Abstract
Intoduction
Psoriasis is an immune-mediated inflammatory skin disease involving abnormal keratinocyte activation and immune infiltration, while the key transcriptional regulators underlying epidermal–immune interactions remain insufficiently defined.
Aim
To identify and validate core immune-related biomarkers linked to psoriasis pathogenesis, emphasising transcription factor JUNB, through integrated bioinformatics and experimental approaches.
Material and methods
Bulk transcriptome datasets (GSE14905, GSE66511, GSE154200) were examined to detect differentially expressed genes (DEGs), while three ML approaches (LASSO, random forest, and XGBoost) were used to identify hub genes. Diagnostic performance was evaluated via ROC curves. Immune infiltration patterns were assessed using CIBERSORT. Single-cell RNA-seq (GSE162183) was used for cellular localization. In vitro experiments were conducted in HaCaT cells stimulated with IL-17A, with JUNB knockdown via siRNA, and expression levels of IL-6, IL-22, and IFNG measured by qRT-PCR.
Results
JUNB and S100A7 were consistently upregulated in psoriasis across datasets and showed good diagnostic performance (AUC > 0.85). Gene set enrichment analysis (GSEA) demonstrated JUNB to be enriched in IL-17, NF-κB, and TNF signalling pathways. Immune analysis indicated JUNB was positively correlated with neutrophils and macrophages. Single-cell analysis confirmed JUNB and S100A7 were primarily expressed in keratinocytes. In vitro, IL-17A stimulation was associated with increased JUNB and S100A7 expression, while JUNB knockdown significantly reduced downstream inflammatory cytokines.
Conclusions
These findings indicate that JUNB is closely associated with IL-17–related inflammatory responses and immune infiltration in psoriasis, providing mechanistic context for its involvement in disease pathogenesis and supporting its potential utility as a diagnostic biomarker.
Keywords: psoriasis, JUNB, machine learning, immune infiltration, IL-17A, single-cell RNA-seq
Introduction
Psoriasis is a common, chronic immune-mediated inflammatory skin disease characterized by epidermal hyperproliferation, abnormal keratinocyte differentiation, and prominent immune cell infiltration [1]. Its pathogenesis involves a complex interplay between genetic susceptibility – most notably the HLA-C*06:02 allele – environmental triggers, and dysregulated cytokine networks dominated by the IL-23/Th17 axis [2]. Excessive production of pro-inflammatory mediators, including IL-17A, IL-22, TNF-α, and IFN-γ, drives sustained keratinocyte activation and plaque formation, while also contributing to systemic comorbidities such as psoriatic arthritis and cardiometabolic disorders [3]. Although biologic therapies targeting IL-17, IL-23, and TNF-α have substantially improved disease control, the upstream transcriptional regulators coordinating immune–keratinocyte interactions remain incompletely understood.
JunB, a core component of the activator protein-1 (AP-1) transcription factor complex, plays a critical role in skin homeostasis, inflammatory signalling, and immune regulation. Emerging evidence has implicated JunB in redox-dependent fibroblast senescence [4], tissue-resident memory T-cell differentiation [5], and the development of psoriasis-like skin inflammation through modulation of cytokine and complement pathways [6]. In keratinocytes, disruption of immune regulatory mechanisms has been shown to elevate JunB expression, promoting epidermal hyperproliferation and neutrophil infiltration in psoriatic skin [7]. Conversely, reduced JUNB expression in lesion-derived dermal mesenchymal stem cells has been linked to altered glucose metabolism, impaired proliferative capacity, and aberrant vascular differentiation, suggesting cell-type-specific roles of JunB in psoriasis pathogenesis [8]. Collectively, these findings suggest that JunB may act as a context-dependent regulator of epidermal–immune interactions within the psoriatic skin microenvironment.
