Regeneration of Spermatogenesis and Production of Functional Sperm by Grafting of Testicular Cell Aggregates in Zebrafish (Danio rerio)¹

Abstract

The self-renewal and differentiation of spermatogonial stem cells (SSCs) is essential for the continuous production of sperm throughout life in male vertebrates. The development of a functional assay to analyze these properties in isolated SSCs remains necessary. In our current study, we have developed a transplantation method for testicular cell aggregates in zebrafish (Danio rerio) in which allogeneic SSCs can undergo self-renewal and differentiation. The immature testes from juveniles are dissociated, aggregated by cultivation, and then transplanted under the abdominal skin of the recipient fish. The grafted aggregates reconstitute the appropriate testicular structures, including the lobule structure, consisting of basement membrane and interstitial steroid-producing cells on the outside, and the cysts, which comprise germ cell clusters and surrounding Sertoli cells. Bromodeoxyuridine incorporation analysis indicated that continuous spermatogenesis is maintained for at least 6 mo in the reconstituted testis. Moreover, when the sperm generated from the aggregates at 3 mo postgrafting were used for artificial insemination, fertilized eggs were obtained that developed sexually mature fish. These results suggest that self-renewal of SSCs takes place in reconstituted testes under the abdominal skin and that their differentiating progeny can develop into functional sperm. Furthermore, allogeneic spermatogonia were also found to proliferate and differentiate into sperm in these grafts. Our method of grafting testicular cell aggregates should thus prove useful not only analyzing the stem cell ability of an individual SSC but also for the production of progeny from cultured SSCs or SSCs of sterile mutants with somatic cell defects.

Introduction

The ability of stem cells to self-renew and to produce differentiating cells makes it possible to regenerate a cell population of a specific lineage in a tissue or organ following transplantation. A method of spermatogonial transplantation into the seminiferous tubules in mice, in which spermatogonial stem cells (SSCs) colonize and initiate spermatogenesis, has previously been established in SSC-defective dominant white spotting (W) mutant mice [1, 2]. This transplantation assay facilitates the identification of SSCs from isolated spermatogonia with valuable results, such as enrichment of SSCs [3, 4], establishment of SSC cultures [5], production of knockout mutants from genetically modified SSCs in culture [6], and identification of the potential and distinct ability of the stem cells in type A spermatogonia [7].

The zebrafish (Danio rerio) is an excellent model organism for investigating gene function through the use of mutants, including those with defects in spermatogenesis. In addition, significant advances have recently been made in cellular-based studies of spermatogonia, including characterization and successful culturing. Type A spermatogonia are defined as a single spermatogonium in a cyst structure surrounded by supporting somatic cells, the Sertoli cells, and containing an SSC population [8]. Morphological analysis has revealed the heterogeneity of type A spermatogonia [9]. The analysis of label-retaining cells via bromodeoxyuridine (BrdU) incorporation and with the replication protein Sld5 (synthetic lethal mutants of dpb11–1 5) has also indicated that type A spermatogonia can have different cell-cycle lengths and that these populations include quiescent cells (Saito and Sakai, unpublished results). However, the precise properties of stem cells in each subpopulation have not been determined. Since the establishment of culture conditions that allow us to grow differentiating spermatogonia [10], refinements that could lead to the establishment of SSC cultures have been investigated. A functional assay system for SSCs will be required to achieve this, but such a system has not been previously established in zebrafish.

A technique for germ cell transplantation into the lobules, in which spermatogonia differentiate or in which SSCs colonize and complete spermatogenesis, has been reported in teleosts (Oreochromis niloticus [11] and Odontesthes bonariensis [12]). Transplantation of germ cells from an adult testis into an embryo has also been shown to successfully transmit the germline and produce functional sperm and eggs [13]. Additionally, alternative transplantation techniques that induce spermatogenesis of SSCs have been reported in Xenopus laevis and mice [14, 15]. In these methods, Xenopus testicular cell aggregates or mouse testicular cells are transplanted into the subcutaneous space of the respective recipient. This leads to the successful reconstitution of testicular structures and the initiation of spermatogenesis beneath the skin of the recipient animal. Although the fertilizing ability of the sperm in these reconstituted testicular structures was not observed in the case of either X. laevis or mice, or for a similar method in pigs [16], the fertilizing ability of spermatids was confirmed by round spermatid injection into oocytes in mice [15], and fairly normal-looking sperm were produced in pigs. In addition, successful graft of testicular grafts has been reported, including the production of sperm able to fertilize eggs in rainbow trout (Oncorhynchus mykiss) [17] and mice [18]. These results imply that appropriate reconstitution of testicular structure, even if it is grafted, makes spermatogenesis possible.

