(A), Genotype-phenotype map to visualize deletion of AZF a, b and c regions in hESC lines, patient samples and patient-derived iPSCs used for this study. A deletion map was constructed for every hESC line, patient fibroblast and iPSC line by testing for the presence of 20 major Sequence-Tagged-Sites (STS) sites by PCR. Vertical black bars (top) represent the STS amplification sites and genes we used to diagnose AZFa, b and c deletions. Grey boxes represent the deleted regions of the Y chromosome. The fertility phenotype (SCO = Sertoli cell only; EMA = early maturation arrest) of each patient is listed (left) and karyotype (right). The Δ symbol indicates the deletion of an AZF region in that cell line. (B), Immunocytochemistry in all patient-derived iPSC lines for nuclear (OCT4, NANOG) and cell surface (SSEA1/3/4, TRA1–60, TRA1–81) markers of pluripotency. (C), In vitro differentiation of 4 patient-derived lines to cells representative of all three germ layers (AFP= α-fetoprotein, βIII-Tub = β-III-tubulin/Tuj1, SMA = smooth muscle actin). (D), In vivo teratoma formation of iAZFΔbc, iAZFΔc & iAZFΔa lines showing evidence of all 3 germ layers: endoderm (e), gut-like epithelium (ep), mesoderm (m), smooth muscle-like mesoderm (sm), cartilage-like (c), ectoderm (e) and neural ectoderm (ne). See also .

(A), Single cell gene expression of VASA-GFP+ iPGCs derived from undifferentiated AZF-intact iPSCs (iAZF1), hESCs (H1) and AZF-deleted iPSCs (iAZFΔc, iAZFΔbc, iAZFΔa). The relative expression of PGC genes PRDM1, PRDM14, DAZL, STELLA (DPPA3), IFITM3, NANOS3 and VASA was measured after normalization in all cells positive for GFP mRNA and plotted individually. Expression levels in undifferentiated single cells from each iPSC line are also shown in each graph. Each vertical bar represents one individual cell and at least 25 cells are represented for differentiated samples and 20 cells for undifferentiated samples. See legend colors for sample identification. A heat index of pluripotency gene expression in 5 differentiated single cells and in undifferentiated iPSCs is shown (right). (B), Immunocytochemistry of GFP+ sorted cells for GFP and PGC proteins VASA, STELLA and DAZL. Nuclei are counterstained with DAPI (blue). Scale bar, 25 µm. (C), Two-component Linear discrimination analysis of all single cell RT-PCR measurements for 27 germ cell genes from all iPSC lines tested. Each single dot represents one single cell measurement (D), Percentage of GFP+ sorted single cells co-expressing 12 additional germ cell-associated genes (see ) as plotted by the number of markers within each individual cell. Dashed, vertical lines indicate the segment of the population of single cells from each line relative to the middle 70–80% of control, iAZF1 population. Co-expression of 9 PGC (blue) and 6 gonocytic/spermatogonial genes (red) was quantified in GFP+ iPGCs and plotted by the number of respective markers within each cell. See also .

(A) Transplantations were performed by independently injecting 22-week old human fetal testicular cells, H1 hESCs or human iPSCs directly into seminiferous tubules of busulfan-treated mouse testes (pink oval) that were depleted of endogenous germ cells. Testis xenografts were analyzed by: 1) Whole mount histology using a human cell-specific antibody or 2) immunohistochemistry using a human cell-specific antibody, NuMA to detect donor cells. (B) Haemotoxylin and eosin histology of normal mouse recipient testes before busulfan treatment, after busulfan-mediated germ cell depletion and 2 months post-transplantation of iPSCs. (C) (i) Human fetal testis (22 week-old) with cytoplasmic VASA expression (green) in germ cells. Nuclei are counterstained with DAPI. Human fetal testis was enzymatically dissociated, fetal germ cells were transplanted and then detected after 2 months by (ii) whole mount staining or (iii) cross-sectional immunohistochemistry for NuMA+/VASA+ cells (white asterisks). (D) Human and mouse immunostaining controls for NuMA antibody specificity. (E) Imunohistochemical analysis of testes xenografts derived from undifferentiated (i–ii) H1 hESC and (iii–iv) iAZF1 iPSCs. All images are merged from NuMA (red), VASA (green) and DAPI-stained nuclei. White rectangles indicate regions of interest of positive co-localization of NuMA and VASA and magnified in panels ii and iv. Red asterisks indicate NuMA+ donor cells adjacent to the basement membrane while white asterisks represent NuMA+/VASA+ donor cells near to the basement membrane. In all panels, dashed white lines indicate the outer edges of spermatogonial tubules. Scale bar, 50µm. See also .

