermatogonia and more advanced germ cells were missing (Fig. 2A–D) (SchransStassen et al., 1999). VA readministration induced the appearance of Kit+ spermatogonia throughout the testes (Fig. 2E), leading to restoration of full spermatogenesis that was

612

MECHANISMS OF DEVELOPMENT

synchronized all over the testes as reported (van Pelt and de Rooij, 1990). In the resultant synchronized testes, the Sertoli cell cycle is also synchronized, showing good coordination with the germ cell cycle (Fig. 2F and G). Altogether, this indicates that the cycle of Sertoli cells is also affected by the artificial alteration of RA signal strength. We therefore analyzed the state of the Sertoli cell cycle during this experimental procedure. Given that the differentiation of Aundiff into Kit+ spermatogonia occurs at stage VII in

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

normal spermatogenesis, we hypothesized that Sertoli cells would also be arrested before stage VII all over the testes in the VAD situation, and that addition of VA would resume their cycle. This hypothesis was also supported by the backward extrapolation from the appearance of stages following VA readministration (van Pelt and de Rooij, 1990), and the finding that Sertoli cells-specific disruption of RARa abolished the cycle and exhibited a uniform gene expression pattern that was similar to those in stages before VII (Vernet et al., 2006a).

MECHANISMS OF DEVELOPMENT

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

613

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ Fig. 1 – Stratified architecture and periodical events in mouse seminiferous epithelium. (A) Architecture of seminiferous tubules and seminiferous epithelium. Inside the long convoluted seminiferous tubules, the seminiferous epithelium exhibits a stratified organization of the differentiating germ cells, which are nourished by Sertoli cells. Bar, 20 lm. (B) Schematic diagram of spermatogenic differentiation program from As, Apr, and Aal spermatogonia to spermatozoa, which takes 35 days. As shown by the open arrows, the As, Apr, and Aal population is maintained while giving rise to A1 spermatogonia. Steps are as follows: spermatogonia (A1–A4, Intermediate (In), and B), primary spermatocytes (preleptotene (PL), leptotene (L), zygotene (Z), early pachytene (EP), late pachytene (LP), diplotene (D)), meiotic divisions and secondary spermatocytes (M), and spermatids (round spermatids (RT) and elongated spermatids (ET)). (C) Diagram of the Sertoli cell cycle, which takes 8.6 days. Different phases of the cycle characterized by expression patterns of stage-specific genes are schematically indicated by different intensities of the shading. (D) Schematic of the temporal coordination of spermatogenic differentiation and the cycle of Sertoli cells. Arrows and shading, such as those from (C) and (D), indicate germ cell differentiation and the Sertoli cell cycle. The Aundiff population persists and periodically gives rise to A1 spermatogonia once every 8.6 days (at stage VII). Subsequent differentiation and regular production of A1 spermatogonia result in the formation of four stratified layers of differentiating cells, progressing from the basal to the luminal side, as indicated. Note that particular combinations of differentiating germ cells appears periodically, which is recognized as the seminiferous epithelial cycle that is divided into stages I to XII, as shown schematically.

Unexpectedly, however, galectin1 and tPA (tissue plasmingen activator), which show stage-specific expression in Sertoli cells (Liu et al., 1995; Timmons et al., 2002), showed variable expression in the tubule sections in VAD testes, similar to that in the normal testes (Fig. 2H, J and O). The spatial relationship between the expression domains of stage-specific genes was also maintained (Supplementary Fig. S1Q–S). Therefore, it was indicated that contrary to our prediction, Sertoli cells underwent cycling in VAD, or were arrested at a random stage of the seminiferous epithelial cycle. Differences in the Sertoli cell phenotype between VAD and Sertoli cell-specific RARa-knockout animals may be due to the low level of residual RA signaling in the former, while eventually all the signals were expected to be lost in the latter. Alternatively, this may be due to the normal establishment of the cycle before becoming VAD (see incomplete VAD testes in Fig. 6B), while the RARa gene was removed before the Sertoli cell cycle was first observed during embryogenesis (Timmons et al., 2002; Vernet et al., 2006a). One day after VA readministration, however, Sertoli cell gene expression showed a pattern similar to that of stage VII to VIII, i.e., uniform expression of tPA with a small portion that is positive for galectin1. Then, it was followed by a synchronous and sequential shift as observed in the normal progression of the seminiferous epithelial cycle (Fig. 2J–S, Supplementary Fig. S1E and F). These results indicate that VA readministration ‘‘resets’’ Sertoli cells to stage VII from any stages. These observations prompted us to explore whether exogenous VA is also capable of resetting Sertoli cells even in a VA-sufficient (VAS) situation. We examined the testes of artificial cryptorchids and of Nanos3 and W/Wv mutants, which contain no germ cells or Aundiff only (Fig. 3) (Geissler et al., 1988; Nishimune et al., 1978; Tsuda et al., 2003). Sertoli cells in these conditions exhibited variable expression of stage-specific genes similar to that in VAD, which then became uniform 2 days after administration of VA (Fig. 3B–G, Supplementary Fig. S1G–S). Additionally, the residual Kitnegative Aundiff spermatogonia in cryptorchid testes, which are genetically wild type, synchronously differentiated into being Kit+ after VA administration (Fig. 3H and I). Therefore, we consider that not only is RA signaling required for differ-

entiation of Aundiff and proper cycling of Sertoli cells but also that its artificial increase can simultaneously elicit differentiation of Aundiff and reset the Sertoli cell cycle to stage VII from any stages, coordinating the timing of these two events. Reset of Sertoli cells in Nanos3 mutant (lacking all the germ cells) indicates that RA signaling act independently from germ cells (probably directly to Sertoli cells). Whilst, the subsequent progression of Sertoli cell cycle was less synchronous in Nanos3 and W/Wv mutant testes relative to VAD wild type testes (not shown), indicating that concurrent progression of germ cell differentiation fine-tunes the cycle of Sertoli cells.

