Germ cell depletion in recipient testis has adverse effects on spermatogenesis in orthotopically transplanted testis pieces via retinoic acid insufficiency

Germ cell depletion in recipient testes is indispensable for successful transplantation of spermatogonial stem cells. However, we found that such treatment had an adverse effect on spermatogenesis of orthotopically transplanted donor testis tissues. In the donor tissue, the frequency of stimulated by retinoic acid (RA) 8 (STRA8) expression was reduced in germ cells, suggesting that RA signalling indispensable for spermatogenesis was attenuated in germ cell-depleted recipient testes. In this context, germ cell depletion diminished expression of testicular Aldh1a2, which is responsible for testicular RA synthesis, while Cyp26b1, which is responsible for testicular RA metabolism, was still expressed even after germ cell depletion, suggesting an alteration of the RA synthesis/metabolism ratio. These observations suggested that RA insufficiency was one of the causes of the defective donor spermatogenesis. Indeed, repetitive RA administrations significantly improved donor spermatogenesis to produce fertile offspring without any side effects. These findings may contribute to improving fertility preservation techniques for males, especially to prevent iatrogenic infertility induced by chemotherapy in prepubertal cancer patients.


Results
Germ cell depletion in recipient testes compromises spermatogenesis in donor testis pieces. Testis pieces collected from mouse pups expressing enhanced green fluorescent protein (EGFP) constitutively (green mice) were transplanted into the testicular interstitium of recipient testes with or without germ cell depletion induced by busulfan (Fig. 1). Although it has been demonstrated that germ cell-depleted testes are more suitable hosts for SSC transplantation than healthy testes with intact germ cells 27,28 , we found that the number of GFRA1 + undifferentiated spermatogonia and the frequency of seminiferous tubules carrying MCAM + spermatogonia were comparable between testis pieces transplanted into busulfan-treated testes and those in control recipients ( Fig. 2a-d, SI Fig. S1a-d). Moreover, we found that the frequency of seminiferous tubules containing PNA + haploid germ cells was reduced significantly in testis pieces transplanted into busulfan-treated recipients (Fig. 2e). Johnsen's score count further revealed that major spermatogenic defects in testis pieces of the busulfan group occurred around the stage of generating haploid germ cells (Fig. 2f,g) 32 . These data suggested that busulfan-mediated germ cell depletion in recipients affects the spermatogenesis of transplanted testis pieces.
Germ cell depletion in recipient testes compromises the retinoic acid supply in donor testis tissue. To elucidate the mechanism of the spermatogenic defects in the transplanted testis pieces, we focused on niche functions. The germline niche allows SSCs to self-renew and differentiate to drive spermatogenesis by supplying soluble factors. GDNF is indispensable for the survival and self-renewal of SSCs, whereas RA induces SSC differentiation toward mature sperm via meiosis [1][2][3] . Testosterone produced by Leydig cells is also indispensable for the progression of meiosis 33 . It has been demonstrated that busulfan-treated testis produce testosterone 34,35 , whereas RA is synthesised by the combination of testicular cells including germ cells 36 . Based on this context, we next focused on RA signalling in transplanted testis tissues and found that the frequency of tubules containing STRA8 + differentiating spermatogonia/pre-leptotene spermatocytes was reduced significantly (Fig. 3a,b, SI  Fig. S2). This observation suggests RA signalling defects in testis pieces transplanted into busulfan-treated testes.
Male germ cells and testicular somatic cells collaboratively synthesise all trans-RA by expressing enzymes that metabolise vitamin A/retinol 36 (Fig. 4a). Quantitative reverse transcriptase-polymerase chain reaction (qRT-PCR) revealed that Aldh1a2, encoding the major RA-synthesising enzyme gene in the testis, was downregulated in busulfan-treated testes (Fig. 4b). Consistent with this finding, we also found downregulation of Cyp26a1 and Cyp26c1 genes (Fig. 4b). Our previous study revealed that RA induces these genes 37,38 . Therefore, the current data strongly suggest that germ cell depletion reduces testicular RA production.
Retinoic acid administration ameliorates spermatogenic functions of testis pieces transplanted into busulfan-treated testes. Based on our observations, we hypothesised that RA administration ameliorates spermatogenic functions of testis pieces transplanted into busulfan-treated testes. To test our hypothesis, busulfan-treated recipients transplanted with testis pieces were administrated RA five times with an 8.5-day interval just prior to sacrifice (Fig. 1).
Taken together, these results support the idea that germ cell depletion in recipient testis results in RA deficiency, and demonstrate that RA administration augments reduced RA signalling in germ cell-depleted testes to recover spermatogenic activity. However, this treatment also exerts adverse effects on GFRA1 + spermatogonia and compromises the arrangement of germ cells in seminiferous tubules, suggesting an abnormality of haploid germ cells to produce the next generation of cells.
To determine whether healthy offspring could be produced from haploid germ cells generated in RA-administrated donor testes, we conducted in vitro microinsemination followed by transfer into oviducts of pseudopregnant females 39 . A transplanted testis piece was collected and immediately immersed in paraffin oil at 4 °C until germ cell preparation. Within 24 h, GFP + haploid germ cells, including round and elongating spermatids, were collected under fluorescence stereomicroscopy and immediately subjected to intracytoplasmic sperm injection. As a result, of the 119 generated embryos, 90 (75.6%) developed to the two-cell stage within 24 h of culture. After transfer into the oviducts of four pseudo-pregnant females, 27 live pups were born (13 males and 14 females, Fig. 6a). These male and female offspring did not show any overt abnormalities and were fertile to produce a healthy next generation (Fig. 6b). These results indicate that RA treatment of the recipient does not exert an adverse effect on the next generation.

