Identification of DEGs in IgAN and SS

The research flowchart is depicted in Fig. 1. DEG analyses were performed using GSE93798 and GSE40611 datasets. In GSE93798, we obtained 1468 DEGs, involving 707 up-regulated and 761 down-regulated genes, while in GSE40611, 142 DEGs were found, containing 123 up-regulated and 19 down-regulated genes, respectively. These results are presented in the volcano plots (Fig. 2A and B). We also labeled the top 10 up-regulated and top 5 down-regulated genes in the heatmaps (Fig. 2C and D). Subsequently, 28 commonly up-regulated and 2 commonly down-regulated DEGs between IgAN and SS were identified using the Venn diagram (Fig. 2E).

Fig. 1

Workflow diagram of the study

Fig. 2

Identification of DEGs in the IgAN and SS. A Volcano plot of GSE93798. B Volcano plot of GSE40611. C Heat map of GSE93798. D Heat map of GSE40611. E Venn diagram showing the common DEGs between the IgAN and SS. F Venn diagram showing the shared genes between the IgAN and SS in the WGCNA modules. DEGs, differentially expressed genes; IgAN, IgA nephropathy; SS, Sjögren’s syndrome; WGCNA, weighted gene co-expression network analysis

To further uncover the latent biological processes and pathways, GO and KEGG analyses were done using the commonly up-regulated DEGs. The results indicated that these DEGs were principally enriched in lymphocyte-mediated immunity, MHC class II protein complex assembly, antigen processing and presentation, virus and staphylococcus aureus infections, and autoimmune diseases (Supplementary File S4).

WGCNA in IgAN and SS

A total of 5753 and 4665 genes were preserved with a meanFPKM value of 7 and 9 in the GSE93798 and GSE40611 datasets, respectively, and the function “Hculst” locked 42 and 28 samples, respectively, for the subsequent processing. The “pickSoftThreshold” function yielded a proper soft threshold value of 10 in IgAN and 7 in SS, respectively (Fig. 3A, B, D, and E). In the GSE93798 and GSE40611 datasets, four and twelve modules with diverse colors were screened out in the WGCNA, respectively (Fig. 3C and F).

Fig. 3

WGCNA in the IgAN and SS. A Scale independent plot in the GSE93798. B Mean connectivity plot in the GSE93798. C Cluster dendrogram in the GSE93798. D Scale independent plot in the GSE40611. E Mean connectivity plot in the GSE40611. F Cluster dendrogram in the GSE40611. G Heat map of the module-trait relationships in the GSE93798. H Heat map of the module-trait relationships in the GSE40611

The relationship between these modules and corresponding disease was computed using the Spearman correlation coefficient heatmap. In the GSE93798, the “MEturquoise” module, including 1435 genes, showed a significantly positive correlation with IgAN (R = 0.94, P = 1e−20), while the “MEblue” and “MEbrown” modules were negatively correlated with IgAN (R = − 0.79, P = 5e−10; R = − 0.87, P = 8e−14) and comprised a total of 1164 genes (Fig. 3G). Similarly, in the GSE40611, the “MEyellow” module, containing 453 genes, showed a high positive correlation with SS (R = 0.86, P = 4e−09), while the “MEred” module had a negative correlation with SS (r = − 0.55, P = 0.003), comprising 246 genes (Fig. 3H).

Afterward, 98 interactively positive-correlated and 38 interactively negative-correlated module genes between IgAN and SS were identified though the Venn diagram (Fig. 2F). The enrichment analyses suggested that the commonly positive-correlated module genes were primarily enriched in the antigen processing and presentation, response to virus, staphylococcus aureus infections, and autoimmune diseases, which were consistent with the results of DEG analysis (Supplementary File S5).

