The postnatal deletion of NIPP1 leads to a loss of male germ cells

In adult mouse testis NIPP1 is expressed in both germ cells and somatic cells, including Sertoli cells (Supplementary Fig. ³). Immunostaining revealed NIPP1 expression in all types of germ cells, except in elongated terminally differentiated spermatids. At the age of 6 weeks, which is two weeks after the administration of tamoxifen, NIPP1 was largely removed from both germ cells and Sertoli cells (Fig. 1c and Supplementary ^3c). This incited us to explore the fate of both cell types in the NIPP1 iKOs. H&E stainings revealed that the diameter of the seminiferous tubules gradually decreased due to a loss of germ cells (Fig. 2a and b). In testis of 9 weeks, i.e. 5 weeks after the administration of tamoxifen, the seminiferous tubules already showed a clear decrease in the number of germ cells (Fig. 2a). The most severely affected tubules showed gross vacuolization and only contained few germ cells. By the age of 12 weeks all seminiferous tubules were histologically agametic and only contained Sertoli cells. However, the number of Sertoli cells per seminiferous tubule (SOX9 staining) was not affected by the deletion of NIPP1 (Fig. 2a and Supplementary Fig. ^3d). These data imply that the testis from tamoxifen-treated iKO mice at 9 and 12 weeks are relatively enriched for Sertoli cells, as independently confirmed by SOX9 immunoblotting (Fig. 2c and d). At 20 weeks the Sertoli-cell only phenotype was still maintained, indicating that no new cycles of spermatogenesis had been started after week 9–12 (not shown). Quantitative RT-PCR for the global germ-cell marker Vasa ^27,28 and stage-specific germ cell markers for spermatogonia (Tacstd1, Plzf, Stra8)^29–33, spermatocytes (Sycp3, Stmn1) ^34,35 and spermatids (Prm1, Tpn1)^36,37, confirmed the progressive loss of these types of germ cells in the iKOs (Fig. 2e–g). Hence, the postnatal deletion of NIPP1 from testis results in the gradual loss of germ cells and culminates within a few cycles of spermatogenesis in a Sertoli-cell only phenotype.

Open in a separate window

Figure 2

The postnatal deletion of NIPP1 in adult mice leads to a loss of germ cells. (a) Testis sections of tamoxifen-injected CTR and iKO mice at the indicated ages were Hematoxylin-Eosin (H&E) stained or immunostained for the Sertoli-cell marker SOX9 (chromogenic DAB detection), with a Hematoxylin counterstain for the visualization of nuclei. Scale bar, 50 μm. (b) Quantification of cross-sectional diameter of the seminiferous tubules. (c,d) The level of SOX9 in total testis extracts from tamoxifen-injected CTR and iKO mice at the age of 6, 9 and 12 weeks was visualized by immunoblotting (c) and quantified (d). GAPDH was used as a loading control. (e–g) qRT- PCR analysis of the indicated genes, including markers for the different types of germ cells, at the indicated ages. Hprt was used as housekeeping gene for normalization. Spg. A, spermatogonia A; Spg. B, spermatogonia B; Spc., spermatocytes; Spt., spermatids. Data are represented as means±SEM (n≥3). *p<0.05; **p<0.01; ***p<0.001.

The postnatal deletion of NIPP1 reduces the proliferation and survival of germ cells

We subsequently examined whether the testis phenotype in the iKOs can be explained by a reduced proliferation and/or an increased apoptosis of germ cells. For these experiments, we used tamoxifen-treated mice of 6 weeks, when a testis phenotype is not yet histologically apparent (Fig. 2). Both BrdU (5′-bromo-2′-deoxyuridine) incorporation assays (Fig. 3a and b) and PCNA (proliferating cell nuclear antigen) stainings (Fig. 3a and c) disclosed a twofold lower proliferation rate of spermatogonia and (pre)leptotene spermatocytes in the iKOs. These conclusions were validated by immunoblotting for PCNA (Fig. 3d and e) and immunostainings for cyclin D2 and histone H3 phosphorylation at Ser10 (H3S10P), which are markers for G1 and mitotic cells, respectively (Supplementary Fig. ^4a–c). Finally, we confirmed that no significant loss of germ cells was detected yet in the iKOs at 6 weeks, as demonstrated by qRT-PCR data (Fig. 2e) and immunostainings for the spermatogonia marker PLZF (promyelocytic leukemia zinc finger)^29,30 (Fig. 3f and g).

