Abstract Background Psoriasis is a chronic, immune-mediated, inflammatory skin disease characterized by abnormal keratinocyte proliferation, in which M1 macrophage polarization plays a critical role. However, the specific biomarkers and mechanisms underlying macrophage polarization in psoriasis remain unclear. Methods We analyzed the psoriasis dataset ([44]GSE14905) to identify differentially expressed genes and applied weighted gene co-expression network analysis to identify key module genes. Macrophage polarization-related (MPR) genes were extracted from the Rummagene database, and MPR genes in psoriasis were identified through Venn analysis. Functional enrichment analysis (GO/KEGG) revealed associated pathways, while six CytoHubba algorithms determined hub genes, with diagnostic potential assessed via ROC curves. Single-gene GSEA further explored biological functions, and single-cell sequencing analysis was performed. Finally, the expression of hub genes and M1 macrophage markers (CD80/CD86) was experimentally validated in psoriasis mouse models. Results Six hub genes (ISG15, RSAD2, IFIT3, OASL, GBP1, and IFIT1) were identified through cytoHubba algorithms. Functional enrichment analysis revealed significant associations between psoriasis-associated macrophage polarization and the RIG-I-like receptor, NOD-like receptor, and cAMP signaling pathways. Experimental validation verified the increased expression of these hub genes and M1 macrophage markers in LPS-stimulated RAW264.7 murine macrophages and IMQ-induced psoriasis animal models. Conclusion Our findings suggest that six interferon-responsive genes (ISG15, RSAD2, IFIT3, OASL, GBP1, and IFIT1) could serve as potential biomarkers for M1 macrophage polarization in psoriasis. Targeting macrophage polarization through IFN pathway inhibition may offer novel therapeutic strategies, particularly for patients with prominent IFN signatures refractory to conventional treatments. Keywords: psoriasis, macrophage polarization, biomarkers, bioinformatics analysis Introduction Psoriasis is a chronic, immune-mediated inflammatory skin disease resulting from a complex interplay of genetic predisposition, environmental triggers, and immune dysregulation.[45]^1 The HLA-Cw6 allele is a well-established genetic risk factor that contributes to disease heterogeneity. These elements converge to disrupt immune homeostasis, driving the characteristic pathology of psoriasis.[46]^2 This genetic background interacts with environmental factors to disrupt immune homeostasis, leading to the characteristic pathological features of psoriasis, which include The pathological features of psoriasis include the rapid proliferation and abnormal differentiation of epidermal keratinocytes, alongside dermal blood vessel growth and the infiltration of immune cells such as neutrophils, macrophages, and T lymphocytes.[47]^3 Clinically, psoriasis manifests as erythema and squamous plaques accompanied by pruritus, pain, and psoriatic arthritis. Notably, patients exhibit an increased risk of systemic complications, including metabolic syndrome and cardiovascular diseases, underscoring its systemic inflammatory nature.[48]^4 Although biologic therapies targeting key cytokines, particularly anti-IL-17/IL-23 monoclonal antibodies, have significantly improved clinical outcomes.[49]^5 The persistent challenges, such as treatment resistance and disease recurrence, underscore the incomplete understanding of immune network heterogeneity and therapeutic target diversity.[50]^6 Macrophages, central effector cells of the innate immune system, exhibit remarkable plasticity driven by the immune micro-environment.[51]^7 During psoriatic inflammation, circulating monocytes migrate into lesional skin and differentiate into functionally distinct subsets. Classically activated M1 macrophages, induced by interferon-γ (IFN-γ) and Toll-like Receptor (TLR) ligands, activate NF-κB and STAT1 to drive pro-inflammatory responses (TNF-α, IL-6, IL-12) and pathogen clearance through ROS/NOS2 production.[52]^8^,[53]^9 In contrast, alternatively activated M2 macrophages, polarized by IL-4, IL-13, or IL-10 through STAT6 and Peroxisome Proliferator-Activated Receptor gamma (PPARγ) activation, mediate tissue repair and immune regulation.[54]^10^,[55]^11 Numerous evidence implicate M1 macrophage polarization as a pivotal driver of psoriatic inflammation.[56]^12 In lesional skin, M1 macrophages secrete IL-23, which synergizes with dendritic cells to activate Th17 cells, thereby amplifying IL-17A production. IL-17A directly stimulates keratinocyte hyperproliferation and the release of pro-inflammatory mediators such as S100A8/A9 and CXCL1/8, creating a feedforward loop that recruits neutrophils and T cells while reinforcing M1 polarization.[57]^13^,[58]^14 Furthermore, lipid peroxidation products enriched in the psoriatic microenvironment, such as palmitic acid, exacerbate inflammation by activating the TLR4-NLRP3 inflammasome axis in macrophages, which sustains M1 polarization and suppresses PPARγ-dependent M2 reparative programs.