Cleveland Clinic Researchers Find Hope in Immune Cells’ Reaction to Immunotherapy

By Irene Yeh 

October 23, 2024 | There is much we still don’t know about the relationship between immune cells and cancer cells and how they interact and influence each other over the course of immune activating cancer therapies. Researchers at Cleveland Clinic are exploring these relationships and reported their results last month in Nature Medicine (DOI: https://doi.org/10.1038/s41591-024-03240-y). 

Why Neoantigens? 

“A lot of patients still don’t respond to immunotherapy,” says Tyler Alban, PhD, co-first author of the study. To understand how tumors are recognized and altered by immune systems in response to immunotherapy, Alban and his colleagues at Cleveland Clinic conducted a study with Bristol Myers Squibb. 

They started their search at neoantigens. In order for immune recognition to occur, tumor cells need to look different from normal cells. As cancer cells evolve, they develop somatic mutations, which create small peptides called neoantigens. These neoantigens allow the immune system to recognize that the cancer cells are different from normal cells. These peptides are detected by the CD8+ T cells, which destroy tumor cells and are the primary target of immune checkpoint therapies, explains Alban. Without the peptides, the CD8 + T cells cannot recognize that the tumor cells need to be disposed of. 

In a previous study, the Cleveland Clinic team had conducted genomic analysis of biopsy samples obtained from patients with melanoma before and during they were treated with nivolumab, a monoclonal antibody used to prevent cancer cells from suppressing the immune system. However, while the results hinted that the predicted neoantigens were reduced more in responders than non-responders, there still was not enough data to conclude an answer.   

Furthermore, “in silico predictions aren’t very reliable,” according to Alban. But a recent TESLA consortium examined immunogenicity by tasking 28 teams to predict immunogenic peptides, resulting in an average success rate of 6% (Nature Medicine, DOI: https://doi.org/10.1038/s41591-024-03240-y). Despite the small rate, it was enough for the team to conduct a similar study focused at a single timepoint pre-therapy. In this study, they highlighted the challenges in immunogenicity prediction. 

The CheckMate 153 Study 

The CheckMate 153 study collected 80 radiographic guided biopsy samples before and during treatment with nivolumab (Nature Medicine, DOI: https://doi.org/10.1038/s41591-024-03240-y). Upon observation, the team determined that people who responded well to nivolumab had a sharp decline in clonal neoantigens. Meanwhile, patients whose cancer did not go into remission still mounted an immunologic response to smaller sub-clonal populations. This is important because it shows that the immune system responds to neoantigens, though the response is insufficient to destroy all tumor clones.   

To identify immunogenic responses, the team had to test neoantigens’ ability to bind to the human leukocyte antigens (HLAs) and subsequently become recognized by a T cell receptor. Current neoantigen prediction tools rely heavily on HLA-binding neoantigens, but they perform poorly on the T cell recognition aspect of immunogenicity, says Alban.   

In addition to the tumor biopsies, the team collected blood samples during the course of treatment to stain for T cells that recognize HLA/neoantigen complexes. The team saw that the samples with the neoantigens that had stronger CD8+ T cell response over time were more likely to be immunoedited out of the tumor, relays Alban.  

“One thing we didn't expect was, some of the non-responders—some of their neoantigens—did have immunogenic T cells that recognized them and were immunoedited, but not all of them,” continues Alban. “Typically, we think of these non-responders as being immune cold and not able to recognize the cancer, but we have evidence here that, at least in some of them do have a T cell response.” 

The Atlas and Future Plans 

The team gathered their data and built the first large-scale atlas of how immune cells react to mutations during cancer immunotherapy, and the data of the study is available in supplementary material in the manuscript. Nadeem Riaz, data scientist Preana Parthasarathy, was a key member of the team and the project, Alban notes.   

“It creates a little bit of hope in our minds that these non-responders aren’t completely cold, so they may potentially respond to the right vaccine or the right CAR T Cell therapy,” says Alban. He states that there is a possibility of redirecting the immune response the right way, thus changing responses. 

Additionally, the atlas will provide clinicians and other medical professionals with a resource that can help develop new and improved antigen prediction tools. Though it is still a small set of immunogenic T cells, it can still be used to develop better machine learning tools to predict immunogenic new antigens, thus developing better vaccines for patients and better understanding of immunogenic responses. Alban states that the team is already on top of that. A few months ago, they collaborated with IBM and published a paper on predicting new antigens using molecular dynamics. There are more plans to further these models with IBM and expand our atlas.  

“We hope to use this data to make more generalizable models in the future that will be more helpful for vaccine design.”