|Home | About | Journals | Submit | Contact Us | Français|
Conceived and designed the experiments: KG KEO. Performed the experiments: KG. Analyzed the data: KG KEO. Contributed reagents/materials/analysis tools: KG KEO. Wrote the paper: KG KEO.
The spermatogenic lineage is established after birth when gonocytes migrate to the basement membrane of seminiferous tubules and give rise to spermatogonial stem cells (SSC). In adults, SSCs reside within the population of undifferentiated spermatogonia (Aundiff) that expands clonally from single cells (Asingle) to form pairs (Apaired) and chains of 4, 8 and 16 Aaligned spermatogonia. Although stem cell activity is thought to reside in the population of Asingle spermatogonia, new research suggests that clone size alone does not define the stem cell pool. The mechanisms that regulate self-renewal and differentiation fate decisions are poorly understood due to limited availability of experimental tools that distinguish the products of those fate decisions. The pluripotency factor SALL4 (sal-like protein 4) is implicated in stem cell maintenance and patterning in many organs during embryonic development, but expression becomes restricted to the gonads after birth. We analyzed the expression of SALL4 in the mouse testis during the first weeks after birth and in adult seminiferous tubules. In newborn mice, the isoform SALL4B is expressed in quiescent gonocytes at postnatal day 0 (PND0) and SALL4A is upregulated at PND7 when gonocytes have colonized the basement membrane and given rise to spermatogonia. During steady-state spermatogenesis in adult testes, SALL4 expression overlapped substantially with PLZF and LIN28 in Asingle, Apaired and Aaligned spermatogonia and therefore appears to be a marker of undifferentiated spermatogonia in mice. In contrast, co-expression of SALL4 with GFRα1 and cKIT identified distinct subpopulations of Aundiff in all clone sizes that might provide clues about SSC regulation. Collectively, these results indicate that 1) SALL4 isoforms are differentially expressed at the initiation of spermatogenesis, 2) SALL4 is expressed in undifferentiated spermatogonia in adult testes and 3) SALL4 co-staining with GFRα1 and cKIT reveals distinct subpopulations of Aundiff spermatogonia that merit further investigation.
Spermatogonial stem cells (SSC) are located on the basement membrane of the seminiferous tubules and drive spermatogenesis throughout adult life by balancing self-renewal and differentiation cell divisions. SSCs are generally considered to be among the cohort of undifferentiated type A spermatogonia (Aundiff), the most primitive cell type in the testis (for review see ). Aundiff are arranged as single cells (Asingle or As), pairs of cells (Apaired or Apr) and chains of four to 16 or 32 cells (Aaligned or Aal) . Among the undifferentiated spermatogonia, the As spermatogonia and some Apr spermatogonia are the most likely candidates to contain stem cell activity, though limited stem cell potential of larger clones has been speculated . It has been estimated that in the mouse, As spermatogonia represent only 0.03% of germ cells in the testes . The pool of adult SSCs is established after birth by the transient population of gonocytes. This cell population is quiescent after birth, but re-enters the cell cycle around postnatal day 3. The gonocytes then migrate to the seminiferous tubule basement membrane where some will establish the pool of SSCs and others will differentiate directly into cKIT-positive differentiating spermatogonia and give rise to the first wave of spermatogenesis. , . This phase of germ lineage development is therefore crucial to initiate and maintain male fertility.
Here we analyzed the expression of the pluripotency factor SALL4, a member of the sal gene family of zinc-finger transcription factors, in the male mouse germline during postnatal development and during steady-state spermatogenesis. Members of this family are highly conserved between species; homologues have been identified in Drosophila , , Xenopus , zebrafish , chicken , mice  and humans . In mice and humans, four sal proteins (SALL 1–4) have been described that are important for normal development and that are implicated in various congenital disorders .