[59]^15 The application of single-cell transcriptomics has recently unveiled the profound heterogeneity within the macrophage compartment in psoriasis, moving beyond the traditional M1/M2 dichotomy. High-dimensional analyses have identified novel, disease-specific macrophage subsets exhibiting distinct transcriptional programs and functional states. For example, pro-inflammatory M1-like subsets in lesions, which exhibit a strong capacity to activate T cells via TNF-α and IL-23 secretion, are driven by m^6A modification-mediated metabolic reprogramming.[60]^16 Refractory lesions are enriched with distinct M1 subsets characterized by CCL20 and STAT1 overexpression, which may evade biological therapies by maintaining IL-23/IL-17 axis activity.[61]^17 These findings highlight the potential value of further investigating macrophage polarization. However, despite the well-documented association between broad interferon-stimulated gene (ISG) expression and a general M1 macrophage phenotype, the existence of a specific, coordinated ISG signature that drives M1 polarization within the unique inflammatory microenvironment of human psoriasis remains largely unexplored. Therefore, our study aims to identify and define such ISG signature and its role in psoriatic inflammation through an integrated bioinformatics and experimental validation approach. This work will provide novel insights into the mechanisms underlying macrophage-mediated psoriatic inflammation and may identify new biomarkers for patient stratification. The study design is summarized in [62]Figure 1. Figure 1. [63]Figure 1 [64]Open in a new tab The flow diagram of this study. Abbreviations: DEGs, differentially expressed genes; WGCNA, weighted gene co-expression network analysis; ROC, Receiver Operating Characteristic; GSEA, Gene Set Enrichment Analysis; EPC, Edge Prediction Centrality; MCC, Matthews Correlation Coefficient; MNC, Modular Network Community. Material and Methods Data Acquisition and Processing Three psoriasis-related datasets ([65]GSE14905, [66]GSE30999, and [67]GSE151177) were obtained from the GEO database ([68]https://www.ncbi.nlm.nih.gov/geo/). [69]GSE14905 and [70]GSE30999 are based on the [71]GPL570 platform for gene expression profiling. [72]GSE151177 is based on the [73]GPL18573 platform for single-cell RNA sequencing. In the [74]GSE14905 dataset, there are 21 non-lesional skin samples from healthy donors and 33 lesional skin samples from individuals with psoriasis. The [75]GSE30999 dataset comprises matched lesional and non-lesional tissue samples from 85 psoriasis patients. The [76]GSE151177 dataset comprises 5 samples of healthy skin and 13 samples of psoriasis lesions for single-cell RNA sequencing analysis. The raw data underwent the following preprocessing steps. Probe IDs were mapped to gene symbols using platform annotation files. The K-nearest neighbors algorithm was applied to impute missing values. The ComBat algorithm was used to harmonize batch effects across datasets. Data were log10-transformed, quantile-normalized, and z-score standardized. After processing, the matrices were employed for further bioinformatics analysis. Identification of Differentially Expressed Genes (DEGs) Statistical analyses were performed in R (v4.2.3). Differential expression analysis of the dataset was conducted using the limma package, with DEGs defined by |log[2]FC| > 1 and a false discovery rate (FDR)-adjusted P-value < 0.05 using the Benjamini-Hochberg method. Visualization was performed with the ggplot2 package, including volcano plots to highlight significance levels and hierarchical clustering heatmaps to display expression patterns. Immune Infiltration Analysis The CIBERSORT algorithm was employed to quantify immune cell infiltration using the LM22 signature matrix.[77]^18 Using the CIBERSORT package, we assessed the proportions of 22 immune cells in the gene expression profiles of the [78]GSE14905 dataset. The differences in immune cell proportions between psoriasis patients and healthy individuals were displayed using ggplot2. Stacked bar plots depicted the overall immune landscape, while violin plots highlighted statistically significant differences (P < 0.05) in specific cell types. Weighted Gene Co-Expression Network Analysis and Module Gene Identification Weighted gene co-expression network analysis (WGCNA) identified psoriasis-associated gene modules from the top 5,000 genes with the highest median absolute deviation. A soft threshold (β) was determined via the pickSoft function, followed by topological overlap matrix construction to minimize spurious correlations. Hierarchical clustering with dynamic tree-cutting defined co-expression modules, with those strongly correlated to psoriasis (MM > 0.8, GS > 0.5) selected for further analysis. Gene Set Enrichment Analysis (GSEA) MPR genes were rank-ordered