Vitrified canine testicular cells allow the formation of spermatogonial stem cells and seminiferous tubules following their xenotransplantation into nude mice

Belgian Malinois (BM), one of the excellent military dog breeds in South Korea, is usually castrated before sexual maturation. Therefore, the transfer of their genetic features to the next generation is difficult. To overcome this, testicular cells from 4-month-old BMs were frozen. Testicular cells were thawed after 3 months and cultured in StemPro-34 medium. Spermatogonial stem cell (SSC) characteristics were determined by the transplantation of the cultured germ cell-derived colonies (GDCs) into empty testes, containing only several endogenous SSCs and Sertoli cells, of immunodeficient mice, 4 weeks after busulfan treatment. Following the implantation, the transplanted cells localized in the basement membrane of the seminiferous tubules, and ultimately colonized the recipient testes. Xenotransplantation of GDCs together with testicular somatic cells conjugated with extracellular matrix (ECM), led to the formation of de novo seminiferous tubules. These seminiferous tubules were mostly composed of Sertoli cells. Some germ cells were localized in the basement membrane of seminiferous tubules. This study revealed that BM-derived SSCs, obtained from the castrated testes, might be a valuable tool for the transfer of BM genetic features to the next generation.

In vivo transplantation of male germ cells has provided the evidence of SSC existence. These cells are recognized by their functional ability to reform spermatogenesis following transplantation and colonization in recipient rodent testes 2,11,14,15 . Xenografts of immature (neonatal or prepubescent) testicular cells are able complete spermatogenesis in the dorsal skin of immunodeficient mice 16 .Testis tissues that retain their normal functions, including normal spermatogenesis and formation of seminiferous tubules, have been observed in the xenografts of the isolated testicular cells, and it was shown that they can produce fertile sperm 17,18 .
Previously, we successfully established spermatogonial GDCs from 2-month-old beagle testes, which contain an abundance of undifferentiated testicular germ cells, and FGF was determined to be an important factor for the proliferation and colony formation of GDCs 19 .
However, a suitable method for the long-term preservation of castrated canine male germ cells has not been established thus far. The objective of this study was to identify the optimal conditions enabling the freezing of canine testicular cells for GDC culture, and to determine the SSC capacity of these GDCs. Here, we report the cryopreservation conditions for canine spermatogonial germ cells, and demonstrate their ability to form GDCs after thawing. Additionally, the GDCs established following the cryopreservation show SSC capacity and de novo testicular tissue formation in immunodeficient mice.

Results
Culturing and characterization of GDCs originating from BM germ cells. Histological analysis of the donated BM testes was performed, and testicular germ and Sertoli cells were observed in the seminiferous tubules of testes originating from both 4-and 5-month-old BMs, (Fig. 1a,c, respectively). The size of seminiferous tubule in 4-month-old BM testis was smaller than in 5-month-old BM testis. PGP9.5 protein-positive spermatogonial germ cells were detected in both 4-and 5-month-old BM testes, and aligned germ cells were located in the basement membrane of the seminiferous tubules (Fig. 1b,d). Gonado-somatic index of 5-month-old BM was significantly higher than that of 4-month-old BM was (Fig. 1e).
GDCs were observed in the primary testicular cell culture originating from 4-and 5-month-old BM testes in StemPro-34 medium, although the GDCs in culture from 5-month-old BM testes were smaller than those from the 4-month-old-derived culture (Fig. 2a,b). GDCs were also observed in the culture of testicular cells from 4-and 5-month-old BM testes grown in DMEM-5% FBS media; however; they exhibited morphologic abnormalities in comparison with the GDCs cultured in StemPro-34 media (Fig. 2c,d). The colonies derived from all BMs were positive for alkaline phosphatase (AP) staining, but the staining intensity of GDCs in 4-month-old BM testes was stronger compared with the 5-month-old (Fig. 2e,f). Based on these results, 4-month-old BM testes and StemPro-34 medium were chosen for further experiments.
Using in-house-developed beagle testicular cell cryopreservation method ( Supplementary Figs 1 and 2), testicular cells originating from 4-month-old BM testes were cryopreserved after a single suspension in StemPro-34 medium without cryoprotective agent (CPA). Three months after the cryopreservation, BM testicular cells were thawed and cultured in StemPro-34 medium. The recovery rate of the cryopreserved BM testicular cells from 4-month-old BM testes was 75.11 ± 3.05%. Cell adhesion to the bottom of the wells was noticeable already on the first day, and again observed on the third day ( Fig. 3a,b). GDCs were first observed on the fifth day, and they continued to grow in size, which is shown in the images taken on the seventh day (Fig. 3c,d). Karyotype analyses of GDCs cultured from the cryopreserved testicular cells in StemPro-34 exhibited 78, XY normal chromosomes (Fig. 3e,f).