Open in a separate window

Figure 3

Ppp1r8 ^−/− testes show a reduced proliferation and survival of germ cells. (a) Testis sections of tamoxifen-treated mice of 6 weeks were immunostained for incorporated BrdU and the proliferation marker PCNA. Nuclei were visualized with propidium iodide (PI) or DAPI, as indicated. Scale bar, 50 μm. (b–c) Quantifications of the stainings as shown in panel (a) and performed as described in Materials and Methods. (d–e) The level of PCNA in total testis extracts from tamoxifen-injected CTR and iKO mice at the age of 6 weeks was visualized by immunoblotting (d) and quantified (e). TBP was used as a loading control. (f) Testis sections of tamoxifen-treated mice of 6 weeks were immunostained for the spermatogonial marker PLZF, and stained by a fluorochrome-based TUNEL assay to visualize apoptosis. Nuclei were visualized by DAPI. Scale bar, 50 μm. (g,h) Quantifications of the stainings as shown in panel (f) and performed as described in Materials and Methods. All bar data are means±SEM (n=4). *p<0.05; **p<0.01; ***p<0.001.

TUNEL assays revealed a 3-fold increased level of apoptosis in germ cells from 7 weeks-old iKO mice (Fig. 3f and h). Apoptosis was particularly increased in germ cells from the basal layers of the seminiferous epithelium, comprising spermatogonia and early meiotic spermatocytes. Stainings for the senescence marker p16^INK4a (Supplementary Fig. ^{4a and d}) and for fibrosis (Supplementary Fig. ^4e) did not disclose differences between the CTRs and iKOs. Also, both groups contained similar proportions of tubules at different stages of development (Supplementary Fig. ^4f). Collectively, our data indicate that the gradual loss of male germ cells after the deletion of NIPP1 is explained by their reduced proliferation and increased apoptosis rather than by the induction of senescence or fibrosis. We also examined whether the loss of germ cells in the iKOs was associated with a destabilization of NIPP1 ligands. However, immunoblotting of testicular lysates of mice of 6 weeks did not disclose a change in the level of three major protein interactors of NIPP1 in the iKOs, i.e. protein phosphatase PP1 or the splicing factors CDC5L and SAP155 (Supplementary Fig. ⁵).

Germ-cell proliferation is also reduced after the deletion of NIPP1 in cultured testis slices

To further explore the origin of the loss of germ cells in the iKOs, we have subsequently investigated whether this could be an indirect effect caused by the deletion of NIPP1 from peripheral tissues. For example, it could be argued that the deletion of NIPP1 from the brain interferes with the level of circulating testosterone, which is required for germ-cell survival³⁸. However, the expression of NIPP1 in the iKOs was only marginally affected in the brain cortex and hypothalamus, and only reduced in a scattered fashion in the cerebellum (Supplementary Fig. ^6a-d). Also, the levels of circulating testosterone at 9 weeks were similar in the CTRs and iKOs and there was no correlation between testosterone levels and testis weight (Supplementary Fig. ^{6e and f}). To further differentiate between intrinsic and extrinsic defects, we aimed to generate a testis-specific NIPP1 mouse KO model by recombination with Cre recombinase under control of the gonocyte-specific Vasa promotor (Supplementary Fig. ^{6g and h}). However, the CRE-recombinase was expressed precociously, resulting in the global deletion of NIPP1, as has been reported for other Cre strains³⁹. As an alternative approach to determine whether the testis phenotype was mediated by the loss of NIPP1 from peripheral tissues, we performed organ cultures using testis that were isolated from tamoxifen-treated mice of 6 weeks (Fig. 4a), or testis isolated at the age of 4 weeks and treated in vitro with hydroxytamoxifen (Fig. 4d). Under in vitro outgrowing conditions, BrdU incorporation (Fig. 4a–c) and the level of PCNA (Fig. 4d–f) were severely decreased in the iKOs, demonstrating that the observed phenotype is a testis-intrinsic defect.

