Spermatogonial Stem Cell Niche and Spermatogonial Stem Cell Transplantation in Zebrafish

1Spermatogenesis is a cellular developmental process by which self-renewing spermatogonial stem cells (SSCs) differentiate into millions of sperm daily,. To sustain this process continuously throughout the male reproductive life span, SSCs reside within a specific microenvironment in the testes called “niche” which regulates their properties, such as, self-renewal, pluripotency, quiescence and their ability to differentiate –. Despite of its crucial importance on SSC fate, the cellular and molecular composition of SSC niche remain unknown for several species of vertebrates. In rodents, the SSC niche has recently been identified within regions of the seminiferous tubules which are adjacent to the interstitial compartment,, preferentially along the branches of the interstitial blood vessels. It has been hypothesized that the cellular and molecular environment near the interstitial compartment promotes SSC renewal, and when SSCs leave these areas, the associated changes in their environment promote SSC differentiation. The proximity of SSC niche to the interstitium perhaps reflects the vascular supply of oxygen, nutrients, or hormones, such as follicle-stimulating hormone (FSH) or luteinizing hormone (LH) which influence Leydig and Sertoli cell functions on SSC self-renewal and also on SSC retention and homing in the niche, –. For example, FSH induces the secretion of GDNF (glial cell-line derived neurotrophic factor), an extrinsic stimulator of SSC self-renewal, produced by Sertoli cells,.

2Currently, the only means to study SSCs and their niche is by exploiting the stem cells' functional properties, such as slow-cycling and quiescent nature through the label-retaining cell (LRC) approach, or by studying SSC functionality and plasticity by transplantation assays. In this context, transplantation techniques developed by Brinster and collaborators, has enabled tremendous progress in the phenotypic and functional investigations of SSCs. Nowadays, SSC transplantation approaches have been developed for a number of species including also teleost fish,. These data have broad implications for understanding the regulation of spermatogenesis, stem cell biology, etiology of male infertility –, and also for advancing biotechnologies such as conservation of valuable genetic stocks, preservation of endangered species, and also as new option for transgenesis,,.

3In anamniote vertebrates (fishes and amphibians), we find the cystic type of spermatogenesis. There are two main differences compared to higher vertebrates. First, within the spermatogenic tubules, cytoplasmic extensions of Sertoli cells form cysts that envelope a single, clonally and hence synchronously developing group of germ cells deriving from a single spermatogonium. Second, the cyst-forming Sertoli cells retain their capacity to proliferate also in adult fish,. Hence, the basic functional unit of the spermatogenic epithelium in fish is a spermatogenic cyst formed by a dynamic group of Sertoli cells surrounding and nursing one synchronously developing germ cell clone. Different clones being in different stages of development generate the typical histological picture of fish testes, where the tubular compartment contains cysts of different sizes with groups of germ cells in different stages of spermatogenesis (see Figures 1, 2 and 3). Detailed histological studies in zebrafish (Danio rerio) have described two subtypes of type A undifferentiated spermatogonia in the testes, designated as Aund* and Aund. It is not known if these two subtypes are separated by mitosis, or represent different stages of the same cell cycle, and moreover, there is no information on the spermatogonial stem cell niche in the zebrafish testis, or on the stemness of Aund* and Aund.

4In the current study, we identified the putative SSCs and their niche by identifying the LRCs in zebrafish testes and using a transgenic zebrafish expressing enhanced green fluorescent protein under the control of the germ cell specific vasa promoter (vasa::egfp). Furthermore, we evaluated the spatial relationship between blood vessels and the SSC niche, using testes expressing enhanced green fluorescent protein under the control of the endothelial cell specific fli promoter (fli::egfp). To confirm the biological activity (re-establishment of function and plasticity) of the potential SSCs, we developed a transplantation assay in zebrafish. Finally, we studied the recipient's testicular microenvironment prior to transplantation as it might influence the behavior of transplanted SSCs.

5Sexually mature zebrafish males and females, and sexually mature transgenic zebrafish males expressing enhanced green fluorescent protein under the control of the germ cell-specific vasa promoter (vasa::egfp) or the endothelial cell-specific fli promoter (fli::egfp) were used. Animal housing and experimentation were consistent with Dutch and Brazilian national regulations and were approved by the Utrecht University and Federal University of Minas Gerais animal use and care committees, respectively.

