|Home | About | Journals | Submit | Contact Us | Français|
After their arrival in the fetal gonad, mammalian germ cells express E-cadherin and are found in large clusters, similar to germ cell cysts in Drosophila. In Drosophila, germ cells in cysts are connected by ring canals. Several molecular components of intercellular bridges in mammalian cells have been identified, including TEX14, a protein required for the stabilization of intercellular bridges, and several associated proteins that are components of the cytokinesis complex. This has led to the hypothesis that germ cell clusters in the mammalian gonad arise through incomplete cell divisions. We tested this hypothesis by generating chimeras between GFP-positive and GFP-negative mice. We show that germ cell clusters in the fetal gonad arise through aggregation as well as cell division. Intercellular bridges, however, are likely restricted to cells of the same genotype.
In mammals, germ cells migrate to the gonad from the base of the allantois. From the time germ cells leave the allantois, they travel as individual cells often connected by long cytoplasmic filaments (Gomperts et al., 1994). However, when germ cells colonize the gonad, they form large clusters of eight or more cells. During this period, germ cells are also undergoing cell divisions (Pepling and Spradling, 1998). Whether the clusters of germ cells in the gonad arise through aggregation or cell division is unclear.
Imaging analyses suggested that filopodial connections between migrating germ cells might lead to their aggregation once they reach the genital ridge (Gomperts et al., 1994; Bendel-Stenzel et al., 2000). When cells from the hindgut mesentery and associated mesonephroi were disaggregated and plated on STO feeder cells, germ cells formed clusters. Furthermore, when half of the germ cells in the cultures were labeled with rhodamine-conjugated dye, clusters were composed of labeled and unlabeled cells (Gomperts et al., 1994), compatible with the idea that clusters can arise through aggregation, rather than solely through cell division. Further investigations showed that blocking E-cadherin, which is expressed by germ cells upon arrival in the gonad (Di Carlo and De Felici, 2000), led to a disruption in germ cell aggregation in slice cultures and in dissociated reaggregation assays (Bendel-Stenzel et al., 2000). However, it was not clear that these in vitro assays reflected aggregation properties in vivo.
Other investigators have noticed the similarity between premeiotic germ cell clusters in vertebrates and those found in invertebrates such as Drosophila. In Drosophila, germ cell cysts contain 16 cells that arise through stereotypic divisions. Divisions in the cyst are incomplete and result in ring canals that form cytoplasmic bridges between all 16 cells in the cluster (reviewed by Pepling et al., 1999). Similar intercellular bridges have been seen in electron microscopy studies of mice, rabbits, rats, hamsters, and humans (Fawcett, 1961; Gondos, 1973; Gondos and Conner, 1973; Pepling and Spradling, 1998). Pepling and Spradling further analyzed these structures in the mouse ovary and showed that intercellular bridges resembling Drosophila ring canals are present between germ cells in the ovary between embryonic day E11.5 and E17.5. Germ cells in clusters also show synchronous cell divisions, which are characteristic of germ cell cysts in Drosophila and medaka (Pepling et al., 1999; Nakamura et al., 2010). In addition, clusters tend to be comprised of an even number of cells, consistent with the idea that clusters arise through cell division rather than random aggregation (Pepling and Spradling, 1998).
In Drosophila, only one of the germ cells in the cluster becomes an oocyte, while the others develop into nurse cells that nourish the oocyte by a directional transport of mitochondria and specific mRNAs through the cytoplasmic bridges (de Cuevas et al., 1997). In mice, there is at least a 3-fold reduction in the number of germ cells in the ovary between the time that their numbers reach a maxima and the time of follicle formation just after birth (Pepling and Spradling, 2001; McClellan et al., 2003). There is currently no clear explanation for this high level of germ cell atresia during fetal ovary development. However, the identification of endoplasmic reticulum and mitochondria within intercytoplasmic bridges between adjacent germ cells in mouse ovaries led to the idea that germ cell atresia could be explained by a selective mechanism similar to the one that operates in Drosophila (Pepling and Spradling, 2001). The idea that processes underlying oocyte selection are conserved is appealing.
