Effects of W (c-kit) gene mutation on gametogenesis in male mice: agametic tubular segments in W^f/W^f testes

A number of independent mutations exist at the dominant-white spotting locus (W), which is located on chromosome 5 in mice. These mutations are known to produce the pleiotropic effects of sterility, macrocytic anemia, and depletion of melanocytes and mast cells (Russell, 1979; Kitamura et al., 1978; Green, 1990). The mechanisms underlying the depletion of erythrocytes, melanocytes and mast cells have been investigated by transplantation of cells and tissues (McCulloch et al., 1964; Russell and Bernstein, 1968; Kitamura et al., 1978) and absence of these three types of cells is attributable to a defect in their precursor cells. In recent years, W locus was determined to be allelic with the c-kit proto-oncogene (Chabot et al., 1988; Geissler et al., 1988; Tan et al., 1990; Reith et al., 1990; Nocka et al., 1989, 1990), which encodes a transmembrane tyrosine kinase receptor (Yarden et al., 1987; Qiu et al., 1988; Majumder et al., 1988).

Most of the W mutant alleles, such as W, W^v, W^J and W⁴⁴, cause severe impairment of fertility due to the almost absence of germ cells in the gonads when in the homozygous or double-dominant heterozygous states (Coulombre and Russell, 1954; Russell, 1979; Geissler et al., 1981). The germ cell defect is first apparent at day 9 of gestation, and results from a reduction in the proliferative capacity of the primordial germ cells (PGC) and a retardation in their migration from the yolk sac splanchnopleure to the genital ridge (Mintz and Russell, 1957; Mintz, 1957). Moreover, several severe alleles, such as W³⁵, W, W⁴⁰, W⁴² and W⁴³, which are almost homozygous lethal in utero, have deleterious effects on postnatal germ cell differentiation even in the heterozygotes (Geissler et al., 1981).

In contrast, a few alleles permit normal fertility in homozygotes. The W^f allele, which occurred spontaneously in a C3H/He stock, is one of the fertile alleles at the W locus of the mouse (Geissler et al., 1979). Homozygous W^f/W^f mice show very extensive white spotting and have macrocytic anemia, but they are fully fertile in both sexes. However, since further physiological and histological analyses have not been conducted, the possibility of undetected effect(s) of the mutation on gametogenesis cannot be ruled out. In fact, we recently demonstrated a suppression of the regenerative capacity of testicular germ cells in adult W^f/+ and W^f/W^f mice by using the experimental cryptorchidism and its surgical reversal; one of the stressed conditions that makes evident the masked effects of mutations at intact condition (Koshimizu et al., 1991).

In this study, the effect of the W^f mutation on spermatogenesis was examined histologically. In W^f/W^f testes, we found some agametic segments of seminiferous tubules lacking germ cells and a decrease in the number of gonocytes at birth. These results indicate that the W^f allele does affect gametogenesis, at least in the male, although the effect is much less than with other W alleles.

C57BL/6-W^f /w^f and congenie +/+ male mice were used in this study. The mutant stock was originally maintained on a C3H/He background at the Institut Pasteur, Paris (Guenet et al., 1979) and the mutant allele was subsequently backcrossed onto C57BL/6 by Dr Kitamura, Osaka University Medical School, Osaka (Nakayama et al., 1990). Mice have been maintained in our animal facilities and were kept in a controlled environment and provided with a commercial diet and tap water ad libitum.

W^f/W^f mice have normal prenatal and postnatal viabilities as compared to +/+. Testes of W^f/W^f mice were reduced in weight (Mean ± s.e.m., 88.3±4.3 mg in +/+ versus 73.8±2.2 mg in wywf; P<0.01). Nevertheless, almost all of the mutant males showed normal fertility (Koshimizu et al., 1991).

Testes were removed and fixed in Bouin’s solution, embedded in paraffin, serially cut at 5 μm, and stained with hematoxylin and eosin. Serial cross-sections were examined histologically, and each seminiferous tubule was followed and reconstructed under the fight microscope. Identification of spermatogenic cell was according to the criteria of Leblond and Clermont (1952) and Oakberg (1956).

