Conservation of spermatogonial stem cell self-renewal signaling between mouse and rat

Abstract

Self-renewal of spermatogonial stem cells (SSCs) is the foundation for maintenance of spermatogenesis throughout life in males and for continuation of a species. The molecular mechanism underlying stem cell self-renewal is a fundamental question in stem cell biology. Recently, we identified growth factors necessary for self-renewal of mouse SSCs and established a serum-free culture system for their proliferation in vitro. To determine whether the stimulatory signals for SSC replication are conserved among different species, we extended the culture system to rat SSCs. Initially, a method to assess in vitro expansion of SSCs was developed by using flow cytometric analysis, and, subsequently, we found that a combination of glial cell line-derived neurotrophic factor, soluble glial cell line-derived neurotrophic factor-family receptor α-1 and basic fibroblast growth factor supports proliferation of rat SSCs. When cultured with the three factors, stem cells proliferated continuously for >7 months, and transplantation of the cultured SSCs to recipient rats generated donor stem cell-derived progeny, demonstrating that the cultured stem cells are normal. The growth factor requirement for replication of rat SSCs is identical to that of mouse; therefore, the signaling factors for SSC self-renewal are conserved in these two species. Because SSCs from many mammals, including human, can replicate in mouse seminiferous tubules after transplantation, the growth factors required for SSC self-renewal may be conserved among many different species. Furthermore, development of a long-term culture system for rat SSCs has established a foundation for germ-line modification of the rat by gene targeting technology.

Stem cells are the foundation for body development and maintenance of many adult tissues. Despite their importance, it has been difficult to study and understand the signaling pathways that regulate stem cell fate, i.e., self-renewal and differentiation. In part, this problem results from the difficulty in developing in vitro systems that allow examination of individual environmental cues and growth factors responsible for differential signaling and responses. In addition, only three adult stem cells, hematopoietic, epidermal, and spermatogonial, have a functional assay that results in complete regeneration of the dependent tissue (1-4). Most other adult stem cells (e.g., neural, muscle, and cardiac) depend on in vivo production of one or more cell types from putative stem cells to confirm the presence of stem cell activity. The absence of well defined and unequivocal phenotypic, biochemical, or morphological markers compounds the difficulty in identifying stem cells, studying their behavior, and determining critical signaling events. However, recent work on spermatogonial stem cells (SSCs) has established a foundation that facilitates examination of signaling in these adult stem cells.

A transplantation assay was developed for the mouse in which cell suspensions are introduced into the seminiferous tubules of a recipient mouse, and spermatogonial stem cell (SSC) activity is identified by the development of colonies of spermatogenesis in the recipient testes, with each colony representing the clonal expansion of an individual stem cell (2, 3, 5). The reconstitution of spermatogenesis in the tubules and production of spermatozoa with donor haplotype represents a qualitative and quantitative assay of stem cell activity (3). This spermatogonial stem cell assay has been used to study a range of characteristics of the stem cell, one of the most important of which is surface phenotype, thereby providing a mechanism for stem cell identification and enrichment. Experiments have demonstrated that Thy-1 is a unique marker for mouse SSC, and the surface phenotype is MHC-1^- Thy-1^lo/+ c-Kit^-, αv-integrin^-/dim α6-integrin⁺ at all postnatal ages (6, 7). Using these surface markers, it is possible to produce testis cell populations that are highly enriched for stem cells (≈1 stem cell in 15 total cells) by using FACS or magnetic-activated cell sorting (MACS) (6, 7). Similar studies in rat have indicated that Thy-1 is also a SSC marker in this species (8). In addition, the rat stem cell is EpCAM⁺ and β3-integrin^- in the 8- to 12-day-old rat. Using cell surface markers, populations of rat testis cells containing 1-2 stem cells in 15 total cells can be generated. These cell populations enriched for SSCs are extremely valuable in combination with the transplantation assay to study biological characteristics of stem cells.

