Modern genomic sequencing techniques reveal that cancers come with a bewildering variety of mutations. Cancers that have the same driving mutation but grow in different environments — skin vs. colon, for example — may respond completely differently to the same therapy.
Conversely, tumors growing in different patients in the same organ may superficially appear similar but are often genetically diverse. With little information about how a tumor will respond to treatment except trial and error, it is challenging for oncologists to identify the optimal therapy for each patient.
To better correlate genetic profiles with cancer treatment and survival, oncologist Dr. John Shen, M.D., was recruited to The University of Texas MD Anderson Cancer Center in 2018. A physician-scientist, he was a postdoctoral fellow and clinical instructor at the University of California, San Diego, and was recruited with the help of a First-Time Tenure-Track Award from CPRIT. In addition to conducting research, Dr. Shen is an attending physician treating gastrointestinal cancers.
“Cancers are in some sense like snowflakes, in that every person’s tumor is unique,” Dr. Shen says. “As an oncologist, I want to know, ‘How do I kill that cancer, what is its weakness?’ It’s fine to say that a particular kind of chemotherapy works 60% of the time, but for an individual patient, all they want to know is, ‘Will it work for me?’”
To get a better handle on this question, Dr. Shen is using functional genomics, a field that aims to understand the functions of genes that are mutated or amplified in each cancer. Colon cancers typically contain 60-100 mutations, but figuring out which of these are key to the cancer’s continued survival is a challenge.
Using CRISPR/Cas9 gene-editing technology in cultured cancer cell lines, in a single experiment Dr. Shen can knock out two genes at a time and try 20,000 gene-knockout combinations to find out which genes are crucial for cell survival. He’s initially targeting genes for which there are known drugs. His next step is to try these experiments in models of human cancers in mice. He hopes to eventually be able to test a patient’s biopsied cells grown in an organoid or mouse.
Dr. Shen is also trying to train computers to identify trends and correlations between the genetic profile of a patient’s cancer and treatment and outcome. So far, machine learning hasn’t succeeded at improving survival for colon cancer but mostly because there isn’t enough data yet for the computer to make accurate predictions. “How long someone lives after cancer treatment is the most important outcome,” he says, “but it’s influenced by so many other things — someone may die in a car crash, or stop therapy because of intolerable side effects.”
The combination of the knockout experiments and machine learning will enable physicians like Dr. Shen to better predict which combinations of drugs will benefit which patient, without having to run the experiment every time. “If we have 10 drugs that are promising, in order to test all the combinations, we’d have to run 45 clinical trials,” he says, “while in our experiments we can easily try all of the possible combinations.”
Dr. Shen’s research is integrated with MD Anderson’s Colorectal Cancer Moon Shot™, which has the infrastructure to collect tumors from patients and grow cancer cells in organoids and mice.
Dr. Shen studied chemistry as an undergraduate at the Massachusetts Institute of Technology, and received his M.D. from Washington University School of Medicine. He came to UCSD as a resident in internal medicine in 2008, and remained there first as a fellow in hematology & oncology, and then as a postdoctoral researcher in genetics.
Read Less
