Abstract

   Psoriasis is a chronic inflammatory skin disorder driven by immune
   dysregulation and excessive cell death. Pyroptosis, a form of
   inflammatory programmed cell death, has not been extensively studied in
   the context of psoriasis despite its importance in inflammation. In
   this study, we systematically analyzed the expression of
   pyroptosis-related genes (PRGs) in psoriasis to identify critical
   players involved in disease progression. Using bioinformatics tools and
   publicly available datasets, we constructed a risk score model based on
   machine learning algorithms, which identified several key hub genes
   including CASP1, CASP5, AIM2, GZMB, GZMA, IL1B, and NOD2. The generated
   risk score model demonstrated robust performance in external validation
   datasets, showing strong predictive power for psoriasis severity and
   immune infiltration. High-risk patients exhibited increased
   inflammatory cell infiltration and worsening clinical symptoms, which
   was consistent with the model’s ability to predict immune response
   dynamics in psoriatic lesions. To further validate our findings, we
   analyzed single-cell RNA sequencing data and demonstrated that the risk
   score was highly correlated with immune cell composition, particularly
   DCs, T cells, and mast cells, indicating that patients with higher risk
   scores have more severe disease and stronger immune infiltration.
   Additionally, we targeted CASP1 and CASP5 using CRISPR-Cas9 delivery
   via lipid nanoparticles (LNPs) to selectively knock out these genes in
   keratinocytes, resulting in significant therapeutic effects in the
   IMQ-induced psoriasis mouse model. Our findings provide comprehensive
   insights into the role of pyroptosis in psoriasis and propose a novel
   gene editing strategy for alleviating the disease.

   Keywords: psoriasis, pyroptosis, immune infiltration, risk score, lipid
   nanoparticles, CRISPR-Cas9

Introduction

   Psoriasis is a chronic, immune-mediated inflammatory skin disease
   characterized by excessive keratinocyte proliferation, epidermal
   hyperplasia, and aberrant immune responses ([37]Raharja et al., 2021;
   [38]Lowes et al., 2014; [39]Gudjonsson et al., 2004). The condition is
   marked by persistent lesions and inflammation, often associated with
   significant morbidity. Although the exact pathogenesis of psoriasis
   remains incompletely understood, immune dysregulation plays a central
   role, with T cell activation and cytokine release being critical
   drivers of disease progression ([40]Li et al., 2020). Type 17 immunity,
   characterized by the overproduction of IL-17 and IL-23, is considered a
   hallmark of the disease, further exacerbating skin inflammation and
   contributing to epidermal hyperplasia ([41]Kagami et al., 2010;
   [42]Fitch et al., 2007).

   Current treatments for psoriasis primarily focus on controlling
   symptoms and reducing inflammation. These include topical therapies
   (such as corticosteroids and vitamin D analogs), phototherapy, and
   systemic treatments (such as immunosuppressive drugs, biologics
   targeting IL-17, IL-23, and TNF-α) ([43]Trémezaygues and Reichrath,
   2011; [44]Dogra and Mahajan, 2013; [45]Roenigk et al., 1998). Although
   these treatments are effective in managing symptoms for many patients,
   they have significant limitations. Topical treatments are often only
   effective for mild-to-moderate cases and can cause skin thinning and
   other side effects with prolonged use ([46]Murphy and Reich, 2011).
   Systemic therapies, while effective for severe cases, carry risks of
   immune suppression and other systemic side effects, and biologics are
   expensive and often require regular injections ([47]Lebwohl and Ali,
   2001). Furthermore, many patients develop tolerance or resistance to
   these therapies over time, necessitating the need for more targeted and
   personalized treatments ([48]Heath et al., 2019).

   In recent years, the role of programmed cell death in psoriasis has
   gained increasing attention ([49]Krüger‐Kra et al., 2006). While
   apoptosis and necrosis have long been implicated in the disease,
   pyroptosis, a form of inflammatory cell death driven by inflammasomes,
   has emerged as an important mechanism in immune-mediated skin disorders
   ([50]Kaštelan et al., 2009; [51]Yao et al., 2024). Pyroptosis is a
   highly inflammatory process that leads to the release of
   pro-inflammatory cytokines like IL-1β and IL-18, which are known to
   contribute to the chronic inflammation observed in psoriasis ([52]Li
   and Jiang, 2023). However, the exact role of pyroptosis-related genes
   (PRGs) in psoriasis pathogenesis remains poorly understood, and much of
   the literature to date has focused on other forms of cell death.

   In this study, we aim to investigate the role of pyroptosis-related
   genes in psoriasis through a comprehensive bioinformatics analysis.
   Using publicly available datasets, we systematically examined the
   expression of PRGs in psoriatic lesions and identified key genes
   involved in disease progression. By constructing a risk score model
   based on machine learning algorithms, we developed a predictive tool
   that can assess disease severity and immune infiltration. Furthermore,
   we used single-cell RNA sequencing to validate the robustness of our
   findings at the cellular level, and to explore the immune
   microenvironment of psoriatic skin.

   Building on these insights, we sought to test the therapeutic potential
   of targeting key pyroptosis-related genes, specifically CASP1 and
   CASP5, in the treatment of psoriasis. We employed CRISPR-Cas9 gene
   editing, delivered via lipid nanoparticles (LNPs), to selectively knock
   out Casp1 and Casp11 (murine homolog of human CASP5) in keratinocytes
   within the IMQ-induced psoriasis mouse model. Our results demonstrate
   that targeting these key inflammasome components can significantly
   reduce disease severity, immune cell infiltration, and skin
   hyperplasia, providing a novel therapeutic strategy for managing
   chronic inflammatory diseases like psoriasis.

