WIN18,446 enhances spermatogonial stem cell homing and fertility after germ cell transplantation by increasing blood-testis barrier permeability

Abstract

Spermatogonial stem cells (SSCs) possess a unique ability to recolonize the seminiferous tubules. Upon microinjection into the adluminal compartment of the seminiferous tubules, SSCs transmigrate through the blood-testis barrier (BTB) to the basal compartment of the tubule and reinitiate spermatogenesis. It was recently discovered that inhibiting retinoic acid signaling with WIN18,446 enhances SSC colonization by transiently suppressing spermatogonia differentiation, thereby promoting fertility restoration. In this study, we report that WIN18,446 increases SSC colonization by disrupting the BTB. WIN18,446 altered the expression patterns of tight junction proteins (TJPs) and disrupted the BTB in busulfan-treated mice. WIN18,446 upregulated the expression of FGF2, one of the self-renewal factors for SSCs. While WIN18,446 enhanced SSC colonization in busulfan-treated wild-type mice, it did not increase colonization levels in busulfan-treated Cldn11-deficient mice, which lack the BTB, indicating that the enhancement of SSC colonization in wild-type testes depended on the loss of the BTB. Serial transplantation analysis revealed impaired self-renewal caused by WIN18,446, indicating that WIN18,446-mediated inhibition of retinoic acid signaling impaired SSC self-renewal. Strikingly, WIN18,446 administration resulted in the death of 45% of busulfan-treated recipient mice. These findings suggest that TJP modulation is the primary mechanism behind enhanced SSC homing by WIN18,446 and raise concerns regarding the use of WIN18,446 for human SSC transplantation.

Spermatogonial stem cells (SSCs) are the only stem cells in the germline with self-renewal activity [1, 2]. The number of SSCs in the testis is very small; they comprise only 0.02 to 0.03% of all germ cells in the testis [2, 3]. SSCs are found on the basement membrane of the seminiferous tubules and continuously produce committed progenitor cells. Differentiating spermatogonia, interconnected by each other via cytoplasmic bridges, differentiate as clusters of germ cell clones. These cells gradually detach from the basement membrane and migrate through the blood-testis barrier (BTB) between Sertoli cells toward the adluminal lumen as they undergo meiotic division to produce haploid cells. The BTB is comprised of several tight junction proteins (TJPs), including OCLN, CLDN3, CLDN5, and CLDN11, and is considered to be essential for meiosis and protection against autoimmunity for haploid germ cells [4]. Although SSCs divide infrequently during steady spermatogenesis, detailed morphological analyses have shown that the probability of self-renewing division increases after testis damage by irradiation and chemicals, and restoration of spermatogenesis is enhanced [2, 5].

In 1994, a spermatogonial transplantation technique was developed [6]. In this technique, SSCs were microinjected into the seminiferous tubules of infertile mouse testes devoid of endogenous germ cells. Transplanted SSCs attached to the Sertoli cells and transmigrated through the BTB before settling on the basement membrane [7]. It is thought that SSCs migrate to a special environment called niches for self-renewal division and reinitiate spermatogenesis [8]. Within 2 weeks of transplantation, SSCs initially divide to form monolayer networks of spermatogonia on the basement membrane [7]. These germ cell colonies begin vertical differentiation at around 1 month, and haploid germ cells are found within 2 months after transplantation. In the most successful case, recipients can sire offspring from donor SSCs by natural mating [9]. This experiment provided evidence that SSCs have a unique potential to reconstitute spermatogenesis, and has opened up new possibilities for the use of SSCs for male fertility treatment.

