Researchers Add Nearly 700 Patient-Derived Models to Cancer Research Resources

An international team led by researchers at MIT, the Broad Institute, Dana-Farber Cancer Institute, the National Cancer Institute and other institutions has created nearly 700 new laboratory models derived from human tumors. The models, representing 25 types of cancer, are being made available to researchers worldwide through the American Type Culture Collection.

The work is described in an open-access paper published in Nature and concludes a 10-year effort known as the Human Cancer Models Initiative. The project was funded primarily by the National Cancer Institute and the United Kingdom’s Wellcome Trust.

Researchers began with more than 2,700 tumor samples collected from consenting patients at hospitals in the United States, the United Kingdom and the Netherlands. About one-third of the samples were successfully converted into models that could be maintained in the laboratory. Most came from commonly diagnosed cancers, including lung, liver and pancreatic tumors, while approximately 150 represented rare cancers such as gallbladder and small-intestinal tumors.

Most of the resulting models are organoids: three-dimensional cell cultures grown in specialized media and a scaffold that supports their development. Unlike traditional cancer cell lines, which generally grow as a single layer in a laboratory dish, organoids can form structures that more closely resemble the tissue from which the tumor cells originated.

After establishing the models—a process that could take up to a year—the researchers compared their genomic sequences, RNA activity and epigenomic modifications with those of the original tumors. The models were also linked to information about the patients, including inherited genetic mutations and treatments they had received.

The initiative was created in part because the approximately 1,000 patient-derived cancer cell lines available when it began did not adequately represent the genetic diversity of cancer. According to the MIT researchers, existing collections also contained relatively few samples from some populations and lacked many rare cancers.

The new models are intended to support larger-scale experiments for identifying drug targets and testing potential treatments. In a companion Nature paper, Broad Institute researchers used high-throughput genome and RNA sequencing to profile more than 300 of the models, and conducted CRISPR loss-of-function screens on more than 100. Those experiments identified potential vulnerabilities in individual models, with the results added to the Cancer Dependency Map, which now contains information on more than 2,000 cancer models.

A separate Nature paper from the Wellcome Sanger Institute characterized another 256 organoids developed through the initiative. The researchers emphasize that the models are experimental tools, not treatments themselves. Although much of the formal initiative is winding down, the teams hope to continue adding models, particularly for pediatric and rare cancers, because the current collection does not yet represent all human tumor types and genetic backgrounds.