Materials and methods

Data collection and preprocessing

   The gene expression data for psoriasis were obtained from publicly
   available datasets, including [53]GSE30999 ([54]Suárez-Fariñas et al.,
   2012) and [55]GSE13355 ([56]Nair et al., 2009), through the Gene
   Expression Omnibus (GEO) database. These datasets were preprocessed and
   normalized using the R package limma ([57]Ritchie et al., 2015). An
   independent validation dataset was obtained from GEO dataset
   ([58]GSE41662) and preprocessed using the same normalization
   procedures.

Pyroptosis-related genes (PRGs) identification

   A curated list of pyroptosis-related genes was compiled from the
   literature ([59]Zeng et al., 2023). Differentially expressed genes
   (DEGs) were identified by comparing lesional and normal skin samples,
   using a |log[2] fold change| > 1 and adjusted p-value <0.05 as
   thresholds. The intersection of the DEGs and the PRG list was analyzed.

Differential gene expression and pathway enrichment analysis

   Identification of differentially expressed genes (DEGs) between
   lesional skin and normal skin samples was performed using DESeq2
   ([60]Love et al., 2014). A threshold of |log[2] fold change| > 1 and
   adjusted p-value <0.05 was applied to select significant DEGs. To
   further understand the biological functions and pathways involved, we
   performed Gene Ontology (GO) and Kyoto Encyclopedia of Genes and
   Genomes (KEGG) enrichment analysis using the clusterProfiler package
   ([61]Wu et al., 2021) in R.

Immune microenvironment characterization

   To evaluate the immune landscape associated with pyroptosis activity in
   psoriasis, we applied the CIBERSORT ([62]Chen et al., 2018) algorithm
   to deconvolute the immune cell composition from bulk gene expression
   data. Gene expression matrices from both [63]GSE13355 and [64]GSE30999
   datasets were first normalized using the limma and edgeR packages in R.
   CIBERSORT was then run using the default LM22 immune cell signature
   matrix, which quantifies 22 distinct human immune cell types, including
   T cell subsets, dendritic cells (DCs), macrophages, mast cells, and
   others. We set the number of permutations to 1,000 and filtered output
   to retain samples with CIBERSORT p-values <0.05, indicating
   statistically reliable estimates. Immune cell fractions were compared
   between high- and low-risk score groups using the Wilcoxon rank-sum
   test, and the relationship between individual immune cell types and
   risk score was evaluated using Spearman correlation. The immune
   infiltration profiles were visualized using heatmaps and violin plots
   generated with ggplot2 and ComplexHeatmap packages.

Risk score model construction

   To develop a pyroptosis-related risk score model for psoriasis, we
   began by identifying differentially expressed pyroptosis-related genes
   (PRGs) from lesional versus non-lesional skin using the [65]GSE13355
   dataset. DEGs were selected using a threshold of adjusted p-value <0.05
   and |log2 fold change| > 1, and were then intersected with a curated
   list of 46 PRGs derived from previous literature. The selected PRGs
   were subjected to least absolute shrinkage and selection operator
   (LASSO) regression using the glmnet package in R. Ten-fold
   cross-validation was employed to select the optimal value of the
   regularization parameter lambda that minimized prediction error.
   Patients were then stratified into high- and low-risk groups based on
   the median risk score. The model’s predictive performance was validated
   in the independent dataset, using Receiver Operating Characteristic
   (ROC) curve analysis with the pROC package. Additionally, we evaluated
   the correlation between the risk score and immune infiltration patterns
   (from CIBERSORT) to assess biological relevance.

Single-cell RNA sequencing analysis

   Single-cell RNA-seq data of psoriatic skin samples were downloaded from
   public repositories ([66]GSE228421) ([67]Francis et al., 2024) and
   processed using the Seurat package ([68]Satija et al., 2015) in R. The
   data underwent standard preprocessing steps, including quality control,
   normalization, and scaling. The cells were clustered based on gene
   expression profiles using graph-based clustering. Immune cell subtypes
   were identified using known marker genes. The relationship between risk
   score and immune cell composition was examined by comparing the
   distribution of cell types in high- and low-risk groups. Uniform
   manifold approximation and projection (UMAP) was used for visualization
   of immune cell clusters. Cell-cell communication networks were inferred
   using CellChat ([69]Jin et al., 2021), and interaction strengths were
   compared between risk groups.

Animal model and LNP-mediated CRISPR-Cas9 delivery

   Female C57BL/6 mice (8 weeks old) were used for the IMQ-induced
   psoriasis model. Mice were treated topically with 5% imiquimod (IMQ)
   cream on their ears for 8 consecutive days to induce psoriasis-like
   skin inflammation. The control group was treated with vehicle
   (petrolatum). Lipid nanoparticles (LNPs) were prepared using a standard
   method with DLin-MC3-DMA as the lipid component ([70]Ermilova and
   Swenson, 2020). LNPs were loaded with gRNAs targeting Casp1 and Casp11
   and delivered topically to the ears of IMQ-treated mice on days 1, 3,
   5, and 7 of the treatment regimen. The gRNA sequences targetting Casp1
   and Casp11 were described before ([71]Yu et al., 2024). Ear thickness
   was measured using a digital micrometer at the indicated time points
   (days 0, 2, 4, 6, 8). PASI scores were assessed based on erythema,
   scaling, and thickness using a previously established scoring system
   ([72]Li et al., 2021). Ears were harvested at the end of the treatment
   and fixed in 10% formalin. Tissues were embedded in paraffin,
   sectioned, and stained with H&E for histopathological examination. The
   C57BL/6J mice were obtained from The Jackson Laboratory. Animal care,
   feeding, housing, and enrichment were accomplished in the Shang Hai
   MODEL ORGANISMS. All the surgeries on the mice were carried out under
   isoflurane anesthesia and carbon dioxide euthanasia was used all the
   time to sacrifice the mice. 3 mice of 8 week old were included in each
   group. All animal-related experiments were reviewed and approved by the
   Shanghai Model Organism Ethics Committee.