One of the biggest problems for SSC studies is the low frequency of donor cell colonization. Although previous morphological studies suggested that 0.02–0.03% of total germ cells act as SSCs [2, 3], transplantation studies have reported significantly lower SSC numbers, ranging from 150 to 3,000 in a testis [10,11,12]. This discrepancy between morphological and functional estimates is considered to be due to the unphysiological transmigration of SSCs through the BTB. This possibility was initially suggested by spermatogonial transplantation into immature testes which lack the BTB before sexual maturation [13]. In these testes, donor cells produced ~ 5-fold more germ cell colonies, and the recipient mice sired progeny as early as 3 months after transplantation, which was significantly faster than transplantation into adults (5–8 months) [9, 14]. The involvement of the BTB was directly demonstrated by increased colonization of donor SSCs in Cldn11 knockout (KO) mice [15]. Cldn11 KO mice are unique because disruption of a single claudin gene impairs tight junction development. Donor cells produced ~3-fold more germ cell colonies in Cldn11 KO testes than wild-type testes. An alternative approach to improve SSC colonization is based on the suppression of the hypothalamic-gonadal axis [16, 17]. Suppression of this hormonal regulation increased the expression of WNT5A [18], which promotes SSC self-renewal [19]. Because fertility of recipients increases in proportion to the number of transplanted SSCs [20], not only the recipient microenvironment but also donor SSC number determines the colonization efficiency and the fertility of recipient mice.

Recently, a dramatic increase in donor cell colonization was reported by administration of WIN18,446 (WIN) [21]. WIN inhibits ALDH1A2, a testis-specific isoform of the enzyme involved in retinoic acid (RA) metabolism [22]. Because RA is essential for meiosis induction, WIN inhibits the development of differentiating spermatogonia and continuous administration results in male infertility [22, 23]. According to this report, however, transient WIN treatment at the time of transplantation not only increased SSC colonization but also restored fertility of recipient mice. This result was unexpected because WIN inhibits spermatogenesis. It also suggested that WIN exerted a long-term positive effect despite transient administration. From a practical point of view, the enhancement of fertility restoration was nevertheless striking, considering the number of SSCs transplanted. For the mechanism of enhanced colonization, the authors suggested that WIN-mediated suppression of germ cell differentiation enhanced donor cell colonization [21]. However, the authors did not consider the possible involvement of TJPs in SSC homing. Therefore, their results could have an alternative explanation. In the present study, we investigated this possibility using Cldn11 KO mice. Additionally, although WIN was proposed to be useful for fertility restoration in humans, we observed that WIN treatment resulted in the death of a significant proportion of recipient mice.

Materials and Methods

Animals and transplantation procedure

For spermatogonial transplantation into WIN-treated mice, 5–6-week-old C57BL/6 (B6) mice (Japan SLC, Shizuoka, Japan) or Cldn11 KO mice (gift from Dr. S. Tsukita, Teikyo University) were first treated with busulfan, as described previously [15, 24]. One month after busulfan treatment, the animals were injected subcutaneously with WIN in dimethyl sulfoxide (DMSO; 2 mg; Cayman Chemical, Ann Arbor, MI, USA). Where indicated, we also administered WIN into wild-type untreated mice with the same amount of WIN (2 mg) to determine WIN toxicity. Donor cells were prepared from C57BL6/Tg14 (act-EGFP-OsbY01) (designated green) mice using a two-step enzymatic procedure with collagenase type IV and trypsin (Sigma, St. Louis, MO, USA). Cells were microinjected into the seminiferous tubules via the efferent duct [24]. For colony counting, approximately 2 × 10⁵ cells were transplanted into the seminiferous tubules. For serial transplantation, recipient testes were collected 2 months after transplantation. After counting the colonies, the testes were dissociated into single cells. Cells collected from a primary recipient testis were transplanted into three secondary recipient testes. For production of offspring, approximately 5 × 10⁵ cells were transplanted. Recipient mice were caged with three wild-type female mice for offspring production. Each injection filled approximately 75–85% of the seminiferous tubules. The Institutional Animal Care and Use Committee of Kyoto University approved all animal experimentation protocols.

Analysis of recipient testes

Recipient mice were sacrificed 2 months after transplantation, and testes were exposed to UV light to count colony numbers. Where indicated, testis samples were collected, and fixed in 10% neutral-buffered formalin, and embedded in paraffin for hematoxylin and eosin staining.

