Identification of novel cytoskeleton protein involved in spermatogenic cells and sertoli cells of non-obstructive azoospermia based on microarray and bioinformatics analysis

Abstract

Background

During mammalian spermatogenesis, the cytoskeleton system plays a significant role in morphological changes. Male infertility such as non-obstructive azoospermia (NOA) might be explained by studies of the cytoskeletal system during spermatogenesis.

Methods

The cytoskeleton, scaffold, and actin-binding genes were analyzed by microarray and bioinformatics (771 spermatogenic cellsgenes and 774 Sertoli cell genes). To validate these findings, we cross-referenced our results with data from a single-cell genomics database.

Results

In the microarray analyses of three human cases with different NOA spermatogenic cells, the expression of TBL3, MAGEA8, KRTAP3-2, KRT35, VCAN, MYO19, FBLN2, SH3RF1, ACTR3B, STRC, THBS4, and CTNND2 were upregulated, while expression of NTN1, ITGA1, GJB1, CAPZA1, SEPTIN8, and GOLGA6L6 were downregulated. There was an increase in KIRREL3, TTLL9, GJA1, ASB1, and RGPD5 expression in the Sertoli cells of three human cases with NOA, whereas expression of DES, EPB41L2, KCTD13, KLHL8, TRIOBP, ECM2, DVL3, ARMC10, KIF23, SNX4, KLHL12, PACSIN2, ANLN, WDR90, STMN1, CYTSA, and LTBP3 were downregulated. A combined analysis of Gene Ontology (GO) and STRING, were used to predict proteins’ molecular interactions and then to recognize master pathways. Functional enrichment analysis showed that the biological process (BP) mitotic cytokinesis, cytoskeleton-dependent cytokinesis, and positive regulation of cell-substrate adhesion were significantly associated with differentially expressed genes (DEGs) in spermatogenic cells. Moleculare function (MF) of DEGs that were up/down regulated, it was found that tubulin bindings, gap junction channels, and tripeptide transmembrane transport were more significant in our analysis. An analysis of GO enrichment findings of Sertoli cells showed BP and MF to be common DEGs. Cell-cell junction assembly, cell-matrix adhesion, and regulation of SNARE complex assembly were significantly correlated with common DEGs for BP. In the study of MF, U3 snoRNA binding, and cadherin binding were significantly associated with common DEGs.

Conclusion

Our analysis, leveraging single-cell data, substantiated our findings, demonstrating significant alterations in gene expression patterns.

Introduction

It is estimated that 15% of couples worldwide are infertile, with male factors contributing to about 50% of cases [1]. There are 10% of these infertile men have azoospermia. Obstructive azoospermia (OA) and non-obstructive azoospermia (NOA) can be categorized as azoospermia [2, 3]. The cause of 60% of azoospermia in men is OA and the cause of 40% is NOA [4]. Mechanical obstructions of the reproductive tract are known to cause OA [5]. These include congenital bilateral vas deferens absence, seminal vesicle atresia, post-vasectomy, ejaculatory duct cysts, infection-induced epididymal obstruction, spinal cord injury, diabetes, or idiopathic obstruction of the reproductive tract [6]. Testicular and primordial germ cells defects may cause NOA. There are many causes of testicular NOA, including (i) spermatogenetic defects, such as microdeletion (microdeletion of one or more Y chromosomes) [7], (ii) genetic defects (e.g., Klinefelter syndrome), and (iii) infection and chemotherapy [8, 9]. As the name implies, pre-testicular azoospermia is caused by extra-gonadal endocrine disorders, such as hypothalamic or pituitary dysfunction, which lead to azoospermia [10, 11]. Up to 70% of NOA cases are idiopathic, which can lead to either meiotic arrest (MA) phenotypes where spermatocytes fail to differentiate beyond primary spermatocytes, or Sertoli cell only (SCO) phenotypes where germ cells are not present beyond spermatogonia in tubules [12, 13]. A SCO phenotype is characterized by the altered migration of primordial germ cells from yolk sacs to gonadal ridges, which results in the absence of spermatogonia, including spermatogonial stem cells (SSCs) [14].

Among the essential structures and environments for successful spermatogenesis, the Sertoli cells are the ‘nurse’ cells of the seminiferous epithelium [15, 16]. Columnar epithelial cells of the seminiferous tubule are polarised, irregularly shaped, and columnar in shape. Sertoli cells contain a large number of organelles in their basal regions. With the exception of cytoskeletal elements and mitochondria, Sertoli cells lack organelles, aside from long apical cytoplasmic processes [17, 18]. The Sertoli cells maintain many cytoskeletal junctions with germ cells, which serve to tether germ cells in place and remodel the spermatogenesis process [19]. Furthermore, Sertoli cells undergo cyclical remodeling to accommodate structural changes and movements of developing germ cells. These functions are achieved by Sertoli cells by utilizing a dynamic cytoskeleton with actin filaments, microtubules, and intermediate filaments. In Sertoli cells, each of these cytoskeletal structures plays an important role [20,21,22].

All of these components are crucial components of eukaryotic cytoskeletons, including actin filaments, microtubules, intermediate filaments, and septin proteins [23]. Each of these elements must be present for eukaryotic cells to divide, move, differentiate, migrate, adhere, and transport intracellularly. Actin filaments are clearly concentrated in certain areas of Sertoli cells when observed by microscopy, especially at adhesion junctions between Sertoli cells at their base, adjacent to spermatids during spermiogenesis, and at the site of spermiation, when mature spermatids are released [17, 24, 25].

A structure known as the ectoplasmic specialization (ES) is located at the junction between the basal aspects of adjacent Sertoli cells [26]. At this site, actin filaments are bundled into hexagonal arrays by the protein espin. A layer of endoplasmic reticulum surrounds the actin array in Sertoli cells [27, 28]. Intercellular adhesion is mediated by proteins found in Sertoli cell membranes. Other actin-related adhesion junctions are interspersed with the ES structures at this site, including adhesins, gap, and tight (occluding) junctions [29]. In addition to actin-based tubulobulbar complexes (TBCs), Sertoli cells at the basal aspect also contain unique endocytic structures that can internalize junctional structures [30, 31].