An important use for relatively pure populations of stem cells is to develop culture conditions that allow an increase in stem cell number, which provides a system to test the effect of environmental cues, including growth factors, on signaling that regulates fate determination of the stem cell. Early studies suggested that contaminating testis somatic cells were not conducive to stem cell maintenance in vitro. Moreover, growth of contaminating testis somatic cells is stimulated by FBS, which is present in many culture media. In fact, FBS severely decreased proliferation of mouse SSCs when cultured with testis somatic feeder cells (6). Taking advantage of purified stem cell populations and a defined serum-free medium, mouse SSC were recently cultured for >6 months (9). The cells double every 5-6 days and, when transplanted to recipient males, result in spermatogenic colonies and reestablish fertility. Because the cell population was highly enriched for stem cells and a defined medium was used, it was possible to identify the growth factor requirements for continuous self-renewal as glial cell line-derived neurotrophic factor (GDNF), GDNF-family receptor α-1 (GFRα1) and basic fibroblast growth factor (bFGF). Other workers have obtained in vitro proliferation of mouse SSC by using less defined media with FBS and a mixture of growth factors, with or without GDNF (10, 11). Because the SSCs of many species replicate in the seminiferous tubules of the mouse, we hypothesized that GDNF, GFRα1, and bFGF, as identified in mice, also would support rat SSC self-renewal in vitro. We describe here experiments that establish a culture system for rat SSCs, which allowed identification of growth factors and signaling mechanisms necessary for self-renewal in this species.

Materials and Methods

Donor and Recipient Animals. Donor testis cells were obtained from Sprague-Dawley (S/D) rat pups (8-12 days postpartum, dpp; day of birth is 0 dpp) carrying a fusion transgene composed of the mouse metallothionein I (MT) promoter driving the expression of the Escherichia coli lacZ structural gene (MT lacZ) (12, 13). Donor-derived spermatogenesis is identified after transplantation into recipient seminiferous tubules by staining with the β-gal substrate, 5-bromo-4-choloro-3-indolyl β-d-galactoside (X-gal). Busulfan-treated NCr nude (nu/nu, Taconic Farms) male mice and S/D male rat pups were used as recipients for transplantation (13). For details, see Supporting Text, which is published as supporting information on the PNAS web site.

Cell Culture. The culture system for SSCs consisted of serum-free medium and mitotically inactivated STO (SIM mouse embryo-derived thioguanine and ouabain resistant) cell feeders (≈5 × 10⁴ cells per cm²) as described in refs. 6 and 9. STO cells (STO SNL76/6 cells) were obtained from A. Bradley (The Wellcome Trust Sanger Institute, London). The serum-free medium for SSCs consisted of minimum essential medium α (Invitrogen catalog no. 12561), BSA (MP Biomedicals catalog no. 810661), and a supplement (Supporting Text; see also Table 2, which is published as supporting information on the PNAS web site). Testis cells enriched for SSCs were prepared from pups by MACS with anti-EpCAM antibody (GZ1; PickCell Laboratories, 1:200 dilution) and magnetic microbeads conjugated to anti-mouse IgG antibody (Miltenyi Biotec, 1:5 dilution) (6). The enriched rat SSCs were cultured on STO feeders in wells of a 12-well plate at densities of 1-2 × 10⁵ cells per well in 1.5 ml of serum-free medium with or without growth factors as indicated. Growth factors used for long-term culture were human GDNF (R & D Systems), rat GFRα1-Fc fusion protein (GFRα1, R & D Systems), human bFGF (BD Biosciences), and mouse leukemia inhibitory factor (LIF, Chemicon). All cultures were maintained at 37°C in a humidified 5% CO₂ atmosphere or a 5% CO₂/5% O₂/90% N₂ atmosphere. The medium was changed every 2-3 days. For further details about culture experiments and characterization of cultured germ cells, see Supporting Text.

Results

Growth Factors Required for Rat Spermatogonial Stem Cell Self-Renewal Are Identical to Those for Mouse. Previous studies of mouse SSC cultures indicated that a testis cell population from pups (4-8 dpp), highly enriched for SSCs, provided the best starting material to study the growth factors for stem cell self-renewal (9). In the rat pup (8 dpp), ≈8% of testis cells are EpCAM⁺, and 90% of the SSCs in the testis are in this fraction (8). One in 8.5 EpCAM⁺ cells is a stem cell, which represents the most concentrated source of stem cells readily available from the rat. Therefore, to obtain an enriched rat SSC population for culture, EpCAM⁺ cells were isolated by using MACS with anti-EpCAM antibody (MACS EpCAM⁺ cells) from MT lacZ S/D rat pups (8-12 dpp), which express the lacZ transgene in differentiating germ cells but not spermatogonia. When examined by flow cytometric analysis (FCA), the MACS EpCAM⁺ cells were highly uniform for forward scatter (FSc; an indicator of cell size) and side scatter (SSc; a measure of cell structural complexity) (Fig. 1A Right). Approximately 96% of the live cell population (propidium iodide^-; PI^-) was in gate 1 (G1; Fig. 1 A Right). The G1 cells uniformly express EpCAM (see below), allowing the establishment of a convenient gating strategy by FSc and SSc to identify the cell population containing rat SSCs. Because studies with mouse SSCs demonstrated that the stem cell surface phenotype, including FSc and SSc, remains constant during long-term culture (9), we used these parameters and the gating strategy to assess the degree of in vitro proliferation of rat SSCs and designated G1 as the “stem cell containing gate” for initial analysis although closely related non-stem cells are also likely in this gate.

