Much like how a shot of espresso might re-energize you to get through the day, cancer immunotherapies have been developed to re-energize anti-tumor immune responses to help clear cancerous cells. When chronically exposed to antigen, T cells, an important component of our adaptive immune response for anti-tumor immunity, can become less responsive, a term called exhaustion. When they reach this hyporesponsive state, they often express characteristic markers on their surface that can make them easily identified by researchers, such as PD-1. The discovery of PD-1 as an exhaustion marker was a major advancement in field of cancer immunology, resulting in a Nobel prize in 2018, and helped lead to the development of immune checkpoint blockade therapies. However, response rates to therapy can be variable, and much of the focus now has turned to methods to improve these treatments, as well as to recognize biomarkers predictive of positive outcomes.
It was with these goals in mind that Dr. Heeju Ryu, now at Sungkyunkwan University in Korea, in Dr. Evan Newell’s lab recently published in Cell Reports Medicine, wherein they wanted to assess whether they could identify biomarkers of response to immunotherapy in Merkel Cell Carcinoma (MCC). MCC, a rare form of skin cancer, is a useful model for studying anti-PD-1 therapies as it is particularly responsive to this treatment and is often driven by Merkel cell polyomavirus (MCPyV). “Because MCC is such a strong example of effective anti-PD-1 therapy, I think it provides a perfect setting for understanding the mechanisms of this response,” Dr. Newell commented. In the case of tumors recognized to be driven by MCPyV, it can be easy to identify tumor-specific T cells as they are often responsive to viral antigens, which further enhances the ability to understand how T cells may interact with the tumor. By using mass cytometry and RNA-seq methods, the team identified two unique populations of CD8 T cells present in the blood which corresponded to different responses to therapy, suggesting that the presence of either of these cells could be utilized as biomarkers in patients with MCC.
To approach this study, owing to a successful long-term collaboration with Dr. Paul Ngheim at the University of Washington, the team used longitudinal samples from a recent clinical trial using pembrolizumab (anti-PD-1) in MCC patients. This trial included a variety of clinical responses, such as complete and partial response (CR and PR, respectively) as well as stable or progressing disease (SD and PD, respectively), and included both virus positive (VP) and negative (VN) individuals. The researchers used a method called tetramer staining, which could allow for the identification of CD8 T cells specific to MCPyV. The team knew that virus specific CD8 T cells would be rare and hard to detect, so they “needed to make sure that our staining and approach was working very well”, noted Dr. Newell. As a testament, then, to taking time to carefully optimize these types of approaches, they were reliably able to detect these rare populations in patients known to be virus positive. Consistent with parallel reports, they found that higher frequencies of MCPyV-specific CD8 T cells at baseline correlated with improved response to treatment. Interestingly, they observed that in CR patients, virus-specific cell frequencies dropped by the end of treatment, something that was not observed in the other patients.
Using high-dimensional approaches with their phenotyping panel, they wanted to further assess whether specific subpopulations of CD8 T cells were associated with these responses. From this analysis, the authors identified unique phenotypes of MCPyV-specific cells by comparing them to unrelated viral-specific CD8 T cells. These cells of interest expressed a variety of surface markers, but of particular interest to the research team was the stable and high expression patterns of CD39, a marker of antigen-specific exhausted CD8 T cells, CLA, a skin-homing receptor, and CD103, a tissue-recirculating marker. Following clustering analysis, they identified clusters predictive of outcome, from which, one distinguished by the co-expression of CD39 and CLA was associated with a positive outcome to treatment. Agnostic to virus-specificity, these cells were also found in higher frequency in VP patients and their frequencies correlated with those of virus-specific cells, indicating that they may represent virus-specific cells. Higher baseline levels of these CD39+CLA+ CD8 T cells were also associated with CR, suggesting that they may be a useful biomarker predictive of favorable outcomes. They next focused in on CD39+CD103+ CD8 T cells and observed that, in contrast to CD39+CLA+ CD8 T cells, they were found in both VP and VN patients and higher frequencies were associated with higher initial tumor burden and poor response to treatment. To help reinforce that these T cell populations represent MCPyV-specific cells, using a variety of analyses such as tetramer staining and T cell receptor (TCR) sequencing of blood-derived and tumor-derived cells, they found that these CD8 T cell populations of interest were likely, in fact, enriched for virus-specific cells, particularly CD39+CLA+ CD8 T cells.
When the research team looked at gene expression within these cell populations by RNA-sequencing, they found that both expressed genes indicative of exhaustion, indicating they are likely less responsive to tumor antigen, despite being in the presence of it. But within the CD39+CLA+ population, which was increased in patients with better clinical response to treatment, they saw genes associated with effector functions, suggesting they still had the capacity to respond. However, the CD39+CD103+ population, which was associated with poor outcomes, had gene profiles consistent with terminal exhaustion, suggesting their capacity to respond could be extremely low, and they might not be able to be re-energized even by treatments targeting exhausted cells, such as anti-PD-1.
Altogether, the team appears to have identified a novel set of biomarkers that can be detected from simple blood draws of patients with MCC. The higher frequency of CD39+CLA+ cells prior to immunotherapy onset, and which decrease after the start of treatment, may represent a population of tumor specific cells which are associated with positive outcomes to treatment. In contrast, CD39+CD103+ CD8 T cells may be terminally exhausted cells that are unable to provide any clinical benefit to patients, and correspondingly are found in higher frequencies in patients with worse clinical outcomes. Dr. Ryu explained that these contrasting predictive cell populations was a surprising finding, “my initial expectation was that tumor-specific T cells would correlate with better clinical outcome. However, it was a more complex picture.”
By utilizing surface markers, such as the ones they identified in the study, Dr. Newell was excited by the feasibility of these results in a clinical setting, stating that this could be used for “the identification of easier to measure (no tetramers required) biomarkers”. Dr. Ryu added that their data also highlighted how “even small populations of these cells may hold significant prognostic value”. The research team is also interested in seeing if these results could be more broadly applicable to other cancer types. However, they recognize this might not be a straightforward question as it would be useful to determine if other surface markers are important in different contexts, or it there are shared phenotypic signatures that could be identified. Finally, they hope that their findings could extend beyond biomarkers for clinical outcomes. Dr. Ryu added, hopeful, that “understanding these T-cell populations might reveal novel targets for boosting anti-tumor immunity. This work could inform the development of cell-based therapies… But it is still a long way to go!”.
This research was funded by the CCSG, Fred Hutch, the National Research Foundation of Korea, the NIH, a Kelsey Dickson Team Science Courage Research Award, and Merck & Co.
Fred Hutch/University of Washington/Seattle Children’s Cancer Center Consortium members Drs. David Koelle, Paul Nghiem, and Evan Newell contributed to this work.
Ryu H, Bi TM, Pulliam TH, Sarkar K, Church CD, Kumar N, Mayer-Blackwell K, Jani S, Ramchurren N, Hansen UK, Hadrup SR, Fling SP, Koelle DM, Nghiem P, Newell EW. Merkel cell polyomavirus-specific and CD39+CLA+ CD8 T cells as blood-based predictive biomarkers for PD-1 blockade in Merkel cell carcinoma. Cell Rep Med. 2024 Feb 20;5(2):101390. doi: 10.1016/j.xcrm.2023.101390. Epub 2024 Feb 9. PMID: 38340724; PMCID: PMC10897544.