Several researches in human testes have demonstrated that Sertoli cell cytoskeletons are critical for spermatogenesis and targetable by toxicants that lead to male reproductive dysfunction [32]. Microtubules (MT) and F-actin cytoskeletons are polarized structures, with plus and minus ends for MT protofilaments and barbed and pointed ends for actin filaments, respectively, adjacent to the basement membrane and the lumen of the seminiferous tubules across the epithelium [33]. In the epithelium, cytoskeletons, such as Microtubules, but also F-actin (but limited to stage VIII tubules), lay perpendicular to the basement membrane, acting as tracks to facilitate transport of cellular cargo [34]. Through apical ectoplasmic specialization (ES), the cytoskeleton also supports the transport of developing haploid spermatids. Microtubule-associated proteins (MAPs) are also involved in cargo and germ cell transport [35]. In these events, the following MAPs are involved: (i) MT-dependent plus end motor proteins (such as KIF15) and minus end motor proteins (such as dynein 1) [36]. In addition, there are microtubule + TIPs (tracker proteins for microtubules plus ends, such as EB1) and -TIPs (targeting proteins for microtubules minus, such as CAMSAP2. Germ cell transport and sperm release may also be mediated by other microtubule and actin cytoskeletal proteins [34]. Therefore, these Microtubule-associated proteins play a crucial role in aiding intracellular protein trafficking and organelle transport (e.g., phagosome) during rodent spermatogenesis [37,38,39,40,41,42]. We examined the distribution of cytoskeleton proteins in the epithelium of the testes of normal or non-obstructive azoospermia men.

In spermatozoa and fertility, cytoskeleton genes (such as Katanin catalytic subunit A1 like 1 (KATNAL1), A-kinase anchor proteins, and other cytoskeleton protein binding genes) can be critically important; however, it is unclear whether cytoskeleton genes are necessary during fertility in vitro, and some studies have examined stage association in the human seminiferous epithelium [43]. The expression of this subfamily of cytoskeleton genes in male germ cells has been studied a few times. In this experimental study, we showed and analyzed the expression of the cytoskeleton genes in spermatogenic cellsand Sertoli cell by microarray and in-silico analysis.

Materials and methods

Human ethics and consent

The experiment was conducted in 2016 utilising testicular material obtained from three adult male patients with different medical histories, as well as three healthy people. The human material tests conducted here were approved by the local ethical committees, namely the Ethics Committee of the Medical Faculty of Heidelberg University in Germany and the Iran National Committee for Ethics in Biomedical Research at Amol University of Special Modern Technologies. The tests were carried out in accordance with the approved number Ir.ausmt.rec.1402.05, and all human participants provided informed written consent.

Testicular biopsies from patients with NOA and controls

The human material studies conducted at this facility received full approval from the respective ethical committees of the University Hospitals of Heidelberg, and Amol University of Special Modern Technologies. All methods were carried out in accordance with relevant guidelines and regulations. For the testicular biopsy, 3 specimens with NOA were obtained, ranging in age from 28 to 45. All of these patients’ causes of infertility were excluded, including a Y-chromosome microdeletion, CFTR gene mutation, antisperm antibodies, orchitis, testicular torsion, and varicocele. Two samples were collected per patient by the Institute for Anatomy and Cell Biology at the University of Heidelberg. Bioethics committees at the University of Heidelberg and the Amol University of Special Modern Technologies approved informed consent from all participants. The transcriptome of the control group (n = 3) was obtained using RNA prepared from normal testicular tissue with preserved spermatogenesis [22, 39, 44].

Human semen collection

Human ejaculate samples were collected from healthy participants. A routine spermatogenic cellsanalysis was performed according to the 2016 WHO guidelines. Based on these results, spermatogenic cellswere classified as either normal or motility-impaired (Table 1). Normal spermatogenic cellssamples were defined as containing > 20 million spermatozoa per milliliter, with over 50% of spermatogenic cellsbeing motile, more than 25% demonstrating rapid, linear progressive movement, and with ≤ 1 lymphocyte per high-power field. A total of 39 normal sperm samples were obtained. Ten of these samples were homogenized and used in experiment 1 for hybridization with human testis cDNA microarrays. The remaining 29 samples were processed individually and used in experiment 2. Semen samples were also collected from 24 patients with poor sperm motility, characterized by less than 40% motile spermatogenic cells, with fewer than 5% showing rapid linear progression, and ≤ 1 lymphocyte per high-power field. The specific parameters of the spermatogenic cellssamples are detailed in Table 1. After liquefaction at room temperature, all spermatogenic cellssamples were washed twice in phosphate-buffered saline (0.1 M PBS) before further processing.

Table 1 Characteristics of patients and sperm samples included in this research

Immunohistofluorescence staining of tissue

We fixed testicular cells from adult human in 4% paraformaldehyde after washing them with PBS. Paraplast Plus was used for dehydrating tissue samples, and microtome devices were used for chopping them at 10 μm thickness. A Superfrost Plus slide was used to mount tissue sections from testis tissue and kept at 25 °C until use. We washed samples with xylene and slowly replaced them with water in ethanol during immunohistofluorescence staining. In order to retrieve antigens from tissue samples, heat was applied to 94 °C for 18 min. 10% serum/0.3% Triton in PBS was used to block nonspecific binding sites in tissue samples. In the following steps, we used secondary antibodies specific for the incubation fluorochromes and treated the labeled cells with 0.2 mg/mL of 4′,6-diamidino-2-phenylindole (DAPI) dye. These samples were subjected to immunofluorescence staining as explained above. Zeiss LSM 700 confocal microscopy and a Zeiss LSM-TPMT camera were used to examining positively labeled testicular cells (Oberkochen, Germany). Our study used the GATA4, Vimentin, UTF1 and VASA antibody from Abcam to identify spermatogenic cells and Sertoli cells expression in immunohistochemistry .

A method for isolating and cultivating sertoli cell

As a Sertoli cell suspension of the cryopreserved testicular tubules, 750 U/mL collagenase Type IV (Sigma, St Louis, MO, USA), 0.25 g/mL dispase II (Roche, Ludwigsburg, Germany), and 5 g/mL DNase were used to enzymatically digest the dissociated tubules for 30 min at 37 °C. It was then supplemented with 10% ES cell-qualified FBS. Using a 100 m cell strainer, a 15-minute centrifugation at 1000 rpm was performed on the cell suspension. After removing the supernatant from the pellet, Ca + + and Mg + + were added to the HBSS solution. After washing the cells, they were plated onto five culture dishes coated with 0.2% gelatin (Sigma) (d = 10) in hGSC medium (human germ stem cell) containing StemPro hESC medium, 1% N2-supplement (Invitrogen, Waltham, MA, USA), 6 mg/mL D + glucose, 5 g/mL bovine serum albumin, and 1% L-glutamine (Sigma) with N2-supplement. A 96-hour incubation was performed at 37 °C in an incubator with 5% CO2. The culture medium was changed after 72 h from 5 milliliters to 4 milliliters of new medium and the cells grew for 4 days. The germ cells adhered to the monolayer of adherent somatic cells were gently washed with DMEM/F12 culture medium containing L-glutamine (PAA) on day 7 before harvesting. A 5 mL pipette was used for pipetting DMEM/F12.

