Germline Stem Cell Activity Is Sustained by SALL4-Dependent Silencing of Distinct Tumor Suppressor Genes

Abstract

Sustained spermatogenesis in adult males and fertility recovery following germ cell depletion are dependent on undifferentiated spermatogonia. We previously demonstrated a key role for the transcription factor SALL4 in spermatogonial differentiation. However, whether SALL4 has broader roles within spermatogonia remains unclear despite its ability to co-regulate genes with PLZF, a transcription factor required for undifferentiated cell maintenance. Through development of inducible knockout models, we show that short-term integrity of differentiating but not undifferentiated populations requires SALL4. However, SALL4 loss was associated with long-term functional decline of undifferentiated spermatogonia and disrupted stem cell-driven regeneration. Mechanistically, SALL4 associated with the NuRD co-repressor and repressed expression of the tumor suppressor genes Foxl1 and Dusp4. Aberrant Foxl1 activation inhibited undifferentiated cell growth and survival, while DUSP4 suppressed self-renewal pathways. We therefore uncover an essential role for SALL4 in maintenance of undifferentiated spermatogonial activity and identify regulatory pathways critical for germline stem cell function.

Keywords: SALL4; germline stem cells; self-renewal; spermatogonia; transcription factors; tumor suppressor genes.

Graphical abstract

Figure 1

Expression of SALL4 in Spermatogonia of Undisturbed and Regenerating Testis (A) Schematic illustrating mouse seminiferous epithelium, spermatogonial hierarchy, and markers of populations. (B–D) Representative whole-mount IF of wild-type (WT) adult seminiferous tubules. Inset in (C) shows low KI67 in SALL4+ A_al–A₁ (arrowheads). Arrowheads in (B) and (D) indicate A_al cysts. Insets in (D) show SALL4 in GFRα1+ A_s and A_pr. Dashed outlines indicate SALL4+ cysts. (E) Representative whole-mount IF of untreated and busulfan-treated WT mice. Images were taken along the tubule then stitched together. Grayscale of each channel within the indicated regions are shown. (F) SALL4 staining intensity from (E) using ImageJ. For controls, SALL4 was measured in GFRα1+ A_s, A_pr, and A_al<4, and GFRα1− A₁–A₂. For busulfan-treated mice, SALL4 was measured in GFRα1+ A_al>4. Mean values ± SEM are shown (n = 4 mice per condition). At least 100 cells were analyzed from controls and 40 from busulfan-treated mice. ^∗p < 0.05. (G) Representative whole-mount IF demonstrating mutually exclusive GFRα1 and RARγ expression in cysts of regenerating tubules from (E). Scale bars, 50 μm. Dotted lines indicate tubule profile. See also Figure S1.

Figure 2

Effects of Acute Sall4 Deletion on Spermatogonial Populations In Vivo (A) Representative whole-mount IF of Z/EG; UBC-CreER tubules 7 days after TAM. (B) Adult Ctrl and Sall4^TAM−KO mice were treated with TAM and harvested at the indicated time points. Lower panels: representative whole-mount IF of seminiferous tubules 7 days post-TAM. (C) Representative whole-mount IF of seminiferous tubules 5 and 7 days post-TAM. Three mice per genotype were analyzed at 5 days and seven per genotype at 7 days. Day 7 control tubules are shown. PreL denotes preleptotene spermatocytes. (D) Representative whole-mount IF of tubules 7 and 14 days post-TAM. Seven mice per genotype were analyzed. Day 14 control tubules are shown. Insets demonstrate PLZF localization and DAPI identifies nuclei. (E and F) Flow cytometry of testis cells from Ctrl and Sall4^TAM−KO mice 7 days post-TAM. Populations no. 1 are undifferentiated cells, no. 2 early differentiating cells, and no. 3 late differentiating cells. Graph shows mean percentage of cells in each population ± SEM. Four mice per genotype were analyzed. ^∗∗∗p < 0.001. (G) Representative flow cytometry of EdU incorporation by undifferentiated cells (PLZF+ c-KIT−) from Ctrl and Sall4^TAM−KO mice 14 days post-TAM as in (E). Only SALL4− cells are included from Sall4^TAM−KO. Percentages of cells EdU+ are indicated. (H and I) Representative whole-mount IF of seminiferous tubules 7 and 14 days post-TAM. Day 7 control tubules are shown. Arrowheads in Sall4^TAM−KO indicate Sall4 null progenitor cysts. Scale bars, 50 μm. Dotted lines indicate tubule profiles. See also Figures S2 and S3.