6Testes from males (n = 5) were fixed in 4% buffered glutaraldehyde at 4°C overnight, dehydrated, and embedded in Technovit 7100 (Heraeus Kulzer, Wehrheim, Germany, http://www.kulzer-technik.de), sectioned and stained according to conventional histological procedures,. The topographical distribution of A spermatogonia was recorded by examining if type A undifferentiated spermatogonia (Aund* or Aund) were adjacent to the interstitial compartment, or contacted one or more tubules (intertubule). The position of 500 type A undifferentiated spermatogonia was counted per animal and expressed as percentage of the total number evaluated; to determine if the distribution of these cells follows a random pattern, the tubular perimeters adjacent to the interstitium, or intertubule were measured using Image J software (National Institutes of Health, Bethesda, Maryland, USA, http://rsbweb.nih.gov/ij), and the values were expressed as percentage of the total tubular perimeter (n = 50 tubules/animal).

7To estimate the cell cycle duration of type A undifferentiated spermatogonia (Aund* and Aund), 10 males were exposed to BrdU (Sigma-Aldrich, St. Louis, MO, USA, http://www.sigmaaldrich.com) dissolved in water (4 mg/ml) for approximately 4 and 10 h. To identify the LRC population, 25 males were pulsed with BrdU dissolved in water (4 mg/ml) for 10 h/day during 3 consecutive days. Animals (n = 5) were sacrificed immediately after the third BrdU pulse, and after 4, 11, 18 and 25 days of chase. Testes were fixed, embedded, and sectioned as described above. BrdU incorporation was detected by immunohistochemistry as described previously. The labeling index of both types of A undifferentiated spermatogonia (Aund* or Aund) was determined by counting the total number of labeled cells out of 300 cells (labeled and non-labeled) during the different periods after chase. After 18 and 25 days of chase, the spatial distribution of labeled Aund* or Aund (LRCs) was determined by counting the number of labeled cells situated near the interstitium, or in the intertubule, or near the testicular capsule. The values were expressed as percentage of positive-BrdU cells in the mentioned regions.

8Testes from transgenic vasa::egfp or fli::egfp zebrafish were fixed in 2% buffered paraformaldehyde for 2 h. For whole-mount examination, vasa::egfp testes were permeabilized with 0.2% PBT (0.2% Triton X-100 in PBS) for 10 min, and subsequently stained in DAPI (Invitrogen Molecular Probes, Carlsbad, CA, USA, www.invitrogen.com) for 5 min, followed by two rinses in PBS for 10 min. vasa::egfp testes (n = 5) were analyzed by CLSM 510 Meta (Zeiss, Jena, Germany, www.zeiss.de/lsm) using 358 nm and 488 nm as excitation wavelengths for DAPI and GFP, respectively. fli::egfp testes were frozen in Tissue-Tek (Sakura Finetek Europe B.V., Leiden, Netherlands, http://www.sakura.eu) and cryosectioned at 10 µm, permeabilized in 0.2% PBT for 10 min, stained with DAPI (Invitrogen) for 5 min, and mounted with a coverslip using an anti-fading Vectashield mounting medium (Vector Laboratories, Burlingame, CA, USA, http://www.vectorlabs.com). Sections were examined under fluorescence microscopy using filters for DAPI and FITC visualization.