Isolation of sertoli cells

In our previous research, we isolated and identified human Sertoli cells [22]. We washed testicular tissues three times and cut them into 0.2 cm pieces. We then treated them with Enzyme I (10 mL DMEM containing 2 mg/mL type IV collagenase and 10 mg/mL DNase I) at 34 °C for 15 min. Incubation was performed at 34 °C for 10–15 min with Enzyme II (4 mg/mL collagenase IV, 2.5 mg/mL hyaluronidase, 2 mg/mL trypsin, and 10 mg/mL DNase I). DMEM/F-12 supplemented with 10% FBS was grown at 34 °C in 5% CO2 at 34 °C with tissue blocks filtered through a 40-mm cell strainer. A surface panning strategy with laminin and lectin was used to select Sertoli cells from pooled cell groups. A dish coated with lectin was used to transfer cells that did not adhere to laminin. A microarray study was conducted with the attached cells. We used human Sertoli cells that were highly pure in Vimentin+, Sox9+, Gata4+, Vasa-, and UTF1+.

Extraction of RNA

A disperser tool T10 Basic Ultra-Turrax (IKA-Werke GmbH & Co. KG) was used to homogenize the testicular specimens in TRI Reagent (Sigma-Aldrich). According to the manufacturer’s instructions, TRI Reagent was used to extract total RNA. A Qiagen RNeasy Plus Mini Kit was used to post-purify the extracted RNA to remove DNA contamination. Agilent 2100 bioanalyzer (Agilent Technologies) was used to determine sample quality, while a NanoDrop ND-1000 spectrophotometer was used to determine RNA quantity. As long as the RNA Integrating Number (RIN) was greater than 7.0, the samples were qualified for microarray analysis.

Analyzing and preparing microarrays

R/Bioconductor was used to perform all statistical analyses relating to gene expression. Raw microarray data were used to determine gene expression levels. The robust multichip average (RMA) method was used to normalize the files. Using Ward hierarchical clustering, an unsupervised analysis was conducted. Using unpaired Student’s t-tests, we selected genes that were differentially changed between study groups. For each differentiating gene, we computed false discovery rates (FDR) to minimize the impact of multiple testing on the Student’s t-test P value.

The enrichment analysis of the GO and KEGG terms is as follows

For the purpose of determining the functions of common DEGs, DAVID (http://david.ncifcrf.gov/) was used. To further analyze the results of the pathway enrichment analysis, the GO pathway enrichment analysis and KEGG pathway enrichment analysis were downloaded to TXT format. A visualization of the results was performed using R version 3.6.2 software. A P value greater than 0.05 was considered statistically significant [45].

Network identification and hub gene identification for the protein–protein interaction

Online tools for analyzing protein–protein interaction (PPI) information can be found at http://string-db.org/. Using the STRING and a minimum interaction score of 0.7, we constructed a PPI network of common DEGs. In order to visualize the PPI network derived from the STRING database, we used Cytoscape software v3.7.1 (https://cytoscape.org/). Based on the degree calculation method, the top 20 genes in the PPI network were considered hub genes in Cytoscape by using the cytoHubba plugin. A WebGestalt (http://www.webgestalt.org/) analysis was conducted for the hub genes to determine their GO and KEGG pathways. A statistically significant difference was defined as a P value of 0.05 [22, 46,47,48].

Gene co-expression analysis

The hub genes were identified through a weighted gene co-expression analysis, using Coxpressdb and Genemania tools to uncover co-expression networks and relationships between these genes. A similarity matrix was first generated by applying Pearson’s correlation analysis to all gene pairings. This matrix was then used to construct a scale-free co-expression network, using an optimized soft threshold power (β) to enhance network robustness. Next, the matrix was converted into a topological overlap matrix (TOM), which quantifies the connectivity of each gene by measuring its overall adjacency to all other genes within the network. The TOM represents how genes are interconnected based on their shared expression patterns. Simultaneously, average linkage hierarchical clustering was performed using a TOM-based dissimilarity metric, and the resulting gene dendrogram was generated with a minimum cluster size of 50 genes. To deepen the analysis, the dissimilarity of module eigengenes was also calculated, allowing for a more detailed examination of the relationships between different gene modules.

Hub gene analysis in the NOA subgroup

To determine whether hub gene expression differed between the two subgroups, the Wilcox test was used. Statistical significance was defined as a P-value of 0.05.

Differential expression analysis and data processing

In non-obstructive azoospermia, 184 transcripts related to the cytoskeleton were evaluated using a microarray. A comparison was then made between these genes and those found in normal cells. |log2 fold change (FC)| > 2 were the default settings for adj. value 0.05 and fold change value (FC) > 2. Visualizing DEGs was achieved using the volcano map and heat map. With the help of R and Python software, volcanic maps and heat maps were created.

Sorting proteins by class and comparing groups

Comparing DEGs between research groups was conducted using the online program ArrayMining. In order to analyze gene ontologies, we loaded the gene list into PANTHER (http://www.pantherdb.org/; accessed on 10 March 2023).

Weighted Gene Co-expression Network Analysis (WGCNA)

The intricate WGCNA algorithm was used to uncover gene co-expression modules associated with hub gene co-expression. Gene co-expression networks were constructed using the WGCNA R package, using both NOA and control DEGs. In summary, genes with similar expression patterns were grouped into a co-expression module utilizing a weighted correlation adjacency matrix and cluster analysis. In order to satisfy the criteria of a scale-free network, a weighted adjacency matrix was generated and an appropriate soft threshold β was calculated. The weighted adjacency matrix was transformed into a topological overlap matrix (TOM), and the corresponding dissimilarity was computed as 1 minus the TOM. The dynamic tree-cutting method was used to identify modules and modules with dissimilarities below 0.3 was combined. The relationship between module eigengene values and immune cell abundance was evaluated using Pearson correlation. Modules that had a high correlation with the majority of gene co-expression were identified as crucial signaling pathway-related modules and chosen for further investigation.

Investigation of gene ontologies (GO)

The functional annotation chart and pathway analysis were performed using the G: profiler web server (https://biit.cs.ut.ee/gprofiler/). In order to consider a term statistically significant, the q-value must be greater than 0.05.

