Differential responsiveness of spermatogonia to retinoic acid dictates precocious differentiation but not meiotic entry during steady-state spermatogenesis

Abstract

The foundation of mammalian spermatogenesis is provided by undifferentiated spermatogonia, which comprise of spermatogonial stem cells (SSCs) and transit-amplifying progenitors that differentiate in response to retinoic acid (RA) and are committed to enter meiosis. Our laboratory recently reported that the foundational populations of SSCs, undifferentiated progenitors, and differentiating spermatogonia are formed in the neonatal testis in part based on their differential responsiveness to RA. Here, we expand on those findings to define the extent to which RA responsiveness during steady-state spermatogenesis in the adult testis regulates the spermatogonial fate. Our results reveal that both progenitor and differentiating spermatogonia throughout the testis are capable of responding to exogenous RA, but their resulting fates were quite distinct—undifferentiated progenitors precociously differentiated and proceeded into meiosis on a normal timeline, while differentiating spermatogonia were unable to hasten their entry into meiosis. This reveals that the spermatogonia responding to RA must still complete the 8.6 day differentiation program prior to their entry into meiosis. Addition of exogenous RA enriched testes with preleptotene and pachytene spermatocytes one and two seminiferous cycles later, respectively, supporting recent clinical studies reporting increased sperm production and enhanced fertility in subfertile men on long-term RA analog treatment. Collectively, our results reveal that a well-buffered system exists within mammalian testes to regulate spermatogonial RA exposure, that exposed undifferentiated progenitors can precociously differentiate, but must complete a normal-length differentiation program prior to entering meiosis, and that daily RA treatments increased the numbers of advanced germ cells by directing undifferentiated progenitors to continuously differentiate.

Retinoic acid induces widespread precocity of progenitors, not differentiating spermatogonia, through steady-state spermatogenesis

Graphical abstract

Introduction

A key requirement for successful germ cell progression through spermatogenesis is provided by the essential micronutrient Vitamin A. Vitamin A deficiency in mammals results in reduced seminiferous tubule size, disrupted spermatogenesis, and impaired sperm production [1, 2]; these effects are reversible, as spermatogenesis resumes following the reintroduction of Vitamin A into the diet [3]. Dietary Vitamin A is converted into a retinyl ester for long-term storage or into retinal/retinaldehyde, which is further oxidized into the morphogen retinoic acid (RA). RA synthesis is achieved by retinaldehyde dehydrogenases (RALDHs), which are active in distinct testicular cells. In the developing testis, Sertoli cells provide the main source of RA, and this role is taken over by spermatocytes in adult testes [4, 5]. Macrophages located just outside the seminiferous tubules, some of which are positions at junctions between peritubular myoid cells, also contain RALDH enzymes and thus may contribute to testicular RA levels [6]. Once synthesized, dogma states that RA binds to a “retinoic acid receptor” (RAR, alpha, beta, and gamma genes) which then heterodimerizes with a “retinoid X receptor” (RXR, also alpha, beta, and gamma genes) to activate or repress transcription at RA response element-containing promoters of target genes [7]. RA signaling has been shown, both in somatic tissues as well as in spermatogonia, to also act through nongenomic means, for example by activating mTORC1 to enhance cap-dependent translation [8–10].

In the mammalian testis, divisions of spermatogonial stem cells (SSCs) produce daughter cells that either remain as SSCs (self-renewal) or proliferate as undifferentiated progenitors that will eventually differentiate in response to RA (Supplementary Figure S1) [11]. The requirement for RA in differentiation is unequivocal, and thus represents a pivotal spermatogonial cell fate detection, which is the commitment to initiate meiosis [12, 13]. Defining the regulatory logic underlying how spermatogonia manage this fate decision is fundamental to our understanding of how proper ratios of SSCs, undifferentiated progenitor, and differentiating spermatogonia are maintained to ensure life-long male fertility.

In human small-scale clinical trials, researchers assessed the ability for exogenous RA to enhance fertility of subfertile or infertile men with low sperm counts. When these men were given the RA analog Isotretinoin, they exhibited significantly improved sperm counts (+35% motile sperm in men with oligoasthenozoospermia) [14], and in some cases even fathering children, either through intercourse or by intracytoplasmic sperm injection [14, 15]. Despite these exciting preliminary findings, the mechanisms by which RA enhanced sperm production and the long-term outomes of continuous RA treatment on the male germline remain unknown.

In a recent report from our laboratory [16], the foundational spermatogonial populations in the neonatal testis were categorized both in vitro and in vivo based on their responsiveness to RA. SSCs could not respond to RA (RA-insensitive), undifferentiated progenitors could respond to RA but had not yet done so, and differentiating spermatogonia had already responded to RA. Based on our results in the developing testis, we proposed these responses were based on a lack of expression of a congante RAR (likely, RARG) in SSCs, and CYP26-mediated RA catabolism in undifferentiated progenitors [16]. Whether these mechanisms function similarly during steady-state spermatogenesis in juvenile and adult testes remains unknown. Broadly speaking, adult spermatogonial subtypes are predicted to manage their response to RA based on two non-mutually exclusive scenarios: (1) differential distribution of RA gradients within the testis; (2) cell type-specific differential expression of cognate RARs.

In the first, non-cell autonomous scenario, a differential gradient of the lipophilic morphogen RA is formed based on the balanced action of enzymes directing its synthesis (e.g. RALDHs) or catabolism (e.g. CYP26s) [4, 17–20]. This concept is supported by reports measuring RA levels, which differ along the mouse seminiferous tubules [5, 19, 21]. Specifically, RA levels are highest in tubule segments containing seminiferous epithelial stages VII–IX in the mouse [13, 22]. Perhaps coincidentally, these stages in mice contain germ cells undergoing key RA-related transitions such as spermatogonial differentiation, meiotic initiation, spermatid elongation, and spermiation [5]. This suggests RA levels themselves are regulatory, and predict RA levels would be limiting in stages I–VI and X–XII. An issue complicating extension of this model to other species—notably primates and humans—is that these key cell transitions do not similarly align in this spatial manner in those species [23–25]. Thus, it is difficult to conceive how these critical events are co-regulated by an RA gradient in primates and humans.