In the present study, we describe the establishment of a new functional assay system for SSCs in zebrafish. The microinjection of dissociated germ cells into recipient lobules seems to be difficult in living recipient zebrafish because of their small body size and small size of the immature or germ cell-depleted testes. The transplantation of testicular germ cells into embryos requires a significant number of cells [13], making it difficult to transplant them into tiny embryos of zebrafish. We thus attempted to develop a transplantation system in zebrafish using aggregates of dissociated testicular cells as has been described in X. laevis [14]. When we used 9th- to 12th-generation sister-brother mating zebrafish pairs as donors and recipients to avoid graft rejection, we were able to transplant a testicular cell aggregate from dissociated testicular cells by centrifugation and cultivation into recipient juvenile zebrafish (Fig. 1). The transplanted aggregates successfully reconstituted zebrafish testicular structures, maintained continuous spermatogenesis that produced functional sperm for more than 3 mo, and allowed allogeneic spermatogonia to proliferate and differentiate. Our results thus suggest that this method will make it possible to analyze the properties of stem cells within isolated spermatogonia in zebrafish.

Fig. 1.

The strategy used to establish a transplantation system in zebrafish for aggregates of dissociated testicular cells. The immature testis was removed from juvenile fish and dissociated enzymatically. The dissociated testicular cells were then centrifuged and cultured to form aggregates, which were then transplanted just under the abdominal skin of a recipient juvenile. When labeled spermatogonia are added into the dissociated testicular cells before aggregation, it is possible to trace their proliferation and subsequent differentiation.

Materials and Methods

Zebrafish

The present study used TM strain zebrafish (generations 9–12), which were generated by sister-brother pair mating from a pair of Tuebingen lines. Vas::EGFP transgenic fish were produced to express the enhanced green fluorescent protein (EGFP) reporter gene under the control of the vas promoter [19] using a Tuebingen/AB (Tu/AB) line. At 5–8 wk postfertilization, male juvenile zebrafish with body lengths of 1.5–1.8 cm were used to obtain immature testes or as recipient animals. The use of these animals for experimental purposes was in accordance with the guidelines of the National Institute of Genetics.

Subcutaneous Transplantation of Testes and Testicular Cell Aggregates

Testes were removed from juvenile zebrafish under a dissecting microscope after anesthetization with 0.01% ethyl p-aminobenzoate (Wako). One day before transplantation, all recipient juveniles were maintained without feeding. The recipients were anesthetized, and an incision of approximately 2 mm was made into their abdominal skin with a razor blade. The tip of a forceps was then inserted between the muscle and skin through this wound to enable transplantation of a donor juvenile testis or a testicular cell aggregate. The recipient fish were reared in 0.4× PBS containing 10 μg/ml of gentamicin (Gibco) without sewing of the wound for 4 days to facilitate wound healing. The fish were then reared normally and, after an appropriate period, were anesthetized to enable removal of the grafted tissues. The removed aggregates or testes were then fixed in 4% paraformaldehyde (PFA) in PBS and acetic acid (10:1, v/v) for 2 h at room temperature and sectioned (thickness, 5 μm) in paraffin for histological or immunohistochemical analysis. In total, we transplanted seven testes from four juvenile TM males and 10 testes from five juvenile vas::EGFP transgenic males [19] individually into 17 juvenile TM males.

Aggregation of Dissociated Testicular Cells

Ten testes from five zebrafish juveniles (TM strain) were used to generate one testicular cell aggregate. The tissues were cut into five fragments and treated with 500 U/ml of collagenase (N-2; Nitta Gelatin, Inc.) in 80% Leibovitz L-15 medium (L-15; Gibco) for 2 h at 28.5°C with pipetting every 20 min. When we made the aggregate containing vas::EGFP testicular cells, dissociated cells from one vas::EGFP testis were mixed with TM testicular cells. The resulting testicular cell suspensions were then diluted seven times with 80% L-15 and centrifuged for 10 min at 190 × g. The supernatant was discarded, and the centrifugation pellet was resuspended in 100 μl of 80% L-15 supplemented with Hepes (10 mM, pH 7.9; Sigma), bovine serum albumin (BSA fraction V; 0.5%; Sigma), penicillin (50 U/ml; Gibco), and streptomycin (50 μg/ml; Gibco). The testes suspensions were then centrifuged for 3 min at 600 × g and incubated for 2 days at 28°C to facilitate the formation of cell aggregates. In total, we transplanted 20 aggregates individually into 20 juvenile TM males as described above.