Statistical analysis

   All statistical analyses were conducted using R (v4.1.2) and Python
   (v3.8). Survival Data are presented as mean ± SD. Statistical
   significance was determined using one-way ANOVA followed by Tukey’s
   post hoc test for comparisons between multiple groups. p-values <0.05
   were considered statistically significant.

Results

Identification and characterization of pyroptosis-related genes (PRGs) in
psoriasis

   To explore the role of pyroptosis in psoriasis, we performed
   differential expression analysis on two independent microarray datasets
   ([73]GSE30999 and [74]GSE13355). As shown in the volcano plots
   ([75]Figure 1A), both datasets revealed widespread transcriptional
   changes between psoriatic and healthy skin, with a large number of
   significantly upregulated genes. Among 46 curated pyroptosis-related
   genes (PRGs), 20 were differentially expressed in at least one dataset,
   and 15 overlapped between both datasets ([76]Figure 1C). These shared
   PRGs—such as IL1B, CASP1, TNF, GSDMD, and AIM2—were predominantly
   upregulated and are known to participate in key inflammatory and cell
   death pathways. Heatmap clustering based on PRG expression profiles
   clearly separated lesional from normal samples in both datasets
   ([77]Figure 1B), Most PRGs were upregulated in psoriasis samples.
   highlighting their diagnostic and biological relevance. Overall, these
   findings indicate that pyroptosis-associated gene dysregulation is a
   reproducible feature of psoriatic lesions, suggesting a potential role
   for pyroptosis in disease pathogenesis.

FIGURE 1.

   [78]A series of images displaying gene expression data analysis. A: Two
   volcano plots for datasets GSE30999 and GSE13355, highlighting
   up-regulated (red) and down-regulated (blue) genes. B: Heatmaps for
   GSE30999 and GSE13355 showing expression levels of genes, with a color
   gradient from blue (low expression) to red (high expression).C: A Venn
   diagram illustrating the overlapping PRG genes between datasets
   GSE30999 and GSE13355, with 1652 genes common to both and distinct
   numbers unique to each dataset.
   [79]Open in a new tab

   Identification and characterization of differentially expressed
   pyroptosis-related genes (PRGs) in psoriasis. (A) Volcano plots
   displaying differentially expressed genes (DEGs) between psoriatic
   lesional skin and normal skin in two publicly available microarray
   datasets: [80]GSE30999 (left) and [81]GSE13355 (right). Red dots
   represent upregulated genes, blue dots represent downregulated genes,
   and gray dots indicate genes without significant changes. DEGs were
   defined using the cutoff of |log[2] fold change| > 1 and adjusted p <
   0.05. Selected PRGs are labeled on the plots. (B) Heatmaps of PRG
   expression in [82]GSE30999 (top) and [83]GSE13355 (bottom). Data were
   normalized and Z-scored by gene across samples. Rows represent
   individual PRGs and columns represent patient samples. Hierarchical
   clustering was applied to both genes and samples using Euclidean
   distance and complete linkage. Red and blue color scales indicate high
   and low expression, respectively. (C) Venn diagram showing the overlap
   of DEGs between the two datasets and the curated list of 46
   pyroptosis-related genes. A total of 1,652 DEGs were shared between
   [84]GSE30999 and [85]GSE13355. Among them, 15 genes overlapped with the
   PRG list and were consistently dysregulated in both datasets.

Functional enrichment reveals immune activation and metabolic suppression in
psoriatic lesions

   To investigate the biological significance of differentially expressed
   genes (DEGs), we performed GO and KEGG enrichment analyses using the
   DEGs shared by both [86]GSE30999 and [87]GSE13355 datasets. GO analysis
   showed that upregulated genes were predominantly enriched in
   immune-related biological processes such as neutrophil activation,
   cytokine-mediated signaling, and regulation of innate immune response
   ([88]Figure 2A, left). In contrast, downregulated genes were associated
   with epidermal differentiation, fatty acid metabolism, and skin barrier
   development ([89]Figure 2A, right). KEGG pathway analysis revealed that
   upregulated genes were significantly involved in inflammatory and
   immune signaling pathways, including IL-17 signaling, TNF signaling,
   NOD-like receptor signaling, and cytokine–cytokine receptor interaction
   ([90]Figure 2B, left). These pathways are well-known to be dysregulated
   in psoriasis and are consistent with its inflammatory nature.
   Conversely, downregulated genes were enriched in metabolic and
   structural pathways such as PPAR signaling, AMPK signaling, Wnt
   signaling, and cytoskeletal regulation ([91]Figure 2B, right),
   indicating impaired metabolic and differentiation programs in psoriatic
   skin. GSEA analysis further demonstrated consistent and significant
   enrichment across both datasets, with notably low adjusted p-values for
   processes such as antimicrobial humoral responses, innate immune
   activation, and defense response to pathogens ([92]Figure 2C).
   Together, these findings support a model in which psoriatic
   inflammation is accompanied by robust upregulation of innate immune
   pathways and suppression of homeostatic epidermal programs.

FIGURE 2.

