Defects in efferent duct multiciliogenesis underlie male infertility in GEMC1-, MCIDAS- or CCNO-deficient mice

Abstract

GEMC1 and MCIDAS are geminin family proteins that transcriptionally activate E2F4/5-target genes during multiciliogenesis, including Foxj1 and Ccno Male mice that lacked Gemc1, Mcidas or Ccno were found to be infertile, but the origin of this defect has remained unclear. Here, we show that all three genes are necessary for the generation of functional multiciliated cells in the efferent ducts that are required for spermatozoa to enter the epididymis. In mice that are mutant for Gemc1, Mcidas or Ccno, we observed a similar spectrum of phenotypes, including thinning of the seminiferous tubule epithelia, dilation of the rete testes, sperm agglutinations in the efferent ducts and lack of spermatozoa in the epididymis (azoospermia). These data suggest that defective efferent duct development is the dominant cause of male infertility in these mouse models, and this likely extends to individuals with the ciliopathy reduced generation of multiple motile cilia with mutations in MCIDAS and CCNO.

Keywords: CCNO; Efferent ducts; Fertility; GEMC1; MCIDAS; Multiciliated cells; P73; Testes; Transcription.

Fig. 1.

Gemc1 loss impairs late stages of spermatogenesis. (A) Example of testes from littermate mice of the indicated genotypes at 3 months (top). Ruler: mm. Testes weight relative to whole body weight at the indicated ages (n=8 for 1 month and n=6 animals/genotype for 2-3 months) (bottom). (B) Schematic of semi-synchronous stages of spermatogenesis in mice (adapted from Comazzetto et al., 2014). (C) PAS staining of developing testes from Wt, Gemc1^+/− and Gemc1^−/− littermates at p0, p7, p9, p14 and p20. Scale bar: 100 µm. (D) PAS staining of p27 and p35 testes. Note thinner seminiferous tubule epithelia in Gemc1^−/−. Scale bars: 200 µm (p27, p35: left panels); 50 µm (p27, p35: right panels). (E) RT-qPCR analysis of Gemc1 expression in testes at the indicated post partum days (n=2 animals for p7, p18, p20, p30; n=3 animals for p12, p22, p27, p35) and plotted relative to the trachea. Actb was used as a normalization control. Data are mean±s.d. (F) RT-qPCR analysis of Gemc1 expression in 1- to 2-month-old testes (n=6) compared with trachea (n=4). Actb was used as a normalization control. (G) Ratio of Gemc1 expression in isolated RS/ES populations compared with germ cell pellets (RT-qPCR, n=8 animals). Actb was used as a normalization control. (H) Comparative abundance of each spermatogenic cell type of control and Gemc1^−/− mice using FACS (Wt, n=10 animals; Gemc1^−/−, n=5 animals). Box and whisker plots (A,F-H): median values (middle bars) and first to third interquartile ranges (boxes); whiskers indicate min to max; dots indicate data points. *P=0.023 and **P=0.0023, unpaired two-tailed t-test. ns, not significant; p, post partum days.

Fig. 2.

Seminiferous tubule and rete testes dilation and SC degeneration in testes of Gemc1, Mcidas or Ccno mutant mice. (A) RT-qPCR analysis of Ccno, Mcidas, Foxj1, Trp73 and Cdc20b in the p27 testes of Wt (n=4) and Gemc1^−/− (n=3) mice. For Trp73 n=5 for both genotypes. Actb was used as a normalization control. (B) A representative western blot of TP73 levels in testes lysates from p27 Wt or Gemc1^−/− (n=2). Vinculin was used as a loading control. (C) Quantification of empty lumen space (n=4 testes/genotype). (D) H&E staining of testes sections from p35-p37 Gemc1^−/−, Mcidas^−/− and Ccno^−/− mice revealed thinning of the spermatogenic cell layer. (E) Examples of rete testes dilation in the indicated genotypes at p35. (F) Vimentin staining of SC intermediate filaments in the testes of the indicated genotype. (G) Quantification of the length of vimentin-positive SC arms (Wt, n=4; Gemc1^−/−, Mcidas^−/− and Ccno^−/−, n=2 animals/genotype). Genotype key as in C. (H) SC-only tubules (n=4 testes/genotype). Results from p30-37 testes are shown. Genotype key as in C. (I) Ac-tub staining of p35 seminiferous tubules of Wt (Gemc1^+/+) and Gemc1^−/− mice. Detached spermatids in Gemc1^−/− indicated by black arrowheads. Data are mean±s.d.; dots indicate data points. ***P<0.0001, unpaired two-tailed t-test. Insets show magnification of boxed areas. Scale bars: 100 µm in D,F; 200 µm in E; 50 µm in I.

Fig. 3.

Gemc1, Mcidas or Ccno deficiency causes sperm agglutination in the ED. (A) Schematic of a vertical section of the testis, rete testis, EDs and epididymis (caput, corpus, cauda). (B) Gross morphology of the p35-p37 epididymides of the indicated genotypes. Ruler: mm. (C) PAS staining of the three major regions of the mouse epididymis (caput, corpus and cauda) from adult mice of the indicated genotype. (D-G) RT-qPCR analysis of the expression levels of the indicated gene in different tissues (TR, trachea; TS, testes; ED, efferent ducts; CA, cauda epididymis), normalized to the trachea (n=4 p35 animals). Actb was used as a normalization control. Red dashed line indicates normalized average in the TR. Box and whisker plots: median values (middle bars) and first to third interquartile ranges (boxes); whiskers indicate min to max; dots indicate data points. (H) RT-qPCR analysis of the expression levels of the indicated gene in the EDs of p35 Wt (n=3) or Gemc1^−/− (n=2) animals. Actb was used as a normalization control. Data are mean±s.d.; dots indicate data points. (I) PAS staining of the p35-p37 EDs of mice of the indicated genotype. Black arrowheads indicate the aberrant accumulation of spermatozoa in the Gemc1^−/− and Ccno^−/− mice compared with Wt. Bottom panels show magnification of boxed areas. Scale bars: 100 µm in C,I (top panels); 50 µm in I (bottom panels).

Fig. 4.

GEMC1, MCIDAS and CCNO are required for ED MCC development. (A) Representative ac-tub staining of p35-p37 EDs of the indicated genotypes. (B) ED sections of the indicated genotypes stained with an antibody against TP73. (C) Overexpression of FLAG-GEMC1, but not Myc-CCNO, in AD293 cells by transient transfection induces TP73 expression. RT-qPCR was used to measure relative mRNA levels (n=4) and a representative western blot of three independent experiments is shown. Actb was used as a normalization control and vinculin as a loading control for western blots. (D) Mice with mutations in miR-449/34, Gemc1, Mcidas, E2f4/5, Trp73 and Ccno exhibit defects in MCC development and a similar phenotypic spectrum that includes dilation of the seminiferous tubules and rete testes, SC degeneration and lack of spermatozoa in the epididymis (azoospermia). We propose that the failure of the EDs and resulting agglutination of spermatozoa contributes directly to fluid backpressure, preventing spermatozoa from entering the epididymis. This potentially occurs in human RGMC patients with MCIDAS or CCNO mutations. (E) Immunofluorescent co-immunostaining of AQP1 and TP73 in EDs of p30-35 animals of the indicated genotypes. Insets show magnification of boxed areas. Scale bars: 100 µm in A,B; 50 µm in B (insets); 200 µm in E.