Research Article

A Vasculature-Associated Niche for Undifferentiated Spermatogonia in the Mouse Testis

See allHide authors and affiliations

Science  21 Sep 2007:
Vol. 317, Issue 5845, pp. 1722-1726
DOI: 10.1126/science.1144885


Mammalian spermatogenesis produces numerous sperm for a long period based on a highly potent stem cell system, which relies on a special microenvironment, or niche, that has not yet been identified. In this study, using time-lapse imaging of green fluorescent protein–labeled undifferentiated spermatogonia (Aundiff) and three-dimensional reconstitution, we revealed a biased localization of Aundiff to the vascular network and accompanying Leydig and other interstitial cells, in intact testes. Differentiating spermatogonia left these niche regions and dispersed throughout the basal compartment of the seminiferous epithelium. Moreover, rearrangement of Aundiff accompanied the vasculature alteration. We propose that the mammalian germline niche is established as a consequence of vasculature pattern formation. This is different from what is observed in Drosophila or Caenorhabditis elegans, which display developmentally specified niche structures within polarized gonads.

Tissue homeostasis is ensured by the normal functioning of stem cells, which are characterized by their ability to self-renew, as well as to differentiate into specialized cell types. Stem cells are tightly linked to their niche or microenvironment, which regulates their behaviors (1, 2). The Drosophila and Caenorhabditis elegans germ lines represent model systems that have provided insights into the stem cell niche. In these organisms, the gonads are polarized: One end opens out of the body wall, and the opposite end harbors a specialized somatic compartment that supports the stem cells. The somatic components of the stem cell niche are developmentally specified, and their elimination causes irreversible stem cell dysfunction (1, 3).

Functional and genetic evidence suggests that the niche is also critical to mammalian spermatogenic stem cells (47); however, the structural basis of this niche is largely unknown. The mammalian testis lacks overall polarity: Seminiferous tubules form loops with both their ends connected to the rete testes, the outlet for sperm from the testis. Stem cells are scattered throughout the loops. The interstitium between the convoluted seminiferous tubules contains the vasculature network and Leydig cells, macrophages, as well as other cells that surround the blood vessels (8, 9) (Fig. 1, A and B, see also fig. S1, A and A′). The basal compartment within the seminiferous tubules (the space between the basement membrane and the Sertoli cell tight junctions) supports all the stages of spermatogonia before meiosis, including the stem cells. However, no specialized niche substructure has been identified (4) (Fig. 1B).

Fig. 1.

Anatomy of the testis and observation of the GFP-labeled spermatogonia. (A and B) Anatomy of the testis and the observation strategy. See text for details. To observe the GFP-labeled Aundiff, a part of a testis of Ngn3/EGFP transgenic mouse (20) was flattened with a cover glass. The GFP-labeled spermatogonia in the outermost tubules were observed under a fluorescence microscope through the tunica albuginea, peritubular myoid cells, and basement membrane. Only the superficial region that covered the basal compartment could be seen (indicated by the upward-pointing arrows). (C and D) (C) A representative tiled image and (C′) a trace of the labeled spermatogonia (green), blood vessels (red), and interstitium (yellow); seminiferous tubule continuity is indicated in blue. (D) and (D′) are magnified images of a part of (C) marked by a rectangle. Labeled Aundiff can be observed, whereas the meshwork of cells weakly positive for GFP in (D) [indicated in gray in (D′)] constitutes the differentiating spermatogonia (diff.). Scale bars, 100 μm.

Unequivocal identification of the spermatogenic stem cells in the mouse testis has also not been achieved. However, a subset of primitive spermatogonia (<1% of total testicular cells) termed “undifferentiated spermatogonia” or “Aundiff” includes stem cells that constitute an as-yet-unidentified subpopulation (10, 11). The Aundiff population persists throughout the seminiferous epithelial cycle, which takes 8.6 days to complete and is divided into stages I to XII (9, 12). The Aundiff population periodically gives rise to the A1 differentiating spermatogonia at a specific stage (VII to VIII) of the cycle (9, 13, 14). A1 spermatogonia subsequently go through six synchronous mitoses [generating A2, A3, A4, intermediate (In), and B spermatogonia, and preleptotene spermatocytes], which are tightly linked to the seminiferous epithelial cycle progression, before undergoing meiosis. In contrast, Aundiff divide asynchronously under a weaker restriction of the cycle (9, 14, 15). The reconstituting colony-forming activity after transplantation is enriched in the Aundiff population (16, 17).