More recently, molecular components of intercellular bridges in mammalian cells have been identified (reviewed by Greenbaum et al., 2011). The first of these was TEX14, a protein that is required for the stabilization of intercellular bridges and essential for spermatogenesis in males (Greenbaum et al., 2006) but not oogenesis in female mice (Greenbaum et al., 2009). Using a proteomics approach on testis fractions enriched for cytoplasmic bridges, several components of the cytokinesis complex were identified and shown to colocalize with TEX14 to cytoplasmic bridges during early spermatogenesis, including three components of the midbody, the mitotic kinesin-like protein 1 (MKLP1/KIF23), RACGAP1 (MgcRacGap), and centrosomal protein 55 (CEP55), as well as several septin proteins (Greenbaum et al., 2007; Iwamori et al., 2010). These elegant studies provided clear evidence that cytoplasmic bridges can form as the result of incomplete cell divisions during germ cell development.
To investigate whether germ cell aggregates arise only through clonal divisions in mice, we produced chimeras between GFP-positive and GFP-negative embryos. Our results indicate that while fetal germ cell clusters arise through both aggregation and clonal divisions, we detected clear evidence of bridges only between germ cells of the same genotype.
Chimeric embryos were allowed to develop to E11.5, E12.5, or E13.5 before dissection. The sex of the embryos was unknown at the time of assembly. Therefore, chimeras could be XX↔XX, XY↔XY, or XX↔XY. To determine the XX or XY constitution of the chimeras, the fetal liver was isolated from embryos at the time of dissection. Fragments of the fetal liver were reserved for immunocytochemistry and for an unsorted control (UN). Cells in the remainder of the fetal liver were dissociated, and fluorescence activated cell sorting (FACS) was used to isolate pure populations of GFP-positive (GP) and GFP-negative (GN) cells. PCR analysis of these two populations was performed with primers that distinguish the X and Y chromosomes (Fig. 1A,D,G). This analysis determined the sex of the GFP-positive and GFP-negative components of each chimera.
To confirm the PCR results, the sex was retrospectively determined by staining chimeric tissue (limb, liver or intestine) with antibodies against histone H3 lysine 27 (H3K27), which strongly labels the inactive X chromosome in XX cells as well as heterochromatin throughout the nucleus (Rougeulle et al., 2004), and co-imaging with GFP. This assay confirmed the PCR results in all chimeras tested (Fig. 1B,E,H).
Immunocytochemistry was performed on whole-mount fetal gonads in XX↔XX, XY↔XY, and XX↔XY chimeras using an antibody against E-cadherin. In most samples tested at E11.5–E13.5, clusters of germ cells contained both GFP-positive and GFP-negative germ cells (Fig. 1C,F,I), which could be detected in Z-planes of the cluster (Fig. 2A). E-cadherin-positive boundaries were detected between clustered germ cells of different XX and XY genotypes (Figs. 1C,F,I and and2A).2A). This result was confirmed with two alternative germ cell markers, MVH and PECAM-1 (Fig. 2B,C).
Control XX and XY gonads at E11.5 and E13.5 were labeled with antibodies against E-cadherin and TEX14, an element of intercellular bridges in mammalian germ cells. TEX14-positive structures were identified between E-cadherin-positive germ cells in both sexes at E11.5 and E13.5 (Fig. 3A), as reported previously (Greenbaum et al., 2009). The antibody against TEX14 was then used in combination with an antibody against E-cadherin on E12.5 and E13.5 XX↔XY, XX↔XX, and XY↔XY chimeras. The vast majority of TEX14-positive structures fell between germ cells of the same genotype in all chimeras (Fig. 3B). Out of more than 600 structures counted in an analysis of two E12.5 XX↔XY, one XX↔XX, and two XY↔XY chimeras, only six appeared to be positioned between germ cells of different genotypes, and only two of the six were convincing by Z-stack analysis (Fig. 3C). None of the bridges between cells of different genotypes fell in the middle of a definitive, straight border between the two germ cells, as was commonly seen between germ cells of the same genotype. This observation, together with their extremely low frequency of occurrence, suggests that these inter-genotype TEX14-positive structures may represent anomalies or imaging artifacts rather than examples of biologically relevant de novo bridge formation.