   [93]Panel A displays circular plots showing enriched Gene Ontology (GO)
   terms for up-regulated and down-regulated genes with categories and log
   fold change. Panel B presents dot plots of KEGG pathway enrichment for
   up-regulated and down-regulated genes, indicating significance and gene
   count. Panel C features ridgeline plots illustrating the distribution
   of various immune-related responses in subjects, with case and control
   comparisons highlighted by color.
   [94]Open in a new tab

   Functional enrichment reveals immune activation and metabolic
   suppression in psoriatic lesions. (A) GO enrichment analysis of
   upregulated (left) and downregulated (right) genes from [95]GSE30999
   and [96]GSE13355. Top enriched biological processes were visualized
   using circular barplots. Upregulated genes were mainly associated with
   immune activation, while downregulated genes were enriched in lipid
   metabolism and epidermal development. (B) KEGG pathway enrichment
   analysis showing the top 10 pathways for upregulated (left) and
   downregulated (right) genes. Upregulated pathways included IL-17, TNF,
   and NOD-like receptor signaling, while downregulated pathways were
   related to metabolism and structural maintenance. (C) GSEA ridge plots
   highlighting immune-related GO terms enriched in psoriatic lesions.
   Curves represent the distribution of genes ranked by expression, with
   color indicating adjusted p-value. Most immune pathways showed positive
   enrichment in lesional samples across both datasets.

Pyroptosis-based risk score model distinguishes psoriatic lesions and
performs well in external validation

   To determine whether pyroptosis-related genes (PRGs) could serve as
   effective biomarkers for psoriasis, we constructed a diagnostic risk
   score model using the datasets. First, we performed LASSO logistic
   regression ([97]Morawska et al., 2024) on the differentially expressed
   PRGs, and identified 8 genes with non-zero coefficients under the
   optimal lambda value determined by ten-fold cross-validation
   ([98]Figures 3A,B). These genes were used to calculate a risk score for
   each sample, which clearly separated lesional (LS) from normal (NL)
   skin ([99]Figure 3C). The risk score was calculated as: Risk score =
   (1.129 × expression of CASP1) + (1.549 × PYCARD) + (0.306 × GZMB) +
   (0.486 × AIM2) + (0.318 × NLRP2) + (1.230 × NOD2) + (0.580 × CASP4) +
   (0.224 × BAK1). Lesional skin (LS) samples showed significantly higher
   risk scores than normal skin (NL). The model demonstrated high
   discriminative power within the training set. To evaluate the
   robustness of this model, we tested it in an independent external
   dataset ([100]GSE41662). All 8 PRGs included in the model—GZMB, AIM2,
   NOD2, CASP1, NLRP2, PYCARD, CASP4, and BAK1—showed significantly
   elevated expression in lesional samples compared to normal controls (p
   < 0.001 for all genes, [101]Figure 3D). ROC curve analysis further
   confirmed the diagnostic value of these genes: most achieved excellent
   classification performance with area under the curve (AUC) values
   exceeding 0.95, including PYCARD (AUC = 0.999), CASP4 (AUC = 0.988),
   and GZMB (AUC = 0.982), while NLRP2 showed moderate predictive value
   (AUC = 0.725) ([102]Figure 3E). Together, these results demonstrate
   that a PRG-based signature can robustly distinguish psoriatic lesions
   from normal skin across independent datasets. The high AUC values
   indicate that these genes not only contribute to disease pathogenesis
   but may also serve as reliable molecular markers for diagnostic
   stratification.

FIGURE 3.

   [103]Five sections of data visualizations labeled A to E. A: Line graph
   showing coefficients vs. log lambda. B: Plot of deviance vs. log
   lambda. C: Risk score curve with samples labeled LS and NL. D: Violin
   plots for genes GZMB, AIM2, NOD2, CASP1, NLRP2, PYCARD, CASP4, BAK1
   showing log expression in red and blue. E: ROC curves for the same
   genes, displaying AUC values ranging from 0.725 to 0.999.
   [104]Open in a new tab

   Pyroptosis-based risk score model distinguishes psoriatic lesions and
   performs well in external validation. (A,B) LASSO logistic regression
   was applied to the training dataset ([105]GSE13355) using expression
   profiles of differentially expressed pyroptosis-related genes (PRGs).
   (A) Coefficient profiles for candidate genes. (B) Ten-fold
   cross-validation identified the optimal lambda value corresponding to
   minimum binomial deviance. (C) Risk scores were calculated based on the
   selected PRGs. Lesional skin (LS) samples showed significantly higher
   risk scores than normal skin (NL), effectively distinguishing the two
   groups. (D) Violin plots showing expression of the selected PRGs in the
   external validation dataset ([106]GSE41662). All genes were
   significantly upregulated in psoriatic lesions (p < 0.001). (E) ROC
   curves evaluating the diagnostic performance of the selected PRGs in
   [107]GSE41662. All genes exhibited high area under the curve (AUC)
   values, indicating strong discriminatory power.

Upregulated PRGs form a dense interaction network

   To investigate the interactions among the 15 upregulated
   pyroptosis-related genes (PRGs) identified in both datasets, we
   constructed a protein–protein interaction (PPI) network using the
   STRING database. The resulting network revealed extensive
   interconnectivity among inflammasome components, caspases, cytokines,
   and effector molecules ([108]Figure 4A). Notably, GZMB and CASP1
   occupied central positions within the network, each displaying multiple
   direct links with other PRGs, suggesting potential roles as core
   regulators of pyroptotic signaling in psoriasis. Hub gene analysis
   using the cytoHubba plugin in Cytoscape further confirmed the
   centrality of GZMB and CASP1 based on degree score. GZMB was connected
   to IL6, TNF, CASP1, IRF1, and GZMA, while CASP1 bridged key components
   of the canonical inflammasome pathway, including AIM2, IL1B, and PYCARD
   ([109]Figure 4B). These findings imply that GZMB and CASP1 may jointly
   orchestrate inflammation and cell death processes in keratinocytes
   during psoriasis progression.

FIGURE 4.

