Results

Identification of FUNDC1-associated DN hub genes

Transcriptomic analysis of the glomerular and tubular datasets identified a total of 5560 DEGs between DN and control groups. Specifically, 2625 DEGs were screened in glomerular tissues, including 1249 upregulated and 1376 downregulated genes, while 3199 DEGs were obtained in tubular tissues, consisting of 3119 upregulated and 80 downregulated genes (Supplementary Materials 1–2). Based on the ranking of |log2 fold change (FC)|, the top 10 upregulated genes (including NOD2, FOXM1, and CD53) and top 10 downregulated genes (including ACTA1, PRTG, and KIT) were selected and visualized via volcano plots and heatmaps (Fig. 2).

Fig. 2: Volcano plot and heatmap for identifying differentially expressed genes (displaying the top 10 genes with upregulated or downregulated expression).

A total of 3611 mitochondrial dysfunction-related genes (MDRGs) were retrieved from the GeneCards database, and 130 FUNDC1-interacting genes were obtained from three gene interaction databases (Supplementary Materials 3–4). After intersecting the DEGs, MDRGs, and FUNDC1-interacting genes, 20 FUNDC1-associated hub genes potentially involved in DN were identified, including DHCR24, EHHADH, GABARAPL2, TUFM, BNIP3, FIS1, HIF1A, ITPR3, PPARGC1A, FUNDC1, USP30, TFEB, PACS2, TBK1, MIEF2, ATG5, MUL1, PRKN, ATG12, HSPA9.

Enrichment analysis

GO and KEGG enrichment analyses of the 20 hub genes identified 87 GO terms (61 BP, 18 CC, and 8 MF) and 8 KEGG pathways (Supplementary Material 5), which were primarily associated with mitophagy, mitochondrial organization, and hyperglycemia-related signaling pathways, suggesting that these genes may participate in pathways related to mitochondrial dysfunction and DN progression. (Fig. 3).

Fig. 3: GO and KEGG enrichment analysis of FUNDC1–DN hub genes.

Identification of key genes by SMR analysis

SMR analysis of these hub genes initially identified two candidate genes (TUFM and EHHADH). After HEIDI testing, only EHHADH remained as a significant candidate gene associated with DN without evidence of heterogeneity (HEIDI P > 0.05). The effect estimate suggested an inverse genetic association between EHHADH expression and DN risk (βSMR = −0.15), consistent with its downregulated expression pattern (log2FC < 0) (Supplementary Material 6). Genomic region analysis indicated that EHHADH is located within the 184.0–186.0 Mb interval on chromosome 3. In this region, eQTL, GWAS, and SMR signals overlapped, suggesting a potential shared genetic signal between EHHADH expression and DN susceptibility. These findings suggest that EHHADH may represent a candidate gene associated with DN in this locus and warrants further investigation (Fig. 4).

Fig. 4: Results of SMR analysis of key gene.

Functional and correlation analyses of key genes

Single-gene GSEA suggested that both FUNDC1 and EHHADH were predominantly enriched in mitochondrial-related processes, including mitochondrial matrix, mitochondrial protein complexes, and monocarboxylic acid catabolic processes (all P < 0.05, |NES| > 1). KEGG-based GSEA further revealed that FUNDC1 and EHHADH were significantly enriched in pathways related to respiratory chain complex I electron transport, thermogenesis, and SNCA/ABETA/PINK1 mutation-associated electron transport dysfunction (all P < 0.05, |NES| > 1) (Supplementary Material 7).

Based on FUNDC1, its hub gene EHHADH, and their top five enriched KEGG pathways, a FUNDC1–hub gene–pathway interaction network was constructed using Cytoscape. Correlation analysis showed a significant positive association between FUNDC1 and EHHADH expression (cor = 0.44, P = 0.00043), indicating a potential coordinated association in DN (Fig. 5).

Fig. 5: Functional and correlation analyses of key genes.

TF regulatory network and cross-compartment transcriptomic assessment of hub gene expression

TF prediction analysis identified 60 TFs regulating EHHADH and 77 TFs regulating FUNDC1, among which 17 TFs were shared (Supplementary Material 8). A TF–hub gene regulatory network was subsequently constructed, highlighting E2F4 as a candidate shared regulator of FUNDC1 and EHHADH (Fig. 6A).

Fig. 6: TF regulatory network of FUNDC1 and EHHADH and their tubular expression evaluation in DN.

To further evaluate the expression pattern of the identified hub genes across renal compartments, GSE294519 was analyzed using the same differential expression pipeline. Consistent with the direction observed in the discovery cohort, EHHADH remained downregulated in tubular tissues from patients with DN and showed nominal statistical significance (P = 0.0101), although the adjusted P value indicated borderline significance (FDR = 0.0549). In contrast, FUNDC1 expression did not differ significantly between DN and control samples (Fig. 6B–Supplementary Material 2).