Discussion
Busulfan-mediated germ cell depletion optimises testicular conditions for SSC transplantation 30,31 . Although this finding suggests that germ cell-depleted testes are also suitable for testis piece transplantation 27 , our present study revealed that germ cell depletion in recipient testes compromised spermatogenesis in orthotopically Testes were collected from neonatal mice constitutively expressing EGFP and transplanted into recipient testes. Untreated recipients were designated as the control group. Some recipients (busulfan group) were administrated with busulfan at a dose of 44 mg/kg b.w. that is sufficient to deplete germ cells completely. In addition, some germ cell-depleted recipients were also administrated retinoic acid at a dose of 750 µg/body five times with an 8.5-day interval (yellow arrowheads) just prior to sacrifice (busulfan + RA group). Transplanted testis pieces were collected after around 146 days (139-154 days) and subjected for immunofluorescence staining.
Scientific RepoRtS | (2020) 10:10796 | https://doi.org/10.1038/s41598-020-67595-1 www.nature.com/scientificreports/ transplanted donor testis pieces at least in part by RA insufficiency. Indeed, repetitive supply of RA improves spermatogenesis in testis pieces transplanted into germ cell-depleted testes. Although there were some negative effects on GFRA + spermatogonia, considering that RA-treated sperms produced healthy offspring that were also fertile, RA treatment was considered to be supportive for spermatogenesis. Generally, recipient conditions affect the prognosis of transplantation. Shinohara and colleagues demonstrated that germ cell depletion releases the germline niche from occupation by resident SSCs 30 . In fact, the degree of germ cell depletion in recipient testes correlates positively with the frequency of SSC colonisation and subsequent germline transmission 31 . Considering the abundant GDNF in germ cell-free tubules and hyperproliferation of transplanted GFRA1 + spermatogonia in Kit W /Kit W-v recipient tubules 40 , germ cell-depleted testes are likely to be suitable recipients for not only SSC transplantation, but also testis piece transplantation. Therefore, it was unexpected that germ cell depletion compromised the spermatogenesis of donor tissues.
Spermatogenesis is a complex process. Briefly, undifferentiated spermatogonia including SSCs produce differentiating spermatogonia that subsequently enter meiotic division after six mitotic divisions to produce haploid www.nature.com/scientificreports/ disrupts spermatogenesis [41][42][43] . In the present study, we focused on RA signalling, because germ cells contribute significantly to RA production by expressing Aldh1a2 that is responsible for production of testicular RA. Quantitative RT-PCR confirmed that germ cell depletion resulted in RA insufficiency. Simultaneously, we found that expression of Cyp26b1 was still maintained even after germ cell depletion. Of the three Cyp26 genes, only Cyp26b1 is indispensable for proper spermatogenesis 44 . Considering these observations, loss of Aldh1a2 and maintained Cyp26b1 expression in germ cell-depleted recipient testes should consume RA in orthotopically transplanted testis pieces. In fact, the frequency of tubules containing STRA8 + germ cells in donor testes was reduced significantly when recipient germ cells were depleted, and repetitive RA treatments rescued this effect to some extent. However, we also found that the present method often induced tubules with an unusual germ cell arrangement with a closed lumen. In this experiment, we employed repetitive RA administrations to compensate for the magnitude of RA signals. As described previously, periodical RA oscillation controls the seminiferous epithelial cycle with 8.6-day periodicity 2,3 . However, our protocol did not exactly follow the physiological RA oscillation. Although administrating the dose employed in the present study was sufficient to upregulate RA-inducible genes in testes with or without germ cells 37,38 , this dose might not be optimal to support proper spermatogenesis under the present experimental conditions. However, it should be considered that RA might cause adverse effects, especially on the undifferentiated spermatogonia population. Indeed, although Agrimson et al. reported that neonatal RA administration has no effect on the population of SSCs 45 , our RA administration protocol caused a reduction of GFRA1 + spermatogonia. These data suggest that testicular conditions and age affect the RA sensitivity of SSCs in the testis. Therefore, optimisation of the timing and dose of RA administration should improve spermatogenesis and SSC maintenance in transplanted testis pieces. The present study did not investigate testosterone. Previous studies have demonstrated that busulfan treatment does not alter testicular testosterone 34,35 , whereas another report demonstrated that RA signal inhibition in Leydig cells results in abnormal spermatogenesis and epididymal functions 46 . The former two reports employed a single 30 mg/kg busulfan administration. However, this procedure might restore spermatogenesis from remnant SSCs 47 . This might also permit RA synthesis in the testis to some extent and lead to testosterone production. However, the present study used a higher dose (44 mg/kg) of busulfan at which germ cell-mediated RA synthesis is considered to be diminished completely. Such treatment might diminish the production of  www.nature.com/scientificreports/ testosterone. Moreover, oestradiol, another sex steroid synthesized from testosterone by CYP19A/aromatase in Leydig cells and indispensable for proper spermatogenesis, might be affected by busulfan 13,14 . Considering these observations, RA insufficiency might also indirectly compromise spermatogenesis via modulated production of sex steroids in Leydig cells. A GnRH analog supports colonisation of SSCs transplanted into infertile rodent testes by suppressing gonadotropin release from the pituitary grand 48,49 . In addition, repetitive administrations of the GnRH analog support recovery from busulfan damage for spermatogenesis 50 . These findings suggest that gonadotropin exerts adverse effects on the recovery process of spermatogenesis. In this context, busulfan treatment distinctly elevates LH and FSH levels in the pituitary and serum of rats 51 . Therefore, gonadotropin suppression in recipient might improve spermatogenesis in orthotopically transplanted donor testis pieces.
Altered expression of several factors might have also contributed to the reduced spermatogenesis in transplanted testis pieces. Our previous study demonstrated that GDNF suppresses Rarg in undifferentiated spermatogonia both in vitro and in vivo, and germ cell depletion upregulates testicular Gdnf 37,38 . Although the number of GFRA1 + spermatogonia was not changed in our present study, Gdnf upregulated in recipient testes might suppress Rarg expression in GFRA1 + spermatogonia of donor testes to suppress production of differentiating spermatogonia. Germ cell-derived PDGFA is also important for Leydig cell functions and proper spermatogenesis 52 . Restructuring of the molecular network by controlling these factors is beneficial for proper spermatogenesis in transplanted testis pieces.
RA causes various developmental defects in vertebrates because it plays fundamental roles in gastrulation, axis formation, and organ formation during development because of cell fate disruption 53 . In addition, RA alters the epigenome status of several cell types 54 . Moreover, alteration of the epigenome status in sperm results in developmental abnormalities of offspring 55 . However, the present study demonstrated that ectopic RA treatment did not cause any developmental defects in offspring.
The microinsemination technique was primarily developed by Yanagimachi and colleagues in 1995 using mice 39 . This technique has already been employed to assist human reproduction, and no overt abnormalities have been reported to date 16,17 . A combined technique including testis piece transplantation, RA treatment, and microinsemination might be practical to restore spermatogenesis, especially in prepubertal cancer patients.
Although it has been reported that recipient conditions affect the spermatogenesis of transplanted testis pieces 56 , most studies have employed subcutaneous transplantation, and other transplantation sites remain to be elucidated. The present study employed intratesticular transplantation and found that depletion of recipient germ cells compromised the RA production/metabolism ratio and reduced spermatogenesis of donor testis pieces transplanted into recipient testis. RA treatment supported the spermatogenesis of donor testis pieces when the size of the donor testis piece was insufficient for the size of the recipient testis after treatment with the anti-cancer drug. These findings may contribute to the treatment of prepubertal cancer patients to recover fertility after anti-cancer therapy. However, there are still many parameters (e.g. transplantation site, supportive molecules, and factors affecting prognosis) to be considered when applying testis piece transplantation to assisted reproductive technology.

