DMRT1 haploinsufficiency leads to secondary infertility in XY male rabbits

Abstract

DMRT1 is a key factor in testis development, where it is involved in sex determination and fertility. Mutations in DMRT1 have been described in humans, with patients presenting 46,XY Disorders of Sex Development (46,XY DSD) or infertility. In a previous study, we demonstrated that DMRT1 is a testis-determining factor in rabbits, with DMRT1^−/− rabbits exhibiting a male to female XY sex reversal. In this study, we show that DMRT1 haploinsufficiency induces secondary infertility, with XY rabbits presenting oligospermia or even azoospermia at 2 years of age. We observed that sperm concentration decreases and sperm anomalies increase in DMRT1^+/− rabbits at adulthood. Furthermore, spermatogenesis is impacted as early as 4 months (the earliest stage where spermatozoa are detected), with dysregulation of genes involved in spermatid maturation and oocyte/spermatozoa fusion, as well as overexpression of genes involved in the mitosis/meiosis transition of spermatogonial stem cells. Finally, DMRT1 haploinsufficiency impacts the earliest stages of germ cell differentiation, with persistent proliferation and pluripotency in the postnatal period. In conclusion, our findings underscore DMRT1 as a crucial factor at various stages of testicular development and reinforce its role in the multiple phenotypes observed in humans.

DMRT1^+/− XY rabbits display a defect in primordial germ cell differentiation, impairments in spermatogonial stem cell renewal, and an impact on spermiogenesis, collectively resulting in secondary infertility in adulthood.

Graphical Abstract

Graphical Abstract.

Introduction

DMRT1(Doublesex and Mab-3 Related Transcription Factor 1) is a member of the evolutionarily conserved DM domain protein family. The DMRT1 gene was first discovered in Drosophila (named Doublesex or dsx) and Caenorhabditis elegans (named Mab-3) [1, 2]. These transcription factors are characterized by the presence of a zinc finger–like DNA-binding domain, namely the DM domain. DMRT1 represents the most conserved factor for its involvement in testis differentiation across the animal kingdom. Moreover, DMRT1, or duplicates of DMRT1, localized on sex chromosomes, have been identified as master sex-determining genes in different vertebrate species such as medaka (Oryzias latipes) [3, 4], xenope (Xenopus laevis) [5], and chicken (Gallus gallus domesticus) [6, 7]. Interestingly, in chickens and certainly in the majority of birds [8], DMRT1 is located on the Z chromosome and is required at double doses (ZZ males, ZW females) in order to promote testis formation by inducing the differentiation of pre-supporting gonadal cells into Sertoli ones; as a consequence, ZZ DMRT1^+/KO chickens developed ovaries [7].

In addition to its primary role in testis determination, DMRT1 remains expressed at all testicular developmental stages where it appears crucial for spermatogenesis and male fertility throughout the entire male reproductive lifespan ([9], for review). In the testis, two cellular types, Sertoli and pluripotent germ cells, continuously express DMRT1. This specific testicular DMRT1 expression in both cell types seems to have emerged during evolution, as it is absent in drosophila [10] and medaka [11, 12] where DMRT1 expression is restricted to somatic supporting cells. However, the somatic and germline expression of DMRT1 has been described in a number of mammals, including mice [13, 14], rabbits [15], humans [16], marsupials [17], also in chickens [6, 13], amphibians [18, 19], and reptiles [20, 21]. The role of DMRT1 in spermatogenesis and male fertility has been primarily investigated in mice where conditional knockout (KO) has enabled the suppression of Dmrt1 expression exclusively in Sertoli cells or in germ cells [22–24]. The bi-allelic loss of function of Dmrt1 in mice indicates a crucial role for this protein in spermatogenesis. Starting from post-natal day 10 (P10) in mice, when meiosis normally begins, the number of germ cells is greatly reduced in KO testes, with no meiotic germ cells present. Thereafter, no germ cells remain visible 4 days later at P14 [25, 26]. Moreover, an elevated number of Sertoli cells have been reported in these P14 testes compared to other germ-cell depleted mutants such as c-Kit, suggesting a role of Dmrt1 in the final step of differentiation of the male supporting lineage [25]. Later on, it was clearly demonstrated that DMRT1 functions as a Sertoli gatekeeper factor in mouse adulthood since its Sertoli-specific ablation leads to female reprogramming of different testicular cell types [23]. Furthermore, Dmrt1 is expressed in spermatogonial stem cells, but not in the meiotic or post-meiotic germ cells, where it regulates the mitosis–meiosis transition [25]. By conditional deletion of Dmrt1 in the germinal lineage, the authors showed that DMRT1 promotes the differentiation and proliferation of spermatogonia and represses Stra8 (Stimulated by retinoic acid 8), which allows the entry of germ cells into meiosis [27]. The germ cells of these mutant mice enter meiosis prematurely, leading to depletion of the spermatogonial pool and non-obstructive azoospermia [27]. A role of DMRT1 in the maintenance and renewal of spermatogonial stem cells (SSCs) has also been nicely documented [24]. Indeed, loss of Dmrt1 in SSCs causes loss of the SSC maintenance factor PLZF (Promyelocytic Leukemia Zinc Finger, or ZBTB16) and differentiation of SSCs [24]. Finally, it has also been shown that mouse DMRT1 controls fetal germ cell proliferation and pluripotency in a dose-sensitive manner [28]. The loss of Dmrt1 in 129Sv mice results in a high incidence of teratomas, demonstrating that Dmrt1 acts as a dose-sensitive tumor suppressor gene that directly controls transcription of the pluripotency regulator Sox2 in germ cells [28].