In the second, cell autonomous scenario, differential reception of the RA signal is thought to occur through binding one or more of the three RARs (RARA, RARB, RARG) [26], which are expressed in cell- and tissue-specific patterns. In the testis, RARB is undetectable, and male knockout mice were fertile [27]. RARA is expressed, albeit at different levels, in both spermatogonia and Sertoli cells, and is not essential for spermatogonial differentiation in global, spermatogonia- or Sertoli cell-specifc KO mouse models [28–30]. RARG is the most likely candidate for managing the spermatogonial RA response, since it is readily detectable in nuclei of RA-responsive undifferentiated progenitor spermatogonia [31]. In addition, when recombinant RARG was expressed precociously in SSCs and very early undifferentiated progenitors under control of the “glial cell line derived neurotrophic factor family receptor alpha 1” (Gfra1) promoter, those cells differentiated early in response to RA [32]. Confounding this issue, however, is that it was shown in mouse KO models that RARG was not essential for spermatogonial differentiation [31]. Therefore, at this point it can be best summarized that RARG is sufficient but not necessary for spermatogonial differentiation, and redundant mechanisms must exist to mediate spermatogonial response to RA.

Here, we examined the means by which spermatogonial responds to RA direct differentiation and meiotic initiation during steady-state spermatogenesis. Our results reveal that the adult mammalian testis contains a well-buffered system that employs multiple modalities to manage the spermatogonial response to RA during steady-state spermatogenesis. In addition, we found that while exogenous RA did not hasten entry into meiosis, as recently reported [33], it did advance undifferentiated progenitor spermatogonia into the 8.6 day differentiation program, leading to increased numbers of spermatocytes 2 weeks later. Ultimately, this buffering system ensures the preservation of long-term spermatogenesis, but could be temporarily overridden to increase sperm output to enhance fertility.

Materials and methods

Animal care and treatments

All animal procedures adhered to the National Research Council Guide for the Care and Use of Laboratory Animals guidelines and were approved by the East Carolina University Animal Care and Use Committee (A3469-01). Mouse strains used for this study were outbred CD-1 (Charles River Laboratories, stock #022) and inbred C57Bl/6 N (Charles River Laboratories, stock #027).

Adult mice (~6–10 weeks of age) were subjected to at least one of the following treatments: (1) 100 μl of sterile RA (R2625 Sigma Aldrich; 10 μg/μl via intraperitoneal (IP) injection; (2) 5 μg/g body weight of the CYP26 inhibitor talarozole (Cat #HY-14531; Medchem Express; 1 μg/μl via oral administration) prepared in sterile vehicle (DMSO); (3) equivalent volume of sterile DMSO via IP injection or oral administration as vehicle-only control. Mice were age-matched prior to receiving treatments.

Synchronized spermatogenesis was performed similar to previous reports as a model for steady-state spermatogenesis [22, 34, 35]. Briefly, WIN 18 446 (bis-(dichloroacetyl)-diamine; Cayman Chemical, #14018) was resuspended in DMSO (final concentration = 100 μg/g of body weight) and given orally to CD-1 mice for 10 consecutive days (postnatal days (P)1–P10) to block spermatogonial differentiation. On P11, mice received one subcutaneous injection of 10 μl exogenous RA (10 μg/μl) to stimulate differentiation. WIN 18 446 treatments may continue post-differentiation if mice are subjected to an “RA-deficient” synchronization schedule. Mice received either one injection of RA at P11 or two injections at either P11 + P14 or P11 + P16. Mice were humanely euthanized via CO₂ asphyxiation followed by cervical dislocation.

Tissue preparation and histological staining

Testes were fixed in either Bouin’s solution or 4% paraformaldehyde (PFA) for 24 h at 4°C and subsequently washed in multiple changes of 1X PBS overnight. Bouin’s-fixed samples were dehydrated through an ethanol series and processed for paraffin embedding using standard methods. Five micrometer tissue sections were subjected to Periodic Acid-Schiff’s staining using conventional methods. Coverslips were mounted with Permount, and images taken using an Axio Observer A1 microscope equipped with an Axiocam 503 color digital camera (Carl Zeiss Microscopy).

Indirect immunofluorescence

Immunostaining was performed as before [10, 35–37]. Briefly, PFA-fixed testes were washed in 1X PBS overnight, soaked in 30% sucrose for 24 h, and frozen in O.C.T. compound (Tissue Tek) prior to sectioning. Five micrometer sections were incubated in blocking solution (1X PBS + 3% BSA + 0.1% Triton X-100) for at least 30 min at room temperature. Thereafter, primary antibodies (Table 1) were diluted with blocking solution and incubated on tissue sections for 1 h at room temperature. One section per slide was incubated in blocking solution without primary antibody as a negative control. After 15 min in washing solution (1X PBS + 0.1% Triton X-100), sections were incubated in blocking solution with corresponding fluorescent secondary antibody (1:500) for 1 h at room temperature. Slides were washed, covered in DAPI with mounting medium (EnQuire BioReagents) or 1:1 1X PBS/Glycerol solution, and coverslipped. All incubation and wash steps were performed in humid opaque containers. Imaging was done using either the Celldiscover7 (Carl Zeiss Microscopy) or Fluoview FV1000 laser scanning confocal microscope (Olympus America).

Table 1.

Antibodies and cell labeling reagents

Antigen	Host	Source	Dilution	Catalog #
GFRA1	Goat	R&D Systems	1:800	AF560
H1F6	Guinea pig	Gift from Handel	1:500	Inselman et al 2003
KIT	Goat	R & D Systems	1:1000	AF1356
Lectin-488		Thermo Fisher Scientific	1:1000	L21409
Lectin-594		Thermo Fisher Scientific	1:1000	L32459
RARG	Rabbit	Cell Signaling	1:500	8965 T
STRA8	Rabbit	Abcam	1:3000	Ab49602
SYCP3–488	Mouse	Abcam	1:200	Ab205846
TRA98	Rat	Abcam	1:1000	Ab82527
ZBTB16/PLZF	Goat	R & D Systems	1:1000	AF2944

Testicular single-cell suspensions

Adult mouse testes were used to make single-cell suspensions, as previously described [16]. Briefly, testes were detunicated and transferred to a solution containing 4.5 ml of 0.25% trypsin (Gibco) and 0.5 ml DNase1 (7 mg/ml, Sigma-Aldrich) and incubated in a 37°C water bath for 3 min. An additional 1 ml DNase1 was added and the mixture triturated to further break up the testes, followed by a 1 min incubation in a 37°C water bath. One milliliter of fetal bovine serum (ATCC) was added to deactivate the trypsin. The cell mixture was filtered through a 40 μm sieve and centrifuged at 500 x g for 7 min. The resulting cell pellet was suspended in DMEM media (Gibco) containing 1X penicillin–streptomycin (Thermo-Fisher Scientific). Cells were plated at a density of 1.4 × 10⁶ cells/well into a 12-well plate containing glass coverslips and allowed to adhere for 2 h at 37°C with 5% CO₂ prior to treatment. Cells were treated with either vehicle control (DMSO) or talarozole (25 nM) for 2 h, followed by an addition of vehicle or RA (0.001, 0.01, and 1.0 μM) for 8 h. These concentrations were determined based on published IC(50) values shown to achieve effective inhibition of the CYP enzymes at the minimum dosage [17]. These cells were cultured for a total of 12 h, and experiments were repeated in triplicate.