During embryo development, Sall4 regulates axis formation  and expression was observed in progenitor populations of brain, neural tube, heart, pituitary gland, somites and limbs, as well as in fetal hepatic stem/progenitor cells , . Proper Sall4 action in these organs during development is crucial, as evidenced by the mutant phenotype. Mutations in the human SALL4 locus cause Okihiro syndrome, a rare autosomal-dominant condition characterized by limb and cardiac malformations, hearing loss, strabismus and postnatal growth retardation , . Several Sall4 mutant mice have been developed , , , . Homozygous mice were embryonic lethal in these models, and haploinsufficiency mimicked the human phenotype in varying degrees, exhibiting modest to severe phenotypes and postnatal death of pups within the first weeks after birth. Although in one study the heterozygous mutants were able to produce small litters when intercrossed or backcrossed to a C57BL/6 background , the testicular phenotypes of the mutants have not been evaluated systematically in these studies. Conditional knock-out mice where Sall4 was deleted in embryonic germ cells failed to establish a robust stem cell pool, resulting in many agametic tubules . Postnatal deletion of Sall4 in Stra8-expressing germ cells did not impair the onset of spermatogenesis, and germ cell differentiation appeared normal in young adult mice. However, increased seminiferous tubule degeneration was observed in older mice, indicating a role of SALL4 for stem cell pool homeostasis .
Hobbs et al. demonstrated that SALL4 can interact directly with PLZF, a transcription factor required for SSC function and proposed that the titration of these two transcription factors regulates their distribution on the chromatin and consequently target gene regulation . To gain additional insights about this novel regulatory mechanism in the context of spermatogenic lineage development, we evaluated SALL4 expression in the developing testis on postnatal days 0, 3, 7 and 14, the period when gonocytes give rise the SSCs and the first round of spermatogenesis. In addition, we correlated SALL4 expression in adult testes with established markers of undifferentiated (GFRα1, PLZF, LIN28) and differentiating (cKIT) spermatogonia as well as with clone size (As, Apr, Aal, etc.) in whole mount preparations of seminiferous tubules.
C57BL/6 mice were obtained from the Jackson Laboratory (Bar Harbor, ME) and maintained in the laboratory animal housing facility at Magee-Womens Research Institute. All animal procedures were approved by the Institutional Animal Care and Use Committees of Magee-Womens Research Institute and the University of Pittsburgh. All procedures were performed in accordance with the NIH Guide for the Care and Use of Laboratory Animals (Assurance # 3654-01).
The following primary antibodies were used for immunoblotting, immunofluorescence staining of testis tissue sections, or whole mount seminiferous tubules: mouse anti-ß actin (No. A1978, Sigma, St. Louis, MO); mouse anti-DAZL (ab17224, Abcam, Cambridge, MA); rabbit anti-SALL4A/B (ab29112, Abcam); mouse anti-SALL4A (ab57577, Abcam), goat anti-GATA4 (SC-1237, SantaCruz); goat anti-PLZF (AF2944, R&D Systems Inc., Minneapolis, MI); goat anti-LIN28 (AF3757, R&D Systems); goat anti-GFRα1 (AF650, R&D Systems); rabbit anti-VASA/DDX4 (ab13840, Abcam); armenian hamster anti-cKIT (H2C7, ).
Testes from PND 0, 3, 7 and 14 mice were dissected, the tunica albuginea removed and collected in chilled RIPA buffer (150 mM NaCl. 0.1% SDS, 1% NP-40, 50 mM Tris pH 7.5, 0.5% Sodium-Deoxycholate) for preparation of whole testis protein lysates. Tissues were disrupted mechanically and incubated on ice for 30 min and lysates were cleared from residual cell debris by centrifugation for 10 min at 10,000×g. The supernatant was collected and protein concentrations were determined by Bradford colorimetric assay (Protein Assay No.500-0006, Biorad, Hercules, CA). Protein (10 µg) was separated by SDS-PAGE on a 10% gel (Biorad Mini-PROTEAN® TGX™ Precast Gel #456-1033) and transferred onto a nitrocellulose membrane using standard techniques. Blots were blocked with 5% Blotto (No. sc-2324, Santa Cruz Biotechnology Inc., Santa Cruz, CA) in 1× Dulbecco’s Phosphate Buffered Saline (DPBS, No. 14200166, Invitrogen, Carlsbad, CA) containing 0.1% Tween-20 (P5927, Sigma) for 1 hour at room temperature. Primary antibodies rabbit anti-SALL4 (11000) and mouse anti-beta actin (110000) were incubated overnight at 4°C. After washing the blots three times for 10 min in DPBS-T, horseradish peroxidase-conjugated anti-rabbit IgG or anti-mouse IgG (15000 dilution) were incubated for 1 hour at room temperature and protein bands were detected with SuperSignal West Pico Chemiluminescent Substrate as instructed (PI34080, Thermo Fisher Scientific, Pittsburgh, PA). ImageJ software (NIH, Bethesda, MD) was used for densitometry to evaluate relative levels of SALL4A, SALL4B and total SALL4 protein during testis development.
