- Record: found
- Abstract: found
- Article: found
Functional assessment of spermatogonial stem cell purity in experimental cell populations
research-article
Publication date (Electronic):
03 April 2018
Journal:
Stem cell research
Read this article at
There is no author summary for this article yet. Authors can add summaries to their articles on ScienceOpen to make them more accessible to a non-specialist audience.
Abstract
Historically, research in spermatogonial biology has been hindered by a lack of validated approaches to identify and isolate pure populations of the various spermatogonial subsets for in-depth analysis. In particular, although a number of markers of the undifferentiated spermatogonial population have now been characterized, standardized methodology for assessing their specificity to the spermatogonial stem cell (SSC) and transit amplifying progenitor pools has been lacking. To date, SSC content within an undefined population of spermatogonia has been inferred using either lineage tracing or spermatogonial transplantation analyses which generate qualitative and quantitative data, respectively. Therefore, these techniques are not directly comparable, and are subject to variable interpretations as to a readout that is representative of a ‘pure’ SSC population. We propose standardization across the field for determining the SSC purity of a population via use of a limiting dilution transplantation assay that would eliminate subjectivity and help to minimize the generation of inconsistent data on ‘SSC’ populations. In the limiting dilution transplantation assay, a population of LacZ-expressing spermatogonia are selected based on a putative SSC marker, and a small, defined number of cells (i.e. 10 cells) are microinjected into the testis of a germ cell-deficient recipient mouse. Using colony counts and an estimated colonization efficiency of 5%; a quantitative value can be calculated that represents SSC purity in the starting population. The utilization of this technique would not only be useful to link functional relevance to novel markers that will be identified in the future, but also for providing validation of purity for marker-selected populations of spermatogonia that are commonly considered to be SSCs by many researchers in the field of spermatogenesis and stem cell biology.
Related collections
Most cited references35
- Record: found
- Abstract: found
- Article: not found
Essential role of Plzf in maintenance of spermatogonial stem cells.
Jose Costoya, Robin M. Hobbs, Maria Barna … (2004)
Little is known of the molecular mechanisms whereby spermatogonia, mitotic germ cells of the testis, self-renew and differentiate into sperm. Here we show that Zfp145, encoding the transcriptional repressor Plzf, has a crucial role in spermatogenesis. Zfp145 expression was restricted to gonocytes and undifferentiated spermatogonia and was absent in tubules of W/W(v) mutants that lack these cells. Mice lacking Zfp145 underwent a progressive loss of spermatogonia with age, associated with increases in apoptosis and subsequent loss of tubule structure but without overt differentiation defects or loss of the supporting Sertoli cells. Spermatogonial transplantation experiments revealed a depletion of spermatogonial stem cells in the adult. Microarray analysis of isolated spermatogonia from Zfp145-null mice before testis degeneration showed alterations in the expression profile of genes associated with spermatogenesis. These results identify Plzf as a spermatogonia-specific transcription factor in the testis that is required to regulate self-renewal and maintenance of the stem cell pool.
- Record: found
- Abstract: found
- Article: not found
Functional hierarchy and reversibility within the murine spermatogenic stem cell compartment.
Manju Sharma, Yo-ichi Nabeshima, Doug R. Braun … (2010)
Stem cells support tissue maintenance by balancing self-renewal and differentiation. In mice, it is believed that a homogeneous stem cell population of single spermatogonia supports spermatogenesis, and that differentiation, which is accompanied by the formation of connected cells (cysts) of increasing length, is linear and nonreversible. We evaluated this model with the use of lineage analysis and live imaging, and found that this putative stem cell population is not homogeneous. Instead, the stem cell pool that supports steady-state spermatogenesis is contained within a subpopulation of single spermatogonia. We also found that cysts are not committed to differentiation and appear to recover stem cell potential by fragmentation, and that the fate of individual spermatogonial populations was markedly altered during regeneration after damage. Thus, there are multiple and reversible paths from stem cells to differentiation, and these may also occur in other systems.