Identification of hub genes between IgAN and SS

The 28 commonly up-regulated DEGs and 98 interactively positive-correlated module genes between IgAN and SS were merged, and 120 genes were acquired for the subsequent research. As shown in Fig. 4A–D, the results of GO and KEGG enrichment analyses highlighted the crucial roles of antigen processing and presentation and response to virus in the biological processes and pathways in both autoimmune diseases. Subsequently, we employed PPI analysis through the STRING site and three algorithms of the plug-in cytoHubba (MCC, MNC, and EPC) in the Cytoscape to investigate the core pathogenic genes. The interaction networks display the top 30 genes that obtained using the MCC, MNC, and EPC algorithms, respectively (Fig. 4D–F). By intersecting the Venn diagrams (Fig. 4G), 24 shared genes were identified as the candidate hub genes, including PSMB8, PSMB9, IFI44, ISG15, CD53, HLA-DQA1, HLA-DQB1, HLA-C, HLA-DMB, HLA-DRA, HLA-DPA1, GZMA, GIMAP7, CXCL9, CXCL10, CMPK2, SAMD9L, PARP14, PARP9, IFIT1, BTK, IRF8, MYD88, and CD48.

Fig. 4

GO, KEGG, and PPI analyses. The A cellular component, B molecular function, C biological process, and D KEGG enrichment analyses of the 120 crosstalk genes. The PPI networks between the top 30 genes in the E MCC, F MNC, and G EPC analyses. H Venn diagram showing the 24 potential hub genes between the IgAN and SS. GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; PPI, protein–protein interaction; MCC, maximal clique centrality; MNC, maximum neighborhood component; EPC, edge percolated component

Validation of hub gene expressions between IgAN and SS

The expression levels of the 24 genes were externally validated in the IgAN dataset Ju CKD and SS dataset GSE7451. The results indicated that the expressions of PSMB8, PSMB9, IFI44, ISG15, and CD53 in the IgAN glomerular samples were obviously higher than those in the samples of healthy controls, respectively (Fig. 5A). More interestingly, the expressions of ISG15 in the glomerular samples exhibited a remarkably positive correlation, while the CD53 showed a negative correlation with the GFR values of corresponding IgAN patients, respectively (R = 0.41, P = 0.039; R = − 0.47, P = 0.016; Fig. 5B). Similar analyses were done in the Ju CKD TubInt. The results demonstrated that the PSMB8, PSMB9, IFI44, ISG15, and CD53 expressions were higher in the IgAN tubulointerstitial samples than those in the controls, respectively, and the PSMB8, PSMB9, and CD53 expressions showed negative correlations with the GFR values (R = − 0.7, P = 0.00016; R = − 0.65, P = 0.00071; R = − 0.53, P = 0.0089), respectively, in the IgAN tubulointerstitial samples (Fig. 5C and D).

Fig. 5

External validation of the hub gene expressions and correlation analyses. The expressions of the five hub genes in the A Ju CKD Glom, B Ju CKD TubInt, and C GSE7451 datasets. Correlation analyses between the expression levels of the five hub genes and the GFR (ml/min/1.73 m.2) using the D Ju CKD Glom and E Ju CKD TubInt datasets. P value < 0.05 was considered statistically significant. *P < 0.05, **P < 0.01, ***P < 0.001

On the other hand, in the GSE7451, the expression levels of PSMB8, PSMB9, IFI44, ISG15, and CD53 were significantly increased in the salivary gland samples of SS than those in the samples of healthy controls, respectively (Fig. 5E). Consequently, according to these comparative results, PSMB8, PSMB9, IFI44, ISG15, and CD53 were selected as the hub genes between IgAN and SS for the subsequent analyses.