Involvement of DMRT1 in testis function and male fertility has also been evidenced in humans. However, by contrast with mice, the very rare mutations reported in humans are heterozygous, and the phenotypes observed seem to be linked to a haploinsufficiency of DMRT1 rather than a loss of function. Depending on the mutation, two types of associated diseases have been reported following DMRT1 haploinsufficiency: 46,XY DSD with gonadal dysgenesis or male infertility ([29], for review). One de novo point mutation (R111G) leading to XY, DSD was well documented in 2015. Although heterozygous, it seems to act as a dominant negative protein, effectively functioning as a loss-of-function mutation [30, 31]. Moreover, DMRT1 mutations associated with male infertility are very rare, as only four putative pathogenic variants have been described in six patients among a cohort of 171 investigated for cryptozoospermia or non-obstructive azoospermia [32]. Finally, a heterozygous P74L mutation has been found in two brothers with Sertoli cell-only (SCO) syndrome (mentioned in [29]), but these cases remain unpublished. A recent study has indicated that a deletion of the first two exons of the human DMRT1 gene (del9p24.3) may be associated with non-obstructive azoospermia [33]. The present case of male infertility may also be attributed to the varicocele detected in the patient. The authors of the present study concluded that this partial deletion of DMRT1 (del9p24.3) is probably the determining factor in the development of azoospermia in the described family by a loss-of-function mechanism [33].

Recently, we have characterized a DMRT1 knockout rabbit line and shown that the bi-allelic loss of function of this gene leads to XY sex reversal and to infertility of both XY and XX animals because the germ cells fail to enter meiosis in DMRT1 KO XX and XY ovaries [15]. The founder animal of this rabbit genome–edited line was obtained in 2013, and since then, many crosses have been carried out to establish and maintain the line and to produce biological material for our studies [15]. From these crosses, mainly involving heterozygous DMRT1 males and females, we noticed a drastic loss of fertility with aging of the animals. Finally, by carefully checking the different crosses, we observed a loss of fertility of the heterozygous males DMRT1^+/− starting at around 2 years of age but being hugely variable across considered animals. A loss of fertility at this age would mean a significant reduction in the lifespan of the animals, 2 years being the first fifth of a rabbit’s life [34]. These observations led us to finely characterize this phenotype of secondary infertility in XY male DMRT1^+/− rabbits. Initially, we showed that 2-year-old testes of DMRT1^+/− rabbits are almost devoid of spermatozoa, with a majority of seminiferous tubules showing a SCO phenotype. Subsequent characterizations of the testicular development at earlier stages, namely, during the first month of age and at 4 months (pre-pubertal age in rabbits), demonstrated that DMRT1 haploinsufficiency is responsible for spermatogenetic defects linked to both Sertoli and germ cells. It appears that different failures are responsible for this degenerative fertility: (1) failures linked to Sertoli cells, which remain unable to support spermatogenesis, clearly detectable during the spermiogenesis process; and (2) failures linked to the germinal lineage, affecting spermatogonia renewal and the transition from mitosis to meiosis.

Materials and methods

Animals

New Zealand rabbits (NZ1777; Hypharm, Roussay, France) were selectively bred at the SAJ (Sciences de l'animal de Jouy) rabbit facility (Jouy-en-Josas, France). All experimental procedures were conducted in accordance with the guidelines established by the French Ministry MENESR (accreditation numbers APAFIS #685 and #21451) and adhered to the recommendations provided by the local committee for ethics in animal experimentation (COMETHEA, Jouy-en-Josas). All researchers involved in animal handling possessed valid animal experimentation licenses issued by the French veterinary services. The DMRT1 mutant rabbit line was established as described [15]. From sexual maturity (6–8 months), heterozygous DMRT1^+/− males and females were mated together or with wild-type animals. The number of mating with or without birth per crosses, as well as the number of pups per litter, was supervised. An animal care technician was present throughout the mating process between the male and female rabbits and reported a successful mating when the male rabbit fell on its side or backwards after mounting the female.

Semen collection and sperm parameter analyses

The semen of rabbits from both genotypes (WT or DMRT1^+/−) was collected using a custom-designed artificial vagina. Two consecutive collections were conducted for each animal (~10 min apart). The volume of ejaculate was determined by pipetting, followed by immediate dilution of sperm (1:20) in GALAP media (IMV Technologies, France), specifically formulated for rabbit semen preservation. Each diluted sample was incubated for 10 min at 37°C before analyzing sperm parameters using a CASA Hamilton Thorne IVOSII device (Hamilton Thorne, France) with a 10× objective lens.

Histological, immunohistological, and in situ hybridization analyses

Following sampling, adult testicular pieces were promptly submerged in either Bouin fixative or paraformaldehyde (4% PFA in 1× PBS solution). After fixation for 24–48 h, the tissues underwent 1× PBS rinses and were then stored at 4°C in 70% ethanol until embedded in paraffin. Tissue sections of 5-μm thickness were prepared using a Leica RM2245 microtome and mounted on Superfrost Plus Slides (J18000AMNZ; Epredia). Prior to staining or experimentation, the sections were deparaffinized and rehydrated through sequential baths of xylene and ethanol (100%, 96%, 70%, 30%, and finally, in H₂O).

Hematoxylin–eosin–saffron (HES) staining was conducted by the @Bridge platform (INRAE, Jouy-en-Josas, France) utilizing an automatic Varistain Slide Stainer (Thermo Fisher Scientific).