- Record: found
- Abstract: found
- Article: found
Mouse Spermatogenic Stem Cells Continually Interconvert between Equipotent Singly Isolated and Syncytial States
Kenshiro Hara, Toshinori Nakagawa, Hideki Enomoto … (2014)
Introduction In mammalian testes, spermatogenic stem cells are responsible for both the continual production of sperm in steady state and regeneration following injury (de Rooij and Russell, 2000; Meistrich and Van Beek, 1993; Yoshida, 2012). However, the dynamics of the stem cell population remain largely unresolved at the single-cell level. The process of spermatogenesis takes place in seminiferous tubules (Figure 1A). All stages of germ cells are nourished by somatic Sertoli cells, which support a prominent network of tight junctions that separate the basal and adluminal compartments and, together with the basement membrane, provide the structural basis of the tubules. The tubules are surrounded by peritubular cells, whereas the intertubular space is made up of a network of blood vessels and interstitial cell types. Spermatogonia (mitotic germ cells that include stem cells) lie in close association with the basement membrane in the basal compartment. When meiosis begins, cells detach from the basement membrane and translocate across the tight junctions, after which they undergo meiotic divisions and spermiogenesis, and mature sperm are released into the lumen. This organization is observed uniformly throughout the entire 1.7 m tubule length that constitutes a single mouse testis (Russell et al., 1990), suggesting that seminiferous tubules lack a discrete anatomically defined niche. In mouse, spermatogonia are divided into “undifferentiated” and “differentiating” populations (Figures 1A and 1B). “Undifferentiated spermatogonia” are found as singly isolated cells (As) or syncytia consisting mainly of 2 (Apr), 4 (Aal-4), 8 (Aal-8), or 16 (Aal-16) cells. The formation of syncytia is due to “incomplete division,” a germline-specific cell division process by which cytokinesis does not complete and cytoplasmic connection between daughter cells persists via intercellular bridges (de Rooij and Russell, 2000; Russell et al., 1990). This process continues through subsequent mitotic and meiotic divisions, resulting in the extension of syncytia from differentiating spermatogonia (A1, A2, A3, A4, Intermediate, and B) to haploid spermatids. Experimentally, posttransplantation colony-formation and regeneration assays established that, whereas Kit-positive (Kit+) differentiating spermatogonia seem to retain some stem cell potential, the vast majority of stem cell function is restricted to Kit-negative (Kit−) undifferentiated spermatogonia (Barroca et al., 2009; Ohbo et al., 2003; Shinohara et al., 2000). Based on the detailed analyses of fixed specimens, it was proposed in 1971 that stem cell activity may be restricted to the population of As spermatogonia, whereas interconnected Apr and Aal syncytia are irreversibly committed to differentiation and no longer contribute to the stem cell pool (Huckins, 1971; Oakberg, 1971). This hypothesis, which has become the prevailing theory, is known as the “As model.” The population of undifferentiated spermatogonia is further divided according to their heterogeneous gene expression (Hofmann et al., 2005; Nakagawa et al., 2010; Sada et al., 2009; Suzuki et al., 2009a; Yoshida et al., 2004, Yoshida, 2012; Zheng et al., 2009). In undisturbed steady-state spermatogenesis, the GFRα1+ subpopulation (mainly As, Apr, and fewer Aal) is thought to reside on the top of the hierarchy (Nakagawa et al., 2010; Sada et al., 2009). As well as maintaining their own population, GFRα1+ cells also give rise to the second, Ngn3+, subpopulation of undifferentiated spermatogonia (comprised of more Aal and fewer As and Apr) (Figure 1B). Previous studies using live imaging and cre-mediated pulse labeling of Ngn3+ spermatogonia have demonstrated that the majority of Ngn3+ cells transfer to the next Kit+ differentiating spermatogonia (Nakagawa et al., 2007, 2010). Intriguingly, these studies also revealed that Ngn3+ cells retain the capability of regaining GFRα1 expression, fragmenting into single cells or shorter syncytia (through breaking of intercellular bridges), and contributing to the long-term stem cell pool (Figure 1B, yellow dotted arrows), suggesting that the entirety of undifferentiated spermatogonia (both GFRα1+ and Ngn3+) may contribute to stem cell activity. However, such “reversion” of Ngn3+ cells takes place only infrequently in steady state but becomes prevalent in regeneration following tissue insult or transplantation. Therefore, Ngn3+ cells have been considered as a reserve population, whereas GFRα1+ cells are thought to represent the primary population responsible for the stem cell function (Nakagawa et al., 2007, 2010; Spradling and Fan, 2010). Such flexible behavior of Ngn3+ cells, especially syncytial fragmentation, questions the premise of the “As model” that syncytia are irreversibly committed to differentiation. It was also shown that, in common with other Ngn3+ spermatogonia, Ngn3+ As cells are strongly biased toward differentiation to Kit+, indicating that not all the As spermatogonia function equally as stem cells (Nakagawa et al., 2010). In addition, based on the occasional contribution of Ngn3+ cells to the long-term stem cell pool, the behavior of the pulse-labeled stem cells was analyzed for months to over a year. The results demonstrated that stem cells are continually and stochastically lost and replaced by their neighbors, through a process of population asymmetry (Klein et al., 2010; Nakagawa et al., 2007). This observation not only further challenged the “As model” but also questioned the cellular basis of stem cell loss and replacement. To summarize, the preceding analyses of how Ngn3+ spermatogonia behave over time (by means of pulse labeling and live imaging) have questioned the validity of the “As model,” at least in its original form. Although As spermatogonia that are GFRα1+ were presumed to be the primary stem cell type (Nakagawa et al., 2010) (Figure 1B, black dotted arrow with asterisk), this conjecture lacks direct experimental support. To understand the stem cell dynamics, therefore, it is essential to dissect the fate behavior of GFRα1+ spermatogonia over time. In this study, we developed knockin mouse models and conducted intravital in vivo live-imaging and pulse-labeling studies