NEW YORK – Findings from a recent biomarker analysis within the Phase III Orient-11 trial of Eli Lilly and Innovent Biologics' sintilimab (Tyvyt) for non-small cell lung cancer add to a growing collection of research suggesting that, across multiple cancer types, expression of the antigen-presenting molecules major histocompatibility complex (MHC) class I and class II may be useful biomarkers for guiding immune checkpoint inhibitor therapy, especially when combined with other established biomarkers.
The Orient-11 biomarker analysis, published in the Journal of Thoracic Oncology, considered several potential biomarkers that might predict survival benefit from the anti-PD-1 agent sintilimab plus chemotherapy as first-line treatment for advanced or metastatic non-small cell lung cancer.
While the researchers probed established NSCLC immunotherapy biomarkers such as PD-L1 expression in their analysis, they also used whole-transcriptome sequencing of baseline tumor samples to analyze antigen presentation pathway expression — including expression of MHC class I (MHC-I) and MHC class II (MHC-II) — in the tumor microenvironment.
The Orient-11 trial randomized 397 patients to receive either chemotherapy plus sintilimab or chemo plus a placebo. Among patients with high MHC-II-related gene expression, the investigators observed significantly longer progression-free survival times and overall survival times with the chemo-sintilimab combination therapy than they did among patients with low MHC-II-related gene expression. Among patients who received only chemo, on the other hand, patients with high and low MHC-II-related gene expression experienced comparable progression-free survival and overall survival times. The researchers observed a similar phenomenon, though to a lesser extent and only for progression-fee survival, with MHC-I-related gene expression.
"These findings support the development of composite biomarkers consisting of PD-L1 and MHC-II expression to enlarge the patient population that benefits from such a combination treatment, but also could potentially spare toxicity from a patient population with limited treatment benefit," wrote the Journal of Thoracic Oncology study's authors, including Li Zhang of China's Sun Yat-sen University Cancer Center.
MHC analysis excluded from industry trials
For the time being, Eli Lilly and Innovent haven't indicated that their MHC biomarker findings from Orient-11 will play into their regulatory strategy for sintilimab in NSCLC. Indeed, sintilimab is already approved for an all-comer population of NSCLC patients in China and has a Biologics License Application under review with the US Food and Drug Administration for the same indication.
That said, the findings are significant in that few other randomized, registration-directed Phase III checkpoint inhibitor immunotherapy trials incorporate MHC biomarker analyses. To the study authors' knowledge, their biomarker study of the Orient-11 trial is the "first to explore the association between gene expression profile in the tumor microenvironment and the efficacy of immune-combination treatment in NSCLC patients based on a large, randomized Phase III trial."
In cases where these large, registration-directed Phase III checkpoint inhibitor studies — including those evaluating blockbuster anti-PD-1/PD-L1 drugs like pembrolizumab (Merck's Keytruda) or nivolumab (BMS' Opdivo) — incorporate biomarker analyses at all, they tend to focus only on biomarkers associated with regulatory approvals, including PD-L1 expression, DNA mismatch repair deficiency, microsatellite instability, or tumor mutational burden (TMB).
The predictive ability of these biomarkers has been validated to the extent that they are incorporated into FDA-approved immunotherapy indications — albeit sometimes controversially — and used across the board to guide treatment strategies for solid cancers.
These biomarkers have presented challenges, in part because there are many patients who fall outside of the FDA-approved cutoff thresholds for PD-L1 or TMB and still respond to the drugs, as well as many patients who fall within the threshold and derive minimal benefit with significant added toxicity.
Researchers like Jessica Dal Col of University of Salerno, who was not involved in the sintilimab study but whose work has focused on the human leukocyte antigen system — including MHC-I and MHC-II — as an immune checkpoint inhibitor biomarker, have expressed frustration over the fact that large Phase III studies seldom incorporate MHC biomarker analyses.
"Few clinical studies have been investigating the potential role of HLA class I and II antigens in predicting the efficacy of immune checkpoint inhibitor-based immunotherapy, and no large clinical cohort analysis of patient population has been performed," Dal Col and colleagues wrote in a 2020 International Journal of Molecular Sciences paper.
In Dal Col's view, this absence of MHC expression analyses in large-scale trials isn't due to the biomarker's novelty since researchers have understood the science for quite a while. "The role of HLA class I antigens in tumor immune escape has been well-known for many years and was probably one of the first mechanisms discovered to contribute to loss of antigenicity in tumor cells," she said, explaining how the mechanism behind the biomarker association makes sense.
"[Checkpoint inhibitor therapy] is based on the activation and re-activation of T cytotoxic lymphocytes capable of recognizing tumor-associated antigens expressed by tumor cells after their processing and presentation through HLA class I complex … this consideration simply and clearly indicates that if the tumor cell does not express HLA class I, its antigens are not presented and thus it is not recognizable by T cytotoxic lymphocytes," she said.
Despite this well-known role of MHC molecules, Dal Col suggested that perhaps unsuccessful attempts to use it as a be-all and end-all single biomarker in past studies may have contributed to its falling to the wayside in late-phase trials.
"Likely in the past, in the different immunotherapeutic approaches used, the evaluation of HLA class I as a single biomarker did not give the expected results, [and] as sometimes happens, we tend to abandon the old road completely for the new one, forgetting the relevance of previous discoveries," she said.