9To deplete endogenous spermatogenesis, we first examined the effect of different temperatures. Overall, the speed of spermatogenesis in teleost fish is significantly influenced by temperature. Hence, males were kept in water at 20°C (n = 15), 27°C (n = 12), 30°C (n = 7), or 35°C (n = 12) for at least one week. Then, animals received one single intraperitoneal injection of 3H-thymidine (Amersham/GE Healthcare, Piscataway, NJ, USA, http://www.gehealthcare.com) (2 µCi/g/BW), and were sacrificed at 2 h, 12 h, 1, 2, 3, 4, 5 and 6 days after thymidine injection. Testes were weighed for calculating the gonadosomatic index (GSI = testes weight/body weight ×100), fixed, embedded and sectioned as above, and prepared for autoradiographic analysis to estimate the duration of meiosis and spermiogenesis as described previously,. Based on these results (Figure S1), 35°C was chosen as optimal temperature to deplete endogenous zebrafish spermatogenesis. Then, 60 zebrafish males were kept at 35°C for one week, and received a single intraperitoneal dose of 30 or 40 mg/Kg/BW of busulfan (Sigma). Testes from males (n = 6) were sampled 2, 4, 6, 10 and 12 days after the injection, weighed, fixed, embedded, sectioned and stained as above. As control group, males (n = 30) received a single intraperitoneal injection of dimethyl sulfoxide and sampled at the same reported periods. To evaluate the optimal dose and the best window of depletion, frequency of spermatogenic cysts, germ cell apoptosis, and Sertoli cell only phenotype were determined for each sampled period. Results were expressed as percentage of the total number of counted structures. To address the composition of type A undifferentiated spermatogonia (Aund* or Aund) in busulfan-depleted testes (n = 10), the percentage of both subtypes was determined by counting the number of Aund* and Aund out of 100–200 type A undifferentiated spermatogonia in each male. As control, the percentage of both subtypes was also determined in normal testes (n = 7).

10Ovoposition was induced according to usual procedures (http://zfin.org). To evaluate if females after ovoposition may be suitable recipients for SSC transplantation, ovaries (n = 5) were fixed, embedded, and sectioned as above, and stained with PAS to determine the number of postovulatory follicles (POFs)/mm2 of tissue. POFs consisting for a great part of granulosa cells remaining after ovulation were considered as a space potentially available for transplanted SSCs. The number of POFs/mm2 from preovulatory females was also evaluated as a control.

11Testes from males (n = 10) were digested with 0.2% collagenase and 0.12% dispase. The obtained cell suspension was immediately submitted to FACS (Fluorescence Activated Cell Sorting) using an inFlux cell sorter (BD Bioscience, San Jose, CA, USA, www.bdbiosciences.com). Vasa is highly expressed in type A undifferentiated spermatogonia, but expression decreases during meiosis and spermiogenesis. Since type A undifferentiated spermatogonia are the largest germ cell type in zebrafish testes (∼10 µm nuclear diameter ), FACS settings were adjusted to sort a cell population displaying large size and high intensity of fluorescence, which should enrich type A undifferentiated spermatogonia. To validate the enrichment, the unsorted and sorted cell fractions obtained by FACS was analyzed under fluorescence microscopy, or fixed, embedded, sectioned, and stained as above to determine the percentage of germ cell type. For SSC transplantation, the FACS-enriched cell fraction was suspended in L-15 medium (Sigma), 5% trypan blue and 10% calf serum. SSCs were transplanted into testes or ovaries through the genital pores, using a glass capillary needle coupled to a peristaltic pump (Figure S2). To optimize this procedure, zebrafish male and female genital pores were analyzed morphologically (diameter and angle) to adjust the settings for the glass capillary needle (Figure S2). The transplantation route was standardized and tested by injecting trypan blue (Figure S2).

12Recipients were sacrificed 2 and 3 weeks (males) or 3 and 4 weeks (females) after transplantation. The gonads were fixed in 2% buffered paraformaldehyde for 2 h, permeabilized, and stained with DAPI before analysis by CLSM, as described above. As positive control, Vasa protein expression was examined by GFP immunodection in vasa::egfp gonad sections or by Vasa immunocytochemistry in wild-type males. The PCR detection of gfp DNA from donor-derived germ cells in male and female recipients was carried out as described previously.

13Males (n = 19) were sampled 10 days after busulfan (40 mg/kg) treatment. A blood sample was collected for quantification of 11-ketotestosterone (11-KT) plasma levels, as described previously. As controls, 11-KT plasma levels were also quantified in zebrafish kept at 27°C (n = 7), or at 35°C (n = 11). In other experiments, carried out at the same time, but published separately, 11-KT plasma levels were measured 2 h after a single injection of recombinant zebrafish Fsh or hCG. Results are expressed as ng 11-KT/ml of plasma. To evaluate testicular 11-KT release in tissue culture, testes were collected from adult zebrafish kept at 27°C (control) (n = 7), or from busulfan-treated zebrafish (n = 7). The two testes of a given fish were incubated in parallel, such that one of them (randomly chosen left or right) served as control (basal) for the contralateral one, which was incubated in the presence of 1 µM of the adenylate cyclase activator forskolin. After incubation the medium was processed for the quantification of 11-KT. Results were expressed as ng 11-KT/mg of tissue.