Methods
Ethics statement. The institutional animal care and use committee of Shinshu University approved all animal experimentation protocols (Approval Nos. 260013, 280120, and 019032). All experiments were performed in precise accordance with the manual provided by the institutional animal care and use committee of Shinshu University.
Testis piece transplantation and treatments. Male C57BL/6NCrSlc mice were purchased from Japan SLC (Shizuoka, Japan). Seven-week-old mice with/without treatment with 44 mg/kg body weight busulfan (Sigma-Aldrich, St Louis, MO, USA) at 3 weeks of age, which depletes germ cell in testes, were used as recipients. Male C57BL/6 Tg14(act-EGFP)OsbY01 'green' mice were used as testis donors 57 . Testis pieces collected from newborn mice (1.5-4.5 days postpartum) were transplanted into recipient testes. Briefly, recipient mice were anesthetised with isoflurane (FUJIFILM Wako Pure Chemical Co., Ltd., Osaka, Japan) and a small cut was made in the tunica albuginea of each animal using fine forceps. A single testis piece was inserted into the testicular parenchyma 25 . In some cases, recipient testes were treated five times with all-trans retinoic acid (750 µg per mouse, Sigma-Aldrich) dissolved in a 10% ethanol-sesame oil (Nacalai Tesque, Kyoto, Japan) with 8.5-day intervals just prior to sacrifice (Fig. 1).
Immunofluorescence staining. Recipient mice were sacrificed at around 146 days after transplantation, and donor recipient testes carrying donor testis pieces were collected. After overnight fixation in 4% paraformaldehyde/PBS (−) at 4 °C followed by overnight immersion in 30% sucrose/PBS (−), testes were embedded in Tissue-Tek OCT compound (Sakura Finetek, Tokyo, Japan) to prepare 8 µm-thick cryosections. After 10 min of treatment with 0.1% Triton-X 100 (Sigma-Aldrich) in PBS (−) at 4 °C followed by blocking with 10% donkey serum (Jackson ImmunoResearch Laboratories, Inc., West Grove, PA, USA), the sections were incubated with primary antibodies overnight at 4 °C. Sections were further incubation with secondary antibodies at room temperature for 1 h. Nuclei were counterstained with Hoechst 33342 (Sigma-Aldrich). Images were captured by confocal laser scanning microscopy (FV1000-D; Olympus, Tokyo, Japan) or All-in-one fluorescence microscopy (BZ-X710, KEYENCE Co., Ltd., Osaka, Japan). Collected images were joined to reconstruct an entire image of testicular cross-sections. Antibodies used in this study were: a rat anti-mouse CD146 monoclonal antibody (anti-mouse MCAM antibody, cat# 134702; clone ME-9F1; Biolegend Inc., San Diego, CA, USA); rhodamineconjugated peanut agglutinin (RL-1072; Vector Laboratories, Burlingame, CA, USA); goat anti-rat GFRA1 ). The standard curve method was applied to determine the absolute quantity. A standard cDNA stock for absolute quantitation was prepared from the PCR products using a FastGene™ Gel/PCR Extraction Kit (Nippon Genetics, Tokyo, Japan). For data evaluation, values of the control group were set to 1.0 after normalisation to Hprt expression levels. Primer sequences are listed in Table 1.

