A bioinformatics analysis to identify shared molecular pathways and hub genes between NAFLD and Gestational Diabetes Mellitus Authors: Nasrin Amiri-Dashatan, Mehdi Koushki, Masoumeh Farahani, Somayeh Chahkandi, Mohammad Salehi, Hossein Chiti, Zohreh Amarloo, Aliasghar Keramatinia, Mohsen Norouzinia Published: 2025 ABSTRACT AbstractAim:This research aimed to identify the possible links between Non-alcoholic fatty liver disease (NAFLD) and gestational diabetes mellitus (GDM) by examining shared genes and pathways through the use of bioinformatics tools.Background:NAFLD presents several potential risk factors for the onset of GDM.Methods:We downloaded relevant microarray datasets from the Gene Expression Omnibus (GEO) database for screening common differentially expressed genes (DEGs) between NAFLD and GDM by GEO2R. Then, we used gene ontology analysis to explore the biological processes and KEGG pathways of NAFLD and GDM occurrence. The hub genes of each disease were screened by analysis of the PPI network, and the common hub genes were identified. We designed a co-expression network and miRNA hub gene regulatory network for selected hub genes by using GeneMANIA and miRNet platforms, respectively.Results:We identified 521 and 185 DEGs for NAFLD and GDM, respectively. 10 shared genes (FOS, CD22, AMZ1, ANGPT2, ATP1B2, STAB2, EGR1, MMP9, CXCL9, and LCN2) were screened among DEGs of two diseases, with analysis revealing enrichment in pathways such as IL-7 and TNF signaling pathways. Then, common hub genes (FOS, MMP9, and CXCL9) were identified via PPI network analysis, which were strongly correlated with both diseases. In addition, has-mir-34a-5p and has-mir-335-5p were detected as shared miRs that target hubs.Conclusion:We found 3 shared hub genes involved in NAFLD and GDM. Our research established a connection between the two diseases and could aid in formulating possible intervention strategies aimed at addressing both disorders by utilizing these risk factors. Key Words: Non-alcoholic fatty liver disease, Gestational diabetes mellitus, Bioinformatics, Protein-protein interaction network, Shared genes Introduction Non-alcoholic fatty liver disease (NAFLD) represents the most common type of chronic liver disease which was more recently termed " metabolic-associated fatty liver disease" (MAFLD) in 2020 (1), and later referred to as "metabolic dysfunction-associated steatotic liver disease" (MASLD) in 2023 (2). The occurrence of NAFLD stands at 24.4% among adults and 7.6% among children, with a significantly elevated prevalence noted in groups impacted by obesity and type 2 diabetes (T2D) (3, 4). It is defined by an abnormal buildup of fat in the liver, and the increasing occurrence of NAFLD poses a significant public health issue, with considerable consequences for healthcare systems and personal health. In addition, NAFLD is becoming a more prevalent cause of cirrhosis and hepatocellular carcinoma, and it is on track to emerge as the leading reason for liver transplantation in the United States (5). At present, there is no efficient pharmacotherapeutic approach available for NAFLD. Consequently, treatment relies on lifestyle changes, which encompass dietary adjustments and increased physical activity, although the most effective type of intervention remains uncertain (6). Gestational diabetes mellitus (GDM) represents the most prevalent medical complication encountered by pregnant women, characterized by any form of glucose intolerance that is either recognized or arises during pregnancy (7). The growing prevalence of obesity has led to a rise in the frequency of GDM and associated complications during both pregnancy and the perinatal period. Several risk factors associated with GDM encompass significant weight gain during early adulthood and prior to 24 weeks of gestation, advanced maternal age, a previous occurrence of GDM, a family history of diabetes, a pre-pregnancy body mass index (BMI) of 30 kg/m2 or greater, and pre-pregnancy polycystic ovary syndrome (PCOS) (8). In recent years, the link between NAFLD and GDM has captivated researchers' interest. In this regard, several observational studies have also suggested that individuals with GDM are at a heightened risk of developing NAFLD (9, 10). A recent investigation by Cho et al. examined the relationship between a previous history of GDM and risk of incident NAFLD and showed that a previous history of GDM is an independent risk factor for NAFLD development (11). Although the previous studies have indicated a correlation between NAFLD and GDM, implying potentially shared molecular pathways