After 3 mo posttransplantation, the aggregate was removed and suspended in 30 μl of Hank saline (0.137 M NaCl, 5.4 mM KCl, 0.25 mM Na₂HPO₄, 0.44 mM KH₂PO₄, 1.3 mM CaCl₂, 1.0 mM MgSO₄, and 4.2 mM NaHCO₃). Collection of unfertilized eggs from a single female of Indian line and the artificial insemination were performed according to the method of Westerfield [20].

BrdU Labeling and Detection

Bromodeoxyuridine treatments were performed in vitro or in vivo to label dividing spermatogonia and spermatocytes in the grafted aggregates or testes. For in vitro BrdU labeling, the grafted aggregates or testes were removed from the recipient fish and treated with 0.1% BrdU labeling reagent (GE Healthcare) in 80% L-15 for 3 h. For in vivo BrdU labeling, recipient fish at 6 mo posttransplantation were maintained in a tank containing 30 ml of 1 mg/ml of BrdU for 24 h.

Paraffin-embedded sections (thickness, 5 μm) of testicular cell aggregates or testes were obtained as described above, and BrdU incorporation was detected immunohistochemically using a Cell Proliferation Kit (GE Healthcare) in accordance with the manufacturer's instructions. During spermatogenesis in teleosts, germ cells form a cell cluster with surrounding Sertoli cells (the cyst) and develop synchronously [8]. Hence, we measured the percentage of cysts with labeled spermatogonia or spermatocytes in the total number of cysts (BrdU labeling index of cysts) to estimate spermatogonia and spermatocyte activity. The average ± SEM were thus obtained for seven grafted and recipient testes.

Detection of 3β-Hydroxysteroid Dehydrogenase/Δ^5–4-Isomerase Activity in Grafted Testicular Cell Aggregates

The detection of 3β-hydroxysteroid dehydrogenase/Δ^5–4-isomerase (3β-HSD) activity in a zebrafish testis was performed as described previously [21] with slightly modification. At 6 mo after transplantation of the testicular cell aggregates, these grafts were removed, fixed overnight in 4% PFA in PBS, washed three times in PBS containing 0.1% Tween-20 (PBST), and then stained in PBST containing 0.33 mg/ml of dehydroisoandrosterone, 2.5 mg/ml of nicotinamide adenine dinucleotide, 167 μl/ml of dimethyl sulfoxide, and 1.67 mg/ml of nitroblue tetrazolium (NBT). The stained tissues were then fixed in 4% PFA, embedded in paraffin, and sectioned (thickness, 5 μm).

Immunodetection of EGFP

To detect germ cells of the vas::EGFP transgenic zebrafish, EGFP immunohistochemistry was performed as described previously [14]. Briefly, sectioned (thickness, 5 μm) specimens of the testicular cell aggregates were immersed in citrate buffer (10 mM trisodium citrate, pH 6.0) and autoclaved (120°C, 5 min) to enable antigen retrieval. The sections were then incubated with 5 μg/ml of a solution of a green fluorescent protein (GFP) mouse monoclonal antibody (BD Clontech) for 1 h. After washing, the sections were further incubated with 2 μg/ml of alkaline phosphatase-conjugated anti-mouse immunoglobulin G antibody (Santa Cruz Biotechnology) for 1 h. Immunoreactivity was detected using NBT/5-bromo-4-chloro-3-indolyl phosphate, toluidine salt (Roche Molecular Biochemicals).

