A total of 30 C57BL/6 male mice were purchased from the experimental animal center of Sun Yat-sen University (Guangzhou, China) and Xiamen University (Xiamen, China), and maintained in specific pathogen-free (SPF) environment (12:12 h light/dark cycle) with free access to food and water. Human semen samples were collected from 145 consecutive male outpatients (aged 30–50 years) for fertility examination with normal semen parameters that meet the standard of the fifth edition of the World Health Organization manual of semen analysis (Cooper, 2010). All patients were abstinent for 2–7 days before examination. The semen samples were obtained by masturbation and allowed to liquefy at 37°C for 20 min. Seminal plasma was obtained after centrifugation (3,000 × g, 15 min). Human testis tissues for immunofluorescent staining analysis were collected from 3 obstructive azoospermia patients by diagnostic testicular biopsy.

Before euthanasia, the muscle strength of limbs was assessed by a wire screen holding test described previously (Carlson et al., 2010). The muscle strength was evaluated by the physical impulse (Fdt) (N·s) = body mass (g) × 0.00980665 N/g × holding time(s). After euthanasia with an intraperitoneal injection of excess pentobarbital sodium, the femurs and tibias were harvest. And the Bone mineral density was measured using a Siemens Inveon Micro-CT scanner (Siemens Medical Solutions United States Inc, Malvern, PA, United States).

The cauda epididymidis (2 mm in length) was minced in 1 ml Biggers, Whitten, and Whittingham medium (BWW) containing 0.5 ml 4% bovine serum albumin (BSA) prewarmed to 37°C, then cut into small pieces and incubated for 10 min at 37°C to release the sperm. Sperm concentration and motility were measured with a computer aided sperm analysis (CASA) system (Hamilton Thorne, Beverly, MA, United States). The average of 5 visual fields was calculated.

Blood samples were collected from mice after euthanasia. Serum was obtained after centrifugation (3,000 × g, 15 min, 4°C), and stored at −70°C until assayed. The testosterone concentrations were measured by electrochemiluminescence immunoassay using Elecsys Testosterone II (05200067190, Roche, Basel, Switzerland) according to the manufacturer’s instructions.

Four mice each of 3 months old and 21 months old were used for RNA sequencing analysis. A pair of bilateral testes, efferent ductules, epididymides and vas deferens of the same mouse was dissected into a 10-cm plate containing 10 ml Phosphate Buffered Saline (PBS). After the removal of any fat and excessive connective tissue, seven regions of the male reproductive tract: testis, efferent ductules, initial segment, caput epididymidis, corpus epididymidis, cauda epididymidis, and vas deferens, were isolated under a dissection microscope as previously described (Nakata and Iseki, 2019). Total RNA was extracted using a RNeasy mini kit (Qiagen, Hilden, Germany) according to the manufacturer’s protocol. The library was prepared using Illumina TruSeq Stranded Total RNA Library Prep Kit (Illumina Inc., San Diego, CA, United States), and paired-end 150 bp RNA-seq was performed on Illumina NovaSeq (6,000 (Illumina Inc.). Reads were aligned with Hisat2 to the mouse genome (Ensemble Mouse GRCm38). Differential gene expression was assessed using DEseq2 R package. High-throughput sequencing of exosomal mRNA in seminal plasma was conducted according to the method reported in our previous study (Xie et al., 2020).

Total RNA was extracted from tissues or semen plasma using a RNAiso Plus (9,109, Takara, Japan) according to the manufacturer’s protocol. Total RNA (2 µg) was used as a template to synthesize cDNA with PrimeScrip RT Master Mix Kit (RR036A; Takara, Japan) at 37°C for 15 min and 85°C for 5 s. The qPCR was performed using the TB Green Premix Ex Taq II (RR820A, Takara, Japan) according to the manufacturer’s instructions. The qPCR program using the LightCycler 480 II real-time PCR machine (Roche, Basel, Switzerland) was then carried out as follows: initial denaturation at 95°C for 30 s and 80 cycles of 95°C for 5 s and 60°C for 20 s; with a final step melting curve of 95°C for 5 s; 60°C for 1 min and 95°C for 5 s. All the primers used are listed in Supplementary Table S1 (Sangon Biotech, Shanghai, China).

The mouse testis, efferent ductules, initial segment, caput, corpus and cauda epididymidis, and vas deferens were fixed in 4% paraformaldehyde. After steps of dehydration, paraffin-embedding, slices at 5-μm thickness were used for hematoxylin and eosin (H&E) staining. β-galactosidase staining in frozen mouse testis and epididymis tissues was performed using a senescence β-galactosidase staining kit (G1073-100T, Servicebio, Wuhan, China) according to the manufacturer’s instructions.