   [110]Diagram A displays a complex network of proteins, including BAK1,
   BAX, IL6, TNF, and others, connected by colored lines representing
   interactions. Diagram B shows a simplified network focusing on GZMB
   connections to proteins like GZMA, CASP1, and IL6, using light blue
   lines.
   [111]Open in a new tab

   Upregulated PRGs form a dense interaction network centered on GZMB and
   CASP1. (A) PPI network of the 15 upregulated PRGs identified in both
   [112]GSE30999 and [113]GSE13355, constructed using the STRING database.
   Edges represent known or predicted protein associations based on
   experimental data, co-expression, databases, and text mining. (B) Hub
   gene analysis using Cytoscape with the cytoHubba plugin identified GZMB
   as the central gene based on degree score. Direct interactions of GZMB
   with other PRGs are shown.

High pyroptosis risk score correlates with increased infiltration of
inflammatory immune cells

   To explore the immunological landscape associated with pyroptosis
   activation, we applied our risk score model to a subset of psoriasis
   patients and observed significantly elevated scores in lesional skin
   compared to normal controls ([114]Figure 5A). Patients were divided
   into five groups (P1–P5) based on ascending risk scores ([115]Figure
   5B), representing increasing pyroptotic level. Immune cell profiling
   using CIBERSORT revealed that patients in high-risk groups (P4 and P5)
   exhibited greater infiltration of proinflammatory cell types, including
   activated dendritic cells, mast cells, CD4 T cells and M2 macrophages
   ([116]Figure 5C). This trend suggests a link between pyroptosis
   activation and local immune dysregulation. Correlation analysis showed
   that several key PRGs—GZMB, CASP5, IL1B, AIM2, and NOD2—were positively
   associated with these inflammatory populations ([117]Figure 5D),
   further implicating pyroptosis in shaping the immune microenvironment
   of psoriatic lesions.

FIGURE 5.

   [118]Panel A displays a scatter plot of risk scores versus sample ID,
   distinguishing two groups labeled L and NL with different colors. Panel
   B shows a similar scatter plot for five samples, P1 to P5, each in a
   unique color. Panel C is a bar chart showing estimated cell type
   proportions for several cell types across three groups. Panel D is a
   scatter plot displaying Pearson correlations for various genes across
   different cell types, with each gene represented by a distinct color.
   [119]Open in a new tab

   High pyroptosis risk score correlates with increased infiltration of
   inflammatory immune cells (A) Risk score distribution in lesional (L)
   and normal (NL) skin samples. Lesional samples showed consistently
   higher scores. (B) Samples were stratified into five groups (P1–P5)
   according to increasing risk scores. (C) Immune cell composition was
   inferred using CIBERSORT. Inflammatory cell types including activated
   dendritic cells, neutrophils, and M1 macrophages were more abundant in
   high-risk groups (P4–P5). (D) Spearman correlation between selected
   PRGs and immune cell types. Proinflammatory genes such as GZMB, IL1B,
   and CASP5 showed positive correlations with inflammatory immune
   subsets.

High pyroptosis risk score is associated with increased infiltration of
inflammatory immune cells in skin tissue at the single-cell level

   To further investigate the cellular composition associated with
   pyroptosis activation in psoriasis, we analyzed a publicly available
   single-cell RNA-seq dataset of psoriatic skin. Patients were stratified
   into high-risk and low-risk groups based on their bulk-tissue
   pyroptosis risk scores, and single-cell data were visualized by UMAP.

   At the broad cell type level, high-risk samples showed a higher
   proportion of skin structural cells (e.g., keratinocytes) and an
   increased proportion of immune cells compared to low-risk samples
   ([120]Figures 6A,B). These reflected the two basic features of skin
   chronic inflammation: Epidermal hyperplasia and immune infiltration. To
   dissect the immune compartment in greater detail, we performed
   subclustering of immune cells. UMAP and composition analysis revealed
   that high-risk lesions were enriched for DCs, T cells and mast cells,
   while low-risk lesions were relatively enriched in macrophages
   ([121]Figures 6C,D). These findings suggest that pyroptosis activation
   correlates with heightened immune cell infiltration and expansion in
   the psoriatic skin microenvironment, supporting the proinflammatory
   nature of the high-risk group.

FIGURE 6.

   [122]Four-part composite image showing UMAP plots and bar charts. Image
   A: UMAP plots comparing high-risk and low-risk samples, with cell types
   distinguished by color. Image B: Bar chart displaying the ratio of
   different cell types in high-risk and low-risk samples. Image C: UMAP
   plots detailing cell lineage distributions for both risk categories.
   Image D: Bar chart showing proportions of cell lineages in high-risk
   versus low-risk samples. Each plot and chart includes a legend
   identifying cell types and lineages by color.
   [123]Open in a new tab

   High pyroptosis risk score is associated with increased infiltration of
   inflammatory immune cells in skin tissue at the single-cell level.
   (A,B) UMAP visualization (A) and cell composition analysis (B) of major
   cell types in high-vs. low-risk psoriatic skin samples, showing
   increased immune, mesenchymal, and endothelial cell fractions in
   high-risk lesions. (C,D) UMAP (C) and barplot (D) of sub-clustered
   immune cell populations. High-risk lesions showed increased
   infiltration of DC lineage cells, macrophages, and proliferating immune
   cells, while low-risk lesions had higher proportions of Langerhans
   cells and mast cells.

