Outer Biosciences, more commonly known as Outer Bio, has released an ex vivo human-skin platform called Yuna. The company plans to use the platform to generate real human-skin data for training an AI system that will allow the discoveries of new skin bioactives.
Long-term ex vivo skin analysis
Developing and testing new skin treatments remains challenging with animal and human trials still requiring significant time and cost. In vitro techniques are good alternatives, but current skin tissue analysis methods are limited by how long the tissue can remain viable outside the body. Some ex vivo skin models have reported viability for about 3 days. This makes it difficult to study processes such as collagen remodeling, chronic inflammation, and cellular senescence, which can take weeks to develop.
Yuna is designed to address this limitation with an ex vivo tissue support system that holds full-thickness human skin and is designed to preserve its function for more than four weeks. This allows researchers to collect data over a longer period and use AI to analyze how different compounds affect specific skin mechanisms. The resulting data could help identify compounds for further evaluation in skincare applications.
The rise of AI-powered skin research
Outer Biosciences was co-founded in 2020 by Michael Polansky (CEO), Kyung-Jin Jang, Chris Hinojosa, Stanley King, and Stefani Germanotta, better known as Lady Gaga.
Polansky’s vision for the company began from the stalled innovation pace in biological research relative to software development. Instead of animal testing or cell culturing, the company ventured into designing a protocol to preserve human skin sourced from plastic surgeries, nonprofit biobanks, and tissue brokers that would be discarded.
The company has raised $23 million as of August 2026 from investors including Wing Ventures, Initialized Capital, and Hawktail. Using Yuna, Outer Bio says its AI model can screen up to 750 million compounds annually using transcriptomic, biochemical, and histological data to identify compounds for further study.
The IP behind Outer Bio’s skin preservation method
The patent describes an ex vivo skin model system for maintaining skin tissue and analyzing its response to different test agents, forming the basis of Yuna. Two key features of the system help support tissue viability. The first is a porous layer within an insert that holds the skin tissue. The porous layer can use structures such as gyroids, Schwarz structures, or cross-shaped patterns to promote the movement of nutrients, fluids, and other reagents into the tissue while allowing waste to be removed.
The system also maintains the skin at an air-liquid interface, where one side of the tissue is exposed to air while the other is in contact with growth media. The insert is designed to fit the tissue sample and can include a UV-curable adhesive film and threaded components to seal the tissue in place. A compression unit applies pressure to secure the tissue while limiting mechanical damage. Together, these components help maintain the tissue within the system while allowing nutrients and reagents to reach the sample.
The system is being tested, and a recent preprint from Outer Bio reports several experiments evaluating its ability to reproduce skin responses. In one experiment, researchers treated the tissue with IL-17 and IL-22 to model psoriasis-related inflammation. The treatment produced inflammation and skin barrier disruption, while tofacitinib reduced these effects. In another experiment, older donor skin was treated with dasatinib or a combination of dasatinib and quercetin for three weeks, with the treatments associated with reduced markers of skin aging and inflammation. The researchers also exposed the tissue to UV radiation and found that SPF 50 sunscreen reduced the resulting damage, demonstrating a potential use of the model for evaluating topical products.
Mapping skin responses with AI
The patent also describes a framework for analyzing how test agents affect skin tissue. The workflow involves collecting multiple types of measurements, including histology, immunostaining, biochemical assays, clinical sensor data, and gene expression. The inclusion of adipose tissue or adipose-derived cells can also allow researchers to study interactions between the skin and underlying adipose tissue, including pathways associated with inflammation and wound healing.
The resulting datasets can be mapped and scored to establish threshold criteria associated with specific biological pathways and skin characteristics. The patent describes applying machine learning to these datasets for applications such as efficacy and toxicity screening, quantitative comparisons with product benchmarks, virtual screening of test agents, and substantiating product claims.
With sufficient training data, the model can also identify relationships between the structure of an active ingredient and its observed effects on skin. These relationships could be used to build a database linking active ingredients with specific biological effects and support the evaluation of additional compounds.
U.S. Pat. App. Pub. No. 2026/002109, titled “Systems and processes for skin tissue growth and analysis,” was filed on March 6, 2023 and published on January 1, 2026. The listed inventors are Chris Hinojosa, Kyung-Jin Jang, and Stanley King.