Somatic cell fate maintenance in mouse fetal testes via autocrine/paracrine action of AMH and activin B

Abstract

Fate determination and maintenance of fetal testes in most mammals occur cell autonomously as a result of the action of key transcription factors in Sertoli cells. However, the cases of freemartin, where an XX twin develops testis structures under the influence of an XY twin, imply that hormonal factor(s) from the XY embryo contribute to sex reversal of the XX twin. Here we show that in mouse XY embryos, Sertoli cell-derived anti-Mullerian hormone (AMH) and activin B together maintain Sertoli cell identity. Sertoli cells in the gonadal poles of XY embryos lacking both AMH and activin B transdifferentiate into their female counterpart granulosa cells, leading to ovotestis formation. The ovotestes remain to adulthood and produce both sperm and oocytes, although there are few of the former and the latter fail to mature. Finally, the ability of XY mice to masculinize ovaries is lost in the absence of these two factors. These results provide insight into fate maintenance of fetal testes through the action of putative freemartin factors.

Fig. 1. Loss of Amh/Inhbb in the XY mice leads to intersex phenotypes with ovotestis that produce oocytes and sperm in adulthood.

A, B The external and internal genitalia of the five months old control and dKO XY mice. C, D Hematoxylin & Eosin staining of a section of control and dKO XY gonads at five months of age. Scale bar = 200 µm. E, G Higher magnification of the dotted lined area in C, D. F, H Immunofluorescence for DMRT1 (Sertoli cells, blue), GCNA1 (germ cells, green), and FOXL2 (granulosa cells, pink) in control and dKO XY gonads. Stars and arrows mark seminiferous tubules and follicles, respectively. Scale bar = 200 µm. I–L Sperm and oocytes were recovered from the control XY (sperm only, n = 2/2), control XX (oocytes only, n = 4/4) and dKO XY mice (both sperm and oocyte, n = 2/4) 46 h after PMSG stimulation. Scale bar = 25 µm.

Fig. 2. Amh/Inhbb dKO XY testis forms normally at E12.5 but becomes ovotestis by E15.5.

A–C Immunofluorescence for SOX9 (Sertoli cells, yellow), GCNA1 (germ cells, blue), and FOXL2 (granulosa cells, pink) and D–F LAMININ (basal membrane, green), NR2F2 (interstitial cells, blue), and FOXL2 (granulosa cell, pink) on control XY (n = 4), dKO XY (n = 4), and control XX (n = 4) gonads at E12.5. A–C Dash lines highlight the boundary between the gonad and mesonephros. Scale bars = 25 um. G Principal Component Analysis (PCA) of the transcriptome of whole gonads at E12.5. H Volcano plot shows differentially expressed genes between the control XY and dKO gonads at E12.5 (2-fold, one-way ANOVA, FDR-corrected p-value < 0.05). I Relative mRNA expression level for genes critical for sex determination of the XY gonads (Sox9 and Fgf9) and XX gonads (Fst and Rspo1), N = 4 per group. Different letters indicate statistical difference (one-way ANOVA, Tukey’s means separation test, p < 0.001). J–O Immunofluorescence for DMRT1 (Sertoli cells, green), GCNA1 (germ cells, blue), and FOXL2 (granulosa cells, pink). M–O higher magnification of inset shown in J–L. P–R LAMININ (basal membrane, green), NR2F2 (interstitial cells, blue), and FOXL2 (granulosa cells, pink) on control XY (n = 4), dKO XY (n = 4), and control XX gonads (n = 4) at E15.5. S Principal Component Analysis (PCA) of the transcriptome of whole gonads at E15.5. T Volcano plot shows differentially expressed genes between the control XY and dKO XY gonad at E15.5 (2-fold, one-way ANOVA, FDR-corrected p-value < 0.05). U Relative mRNA expression level for genes in the XY gonads (Sox9 and Dhh) and XX gonads (Fst and Rspo1) N = 4 per group. Different letters indicate statistical difference (one-way ANOVA, Tukey’s means separation test, p < 0.001). I, U depict means ± SE of log intensity values. Source data are provided as a Source Data file. All data were collected from biologically independent samples.

