This study integrates primary cell CRISPR screening with an AI-based prioritization strategy, providing a scalable paradigm for mechanistic investigation and target discovery in psoriasis, and particularly inspiring systematic exploration of non-canonical signaling pathways in inflammatory diseases.
Literature Overview
This article, 'AI-guided CRISPR screening reveals therapeutic targets in psoriasis,' published in Nature Communications, systematically investigates the use of a genome-wide CRISPR knockout screen in primary human keratinocytes to identify regulators of IL17RA expression, combined with the language model-based AI framework VirtualCRISPR for high-confidence candidate gene prioritization. The authors further validated the functional roles of ALOX5 and OXTR as upstream regulators and demonstrated the therapeutic potential of their inhibitors in mouse models. The study not only uncovers novel therapeutic targets but also establishes an efficient pipeline from functional genomic screening to drug repurposing.Background Knowledge
Psoriasis is a chronic inflammatory skin disease affecting over 125 million people worldwide, with its core pathological mechanism involving aberrant activation of the IL-17/IL-17RA signaling axis. Although clinically used biologics (e.g., anti-IL-17 or anti-IL-17RA antibodies) are effective, they require systemic administration, are costly, and carry immunogenicity risks. Topical treatments such as corticosteroids lead to skin atrophy and rapid tolerance with long-term use. Therefore, developing topically applicable small-molecule drugs is an urgent need.
However, keratinocytes—the primary effector cells in psoriasis—have poorly understood mechanisms regulating IL17RA expression, and the genetic manipulation of primary human keratinocytes is challenging, limiting the application of genome-wide screens. In addition, traditional screening approaches often focus on known pathways, making it difficult to discover truly novel targets. To address this bottleneck, the study introduces the AI model VirtualCRISPR to assess the 'novelty' of screening hits, enabling prioritization of candidates with strong functional support but low prior literature evidence, thereby offering a fresh perspective on psoriasis therapy.
Research Methods and Experiments
The research team performed a genome-wide CRISPR-Cas9 knockout screen in primary human epidermal keratinocytes (HEKa), using FACS to sort the top and bottom 5% of cells based on surface IL17RA expression, followed by MAGeCK analysis to identify regulatory genes. To overcome the low transduction efficiency of HEKa cells, the team optimized infection conditions and avoided chemical enhancers that inhibit proliferation. The screen identified 472 genes that reduce IL17RA expression. Subsequently, the AI model VirtualCRISPR—trained on the BioGRID-ORCS database to predict the likelihood that gene knockout in keratinocytes affects IL17RA expression—was applied. By comparing experimental enrichment scores with AI-predicted probabilities, the researchers identified high-priority, highly novel candidates exhibiting strong experimental enrichment but low AI prediction.
In vitro validation involved CRISPR-Cas9 knockout of ALOX5, OXTR, and CYP3A5 in HEKa cells from independent donors, with flow cytometry confirming significant downregulation of surface IL17RA expression. Multi-omics analyses (proteomics, cytokine secretion) further revealed that knockout of these genes suppressed the release of inflammatory cytokines (e.g., CXCL1, IL-6, IL-1β) and induced differentiation-related programs. In vivo, using an IMQ-induced mouse model of psoriasis, topical application of the FDA-approved ALOX5 inhibitor zileuton and the selective OXTR antagonist cligosiban significantly reduced skin inflammation, with efficacy comparable to systemic anti-IL-17RA antibody treatment.Key Conclusions and Perspectives
Research Significance and Prospects
This study provides a generalizable framework for target discovery in complex inflammatory diseases: integrating functional genomics in primary cells, AI-driven prioritization, and multi-omics validation. This strategy can be widely applied to other skin disorders or chronic inflammatory conditions, accelerating the translation from gene to therapy.
In terms of drug development, the study supports the repurposing of approved drugs (e.g., zileuton) for topical psoriasis treatment, shortening development timelines. Moreover, as a G protein-coupled receptor, OXTR is highly druggable, providing a solid foundation for developing novel small-molecule inhibitors.
Conclusion
This study innovatively integrates AI with genome-wide CRISPR screening to systematically identify ALOX5 and OXTR as novel regulatory nodes controlling IL17RA signaling in primary human keratinocytes. Functional validation demonstrates that inhibition of these targets significantly suppresses inflammatory outputs through cell-autonomous mechanisms. More importantly, topical application of their inhibitors achieves therapeutic efficacy comparable to systemic antibody treatment in mouse models, offering psoriasis patients a potential localized, low-cost, and well-tolerated treatment strategy. This work not only deepens our understanding of psoriasis pathogenesis but also demonstrates an efficient path from functional genomics to clinical translation, laying the foundation for precision therapy in inflammatory skin diseases.