Testes from neonatal (postnatal day (PND) 0 and PND 3), juvenile (PND 7 and PND 14) and adult (>6 weeks of age) mice were dissected and fixed in 4% PFA for 2 hours (neonatal testes) or overnight (juvenile and adult testes) at 4°C. Fixed tissues were processed for paraffin embedding and 5 µm serial sections were collected. Sections were deparaffinized in xylene (2×15 min), rehydrated in a graded ethanol series (2×100% for 10 min, 1×95% for 5 min, 1×80% for 5 min, 1×70% for 5 min, 1×50% for 5 min, 1×25% for 5 min) and rinsed in 1× DPBS. Slides were then incubated in sodium citrate antigen retrieval buffer (10 mM Sodium Citrate, 0.05% Tween-20, pH 6) for 30 min at 97.5°C and allowed to cool to room temperature. After rinsing twice in 1× DPBS containing 0.1% Tween-20 (DBPS-T), unspecific binding sites in tissue sections were saturated by incubation with blocking buffer (1× DPBS containing 3% bovine serum albumin, 0.1% Triton X-100 and 5% normal serum from the host species of the secondary antibody) for 30 min at room temperature. Primary antibodies were diluted in blocking buffer and added to tissue sections for 90 min at room temperature. Isotype matched normal IgG at a comparable concentration was used instead of primary antibody in negative controls (see Figures S1 and S2). After washing the slides in DPBS-T 3× for 5 min, samples were incubated with the appropriate secondary antibody (goat IgG or donkey IgG) conjugated to AlexaFlour-488 or AlexaFluor-568 dye (all Molecular Probes, Invitrogen) for 45 min at room temperature. Sections were washed 3×5 min with DPBS-T and 1×5 min with DPBS before mounting with Vectashield Mounting Medium with DAPI (H-1200, Vector Laboratories, Burlingame, CA).
Testes from adult C57/Bl6 male mice were collected in chilled DPBS and the tunica albuginea was carefully removed. The seminiferous tubules were gently teased apart with fine forceps and briefly digested with 1 mg/ml DNase I (No. DN-25, Sigma) and 1 mg/ml Collagenase IV (No. C5138, Sigma) in DPBS. Tubules were fixed overnight in 4% PFA at 4°C and then extensively washed 3–4 times with DPBS at 60-minute intervals. For SALL4, PLZF and LIN28 staining, tubules were dehydrated in a graded methanol series (25%, 50%, 75%, 95% and 2× in 100% MeOH) for 10 min at room temperature. Samples were permeabilized for 3 hours at room temperature in a 411 mix of MeOH/DMSO/H2O2. Sequential rehydration was by incubation in 50% MeOH (10 min) and 25% MeOH (10 min) and DPBS (2×10 min). Samples were blocked with ice cold PBSMT blocking buffer (1×DPBS +2% blotto milk powder +0.5% Triton X-100) for 2×15 min and 1×2 hours and incubated in primary antibody at 4°C overnight. Isotype matched normal IgG at a comparable concentration was used instead of primary antibody as a negative control (see Figure S3).
