Immuno-oncology research encompasses the complex and dynamic relationship between a patient, their immune system, the tumor, and the microenvironment surrounding the tumor. A personalized medicine approach—leveraging biomarkers from genomics, proteomics, metabolomics, or a combination—can anticipate, harness, and modulate immune responses to a cancer, potentially delivering cures for each unique patient.

Most predictive biomarker development in immuno-oncology focuses on immune checkpoint inhibitors (ICIs). ICIs have succeeded cancer vaccines, which were generally disappointing in clinical trials, where they showed poor efficacy and high toxicity. The first ICI, the anti-CTLA-4 antibody ipilimumab, was approved by the FDA in 2011. Ipilimumab was quickly followed by the anti-PD-1 antibodies pembrolizumab and nivolumab and the anti-PD-L1 antibodies atezolizumab and durvalumab. These drugs are now widely used both as single agents and in combination with chemotherapies for about 50 cancers, and thousands of clinical trials assessing T-cell modulators are currently underway.

ICIs present unique challenges to the development of predictive and prognostic biomarkers. ICIs target T-cell responses, which are not simple on-off switches for an antitumor response. Although T cells do attack tumor cells that express certain tumor-specific antigens, multiple signals are required for this attack, and the T cells must overcome defenses mounted by tumor cells. Also, T cells can run afoul of inhibitory pathways, leading to short-lived T-cell responses and spared tumors.

ICIs could be more successful if their use were guided by predictive biomarkers, that is, by biomarkers capable of providing a readout on a tumor, facilitating the selection of therapeutic agents, and tracking the evolution of the tumor’s response to therapy. Predictive biomarkers could even indicate whether agents should be deployed singly or in combination.

Predictive biomarker–based tests have been approved by the FDA as clinical diagnostics for pembrolizumab. There are, for example, clinical diagnostics for pembrolizumab that assess tumors for PD-L1-immunohistochemistry (PD-L1-IHC), microsatellite instability-high (MSI-H), or tumor mutational burden-high (TMB-H). In several instances, the clinical diagnostics were developed by companies working in partnership with pembrolizumab’s manufacturer, Merck & Co.

Mohini Rajasagi, PhD, senior director, Oncology Translational Science and Clinical Biomarkers, Merck & Co., says that her company has two patient-centric goals for pursuing biomarkers within translational oncology: ‘’The first is to select patients most likely to derive maximum benefit from pembrolizumab and our other oncology assets. The second fits under our umbrella of exploratory biomarkers.

“Exploratory biomarkers, which include molecular correlates and gene signatures, allow us to continue to advance our understanding of the mechanism of action for pembrolizumab and other oncology assets. Such exploration may help identify patients that may benefit from pembrolizumab monotherapy versus a combination of drugs.”

Finding the sweet spot for T-cell activation

At IGM Biosciences, IgM antibodies are being engineered to realize various therapeutic applications. For example, the company is developing IgM antibodies that could serve as therapeutic bispecific T-cell engagers.

Cancer Biomarkers diagram
At IGM Biosciences, the lead product candidate is an engineered bispecific IgM antibody called IGM-2323. Compared with bispecific IgG antibodies, IGM-2323 may limit supraphysiologic stimulation of T cells, potentially improving safety and tolerability, and preserving and strengthening T-cell activation.

Naturally occurring IgM antibodies possess 10 binding domains, whereas IgG antibodies possess 2 binding domains. In general, IgM antibodies have higher avidity (overall or total binding strength) than IgG antibodies, albeit lower affinity (binding strength on a per-bond basis). IGM’s IgM antibodies, however, are engineered to possess high avidity and high affinity, resulting in stronger binding to a cell surface target.

IGM’s goal with T-cell engagers is to kill cancer cells by T-cell-directed cellular cytotoxicity, complement-dependent cytotoxicity, and interferon gamma (IFN-γ)-dominant immune stimulation. The immunity induced by IgM’s bispecific T-cell engagers differs from conventional endogenous immunity, where antigen presentation by an MHC peptide complex culminates in T-cell activation.

“With bispecific T-cell engagers, what you’re really leveraging is synthetic immunity,” explains Genevive Hernandez, PhD, director and head of clinical biomarkers at IGM Biosciences. “With synthetic immunity, the activation of the T cell is induced by an antibody that recognizes a target expressed on the surface of a cell and at the same time binds to CD3 expressed on T cells.”