14Testes from males kept at 27°C (control) (n = 7), or at 35°C (n = 5), or treated with busulfan (n = 12) were snap frozen in liquid nitrogen and stored at −80°C until RNA extraction. Total RNA was extracted from testes using the RNAqueous®-Micro Kit (Ambion, Austin, TX, USA, http://www.ambion.com). Further processing to determine the threshold cycle (Cq) values of the reference endogenous control gene elongation factor 1-alpha (ef1α) and β-actin1, as well as of insulin-like 3 (insl3), steroidogenic acute regulatory protein (star), and cytochrome P450, family 17, subfamily A, polypeptide 1 (cyp17a1), androgen receptor (ar), anti-Müllerian hormone (amh), gonadal soma-derived growth factor (gsdf), insulin growth factor 1a (igf1a) and 1b (igf1b), and germ cell genes piwil1 (spermatogonia), and synaptonemal complex protein 3 (sycp3l) (spermatocytes) by qPCR analysis was performed as reported,,. No significant differences (P>0.05) were found among the mean β-actin1 and ef1α Cq values in the different groups (Figure S3) thus validating β-actin1 and ef1α as suitable references for the current experiments. Then, relative mRNA levels of the selected genes were normalized to β-actin1 and ef1α, and expressed as fold of relative control (27°C) mRNA levels. Nomenclature of zebrafish proteins, mRNAs and genes are according to ZFIN (http://zfin.org) rules.

15Significant differences between two groups were identified using Student test (paired and unpaired) (P<0.05). Comparisons of more than two groups were performed with one-way ANOVA followed by Student-Newman-Keuls test (P<0.05). Graph Pad Prism 4.0 (Graph Pad Software, Inc., San Diego, CA, USA, http://www.graphpad.com) was used for all statistical analysis.

16Analyzing the topographical distribution of type A spermatogonia showed that 75% of these cells are situated adjacent to the interstitial compartment (Figure 1A,C,D) although this compartment represents only 1/3rd of the total perimeter of the spermatogenic tubules (Figure 1B). Quantification of the mitotic index of these cells was achieved by determining the BrdU labeling index during three successive 10 h long exposures to BrdU during three consecutive days. The results suggested that both subtypes of type A undifferentiated spermatogonia (Aund* and Aund) have a long cell cycle lasting at least 10 h but less than 30 h (Figure 2A). To confirm the existence of slow-cycling cells (stem cell candidates), testes were examined for BrdU-positive type Aund* and Aund spermatogonia at different periods (up to 25 days) after the final BrdU exposure; putative stem cells are considered to be part of the LRC population. During this chase period, two distinct patterns of BrdU staining were found among the two subtypes of type A spermatogonia (Figure 2A). Aund rapidly lost the label until the 18th day of chase, when the labeling index stabilized at ∼10% for this cell type (Figure 2A). On the other hand, the percentage of BrdU-positive type Aund* spermatogonia varied in a statistically not significant manner around 30% during the complete chase period (Figure 2A). Apart from some somatic elements, type A undifferentiated spermatogonia (Aund* and Aund) were the only BrdU retaining germ cells after 18 and 25 days of chase. Intriguingly, most of the labeled spermatogonia (71% of Aund* and 80% of Aund) were situated adjacent to the interstitium (Figure 2B,C–H). A similar pattern was found in vasa::egfp testes examined under CLSM where the strongest expression of vasa was observed adjacent to the interstitium (Figure 3 and Video S1). To further study the spatial relation between the vasculature and undifferentiated spermatogonia, testes of transgenic zebrafish expressing GFP in endothelial cells (fli::egfp) were analyzed. Capillaries surround the seminiferous tubules (Figure 3F), and type A undifferentiated spermatogonia were often found in close association with endothelial cells (Figures 2E-inset; 3G–I). A combination of the above results is illustrated schematically in Figure S4, showing a hypothetical spermatogonial stem cell niche in zebrafish.