Tracer experiment

Sulfo-NHS-LC-biotin solution (7.5 mg/ml; Thermo Fisher Scientific, Waltham, MA, USA) was prepared in phosphate-buffered saline (PBS) containing 1 mM CaCl₂, as described previously [25]. Approximately 10 μl of freshly prepared solution was microinjected into the interstitium of adult (> 8-week-old) testes using 30 G disposable needles. After 30 min, the testes were collected and fixed in 4% formalin.

Immunostaining

Testis samples were fixed in 4% paraformaldehyde for 2 h at 4°C, embedded in Tissue-Tek OCT compound (Sakura Finetek, Tokyo, Japan) for cryosectioning. Sections of 8 μm thickness were made. To block non-specific antibodies, sections were treated with 3% bovine serum albumin (BSA) and 10% goat serum in PBS supplemented with 0.1% Tween 20 (PBST) for 1 h at room temperature. The sections were then incubated with primary and secondary antibodies with 0.5% BSA in PBST, overnight and for 1 h, respectively. Sections were washed with PBST. Rhodamine-labeled peanut agglutinin (PNA)(Vector Laboratories, Burlingame, CA, USA) was used to detect the acrosome. Antibodies used are listed in Supplementary Table 1. Hoechst33342 (Sigma) was used for counterstaining.

Real-time polymerase chain reaction (PCR) analysis

Total RNA was isolated using TRIzol reagent (Invitrogen, Carlsbad, CA, USA). First-strand cDNA was produced using a Verso cDNA synthesis kit for reverse transcription-PCR (RT-PCR) (Thermo Fisher Scientific). For real-time PCR, the StepOnePlus^TM real-time PCR system (Applied Biosystems, Warrington, UK) and Power SYBR Green PCR Master Mix (Applied Biosystems) were used according to the manufacturers’ protocols. Transcript levels were normalized to those of Hprt. The PCR conditions were as follows: 95°C for 10 min, followed by 40 cycles at 95°C for 15 sec and 60°C for 1 min. Each PCR was performed at least in triplicate. Primers used are listed in Supplementary Table 2.

Western blotting

Samples were separated by SDS-PAGE, transferred to Hybond-P membranes (Amersham Biosciences, Buckinghamshire, UK), and incubated with primary antibodies. The antibodies used in the experiments are shown in the Supplementary Table 1.

Statistical analyses

Significant differences between means for single comparisons were determined by Student’s t-test. Multiple comparisons were carried out using ANOVA followed by Tukey’s honestly significant difference test.

Results

Analysis of WIN-treated mice after busulfan injection

We first examined the effects of WIN in busulfan-treated testes. To determine the impact of WIN on testicular somatic cells, we analyzed the role of WIN in empty seminiferous tubules before SSC transplantation. Wild-type mice were administered busulfan to deplete SSCs. One month after busulfan treatment, germ cells were depleted in almost all seminiferous tubules. At this point, WIN was administered for 13 days, which was previously used to enhance fertility recovery experiments [21]. When the testes were analyzed at this point, the size of the testis was comparable between WIN-treated and control testes (Fig. 1A). However, histological sections revealed smaller seminiferous tubules in WIN-treated testes (Figs. 1B–D), which suggested that WIN treatment influenced the testis microenvironment.

Fig. 1.

Enhanced self-renewal factor expression after WIN treatment. (A) Testis weight (n = 8). (B) Histological appearance of WIN-treated testes. (C, D) Area (C) and perimeter (D) of the WIN-treated seminiferous tubules (n = 118 for WIN; n = 108 for control). (E) Real-time PCR analysis of cytokines (n = 4 for WIN; n = 3 for control). (F) Western blot analysis of WIN-treated testis (n = 3). (G) Immunostaining of WIN-treated testes. Bar = 100 μm (B), 30 μm (G). Stain: hematoxylin and eosin (B), Hoechst33342 (G).