High pyroptosis risk score is associated with enhanced immune cell–cell
communication in psoriatic skin

   To explore how pyroptosis activation influences immune cell
   interactions in the psoriatic microenvironment, we inferred
   intercellular communication networks among immune subsets based on
   known ligand-receptor pairs. In low-risk patients, cell-cell
   interactions were relatively sparse and concentrated around DC lineage,
   macrophages, and T cells ([124]Figure 7A). In contrast, high-risk
   patients exhibited a much more interconnected immune network, with
   intensified interactions particularly involving proliferating T cells,
   proliferating DCs, and mast cells ([125]Figure 7B).

FIGURE 7.

   [126]Two network diagrams labeled A and B show interactions between
   cell types. Diagram A, titled "Low-risk," and Diagram B, titled
   "High-risk," both feature nodes for macrophages lineage, Langerhans
   cell, DC lineage, mast cell, proliferating DC cell, T lineage, and
   proliferating T cell. Multicolored lines connect these nodes,
   illustrating different interaction patterns in low and high-risk
   scenarios.
   [127]Open in a new tab

   High pyroptosis risk score is associated with enhanced immune cell–cell
   communication in psoriatic skin (A,B) Predicted intercellular
   communication networks among immune cell subsets in low-risk (A) and
   high-risk (B) psoriatic skin. Edges represent predicted ligand–receptor
   interactions between cell types; edge width reflects interaction
   strength. High-risk samples displayed more extensive and stronger
   immune cell connectivity, particularly involving proliferating T cells,
   DCs, and mast cells.

   These results suggest that pyroptosis activation is not only associated
   with immune infiltration, but may also amplify immune cell crosstalk,
   potentially exacerbating local inflammation and epidermal damage.

Targeted delivery of CRISPR-Cas9 to keratinocytes alleviates psoriasis-like
skin inflammation in vivo

   To explore the therapeutic potential of targeting these hub genes in
   alleviating psoriasis, we used Lipid Nanoparticles (LNPs) to deliver
   gRNAs targeting Casp1 and Casp5 to keratinocytes in the IMQ-induced
   psoriasis mouse model ([128]Figure 8A). Casp1 and Casp5 were selected
   as targets for gene knockout because they are key players in the
   pyroptosis pathway and also enriched in our hub gene analysis. Also,
   AIM2 and NOD2 are key players in pyroptosis (hub genes) identified
   through our analysis in psoriasis. Both of these genes play significant
   roles in activating the inflammasome, with CASP1 being a common
   downstream effector. Upon activation, both AIM2 and NOD2 trigger the
   formation of inflammasomes that ultimately lead to CASP1 activation,
   which processes pro-inflammatory cytokines such as IL-1β, which is also
   a key player in psoriasis through our former analysis. Given this
   critical role of CASP1 in amplifying inflammation, and the fact that
   CASP5 is similarly involved in downstream pyroptotic signaling,
   targeting CASP1 and CASP5 emerges as the most effective strategy to
   mitigate the inflammatory response in psoriasis. So these genes were
   selected because they are critical upstream regulators of pyroptosis,
   making them more likely to have a broad therapeutic impact compared to
   targeting downstream effectors like GZMB.

FIGURE 8.

   [129]Diagram illustrating a study on a mouse psoriasis model. Panel A
   shows lipid nanoparticles (LNP) targeting pyroptosis with Cas9 mRNA and
   gRNAs targeting Casp1 and Casp11. Panel B presents histological
   sections of mouse ear skin for control, IMQ treatment, and IMQ+LNP
   treatment, showing varying thickness. Panel C contains two line graphs
   comparing ear thickness and PASI scores over eight days, with three
   treatments: control (Ctrl), IMQ, and IMQ+LNP. The IMQ treatment shows a
   significant increase in both metrics compared to other groups.
   [130]Open in a new tab

   Targeting pyroptosis in skin using LNP-delivered CRISPR-Cas9 to
   alleviate psoriasis in a mouse model (A) Schematic illustration of
   LNP-mediated CRISPR-Cas9 delivery targeting pyroptosis in skin. LNPs
   are loaded with gRNAs targeting Casp1 and Casp11 to knockout these
   genes in keratinocytes in the psoriatic mouse model. The LNP
   formulation is applied topically to the skin. (B) Representative H&E
   stained sections of mouse ear tissues from the three experimental
   groups: Ctrl, IMQ, and IMQ + LNP-Cas1/11 gRNA. Control group shows
   normal skin architecture, IMQ treatment results in epidermal
   hyperplasia and inflammatory cell infiltration, while the LNP + IMQ
   group shows a reduction in both epidermal thickness and immune cell
   infiltration after Casp1/11 knockout. (C) Ear thickness (left) and PASI
   score (right) measurements over the course of 8 days. The IMQ group
   shows significant thickening and increased PASI score, while the IMQ +
   LNP-Cas1/11 gRNA group exhibits reduced ear thickness and PASI scores
   compared to the IMQ group, indicating the therapeutic potential of
   Casp1/11 knockout. Data represent mean ± SEM, with statistical
   significance indicated by p values (****: p < 0.0001, *: p < 0.05).

   LNPs (lipid nanoparticles) were chosen as the delivery vehicle due to
   their ability to efficiently encapsulate and deliver CRISPR-Cas9
   components to keratinocytes both in animal model and human ([131]Zeng
   et al., 2025; [132]Bolsoni et al., 2023; [133]Lee et al., 2025).
   Keratinocytes are the primary cell type involved in psoriasis
   pathogenesis, and they turnover rapidly, minimizing the risk of
   off-target mutations in long-lived cells. Topical delivery of LNPs to
   the skin allows for targeted delivery to epidermal keratinocytes,
   ensuring high local concentrations of CRISPR-Cas9 components without
   significant systemic exposure.