Microinsemination.
To assess the integrity of haploid germ cells in donor testes treated five times with retinoic acid injection, round and elongated spermatids were collected and subjected to intracellular sperm injection (ICSI) 39 . In brief, donor testes were recovered and transported overnight at 4 °C before ICSI using a previously published method for transportation of epididymides 58 . Round and elongated spermatids were collected from donor testis pieces by simple repeated pipetting of seminiferous tubules and injected into oocytes of B6D2F1 mice (Japan SLC) using a Piezo-driven micropipetter (PrimeTech, Tokyo, Japan) as described previously 59 . Resultant zygotes were cultured for 24 h and then transferred into the oviducts of pseudopregnant female mice (ICR strain, Japan SLC). Offspring were delivered by caesarean section.

Statistical analyses.
Results are presented as means ± SEM. Significance of differences between means for single comparisons was determined by the Student's t test. P < 0.05 was considered statistically significant.

Data availability
There are no datasets for this manuscript to share via public repository.

Gene
Forward primer Reverse primer Aldh1a1  CTC CTG GCG TGG TAA ACA TT  CCA TGG TGT GCA AAC TCA AC   Aldh1a2  CCA TGA CTT CCA GCG AGA T  TCT GAG TTC TGC CAT TCA TTG   Aldh1a3  GTG GAG TTC GCC AAG AAG AG AGA AGA CCG TGG GTT TGA TG   Cyp26a1  TGA CCC GCA ATC TCT TCT CT  GAG GAG CTC TGT TGA CGA TTG   Cyp26b1  TGG TCA CTG GTT GCT ACA GG  TGG GCA GGT AGC TCT  www.nature.com/scientificreports/