For each parameter, statistical significance of differences between the two genotypes was determined by Student’s r-test or Chi-square test.

The tunica albuginea, rete testis and seminiferous tubules were all well developed in the testes of W^f/W^f mice as in +/+ mice. As shown in Fig. 1, the seminiferous system was composed mainly of gametic tubules with a small number of agametic tubules. The gametic tubules were populated with spermatogenic cells of all stages, indicating normal spermatogenesis and well-established spermatogenic cycle. In these tubules, no difference was found in the tubule diameter and the number of germ cells at each stage, when compared to +/+ seminiferous tubules (data not shown).

In all W^f/W^f mice examined (more than 40 mice), some agametic tubular cross-sections could be observed. Agametic tubules were reduced in diameter and had very few germ cells; those present, either sperm and/or other types of germ cells, in the central part of the lumen seemed to be an overflow from flanking gametic segments (Fig. 1C,D,E). Although the proportion of agametic tubular cross-sections in the largest testicular sections varied among individuals (1% to more than 10%), there was no difference in the frequency of agametic tubules observed among different age groups (Fig. 2). Such agametic tubules were never observed in +/+ mice. The somatic components of the testis, such as Sertoli cells and Leydig cells, showed normal size, shape and number when examined under the light microscope.

We reconstructed the whole seminiferous tubules from complete serial cross-sections under the light microscope, and found that a group of agametic crosssections existing in a limited area was derived from a single tubular segment. The number of agametic segments per testis was variable from one to more than ten (Means±s.e.m., 4.6±0.8). These segments could be found in any part of the testis. Furthermore, all of the agametic segments were connected at both ends with normal gametic tubules that opened into the rete testis (Fig. 3B). We never observed gametic and agametic tubular segments alternately in a single tubule (Fig. 3C), and agametic segments never opened directly into the rete testis (Fig. 3D). ‘Intermediate’ segments, less than 0.2 mm in length, containing a few spermatogonia but no differentiated germ cells were observed in the flanking portion of gametic and agametic segments. The length of agametic segments was estimated for reconstructed tubules of 20-, 40- and 60-day-old mice. The length varied from 0.7 to 6 mm, 0.8 to 14 mm and 0.7 to more than 20 mm, respectively (Fig. 4). Although the maximum length of agametic segments was usually longer in older testes, the minimum length was almost constant (0.7-0.8 mm).

To understand the developmental origin of the agametic segments, we examined perinatal and prepuberal testes of W^f/W^f mice. In seminiferous cords of 1-day-old mice, two different types of cells are observed, large round cells at the center of the lumen and smaller cells arranged at the periphery of the tubules (Fig. 5A). The former are the gonocytes, the progenitors of type A spermatogonia, and the latter are immature Sertoli cells (Clermont and Perey, 1957; Gondos, 1970; Rugh, 1991). Although no difference was detected in the number of immature Sertoli cells per tubule between +/+ and W^f/W^f testes, the number of gonocytes in W^f/W^f testes was significantly (P<0.01) smaller than that in +/+ testes (Table 1, Fig. 5B) and estimated to be approximately 50% of the control value.

In contrast, the first wave of spermatogenesis, characterized by defined accumulation of germ cells at progressively more advanced stages of differentiation (Nebel et al., 1961; Bellv’e et al., 1977; Rugh, 1991), was not impaired by the W^f mutation (Fig. 1C). There was no difference between genotypes in the proportion of differentiated tubules which both have meiotic germ cells in 12-day-old testes and haploid germ cells in 24-day-old testes (Table 2).