and genes, it is still less explored. We chose the NAFLD and GDM for bioinformatics analysis of shared genes and molecular pathways because these diseases have interconnected metabolic dysfunctions, particularly related to insulin resistance and lipid metabolism. Therefore, studying shared genes and pathways helps identify molecular mechanisms and hub genes involved in processes like lipid catabolism, inflammation, insulin signaling, and cellular stress responses that contribute to both diseases. In this regard, bioinformatics analysis could be applied to disclose the pathogenesis and crosstalk mechanisms and genes between different diseases at the transcriptomic level (12-16). We hypothesize the existence of common genes between NAFLD and GDM, and then to address these issues, we applied bioinformatics analysis to reveal potential connections between NAFLD and GDM by analyzing shared key genes and molecular pathways. Methods Data sources Gene expression profiles were searched in the GEO database using the keywords “Liver Diseases” OR “Non-Alcoholic Fatty Liver Disease” OR “NAFLD” to select the NAFLD-related dataset and “Gestational Diabetes Mellitus” OR “GDM” for the GDM-related dataset. Inclusion criteria for the datasets: (1) gene expression analysis must include both case and control groups; (2) the samples used for analysis is tissue (liver for NAFLD, and Placenta for GDM); (3) raw or processed data must be available for re-analysis. Exclusion criteria included the following: (1) incomplete datasets (2) Lack of detailed clinical descriptions or phenotypic characteristics of patients or samples. After the screening, two datasets were downloaded, including GSE126848 and GSE154414, related to NAFLD and GDM diseases, respectively. The flowchart of the study is shown in Figure 1. The study protocols. DEGs: Differentially Expressed Genes; GO: Gene Ontology; KEGG: Kyoto Encyclopedia of Genes and Genomes; PPI: Protein- Protein Interaction. Identification of differentially expressed genes (DEGs) by the GEO2R tool The analysis of each dataset according to the differential expression was performed using the GEO2R analysis tool. The datasets were normalized, and Benjamini & Hochberg (FDR) was performed to adjust the P-values. The DEGs were determined with criteria as follows: adjusted p-value less than 0.05 and |log fold change (FC)| larger than 1.5. The results of DEGs were displayed using a volcano plot. Identification of common signature genes by Venn diagram The DEGs associated with NAFLD and GDM were identified in every dataset separately. Then, the common up- and down-regulated genes between the two datasets were analyzed using the Venny online platform to find the intersection parts between the datasets. Enrichment gene ontology and pathway analysis To describe the biological process (BP) and associated pathways of the DEGs of each dataset, the gene ontology (GO) enrichment analysis and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis were done by ShinyGO 0.80 (http://bioinformatics.sdstate.edu/go/). GO terms and also KEGG analysis data were screened based on p-value < 0.05. Protein-protein network analysis and hub gene selection We further predicted the protein-protein interaction (PPI) network using all the DEGs in each dataset, including upregulated and downregulated genes from the STRING (https://string-db.org) database, aiming to provide more comprehensive information for the key genes. The visualization and analysis of the constructed network were performed by Cytoscape software version 3.7.0. Then, the top key genes with the highest degree index were identified through the Cytohubba plugin. To further evaluate internal interaction within commonly selected hub genes shared between two diseases, we predicted a co-expression network by using the GeneMANIA (https://genemania.org/) platform. Construction of miRNA‑hub gene regulatory network for common hubs The miRNet (www.mirnet.ca) serves as a comprehensive online platform aimed at facilitating the integrative analysis of miRNAs-centric regulatory networks. This database aggregates a broad range of experimentally and predicted miRNA-target interactions to construct a network. In this analysis, we predicted a miRNA-hub gene regulatory network using the miRNet database (https://mirnet.ca) that targets common shared hub genes between GDM and NAFLD. Results Identification of DEGs for NAFLD dataset GSE126848 includes 15 NAFLD and 14 healthy control samples, of which five samples are from the NAFLD group (including GSM3615296, GSM3615298, GSM36152300, GSM3615301, and GSM3615307), and one