We carried out real-time PCR to examine the impact of WIN on cytokines involved in SSC self-renewal and homing. Gdnf and Fgf2 are established self-renewal factors for SSCs [8]. Gdnf and Cxcl12 can attract SSCs to germline niches [26,27,28,29]. WIN treatment significantly decreased the expression of Fgf2 and Cxcl12 (Fig. 1E). Although no statistical difference was observed, Gdnf expression was also downregulated. To confirm the downregulation of self-renewal factors, we next examined protein levels by Western blotting (Fig. 1F). Quantification of stained bands showed a significant increase in FGF2 levels in WIN-treated testes. Although GDNF expression was also increased, the difference was not statistically significant. CXCL12 expression was comparable between WIN-treated and control testes.

To examine the distribution of cytokines, we performed immunostaining of WIN-treated testes (Fig. 1G). In control testes, FGF2 signals were found in the peripheral areas of the seminiferous tubules, including interstitial tissues. WIN administration did not influence the basic staining pattern, but FGF2 exhibited more intense staining. GDNF and CXCL12 signals did not show apparent changes. These results suggest that the WIN treatment increases the concentration of FGF2.

Functional analysis of SSCs in WIN-treated testes by serial transplantation

A previous study showed an increased number of GFRA1⁺ spermatogonia in response to WIN treatment, and suggested that suppression of differentiation served to increase SSC numbers [21]. However, this was based on the immunohistochemical analysis. Because stem cells are defined by functional assay only, we carried out spermatogonial transplantation experiments to directly examine the impact of WIN on SSC self-renewal directly. After administering WIN into busulfan-treated mice for 2 days, donor testis cells from green mice were microinjected into the seminiferous tubules (Fig. 2A). The same number of cells was also injected into control DMSO-treated mice. The animals received further WIN treatment for 11 days, including on the day of cell transplantation. Two months after transplantation, the recipient mice were killed and donor cell colonization levels were analyzed by colony counting under UV light (Fig. 2B). The number of colonies generated in WIN-treated and control mice were 16.4 and 6.5 per 10⁵ cells, respectively (Fig. 2C). The difference was statistically significant.

Fig. 2.

Serial transplantation of SSCs in WIN-treated testes. (A) Experimental scheme. (B) Macroscopic appearance of recipient testes. (C) Colony count in the primary recipients (n = 8). (D) Total increase in colony numbers (total regenerated colony number × 10)/(primary colony number used for serial transplantation)(n = 8). (E) Immunostaining of the secondary recipients. Bar = 1 mm (B), 30 μm (E), Stain: Hoechst33342 (E).

To examine the degree of self-renewal in the primary recipient testes, we next collected testis cells from primary recipients. Dissociated testis cells were then microinjected into the seminiferous tubules of additional busulfan-treated mice (secondary recipient). The secondary recipients were euthanized 2 months after transplantation and the total number of colonies generated from primary colonies was calculated. Assuming that 10% of transplanted SSCs can colonize [7] and that each colony is produced by one stem cell [30], significantly more secondary colonies were generated by control testis cells. The numbers of secondary colonies generated from a single primary colony in WIN-treated and control testes were 11.8 and 28.5, respectively (Fig. 2D). This difference was statistically significant. Immunostaining of the recipient testes revealed SYCP3 and PNA expression (Fig. 2E), indicating that the transplanted SSCs underwent normal spermatogonial differentiation and maturation into haploid cells. These results suggest that WIN induces the loss of SSCs or limits self-renewal division after transplantation in busulfan-treated testes.