   Histological analysis of H&E-stained skin sections revealed that IMQ
   treatment caused significant epidermal hyperplasia and inflammatory
   cell infiltration ([134]Figure 8B, middle). In contrast, the IMQ + LNP
   group showed reduced epidermal thickness and fewer infiltrating immune
   cells, indicating that targeting Casp1 and Casp11 (murine homolog of
   CASP5) successfully mitigated the inflammatory response in the skin
   ([135]Figure 8B, right). The control group exhibited normal skin
   structure with no noticeable inflammation ([136]Figure 8B, left).

   Quantitative measurements of ear thickness ([137]Figure 8C, left) and
   PASI scores ([138]Li et al., 2021) ([139]Figure 8C, right) confirmed
   these histological findings. The IMQ group showed progressive increases
   in both ear thickness and PASI scores, consistent with the severe
   inflammation associated with psoriasis. In contrast, the IMQ + LNP
   group exhibited significantly lower ear thickness and PASI scores,
   indicating a marked reduction in disease severity. The control group
   showed no change over time, with consistent baseline values.

   These results demonstrate that CRISPR-Cas9-mediated knockout of Casp1
   and Casp11 in keratinocytes, using LNPs for targeted delivery,
   effectively reduces the inflammatory response and mitigates psoriasis
   symptoms in the mouse psoriasis model, highlighting the potential of
   this strategy as a novel therapeutic approach for inflammatory skin
   diseases.

Discussion

   In this study, we systematically investigated the role of
   pyroptosis-related genes (PRGs) in psoriasis and developed a machine
   learning-based risk score model that accurately predicts disease
   severity and immune infiltration. Our analysis identified CASP1, CASP5,
   AIM2, NOD2, GZMB, GZMA, and IL1B as key hub genes in the inflammatory
   pathways driving psoriasis pathogenesis. Notably, CASP1 and CASP5 were
   highlighted as central downstream effectors of pyroptosis, making them
   attractive therapeutic targets for modulating skin inflammation and
   hyperplasia in psoriasis.

   Our risk score model, based on the expression of these PRGs,
   demonstrated robust predictive power in external validation datasets,
   effectively distinguishing high-risk from low-risk patients. High-risk
   patients, as indicated by the risk score, exhibited significantly more
   immune cell infiltration and worsening disease severity, consistent
   with the characteristic type 17 immunity observed in psoriasis. These
   findings underscore the potential of pyroptosis as a key driver of
   inflammation in psoriasis, offering new insights into the molecular
   mechanisms underlying the disease.

   Single-cell RNA sequencing further validated the robustness of our risk
   score model by revealing a strong correlation between immune cell
   composition and the risk score. High-risk patients showed an enriched
   population of activated dendritic cells (DCs), T cells, and mast cells,
   while macrophages were more prominent in low-risk lesions. This finding
   aligns with previous studies showing that DCs and T cells play a
   pivotal role in psoriasis inflammation, while macrophages are more
   involved in the tissue repair process. The immune microenvironment of
   psoriatic lesions, characterized by heightened immune infiltration,
   appears to be regulated by pyroptosis and inflammasome activation,
   making these pathways critical therapeutic targets.

   One of the most significant findings of this study is the successful
   application of LNP-mediated CRISPR-Cas9 gene editing to target CASP1
   and CASP5 in keratinocytes. Both CASP1 and CASP5 can independently
   induce pyroptosis, and since both genes were identified as key
   regulators in our analysis, we chose to target both to minimize
   functional redundancy. By selectively knocking out these genes, we were
   able to significantly reduce the severity of psoriasis in the
   IMQ-induced mouse model, as evidenced by decreased ear thickness, lower
   PASI scores, and reduced immune cell infiltration. CASP1 has long been
   recognized as a central mediator of IL-1β and IL-18 processing, two
   cytokines that contribute to the chronic inflammation in psoriasis.
   CASP5, although less studied, appears to play a similar role in
   regulating inflammasome-induced pyroptosis, and our results suggest
   that targeting both CASP1 and CASP5 may offer a more comprehensive
   approach to mitigating inflammation.

   The use of LNPs for targeted delivery of CRISPR-Cas9 components to
   keratinocytes offers a promising, non-invasive method for gene editing
   in skin diseases. The topical application of LNPs ensures high local
   concentrations of the CRISPR system without systemic exposure,
   minimizing potential off-target effects. Compared to existing topical
   agents such as corticosteroids or vitamin D analogs, our approach
   offers the advantage of targeting upstream inflammatory machinery
   (e.g., inflammasome components) rather than downstream cytokines.
   Furthermore, unlike biologics that require repeated systemic
   administration and may cause immunosuppression or loss of response over
   time, CRISPR-Cas9-mediated knockout via LNPs is localized, durable, and
   potentially administered at much lower frequency. This strategy
   represents a significant step forward in the development of gene
   therapy for inflammatory skin diseases, offering a novel method for
   modulating cell death pathways and restoring immune homeostasis in
   psoriasis.

   While our study provides compelling evidence for the role of pyroptosis
   in psoriasis and the potential of CASP1 and CASP5 as therapeutic
   targets, several questions remain to be addressed. Although we have
   taken measures to ensure the specificity of our gRNAs and minimize
   off-target effects, the possibility of off-target cleavage remains a
   concern. Methods such as next-generation sequencing (NGS) can be
   employed to assess editing fidelity and identify any unintended genetic
   modifications in the target tissue. Additionally, immune reactions to
   both the LNP carrier and the Cas9 protein may arise ([140]Chew, 2018),
   potentially limiting the safety and efficacy of this approach. Future
   studies should explore the long-term effects of CRISPR-mediated gene
   editing in psoriasis models, particularly with regard to immune
   tolerance and off-target effects ([141]Han et al., 2020). In addition,
   although the gRNAs used for Casp1 and Casp11 knockout were carefully
   designed for high specificity, potential off-target effects remain a
   consideration for translational applications. To mitigate this, future
   studies may incorporate strategies such as high-fidelity Cas9 variants,
   optimized gRNA design algorithms, and LNP formulations with enhanced
   tissue and cell-type specificity to further improve safety and
   precision. Additionally, the broader applicability of our LNP-based
   CRISPR delivery system to other chronic inflammatory diseases and
   autoimmune conditions should be investigated.