by differential expression levels between normal and psoriasis groups to generate a sorted gene list. The enrichment score (ES) for the predefined MPR gene set was derived along this list, yielding the running enrichment score (RES) curve. Significance was assessed via a permutation test to obtain P-values, which were adjusted for multiple testing by Benjamini-Hochberg FDR, resulting in adjusted P-values. This evaluated coordinated distribution shifts of the MPR gene set within the ranked list. For each sample, a single-sample GSEA Score was calculated independently, quantifying the level of the MPR gene set per sample. Identification of Common Genes Associated with Macrophages A list of MPR genes was obtained from the Rummagene database (version 4.14, [79]https://rummagene.com/), using the search term of macrophage polarization. Intersection of DEGs, MPR genes, and essential module genes from WGCNA identified shared polarization-associated genes, visualized via a Venn Diagram. Subsequently, functional enrichment analysis encompassing Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) was carried out to evaluate the functions and pathways related to the overlapping genes.[80]^19 Functional enrichment analysis (GO/KEGG) was conducted using clusterProfiler, with results visualized via ggplot2. Protein-Protein Interaction (PPI) Network Construction The common genes were submitted to the STRING database ([81]https://string-db.org)[82]^20 to predict functional interactions among proteins and develop PPI networks using a confidence threshold of PPI score ≥ 0.4. Subsequently, the data from the network were imported into the Cytoscape software, in which six distinct algorithms (Degree, Closeness, Stress, EPC, MCC, and MNC) were utilized to assess the connectivity level of the MPR DEGs.[83]^21 Hub genes were determined as the overlapping top 6 genes across these algorithms. ROC Analysis of Hub Genes Hub gene diagnostic efficiency was assessed via ROC curves utilizing the “pROC” package, and validation was performed on the [84]GSE30999 dataset. The expression levels of these hub genes between the normal and psoriasis groups were presented using violin plots created by the “ggpubr” package. Single-Cell Sequencing Analysis Single-cell RNA sequence analysis was performed using [85]GSE151177 to assess hub gene expression patterns. The integration of the data, calculation of mitochondrial and erythrocyte ratios, and filtering of the dataset were performed using the “Seurat” package, applying criteria that specified the number of detected genes should be between 100 and 7000, with mitochondrial content remaining below 20% and erythrocyte content under 5%. The dataset underwent standardization and normalization through the use of “Harmony” for integration and de-batching, which was subsequently followed by downscaling for clustering via t-distributed Stochastic Neighbor Embedding (t-SNE). The clusters were then annotated utilizing the CellMarker[86]^22 and PanglaoDB databases,[87]^23 along with the original publication linked to the dataset.[88]^24 The visualization and display of average expression levels of hub genes were carried out using the “FeaturePlot” package. Cell Cultures and Treatment RAW264.7 murine macrophages, obtained from the American Type Culture Collection (ATCC, Manassas, VA, USA), were cultured in Dulbecco’s modified Eagle’s medium (DMEM; Gibco) supplemented with 10% fetal bovine serum (FBS; Hyclone), 2 mM L-glutamine, 100 U/mL penicillin, and 100 μg/mL streptomycin. The cells were maintained at 37 °C in a humidified incubator with 5% CO[2]. For experimental treatments, cells were stimulated with LPS (1 μg/mL; Sigma-Aldrich) for 12 hours before lysis and subsequent real-time quantitative PCR (qPCR) analysis. Imiquimod (IMQ)-Induced Psoriatic Mouse Model Female BALB/c mice (8-week-old), obtained from the Model Animal Research Center of Kunming Medical University, were allocated to random groups and housed in SPF conditions, experiencing a 12-hour light/dark cycle, with free access to both food and water. All procedures involving animals were approved by the Institutional Animal Ethics Committee at Kunming Medical University. Psoriasis was induced in mice by daily topical application of 5% IMQ cream (62.5 mg) on shaved backs for 6 days, with Vaseline-treated controls. Disease severity was assessed using the Psoriasis Area and Severity Index (PASI). Mice were euthanized by cervical dislocation, and spleen index was calculated as spleen weight (mg) / body weight (g). Hematoxylin-Eosin (H&E) Staining H&E staining was employed to evaluate skin damage in the mouse model. Skin tissue underwent fixation in 4% paraformaldehyde, followed by dehydration using ethanol and embedding in paraffin. The tissue was sectioned into 4 μm slices, stained with hematoxylin and eosin, and scrutinized under a microscope. Subsequently, the sections were dehydrated, sealed, and examined microscopically. RNA Extraction