Altered JunB expression in psoriatic lesions has previously been reported at the tissue level, including immunohistochemical and transcript-based analyses [9]. However, these earlier studies were largely descriptive and did not systematically address the immune microenvironmental context, cell-type–specific localization, or functional relevance of JunB across multiple biological layers. In parallel, S100A7 is consistently overexpressed in psoriatic lesions, and accumulating evidence from immunohistochemical and functional studies supports its contribution to disease pathogenesis. Nevertheless, the potential relationship between JunB, S100A7, and Th17-associated inflammatory signalling remains largely unexplored.
In this study, we employed an integrated transcriptomic framework combining bulk and single-cell analyses, immune microenvironment profiling, and functional validation to characterize the role of JUNB in psoriasis. Using IL-17A–stimulated HaCaT keratinocytes with targeted JUNB knockdown, we further examined its cellular functional relevance. This systematic approach aimed to elucidate the biological significance of JUNB within the psoriatic immune milieu and to define its potential relevance as a diagnostic-associated molecular feature.
Material and methods
Data download and processing
Three psoriasis-related bulk transcriptome datasets (GSE14905, GSE66511, and GSE154200) were downloaded from the Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/geo/). These datasets included gene expression profiles from psoriatic lesions and matched non-lesional or healthy control skin samples. Samples from healthy or non-lesional skin were assigned to the control group, whereas samples from psoriatic lesions were assigned to the psoriasis group. In addition, a psoriasis single-cell RNA sequencing (scRNA-seq) dataset (GSE162183) was obtained from the GEO database, comprising psoriatic and normal skin tissues at single-cell resolution. This dataset was used for cell-type–specific gene expression analysis of candidate hub genes (Supplementary Table S1). All bulk transcriptome datasets were normalized and log2-transformed prior to analysis. Batch effects across datasets were corrected using the ComBat function implemented in the sva R package when integrating data from different platforms or batches.
Differential expression analysis
Differential expression analysis was independently performed for each bulk transcriptome dataset (GSE14905, GSE66511, and GSE154200) using the limma R package. Linear models were fitted to the normalized gene expression data, followed by empirical Bayes moderation of the standard errors. Genes with an absolute log2 fold change (|log2FC|) greater than 1 and a p-value < 0.05 were defined as differentially expressed genes (DEGs). Overlapping DEGs among the three datasets were identified and used for subsequent analyses.
Functional enrichment analysis and protein–protein interaction (PPI) network construction
Gene Ontology (GO) enrichment analysis, including biological process (BP), cellular component (CC), and molecular function (MF) categories, as well as Kyoto Encyclopaedia of Genes and Genomes (KEGG) pathway analysis, were performed on the identified DEGs using the clusterProfiler R package. Enriched terms and pathways with a false discovery rate (FDR) < 0.05 were considered statistically significant. A protein–protein interaction (PPI) network was constructed using the STRING database (https://string-db.org/) with the interaction confidence score set to 0.900 (highest confidence). Disconnected nodes were removed from the network. The resulting PPI network was visualized and further analysed using Cytoscape software to identify hub genes and key network modules.
Machine learning-based hub gene selection
To identify robust hub genes associated with psoriasis, three machine learning algorithms were applied to the differentially expressed genes (DEGs) derived from the GSE14905 and GSE66511 datasets. Least absolute shrinkage and selection operator (LASSO) logistic regression was performed using the glmnet R package, with ten-fold cross-validation applied to determine the optimal penalty parameter (λ.min). Genes with non-zero coefficients were retained as candidate features. Random forest (RF) analysis was conducted using the randomForest package, and candidate genes were ranked based on the mean decrease in accuracy (%IncMSE). In parallel, extreme gradient boosting (XGBoost) was implemented using the xgboost package to rank gene importance within an ensemble learning framework. Candidate genes identified consistently by all three methods in both datasets were defined as hub genes and selected for subsequent analyses.
GSEA
To explore the potential biological pathways associated with the identified hub genes, gene set enrichment analysis was performed using the clusterProfiler R package. For each hub gene, samples from the bulk transcriptome datasets were stratified into high- and low-expression groups based on the median expression level. The analysis was conducted against the Molecular Signatures Database (MSigDB) version 7.0, including both hallmark and KEGG pathway gene sets. Enrichment scores were calculated to evaluate whether predefined sets of genes showed statistically significant differences across the two groups. Pathways meeting the threshold of adjusted p < 0.05 were regarded as enriched, and the findings were illustrated by enrichment plots and bar graphs.
