A new study from Northwestern Medicine investigators, in collaboration with Biohub, a program of the Chan Zuckerberg Initiative, has identified potential drug targets for treating psoriasis, according to the study published in Nature Communications
The study was led by Shana Kelley, PhD, the Neena B. Schwartz Professor of Chemistry, Biomedical Engineering, and Biochemistry and Molecular Genetics, president of Biohub, and Abdalla M. Abdrabou, ’23 PhD, leader of the Functional Immunogenomics group at Biohub
More than 8 million people in the U.S. have psoriasis, according to the National Psoriasis Foundation. The chronic autoimmune disease causes skin cells to multiply up to ten times faster than normal, resulting in thick, red or discolored patches of skin
In the current study, scientists combined genome-wide CRISPR screening with an artificial intelligence tool to uncover druggable targets that had not previously been linked to a psoriasis signaling pathway
“We wanted to understand what first sets psoriasis inflammation in motion,” said Abdrabou, who was a co-corresponding author of the study
The team focused on IL17RA, a receptor known to play a central role in psoriasis inflammation and is already the target of several effective drugs. While current therapies can improve symptoms, they work by dampening the entire immune system and can be expensive or unsuitable for some patients
To search for new therapeutic possibilities, investigators conducted a genome-wide CRISPR knockout screen in human epidermal keratinocytes, the main cell type involved in psoriatic skin lesions
The team then used an AI-based tool called VirtualCRISPR to prioritize the most promising and novel findings from among the hundreds of genes identified in the screen
“The AI offered a prioritization algorithm,” Abdrabou said. “It analyzed patterns across the medical literature and large biological datasets, then organized the genes based on which ones appeared most novel and least explored.”
A major challenge of the study was performing large-scale genetic screens in keratinocytes, said Mushaine Shih, PhD, co-first author of the study and a postdoctoral fellow in the Functional Immunogenomics group at Biohub
“Previously, no one had been able to systematically screen these cells genetically because they’re very hard to screen,” Shih said. The team developed a new centrifugation-based method that allowed them to perform genome-scale CRISPR screening in the cells
The screen identified hundreds of genes that influenced IL17RA expression, but choosing which ones to investigate further posed another challenge. Traditionally, investigators must sift through massive datasets and rely on an extensive literature review to choose genes for further investigation
The AI algorithm helped the team identify two standout targets: ALOX5, already targeted by an FDA-approved asthma drug, and OXTR, the receptor that mediates the effects of oxytocin, which is best known as the “love hormone” for its roles in childbirth and social bonding
Both genes showed little prior connection to IL17RA regulation, making them good candidates for further study
“This study is also a proof-of-concept of how AI can help as a scientist,” said Chenlin Zhao, ’26 PhD, a postdoctoral scholar in the Kelley lab and co-first author of the study. “Previously, it might take weeks, months, to decide which gene to pursue from a CRISPR screening. But now, with AI’s help, I can focus my time and re
The team of scientists then confirmed that disabling the genes in human keratinocytes reduced IL17RA expression and inflammation associated with psoriasis
Next, the team repurposed existing drugs that target the two pathways, reformulating them into a gel that could be tested on a mouse model of psoriasis
The investigators found that the topical treatments significantly reduced psoriasis severity, normalized skin thickness and achieved results comparable to existing therapies, which served as the control
Beyond psoriasis, the investigators say the work demonstrates how AI has the potential to accelerate biomedical research. The team plans to apply the same AI-guided approach to study other diseases, including neuroimmunological and neurodegenerative disorders
“In a sense, we didn’t just build a tool for one skin disease, we created a tool that can be universally applied,” Abdrabou said. “We created a tool that turns years of research into a faster path to discovery, bringing Biohub closer to its mission of curing or preventing all disease through AI-powered biology.”
The study was supported by Biohub and funding from the National Cancer Institute (5R01CA295888) and the National Institute of Allergy and Infectious Diseases (1R01AI181023)