Though MHC markers have been excluded from many industry-sponsored registration-directed trials, a growing compendium of recent research has begun to incorporate these biomarkers into earlier studies across different tumor types. Just last month, for instance, a group of researchers published a paper in Clinical Cancer Research in which they detailed an analysis of samples taken from early-phase breast cancer studies and suggested that MHC-II expression does a better job of predicting response to checkpoint inhibition in early-stage HER2-negative breast cancer than does PD-L1 expression.
Meanwhile, in Dal Col and colleagues' International Journal of Molecular Sciences paper, researchers highlighted several retrospective analyses of tumor samples taken from patients participating in these trials, including from the BMS-sponsored CheckMate-064 and CheckMate-069 studies evaluating nivolumab and BMS' anti-CTLA4 drug ipilimumab (Yervoy) for patients with metastatic melanoma. The retrospective analysis showed that MHC-I expression was a reliable predictive biomarker of tumor response to the anti-CTLA4 agent but not to the anti-PD-1 agent, whereas MHC-II expression might represent a useful predictive biomarker for nivolumab but not ipilimumab.
Dal Col explained why the two HLA classes could have different response associations with monotherapy and combination immunotherapy; all human cells, including tumor cells, express class I, she said, making class I a fundamental part of the effector phase of the immune system's T cytotoxic lymphocytes recognizing and killing tumor cells. This is the phenomenon that anti-PD-1 monotherapy is designed to correct. However, the other phase of immune response targeted by immune checkpoint inhibitor therapy — particularly anti-CTLA4 drugs — is the "priming" phase, which is necessary for the T cell to interact with the antigen-presenting cells that express HLA class II. "So, both class I and II are important," she said. "It depends on which phase of the immune response we chose as [an immune checkpoint inhibitor] therapy target."
Elena Shklovskaya of Macquarie University in Australia, together with colleagues, also authored recent studies focused on this phenomenon. Specifically, in a study published last year in Cancers, and a follow-up review published more recently in the International Journal of Molecular Science, Shklovskaya and colleagues detailed how — and the potential mechanistic reasons why — downregulation or loss of MHC-I molecules on melanoma cells is associated with lack of response to anti-PD-1 monotherapy but not necessarily anti-PD-1 plus anti-CTLA4 combination immunotherapy.
The researchers used flow cytometry to evaluate the expression of the antigen-presenting molecule MHC-I on melanoma cells from samples collected from 36 stage III and IV melanoma patients treated with immune checkpoint inhibitors at Melanoma Institute Australia. The fact that patients with low MHC-I-expressing melanoma did not benefit from anti-PD-1 monotherapy but did derive benefit from combination immunotherapy led the researchers to conclude that their study "provides a rationale for the selection of combination immunotherapy as the first-line treatment in MHC-I low melanoma."
While Shklovskaya and colleagues' and Dal Col and colleagues' research, along with the new Orient-11 biomarker analysis, collectively support using MHC expression to inform immunotherapy decision-making, the researchers behind these studies all ultimately suggested that the best strategy would be to take a composite approach, using this biomarker alongside more established biomarkers like PD-L1 expression and TMB to complement their predictive capabilities rather than replace them.
"I don't ever think there will be a time when one biomarker will allow a clinician to guide their clinical decisions in terms of their choice of immunotherapy," Shklovskaya said. "We also need to take other things into consideration."
Dal Col echoed this sentiment. "Sure PD-L1 and TMB are biomarkers strictly connected to ICI immunotherapy, but the biology is more complex," she said. "Often, we can obtain the right answer if we take into consideration all components of a pathway and all steps of a specific process."
The researchers' support for multi-biomarker approaches mirrors what Zhang and colleagues concluded in their analysis from the Orient-11 study in NSCLC: that their findings support the value of composite biomarker analysis.
Next steps, further research
Absent prospective MHC analyses in large-scale randomized immunotherapy trials, Dal Col and Shklovskaya both felt the best thing for researchers to do in the meantime is to continue publishing studies validating the biomarker approach to the extent they can.
"The simplest thing that we can do in the near future is to perform retrospectives studies and in silico studies based on transcriptomic or proteomic data and try to obtain information regarding HLA class I and II expression's [correlation] with immunotherapy response," Dal Col said.
For their part, Shklovskaya and her team are currently taking their observations from the retrospective analysis of melanoma patients to animal model studies. Within their published research — particularly Shklovskaya's recent International Journal of Molecular Sciences review — these researchers also began to lay out potential therapeutic strategies for overcoming MHC-I deficiency and restoring response to checkpoint inhibitor monotherapy. These strategies, including combination therapies with targeted agents, would need further validation and represent an area ripe for further research to improve patients' outcomes on immunotherapy.
Whatever the next steps may be toward encouraging the incorporation of these biomarkers into larger scale clinical trials, there is a current consensus that the field needs to refine its approach to selecting patients for checkpoint inhibitor immunotherapy. These treatments, while lifesaving for many patients, still bring heightened risks and minimal benefit to a significant population of patients. This is especially true of checkpoint inhibitor combinations, which have high adverse event risks and can be financially taxing.
"To have an informed clinical decision is better than what we currently do in clinical trials, which is often randomizing patients into receiving whatever combination therapy is being offered, particularly as a second or third line of treatment," Shklovskaya said. "We have to see that the combination immunotherapy is much more toxic than monotherapy with a higher number of patients experiencing adverse events. [We should] make a decision early on as to whether a patient should receive monotherapy versus combination immunotherapy."