Pediatric cancers tend to have quiet genomes, with few genetic culprits underlying processes that go wrong in cancer. When starting his research career, Charles W.M. Roberts, MD, PhD, St. Jude Comprehensive Cancer Center director, focused on understanding the mechanisms behind the biology of cancer cells. The goal was to find their vulnerabilities. Early in Roberts’ work, he began studying mutations in SMARCB1/INI1 in pediatric rhabdoid tumors. This rare cancer develops in the kidney, brain, liver or other soft tissues in the body. Occurring in babies and toddlers, these cancers are lethal in most cases. As a pediatric oncologist and a scientist, Roberts set up his laboratory to investigate the biology of SMARCB1/INI1.

As the only mutation driving rhabdoid tumors, SMARCB1/INI1 was undoubtedly important. But Roberts was also interested in this mutation for another reason. Previous research on a rare early childhood eye cancer called retinoblastoma had revealed mechanisms that were broadly applicable to many types of cancer. Roberts hypothesized that SMARCB1/INI1 mutations in rhabdoid tumors might also be a “canary in a coal mine.”

SMARCB1/INI1 is a part of the SWI/SNF chromatin remodeling complex. This means it plays a role in how other genes are turned on and off. Research from Roberts’ lab proved that SMARCB1 was the key culprit in these cancers. His hypothesis proved correct: we now know that SWI/SNF complex mutations are present in more than 20% of all cancers of adults and children. A single mutation in a rare pediatric cancer again alerted scientists to this broader mechanism driving cancer.

In this interview, Roberts describes how his research has been shaped by these findings, identifying strategies for treating cancers driven by this complex.

Q: What can you tell us about targeted therapies for EZH2 and your early work on this target?

A: In 2010, we found that the function of SWI/SNF is counter-balanced by a gene called EZH2. Due to the absence of SMARCB/INI1 in rhabdoid tumors, the activity of EZH2 was unchecked. We found that shutting down EZH2 could stop the growth of cancers with mutations in SMARCB1/INI1. Our paper on this finding in Cancer Cell provided scientific stimulus to encourage a pharmaceutical company to begin testing their newly developed EZH2 inhibitor in SMARCB1/INI1-mutant cancers. The effect was promising, and the resulting drug, tazemetostat, was tested in clinical trials. Ten years after our discovery, this drug was approved by the Food and Drug Administration to treat older teens and adults with epithelioid sarcoma who have mutations in SMARCB1/INI1. The approval of tazemetostat shows what is possible when you pursue the mechanisms driving cancer and use them to develop treatments.

Q: How does collaboration influence your progress?

A: After identifying EZH2 as a potential therapeutic target, our lab wanted to do more to find possible weaknesses in pediatric cancers. We launched a close collaboration with colleagues Kimberly Stegmaier, MD, Todd Golub, MD, and others at Dana-Farber Cancer Institute and the Broad Institute. Together, we pursued large-scale efforts to screen pediatric cancer cell lines from a variety of cancers. Through this approach, we are looking for previously unrecognized genetic vulnerabilities and drug sensitivities. The effort now includes genetic screens in more than 100 pediatric cancer cell lines, which we can compare to more than 500 adult cancer cell lines.

Through another collaboration with Stuart Schreiber, PhD, at the Broad Institute, we were able to add 12 rhabdoid cell lines to 900 other cell lines that were then all tested against 430 drugs. We marry the robust data from these efforts with our continued study of the mechanisms that underlie SWI/SNF complex mutations and cancer. With this approach, we are identifying potential therapeutic vulnerabilities at a rapid rate.

Q: What other technological advances in the past decade have helped speed up cancer research?

A: Certainly, CRISPR-Cas9 genome-wide screening has had a large impact. We used this approach to better identify new vulnerabilities in rhabdoid cell lines. We and others found that BRD9 is such a vulnerability. Published in Nature Communications, this was a bit of a “eureka” moment, because we found that BRD9 is itself a subunit of the SWI/SNF chromatin remodeling complex. Understanding how loss of SMARCB1/INI1 causes BRD9 to become essential in the SWI/SNF complex may make it possible to pursue drugs that target BRD9 in these cancers.

Q: You’ve mentioned several targets that chemists can then go on to design drugs to inhibit. Are there other ways to exploit new vulnerabilities?

A: We have also found targets for which drugs already exist. This is the case with our work on MDM2 and MDM4, targets related to the SWI/SNF complex. We recently published work in Cancer Research showing that malignant rhabdoid tumors are vulnerable to inhibition of MDM2 and MDM4. Drugs in development that inhibit these targets are now being brought forward quickly into clinical trials, which are in development or underway at St. Jude and elsewhere.

We also recently found that receptor tyrosine kinases (RTKs) can also be targets for rhabdoid tumors. While the mechanism remains poorly understood, the vulnerabilities came through strongly. This work, which appeared in Cell Reports, showed that rhabdoid tumor cell lines highly express and activate different RTKs, creating dependency without mutation or amplification. This opens the door for using these drugs as possible treatments in patients.

Most recently, we published a paper in Clinical Cancer Research showing that every rhabdoid tumor cell line we tested was highly sensitive to a drug called homoharringtonine. This drug is currently approved to treat another type of cancer called chronic myeloid leukemia. We found that low expression of the gene BCL2L1, which is involved in cell death, could help predict sensitivity to homoharringtonine.

Q: What is next for your lab?

A: While a tremendous amount of work has been done to identify therapeutic opportunities against rhabdoid tumors (summarized with postdoctoral researcher Priya Mittal, PhD, in Nature Reviews Clinical Oncology), every day we learn new things about the biology of this disease.

When I started our lab, we faced a question that still drives our work today: How do you fix a missing gene? For SMARCB1/INI1, the mutation can’t be drugged because the gene is missing. This story is common across pediatric cancers, where the few recurrent mutations present in a cancer may not be targetable.

We now know that this scenario of SWI/SNF mutations is remarkably prevalent in many types of cancer. My passion continues to be to understand how SWI/SNF functions, what goes wrong when one of its constituent genes is mutated that results in cancer formation, and ultimately how we can use this information to cure cancers. At St. Jude, we continue to pursue these questions, making discoveries that hold the potential for helping treat childhood cancer.