Open in a separate window

Figure 4

The deletion of NIPP1 leads to a reduced proliferation of germ cells in cultured testis slices. (a) Testis from tamoxifen-treated CTR and iKO mice were isolated at 5 weeks and 3 days, and cultured for 4 days. After incubation for 6h with BrdU, testis sections were H&E stained and immunostained for NIPP1 and incorporated BrdU. Nuclei were visualized with propidium iodide (PI) or DAPI as indicated. Scale bar, 50µm. (b,c) Quantification of stainings for NIPP1 (b) and incorporated BrdU (c) as shown in panel (a). (d) Organ culture of testis slices that were isolated from non-treated CTRs and iKOs of 4 weeks. (Z)-4-Hydroxytamoxifen (4-OHT) was added to the slices for 96hours to delete the floxed Ppp1r8 allele. Subsequently, testis sections were H&E stained, and immunostained for NIPP1 and PCNA. Nuclei were visualized by DAPI. Scale bar, 50 μm. (e,f) Quantification of stainings as shown in panel (d). All data in this figure are represented as means±SEM (n=4). *p<0.05; **p<0.01; ***p<0.001.

Organ and cell culture

Testis organ culture was performed as described^60,61. Briefly, testis from adult male mice were decapsulated by removal of the tunica albuginea and subsequently divided with forceps into 4 pieces of approximately 2 mm diameter. The 4 pieces from each testis were placed in one hexahedron of agarose gel and soaked in the organ culture medium containing 2x alpha-MEM (Sigma), 10% knockout serum replacement KSR (Sigma), 100 U/ml penicillin, 100µg/ml streptomycin and 7% w/v sodium bicarbonate. Testis pieces were maintained in culture for 4 days and 10µM BrdU (Sigma) was added to the medium 6hours before harvesting. In vitro deletion of the floxed Ppp1r8 allele was obtained by addition of 2µM (Z)-4-Hydroxytamoxifen (OHT) (Sigma-Aldrich) to the medium for 96h. Finally, the 4 pieces of one agarose hexahedron were collectively embedded in paraffin and all 4 pieces were included for the quantification of this experimental condition.

GFRA1-enriched cells (undifferentiated spermatogonia) were isolated from 12 days-old CTR and iKO mice, as described by Guan et al. (2009)⁴¹. Briefly, testicular cells were isolated by an enzymatic procedure using collagenase IV (Worthinghton biochem corp.) and dispase (Invitrogen), followed by enrichment of GFRA1-enriched spermatogonia through laminin (Sigma) selection. This laminin selection method is based on the expression of specific surface markers on SSCs, such as α6 integrin (Itga6) and β1-integrins (Itgb1). Since some of these receptors bind the extracellular matrix molecule laminin, laminin can be used to enrich testicular cells for SSCs, as reported^41–43. After enrichment, cells were co-cultured with freshly prepared MEF (mouse embryonic fibroblasts) as feeding cells. The MEFs had been pre-incubated with Mitomycin C (Sigma) and maintained in spermatogonia stem cells (SSCs) growth medium, containing MEM alfa medium (Thermo fisher scientific) and supplemented with 10% FBS, 100 U/ml penicillin, non-essential amino acids (Thermo fisher scientific), β-Mercaptoethanol (Sigma), N2-1 supplement (Thermo fisher scientific), recombinant GDNF (R&D systems; 4 ng/ml) and recombinant LIF (Sigma; 10³ U/ml). NIPP1 deletion was induced by addition of 1µM 4-OHT to the medium for 72h. For the proliferation assay, 15µM BrdU was added to the medium 5hours before harvesting cells. The purity of the cells after laminin enrichment was assessed by quantifying the relative number of GFRA1+ cells in images from GFRA1 and DAPI stained cells. Two independent isolations after laminin selection were scored by selecting 8 images randomly and scoring >180 cells for each isolation.

Immunohistochemistry

Testes fixed in Bouin’s (6h) or 4% PFA (24h) were embedded in paraffin and sectioned at a thickness of 6µm. Testis sections were stained with Hematoxylin and Eosine (H&E) or the fibrosis marker Sirius red according to standard protocols. DNA was visualized by DAPI or Propidium Iodide (PI), as indicated. For the clarity of the pictures, the blue color of DAPI was often converted to red by image processing. TUNEL assays were performed using the In-Situ-Cell-Death-Detection-Kit and Fluorescein (Roche) on paraffin-embedded testis tissue following the instructions of the manufacturer. Testis sections were immunostained using the antibodies listed in Supplementary Table ³ and using the TSA^TM Fluorescein, TSA^TM Biotin system (PerkinElmer) or chromogenic detection using DAB (3,3′-diaminobenzidine). Detailed immunostaining protocols are available on request. The immunofluorescence images were acquired with a Leica TCS SPE laser scanning confocal system mounted on a Leica DMI 4000B microscope and equipped with a Leica ACS APO 20X objective. Quantifications shown in, Fig. 3b,g and h, Supplementary Fig. ^3d and Supplementary Fig. ^4b,c were performed by counting the number of cells stained with the indicated cell markers per seminiferous tubule (30 randomly selected tubules per mouse, n=4). Leica MM AF 2.1 software was used for quantification (10 pictures of 10x objective magnification were randomly selected per mouse testis) of the relative expression of PCNA (Figs 3c and ​and4f),4f), NIPP1 (Fig. 4b and e, and Supplementary Fig. ^1f), SOX9 (Supplementary Fig. ^6b), and p16 (Supplementary Fig. ^4d) over DAPI (nucleus) signal. Quantification for incorporated BrdU and apoptopic cells in the GFRA1-enriched cell populations (Fig. 6g and h) were performed by counting the number of positive cells relative to the total number of cells in the colony (n=3; eight colonies for each condition were analysed).