Paraffin slides of the mouse testis, epididymidis, and human testis tissues were deparaffinized in xylene and then rehydrated in graded alcohol solutions. The endogenous peroxidase activity was inhibited by incubating in 3% H₂O₂, and the antigenicity was recovered with 1 × Tris-EDTA buffer (pH 9.0, G1203, Servicebio, Wuhan, China). The samples were permeabilized with 0.3% Triton-X100 for 10 min, then incubated with 5% Bovine Serum Albumin for 30 min. For immunohistochemistry staining in mouse testis, the primary antibody included anti-Pla2g2d (1:100, PA5-93147, Invitrogen, United States) was used. PBS buffer was used as negative control. After washed with PBS then incubated with HRP Conjugated Goat Anti-Rabbit IgG (1:100, CW0103, CWBIO, Beijing, China) at room temperature for 30 min, the sections were counterstained with Hematoxylin Staining Solution (ZLI-9610, ZSGB-BIO, Beijing, China). For immunofluorescent staining in mouse epididymis and human testis, the primary antibody included anti-Pla2g2d (1:100, PA5-93147, Invitrogen, United States) and anti-PRM2 (1:100, 14500-1-AP, Proteintech Group, Inc, China) were used. After washed with PBS then incubated with CY3 Conjugated AffiniPure Goat Anti-rabbit IgG (1:100, BA1032, Boster Biological Technology co.ltd, Wuhan, China) at room temperature for 30 min, the sections were counterstained with DAPI (10 μg/ml, C1002, Beyotime Biotechnology, Shanghai, China). Immunofluorescent staining in human sperm was conducted according to previous study (Liu et al., 2015). Images were captured with a Leica inverted fluorescence microscope (Leica, Wetzlar, Germany) or an Olympus inverted microscope (Olympus, Tokyo, Japan).

Human seminal plasma was obtained after centrifugation (3,000 × g, 15 min, 4°C). The supernatant was centrifuged at 12 000 × g for 30 min at 4°C to remove cellular debris. The PLA2G2D levels of 40 samples from different ages (30–50 years) were determined using an ELISA kit according to the manufacturer’s instructions (LS-F4416, LifeSpan BioSciences Inc., Seattle, WA, United States).

Gene Ontology (GO) enrichment analysis was performed using the Gene Ontology resource database (http://geneontology.org/) and visualized using R language packages. The GO terms with a FDR-adjusted P-value less than 0.01 were thought significant enriched. The PPI network was constructed using STRING database (https://www.string-db.org/) and visualized by Cytoscape software (https://cytoscape.org/). Upset plot and GO Correlation Network plot were performed using the OmicStudio tools at http://www.omicstudio.cn/.

SPSS 26.0 statistical software (IBM Corp., NY, United States) was used for statistical analysis. All data were presented as the mean ± standard deviation (SD) obtained from at least four independent experiments. Comparisons between groups were performed using Student’s t-test, and p < 0.05 was considered statistically significant.

To determine whether 21-month-old male mice are a suitable model for the natural reproductive aging, we comprehensively studied phenotypes associated with reproductive aging as well as systemic aging. The 21-month-old male mice exhibited physical features of increased body weight and loss of hair and luster (Figure 1A). Both testis and epididymis indices (organ weights relative to the mouse body weight) were significantly decreased (p < 0.05) (Figures 1B,C). However, no evident morphological differences were detected between the young and elderly mice, except for a decrease in plumpness with age (Figure 1B). Noticeably, the initial segment had the richest blood supply across the epididymis at both ages (Figure Figure1B). Furthermore, the limb muscle strength, a simple indicator of systemic aging and the bone mineral density of the femur decreased significantly with age (p < 0.05) (Figures 1D,E). The serum testosterone concentration, epididymal sperm concentration, and motility decreased remarkably in the elderly group (p < 0.05) (Figures 1F–H). Collectively, the 21-month-old male mice showed characteristics of both systemic and reproductive aging.

Previous analyses paid little attention to the histological characteristics of male reproductive tract aging, and it is unclear which region responsible for reproductive function arises age-related changes. To investigate the biological basis for the reproductive aging in mice, we further performed histological analyses on the testis and epididymis by Hematoxylin and eosin (HE) staining and β-Galactosidase (β-Gal) staining. The testes of elderly mice had dilated and empty lumens of the seminiferous tubules, as well as disordered and thinner seminiferous epithelium (Figure 1I, left panel). In the epididymis, dilatation of the lumen was also observed with age. Notably, the arrangement of ducts in the initial segment and caput epididymidis was compact in the young group; in contrast, it was much looser in the elderly group (Figure 1I, right panel). Moreover, β-galactosidase (β-Gal) in the testis and epididymis showed strong positive staining in the testicular interstitium, the initial segment of the epididymis, and a small area of caput epididymidis connected to the initial segment in aged mice (Figure 1J). In addition, we also detected non-specific histological changes in the efferent ductules and vas deferens, which showed an increased duct diameter, thickening duct wall with age (Supplementary Figure S1).

