Downregulation of KIF2C and TEKT2 is associated with male infertility and testicular carcinoma

Background: Genetic factors are important in spermatogenesis and fertility maintenance, and are potentially significant biomarkers for the early detection of infertility. However, further understanding of these biological processes is required. Methods: In the present study, we sought to identify associated genes by reanalyzing separate studies from Gene Expression Omnibus datasets (GSE45885, GSE45887 and GSE9210) and validation datasets (GSE4797, 145467, 108886, 6872). The differential genes were used the limma package in R language. Gene ontology and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analyses were performed by the clusterprofier package. The protein-protein interaction network was constructed by the STRING database. The interaction between mRNA and TF was predicted by miRWalk web. At last, The Cancer Genome Atlas data were used to identify hub gene expression levels in GEPIA web. Results: The results showed that 27 shared genes associated with spermatogenesis. We effectively screen out two genes (KIF2C and TEKT2) and both validated by GSE4797, 145467, 108886 and 6872. Among 27 shared genes, KIF2C and TEKT2 both down-regulated in spermatogenesis. The network of TF-miRNA-target gene was established, we found KIF2C-miRNAs (has-miR-3154, 6075, 6760-5p, 1251-5p, 186-sp)-TFs (EP300, SP1) might work in spermatogenesis. Conclusions: Our study might help to improve our understanding of the mechanisms in spermatogenesis and provide diagnostic biomarkers and therapeutics targets.

AGING main causes of male infertility: spermatogenic function defects; ductal obstruction or disorder; hypothalamicpituitary axis disorder. A patient with azoospermia is at the highest risk of carrying genetic abnormalities.
With the biotechnology's improving, we can immediately identify expression diversifications at the transcription level, which beneficially contribute to infertile men. A lot of studies concentrated on differently expressed gene (DEG) analysis have found a quantity of possible molecular goals and diagnostic biomarkers for infertility at the transcription level. Agnieszka et al. discovers that genes, for example, GGN, GSG1, ADCY10, and GTSF1L are down regulated in human beings with azoospermia. Ribosomal protein S3 (RPS3) is recognized by Zhang et al., RPS5, RPS16, RPS23 and RPS6 were downregulated in teratozoospermia [12,13].
As the study of male infertility is still insufficient, our goal is to analyze whether there are rare potential disease-related genes associated with infertility, and provide clinical evidence [14]. In this study, a comprehensive bioinformatics method was used to study the related genes of spermatogenesis.

Fetching testicular tissue microarray data sets from GEO
The data sets were accessed from GEO (http://www.ncbi.nlm.nih.gov/geo/) in the National Center for Biotechnology Information Database (NCBI) utilizing the accession numbers GSE45885, GSE45887, and GSE9210. Data sets GSE45885 was associated by azoospermia and GSE45887 were presented by Agnieszka Malcher and based on the GPL6244 platform ([HuGene-1_0-st] Affymetrix Human Gene 1.0 ST Array [transcript (gene) version, Affymetrix, Inc., USA]). The information of testicular samples in GSE45887 and GSE45885 was accessed from the issued literature (Table 1). The research procedure appeared in Figure 1.

Microarray data preprocessing
In this study, the original CEL profiles were imported into R (version 3.5, https://www.r-project.org/), and the profiles background was corrected and normalized using the Affy R-package (Bioconductor version 3.6). Affybatch's mas5calls method returns expression sets corresponding to specific genes through multiple probes.

Differentially expressed gene selection
DEGs were used the limma package (version 3.6). |log2 fold change|>1 and P < 0.05 were set at the cutoff values.

Functional annotation and pathway analysis of DEGs
Go analysis is composed of molecular function (MF) and cellular compartment (CC) and biological process (BP) are usual processes for large-scale genomic data's functional annotation. To learn more the mechanism of DEGs that be involved in infertility, we used clusterprofiler R-package to analyze the enrichment of KEGG pathways and GO (version 3.16). In these analyses, P < 0.05 was considered statistically significant.

Protein interaction network and module analyses
The STRING database (http://string-db.org), A proteinprotein interaction network that was composed of upregulated and down-regulated DEGs was built, with a cutoff score more than 0.4. Using the clusterone add-in of Cytoscape v3.6.1 to pick the significant modules from the PPI network (https://cytoscape.org/) with P < 0.01 showed statistical importance. The K-core analysis and degree/betweenness/closeness centrality were executed by two add-ins CentiScaPe and Molecular Complex Detection (MCODE) in Cytoscape to illuminate the modules and most significant nodes in the network.

TF-miRNA-target gene network construction
Interactions between differentially expressed miRNAs and differentially expressed mRNAs and expressed miRNAs were forecast utilizing miRWalk 3.0 (http://mirwalk.umm.uni-heidelberg.de/), and a mark 0.95 was regarded to be the cutoff principle for the estimate analysis in miRWalk. The target mRNAs that was involved in all of these databases were only picked for the further analysis. The interaction between mRNA and TF was predicted by using miRWalk 3.0 and the mark 0.4 was considered as cutoff standard for the estimate analysis in the experimental module of LncBase. After the predicted marks were intersected with DEGs, miRNAs, TFs and mRNAs were picked for further analysis. Cytoscape software (version 3.6.1) was used to visualize the regulatory network.

Validation and expression of hub-gene in male infertility and testicular carcinoma
At last, the expression levels of hub genes showed in GEPIA (based on TCGA data) (http://gepia.cancerpku.cn). Next, validation of hub-gene in male infertility by GEO database (GSE4797, 145467, 108886, 6872) ( Table 1). P < 0.05 was viewed to show a statistically important difference in these analyses.

Available of data and materials
The datasets analyzed for this study can be found in the GEO datasets (https://www.ncbi.nlm.nih.gov/gds) and TCGA.