sample from the control group (including GSM3615328) was excluded from further analyses due to non-normality of data. The analysis consisted of 10 and 13 samples from the NAFLD and healthy control groups, respectively. The volcano plot and UMAP graph of the dataset after analysis are presented in Figure 2a, b. 304 up- and 217 downregulated genes were recognized in GSE126848. Volcano plots and UMAP graph of the studied datasets. a) GSE126848, and b) GSE154414. Red dots: upregulated genes. Blue dots: downregulated genes. Black dots: genes with no significant differences in the expression level. The nbrs parameter in a UMAP plot refers to the "n_neighbors" parameter, which controls the number of nearest neighbors used by UMAP to determine the local structure of the data. Identification of DEGs for the GDM dataset The GSE54414 dataset was chosen for analysis, consisting of 4 GDM and 4 control samples. The volcano plot and UMAP graph of the GDM dataset analysis are shown in Figure 2c, d. We detected a total of 129 up- and 56 downregulated genes in this analysis. Identification of shared genes A total of 10 shared DEGs between NAFLD and GDM (including ANGPT2, FOS, EGR1, LCN2, ATP1B2, CXCL9, CD22, AMZ1, STAB2, and MMP9) were detected by Venn diagram analysis (Figure 3). The list of total integrated differentially expressed genes, including up- and downregulated genes along with their descriptions, is shown in Table 1. Venn diagram showing shared DEGs between the selected GEO datasets for NAFLD and GDM. Table 1List of shared DEGs between NAFLD and GDM diseases.NO.Gene nameDescriptionGene Expression Status1FOSFos proto-oncogene, AP-1 transcription factor subunitDown (GDM & NAFLD)2CD22CD22 moleculeUp (GDM), Down (NAFLD)3AMZ1archaelysin family metallopeptidase 1Up (GDM), Down (NAFLD)4ANGPT2angiopoietin 2Down (GDM & NAFLD)5ATP1B2ATPase Na+/K+ transporting subunit beta 2Down (GDM & NAFLD)6STAB2stabilin 2Up (GDM), Down (NAFLD)7EGR1early growth response 1Down (GDM & NAFLD)8MMP9matrix metallopeptidase 9Down (GDM & NAFLD)9CXCL9C-X-C motif chemokine ligand 9Up (GDM & NAFLD)10LCN2lipocalin 2Up (GDM & NAFLD) List of shared DEGs between NAFLD and GDM diseases. Results of enrichment gene ontology and pathway analysis The Gene Ontology analysis of the DEGs was carried out by the ShinyGO tool, and the results are shown in Table 2. GO analysis of the NAFLD dataset indicated that in BP, the DEGs were firstly enriched for the GO terms including “inflammatory response”, “cell adhesion”, and “cell migration”. GO analysis discloses that the GDM DEGs were significantly enriched in the “ion homeostasis”, “cell death”, “response to external stimulus”, “immune system”, and “response to stress” (Table 2). As shown in Figure 4a, the significantly enriched KEGG pathways for NAFLD DEGs were associated with “viral protein interaction with cytokine and cytokine receptor (p-value: 2.65E-04)”, “protein digestion and absorption (1.23E-03)”, and “cytokine-cytokine receptor interaction (2.65E-04)”. In addition, “hematopoietic cell lineage (8.72E-03)”, “IL-17 signaling pathway (2.94E-02)”, “viral protein interaction with cytokine and cytokine receptor (2.94E-02)”, and “pathways in cancer (3.42e-02)” are associated with GDM DEGs (Figure 4b). The results of KEGG pathway analysis of shard DEGs are presented in Figure 4c. Table 2List of top 10 significant biological processes enriched for NAFLD and GDM DEGs.Enrichment FDRnGenesPathway GenesFold EnrichmentPathwayGO (Biological Process) of NAFLD DEGs2.54E-076217292.427476292GO:0007155 cell adhesion1.17E-065716022.40863307GO:0016477 cell migration2.37E-06398922.959770611GO:0006954 inflammatory response5.49E-065918062.211528632GO:0048870 cell motility5.49E-065918062.211528632GO:0051674 localization of cell9.68E-068833001.805207101GO:0070887 cellular response to chemical stimulus1.14E-058732691.801617671GO:0010033 response to organic substance1.15E-056722542.012237285GO:0006928 movement of cell or subcellular component1.51E-058129901.83388514GO:0007166 cell surface receptor signaling pathway1.58E-056119922.072997612GO:0040011 locomotionGO (Biological Process) of GDM DEGs3.27E-075130732.531578262GO:0009605 response to external stimulus2.30E-063820312.854022649GO:0006952 defense response2.30E-06227144.700112045GO:0055065 metal ion homeostasis4.82E-064023202.63GO:0008219 cell death4.82E-06227724.346994819GO:0009617 response to bacterium5.39E-065944422.034326401GO:0006950 response to stress5.39E-06227944.226549118GO:0055080 cation homeostasis6.59E-06228094.148182942GO:0098771 inorganic ion homeostasis7.29E-06228194.097533578GO:0050801 ion