We performed transcriptome profiling of seven well-defined anatomical regions across the male reproductive tract of 3-month-old adult mice (Figure 2). Genes with fold change larger than or equal to 4 relative to the expression levels in the other 6 regions and p < 0.05 were defined as region-specific genes. A principal component analysis (PCA) of the RNA-seq data is showed in Figure 2A and Supplementary Table S2. Interestingly, the unsupervised hierarchical classification, based on region-specific genes, showed a relative affinity between the transcriptome profiles of the initial segment and the caput, corpus, and cauda, and between the vas deferens and epididymal regions. The transcriptome features of the efferent ductules were more similar to those of other post-testicular tube-like structures than to those of the testis (Figure 2B; Supplementary Table S3). These results were consistent with the anatomical distribution of the seven regions across the male reproductive tract. The embryonic origin of the male reproductive tract may explain this coincidence, since the tubular structures of male reproductive duct such as the vas deferens and epididymis are well-known to be derived from the mesonephric duct (also known as Wolffian duct), whereas the testis derived from the primordial genital ridge(Barsoum and Yao, 2006).

We identified a series of region-specific genes of male mouse reproductive tract (Figures 2B,C), which were in broad agreement with those of previous RNA-seq and microarray analyses of murine transcriptome profiles, such as ADAM metallopeptidase domain 28 (Adam28) in initial segment, Transmembrane Epididymal Protein 1 (Teddm1b) in caput, and serine peptidase inhibitor, Kunitz type 3 (Spint3) in corpus epididymis (Johnston et al., 2005; Shi et al., 2021) (reviewed in Supplementary Table S4). GO analysis revealed the functional characteristics of the region-specific genes (Figure 2D; Supplementary Table S5). For example, in the biological process (BP) category, genes expressed in the testis were enriched for the terms “spermatogenesis” and “spermatid development,” genes in the efferent ductules were enriched for the terms “transmembrane transport” and “water homeostasis,” and genes in the caput epididymidis were enriched for “defense response.” The top 20 genes specifically expressed in each region are shown in Figure 2E; Supplementary Table S6. Overall, our bulk RNA-seq analysis provides reliable datasets and shows the transcriptome landscape of the male reproductive tract at an anatomical resolution.

As region-specific mRNA or protein detection in the seminal plasma enables the noninvasive diagnosis of azoospermia (Li et al., 2012; Xie et al., 2020), we verified the region-specific genes expressed in the mouse testis and four regions of the epididymis, as well their potential uses as seminal markers for the male reproductive tract in humans (Figure 3A). First, we filtered 19 testis-specific and 64 epididymal region-specific genes out of 100 region-specific genes ranked top 20 in these five regions that had conserved orthologs in Homo sapiens by browsing the GeneCards database (Figure 3B). Then, 15 and 42 genes plotted in the protein-protein interaction (PPI) network were further picked out using STRING database (Figure 3C). Subsequently, we confirmed 13 and 11 out of these genes with strict testis- or epididymis-specific expression pattern across the human organs by Human protein atlas database (Figure 3D). Most of the products of these 13 testis-specific genes were germ cell-specific and non-secretory proteins, while all 11 epididymis-specific gene products were secretory proteins likely to be detected in human seminal plasma (Figure 3D). Therefore, we further sequenced the human seminal exosomal RNA profiles, compared seminal RNA-seq data with these 13 testis-specific genes, and obtained nine testis-specific genes present in human seminal plasma (Figure 3E; Supplementary Table S7). Additionally, we compared 11 epididymis-specific genes with two prior proteomes of human seminal plasma (Batruch et al., 2011; Rolland et al., 2013), yielding five candidate genes detected in both proteomic datasets (Figure 3F; Supplementary Table S7). According to our RNA-seq data, these 14 region-specific genes in mice showed typical region-specific expression patterns (Figure 3G) and had invariant expression levels in 3-month-old and 21-month-old mice (Supplementary Figure S2).

We validated the expression patterns of these 14 region-specific genes in the testis and epididymis by qPCR (Supplementary Figure S3). Except for Transition protein 1 (Tnp1) and Spint3, the mRNA expression levels of the other 12 genes (eight for the testis and four for epididymal regions) were significantly enriched in the corresponding regions of the mouse testis and epididymis. The expression of Protamine 2 (PRM2) in the human testis was also validated by immunofluorescence staining, showing exclusive expression within the seminiferous tubules and sperm (Supplementary Figure S4). Collectively, we screened and validated 12 region-specific genes that are potential markers for the testis and the four regions of the epididymis (Figure 3H). This finding was a windfall benefit of this transcriptomics analysis, meanwhile, it also verified the reliability of our sequencing data.