Figure 3

SALL4 Is Required for Long-Term Maintenance of Spermatogonial Stem Cell Activity (A) Representative whole-mount IF of seminiferous tubules from Ctrl and Sall4^TAM−KO mice 7 and 30 days post-TAM. Images were taken along the tubule and stitched together. Seven mice per genotype were analyzed at 7 days and three per genotype at 30 days. Day 7 control tubules are shown. Regions of germ cell-depleted and recovering Sall4^TAM−KO tubules at 30 days are indicated. Dashed outline in “b” indicates a GFRα1+ KI67+ Sall4-deleted regenerative A_al. Arrowheads in “d” mark a Sall4-retaining GFRα1+ A_s. Dotted lines indicate tubule profiles. (B) Quantification of GFRα1+ cells/cysts that were A_al4 and A_al8 from the 7-day time point of (A). Three mice per genotype were analyzed and >2 cm of tubules scored per sample. Mean values ± SEM are shown. (C) Quantification of PLZF+ cells expressing SALL4 in Ctrl and Sall4^TAM−KO testis sections at indicated time points post-TAM. Three mice per genotype were analyzed at each time point and 100 PLZF+ cells scored per sample. Mean values ± SEM are shown. (D) Testis weights of Ctrl and Sall4^TAM−KO adult mice 30 days post-TAM. Horizontal bars represent the mean. Three Ctrl and five Sall4^TAM−KO mice were analyzed. (E) Representative IF of Ctrl and Sall4^TAM−KO testis sections 30 days post-TAM. Inset shows DAPI stain of a tubule portion. (F) Percentage of PLZF/SALL4+ cells from Ctrl testis sections and PLZF+ SALL4− (KO) cells from Sall4^TAM−KO sections KI67+ 14 and 30 days post-TAM. Mean values ± SEM are shown. Four mice per genotype were analyzed at each time point and 100 PLZF+ cells scored per sample. ns, not significant. (G) Representative IF of Ctrl and Sall4^TAM−KO testis sections 7 and 30 days post-TAM. Arrowheads in insets refer to GFRα1+ cells scored as KI67+ and KI67−. Three mice per genotype were analyzed at each time point. (H) Percentage of GFRα1+ cells KI67+ from (G) scored as in (F). Mean values ± SEM are shown. Three mice per genotype were analyzed at each time point and ≥30 GFRα1+ cells scored per sample. ^∗∗p < 0.01, ^∗∗∗p < 0.001; not significant (ns) p > 0.05. Scale bars, 50 μm. See also Figure S4.

Figure 4

Analysis of Plzf KO and Plzf/Sall4 Double KO Spermatogonial Phenotype (A) Representative whole-mount IF of seminiferous tubules from WT and Plzf^−/− mice. Images taken along the tubule were stitched together. Areas of intact and germ cell-depleted (degenerated) seminiferous epithelium are indicated. (B–D) Representative whole-mount IF of Plzf^−/− and Plzf^−/−; Sall4^TAM−KO tubules 10 days post-TAM. Three Plzf^−/− and four Plzf^−/−Sall4^TAM−KO mice were analyzed. Arrowheads in (C) indicate Sall4-expressing (upper panels) and Sall4 KO (lower panels) undifferentiated cells. Dashed line in (D) indicates regenerative GFRα1+ A_al. Scale bars, 50 μm. Dotted lines indicate tubule profiles.

Figure 5

Identification of SALL4 Targets Using Cultured Undifferentiated Spermatogonia (A) Method for generating cultures of undifferentiated spermatogonia. Spermatogonia were enriched from Sall4^TAM−KO testis cell suspensions by αCD9 magnetic selection and plated onto mitotically inactivated mouse embryonic fibroblasts (MEF). Colonies formed within 1–2 weeks and passaged cells were treated with vehicle (Veh) or 4-hydroxytamoxifen (TAM) then analyzed 4 days later. (B) Representative IF of Veh and TAM-treated Sall4^TAM−KO cultures as in (A). Grayscale insets show PLZF localization. (C) Representative images of Veh and TAM-treated Sall4^TAM−KO cultures as in (B). Insets show higher magnification of indicated areas. Mean numbers of cleaved-PARP+ cells per colony are in upper graph. Fifty colonies scored per culture and condition. Lower graph shows relative KI67 levels quantified using ImageJ. One hundred cells scored per culture and condition. Mean values ± SEM from three independent cultures are shown. (D) qRT-PCR of SALL4-regulated genes from microarray analysis. Sall4^TAM−KO cultures were treated with Veh or TAM as in (B). mRNA levels are corrected to b-actin and normalized to Veh-treated sample. Levels of germline marker Ddx4/Vasa and undifferentiated cell marker Pou5f1/Oct4 are included as controls. Five independent cultures were analyzed. Mean values are shown ± SEM. (E) Western blot of three independent Sall4^TAM−KO cultures treated as in (B). Molecular weights (kDa) are indicated. VASA and β-ACTIN were used as loading controls. (F and G) Representative IF of Sall4^TAM−KO cultures treated as in (B). Dashed line indicates Sall4-deleted cells expressing high DUSP4 levels. (H) Representative flow cytometry of Ctrl and Sall4^TAM−KO testis cells 7 days post-TAM. Percentages of EpCAM+ cells within c-KIT− α6-integrin+ undifferentiated gate are indicated. (I) qRT-PCR of SALL4-regulated genes in undifferentiated cells isolated as in (H). mRNA levels are corrected to b-actin and normalized to Ctrl. Ddx4/Vasa and Pou5f1/Oct4 are included as controls. Mean values ± SEM are shown (n = 6 mice per genotype). ^∗p < 0.05; ^∗∗p < 0.01, ^∗∗∗p < 0.001. Scale bars, 50 μm. See also Figure S5 and Table S1.

