Why we are building Outer Bio
Today I'm excited to introduce Outer Bio and the platform we've spent the past six years building: Yuna, a platform that keeps ex vivo human skin alive and measurable for four weeks.
Two things led us here.
The first is discovery. Skin is one of the largest and most visible fields in medicine. It is the organ where aging and disease often appear first, and the one most people actively want to improve. And yet genuinely new solutions are strikingly rare. In thirty years, topical skincare has produced exactly one FDA-approved mechanism for visible skin aging: retinoids. One. The cosmetic side of the industry has fared no better; most products recycle the same limited set of active ingredients and stretch their claims well past the evidence, with little proof they change skin over time. Not for lack of demand, money, or talent. For lack of an instrument that could study human skin as it actually lives–in full, and over time. We came to believe that there is an enormous amount left to discover, and that it is finally within reach.
The second emerged more recently (and may well be the more consequential of the two). The same living-skin data we generate for discovery is exactly what AI-driven biology is starving for. AI has become remarkably good at reasoning about life sciences, but mostly in the few areas where the data to train and check it already exists: protein structure, genomics. Almost everywhere else, models can't generate or otherwise access the data that would guide their own learning and show whether their reasoning is any good. A model can propose; only an experiment can decide. And in human biology, the experiments a model most needs to learn from often can't be run at all.
Both objectives run into the same wall. Take human skin out of the body and it dies in the lab within several days. The biology that determines whether a treatment actually works (collagen remodeling, chronic inflammation, cellular senescence) takes weeks. The field has been trying to read a living tissue from cells that were already dying, with no instrument purpose-built to maintain its experimental viability.
So we spent years building one, and assembled the biologists, AI researchers, and engineers it would take.
Yuna holds full-thickness human skin alive and measurable for four weeks: epidermis, dermis, multiple resident cell types, resident immune signaling intact. It does so in enough detail to be worth training a model on. It is, by design, a data-generation engine, the part of our system that produces clinically relevant human evidence, kept separate from the models that learn from it. We started with skin because it is where this problem is solvable today. It is not where this ends. Every organ needs its own instrument, and they will have to be built one at a time.
Our foundational work is now public. In the preprint we posted on bioRxiv, we put three interventions with known clinical answers through the platform and reported what happened. Sustained IL-17/IL-22 stimulation produced a psoriasis-like state in human skin across three weeks, a chronic process that cannot exist inside the field's three-to-seven-day window. Tofacitinib, a JAK inhibitor used in the clinic, reversed it at every time point. Dasatinib and quercetin attenuated senescence and inflammaging signatures in aged donor tissue. Topical sunscreen mitigated UVB injury at the histological, biochemical, and transcriptional levels. That is living human tissue doing what it does in a person: not a reconstructed model that translates poorly, not a simplistic cell line.
The discovery engine we're building on top of Yuna, and the portfolio coming out of it, are advancing. We’ll be sharing more about it all in the coming weeks.
We plan to work in the open and share our progress. We hope you'll follow along.