of GFRα1+ spermatogonia at a single-cell resolution. On the basis of the unexpected behavior of GFRα1+ cells observed in these studies, we propose an alternative theory of the identity and behavior of mouse spermatogenic stem cells. Results Stem Cell Function of GFRα1+ Spermatogonia GFRα1+ spermatogonia lie scattered unevenly on the basement membrane of seminiferous tubules (Figures 1A and 1C). Intriguingly, however, their local density over a prolonged tubule length (several millimeters) was found to be remarkably constant in adult mice, with 17 ± 1 “units” of GFRα1+ spermatogonia per mm of tubule. Here, we define “a unit” as either “an As spermatogonium” or “a single syncytium consisting of multiple spermatogonia.” Their composition was also conserved between individuals, not only for As, Apr, and Aal-4, which comprised some 50%, 30%, and 10% of total units, respectively, but also for the small numbers of Aal-8 and “odd-sized” syncytia (e.g., 3-, 5-, and 6-cell chains designated hereafter as Aal-3, Aal-5, and Aal-6), which together constitute the remaining 10% (Figure 1D). To trace the fate of GFRα1+ spermatogonia, we developed a knockin mouse model that enables the pulse labeling of GFRα1+ cells with persistent GFP expression, without disturbing the tissue architecture, following a single administration of 4OH-tamoxifen to GFRα1-CreER T2 ;CAG-CAT-EGFP mice (Figures 2A and 2B; Figures S1A and S1B available online). After partial (∼20%) pulse labeling of this population, long-term chase (months to over a year) revealed that labeled cells formed contiguous patches in which GFP+ cells were found in all stages of differentiation (Figure 2C). Importantly, the percentage of GFP+ cells out of the total GFRα1+ spermatogonia remained constant over the same extended period (Figure 2D). This indicates that the initially labeled GFRα1+ population (comprised of around 6,000 units = some 20% of the 30,000 total GFRα1+ units per testis) continually reproduced approximately the same number of GFRα1+ spermatogonia and gave rise to differentiating descendants that lost GFRα1 expression. This finding also shows that the influx of cells from outside the GFRα1+ compartment (i.e., Ngn3+, Kit+, or other unknown cells) is minimal in this undisturbed condition, consistent with previous reports (Nakagawa et al., 2007, 2010). To conclude, in steady state, the population of GFRα1+ spermatogonia function as the stem cells. Intricate Clonal Fate Behavior of Pulse-Labeled Individual GFRα1+ Spermatogonia Then we were motivated to investigate how individual GFRα1+ spermatogonia behave to achieve such population-level stem cell dynamics, using the same mouse model. With a lower dose of 4OH-tamoxifen, label was introduced into GFRα1+ cells sparsely (one labeled unit per ∼6 mm tubule length on average), so that the fate of each “clone” (defined as a cohort of cells derived from a single labeled unit, regardless of whether or not they have split into multiple units) can be analyzed (Figure 2E). Over the time course of 2–20 days postinduction, the constituent units of individual clones were scored according to their GFRα1 expression and unit length by whole-mount immunostaining of seminiferous tubules. Shortly (2 days) after induction, the majority of labeled clones contained a single GFRα1+ unit (Figures 2F–2K). However, intriguingly, the subsequent clonal fates did not follow a regular pattern, but were highly variable between clones: for example, at day 14, some clones were observed to contain multiple GFRα1+ units only (where the GFRα1+ unit number increased without producing differentiation-destined GFRα1− progeny) (Figure 2G); some contained one GFRα1+ and multiple GFRα1− units (where the GFRα1+ unit number was maintained, while producing GFRα1− progeny) (Figure 2H); and others contained GFRα1− units only (where the GFRα1+ unit was lost altogether) (Figure 2I). The degree of clonal variation in the number of GFRα1+ and GFRα1− units (and cells) broadened progressively over time (Figures 2J, 2K, and S2; Table S1). Despite the apparent variability in the individual clonal fates, the average of more than 100 clones at each time point recovered a conventional steady-state stem cell behavior through this period: in particular, the average number of GFRα1+ units per clone remained close to one, whereas the number of GFRα1− units (and, more dramatically, GFRα1− cells) per clone steadily increased (Figures 2L and 2M). Further, the composition of GFRα1+ units with different lengths across numerous clones remained largely constant over time, commensurate with that of the total GFRα1+ population observed by immunostaining of fixed samples (Figure 2N). Altogether, these findings are consistent with the dynamics of population asymmetry, in which maintenance of stem cells and production of differentiating descendants are balanced at the population level through continuous loss and replacement of stem cells (Klein and Simons, 2011). Cell Division and Syncytial Fragmentation of GFRα1+ Spermatogonia Observed by Live Imaging We then investigated the behavior of GFRα1+ spermatogonia by means of in vivo live imaging of GFRα1-EGFP knockin mouse testis (Uesaka et al., 2007) (Figure S1C), exploiting a procedure reported previously (Yoshida et al., 2007). Theoretically, the observed increase in cell number per pulse-labeled clone indicates the process of cell division (either complete or incomplete), whereas the increase in unit number per clone provides evidence of complete division (As → 2 × As) and/or syncytial fragmentation. Indeed, all of these expected processes were observed directly in the live imaging. Since continuous live imaging was feasible up to ∼3 days, the average rates of these processes were measured by collecting data from multiple time courses (Figure 3A). Unexpectedly, following a total of ∼8,000 hr of observation, only two cases out of 35 divisions of GFRα1+ As cells were found to be complete, leading to the generation of two As spermatogonia (translating to a rate of once per 5–6 months), whereas the vast majority of divisions were incomplete and gave rise to one Apr (Figures 3A–3C; Movies S1 and S2). Within a syncytium (Apr and Aal), cell division was always incomplete and synchronous, leading to the doubling of syncytial length (e.g., Apr→Aal-4). Of particular note, fragmentations of GFRα1+ syncytia were observed at a frequency much higher than that of Ngn3+ syncytia (estimated at around once per 4 months per bridge) (Nakagawa et al., 2010) and even comparable to that of cell division (Figures 3A and 3D; Movie S3). This effectively replenishes the shorter units lost through incomplete divisions. Because of the half-life of the EGFP protein (2–3 days), the live-imaging study could not resolve their transition to Ngn3+ cells, based on their loss of EGFP fluorescence during the filming time available. However, the clonal fate of pulse-labeled GFRα1+ spermatogonia 2 days postinduction indicated that the GFRα1+→Ngn3+ transition occurred in all categories of As, Apr, and Aal GFRα1+ spermatogonia (Figure S3A), consistent with previous live imaging of Ngn3-EGFP mouse testes (Nakagawa et al., 2010). The death of GFRα1+ units was observed only rarely (Figure 3A). Altogether, these observations indicate that GFRα1+ cells continually change their states between As, Apr, and Aal spermatogonia through a combination of incomplete division and syncytial fragmentation, while giving rise to Ngn3+ cells from all of these states. Rates of Incomplete Division and Syncytial Fragmentation of GFRα1+ Spermatogonia In addition to these qualitative implications, the live-imaging study further provides quantitative insight into the dynamics of GFRα1+ cells. First, the rate of cell division (essentially incomplete) appeared to be independent of unit length because As, Apr, and Aal-4 syncytia all divide at around once per 10 days (Figure 3A). Second, the average fragmentation frequency of Apr (one bridge) was around once per 20 days, whereas that of Aal-4 (three bridges) was proportionately higher at around once per 7 days (Figure 3A), suggesting that each bridge breaks around once per 20 days, independent of unit length. Intriguingly, the fragmentation of Aal-4 syncytia provided, instead of a regular pattern, fragments involving all possible permutations, viz. 4xAs, 2xApr, 2xAs+Apr, or As+Aal-3 (Figure 3A), at frequencies compatible with stochastic breakdown of intercellular bridges once a syncytium is licensed for fragmentation (Figure S3B). Therefore, incomplete cell division and syncytial fragmentation of GFRα1+ spermatogonia appear to occur at constant rates, independent of the unit length. Active Movement of GFRα1+ Spermatogonia around the Vasculature-Associated Region In common with the entire population of undifferentiated spermatogonia (Chiarini-Garcia et al., 2001; Yoshida et al., 2007), the GFRα1+ subpopulation tends to localize near the vasculature and interstitium surrounding the tubules (Figures S3C, S3C′, and S3F–S3G′). Moreover, from the live-imaging study, it was apparent that GFRα1+ spermatogonia were in constant movement in the basal compartment (Figures 3E, 3E′, and 3F; Movie S4). Whereas the majority preferentially moved within the vasculature-proximal region, others migrated from one such region to another. Without showing any systematic pattern, cells were seen to actively weave their way through the ordered network of immobile Sertoli cells (Figures 3G, 3G′, and 3H; Movie S5), over a range of approximately 20–150 μm within a single day. This contrasts with the behavior of Ngn3+ spermatogonia, which are less motile in the vasculature-associated region, before actively spreading over the basal compartment on transition into A1 spermatogonia (Yoshida et al., 2007). Synthesis of a Minimal Biophysical Modeling Scheme Considering the aforementioned observations of highly variable clonal fates, continual conversion between the states of As, Apr, and Aal and active movement in the tissue, the behavior of GFRα1+ spermatogonia may seem unconstrained. However, the observation that incomplete division and syncytial fragmentation occur at constant rates, independent of unit length, may suggest simple rules underlying such complex behaviors. We were motivated, therefore, to try to capture the dynamics of GFRα1+ spermatogonia using a biophysical modeling scheme. In particular, we synthesized a model relying solely on parameters inferred from live imaging (the foregoing two rates), as well as their density and localization in seminiferous tubules. Then the validity of the model was evaluated by testing whether it was able to predict the wide range of independent data, including the intricate clonal fates revealed by the pulse-labeling study. Here, the model was designed to be as simple as possible, with minimal parameters, so that we could capture the basic principles of the dynamics of GFRα1+ spermatogonia, as described below. For further details of the modeling scheme, see the Supplemental Experimental Procedures. As seen above, although the GFRα1 units are scattered and moving around the tissue, their local density (viz. pool size) is maintained largely constant. In formulating the model, we aimed to capture the dynamics of GFRα1+ units under the condition of such a constant density. To reflect this, we considered a modeling scheme in which the basal compartment of the seminiferous tubules was divided into domains that accommodate one GFRα1+ unit each (Figure 4A). Based on their measured average density (17 units/mm tubule) and affinity to the vasculature (Figures S3C, S3C′, and S3F–S3G′), whose average number around a tubule is 5.2 (Klein et al., 2010), we divided the circumference into five domains of 1/3 mm in length (Figure 4A). In this scheme, each domain corresponds to the approximate territory of a single GFRα1+ unit. Then to reflect our live-imaging observations, all the constituent GFRα1+ spermatogonial units were allowed stochastically to undergo incomplete cell division that doubles the unit length, at a constant rate (defined as D), independent of unit length (Figure 4B). In addition, intercellular bridges were allowed to break stochastically, leading to the fragmentation of a syncytium into multiple units, at a constant rate (defined as F) per bridge throughout all the GFRα1+ syncytia (Figure 4C). The pattern of fragmentation was set to occur randomly among the bridges (with any bridge breaking with a probability of 50%), consistent with observation (Figure S3B). To achieve a constant density of GFRα1+ units, the genesis of a new GFRα1+ unit by fragmentation was set to accompany the GFRα1+ → Ngn3+ transition in any one of the neighboring domains and the replacement of the lost GFRα1+ unit by the newly formed GFRα1+ unit (Figure 4C). Such translocation between domains was consistent with the observed movements of GFRα1+ cells (Figures 3E–3H). To reduce the complexity of the model, we did not include the infrequent process of As → 2 × As complete division (Figures 3A and 3B), which is in any case implicit in the sequential occurrence of an As→Apr division, followed by Apr → 2 × As fragmentation. Nor did we allow for the death of GFRα1+ spermatogonia or Ngn3+ → GFRα1+ reversion, reflecting the low frequency of these processes (Figure 3A) (Nakagawa et al., 2010). Reflecting their low fragmentation frequency, we could assume that GFRα1− (Ngn3+ and more advanced) units do not fragment but simply accumulate at each lattice site. Finally, GFRα1− units were also set to follow a low rate of death (once per 30 days), consistent with observation (Huckins and Oakberg, 1978). However, within the framework of the model, death of GFRα1− units does not, in any case, affect the dynamics of the GFRα1+ population. Model Prediction of the In Vivo Dynamics of GFRα1+ Spermatogonia We then questioned whether the model had the capacity to predict the wide and complex range of independent in vivo measurements. In this scheme, the in silico dynamics of GFRα1+ spermatogonia is fully specified by just two parameters: the rates of cell division (D) and fragmentation (F). First, within the framework of the model, the GFRα1+ population is predicted to converge to steady state, in which GFRα1+ units acquire a particular composition that is independent of the initial condition but depends uniquely on the ratio D/F (colored lines in Figure 4D). Intriguingly, using the rates of D ( = once per 10 days) and F ( = once per 20 days per bridge) inferred from live imaging, the model faithfully recapitulated the steady-state composition of GFRα1+ spermatogonia measured in vivo (Figure 4D). The steady state is recovered rapidly (largely within 10 days, corresponding to one round of cell division on average), even from such an extreme initial condition in which all GFRα1+ units are As (Figure 4E). In addition to the proportions of GFRα1+ As, Apr, and Aal-4 spermatogonia, the model correctly predicted the near-absence of GFRα1+ units larger than eight, as well as the small number of “odd-sized” units (Aal-3, Aal-5, Aal-6, and Aal-7), which was already a nontrivial test of the validity of the modeling scheme. Second, we evaluated the extent to which the model can predict the detailed clonal fate dynamics of GFRα1+ units scored in the pulse-labeling experiment over 20 days postinduction (Figures 2E–2N; Figure S2; Table S1), with the same rates of D and F. In common with the in vivo observation (Figures 2J, 2K, and S2), in silico clones also followed variable fates. In the model, once the GFRα1+ → Ngn3+ transition occurs in all the units of a clone, such a clone never returns to the GFRα1+ compartment. As a result, the fraction of clones that retained at least one GFRα1+ unit (surviving rate) progressively diminishes over time (line in Figure 4F), which quantitatively recapitulated the in vivo measurements (squares in Figure 4F). As a consequence of the steady-state dynamics, it was also predicted that the average number of GFRα1+ units (and cells) of the “surviving” clones progressively increase, so that the average number of GFRα1+ units across all clones remains close to one. Indeed, these predictions quantitatively captured the in vivo observations (Figures 2L, 4G, and S4A). More significantly, the model prediction showed excellent agreement with the in vivo distribution of GFRα1+ units (and cells) at all data points over the time course of 20 days (Figures 2J, 4H, S2, and S4B). Finally, the average and distribution of the number of GFRα1− units per clone (i.e., units that had exited GFRα1+ compartment and transited to Ngn3+ and then more advanced spermatogonia) were also accurately predicted by the model (Figures 2K, 2L, 4I, and S4A). To summarize, from an in silico modeling scheme that was synthesized solely from the local density and distribution of GFRα1+ units in seminiferous tubules and the rates of cell division and syncytial fragmentation inferred from the live-imaging study, we were able to accurately predict both the steady-state tissue composition of spermatogonial units (an independent measurement) and the intricate fate behavior of spermatogonial units (obtained from a totally independent pulse-labeling study). These findings provide strong support for the validity of the simple modeling scheme in capturing the in vivo steady-state behavior of GFRα1+ population, at least over the 20 day time course. Long-Term Dynamics of GFRα1+ Population We then assessed the capacity of the same model to predict the long-term (over months) clonal behavior in steady state. On this timescale, clones derived from the pulse-labeled single GFRα1+ units evolved into large (sometimes fragmented) patches in which the labeled GFRα1+ cells were overwhelmed in number by differentiating labeled cells. Because a count of unit(cell) number is unfeasible in this phase, we characterized clone size by the patch length along the axis of the tubule (Figures 5A and 5B). In common with the number of GFRα1+ units(cells) in each clone seen in the short-term pulse-labeling study up to 20 days, at 2, 3, 6, 10, and 14 months postinduction, the patch length showed a variable size distribution, with the average length increasing over time (Figures 5C–5E). In parallel, the number of surviving patches per testis decreased (Figure 5F). Significantly, extrapolation of the in silico dynamics to these longer times correctly predicted the average and distribution of patch length and surviving rate of GFRα1+ unit-derived clones, using the same rates of cell division and fragmentation, D ( = once per 10 days) and F ( = once per 20 days per bridge) (Figures 5D–5F). In previous studies, pulse labeling of Ngn3+ spermatogonia was used to trace the fate of surviving stem cell clones over a 14 month time course (Klein et al., 2010; Nakagawa et al., 2007). Because the majority of Ngn3+ cells are destined for differentiation, these studies relied on the premise that the few Ngn3+ spermatogonia that had transited back into the GFRα1+ compartment behaved without distinction from the “innate” GFRα1+ cells (Nakagawa et al., 2010). This premise is strongly supported by the observation that the average and distribution of patch lengths as well as the rate of surviving clones are all consistent between GFRα1+ and Ngn3+ cell-derived clones and predicted by