Fig. 1.

Patterns of FSc and SSc by FCA for fresh MACS EpCAM⁺ and 1-week-cultured MACS EpCAM⁺ cells. (A) Fresh unfractionated rat pup testis cells and MACS EpCAM⁺ cells. The MACS EpCAM⁺ cells are uniform for FSc and SSc. Almost all cells (96.3 ± 0.3%; mean ± SEM, n = 3) in the MACS EpCAM⁺ cell population are in G1, designated the stem cell (containing) gate. Before analysis, cells were stained with PI, and only PI^- cells (live cells) were analyzed. (B) Effect of growth factors on germ cell clump formation. MACS EpCAM⁺ cells (10⁵ cells per well of a 12-well plate) were cultured on STO feeders in mouse serum-free culture medium supplemented with various combinations of growth factors (GDNF, GFRα1, bFGF, and LIF). At 7 days in culture, all cells in the wells were harvested and analyzed by FCA. The cell number in G1, the stem cell gate, was as follows: GDNF = 1.9 ± 0.3 × 10⁴, GDNF + GFRα1 = 6.4 ± 0.9 × 10⁴, GDNF + bFGF = 5.9 ± 0.8 × 10⁴, GDNF + LIF = 2.1 ± 0.3 × 10⁴, GDNF + GFRα1 + bFGF = 10.0 ± 1.4 × 10⁴, and GDNF + GFRα1 + bFGF + LIF = 11.5 ± 2.1 × 10⁴ (mean ± SEM, n = 5). The FSc/SSc patterns for GDNF + GFRα1, GDNF + bFGF, GDNF + LIF, and GDNF + GFRα1 + bFGF are not shown. When 10⁵ cells were placed in wells, the cell number found by FCA in comparable samples in the stem cell gate of fresh MACS EpCAM⁺ cells was 8.2 ± 0.4 × 10⁴ (mean ± SEM, n = 3, 82% of 10⁵ cells). The 18% not in the gate represent cells outside the gate (see A) and PI⁺ cells detected by FCA but not trypan blue staining used to establish the 10⁵ live cells for culture. FCA always identifies more PI⁺ putative dead cells than trypan blue staining. The number of cells in the stem cell gate cultured with four factors is significantly higher than that of GDNF alone or the two-factor combinations (P < 0.01). There is no significant difference between three-factor and four-factor combinations.

