insitro Completes First AI-Enabled Human Genetics Study of Brown Adipose Tissue, Shares Differentiated Targets with Anti-Obesity Effects

AI-enabled brown adipose tissue (BAT) phenotype unlocks population-scale human genetic discovery

BAT-01 knockdown drives 15% body-weight reduction in preclinical models through peripheral adipose beiging, preserving lean mass

Key findings presented at the Keystone Symposia on Obesity Therapeutics

SOUTH SAN FRANCISCO, Calif.--(BUSINESS WIRE)--insitro, the AI therapeutics company built on causal biology, today reported research demonstrating that artificial intelligence enables population-scale genetic analysis in tissues that have historically been challenging to study in large cohorts. In a first-of-its-kind genome-wide association study (GWAS) of brown adipose tissue (BAT) quantified using computer vision, insitro identified genetic loci associated with BAT biology and subsequently validated a prioritized target, BAT-01, the modulation of which produced significant weight loss in diet-induced obese mice while preserving lean mass potentially through a peripheral mechanism distinct from centrally acting appetite suppressants. The findings were presented by David Lloyd, Ph.D., Senior Vice President, Metabolic Disease and Translational Pharmacology, at the Keystone Symposia on Obesity Therapeutics.

BAT has become increasingly implicated in metabolic health, but its anatomical diffuseness and functional heterogeneity have limited understanding of its genetic regulation in humans.

“To power human genetics, you need tens of thousands of people – but brown fat measurements have historically required specialized approaches such as PET scans, which are hard to acquire at scale,” said Daphne Koller, Ph.D., founder and CEO of insitro. “AI enables these measurements to be derived from broadly available MRIs, unlocking a first-ever analysis of human genetics for BAT. Our results provide strong evidence for the important role that BAT plays across multiple metabolic health outcomes, and potentially reveal novel peripheral mechanisms for fat reduction, orthogonal to the central appetite pathways dominating current obesity therapies.”

Leveraging its ClinML™ platform, insitro developed a machine learning-derived BAT imaging phenotype from UK Biobank Dixon MRI fat-signal fraction maps (n=69,598), using the delta between abdominal and supraclavicular adipose fat-signal fraction, to estimate brown fat content. Before launching the GWAS, the team established biological specificity through multiple validations. The phenotype showed seasonal variation consistent with BAT, with the strongest signal observed during late-winter months – a pattern absent in broader adiposity measures.

Phenome-wide association analyses demonstrated correlations with body composition, lipid profiles, glucose homeostasis, and vascular health aligned with established BAT biology. In addition, a BAT polygenic risk score showed causal associations across multiple cardiometabolic trait categories consistent with BAT’s known metabolic benefits.

Consistent with the novelty of the phenotype, the GWAS identified multiple genes that were unique to this BAT GWAS and were not identified in previous genetic studies of obesity.

Building on these human genetic findings, insitro’s CellML™ platform was used to screen genetically supported targets in primary human adipocytes using high-content imaging, transcriptomics, and functional assays to assess beige/brown-like character and lipid mobilization, prioritizing BAT-01 for in vivo evaluation.

In diet-induced obese mice, BAT-01 knockdown via fat-targeting siRNA produced a 15% reduction in body weight over four weeks alongside a 25% reduction in fat mass and preservation of lean mass, without impact on caloric intake. BAT-01 knockdown also increased Ucp1 and decreased Leptin gene expression in white-adipose depots, findings consistent with induction of a beige-like phenotype.

“This is the difference between discovery driven by AI and human genetics, and discovery driven by trial and error,” said Lloyd. “Starting with scalable human phenotypes and genetic support allows us to move into functional validation with far more confidence and conviction. These preclinical results point to BAT-linked targets that promote fat loss and cardiometabolic health through selective peripheral targeting while avoiding appetite suppression, and will open new paths for differentiated obesity therapies.”

insitro is evaluating additional BAT-linked genes from the GWAS using CellML™ and in vivo studies toward a uniquely differentiated pipeline of targets for obesity and potentially other cardiometabolic diseases.

About insitro

insitro is the physical AI company unlocking causal human biology, founded and led by AI pioneer Daphne Koller. By generating the world's largest integrated multi-modal corpus of human and cellular data, we’ve built the Virtual Human™ – a genetically anchored causal AI engine that reveals how disease begins, progresses, and can be resolved. Our platform enables us to precisely identify causal genetic drivers and deploy our TherML™ AI platform to design optimal medicines, advancing a broad pipeline of therapeutics for neuroscience and metabolic diseases. This industrialized architecture creates a self-learning loop: with every biology we onboard, our predictive models grow smarter, accelerating discovery across scales of biology. Backed by ~$800M in capital from world-class investors like a16z, ARCH, Blackrock, Casdin, CPP, Foresite, GV, Softbank, Temasek, Third Rock, T. Rowe Price – including ~$150M in revenue from partnerships with BMS, Lilly, and Gilead – insitro is rebuilding drug discovery from an unpredictable journey into an industrialized, repeatable process with scalable impact for patients and the world.