2.2. cells

Interference in the effect of VA by differentiating germ

The above mentioned function of exogenous VA was observed in testes that contained no germ cells or only Aundiff. However, when VA was administered to normal wild-type mice that carried differentiating germ cells, no ubiquitous induction of Kit+ spermatogonia or synchronization of the Sertoli cell cycle was observed (Fig. 4A–E). We then analyzed this finding in greater detail; we observed that Kit+ spermatogonia were induced but this was limited to stages VII and VIII in which Aundiff differentiation normally occurs (Fig. 4F). The Sertoli cell gene expression was also weakly affected by the direction of ‘‘being reset to stage VII’’ (i.e., galectin1 and tPA were upregulated while AR was downregulated); however, stage-related oscillation was clearly preserved (Fig. 4G–I). These effects did not persist and the resultant tissue was undistinguishable from untreated testes after a prolonged period, in contrast to the testes without differentiating germ cells (data not shown). Therefore, we conclude that differentiating germ cells interfere with the effect of exogenous VA, pushing back Sertoli cells and germ cells to their original stages. Similar observation has been reported recently: RA administration into premature mice that lacks differentiating germ cells, but not into adult, can elicit

614

MECHANISMS OF DEVELOPMENT

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

Fig. 2 – Effect of exogenous VA administration on spermatogonial differentiation and Sertoli cell cycle in VAD mouse testes. (A) The experimental schedule. VAD mice that had been given a VAD diet were injected with VA, after which they were put on a normal (VA-sufficient) diet. Testis samples were harvested at the indicated time points. (B and C) PAS-hematoxylin-stained cross-sections of a VAD mouse testis, at a low power view (B) and as a magnification of the region in the rectangle (C). Arrows and arrowheads indicate residual Aundiff spermatogonia and Sertoli cells, respectively. (D and E) In situ hybridization for Kit in VAD (D) and 2 days after VA injection (E). Seminiferous tubules are outlined. Kit-positive cells outside the tubules are interstitial cells. (F and G) PAS-hematoxylin-(F) galectin1-stained sections of a testis 35 days after VA injection. Arrowheads indicate secondary spermatocytes, which are a characteristic mark of stage XII. Roman numerals indicate the stage of the seminiferous epithelial cycle, as judged from the PAS-H staining on the adjacent section. The stages were synchronized within a testis but varied among individuals, as previously reported (van Beek and Meistrich, 1991) (H and I). In situ hybridization for galectin1 in VAD (H) and 2 days after VA injection (I) (J–S). In situ hybridization of galectin1 (J–N) and tPA (O–S), 0, 1, 2, 3, and 4 days after VA injection. Note that the gene expression pattern recapitulates its normal progression observed in stages VII–VIII, IX–X, XI–XII, and XII–I, at 1, 2, 3, and 4 days, respectively, following VA injection (compare with Fig. S2K and S2L). Seminiferous tubules harboring intact Sertoli cells are outlined. Scale bars indicate 200 lm (B, D, E, and G–I) or 50 lm (C, F, and J–S).

MECHANISMS OF DEVELOPMENT

the synchronization of seminiferous epithelial cycle (Snyder et al., 2011). We then assessed whether the interference by differentiating germ cells in the VA effect was due to local or remote action, by using mosaic testes in which some segments of the seminiferous tubules harbored full spermatogenesis while others contained only Sertoli cells. This was achieved by killing off a majority of spermatogonia by a marginal dose of busulfan, followed by 2 months of recovery that allowed the surviving stem cells to restore full spermatogenesis in a patchy manner (Fig. 4J). In these testes, galectin1 expression varied in segments both with and without differentiating germ cells (Fig. 4K–M and Q). In response to the VA administration, eventually all the segments lacking differentiating germ cells became positive for galectin1. In contrast, segments with full spermatogenesis showed minimal upregulation of galectin1 in a stage-dependent manner as observed in normal testes (Fig. 4N–R; compared with Fig. 4G). Therefore, we concluded that differentiating germ cells located in close proximity would interfere with the effects of exogenous VA on the differentiation of Aundiff and resetting of the Sertoli cell cycle.

2.3. Control of RA metabolism by multiple cell types during seminiferous epithelial cycle The above findings indicate that differentiating germ cells interfere with the RA functions that control Aundiff differentiation and the Sertoli cell cycle, in a manner depending on their step of differentiation. There are two possible mechanisms for this process, which are not mutually exclusive. First, the regulation of the environmental RA concentration by modulating its metabolism; second, the action of differentiating germ cells on Aundiff and Sertoli cells, directly or indirectly,

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

615

via other signaling molecule(s), which in turn interfere with their response to RA. While the second possibility warrants further investigation, in this study, we challenged the first possibility in order to identify the involvement of RA in the seminiferous epithelial cycle. RA is converted from VA, whose primary form is retinol (Rol), by the cooperation of multiple gene products. After being absorbed from the intestines, Rol circulates as a retinol-binding protein (RBP)-bound form, and is taken up into cells in a Stra6-dependent fashion that recognize the Rol– RBP complex (Kawaguchi et al., 2007). Rol is subsequently metabolized into retinal (Ral) and then RA by Aldh (aldehyde dehydrogenase) activity. RA is then oxidated into inactive forms by Cyp26 (cytochrome P450 family 26) enzymes. Alternatively, Rol will be esterized by Lrat (lecithin:retinol acyltransferase) and stored in lipid droplets with the aid of Adfp (adipose differentiation-related protein), preventing the RA synthesis (Supplementary Fig. S2A) (Chung and Wolgemuth, 2004; Imamura et al., 2002). It is likely that RA permeates cell membranes to move across neighboring cells and that Ral is effectively transported between cells by an unknown mechanism (Chung and Wolgemuth, 2004; Noy, 1992). In testes without differentiating germ cells, Sertoli cells widely expressed Aldh1a1 and Aldh1a2 genes, which encode the rate-limiting enzyme of RA synthesis (Supplementary Fig. S5A–D), while genes for storage (Adfp or Lrat) were not activated (Supplementary Fig. S5E–H). Stra6 expression varied among seminiferous tubule segments (Supplementary Fig. S5I and J). Therefore, we speculate that, after intraperitoneal injection of VA (Rol), the excess dose of circulating free Rol diffused into Sertoli cells in a Stra6-independent manner and was converted into RA by Aldh activity, and that the increased RA elicited both the reset of Sertoli cell cycle and the differentiation of Aundiff spermatogonia simultaneously.