IGM has a bispecific T-cell engager in Phase I clinical development for non-Hodgkin’s lymphoma and other CD20-expressing malignancies. This bispecific T-cell engager, which is designated IGM-2323, is also in several preclinical programs in other cancers. Biomarkers for these therapeutics need to address three aspects of the system: the T cells, the tumor, and the tumor microenvironment. “We are looking for biomarkers that let us know that the T cells have been ‘tickled’ or activated,” says Hernandez, “and what they do when activated.”

That process of activation involves the production of many different types of cytokines, which are important for T-cell activity but may also bring about unwanted toxicity in the form of cytokine release syndrome or even lead to less immune function over time. The IFN-γ-dominant pattern of cytokines observed with IGM-2323 suggests that there is a “sweet spot” for T-cell activation without overstimulation.

“When I was first doing work on T cells, it was always about inducing the maximum response,” Hernandez recalls. “But now the field is starting to have a greater appreciation that the maximum response is not always the best response. There is really something to being able to tailor T-cell responses.”

Real-time data with CTCs

Biomarker strategies typically focus on small and large molecules found in blood, serum, or tissue biopsy. A complementary and potentially powerful approach is being explored by Catherine Alix-Panabières, PhD, associate professor and director, Laboratory of Rare Human Circulating Cells (LCCRH), University Medical Center of Montpellier, France.

At the University Medical Center of Montpellier, researchers are working to improve patient stratification for immunotherapy by deriving more information from liquid biopsies. For example, they are looking at circulating tumor cells (CTCs), PD-L1-positive CTCs, circulating tumor DNA, exosomes, and tumor-educated platelets at baseline and during immunotherapy for non-small cell lung cancer.

Alix-Panabières is focusing on circulating tumor cells (CTCs). She says that a liquid biopsy measuring CTCs could help stratify patients receiving cancer immunotherapy and provide real-time data on how patients are responding. Such information could enable a quick shift to a different form of treatment.

Alix-Panabières and her team reported for the first time in 2015 that CTCs can express PD-L1 and escape the immune system in breast cancer. The researchers followed up by demonstrating the prognostic significance of PD-L1 expression on CTCs in patients with head and neck squamous cell carcinoma. Alix-Panabières is currently carrying out clinical trials using CTCs and PD-L1-positive CTCs as prognostic biomarkers in metastatic breast cancer and metastatic non-small cell lung cancer (NSCLC).

“The key message of these two projects was that there was a clear discrepancy between the detection of PD-L1 in the tissue biopsy and liquid biopsy,” Alix-Panabières maintains. “In both cases, only the liquid biopsy could predict the clinical outcome of the patients, whereas PD-L1 expression in the tissue biopsy did not.”

Her group is bringing in additional data from circulating tumor DNA (ctDNA) and exosomes from the corresponding plasma of these studies as well as additional blood samples collected during recruitment for a new clinical trial in metastatic NSCLC looking at CTCs, PD-L1-positive CTCs, ctDNA, exosomes, and tumor-educated platelets at baseline and during immunotherapy.

According to Alix-Panabières, there is an urgent need for interventional clinical trials focusing on the clinical utility of liquid biopsy: “We need to propose personalized medicine to improve the strategy for cancer management—which drug, at which moment, for which patient?”

A unique tumor signature

Many patients on ICIs could benefit from a complementary therapy that is mechanistically distinct. Such a complementary therapy could be effective against an ICI-resistant tumor or induce a stronger response right from the beginning of therapy.

A candidate for such a complementary therapy is JTX-8064, a humanized IgG4 monoclonal antibody developed by Jounce Therapeutics. JTX-8064 is designed to specifically bind to leukocyte immunoglobulin-like receptor B2 (LILRB2/ILT4), an immune-inhibitory protein expressed on the surface of myeloid cells. Binding the receptor inhibits cell activation and promotes immunosuppression.

“JTX-8064 is designed to reprogram immunosuppressive, or M2, macrophages into immunostimulatory, or M1, macrophages, to enhance or restore antitumor immune activity,” says Johan Baeck, MD, senior vice president, Clinical Development and Medical Affairs, Jounce Therapeutics. “We view JTX-8064 as a macrophage checkpoint inhibitor with the potential to reverse PD-L1 inhibitor resistance as well as to further improve outcomes in PD-L1-inhibitor-sensitive tumors.”

Jounce presented preclinical data about JTX-8064 at a virtual meeting, specifically, the annual meeting of the American Association for Cancer Research, which was held last April. The data showed that when JTX-8064 is given in combination with a PD-1 inhibitor, resistance to the PD-1 inhibitor is reversed.