In situ hybridization (ISH) was conducted on paraffin sections, utilizing the RNAscope ISH methodology (ACD; Bio-Techne SAS, Rennes, France) in cases where a reliable antibody was unavailable for characterizing the target protein. In brief, 5-μm sections of PFA 4% fixed tissue were subjected to labelling using the RNAscope 2.5HD assay-brown kit (322300; ACD), along with 1000-nucleotide-long probes designed and manufactured by the provider (Supplementary Table S1). Brown labeling was visually detected as a signal, and hybridization was deemed positive upon observation of at least one dot within a cell.

Immunohistochemistry (IHC) was conducted on paraffin-embedded sections using the ImmPRESS polymerized reporter enzyme staining system (Vector Laboratories). First, slides were exposed to a citrate buffer (pH 6.0; H-3300, Vector Laboratories), and endogenous peroxidases were inactivated with a 0.3% H₂O₂ solution (H1009; Sigma-Aldrich). Subsequently, sections were blocked with 2.5% normal horse serum (from ImmPRESS anti-goat kit MP-7405, ImmPRESS anti-rabbit kit MP-7401, or ImmPRESS anti-mouse kit MP-7402; Vector Laboratories) for 30 min at room temperature. Primary antibodies (Supplementary Table S2) were then applied and incubated for 1 h and 30 min at 37°C. Following washes with PBS 1×, sections were treated with secondary antibody reagents (supplied in ImmPRESS kits) for 30 min at room temperature. Finally, these antibodies were revealed with DAB enzymatic reaction (SK-4100; Vector Laboratories), followed by brief hematoxylin counterstaining. Slides were dehydrated and mounted with Eukit (CL04.0503.0500; Chem-Lab NV). A 3DHISTECH panoramic scanner was used for image acquisition at the @Bridge platform (INRAE, Jouy-en-Josas, France), and specific structures were quantified with ImageJ software (Cell Counter plugin).

Immunofluorescence was performed on rehydrated sections, and the epitopes were unmasked as described above, using a citrate-based unmasking solution in a pressure cooker. For studies on DNA methylation, the DNA was denatured for 15 min using HCl 2 N. The sections were then permeabilized by incubating them with 0.5% Triton and 1% BSA for 1.5 h. Following an overnight incubation at 4°C with the primary antibodies (Supplementary Table S2), and a 1-h incubation at room temperature with secondary antibodies (Supplementary Table S2), slides were mounted in Vectashield mounting medium (H-1400; Vector Laboratories) containing DAPI. Images were acquired with an Apotome Zeiss camera at the MIMA2 platform (INRAE, Jouy-en-Josas, France).

Total RNA extraction

The testes of postnatal or adult rabbits were collected and immediately snap-frozen in liquid nitrogen before being stored at −80°C until extraction. Total RNA was extracted from each sample using Trizol reagent (15596018; Life Technologies), purified with the RNeasy MiniKit (74104; QIAGEN) according to the manufacturer’s protocol, and subsequently treated with DNase (10223460; QIAGEN).

RNA sequencing and bioinformatics analysis

Total RNAs were extracted from testes of wild-type or DMRT1^+/− rabbits at either 4 months or 1-day post-partum (dpp). The integrity of the total RNA was assessed using an Agilent 2100 Bioanalyzer (Matriks, Norway), and samples with a RNA integrity number >9 were selected for RNA sequencing. The present study made use of the facilities and expertise of the I2BC High-throughput Sequencing Platform (https://www.i2bc.parissaclay.fr/sequencing/ng-sequencing, Université Paris-Saclay, Gif-sur-Yvette, France) for orientated library preparation (Illumina Truseq RNA Sample Preparation Kit) and sequencing (paired-end 50–35 bp; NextSeq500 on 1 dpp gonads/paired-end 50 bp; NextSeq2000 with gonads on 4-month-old gonads). Each sample yielded >35 million paired-end reads. Demultiplexing and adapter removal were performed at the I2BC High-throughput Sequencing Platform using bcl2fastq2–2.18.12 and Cutadapt 3.2, respectively, from 1 dpp gonads and bcl-convert 4.1.5 and Cutadapt 3.2, respectively, from 4-month-old gonads. Only reads >10 base pairs were included in the subsequent analysis. Quality control of the raw RNA-Seq data was conducted using FastQC v0.11.5.

Reads were mapped to all genes in the best-annotated rabbit genome (OryCun2.0, Oryctolagus cuniculus, Ensembl version 106) where we had improved the transcriptome (by extending the 5′ and 3’-UTRs of genes; adding or fixing sequences of 22 gonadal differentiation marker genes) as previously described [15]. Then, after mapping with STAR version 2.5.1b [35], reads were counted using FeatureCounts version 1.4.5 [36]. Data normalization and single-gene level analyses of differential expression were performed using DESeq2 [37]. Differences were considered to be significant for Benjamini–Hochberg adjusted p-values <0.05, and FC <1/1.5 or FC >1.5 [38]. RNA-seq data were deposited via the SRA Submission portal (http://www.ncbi.nlm.nih.gov/bioproject/1154456), BioProject ID PRJNA1154456.

Statistics

The statistical analysis was conducted utilizing GraphPad Prism 10 Software (GraphPad Software Inc., La Jolla, CA, USA). Due to the small sample sizes within groups, comparisons between values were conducted using the Mann–Whitney test for non-parametric data. A significance level of p <0.05 was used for determining statistical significance.