Fig. 3 – Effect of exogenous VA in vitamin A-sufficient testes in the absence of differentiating germ cells (A–I). The experimental schedule (A). In situ hybridization of Nanos3-mutant (B and C), W/Wv (D and E), and cryptorchid (F–I) mouse testes maintained with VAS diet, before (no treatment) or 2 days after VA injection, for galectin1 (B–G) and Kit (H and I). Seminiferous tubules are outlined. Scale bars indicate 200 lm.

616

MECHANISMS OF DEVELOPMENT

In normal testes, these RA metabolism-related genes are expressed separately in multiple cell types (including differentiating germ cells and Sertoli cells) in a beautifully coordi-

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

nated manner with the seminiferous epithelial cycle, as shown by the preceding work and our current analyses (Vernet et al., 2006b) (Supplementary Figs. S2, S3, S4 and Table

MECHANISMS OF DEVELOPMENT

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

617

ƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒƒ Fig. 4 – Interference in the effect of exogenous VA by differentiating germ cells. (A–I) Administration of VA into normal testes. (A) The experimental schedule (B–E). In situ hybridization images of normal wild-type mice for Kit and galectin1, before treatment or 2 days after VA injection, as indicated (F–I). Number of Kit+ spermatogonia per Sertoli cells (F), and the proportion of seminiferous tubule sections highly positive for galectin1 (G), tPA (H), and AR (I) to the total tubule sections in each stage, based on the in situ hybridization performed on three testis sections. ** and * indicate P < 0.05 and P < 0.01 (v2-test). Raw counts are summarized in Table S2 in the supplementary material. Roman numeric indicates stages of seminiferous epithelium judged from the combination of germ cells. Scale bars indicate 200 lm (J–Q). Local effect of differentiating germ cells that interfered with the VA effect. (J) The experimental schedule. Wild-type VAS mice were injected with busulfan (30 mg/kg) to eliminate a majority of spermatogenic stem cells, after which they experienced patchy regeneration of complete spermatogenesis from surviving stem cells in 2 months. This was followed by VA injection 2 days before sacrifice. (K–P) In situ hybridization images of the treated testes hybridized for galectin1, before (K–M) or after (N–P) VA administration. L–M and O–P are higher-magnification images of the indicated areas in K and M, as indicated. Closed circles and asterisks indicate examples of the tubules with and without differentiating germ cells, respectively. Scale bars indicate 200 lm. (Q) Percentages of galectin1-high seminiferous tubules (classified by the presence (+) or absence () of differentiating germ cell layers) before (VA) and after (+VA) VA injection. *P < 0.001 (v2-test). (R) Percentages of galectin1-high seminiferous tubules in tubule sections harboring full spermatogenesis before (VA) and after (+VA) VA injection, classified by stages as indicated. Counts of seminiferous tubule cross sections are summed from four testis specimens; the absolute numbers of counted cross sections are indicated in each column for (Q) and (R).

S4). Their expression pattern suggests that endogenous RA concentration is suppressed in the early stages (I to VI) due to the prominent expression of storage genes (Adfp and Lrat) and the lower expression of RA-synthesizing enzymes (Supplementary Fig. S2I). In contrast, in the late stages (stage VII onward), while the storage genes are turned off, Stra6 is upregulated in Sertoli cells to promote uptake of RA precursor, and Aldh1a2 mRNA and protein become prominent in late pachytene spermatocytes, suggesting that RA synthesis is promoted (Supplementary Fig. S2E, J and S4). The delayed up-regulation of Cyp26a1 from that of Stra6 (Supplementary Fig. S2) appears to cause sharp increase and subsequent decrease of local RA concentration and to restrict an effective time window of RA action. It is frustrating that such a predicted oscillation of RA concentration cannot be tested directly due to current technical limitations, although this hypothesis is supported by a number of indirect evidences as discussed later. Cyp26b1 expressed in peritubular myoid cells (Supplementary Figs. S2H and S3H) is considered to restrict the RA action from being disturbed by the neighboring tubules or interstitium as discussed in Vernet et al., 2006b. These observations suggest that differentiating germ cells are involved in the regulation of RA signaling by expressing key players of RA metabolism. In early stages, round spermatids that express the storage genes (Lrat and Adfp) may contribute to the suppression of RA synthesis and to the blockade of the effect of exogenously added VA. On the other hand, high Aldh1a2 expression in late pachytene spermatocytes appears to drive the increase in RA concentration in late stages, in cooperation with the coordinated induction of Stra6 in Sertoli cells (Supplementary Fig. S2).

2.4. genes

Feedback regulation of the RA metabolism-related

The abovementioned orchestration of RA metabolism-related gene expression among multiple cell types prompted us to investigate how their expressions are regulated. Moti-