The GDCs were collected on the seventh day, and real-time RT-PCR was performed, in order to identify the expression levels of stem cell and undifferentiated cell marker genes. As a result, PGP9.5, GFRα-1, and PLZF expression levels in GDCs were significantly (P < 0.05) higher than that in the 4-month-old BM testes were ( Fig. 4a-c). No significant differences were observed in Oct4, Nanog, and GATA4 expression levels between GDCs and 4-month-old BM testes ( Fig. 4d-f).

Xenotransplantation of GDCs into immunodeficient mice. Germ cell-depleted seminiferous tubules
were observed in busulfan-treated mice 1 month after the treatment (Fig. 5a), and red fluorescent-labelled GDC cells were transplanted into them. No morphological differences were observed between non-GDC and GDC-injected testes (Fig. 5b). The examination of the seminiferous tubules of the injected mice by fluorescence microscopy showed that red-labelled cells were present in the seminiferous tubules ( Fig. 5c-e). Whole mount immunohistochemical staining of red-labelled-GDC-injected mice detected PGP9.5-positive cells (male BM germ cells) in the seminiferous tubules, localized in their basement membranes ( Fig. 5f-h). The seminiferous tubules of non-GDC-injected immunodeficient mice were stained with anti-PGP9.5 antibody, and PGP9.5-positive cells were not observed ( Fig. 5i-k). PGP9.5-stained cells were not observed in the negative control samples either ( Fig. 5l-n).
In order to determine the tissue regeneration capability of GDCs, the cryopreserved GDC cells were thawed and mixed with non-germ cells used as feeder layer for GDC culture, and matrigel, and afterward xenografted into the dorsal skin of immunodeficient mice. We observed tissue growth in these mice (Fig. 6a), and 20 weeks after the xenotransplantation, when the mice were sacrificed, the proliferated tissue was observed, with skin blood vessels connected with these newly grown tissues (Fig. 6b). The newly formed tissues were microscopically observed, and we noticed that blood vessels surrounded the tissues, and the putative seminiferous tubules could be seen clearly in the tissue sections (Fig. 6c). Additionally, no differentiating germ cells were observed in the tubules; only Sertoli-like cells were observed (Fig. 6d). To detect the germ cells in the seminiferous tubules, tissue sections were stained with anti-PGP9.5 antibody. Small populations of the PGP9.5-positive cells were observed, and they were shown to be located in the basement membrane of tubules ( Fig. 6e-g).

Discussion
Cryopreservation of mammalian SSCs using CPAs has been studied previously, and it was shown that cryopreservation with DMSO, 200 mM trehalose, and 2.5% PEG represents an efficient method in mice 12,13,20 . Additionally, undifferentiated male germ cells from pigs and bovine SSCs can be efficiently cryopreserved in the presence of 200 mM trehalose 21,22 . In our study, the highest recovery rate and colony number of the testicular cell cultures were shown for cells grown in non-CPA StemPro-34 medium; the use of CPAs resulted in a decreased recovery rate of cryopreserved testicular cells ( Supplementary Figs 1 and 2). These data suggest that the type of freezing medium is more important than the choice of CPA for the cryopreservation of canine testicular cells.
Our previous studies showed that the testes of 2-month-old beagles had an abundance of undifferentiated germ cells, and that GDCs can be derived from these testes using StemPro-34 and DMEM supplemented with FBS media. Porcine GDCs can be cultured in StemPro-34 medium as well 11,19 . Here, we showed that BM GDCs cannot be formed in DMEM supplemented with 5 and 10% of FBS. However, StemPro-34 medium is sufficient for the culturing of GDCs derived from dogs and pigs. This indicates that StemPro-34 medium is widely acceptable basic medium for the derivation and proliferation of mammalian SSCs.
SSC transplantation into the seminiferous tubules of germ cell-depleted immunodeficient mice, in order to determine spermatogonial characteristics of domestic animals, has been reported. Pig-derived germ cells were able to colonize the seminiferous tubules of immunodeficient mice, but the subsequent stages of donor-derived spermatogenesis were not observed 23 . Our previous investigations demonstrated that the xenotransplantation of porcine GDCs to immunodeficient mice can lead to a successful colonization and localization of the GDCs at the seminiferous tubule basement membrane in the recipient testes 11 . Human SSCs survived in mouse testes for at least 6 months and proliferated during the first month after transplantation. No human differentiated spermatogonia were identified, and meiotic differentiation did not occur in mouse testes 24 , although the transplantation of mouse SSCs to immunodeficient mice resulted in the completion of spermatogenesis and successful offspring production 3,25 . Here, the seminiferous tubules of immunodeficient mice were colonized by GDC cells, but spermatogenesis was not completed; only the undifferentiated male germ cells were detected in the tubules. These results suggest that the microenvironment of mouse seminiferous tubules is not optimal for the domestic animal-derived germ cell transplantation.