Results

Progression of Spermatogenesis Following Subcutaneous Transplantation

We first examined whether spermatogenesis is maintained in the normal testicular tissue after subcutaneous transplantation. Inbred strains are useful in this regard because they avoid rejection, but inbred strains have not yet been established in zebrafish. We achieved this, however, using TM strain zebrafish (generations 9–12) that were generated by sister-brother pair mating from a pair of Tuebingen lines. The immature testes from TM juveniles (5–8 wk after fertilization and 1.5–1.8 cm in body length) were used as the donors. Seven immature testes were removed (Fig. 2A) and transplanted under the abdominal skin of TM juvenile recipients. Four weeks later, the grafted testes and testes of the recipient were collected and treated with BrdU for 3 h. In all of the grafted testes (Fig. 2B) and recipient testes (Fig. 2C), abundant sperm and cysts containing spermatogenic cells at various stages were observed. Using a BrdU antibody to confirm the activity of spermatogonia and spermatocytes, we observed many cysts consisting of type A spermatogonia, type B spermatogonia, or primary spermatocytes that incorporated BrdU in the grafted testis to the same levels as in the recipient testis (Fig. 2, E and F). The BrdU labeling index of the cysts was measured at 42.6% ± 1.37% in the grafted testes and 43.1% ± 3.87% in the recipient testes, a difference that is not significant (t-test, P > 0.05) (Fig. 2G). These results suggest that the subcutaneous environment can support normal spermatogenesis in zebrafish. In contrast, when we transplanted the testis of vas::EGFP transgenic juveniles (Tu/AB) under the skin of TM juveniles, the grafted testes had degraded at 4 wk after transplantation in all 10 recipients (Fig. 2D). This finding indicated that the immune system of TM recipient fish had generated a response to allogeneic testes but not TM testes and, therefore, that sister-brother mating to generation 9 is sufficient to prevent host rejection in zebrafish.

Fig. 2.

Maintenance of spermatogenesis following the subcutaneous transplantation of a zebrafish testis. The zebrafish testis is organized into lobules surrounded by a basement membrane (dotted line in A–C). A) Section of a juvenile testis stained with hematoxylin and eosin. The testes of TM or Tu/AB juvenile fish were transplanted beneath the skin of TM strain recipients. Four weeks after transplantation, the grafted testes were removed and analyzed histologically. B and E) Serial sections of the grafted TM strain testis stained with hematoxylin and eosin (B) and anti-BrdU antibodies (E). C and F) Serial sections of the recipient testis stained with hematoxylin and eosin (C) and anti-BrdU antibodies (F). Note that type A spermatogonia (arrowheads and insets of E and F) contain the SSC population and that type B spermatogonia (arrows in E and F) incorporate BrdU in both the recipient testis and grafted testis. D) Section of a grafted Tu/AB testis stained with hematoxylin and eosin. Note that this transplanted tissue had degraded in the TM strain. G) Average BrdU labeling index of the cysts in grafted and recipient testes. Error bars represent the SEM. AG, type A spermatogonium; BG, type B spermatogonia; PC, primary spermatocytes; ST, spermatids; S, sperm. Bars = 30 μm.

Proliferation and Differentiation of Spermatogonia from a Testicular Cell Aggregate after Transplantation

We next examined whether the dissociated zebrafish testicular cells could reconstitute the testicular structure and maintain spermatogenesis in the recipient fish. We used the immature testes of TM juveniles to produce the cell aggregates. In the immature testes, clear lobules containing numerous large cysts were observed (Fig. 3A). After enzymatic dissociation, the immature testicular cells were centrifuged and cultured for 2 d to form aggregates (Fig. 3, B and C). Immediately following aggregation, lobule and cyst structures were not observed (Fig. 3D). We then transplanted 20 aggregates under the abdominal skin of 20 TM juveniles, and 12 grafted aggregates were found to have been maintained in the recipient fish at 3 and 6 mo after transplantation (Table 1). Within these aggregates, clear lobules and numerous cysts comprising spermatogenic cells at various stages were observed (Fig. 3, E and F). In the cyst structure of a teleost, Sertoli cells surround the germ cell clusters [8]. The formation of well-developed cysts suggests that Sertoli cells and germ cells had formed the appropriate arrangements in the transplanted aggregates. In addition, at 6 mo after transplantation, we observed 3β-HSD activity in the interstitial region between the reconstituted lobules that was comparable to that of a normal testis (Fig. 4). This activity is a marker of steroidogenic cells such as the Leydig cells that localize in the interstitial connective tissue around the lobules [21, 22]. Hence, these results indicate that dissociated testicular cells containing germ cells and somatic cells maintain the ability to reconstitute a lobule and cyst structure even in a subcutaneous environment.

Table 1.