Results

Accelerated reproductive aging in DMRT1^+/− rabbits

During our investigations on the rabbit line carrying the DMRT1 mutation [15], we noted a decline in fertility in heterozygous male animals. To characterize this unexpected phenotype, we recorded the number of unsuccessful crosses (those not resulting in pregnancy and birth), distinguishing between the two male types of crosses made (Table 1A). The reference values for wild-type (WT) parents are derived from crosses carried out in our animal facility (SAJ experimental unit of Jouy-en-Josas, France) for the production of experimental New Zealand NZ1777 rabbits used in all our studies [39, 40]. When comparing the number of crosses between XX WT and XY WT or XY DMRT1^+/− with a mean age of 17.5 months for XY WT rabbits and 16 months for XY DMRT1^+/− rabbits (Table 1A), we noticed an increase in unsuccessful crosses (20.3% vs. 61.3%, respectively) and conjointly a decrease in successful crosses leading to births (79.7% vs. 38.1%, respectively) when using male heterozygous animals (Table 1A).

Table 1.

DMRT1 ^+/− male rabbits have lower fertility. (A) Number of successful and unsuccessful mating, depending on the genotype of the XY parents

A	Wild-type males × wild-type females	DMRT1+/− males × wild-type females
Matings with birth	106 (79.7%)	8 (38.1%)
Matings without birth	27 (20.3%)	13 (61.9%)
Total number of matings	133	21
B	Wild-type males × wild-type females	DMRT1+/− males × wild-type females
Number of distinct breeding males	14	11
Number of distinct breeding females	20	11
Total number of litters	106	8
Total number of pups	854	48
Average number of pups per litter	6.7 ± 4.4	3.5 ± 0.8

The mean age of the wild-type males used is 17.5 months (ranging from 6 to 36 months) and 16 months for the DMRT1^+/− males (ranging from 6 to 32 months). (B) Number of different males and females used and obtained pups per litter depending on the XY genotype. The mean age of the wild-type males used is 17.5 months (ranging from 6 to 36 months) and 16 months for the DMRT1^+/− males (ranging from 6 to 32 months)

Moreover, we analyzed the number of pups per litter by comparing the two types of crosses (Table 1B). Here again, the average number of pups per litter was almost halved in heterozygous XY (3.5 ± 0.8) compared with WT XY (6.7 ± 4.4) (Table 1B).

In parallel, the fertility of the heterozygous female rabbits was also observed. Interestingly, the number of unsuccessful crosses increased slightly when the female carried the DMRT1 mutation compared to XX WT (25% vs. 20.3%, respectively) (Supplementary Table S3). This drop in the number of pups per litter is also noticeable when the female is heterozygous compared to WT XX (Supplementary Table S3) (3.7 ± 1.6 vs. 6.7 ± 4.4). Furthermore, when both parents carried the mutation, there was an increase in unsuccessful crosses (48.1% vs. 20.3% in WT rabbits) and a consequent decrease in successful crosses that resulted in births (51.9% vs. 79.7%) (Supplementary Table S3). The average number of pups per litter was also reduced by more than half when both parents were heterozygous (2.8 ± 3.1 vs. 6.7 ± 4.4 for WT parents) (Supplementary Table S3). Although the number of crosses involving a single carrier animal is low, the possible reduction in DMRT1^+/− female fertility is currently under investigation by evaluating the number of ovarian follicles at different ages.

All these observations suggest that DMRT1 haploinsufficiency could lead to reduced fertility, including in XY DMRT1^+/− male carriers. To further characterize this phenotype using non-invasive approaches, we analyzed the semen of these XY DMRT1^+/− males in comparison with XY age-matched wild-type.

Sperm parameters of 10 WT and 10 DMRT1^+/− male rabbits (sample details in Supplementary Table S4) were assessed using the IVOS II CASA system on semen ejaculates (Figure 1). First, as the computer-assisted sperm analysis (CASA) failed to correctly count the number of spermatozoa due to the presence of numerous particles in the rabbit sperm linked to the petroleum jelly used for sampling (Supplementary Fig. S1), we manually counted the spermatozoa using a Malassez cell. We noticed that the concentration of spermatozoa was significantly decreased in DMRT1^+/− semen samples compared to WT ones (Figure 1A). Furthermore, the CASA system detected significant differences between DMRT1^+/− and WT ejaculates. Motility and progressive motility of DMRT1^+/− spermatozoa are greatly decreased compared to WT (Figure 1B and1C). Moreover, sperm anomalies such as bent tail, coiled tail, and the presence of a distal droplet were found to be increased (Figure 1D–F). All these results encouraged us to further study the testes of these DMRT1^+/− rabbits.

Figure 1.

Sperm analysis of wild-type and DMRT1^+/− rabbits. (A) Concentration of sperm (sperm number/mL), measured using a Malassez cell. Wild-type n = 10; DMRT1^+/− n = 10. Motility and morphometric parameters of the sperm from wild-type and DMRT1^+/− rabbits were obtained by computer-assisted sperm analysis. Each point or square represents the average of 2 ejaculates: (B) total motility and (C) progressive motility of the sperm was decreased in mutants; percentages of (D) bent tail, (E) coiled tail, and (F) distal droplets were increased in mutants. Wild-type n = 10; DMRT1^+/− n = 10 or 9. The median is shown by the wide horizontal line. Mann–Whitney test: ^*p-value <0.05; ^**p-value <0.005; ^***p-value <0.0005; ^****p-value <0.00005.