Interaction network construction for target transcript proteins

In order to identify the associations between candidate target genes, we use the STRING database for PPIs, and Cytoscape for visual analysis of the PPI network (accessed on 16 June 2023). A node’s size is indicated by CytoHubba using its degree value. By using plug-in MCC algorithms, we select the top 50 core genes for upregulation or downregulation.

Data collection in scRNA-seq

The Male Health Atlas (MHA) [49], Version One, is a comprehensive dataset including 258,428 unique scRNA profiles (GSE216907 and GSE235324). It covers two species (Homo sapiens and Mus musculus) and five organs/tissues (testis, epididymis, vas deferens, corpus cavernosum, prostate), encompassing eight cell types (Table 2). Key datasets include:

Table 2 Related GEO dataset information of microarray and single cell of NOA in spermatogenic cells and sertoli cell

• Human Testis Development Atlas: Testicular samples across various ages.

• Human Germ Cell Lineage Atlas: Germ cell progression from spermatogonial stem cells to spermatids.

• Human Testis NOA Atlas: Samples from NOA patients and individuals with normal spermatogenesis, classified by etiology.

Data processing utilized limma and Seurat R packages for bulk RNA-seq and scRNA-seq, respectively. Techniques included cell filtration, normalization, dimensionality reduction, and clustering. Batch effect elimination used canonical-correlation analysis (CCA). Quality control retained 61,622 cells from 68,066, analyzing them using UMAP and selecting a resolution for further study. Cell type annotation leveraged marker genes from prior research. The scATAC-seq data, processed using 10X Genomics Cell Ranger ATAC and Signac package, resulted in 750 cells meeting stringent quality standards. Annotation of scATAC-seq cell types was aligned with scRNA-seq cell types, employing the FindTransferAnchors function. Spatial transcriptome data, filtered based on genes and UMIs, were visualized using the Seurat package (refer to Table 3 and Supplementary 3).

Table 3 Infertile men with non-obstructive azoospermia have genes that down or up their expression in spermatogenic cells

Data integration and analysis

For integrating data from different donors and scRNA-seq technologies, we used an anchor-based strategy (GSE45885, GSE9210, GSE108886, GSE145467, GSE216907, and GSE235324). This unsupervised method identifies ‘anchors’ representing shared biological states for data integration, using 35 dimensions for weighting. The integrated dataset was scaled and underwent dimensional reduction via PCA, with the UMAP method applied for non-linear dimensional reduction and dataset exploration. UMAP utilized 35 dimensions based on principal component variance contribution. Cell clustering used the graph-based method in the Seurat package, with dimensionality and resolution set at 35 and 0.3. Identification of specific markers for testicular germ and somatic cells, derived from existing literature, facilitated cell type categorization within clusters.

Cell–cell interactions in scRNA-seq

The markers used for cell type identification were sourced from references. We adhered to the same pre-processing and classification techniques outlined in the mouse syngeneic investigations. We first extracted T cells from the whole dataset to classify T cell subgroups. The detected T cells were further categorized into separate subgroups according to particular marker expression. CD8 + cells were characterized by the presence of CD8A + and CD8B + markers and the lack of CD4 − markers. T-helper cells were identified by the lack of CD8 and CD8B markers, and the presence of CD4+, FOXP3−, and CD25 − markers. Tregs (regulatory T cells) were characterized by the lack of CD8A − and CD8 + markers while exhibiting CD4+, FOXP3+, and CD25 + markers. This categorizing procedure adhered to the previously outlined methodology.

Predicting target transcripts and constructing miRNA regulatory networks

In order to identify the target genes of differentially expressed genes and miRNAs, TargetScan (https://www.targetscan.org/vert_80/; Whitehead Institute for Biomedical Research, Cambridge; accessed on 25 March 2022), miRTarBase (https://mirtarbase.cuhk.edu.cn/ Institute of Bioinformatics and Systems Biology, Germany; accessed on 25 March 2022), and StarBase (https://starbase.sysu.edu.cn/; Hertford, accessed on 26 June 2022) databases were used. Microarray differential genes and target genes predicted by all three databases were used to identify candidate miRNA database (GSE224511). To build the miRNA–mRNA regulatory network, miRNA and mRNA were linked. Cytoscape software (version 3.8.1) was used to visualize the miRNA–mRNA regulatory network.

Analyses of statistics

For statistical analysis, we used R software version 4.2.2. We compared the two groups using the Wilcox test. A statistically significant difference was defined as a P value of 0.05.

Results

Spermatogenic cellsand sertoli cell and spermatogenic cellsIsolation and immunohistofluorescence staining

We first examined vimentin expression in the human seminiferous tubules by immunohistochemistry (IMH). For the distinction between germ cells and Sertoli cells, we used GATA4 and Vimentin-specific markers. According to IMH analysis, Sertoli cells display high GATA4 and Vimentin expression in the basal compartment, whereas differentiating germ cells display low expression. In order to differentiate undifferentiated spermatogonia from Sertoli cells, we used UTF1 and VASA-specific markers. In Sertoli cells, UTF1 and VASA were either low or non-expressed, whereas they were highly expressed in spermatogonia (Fig. 1).

Fig. 1

Specific genes expression in the non-obstructive azoospermia’s seminiferous tubules by immunohistochemistry (IMH). We used GATA4 and Vimentin-specific markers to the distinction between germ cells and Sertoli cells. According to IMH analysis, Sertoli cells expressed high GATA4 and Vimentin expression in the basal compartment, whereas differentiating germ cells express low expression. In order to differentiate spermatogonia from Sertoli cells, we used UTF1 and VASA-specific markers. In Sertoli cells, UTF1 and VASA were either low or non-expressed, whereas they were highly expressed in spermatogonia (Scale bar: 50 μm)

NOA versus normal gene expression profiles

A comparison was made between gene expression profiles in various cell groups and normal cases based on microarrays from various cell groups. Based on a comparative analysis of gene expression profiles between infertile and control subjects, Supplemental 1 highlights 42 differentially expressed genes (RefSeq annotated) in spermatogenic cellsand Sertoli cells (774 cytoskeleton genes). Figure 2 shows a Volcano plot and heat map (raw microarray data in Supplementary 1 for spermatogenic cellsand Supplementary 2 for Sertoli cell).