In general, most mutations at the W locus affect the hematopoietic stem cells, melanocytes, mast cells and germ cells. However, in some W mutations, such as W^f, W^sh and W⁴¹, fertility is not affected, even in the homozygotes, although severe impairments of pigmentation and/or hematopoiesis are observed (Guenet et al., 1979; Geissler et al., 1981; Lyon and Glenister, 1982). Although several possible explanations for such a non-parallel display of mutant characteristics have been discussed (see Discussion in Reith et al., 1990; Nocka et al., 1990), little information on the effects of these mutations on germ cell development was available. In this study, we demonstrate for the first time that homozygous W^f/W^f mice, even though they are fully fertile, had impaired gametogenesis. In W^f/W^f testes, a small number of agametic segments of seminiferous tubules were observed (Fig. 1). These segments lacked all stages of spermatogenic cells, reflecting a complete absence of stem cells. Furthermore, the number of gonocytes at birth was significantly reduced in mutants (Table 1, Fig. 5), indicating either a reduction in the number of PGC colonizing the W^f/W^f gonad or in their subsequent proliferation and/or survival. It has already been reported that c-kit transcripts are expressed not only in PGC in fetal testis (Orr-Utreger et al., 1990; Keshet et al., 1991; Manova and Bachvarova, 1991) but also in germ cells in the postnatal testis (Manova et al., 1990; Sorrentino et al., 1991). Although cytological evidence has suggested that the W (c-kit) and SI (c-kit ligand, for review, see Witte, 1990) genes play an important role in spermatogenesis even after birth (Nishimune et al., 1980; Tajima et al., 1991; Koshimizu et al., 1991; Sawada et al., 1991; Yoshinaga et al., 1991), the presence of agametic segments of tubules in W^f/W^f seems to result from a prenatal reduction in germ cells and not impaired spermatogenesis after birth. At first, W^f/W^f mice had normal fertility as compared to +/+ and well-established spermatogenesis through most portions of the testis (Fig. 1). Secondly, agametic tubules in W^f/W^f mice were observed at all ages examined and the proportion of these agametic tubules did not increase with age (Fig. 2). Furthermore, no retardation of the first wave of spermatogenesis was observed in prepu-beral W^f/W^f mice (Fig. 1D, Table 2). In addition, we previously reported the normal proliferative activity of type A spermatogonia in this mutant testis (Koshimizu et al., 1991). Judging from these results, it is suggested that W^f mutation, like other W series of mutants, primarily affects the migration and/or proliferation of PGC, resulting in a decrease in the number of stem cells, but hardly impairs spermatogenesis after birth. We have obtained similar results in W^sh/W^sh (data not shown), another fertile W mutant mice (Lyon and Glenister, 1982). These results indicate that the W (c-kit) gene is more indispensable for the development of PGC than that of postnatal germ cells. A similar result was also reported by Baker and McFarland (1988) for the hematopoietic cell lineage in W⁴⁴/W⁴⁴ mice (Geissler et al., 1981) which are nonanemic but stemcell-deficient. During hematopoiesis, an erythroid progenitor for in vitro burst (BFU-E) is not a primary site of mutant gene action and a progenitor less mature than BFU-E is one of the cells most severely affected.

The important conclusion from our study is that the W (c-kit) gene has its primary effect on PGC, but has less effect on more differentiated stages of germ cells. Such phenomenon is not observable in other W mutants, such as W, W^19H, W³⁷ and W⁴², in which c-kit kinase activity is severely reduced (Tan et al., 1990; Reith et al., 1990; Nocka et al., 1989, 1990), because of death in utero or the complete loss of germ cells in homozygous condition. The residual level of c-kit kinase activity present in W^f may be sufficient for postnatal germ cell development, but inadequate for normal development of PGC, melanogenesis and hematopoiesis. In fact, W⁴¹, the phenotypes of which are very similar to W^f (Geissler et al., 1981), has a point mutation within the kinase domain of c-kit that partially impairs the kinase activity in cultured mast cells (Nocka et al., 1990). The molecular characterization of the W^f allele, which is unknown at present, will be useful in elucidating the mechanism of the W (c-kit) gene action on gametogenesis.

We thank Dr Yukihiko Kitamura, Osaka University Medical School, for providing C57BL/6-W^f/W^f mice. This work was supported by a Grant-in-Aid from the Ministry of Education, Science and Culture of Japan.