17Zebrafish were exposed to different temperatures to optimize the treatment with busulfan (Figure 4A, Figure S1). Higher temperatures accelerated spermatogenesis (Figure S1), whereas at 35°C, spermatogenesis did not progress beyond metaphase I and showed abnormalities, such as massive apoptosis in particular amongst spermatocytes, absence of sperm, and a significant decrease in the GSI (Figure 4A,E; Figure S1). Both doses of busulfan (30 or 40 mg/Kg/BW) tested at 35°C induced further and a more general germ cell apoptosis (Figure 4D–H) and a progressive GSI decrease, which reached its lowest value 10 days after injection (Figure 4B,C). However, only the higher dose suppressed efficiently endogenous spermatogenesis, resulting in 88% of the spermatogenic tubules showing a Sertoli cell only appearance, i.e. all germ cells were missing (Figure 4D,I). Analyzing the composition of type A spermatogonia in the 12% of tubules where some germ cells remained and comparing this to normal testes demonstrated that the ratio between Aund* and Aund shifted from 1∶1 to 1∶3 in busulfan-depleted testes (Figure 1E). The preponderance of Aund might be associated with the observation that 12 days after injection, most tubules showed the first type B spermatogonia and/or spermatocytes again, i.e. showed a fast recovery of endogenous spermatogenesis after treatment with 40 mg/Kg/BW of busulfan at 35°C (Figure 4D).

18To prepare female zebrafish recipients for SSC transplantation, ovoposition was induced; a pilot study indicated a high mortality among females when applying the combination of high temperature and busulfan. The increased number of POFs showed that ovoposition created spaces and “free” somatic cells (follicle cells from POFs), potentially suitable for receiving injected germ cells, and to support transplanted SSCs development, respectively (Figure S5).

19Using vasa::egfp transgenic zebrafish as donors, a testicular cell suspension was obtained and subsequently submitted to FACS, in order to enrich transplantable SSC (Figure 4J,K) by sorting for big cells carrying high fluorescence (Figure 4L,M). Indeed, after sorting, we observed an enrichment of type A undifferentiated spermatogonia (Figure 4K–N). The sorted cells (8.103−4.104 cells/µl) were injected through the genital pore (Figure 4O, Figure S2). Since females have a prominent belly, direct injections into ovaries (via lateral body wall) were also successfully performed (Figure S2). After transplantation, zebrafish recipient males and females were placed in water of 27°C.

20After two weeks of transplantation, donor cells colonized the recipient's seminiferous epithelium, and formed clusters, which were situated near the interstitial compartment (Figure 5A,B). Using CLSM, we found that these clusters were composed of ∼8 cells/cyst (Figure 5C,D; Video S2). Three weeks after transplantation, donor-derived cysts had increased in number and size, and were found at different stages of spermatogenesis (e.g. differentiating type A spermatogonia, type B spermatogonia and spermatocytes) along the recipient's seminiferous epithelium (Figure 5E,F; Video S3). Donor type A undifferentiated spermatogonia (Aund* and Aund) were also found in the recipient seminiferous epithelium (Figure 5E,F). The vasa expression pattern in donor-derived germ cells was the same as observed in vasa::egfp testes immunostained for GFP, or in wild-type testes immunostained for Vasa (Figure S6). The transplantation efficiency was considered not high (30%) as colonization and donor-derived spermatogenesis was observed in about 3 out of 10 recipients. With regard to SSC transplantation into females, cell clusters derived from transplanted SSC were found in recipient ovaries after three weeks of transplantation (Figure 6A, Video S4). Small GFP-positive oocytes at an early stage of oocyte development were also found in recipient females (Figure 6B,C). Surprisingly, these male-derived germ cells progressed into oocyte development, and gave rise to advanced oocytes after one month of transplantation (Figure 6D–F; Video S5). Clusters of GFP-positive cells, possibly clones of oogonia, were also found near male-derived oocytes (Figure 6G,H). The presence of GFP-positive cells in both transplanted males (three weeks after transplantation) and females (one month after transplantation) was also confirmed by PCR analysis (Figure S7).