Fig. 2

Microarray spermatogenic cellsprofile datasets and heat map of DEGs. A normalization spermatogenic cellsand Sertoli cell data of microarray spermatogenic cellsprofile (B) Volcano plot of the DEGs in microarray spermatogenic cellsprofile identified 771 (12 upregulated and 6 downregulated) common DEGs and (C) Volcano plot of the DEGs in spermatogenic cellsprofile identified 774 (4 upregulated and 20 downregulated) common. We use the FDR of 0.05 and the |FC| of 1.5 as the cut-off criteria. The false discovery rate is the false discovery rate; the fold change is the fold change. D Heat map of DEGs of spermatogenic cells. E Heat map of DEGs of Sertoli cells. Red represents upregulation, and blue represents downregulation

Microarray analysis was performed on whole sequencing (about 774 transcripts). In three cases of non-obstructive azoospermia, 12 genes were upregulated while the expression of six genes was downregulated (Figs. 2 and 3). Three cases of non-obstructive azoospermia were studied using microarrays. They showed upregulations of transducin beta like 3 (TBL3), MAGE family member A8 (MAGEA8), keratin-associated protein 3 − 2 (KRTAP3−2), keratin 35 (KRT35), versican (VCAN), myosin XIX (MYO19), fibulin 2 (FBLN2), SH3 domain containing ring finger 1 (SH3RF1), actin-related proteins 3B (ACTR3B), stereocilin (STRC), thrombospondin 4 (THBS4) and catenin delta 2 (CTNND2). The gene expression levels of netrin 1 (NTN1), integrins alpha 1 and beta 1 (ITGA1 and GJB1), capping actin subunit of muscle Z-line alpha 1 (CAPZA1), septin 8 and golgin like 6 (GOLGA6L6) were lower versus normal spermatogenic cells (Fig. 3; Table 3).

Fig. 3

Compared with controls, non-obstructive azoospermia men’s gene expression was down- or up-regulated. A box plot of obstructive azoospermia’s spermatogenic cellsis down-/up-regulated and (B) box plot of obstructive azoospermia’s Sertoli cells are down-/up-regulated

Microarray analysis of three different NOA Sertoli cell cases revealed 4 upregulated genes and 20 downregulated genes (Figs. 2 and 3). Based on microarray analysis of three cases of non-obstructive azoospermia, it was found that the adhesion molecule 3 of the nephrin family (KIRREL3), tubulin tyrosine ligase 9 (TTLL9), ankyrin repeats and SOCS boxes containing 1 (ASB1) were upregulated, as well as the GRIP domains containing 5 (GDPR5). And desmin (DES), erythrocyte membrane protein band 4.1 like 2 (EPB41L2), potassium channel tetramerization domain containing 13 (KCTD13), kelch like family member 8 (KLHL8), TRIO and F-actin binding protein (TRIOBP), extracellular matrix protein 2 (ECM2), dishevelled segment polarity protein 3 (DVL3), armadillo repeat containing 10 (ARMC10), SH3 domain binding protein 4 (SH3BP4), kinesin family member 23 (KIF23), sorting nexin 4 (SNX4), kelch like family member 12 (KLHL12), protein kinase C and casein kinase substrate in neurons 2 (PACSIN2), anillin (ANLN), WD repeat domain 90 (WDR90), stathmin 1 (STMN1), sperm antigen with calponin homology and coiled-coil domains 1 like (CYTSA, SPECC1L), latent transforming growth factor beta binding protein 3 (LTBP3) and spectrin beta non-erythrocytic 1 (SPTBN1) were downregulated versus the normal’s Sertoli case (Fig. 3; Table 4).

Table 4 Genes down/up the expression in sertoli cell of nonobstructive azoospermia infertile men

Sorting proteins by class and comparing groups

Using PANTHER (Protein Analysis Through Evolutionary Relationships) transcript analysis, it was determined that differentially expressed RNAs covered spermatogenic cellscytoskeleton genes located throughout the cell, including organelles, membranes, and extracellular matrix (Fig. 4). The PANTHER server showed that spermatogenic cellscytoskeleton genes were involved in gene-specific transcriptional regulator (PC00264) 2.9%, cytoskeletal protein (PC00085) 37.1%, scaffold/adaptor protein (PC00226) 17.1%, cell junction protein (PC00070) 5.7%, cell adhesion molecule (PC00069) 17.1%, structural protein (PC00211) 5.7%, extracellular matrix protein (PC00102) and 14.3% (Fig. 4a).

Fig. 4

Identify and sort a hub gene module that exhibits a correlation with NOA in Sertoli cells and spermatogenic cells. A Identify and correlation between PPI, WGCNA and GO enrichment, B sorting and correlation between PPI, WGCNA and GO enrichment, C Branches of the cluster dendrogram of the most connected genes gave rise to 3 gene co-expression modules, D Heat map of the correlation between module eigengenes and phenotype, E Intergenic connectivity of spermatogenic cells’s genes in turquoise module and (F) Intergenic connectivity of Sertoli cells’s genes in turquoise module

The PANTHER server showed that Sertoli cell’s cytoskeleton genes were involved in cytoskeletal protein (PC00085) 40.8%, scaffold/adaptor protein (PC00226) 38.8%, cell junction protein (PC00070) 4.1%, cell adhesion molecule (PC00069) 4.1% and extracellular matrix protein (PC00102) 8.2% (Fig. 4b).

Gene Co-expression modules for signaling pathway

The DEGs between NOA and control groups were used in the WGCNA. The soft-thresholding method was used to identify DEG co-expressed gene modules (Fig. 4c and d). In the end, the DEGs were grouped into modules based on their colors: gray, blue, and light blue. Out of these gene modules, the blue module was shown to be prominently involved in the FAK (focal adhesion kinase) plays a key role in integrin signaling. The light blue module had a strong correlation with the GTP-binding proteins regulation, whereas the gray module was shown to be engaged in the Rho kinase phosphorylates signaling pathway. Collectively, these two modules had a significant impact on the activity of the Rac and Cdc42 is the protein-serine/threonine kinase. They were identified as crucial gene modules that were dysregulated at the transcriptional level, as shown in Fig. 4e and f.

Analysis of GO enrichment

Biological process (BP) and molecular function (MF) enriched the DEGs of spermatogenic cells. DAVID online analysis tool was used to analyze GO enrichment of the common DEGs. As a result, common DEGs were significantly associated with mitotic cytokinesis, cytoskeleton-dependent cytokinesis, stem cell proliferation, Rho protein signal transduction, supramolecular fiber organization, negative regulation of Ras protein signal transduction, positive regulation of protein localization to the cell periphery, and positive regulation of adhesion between cells and substrates. The results showed that common DEGs played a significant role in GTPase, small GTPase, tubulin, Gap Junction, glutathione transmembrane transport, and gap junction hemichannel activity (Fig. 5).