Because we hypothesized that mouse and rat SSCs would need similar growth factors, we first tested the effect of various single growth factors on rat SSC proliferation in vitro, including GDNF, GFRα1, and bFGF, that support mouse SSC self-renewal. Eight growth factors, GDNF (40 ng/ml), GFRα1 (300 ng/ml), bFGF (1 ng/ml), LIF (10³ units/ml), stem cell factor (SCF; 10 ng/ml), EGF (10 ng/ml), insulin-like growth factor-1 (10 ng/ml), and Noggin (300 ng/ml), were examined based on our previous studies (9). MACS EpCAM⁺ cells were placed in mouse serum-free culture medium on STO feeders with individual growth factors and cultured for 1 week. Within 1 week of culture, we observed germ cell clumps, which were similar to mouse SSCs in culture (9) only in medium supplemented with GDNF or bFGF (Fig. 4, which is published as supporting information on the PNAS web site). However, GDNF was obviously better than bFGF, because cell clumps in the bFGF culture were small and chains of germ cells were present indicating differentiation (Fig. 4). In addition, FCA showed that the GDNF cultures contained twice the number of cells in the stem cell gate as the bFGF culture (data not shown). This result suggested that the growth factors necessary for replication of SSCs are conserved between the rat and mouse. Because GFRα1 and bFGF show a synergistic effect on mouse SSC self-renewal in the presence of GDNF (9), we next examined the effects of combinations of GDNF, GFRα1, and bFGF on rat SSCs in culture. Although we did not see an obvious effect of LIF on rat SSC proliferation, LIF was examined again because it has been demonstrated to be essential for mouse ES cell self-renewal and supports in vitro proliferation of mouse primordial germ cells (PGCs) (14, 15). In these experiments, 10⁵ MACS EpCAM⁺ cells were placed in individual wells of a 12-well plate on STO feeders, cultured 7 days, and the cells of each well were analyzed by FCA to compare the cell number in the stem cell gate (Fig. 1B). In wells containing GDNF alone, only ≈23% (1.9 × 10⁴/8.2 × 10⁴) of the cells initially placed in culture remained in G1 after 1 week (see Fig. 1B legend). The addition of GFRα1 or bFGF with GDNF resulted in retention of 72-78% of the G1 cells originally placed in culture. The addition of LIF to GDNF did not alter the percent of cells in G1. When GDNF, GFRα1, and bFGF, or these three factors plus LIF were used, the number of cells in G1 was increased 20-40% after 7 days of culture. Thus, three or four growth factors supported proliferation of rat SSCs in vitro significantly better than GDNF alone or GDNF in combination with another factor (Fig. 1B legend). As expected from the difference in cells present in G1, the germ cell clumps were more numerous and larger in the presence of three or four factors than with one or two factors (data not shown). These results demonstrate a dramatic effect of growth factors (GDNF, GFRα1, and bFGF) on rat SSC proliferation. Although the key growth factors essential for germ cell clump formation were determined and a significant increase in cell number in the stem cell gate was obtained, germ cell clumps disappeared after subculturing. To achieve continuously proliferating germ cell clumps in vitro, modifications were made in medium composition, culture environment, and subculture techniques but not growth factors (Supporting Text and Table 2; see also Figs. 5-8, which are published as supporting information on the PNAS web site). The modified culture method also supported proliferation of clump-forming germ cells from Long-Evans rat pups, Fisher 344 rat pups, Brown Norway rat pups, and S/D neonatal and adult rats (data not shown). The modified medium for rat SSCs was designated rat serum-free medium. We also cultured cells in a 5% O₂/5% CO₂/90% N₂ atmosphere because low oxygen concentration benefits mouse embryo development in culture and decreases differentiation of human ES cells (16, 17).

Cultured Rat Spermatogonial Stem Cells Increase in Number. During mouse SSC culture, continuous proliferation of clump-forming cells results in a doubling of SSC activity every ≈5.6 days (9). Using the transplantation assay for SSCs, we investigated the doubling time of cultured rat SSCs. At the beginning of culture, 10⁵ MACS EpCAM⁺ cells from MT lacZ transgenic rat pups were placed in each well of a 12-well plate on STO feeders in rat serum-free medium with growth factors and 5% oxygen atmosphere. Two separate experiments were performed in which GDNF, GFRα1, bFGF, and LIF were used, and one experiment was performed in which LIF was omitted from the growth factor combination. The wells were subcultured at 6- to 8-day intervals with a 1:1-2 dilution by pipetting the clumps free of the feeder layer and digestion with 0.01% trypsin, followed by plating on fresh STO feeders. With this procedure, the clumps were dissociated into mostly single cells, and the majority of contaminating somatic cells appeared to be removed at each subculture. Tightly connected cell clumps re-formed after each passage (Fig. 2A). At 1-month intervals for 5 months, cultured cells were transplanted into the seminiferous tubules of busulfan-treated nude mouse testes to determine the ability of the germ cell clumps to form colonies of spermatogenesis. Before culture, fresh MACS EpCAM⁺ cells were transplanted to establish the stem cell activity in the starting population of cells. Two months after transplantation, the recipient testes were incubated with X-gal, and individual donor cell-derived colonies of spermatogenesis were stained blue (Fig. 2B). Each colony represents the clonal expansion of a single stem cell and differentiating germ cells (18). The number of colonies produced from each experimental group at the individual time points, per 10⁵ cells of MACS EpCAM⁺ cells originally placed in culture, was determined (Fig. 2C). In experiment 3, transplantation to recipient testes was performed only at months 3, 4, and 5. The transplantation assay confirmed that the increase in germ cell clumps reflected an increase in SSCs (Fig. 2C).