Biochemical procedures

Testes were homogenized using a dounce homogenizer (Sigma) and incubated in lysis buffer for 20min at 4°C. The lysis buffer comprised modified RIPA buffer (50mM Tris-HCl at pH 7.4, 1% Triton-X 100, 0.2% sodium deoxycholate, 0.2% sodium dodecyl sulfate (SDS), 1mM EDTA, 0.3M NaCl) or SDS-lysis buffer (50mM Tris-HCl at pH 7.4, 2% SDS, 12% glycerol, 100mM β-mercaptoethanol). Lysis buffers were supplemented with 20mM NaF, 5µM leupeptin, 0.5mM phenylmethanesulfonyl fluoride, 0.5mM benzamidine and 1mM orthovanadate. The lysates were centrifuged for 5min at 1800 × g and the supernatant was used for immunoblotting. The testis extracts for immunoblotting in Figs 1e, ​,2c,2c, ​,3d3d and Supplementary Fig. ⁵ were prepared with modified RIPA, while SDS-lysis buffer was used in Supplementary Fig. ^7c. Immunoblotting was performed following 10% SDS- PAGE with the indicated antibodies (Supplementary Table ³). Immunoblots were visualized using ECL reagent (Perkin Elmer) in an ImageQuant LAS4000 imaging system (GE Healthcare) (see Supplementary Fig. ⁸ for uncropped blots) and were quantified using ImageQuant TL software (GE Healthcare).

RNA sequencing and Quantitative Reverse Transcriptase PCR (qRT-PCR)

Total RNA was isolated from 40mg of snap-frozen mouse testis using the GenElute^TM. Mammalian Total RNA Miniprep kit (Sigma-aldrich). RNA integrity of the samples used for RNA sequencing was assessed using a Bioanalyser 2100 (Agilent). Library preparation, sequencing and statistical analysis of the RNA sequencing data were performed by VIB Nucleomics Core, as detailed in the Supplementary information.

Complementary DNA (cDNA) was synthetized from 2µg of total RNA using oligo dT primers (Sigma-aldrich) RevertAid Premium Reverse Transcriptase and RiboLock RNase inhibitor enzymes (Fermentas). About 1.2% of the cDNA was amplified by PCR in duplicate using SYBR Green qPCR Mix (Invitrogen) and a Rotorgene detection system (Corbett Research). To compare the relative amount of target genes in different samples, values were normalized to the housekeeping gene Hprt (Hypoxanthine-guanine phosphoribosyltransferase) or a cell-type specific gene as indicated in the Figures. qRT-PCR primers are listed in Supplementary Table ⁴.

Statistical analysis

All statistical analysis was performed using GraphPad Prism software. Two-way unpaired (Figs 1–5 and 6i–l, Supplementary Figs ¹–³ and ⁴–⁶) or paired (Fig. 6e–h) student’s t-test, and Pearson’s correlation test (Supplementary Fig. ^6f) were used.

We acknowledge Dr. Karel De Gendt for expert advice and valuable discussions. We thank Maud De Meyer, Annemie Hoogmartens and Fabienne Withof for technical assistance. Prof. M. Baes (KU Leuven, Belgium) is acknowledged for the donation of Ubc-CreERT2 mice, and Prof. DH. Castrillon (University of Texas Southwestern Medical Center, USA) for providing the Vasa-Cre mice. RNA sequencing library preparation, sequencing and statistical analysis was performed by the VIB Nucleomics Core (www.nucleomics.be). This work was financially supported by the Fund for Scientific Research - Flanders (grant G0A6213N). The FCT/Portuguese Ministry of Science and Technology provided a doctoral fellowship for M.F.