the same model at these long times (Figures 5D–5F). To summarize, the minimal biophysical model, which was synthesized from the very short-term observations in live imaging (up to 3 days), is capable of capturing the fate behavior of GFRα1+ spermatogonia from the short term (up to 20 days) to the long term (up to times comparable with the life span of mice). Dynamics of GFRα1+ Spermatogonia Following Tissue Insult We then turned to investigate the dynamics of GFRα1+ spermatogonia following a strong perturbation from steady state. To do so, we analyzed a partial germ cell depletion model induced by a moderate dose of busulfan (10 mg/kg). In this condition, testis shows acute and massive germ cell death, causing the reduction of the number of GFRα1+ units to a minimum of around one-third of the steady-state value by postinsult day 10. Then the number of GFRα1+ spermatogonia gradually comes back to their preinsult level in about 2 months (Nakagawa et al., 2010). Experimentally, we pulse labeled the GFRα1+ cells on postinsult day 10 and analyzed their clonal fate in the following recovery phase (Figure 6A). Although the rate of Ngn3+ → GFRα1+ reversion increases substantially following insult (Nakagawa et al., 2010), the appearance of GFRα1+ cells through this process was limited during the observed period (Figure S5A). As with steady state, in regeneration, the size of individual pulse-labeled clones diverged over time in the number of both constituent GFRα1+ and GFRα1− units (Figures 6B and 6C). Interestingly, a significant portion of clones lost all GFRα1+ progeny even in regeneration. However, in contrast to steady state, the total number of GFRα1+ units (viz. the average number of GFRα1+ units per clone) increases because of a tilt in the overall balance of production of GFRα1+ and GFRα1− units toward the former (Figure 6D). Then we questioned whether the same biophysical model could capture the clonal dynamics in regeneration, after seeding the lattice with GFRα1+ units in proportion to their observed density and syncytial composition. As the simplest adaptation of the model, we allowed syncytial fragmentation to be uncompensated by loss of a neighboring GFRα1+ unit, when the fragment migrates into a vacant domain. Under these conditions, we found that the predictions of the model showed a remarkable agreement with the measured clone dynamics, including the clone survival rate (Figure 6E), and the average and distribution of clone sizes (for both GFRα1+ units[cells] and GFRα1− units) (Figures 6D, 6F–6H, and S5B), over the wide range of time points for 18 days, if we made a minimal adjustment of the rate parameters. In particular, we found an optimal fit of the model to the data when the rates of cell division and fragmentation were both increased to around once per 8 days (from once per 10 days in steady state) and once per 10 days per bridge (from once per 20 days), respectively, while introducing a significant decrease of the death rate of GFRα1− cells (once per 160 days, from once per 30 days). Using these parameters optimized from the short-term clonal behavior (up to 18 days), the model also accurately predicted the recovery in the tissue density of GFRα1+ units and cells up to 2 months following insult, when the regeneration process largely completed (Figure S5C). These results provide further support for the general validity and predictive power of the modeling scheme. Moreover, these results suggest that the dynamics of GFRα1+ cells in regeneration following tissue injury is not based on a distinct program but follows the same pattern of stochastic rules as that seen in steady state. Discussion Motivated by recent observations that question the validity of the prevailing “As model,” this study explored an alternative theory of the identity and behavior of mouse spermatogenic stem cells and conducted single-cell-resolution analyses of the behavior of GFRα1+ spermatogonia. The live-imaging study revealed that, in the GFRα1+ compartment, practically all cell divisions are incomplete, whereas syncytial fragmentation occurs rather frequently, and that these processes follow constant rates that are independent of unit length. Based on these two measured rates, as well as cell density and localization in seminiferous tubules, we developed a minimal biophysical model to describe the dynamics of GFRα1+ spermatogonia. This model could predict the range of complex data obtained from independent measurements, from the steady-state composition of GFRα1+ units to the wide range of intricate clonal fate behaviors of pulse-labeled GFRα1+ cells, both in steady state and in regeneration. Given the ability of such a highly simplified model to predict the complex in vivo behavior, we concluded that the principles that define the dynamics of GFRα1+ compartment have been successfully resolved. On the other hand, the contribution of rare events, such as cell death or possible deviation from stochasticity, should be small enough to capture the overall dynamics of stem cells, although these factors should affect the detailed behavior of stem cells. Figure 7A illustrates the stem cell dynamics proposed in this study, in which GFRα1+ units continuously extend via incomplete division and fragment via intercellular bridge breakdown, while giving rise to Ngn3+ progeny. In this scheme, individual GFRα1+ cells constantly change their state reversibly between single cells and variable lengths of syncytia; Figure 7B represents a typical example of such stochastic fate behavior predicted in silico. Yet, through this process at the population level, the number and composition of GFRα1+ cells, as well as the production rate of Ngn3+ cells, were kept constant. Therefore, we propose that the entirety of GFRα1+ spermatogonia comprises a single “stem cell pool.” Onset of Ngn3 expression represents the exit from the stem cell compartment toward differentiation. Although this transition does not indicate the loss of reversibility, Ngn3+ cells show a pronounced differentiation bias. Indeed, in steady state, the vast majority of pulse-labeled Ngn3+ cells differentiate into spermatozoa and disappear from the tissue after a couple of months. When cells become Kit+, they appear to further decrease in potential to return into the GFRα1+ compartment (Figure 7A, black broken arrows) (Barroca et al., 2009; Nakagawa et al., 2007, 2010). In parallel, the frequency of (incomplete) cell division