Spermatogonial stem cell transplantation was first described by Ralph Brinster and colleagues in 1994, who demonstrated that SSCs could be isolated and transplanted to regenerate spermatogenesis in infertile recipient mice [75, 76]. SSC transplantation has now been reported in mice, rats, pigs, goats, bulls, sheep, dogs, and monkeys, and donor-derived progeny has been produced by natural breeding in mice, rats, goats, and sheep [77–88]. SSCs from donors of all ages, newborn to adult, are competent to regenerate spermatogenesis [78, 89], and SSCs can be cryopreserved and retain spermatogenic function upon thawing and transplantation [85, 90, 91]. Thus, it appears feasible that a testicular tissue biopsy (containing SSCs) could be obtained from a prepubertal boy prior to gonadotoxic therapy, frozen, thawed at a later date, and transplanted back into his testes to regenerate spermatogenesis. If spermatogenesis from transplanted cells is robust, this approach may restore natural fertility, allowing survivors to achieve pregnancy with their partner by natural intercourse and have biological children.

Radford and colleagues already reported cryopreserving testicular cells for 11 adult non-Hodgkin’s lymphoma patients in 1999 and subsequently reported transplanting autologous frozen and thawed testis cells back into the testes of seven survivors [92, 93]. The fertility outcomes for patients in that study have not been reported, and even if the men fathered children, it would not be possible to ascertain whether the sperm arose from transplanted stem cells or surviving endogenous stem cells. This uncertainty will always plague the interpretation of human SSC transplant studies where it is not ethically possible to genetically mark the transplanted cells because the genetic modification would be transmitted to progeny. Therefore, large epidemiological datasets generated over decades will be required to prove the fertility benefit of SSC transplantation. Nonetheless, this study demonstrates that patients are willing to pursue experimental stem cell-based options even when there is no guarantee of a fertile outcome. There are no published reports of SSC transplantation in humans since Radford’s follow-up report of his non-Hodgkin’s lymphoma patients in 2003 [93].

Based on our experiences at the Fertility Preservation Program in Pittsburgh [54] and published reports [56, 59], it is reasonable to expect that 50–500 mg of testicular tissue can be obtained by wedge biopsy or needle biopsy from a single testis of a prepubertal boy. This is a small amount of tissue relative to size of adult human testes that can range from 11 to 26 g in size [94]. It is widely believed that the number of stem cells in biopsies from prepubertal boys will be small and that SSCs will have to be expanded in culture prior to transplant. Conditions for maintaining and expanding rodent SSCs in culture are well established and SSCs maintained in long-term culture remain competent to regenerate spermatogenesis and restore fertility [95–100].

If cultured human SSCs function similar to cultured rodent SSCs, it should be feasible to expand a few stem cells obtained from the testis biopsy of a prepubertal boy to a number sufficient to produce robust spermatogenesis upon transplantation back into his testes when he is an adult. Several studies have reported culturing human SSCs [57, 58, 101–109], including two studies in which cultures were established from the testes of prepubertal patients [58, 101]. Human SSC cultures have been evaluated by quantitative PCR or immunocytochemistry for spermatogonial markers or xenotransplantation into mouse testes. Strategies to isolate and culture human spermatogonia have been unique to each study, and to date no approach has been independently replicated in another laboratory. Also, the field is frustrated by the lack of a functional assay to test the full spermatogenic potential of cultured human cells.

A testicular biopsy obtained from a cancer patient could harbor malignant cells, especially for patients with leukemia. Kim and colleagues [110] reported that 20 % of boys with acute lymphocytic anemia had malignant cells in their testicular tissue prior to the initiation of oncologic treatment. Jahnukainen and colleagues [111] reported the transmission of leukemia after transplantation of testis cells from terminally ill leukemic rats into the testes of non-leukemic recipients. The same group further demonstrated that transplantation of as few as 20 leukemic cells was sufficient for disease transmission, leading to terminal leukemia within 3 weeks.