and Real-Time Quantitative PCR RNA isolation from skin tissues was carried out with Trizol reagent (Invitrogen, USA). Subsequently, the RNA underwent reverse transcription to synthesize cDNA utilizing a reverse transcription reagent (Takara, RR820A). For real-time qPCR, fluorescent labeling was conducted using SYBR Green (Takara, RR047A). The expression levels of the hub genes relative to ACTB were determined through the 2-ΔΔCt method. [89]Table 1 contains the detailed sequences of the primers used. Table 1. Primer Sequences for Quantitative Real-Time Amplification Gene Forward (5′ to 3′) Reverse (3′ to 5′) ISG15 GGTGTCCGTGACTAACTCCAT TGGAAAGGGTAAGACCGTCCT RSAD2 TGCTGGCTGAGAATAGCATTAGG GCTGAGTGCTGTTCCCATCT IFIT3 CCTACATAAAGCACCTAGATGGC ATGTGATAGTAGATCCAGGCGT IFIT1 CTGAGATGTCACTTCACATGGAA GTGCATCCCCAATGGGTTCT GBP1 ACAACTCAGCTAACTTTGTGGG TGATACACAGGCGAGGCATATTA OASL CCATTGTGCCTGCCTACAGAG CTTCAGCTTAGTTGGCCGATG ACTB GGCTACAGCTTCACCACCACAG GGAACCGCTCGTTGCCAATAGT [90]Open in a new tab Abbreviations: MPR, Macrophage Polarization-Related; GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; GSEA, Gene Set Enrichment Analysis; PPI, Protein-Protein Interaction; ROC, Receiver Operating Characteristic; AUC, Area Under the Curve; IMQ, Imiquimod; PASI, Psoriasis Area and Severity Index; H&E, Hematoxylin and Eosin; IFN, Interferon; IL, Interleukin; TNF-α, Tumor Necrosis Factor-alpha; STAT, Signal Transducer and Activator of Transcription; ISGs, interferon-stimulated genes; TLR, Toll-like Receptor; NLRP3, NLR Family Pyrin Domain Containing 3; JAK, Janus Kinase; IRF, Interferon Regulatory Factor. Western Blot (WB) Analysis Proteins from skin tissue were obtained through homogenization in RIPA lysis buffer (Beyotime, China). The protein samples underwent denaturation, were separated using SDS-PAGE, and subsequently transferred onto PVDF membranes. To block the membranes, a solution of 5% non-fat milk in TBST was used for 2 h at room temperature. The membranes were incubated overnight at 4°C with primary antibodies against CD80 (MCE, YA3386; 1:1000) and CD86 (MCE, YA6341; 1:1000). Membranes were treated with HRP-conjugated secondary antibodies for 2 hours at room temperature. Densitometry analysis with ImageJ software (NIH, USA) was utilized to quantify band intensity. Statistical Analysis Statistical analyses were performed using SigmaStat (v3.5). Continuous variables were analyzed using the Wilcoxon rank-sum test or Student’s t-test, selected based on data distribution normality. All statistical tests were two-sided, with P < 0.05 being considered statistically significant. Results Identification of DEGs The [91]GSE14905 dataset was analyzed to identify 820 upregulated and 794 downregulated genes ([92]Figure 2A and [93]Supplementary Table 1), and a heatmap displays expression patterns of the top 25 DEGs ([94]Figure 2B). Immune infiltration analysis using CIBERSORT revealed microenvironmental remodeling in psoriatic lesions. Box plots revealed increased immune cell infiltration in psoriasis, particularly T cells, macrophages, mast cells, and NK cells ([95]Figure 2C). Moreover, the stacked bar plots exhibited 22 immune cell ratios ([96]Figure 2D), and the ratio of M1 macrophages was significantly higher in the psoriasis group. Given the observed significant enrichment of M1 macrophages in psoriatic lesions, we next sought to specifically investigate the molecular signature associated with macrophage polarization. Based on these findings, MPR genes from the Rummagene database were analyzed by GSEA and single-sample GSEA (ssGSEA). Running enrichment score (RES) showed that MPR genes were significantly enriched in the top 5000 genes (P. adjusted =1.625e-08) ([97]Figure 2F). ssGSEA scores were higher in psoriatic lesions than in normal skin, indicating increased macrophage polarization activity in psoriasis ([98]Figure 2E). These results suggest that macrophage polarization plays an important role in psoriasis. Figure 2. [99]Figure 2 [100]Open in a new tab Analysis of DEGs in individuals with psoriasis and assessment of immune infiltration. (A) Volcano plot of DEGs between psoriatic (red) and normal skin (blue). (B) Heatmap of top 30 DEGs in [101]GSE14905. (C) Box plot showing 22 immune cell type proportions. (D) Stacked bar chart of major immune subsets. (E) Heatmap showing MPR gene expression profiles and corresponding ssGSEA score. Red and blue indicate high and low expression levels. (F) The plot demonstrates significant macrophage polarization gene set enrichment, the red line presents the running enrichment score trace, and the black bar indicates the position of the gene in the ranked list. The metric plot below shows upregulation (positive values) and downregulation (negative values) across the ranked dataset. *P < 0.05, **P < 0.01, ***P < 0.001, ns = not significant (p>0.05) compared with normal group. Establishment of WGCNA and Module Gene Identification WGCNA was employed to investigate gene clusters