Psoriasis is the rare immune-mediated indication where the central pathway is widely agreed on. The IL-23/IL-17 axis is the textbook driver, and five FDA-approved biologics in routine clinical use disrupt it directly: secukinumab, ixekizumab, and brodalumab target IL-17 signalling; guselkumab and tildrakizumab target IL-23 (Hawkes, Yan, Chan & Krueger 2018). Response rates are high. Patients on these agents reach PASI-90 in trial settings at rates that were unimaginable a decade ago.
Which raises the question: where does the next round of psoriasis target work sit?
Inside the Euretos AI Platform, that question has three components. The disease has an established immune driver. It has a well-characterised epithelial response (the keratinocyte programme that produces the visible plaque). And it has a longer list of comorbidities — psoriatic arthritis, cardiovascular risk, metabolic syndrome — that pull the molecular conversation outside the skin. Each of those is a cell-type-resolved question with a different rank order of candidate targets.
The pathway is well-described. IL-23 produced by myeloid antigen-presenting cells in psoriatic skin maintains a feed-forward loop with IL-17-producing T-cell populations, principally Th17 cells and IL-17-producing γδ T-cells. The IL-17 cytokines (predominantly IL-17A and IL-17F) act on keratinocytes, drive a chemokine and antimicrobial-peptide response, and recruit additional inflammatory cells back into the lesion. Reviews of the field place this pathway as the dominant therapeutic axis (Hawkes et al. 2018), and the existing anti-IL-23 and anti-IL-17 biologics are the empirical confirmation.
The platform’s view of psoriasis recovers this pathway — the integrated knowledge graph anchors IL23, IL23R, IL17A, IL17F, IL17RA, RORC and the canonical Th17 gene programme as top-tier evidence. None of this is novel; it is the validation that the platform’s evidence integration matches established biology.
The interesting next-question is the response side. IL-17 acts on keratinocytes; the keratinocyte response is what produces the clinical phenotype. Yet most therapeutic effort has been spent on the upstream cytokines, not on the cells that actually generate the plaque.
A target-discovery query restricted to keratinocyte-expressed genes within the psoriatic-skin context surfaces a different shortlist. The integrated platform view brings together genetic association data (the well-characterised psoriasis GWAS), keratinocyte-specific expression from single-cell skin atlases, and perturbation evidence from keratinocyte cell-line experiments. Candidates that sit at the top of that list are typically downstream of IL-17 signalling — transcription factors, kinases, and chemokines that the keratinocyte uses to amplify the inflammatory response and that an upstream biologic does not always fully suppress.
This is where the platform’s evidence integration earns its weight. A literature-only query for psoriasis will return the IL-23/IL-17 cytokines because that is what the field has written about. A primary-data query that filters for keratinocyte expression returns a different slice — the response-side genes that may matter for the patients who fail or partially respond to the existing biologics.
Psoriasis is associated with multiple systemic comorbidities: joint involvement, cardiovascular disease, metabolic syndrome, mental-health burden (Hawkes et al. 2018). Whether those comorbidities reflect shared underlying biology or downstream consequences of chronic inflammation is a question that single-tissue analysis cannot answer.
The AI Platform’s Indication Selection capability is built for that question. By pivoting from a target back to the cross-disease landscape, a researcher can ask whether a candidate that scores well in psoriatic skin also scores in psoriatic arthritis, in atopic dermatitis, in inflammatory bowel disease, or in the cardiovascular-event signal that runs alongside chronic inflammation. The platform answers that across its full integrated knowledge base of more than 275 public databases.
For the IL-23 axis, the cross-disease answer is already partly in clinic — guselkumab and risankizumab have shown activity beyond psoriasis. For the keratinocyte-response side, the cross-disease pattern is less explored, and there is a real research question about which downstream targets are private to skin and which are shared with other epithelial inflammatory diseases.
The pattern across this case study is consistent with the way Euretos’s translational research customers use the platform: validate the established pathway first (because the platform’s integration must match what the field already knows), then push into the parts of the disease where the cell-type axis or the cross-disease axis surfaces something the literature alone does not. For psoriasis, that is the keratinocyte response and the comorbidity overlap. For other indications, the structure of the question is the same; the answer is different.
The next case study in this series will look at amyotrophic lateral sclerosis, where the cell-type cast is a mix of motor neurons and the surrounding glia, and the genetic anchor is heterogeneous in a way that psoriasis is not.