Researchers Devise Model of How Tumors Respond to Immune Checkpoint Inhibitor Drugs

Using technology that shapes tumor tissue into minuscule spheres, researchers at Dana-Farber Cancer Institute have devised a technique for determining how a tumor responds to immunotherapy agents known as PD-1 inhibitors, and whether these drugs may be more effective when teamed with other therapies. Researchers used the technique to successfully identify a drug combination that was effective in animals with various types of tumors.

The technique, described in a study posted today on the Cancer Discovery website, has the potential to answer some of the most pressing questions surrounding treatment with PD-1 inhibitors, the authors state. While the drugs, part of a class of agents known as immune checkpoint inhibitors, have sparked long-lasting remissions in some cases of melanoma, lung cancer, kidney cancer, bladder cancer, and other solid tumors, they’re effective in a minority of patients. Oncologists are searching for more reliable ways of predicting which patients are likely to benefit from such drugs.

And while combinations of checkpoint inhibitors and other drugs are thought to have immense potential, the number of possible combinations ranges into the hundreds. Testing each combination in enough patients to achieve statistical validity is a mammoth, costly, and time-consuming undertaking.

PD-1 is a protein molecule that some cancer cells use to blend in with neighboring normal cells, and thereby avoid being attacked by the immune system, which is programmed to eliminate foreign or diseased cells. Drugs that block PD-1 or other immune checkpoint molecules essentially strip cancer cells of their disguise, laying them open to immune system attack.

“Tumors exist in a state of constant interactions with the immune system – and whether a checkpoint inhibitor succeeds or fails in killing tumor cells depends to a large degree on those interactions,” said David Barbie, MD, the co-senior author of the study with Dana-Farber colleague Cloud Paweletz, PhD, and David Dornan of Gilead Sciences. “Until now, our ability to understand the dynamic between tumors and the immune system ­– and how checkpoint inhibitors affect that dynamic – has been quite limited. We’ve had to rely on measurements of substances in the blood or on studies of preserved tumor tissue.

“In the new study, our goal was to create a model system that captures, outside the body, the interplay between cancer cells and the immune system. Such a model could help scientists identify substances, or ‘biomarkers,’ within a tumor that indicate whether a particular checkpoint inhibitor is likely to be effective against that tumor. It could also help identify new inhibitors, or combinations of inhibitors and other drugs, for patients who are resistant to PD-1 inhibitors alone.”

The model created by Barbie and his colleagues harnesses what’s known as three-dimensional microfluidic culture technology. After a tumor tissue sample is collected, it is immersed in enzymes that “digest” it into smaller pieces. The pieces coalesce into tiny roundish structures known as spheroids, which pass through a filter and are collected and stored in a medium that keeps the cells alive.

“Each spheroid is a mini-ecosystem,” Paweletz remarks, “containing not only tumor cells but the immune system cells with which they interact.” The spheroids, each about the size of a dust particle, can then be treated with checkpoint inhibitors, alone or in combination with other drugs, to study the effect – on the tumor cells themselves and on substances called cytokines that are secreted by the immune system cells. Changes in cytokines may provide an early indication of whether a drug or drug combination is proficient at killing tumor cells.

To test whether the model faithfully reflects what happens in the body when a tumor is treated with a PD-1 inhibitor, researchers selected a strain of “syngeneic” mice – animals implanted with tumor tissue that originally grew in mice with the same genetic background. This particular strain of mice was chosen because, as in human patients, some of the animals’ tumors shrank in response to PD-1 inhibitor treatment while others didn’t.

When the animal and the spheroid tumor models were treated with a PD-1 inhibitor, they responded in tandem: when a mouse tumor succumbed to the drug, the spheroid made from that tumor did as well; when a mouse tumor resisted the drug, the spheroid also did. Researchers also found that following treatment, the animal tumors and the spheroids harbored the same types of immune system cells.

“These results showed that the effect of the drug on the spheroid models closely mirrored its effect on tumors in animals,” Barbie remarks.

The investigators then used the model to test a combination of a PD-1 inhibitor and a novel compound called TBK1/IKKε. The model predicted that the combination would outperform the PD-1 inhibitor alone at controlling tumor growth and extending the mice’s survival – which is precisely what researchers found when they duplicated the experiment in the animal models.

In the next phase of the study, the researchers made spheroids from human tumors rather than mouse ones. As in the mouse studies, the spheroid models responded to PD-1 blockers much as the tumors themselves had in patients. The PD-1 blockers spurred the spheroids to produce certain cytokines – substances that beckon immune system cells to the site or infected or diseased tissue – that gave important clues as to how a tumor may or may not respond in the body.

“Our findings offer the first demonstration that it is possible to determine tumors’ responsiveness or resistance to PD-1 inhibitors in an experimental model,” said the study’s lead author, Russell Jenkins, MD, PhD, of Dana-Farber and Massachusetts General Hospital (MGH).

The researchers note that the technique is limited by the ability to obtain enough tumor tissue to be tested – roughly a cubic centimeter, or the size of a small jelly bean – to keep the tissue alive, and by the inherent challenges of working with live cells.

More work is needed to refine the technique into a practical test that could be used in the clinic, but the benefits of such a test are clear. “For patients, such a test could indicate which PD-1 or other immune checkpoint inhibitor would most likely be effective for them,” Barbie observes. “For patients who don’t respond to a PD-1 inhibitor alone, the test could point them toward clinical trials of combinations of PD-1 inhibitors and other agents that hold the greatest potential.”

Financial support was provided by the Robert A. and Renée E. Belfer Foundation; a National Cancer Institute Training Grant (T32CA009172-41); a John R. Svenson Fellowship; the Gloria T. Maheu and Heerwagen Family Funds for Lung Cancer Research; the Rising Tide Foundation; the Expect Miracles Foundation; the Robert A. and Renée E. Belfer Foundation; the National Institute of General Medical Sciences (grant T32 GM008313); the National Science Foundation (grant DGE1144152); a Caja Navarra Foundation Excellence Doctoral Mobility Grant; an Association of Friends of the University of Navarra Doctoral Mobility Grant; the National Cancer Institute (grants R01 CA190394-01, P01CA120964, 5R01CA163896-04, 5R01CA140594-07, 5R01CA122794-10, and 5R01CA166480-04); the Gross-Loh Family Fund for Lung Cancer Research; the Susan Spooner Family Lung Cancer Research Fund at Dana-Farber Cancer Institute; and the Elizabeth and Michael Ruane Fund. Additional funding was provided by National Institute of Health grants P01 CA114046, P01 CA025874, P30 CA010815, and R01 CA047159; and by the Dr. Miriam and Sheldon G. Adelson Medical Research Foundation and the Melanoma Research Foundation. The support for Shared Resources used in this study was provided by Cancer Center Support Grant CA010815 to The Wistar Institute. Additional support was from a Stand Up To Cancer-American Cancer Society Lung Cancer Dream Team Translational Research Grant (SU2C- AACR-DT17-15).