The Aundiff population contains Asingle (As, isolated cells), Apaired (Apr, pairs of interconnected cells), and Aaligned (Aal, chains of 4, 8, 16, or occasionally 32 cells) spermatogonia (9, 13). Visualization of the chains requires whole-mount preparations of the seminiferous tubules, a procedure that disturbs the testicular organization and eliminates the vasculature and cells of the interstitial compartment. In paraffin sections, Aundiff show primitive nuclear morphology lacking prominent heterochromatin condensation; however, distinguishing Aundiff from differentiating A spermatogonia (A1 to A4) is difficult (13). Chiarini-Garcia et al. morphologically identified Aundiff in plastic-embedded thin sections and reported their preference for an area facing the interstitium in the mouse and rat testes (18, 19).

In this study, we analyzed the localization and behaviors of Aundiff in undisturbed testes. For that purpose, Aundiff were labeled with green fluorescent protein (GFP) under the control of the Ngn3 (neurogenin3) regulatory sequence (20, 21). Ngn3-positive spermatogonia constitute a major, although not the entire, fraction of Aundiff and give rise to all of the spermatogenic cells in the mature testis (11, 20, 22).

Seminiferous epithelial stage–related distribution of the labeled spermatogonia. Our strategy for observing the GFP-labeled spermatogonia located just beneath the tunica albuginea, the outer capsule of the testis, is shown (Fig. 1, A and B) and described (21). By this method, the labeled Aundiff (green in Fig. 1, C and D) were detected because of their strong GFP signal, while the background auto-fluorescence allowed us to recognize the position of seminiferous tubules (blue), blood vessels (red), and interstitial cells (yellow). The differentiating progeny of Aundiff that remained in the basal compartment (A1 to B spermatogonia) were also observed within the same two-dimensional image before cells moved to the adluminal compartment and entered meiosis, as long as they retained the residual GFP fluorescence (gray in Fig. 1D).

The GFP-labeled spermatogonia showed variable distribution patterns relative to the local seminiferous epithelial stages that we determined (fig. S1 and Fig. 2). At stages IV to VI, they exhibited a strong preference for the region adjacent to the blood vessels and the interstitium (Fig. 2, B to E). The preference became weaker at stages VII to VIII, when the Aundiff to A1 transition occurs, and was mostly lost in stage IX, when A1 give rise to A2 after the first synchronous division. This is compatible with previous observations on plastic-embedded sections (18, 19). Aundiff showed a strong correlation with intertubular vessels (arterioles and venuoles) that usually accompany interstitial cells, but not with capillaries or vessels without surrounding interstitial cells. Weakly stained differentiating spermatogonia were detectable usually up to A3 or A4 (stages I or II to III), or occasionally up to B at stage VI.

Fig. 2.

Seminiferous epithelial stage–related distribution of the GFP-labeled spermatogonia. (A, A′) Tiled multiple fluorescence images (A) and their traces, similar to Fig. 1C′. The position of the GFP-positive spermatogonia is indicated by dots (A′). Roman numerals indicate the seminiferous epithelial stages of each region, determined as described in fig. S1. Scale bar, 100 μm. (B to E) Quantification of the GFP-labeled spermatogonia located close to (<40 μm) and distant from (>40 μm) the interstitium. A cut-off of 40 μm was selected because it was one-fourth the average width of the observed seminiferous tubule region. The observed seminiferous tubule surfaces were divided: B, stage IV; and C, stages VIII and IX. The GFP-positive spermatogonia in the close (pink) and distant (white) areas were counted. (D) Summary of the GFP-positive spermatogonia density in each area, classified by the seminiferous epithelial stages. (E) Their preference for the close area [(density in close area)/(density in distant area), which was expected to be 1 in the case of random distribution]. Results from 1155 GFP-positive spermatogonia in 77 seminiferous tubule segments are summarized.