External validation and ROC curve analysis
The diagnostic capacity of the identified hub genes was validated using an independent bulk transcriptome dataset (GSE154200) obtained from the GEO database. Expression values for each hub gene in the validation dataset were extracted and pre-processed in the same manner as the discovery datasets. Group comparisons between psoriasis and control samples were performed with Student’s t-test, with a threshold of p < 0.05 set for significance.
To further evaluate the discriminatory capacity of the hub genes, receiver operating characteristic (ROC) curves were generated, with the area under the curve (AUC) values computed through the pROC package in R. ROC analyses were performed for each dataset included in the study to evaluate the stability as well as generalizability of these genes.
Immune cell infiltration assessment
Distributions of immune cell subtypes across psoriasis versus control samples were estimated via the CIBERSORT tool (https://cibersort.stanford.edu/) applying the LM22 reference matrix (22 immune cell types) and 1,000 permutations to ensure robust deconvolution. This method supports vector regression to deconvolute the bulk transcriptome expression matrix into immune cell composition profiles based on a reference signature matrix (LM22), which includes 547 gene markers distinguishing 22 human immune cell phenotypes, such as T cells, B cells, plasma cells, and myeloid subsets. The analysis was performed with 1,000 permutations to ensure robust estimation.
The output from CIBERSORT was used to compare the infiltration levels of each immune cell type between groups using the Wilcoxon rank-sum test. In addition, Spearman’s correlation analysis was conducted to evaluate the association between the expression levels of hub genes and the relative abundance of each immune cell type. Statistical significance was defined as p < 0.05.
Processing of single-cell RNA-seq data
The scRNA-seq dataset GSE162183, containing psoriasis and normal skin samples, was obtained from the GEO database. Raw count matrices were analysed using the Seurat R package. Cells with fewer than 100 detected genes, fewer than 500 unique molecular identifiers (UMIs), or greater than 20% mitochondrial gene content were discarded. Doublets were identified and excluded using the DoubletFinder package.
Gene expression data normalization was conducted using the LogNormalize method (scale factor = 10,000), followed by identification of variable genes with Find-VariableFeatures. Cell cycle scores were calculated using CellCycleScoring. Data scaling was performed with the ScaleData function to regress out variations related to mitochondrial and ribosomal gene content as well as cell cycle effects. Dimensionality reduction was performed using principal component analysis (PCA), and batch variation across samples was adjusted using Harmony.
Uniform Manifold Approximation and Projection (UMAP) was applied for dimensionality reduction. Cell clustering was performed using the FindNeighbors and FindClusters functions. Cell annotation was conducted with the SingleR package, with reference marker genes obtained from the CellMarker and PanglaoDB databases and verified using published literature. Marker genes for each cluster were identified using the FindAllMarkers function, and hub gene expression was visualized using the FeaturePlot and DotPlot functions in Seurat.
Cell culture and experimental grouping
Human immortalized keratinocyte HaCaT cells were obtained from American Type Culture Collection (ATCC, C6282) and cultured in Dulbecco’s Modified Eagle Medium (DMEM; Yujia, Cat# C6282C) supplemented with 10% foetal bovine serum (FBS; Yujia, Cat# C6282F) and 1% penicillin/streptomycin (Yujia, Cat#C0222) at 37°C in 5% CO2. Cells were used within a limited passage range to ensure experimental reproducibility. For Experiment 1, cells received recombinant human IL-17A (PeproTech, USA) at 0, 50, and 100 ng/ml for 24 h (G1–G3). For Experiment 2, cells underwent transfection with negative control siRNA (si-NC) or JUNB-targeting siRNA (si-JUNB) (GenePharma, China) for 24 h via Lipofectamine 3000 (Invitrogen, USA) following the supplier’s guidelines (G4–G5). For Experiment 3, cells underwent 24 h transfection with si-NC or si-JUNB, after which they were exposed to 100 ng/ml Il-17A for another 24 h (G6–G8). All experimental treatments were performed in triplicate.