Impaired barrier function of the BTB

Transplantation experiments confirmed increased colonization in the primary recipients by WIN treatment. Although CXCL12 expression was not significantly increased, GDNF has been shown to be involved in SSC homing [27, 28]. Therefore, it is possible that induction of SSC migration by GDNF promotes SSC homing. Because previous studies suggested a critical role of the BTB in SSC homing [13, 15], we also examined the expression of the TJPs involved in BTB formation. The BTB is composed of several transmembrane molecules, including OCLN, CLDN3, CLDN5, and CLDN11 [4]. Real-time PCR did not show significant changes in Ocln, Cldn3, Cldn5, or Cldn11 by WIN treatment (Fig. 3A). However, Western blot analysis showed increased expression of all claudin proteins (Fig. 3B). OCLN did not change significantly.

Fig. 3.

Impaired BTB following WIN treatment. (A, B) Real-time PCR (A; n = 4 for WIN; n = 3 for control) and Western blot (B; n = 3) analyses of TJPs. (C) Immunostaining of WIN-treated testes. The arrow indicates a lack of signals in WIN-treated testes. (D) Functional assessment of the BTB. WIN-treated testes underwent interstitial injection of biotin. Thirty minutes after microinjection the testes were fixed and sectioned. Biotin was detected by fluorescein isothiocyanate-conjugated streptavidin (red). At least 121 tubules were counted. Bar = 20 μm (C), 100 μm (D). Stain: Hoechst33342 (C).

Because these results suggested abnormal regulation of BTB by WIN, we examined the expression of TJPs by immunostaining. The staining pattern of CLDN3 did not change by WIN treatment. However, CLDN5 and CLDN11 expression differed compared with wild-type mice (Fig. 3C). In the control testes, signals of these proteins were found in all Sertoli cells on the basement membrane. However, not all Sertoli cells showed expression of CLDN5 and CLDN11 after WIN treatment. Although OCLN expression levels did not change by Western blotting, we occasionally noted a lack of OCLN staining, suggesting its regulation by WIN treatment.

These abnormal TJP expression patterns raised the possibility that WIN treatment compromises the barrier function of the BTB. To test this directly, we analyzed the permeability of the BTB by injecting biotin into the interstitial tissue of busulfan-treated mice (Fig. 3D) Thirty minutes after biotin microinjection, we collected the injected testes and the biotin distribution was examined. Although control mice did not exhibit biotin in the seminiferous tubules, more than 80% of the seminiferous tubules contained biotin in WIN-treated mice. These results suggest that WIN impairs the BTB function.

Transplantation into Cldn11 KO mice

The increased permeability of the BTB induced by WIN treatment suggested that WIN-mediated enhancement of fertility restoration occurred as a result of increased SSC homing. If WIN treatment enhances SSC self-renewal by suppressing differentiation, the number of colonies would be still larger in mice with impaired BTB function. However, if WIN treatment enhances SSC colonization by increasing the permeability of the BTB, the number of colonies would be comparable between WIN-treated mice and control mice lacking the BTB. To distinguish between these possibilities, we used Cldn11 KO mice [31], which lack the BTB and are congenitally infertile due to defective differentiation beyond the spermatocyte stage (Fig. 4A). Based on a protocol used in a previous study [21], we removed endogenous SSCs by busulfan treatment and administered WIN for 2 days before spermatogonial transplantation. WIN treatment was continued for 11 days after transplantation. We used DMSO-treated Cldn11 KO mice to assess the impact of WIN on Cldn11 KO mice.

Fig. 4.

Spermatogonial transplantation in Cldn11 KO testes. (A) Histological appearance of Cldn11 KO testis. (B) Macroscopic appearance of Cldn11 KO testes after spermatogonial transplantation. (C) Colony count (n = 8). (D) Immunostaining of recipient testis. Bar = 100 μm (A), 1 mm (B), 20 μm (D). Stain: hematoxylin and eosin (A), Hoechst33342 (D).

Analysis of recipients showed that increased donor cell colonization does not occur in Cldn11 KO mice (Figs. 4B and C). To examine whether WIN influences subsequent differentiation, we carried out immunostaining of the recipient testes. Consistent with our previous study [15], immunostaining of recipient mice showed generation of haploid cells in Cldn11 KO mice despite the disruption of the BTB (Fig. 4D). Taken together, these results suggest that increased permeability of the BTB enhanced donor SSC colonization by WIN.