21The relatively low efficiency of colonization after transplantation triggered studies on the recipient's testicular microenvironment prior to transplantation. The plasma levels of 11-KT were ∼3 times higher in fish kept at 35°C (with or without busulfan injection) than in non-treated zebrafish kept at 27°C (Figure 7A). The temperature-induced elevation of 11-KT plasma levels was similar to the stimulatory effect of hCG (10 IU/g) or recombinant zebrafish Fsh (100 ng/g). Also, testicular 11-KT release in primary culture was higher from tissue of males exposed to 35°C and busulfan than from testis tissue of control males kept at 27°C, in particular with regard to basal release (7-fold higher), while the difference in forskolin-stimulated androgen did not reach statistical significance (Figure 7B). Quantifying the expression of selected genes revealed that transcript levels of steroidogenesis and androgen signaling genes (star, ar, cyp17a1) had increased 3-fold in busulfan-depleted testes (Figure 7C). In most cases, exposure to 35°C alone did not change gene expression. There are two interesting exceptions. The transcript levels of amh were significantly down-regulated at 35°C and in busulfan-depleted testes, while igf1b transcript levels were strongly up-regulated. Expression of germ cell-specific genes, such as, piwil1 and sycp3l significantly decreased only on busulfan-treated animals (Figure 7C), while gsdf and insl3 mRNA levels did not change among the groups (Figure 7C).

22Using a BrdU pulse-chase approach, we identified the LRC population in zebrafish testis. Within this category of testicular cells, somatic cells were found (e.g. Sertoli and interstitial cells), but type A undifferentiated spermatogonia (Aund* and Aund) were the only germ cells able to retain the label after 18 and 25 days of chase, constituting therefore the putative SSCs. Two distinct patterns of BrdU dilution were found among the two subtypes of type A spermatogonia in zebrafish. The BrdU dilution kinetics for type Aund showed a rapid decline that stabilized at the 18th day after chase, while the percentage of BrdU-positive Aund* spermatogonia remained constant. Applying mathematical modeling analysis to hematopoietic stem cells (HSCs), a similar biphasic profile of BrdU dilution was described. The initial, rapid decline was attributed to the faster turnover of activated cells, while the subsequent, decelerated decline resulted from the slow turnover of quiescent cells. This suggested the presence of a heterogeneous population of HSCs which can be activated or deactivated according to the systemic needs. Likewise, the biphasic profile found in zebrafish testes might be attributed to functional differences between the two subtypes of spermatogonia, in which Aund would be the more rapidly dividing, “active” population (related to differentiation and rapid proliferation, as indicated by the more rapid loss of BrdU), and Aund*, the slow-cycling, “reserve” population (related to slow self-renewal proliferation, as indicated by the relatively stable BrdU labeling index). Also in the human testis, two types of single A spermatogonia (pale and dark) are present, and might play distinct roles as “reserve” versus “active” stem cell, respectively,. In another teleost fish, in female medaka (Oryzias latipes), a BrdU pulse-chase experiment also revealed two distinct populations of oogonial stem cell, one being fast and the other a slow-dividing cell population. Across the vertebrates, it seems that tissues with constant demand for stem cells to proliferate and generate differentiated cells, two populations of stem cells are required; one to rapidly expand and produce differentiated cells, and another which rarely divides and functions as “reserve” in case of insults.

23According to the “immortal strand hypothesis”, stem cells divide asymmetrically, since a hypothetical mechanism would sort chromatids containing the original template DNA strands in one daughter cell, passing therefore the new strands to the other daughter cell. Hence, in a stem cell self-renewal division, the new stem cell would inherit the original templates, the other daughter cell, destined to differentiate, would inherit the new strands. Recently, it has been demonstrated that the “immortal strand hypothesis” is not exclusive of self-renewing cells, but also of divisions in which daughter cells adopt divergent fates. In this case, daughter cells inheriting the original templates would also retain the more immature phenotype, whereas daughter cells inheriting the newer templates acquired a more differentiated phenotype. The BrdU dilution for Aund suggests the existence of two types of template DNA strand segregation, one random type, possibly reflecting the rapid loss of the label up to the 18th day of chase, and one non-random type (asymmetrical division), which could explain the retention of BrdU after 18 and 25 days of chase. We speculate that the asymmetrical division of Aund might prevent its exhaustion during spermatogenesis, maintaining, therefore, a subset of immature Aund. An asymmetrical division of type A undifferentiated spermatogonia was suggested previously in zebrafish based on morphological evidence.