Samples were washed with ice-cold PBSMT (2×15 min, 5×1 hr) and incubated in AlexaFluor-conjugated secondary antibody overnight at 4°C. Samples were washed again with ice-cold PBSMT (2×15 min, 5×1 hr) followed by DPBS (2×10 min) and mounted on raised coverslips with Vectashield mounting medium containing DAPI nuclear stain. For co-staining of SALL4 with the surface antigens GFRα1 or cKIT, tubules were only shortly dehydrated in DPBS/10% MeOH/0.1% Triton X-100 at 4°C for 60 min and blocked with PBSMT at 4°C for 60 min. Primary antibodies were diluted in PBSMT and incubated overnight at 4°C. Samples were washed 6×15 min at room temperature with DPBS/0.1% Triton X-100 (PBSTx) and incubated with donkey anti-goat Alexa568 (No. A-11057, Molecular Probes, Invitrogen) or goat anti-armenian hamster Rhodamine-Red-X (No. 127-295-160, Jackson Immuno Research, West Grove, PA) conjugates for 1 hour at 37°C. Samples were washed 6×15 min at room temperature with PBSTx and mounted on raised coverslips with Vectashield mounting medium containing DAPI.
Sections and tubules where imaged using an Eclipse 600 fluorescent microscope (Nikon USA, Melville, NY) equipped with an X-Cite 120 fluorescence source (EXFO Life Sciences, Ontario, Canada) and a Spot RT 2.3.1 slider camera (Diagnostic Instruments, Sterling Heights, MI). A dual-emission FITC/TRITC filter was used to observe green and red fluorescent signals. Image analysis was performed using MetaVue software (Molecular Devices, Sunnyvale, CA).
Stained tissue sections from at least three adult animals were used for each co-staining experiment and a minimum of 100 circular cross sections per animal was evaluated for quantification of co-expression patterns. Whole mount tubules were mounted in small fragments of several mm in length and a minimum of 500 spermatogonial clones were scored in at least three animals for each experiment. Spermatogonial clones were defined according to Tegelenbosch and de Rooij . The combined number of all single- and/or double-stained spermatogonia was used as the denominator to calculate cell ratios.
Previous work by Wang et al. suggested that the Sall4 gene is expressed in pre-meiotic spermatogonia in the postnatal day 7 mouse testis . These observations were confirmed in the recent studies of Eildermann et al.  and Hobbs et al. . The first week after birth is critical for male germline development. During this time, gonocytes re-enter mitosis and migrate to the seminiferous tubule basement membrane where some will establish the pool of SSCs and others will differentiate directly into cKIT-positive differentiating spermatogonia and give rise to the first wave of spermatogenesis .
The mammalian Sall4 gene encodes three alternative splice variants (SALL4A, SALL4B, SALL4C). SALL4A and SALL4B were recently shown to have isoform-specific target genes and to play different roles for pluripotency maintenance and differentiation of mouse embryonic stem cells . In this context we wanted to determine which of these isoforms is present during the establishment of the spermatogonial stem cell pool and the initiation of spermatogenesis in the postnatal mouse testis. Germ cells in sections from PND 0, 3, 7, and 14 were identified by staining for either DAZL or VASA, two ubiquitous germ cell markers, that were confirmed by double-immunohistochemistry to be expressed in all germ cells at the selected ages (data not shown). For co-staining experiments, we used a polyclonal antibody that recognizes both SALL4A and SALL4B in combination with anti-DAZL antibody, and we also used a monoclonal antibody that specifically recognizes SALL4A in combination with anti-VASA antibody. The data in Figure 1 indicate that while the SALL4A/B antibody stains DAZL+ germ cells on postnatal days 0, 3, 7 and 14 (Fig. 1A–D); the SALL4A-specific antibody stains VASA+ germ cells only on days 7 and 14 (Fig. 1E–H). These results were also confirmed by Western blot with the SALL4A/B antibody that distinguishes SALL4A and SALL4B isoforms based on size (Fig. 1I).
We next tested the hypothesis that SALL4A and SALL4B might be expressed in different subpopulations of spermatogonia in the developing mouse testis by co-staining with the SALL4A/B and SALL4A antibodies. However, the data in Figure 2 indicate a complete overlap of staining with these two antibodies on days 7 and 14 of development, corresponding to times when the spermatogonial stem cell pool is established, and spermatogenesis is initiated. Therefore, SALL4A and B do not appear to mark different subpopulations of spermatogonia at these stages of development. To confirm the spermatogonial subtypes that express SALL4 in postnatal day 7 and 14 testes, we co-stained with the SALL4A/B antibody and LIN28 or PLZF, which are established markers of undifferentiated spermatogonia. We observed nearly complete overlap of SALL4 expression with both LIN28 and PLZF (Fig. 3). Therefore, SALL4 expression appears restricted to LIN28+ and PLZF+ undifferentiated spermatogonia at these stages of mouse testis development.