Reconstruction of lobules and spermatogenesis in grafted zebrafish testicular cells.

Type of grafted tissue	Recipient	No. of grafts	Rearing period (wk)	No. of grafts recovered from recipients	No. of grafts having testicular structure & spermatogenic cells
TM juvenile testis	TM juvenile	7	4	7 (100%)	7
Tu/AB juvenile testis	TM juvenile	10	4	3 (30%)	0
TM aggregate	TM juvenile	14	12	7 (50%)	7
6	24	5 (83.3%)	5

Fig. 3.

Reconstitution of the zebrafish testicular structure and regeneration of spermatogenesis from dissociated testicular cells. A) Section of a juvenile zebrafish testis stained with hematoxylin and eosin. B) Dissociated testicular cells of a juvenile testis. C and D) A testicular cell aggregate generated from the dissociated testicular cells shown in B (C) and a stained section (D). Note that a lobule and cyst are not detectable in D. E and H) Serial sections of a grafted aggregate at 3 mo after transplantation stained with hematoxylin and eosin (E) and anti-BrdU antibodies (H). Note that the lobule and cyst structures have been reconstituted. F and I) Serial sections of a grafted testicular cell aggregate at 6 mo posttransplantation stained with hematoxylin and eosin (F) and anti-BrdU antibodies (I). Type A spermatogonia (arrowheads in H and I) and type B spermatogonia (arrows in H and I) were labeled with BrdU. G and J) Serial sections of a grafted testicular cell aggregate that has been removed from a recipient fish labeled with BrdU at 6 mo posttransplantation and reared normally for a further 2 wk stained with hematoxylin and eosin (G) and anti-BrdU antibodies (J). Note that BrdU-labeled spermatogonia or primary spermatocytes at 6 mo after transplantation have differentiated into sperm in the reconstituted testis (I and J). AG, type A spermatogonium; BG, type B spermatogonia; PC, primary spermatocytes; S, sperm. Bars = 20 μm (A, D, and E–J), 50 μm (B), and 200 μm (C).

Fig. 4.

Detection of 3β-HSD activity in interstitial cells of a reconstituted zebrafish testis. A and B) Detection of 3β-HSD activity in grafted testicular cell aggregates at 6 mo posttransplantation with (A) or without (B) hematoxylin-and-eosin staining. C and D) Detection of 3β-HSD activity in a section of a normal testis with (C) or without (D) hematoxylin-and-eosin staining. E and F) The negative control of the detection procedure of 3β-HSD activity in section of a normal testis with (E) or without (F) hematoxylin-and-eosin staining. In negative control, the detection was performed without nicotinamide adenine dinucleotide in detection solution. Note that 3β-HSD activity (arrows) can be within the interlobular spaces of both the grafted aggregate (B) and a normal zebrafish testis (D). The lobules are indicated by the dotted line. Bars = 20 μm.

To investigate the proliferative activity of the spermatogonia in the grafted aggregates, BrdU was incorporated over a 3-h incubation just after removal of the aggregates from the recipients. Three or six months after transplantation of the aggregates, BrdU was detected in spermatogonia at all stages, including type A spermatogonia and primary spermatocytes in the aggregates (Fig. 3, H and I). In addition, when we maintained the recipient fish (at 6 mo posttransplantation) in fish water containing 1 mg/ml of BrdU for 24 h and then reared them normally for 2 wk to enable the differentiation of later-stage spermatogonia into sperm [9, 10], BrdU was detected in the sperm (Fig. 3, G and J). This result indicates that the later-stage spermatogonia or primary spermatocytes within the grafts underwent meiosis and spermiogenesis within 2 wk. This evidence of proliferation of spermatogonia and their subsequent differentiation into sperm, even at 3–6 mo posttransplantation, indicates that continuous spermatogenesis occurs (i.e., more than a single round of differentiation from type A spermatogonia to sperm, which takes less than 1 mo, in the reconstituted testis).

Finally, we investigated the fertilization ability of the sperm produced in the grafted aggregates. We removed the grafts at 3 mo posttransplantation and artificially inseminated unfertilized eggs with the sperm therein. Although the fertilization efficiency of the sperm derived from the grafted aggregates was lower than normal, fertilized eggs were successfully obtained (Table 2). Moreover, these embryos completed embryogenesis normally and became fertile adult fish. These results indicate that functional sperm were, indeed, produced in the grafted testicular cell aggregates.