vated by the fact that some of these genes are known to be under feedback regulation by RA signaling in other cell types (Balmer and Blomhoff, 2002; Bouillet et al., 1995; Elizondo et al., 2000; Loudig et al., 2000, 2005), we examined whether similar regulation also occurred in cell types comprising seminiferous tubules. In these experiments, the level of RA signaling was manipulated in vivo by the depletion and administration of VA (Figs. 5 and 6). First, the expression of RA metabolism-related genes in Sertoli cells was analyzed in VAD testes before and after VA readministration by in situ hybridization and RT-qPCR. Artificial cryptorchid testes in the normal level of VA (VAS), which harbored the same cellular composition as VAD, were also examined for comparison (Fig. 5A and B). It was revealed that Stra6 was suppressed in VAD, and upregulated by VA readministration, indicating positive regulation by RA. In contrast, Aldh1a1 expression was elevated in VAD and downregulated after the addition of VA. Expression of Aldh1a2, which was not detected in VAS Sertoli cells with or without germ cells, became detectable in VAD, and downregulated when VA was readministered (Fig. 5B). Thus, in Sertoli cells, Aldh1a1 and Aldh1a2 are under negative regulation by RA. Cyp26a1 was also upregulated by VA readministration, indicating positive regulation. The feedback relationship between RA metabolism and the expression of related genes in Sertoli cells is summarized in Fig. 5C. We then examined the regulation of RA metabolism-related genes in spermatocytes and spermatids when RA-signaling was artificially reduced and elevated. For that purpose we used ‘‘incomplete VAD’’ testes, which were harvested from animals maintained in VAD diet about 3 weeks prior to becoming completely devoid of differentiating germ cells. Such ‘‘incomplete VAD’’ testes had lost differentiating spermatogonia due to the blockade of Aundiff-to-A1 differentiation, but still retained spermatocytes and spermatids undergoing differentiation (Fig. 6). galectin1 expression in Sertoli cells was downregulated in these incomplete VAD testes and regained after VA re-administration, indicating that the RA signal strength was manipulated as expected (Fig. 6B, bot-

618

MECHANISMS OF DEVELOPMENT

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

Fig. 5 – Effect of altered VA levels on the expression of RA metabolism-related genes in Sertoli cells (A and B). (A) The experimental schedule. Testes from VAD mice were processed for in situ hybridization to detect the RA metabolism-related genes expressed in Sertoli cells, before and 2 days after VA injection, along with artificial cryptorchid testes prepared under a VAS situation. The scale bars indicate 200 lm. The right column shows the results of quantitative RT-PCR shown as relative values versus Gapdh. *P < 0.05 and **P < 0.002 (t-test). (C) A schematic of the regulations of RA metabolism-related genes in Sertoli cells, based on the results of (B).

MECHANISMS OF DEVELOPMENT

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

619

Fig. 6 – Effect of altered VA levels on the expression of RA metabolism-related genes in germ cells (A and B). (A) The experimental schedule. (B) Testes from ‘‘incomplete VAD’’ mice were examined for expression of RA metabolism-related genes in germ cell types, before and after VA injection. Normal (VAS) testes were also analyzed. Roman numerals indicate the seminiferous epithelial stage, as judged from the PAS-H-stained germ cell morphology. Note that the stage-specific expression of these genes is largely unaffected. The right column shows the results of quantitative RT-PCR indicated as relative expression to Gapdh. Note the similar level of expression across all conditions, although some comparisons show statistical differences. *P < 0.05 and **P < 0.001 (t-test). (C) A schematic of the regulations of RA metabolism-related genes in germ cells, based on the results of (B). Note that RA does not play a major role in their expression.

tom panels). Normal (VAS) testes were also analyzed for comparison. In these conditions, we did not observe any obvious changes in the expressions of Adfp, Lrat, or Aldh1a2 by in situ hybridization and RT-qPCR (Fig. 6B). These results sug-

gest that the expression of these genes may be under the control of other mechanisms than local RA, perhaps related to the steps of differentiation (Fig. 6C). Late pachytene spermatocytes appear to receive and respond to RA, given that they

620

MECHANISMS OF DEVELOPMENT

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

Fig. 7 – Model of the temporal control of spermatogonial differentiation and Sertoli cell cycle. (A) A schematic representation of predicted RA metabolism in the presence of round spermatids that express high levels of Adfp and Lrat, and in the absence of late pachytene spermatocytes that express high levels of Aldh1a2. Pathways shown in gray will be downregulated, while pathways shown in black will act prominently, resulting in a decrease in RA concentration. This situation is expected to occur in early stages (I–VI). (B) The prediction of local retinoid metabolism in the absence of round spermatids and in the presence of late pachytene spermatocytes. This situation will occur in late stages (VII to XII). (C) A simplified diagram representing the presumptive roles of round spermatids (RT) and late pachytene spermatocytes (LP) on the control of differentiation of Aundiff spermatogonia and progression of the Sertoli cell cycle to stage VII. RT and LP, by down-and up-regulating the local RA concentrations, respectively, send the ‘‘wait’’ and ‘‘go’’ signals to these two periodical events. (D) An integrated model that summarizes the temporal orchestration of germ cell differentiation and Sertoli cell cycle, coordinated by means of the ‘‘wait’’ and ‘‘go’’ signals (red and green arrows, respectively).

are the very prominent cell type of Aldh production (Supplementary Figs. S2–S4) and an artificial RA-responsive reporter gene (RAREhspLacZ; Rossant et al., 1991) is activated in this cell type (Supplementary Fig. S6). It is interesting that the expression of the same Aldh1a2 locus is regulated differently between Sertoli and germ cells.

2.5. Superior role of differentiating germ cells in the local RA metabolism, and in the timed Aundiff differentiation and Sertoli cell cycle Our findings related to the expression and regulation of RA metabolism-related genes (Figs. 5C and 6C) indicate that the presence of particular stages of differentiating germ cells directs the modes of local RA metabolism, while Sertoli cells can flexibly adjust to the germ cell-oriented modulation of RA metabolism (Fig. 7). As shown in Fig. 7A, the presence of round spermatids that are Adfp-high, Lrat-high, and Aldh1a2-low, would make the storage pathway prominent, while the local RA concentration would become suppressed. The low RA concentration, in turn, would downregulate the expression of Stra6 and Cyp26a1, and upregulate Aldh1a1 (and perhaps also Aldh1a2) in Sertoli cells as a result of feed-

back regulations. In contrast, as shown in Fig. 7B, late pachytene spermatocytes that are Aldh1a2-high but do not express high levels of storage genes, would raise the local RA concentration. High concentrations of RA would then affect Sertoli cell gene expression in the opposite direction (e.g. upregulate Stra6 and Cyp26a1 and downregulate Aldh1a1). Intriguingly, the above prediction suitably matched the observations in the normal seminiferous epithelium of the early and late stages, respectively (Fig. 7A and B, Supplementary Fig. S2I and J) (Vernet et al., 2006b). Thus, it is suggested that spermatocytes and spermatids play crucial roles in regulating the local RA metabolism. It should be noted, however, that this model does not exclude other intercellular communication mechanisms that achieve the coordinated expression of RA metabolism-related genes. To summarize, we propose a model wherein round spermatids and late pachytene spermatocytes modulate the local RA metabolism so as to down-and up-regulate the RA concentration, respectively. The resultant decrease and increase in local RA concentration can be considered as ‘‘wait’’ and ‘‘go’’ signals, respectively, to Aundiff and Sertoli cells (Fig. 7C). As shown in Fig. 7D, this idea satisfactorily explains the chronological orchestration between the scheduled appearance of

