Dr. Christopher Kemp is confident that the answer to personalized cancer treatment is written in the molecules inside each patient’s tumor; we just need to figure out how to separate the wheat from the chaff.
As many are learning in the field of precision oncology — which aims to match targeted cancer therapies to a tumor’s unique set of DNA mutations — this is easier said than done. Tumors have so many mutations, it is difficult to figure out which are important (and thus potentially “druggable”) and which are just along for the ride. Even when the cancer-driving mutation is known, a drug may not exist that targets that particular gene. And sometimes, even if that drug exists, it doesn't work to shrink patients’ tumors as researchers had hoped.
This is not to say that cracking tumors’ secret genetic code has not led to major advances. Some targeted therapies for tumors bearing specific mutations are showing benefit for patients already — such as Perjeta, for patients with HER2-positive breast cancer — and others are in the works. But for some cancers, understanding the genetic makeup of the tumor may not be enough, said Kemp, a biologist at Fred Hutchinson Cancer Research Center.
“Cancers have many vulnerabilities that can be targeted with drugs, and these vulnerabilities could not be found by DNA sequencing alone.” he said. “Cancers have many Achilles’ heels.”
With the arrival of new technologies, it is now possible to find those hidden vulnerabilities and exploit them for better cancer treatments, Kemp said. We don’t need to wait years or decades for cancer genetic studies and drug development to catch up to the promise of precision medicine — a promise that has so far failed to deliver for many cancer patients.
The approach adopted by Kemp falls under the category of techniques known as “functional genomics.” Rather than just sequencing tumors’ genomes, the method involves pairing that DNA sequencing with high throughput drug testing and a third method that pinpoints cancer’s vulnerabilities, all using the same patient-derived tumor cells grown in the lab. These so-called high throughput screens are miniaturized and automated, meaning they can run hundreds or thousands of different experiments simultaneously, yielding insight into a patient’s tumor in a matter of weeks, Kemp said, which is years faster than older techniques.
“Screens make you smart,” said Kemp. “They give you all the inside information you need about what you want to target and what are the vulnerabilities of any given cancer. How can that not be a good idea?”
This triple whammy approach to precision oncology is already yielding new therapeutic targets for some cancers. Kemp has projects in the works with Fred Hutch, University of Washington, and other national and international collaborators studying breast cancer, head and neck cancer, pancreatic cancer, and ovarian cancer. In what may be the closest-to-clinical-impact example of the technique, an early-stage clinical trial led by Fred Hutch’s Dr. Eduardo Méndez is pairing an existing drug with patients with a certain type of head and neck cancer. Many of the patients in that trial are responding very well to the drug, Méndez said — and most of those who completed the experimental treatment saw their tumors shrink enough so they could receive surgery for which they were previously ineligible.
This is a drug nobody would have thought to try on these patients. It was only through evidence from their screens that pinpointed this drug among hundreds of possibilities, the researchers said. That’s why they are convinced that their approach can be more broadly applied to other patients and cancers in the near future.
Saturday, a group of poker aficionados will gather at the Hutch to help fulfill that goal. April 1 will mark the cancer center’s second “Poker Masters” event, a fundraising tournament hosted by poker legend Phil Gordon to benefit the functional genomics research. The poker event was championed by Fred Hutch’s Dr. V.K. Gadi. Gadi liked the idea of a breast cancer fundraising event with a fun and inclusive vibe, he said — and he also just loves the game.
“I love the math and the challenge of having incomplete information and trying to use that incomplete information about what’s going on around you to make decisions. I see tons of parallels between that and what I do in my clinic every day,” he said. “It’s a beautiful game.”
The functional genomics approach using patient derived tumor cells is a collaborative effort on a grand scale, Kemp said.
“This can’t be done in any one lab,” Kemp said. “To make this happen requires buy-in from clinicians, patients, researchers and computational scientists.”
Gadi and Méndez are oncologists, seeing patients with breast cancer and head and neck cancer respectively, as well as conducting clinical research. Together with Kemp and Dr. Carla Grandori, a former Fred Hutch cancer researcher, the team founded the spin-off biotech company SEngine Precision Medicine in 2015 to bring this targeted approach directly to patients to determine the best treatment option for their particular tumor.
For now, the work is still mainly in research mode. In its latest incarnation, the methodology uses live, patient-derived cancer cells in petri dishes and what are known as tumor “organoids,” 3-D clusters of tumor cells grown in the lab that mimic actual human cancers better than cancer “cell lines,” which are cancer cells that originally came from a patient but have spent so long growing and dividing in petri dishes that they bear little resemblance to the original cancer.
The reliance of much of cancer research on such cell lines is one of the reasons so many promising approaches to treatment fail to translate to actual clinical benefit, Gadi said.
“A lot of the drug discovery and trying to understand the genetic underpinnings of cancers have led to a lot of false leads,” he said. “You can spend $1 billion developing a drug and at the end of the day it doesn’t work, because it just didn’t matter in people’s cancers as opposed to this theoretical cell line.”