Previous studies noted that clusters of embryonic germ cells undergo synchronized cell divisions, a process that may involve communication through intercellular bridges (Pepling and Spradling, 1998). We predicted that if bridges were only present between clonally related germ cells and were required to coordinate the cell cycle, synchronous divisions should be restricted to cells of the same genotype, even within mixed clusters. Metaphase cells were identified by DNA condensation. Five chimeric gonads (two E12.5 XX↔XY, two E12.5 XY↔XY, and one E11.5 XX↔XY) that contained both GFP-positive and GFP-negative germ cells were evaluated in this analysis. We counted fifteen small groups of two or more synchronously dividing germ cells adjacent to non-dividing germ cells of the same or opposite genotype (Fig. 4A,B). In fourteen of these cases, the mitotic germ cells were of the same genotype, suggesting that synchronicity is generally restricted to clonally related cells. In only one instance, we observed four mitotic GFP-positive cells adjacent to three mitotic GFP-negative cells (Fig. 4C). This latter case can be interpreted as either a stochastic coincidence or indicative that synchronicity of germ cell divisions does not depend on intercellular bridges. However, we did not observe any cases in which all germ cells of a large cluster were synchronously dividing: clusters containing mitotic cells always included non-dividing cells as well. In fact, in fifteen clusters, a single mitotic germ cell was surrounded by non-mitotic cells of the same or opposite genotype (non-synchronous division). This pattern may reflect the aggregative nature of the clusters, in which not all cells of a given genotype within a cluster are clonally related or linked by intercellular bridges.
These results clearly demonstrate that germ cell clusters do not arise solely through incomplete divisions, but must also form through aggregation. Female germ cells are known to form cyst-like structures in the fetal mouse ovary and undergo synchronous cell divisions, similar to Drosophila germ cell cysts, which arise through incomplete divisions (Pepling and Spradling, 1998). Our analysis determined that germ cell clusters in the fetal gonad are not uniformly clonal. Germ cell clusters that contain germ cells of different genotypes (i.e., GFP-positive and GFP-negative) cannot have arisen by cell division. The mechanisms that promote germ cell aggregation as germ cells arrive in the fetal mouse gonad are presently unknown, although they likely involve E-cadherin.
Intercellular bridges within the mixed germ cell clusters were frequently observed between cells of the same genotype and only rarely found between cells of different genotypes (2/600 bridges), suggesting that if bridges actually occur between non-clonal cells, they are unlikely to have biological relevance within germ cell aggregates in vivo. This result supports previous work showing that intercellular bridges between mammalian spermatogonia contain markers of the cleavage furrow and likely form as a consequence of incomplete cell division (Burgos and Fawcett, 1955; Braun et al., 1989; Greenbaum et al., 2007; Greenbaum et al., 2011). Within the mixed germ cell clusters, we observed many synchronous cell divisions among germ cells of the same genotype but only one instance of synchronicity among germ cells of different genotypes. Our sample size for this analysis was small, but the cell division patterns observed are consistent with the idea that factors passed through intercellular bridges may coordinate the cell cycle across germ cell clusters. However, we cannot exclude the possibility that germ cell divisions are coordinated by signals from their local environment.
