For the 125 million people living with psoriasis, current biologic therapies represent a genuine breakthrough — but they require injections or infusions, carry immune suppression risks, and cost tens of thousands of dollars annually. The possibility of a topical small-molecule alternative that matches biologic efficacy would reshape treatment access and safety profiles. New findings from a genome-wide screen suggest that goal may now be within reach.
Using a genome-wide CRISPR knockout screen in primary human epidermal keratinocytes — the skin cells most directly implicated in psoriatic inflammation — researchers systematically identified genes that regulate surface expression of IL-17 receptor A (IL17RA), the central molecular target of leading biologics like ixekizumab and secukinumab. To prioritize thousands of candidate hits, the team deployed VirtualCRISPR, a large-language-model framework trained on functional genomics data, which efficiently surfaced two previously overlooked regulators: ALOX5, the enzyme 5-lipoxygenase, and OXTR, the oxytocin receptor. Both operate through distinct intracellular mechanisms. Critically, topical application of zileuton (an existing ALOX5 inhibitor) and cligosiban (an OXTR antagonist) suppressed imiquimod-induced psoriasiform dermatitis in mice with efficacy comparable to systemic anti-IL17RA antibody treatment.
The finding is notable on several levels. ALOX5 is well-characterized in leukotriene biology, and zileuton is already FDA-approved for asthma — a regulatory head-start that could accelerate dermatological repurposing studies. OXTR's role in keratinocyte biology is far less explored, making it a more speculative but potentially novel target class. The AI-assisted prioritization pipeline itself represents a methodological contribution: coupling language-model screening to wet-lab validation could compress early-stage drug discovery timelines significantly. The main caveat is the reliance on a mouse imiquimod model, which imperfectly recapitulates human psoriasis immunopathology. Human clinical validation remains a critical and distant next step. Still, as a proof-of-concept for AI-augmented functional genomics in dermatology, this is a genuinely promising advance.