Fig. 2.

Expansion of rat SSCs in culture. MACS EpCAM⁺ cells were cultured in rat serum-free medium supplemented with four growth factors (GDNF, GFRα1, bFGF, and LIF) or three growth factors (GDNF, GFRα1, and bFGF) on STO feeders in a 5% oxygen atmosphere. Clump-forming cells were subcultured by pipetting followed by 0.01% trypsin digestion. (A) EpCAM⁺ rat germ cells formed clumps (arrows) after subculturing and continuously proliferated in vitro. (Scale bar: 100 μm.) (B) Macroscopic appearance of recipient testis 2 months after transplantation with 5-month-cultured rat SSCs from MT lacZ rat pup testes. Each blue stained stretch of seminiferous tubule indicates donor cell-derived spermatogenesis from an individual stem cell. (Scale bar: 2 mm.) (C) Fresh EpCAM⁺ cells and cultured cells were transplanted into recipient nude mouse testes. The number of donor-derived spermatogenic colonies per 10⁵ cells originally seeded in culture is shown. Cultured germ cells with four factors were transplanted at 1-month intervals for 5 months. The culture with three factors was transplanted at only 3-5 months. The transplantation assay indicated an increase of rat SSCs in culture with three or four factors for 5 months. Data are shown as means ± SEM for 6-12 recipient testes per time point. Error bars for most points are within the symbol.

In experiments 1 and 2, with four growth factors, the increase in colony number diverged after 2 months, and at 5 months in culture, 10⁵ MACS EpCAM⁺ cells had generated 10.8 ± 2.3 × 10⁶ and 7.1 ± 1.8 × 10⁴ (mean ± SEM, n = 8 and 12 testes, Fig. 2C) colonies, respectively. In experiment 3, with three growth factors after 5 months, 6.4 ± 1.4 × 10⁶ (mean ± SEM, n = 8 testes, Fig. 2C) colonies were produced. Before culture, 10⁵ MACS EpCAM⁺ cells generated 613.8 ± 97.6 (mean ± SEM, n = 8 testes) colonies. Therefore, in experiment 1, stem cell activity increased 1.76 × 10⁴-fold (10.8 × 10⁶/613.8), and a single stem cell produced >17,000 copies in 5 months. In this experiment, stem cell number doubled every 10.6 days (150 days/log₂1.76 × 10⁴). Similar calculations indicated that doubling time was 21.9 days in experiment 2 conducted under identical conditions. In experiment 3, which was maintained with only three growth factors, stem cell expansion was similar to experiment 1 with four growth factors, and stem cell doubling time was 11.2 days. Different experimental groups of stem cells differ in growth characteristics, perhaps related to the type and number of contaminating somatic cells in the original MACS EpCAM⁺ cell population.

Molecular Characteristics in Rat Spermatogonial Stem Cells Proliferating in Vitro. Proliferation of rat germ cell clumps depended on the presence of GDNF. Therefore, we looked for expression of receptor molecules, which are detected in mouse SSCs (9), necessary to bind the growth factor ligand (19). Immunocytochemistry for the c-Ret receptor tyrosine kinase, the major signal transducer for GDNF in other systems, demonstrated strong expression on all cells in germ cell clumps after 10 months' culture (Fig. 3A). In addition, neural cell adhesion molecule (NCAM), which has been identified as an alternative GDNF receptor (20), was expressed by cultured rat SSCs as well (Fig. 3A). Flow cytometry of clump-forming cells also showed significant expression of GFRα1 (Fig. 3B).

Fig. 3.

Rat SSCs express GDNF-receptor molecules and Oct-4 transcriptional factor. (A) Immunocytochemistry for c-Ret receptor tyrosine kinase and NCAM. All clump-forming germ cells express both GDNF receptors. (B) FCA for GFRα1 expression. Cells in the stem cell gate express GFRα1. Closed (red) and open histograms represent GFRα1-stained cells and isotype-stained control cells, respectively. (C) Immunocytochemistry for Oct-4. All germ cell clumps express Oct-4. The Oct-4 immunostaining is localized to the nucleus. (Scale bar: 100 μm.)