Because infertility is not life threatening and fertility treatments are elective, it is essential that risk of cancer recurrence after transplant be reduced to zero. Fluorescence-activated cell sorting (FACS) and magnetic-activated cell sorting (MACS) strategies to isolate and enrich therapeutic spermatogonia while removing malignant contamination have been explored with mixed results. To date, transplantable human spermatogonia have been recovered in the Ep-CAM^lo, THY-1^lo, CD49f⁺, SSEA4⁺, GPR125⁺, and CD9⁺ fractions of human testis cells [94, 102, 105, 112–114].

Fujita and coworkers isolated germ cells from the testes of leukemic mice in the forward scatter high and side scatter low fraction (positive selection), which was then further divided into fractions that were CD45/MHC class I antigens (H-2K^b/H-2D^b) double positive and CD45/MHC class I double negative cells. All recipient males injected with the CD45⁺/MHC class I⁺ cells developed terminal leukemia within 40 days. All mice injected with CD45^–/MHC class I^– cells survived for 300 days without onset of leukemia and produced donor-derived offspring [115]. In a subsequent study, the same group reported that seven out of eight human leukemic cell lines expressed the cell surface antigens CD45 and MHC class I [116]. In a rat model of Roser’s T-cell leukemia, Hou and colleagues concluded that single parameter selection using either leukemic (CD4 and MHC Class I) or SSC (Ep-CAM) markers was not sufficient to eliminate malignant contamination [117], but malignant contamination was successfully removed using a combination of leukemia and SSC markers (plus/minus selection) [114, 118]. Using similar positive/negative selection strategies, Hermann and colleagues isolated VASA+ germ cells in the THY-1⁺/CD45^– fraction of leukemia-contaminated prepubertal nonhuman primate testis cells [118], and this fraction did not produce tumors in mice. Dovey and colleagues contaminated human testis cells with MOLT-4 acute lymphoblastic leukemia cells and demonstrated by xenotransplantation that the Ep-CAM^lo/HLA-ABC^–/CD49e^– fraction was enriched 12-fold for transplantable human SSCs and was devoid of malignant contamination [114]. Collectively, these results are encouraging, but caution is still warranted as Geens and colleagues concluded, using EL-4 lymphoma contaminated mouse and human testis cells, that FACS- and MACS-based methods were insufficient to remove malignant contamination [119].

It will not be possible to perform comprehensive in vivo testing on patient samples because this would limit the amount of sample available for fertility therapy. More sensitive PCR-based methods have been described for detection of minimal residual disease (MRD), and this approach has identified malignant contamination in many ovarian tissue samples that were preserved for leukemia patients, even after negative histology and immunocytochemistry examination [120, 121]. However, in one of those studies, Dolmans and colleagues obtained disparate results from histology, qRT-PCR, and xenografting of ovarian tissues from leukemia patients. Quantitative RT-PCR to detect MRD revealed the possibility of malignant contamination in 9 of the 16 samples that was not detected by histological examination. However, when those ovarian tissues were grafted into recipient mice, only five of the nine samples with positive MRD had evidence of leukemic cells 3 months after transplantation [120]. Were the MRD results in the other four cases nefarious, or were they accurate and the leukemic cells simply failed to survive freezing, thawing, and grafting? In the absence of a definitive and practical test of malignant contamination, alternatives to autologous transplantation are needed for patients with hematogenous cancers, testis cancers, or cancers that metastasize to the testes.