   While our findings in the IMQ mouse model are promising, translating
   these results to human clinical applications requires several key
   steps. First, we plan to transition from the murine model to humanized
   skin models, such as human epidermis-on-a-chip or skin grafts, to more
   closely simulate the human immune response and skin physiology. These
   models can be treated with LNP-mediated CRISPR-Cas9 gene editing
   systems to assess delivery efficiency, editing fidelity, and immune
   responses in a human context. Another promising approach involves the
   use of ex vivo human skin organoids, which offer a robust system for
   testing the effects of gene editing on human skin inflammation, cell
   turnover, and immune modulation. These advanced models will provide
   critical insights into the safety, efficacy, and long-term effects of
   our CRISPR-based therapy in human skin, paving the way for clinical
   trials and therapeutic application in psoriasis and other chronic
   inflammatory diseases.

   Recent studies have increasingly highlighted the involvement of
   regulated cell death pathways in psoriasis beyond classical apoptosis.
   For instance, necroptosis and ferroptosis have both been implicated in
   keratinocyte dysfunction and immune activation in psoriatic lesions,
   suggesting that multiple forms of cell death may interact to shape
   disease pathogenesis ([142]Duan et al., 2020; [143]Shou et al., 2021).
   Moreover, the therapeutic potential of modulating inflammasome pathways
   has gained traction, with several studies exploring NLRP3 inhibitors or
   IL-1β blockade in preclinical models of psoriasis and other chronic
   inflammatory diseases ([144]Tsuji et al., 2020). Our work contributes
   to this growing field by integrating bioinformatics-guided target
   discovery with precise, keratinocyte-focused genome editing, offering a
   proof-of-concept for rationally designed, cell-type-specific
   interventions in skin inflammation.

   In conclusion, our study provides new insights into the molecular
   mechanisms of psoriasis by identifying key pyroptosis-related genes and
   demonstrating the potential of targeted gene editing as a therapeutic
   approach. The CRISPR-Cas9/LNP strategy represents a promising tool for
   precision medicine, offering a targeted approach to modulating the
   immune response and cell death pathways in chronic inflammatory
   diseases like psoriasis.

Conclusion

   In this study, we provide a comprehensive analysis of
   pyroptosis-related genes (PRGs) in psoriasis, identifying key molecular
   players such as CASP1, CASP5, AIM2, NOD2, and IL1B that drive the
   inflammatory response and disease progression. Our findings underscore
   the critical role of pyroptosis in psoriasis pathogenesis and
   demonstrate that targeting CASP1 and CASP5 through CRISPR-Cas9 gene
   editing can effectively alleviate disease symptoms and reduce
   inflammation in the IMQ-induced mouse model. By constructing a machine
   learning-based risk score model and validating it across external
   datasets, we established a robust tool for predicting disease severity
   and immune infiltration in psoriasis. Our single-cell RNA sequencing
   analysis further validated the model, highlighting the strong
   association between high-risk scores and increased immune cell
   infiltration, specifically DCs, T cells, and mast cells. The use of
   lipid nanoparticles (LNPs) for targeted CRISPR delivery to
   keratinocytes represents a promising therapeutic strategy for psoriasis
   and other chronic inflammatory skin diseases. This approach offers a
   non-invasive, highly localized method for gene editing, minimizing
   systemic effects and enhancing therapeutic efficacy. Taken together,
   our findings offer new insights into the role of pyroptosis in
   psoriasis and pave the way for gene-based therapies targeting key
   inflammasome components, providing a foundation for the development of
   novel treatments for inflammatory diseases.

Funding Statement

   The author(s) declare that no financial support was received for the
   research and/or publication of this article.

Data availability statement

   The original contributions presented in the study are included in the
   article/supplementary material, further inquiries can be directed to
   the corresponding author.

Ethics statement

   The animal study was approved by Shanghai Model Organism Ethics
   Committee. The study was conducted in accordance with the local
   legislation and institutional requirements.

Author contributions

   GX: Data curation, Formal Analysis, Investigation, Methodology, Project
   administration, Resources, Software, Validation, Visualization, Writing
   – original draft, Writing – review and editing. GM: Data curation,
   Formal Analysis, Investigation, Methodology, Software, Validation,
   Writing – original draft, Writing – review and editing. JaS:
   Investigation, Methodology, Software, Writing – original draft, Writing
   – review and editing. XY: Methodology, Software, Writing – original
   draft, Writing – review and editing. JeS: Methodology, Software,
   Writing – original draft, Writing – review and editing. BG:
   Conceptualization, Data curation, Investigation, Methodology, Project
   administration, Software, Supervision, Writing – original draft,
   Writing – review and editing.

Conflict of interest

   The authors declare that the research was conducted in the absence of
   any commercial or financial relationships that could be construed as a
   potential conflict of interest.

Generative AI statement

   The author(s) declare that Generative AI was used in the creation of
   this manuscript. Generative AI was used: GPT-4 for language refinement
   and polishing of the manuscript.

Publisher’s note

   All claims expressed in this article are solely those of the authors
   and do not necessarily represent those of their affiliated
   organizations, or those of the publisher, the editors and the
   reviewers. Any product that may be evaluated in this article, or claim
   that may be made by its manufacturer, is not guaranteed or endorsed by
   the publisher.

References