increases, whereas syncytial fragmentation becomes more infrequent (Huckins and Oakberg, 1978; Nakagawa et al., 2010). As a result, the bulk of Ngn3+ and Kit+ spermatogonia extend their unit length unidirectionally as differentiation progresses (Figure 7A). In mouse spermatogenesis, syncytial fragmentation was proposed based on the odd-numbered spermatogonial units observed after irradiation (Erickson, 1981) and directly filmed for the first time in Ngn3+ spermatogonia (Nakagawa et al., 2010). Given its low frequency in steady state, this process has been considered functionally significant in regeneration, where fragmentation appears to occur much more frequently (Nakagawa et al., 2010; Spradling and Fan, 2010). However, the current study revealed that, in the GFRα1+ compartment, syncytial fragmentation regularly occurs in steady state at a frequency comparable with that of cell division and that this process plays a fundamental role in the maintenance of stem cell pool and the continuity of spermatogenesis. Within the modeling scheme established in this study, the maintenance of the GFRα1+ population relies on the balance between their multiplication (syncytial fragmentation) and loss (transition into Ngn3+), which are locally coordinated. This provides a cell-level explanation for the population asymmetry of mouse spermatogenic stem cells that was first discovered by large-scale (millimeters of patch length) and long-term (over months of chase period) clonal fate analysis of pulse-labeled Ngn3+ cells (Klein et al., 2010; Nakagawa et al., 2007). Supporting this idea, the rate of stem cell loss and replacement and the long-term scaling property of the clonal patch length distribution obtained by Klein et al. (2010) agree quantitatively with those predicted by the current biophysical model (Figure 5D; Supplemental Experimental Procedures). In tissues without formation of syncytia, population asymmetry is typically achieved by local correlation of the division and loss (differentiation) of stem cells (Klein and Simons, 2011). In mouse spermatogenesis, the striking capacity of the model to capture the clonal fate behavior suggests that syncytial fragmentation may be linked with loss (transition to Ngn3+). However, the current study can address neither the causal relationship between fragmentation and loss nor the mechanism that coordinates these processes. To answer these questions, it is important to reveal the (yet unknown) mechanism that keeps the local density of GFRα1+ units constant in seminiferous tubules, a feature that was built into the construction of our current modeling scheme. This study highlights a long-held question in germ cell research: what is the biological significance of intercellular connection? After meiosis, it is established that the connection ensures the formation of equivalent gametes from haploid spermatids, utilizing the shared cytoplasmic gene products, including those from X and Y chromosomes (Braun et al., 1989). However, the role of the connection remains an open question for diploid spermatogonia. Given the theory of equipotent stem cell pool composed of the entire GFRα1+ spermatogonia, the connection appears to be unrelated to their stem cell potential. However, Ngn3+ and Kit+ spermatogonia harbor more stable intercellular bridges, suggestive of some unknown role of intercellular connection in this differentiation-destined compartment. Further investigations are warranted to address this fundamental question. Another important finding of this study is that the seemingly complex dynamics of the GFRα1+ compartment can be effectively described only by the rates of incomplete cell division (D) and syncytial fragmentation (F), the governing parameters of the modeling scheme. This notion is strengthened by the analysis of regeneration (Figure 6), which suggests that these rates are regulated in parallel throughout the GFRα1+ compartment, independent of their unit length, so that GFRα1+ spermatogonia can rapidly recover their pool size. Interestingly, the steady-state composition of GFRα1+ units with different lengths sensitively reflects the ratio of these rates, D/F (Figure 4D). In this scheme, the generation and frequency of spermatogonial units are simply a reflection of this ratio, including not only As, Apr, Aal-4, Aal-8, etc., that are considered “regular” in the “As model” but also those with 3, 5, 6, 7 cells, etc., that are often considered “irregular.” Together with the mutual interchange of these morphologically different states, we propose that all GFRα1+ units are equally regular and, in particular, that As is not a special entity. We also propose here the nomenclature of Aal-3, Aal-5, Aal-6, Aal-7, etc., without distinction from As, Apr, Aal-4, Aal-8, etc. Although the results of this study fully support the theory of a single stem cell pool comprised of functionally equivalent As and syncytial GFRα1+ spermatogonia (Figure 7A), one may also conjecture that a small population of slow-cycling GFRα1+ As cells that always undergo complete division act as the “true” stem cells. If this were the case, then one would expect a transfer of the clonal dynamics from that represented by the behavior of short-lived cells (which interconvert between As and syncytia) to that of the slow-cycling As population over a long time course. However, such transfer was not observed in the fate behavior of GFRα1+ cells over extended timescales, from days to over a year. Therefore, although we do not rule out the presence of such a slow-cycling compartment, we can conclude that their contribution (would they exist) is not essential to maintain life-long spermatogenesis in mouse (for details, see the Supplemental Experimental Procedures). It remains open, however, in other animals with longer longevities, whether the same scheme of a single stem cell pool can be extrapolated to the years- or decades-long spermatogenesis or whether some slow-cycling population plays a significant role. Interestingly, primate testes host a large number of immature spermatogonia, Adark, which appear not to have rodent counterparts (Hermann et al., 2010). Elucidating the roles of this population would be warranted to address this interesting question. In this study, the live-imaging observation also revealed a unique property of GFRα1+ spermatogonia: the active movement over the seminiferous tubules. To orchestrate the local coupling of syncytial