associated with psoriasis in the [102]GSE14905 dataset. After removing duplicates and missing values, we selected the top 5,000 genes based on median absolute deviation for network construction. Sample clustering confirmed data integrity without outliers ([103]Figure 3A). Scale-free topology analysis determined that a soft threshold of 19 was optimal, with an R² value of 0.89 ([104]Figure 3B). With a minimum module size set to 75, hierarchical clustering identified 14 modules of co-expression using this parameter. ([105]Figure 3C). Heatmaps illustrating the relationships between modules and traits evaluate the correlations between each module and psoriasis. ([106]Figure 3D). Blue module (3,230 genes) showed a strong positive association with psoriasis severity (r = 0.91, P = 1×10^−²¹) ([107]Figure 3E), and pink module (333 genes) exhibited a negative correlation (r = −0.82, P = 6×10^−^14) ([108]Figure 3F). These findings were visualized through heatmap analysis, with unassigned genes categorized as gray. Figure 3. [109]Figure 3 [110]Open in a new tab WGCNA identifies functional modules associated with psoriasis. (A) Hierarchical clustering of samples based on gene expression. (B) Determination of optimal soft-thresholding power. (C) Gene clustering and dynamic module identification. (D) Module-trait relationship heatmap, positive correlation (red), negative correlation (blue). (E) Gene significance-module membership correlation in the pink module. (F) Gene significance-module membership correlation in the blue module. Macrophage Polarization-Related Genes and Enrichment Analysis A total of 26 commonly identified genes related to macrophage polarization were identified ([111]Figure 4A). Using GO and KEGG enrichment analysis to investigate the function of the 26 common genes in psoriasis. GO_BP indicated notable enrichment in processes such as the defense response to viral infections, reactions to viruses, and the replication of viral genomes. GO_CC results showed significant enrichment in the secretory granule lumen, cytoplasmic vesicle lumen, and reticulum lumen. GO_MF results highlighted significant enrichment in peptide regulator activity, cytokine receptor binding, and ubiquitin-like protein ligase binding ([112]Figure 4B and [113]Supplementary Table 2). KEGG analysis revealed that the common genes exhibited the highest enrichment in the NOD-like receptor signaling pathway, RIG-I-like receptor signaling pathway, and NF-kappa B signaling pathway, and so on ([114]Figure 4C and [115]Supplementary Table 3), and further visualizes interconnectedness between these pathways in [116]Figure 4D. Figure 4. [117]Figure 4 [118]Open in a new tab Functional annotation and pathway enrichment analysis. (A) Venn diagram showing the overlap of genes from three distinct modules. (B) Top 10 GO pathway enrichment pathways. (C) Top 25 KEGG pathway enrichment pathways. (D) Gene-pathway interaction network of core module genes linked to key signaling pathways. Discovery of Hub Genes for Macrophage Polarization in Psoriasis In the [119]GSE14905 dataset, we explored the functional interactions among 26 MPR genes. The interaction network of these genes contained 26 nodes and 75 edges ([120]Figure 5A). CytoNCA analysis identified IL-1β, ISG15, CXCL8, SERPINA1, and CMPK2 as top centrality genes, ordered by degree centrality value ([121]Figure 5B). Hub genes were determined by intersecting the top six genes from six CytoHubba algorithms ([122]Supplementary Table 4). As shown in [123]Figure 5C, Interferon-Stimulated Gene 15 (ISG15), radical s-adenosyl methionine domain containing (RSAD2), interferon-induced protein with tetratricopeptide repeats 3 (IFIT3), interferon-induced protein with tetratricopeptide repeats 1 (IFIT1), 2′-5′-oligoadenylate synthetase like (OASL), guanylate-binding protein 1 (GBP1) were identified as hub genes. The expression levels of these hub genes were significantly upregulated in psoriasis ([124]Figure 5D). Moreover, the six hub genes showed high AUC values (AUC > 0.9) ([125]Figure 5E). ROC analysis confirmed their diagnostic efficiency with optimal sensitivity and specificity in the [126]GSE30999 dataset ([127]Figure 5F), and hub genes significantly increased in psoriatic skin ([128]Figure 5G), consistent with [129]GSE14905 data. Figure 5. [130]Figure 5 [131]Open in a new tab Hub gene expression and PPI network analysis related to psoriasis. (A) Network of hub genes based on PPI. (B) Gene network with the top 20 degree centrality scores. (C) The UpSet plot shows hub genes from six algorithms, with bar height as intersection size and dots indicating algorithm combinations. (D and E) Violin plots and ROC curves for hub genes related to the dataset [132]GSE14905. (F and G) Violin plots and ROC curves for hub genes associated with the dataset [133]GSE30999. *P < 0.05, **P < 0.01, and ***P < 0.001 compared with normal group. The Biological Function of Hub Genes The results of GSEA analysis of six