The Critical Biomarkers Identification of Insulin Signaling Involved in Initiating cAMP Signaling Mediated Salivary Secretion in Sjogren Syndrome: Transcriptome Sequencing in NOD Mice Model

Decreased Salivary Secretion and Increased Water Consumption Rates Were Observed in NOD Mice Model

To evaluate phenotypic characteristics of SS, we compared between NOD mice and ICR mice. The salivary secretion of the mice was monitored at 0, 3, 6, 9, 12, 15, and 18 days, respectively, as shown in Fig. 1A. Compared with the ICR mice, the salivary secretion of the NOD mice was significantly reduced, and the salivary secretion index was 0.19-2.83 μg/g. In the statistics of tear secretion in mice, the results also suggested that compared with the ICR mice, the tear secretion of the NOD mice was significantly reduced, and the tear secretion index was 0.04-0.25 μg/g (Supplementary Fig. 1). To further illustrate the results, we dynamically monitored the water consumption of the NOD and ICR mice over 18 days. The results showed no significant change in water consumption in each stage of ICR, and the water consumption index was 0.32-0.68 ml/g. While the water consumption in the NOD mice increased to 3 times that of the ICR mice, the water consumption increased slowly. It remained at a high level during the observation period (Fig. 1B). These results suggest that NOD mice have decreased salivary secretion and increased water consumption, suggesting symptoms of xerostomia.

Fig. 1figure 1

Results of NOD mice model measuring salivary secretion-related indicators. A Monitoring of salivary secretion index between NOD and ICR mice. B Measurement of water consumption between NOD and ICR mice. C Histological images (H&E stained) of mice submandibular gland tissues. D Immunohistochemistry images of AQP5 of mice submandibular gland tissues

Acinar Destruction and Decreased AQP5 Expression in the Submandibular Gland of NOD Mice

Since salivary secretion is impaired in NOD mice model, we next investigated the expression of a critical protein AQP5 involved in submandibular gland salivary secretion. The results of H&E staining indicated that, compared with the ICR mice, the histology of the submandibular gland tissue in the NOD mice showed acinar destruction and basement membrane changes (Fig. 1C). Studies have shown that acinar cell apical and basolateral membranes show positive AQP5 labeling in mice [31, 32]. In addition, the expression of AQP5 was also detected in intercalated ducts [33]. As shown in Fig. 1D, AQP5 immunostaining in the ICR mice was distributed on the acinar cell membrane. In the NOD mice, AQP5 immunostaining of the cell membrane and apical membrane of the acinus was severely decreased. These data suggest that acinar destruction and AQP5 expression are reduced in the submandibular gland tissue of NOD mice model.

Processing of Transcriptome Sequencing

To better understand the molecular mechanisms of salivary secretion regulation, we performed transcriptome sequencing analysis of submandibular gland tissues extracted from 6-week-old ICR and NOD mice model using RNA-seq. The dataset was preprocessed using the “Affy” package in R to remove systematically biased genes in the original data. Figure 2A shows gene expression before and after normalization. Based on the normalized data, the NOD and ICR mice were completely distinguished by PCA analysis (Fig. 2B). The DEGs were screened using the “limma” package in the R language, with “P-value < 0.05 and |log2FC|>1” as filter conditions. 834 DEGs were obtained, including 505 up-regulated and 329 down-regulated genes (Fig. 2C). In addition, a heatmap was used to show the expression of all DEGs (Fig. 2D).

Fig. 2figure 2

Processing of transcriptome sequencing. A Normalization of the transcriptome sequencing data. The black line in each box represents each data group’s median, which determines the degree of normalization of the data through its distribution. The upside is the expression value data before normalization, and the normalized expression value data is the downside. B PCA diagram of samples based on expression abundance. C The volcano of DEGs. The blue points indicate the screened down-regulated DEGs, the red points indicate the screened up-regulated DEGs, and the black points indicate genes with no significant differences. D The heatmap for all DEGs. All DEGs are screened based on P-value < 0.05 and |fold change| > 1. DEGs differentially expressed genes; PCA, Principal Component Analysis

GO and KEGG Enrichment Analysis

To further understand the molecular functions and pathways involved in DEGs, we performed an enrichment analysis using the DAVID database. The GO enrichment results showed that these DEGs play critical roles in regulating water transport, insulin secretion, and cell-cell signaling (Fig. 3A-C). Further KEGG enrichment analysis showed that insulin signaling is essential in regulating DEGs. Furthermore, salivary secretion and cAMP signaling significantly enriched KEGG (Fig. 3D).

Fig. 3figure 3

GO and KEGG enrichment analysis. A-C GO enrichment analysis of the DEGs. D KEGG enrichment analysis of the DEGs. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; DEGs, differentially expressed genes

Identification of Hub Genes

First, an integrated analysis of insulin signaling, cAMP signaling, and salivary secretion related genes and DEGs was performed using the Venn tool, and overlapping parts of the four datasets were taken. In our study, 42 overlapping DEGs were identified, including 26 up-regulated DEGs and 16 down-regulated DEGs (Fig. 4A).

Fig. 4figure 4

Analysis and screening of hub genes. A Venn diagram of 42 overlapped DEGs of insulin signaling, cAMP signaling, salivary secretion, and DEGs. B The PPI network of overlapped DEGs was constructed in Cytoscape. C MCODE analysis. D Degree score sorting. DEGs, differentially expressed genes

To establish protein-protein interactions, we constructed a PPI network for overlapping DEGs using the STRING database, consisting of 42 nodes and 37 edges. After exporting the TSV file, Cytoscape filters out non-interacting proteins and visualizes them (Fig. 4B). In addition, the algorithm of the MCODE plug-in was used for analysis, and the highest-scoring module was selected as the hub module (Fig. 4C). The Degree plug-in identified the network of the top five genes in cytoHubba (Fig. 4D). Finally, we integrated two key networks of genes (REN, A2M, SNCA, KLK3, TTR, and AZGP1) as our hub genes, which are also considered to be initiating regulators of insulin signaling.