Bridges were present between germ cells in both the testis and ovary from E11.5 to E13.5, in accord with previous observations (Pepling and Spradling, 1998). Maintenance of these structures suggests that they could play an important role in germ cell biology. Consistent with this prediction, elimination of TEX14 caused spermatogenic arrest at the first meiotic division, resulting in male sterility (Greenbaum et al., 2006). Although loss of TEX14 also eliminated bridges at all stages of ovarian development and led to a reduced number of oocytes in the neonatal ovary, it did not affect fertility in females (Greenbaum et al., 2009). However, regulated disassembly of intercellular bridges may be involved in oocyte nest breakdown and follicle formation in the postnatal ovary, as a correlation was noted between the presence of multioocyte follicles and the persistence of intercellular bridges in pups exposed to phytoestrogens (Jefferson et al., 2006). Whether intercellular bridges play a role in coordinating germ cell development at fetal stages is currently unknown. However, clustering of germ cells in the fetal gonad does not depend on clonal cell divisions or on intercellular bridges connecting all cells in the cluster.
To test whether germ cell clusters arise through incomplete divisions or through a process of aggregation, we examined germ cell clusters in chimeric mice. Chimeras were produced between embryos derived from wild type CD1 × CD1 matings and embryos derived from matings between CD1 and mice expressing EGFP from the β-actin promoter and CMV intermediate early enhancer (FVB.Cg-Tg(GFPU)5Nagy/J (EGFP)). Embryos were flushed from the oviducts of pregnant females on embryonic day E2.5. The zona pellucida was removed, and embryos were assembled in pairs (one from each mating) in shallow wells, cultured overnight to the blastocyst stage, then transferred to E2.5 pseudopregnant female hosts (Nagy et al., 2003). All mice were housed in accordance with NIH guidelines, and experiments were conducted with the approval of the Duke University Medical Center IACUC.
Liver tissue was incubated in 1 ml of Trypsin-EDTA (1X) at 37°C for 15–20 minutes and passed through a 27.5 gauge syringe four times to dissociate the tissue and separate the clumps of cells. Cells were centrifuged at 5000 rpm for 3 minutes, washed two times in PBS, resuspended in 500 μl of PBS, placed on the filter of a 5-ml glass conical tube (BD Falcon REF 352235), and briefly centrifuged. GFP-positive and negative cells were sorted using the Beckman-Coulter MoFlo or FACSVantage Flow Cytometry Shared Resource (Duke Comprehensive Cancer Center).
DNA was extracted from GFP-positive and GFP-negative cells and analyzed by PCR for the presence of the X and Y chromosomes as previously described (Mroz et al., 1999).
Samples were fixed in 4% paraformaldehyde, and immunocytochemistry was performed on whole-mount tissue by methods published in detail elsewhere (Tang et al., 2008) using the following primary antibodies: rat anti-mouse E-cadherin (Zymed; 1:200 or Invitrogen 131900; 1:250), rabbit anti-H3K27 (Millipore 07-449; 1:150), guinea pig or goat anti-TEX14 (a kind gift from Martin Matzuk; 1:1000), rabbit anti-PECAM-1 (BD 553370; 1:500), chicken anti-GFP (Aves lab GFP1020; 1:500), and rabbit anti-mouse VASA homolog (MVH; Abcam ab13840; 1:500). Primary antibody staining was revealed by Cy3 donkey anti-rabbit, DyLight 488 donkey anti-chicken, Cy3 donkey anti-guinea pig, Cy3 donkey anti-rat (Jackson Immunoresearch), and Alexa Fluor 647 donkey anti-goat (Invitrogen) secondary antibodies. Nuclei were stained with DAPI as needed. Samples were mounted in DABCO as described (Karl and Capel, 1998) and imaged on a Zeiss LSM510 or LSM710 confocal microscope.
Thanks to members of the laboratory for their assistance with the manuscript. We are grateful to Martin Matzuk for providing the antibody against TEX14. This work was funded by a grant to BC from NICHD (HD39963).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.