Jounce has developed a proprietary gene signature based on tumor-associated macrophages (TAMs) by analyzing single-cell RNA sequencing data collected in experiments in which seven solid tumor indications were evaluated. Subsequently, the TAM signature and an IFN-γ signature were used to identify tumor types that may be most susceptible to JTX-8064 treatment, alone or in combination with PD-1 inhibitors. In its INNATE Phase I clinical trial, Jounce is assessing JTX-8064 as a monotherapy and in combination with the anti-PD1 agent, pimivalimab, in solid tumors.

The predictive profile

Evisa Gjini-Bood, PhD, director, Solid Tumor Team, Translational Medicine, Bristol Myers Squibb, is working to identify patients for whom immunotherapy would provide the best chance of a positive outcome in high-grade, high-risk prostate cancer. Her group is using deep immune and tumor profiling in high-powered patient cohorts to probe for biomarkers of resistance.

Unlike cancers that are immunologically “hot,” prostate cancer is “cold.” It has proven to be insensitive to recently developed immunotherapies.

“Our internal work and wider efforts to profile the immune and tumor characteristics of prostate cancer” suggested that some patients were immunocompetent and more likely to respond to immunotherapy, says Gjini-Bood. In fact, her team’s investigation turned up a subset of patients with a favorable biomarker profile—moderate CD8 infiltration, sustained major histocompatibility complex class I expression, and lower cancer-associated fibroblast numbers.

These results were obtained from an evaluation of pretreatment, formalin-fixed, paraffin-embedded resection samples in a primary prostate cancer cohort. In addition to whole-exome DNA and RNA sequencing, Gjini-Bood and colleagues used mass spectrometry–based proteomic profiling. The proteomic profiling technology allows the expression of around 5,000 proteins to be detected.

The scientists also considered the spatial context of stromal and immune cell types in relation to the tumor through immunohistochemistry and immunofluorescence protein epitope imaging techniques. Those techniques are flexible enough to identify defined genes, pathways, or cell type biomarkers when contextualized with patient outcomes.

Gjini emphasizes the importance of big data integration, from either tissue or liquid biopsy, for biomarker profiling: “As we can now appreciate, even in the ‘hottest’ tumors, there is no ‘perfect’ biomarker. But the predictive value of a biomarker may increase when combined with multiple candidates across platforms to create a more predictive profile.”

Bringing biomarkers to patients

The field of immuno-oncology is still very young, given that the FDA approved ipilimumab just 10 years ago. “Our time horizon using these clinically is fairly short,” says Genevieve Boland, MD, PhD, section head, Melanoma/Sarcoma Surgery Program, Massachusetts General Hospital (MGH). “I think what makes this exciting is we’re learning in real time.”

Boland notes that she participates in both translational research and patient care. The latter activity, she suggests, brings her closer than many researchers to the lived experience of the patients on the receiving end of cancer immunotherapies.

research on the molecular profiling of melanoma
At Massachusetts General Hospital, Genevieve Boland, MD, PhD, leads research on the molecular profiling of melanoma, the characterization of molecular and immunological changes that occur during immunotherapy, and the identification of circulating biomarkers of cancer. Tumor-banked samples facilitate the multimodal study of tumor-immune interactions. Liquid biopsies permit plasma-based approaches including extracellular vesicle analysis and plasma proteomics.

Her group is searching for biomarkers of immunotherapy response and resistance in melanoma using a combination of modalities including whole-exome sequencing, RNA sequencing, assay for transposase-accessible chromatin with high-throughput sequencing (ATAC-seq), chromatin immunoprecipitation followed by sequencing (ChIP-seq), and multiplex imaging. Using a variety of techniques, Boland says, helps reveal patterns that would otherwise remain obscure.

“As we get down to a single-cell resolution, from a sequencing standpoint, that doesn’t really give you any architecture or any understanding of how the cells are interacting,” Boland explains. Looking forward, she plans to incorporate more spatial transcriptomics and to explore plasma-based approaches including extracellular vesicle analysis and plasma proteomics.

“You can either get lost in the weeds of all this data, or you can try to use it to see what patterns emerge via multiple different ways of looking,” she advises. “If you see consistent patterns across the different platforms, it becomes somewhat reassuring that you’re really detecting some sort of biologically meaningful process.”

One of the important findings of Boland’s work, carried out in conjunction with Nir Hacohen, PhD, at MGH, was that loss of antigen processing and presentation by tumor cells was an important mechanism of resistance in metastatic melanoma. Boland isn’t yet able to translate this and other markers of resistance into use at the bedside just yet, but the team at MGH is working toward developing a CLIA-certified immuno-oncology biomarker panel that can be run through a pathology laboratory.