MECHANISMS OF DEVELOPMENT

differentiating germ cells (spermatocytes and spermatids) and the timing of differentiation of Aundiff as well as the cycle of Sertoli cells. The regulation of RA signaling between germ cells and Sertoli cells alone can explain the periodical coordination during the seminiferous epithelial cycle, although involvement of other signals cannot be excluded. To evaluate this proposed model it is essential to directly measure the local RA concentration, but it is difficult currently. However, expression of a series of genes that are regulated by RA signaling evidences the oscillation of local RA. First, expression of Sertoli cell genes that have been shown, in this study, to be under positive (Stra6 and cyp26a1) or negative (Aldh1a1) regulation of RA signaling (Supplementary Fig. S2) strongly support the proposed RA oscillation as described earlier. Second, upregulation of Stra8, an RA-induced gene, is observed in multiple steps of germ cell differentiation that occur in stage VII and VIII, i.e., Aundiff to A1 differentiation and meiosis entry from B spermatogonia ((Anderson et al., 2008; Mark et al., 2008; Zhou et al., 2008) and Supplementary Fig. S6). Third, RAREhsplacZ, an artificial reporter gene carrying retinoic acid response elements (Rossant et al., 1991) is highly activated after stage VII (Supplementary Fig. S6A), although this reporter need to be interpret carefully in seminiferous tubules because it is strongly activated in spermatocytes but silent in spermatogonia and Sertoli cells, even when VA was re-administration into VAD animal (Supplementary Fig. S6B and C). This is probably at least in part due to the DMRT1 activity, because DMRT1 mutant upregulates this reporter in spermatogonia as recently suggested (Matson et al., 2010).

2.6. RA may control germ–Sertoli cell coordination in multiple steps In addition to the differentiation of Aundiff spermatogonia, VAD affects multiple steps of spermatogenesis, including entry into meiosis, onset of spermatid transformation, and spermiation (Huang and Marshall, 1983). Intriguingly, these events all happen during stages VII and VIII in mice (Supplementary Fig. S7) (Hogarth and Griswold, 2010; Oakberg, 1956a; Russell et al., 1990b). Thus, it is possible that these events are also regulated simultaneously by the local RA concentration. These multiple checkpoints could be beneficial for Sertoli– germ coordination during the long process of differentiation. This hypothesis is strengthened by the fact that these events occur within a limited time window of the cycle and that the entire differentiation process corresponds to 4 rounds of seminiferous epithelial cycles, in all the species described, regardless of the length of the seminiferous epithelial cycle (Supplementary Fig. S7).

2.7.

Using this model to explain the puzzling results

The model proposed in this study can explain several puzzling findings. First, differentiating germ cells dominate Sertoli cells in regulating stage VII progression: that is, although Sertoli cells appear to have a cycling program of their own, it becomes apparent only when differentiating germ cells are absent. This explains the results of the rat-to-mouse transplantation, where transplanted rat germ cells showed a ratproper differentiation cycle (Franca et al., 1998). We speculate

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

621

that the mouse Sertoli cells cycled in a coordinated manner with the rat germ cells, although the state of the Sertoli cell cycle was not analyzed in that study. Second, this model accounts for the essentially normal combination of differentiating germ cells between Sertoli cells that do not show any sign of cycling due to Sertoli cell-specific deletion of RARa (Vernet et al., 2006a). Given that the particular stages of differentiating germ cells express a particular set of RA-related genes (presumably according to the germ cellintrinsic program), they possibly adjust the RA concentration as necessary to a certain extent, even in the absence of the cooperation of Sertoli cells. This may be enough to instruct the correct timing of Aundiff differentiation. However, the overall integrity of spermatogenesis was affected in the Sertoli cellspecific RARa knockout mice, indicating that the normal cycling of Sertoli cells is essential for efficient spermatogenesis.

2.8.

Perspective

This study suggests that the differentiation timing of Aundiff is controlled by the local regulation of RA metabolism, in which differentiating germ cells dominate in fine-tuning the Sertoli cell-intrinsic cyclic program, so that the seminiferous epithelial cycle progress in a orchestrated manner. This is based on the feedback regulations of RA metabolism-related genes expressed in Sertoli cells but not for those expressed in germ cells. The nature of the Sertoli cell-intrinsic cycle, which becomes apparent in case of the lack of differentiating germ cells, remains an interesting open question. However, it may be intriguing to hypothesize that the positive and negative feedback loops between RA and related genes (Fig. 5C) drives the cycle by themselves, because these are theoretically able to compose an oscillating system (Murray, 2002). This study proposes a novel mechanism for explaining the control of the timing of stem/progenitor cell differentiation. Given that the differentiation of germ cells progresses in a temporally programmed manner toward spermatozoa, this mechanism ensures periodical differentiation at regular intervals. Similar mechanisms might be involved in other stem cell systems.

3.

Experimental procedures

3.1.

Mice

C57BL/6 and W/Wv mice were purchased from Japan SLC. Nanos3+/Cre mice (Suzuki et al., 2008; Tsuda et al., 2003) were interbred to yield Nanos3Cre/Cre individuals (designated as Nanos3-mutant in the text), which lack the functional allele of Nanos3 and exhibit a complete loss of germ cells in the embryonic stage (Tsuda et al., 2003). RAREhspLacZ transgenic mice has been described (Rossant et al., 1991). In this study, adults were defined as mice of 8 weeks of age. Unless mentioned, mice were fed with NFM diet (Oriental Yeast), designated as ‘‘normal’’ or ‘‘VAS (vitamin A-sufficient)’’ diet. All mice were maintained, mated, operated on, and sacrificed with the permission of The Institutional Animal Care and Use Committee of National Institutes of Natural Sciences and Animal Research Committee of Kyoto University.