Fertility of WIN-treated mice after spermatogonial transplantation

To confirm the impact of WIN on fertility in the previous study, we collected donor testis cells from mature green mice, and single cell suspensions of donor cells were microinjected into the seminiferous tubules of WIN-treated mice. Approximately 1 month after transplantation, the recipient mice were housed with three wild-type females to monitor the fertility restoration. The fertility of 10 WIN-treated and 9 control DMSO-treated mice was examined. As early as 174 days after transplantation, the first offspring were born from WIN-treated mice (Fig. 5A). During the 6-month experimental period, 4 of 10 (40%) WIN-treated mice sired offspring, while no offspring were born from the 9 control mice. As many as three litters were born from a WIN-treated recipient male. Of the total of eight litters, five from three recipient mice showed EGFP fluorescence. These results are consistent with a previous study that showed enhanced fertility of WIN-treated recipient mice [21].

Fig. 5.

Restoration of fertility by WIN treatment. (A) Offspring from a WIN-treated recipient mouse. (B) Macroscopic appearance of WIN-treated testis after transplantation. (C) Testis weight (n = 20 for WIN; n = 18 for control). (D) Histological section of testes. (E) Number of tubules showing spermatogenesis. At least 968 tubules were counted. (F) Histological section of epididymis. Bar = 1 mm (B), 200 μm (D, F). Stain: hematoxylin and eosin (D, F).

Six months after transplantation, we sacrificed all recipient mice and recovered their testes and epididymides. As expected from the increased fertility, WIN-treated testes were significantly larger than control testes (Figs. 5B and C). To evaluate the degree of colonization, we prepared histological sections of the testes and counted the number of seminiferous tubules exhibiting spermatogenesis (Fig. 5D). While 24.1% of the seminiferous tubules showed spermatogenesis in control mice, 58.5% of the tubules exhibited spermatogenesis in WIN-treated testes (Fig. 5E). The difference between the two types of recipients was significant. We also analyzed the epididymis to examine potential fertility (Fig. 5F). Among the control DMSO-treated mice, only 3/18 (16.7%) of the epididymides contained sperm. However, 16/20 (80.0%) of WIN-treated mice contained sperm, including 6 mice that remained infertile. These results suggest that WIN treatment is an effective method for restoration of fertility.

Increased mortality of mice after WIN treatment

In preparing the recipient mice for WIN-treated animals, we noticed that WIN treatment is toxic to busulfan-treated mice. In total, we administered WIN to 20 busulfan-treated mice. Six mice died before spermatogonial transplantation. After transplantation, three more mice died within 10 days during WIN administration (Fig. 6A). Therefore, 9 of 20 (45.0%) busulfan-treated mice died by WIN administration. As for control mice, none of the 14 mice died during the same period. Only one control mouse died 2 weeks after transplantation. These results demonstrated that WIN treatment is toxic to busulfan-treated mice.

Fig. 6.

Mortality of WIN-treated mice. (A) Busulfan-treated mice (n = 20 for WIN; n = 14 for control). (B) Wild-type mice (n = 12).

Because no such toxicity was reported in previous studies, we confirmed these previous observations using wild-type mice without busulfan. We administered WIN to 3-week and 12-week-old wild-type mice. For each age, 12 mice were used to evaluate the toxicity. Although two 3-week-old mice died within 2 days after WIN treatment, none of the remaining animals died despite injection of WIN for the same periods of time (Fig. 6B). These results suggest that busulfan increases the toxicity of WIN, resulting in the death of busulfan-treated recipient mice.