To investigate the expression of SALL4 during steady-state spermatogenesis in the adult mouse testis, we co-stained tissue sections with SALL4 and the established spermatogonia markers, PLZF and LIN28, and the Sertoli cell marker GATA4 (Fig. 4). Co-expression analysis showed that SALL4 co-localizes with PLZF in 13.2 cells per tubular cross section (86.7%±1.9 overlap) and only occasionally spermatogonia were observed that expressed either SALL4 alone (0.6 cells/cross section; 3.8%±0.9) or PLZF alone (1.5 cells/cross section; 9.6%±1.3). Similarly, co-expression of SALL4 with LIN28 was found in 10.2 cells/cross section (66.3%±6.3), while 2.9 cells/cross section were SALL4 positive only (17.3%±4.7) and 2.7 cells/cross section were positive for LIN28 alone (16.5%±1.7). In contrast, SALL4 was not expressed in GATA4-positive Sertoli cells (16.6 cells/tubule cross section). Overall, 829 SALL4-positive spermatogonia, 883 PLZF-positive spermatogonia, and 778 LIN28-positive spermatogonia were observed per 1000 Sertoli cells in the adult testis based on these cell counts in adult mouse testicular sections (Fig. 4L). In summary, these data demonstrate that SALL4 is expressed by the majority of PLZF+ and LIN28+ undifferentiated spermatogonia during steady-state spermatogenesis.
To correlate SALL4 expression with spermatogonial clone size (As, Apr, Aal), we performed SALL4 immunohistochemistry in whole mount preparations of seminiferous tubules (Fig. 5A–D). We found that SALL4 positive cells were arranged as Asingles (31.4%±2.2 of SALL4 positive clones), Apairs (26.3±1.6%), chains of Aal4 (19.5±1.3%), Aal8 (8.1±0.9%), Aal16 (2.9±0.3%) and sometimes in longer chains of Aal>16 spermatogonia (1.3±0.3%) (Fig. 5E). Clones of 3 SALL4+ cells were occasionally observed. Otherwise odd clone sizes that did not correspond to population doubling divisions (2, 4, 8, 16) were rarely observed.
It was previously reported that the Aundiff population is heterogeneous , . In order to determine the distribution of SALL4 in spermatogonial clones in relation to other known markers, we used whole mount immunohistochemical co-staining (Fig. 6). In accordance with the data from tissue section immunostaining, we found that SALL4 expression substantially overlapped with both PLZF and LIN28. In particular, PLZF was co-expressed with SALL4 in 82.7±2.2% of As spermatogonia, 86.3±0.7% of Apr spermatogonia and 90.3±2.0% of Aal4 spermatogonia. Co-expression further increased in chains of eight and 16 cells (96.6±1.7% and 98.7±1.3%, respectively) and was complete in all clones observed with more than 16 cells. SALL4 and LIN28 were co-expressed in 77.9±3.2% of As and 88.3±1.1% of Apr spermatogonia. As with PLZF, co-expression of SALL4 and LIN28 increased in Aal4 and Aal8 (95.4±1.9% and 97.2±0.7%, respectively) and was complete in all chains containing 16 or more spermatogonia. Asymmetrical clones of Apr and Aal spermatogonia, as previously reported by others , , , were observed infrequently (<1% of clones, see Fig. 6D and H) and were counted as double stained clones because it is not possible to know whether the asymmetry is biologically relevant or an artifact caused by tubule preparation or staining procedure.
Thus, it appears that most Aundiff express SALL4, but some minor populations of SALL4-only, PLZF-only and LIN28-only expressing cells were present in the seminiferous epithelium. In this regard it is interesting to note that heterogeneity of expression was greater in As and Apr spermatogonia than in Aal spermatogonia. The biological significance of these subpopulations for spermatogonial self-renewal and differentiation remains unclear, but this observation is consistent to reports by others and will require further investigation , , .