Quantitative reverse transcriptase polymerase chain reaction

Total RNA was isolated from tissues using the miRNeasy Mini Kit (Qiagen) following the manufacturer’s protocol and quantified using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific). Quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) was performed using the iTaq Universal SYBR Green One-Step Kit (Bio-Rad) using a StepOnePlus real-time PCR thermocycler (Applied Biosystems). Message abundance was normalized to the average C_t values of the housekeeping gene Rpl19 and expressed as 2^−ΔC_t. Primer sets are listed in Table 2.

Table 2.

qRT-PCR primers

Primer name	Sequence (5′ to 3′)
Rpl19-F	GAAATCGCCAATGCCAACTC
Rpl19-R	TCTTAGACCTGCGAGCCTCA
STRA8-F	GAGGTCAAGGAAGAATATGC
STRA8-R	CAGAGACAATAGGAAGTGTC

Statistics and data analyses

Student t-test, one-way analysis of variance, and Tukey test were utilized to determine statistical significance at P < 0.05. For in vivo assays, at least three mice were used per age per treatment. For adult in vivo experiments, the average number of tubule cross-sections quantified per testis exceeded 100. Quantification was done by counting the number of marker(s) + germ cells per tubule perimeter. This method was chosen to minimize the effects of varying tubule cross-section presentations (e.g. transverse versus oblique versus longitudinal) in data collection. The perimeters of seminiferous tubules were manually traced. Negative controls and background intensity were used to visually distinguish marker+ from marker− cell types. Counting of cells was first done automatically using QuPath, then verified (and/or adjusted) through manual counts. For single-cell suspension experiments, trials were performed thrice. QuPath, Microsoft Excel, and GraphPad Prism were used for data collection. GraphPad Prism and ImageJ were both used for formatting figures and data.

Results

Exogenous RA induces widespread spermatogonial response across the testis

The extent to which male germ cells undergoing adult steady-state spermatogenesis are capable of responding to RA was first determined by examining testes of mice treated once with 100 μl of vehicle or 10 μg/μl RA via IP injection. CD-1 mice were euthanized 12, 24, and 48 h after treatment, and expression of “stimulated by RA gene 8” (STRA8) protein was used as an established proxy for RA-responsive germ cells within the testis [8, 16, 38]. Immunostaining revealed significantly the increased numbers of STRA8+ spermatogonia in RA-treated testes 12 h after treatment, with ~80% of tubules containing STRA8+ spermatogonia (~3.5 fold increase over vehicle-treated controls (Figure 1A, D, G; Supplementary Figure S2)). Twenty-four hours post-RA, numbers of STRA8+ spermatogonia were significantly lower than at 12 h post-RA, but greater than in vehicle-treated testes (Figure 1B, E, G; Supplementary Figure S2). STRA8 was undetectable 48 h later, as both the numbers of STRA8+ spermatogonia and the percentage of tubules with STRA8+ spermatogonia in RA-treated testes were not different from vehicle-treated controls (Figure 1C, F, G; Supplementary Figure S2). To ensure these results were not strain-dependent, treatments were repeated in adult mice on a C57Bl/6 N background. Immunostaining revealed similar STRA8 protein levels in both vehicle- and RA-treated testes within the 48 h acute response window (Supplementary Figure S3) to those evidenced in CD-1 mice (Figure 1). Collectively, these findings reveal that RA-responsive spermatogonia are located throughout adult testes, supporting the concept of differential localized RA concentrations along the length of the seminiferous tubules.

Fig. 1.

Exogenous RA drives transient STRA8 expression in the adult testis. (A–G) Indirect immunofluorescence (IIF) for RA response marker (STRA8) in green, a pan germ cell marker (TRA98) in red, and pan nuclei marker (DAPI) in blue in vehicle-treated (A–C) and RA-treated testes (D–F) of adult mice. Treatment and time-of-sacrifice post-treatment indicated in each image. The yellow arrows indicate tubules with spermatogonia responding to exogenous RA. Scale bar = 100 μm. G) Quantification of the number of STRA8+ spermatogonia per tubule perimeter (mm). *P < 0.05 versus 12 h vehicle-treated and #P < 0.05 versus 12 hr RA-treated. (H, I) Staging of seminiferous tubules using lectin (green) and DAPI (blue) to identify locations of STRA8+ germ cells (red). Scale bar = 500 μm. (I) Distribution of seminiferous tubule stages containing STRA8+ germ cells per treatment using criteria established in literature and grouped by like characteristics: Group 1 = Stages I–III; Group 2 = Stage IV–VI; Group 3 = Stages VII–VIII; Group 4 = Stage IX; Group 5 = Stages X–XII. (J) Location of spermatogonia responding to exogenous RA are concentrated in stages prior to the onset of meiosis and differentiation (indicated by the broken red circle). The diagram is created using BioRender.com.

The precise timing of mouse spermatogenesis enables subdivision of the adult seminiferous epithelia into 12 specific stages (I–XII), with each tubule section containing an expected profile of specific germ cell types [39]. Mouse seminiferous tubule stages VII–IX were previously shown to have higher levels of RA than other stages [19], and it is during stage IX that differentiation occurs to form type A₁ spermatogonia [40, 41]. Using the established staging criteria [39], fluorescent lectin labeling to visualize spermatid acrosomes, and immunostaining for STRA8, the stages containing RA-responsive germ cells were identified. For simplicity of data interpretation, the 12 stages were grouped together based on like characteristics. In vehicle-treated testes, STRA8+ germ cells were identified only in stages VII–VIII tubules. In comparison, all tubule stage groupings 12 h after RA treatment contained STRA8+ germ cells (Figure 1H, I; Supplementary Figure S4). These RA-responsive germ cells in RA-treated testes (versus vehicle-treated controls) were predominantly in stages I–VI (Figure 1J).