Fig. 3. Single cell mRNA sequencing analysis of control XX (ovary), control XY (testis) and dKO XY (ovotestis) at E15.5.

A, B Genotype distribution and cell type clustering based on unbiased clustering of transcriptomic similarity. Thirteen cell clusters were identified. C Dot plots of the expression of genes used to identify each cluster. The rectangle highlights both XY and XX supporting cell populations. D Cell type distribution of the expression of Xist (X or female marker), Uty (Y or male cell marker), Sertoli cell marker (Dhh) and granulosa cell marker (FoxL2). Source data are provided as a Source Data file. N = 2 biological independent samples per genotype.

Fig. 4. Feminization of the dKO XY supporting cells at the single cell level.

A Single cell mRNA sequencing analysis and clustering of the supporting cells based on their genotypes: control XY (blue), dKO XY (green), and control XX (pink) gonads at E15.5. Each dot represents the transcriptome of a single cell. B Cell clustering based on their unique transcriptomes: one population in the XY control, two subpopulations in the dKO XY, and three subpopulations in the control XX supporting cell populations. C Trajectory of transcriptomic changes (pseudotime analysis) among supporting cells. Control XY supporting cell in the far left was arbitrarily designated as the starting point (black dot) of the differentiation. Arrows and the gradient of colors (yellow to dark blue) indicate the direction and progression of transcriptomic changes. D Transcriptome trajectory of six representative genes critical for sex determination of the XY gonads (Dhh and Dmrt1) and XX gonads (Runx1, Wnt4, Rspo1, and Foxl2). The X axis represents the arbitrary trajectory of transcriptome changes from a single control XY supporting cells toward control XX supporting cells (dark blue). Y axis indicates the gene expression level in each cell. The black lines represent the average expression of the genes over the progression. Source data are provided as a Source Data file. N = 2 biologically independent samples per genotype.

Fig. 5. Autocrine/Paracrine action of AMH and activin B on Sertoli cells.

A Statistical prediction of cell-cell communication via ligand and receptor interaction among germ cells, Sertoli cells, and Leydig cells from the single cell RNAseq dataset using CellPhoneDB. Each square in the 3 × 3 grid represents interaction between two cell types (germ vs germ, Sertoli vs Leydig, etc.). The number inside the square is the number of ligand/receptor pairs. Interactions for AMH, activin B and their receptors were significantly enriched among Sertoli cells (black outlined square; source data are provided as a Source Data file; n = 2 biologically independent samples per genotype). B Dot plots of the expression of Amh, Inhbb, and their receptors in germ cells, Sertoli cells, and Leydig cells. C Proposed model of the autocrine/paracrine action of AMH and Activin B. Sertoli cells secrete AMH and Activin B which act on an autocrine and paracrine way to block the female fate in the wild type Sertoli cells. In the absence of Amh and Activin B in the dKO, the cells become feminized at the testis poles.

Fig. 6. XY environment without Amh and Inhbb fails to masculinize the fetal ovary.

A Mouse freemartin model was created by transplanting wildtype E13.5 ovaries under the kidney capsule of adult (five to seven months old) wild type, single Amh, single Inhbb KO, or dKO XY mice. Transplants were recovered 21 days after. B Immunofluorescence of Sertoli cell marker SOX9 (green) and granulosa cell marker FOXL2 (pink) was used to classify the transplanted tissue as testis (SOX9 + /FOXL2-), ovotestis (SOX9 + /FOXL2 + ), or ovary (SOX9-/FOXL2 + ). Asterisks and circles indicate seminiferous tubules and follicles, respectively. C The numbers and percentage of fetal ovary transplants that were masculinized with testis or ovotestis structure (blue) or remained as ovary (pink) in controls, Amh single knockout, Inhbb single knockout, or dKO XY recipients. All data from biologically independent samples. Scale bar = 25 µm. Source data are provided as a Source Data file. Mouse image was created with BioRender.com.