NIH backs human-based modeling to map sex-specific hormone biology
The National Institutes of Health is committing $21 million to a new research program designed to improve how hormone biology is modeled, simulated, and ultimately translated into care. The initiative, named the Computational Modeling of Hormone Homeostasis Initiative, centers on a problem that has long complicated therapeutics: male and female physiology does not respond to drugs in identical ways, and hormone-driven variability can alter both benefit and toxicity.
The clinical logic is straightforward. Hormones help regulate far more than reproduction, shaping signaling across the lungs, bones, immune system, and other organs. When that regulatory network shifts during puberty, the menstrual cycle, pregnancy, and menopause, the effects can ripple through multiple systems at once. Traditional mathematical and animal models have not captured those differences with enough fidelity to support sex-specific drug development, and the new NIH program is meant to close that gap with human-based, data-driven computational tools.
Computational tools are being built for hormone homeostasis across the lifespan
The funded projects will advance models that simulate human hormone homeostasis across the lifespan, with an explicit goal of testing drugs and therapeutics for both effectiveness and toxicity. The technical approaches span interactive machine learning, multi-scale models, AI-enhanced multi-scale models, physics-informed models of mechanical stress response and tissue remodeling, and virtual replicas of biological systems known as digital twins. In practical terms, the program is trying to create a better digital stand-in for real human physiology, one that can incorporate sex as a biological variable rather than treat it as a nuisance factor.
That design choice matters because endocrine signaling is not static. Hormone levels rise and fall across a person’s life, and the tissues that receive those signals often respond differently depending on developmental stage and sex-specific biology. A model that ignores those shifts can miss how a therapy behaves in real patients, especially when dosing decisions intersect with changing physiology. By building simulation tools around human biological data, investigators aim to improve the preclinical evaluation of therapies before they reach patients.
Nicole Kleinstreuer, PhD, NIH Deputy Director for Program Coordination, Planning, and Strategic Initiatives, said current mathematical and animal models do not sufficiently capture variations between men and women. She framed the new effort as essential to understanding sex-specific biology and to developing more individualized approaches to therapeutics. Her comments point to a persistent translational bottleneck: when model systems flatten biologic sex differences, the clinic inherits uncertainty about dosing, adverse effects, and response patterns.
Sex as a biological variable moves from concept to computational architecture
The initiative also aligns directly with NIH efforts to account for sex as a biological variable, or SABV. Janine Clayton, MD, FARVO, NIH Associate Director for Research on Women’s Health, described the program as a way to expand understanding of the relationship between hormones and therapeutics using a lifespan perspective. Her view is not simply administrative; it is methodological. If sex-based differences are built into the modeling framework from the outset, the resulting tools may better support drug and biologics testing, and may do so earlier in the development pipeline.
The bedside implications are concrete even though the work remains preclinical. Better models could help identify when a drug is likely to underperform in one sex, when toxicity risk rises because of hormonal context, or when a therapy’s effect changes during reproductive transitions. In endocrine practice, those questions are rarely academic. They influence whether a treatment is dose-adjusted, deferred, or replaced, especially in patients whose hormone milieu is changing rapidly.
The broader scientific ambition is to move beyond coarse comparisons and toward systems-level simulation. Human hormone homeostasis is governed by feedback loops, tissue cross-talk, and mechanical and inflammatory inputs that can differ across sexes and across life stages. Digital twins and physics-informed models are attractive because they can integrate that complexity more naturally than conventional linear frameworks. The hope is not to replace experiments, but to make them smarter.
A federal investment aimed at translational speed, not just theory
Funding for the initiative comes through awards OD042764, OD042817, OD042825, OD042813, OD042820, and OD042823. Those grants support a program that is meant to accelerate bench-to-bedside translation by improving how therapies are evaluated before clinical deployment. The NIH language is unmistakably translational: the goal is better evaluation of new treatments and more informed clinical decisions, not simply more elegant computational math.
A useful way to read the program is as an attempt to widen the evidence base available to researchers before human trials begin in earnest. If models can better approximate sex-specific hormone dynamics, investigators may be able to refine candidate therapies with greater precision and fewer blind spots. That could be especially useful in areas where hormonal state interacts with immune function, tissue remodeling, or organ-level stress responses.
The funding announcement also reflects a larger shift in biomedical research. Computational methods are no longer being asked only to describe biology after the fact; they are being positioned to predict how biology will behave under stress, treatment, and physiologic change. For endocrinology, that means hormone biology is becoming a test case for how AI, mechanistic modeling, and human data can be combined in service of more individualized therapeutics.
The unanswered question is how quickly these models will achieve enough fidelity to influence actual drug development decisions. Their value will depend on whether they can reliably translate complex human endocrine biology into predictions that hold up across ages, reproductive states, and organ systems. For now, the NIH has placed a major bet on the idea that better sex-specific simulation can improve the quality of the therapies that eventually reach patients.