Biased localization of Aundiff and their migration during the transition to differentiating spermatogonia. To more comprehensively visualize the locations and movements of spermatogonia during the Aundiff to A1 transition, we performed time-lapse imaging. A representative compilation of images of a continuous tubule segment are shown (Fig. 3), with a portion of them highlighted (Fig. 3B and movie S1). At the beginning of the imaging (around stage VII), labeled spermatogonia clustered at a position adjacent to the vessels and interstitium (star). During ∼30 to 70 hours, two chains of interconnected eight GFP-positive Aundiff (Aal-8, indicated by orange and yellow asterisks in Fig. 3B) migrated out of this region and, thereafter, divided into 16-cell chains with a similar timing (at 76 hours and 73 hours). On the basis of the stage-specific distribution pattern of the GFP-labeled cells (Fig. 2), we determined that these divisions were the A1 to A2 division, i.e., the first synchronous division of the differentiating spermatogonia that occurs at stage IX. Similar behaviors were consistently observed of other spermatogonia undergoing the Aundiff to A1 transition, which led us to conclude that this transition event coincides with migration away from their original position near the intertubular vessels. The migrating spermatogonia divided infrequently [usually >40 hours (Fig. 3B and movie S1), sometimes >70 hours], compatible with previous studies using fixed testes (13, 14). The spermatogonia continued to migrate beyond the A1 to A2 division until they were evenly distributed throughout the basal compartment—this occurred around the A2 to A3 division in stage XI.

Fig. 3.

Time-lapse observation of the GFP-labeled spermatogonia during the transition from undifferentiated to differentiating spermatogonia. (A, and A′) Composition of multiple fluorescence images at the beginning of the imaging (A) that includes a continuous seminiferous tubule segment represented in (A′) (as in Fig. 2A′). Rectangles in (A) indicate the positions of the images in the rest of the figure. Gray ovals in (A′) indicate the sites from which GFP-positive Aundiff migrated during their transition to differentiating spermatogonia. The smaller blue circles indicate the blood vessels branch points, some of which (indicated by asterisks) have branches running inward into the testis (perpendicular to the image plane). Scale bar, 100 μm. (B) Selected frames of the record of the area shown in (A) (also see movie S1). Numbers indicate the approximate elapsed time in hours. Orange and yellow asterisks indicate two syncytial clones that divided during the imaging. Individual cells in these clones are indicated by arrowheads in some frames; some cells of the orange clone are behind a blood vessel in frame 84 hours. Mitoses were observed at 30 hours and 76 hours in the orange clone (4 to 8 and 8 to 16 cells, respectively), and at 73 hours in the yellow clone(8 to 16 cells). Star, a blood vessel branch point, corresponds to the star in (A). (C to F) Typical examples of the regions wherein Aundiff are located before they transit to differentiating spermatogonia. Arrowheads indicate clustered GFP-positive Aundiff. (C) to (E) Magnified images of the areas presented in (A); (F) a different observation.

Gray ovals in Fig. 3A′ indicate the positions from which the labeled Aundiff migrated out and transited into A1. Representative examples with higher magnifications are also shown (Fig. 3, C to F). Within the zone facing the interstitium and blood vessels, the labeled spermatogonia exhibited a further preference for the branch points of the blood vessels, as indicated by blue circles in Fig. 3A′. Characteristically, labeled Aundiff spermatogonia were often found in the inner or outer corner of turning tubules, which were also associated with vascular branching (Fig. 3, E and F). It is noteworthy that such branch points always coincided with abundant interstitium. In some instances, Aundiff in neighboring seminiferous tubules were localized on the opposite side of the branching vessels (Fig. 3, C and D). The majority of such branch points accompanied the Aundiff clusters, although, in some cases, the migration of the Aundiff spermatogonia away from the surface of the exposed seminiferous tubule hid them from view.

Vasculature-related localization of Aundiff in inner seminiferous tubules. The above findings were from the superficial area of the seminiferous tubules that contacts with tunica albuginea, and the GFP-labeling was dependent on the Ngn3-regulatory sequence and might not represent the entire Aundiff population (11, 20). Therefore, to obtain more universal information, we reconstructed three-dimensional images from paraffin-embedded serial sections after cell-type determination by examining nuclear morphology after hematoxylin staining (9) (Fig. 4). Because of the uncertainty in distinguishing Aundiff from A1 to A4 differentiating spermatogonia, we restricted our analysis to stages IV to VIII, stages that contain Aundiff and nascent A1 spermatogonia, but lack more differentiated type-A spermatogonia.