The progression of spermatogonia from As to Aal is thought to be indicative of a cell fate shift towards differentiation and thus, the spermatogonial stem cell potential to self-renew is primarily attributed to As and sometimes Apr spermatogonia. We examined this phenomenon further by evaluating the co-expression of SALL4 with GFRα1, one of the most restricted markers for undifferentiated spermatogonia .
Glial cell line-derived neurotrophic factor (GDNF) is essential for SSC maintenance and self-renewal both in vitro and in vivo ,  and its receptor GFRα1 is expressed on a subset of undifferentiated spermatogonia . Here, we used double-immunofluorescence to investigate if SALL4 is present in the same spermatogonial population as this putative SSC marker (Fig. 7). As expected, GFRα1 expression was restricted primarily to As, Apr and some Aal4 spermatogonia. It was rarely observed in Aal8, and was never observed in chains larger than 8.
Substantial heterogeneity between SALL4 and GFRα1 was observed in As and Apr spermatogonia. In particular, 51.7±7.8% of As co-expressed both markers, but 30.8±7.4% of As were SALL4 positive only and the remaining 17.6±3.9% As spermatogonia expressed only GFRα1. Similarly, 44.3±7.2% of the Apr spermatogonia co-expressed SALL4 and GFRα1, 49.7±7.0% expressed SALL4 only and 6±0.3% expressed GFRα1 only. In Aal4, only 19.6±2.9% of the cells were SALL4+/GFRα1+ and the dominant phenotype was SALL4+ only (79.2±3.0%). GFRα1/SALL4 co-expression was observed in only 4% of Aal8 and GFRα1 expression was not observed in chains of 16 or more aligned spermatogonia, resulting in a singular SALL4+ phenotype. In rare events, asymmetrical Apr and Aal4 were observed (<1%, see Fig. 7).
Next, we investigated the expression of SALL4 in spermatogonia poised to undergo differentiation. The cKIT receptor tyrosine kinase is upregulated during the transition of undifferentiated spermatogonia into differentiating A1 spermatogonia and remains expressed during the amplifying pre-meiotic spermatogonial divisions . cKIT expression (SALL4+/cKIT+ and SALL4−/cKIT+) was occasionally observed in the smallest As (4.1%) and Apr (14.3%) clones, and expression increased markedly in Aal4, when 39.8% of spermatogonia expressed cKIT (Fig. 8). However, at this point the majority of Aal4 (60.3±8.7%) were SALL4+ only. At the stage of Aal8, cKIT expression became more prominent, with increasing numbers of SALL4+/cKIT+ clones (40.7%) and cKIT+ only clones (12.3%) being observed. Most Aal16 spermatogonia expressed cKIT (79.0%), but it is important to note that a substantial number of Aal16 were SALL4+/cKIT- (21.0±5.14%) (Fig. 8B). Taken together, these results indicate that SALL4 expression declines in larger Aundiff clones coincident with increasing cKIT expression. Asymmetrical clones were not observed.
The expression data from all whole mount immunostaining experiments are summarized in Figure 9.
SALL4 participates in transcriptional regulatory networks during fetal development of various organs  and is critical for cell fate decisions and lineage specification . In contrast to most organs, expression in the germline is maintained after birth  and is restricted to pre-meiotic germ cells in rodents, non-human primates and humans , . Although the exact function of SALL4 in spermatogonia is still unclear, increasing knowledge from studies of embryonic stem cells and hematopoietic stem cells may provide clues about possible mechanisms of SALL4 action in germ cells , , , . Here, we analyzed the expression pattern of SALL4 during spermatogenic lineage development and steady-state spermatogenesis in relation to clone size and established markers for undifferentiated and differentiating spermatogonia.