RNA extraction and quantitative real-time PCR (qRT-PCR)
Total RNA was isolated from HaCaT cells in each experimental group via TRIzol reagent (Invitrogen, USA) following the supplier’s protocol. RNA concentration and purity were measured with a NanoDrop 2000 (Thermo Fisher Scientific, USA). cDNA synthesis was conducted with PrimeScript™ RT Master Mix (Takara, Japan) with 500 ng total RNA as input, and quantitative PCR was carried out with TB Green® Premix Ex Taq™ II (Takara, Japan) on an ABI 7500 platform (Applied Biosystems, USA). GAPDH served as the reference gene, and relative mRNA expression levels were determined via the 2–∆∆Ct method. The primer sequences applied in this study are summarized in Supplementary Table S2.
Statistical analysis
Data analysis was conducted via GraphPad Prism 10.4 (GraphPad Software, USA) and R software. Results are expressed as mean ± standard deviation (SD) unless otherwise specified. Pairwise comparisons were assessed via Student’s t-test for normal data or Wilcoxon rank-sum test for non-normal data. Multiple group comparisons were performed through one-way analysis of variance (ANOVA) with Tukey’s post-hoc test. Spearman’s rank correlation was applied to evaluate associations. Multiple testing was corrected with the Benjamini–Hochberg false discovery rate (FDR), with FDR < 0.05 regarded as statistically significant. Similarly, p < 0.05 indicated significance.
Results
Detection of differentially expressed genes in psoriasis
Differential expression analysis was performed independently in two psoriasis bulk transcriptome datasets. In both GSE66511 and GSE14905, numerous genes were differentially expressed between psoriatic lesions and control skin, with a predominance of upregulated genes (Figures 1 A, B). Intersection analysis identified 256 DEGs consistently shared between the two datasets (Figure 1 C). Functional enrichment analysis of the shared DEGs revealed significant enrichment in cell cycle–related and immune-associated pathways, including mitotic nuclear division and defence response to virus in Gene Ontology analysis, as well as IL-17 signalling, cytokine–cytokine receptor interaction, and cell cycle pathways in KEGG analysis (Figures 1 D, E). A protein–protein interaction network constructed from these shared DEGs comprised 117 interconnected genes, with several highly connected nodes observed (Figure 1 F).
Figure 1.
Identification and functional analysis of DEGs in psoriasis. A – Volcano plot displaying DEGs from the GSE66511 dataset. Red/blue dots denote markedly upregulated or downregulated genes, respectively, based on |log2FC| > 1 and p < 0.05. B – Volcano distribution of DEGs within the GSE14905 dataset. C – Venn diagram showing 256 common DEGs identified in both datasets. D – GO enrichment analysis of the intersecting DEGs. E – KEGG pathway mapping of the intersecting DEGs. F – PPI network of the intersecting DEGs constructed using the STRING database. Node colour indicates connectivity (degree), with yellow representing higher connectivity and blue representing lower connectivity
Machine learning-based screening of hub genes and diagnostic performance evaluation
To identify candidate hub genes with diagnostic-associated relevance in psoriasis, three machine learning approaches – LASSO logistic regression, random forest (RF), and extreme gradient boosting (XGBoost) – were applied to the DEGs from two independent discovery datasets (GSE66511 and GSE14905). Across both datasets, each algorithm consistently selected a restricted subset of candidate genes, and intersection analysis across the three models identified JUNB and S100A7 as common hub genes (Figures 2 A–F), indicating stability of feature selection across algorithms and cohorts. External validation using the independent dataset GSE154200 confirmed that JUNB and S100A7 were significantly upregulated in psoriatic lesions compared with controls (Figure 2 G). Receiver operating characteristic (ROC) curve analyses demonstrated reproducible discriminatory performance for both genes in GSE66511 (JUNB AUC = 0.991; S100A7 AUC = 0.973; Figure 2 H), GSE14905 (JUNB AUC = 0.853; S100A7 AUC = 0.899; Figure 2 I), and GSE154200 (JUNB AUC = 0.729; S100A7 AUC = 0.799; Figure 2 J), supporting their potential value as diagnostic-associated molecular signatures rather than standalone clinical classifiers.