key genes, ISG15, RSAD2, IFIT3, IFIT1, OASL, and GBP1, revealed the association patterns of these genes with immune-related pathways. These gene sets showed similar enrichment patterns in the sorted dataset, especially in the NOD-like receptor signaling pathway and RIG-I-like receptor signaling pathway, the cAMP signaling pathway, and Cytokine-cytokine receptor interaction showed an obvious enrichment trend ([134]Figure 6A–F). The overall data showed that all six genes had significant associations with immune-related pathways. Figure 6. [135]Figure 6 [136]Open in a new tab Functional annotation and pathway enrichment analysis of hub genes. (A–F) Single-gene GSEA analysis of ISG15, RSAD2, IFIT3, IFIT1, OASL and GBP1. The x-axis represents the ranked position of genes in the ordered dataset, and the y-axis indicates the running enrichment score. The grey curve represents the cumulative distribution of enrichment scores of all differentially expressed genes, and the colored vertical bars indicate the specific positions of genes in each pathway in the sequencing. The Correlation Between Hub Genes and Macrophages To explore the functional relevance of hub genes to immune microenvironment regulation, we analyzed the correlation between hub genes and 22 immune cell subsets using previously generated immune infiltration data. It was found that ISG15 expression was significantly positively correlated with M1 Macrophages (R = 0.37, P = 0.035) and negatively correlated with M2 macrophages (R = −0.35, P = 0.045) ([137]Figure 7A and [138]B). RSAD2 showed positive correlation with M1 macrophages (R=0.36, P=0.04) but no M2 association ([139]Figure 7C and [140]D). IFIT3 correlated positively with M1 (R=0.35, P=0.045) and negatively with M2 macrophages (R = −0.35, P = 0.051) ([141]Figure 7E and [142]F). GBP1 was significantly positively correlated with M1 macrophages (R = 0.41, P = 0.018) and showed no significant correlation with M2 macrophages ([143]Figure 7G and [144]H). IFIT1 and OASL exhibited no significant M1 and M2 macrophage associations. ([145]Figure 7I and [146]L). Figure 7. [147]Figure 7 [148]Open in a new tab Scatter plots demonstrating the correlation between hub genes and macrophage subtypes. Scatter plots showing correlations of (A and B) ISG15, (C and D) RSAD2, (E and F) IFIT3, (G and H) IFIT1, (I and J) OASL, and (K and L) GBP1 with M1 and M2 macrophages. scRNA-Seq Analysis of Psoriasis Hub Genes To preliminarily investigate the cellular microenvironment and assess macrophage abundance in psoriatic lesions, we performed t-SNE visualization on a public scRNA-seq dataset. Following quality control and batch correction, 23 cell clusters were identified by graph-based clustering and further annotated into 8 cell types based on top marker genes, including keratinocytes, T cells, dendritic cells, macrophages, and fibroblasts ([149]Figure 8A and [150]B). Focusing on macrophage subpopulations, we extracted the macrophage subset to reveal expression patterns of six hub genes (ISG15, RSAD2, IFIT3, OASL, GBP1, IFIT1) through t-SNE visualization ([151]Figure 8C). Notably, psoriatic lesions exhibited a higher proportion of cells with elevated expression of these genes compared to normal skin ([152]Figure 8D–I). Among these, ISG15, GBP1, IFIT1, and RSAD2 were significantly upregulated in psoriatic macrophages ([153]Figure 8J–M), while OASL and IFIT3 showed no significant changes ([154]Figure 8N and [155]O). Figure 8. [156]Figure 8 [157]Open in a new tab Single-cell RNA sequence analysis of hub genes in psoriasis. (A) t-SNE visualization illustrates the diverse cell populations identified in normal and psoriasis samples, with each cluster representing a unique cell type. (B) Bubble plots visualize the expression intensity of marker genes in each cell cluster (bubble size represents the proportion of expressing cells, and the depth of color represents the average expression level). (C) t-SNE plots showing the expression intensity of ISG15, RSAD2, IFIT3, OASL, GBP1, and IFIT1 in macrophage, with color intensity indicating expression levels. (D-I) The stacked bar charts exhibited the cell ratio of ISG15, RSAD2, IFIT3, OASL, GBP1, and IFIT1 in macrophages. (J–O) Violin plots contrast the expression of six hub genes in macrophage cells between normal and psoriasis samples. LPS Stimulation Promotes M1 Polarization and Hub Gene Expression in RAW264.7 Macrophages To validate the correlation between the hub genes identified through bioinformatics analysis and M1 macrophage activation, we conducted in vitro cell experiments. RAW264.7 murine macrophage cell line was stimulated with 1 μg/mL LPS to induce polarization toward the M1 phenotype. Compared with the control group, the mRNA expression levels of classical M1 macrophage surface markers CD80 and CD86 were significantly upregulated after LPS stimulation ([158]Figure 9A and [159]B), successfully establishing the M1 polarization model. Based on this model, we found