Oct-4, a member of the POU transcription factors, is critical for self-renewal and maintenance of pluripotency in ES cells, and the expression is high in the early embryo, PGCs, and mouse SSCs (9, 15, 21). Associated with the differentiation of ES cells, Oct-4 expression is at first decreased and then totally lost. We found that Oct-4 is expressed in rat SSCs, and the staining intensity was similar to that in mouse ES cells (Fig. 3C). Mouse PGCs and ES cells also express high levels of alkaline phosphatase (AP), although its precise role in these stem cells is unclear, and expression is lost when these cells differentiate (22). Mouse SSCs in culture displayed lower AP activity than that of ES cells (9). Likewise, rat SSCs showed low AP activity, as seen in mouse SSCs (data not shown).

Cultured Rat Spermatogonial Stem Cells Transplanted to Recipients Generate Progeny with Donor Haplotype. During long-term culture, germ cell clumps were periodically transplanted to recipient nude mice, and colonies of rat spermatogenesis were produced to measure stem cell proliferation (see Fig. 2). These colonies typically contained spermatozoa (data not shown). However, to confirm further the ability of cultured stem cells to produce functional rat spermatozoa and to assess the contribution of donor stem cells to recipient fertility, germ cell clumps were transplanted to busulfan-treated S/D rat pup testes. Recipients received stem cells cultured for 2, 3, 4, or 7 months (Table 1). At 2 months, cells from two separate culture dates (A2599/A2600 and A2606-2609) were transplanted, and at 3 months, cells from experiment 2 in Fig. 2 were transplanted. Recipients at 4 and 7 months received stem cells from experiment 1 in Fig. 2. Cells transplanted to rat recipients from all four culture dates generated colonies of spermatogenesis (data not shown). Although stem cells from one culture date (2 months, A2599/A2600) did not generate progeny, stem cells from the other three culture dates produced progeny (Table 1). Testes from all infertile recipients were small (267.4 ± 18.7 mg, n = 18; mean ± SEM) but contained blue colonies of spermatogenesis (except the left testis of A2600). Testes from fertile recipients weighed 839.2 ± 59.1 mg (n = 10, data from rat recipients that received cell cultured 2 and 3 months). Small testes in recipients sometimes result from busulfan treatment, and in these males, spermatogenesis is not sufficient to generate progeny (13).

Table 1. Progeny produced from cultured SSCs.

Donor cell culture period, mo	Recipient rat no.	Transgenic/total pups analyzed*	Transplant to first transgenic pup,† days
2	A2599	–	–
	A2600	–	–
	A2606	–	–
	A2607	11/13	87
	A2608	12/13	116
	A2609	3/10	118
3	A2692	–	–
	A2693	7/7	114
	A2694	–	–
	A2695	9/11	97
4	A2685	6/12‡	116
	A2686	–	–
	A2687	5/14	118
	A2688	–	–
	A2689	8/16	160
7	A2856	–	–
	A2857	13/30§	105
	A2858	1/25¶	104
	A2859	–	–

^*The testes of male progeny were evaluated for lacZ expression by staining with X-Gal. Testes are the only tissue that can be reliably stained blue in the MT LacZ transgenic rat model. Transgene presence in some progeny from 4- and 7-month transplantations was detected by PCR with DNA from both males and females

^†All fertile recipients produced transgenic progeny in the first litter. Testes of all infertile recipients contained blue colonies of donor cell-derived spermatogenesis (except the left testis of A2600)

^‡One of 3 by staining with X-Gal and 5 of 9 by PCR

^§Eight of 13 by staining with X-Gal and 5 of 17 by PCR

^¶All data by PCR

Busulfan-treated rat pup recipients that became fertile retained residual endogenous spermatogenesis, which is critical for the maintenance of testis physiological stability in rats (13). Thus, only a fraction of progeny resulted from donor cell-derived spermatozoa. In these experiments, donor cells were obtained from male rats homozygous for the lacZ gene. Therefore, all donor cell-derived spermatozoa will carry the lacZ gene, and eggs fertilized by these spermatozoa will generate transgenic progeny. To assess cultured stem cell contribution to recipient progeny, the testes of newborn pups were incubated with X-Gal, which stained the testes from donor cell-derived male progeny blue. Transgene presence in some progeny from 4- and 7-month transplantations was detected by PCR with DNA from both males and females. Overall, 53% (10 of 19) of recipients were fertile, and 50% (75 of 151) of the progeny tested by staining or PCR carried the transgene, confirming their origin from cultured donor stem cells (Table 1). The first transgenic pups were born 113.5 ± 6.1 days (mean ± SEM, n = 10) after stem cell transplantation. Because the cells transplanted at 7 months came from experiment 1 in Fig. 2, one can estimate the stem cell number generated based on previous calculations. Thus, after 210 days of culture, the initial stem cells would have doubled ≈20 times, and each original stem cell produced >10⁶ copies in this period. These descendant stem cells generated both male and female progeny. There were no obvious defects in progeny from cultured cells, and both male and female progeny were fertile.