Figure 2.
Machine learning-based discovery and verification of hub genes in psoriasis. A – Coefficients of genes generated by the LASSO logistic regression model in the GSE66511 dataset. B – Optimization of the tuning parameter (λ) by ten-fold cross-validation in the LASSO framework, with λ.min representing the value that minimized binomial deviance. C – Variable importance ranking of the top 10 genes from the RF model in GSE66511. D – Feature importance scores of the top genes estimated by the extreme gradient boosting (XGBoost) model in GSE66511. E – Venn diagram showing the intersection of genes identified by the LASSO, RF, and XGBoost models in GSE66511. F – Venn diagram showing the intersection of genes identified by the three models in GSE14905. G – Box plots showing the expression levels of JUNB and S100A7 in psoriasis and control samples from the external validation dataset GSE154200. *p < 0.05 was considered statistically significant H–J – ROC curve analyses evaluating the diagnostic performance of JUNB and S100A7 in GSE66511 (H), GSE14905 (I), and GSE154200 (J). AUC. Data are presented as mean ± SD. *p < 0.05 was considered statistically significant
GSEA of hub genes
To investigate the underlying molecular mechanisms through which hub genes affect psoriasis development, GSEA was performed for JUNB and S100A7 in both GSE14905 and GSE66511 datasets. Across both datasets, JUNB was consistently associated with enrichment in immune pathways such as the IL-17, NF-κB, and TNF signalling pathways (Figures 3 A, C). By contrast, S100A7 was primarily associated with DNA replication and metabolic processes such as galactose and nucleotide metabolism (Figures 3 B, D). These results suggest that JUNB and S100A7 may participate in psoriasis pathogenesis by modulating immune-inflammatory responses and cell proliferation-related pathways.
Figure 3.
GSEA of hub genes JUNB and S100A7 in psoriasis. A – GSEA plots for JUNB in the GSE14905 dataset. B – GSEA plots for S100A7 in the GSE14905 dataset. C – GSEA plots for JUNB in the GSE66511 dataset. D – GSEA plots for S100A7 in the GSE66511 dataset. Pathways were regarded as significantly enriched when FDR < 0.05
Immune infiltration analysis by CIBERSORT
Immune cell composition in psoriatic and control skin samples was estimated using CIBERSORT based on bulk transcriptomic data. In the GSE14905 dataset, psoriasis lesions showed significantly increased proportions of macrophages (M0), resting NK cells, activated CD4 memory T cells, and follicular helper T cells, accompanied by reduced fractions of resting mast cells, activated NK cells, and plasma cells (Figures 4 A, B). Correlation analysis revealed that JUNB expression was inversely associated with regulatory T cells, while S100A7 expression showed positive correlations with multiple pro-inflammatory immune subsets, including CD4 memory T cells, macrophages, activated dendritic cells, and neutrophils.
Figure 4.
Immune infiltration analysis by CIBERSORT in psoriasis. A, B – Relative proportion and intercorrelation among immune cell subtypes in the GSE14905 dataset. C, D – Immune cell abundance and correlations of JUNB/S100A7 with immune subsets in the GSE66511 dataset. Positive versus negative correlations are shown in red and blue, respectively; *p < 0.05, **p < 0.01
Consistent immune infiltration patterns were observed in the GSE66511 dataset, where psoriasis samples exhibited increased activated dendritic cells, M2 macrophages, and neutrophils, together with decreased resting dendritic cells, macrophages M0, and resting mast cells (Figures 4 C, D). In this dataset, both JUNB and S100A7 expression levels were positively correlated with macrophage and neutrophil populations and negatively correlated with resting immune cell subsets.