that the mRNA expression levels of previously screened hub genes (ISG15, GBP1, IFIT1, IFIT3, RSAD2, and OASL) were markedly increased in the model group ([160]Figure 9C and [161]H). These results experimentally confirm in vitro that M1 macrophage activation is associated with the activation of the interferon signaling pathway, indicating that the hub genes screened through bioinformatics are reliable indicators of the M1 polarization state. Figure 9. [162]Figure 9 [163]Open in a new tab Expression of M1 markers and hub genes in macrophages after LPS-induced M1 polarization. The relative mRNA expression levels of the genes (A) ISG15, (B) GBP1, (C) IFIT1, (D) IFIT3, (E) RSAD2, (F) OASL, (G) CD80, and (H) CD86 between control and model groups. All data are shown as means ± SEM. IMQ Application Induces Psoriasis-Like Skin Lesions and Inflammatory Responses in Mice 5% IMQ was used to induce a psoriasis-like mouse model. Compared with the control group, model mice exhibited typical psoriasis-like skin lesions, including erythema, scaling, and skin thickening ([164]Figure 10A). PASI scoring revealed significantly elevated erythema scores, increased scaling, and epidermal thickening starting from day 3 ([165]Figure 10B–E), along with a gradual decrease in body weight ([166]Figure 10F). Additionally, splenomegaly was observed in model mice ([167]Figure 10G and [168]H). H&E staining demonstrated marked acanthosis, elongated epidermal ridges, and dense inflammatory infiltration in the dermis of model mice ([169]Figure 10I and [170]J). As expected, IMQ treatment significantly induced the expression of canonical inflammatory cytokines IL-17, IL-1β, TNF-α, IL-22, IL-23, IL-17a, and IL-17f, confirming the successful establishment of a robust inflammatory response ([171]Figure 10K and [172]P). Figure 10. [173]Figure 10 [174]Open in a new tab Phenotypic characteristics and expression of related inflammatory factors in psoriasis mice. (A) The representative manifestations of the back skin. (B–E) skin inflammatory scores, including (B) PASI total score, (C) scales, (D) thickness, and (E) erythema. (F) The weight change. (G) The representative manifestations of the spleen. (H) The spleen index, defined as spleen weight (mg)/body weight (g). (I) The typical H&E staining pictures of tissue slices at 40× magnification. (J) The Baker score of H&E staining. (K–P) the gene expression of key inflammatory factors, including IL-1β, TNF-α, IL-22, IL-23, IL-17α, and IL-17f. All data are shown as means ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 and ns = not significant (p>0.05) compared with control group. Hub Gene Expression in the Psoriasis Model The expression of these six hub genes was analyzed in the IMQ-induced psoriasis model. Compared with controls, ISG15, GBP1, IFIT1, IFIT3, and RSAD2 were significantly upregulated in the model group ([175]Figure 11A–E), while OASL showed no significant difference in expression between the two groups ([176]Figure 11F). Additionally, the mRNA levels of M1 macrophage markers CD80 and CD86 were significantly elevated in psoriatic mice ([177]Figure 11G and [178]H). Corresponding increases in CD80 and CD86 protein expression were further confirmed by Western blot analysis ([179]Figure 11I–K). Figure 11. [180]Figure 11 [181]Open in a new tab Expression of hub genes in the IMQ-induced psoriasis mouse model. Relative mRNA expression levels measured by qPCR for the genes (A) ISG15, (B) GBP1, (C) IFIT1, (D) IFIT3, (E) RSAD2, (F) OASL, (G) CD80, and (H) CD86 between control and model groups. (I) Representative Western blot images of CD80 and CD86. (J and K) Quantification analysis of CD80 and CD86 protein expression levels. All data are shown as means ± SEM, and ns = not significant (p>0.05) compared with the control group. Discussion The skin, a barrier organ, is inhabited by tissue-resident macrophages that are crucial for defense, homeostasis, and tissue repair.[182]^25 In psoriasis lesions, macrophages secrete numerous cytokines like IL-1β, IL-6, and TNF-α. These cytokines activate T cells and drive the abnormal proliferation of keratinocytes, leading to epidermal thickening and scale formation.[183]^12^,[184]^26 While ISGs are well-characterized for their antiviral functions, their specific role in coordinating immune pathologies such as psoriasis remains less explored. Here, using an integrated bioinformatics and experimental validation approach, we identify a novel signature of six ISGs (ISG15, RSAD2, IFIT3, OASL, GBP1, and IFIT1) that are collectively and specifically associated with M1 macrophage polarization in the psoriatic microenvironment—a context distinct from their classical antiviral role. This signature, demonstrating high diagnostic value (AUC > 0.9), may represent a previously underappreciated axis driving inflammation. Functional enrichment analyses revealed their involvement in key innate immune pathways, including NOD-like receptor, RIG-I-like receptor, and NF-κB signaling. Notably, all six genes were enriched in the RIG-I-like receptor pathway, which mediates viral RNA sensing and activates type I IFN responses. Crucially, this persistent type I IFN signaling cascade, evidenced by elevated ISG expression in psoriatic M1 macrophages, emerges as a central driver of their polarization.