Testis Molecular Pathways in CAIS Unveil Testosterone/Estradiol on Germ Cell Tumor Risk in Non-Obstructive Azoospermia

Abstract

Context: Non-obstructive azoospermia (NOA) is the most severe form of male infertility, affecting 1% of all men, with a clinical picture characterized by no sperm production, hyalinization of the basal membrane of the seminiferous tubules, primary hypogonadism, and earlier onset of age-related comorbidities compared with fertile men. NOA is also characterized by etiologic heterogeneity and the non-genetic form has higher incidence of testicular germ cell cancer (TGCC) compared to the forms with genetic abnormalities.

Objective: We aimed to establish molecular pathways in the testicular somatic cells that are either shared or specific for non-genetic and genetic forms of NOA, such as complete androgen insensitivity syndrome (CAIS) and Klinefelter syndrome (KS).

Methods: We performed single-cell RNA sequencing of the testicular somatic cells of an individual with CAIS, and data integration with published scRNA-seq datasets of testis with normal spermatogenesis, NOA, KS, and germinal testicular cancer. Detailed clinical data of the CAIS patient, testosterone and estradiol levels in age-matched men (120 fertile, 155 infertile, 116 NOA, 18 KS, and 343 with TGCC) were analyzed.

Results: In all conditions, Leydig cells are immature and senescent, but those of NOA associated with primary hypogonadism depict the highest expression of transcripts associated with the seminoma microenvironment, including estrogen-responsive genes. An oncological transcriptional signature in the Leydig cells has been confirmed at the systemic levels by showing a prognostic role of the decreasing testosterone/estradiol ratio for TGCC in men with non-genetic NOA.

Conclusion: This study offers molecular insights into the prediction of TGCC in persons with NOA and eligibility for the use of aromatase inhibitors.

Keywords: androgen insensitivity syndrome; hormones; male infertility; scRNA-seq; testis tumor.

Figure 1.

Testis histology. Representative morphological analysis of testicular parenchyma with normal spermatogenesis (spermatozoa at different stages of maturation in the seminiferous tubules) and clusters of LEY cells in the parenchyma obtained from a case of obstructive azoospermia due to mutation of the cystic fibrosis transmembrane conductance regulator (CFTR) gene (A), and histological analysis of testes from NOA (B), KS (C), and CAIS (D). *LEY cells. Scale bar: 50 μm.

Figure 2.

ScRNA-seq of the testis from an individual with Morris syndrome. A) UMAP plot at resolution of 0.4 representing testis cells from a tissue sample of CAIS patient. B) UMAP plots showing the expression patterns of selected marker genes used to identify testis cell types. C) Relative abundance of cell populations in CAIS sample. D) Heatmap showing the expression signature of the top 10 expressed genes in each cluster. E) Selection of significant pathways from Supplementary File S1 for each cluster (27).

Figure 3.

Integration of CAIS sample with NOA, KS, and CTL from the literature. A) UMAP plot resulting from the integration of somatic testis cells of CAIS with 3 samples of NOA, 2 samples of KS, and 9 samples of normal CTL reported in Table S2 (27). B) Relative abundance of cell populations in CTL, NOA, CAIS, and KS samples. C) Violin plots compare expression levels of Stage A–C signatures of LEY cell maturation in CAIS, NOA, and KS vs healthy donor (2-sided Mann Whitney test). Signatures were established by combining 3 datasets reporting single-cell RNA seq of neonatal, prepubertal, and adult human testis analysis recently detailed (12); for the CAIS sample all LEY phenotypes (mLEY, iLEY, and LEY/MYD) were pulled together. D) Heatmaps showing the average expression of the genes in the 3 stages signatures of LEY with hierarchical clustering of genes. Heatmaps were produced with the R library pheatmap v1.0.

Figure 4.

Re-clustering of LEY and MYD cells. A) UMAP plot showing the re-clustering of integrated LEY and MYD cells from CAIS, NOA, KS, and CTL samples. B) Relative abundance of cell subpopulations in LEY-MYD re-clustering. C) Top 10 most significant pathways enriched from the marker genes of the 8 clusters derived from LEY-MYD re-clustering (Supplementary File S3) (27). D) UMAP plots highlighting the 3 clusters of LEY cells in CAIS samples (LEY/MYD, iLEY, and mLEY) across LEY-MYD sub-clusters of cells.

Figure 5.

Senescence phenotype of LEY cells in NOA, KS, and CAIS. A) Venn diagram of the intersection of up-modulated pathways enriched in LEY cells for the comparisons CAIS vs CTL, NOA vs CTL, and KS vs CTL. The pathways were filtered out for p.adj value higher than 0.05, and at least 5 modulated genes (Supplementary Files S3 and S4) (27). B) The pathways passing the gene ratio filter >10% were used to plot the heatmap. Pathways were placed in a custom order to highlight pathways shared among conditions or specific to one condition. C) Venn diagram of the intersection of downmodulated pathways enriched in LEY cells for the comparisons CAIS vs CTL, NOA vs CTL, and KS vs CTL. The pathways were filtered out for p.adj value higher than 0.05, and at least 5 modulated genes (Supplementary Files S5 and S6) (27). D) The pathways passing the gene ratio filter >20% were used to plot the heatmap. Pathways were placed in a custom order to highlight pathways shared among conditions or specific to one condition; pathways shared between NOA and KS are not reported because they did not pass the gene ratio filter. E) Quantification of the circularity of the nuclear envelope of LEY cells, as proxy for nuclear envelope deformation in NOA, KS, and CAIS. CTL, n = 4 independent donors both for a total of 97 LEY cells; NOA, n = 4 independent donor for a total of 86 LEY cells; KS, n = 3 independent donors both for a total of 62 LEY cells; CAIS, n = 1 independent donor for a total of 54 LEY cells. Red bars represent mean ± SEM; ANOVA test. F) Representative images used for the quantification of the circularity of the nuclear envelope of LEY cells. LEY cells were identified by immunofluorescence staining of lineage-specific marker CALB2, and chromatin by DAPI. Two representative images of the testis parenchyma with normal spermatogenesis from 2 independent donors showing the regular shape of the nuclear envelope of LEY cells, and one image representative of the testis parenchyma of NOA, KS, and CAIS showing the presence of several nuclei of LEY cells with indented nuclear envelope. Sperm cells (Sp), meiotic cells (Me), peritubular myoid cells (My), and Sertoli cells (Se) that are recognized for the presence of a characteristic DAPI-negative condensed nucleolus.

Figure 6.

LEY cells from men with NOA express gene signature associated with seminoma microenvironment. Heatmap showing the expression level of the most differentially expressed genes by LEY cells in the context of seminoma, among LEY cells in our cohorts (A). Violin plot showing the average expression level of the GCT signature of LEY cells for condition, with statistical analysis for each group vs NOA condition (B).

Figure 7.

Sex hormones. Peripheral level of T, E2, and T/E2 ratio were established in the blood sample of fertile and infertile men in the absence of any neoplastic disease (A-C), and in the NOA with TGCC and TGCC alone (D-F); the data for the cohort “Men TGCC” are from the pool of data from OSR dataset and publicly available dataset (43). Red lines show median and interquartile range. Dotted lines show reference value. ROC-derived curve, showing the AUC, 95% CI, statistical significance, and the optimal Youden index (G). Decision curve analysis showing the standardized net benefit of T/E2 ratio value of 130 on the prediction of TGCC in NOA men (H).