Figure 6

SALL4 Targets Regulate Activity of Cultured Undifferentiated Spermatogonia (A–C) Analysis of SALL4 binding to Foxl1 (A), Dusp4 (B), and Sall3 (C) in WT cultured undifferentiated cells by ChIP-qPCR. Top panels depict promoter regions and first exon (E1) of genes. Arrows indicate transcription start sites (TSS) and red bars ChIP amplicons. Blue lines are SALL4 binding motifs and green lines PLZF motifs from cultured spermatogonia. Orange lines are SALL4 motifs from ESCs. Graphs show relative enrichment of amplicons from four independent lines normalized to negative control region of H1foo not targeted by SALL4 (-ve Ctrl). IgG controls are included. Mean values ± SEM are shown. (D) Identification of SALL4 interacting proteins in undifferentiated spermatogonia. Lysates were incubated with magnetic beads conjugated to SALL4 antibody or IgG (control). Immunoprecipitated proteins were identified by MS. (E) Summary of SALL4-associated proteins from three combined runs from (D). (F) Confirmation of SALL4-interacting proteins by SALL4 IP from WT cultures and WB. Non-specific IgG control is shown. (G) WT cultured undifferentiated cells infected with lentivirus containing Myc-tagged Foxl1 and Dusp4. Cells were passaged 2 days after infection and allowed to form colonies for 10 days before analysis. Infection efficiency was 30%–40%. Representative IF of infected cells is shown. Insets: uninfected control cells. (H) Relative mean colony size of infected cells (+) versus uninfected cells (−) in cultures from (G). Mean values ± SEM are shown (n = 4 lines of infected cells). GFP-infected cells were included as controls. (I) Representative IF of WT cultures infected with lentivirus containing Myc-tagged Foxl1 or GFP control constructs as in (G). Cells were stained for Myc-tag or GFP to confirm infection (inset). (J) Relative mean KI67 levels in infected cells from (I) quantified with ImageJ. One hundred cells were quantified from each of six infected WT cultures. Mean values ± SEM are shown. (K) Representative IF of WT cultures infected with lentivirus containing Myc-tagged Foxl1 and GFP control constructs as in (G). Cleaved caspase-3 staining identified apoptotic cells. (L) Graph indicates mean number of cleaved caspase-3+ cells per infected spermatogonial colony ± SEM from cultures of (K). One hundred colonies from each of three infected WT cultures were analyzed. (M) WB of independent WT cultures (no. 1 and no. 2) infected with lentivirus containing Myc-tagged Dusp4 and GFP as in (G). VASA and β-ACTIN were used as loading controls. (N) Graph indicates relative levels of P-JNK from (M) corrected to total JNK and normalized to GFP-infected cells. Four independent WT cultures were analyzed. Mean values ± SEM are shown. ^∗p < 0.05; ^∗∗p < 0.01, ^∗∗∗p < 0.001. Scale bars, 50 μm. See also Figure S6.

Figure 7

SALL4-Dependent Pathways Maintaining Undifferentiated Cell Activity SALL4 silences Foxl1 and Dusp4 by binding promoter regions and recruiting the NuRD co-repressor. FOXL1 can inhibit spermatogonial proliferation and survival via multiple targets. DUSP4 inhibits JNK, which is required for self-renewal downstream reactive oxygen species (ROS) and non-canonical WNT. Upon Sall4 deletion, FOXL1 and DUSP4 accumulate to block proliferation plus survival and suppress JNK-dependent self-renewal, resulting in progressive stem cell failure.