fragmentation and loss (transition into Ngn3+) of sparsely distributed GFRα1+ spermatogonia, it is vital that GFRα1+ cells are able to freely relocate. Failure to do so would elicit progressive unevenness in the local density of GFRα1+ cells, which would eventually compromise the integrity of the tissue. Indeed, the high motility of stem cells observed in mouse spermatogenesis may be paradigmatic of systems in which the stem cell niche is “facultative” or “open” (Morrison and Spradling, 2008; Stine and Matunis, 2013). This shows a stark contrast with other systems supported by “definitive” or “closed” niche, such as Drosophila germline, mammalian intestinal crypt, or hair follicles, where stem cells lie in close association with each other and remain attached to a localized niche structure, and their movement must be limited within the niche region (Fuller and Spradling, 2007; Lin and Spradling, 1997; Sheng and Matunis, 2011). In future studies, it will be important to understand how the movement of stem cells is controlled and regulated by the interaction with such niche environments. Finally, although incomplete division and syncytial fragmentation are germ cell specific, this study may provide important insights for other stem-cell-supported systems. In particular, we show that stem cells can be defined, not as a particular cell type, but as a heterogeneous population in which cells continually interconvert between different states. Indeed, such “dynamical heterogeneity” resonates with the recent live-imaging study of hair follicle stem cells, which show that self-renewal potential may be correlated with position within the stem cell niche (Rompolas et al., 2013). Experimental Procedures Animals GFRα1 CreERT2 knockin mice were generated as described in Figure S1A. GFRα1 EGFP (Uesaka et al., 2007), CAG-CAT-EGFP (Kawamoto et al., 2000), and GATA1-EGFP (Suzuki et al., 2009b) alleles were as previously described. All the mice used in this study were heterozygous for one or two of these alleles and simply indicated by their allelic name(s), with the background of C57BL/6 (from Japan CLEA and Japan SLC and used as wild-type animals). Although mice carrying one nonfunctional knockin allele of the GFRα1 gene, GFRα1 CreERT2 , and GFRα1 EGFP were used for pulse-labeling and live-imaging experiments, respectively, such heterozygosity neither affected the total density and composition of GFRα1+ units nor the overall integrity of spermatogenesis, over 1 year postlabeling (Figures S1D–S1I). All animal experiments were conducted with approval of The Institutional Animal Care and Use Committee of National Institutes of Natural Sciences, unless specifically mentioned. Whole-Mount Immunofluorescence of Seminiferous Tubules Immunostaining of whole-mount seminiferous tubules was performed as previously described (Nakagawa et al., 2010) using anti-GFRα1 Ab (1:1,000 dilution; R&D Systems), anti-GFP Ab (1:300 dilution; Invitrogen), and anti-Kit Ab (1:200 dilution; BD Biosciences). Observation and photography were performed with a BX51 upright fluorescence microscope equipped with a DP72 CCD camera (Olympus). Spermatogonia were judged as belonging to a syncytium when, based on a continuous GFRα1 or GFP staining using a 60× water immersion objective lens, the cell-cell connection was visually detected. To measure the lengths of the patches of GFP+ cells, M205C stereomicroscope with a DFC490 CCD camera (Leica) was used. Pulse Labeling of GFRα1+ Spermatogonia Three-month-old GFRα1-CreER T2 ; CAG-CAT-EGFP mice were injected intraperitoneally with 2.0 mg (for a higher labeling efficiency) or 0.25–0.35 mg (for clonal labeling) per individual of 4OH-tamoxifen (Sigma) dissolved in ethanol, in dimethyl sulfoxide, and then in sesame oil (Nakalai Tesque). For clonal lineage tracing, their testes were processed for whole-mount immunofluorescence. To induce regeneration, mice were intraperitoneally injected with busulfan (10 mg/kg) as described previously (Nakagawa et al., 2010) prior to 4OH-tamoxifen administration. Intravital Live Imaging Live-imaging of the testes of 4- to 5-month-old GFRα1 EGFP or GFRα1 EGFP ;GATA1-EGFP mice under anesthesia was performed as described before, using epifluorescence IX61WI microscope (Olympus) (Yoshida et al., 2007). Time-lapse images were captured at the rate of one flame per 30 min using the Andor iXon EM-CCD camera controlled by Metamorph software (Molecular Devices). Movies were constructed by Metamorph software, and the trajectories of spermatogonia and Sertoli cells were manually extracted from movies using Metamorph and ImageJ software. An intercellular bridge was deemed to be intact if the cells remained within 30 μm for more than 12 hr (Supplemental Experimental Procedures). Although only GFRα1+ cells located in the superficial region of the testis were filmed, their representativeness is supported by the agreement of the densities and compositions of GFRα1+ units measured by live-imaging and whole-mount immunostaining studies (the latter represents the average values of the entire tubules) (Figures S3C–S3E). It is further consolidated by the correspondence between rates of cell division and syncytial fragmentation measured from live-imaging data and the range of fate behavior of pulse-labeled GFRα1+ cells found evenly throughout the testis (see text). Author Contributions K.H., B.D.S., and S.Y. designed the experiments, performed data analysis, and wrote the manuscript. K.H. and T.N. performed in vivo experiments. B.D.S. performed in silico analyses. H.E., M.S., and M.Y. generated genetically modified animals.
Author and article information
Journal
Journal ID (nlm-journal-id): 101316957
Journal ID (pubmed-jr-id): 35616
Journal ID (nlm-ta): Stem Cell Res
Journal ID (iso-abbrev): Stem Cell Res
Title:
Stem cell research
ISSN
(Print):
1873-5061
ISSN
(Electronic):
1876-7753
Publication date Nihms-submitted: 11
June
2018
Publication date
(Electronic):
03
April
2018
Publication date
(Print):
May
2018
Publication date PMC-release: 27
February
2019
Volume: 29
Pages: 129-133
Affiliations
Author notes
Article
Manuscript ID: NIHMS974266DOI: 10.1016/j.scr.2018.03.016
PMC ID: 6392036
PubMed ID: 29660605
SO-VID: 5bb9dbf4-ead7-4181-9d0f-8e4aa22ab710
License:
This is an open access article under the CC BY-NC-ND license ( http://creativecommons.org/licenses/by-nc-nd/4.0/).
History
Categories
Subject:
Article
ScienceOpen disciplines: Molecular medicine