Discussion

A culture system that supports continuous replication of rat SSCs in vitro has been developed by using FCA to establish a stem cell gate for assessing the effect of alterations in culture characteristics on SSC number. In initial experiments, a S/D rat testis cell suspension highly enriched for SSCs by MACS selection for EpCAM⁺ cells was placed on STO feeders in mouse serum-free culture medium that was developed for continuous in vitro proliferation of mouse SSCs (9). When MACS EpCAM⁺ cells were cultured in mouse serum-free culture medium with various single growth factors, only GDNF and bFGF supported initiation of germ cell clump formation. FCA showed that the addition of GFRα1 to GDNF significantly increased the cell number in the stem cell gate, indicating that a soluble form of GFRα1 potentiated the GDNF signaling pathway as previously demonstrated for mouse SSCs in culture (9). Furthermore, when bFGF was added to the GDNF/GFRα1 culture condition, an additional increase of the cell number in the stem cell gate was achieved. Thus, three growth factors (GDNF, GFRα1, and bFGF) supported proliferation of rat SSC in vitro, and stem cells continued to self-renew for >7 months after modification of the basal medium and subculture method. These growth factors are the same as found necessary for long-term mouse SSC self-renewal in vitro. Although LIF is essential for mouse ES cell self-renewal (15), both short- and long-term experiments with rat SSCs failed to demonstrate an effect on stem cell proliferation. Moreover, LIF showed no effect on mouse SSC replication in our previous studies (9). Thus, the function of LIF on self-renewal of ES cells and SSCs appears to be different.

It recently has been shown that bFGF strongly promotes self-renewal of undifferentiated human ES cells in vitro (23), and bFGF is a potent growth factor for proliferation of mouse PGCs (24, 25). However, when MACS EpCAM⁺ cells were cultured with bFGF alone, the culture supported only small germ cell clumps. On the other hand, the culture with GDNF alone supported larger clumps and generated more cell clump forming germ cells than that of bFGF alone, suggesting that GDNF is a key signaling molecule for self-renewal of rat SSCs as demonstrated in mouse SSCs (9). There are two identified signal transducers for GDNF, c-Ret, and NCAM (19, 20), and both molecules are expressed by rat and mouse SSCs (9). GFRα1 is thought to be an essential component of the ligand receptor complex for GDNF activation of both c-Ret and NCAM, and GFRα1 was detected on both cultured rat and mouse SSCs (9). Therefore, at least, two receptor signaling pathways, c-Ret and NCAM activated, exist in SSCs. Although NCAM is a homophilic and heterophilic adhesion molecule, it is involved in various biological functions (26). Therefore, NCAM may have more than one role in SSC biology. For example, differential signaling between NCAM and c-Ret may regulate the balance between stem cell self-renewal and differentiation. Interestingly, it has been reported that NCAM-mediated cell-cell adhesion leads to activation of FGF-receptor tyrosine kinase, which is required for NCAM-induced neurogenesis (27). These studies indicate cross-talk occurs between NCAM binding and the FGF signaling pathway. In addition, bFGF promotes GDNF/GFRα1 expression and subsequent activation of c-Ret in neuronal cells (28). Thus, it is very likely that crucial interaction(s) between GDNF and bFGF signaling pathways exists in self-renewal of SSCs. Furthermore, although GFRα1 antibody staining can be detected on cultured rat SSCs and mouse SSCs (9), the addition of a soluble form of GFRα1 showed significant enhancement of clump-forming cell proliferation in our present and earlier studies (9). These results suggest that quantitative and/or qualitative differences exist in cell activation through the membrane bound form or soluble form of GFRα1, which also may be involved in fate determination.