Analysis of DEGs
The expression description information was preprocessed and then analyzed with the Affy package in R language. The entire gene expression was examined.
The RNA expression levels are revealed in Figure 2. Differences were revealed by the hierarchical cluster analysis in distribution between normal samples and azoospermic. The consequences showed that grouping was reasonable, and further analysis was undergone by the data with success. Microarray data from the normal samples were compared with those from the azoospermic samples and a sum of 1396 DEGs were discovered.

Protein interaction network and module analysis
To investigate shared genes' relationship between modules, PPI networks were built by using Cytoscape software that was based on the STRING database.
Moreover, the k-core analysis was executed to discover the hub genes and cardinal clusters of PPI networks. By Cytoscape-MCODE analysis, parameters are set as follows: Degree Cutoff: 5, Node Score Cutoff: 0.2, K-Core: 5. 10 clusters of 338 spermatogenesis-associated genes ( Figure 4) were identified by us. The hub genes in cluster 1 (Figure 4)

Expression of hub-gene in testicular carcinoma
A total of 12 genes (ANLN, CCNB1, CENPF, COIL, CYCS, KIF2C, KNSTRN, LELP1, OAZ3, SRPK2, TEKT2, WDTC1) were identified as hub genes. We used TCGA data of testis cancer to validate the hub gene expression with the online tool of GEPIA. All of the hub genes are expressed differently in normal and cancer tissues of testis cancer by the criterion of |logFC|>1 and p < .01 ( Figure 6).

Validation of hub-gene in male infertility
Through validation of hub-gene (A total of 12 genes were identified as hub genes in testicular carcinoma) in male infertility by GSE4797 set, we found expression of KIF2C between more than or equal to John score 5 and less than 5 existed significant difference. The expression of TEKT2 decreased with reduction of John score in the Figure 7. The Figure 8 showed the expression of KIF2C and TEKT2 both decreased in three datasets from GEO database. Furthermore, we found KIF2C might act in infertility and testis cancer through DEG, functional annotation and pathway, protein interaction network and module analysis.

DISCUSSION
With the development of science and technology, the molecular mechanism of azoospermia has aroused great interest. The study of azoospermia depends on human studies, animal models, organ culture models and cell lines [9,15]. It has been confirmed that the number of genes and proteins related to male infertility has increased [9,16]. However, how these biological processes are regulated at the molecular level remains to be elucidated. Therefore, it is necessary to further study the pathogenesis of azoospermia at the molecular level.

AGING
In this study, we identified the genes associated with spermatogenesis and systematically constructed a comprehensive framework of genes and miRNAs.
KIF2C (also known as the mitotic centromereassociated kinesin, MACK) is a member of the kinesin-13 family of microtubule (MT)-depolymerizing kinesins, which is critical in the regulation of microtubule dynamics. During cell division, KIF2C inhibits the wrong connection between MT and chromosome [19,20]. The function of KIF2C in interphase cells is not obvious, although its main localization in nucleus suggests that KIF2C may work in nuclear processes. KIF2C promotes the formation of AGING DNA damage foci, which may involve the migration and aggregation of DSBs (DNA Double Strand Break) [21][22][23][24][25]. We found KIF2C might work in the testis cancer and spermatogenesis.
Tektins (TEKTs), the proteins of the microtubules in Cilia, Flagella, Basal bodies and centrioles [26][27][28], have been found in various animals, including Filariae, including silk-worms [29], mice [26,30] and humans [31,32]. They were originally isolated from sea urchins and are a group of proteins: TEKT-A, -B and -C [33,34]. On the other hand, five types of TEKTs have been identified in mammals. TEKT2, which is similar to Tektin-t, locates in the main part of human spermatozoa but no immune signal was detected in the middle or at the end of the human sperm. Tektin2, a membrane protein, is responsible for sperm flagellum movement. Previous studies show that CatSper and tektin are associated with male infertility because they play an important role in sperm motility [35]. Tektin2 is essential for the integrity of motilin arm in sperm flagellum. Lack of tektin2 can lead to impaired sperm motility and male infertility [36]. The low expression of tektin2 mRNA was observed in frozen spermatozoa, suggesting that the decrease of sperm motility after cryopreservation may be due to the transcriptional damage of some sperm motility related genes [37].
Our study also has some limitations. First, more samples could be included in this study and we assessed our results based on published observations. Further in vitro and/or in vivo experiments would need to be carried out to test reliability of our results. This might reduce the error caused by individual differences of patients.

CONCLUSIONS
We applied DEG analysis to identify genes associated with azoospermia in this study. Then, through a system biology framework for a comprehensive and systematic biological function-and network-based analysis of azoospermia, we found 27 hub genes and test on the expression of Hub-gene in testicular carcinoma (found 12 hub genes were different in testicular carcinoma). Furthermore, we made the validation of hub-gene (A total of 12 genes were identified as hub genes in testicular carcinoma) in male infertility by GSE4797, 145467, 108886 and 6872 and found TEKT2 and KIF2C might work in infertility. The network of TF-miRNA-target gene was established and we found KIF2C-miRNAs (has-miR-3154, 6075, 6760-5p, 1251-5p,186-sp)-TFs (EP300, SP1) might work in spermatozoa of infertile men. Our study might help to improve our understanding of the mechanisms in azoospermia and provide diagnostic biomarkers and therapeutics targets.

AUTHOR CONTRIBUTIONS
HMC and ZW conceived the study, performed bioinformatic analysis, analyzed and interpreted the data and wrote the manuscript. FW and ZYL contributed to resources, revised the manuscript. XFL and JQH conceived the study, supervised the study, analyzed and interpreted the data, gave the final approval of the manuscript.