We set out to characterize age-related transcriptome changes in each individual region of male reproductive tract by comparing the datasets from the 3-months and 21-months mice. The genes with at least 1.5-foldchange (upregulated or downregulated by FC ≥ 1.5) were defined as age-related differentially expressed genes (DEGs). The heatmaps of age-related differentially expressed genes (DEGs) showed intra-group clustering at the transcript level (Figures 4A–G). Numbers of age-related DEGs also showed region-dependent patterns (Figure 4H; Supplementary Table S8). Surprisingly, the testis showed a limited number of age-related DEGs (Figures 4A,H). There were 2926 upregulated and 2586 downregulated age-related DEGs in all seven regions of the male reproductive tract (Figures 4I,J). We also assessed overlap in gene expression changes among all seven regions. Five genes (Pla2g2d, Cd209 antigen b, Cd209 antigen f, Cd209 antigen g, and C-C motif chemokine ligand 8) were commonly upregulated in all seven regions, while two genes (Elastin and Lysyl oxidase) were commonly downregulated at 21 months (Figures 4I,J; Supplementary Table S9).

To interpret age-related functional changes, we analyzed the biological processes involved with these age-related DEGs. Of note, GO enrichment analysis revealed that the upregulated DEGs were largely associated with BP terms related to the immune response (Figure 4K; Supplementary Table S10), such as “Positive regulation of cytokine production” in testis, “T cell activation” and “leukocyte cell-cell adhesion” in post-testicular regions. Downregulated DEGs were mainly enriched for “Extracellular matrix organization,” “Angiogenesis,” and “Collagen fiber organization” (Figure 4L; Supplementary Table S9), suggesting that the structures of reproductive organs may undergo degenerative changes. In general, the functional analysis revealed increased immune response activities in the aging male reproductive tract.

Since the testis and epididymis are two of the most important organs for reproductive function, we conducted GO enrichment analysis of age-related DEGs in the testis and epididymis. The upregulated DEGs enriched for functions related to “T cell activation” (Figure 5A). According to the UpSet plot (Figure 4I), Pla2g2d was the only common upregulated DEG related to “T cell activation” in all five regions (Figure 5B; Supplementary Table S11). The expression patterns of Pla2g2d with age in the testis and four regions of the epididymis in mice were validated by qPCR (Figures 5C,D). And due to the interference of testicular autofluorescence, immunofluorescence staining of Pla2g2d could not be done in the testis. Therefore, we carried out immunohistochemistry (IHC) for Pla2g2d in the mouse testis, which confirmed the upregulated expression level with age (Figure 5E). The immunofluorescence staining of Pla2g2d in the four regions of mouse epididymis shows the increasing protein expression trend of Pla2g2d (Figures 5F–I), which is consistent with our RNA-seq data.

To determine whether human PLA2G2D is also upregulated in seminal plasma with age, we collected 145 human semen samples from consecutive outpatients for fertility examination from 30 to 50 years old with normal semen parameters, which all met the standard of the fifth edition of the WHO manual of semen test. As validated by qPCR and ELISA, PLA2G2D expression at both the mRNA and protein levels in human semen plasma remarkably increased with age (Figures 5J,K). It is worth noting that the cut point of the elevated expression level of PLA2G2D appears at the age of 35 years (p < 0.05) (Figures 5J,K). The level of PLA2G2D mRNA is almost undetectable until the age of 35, and surges since then (Figure 5J), similar result was also demonstrated at the protein level (Figure 5K). These results suggest that Pla2g2d plays a critical role in the aging of the testis and epididymis and could be a promising biomarker of male reproductive aging in humans.

Previous histological analyses showed that epididymal aging occurs in a region-dependent manner (Figure 1J); accordingly, we focused on aging in the epididymis for a detailed analysis. The upregulated DEGs related to the top biological processes with the maximal -log₁₀ (adjust P-value) in each region of the epididymis were analyzed (Figures 6A–D; Supplementary Figure S5). The STRING database was further used to construct Protein–protein interaction (PPI) networks of the upregulated DEGs involved in these top biological processes in each region of the epididymis (Figures 6E–H). By qPCR, we validated the expression patterns of several hub genes with high degree values in the PPI network, including Protein tyrosine phosphatase receptor type C (Ptprc) in the initial segment, Lymphocyte protein tyrosine kinase (Lck) in the caput epididymidis, Microtubule associated protein tau (Mapt) in the caput epididymidis, and Interferon induced protein with tetratricopeptide repeats 3 (Ifit3) in the cauda epididymidis. All four genes exhibited increased expression with age (p < 0.05) (Figure 6I), in accordance with the RNA-seq data.