Within the testis, “RA receptor gamma” (RARG) is restricted to spermatogonia and is thus predicted to be a primary determinant in WT testes for RA responsiveness [16, 42]. Therefore, to identify potential RA-responsive spermatogonia, RARG was immunolocalized in adult testes. Surprisingly, the numbers of RARG+ spermatogonia increased ~2.5 fold compared with vehicle-treated controls 12 h post-RA treatment (Figure 2A–E; Supplementary Figure S5). This increase in RARG+ spermatogonia was retained at 24 h post-RA, and higher than vehicle-treated samples, but significantly less than at 12 h post-RA treatment (Figure 2B, C, E; Supplementary Figure S5). Similar to the transient increase in STRA8+ germ cells, numbers of RARG+ spermatogonia decreased 48 h post-RA treatment, comparable with vehicle-treated testes (Figure 2D, E; Supplementary Figure S5). To determine the locations of RARG+ germ cells, staging of tubules was performed by immunostaining for RARG, along with lectin and DAPI. RARG+ spermatogonia were predominantly housed within stages VI–IX of vehicle-treated testes, in agreement with prior reports of RARG expression in late undifferentiated progenitor and type A₁ differentiating spermatogonia [31, 42]. In comparison, RARG+ spermatogonia were present 12 h after RA treatment through testes in all tubule stage groupings (Figure 2F–J).

Fig. 2.

Spermatogonia respond to RA through RARG signaling during steady-state spermatogenesis. (A–D) IIF was done to quantify the number of germ cells (labeled in red) expressing the RA receptor RARG (cells in green) in adult mice. DAPI labels all cell nuclei in blue. Treatment and time-of-sacrifice post-treatment indicated in each image. The yellow circles encompass RARG+ germ cells. The white box corresponds to the zoomed-in region (see insert). DAPI has been removed from insert. (E) Quantification of the number of RARG+ spermatogonia per tubule perimeter (mm). *P < 0.05 versus vehicle-treated, and #P < 0.05 versus 12 h RA. (F–J) Staging of tubules was performed by co-immunostaining RARG (green), lectin (red), and DAPI (blue) to identify the location of RARG+ spermatogonia RA-treated (F–J). Stage of tubule is indicated in each image. The yellow arrows indicate RARG+ spermatogonia. The white box corresponds to the zoomed-in region (see insert). Scale bar for all images = 100 μm.

Since stages I–VI contain SSCs (predominantly type A_s) as well as undifferentiated progenitor (types A_pr, and A_al) and differentiating (types A₄ – B) spermatogonia [39, 41, 43, 44], the specific subtypes were identified that responded to RA. First, we verified adult SSCs were RA insensitive, as in the developing testis [16], by co-immunostaining for STRA8 and “Glial cell line derived neurotrophic factor family receptor alpha 1” (GFRA1), the most restricted fate marker for SSCs and early progenitors [45]. Indeed, we did not observe GFRA1+/STRA8+ spermatogonia 12 h post-RA (Supplementary Figure S6A). These results were corroborated by staining of testes 12 h post-RA with RARG and GFRA1—indeed, GFRA1+ germ cells were predominantly RARG- (Supplementary Figure S6B). Next, the numbers of RA-responsive undifferentiated progenitors were identified by co-immunostaining for both STRA8 and “zinc finger and BTB domain containing 16” ZBTB16 (also termed PLZF), an established marker of undifferentiated progenitor spermatogonia [46]. Twelve hours post-RA, there was a ~ 3.5-fold increase in STRA8+/ZBTB16+ undifferentiated spermatogonia (Figure 3A–E; Supplementary Figure S7), and ~80% of tubules contained these cells. By 24 h post-RA, numbers of STRA8+/ZBTB16+ spermatogonia were comparable with vehicle-treated controls (Figure 3C, E; Supplementary Figure S7). Staging these tubules 12 h after RA revealed that these RA-responsive undifferentiated progenitors were present in all tubule stages throughout the testis (Figure 3F–J). To quantify adult RA-responsive differentiating spermatogonia, co-immunostaining was performed using markers for STRA8 along with “kit proto-oncogene receptor tyrosine kinase” (KIT), a bona fide marker expressed at high levels in differentiating spermatogonia and a lesser extent in preleptotene spermatocytes [47, 48]. Twelve hours post-RA, there were significant increases in numbers of STRA8+/KIT+ germ cells (Figure 4A–E; Supplementary Figure S8). By 24 h post-RA, the numbers were higher than vehicle-treated controls, but significantly lower than 12 h RA-treated testes. By 48 h post-RA, the numbers of STRA8+/ KIT+ germ cells were comparable between RA- and vehicle-treated testes (Figure 4A–E; Supplementary Figure S8). Staging these tubules 12 h after RA revealed RA-responsive KIT+ germ cells (likely differentiating spermatogonia) in all tubule stages (Figure 4F–J). Collectively, these results reveal that both undifferentiated progenitor and differentiating spermatogonia throughout the adult testis have the capacity to respond to RA.

Fig. 3.

Transient RA-responsive spermatogonia are predominantly undifferentiated progenitors. (A–E) IIF was done to quantify the number of progenitors (ZBTB16+ cells in red) responding to RA (STRA8+ cells in green) in adult mice. DAPI labels all cell nuclei in blue. Treatment and time-of-sacrifice post-treatment indicated in each image. The yellow circles encompass STRA8+ ZBTB16+ cells. The white box corresponds to the zoomed-in region (see insert). DAPI has been removed from insert. E) Quantification of the number of STRA8+ progenitors spermatogonia per tubule perimeter (mm). *P < 0.01 versus vehicle-treated, and #P < 0.01 versus 12 hr RA. (F–J) Staging of tubules was performed through co-immunostaining of ZBTB16 (red), STRA8 (green), DAPI (blue), and lectin (magenta) to identify the location of STRA8+ progenitors in RA-treated testes euthanized 12 h post-treatment. Stage of tubule is indicated in each image. The yellow arrows indicate STRA8+ progenitors. The white box corresponds to the zoomed-in region (see insert). DAPI has been removed from insert. Scale bar for all images = 100 μm.