Table 2.

Number of eggs successfully fertilized with sperm taken from a grafted aggregate.

Total no. of eggs	No. of fertilized eggs	No. of normal individuals at
		Hatching	Maturity
140	1	1	1
225	11	7	5
120	1	1	1

Development of Allogeneic Spermatogenesis in Grafted Testicular Cell Aggregates

It has been reported for X. laevis that allogeneic SSCs are capable of self-renewal and differentiation in grafted aggregates of testicular cells [14]. The induction of allogeneic spermatogenesis in such grafts enables the analysis of SSC self-renewal and the production of functional sperm from other lines, transgenic animals, and mutants. We thus investigated whether the spermatogonia from vas::EGFP transgenic fish (Tu/AB), which specifically express EGFP in their germ cells under the control of the vas promoter, could proliferate and differentiate in grafted aggregates of testicular cells. Dissociated testicular cells from vas::EGFP transgenic juvenile fish and TM juveniles were mixed at a 1:10 ratio to form aggregates, which were then transplanted under the abdominal skin of TM juveniles. Immediately after aggregation, a few EGFP-positive cells could be observed as a single cell or a small cell mass (Fig. 5, A and B). At 3 mo after transplantation, numerous EGFP-positive sperm and cysts consisting of various-stage spermatogenic cells were observed in the grafted aggregates (Fig. 5, C and D). Vasa gene is not expressed in haploid spermatids and sperm in zebrafish, but we were able to detect the EGFP signals in the haploid germ cells when we used anti-GFP antibody, probably because of the stability of EGFP protein itself. This result indicates that the allogeneic spermatogonia derived from vas::EGFP transgenic fish were not rejected but, rather, successfully proliferated and differentiated into sperm in aggregates grafted into the TM strain.

Fig. 5.

Continuous spermatogenesis of allogeneic spermatogonia within a reconstituted zebrafish testis. Testicular cells from a vas::EGFP transgenic Tu/AB line were mixed with TM strain testicular cells before forming the aggregate, which was then transplanted subcutaneously. A and B) Serial sections of a testicular cell aggregate before transplantation stained with hematoxylin and eosin (A) and anti-GFP antibodies (B). EGFP-positive cells (arrowheads in B) were observed sporadically. C and D) Serial sections of a grafted testicular cell aggregate at 3 mo posttransplantation stained with hematoxylin and eosin (C) and anti-GFP antibodies (D). Note that the EGFP positive sperm (arrows in D) were observed in the grafted aggregate. The reconstituted lobule is indicated by a dotted line (C). AG, type A spermatogonium; BG, type B spermatogonia; PC, primary spermatocytes; S, sperm. Bars = 20 μm.

Discussion

In our present study, we describe the successful reconstitution of a testis from dissociated testicular cells following aggregation and transplantation under the abdominal skin in zebrafish. Only SSCs maintain continuous spermatogenesis, and it has been shown that N-ethyl-N-nitrosourea-mutagenized progeny can be obtained after 3 wk of the treatment for male zebrafish [23]. In addition, following a sublethal chemical dose that damages zebrafish testis in which only type A spermatogonia have survived, spermatogenesis is restored within 1 mo after this exposure (Saito and Sakai, unpublished results). These findings suggest that less than 1 mo is required for type A spermatogonia to differentiate into sperm in zebrafish. In our current transplantation system, we observed continuous spermatogenesis that produced functional sperm for more than 3 mo after grafting, indicating that SSC self-renewal and differentiation can still occur in a reconstituted zebrafish testis. As far as we know, this is the first report of the successful production of progeny by the sperm in a reconstituted testis from a vertebrate. Similar methodologies that have been reported in X. laevis, mice, and pigs did not yield progeny from viable sperm [14–16]. Furthermore, proliferation and differentiation of allogeneic spermatogonia (vas::EGFP) into sperm was also observed in our reconstituted testis of the TM strain. Our methodology will therefore likely prove useful not only in analyses of the stem cell properties of individual SSCs but also in the production of viable progeny using cultured SSCs or SSCs isolated from sterile mutants that have defects in spermatogenesis caused by testicular somatic cell abnormalities.