Our results show that SALL4 is expressed by a subpopulation of quiescent gonocytes at PND0, but is expressed by all gonocytes at PND3, which coincides with re-entry into the cell cycle and migration to the seminiferous tubule basement membrane . A similar pattern has been reported for neonatal and postnatal germ cells in the marmoset . The pool of perinatal gonocytes has two destinies; differentiate to produce the first round of spermatogenesis and establish the initial pool of spermatogonial stem cells , . Perhaps the heterogeneous SALL4 expression at PND0 distinguishes these two populations. However, by day 7 in our study, when gonocytes reach the basement membrane, SALL4 was expressed in all germ cells. Moreover, the observation that two SALL4 splice variants are expressed differentially during this specific time when the SSC pool is established may suggest that the SALL4A and SALL4B isoforms possess different regulatory functions in germ cells. To our knowledge, this is the first study that identified differential expression of SALL4 isoforms in undifferentiated spermatogonia. It was previously shown that SALL4A and SALL4B can form homo- or heterodimers with distinct DNA target regions regulating either pluripotency genes (SALL4A/B heterodimers, SALL4B/B homodimers) or differentiating/patterning genes (SALL4A/A homodimers) . Thus, the presence of SALL4B alone (in gonocytes), or SALL4A and SALL4B (in PND7 spermatogonia) may have implications for the regulation of target genes, perhaps influencing the fate toward stem cell or differentiation to produce the first round of spermatogenesis. However, we were not able to identify subpopulations of spermatogonia at PND7 or 14 with the phenotypes SALL4A only or SALL4B only; rather, all spermatogonia had the SALL4A/B co-stained phenotype. It is possible that the ratio of SALL4A to SALL4B changes on a per cell basis, but this could not be unequivocally confirmed by immunohistochemistry.
We then investigated SALL4 expression during steady-state spermatogenesis. Our data confirmed previous reports of SALL4 co-localization with PLZF, an established marker for undifferentiated spermatogonia, in tissue sections . We further show that SALL4 is almost completely co-expressed with a second spermatogonia marker, the RNA-binding protein LIN28. In contrast, GATA4-expressing Sertoli cells did not express SALL4. We observed 16.6 Sertoli cells per tubular cross section, which is comparable to data reported by Oatley and coworkers , and using this approach we calculated that approximately 800 SALL4-positive spermatogonia per 1000 Sertoli cells are present in the adult mouse testis. It is important to note that the size of the SALL4-positive spermatogonial population was similar to the population of PLZF-positive and LIN28-positive spermatogonia. Because of the extensive overlap between SALL4, PLZF, and LIN28 expression, and almost identical population size relative to Sertoli cell number, we propose that SALL4 is expressed in undifferentiated spermatogonia.
Studies in tissue sections do not provide information about the topographical arrangement of spermatogonial clones (As, Apr, Aal), which might be related to stage of differentiation. It is therefore important to correlate the molecular phenotype of spermatogonia with their clonal organization by whole-mount seminiferous tubule immunofluorescence staining. Using this approach, we observed SALL4 in As, Apr and Aal4,8,16 spermatogonia. Occasionally, clones with an odd number of cells (Aal3, Aal6, Aal12) were observed. These could be the result of clone fragmentation, a phenomenon that has been described by Nakagawa et al.  and that was proposed as a mechanism to replenish the spermatogonial pool. Further, extensive co-localization studies in whole-mount samples revealed significant co-expression of SALL4 with PLZF and LIN28 in the majority of Aundiff spermatogonia.
Germ cell differentiation is regulated by endocrine and paracrine factors that signal either to the somatic component of the seminiferous tubules, the Sertoli cells, or directly to spermatogonial stem cells to elicit self-renewing or differentiation divisions. A key factor for SSC survival and renewal both in vivo and in vitro is glial cell line-derived neurotrophic factor (GDNF) ,  that is secreted by Sertoli cells  and binds to the RET/GFRα1 receptor complex  on undifferentiated spermatogonia including SSCs . Accordingly, the presence of GFRα1 on the cell surface of putative SSCs is well characterized and was found to be restricted predominantly to As, Apr and Aal4 . In contrast to the PLZF and LIN28 results, we observed a high degree of heterogeneity among As spermatogonia in SALL4 and GFRα1 co-staining experiments. Grisanti et al. reported that 10% of the PLZF-expressing As spermatogonia are GFRα1-negative and that the stem cell activity in this population is greater than in GFRα1 positive cells, as determined by transplantation assay . In the study mentioned above, Grisanti et al. also observed that ~5% of PLZF-positive Apr spermatogonia had asymmetrical GFRα1 expression. In our study, we rarely observed asymmetrical clones (either SALL4 asymmetric or GFRα1 or LIN28 asymmetric, <1% of clones). The reason that we observed slightly fewer asymmetric clones than previously reported could be that Grisanti et al. made counts from images, while we counted clones directly on the microscope, which allowed us to follow clones through different focal planes.