Fig. 4.

Transient STRA8+ spermatogonia throughout the testis can enter differentiation in response to exogenous RA. (A–E) IIF was done to quantify the number of spermatogonia that undergo differentiation (KIT+ germ cells in green) due to the RA signal (STRA8+ cells in red) in adult mice. TRA98, a pan germ cell marker, is shown in blue. Treatment and time-of-sacrifice post-treatment indicated in each image. The yellow circles encompass STRA8+ KIT+ germ cells. The white box corresponds to the zoomed-in region (see insert). TRA98 has been removed from insert. E) Quantification of the number of STRA8+/KIT+ germ cells per tubule perimeter (mm). *P < 0.05 versus vehicle-treated, and #P < 0.05 versus 12 hr RA. (F–J) Staging of tubules was performed through co-immunostaining of KIT (red), STRA8 (green), DAPI (blue), and lectin (magenta) to identify the location of STRA8+ germ cells in RA-treated testes euthanized 12 hr post-treatment. Stage of tubule is indicated in each image. The yellow arrows indicate STRA8+ KIT+ germ cells. The white box corresponds to the zoomed-in region (see insert). DAPI has been removed from insert. Scale bar for all images = 100 μm.

CYP26-mediated catabolism partially mediates the male germ cell RA response in the adult testis

ZBTB16+ germ cells were shown to be RA-responsive in all tubule stages 12 h post-RA treatment (Figure 3), but were somehow protected from differentiating in the unaltered WT testis until stages VII–IX. We previously showed that, in the developing testis, CYP26-based catabolism strongly protected undifferentiated progenitor spermatogonia from precociously differentiating in response to RA [16]—the number of STRA8+ germ cells in RA-treated testes was identical to the number of STRA8+ germ cells in mice treated with talarozole, a specific and potent CYP26 inhibitor [49]. Here, we examined the extent to which CYP26-mediated catabolism also modulated responsiveness of adult undifferentiated progenitor spermatogonia. To assess this, adult mice were treated with vehicle-alone, vehicle + RA, or the specific and potent CYP26 inhibitor talarozole [49]. Immunostaining of testes of mice 12 h after treatment revealed that talarozole treatment increased the numbers of STRA8+/ZBTB16+ undifferentiated progenitor spermatogonia ~2-fold compared with vehicle-treated controls (Figure 5A–D). Interestingly, talarozole treatment increased numbers of STRA8+ undifferentiated progenitor spermatogonia (as well as the percentage of STRA8+/ZBTB16+ tubules), but not to the same extent as RA-treated testes (Figure 5A–E). To validate these in vivo results, adult testes were cultured as single-cell suspensions and treated with vehicle-alone or talarozole with increasing RA concentrations (0.001–1 μM). qRT-PCR revealed that the Stra8 mRNA levels were significantly higher in talarozole + RA-treated samples compared with those treated with vehicle + RA (Figure 5F). Collectively, these results reveal that RA catabolism is involved in regulating RA levels and thus germ cell RA response during adult steady-state spermatogenesis, however RA catabolism does not provide the same level of protection in the adult as in the developing testis.

Fig. 5.

RA catabolism limits the number of progenitors able to respond to RA during steady-state spermatogenesis. (A–E) IIF was done to quantify the number of STRA8+ progenitors post-treatment in adult mice. STRA8+ cells are labeled in green and ZBTB16+ progenitors are labeled in red. Treatment denoted in each image. Talarozole = CYP26B1/RA catabolism inhibitor. The yellow circles indicate STRA8+ progenitors. Scale bar for all images = 100 μm. (D, E) Quantification for the number of STRA8+ progenitors per tubule diameter (D) and the percentage of tubules that contain STRA8+ progenitors (E). * P < 0.01 versus vehicle-treated testes and # P < 0.05 versus vehicle + RA-treated testes. (F) Whole testis single-cell suspensions were treated with vehicle or talarozole with increasing concentrations of RA. Samples were collected after 12 h for qRT-PCR of Stra8 mRNA expression. *P < 0.05 versus Vehicle only, #P < 0.01 versus same treatments without talarozole (e.g. talarozole + vehicle versus vehicle only).

A single pulse of RA caused undifferentiated progenitor spermatogonia to precociously differentiate but still enter meiosis ~8 days later, on the expected timeline

We next examined whether the response to RA by the undifferentiated progenitors was transient or if they became committed to differentiate and ultimately enter meiosis. To distinguish between these possibilities, adult mice were injected with vehicle-alone or RA (=day 1) and euthanized 8 days later (=day 9) (Figure 6A). This timeline was selected because the length of one cycle of the seminiferous epithelium (I–XII) in mice is 8.6 days [40], and 8 days after responding to exogenous RA, spermatogonia committed to differentiation would be expected to be entering meiosis as preleptotene spermatocytes [40]. Co-immunostaining was done to identify preleptotene spermatocytes based on strong nuclear STRA8 signal and punctate staining for the spermatocyte marker “synaptonemal complex protein 3” (SYCP3) [50]. Compared with vehicle-treated controls, testes of RA-treated mice had ~30% more preleptotene spermatocytes, which were present in more tubules (~1.8-fold increase) (Figure 6B–E). These results suggest undifferentiated progenitors that responded to exogenous RA committed to and completed differentiation and entered meiosis on an expected timeline.

Fig. 6.

RA-responsive progenitors progress through differentiation and enter meiosis. (A) Adult mice were treated once with vehicle or RA (D1) and euthanized on D9. A_undiff = undifferentiated progenitors and PL. Spc. = preleptotene spermatocytes. The diagram is created using BioRender.com. (B, C) Immunofluorescence was performed to identify preleptotene spermatocytes, which co-express STRA8 (green) and SYCP3 (red). TRA98, a pan germ cell marker, is in blue. The yellow asterisks indicate tubules with preleptotene spermatocytes. Scale bar for all images = 100 μm. (D, E) Quantification of the number of preleptotene spermatocytes per tubule perimeter (mm) (D) and the percentage of tubules with preleptotene spermatocytes (E). *P < 0.05 versus vehicle-treated testes.