622

3.2.

MECHANISMS OF DEVELOPMENT

Generation of VAD model and readministration of VA

Vitamin A-deficient (VAD) mice were obtained according to the previously described protocol (van Pelt and de Rooij, 1990). Briefly, C57BL/6 females were fed a VAD diet (AIN-93G based, Japan CLEA) for at least 4 weeks before being crossed with males. During pregnancy and nursing, and after weaning, mothers and offspring were fed with VAD diets until the offspring showed VAD symptoms (at 11–16 weeks of age), as judged from reduction in body weight. ‘‘Incomplete’’ VAD mice were used at 9–12 weeks regardless of their body weight; VAD status was verified histologically based on the loss of differentiating spermatogonia and presence of spermatocytes and spermatids before analyses. To resume spermatogenesis, VAD and ‘‘incomplete’’ VAD mice were injected intraperitoneally with VA (0.5 mg retinol acetate (Sigma) dissolved in 25 lL of ethanol and mixed with 75 lL of sesame oil (Nacalai), and the diet was changed to normal until the animals were sacrificed.

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

and BM purple substrate (Roche), followed by counterstaining with Nuclear Fast Red. Templates used for probe preparation are summarized in Table S1. Corresponding sense probes provided no significant signals for all the genes tested (data not shown).

3.7.

Immunofluorescence was performed as described (Hara et al., 2009) with some modifications. Briefly, paraffin-embedded sections of PFA-fixed testes were deparaffinized and subboiled with sodium citrate buffer (pH6.0) for 10 min. Specimens were then incubated with the blocking reagent (Perkin–Elmer) and with anti-Aldh1a2 rabbit pAb (Abcam cat# ab75674; 1/4000 diluted with Can Get Signal Immunostain (TOYOBO)). Immunoreaction was visualized by Alexa Fluor 594-conjugated antirabbit IgG antibody (Invitrogen).

3.8. 3.3.

Artificial cryptorchidism

Two-month-old C57BL/6 male mice received operations to generate artificial cryptorchidism according to previously established methods (Nishimune et al., 1981). Two months later, some mice were experimentally injected with VA.

3.4.

Generation of patchy spermatogenesis

Busulfan (Wako) was initially dissolved in dimethyl sulfoxide (DMSO; Nacalai) and diluted further with an equal volume of PBS to obtain a final concentration of 4 mg/mL. Then, 2month-old C57BL/6 mice were interperitoneally administered with busulfan (30 mg/kg body weight). Two months later, when the tubules showed patchy spermatogenesis, they were anesthetized with Avertin and one of the testes was removed. VA was then injected, and the other testes were collected when the mice were sacrificed 2 days after VA injection.

3.5.

Histology and observation

Testes were fixed either by perfusion using 4% paraformaldehyde (PFA) in PBS or by immersion using 4% PFA in PBS or Bouin’s fixative (Muto Pure Chemicals) for several hours. After removal of the tunica albuginea, testes were immerged in the same fixative overnight, after which they were paraffinembedded and serial-sectioned for PAS-hematoxylin (PAS-H) staining or in situ hybridization (ISH). Stained sections were observed under a BX51 (Olympus) or a DMRBE (Leica) upright microscope. Stages of the seminiferous epithelium were determined using PAS-H-staining sections, according to previously established methods (Oakberg, 1956a; Russell et al., 1990a).

3.6.

Immunofluorescence

Quantitative RT-PCR

Total RNA was purified from testis samples using TRIzol reagent (Invitrogen) with DNaseI (Invitrogen) treatment. Random- and oligo d(T)-primed cDNA was synthesized from 1 lg of total RNA using SuperScript III First-Strand Synthesis SuperMix for qRT-PCR (Invitrogen). Primers were designed using Primer3 Plus software (http://www.bioinformatics.nl/cgi-bin/ primer3plus/primer3plus.cgi) as summarized in TableS2. qPCR was performed using LightCycler 480 System II (Roche) and SYBR qPCR Mix (TOYOBO) according to the manufacturer’s instructions. Expression of a particular gene was determined from three mice, each of which was measured as triplicate and averaged. The value was normalized using the expression of Gapdh.

Acknowledgments We are grateful to Y. Saga for providing the Nanos3+/Cre mice, J. Rossant for RAREhspLacZ mice, to A. Kakizuka for support, to T. Nakagawa for discussions and technical advice, to M. Fuller, T. Ogawa, T. Fujimori, Y. Kitadate, K. Hara, and H. Mizuguchi-Takase for comments, and to Y. Sukeno, Y. Kuboki, R. Ichikawa, K. Inada, K. Ikami, and M. Tokue for encouragements. This work was partly supported by a Grant-in-Aid for Scientific Research (KAKENHI) on Innovative Areas, ‘‘Regulatory Mechanism of Gamete Stem Cells,’’ the Uehara Memorial Foundation, the Nestle´ Nutrition Council, and the Mitsubishi Foundation, and National Institute for Basic Biology.