Fig. 5

DEGs are regulated in spermatogenic cells according to their biological process, molecular function, and PPI. A Genes that are up- or down-regulated are activated through a biological process, B Gene expression is regulated by multiple biological processes and C According to the cytoHubba degree score, both upregulated and downregulated DEGs showed meaningful co-expression

Common DEGs of Sertoli cells were enriched in MF and BP. Common DEGs were significantly correlated with the assembly of cell-cell junctions, the organization of cell-cell junctions, and adhesion between cell-matrix junctions for BP. A significant association was found between common DEGs and collagen binding, cell-matrix adhesion mediator activity, U3 snoRNA binding, MAP-kinase scaffold activity, actin binding, and cadherin binding in MF (Fig. 6).

Fig. 6

In Sertoli cells, DEGs are up-or down-regulated in biological processes, molecular functions, and PPIs. A Activation of up- or down-regulated genes through a biological process, B Biological processes that regulate gene expression, C A meaningful co-expression was found between the DEGs that were up-regulated and down-regulated (hub genes in cytoHubba Degree score)

Hub gene analysis and PPI network analysis

Based on the STRING database, a PPI network of DEGs was constructed with 24 nodes and 11 edges, as shown in Fig. 5A. The top hub genes using the Degree algorithm of the Cytohubba plug-in are shown in Fig. 4B; Table 5. We analyzed 20 hub genes using WebGestalt for GO and KEGG enrichment. GO enrichment analysis in Fig. 5 mainly focuses on metabolic processes, cellular component organization, biological regulation, cytosol, cytoskeleton, membrane, and protein binding. It is possible for these genes to interact with other proteins or co-express with them. With the Centiscape plugin, NOA Sertoli cells, aurora kinase A (AURKA), NDC80 kinetochore complex component (NDC80), tight junction protein 1 (TJP1), and cullin 3 (CUL3) were highly and significantly co-expressed with up- or down-regulated genes (GJA1, DES, KIF23, KLHL12, and ANLN).

Table 5 KEGG pathway enrichment analysis of the common DEGs

As shown by PPI of NOA spermatogenic cells, Ski2-like RNA helicase 2 (SKIV2L2), BOP1 ribosomal biogenesis factor (BOP1), the capping actin protein beta of muscle Z-line (CAPZB), and ribosomal RNA processing 9, U3 small nucleolar RNA binding (RRP9) are highly and significantly interacted and co-expressed with CAPZA1, TBL3, FBLN2, ACTR3B and CTNND2 (Figs. 5 and 6).

Gene Co-expression analysis

The height of the sample clustering had been established at 20, after excluding the 5 outlier samples from further analysis. Using the criteria of MM > 0.9 and GS > 0.3, a total of 41 hub genes were identified from the modules CAPZA1, TBL3, FBLN2, ACTR3B, and CTNND2, together with their co-expression partners GJA1, DES, KIF23, KLHL12, and ANLN (Fig. 7).

Fig. 7

Gene co-expression analysis and PPI. A PPI analysis of up-or down-regulated in biological processes, and molecular functions, B Gene co-expression analysis of CAPZA1, TBL3, FBLN2, ACTR3B, and CTNND2

Pathway enrichment analysis using KEGG and reactome

The DAVID database analyzed the KEGG pathway of the common upregulated and downregulated DEGs. In the NOA Spermatogenic cells, DEGs are mainly involved in trafficking and regulation of gap junctions, extracellular matrix organization, and keratinization of membranes. There were mainly DEGs in Sertoli cells that interacted with Nephrin family members, with RHO GTPases, and with Cell-Cell communication as well as Signal Transduction. As can be seen in Fig. 3; Table 4, these results are presented in graphs and Table 4, and Supplementary 3 for spermatogenic cellsand Supplementary 4 for Sertoli cells, respectively.

Identification of a hub gene from scRNA-seq

In our study, we examined gene expression data from the GSE45885 dataset, which comprised 29 samples. This included 3 from healthy individuals and 26 from patients with NOA. To delve into NOA’s heterogeneity, we conducted unsupervised clustering on 28 samples—21 from healthy individuals and 7 from NOA patients (detailed in Supplementary 5). This approach delineated four distinct clusters representing cell types crucial to spermatogenesis: spermatogonia (SPG), spermatids/sperm (SPT), spermatocytes (SPS), Sertoli cells (SC), Leydig cells (LC), peritubular myoid cells (PTM), endothelial cells (EC), vascular smooth muscle cells (SMC), macrophages (MAC), mast cells (MC), and T-cells (T). This clustering not only differentiated between healthy and NOA samples but also revealed intrinsic heterogeneity among the NOA samples. Subsequent principal component analysis (PCA), illustrated in Figs. 8 and 9, corroborated these findings. Our single-cell transcriptomic analysis highlighted nine hub genes: EPB41L2, KCTD13, ARMC10, SH3BP4, SNX4, and STMN1 were significantly associated with NOA Sertoli cells and exhibited downregulation; TBL3, MYO19, ACTR3B, THBS4, and CTNND2 showed upregulation; GOLGA6L6 and CAPZA1 were downregulated in NOA spermatogenic cells.

Fig. 8

Identification of a Sertoli cell’s NOA-specific hub gene in single cell. A The 29 samples were split into eleven clusters. The analysis of 21 normal and 7 non-obstructive azoospermia (NOA) samples revealed four distinct clusters, highlighting the diversity of cell types involved in spermatogenesis. These clusters included spermatogonia (SPG), spermatids/sperm (SPT), spermatocytes (SPS), Sertoli cells (SC), Leydig cells (LC), peritubular myoid cells (PTM), endothelial cells (EC), vascular smooth muscle cells (SMC), macrophages (MAC), mast cells (MC), and T-cells (T). Single-cell transcriptomic analysis identified nine hub genes – (B) EPB41L2, C KCTD13, D ARMC10, E SH3BP4, F SNX4 and G STMN1 - as significantly associated with NOA Sertoli cell and were found to be downregulated

Fig. 9

A TBL3, B MYO19, C ACTR3B, D THBS4, and E CTNND2 were found to be upregulated, and F GOLGA6L6 and G CAPZA1 were found to be downregulated in Spspermatogenic cellserm of NOA

Cell-cell communication network

We observed that several interactions had a significant connection with a certain phenotype. The interactions included the same receptor but differed in the ligands used. This finding initiated an investigation into whether a specific ligand or receptor is accountable for the relationship, rather than the physical interaction itself. To examine this idea, we calculated Spearman correlations between the expression of the receptor alone or the expression of the ligand independently and the rate of cancer. Our data reveal that interaction scores exhibiting a significant connection with the phenotype are often linked to either a robust correlation with receptors or a pronounced correlation with ligands. Since interaction scores are influenced by the expression levels of both receptors and ligands, it is logical to deduce this consequence, since these two elements are interrelated and not mutually exclusive. Nevertheless, there were particular instances where the interaction score demonstrated a notable correlation with the tumor growth rate, despite a lack of strong correlation between the receptor and the ligand, especially in the region of the plot where their expression varied from − 0.5 to 0.5 (Fig. 10).

