Lady Gaga-backed biotech firm uses living human skin and AI to find new skincare ingredients

A biotechnology company co-founded by Lady Gaga is using living human skin and artificial intelligence to search for new skincare ingredients. Outer Biosciences, led by Gaga's fiancé Michael Polansky, has developed a laboratory system that can keep donated, full-thickness human skin alive outside the body for more than 30 days. The company then uses machine learning to identify chemicals that might affect processes such as skin ageing, inflammation and repair, before testing the strongest candidates on the living tissue.

Gaga, whose legal name is Stefani Germanotta, sits on the company's board alongside Polansky and investor Sara Choi. Her involvement was first made public in a Rolling Stone interview last year, in which Polansky said the company had been her idea. Outer Biosciences has deliberately kept a low profile because it does not intend to sell skincare directly to consumers. The company instead plans to work with cosmetics businesses, ingredient manufacturers and pharmaceutical companies, which could eventually use ingredients discovered through its research in their own products.

One of the main limitations of laboratory skin research is time. Human skin removed from the body has conventionally remained suitable for experiments for roughly a week. That gives researchers a relatively short period in which to observe what happens after an ingredient is applied. Outer Biosciences says its system, called YUNA, can maintain full-thickness human skin for more than 30 days. The tissue retains its epidermis and dermis as well as several resident cell types and immune-related biological activity. That longer window allows researchers to follow slower changes, such as collagen production and breakdown, chronic inflammation, cellular ageing and recovery from damage. The skin can also be exposed to controlled UV radiation, dryness, irritants and physical injury, allowing researchers to test what happens both during damage and while the tissue recovers.

The company sources de-identified adult skin donated for research with informed consent. Most comes from tissue left over after elective surgery that would otherwise be discarded. Its suppliers include hospitals, research centres, biobanks and specialist tissue organisations across North America. Outer Biosciences says it does not take tissue that could be used for transplantation. The company also uses skin from donors of different ages, racial backgrounds and Fitzpatrick skin types, rather than relying on a small number of samples with similar characteristics. Its current tissue bank includes donors aged between 20 and 75 and all six Fitzpatrick skin types.

The artificial intelligence is used to reduce the number of chemicals scientists need to test physically. Outer Biosciences says its machine-learning system can computationally screen about 750 million compounds each year, predicting which are most likely to have a useful biological effect and which might present toxicity problems. Promising candidates are then assessed for practical issues such as whether they can be manufactured before moving into automated cell tests. Only those that perform well at those stages are tested on the living human skin. Each experiment generates biological information that is fed back into the machine-learning model, which is intended to improve subsequent predictions. The company says more than 30,000 data points can be collected for each compound, including information about changes in gene activity, proteins and tissue structure.

Polansky told TechCrunch that the company initially relied on a much slower process of searching scientific literature and testing possible compounds. With machine learning added to the process, he said it is now producing a new candidate roughly every six weeks. Six are currently in development and four appear likely to reach commercialisation.

Outer Biosciences does not plan to manufacture finished creams or serums itself. Its business is to discover and test molecules before licensing or selling successful ingredients to companies that can manufacture them at scale and formulate them into finished products. The company is also offering research services to beauty and pharmaceutical businesses that want to test existing ingredients or investigate how particular compounds behave in human skin. This could be particularly useful to skincare companies trying to move beyond a relatively familiar group of ingredients such as retinoids, vitamin C, niacinamide and hyaluronic acid.

Outer Biosciences wants to find previously unidentified molecules and biological mechanisms rather than simply develop another formulation containing an established cosmetic active. Its research is not confined to cosmetics. Current work covers skin ageing, pigmentation, inflammation, wound healing, the skin barrier and adverse skin reactions caused by medicines. In May, researchers published a preprint describing the use of the long-lasting human skin model to study chronic inflammation and environmental stress. Work using the technology has also been presented at meetings including the Society for Investigative Dermatology and Society of Toxicology. A preprint has not been through the full peer-review process, however, and results produced in isolated human tissue do not replace clinical trials involving people. The system cannot reproduce everything that affects skin in a living person, including circulation, hormones, metabolism and long-term behaviour. Its potential value is earlier in the development process: helping researchers decide which compounds are worth taking into more expensive human studies.

Rather than asking artificial intelligence to design a skincare formula by itself, the company is using software to narrow millions of possibilities down to a manageable number, then asking living human tissue whether those predictions actually work.