Testicular cells (including germ cells, Sertoli cells, peritubular myoid cells, and Leydig cells) have the remarkable ability to reorganize to form normal looking seminiferous tubules when grafted under the skin of recipient mice [122–126]. Ina Dobrinski and colleagues disaggregated neonatal pig and sheep testis cells, pelleted them by centrifugation, and grafted them under the skin of immune-deficient mice. When grafts were recovered between 16 and 41 weeks after transplant, cells had reorganized to form seminiferous tubules with complete spermatogenesis [125, 126]. In a remarkable extension of this approach, Kita and colleagues [124] mixed fetal or neonatal testis cells from mice or rats with GFP⁺-cultured mouse germline stem cells and growth factor-reduced Matrigel and grafted under the skin of immune-deficient mice. Seven to 10 weeks after grafting, seminiferous tubules with complete spermatogenesis originating from both intrinsic germ cells and cultured (GFP⁺) germ cells were observed. Tubules were dissected and GFP⁺ round spermatids were recovered, injected into mouse oocytes. The resulting embryos were transferred to recipient females and gave rise to ten mouse pups, including four with the GFP transgene. It may be feasible to build a human testis from disaggregated human testis cells, but this has not been reported to our knowledge. The human experiment may be complicated by limited availability of fetal, neonatal, or prepubertal human testis cells. It does not appear that anyone has tried to “build a testis” from disaggregated adult testis cells for any species. One day it may be possible to “build a testis,” in vitro or in vivo, on the scaffold of a decellularized human testis [127].

Testicular tissue grafting may provide an alternative approach for generating fertilization-competent sperm from small testicular biopsies. In contrast to the SSC transplantation method in which SSCs are removed from their cognate niches and transplanted into recipient seminiferous tubules, grafting involves transplantation of the intact SSC/niche unit in pieces of testicular tissue. Honaramooz and colleagues reported that grafted testicular tissue from newborn mice, rats, pigs, and goats, in which spermatogenesis was not yet established, could mature and produce complete spermatogenesis when xenografted into nude mice [128]. The same group later reported the production of live offspring from sperm obtained from mouse testicular tissue grafts [129]. Fertilization-competent sperm was also produced from xenografts of prepubertal nonhuman primate testicular tissue transplanted into mice [130]. These results suggest that it may be possible to obtain fertilization-competent sperm by xenografting small pieces of testicular tissue from a prepubertal cancer patient under the skin of mice or other animal recipients such as pigs that are already an established source for human food consumption, replacement heart valves [131, 132], and potentially other organs [133]. Xenografting would also circumvent the issue of malignant contamination. However, the xenografting approach raises concerns about xenobiotics because viruses from mice, pigs, and other species can be transmitted to human cells [134, 135]. There is no evidence to date that xenografted human testicular tissue can produce spermatogenesis or sperm in mice [136–141]. However, there is reason for optimism because Sato and colleagues observed primary spermatocytes 1 year after xenografting testicular tissue from a 3-month-old boy that clearly did not have spermatocytes at the time of transplantation [140]. Xenografting of human testicular tissue to species other than mice has not been tested to our knowledge.

If malignant contamination of the testicular tissue is not a concern, autologous testicular tissue grafting can be considered. Luetjens and colleagues demonstrated that fresh autologous testicular tissue grafts from prepubertal marmosets could produce complete spermatogenesis when transplanted into the scrotum but not under the skin [142]. Frozen and thawed grafts did not produce complete spermatogenesis in that study, but those grafts were only transplanted under the skin. Therefore, additional experimentation is merited. Testicular tissue grafting will not restore natural fertility but could generate haploid sperm that can be used to fertilize oocytes by ICSI.

Sato and colleagues reported that intact testicular tissues from newborn mice (2.5–3.5 days old) could be maintained in organ culture and mature to produce spermatogenesis, including the production of fertilization-competent haploid germ cells [143, 144]. Testicular tissues from neonatal mice were minced into pieces (1–3 mm³) and placed in culture at the gas/liquid interface on a slab of agarose that was soaked in medium. Haploid round spermatids and sperm were recovered from the tissue after 3–6 weeks in culture and used to fertilize mouse eggs by ICSI. The resulting embryos were transferred to pseudopregnant females and gave rise to healthy offspring that matured to adulthood and were fertile. If testicular tissue organ culture can be translated to humans, it will provide an alternative to autologous SSC transplantation, autologous grafting, and xenografting in cases where there is concern about malignant contamination of the testicular tissue. The same authors were also successful to produce haploid germ cells in organ culture of frozen and thawed testicular tissues, which is particularly relevant to the cancer survivor paradigm. However, the fertilization potential of those sperm was not tested [143].