Discussion

Low fertility of recipient mice has been an issue in spermatogonial transplantation since the development of this technique [9]. Because the migration of transplanted SSCs to their niches is a non-physiological process, it is not surprising that only 5–10% of the transplanted SSCs can make germ cell colonies [7, 32]. The fertility of recipients is influenced by the number of transplanted SSCs, the age of the recipients, and the activity of the hypothalamic-gonadal axis [16, 17, 20]. In a significant proportion of recipient mice, fertility can be restored by optimizing these three factors. Superiority of immature recipients is considered to be due to lack of the BTB before sexual maturation. Inhibition of the hypothalamus-gonadal axis also disrupts the BTB and upregulates WNT5A, which stimulates SSC self-renewal [17,18,19]. By optimizing these parameters, it is now possible to obtain offspring with good efficiency. In our previous experiments, for example, three of four recipient mice were able to sire donor-derived offspring by transplantation of cultured SSCs and administration of acyline, which is a GnRH analogue [17]. The offspring were born as early as 88 days after transplantation. Therefore, fertility restoration can now be achieved relatively efficiently in mice and accumulating evidence suggests that the BTB modulation is critical for fertility restoration.

However, this strategy in mice may not be easily extrapolated to other species. There are at least two issues for further improvements. The first issue is the availability of donor cells. At present, enriched populations of SSCs are most readily obtained by culturing SSCs, by adding GDNF and FGF2 to testis cells [33]. These cells are enriched for SSCs by ~100-fold compared with wild-type testis cells. However, SSC cultures are available only in rodents; there is currently no effective method to obtain a sufficient number of SSCs in non-rodent animals for fertility restoration [26]. The second issue is the preparation of host animals. For successful spermatogonial transplantation into immature animals, it is necessary to remove endogenous germ cells before BTB formation [34]. Although the use of genetic mutant mice is possible, such mutant recipients are not available in most other animal species. Removal of endogenous germ cells by cytotoxic chemicals also disturbs the endocrine milieu in recipient testes and inhibits donor cell differentiation [35]. Therefore, it will be necessary to develop an effective strategy to improve fertility of recipient animals for non-rodent transplantation experiments.

In this context, enhancement of colonization by WIN in a previous study was encouraging. It was also proposed that the enhanced colonization was caused by the suppression of spermatogonial differentiation [21]. The authors observed considerable levels of cell conversion from a differentiation-primed GFRA1^– state into a GFRA1⁺ state, which was referred to as “reversion”. Because this event was frequently observed after transplantation, the authors suggested that selective settlement of GFRA1⁺ cells was unlikely. However, several studies have suggested that GFRA1 is not always expressed in SSCs [36,37,38,39]. In fact, some SSCs were independent of GDNF and could undergo self-renewal division without GDNF [39]. Moreover, GFRA1 expression depends on the cell cycle. For example, pulse-chase experiments using BrdU in Purkinje cells showed that GFRA1 is upregulated in the late G1 phase [40]. Undifferentiated spermatogonia have a long cell cycle. A_single, A_paired and A_aligned spermatogonia divide only once every 10, 12, and 13 days, respectively [8]. However, SSCs double by every 5.6 days immediately after transplantation [41]. Therefore, the effect of cell cycle-dependent expression may not be negligible.

We observed striking differences in mRNA and protein levels in our analysis of WIN-treated testes. Specifically, while we detected reduced levels of Fgf2 mRNA using real-time PCR, Western blot analysis and immunostaining revealed increased FGF2 protein levels. This discrepancy in mRNA and protein levels was recently noted in our analysis of busulfan-treated testes [34]. However, it is not restricted in testis somatic cells, as similar observations have been reported in several cell types [42]. It is likely that FGF2 is regulated not only by transcription rates but also by additional control mechanisms, such as transcript stability, translational regulation, and protein degradation. Moreover, RA has also been shown to either increase or reduce translation without altering mRNA levels, depending on the cell type [43, 44]. Therefore, FGF2 may be regulated in a similar manner by RA. Further studies are needed to study this mechanism, which will be instrumental for understanding the impact of RA on spermatogenesis.