In the IHC analyses, the expressions of PSMB8 and PSMB9 were significantly increased in the glomeruli and tubulointerstitium of IgAN than those in the controls, respectively (Fig. 6A–C). As for the patients with renal involvement in SS, the expression levels of PSMB8 and PSMB9 were mainly elevated in the tubulointerstitium (Fig. 6A, D, and E). Furthermore, the PSMB8 and PSMB9 levels in the tubulointerstitium were negatively correlated with the eGFR values of patients with renal involvement in SS and IgAN, respectively, and the proteinuria levels of patients with IgAN, respectively (Supplementary File S6). As for the Oxford classification of IgAN (M [Mesangial hypercellularity], E [Endocapillary proliferation], S [Segmental sclerosis], T [Tubular atrophy and interstitial fibrosis], and C [Crescentic] scores), the PSMB8 and PSMB9 levels in the tubulointerstitium were markedly increased in the T1/2 groups than those in the T0 groups, respectively (Supplementary File S7). Additionally, we also observed that the PSMB9 levels in the glomeruli were elevated in the M1, S1, and C1 groups, respectively, though the lack of statistical significance in our data (Supplementary File S7).

Fig. 6

External validation of the PSMB8 and PSMB9 expressions using the IHC methods. A Photomicrographs of PSMB8 and PSMB9 in the glomeruli and tubulointerstitium of IgAN, SS, and controls. Quantifications of the PSMB8 expressions in the B glomeruli and C tubulointerstitium of IgAN and controls. Quantifications of the PSMB9 expressions in the D glomeruli and E tubulointerstitium of IgAN and controls. P value < 0.05 was considered statistically significant. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Immune cell infiltration and correlation analyses

The immune cell infiltration analyses were done in different samples. The distributions of 28 types of immune cells in the GSE93798 and GSE40611 samples were identified and are presented in the box plots in Fig. 7A and B. We observed that various types of immune cells were significantly activated both in the IgAN and SS samples, including effector memory (EM) CD8 T cell, T follicular helper (Tfh) cell, type 1 T helper cell, regulatory T cell, natural killer T cell, central memory CD4 T cell, immature B cell, activated B cell, natural killer cell, activated dendritic cell, and myeloid-derived suppressor cell (MDSC).

Fig. 7

Immune infiltration analyses associated with the IgAN and SS. A The box plot showing the distribution of 28 immune cells in the GSE93798. B The box plot showing the distribution of 28 immune cells in the GSE40611. C Correlation analysis between immune cells and the hub genes in the GSE93798. D Correlation analysis between immune cells and the hub genes in the GSE40611. P value < 0.05 was considered statistically significant. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.001

The correlations between the infiltrated immune cells and the expressions of hub genes are depicted in Fig. 7C, D, and Supplementary File S5. The EM CD8 T cell, Tfh cell, central memory CD4 T cell, activated B cell, activated dendritic cell, and MDSC were positively correlated to the expressions of PSMB8 and PSMB9, respectively. Most of the activated immune cells mentioned above, except for nature killer cell and nature killer T cell, exhibited remarkably positive correlations with the expressions of CD53 in IgAN and SS, respectively. More interestingly, we noticed that the EM CD8 T cell and Tfh cell showed significantly positive correlations with the expressions of all the five hub genes, respectively, and activated B cell and activated dendritic cell were positively related to the PSMB8, PSMB9, ISG15, and CD53, respectively, in the two diseases. Therefore, these data indicate that the EM CD8 T cell, Tfh cell, activated B cell, and activated dendritic cell may exert dominant effects in the pathogenesis of IgAN and SS.

TF prediction and validation

According to the NetworkAnalyst, we have identified twenty-eight TFs that may participate in the transcriptional regulatory of our five crosstalk genes, including STAT1, FOXC1, USF1, CREB1, YY1, POU2F2, GATA2, GATA3, TP53, MAX, etc. The transcriptional regulatory network is presented in the Supplementary File S6A. As shown in the Supplementary File S5B, the expressions of STAT1 were remarkably elevated in SS (GSE7451) and IgAN tubulointerstitium (Ju CKD TubInt) than healthy controls. We also observed that the expression of STAT1 showed a significantly negative correlation with GFR values in the tubulointerstitium of IgAN (R = − 0.44; P = 0.033; Supplementary File S5C). Together, these results demonstrate a possible role of STAT1 in the transcriptional regulatory of ISG15 and CD53 both in the IgAN and SS.