RA exposure does not hasten differentiating spermatogonia to precociously enter meiosis

During steady-state spermatogenesis, the RA-responsive spermatogonial pool contains both undifferentiated progenitors and differentiating spermatogonia, which had already responded to RA. It was previously reported that differentiating spermatogonia, in mice that received a second bolus of exogenous RA, would exit the differentiation program and enter meiosis [33]. To test whether differentiating spermatogonia at the midpoint or near the end of differentiation were indeed awaiting a second pulse of RA for meiosis, an in vivo model of synchronized adult steady-state spermatogenesis was used (Supplementary Figure S9).

Spermatogenesis was synchronized by treating mice from P1–P10 with WIN 18 446, a potent and selective inhibitor of RA synthesis (Supplementary Figure S9) [35, 51]. On P11, a single bolus of RA was injected to initiate differentiation. Groups of mice with synchronized spermatogenesis were then subjected to one of the following treatments: (1) RA once on P11, to induce differentiation and formation of type A₁ spermatogonia at the beginning of the differentiation program; (2) RA on P11 and P14, when testes contained type A₃ spermatogonia near the midpoint of the differentiation program; (3) RA on P11 and P16, when testes contained type Intermediate (In) spermatogonia near the end of the differentiation program [35]. To ensure the synchronized model properly mirrored events in steady-state spermatogenesis, co-immunostaining for STRA8 and KIT was done to detect RA-responsive spermatogonia. As in the adult testes (Figure 4), differentiating spermatogonia were STRA8+/KIT+ 24 h post-RA (Supplementary Figure S10A, B). To verify STRA8 was induced specifically in response to the second exogenous pulse of RA, 2X RA-treatment experiments were repeated but mice were subjected to RA-deficient synchronization—mice were continuously treated with WIN 18 446 after P11 to ensure no endogenous RA was produced within the testis and the only RA used for germ cell progression was provided via injection (Supplementary Figure S10A, C). In these mice, differentiating spermatogonia were also STRA8+ 24 h post-RA treatment. The histology of P19 testes (with preleptotene spermatocytes as the most abundant germ cell type) [35] that were treated with a single pulse of RA on P11 was compared with those that received second pulses at P14 or P16. If the second pulse of RA had hastened meiosis, then the majority of germ cells in those testes should be advanced beyond preleptonema. Surprisingly, there were no apparent differences in the germ cell complement in testes from mice receiving only the single RA injection with those receiving a second, revealing no evidence for precocious meiotic entry (Figure 7A, B; Supplementary Figure S11). To verify these results, spermatocyte identity was assessed by co-immunostaining for STRA8 (an established marker of preleptotene spermatocytes [33, 38]) and SYCP3, which exhibits a characteristic pattern during each stage of meiosis. In this steady-state synchronization model with a single pulse of RA at P11, testes were enriched with spermatocytes in preleptonema at P19, leptonema/zygonema at P21, and early pachynema at P23 [35]. If a second pulse of RA at P14 or P16 indeed advanced entry into meiosis, then spermatocytes in leptonema, zygonema, and/or pachynema would predominate at P19. However, testes from mice receiving the second pulse of RA at either P14 or P16 did not contain significantly more spermatocytes beyond preleptonema than controls receiving a single pulse at P11 (Figure 7C). To verify that this lack of advancement of differentiated spermatogonia into meiosis phenotype was not unique to the steady-state synchronization model, adult mice were treated with vehicle or RA and euthanized 24, 48, and 72 h later and histologically examined for accumulation of preleptotene or leptotene spermatocytes, which would indicate a truncation of differentiation and hastened entry into meiosis. In comparing the testes of RA-treated mice with vehicle-treated mice, there were no apparent differences in the number of spermatocytes within the tubules and no enrichment of advanced spermatocytes (e.g. more leptotene, zygotene spermatocytes) was evidenced following RA treatment (Supplementary Figure S12). Collectively, these findings reveal that a single pulse of exogenous RA neither shortened the differentiation timeline nor hastened meiotic entry or progression.

Fig. 7.

Exogenous RA does not shorten the differentiation timeline in the synchronized steady-state model. (A) Timeline denoting treatment schedule and germ cells that predominate each postnatal age. WIN 18 446 = RA synthesis inhibitor. Undiff = undifferentiated progenitors. PL = preleptotene spermatocytes. Lep = leptotene spermatocytes. Zyg = zygotene spermatocytes. Pach = pachytene spermatocytes. RA was exogenously administered on P11 only, P11 and P14, or P11 and P16. (B) Periodic acid Schiff’s histological staining was performed to phenotype spermatocytes. Age and treatment schedule are denoted in the image. Scale bar for all images = 100 μm. (C) Quantification of the germ cell type breakdown at P19. The colors correspond to legend below the X-axis.

Continuous RA administration enriches production of advanced spermatogenic cells

We next examined whether multiple pulses of RA were required for spermatogonia to precociously exit differentiation and enter meiosis. To test this, groups of adult mice were injected daily with vehicle-alone or RA for eight consecutive days (D1–D8) and euthanized on D9 or D17 (Figure 8A). On this treatment schedule, germ cells are predicted to complete one seminiferous cycle by D9 and two cycles by D17. Co-immunostaining was done to detect both SYCP3 and “H1.6 linker histone, cluster member” (HIF6, also termed H1T), the latter of which is detectable in spermatocytes from mid-pachynema onward and in round and elongating spermatids [52, 53]. At D9, there were no significant differences in numbers of H1F6+ spermatocytes in testes from mice receiving vehicle-alone or daily pulses of exogenous RA (Figure 8A–C)—demonstrating differentiating spermatogonia were unable to precociously exit differentiation and enter meiosis despite administration of multiple dosages of RA. At D17, after two seminiferous epithelial cycles, significantly more H1F6+ spermatocytes (16% increase) were observed in testes from mice receiving eight daily doses of RA (Figure 8A–C). This reveals that daily RA exposure provided a continual stimulation of undifferentiated progenitor spermatogonia to differentiate, thus forming a larger-than-normal cohort of germ cells proceeding together through spermatogenesis.

Fig. 8.