24The LRC assay also revealed the precise location of stem cells, and consequently the characterization of the neighboring cells that form the stem cell niche –. As in rodents,, we found that slow-cycling BrdU-labeled type A undifferentiated spermatogonia (Aund* and Aund) were preferentially situated in regions of the seminiferous tubules adjacent to the interstitial compartment. This observation suggests that elements of the SSC niche are probably conserved across vertebrates. The preferential location of SSC close to the interstitial compartment in rodents may be related to a locally high concentration of androgens which inhibit spermatogonial differentiation. 11-KT is the main androgen in zebrafish and other teleost fish, and is able to support full spermatogenesis in testis tissue culture of juvenile Japanese eel (Anguilla japonica) or adult zebrafish. Moreover, induction of androgen insufficiency in adult male zebrafish inhibited the differentiation of type A to type B spermatogonia. Although the possible role(s) of androgen signaling in regulating zebrafish SSC activity still have to be clarified, the available evidence in fish strongly suggests that, different from mammals, androgens stimulate spermatogonial differentiation. On the other hand, estrogens stimulated SSC self-renewal in Japanese eel, so that Leydig cells may be relevant as a source for aromatizeable androgens, thereby contributing directly and/or indirectly to the zebrafish SSC niche. Moreover, Leydig cell paracrine signaling might influence SSC behavior, since colony stimulating factor 1 (Csf1) produced by Leydig and myoid cells is an extrinsic stimulator of SSC self-renewal in mice.

25Endothelial cells support the expansion of normal and malignant stem cells not only by delivering oxygen and nutrients, but also by the paracrine release of endothelial cell growth factors and trophogens, which are referred to as “angiocrine factors”. Thus, endothelial cells have been pointed out as an important element of the stem cell niche in several systems, including the rodent SSC niche,. In the zebrafish model, the distribution of SSC candidates (Aund* and Aund) close to blood vessels might indicate an involvement of endothelial cells in regulating SSC function. While there is no information on angiocrine signaling in the zebrafish testis, it has been shown that platelet-derived endothelial cell growth factor (PD-ECGF) induced SSC self-renewal in Japanese eel,.

26The functional capacity of the identified SSC candidates in zebrafish was investigated by transplantation assays. In this assay, SSC candidates were transplanted into zebrafish in which endogenous spermatogenesis had been depleted by busulfan, similar to the technique developed for rodents and others species –,. It seems that similar to tilapia, zebrafish spermatogenesis is more sensitive to busulfan at elevated temperatures (35°C), resulting in 88% of spermatogenic tubules showing Sertoli cells only after 10 days of treatment. Interestingly, spermatogenesis recovered quickly after having passed the nadir of busulfan-induced depletion. It has been shown previously that proliferation of surviving, undifferentiated spermatogonia is greatly enhanced when the number of differentiating spermatogonia is reduced following busulfan exposure,. In addition, differentiated spermatogonia can act as “potential stem cells”, shifting their nature of transit-amplifying cells to self-renewal in order to rapidly recover spermatogenesis in depleted testes. Analyzing the composition of type A spermatogonia in normal versus busulfan-depleted testes in zebrafish, we have demonstrated a relative increase in the population of Aund, and a relative decrease in the population of Aund* spermatogonia in busulfan-depleted testes. Then, considering Aund as the “active” stem cell, our results suggest that the spermatogenic recovery after busulfan treatment might be associated with a preponderance of the active stem cell population Aund.