Fig. 4.

Three-dimensional reconstitution of the seminiferous tubule, surrounding vasculature, and Aundiff localization. (A to C) Tracing of the components in each section. (A) A microscopic image of a serial section stained with periodic acid–Schiff and hematoxylin. The seminiferous tubule of interest was traced; its peripheral regions facing the interstitium (yellow), those facing the neighboring seminiferous tubules (blue), and the blood vessels (except capillaries; red) are shown. Traces of seminiferous tubules at stages from IV to VIII show Aundiff and nascent A1 spermatogonia as green dots. Scale bar, 100 μm. (B), traced image and (C), a magnified view of an Aundiff spermatogonium indicated by the arrowhead in (A), with a pale homogeneous staining of the nuclear materials, and its comparison with a differentiating B spermatogonium found in the same section (inset). (D to G) Three-dimensional reconstituted images. (D) and (E) Entire views and [(F) and (G)] closer images with [(E) and (G)] and without [(D) and (F)] blood vessels. Seminiferous epithelial stages are indicated, with a small modulation in their order, as described (asterisks) (9). Arrows in (G) indicate the examples of blood vessel branch points associated with Aundff clustering.

Three-dimensional reconstruction allowed us to follow the convoluted seminiferous tubules and the surrounding vasculature network (Fig. 4, D to G): The yellow stripes and the blue areas represent the regions facing the interstitium and the neighboring tubules, respectively. The Aundiff (green) showed a clear correlation with the yellow stripes running beneath the blood vessels, with an additional preference for the blood vessel branch points (arrows in Fig. 4G). In stages IV to VI, a small number of Aundiff spermatogonia were tightly clustered near the interstitium. In stages VII to VIII, they increased in number and were more dispersed than in stages IV to VI, similar to what we had observed during the Aundiff to A1 transition with live GFP-labeling (Figs. 1, 2, 3).

Rearrangement of Aundiff accompanies the vasculature pattern alteration. To assess the functional link between the vasculature pattern and Aundiff localization, seminiferous tubule fragments from donor mice (marked with ubiquitous GFP) were freed of the attaching vessels and interstitium, then transplanted beneath the tunica albuginea of a recipient testis (Fig. 5). After 3 months, the grafts survived in an unusual coiled configuration, and a significant part of them harbored donor-derived spermatogenesis with multilayered organization and an apparently normal seminiferous epithelial cycle (Fig. 5, A to E). Therefore, normal functioning of the stem cells was restored in the grafts.

Fig. 5.

Localization of spermatogonia in the transplanted seminiferous tubule fragment. (A and B) A GFP-positive seminiferous tubule fragment that survived in the host testis 3 months posttransplantation beneath the tunica albuginea. (A) An overlay of the bright-field and fluorescence images; (B) a closer fluorescence graft image; and (C) a section of the graft, hybridized in situ with the enhanced GFP probe. (D) A hematoxylin-and-eosin–stained section showing the donor seminiferous tubules harboring multiple germ cell layers. e, elongated spermatids; r, round spermatids; c, spermatocytes; g, spermatogonia; and i, interstitium. (E) Spermatogonia with the type-A morphology observed in the graft. (F) A section of the donor seminiferous tubules hybridized with the Ngn3 probe. Arrowhead indicates an Ngn3-positive spermatogonium. (G) A magnified image. Dots, the basement membrane; and i, interstitium. Scale bars, 100 μm. (H to M) Three-dimensional reconstitution of the same graft; [(J) to (M)], at higher magnifications. Most of the green dots represent the positions of spermatogonia with the type-A morphology, whereas some of them were identified on the basis of their Ngn3 expression determined by in situ hybridization [(F) and (G)]. Roman numerals indicate the approximate seminiferous epithelial stages, and arrows indicate the blood vessel branch points associated with Aundiff clusters. The region marked “def” contains defective spermatogenesis.