Consistent RA treatment enriches a generation of germ cells proceeding through spermatogenesis. (A) Adult mice were treated daily for 8 days with vehicle or RA (D1–8) and euthanized on D9 or D17. A_undiff = undifferentiated progenitors, A_diff = differentiated spermatogonia, and Pach. Spc. = pachytene spermatocytes. The diagram is created using BioRender.com. (B, C) Mid-and late-stage pachytene spermatocytes were quantified by co-immunostaining for SYCP3 (green) and H1F6 (red). DAPI stains all nuclei in blue. (C) Quantification for H1F6+ spermatocytes per tubule perimeter. *P < 0.05 versus vehicle. (D–F) Immunofluorescence to quantify the number of progenitors (ZBTB16+ germ cells) on D17 after consistent vehicle (D) or RA (E) treatments. ZBTB16+ cells are labeled in green, TRA98+ germ cells in red, and all nuclei (DAPI) in blue. (F) Quantification of the number of ZBTB16+ progenitors per tubule perimeter. NS = not-significant (P < 0.05). Scale bar for all images = 100 μm.

A concern with repeated dosages of RA is the long-term effect on the germline, with the potential that transiently enriched numbers of differentiating spermatogonia would occur at the expense of the subsequent germ cell cohort or the self-renewing SSC pool. Thus, we wanted to ensure the germline was preserved in the presence of RA injections for 8 days. To assess this, numbers of ZBTB16+ undifferentiated progenitor spermatogonia were quantified in D17 testes after daily injection of vehicle-alone or RA. Since SSCs are RA-insensitive (Supplementary Figure S6 and [16]), we predicted they would not differentiate in response to these daily RA pulses and thus continue to generate undifferentiated progenitor spermatogonia. Indeed, the numbers of undifferentiated progenitor spermatogonia were similar between vehicle- and RA-treated testes at D17 (Figure 8D–F). Altogether, these results reveal that consistent RA exposure can enrich a generation of germ cells without adversely affecting the SSCs, and provide further evidence that the RA insensitivity is a key feature of these cells.

Discussion

Here, we expand on our prior results in the developing testis [16] to examine the relationship between differential responses to RA and spermatogonial development during steady-state spermatogenesis. We discovered that undifferentiated progenitor spermatogonia outside seminiferous epithelium stages VII–VIII precociously responding to exogenous RA both committed to and completed differentiation before entering meiosis ~8 days later as preleptotene spermatocytes. CYP26-mediated RA catabolism played a role, albeit not a primary one as in the neonatal testis, in limiting the exposure of undifferentiated progenitor spermatogonia to RA. Differentiating spermatogonia were also RA-responsive, but their response did not shorten the 8.6 day differentiation timeline, and it did not hasten their entry into or progression through meiosis. Daily exposure for an entire 8 day seminiferous epithelial cycle increased the numbers of advanced germ cells, presumably by directing undifferentiated progenitors to continuously commit to differentiate and enter meiosis, and this treatment did not appear to adversely affect SSCs. Results here provide an explanation for recent findings from small-scale clinical trials, in which repeated treatment with retinoids transiently increased sperm output and enhanced male fertility [14, 15].

Our results revealed that exogenous RA drives widespread germ cell response throughout the testis. We chose to focus on the acute response to RA and detailed the degree of response over 48 h. This response was indeed transient, with STRA8 signal diminishing over time, such that only germ cells in specific tubule stages (VII–VIII) expressed STRA8. Due to this acute window, we noted the presence of RA-responsive progenitors and differentiating spermatogonia in all seminiferous tubule stage groupings. These results contrast with a recent report of STRA8+ undifferentiated progenitors post-RA injection only in stages II–VI and not in later stages [33]. This discrepancy is likely based on the different time of analyses—here, we assessed germ cell response at multiple timepoints post-treatment, while in that report it was 24 h post-treatment. Also noteworthy, we evidenced this transient spermatogonial STRA8 expression in multiple genetic backgrounds—both CD-1 and C57Bl/6 N as well as on a CD-1xC57Bl/6 J mixed background (data not shown).

STRA8 expression is commonly used in the field as a proxy for RA signaling [33, 38, 54–57]. Although the specific function of this putative transcription factor remains unclear, STRA8 is unique to germ cells and to vertebrates, and has been associated with expression of pluripotency-associated genes and meiotic initiation [56, 58]. Interestingly, STRA8 is expressed in both type A₁ differentiating spermatogonia and preleptotene spermatocytes entering meiosis [33, 54, 55], although it is not required for either process. Indeed, male germ cells in Stra8 KO mice completed spermatogonial differentiation and, on a mixed genetic background arrested in early pachynema, while on those on a congenic C57Bl/6 N background arrested in preleptonema following meiotic DNA replication [54, 55]. Thus, STRA8 is required for progression through meiosis, and serves as an effective protein marker (as demonstrated here) to identify RA-responsive germ cells.

Here, we observed ~25% of ZBTB16+ undifferentiated spermatogonia remained STRA8- in response to a pulse of exogenous RA in the adult. This number closely approximates the percentage of GFRA1+ undifferentiated spermatogonia in the adult testis [59, 60]. Since we did not observe any GFRA1+/STRA8+ undifferentiated spermatogonia post-RA, this reveals that GFRA1 can serve as a reliable marker of RA-insensitive undifferentiated spermatogonia. This is an important point—RA-insensitivity is a functional feature of SSCs, which must avoid responding to RA in order to remain undifferentiated. This reveals that approximately one-quarter of undifferentiated spermatogonia (GFRA1+ spermatogonia) display a key functional feature of SSCs. These results further suggest, as in the neonatal testis, there are distinctions in how adult spermatogonia respond to RA: SSCs and early undifferentiated progenitors cannot respond to RA (=RA-insensitive), mid-late undifferentiated progenitor spermatogonia are poised to respond (=RA-responsive), and differentiating spermatogonia can respond again (=RA-responded). Several studies have investigated subcategories of undifferentiated progenitors based on their expression of specific markers (e.g. “neurogenin 3” (Neurog3/Ngn3) and developmental pluripotency-associated 3 (Dppa3) [32, 61, 62]. In particular, Dppa3 was reported to reliably identify late-stage undifferentiated progenitors that preferentially differentiated in response to RA [61].