The tumor organoid method is still early stage and takes a lot of work to develop because each tumor type requires different methods, Gadi said, underscoring the importance of obtaining live tumor biopsies that can be grown in the lab to further perfect the method.
In a proof-of-concept test of their method, Grandori and SEngine scientists teamed up with researchers from Fred Hutch and the Englander Institute for Precision Medicine at Weill Cornell Medicine. The team recently completed a study, published online in the journal Cancer Discovery, that combined DNA sequencing and drug screening of tumor organoids developed from four patients’ tumors, two with late-stage uterine cancer and two with late-stage colon cancer.
In that study, the researchers tested 120 FDA-approved drugs against the four tumor organoids, and then used mouse avatar models of the patients’ tumors to see how the best hits from that screen fared in a living model of cancer. In the end, the study uncovered four combinations of different drugs that they predicted should work best for each of the four patients — all originally approved for use in different cancer types, such as breast or lung cancer. This new approach to precision medicine will be presented at the 2017 American Association for Cancer Research meeting, which starts this weekend in Washington D.C., by Kemp and Weill Cornell Medicine’s Dr. Mark Rubin on behalf of the group.
This type of precision oncology research is in early stages; its main purpose is to suggest new paths for clinical trials. That’s why the cancer genome sequencing part of the work is important, to find other patients whose tumors might respond similarly to the drug or drugs their organoid screens pull out.
For Kemp, who has had two parents die of cancer, one from colorectal and one from kidney cancer, this work is both deeply fulfilling and deeply frustrating. Even though their drug screens found a potential therapy much faster than older methods, one of the two colon cancer patients died while the study was ongoing.
“That frustrates the hell out of me,” Kemp said. “The only reason we couldn’t do this sooner is we lack the resources and manpower, not because we lack the technology.”
The other arm of their approach to functional genomics is what’s known as a “synthetic lethal” screen, which identifies cancer’s Achilles’ heel (or heels) in a very different way than just identifying cancer-driving mutations. In a synthetic lethal screen, the researchers take the cancer cells and pick off all or some of their genes, one at a time, in an effort to see which genes the tumors rely on for their growth. They then do the same screen in healthy cells to find the genetic weak points that are unique to cancer. The end result? A set of genes that, if removed, could kill cancer cells — but not healthy tissue.
The ideal end result? A new gene target identified for which a drug already exists.
That’s just what Méndez, Kemp, Grandori and their colleagues found when they did this kind of screen for head and neck cancer. Most head and neck cancers carry mutations in a gene known as p53, which Méndez describes as cells’ brakes against rapid growth when DNA is damaged, as it always is in cancer. When cells disable that braking system, they’re able to grow out of control. But understanding the loss of p53 in these cancers didn’t lead to any new targeted therapies for the disease, Méndez said, because it’s impossible to drug what isn’t there.
Instead, Méndez and his team used the technique developed by Kemp and Grandori to look for weak points in head and neck cancer cells with mutant p53. They pulled out several hits, but one — a gene known as Wee1 — turned out to be the target of a cancer drug known as AZD1775 which is being developed by the pharmaceutical company AstraZeneca.
Working with the company, Méndez opened an early-stage clinical trial in 2015 to test the drug on any head and neck cancer patient whose tumors were too large (or otherwise too complex) to be surgically removed. They recruited 11 patients who received AZD1775 as a front-line treatment in combination with a standard chemotherapy, as an attempt to shrink the tumors enough to allow the patients to receive potentially lifesaving surgeries.
The Phase 1 trial was designed to test the drug’s safety and establish a maximum dose for a larger trial, but the researchers saw some promising results on the patient benefit side as well. Of the eight patients who completed the treatment, seven saw their tumors shrink enough that they were eligible for surgery. Some of the patients are more than a year out from treatment, Méndez said, and two have no signs of their cancer after the surgery. That 90 percent rate of patients who became eligible for surgery after receiving the experimental drug combination stands in contrast to the standard treatment for head and neck cancer patients — typically, only 30 to 40 percent of patients ineligible for cancer see their tumors shrink enough to progress to surgery using traditional treatment, Méndez said.
He’s hoping to start a larger, Phase 2 trial very soon — if they can find the funding.
Results like these give Kemp hope that he and his colleagues are onto something, he said.
“I’m very optimistic,” Kemp said. “We have in our hands the ability to peer into a patient’s cancer and figure out where the weaknesses lie. We are all looking forward to what we can accomplish next to realize the promise of precision medicine.”
Fred Hutch staff writer Sabrina Richards contributed to the reporting of this story.
Rachel Tompa is a former staff writer at Fred Hutchinson Cancer Center. She has a Ph.D. in molecular biology from the University of California, San Francisco and a certificate in science writing from the University of California, Santa Cruz. Follow her on Twitter @Rachel_Tompa.
Every dollar counts. Please support lifesaving research today.
For the Media