Although we have made a lot of progress in understanding the role of PI3K in insulin signaling and cancers over the past 30 years, it was not always smooth sailing. There was considerable skepticism of our claim in the 1980s that a lipid kinase was activated by cancer-causing genes. Most researchers in this field were virologists or molecular biologists and had little or no experience working with membrane lipids. The leading laboratories in the field published papers arguing that our findings were not correct, and this made it difficult to publish our work or obtain grants to support it. The graduate students from my lab and Tom Roberts’ lab, Malcolm Whitman and David Kaplan, visited a few of the skeptical scientists and showed them how to do the lipid kinase assay. After that, they were able to reproduce our results and became supporters of the discovery. The lipid chemists were also skeptical that they could have missed this family of lipids over more than 30 years of research, but they ultimately were able to reproduce our results. Yet there were three to four years when funding and publishing this work was difficult.
Figuring out all the cellular events controlled by the PI3K-generated lipids is ongoing. Thirty years later, there’s much more to be discovered. We know a lot — the broad strokes, the major players — but there’s still a lot of subtleties to how this signaling network is regulated and what goes wrong in diseases such as diabetes and cancers.
Cancer tells us a lot about how it works, because mutations that arise are almost invariably affecting some step in growth regulation. By just looking at all the mutations in cancers, we can begin to make sense of it.
As I noted, the gene encoding PI3K, PIK3CA, is the most frequently mutated oncogene across all types of cancers and particularly in women’s cancers. Approximately 30% of breast cancers and 50% of endometrial cancers have PIK3CA mutations.
In 2009, an organization called Stand Up To Cancer, in affiliation with its scientific partner, the American Association of Cancer Research, issued a proposal to fund what they call “dream teams.” At that time I was at Beth Israel Deaconess Medical Center, associated with Dana-Farber Cancer Institute, and I put together a dream team of world-renowned cancer researchers from major institutions throughout the country, including Dr. Ramon Parsons, then at the Herbert Irving Comprehensive Cancer Center at Columbia and NewYork-Presbyterian. We were awarded more than $12 million to evaluate the use of PI3K inhibitors for treating women’s cancers. We went to pharmaceutical companies that were developing these drugs and said, “We can help you design the trials that are more likely to get your drug approved.”
We played a role in designing the phase Ib trial for a Novartis drug called alpelisib for estrogen receptor-positive breast cancer. Forty percent of the patients with these cancers have mutations in PIK3CA.
Along the way we had to resolve issues related to the enzyme’s dual roles in insulin signaling and cancer. If you give a PI3K inhibitor, it hits the enzyme not only in the tumor but also in the liver, muscle, and fat cells, promoting insulin-resistance and diabetes. Being aware that high insulin levels could further activate PI3K in the tumor and drive tumor growth, we insisted that patients on PI3K inhibitors not be given insulin or other drugs that increase insulin production in the pancreas.
Eating a very low carbohydrate diet — limiting both sugar and starch — could be a way to improve responses to these drugs. In studies where we gave a PI3K inhibitor to mice engineered to develop pancreatic, bladder, endometrial, and breast cancers, we put them on a ketogenic diet and their tumors melted away.