[185]^27 Moreover, this pathway intersects with the IL-23/IL-17 axis,[186]^28^,[187]^29 potentially enabling IL-17A to amplify M1 polarization by upregulating pro-inflammatory cytokines.[188]^30 The expression of this ISG signature is likely triggered by exposure to endogenous damage-associated molecular patterns (eg, self-nucleic acids) in the lesional microenvironment, which perpetuates JAK-STAT-mediated IFN signaling.[189]^31–33 Each hub gene contributes distinct pathogenic mechanisms through interferon-dependent pathways. In particular, ISG15, as a key interferon-stimulated gene, regulates macrophage polarization through multiple mechanisms. Upon viral infection, bacterial infection, or certain autoinflammatory conditions, the protein levels of ISG15 in macrophages are dramatically elevated.[190]^34 ISG15 may regulate TBK1/IRF3 activity, enhancing pro-inflammatory cytokine production and M1 macrophage polarization.[191]^35^,[192]^36 Furthermore, ISG15 promotes keratinocyte proliferation through HIF-1α signaling, and knockdown of ISG15 inhibits keratinocyte proliferation.[193]^37 GBP1, a core member of the IFN-induced GBP family, activates NLRP3 inflammasome to drive IL-1β-dependent inflammation,[194]^38 and serves as a defining marker of IFN-γ-polarized macrophages.[195]^39^,[196]^40 Additionally, GBP1 expression is undetectable in normal skin but markedly upregulated in the vasculature of inflammatory skin diseases like psoriasis.[197]^41 IFITs (including IFIT1 and IFIT3) are classically recognized for their antiviral roles by binding viral RNAs to restrict replication.[198]^42 IFIT1 and IFIT3 overexpression correlates with M1 macrophage pro-inflammatory and antiviral states.[199]^43 IFIT3 is typically the most highly expressed member within this family and plays a significant regulatory role in the expression and stability of IFIT1.[200]^44 RSAD2, an inducible ISG, mediates innate immune responses through TLR7 and TLR9 signaling to enhance type I interferon production.[201]^45 Notably, RSAD2 has been identified as a key gene in SLE, highlighting its crucial involvement in autoimmune disorders. In summary, our work identifies a signature of six ISGs as key contributors to M1 polarization in psoriasis. Within the characteristic IFN-dominated microenvironment of psoriatic lesions, these genes mediate an ISG response that sustains inflammation. Although type I IFN activation is a common feature in autoimmunity, its specific linkage to macrophage polarization through this coordinated ISG signature may represent a distinctive hallmark of psoriatic pathogenesis. Several limitations of our study warrant consideration. Our reliance on the IMQ-induced mouse model, with its inherent amplification of type I interferon signaling, necessitates caution in extrapolating these findings to all psoriasis subtypes and may lead to an overestimation of this pathway’s contribution. Additionally, while our scRNA-seq analysis provided valuable visualization and hypothesis generation, it lacked the depth of subclustering and differential expression analysis across larger cohorts needed to fully elucidate macrophage subset-specific functions. The most critical constraint is the absence of functional genetic experiments, which prevents definitive conclusions regarding the causative roles of the identified genes in driving polarization, rather than them being a mere consequence of inflammation. To address these limitations, future research should prioritize: (1) employing spatial transcriptomics to delineate the cell-specific expression patterns of this ISG signature within human psoriatic lesions; (2) conducting comprehensive in vitro loss-of-function studies in relevant human macrophage models to establish causality; and (3) correlating the expression of this signature with clinical responses to targeted therapies in longitudinal patient cohorts to assess its utility as a predictive biomarker. Conclusion In conclusion, we identify six interferon-stimulated genes (ISG15, RSAD2, IFIT3, OASL, GBP1, and IFIT1) as potential key regulators associated with M1 macrophage polarization in psoriasis. Beyond their traditional antiviral roles, these genes are co-opted to sustain IFN-driven inflammation, positioning them as promising diagnostic biomarkers and therapeutic targets. Selective inhibition of these genes may disrupt the M1 polarization cascade, offering a novel approach to mitigate psoriatic pathology. To fully assess the clinical implications of targeting this pathway, future studies employing functional assays in relevant cell models and longitudinal cohorts are warranted. Acknowledgments