Adult DMRT1^+/− rabbits exhibit spermatogenesis arrest

First, the weight of DMRT1^+/− and WT testes was measured and related to the total weight of the rabbit (Figure 2A). In adulthood, the testis-to-body weight ratio decreased drastically in DMRT1^+/− testis, becoming more than three times lower than that WT at 1 to 2 years of age (Figure 2A). Thereafter, we observed the structure of 2-year-old DMRT1^+/− and WT testes by conventional histology following HES staining of the sections (Figure 2B–G). In control WT rabbit testes, all stages of spermatogenic germ cells could be observed, from spermatogonia to mature sperm (Figure 2B and C). The seminiferous tubules were observed to contain germ cells at various stages of meiosis, depending on the specific stage of spermatogenesis. Notably, some tubules at stage XII harbored many spermatozoa in their luminal space (Figure 2C). Conversely, seminiferous tubules from DMRT1^+/− rabbits may be entirely devoid of spermatozoa (Figure 2D–G) or contained a minimal number of spermatozoa, as evidenced by the results obtained in the CASA experiments. Some of them displayed a Sertoli cell–only phenotype, where some supporting cells seemed to detach from the basal lamina and ended up in the lumen of the tubule (Figure 2E). In other tubules, meiotic germ cells were visible but did not progress beyond the spermatocyte stage, and no round spermatids were detectable in these testes (Figure 2G). In conclusion, XY DMRT1 heterozygous male rabbits exhibit a phenotype of secondary infertility at ~2 years of age, characterized by oligospermia or non-obstructive azoospermia depending on the individual.

Figure 2.

Spermatogenesis defects in DMRT1^+/− rabbits. (A) Testis on body weight ratio in wild-type and DMRT1^+/− rabbits, in adult (1 to 2 years old, wild-type n = 5; DMRT1^+/− n = 13). The median is shown by the wide horizontal line. Mann–Whitney test: ^***p-value <0.0005; ns: non-significant. (B, C) HES staining on 2-year-old wild-type testes. (D–G) HES staining on DMRT1^+/− testis on two different adult rabbits.

Spermatogenesis defects are visible as early as 4 months in XY DMRT1^+/− rabbit testes

In order to characterize the origin of infertility, we chose to study the testes of DMRT1^+/− rabbits at the age of 4 months, a stage corresponding to the establishment of complete spermatogenesis and the appearance of the first spermatozoa. Supplementary Figure S2 shows the key stages of testicular differentiation and the early stages of spermatogenesis. Unlike in adulthood, the testis-to-body weight ratio is not decreased in DMRT1^+/− rabbits at 4 months of age (Figure 3A). At this early puberty stage, while the testes of control animals enclose a few seminiferous tubules containing spermatozoa at luminal level (Figure 3B–C), no such tubules were visible in DMRT1^+/− rabbits (Figure 3D–G). In addition, many seminiferous tubules with elongated spermatids were visible in control animals (Figure 3H), whereas their number was greatly reduced in the DMRT1^+/− ones (15% vs. 70%). Furthermore, many seminiferous tubules appeared abnormal in DMRT1^+/− rabbit testes (32% vs. 3% in control testes), clearly demonstrating early dysfunction of spermatogenesis (Figure 3I). These tubules appeared disorganized at the germ cell layers corresponding to different meiosis stages and contained numerous malformed germ cells at the luminal level (Figure 3D–G). These histological observations indicate that from puberty onwards, DMRT1^+/− rabbit testes exhibit defective spermatogenesis, notably characterized by impaired meiosis and spermatid formation followed by spermiogenesis defects.

Figure 3.

Defect in spermatogenesis from the appearance of the first spermatozoa in DMRT1^+/− rabbits at 4 months. (A) Testis on body weight ratio in wild-type and DMRT1^+/− rabbits, at adult age (in gray) and at puberty (in black) (4-month-old, wild-type n = 9; DMRT1^+/− n = 5). (B, C) HES staining on wild-type testis. (D–G) HES staining on DMRT1^+/− testis on two different rabbits. The graphs represent the percentage of (H) tubes with elongated spermatids or (I) abnormal tubes, and each dot represents one animal (n = 4 for wild-type and DMRT1^+/−, number of tubules counted from 206 to 941 in WT and 591 to 1290 in DMRT1^+/−). The median is shown by the wide horizontal line. Mann–Whitney test: ^*p-value <0.05; ns: non-significant.