Xenografting of mammalian immature testes into the dorsal skin of immunodeficient mice allows the completion of spermatogenesis in vivo 16 . However, the reports on xenotransplantation of in-vitro cultured or freeze-thawed testicular cells are rare. Functional testis tissues were observed 41 weeks following the implantation of isolated neonatal porcine testis cells under the skin of immunodeficient mice; somatic cells and germ cells reorganized into structures that were remarkably similar, both morphologically and physiologically, to normal testis tissue 26 . Additionally, complete peccary spermatogenesis, together with the production of fertile sperm, were observed in the tissues formed from testicular cell suspension xenografts 8 months post-grafting 18 . The presence of a basement membrane, a histologically normal interstitium, containing putative Leydig cells, the establishment of tubule lumen, and the integration of few putative spermatogonia into the seminiferous epithelium were observed in the xenografts of dissociated immature rat testicular cells 27 . Xenografting cells isolated from sheep testis tissues led to a complete spermatogenesis 40 weeks post-grafting 28 . In addition to rat testicular cells grafts, complete spermatogenesis has been achieved using cells originating from pig and sheep. We obtained reconstructed seminiferous tubules by the xenotransplantation of cultured BM GDCs and testicular somatic cells conjugated with ECM into the dorsal skin of immunodeficient mouse model. However, differentiated spermatogenic germ cells were not identified, and the majority of the observed cells were Sertoli cells; a few germ cells existed in the tubules. Similar morphology has been reported for dog-derived xenotransplants. Seminiferous tubules of cryptorchidic dogs are lined only by Sertoli cells, and these cells in atrophic tubules with impaired spermatogenesis often have large empty intracytoplasmic vacuoles 29 . In cryptorchidic mice, abdominal heat stress leads to the apoptosis of the testicular germ-cells 30 . Germ cell degeneration has also been observed in heat-treated pufferfish; DMRT1-expressing Sertoli cells represented the majority of cells in testicular tubules 31 . Furthermore, azoospermia was observed during suspension of the testicle to the skin at the scrotal neck for 1 year, but it was shown that short-term scrotal hyperthermia in dogs does not cause substantial changes in sperm quantity and quality 32,33 . Long-term heat treatment (>1 month) induces germ cell and spermatogenesis degeneration, and therefore, it is possible that the higher temperature (~36-37.5 °C) of nude mice caused germ cell depletion in the reconstructed BM testes following the xenotransplantation, because canine testis temperature is generally around 32-33 °C [33][34][35] .
Another possible explanation is that in vitro cultured BM GDCs are not fully functional in the reconstructed tubules, and that they are not capable of self-renewal and spermatogenesis, which leads to a possibility that the majority of the grafted BM GDC cells were degraded. Previous successful xenograft studies used freshly isolated cells from neonatal porcine testes 26 and fragments of testis tissues, which contain functional spermatogonia and supporting physical microenvironment 18,28 . In support of this hypothesis, the xenografts of rat testicular cells that formed spherical aggregates on ECM coated dishes did not result in spermatogenesis, although seminiferous tubules were reconstituted 27 . Our research and previous studies suggest that the supporting environment plays an important role in germ cell differentiation. In conclusion, StemPro-34 medium with DMSO was demonstrated to be optimal for the cryopreservation of canine testicular cells. Using this approach, germ cell characteristics are maintained in GDC culture after thawing. The transplanted BM GDC cells can colonize seminiferous tubules of the recipient mouse model. We showed that, following the xenotransplantation of these cells and testicular somatic cells with ECM into the dorsal skin of the recipient mice, very small population of germ cells is able to localize in the basement membrane of reconstituted seminiferous tubules. We conclude that the body temperature of recipient animals and the supporting GDC cell microenvironment should be taken into consideration during the xenotransplantation of GDC cells.
Karyotyping. The cryopreserved and primary testicular beagle and BM cells were incubated with 100 μ l of colcemid solution (Irvine Scientific, Santa Ana, CA, USA, 9311) in StemPro-34 medium for 3 h at 37 °C, and the cells were treated with 1% citrate. Following the incubation and treatment, the cells were lysed and fixed in a methanol:glacial acetic acid (3:1) solution. G-band formation was analysed on each chromosome.