Collectively, these results demonstrate that the expression of JUNB and S100A7 is closely associated with altered immune cell infiltration patterns in psoriatic lesions across independent datasets.
Single-cell RNA-seq validation and spatial expression mapping
To validate the cellular localization of hub genes within the psoriatic skin microenvironment, we analysed the single-cell RNA-sequencing dataset GSE162183, which comprised 20,147 high-quality cells after quality control. Unsupervised clustering identified 13 transcriptionally distinct clusters that were subsequently annotated into nine major cell types based on canonical markers and reference databases, including keratinocytes, fibroblasts, endothelial cells, smooth muscle cells, CD8+ T cells, monocytes, mast cells, epithelial cells, and melanocytes (Figures 5 A–C).
Figure 5.
Single-cell RNA-seq validation and cell-type localization of JUNB and S100A7 in psoriasis. A – Single-cell UMAP projection from the GSE162183 dataset. B – Cell cluster annotation based on canonical marker expression and SingleR prediction. C – Dot plot showing the expression of representative marker genes across different cell types. Dot size represents the percentage of cells expressing the gene; colour intensity reflects average expression. D – Bar plot comparing the proportions of each cell type between healthy and psoriatic samples. E – UMAP feature plots displaying the expression of hub genes JUNB and S100A7 at the single-cell level
Comparison of cell-type proportions between normal and psoriatic skin revealed an increased representation of CD8+ T cells and monocytes in psoriatic lesions (Figure 5 D), consistent with immune cell enrichment observed in psoriatic lesions.
FeaturePlot and DotPlot analyses demonstrated that both JUNB and S100A7 were predominantly expressed in keratinocytes and epithelial cells, with relatively limited expression in immune cell populations (Figure 5 E). These findings confirm the epidermal localization of JUNB and S100A7 at single-cell resolution and are consistent with their upregulation observed in bulk transcriptomic analyses.
In vitro experimental validation of JUNB function
To examine the responsiveness of JUNB to IL-17A stimulation in keratinocytes, qRT-PCR experiments were performed in HaCaT cells treated with IL-17A and JUNB-targeting siRNA. IL-17A treatment led to a progressive increase in JUNB and S100A7 expression with higher IL-17A concentrations. Compared with unstimulated controls, JUNB expression increased by approximately 1.8-fold at 50 ng/ml IL-17A and reached about a 3.2-fold increase at 100 ng/ml (p < 0.01). Similarly, S100A7 expression increased by approximately twofold at 50 ng/ml and further increased at 100 ng/ml IL-17A (p < 0.01) (Figure 6 A).
Figure 6.
In vitro validation of JUNB function in IL-17A–stimulated HaCaT cells. A – qRT-PCR analysis of JUNB and S100A7 levels in HaCaT cells exposed to IL-17A at 0, 50, and 100 ng/ml for 24 h. B – Knockdown efficiency of JUNB in HaCaT cells transfected with si-JUNB versus si-NC. C – JUNB, IFNG, IL-6, and IL-22 expression levels in NC, IL-17A-treated, and IL-17A + si-JUNB groups. Results are shown as mean ± SD of triplicate experiments. **p < 0.01, ***p < 0.001
Efficient knockdown of JUNB was achieved by siRNA transfection, resulting in a reduction of JUNB mRNA levels exceeding 70% compared with the negative control (p < 0.001) (Figure 6 B). Following IL-17A stimulation, expression levels of inflammatory cytokines, including IFNG, IL-6, and IL-22, were significantly elevated. JUNB silencing partially attenuated the IL-17A–induced upregulation of these cytokines, with reductions of approximately 40–50% relative to IL-17A treatment alone (p < 0.01) (Figure 6 C).
Together, these results indicate that JUNB expression in keratinocytes is responsive to IL-17A stimulation and is associated with the regulation of IL-17A–induced inflammatory gene expression.