Here, we found that only half of the As population co-expressed GFRα1 and SALL4, while the other half consisted of SALL4 only or GFRα1 only cells. The biological function and the actual stem cell potential in these different spermatogonial populations is not known and will have to be investigated in future studies, perhaps using transgenic tools.
We next performed co-staining experiments for SALL4 and the cKIT receptor tyrosine kinase, which is a marker for differentiating germ cells. cKit becomes upregulated as Aal transition to A1 spermatogonia, possibly through de-repression of cKit gene expression in the absence of PLZF . In this study, we observed cKIT expression in a minor fraction of As and Apr, which increased in larger clones and was associated with a corresponding decrease in SALL4 expression in larger clones. Interestingly, a very small proportion of Apr and Aal4 clones expressed cKIT only (without SALL4), suggesting that the Aal to A1 transition might sometimes occur from small clones, as previously described in the Chinese hamster . While the majority of Aal16 clones expressed cKIT (either with or without SALL4), we did observe Aal16 SALL4+ clones that were cKIT negative.
Hobbs et al. recently demonstrated a direct interaction of SALL4 and PLZF in transfected HEK293 cells and reported that the two factors can displace each other from their cognate chromatin locations depending on their relative protein levels . PLZF is a repressor of the differentiation factor, cKIT, and the authors proposed that SALL4 might act as an important spermatogonial differentiation factor that functions by removing PLZF from its cognate targets (e.g. cKIT). Does this suggest that the SALL4+/PLZF- As and Apr spermatogonia express cKIT and are destined to differentiate? In this regard, triple staining experiments would be very informative, but are difficult to execute due to availability of compatible antibodies and differences in optimal staining conditions for each marker.
Taken together, our data from cross-section and whole mount preparations of seminiferous tubules corroborate earlier reports of molecular heterogeneity among Aundiff spermatogonia, which may indicate functional heterogeneity among clone sizes , , , . The heterogeneity can only be observed when co-staining with more than one marker is performed, therefore caution is required for the interpretation of experiments that use only a single marker as a denominator for Aundiff spermatogonia. In this study, we report that SALL4 is a molecular marker for undifferentiated As, Apr and Aal spermatogonia, a population that includes spermatogonial stem cells, and that it can be used in combination with other molecular markers to dissect spermatogonial lineage development and steady state spermatogenesis in the postnatal testis.
Immunostaining of tissue sections with anti-SALL4 and anti-DAZL IgG. Fluorescent signal was specific for SALL4 and DAZL (A) and no fluorescent staining signal was observed in Isotype controls (B). Scale bar=10 µm.
Immunostaining of adult mouse testis sections with spermatogonial markers. Fluorescent signal was specific for SALL4, PLZF and LIN28 (A, C) and no fluorescent staining signal was observed in Isotype controls (B, D). Scale bar=10 µm.
Immunostaining and controls of whole mount seminiferous tubules. Fluorescent signal was specific for SALL4, PLZF, LIN28, GFRα1, or cKIT (left column) and no fluorescent staining signal was observed in Isotype controls (right column). Scale bar=50 µm.
The authors would like to thank Hitomi Suzuki for support with cKIT staining in whole mount preparations of seminiferous tubules and Mahesh Raundhal for assistance with immunofluorescence staining. Tissue embedding and sectioning was performed in the Histology Core of Magee-Womens Research Institute. We are grateful to Julia Donohue and Brandon Chaney for help with animal breeding.
Financial support was provided by United States National Institutes of Health grants HD055475, HD061289 and HD008610 (KEO), a Lalor Foundation Postdoctoral Fellowship (KG) and Magee-Womens Research Institute and Foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.