For cancer patients, the promise of new immune-modulating drugs like the one that apparently helped former President Jimmy Carter comes with a sobering downside: very few get any benefit from them.
But if a new study published on Thursday is right, physicians might be able to figure out which patients those are, sparing others an expensive but useless treatment.
The research also offers clues for how to make a promising but unproven treatment, personalized cancer vaccines, more likely to succeed.
The key to both — identifying patients likely to respond to the new immunotherapy drugs and producing tumor-attacking, individualized vaccines — lies in deciphering the crazy quilt of mutations a particular patient has.
Only about one-fifth of cancer patients respond to immunotherapies like Keytruda, which is credited with helping Carter survive an advanced form of skin cancer. What makes “responders” different, previous studies in melanoma and lung cancer have suggested, is that they have a huge number of mutated genes producing molecules that find their way to the surface of the tumor cell. There, the aberrant molecules, known as neoantigens, stick out like pushpins in a corkboard.
Because those neoantigens are newcomers, the immune system should recognize them as foreign and attack, destroying the cancer cell — which is where drugs like Merck’s Keytruda come in. They lift molecular blockades that tumors use to keep the immune system’s T cells out. As a result, T cells charge in and destroy the tumor.
The reason that doesn’t happen in some 80 percent of cancer patients is that their tumors don’t have enough neoantigens to attract T cells. The total number of neoantigens roughly predicts whether cancer patients respond to immune-blockade drugs like Keytruda and Opdivo, a similar agent from Bristol-Myers Squibb, but the connection isn’t perfect.
“You see cases with a lot of neoantigens who don’t respond, and some with few neoantigens who do respond,” said Dr. Eliezer Van Allen, a clinician-scientist at the Dana-Farber Cancer Institute in Boston and an author of the new study, which was published in Science.
The number of neoantigens is only part of the story. This study goes beyond previous ones in showing that when neoantigens are found throughout a tumor rather than in some cells only, patients have the best shot at benefitting from immunotherapy. And that’s true even if there are relatively few neoantigens.
For instance, 12 out of 13 study participants whose lung cancers responded to Keytruda not only had a lot of neoantigens (more than 70), most of those neoantigens were present throughout the tumor. They attracted tumor-attacking T cells, which Keytruda let in, destroying enough of the cancer to send the patients into remission.
In contrast, 16 of 18 lung cancer patients who did not benefit from Keytruda either did not have many neoantigens or had neoantigens that were present in only some tumor cells. Even a patient with a huge number of neoantigens relapsed after only two months; more than 80 percent of his mutations were found in only some tumor cells.
“The tumors that we think will respond the best [to Keytruda and similar drugs] have a certain neoantigen burden, but those neoantigens have to be in almost every tumor cell,” said Dr. Charles Swanton, a cancer geneticist at the Francis Crick Institute in London, who, with immunologist Sergio Quezada of University College London, led the study. Swanton, Quezada, and three other coauthors also filed several patents that cover methods for identifying neoantigens and predicting the prognosis of cancer patients accordingly.
The scientists got similar results in melanoma: The fewer total neoantigens, and the more scattered around a tumor they were, the less patients responded to Yervoy, an immunotherapy drug from Bristol-Myers Squibb.
Notably, melanoma patients who had already received chemotherapy drugs or radiation had large numbers of neoantigens, apparently induced by those treatments, but each one was found in only a few tumor cells. Standard cancer therapies “may be causing changes in tumors that might not be helpful” in terms of making patients respond to subsequent immunotherapy, Swanton said.
The evidence is too preliminary for doctors to act upon, but the findings suggest that some patients might be better off skipping slash-and-burn chemo and going right to immune-modulating agents. Plus, physicians may be able to tell, by analyzing neoantigens in cells taken via standard biopsy, whether a patient has a good chance of being helped by those drugs.
The study results can also “be used to inform the development of personalized cancer vaccines,” said Van Allen. These experimental treatments, which are tailor-made to a patient’s neoantigen profile, are being tested in clinical trials now. They are designed to stimulate the immune system to attack neoantigens on tumors, but scientists hadn’t been sure which neoantigens would make the best vaccines.
“This is a very important paper,” said Dr. Elizabeth Jaffee, of the Sidney Kimmel Comprehensive Cancer Center at Johns Hopkins, who was not involved in the study. “It addresses a critical question in this field: Are all [neoantigens] equally important to the immune system or is there a way to sort this out?”
The suggestion that the best neoantigens are found throughout the tumor “makes the case that we need to sample more of the tumor before predicting which expressed mutations are most relevant for immune targeting,” Jaffee said.