Mouse testis xenografts were analyzed by immunohistochemistry using the human cell-specific antibody, NuMA and an antibody recognizing the pan-germ cell marker, VASA, marking PGCs, gonocytes and spermatogonia. Cross-sections of testes xenografts derived from (A) 22-week-old human fetal testis cells; (B) iAZF1; (C) iAZF2; (D) iAZFΔc; (E) iAZFΔbc and (F) iAZFΔa cell lines. All images are merged from NuMA (red), VASA (green) and DAPI-stained nuclei. Arrows indicate cells or clusters with positive co-localization of NuMA and VASA. Red asterisks indicate NuMA+ donor cells adjacent to the basement membrane while white asterisks represent NuMA+ donor cells near to the basement membrane. In all panels, dashed white lines indicate the outer edges of spermatogonial tubules. Scale bar, 50µm. (G), Percentage of tubules positive for NuMA+/VASA+ cells were calculated across multiple cross-sections (relative to total number of tubules). (H), For each positive tubule, the ratio of NuMA+/VASA+ cells per tubule was determined. (I), Relative germ cell forming potential calculated by multiplying fraction of positively stained tubules with number of VASA/NUMA co-stained cells for each sample. Data are represented as mean +/− SD of replicates. In each graph, significant differences in percentages/ratios between controls (iAZF1 or iAZF2) and AZF-deleted lines were determined by one-way analysis of variance (ANOVA). *, P<0.05; **, P<0.001, ns, non-significant. See also .

Mouse testes xenografts derived from (A) H1 hESCs (B) 22-week old Human fetal testicular cells, (C) iAZF1 and (D–F) AZF-deleted human iPSCs were stained in adjacent, serial cross sections for the PGC markers DAZL and STELLA and for the gonocyte/spermatogonial markers UTF1, PLZF and DAZ in NuMA+VASA+ regions. For each xenograft, low-magnification regions of the seminiferous tubule containing NuMA+VASA+ cells are shown and boxed regions of interest are enlarged and shown to the right. In NuMA+/VASA+ donor cells, immunostaining for germ cell proteins was shown as follows: Panel 1, Positive immunostaining for DAZL (red) and STELLA (green); Panel 2, UTF1 (green), UTF1+DAZL, PLZF (green) or PLZF+VASA. (G), Mouse testes xenografts derived from H1 hESCs, iAZF1 and human fetal testicular cells showing NuMA+ cells that co-expressed DAZ in boxed regions and enlarged in lower panels. In all panels, dashed white lines always indicate the outer edges of spermatogonial tubules. Wherever shown, nuclei are counterstained with DAPI (blue). Scale bar, 50 µm. (H) Summary of positive (green) and negative (red) germ cell protein expression across all xenografts tested. See also .

(A) Human fetal (22wk) testis, (B) human adult testes and (C–G) mouse testis xenografts derived from human fetal testicular cells and iPSCs were stained for VASA and 5-methylcytosine (5-mC) in NuMA+ regions. (A), Cross-section of a human fetal testis (22 wks) with positive immunostaining for 5-mC alone or VASA and 5-mC together. Boxed region in right panel is enlarged in two separate panels to the left of the main panel. (B), Cross-section of a human adult testis with positive immunostaining for 5-mC alone or VASA and 5-mC together. (C–G) Cross-sections of mouse testes transplanted with 22-week old Human fetal testicular cells (C) and undifferentiated iPSC donor cells from each line (D–G). For each xenografted line, immunostaining for NuMA+/VASA+ cells (top), 5-mC alone (middle) and 5-mC together with VASA (bottom) is shown. White dotted circles indicate NuMA+/VASA+ cells with complete loss of 5-mC signal; yellow arrowheads indicate NuMA+/VASA+ positive cells with positive 5-mC signals. (H), Positive 5-mC signals in NuMA+ cells outside tubules (top) and in undifferentiated iPSCs prior to transplantation (bottom). (I), Quantification of 5-mC positive and negative signals in Human testes (left columns) and in NuMA+/VASA+ cells of testes xenografts (middle columns). 5-mC patterns in extratubular NuMA+ cells is shown (right column). Scale bar, 50 µm. In all panels, dotted lines indicates edges of spermatogonial tubules. See also .

Schematic summarizing major findings of this study. iPSCs were derived from AZF-intact and AZF-deleted patients using the OCT4, SOX2, KLF4 and CMYC cocktail (OSKM). When transplanted, undifferentiated male iPSCs specifically differentiate to PGC-like and gonocyte-like germ cells inside the mouse spermatogonial tubule environment, presumably in the Sertoli cell niche. However, outside the tubule, all patient-derived iPSCs and hESCs remain undifferentiated as primitive tumors. AZF-deleted iPSCs appear to have a lower potential to make germ cells in vivo as compared to AZF-intact iPSCs and appear to be restricted to forming PGC-like cells. See also .