To better characterize the testes of DMRT1^+/− rabbits at 4 months, we conducted transcriptomic analysis using bulk RNA sequencing (Figure 4). The heatmap representation illustrates 1210 differentially expressed genes between WT and DMRT1^+/− rabbit testes (adjusted p-value <0.05, and fold change (FC): FC < 1/1.5 or FC > 1.5) (Figure 4A). A DAVID analysis was performed to determine which signaling pathways are dysregulated in DMRT1^+/− rabbit testes (Figure 4B). Consequently, the 779 over-expressed genes in DMRT1^+/− testes are involved in the regulation of cell differentiation, cell adhesion, and cell proliferation. Meanwhile, the 431 under-expressed genes in mutant rabbits are associated with sexual reproduction and gamete generation (Figure 4B). Among the 1210 differentially expressed genes, a panel of 25 genes was selected on the basis of the following criteria: (1) the genes are well known and characterized for their gonadal or testicular function or (2) the genes were highlighted by the previous DAVID analysis (Figure 4C). Among the over-expressed gene clusters, we identified Sertoli cell differentiation genes such as SOX9 ((Sex Determining Region Y)-Box 9) and AMH (Anti-Mullerian Hormone), as well as meiotic germ cell differentiation genes like STRA8 and MEIOSIN. In the under-expressed gene clusters, we observed genes involved in oocyte/spermatozoa fusion such as ADAM7 and SPACA3 (Sperm Acrosome Associated 3), and genes important during spermatid maturation such as SPEM1 (Spermatid Maturation 1) and SPEM2 (SPEM Family Member 2) (Figure 4C). These results corroborate the histological analyses and confirm impaired spermatid maturation followed by spermiogenesis failures. Furthermore, the overexpression of key meiotic genes suggests dysregulations in the SSC pool with a disrupted balance between differentiation and renewal in favor of meiosis entry. To confirm this result, we conducted STRA8 and MEIOSIN mRNA in situ detection using the RNAscope technology on 4-month-old testes from WT and DMRT1^+/− rabbits (Figure 5). As expected, the percentage of STRA8-positive seminiferous tubules tends to be higher in XY DMRT1^+/− testes compared to WT (Figure 5A), while the percentage of MEIOSIN-positive seminiferous tubules is significantly higher (42% vs. 30%) in DMRT1^+/− testes than in wild-type (Figure 5B).

Figure 4.

Deregulation of gene expression in DMRT1^+/− rabbits at 4 months. (A) Heatmap representation of 1210 deregulated genes (DEGs) (adjusted p-value < 0.05 and FC < 1/1.5 or FC > 1.5), obtained by RNA-seq analyses (DEseq2). (B) These DEGs were submitted to DAVID analyses. Their biological significance was explored by GO term enrichment analyses including biological process, molecular function, and cellular component and by KEGG pathway. Ten GO terms are presented on each of these graphs (left: over-expressed genes, right: under-expressed genes in rabbit mutant testes). The red dashed line corresponds to an enrichment score >1.3. (C) Heatmap representation of 25 deregulated genes (adjusted p-value < 0.05, and FC < 1/1.5 or FC > 1.5).

Figure 5.

Meiosis defect in DMRT1^+/− rabbits. In situ hybridization (RNAscope technology) of (A) STRA8 or (B) MEIOSIN in wild-type and DMRT1^+/− testis (brown) at 4 months old. The graphs represent the percentage of STRA8-positives tubes or MEIOSIN-positives tubes, respectively. The median is shown by the wide horizontal line, and each dot represents one animal (n = 4 for wild-type and DMRT1^+/−, number of tubules counted from 643 to 1240 in WT and 291 to 739 in DMRT1^+/− for STRA8 and 604 to 1087 in WT and 304 to 788 in DMRT1^+/− for MEIOSIN). Mann–Whitney test: ^*p-value < 0.05; ns: non-significant.

In conclusion, XY DMRT1^+/− rabbit testes exhibit spermatogenesis defects as early as 4 months, characterized by different defects such as an increase of spermatogonia committed to differentiation, failure in the meiotic process, a significant decrease in spermatid formation, and subsequent spermiogenesis impairment.

Loss of pluripotency of the germ cells is slightly delayed in XY DMRT1^+/− neonatal testes

As DMRT1 haploinsufficiency is detrimental for spermatogenesis as early as 4 months of age, we then focused on the neonatal period, corresponding to XY germ cell differentiation steps such as their relocation to the base of the seminiferous tubules and the re-methylation process of their DNA. We first performed HES staining on WT and DMRT1^+/− testicular sections at different stages: 3, 6, 10, and 25 days post-partum (dpp) (Figure 6). At 3 dpp, germ cells were predominantly located in the center of the seminiferous tubules (Figure 6A and E), while at 6 dpp they began to relocate to the base of the cords in both testes (WT and DMRT1^+/−), but with a slight delay in the heterozygous ones (Figure 6B and F). At 10 dpp, the majority of WT germ cells were relocated to the base of the seminiferous cords. However, many DMRT1^+/− germ cells were not correctly relocated at this 10 dpp stage, and some appeared to display a pycnotic nucleus, suggesting the possibility of degenerative processes via apoptosis (Figure 6C and G). At 25 dpp, WT and DMRT1^+/− germ cells seemed correctly positioned at the base of the seminiferous cords, thus constituting the pool of spermatogonia (Figure 6D and H).

Figure 6.

Establishment of spermatogenesis in wild-type or DMRT1^+/− rabbits at 3, 6, 10, and 25 dpp (days post-partum). (A–D) HES staining on wild-type testis. (E–H) HES staining on DMRT1^+/− testis. (G) Black arrows indicate abnormal germ cells in the seminiferous cords of the DMRT1^−/− testis.