Discussion
In this study, we integrated bulk transcriptomics, machine learning, immune deconvolution, single-cell RNA sequencing, and in vitro functional assays to investigate the role of JUNB in psoriasis. Rather than reiterating diagnostic performance, our findings place JUNB within an IL-17-associated keratinocyte–immune interaction context, offering mechanistic context into its potential involvement in psoriatic inflammation.
JUNB, a key component of the AP-1 transcription factor complex, has been implicated in immune regulation, tissue remodelling, and keratinocyte biology in prior studies [10–12]. Earlier work demonstrated that JunB can modulate cutaneous inflammation through cytokine- and complement-related pathways [13], while additional studies highlighted its roles in fibroblast niche regulation [4] and tissue-resident memory T-cell differentiation [14]. Consistent with these observations, our single-cell analysis confirmed that JUNB is predominantly expressed in keratinocytes in psoriatic skin and is closely associated with IL-17–related inflammatory signalling, in line with AP-1–driven Th17 responses reported in psoriasis [15]. By integrating multi-level datasets, our study extends these prior findings by situating JUNB within a broader immune–epidermal network rather than a single cellular context.
S100A7, a well-characterized psoriasin, has been widely linked to epidermal barrier dysfunction and immune cell recruitment [16, 17]. Our results demonstrate that S100A7 expression is closely associated with macrophage-, dendritic cell-, and neutrophil-enriched immune infiltration patterns, supporting previous reports of its involvement in neutrophilic inflammation in psoriatic lesions [18–20]. The parallel immune association profiles of JUNB and S100A7 suggest that these molecules may converge on shared inflammatory circuits within the psoriatic microenvironment.
Importantly, immune deconvolution analysis revealed that JUNB expression was associated with both pro-inflammatory immune subsets (including neutrophils and macrophages) and reduced regulatory immune populations, underscoring its potential relevance to immune imbalance in psoriasis. Functional validation further demonstrated that IL-17A stimulation induced JUNB expression in keratinocytes, while siRNA-mediated JUNB knockdown attenuated the expression of key inflammatory cytokines, including IFNG, IL-6, and IL-22. These findings support a role for JUNB as a transcriptional mediator within IL-17–driven inflammatory responses, although causality at the tissue or organismal level remains to be established.
Despite its strengths, this investigation presents several limitations. First, functional experiments were conducted exclusively in HaCaT cells, which may not fully recapitulate the complexity of primary psoriatic keratinocytes or stromal cell interactions. Second, although single-cell RNA sequencing provided cell-type–specific localization, additional spatial or protein-level validation would further refine tissue-level interpretation. Third, post-transcriptional and epigenetic regulatory mechanisms influencing JUNB expression were not explored. Finally, while our findings provide mechanistic insight, the translational and biological relevance of JUNB-associated regulatory mechanisms requires further validation in in vivo psoriasis models and advanced experimental systems.
Conclusions
This study identified JUNB and S100A7 as epidermal-associated genes consistently upregulated in psoriatic lesions through integrated transcriptomic and machine learning analyses. Multi-level evidence from immune deconvolution, single-cell RNA sequencing, and in vitro assays supports an association between JUNB expression and IL-17A–responsive inflammatory gene regulation in keratinocytes. Together, these findings position JUNB as a mechanistically informative marker linked to epidermal–immune interactions in psoriasis and provide a theoretical framework for future studies investigating its biological and translational relevance.
Supplementary Material
Funding Statement
Funding No external funding.
Ethical approval
Not applicable.
Conflict of interest
The authors declare no conflict of interest.