homeostasis1.07E-054731832.25239711GO:0002376 immune system proc.GO (Biological Process) of Shared DEGs2.2E-02744243.6GO:0006950 response to stress2.2E-02629904.6GO:0007166 cell surface receptor signaling pathway2.2E-02620344.5GO:0051239 regulation of multicellular organismal process2.3E-02630734.5GO:0009605 response to external stimulus2.6E-02631834.3GO:0002376 immune system process3.7E-02645523GO:0030154 cell differentiation3.7E-02645763GO:0048869 cellular developmental process3.7E-02646173GO:0048731 system development3.9E-02648232.8GO:0006810 transport4.2E-02649872.8GO:0051234 establishment of localizationGO: Gene Ontology; FDR: False Discovery Rate; DEGs: Differentially Expressed Genes List of top 10 significant biological processes enriched for NAFLD and GDM DEGs. GO: Gene Ontology; FDR: False Discovery Rate; DEGs: Differentially Expressed Genes Significant KEGG pathways enriched for a) NAFLD DEGs, b) GDM DEGs, and c) Shared genes. Exploration of hub genes The STRING database was used to create a PPI network of DEGs for NAFLD (with 358 nodes and 932 edges), and GDM DEGs (with 110 nodes and 319 edges). Then, the constructed networks were visualized and analyzed in Cytoscape software. The top 20 hubs of every network were screened based on node degree by CytoHubba (Table 3). Additionally, we predicted the PPI network of common genes between datasets with STRING, and three of the genes were not related to the network and were left out in the subsequent analysis. Interestingly, we found three common hubs between NAFLD and GDM networks, including MMP9, CXCL9, and FOS. The interaction network of hubs for every dataset was presented in Figure 5a-c. To obtain better insight into the potential functions of these shared hub genes, we assessed their co-expression network using the GeneMANIA database. The expression of MMP9, CXCL9, and FOS was closely related to 20 genes, including JUN, TIMP1, IGFBP7, CXCL10, LCN2, JUNB, CXCL11, PPBP, CXCL2, KDM6B, CCL2, TGFB1, CXCR3, JUND, ATF4, FOSB, CXCL12, XCL1, ELK1, and CCL13 (Figure 5d). Interestingly, CXCL10 and LCN2 were among the identified hubs (Table 3, highlighted in yellow). Table 3List of top 20 genes in the network of DEGs for NAFLD, GDM, and shared genes ranked by Degree method I CytoHubba.Top 20 in NAFLD network ranked by Degree methodTop 20 in GDM network ranked by Degree methodIntegrated genes between NAFLD & GDMNameScoreDegreeNameScoreDegree Name ScoreMMP98040CCL56231MMP910CSF1R8040MMP95226CXCL96CD1637839FN15226LCN24FCGR3A7638CCL34824EGR14CXCL106834CXCL94422FOS4FCGR3B6633GZMB4020CD222VCAM16432FOS3417ANGPT22MRC16231CAMP3417CXCL95628CD3216LYVE15427LCN23216SERPINE15226LTF3015KDR4824S100A123015FOS4623CD383015C1QC4623GNLY2814CCR14623IL2RB2613CDK14422CD3E2613SPP14422CXCR12412IL10RA4422HP2211FLT14221H3-52010C1QB4221MMP1189Colored genes represent common hubs between the three groups listed in the table List of top 20 genes in the network of DEGs for NAFLD, GDM, and shared genes ranked by Degree method I CytoHubba. Colored genes represent common hubs between the three groups listed in the table Interaction of top 20 hub genes for a) NAFLD, b) GDM, c) common genes between DEGs of NAFLD and GDM datasets, and d) Hub genes and their corresponding co-expression genes. Colored nodes show the genes, and edges represent protein-protein associations and darker colors of nodes indicate higher levels of node degree. The network of miRNAs and selected hub genes is presented in Figure 6. Based on the miRNAs, a miRNA-hub gene regulatory network was constructed with 283 nodes (miRNA: 280; hub gene: 3) and 315 edges. Notably, FOS (Degree: 239, Betweenness: 37400.87) targeted 239 miRNAs, MMP9 (Degree: 66, Betweenness: 10410) targeted 66 miRNAs, and CXCL9 (Degree: 9, Betweenness: 585.1264) targeted 9 miRNAs. In this analysis, we found that the two miRNAs, including has-miR-34A-5P and has-miR-335-5P, are shared between hub genes (Figure 6). MiRNA—hub gene regulatory network. The red circle nodes represent the common hub genes between GDM and NAFLD; the blue square nodes represent shared miRNAs between hub genes. Discussion Previous observational investigations have disclosed a potential correlation between NAFLD and GDM but the relationship between them regarding gene expression has garnered less focus. It is established that NAFLD is related to impaired glucose tolerance, type 2 diabetes, and metabolic syndrome (4, 17). On the other hand, GDM is also recognized as the most prevalent medical complication among pregnant women, denoting any form of glucose intolerance that is either detected or arises during the course of pregnancy (17). In recent years, the link between NAFLD and GDM has captivated researchers' interest, nevertheless, the molecular mechanisms that explain these relationships remain unclear. Current investigations have revealed that women with NAFLD during early pregnancy have a remarkably increased incidence of GDM during pregnancy (18). Whereas, women who have experienced GDM exhibit a markedly higher likelihood of developing NAFLD in the future (9, 10). Therefore, this study aimed to reveal the link between them through bioinformatics analysis of gene expression profiles. In the present study, GO analysis revealed that DEGs of NAFLD were related to cell adhesion, cell migration, and inflammatory response. There is some evidence that some of the cell adhesion molecules (CAMs) play roles in NAFLD development. For example, integrins represent a group of heterodimeric transmembrane receptors involved in cell adhesion, which play a crucial role in regulating different aspects of the onset and progression of NAFLD (19). In addition, Abo Halima et al. in another study concluded that the significant predictors of bad outcome in patients with NAFLD and NASH were higher VCAM-1 levels, and presented VCAM-1 as non-invasive diagnostic tools for diagnosis of NAFLD degree of fibrosis (20). Regarding the GO of GDM, DEGs were enriched in response to external stimulus, and defense response. Significant KEGG pathways enriched for both NAFLD and GDM DEGs were broadly related to cytokines. Cytokines, which are a category of signaling proteins, have a crucial function in both NAFLD and GDM, frequently serving as mediators of inflammation and insulin resistance. In this regard, numerous research findings suggest that inflammatory mediators could act as potential biomarkers for patients with NAFLD, and are anticipated to offer new perspectives on the aetiology of NAFLD, along with facilitating early diagnosis and intervention (21, 22). In addition, the IL-7 signaling pathway was identified as a significant pathway for GDM. Previous investigations also demonstrated that immune cells and some of the cytokines might play a critical role in GDM (23, 24). In the GEO dataset analysis, we found common hub genes (i.e. MMP9, CXCL9, and FOS) between the two diseases, which were consistent with the differently expressed genes between them. These results can indicate the high importance of these three genes in the molecular mechanism of the studied diseases, which requires more focus on them. The FOS proto-oncogene, commonly known as c-fos, is a gene that plays an important role in cell growth, differentiation, and development. A previous bioinformatics study has introduced this gene as a vital gene involved in liver diseases. Feng et al. demonstrated the importance of FOS as a hub gene in NAFLD/NASH conditions (25). A bioinformatics analysis has shown that FOS and FOSB are among NASH key genes, as well as IL-17 signaling pathways that were enriched in KEGG pathway analysis (26). It has been established that the expression of c-Fos in mice's liver increases inflammation (27). Matrix metalloproteinase-9 (MMP9) has also been one of the common hub genes between studied diseases. MMP-9 plays a main role in chronic liver disease pathogenesis through participating in the processes of inflammation. Kurbatova et al. have reported that MMP-9 is involved in the pathogenesis of NAFLD already at the early stages and its level in blood could be related to the presence and severity of inflammation in the liver parenchyma (28). In another study, Coilly et al. have introduced MMP9 as a predictive marker for poor prognosis in patients with NAFLD using data mining approaches and gene expression analysis (29). Also, the dysregulation of MMPs has reported in placentas of patients with GDM and also NAFLD in other studies (28-30). Therefore, MMP9 as a blood marker could help with the revelation of NAFLD and/or GDM relationship. Chemokine (C-X-C motif) ligand 9 (CXCL9) is another hub gene in common with NAFLD and GDM that we have identified in our analysis. Research indicates that CXCL9 is involved in various pathological processes, such as tumor growth, immune response, and inflammation (31). Moreover, a recent investigation revealed that CXCL9 interferes with the Treg/Th17 equilibrium in a mouse model of metabolic-associated fatty liver disease (MAFLD) through the activation of the p-JNK pathway (32). Furthermore, bioinformatics analysis has also shown that CXCL9 is up-regulated in both humans and animals suffering from NAFLD and NASH disorders (33, 34). CXCL9 plays a crucial role in chronic liver inflammation, as