To further characterize the stages of differentiation and relocation of male germ cells immediately after birth, we studied the expression of DMRT1, Ki67, and OCT4 (POU5F1 gene) by IHC, comparing WT and DMRT1^+/− testes (Figure 7). As expected, DMRT1 is detected in Sertoli and germ cells before birth at 28 dpc and just after birth at 3 dpp, but its germinal expression decreases between 6 and 16 dpp, a period corresponding to the relocation of the germ cells from the luminal to the basal compartment of the seminiferous tubules (Figure 7A and Supplementary Figure S3). In DMRT1^+/− testes, the expression of DMRT1 appears to persist at 6 dpp in germ cells located in the center of the cords, which are not yet relocated. Finally, DMRT1 is re-expressed at 25 dpp in spermatogonia of both WT and DMRT1^+/− testes, with an apparently strongest intensity compared to Sertoli cells (Figure 7A). Interestingly, the proliferation marker Ki67 exhibits a similar expression profile to that of DMRT1 in the germinal lineage (Figure 7B). Indeed, Ki67 germinal expression decreases from 6 to 10 dpp, during the relocation process, then it is re-expressed at 25 dpp (Figure 7B). Remarkably, DMRT1^+/− germ cells seem to remain proliferative for a longer period than those of WT testes, as the number of Ki67-positive germ cells is significantly higher in DMRT1^+/− testes than in WT at 3 dpp, and more of them remain positive at 10 dpp, if they have not relocated to the base of the cords (Figure 7B). These observations on germ cell proliferation led us to check the expression of OCT4, encoded by the POU5F1 gene, to appreciate the pluripotency status of the germinal lineage. Very clearly and interestingly, pluripotent OCT4-positive germ cells remain visible up to 25 dpp in the DMRT1^+/− testes, whereas none of these cells are detectable in WT testes after 3 dpp (Figure 7C). Consistently, germ cells remaining positive for OCT4 are located at the center of the seminiferous cords (Figure 7C). Finally, we assessed the germ cell DNA re-methylation process in WT and DMRT1^+/− testes. While some germ cells were slightly positive for 5mC at 6 dpp in WT testes, none were 5mC-positive in DMRT1^+/− testes (Supplementary Figure S4). This observation certainly reflects only a small delay in the re-methylation process since at 10 dpp, all germ cells are positive for 5 mC in both conditions (Supplementary Figure S4).

Figure 7.

Defects in the establishment of the spermatogonial stock in DMRT1^+/− rabbits during the post-natal period. Immunohistochemistry of (A) DMRT1, (B) Ki67, and (C) OCT4 in wild-type and DMRT1^+/− testis (brown) at 3 dpp, 6 dpp, 10 dpp, and 25 dpp. The graphs represent the density per mm² of cells positive for Ki67 or OCT4, respectively, with between 3 and 4 areas per individual (n = 3 for wild-type and DMRT1^+/−). Mann–Whitney test: ^**p-value < 0.005; ^***p-value < 0.0005; ^****p-value < 0.00005; ns: non-significant.

To better characterize the impact of DMRT1 haploinsufficiency on the establishment of the SSC pool, we performed a transcriptomic analysis at 1-day post-partum using bulk RNA-seq. The heatmap representation illustrates the 138 differentially expressed genes between WT and DMRT1^+/− rabbit testes (adjusted p-value < 0.05, FC < 1/1.5 or FC > 1.5) (Supplementary Figure S5A). Of the 75 genes over-expressed in XY DMRT1^+/− rabbit testes, a DAVID analysis to identify deregulated signaling pathways in mutant rabbit testes indicates that these genes are related to cell cycle and meiosis (Supplementary Figure S5B). Conversely, the 63 genes under-expressed in mutant rabbit testes are involved in reproduction, cell adhesion, and the cell cycle.

In conclusion, XY DMRT1^+/− rabbit testes experience some differences compared to WT ones in the establishment of the SSC pool occurring after birth. A higher proportion of DMRT1^+/− germ cells failed to lose their pluripotency and did not properly relocate to the basal lamina of the seminiferous tubules. Do these early differences influence the number of SSCs well before puberty, or is it a reflection of the disturbance in the pluripotency/differentiation balance that will persist throughout reproductive life? This question remains challenging to answer in this rabbit model, where the loss of a dose of DMRT1 is present in all testicular cells and from the earliest developmental stages where DMRT1 is expressed.

Discussion

Our study describes an unexpected phenotype linked to haploinsufficiency of the DMRT1 gene in male XY rabbits. Specifically, the loss of a single DMRT1 allele leads to infertility in male animals as they age. We show here that this infertility, which becomes more severe with age, results in a substantial decline in sperm count, notable alterations in sperm parameters, and even a spermatogenesis arrest ~2 years of age. This corresponds to the first fifth of a rabbit’s life, thereby significantly reducing the reproductive lifespan of the animals [34]. This arrest in spermatogenesis appears to be due to a combination of factors that become apparent during the post-natal period, when the stock of spermatogonial stem cells (SSCs) is set up. Subsequently, the renewal/commitment balance of SSCs appears to be disturbed from the age of 4 months, when the first spermatozoa are normally produced in rabbits. Furthermore, disruption to germ cell meiosis has been observed at various stages, and DMRT1 haploinsufficiency also affects the proper functioning of Sertoli cells, with clear repercussions on spermiogenesis.

As will be discussed subsequently, the various roles of DMRT1 in male fertility have been very clearly established in mice, particularly due to the extensive work conducted by David Zarkower’s team [29]. The mouse model is absolutely essential for advancing our understanding of the molecular and cellular mechanisms associated with a major and highly conserved transcription factor such as DMRT1. The complex genetic models, the numerous immunological tools, and the various high-throughput technologies enable the precise characterization of the gene cascades controlled by DMRT1 in different cell types in the mouse. However, the mouse model only partially reproduces the observation made in human genetics. First, sensitivity to gene dosage is much less marked in mice than in humans. Second, the loss of function of a murine gene in a specific cell type at a given moment in development, as may be the case in conditional knockout models, may not reflect the situation observed in human genetics.