Xenografting of BM GDC cells into the dorsal skin of immunodeficient mice. Single cells were
separated from GDCs with 0.25% trypsin. Dorsal skin of 3 immunodeficient mice was injected with 5 × 10 5 GDC cells, combined with 4.5 × 10 6 feeder somatic BM testicular cells thawed from the cryopreserved cells (GDC cells:BM testicular cells = 1:9), 50 μ l StemPro-34 medium and 50 μ l matrigel (BD Bioscience, Franklin Lakes, NJ, USA), using 23 gauge needle and 1-ml syringe. Immunodeficient mice xenografted with GDC cells and testicular somatic cells were sacrificed 20 weeks after the procedure and the tissues formed by GDC cells were collected. These tissues were observed under a microscope (Nikon, Tokyo, Japan), and afterward used for histology and immunohistochemistry.
Histology and immunohistochemistry. Six-micrometre-thick slide sections of 4-and 5-month-old BM testes, busulfan-treated immunodeficient mice testes, and xenograft tissues were deparaffinized with xylene and treated with 100 to 50% ethanol. The sections were stained with hematoxylin and eosin and examined under a microscope (Nikon, Tokyo, Japan). For immunohistochemical analyses, the slides were incubated with a target unmasking fluid (Accurate Chemical & Scientific Corporation, Westbury, NY, USA) for 15 min, using a microwave oven to retrieve the antigens. The slides were washed thrice with PBS and blocked with 10% normal goat serum (v/v). For each PGP9.5 staining, a section of 4-or 5-month-old BM testis was incubated with anti-PGP 9.5 antibody (1:100; AbD Serotec, Raleigh, NC, USA), at 4 °C overnight, and then washed 3 times with PBS. Negative control slides were incubated with 1% BSA. Subsequently, slides were incubated with horse radish peroxidase-conjugated secondary antibody (1:500; Santa Cruz Biotechnology, Dallas, TX, USA) for 1 h, at RT (25 °C), followed by incubation in 3,3′ -diaminobenzidine (Vector Laboratories, Burlingame, CA, USA). The slides were washed with PBS and then observed under a light microscope (Nikon, Tokyo, Japan). To detect PGP9.5, the tissue slides originating from xenografts were stained with anti-PGP 9.5 antibody (1:100; AbD Serotec, Raleigh, NC, USA) at 4 °C overnight, and then washed 3 times with PBS. As previously described, negative controls were incubated with 1% BSA. These slides were incubated with anti-mouse Alexa 568 and anti-rabbit Alexa 488 antibodies (both 1:500; Invitrogen, Carlsbad, CA, USA) against PGP 9.5, respectively, for 1 h at 25 °C (room temperature), followed by incubation with DAPI (Vector Laboratories, San Francisco, CA, USA). Afterward, the slides were washed with PBS and observed using an excitation filter of 450-560 nm and 200× magnification, under a fluorescence microscope (Nikon, Tokyo, Japan).
Real-time PCR. Total RNA from the cultured GDCs, originating from 4-or 5-month-old BM testes, was isolated using the RNeasy Mini Kit (Qiagen, Venlo, the Netherlands). cDNA templates were prepared from 1000 ng total RNA using Maxime RT Premix kit (iNtRON Biotechnology, INC., Seongnam, Korea). cDNA synthesis conditions were as follows: each cycle for 60 min at 94 °C, while the inactivation was performed for 5 min at 95 °C. Gene-specific primers used for beagle sample analysis were applied in the experiments with BM GDCs and testes as well 19 . Real-time PCR was performed with glyceraldehyde 3-phosphate dehydrogenase (GAPDH), GFRα-1, PLZF, Oct4, Nanog, and GATA-binding protein 4 (GATA4) primers, using Rotor-Gene Q (Qiagen, Venlo, the Netherlands). Real-time PCR conditions were as follows: 40 cycles of 20 s at 94 °C, 20 s at 60-62 °C, and 20 s at 72 °C, for all genes.
Statistical analyses. Statistical analyses of recovery rate and colony number were performed using one-way nested analysis of variance (ANOVA) with Tukey test. Gonado-somatic indices and real-time PCR results were analysed using unpaired t-test with Welch's correction analysis of variance. All analyses were performed using GraphPad Prism 4. Quantitative gene expression differences between colonies and testis tissue were calculated with Ct and delta-delta-Ct obtained in comparison with the GAPDH Ct value. The null hypothesis was rejected when the P-value was < 0.05.