Diagnostic Effectiveness of the Biomarkers

To ensure the reliability of transcriptome sequencing analysis results, we used the dataset GSE40611 from the GEO database to examine the expression of hub genes between the pSS and the control. The number of samples in the pSS group was 17, and the number in the control group was 18. The results showed that the expression levels of hub genes REN and A2M were significantly higher than in the control group (P < 0.001). In comparison, the expression levels of the hub genes SNCA, KLK3, TTR, and AZGP1 in the pSS were significantly lower than those in the control group (P < 0.05) (Fig. 5A). We further demonstrated the expression of hub genes between pSS and control groups through clustering heatmaps (Fig. 5B). In addition, we performed ROC analysis to detect the diagnostic validity of hub genes as pSS biomarkers, and “AUC > 0.7” is considered to have good sensitivity for pSS diagnosis. As shown in Fig. 5C, in pSS, the AUC values for REN, A2M, SNCA, KLK3, TTR, and AZGP1 are 0.704, 0.709, 0.827, 0.827, 0.703, and 0.752, respectively.

Fig. 5figure 5

Diagnostic Effectiveness of the Biomarkers. A Expression levels of the hub genes between pSS and control samples in GSE40611. B The expression heatmap of the hub gene in GSE40611. C ROC analysis of hub genes. ROC, Receiver Operator Characteristic Curve

Construct a Regulatory Relationship between Hub Genes and cAMP Signaling and Salivary Secretion

To investigate the relationship between hub genes and cAMP signaling and salivary secretion, we analyzed them using Venn and Cytoscape. First, we integrated cAMP signaling-related genes and DEGs, took their overlapping parts, and showed the up- and down-regulated overlapping genes through cluster analysis heatmaps (Fig. 6A). Next, we constructed regulatory networks between upregulated hub genes (REN and A2M) and downregulated hub genes (SNCA, KLK3, TTR, and AZGP1) and overlapping genes (Fig. 6B), respectively. From this, we can see that the hub gene plays a crucial role in regulating the cAMP signaling gene. In addition, we also integrated salivary secretion-related genes and DEGs (Fig. 7A) and constructed a regulatory network between hub genes and overlapping genes (Fig. 7B). The results also suggest that hub genes are crucial in regulating salivary secretion genes. Finally, based on these 42 co-expressed genes, we established the regulatory relationship between the hub genes, cAMP signaling, and salivary secretion. CATSPER3, DCPS, OTOF, AGR2, and FOXC2 are the top five-degree genes in cAMP signaling (Fig. 8A), while ASCL2, SLC52A3, MMP12, AGR2, and MMP27 are the top five-degree genes in the salivary secretion (Fig. 8B).

Fig. 6figure 6

Construct a regulatory relationship between hub genes and cAMP signaling. A The cluster heatmap showed the up- and down-regulated overlapping genes (cAMP signaling and DEGs common genes). B The regulatory network between the hub and overlapping genes. The red triangle represents the upregulation of common DEGs, the green triangle represents the downregulation of common DEGs, the red circle represents the upregulation of hub genes, and the green circle represents the downregulation of hub genes

Fig. 7figure 7

Construct a regulatory relationship between hub genes and salivary secretion. A The cluster heatmap showed the up- and down-regulated overlapping genes (salivary secretion and DEGs common genes). B The regulatory network between the hub and overlapping genes. The red triangle represents the upregulation of common DEGs, the green triangle represents the downregulation of common DEGs, the red circle represents the upregulation of hub genes, and the green circle represents the downregulation of hub genes

Fig. 8figure 8

The regulatory relationship between the hub genes, cAMP signaling, and salivary secretion. A Construction of regulatory networks between upregulated hub genes and DEGs of cAMP signaling and salivary secretion. B Construction of regulatory networks between downregulated hub genes and DEGs of cAMP signaling and salivary secretion. The red triangle represents the upregulation of common DEGs, the green triangle represents the downregulation of common DEGs, the red circle represents the upregulation of hub genes, the green circle represents the downregulation of hub genes, and the blue circle represents the pathway

Experimental Verification of Hub Genes

To further validate the results of our transcriptome sequencing analysis, we treated A253 cells with IFN-γ at 100 nM and used PBS as the control group. Changes in the expression of hub genes at the mRNA level are then detected. The results showed that after IFN-γ treatment, the expression of REN and A2M was increased at the mRNA level, while the expression of SNCA, KLK3, TTR, and AZGP1 was decreased at the mRNA level (Fig. 9A). In addition, we selected submandibular gland tissue from NOD and ICR mice, respectively. The results also showed that the expression of hub genes REN and A2M was elevated at the mRNA level. In contrast, the expression of SNCA, KLK3, TTR, and AZGP1 was decreased at the mRNA level (Fig. 9B), which further verified the reliability of the results. These data suggest that the hub gene has some validity in diagnosing SS, which is consistent with the results predicted by sequencing analysis.

Fig. 9figure 9

Experimental verification of hub genes. A The mRNA levels of REN, A2M, SNCA, KLK3, TTR, and AZGP1 were detected after cells were treated with 100 nM IFN-γ. B The mRNA levels of REN, A2M, SNCA, KLK3, TTR, and AZGP1 were detected between NOD and ICR mice submandibular gland tissue

留言 (0)

沒有登入
gif