indicated by these findings. In case of GDM, among immune activation biomarkers, the network of chemokines plays a significant role in the pathogenesis of GDM (35). In a recent meta-analysis, Liu et al. have reported a significant increase of CXCL9 along with some other chemokines in the GDM patients compared with the controls (36). Consequently, it is essential to regard these chemokines as possible targets or biomarkers in understanding the pathophysiology of GDM development. According to the findings of the current study, NAFLD and GDM share a complex molecular interaction driven by the convergence of inflammatory signaling, cell adhesion dynamics, and cellular stress responses, which collectively form their shared pathophysiology. Central to this association is chronic low-grade inflammation orchestrated through dysregulated cytokine signaling pathways that impair insulin action and promote metabolic disturbances in both liver and gestational tissues. In NAFLD, innate immune receptors identify damage-associated molecular patterns (DAMPs) released from stressed hepatocytes, initiating inflammatory cascades via Toll-like receptor (TLR) pathways and activating NF-κB signaling. This response not only fuels local hepatic inflammation but also contributes to systemic insulin resistance (37). Similarly, in GDM, increased level of inflammatory mediators and cytokines disrupt insulin signaling and glucose homeostasis in the placenta and maternal tissues, emphasizing a shared inflammatory environment (38). Cell adhesion molecules (CAMs) play a critical role in this process by regulating immune cell recruitment and remodeling of tissues. In NAFLD, integrins and various CAMs mediate hepatic infiltration of immune cells, fostering fibrosis and liver injury (39). Similarly, in GDM, altered adhesion contributes to abnormal placental immune cell dynamics and compromises glucose regulation, signifying a conserved mechanistic axis involving cell adhesion in both conditions. The AP-1 transcription factor complex, with FOS as a key component, modulates gene expression associated with apoptosis, inflammation, and lipid metabolism. Altered expression of FOS, as found in both NAFLD and GDM, likely disrupts this regulatory equilibrium, promoting disease progression (40, 41). Additionally, matrix remodeling enzyme MMP9 contributes by modifying the extracellular environment, facilitating immune cell extravasation and tissue fibrosis in NAFLD, while also being dysregulated in GDM placentas indicating shared pathological remodeling processes (28, 42). Chemokine CXCL9 connects inflammatory signaling and immune cell trafficking in both conditions, sustaining local immune activation and metabolic disturbance (36). Collectively, these pathways create an interdependent network wherein metabolic stress triggers immune and inflammatory signaling; cell adhesion molecules orchestrate immune cell migration and tissue remodeling; and stress-responsive transcriptional programs perpetuate inflammation and metabolic dysregulation. This integrated framework explains the molecular basis of the epidemiologically observed bidirectional risk between NAFLD and GDM and highlights potential targets for diagnostic and therapeutic strategies aimed at these convergent pathways. We propose that the presence of insulin resistance triggers systemic low-grade inflammation, which is further propagated by common molecular mediators like CXCL9. This inflammation simultaneously promotes hepatic steatosis typical of NAFLD and disrupts glucose metabolism during pregnancy, resulting in gestational diabetes mellitus (GDM). Important hub genes such as CXCL9, MMP9, and FOS serve as essential nodes in this interaction by facilitating immune cell recruitment, extracellular matrix remodeling, and inflammatory signaling, thereby connecting the pathophysiology of NAFLD and GDM at the molecular level. The results obtained could also have potential clinical benefits, including CXCL9 as a biomarker for risk stratification and diagnosis, therapeutic targeting of shared inflammatory pathways. Additionally, the shared molecular mechanisms revealed by this study suggest the benefit of integrative clinical monitoring in pregnant women with GDM, including liver function tests and inflammatory marker panels featuring CXCL9 and MMP9. Altogether, these findings underscore the importance of an integrated inflammation-driven network involving cytokine signaling, immune cell recruitment, and tissue remodeling pathways that simultaneously contribute to NAFLD and GDM development. The overlapping molecular signatures