A three-dimensional reconstruction of a graft (Fig. 5, H to M) is represented in the same manner as in Fig. 4. The vasculature was established in good coordination with the abnormal shape of the graft (Fig. 5, B, H, and I), which suggested that the vasculature pattern was different from that present in the donor testis. Interstitial cells including the Leydig cells were also established surrounding the vasculature (Fig. 5D). The green dots represent the spermatogonia with a primitive morphology common to Aundiff and A1 to A4 spermatogonia (Fig. 5E), some of which were Ngn3-positive (Fig. 5, F and G). If we assume that the seminiferous epithelial cycle is normal in the graft, the green dots represent Aundiff localization before the transition into A1 (stages V to VII). They preferentially localized to interstitial zones, and even to the branch points of the vessels (Fig. 5, J to M). In stages ∼VIII to IX, they were no longer restricted to the interstitium and were also found in regions facing neighboring tubules (Fig. 5, H and I).

Therefore, a normal pattern of Aundiff localization reappeared when the surrounding vasculature and interstitium were reorganized. The intriguing correlation between the vessels, the shape of the graft, and the Aundiff localization suggest that the vasculature and interstitium play a causal role in the Aundiff relocation, although it is not excluded that the Aundiff may have also affected the vasculature to some extent.

A microenvironmental niche for Aundiff linked to the vasculature. In summary, in the mature mouse testis, Aundiff are preferentially located in restricted portions of the basal compartment within the seminiferous tubule, which are adjacent to the blood vessels and interstitium and are particularly close to the blood vessel branch points. On making the transition into differentiating spermatogonia, they migrate out of these areas and disperse over the entire basal compartment of the seminiferous epithelium. These limited regions can be considered the microenvironmental niche for Aundiff. The actual stem cells, which constitute an as-yet-unknown Aundiff subpopulation (11), might also reside in these regions, although it is still possible that the actual stem cells are localized differently and that only their immediate Aundiff progeny prefer this microenvironment. Aundiff spermatogonia do not behave uniformly in these niche regions. In some instances, they migrate out and transit into A1; in other instances, they remain within the niche. This might reflect heterogeneity within the niche environments or within the Aundiff population.

Our study raises the question of why Aundiff localization recapitulates the vasculature pattern. One straightforward possibility is that blood vessels recruit Aundiff. Recently, a “vascular niche” has been described for several other stem and/or progenitor cells (2326). A similar interaction might be possible in the testis. However, unlike cells in the brain, bone marrow, and tumors, spermatogonia are not in direct contact with the vasculature.

Leydig cells are strongly associated, anatomically, with the blood vessels. Moreover, Leydig cells have been also proposed to originate from cells in blood vessels such as smooth muscle cells or pericytes (27). Therefore, the Leydig cells (or other types of interstitial cells) might contribute to the niche microenvironment by producing androgens or other factors. This is compatible with the observation that Aundiff are preferentially associated with blood vessels surrounded by Leydig cells. The preference for the branch points of blood vessels might reflect the importance of Leydig or other interstitial cells, because such regions are usually rich in these cells.

Vasculature-oriented niche specification in the mouse testis. Regardless of the identity of the primary somatic niche cell type, this study has shown that the niche for Aundiff, and possibly that for the stem cells, is associated with the vasculature pattern. The grand design of seminiferous tubules (testis cords in the fetal testis) and testicular vessels are established by developmental processes linked to sex differentiation (28). In contrast, the complex packaging of the highly convoluted seminiferous tubules (29) and the networking of the vasculature running between the tubules are not likely to be predetermined. Rather, the intrinsic pattern-forming activity of the vessels (30) appears to play an important role in the establishment of the local vasculature that surrounds the seminiferous tubules. Therefore, niche formation is unlikely to be a developmentally programmed process and is more likely to be an indirect consequence of vasculature formation. This conclusion is further supported by our grafting data, which showed that Aundiff rearrangement accompanied vasculature alteration (Fig. 5).

A vasculature–oriented niche formation has several attractive features. It allows niches to form at regular intervals all through the seminiferous tubule loops, whatever their length. Additionally, new niches could be added as the tissue increases in size or following tissue damage and revascularization. When the niche is regenerated, spermatogenesis could be reestablished after intratubular migration of the stem cells from neighboring unaffected regions. These features make a good contrast to that of the Drosophila or Caenorhabditis elegans germline niche, which is totally dependent on the developmentally determined somatic cells and, once damaged, niche cells never regenerate. Such niche plasticity could be particularly beneficial for continuous spermatogenesis during the long reproductive periods of mammals.

Supporting Online Material

Materials and Methods

Fig. S1


Movie S1

References and Notes

View Abstract

Navigate This Article