In contrast, the rabbit model presented here is more analogous to human genetics, where a patient harbors a mutated DMRT1 allele either inherited from one of their parents or a de novo mutation that occurred during early development. It is noteworthy that the observed failures in the DMRT1^+/− rabbit testes are consistent with the established roles of DMRT1 in mice [29]. Chronologically, the initial dysfunction observed in DMRT1^+/− rabbit testes is detectable within 10 days of birth and concerns the relocation of male SSCs from the lumen of the seminiferous tubules to their base, where the spermatogenic niche will form. A role for Dmrt1 in this radial migration of the germ cells to the periphery of the seminiferous tubule has been clearly documented in mice and found to be the result of a cell-autonomous process, directly linked to Dmrt1 germinal expression [22]. Moreover, concomitant with their radial migration, the SSCs leave their pluripotent state characterized by the extinction of POU5F1 gene expression (encoding OCT4) and by a transient stop in their proliferation. This appears to be directly linked to a decrease in DMRT1 expression in rabbits. In mouse studies, the germinal expression of Dmrt1 has been reported to be essential for mitotic reactivation and the survival of germ cells during the first week of life [22]. The persistence of germ cell pluripotency marker expression and the link between DMRT1 and male germ cell proliferation have also been well documented from mice fetal stages, where conditional loss of Dmrt1 function in the germ line has been shown to result in the appearance of testicular teratomas [28]. Interestingly, in this latest study, the authors have concluded that the genetic link between testicular dysgenesis, pluripotency regulation, and susceptibility to teratomas is highly sensitive to genetic background and gene dosage [28]. This finding serves to reinforce the importance of the expression level of the DMRT1 gene and the effects of its haploinsufficiency described here. The sensitivity to genetic background described in mice must also be related to the strong variability observed between different XY rabbit individuals as to the stage of appearance and severity of the secondary infertility phenotype. It is notable that our rabbit lines are not congenic, and that the inter-individual genetic variability is more pronounced than observed in the mouse model.

Subsequently, during the spermatogenetic process, Dmrt1 remains crucial in maintaining the SSC pool by acting on the balance between renewal and differentiation through the meiotic program initiation. DMRT1 favors SSC renewal by directly repressing Stra8, thus preventing meiosis engagement, and activating Sohlh1 which promotes spermatogonial development [27]. Interestingly, in DMRT1^+/− rabbit testes from 4 months of age, we observe an increasing number of meiotic STRA8-positive germ cells. We also showed a significant elevated number of MEIOSIN-positive cells, suggesting a potential link between DMRT1 in the control of MEIOSIN gene expression. Finally, regarding the germinal expression of DMRT1, we confirm its complete down-regulation during the zygotene and pachytene stages of the meiotic process, in accordance with previous human studies [41, 42]. However, despite this early termination of expression during meiosis, the spermatogenetic process appears to be highly disrupted in DMRT1^+/− rabbit testes during the final stages of spermiogenesis, when round spermatids differentiate into spermatozoa. These spermiogenesis defects might be due to the primary cell-autonomous role of DMRT1 in SSCs, but they should more certainly reflect the inability of supporting Sertoli cells to complete spermatogenesis, as shown by conditional Dmrt1 ablation in Sertoli cells in mice [22].

The main outcome of this work in rabbits is that the dose, and thus the level of expression of the DMRT1 gene, crucially influences the maintenance of testicular homeostasis and function for male gamete production. This is achieved through DMRT1 exerting critical control over different steps of spermatogenesis and Sertoli cell integrity. The secondary infertility phenotype observed in rabbits following haploinsufficiency of DMRT1 in rabbits was certainly unexpected, given the data obtained in mice. However, it is in perfect agreement with the roles of this gene described in humans and birds, where the dose of DMRT1 has been proved to be crucial for its functions. These discrepancies between humans and mice regarding gene dosage are recurrent concerning genes involved in gonadal differentiation. Indeed, many genes encoding transcription factors, such as NR5A1 (Nuclear Receptor Subfamily 5 Group A Member 1, or SF1, Steroidogenic Factor 1) [43–46], WT1 (Wilms Tumor 1) [47, 48], FOXL2 (Forkhead Box L2) [49, 50], or SOX9 [51, 52] as examples, have been demonstrated to induce DSD in humans due to haploinsufficiency. In contrast, a complete loss of function is necessary to obtain similar phenotypes in mice. This evolutionary difference in gene-dosage sensitivity between humans and mice undoubtedly influences various genetic parameters. In mice, it increases reproductive robustness, whereas in humans, it reduces the genetic load of deleterious mutations, where, in the case of DMRT1, these mutations will be purged by spermatogenic failures [53].

In conclusion, our results reinforce previous studies in human genetics showing that specific heterozygous mutations in the DMRT1 gene can be responsible for male infertility [29, 32]. Furthermore, our findings highlight that the expression level of DMRT1 is crucial for the maintenance of diverse testicular functions throughout the reproductive lifespan. This is consistent with its pivotal testis-determining role in birds, rabbits, and humans [6, 7, 15, 30].

Conflict of interest

The authors have declared that no conflict of interest exists.

Author contributions

IB: investigation, formal analysis, methodology, validation, writing—original draft, writing—review and editing. ED, AD, MA, DT, CLD: contributed to the data acquisition, methodology, writing—review and editing. AF, LJ, BM-P: bioinformatics analysis. LJ: data curation. GV, MP: conceptualization, methodology, writing—review and editing. BM-P, EP: conceptualization, supervision, funding acquisition, writing—original draft, writing—review and editing.

Data availability

The datasets generated for this study can be found in the sequence read archive at https://www.ncbi.nlm.nih.gov/sra/PRJNA1154456.