and common hub genes suggest a pathophysiological nexus where chronic low-grade inflammation and immune dysregulation act as pivotal mechanisms linking these two metabolic disorders. This integrated pathway perspective not only advances understanding of their interconnected etiology but also identifies promising molecular targets for early diagnosis and therapeutic intervention aimed at disrupting this pathogenic crosstalk. Interestingly, we found that some shared genes between NAFLD and GDM (e.g., CD22, STAB2) are regulated in opposite directions. This could have several reasons include tissue-specific roles for these genes, distinct regulatory mechanisms, and biological complexity and compensatory effects. These genes may play different roles based on the tissue context such as liver in NAFLD and placenta in GDM, leading to contrasting regulatory model. Gene functions and expression controlling can be different by cellular environment, which can account for the differing expression trends even though the genes are part of shared molecular networks (43). Furthermore, various upstream regulatory factors such as epigenetic modifications, or signaling pathways may uniquely control the expression of these genes in each pathology condition (44). Finally, the contradictory regulation may suggest that these genes play different roles in the both conditions. Overall, paradoxical regulation of gene expression across various diseases emphasizes the complexity of common molecular pathways in multifactorial diseases. It also highlights the importance of source of samples in interpreting of bioinformatics-based analysis findings and indicates further experimental validation of results is needed to decipher the functional consequences of such expression patterns. Additionally, in the current investigation, the miRNA-hub gene regulatory network that regulates GDM and NAFLD was constructed. Our results showed that these common hub genes, along with shared miRNAs might be involved in the progression of studied disease. Investigation data from such bioinformatics studies could pave the way for improving early diagnosis, personalized therapy, and preventive care in patients with NAFLD and GDM through shared potential molecular targets and pathways identified by gene expression and network analyses. Limitations This study has several limitations. First, our findings depend on the selected public datasets and the statistical thresholds applied. The use of a stringent cutoff of |logFC| > 1.5 for differential expression, while reducing false positives, may exclude potentially significant genes with more moderate alterations, thereby affecting the overall comprehensiveness of the findings. Second, the datasets analyzed derive from different tissue types, liver tissue for NAFLD and placental for GDM, resulting in biological heterogeneity. This difference in tissue origin may complicate direct comparisons and interpretations of common molecular pathways and hub genes, as gene expression patterns can differ significantly across organs. Finally, the conclusions are based exclusively on bioinformatics correlation analyses without experimental validation. Therefore, the identified shared genes and pathways indicate associations but do not establish causal relationships. Functional studies are required to confirm these molecular links and elucidate their roles in the pathogenesis of NAFLD and gestational diabetes. Conclusion This study elucidates a common molecular framework linking NAFLD and GDM, underscoring chronic low-grade inflammation, dysregulated cytokine signaling, and cell adhesion dynamics as central factors in their pathophysiology. We identified three key and common hub genes (FOS, MMP9, and CXCL9), which emerge as critical mediators bridging immune cell recruitment, extracellular matrix remodeling, and inflammatory processes that drive disease progression in both conditions. Clinically, our results support ffor integrative monitoring methods and suggest that targeting common pathways could lead to novel and effective diagnostic and therapeutic strategies against NAFLD and GDM. This integrated analysis increases understanding of NAFLD and GDM bidirectional relationship and opens new insights for precision medicine strategies focused on disrupting their pathogenic crosstalk. Acknowledgment We thank the Shahid Beheshti University of Medical Sciences and Metabolic Diseases Research Center of Zanjan University of Medical Sciences, Zanjan, Iran that supported this study. Conflict of interests There is no conflict of interest for authors of this article. References