To directly confirm whether such “reversion” increases the number of SSCs, we performed serial transplantation experiments. Consistent with a previous study, the number of colonies in primary recipients was greater in WIN-treated animals. However, serial transplantation revealed a smaller number of secondary colonies in WIN-treated mice. This occurred despite the increased expression of FGF2. Given the established roles of FGF2 in promoting SSC self-renewal, these results appear counterintuitive. However, increased FGF2 expression alone may not necessarily allow self-renewal stimulation. Because the number of SSCs in vitamin A-deficient mice was reduced to 12.5% of that in untreated adult testes [45], these results suggest that WIN administration prevented expansion of SSCs despite increased colonization and FGF2 expression. In addition, the impact of WIN is not reversible despite transient administration [23]. Therefore, although WIN was administered only transiently and increased SSC colony number, such damages in testis somatic cells by WIN might have impaired self-renewal. Thus, the increased number of GFRA1⁺ cells after transplantation alone does not necessarily reflect functional SSC numbers.

Another possible mechanism underlying the increased colony numbers is enhanced SSC homing. Several previous studies have suggested that inhibition of RA signaling enhances the permeability of the BTB [46,47,48,49]. Genes associated with the regulation of BTB formation and maintenance have been shown to be misregulated in RA-deficient models [46]. In particular, RA signaling is required for appropriate expression of OCLN in stage IX-XII of the seminiferous tubules [46]. Another study demonstrated diffuse staining of OCLN and CLDN11 in VA-deficient mice [47]. Indeed, when BTB integrity was examined by a biotin permeability assay, the number of tubules containing biotin was significantly increased, which we also confirmed in the present study. Therefore, WIN impairs the BTB and increases the permeability. Although this possibility was not investigated in the previous study, it was equally possible that a larger number of SSCs penetrated the BTB and formed more colonies following WIN treatment.

Our transplantation experiments using Cldn11 KO mice showed that the beneficial effects of WIN on SSC homing depend on the BTB. We previously showed an increased colonization of SSCs in Cldn11 KO mice [15]. In that study, loss of the BTB increased the SSC homing efficiency by > ~3-fold. This experiment confirmed our earlier observation in pup recipient experiments that the BTB is a major impediment to SSC transplantation [13]. Because the BTB develops by 12–14 days after birth, SSC transplantation before this time point increases donor cell colonization dramatically. In the current study, we used Cldn11 KO testes and found that enhanced colonization does not occur despite WIN treatment. If reversion occurred in these mice, an increased number of colonies should have been observed even in Cldn11 KO testes. Therefore, our results suggest that the enhancement of donor cell colonization in wild-type mice seen after WIN treatment was caused by the enhanced permeability of the BTB, rather than suppression of differentiation.

The most striking observation in this study was the toxicity of WIN treatment. This result was unexpected; no previous studies have suggested this possibility. Because untreated mice did not show high mortality, WIN toxicity became evident only after busulfan treatment. Busulfan is extensively metabolized in the liver, with approximately 2% of unchanged drug excreted in the urine. Initial metabolism occurs primarily via conjugation with endogenous glutathione [50]. Although we do not yet know why animals treated with busulfan were more sensitive to WIN treatment, busulfan treatment may decrease total glutathione in the liver and make the animals more sensitive to oxidative stress [51]. Consistent with this, WIN decreases vitamin A in the liver [52]. ALDH2 deficiency is known to increase oxidative stress and contribute to vascular dysfunction and structural remodeling in hypertension [53]. Therefore, we speculate that synergy between busulfan and WIN increases oxidative stress, resulting in high toxicity.

Our study confirmed that WIN is beneficial for fertility restoration after spermatogonial transplantation in mice. Although it has been suggested that transient suppression of donor cell differentiation is responsible for enhanced colonization, the present study suggests enhanced SSC homing occurred due to increased permeability of the BTB. Moreover, the toxicity of WIN seen during recipient preparation raises concerns regarding the use of this chemical in clinical spermatogonial transplantation studies. More extensive animal experiments are required to confirm its utility for practical utility.

Conflict of interests

The authors declare no conflict of interest.