References
- 1.Boehncke WH, Schön MP. Psoriasis. Lancet 2015; 386: 983–94. [DOI] [PubMed] [Google Scholar]
- 2.Griffiths CEM, Armstrong AW, Gudjonsson JE, Barker JNWN. Psoriasis. Lancet 2021; 397: 1301–15. [DOI] [PubMed] [Google Scholar]
- 3.Rendon A, Schäkel K. Psoriasis pathogenesis and treatment. Int J Mol Sci 2019; 20: 1475. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Maity P, Singh K, Krug L, et al. Persistent JunB activation in fibroblasts disrupts stem cell niche interactions enforcing skin aging. Cell Rep 2021; 36: 109634. [DOI] [PubMed] [Google Scholar]
- 5.Smith NP, Yan Y, Pan Y, et al. Resident memory T cell development is associated with AP-1 transcription factor upregulation across anatomical niches. BioRxiv 2023; 2023.09.29.560006. [Google Scholar]
- 6.Uluçkan Ö, Guinea-Viniegra J, Jimenez M, Wagner EF. Signalling in inflammatory skin disease by AP-1 (Fos/Jun). Clin Exp Rheumatol 2015; 33 (4 Suppl 92): S44–9. [PubMed] [Google Scholar]
- 7.Gou C, Ni W, Ma P, et al. The endoribonuclease N4BP1 prevents psoriasis by controlling both keratinocytes proliferation and neutrophil infiltration. Cell Death Dis 2021; 12: 488. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Zhao X, Xing J, Li J, et al. Dysregulated dermal mesenchymal stem cell proliferation and differentiation interfered by glucose metabolism in psoriasis. Int J Stem Cells 2021; 14: 85–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Park CC, Kim KJ, Eoo SY, et al. Comparison of the expression profile of JunB, c-Jun, and S100A8 (Calgranulin A) in psoriasis vulgaris and guttate psoriasis. Ann Dermatol 2009; 21: 35–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Uluçkan Ö, Jiménez M, Roediger B, et al. Cutaneous immune cell-microbiota interactions are controlled by epidermal JunB/AP-1. Cell Rep 2019; 29: 844–59.e3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ren FJ, Cai XY, Yao Y, Fang GY. JunB: a paradigm for Jun family in immune response and cancer. Front Cell Infect Microbiol 2023; 13: 1222265. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Pi Z, Qiu X, Liu J, et al. Activating protein-1 (AP-1): a promising target for the treatment of fibrotic diseases. Curr Med Chem 2023; doi: 10.2174/0929867330666230209100059. [DOI] [PubMed] [Google Scholar]
- 13.Zhang X, Jin JY, Wu J, et al. RNA-Seq and ChIP-Seq reveal SQSTM1/p62 as a key mediator of JunB suppression of NF-κB-dependent inflammation. J Investig Dermatol 2015; 135: 1016–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Koizumi SI, Sasaki D, Hsieh TH, et al. JunB regulates homeostasis and suppressive functions of effector regulatory T cells. Nat Commun 2018; 9: 5344. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Fu L, Li M, Wang P, et al. Tanshinone IIA, a component of the self-made Xiao-Yin decoction, ameliorates psoriasis by inhibiting IL-17/IL-23 and PTGS2/NF-κB/AP-1 pathways. Skin Res Technol 2024; 30: e13577. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Kurtović M, Pitesa N, Conkas J, et al. GLI transcriptional targets S100A7 and KRT16 show upregulated expression patterns in epidermis overlying the tumor mass in melanoma samples. Int J Mol Sci 2024; 25: 6084. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Granata M, Skarmoutsou E, Gangemi P, et al. S100A7, Jab1, and p27kip1 expression in psoriasis and S100A7 CRISPR-activated human keratinocyte cell line. J Cell Biochem 2019; 120: 3384–92. [DOI] [PubMed] [Google Scholar]
- 18.Dajnoki Z, Somogyi O, Medgyesi B, et al. Primary alterations during the development of hidradenitis suppurativa. J Eur Acad Dermatol Venereol 2022; 36: 462–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Lu Z, Zheng S, Liu C, et al. S100A7 as a potential diagnostic and prognostic biomarker of esophageal squamous cell carcinoma promotes M2 macrophage infiltration and angiogenesis. Clin Transl Med 2021; 11: e459. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Onderdijk AJ, Baerveldt EM, Kurek D, et al. IL-4 downregulates IL-1β and IL-6 and induces GATA3 in psoriatic epidermal cells: route of action of a Th2 cytokine. J Immunol 2015; 195: 1744–52. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.