We also noted differentiating spermatogonia became STRA8+ after administration of exogenous RA during steady-state spermatogenesis, both in the adult testis and the synchronized testis model. Despite having the capacity to respond to RA, differentiating spermatogonia did not hasten their progression through differentiation and into meiosis, which suggests RA does not play an instructive role in meiotic initiation [5, 33, 63, 64]. To confirm these findings, adult mice were administered RA daily over the length of one seminiferous cycle, and there were no apparent differences compared with vehicle-treated mice in the progression of these already differentiating spermatogonia into meiosis. In light of these results, we propose the following: (1) spermatogonia must complete the 8.6 day differentiation program prior to entering meiosis, regardless of additional exogenous RA; (2) germ cells’ ability to respond to RA is acquired by undifferentiated progenitor spermatogonia, retained throughout differentiation, and then lost as they enter leptonema.

Our results support existence for both cell autonomous and non-cell autonomous mechanisms governing male germ cell fate during steady-state spermatogenesis. Considering very little circulating RA is used by germ cells [65], and the criticality of the RA response for spermatogenesis, it makes sense that multiple systems would be in place to ensure proper RA response and thus homeostasis within the testis. The addition of exogenous RA drove widespread response (indicated by STRA8 reactivation) of multiple germ cell types, while advancement of spermatogenesis only occurred in a single specific cell type—undifferentiated progenitor spermatogonia. Our results also revealed that adult testes employ some level of CYP26-mediated catabolism to decrease the RA levels. However, this RA catabolism was not the singular mechanism regulating RA response during steady-state spermatogenesis. Considering the degree of RA response in terms of the number of RA responding undifferentiated progenitor spermatogonia was reduced in talarozole- versus RA-treated mice, this further supports the concept that multiple mechanisms regulate the RA response during steady-state spermatogenesis. This markedly differed from results in the developing testis, where RA catabolism conferred a near-complete degree of protection of undifferentiated progenitor spermatogonia from RA [16]. This aligns with results from mice with Sertoli cell-specific deletion of Cyp26b1, which encodes the CYP26 enzyme active in the testis—those mice had severe disruptions in spermatogenesis in younger mice that improved as the mice aged [20, 66].

There were remarkable similarities in germ cell responsiveness to RA both in vivo and in vitro, where the testis architecture was disrupted. This supports the concept that germ cells themselves have intrinsic mechanisms regulating RA responsiveness. The molecular nature of this intrinsic response is currently unclear, although it seems reasonable to predict a primary determinant of RA response would be expression of a cognate RA receptor. Indeed, RARG was predominantly undetectable in GFRA1+ undifferentiated spermatogonia (which were RA-insensitive) but readily detectable in RA-responsive GFRA1− undifferentiated progenitors and differentiating spermatogonia. RARG expression was also evident throughout the testis in spermatogonia 12 h post-RA treatment. However, confounding this straightforward model is the observation that spermatogonia in Rarg KO differentiated and completed spermatogenesis to form sperm in nearly all tubules in young adult mice, and testicular defects were only observed as these mice aged [42]. While it is possible that RARA provides a compensatory mechanism in those mice, its levels are quite low in spermatogonia. Altogether, RARG may be sufficient, but not essential for bestowing spermatogonial RA responsiveness. Of note, RARG mRNA and protein are both abundantly expressed in human testes [67, 68], but to the best of our knowledge the RA response in human spermatogonia has not been examined.

A recent approach has been employed to treat human male subfertility and infertility patients with retinoids. This was based on observations that testicular concentrations of RA were significantly reduced in men with abnormal sperm production compared with their fertile counterparts [69], and infertile men had significantly lower levels of RA synthesis enzymes (e.g. RALDH2), which corresponded with reduced germ cell numbers in the testis [18]. Men diagnosed with azoospermia (no sperm in ejaculate) and oligoasthenozoospermia (low sperm count and poor sperm motility) were given long-term treatments with isotretinoin (13-cis-RA) and experienced significant improvements in fertility parameters [14, 15]. These treatments increased numbers of motile sperm in men with oligoasthenozoospermia, resulting in pregnancies and live offspring. With these clinical results in mind, our results revealed that RA administration (once or daily for 8 days) in mice can induce precocious spermatogonial differentiation, resulting in significantly more preleptotene spermatocytes 9 days later and pachytene spermatocytes 17 days later. These RA-treated mice, if aged for another few weeks, would likely have increased sperm counts over their vehicle-treated control counterparts, matching the results from these clinical studies. Considering the remarkable conservation in spermatogenesis between mouse and human spermatogenesis [70], the results presented here in mice provide an explanation for those from isotretinoin-based human clinical trials.

In designing long-term therapies, quality of health and the potential for therapy-induced toxicity are important considerations. Fortunately, the half-life of RA in the mammalian testis is <90 min [71, 72], therefore RA is either utilized by the cell or catabolized. In the human clinical trials, all men receiving long-term isotretinoin treatments experienced minimal side effects (e.g. dry skin, chapped lips) that were easily managed [14, 15]. Additionally, the isotretinoin concentrations administered to patients was several magnitudes lower than the dose range associated with teratogenicity, and the exposure of a pregnant female partner to isotretinoin from the treated partner’s semen was negligible [14]. At this time, we propose an alternative therapeutic approach, in which high amounts of daily RA might be given to patients for a limited time (a few weeks) instead of the full spermatogenic cycle (~2–3 months). This shortened treatment regimen would increase numbers of spermatogonia undergoing differentiation in one seminiferous cycle while minimizing potential side effects associated with long-term RA exposure. Considering SSCs are RA non-responsive, this temporary treatment regimen could be re-implemented as necessary without hindering spermatogenesis long-term or inducing systemic side effects.

In summation, the findings here reveal that adult testes employ a complex and multilayered control system to modulate RA levels that limit spermatogonial progression through spermatogenesis in response to high levels of RA. This ensures normal germ cell development and ultimately preserves long-term fertility in male mammals. The addition of exogenous retinoids temporarily overcomes this system, and presents an attractive strategy to increase chances of successful fertilization for men struggling with subfertility or infertility.

Authors’ contributions

Conceptualization was by TAJ, OK, and CBG; investigation was by TAJ, BAN, OK, EVH, and TM; the first draft of the manuscript was prepared by TAJ and edited by coauthors; funding acquisition, project administration, and final editing were the responsibility of CBG.

Data availability

All data from this study are incorporated into the article and its online supplementary material.

Conflict of interest

The authors have declared that no conflict of interest exists.

Funding

This project was supported by grants R21HD105963 and R01HD090083 (to CBG) from the Eunice Kennedy Shriver National Institute of Child Health and Human Development (NIH/NICHD) and by a fellowship from the Male Contraceptive Initiative (to OK).