Fig. 10

Cell-cell communication analysis. A network cell-cell communication related to cytoskeleton protein, B Heatmap of cell-cell communication related to cytoskeleton protein

The following candidate microRNAs have been isolated and selected

A total of four databases were analyzed in order to predict miRNA targets for differentially expressed mRNAs in infertile spermatogenic cells. We isolated and selected the most relevant microRNAs (Fig. 9A) after identifying 18 genes, including TBL3, MAGEA8, KRTAP3-2, KRT35, VCAN, MYO19, FBLN2, SH3RF1, ACTR3B, STRC, THBS4, CTNND2, NTN1, ITGA1, GJB1, CAPZA1, SEPTIN8 and GOLGA6L6. A number of microRNAs were observed with more clarity than other microRNAs, including hsa-miR-4763-3p, hsa-miR-4787-3p, hsa-miR-3162-3p, hsa-miR-7107-5p, hsa-miR-210-5p, and hsa-miR-6756-5p. A number of genes, such as GJB1, NTN1, ITGA1, TBL3, MYO19, and SEPTIN8, are up-regulated or down-regulated by these microRNAs (Fig. 11a and c).

Fig. 11

The picture shows that the most significant microRNAs associated with the up/ down regulation genes were selected on the Manhattan diagram. A most significant microRNAs associated with the up/down-regulated genes in spermatogenic cells of NOA. B most significant microRNAs associated with the up/down-regulated genes in the Sertoli cell of NOA. C Genes and miRNA interaction in spermatogenic cellsof NOA. D Genes and miRNA interaction in Sertoli cell of NOA

Following the identification of hub 24 genes, KIRREL3, TTLL9, GJA1, ASB1, RGPD5, DES, EPB41L2, KCTD13, KLHL8, TRIOBP, ECM2, DVL3, ARMC10, SH3BP4, KIF23, SNX4, KLHL12, PACSIN2, ANLN, WDR90, STMN1, CYTSA, LTBP3 and SPTBN1, we isolated and selected the most relevant microRNAs (Fig. 9A). In terms of microRNA composition, the hsa-miR-6089, hsa-miR-1268a, hsa-miR-4763-3p, hsa-miR-339-5p, hsa-miR-3663-5p, and hsa-miR-619-5p were observed as more distinct than other microRNA types. There is evidence that these microRNAs are involved in upregulation and downregulation of PACSIN2, ASB1, STMN1, TTLL9, KIRREL3, and DVL3 (Fig. 11b and d) (Supplementary 6 for spermatogenic cellsanalysis and Supplementary 7 for Sertoli cell analysis.

Discussion

Today, we know that the testis has tight, anchoring, and gap junctions. In addition to the ectoplasmic specialization, the tubulobulbar complex is another anchoring junction type found in the testis [28, 50]. Researchers are currently focused on identifying the regulatory molecules that open and close junctions, as this information will be useful in elucidating the mechanism of germ cell movement [51, 52]. Sertoli cell tight junction disassembly is induced by cytokines, which inhibit tight junction protein production. GTPases, kinases, phosphatases, and protease inhibitors are also involved [26, 53,54,55].

There are six different types of actin filaments in mammalian tissues [56]. Actin filaments are made up of actin monomers of 42 kDa, each of which polymerizes into a linear chain with a helical twist of approximately 8 nm [57]. Besides contributing to cell structure and contractility in Sertoli cells, actin appears to also play a role in motility-associated processes at the cell periphery [58]. All eukaryotic cells, including Sertoli cells, coexist in equilibrium with two pools of actin, monomeric G-actin and polymeric F-actin [59]. When one pool predominates, actin filaments are polymerized or depolymerized, affecting the overall state of the cytoskeleton. Consequently, cell shape and motion are affected [59]. Additionally, actin filaments are important components of ectoplasmic specializations and tubulobulbar complexes, as well as modified anchoring junctions between Sertoli cells and germ cells, indicating that both of these junctions facilitate the movement of germ cells during development. T-actin 1 and T-actin 2 are unique actins found in mouse haploid germ cells, and they share 40% homology with the other actins. During spermiogenesis, T-actins 1 and 2 are restricted to the cytoplasm of spermatids, and T-actins 1 and 2 are restricted to the heads and tails of sperms (84), which suggests that both actins play an important role in germ cell morphogenesis [48, 60].

The intricate cellular composition revealed by our analysis underscores the complex nature of spermatogenesis. The identification of distinct cell clusters, including spermatogonia, spermatids, spermatocytes, and various supporting cells, reflects the multifaceted process of sperm development and maturation. Particularly noteworthy is the discovery of nine hub genes significantly associated with NOA in Sertoli cells. The downregulation of genes such as EPB41L2, KCTD13, and ARMC10 in NOA samples may offer new insights into the pathophysiology of this condition, which remains poorly understood. Conversely, the upregulation of genes like TBL3 and MYO19 in sperm from NOA patients could indicate compensatory mechanisms or altered cellular pathways. These findings contribute to the growing body of research on male infertility, particularly in understanding the genetic underpinnings of NOA. Future studies should aim to elucidate the functional roles of these genes in spermatogenesis and explore their potential as therapeutic targets [39, 44, 61,62,63]. This research also highlights the importance of single-cell transcriptomic analysis in uncovering cellular heterogeneity and gene expression patterns in complex tissues, providing a valuable framework for studying other intricate biological processes. Most studies around the KCTD family, including KCTD13, primarily focus on their roles in neurodevelopmental disorders, cancer, and other physiological processes. For instance, KCTD13 has been implicated in conditions like autism and intellectual disability, particularly in relation to the 16p11.2 deletion syndrome. The gene’s role in regulating membrane channel activities and modulating distinct GTPases has been explored in this context [64].