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.mod.2011.12.003

in situ hybridization

ISH using digoxygenin (DIG)-labeled RNA probes was performed in a stringent condition as described previously (Yoshida et al., 2001). After hybridization, DIG was visualized using an alkaline phosphatase-conjugated anti-DIG antibody

R E F E R E N C E S

Anderson, E.L., Baltus, A.E., Roepers-Gajadien, H.L., Hassold, T.J., de Rooij, D.G., van Pelt, A.M., Page, D.C., 2008. Stra8 and its

MECHANISMS OF DEVELOPMENT

inducer, retinoic acid, regulate meiotic initiation in both spermatogenesis and oogenesis in mice. Proceedings of the National Academy of Sciences of the United States of America 105, 14976–14980. Balmer, J.E., Blomhoff, R., 2002. Gene expression regulation by retinoic acid. Journal of Lipid Research 43, 1773–1808. Barroca, V., Lassalle, B., Coureuil, M., Louis, J.P., Le Page, F., Testart, J., Allemand, I., Riou, L., Fouchet, P., 2009. Mouse differentiating spermatogonia can generate germinal stem cells in vivo. Nature Cell Biology 11, 190–196. Bouillet, P., Oulad-Abdelghani, M., Vicaire, S., Garnier, J.M., Schuhbaur, B., Dolle, P., Chambon, P., 1995. Efficient cloning of cDNAs of retinoic acid-responsive genes in P19 embryonal carcinoma cells and characterization of a novel mouse gene, Stra1 (mouse LERK-2/Eplg2). Developmental Biology 170, 420–433. Chung, S.S., Wolgemuth, D.J., 2004. Role of retinoid signaling in the regulation of spermatogenesis. Cytogenetic and Genome Research 105, 189–202. de Rooij, D.G., 1998. Stem cells in the testis. International Journal of Experimental Pathology 79, 67–80. de Rooij, D.G., Russell, L.D., 2000. All you wanted to know about spermatogonia but were afraid to ask. Journal of Andrology 21, 776–798. Elftman, H., 1950. The sertoli cell cycle in the mouse. Anatomical Record 106, 381–392. Elizondo, G., Corchero, J., Sterneck, E., Gonzalez, F.J., 2000. Feedback inhibition of the retinaldehyde dehydrogenase gene ALDH1 by retinoic acid through retinoic acid receptor alpha and CCAAT/enhancer-binding protein beta. Journal of Biological Chemistry 275, 39747–39753. Franca, L.R., Ogawa, T., Avarbock, M.R., Brinster, R.L., Russell, L.D., 1998. Germ cell genotype controls cell cycle during spermatogenesis in the rat. Biology of Reproduction 59, 1371– 1377. Geissler, E.N., Ryan, M.A., Housman, D.E., 1988. The dominantwhite spotting (W) locus of the mouse encodes the c-kit protooncogene. Cell 55, 185–192. Hara, K., Kanai-Azuma, M., Uemura, M., Shitara, H., Taya, C., Yonekawa, H., Kawakami, H., Tsunekawa, N., Kurohmaru, M., Kanai, Y., 2009. Evidence for crucial role of hindgut expansion in directing proper migration of primordial germ cells in mouse early embryogenesis. Developmental Biology 330, 427–439. Hogarth, C.A., Griswold, M.D., 2010. The key role of vitamin A in spermatogenesis. Journal of Clinical Investigation 120, 956–962. Huang, H.F., Marshall, G.R., 1983. Failure of spermatid release under various vitamin A states – an indication of delayed spermiation. Biology of Reproduction 28, 1163–1172. Imamura, M., Inoguchi, T., Ikuyama, S., Taniguchi, S., Kobayashi, K., Nakashima, N., Nawata, H., 2002. ADRP stimulates lipid accumulation and lipid droplet formation in murine fibroblasts. American Journal of Physiology – Endocrinology and Metabolism 283, E775–83. Kawaguchi, R., Yu, J., Honda, J., Hu, J., Whitelegge, J., Ping, P., Wiita, P., Bok, D., Sun, H., 2007. A membrane receptor for retinol binding protein mediates cellular uptake of vitamin A. Science 315, 820–825. Kluin, P.M., Kramer, M.F., de Rooij, D.G., 1984. Proliferation of spermatogonia and sertoli cells in maturing mice. Anatomy and Embryology (Berlin) 169, 73–78. Leblond, C.P., Clermont, Y., 1952a. Definition of the stages of the cycle of the seminiferous epithelium in the rat. Annals of the New York Academy of Sciences 55, 548–573.

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

623

Leblond, C.P., Clermont, Y., 1952b. Spermiogenesis of rat, mouse, hamster and guinea pig as revealed by the periodic acid-fuchsin sulfurous acid technique. American Journal of Anatomy 90, 167–215. Liu, Y.X., Du, Q., Liu, K., Fu, G.Q., 1995. Hormonal regulation of plasminogen activator in rat and mouse seminiferous epithelium. Biological Signals 4, 232–240. Lok, D., Weenk, D., De Rooij, D.G., 1982. Morphology, proliferation, and differentiation of undifferentiated spermatogonia in the Chinese hamster and the ram. Anatomical Record 203, 83–99. Loudig, O., Babichuk, C., White, J., Abu-Abed, S., Mueller, C., Petkovich, M., 2000. Cytochrome P450RAI (CYP26) promoter: a distinct composite retinoic acid response element underlies the complex regulation of retinoic acid metabolism. Molecular Endocrinology 14, 1483–1497. Loudig, O., Maclean, G.A., Dore, N.L., Luu, L., Petkovich, M., 2005. Transcriptional co-operativity between distant retinoic acid response elements in regulation of Cyp26A1 inducibility. Biochemical Journal 392, 241–248. Mark, M., Jacobs, H., Oulad-Abdelghani, M., Dennefeld, C., Feret, B., Vernet, N., Codreanu, C.A., Chambon, P., Ghyselinck, N.B., 2008. STRA8-deficient spermatocytes initiate, but fail to complete, meiosis and undergo premature chromosome condensation. Journal of Cell Science 121, 3233–3242. Matson, C.K., Murphy, M.W., Griswold, M.D., Yoshida, S., Bardwell, V.J., Zarkower, D., 2010. The mammalian doublesex homolog DMRT1 is a transcriptional gatekeeper that controls the mitosis versus meiosis decision in male germ cells. Developmental Cell 19, 612–624. Morales, C., Griswold, M.D., 1987. Retinol-induced stage synchronization in seminiferous tubules of the rat. Endocrinology 121, 432–434. Murray, J.D., 2002. Mathematical biology, 3rd ed. Springer, New York. Nakagawa, T., Nabeshima, Y., Yoshida, S., 2007. Functional identification of the actual and potential stem cell compartments in mouse spermatogenesis. Developmental Cell 12, 195–206. Nakagawa, T., Sharma, M., Nabeshima, Y., Braun, R.E., Yoshida, S., 2010. Functional hierarchy and reversibility within the murine spermatogenic stem cell compartment. Science 328, 62–67. Nishimune, Y., Aizawa, S., Komatsu, T., 1978. Testicular germ cell differentiation in vivo. Fertility and Sterility 29, 95–102. Nishimune, Y., Haneji, T., Aizawa, S., 1981. Testicular DNA synthesis in vivo: changes in DNA synthetic activity following artificial cryptorchidism and its surgical reversal. Fertility and Sterility 35, 359–362. Noy, N., 1992. The ionization behavior of retinoic acid in lipid bilayers and in membranes. Biochimica et Biophysica Acta 1106, 159–164. Oakberg, E.F., 1956a. A description of spermiogenesis in the mouse and its use in analysis of the cycle of the seminiferous epithelium and germ cell renewal. American Journal of Anatomy 99, 391–413. Oakberg, E.F., 1956b. Duration of spermatogenesis in the mouse and timing of stages of the cycle of the seminiferous epithelium. American Journal of Anatomy 99, 507–516. Ohbo, K., Yoshida, S., Ohmura, M., Ohneda, O., Ogawa, T., Tsuchiya, H., Kuwana, T., Kehler, J., Abe, K., Scholer, H.R., Suda, T., 2003. Identification and characterization of stem cells in prepubertal spermatogenesis in mice. Developmental Biology 258, 209–225. Perey, B., Clermont, Y., Leblond, C.P., 1961. The wave of the seminiferous epithelium in the rat. American Journal of Anatomy 108, 47–77.