We demonstrate that zebrafish testicular cell aggregates, which comprise disordered arrangements of immature testicular cells, can successfully reconstruct appropriate testicular structures, such as lobules, cysts, and interstitial Leydig cells in the graft recipients. On the other hand, we have not thus far succeeded in reconstituting testicular structures from the adult testicular cells (data not shown). In other species, the reconstitution of a testis and the onset of spermatogenesis have also been observed with the use of immature testicular cells but not adult testicular cells [14–16]. Rat immature testicular somatic cells, such as Sertoli cells and Leydig cells, recognize their appropriate position in vitro, and they become mature after transplantation into nude mice [24, 25]. It thus seems that the immature testicular cells have the potential to adopt the correct alignments and form a viable structure, which thereafter allows stem cells to self-renew and differentiate. Therefore, it will be interesting in the future to examine which cell type contributes to reconstituting testicular structure. Cografting of germ cells and somatic cells from donors at different developmental stages after chemical or genetic labeling will lead to elucidating cellular mechanisms for de novo morphogenesis of a specific phase of testis development, as discussed in Rodriguez-Sosa and Dobrinski [26].

To avoid a deleterious response from the recipient immune system against the transplanted cells, we used of TM strain (generations 9–12) fish generated by the mating of sister-brother pairs. When the testis of a vas::EGFP transgenic Tu/AB line was transplanted into TM recipients, all of the grafts degenerated. When we grafted aggregates of TM testicular cells mixed with Tu/AB testicular cells, however, the spermatogonia survived and underwent spermatogenesis even though they were alloantigeneic. As in mammals, allograft rejection in teleosts is also caused by a cell-mediated immune system (mainly T cells) [27, 28]. Because germ cells are surrounded by Sertoli cells in the teleost testis, the allogeneic spermatogonia should be protected from any contact with T cells. Hence, this method allows us to trace the fate of mixed germ cells in the aggregate after genetic labeling, even if germ cells are derived from mutants or transgenics of different lines. It has been reported recently that xenogeneic SSCs colonize and complete spermatogenesis following direct germ cell transplantation into the lobules in Odontesthes bonariensis [12]. In other teleosts, such as rainbow trout and zebrafish, methods to transplant primordial germ cells (PGCs) into embryos have already been established, and xenogeneic PGC transplantations between salmonids or between zebrafish and pearl danio (Danio albolineatus) have been reported [29–31]. Our present technique could also provide an opportunity to induce spermatogenesis from the xenogeneic SSCs of immature fish of endangered species related to zebrafish.

In mice, it has been recently reported that undifferentiated type A spermatogonia (from isolated single type A spermatogonia to 4–16 cells connected) exhibit differing degrees of stem cell potential, allowing these cells to be referred to as actual or potential stem cells [7]. The expression patterns of genes such as Nanos2, Gfra1, Nanos3, and Neurog3 (neurogenin 3) also differ in these cells [32, 33]. In zebrafish, morphological analyses have revealed that type A spermatogonia, surrounded by Sertoli cells as a single cell, can be divided into two subgroups [9]. In addition, we have also recently noted the existence of quiescent cells among the type A spermatogonia (Saito and Sakai, unpublished results). Hence, we speculate that the subset of type A spermatogonia with stem cell abilities will be identified in the future through a combination of our present reconstitution system and cell sorting techniques. Furthermore, our current technique has great utility as an assay to identify the stem cell ability of in vitro-cultured spermatogonia. In mice, targeted gene knockouts have been produced through a combination of SSC culture and the transplantation of genetically modified SSCs to induce their differentiation into functional sperm [6]. Assay systems for stem cell ability will also help to improve the zebrafish male germ cell culture systems, because currently, spermatogonia but not SSCs can be grown [10]. The combination of effective zebrafish SSC culture methods and techniques to produce functional sperm from SSCs will also makes it possible to establish a germ cell-mediated gene transfer system in zebrafish.

Note Added in Proof

The TM strain was lost due to embryonic lethality at the 13 generation.

Acknowledgments

We thank Ms. M. Kojima, R. Maeda, and N. Suzuki for maintaining the zebrafish stocks.

Author notes

¹

Supported by the Program for the Promotion of Basic Research Activities for Innovative Biosciences (BRAIN) from the Bio-Oriented Technology Research Advancement Institution of Japan, and partially by a Grant-in-Aid (to N.S.) and a Grant-in Aid for Young Scientists (to M.S.) from the Ministry of Education, Culture, Sport, Science, and Technology, Japan.