Testicular gap junctions were found to be severely altered in human pathological testes with impaired spermatogenesis in previous morphological studies [65]. When the feminized testis was examined by freeze-fracture, no gap junctions were observed between Sertoli cells, but these membranous structures were frequently observed between Leydig cells [66, 67]. It has been reported that infertile azoospermic and oligospermic patients have atypical gap junctions in their seminiferous tubules. It has been reported in two other studies that men with seminiferous tubules exhibiting SCO phenotype have reduced gaps junction-like specializations in their cell membranes [68]. It was found that specific connexins (Cx) 43 protein was not present within seminiferous tubules of patients with SCO syndrome. In these patients, the disappearance of Cx43 protein was accompanied by a decrease in testicular Cx43 mRNA levels, while the protein and transcript levels of Cx43 remained. Based on these findings, Cx43 alteration could be used as a biological marker of Sertoli cell dysfunction due to a defect in Sertoli cell maturation [69].

The effects of carcinogens and oncogenes are associated with impaired gap junctional intercellular communication (GJIC) and Cx dysfunction. The presence of this feature has been reported in numerous neoplastic tissues [70]. Cx43 was observed to be absent from human testes infiltrated with carcinoma-in-situ or seminoma, whereas its presence was observed in normal control testes by immunohistochemistry. Pure testicular seminomas showed normal Cx43 mRNA and protein levels [71]. The fine immunolocalization analysis of seminoma cells, however, showed that Cx43 was, in fact, localized at the plasma membrane, rather than the cytoplasm. Testicular levels of Cx43 are reduced in transgenic mice that develop Leydig cell tumors, and the protein is localized abnormally. We propose that Cx43 delocalization is an early sign of uncontrolled cell proliferation in pathological testes based on our analysis of behavior of Cx43 during the early and advanced stages of Leydig cell tumorigenesis. Testicular tumoral tissues up-regulated two Cxs (Cx26 and Cx40) while Cx43 was dramatically affected. In normal human testis, Cx26 was undetectable, but in infiltrated tubules with spermatogonial arrest or only CIS, it displayed strong intercytoplasmic Sertoli cell staining. Cx26 expression was shown to be inversely correlated with the grade of malignancy in human breast in previous research. Testicular seminoma had higher levels of Cx40 mRNA than normal human testes, according to a recent study [72]. Despite the fact that the reasons for altered expression of Cx26 and Cx40 in tumoral tests are unclear, it has been hypothesized that these two CX could be considered as potential new diagnostic markers for testicular germ cell tumors [73].

Although the critical steps regulated by these growth factors are not well understood, lab data indicate that extracellular matrix (ECM) is key to cell proliferation, meiosis and differentiation of germ cells during spermatogenesis [74]. There are two parts of the ECM that are specialized: the lamina propria and basement membrane. Embedded in a framework of mostly collagen fibers, the lamina propria constitutes the classical basement membrane that separates and mediates cell-to-cell interactions between myoid and Sertoli cells [75]. The ECM regulates the differentiation of Sertoli and myoid cells as well as the development of germ cells [76]. It is well known that laminin, fibronectin, and Type-IV collagen are components of the basement membrane with intrinsic biological and structural activities. As well as this, Tripp & Lamb have demonstrated that the ECM contains growth factors like TGF-beta, IGF-1, EGF, TGF-alpha, and bFGF. Signaling from Sertoli cell receptors to ECM can also be mediated by complex cellular mechanisms [77]. Sertoli cells, along with their basement membranes and peritubular myoid cells, can be viewed as one functional unit based on these characteristics [78]. During the seminiferous epithelial cycle, Sertoli and germ cells, in particular spermatogonia, are dependent on the basement membrane for both structural and hormonal support [79]. Because Sertoli cells are physically in contact with the basement membrane, it is not surprising that ECM regulates spermatogenesis, particularly spermatogonia and Sertoli cells [19]. A tunica propria is also formed by the collagen network underneath the basement membrane, as well as the layers of myoid cells. Through the interaction of the seminiferous epithelium with Leydig cells in the interstitial space, the seminiferous tubule establishes spermatozoa through the seminiferous tubule and tunica propria [80]. There may even be cross-talk between the seminiferous epithelium, myoid cells, and interstitial cells within the tunica propria due to the basement membrane and collagen network [81]. In recent studies, the ECM has been demonstrated to play a critical role in supporting the function of both Sertoli and germ cells in the seminiferous epithelium, as well as the dynamics of the BTB [82]. Using the adult rat testis as a model, we summarize some of the latest findings in the field regarding ECM function in spermatogenesis.

In addition to membrane compartmentalization and vesicle trafficking, septins bind GTP and participate in mitosis and cytoskeletal remodeling [83]. The testis expresses two of the 14 members of the septin family in mammals [84]. Mutations in Septin4 or Septin12 in mice result in male sterility (defective annulus or bent neck of sperm). A higher proportion of infertile men have sperm with abnormal expression patterns for SEPT4, 7 and 12 [85]. Molecular mechanisms of septin filament assembly and disassembly, and the SEPT-related complex’s role in spermatogenesis require further study. During spermiogenesis, SEPT12 is expressed in the nucleus, the neck, mitochondria, and the annulus of sperm. SETP12 and SEPT7 signals tend to be lost in sperm with abnormal heads, necks, and tails. A subset of patients with asthenozoospermia also displayed disorganized annulus/SEPTIN rings. Both motility and morphological defects of sperm may be caused by defect in SEPT synthesis, degradation, or dysfunction (Fig. 12).

Fig. 12

Schematic drawing illustrating the junction Sertoli-Sertoli and Sertoli-spermatogenic cells of germ cell movement in the seminiferous epithelium. Shown is the cytoskeleton defect gene involved in spermatogenic cells and Sertoli cells

Conclusion

In this study, the microarray analysis of three human NOA cases revealed distinct expression patterns in both spermatogenic cellsand Sertoli cells, highlighting potential molecular mechanisms underlying NOA. Several genes were found to be upregulated in spermatogenic cells, including TBL3, MAGEA8, and VCAN, while others, such as NTN1, ITGA1, and GJB1, were downregulated. Similarly, in Sertoli cells, genes like KIRREL3 and GJA1 showed increased expression, whereas DES, TRIOBP, and KIF23 were downregulated. These differential expression profiles indicate disruptions in key biological processes and molecular pathways in NOA patients. These findings offer new insights into the molecular interactions and regulatory networks involved in NOA. The identification of key pathways, such as those related to cell adhesion, cytokinesis, and cytoskeleton dynamics, may serve as potential therapeutic targets for NOA treatment. Future studies are warranted to validate these candidate genes and pathways, providing a clearer understanding of the pathophysiology of NOA and paving the way for novel clinical interventions.