624

MECHANISMS OF DEVELOPMENT

Rossant, J., Zirngibl, R., Cado, D., Shago, M., Giguere, V., 1991. Expression of a retinoic acid response element-hsplacZ transgene defines specific domains of transcriptional activity during mouse embryogenesis. Genes and Development 5, 1333–1344. Russell, L.D., Ettlin, R.A., SinhaHikim, A.P., Clegg, E.D., 1990a. Histological and histopathological evaluation of the testis, 1st ed. Cache River Press, Clearwater, FL. Russell, L.D., Ren, H.P., Hikim, I.S., Schulze, W., Hikim, A.P.S., 1990b. A comparative-study in 12 mammalian-species of volume densities, volumes, and numerical densities of selected testis components, emphasizing those related to the sertoli-cell. American Journal of Anatomy 188, 21–30. Sada, A., Suzuki, A., Suzuki, H., Saga, Y., 2009. The RNA-binding protein NANOS2 is required to maintain murine spermatogonial stem cells. Science 325, 1394–1398. Schrans-Stassen, B.H., van de Kant, H.J., de Rooij, D.G., van Pelt, A.M., 1999. Differential expression of c-kit in mouse undifferentiated and differentiating type A spermatogonia. Endocrinology 140, 5894–5900. Shinohara, T., Orwig, K.E., Avarbock, M.R., Brinster, R.L., 2000. Spermatogonial stem cell enrichment by multiparameter selection of mouse testis cells. Proceedings of the National Academy of Sciences of the United States of America 97, 8346– 8351. Snyder, E.M., Davis, J.C., Zhou, Q., Evanoff, R., Griswold, M.D., 2011. Exposure to retinoic acid in the neonatal but not adult mouse results in synchronous spermatogenesis. Biology of Reproduction 84, 886–893. Suzuki, H., Tsuda, M., Kiso, M., Saga, Y., 2008. Nanos3 maintains the germ cell lineage in the mouse by suppressing both Baxdependent and -independent apoptotic pathways. Developmental Biology 318, 133–142.

1 2 8 ( 2 0 1 2 ) 6 1 0 –6 2 4

Timmons, P.M., Rigby, P.W., Poirier, F., 2002. The murine seminiferous epithelial cycle is pre-figured in the Sertoli cells of the embryonic testis. Development 129, 635–647. Tsuda, M., Sasaoka, Y., Kiso, M., Abe, K., Haraguchi, S., Kobayashi, S., Saga, Y., 2003. Conserved role of nanos proteins in germ cell development. Science 301, 1239–1241. van Beek, M.E., Meistrich, M.L., 1991. Stage-synchronized seminiferous epithelium in rats after manipulation of retinol levels. Biology of Reproduction 45, 235–244. van Pelt, A.M., de Rooij, D.G., 1990. Synchronization of the seminiferous epithelium after vitamin A replacement in vitamin A-deficient mice. Biology of Reproduction 43, 363–367. Vernet, N., Dennefeld, C., Guillou, F., Chambon, P., Ghyselinck, N.B., Mark, M., 2006a. Prepubertal testis development relies on retinoic acid but not rexinoid receptors in sertoli cells. EMBO Journal 25, 5816–5825. Vernet, N., Dennefeld, C., Rochette-Egly, C., Oulad-Abdelghani, M., Chambon, P., Ghyselinck, N.B., Mark, M., 2006b. Retinoic acid metabolism and signaling pathways in the adult and developing mouse testis. Endocrinology 147, 96–110. Yoshida, S., Ohbo, K., Takakura, A., Takebayashi, H., Okada, T., Abe, K., Nabeshima, Y., 2001. Sgn1, a basic helix-loop-helix transcription factor delineates the salivary gland duct cell lineage in mice. Developmental Biology 240, 517–530. Yoshida, S., Sukeno, M., Nakagawa, T., Ohbo, K., Nagamatsu, G., Suda, T., Nabeshima, Y., 2006. The first round of mouse spermatogenesis is a distinctive program that lacks the selfrenewing spermatogonia stage. Development 133, 1495–1505. Zhou, Q., Nie, R., Li, Y., Friel, P., Mitchell, D., Hess, R.A., Small, C., Griswold, M.D., 2008. Expression of stimulated by retinoic acid gene 8 (Stra8) in spermatogenic cells induced by retinoic acid: an in vivo study in vitamin A-sufficient postnatal murine testes. Biology of Reproduction 79, 35–42.