The dependence of rat SSCs on GDNF, GFRα1, and bFGF for continuous proliferation in culture is identical to the situation in the mouse and reflects a remarkable similarity in these two species that diverged 12-24 million years ago (MYA; ref. 29). Previous studies demonstrated that rat SSCs transplanted to mouse seminiferous tubules generated long-term rat spermatogenesis, indicating a conservation of stem cell self-renewal factors, as well as germ cell differentiation factors, between the two species (12). Moreover, SSCs from a wide range of species, including rabbit, pig, baboon, and human, that have diverged from the mouse 50-100 MYA will replicate after transplantation to mouse seminiferous tubules (5). Thus, the self-renewal signaling pathway for these stem cells is also likely to have been conserved and be the same as those identified for the mouse and rat. However, in these more distantly related species, germ cell differentiation factors are not conserved, because stem cells replicate after transplantation to the mouse, but germ cell differentiation and spermatogenesis do not occur (5). Confirming the conservation of SSC self-renewal signaling pathways for these distantly related species will require development of culture conditions specific to each species based on the growth factors necessary for in vitro self-renewal of mouse and rat SSCs.

In a previous study, we found that cultured mouse SSCs share several unique characteristics with ES cells and PGCs (9). Mouse SSCs express Oct-4, normally expressed in pluripotent cells, such as inner cell mass of the blastocyst or ES cells (21). AP activity also is present in mouse SSCs, and is the first indication of PGC differentiation (30). Likewise, cultured rat SSCs express Oct-4 and have AP activity, indicating SSCs, PGCs, and ES cells have metabolic or cell signaling characteristics in common. Although the biological function of AP is not clear, it has been demonstrated that Oct-4 level regulates stem cell fate decisions regarding self-renewal or differentiation in early mouse embryos and ES cells (15). Because Oct-4 also is expressed specifically in undifferentiated spermatogonia in the testis (31), it likely has a similar role in fate determination of SSCs. Because mouse ES cells produce tumors when injected into mice (32, 33), we examined the effect of injecting rat SSCs into nude mice. Unlike mouse ES cells, rat SSCs never generated tumors, teratocarcinomas, or seminomas in immunodeficient mice (i.p. and s.c. for two nude mice, data not shown). This finding is consistent with a previous result from mouse SSCs (9), suggesting that proliferation and differentiation of long-term cultured SSCs is regulated tightly like endogenous SSCs in the testis, despite sharing several important characteristics with ES cells.

Because the rat is a valuable model in many areas of biomedical research, a method to produce targeted modifications in the germ line has been a high priority. Clearly, the ability to continuously culture rat SSCs establishes the necessary system in which to apply homologous recombination techniques to alter specific genes. Using the culture system described, 10⁶ stem cells could be produced in 7 months from a single targeted cell. This number of stem cells transplanted to a recipient male will generate transgenic progeny in 4 months, which can be mated 3 months after birth to generate homozygous mutant offspring. Thus, from the time of homologous recombination in a single stem cell, ≈14 months (7 + 4 + 3 months) are needed to generate homozygous mutant rats for study, which is perhaps twice the time ideally required to generate similar mutant mice. However, the rat SSC system has several important features. First, it is likely to be faster and more efficient than the recently described nuclear transplantation approach (34). Second, the doubling time for in vitro rat SSC replication probably can be reduced from ≈11 days to 5-6 days, the doubling time required for mouse SSCs, because transplanted rat SSCs support donor-derived spermatogonial colony expansion at more than twice the rate found for mouse SSCs (35). Third, the transplantation of genetically modified rat SSCs to a recipient allows the appropriate targeted male and female progeny to be selected at birth (e.g., by DNA analysis), and each will carry the targeted gene in all their germ cells.

A particularly exciting aspect of the development of a culture system that establishes an approach for gene targeting in the rat germ line is the high probability that it now can be extended to other species. The conservation of GDNF signaling in mouse and rat SSCs as the essential pathway to stimulate in vitro self-renewal and the previously demonstrated ability of SSC of many species to proliferate in mouse seminiferous tubules suggest strongly that a similar culture system can be used to obtain continuous proliferation of SSCs of many mammalian species. Development of a robust culture system as now established for the mouse and rat will allow a wide range of genetic modification in these species. Moreover, continuous in vitro proliferation of SSCs of any species lays the foundation for the development of systems to support germ cell differentiation in vitro. Thus, the progressive development of culture systems for SSCs of other mammalian species is imminent and will provide the basis for a wide range of studies on the biology of the stem cell, in vitro differentiation of germ cells, and modification of germ lines.