27The treatment with busulfan induced an increase in androgen plasma levels as well as an increase in testicular androgen release. This might reflect a suppression of the negative feedback on gonadotropin release. In rat, the negative feedback on FSH release exerted by germ cells via Sertoli cell-derived inhibin is transiently eliminated after busulfan treatment, explaining the increase in circulating FSH levels between 6 and 10 weeks after busulfan injection,. In mammals, androgen plasma levels were not elevated in busulfan-treated animals –. However, in fish, also Leydig cells express the receptor for Fsh, which is a strongly steroidogenic gonadotropin in fish (e.g.,, ). Hence, we attribute the increased androgen levels after busulfan treatment to higher Fsh plasma levels. Consistent with the elevated androgen production, busulfan treatment increased the mRNA levels of transcripts encoding proteins involved in steroidogenesis, such as star and cyp17a1. The down-regulation of amh mRNA expression in busulfan-depleted testes also seems coherent with higher androgen levels, since androgens down-regulated amh expression in juvenile Japanese eel testis and stimulated spermatogonial differentiation towards meiosis. Moreover, since Sertoli cells also proliferate in the adult fish testis, in particular those associated with expanding spermatogonial cysts,, the busulfan treatment may have depleted the proliferating Sertoli cell population as well, thereby contributing to lower levels of amh mRNA, and possibly also to the arrest in spermatogenesis. With regard to the expression of igf gene family members, a strong up-regulation in particular of igf1b mRNA levels has been observed. This is interesting in the context of ongoing work showing that recombinant zebrafish Fsh increases igf1b levels in adult testis tissue culture (unpublished data). Spermatogonial proliferation was stimulated in newt testis in primary tissue culture by exposure to Igf1. Since Sertoli cells also express a factor (similar to platelet-derived endothelial cell growth factor) that stimulates SSC self-renewal in Japanese eel, busulfan-induced changes in the number and/or physiological state of Sertoli cells might have resulted in a more generalized imbalance of factors released by Sertoli cells and relevant for SSC self-renewal and differentiation. Taken together, our data suggest that busulfan treatment might create a microenvironment in zebrafish testes where the balance of somatic (mainly Sertoli) cell-derived factors has shifted to a reduced level of factors supporting self-renewal and/or an elevated level of factors stimulating differentiation. While this might facilitate the fast recovery after the severe loss of germ cells induced by busulfan, it may also influence the behavior of transplanted cells. Thus, transplanted SSCs may differentiate rather than self-renew, possibly limiting the efficiency of donor-colonization (30%) in the recipient seminiferous tubules; however, a relatively low efficiency of donor-derived colonization is not exceptional for zebrafish and has been observed in other vertebrates as well. Future studies will aim at weakening the pro-differentiation or strengthening the self-renewal environment, for example by an estrogen treatment of the recipients.

28Despite of the limited efficiency, we have taken important steps towards the development and standardization of SSC transplantation assay to confirm the “stemness” of SSC candidates in zebrafish. Transplantation of a FACS-enriched fraction of Aund* and Aund into busulfan-depleted testes showed that stem cell candidates were able to colonize, self-renew, and to differentiate in the recipient seminiferous tubules. This confirms the presence of stem cells in the transplanted pool of type A undifferentiated spermatogonia. Moreover, spermatogonial colonization started from regions within the recipient's seminiferous tubules which are adjacent to the interstitial compartment, the putative SSC niche in zebrafish testes. This might indicate that the SSC niche creates a cellular and molecular environment which is optimal for colonization and development of transplanted SSCs. Since zebrafish have been used as a model for understanding the mechanisms of testicular germ cell tumors, and most of these tumors have an unknown etiology, a standard experimental design including reciprocal transplantation of germ cells from affected donors to wild-type testes and vice versa could be applied to evaluate whether the defect is on Sertoli cells or intrinsic to the germ cells. Another important application of germ cell transplantation is transgenesis through the male germ line using transplantation of transfected germ cells,. This might be an option to create transgenic animals in a shorter time than is possible with conventional methods.

29Finally, we have demonstrated the plasticity of the transplanted spermatogonia by placing the SSC candidates into a different microenvironment, i.e. into an ovary. SSC candidates were able to colonize recipient ovaries and differentiate into female germ line cells.

30We present in the current work characteristics of zebrafish SSC candidates (Aund* and Aund), their niche, and their functional capability to both self-renew and differentiate in recipient testes, as well as their remarkable plasticity in recipient ovaries. To our knowledge, this is the first work to report on a SSC niche in fish. As in rodents, SSC location in zebrafish was characterized within areas of the seminiferous tubules which are opposite to the interstitial compartment, suggesting the influence of interstitial elements on SSC self-renewal and maintenance. We developed and standardized SSC transplantation techniques in male and female zebrafish, showing donor-derived spermatogenesis and male-derived oocytes, respectively, after transplantation. Thus, we introduced SSC transplantation in zebrafish as a promising technique for the study of SSCs.