Predicting Response to Immunotherapy: PD-L1 as a Biomarker and Beyond

Immunotherapy is a type of treatment that boosts the ability of the immune system to find and kill cancer cells. Immunotherapy drugs such as ipilimumab, nivolumab, pembrolizumab, relatlimab, and atezolizumab are FDA approved for use in advanced melanoma and have revolutionized the treatment of these cancers.

Some of these immunotherapy drugs (nivolumab, pembrolizumab, and atezolizumab) target molecules called PD-1 or PD-L1. When PD-1 (expressed on immune cells) binds to PD-L1 (expressed on tumor cells), the immune cells receive a message to ignore the tumor cells, leading to tumor growth. Immunotherapy drugs that inhibit PD-1 or PD-L1 block this message and thus take the brakes off immune cells, allowing immune cells to do their job, which is to recognize and kill cancer cells. When used alone or in combination with other drugs such as BRAF inhibitors, immunotherapy drugs can be effective in treating advanced melanoma.

However, not all patients respond to PD-1/PD-L1 inhibitors. Biomarkers, which are features in the body that can be measured and give information about a person’s health, are needed that predict a person’s response to treatment. A test that measures biomarkers may predict who will and will not respond to PD-1/PD-L1 inhibitors and would be very valuable for determining who should and should not receive these treatments.

The presence of PD-L1 in a tumor was thought to be a biomarker that would predict a good response to PD-1/PD-L1 inhibitor drugs. So initially, researchers designed a test called PD-L1 immunohistochemistry (IHC) to look for the presence of PD-L1 in tumor tissue. However, PD-L1 IHC results may not provide enough information to predict a response to these drugs in melanoma.¹ Biomarkers that go beyond PD-L1 expression are needed.²

Melanomas have features beyond PD-1/PD-L1 expression that affect the immune system’s response to the cancer. Researchers need to understand how these multiple features interact to determine response to immunotherapy and then develop a test that encompasses multiple biomarkers to better predict a response to PD-1/PD-L1 inhibitors.

• One possible biomarker beyond PD-L1 IHC is tumor mutational burden (TMB), which is the number of mutations in the cancer cell’s DNA. More mutations produce more abnormal proteins that the immune system may identify as foreign.3 The immune system, boosted by PD-1/PD-L1 inhibitors, can invoke a response against these abnormal proteins that results in tumor cell death.
• The presence or absence of one or more immune system molecules (for example, CD8, FoxP3, CD68, and CD163) in the tumor—as well as their level of expression—may be a biomarker of response to PD-1/PD-L1 inhibitors.³,⁴,⁵
• Another possible biomarker is features of the gut microbiome, which is the various species of bacteria in the digestive system and the molecules the bacteria produce. Certain gut bacteria species are associated with a response to immunotherapy in metastatic melanoma.³‍
• Combinations of biomarkers that include PD-L1 IHC, TMB, immune system molecules, gut microbiome features, and others—as well as their level of expression—may be more sensitive for identifying responders and non-responders.³,⁶

MRA is playing an important role in research to identify combinations of biomarkers and tests for these biomarkers to predict a response to immunotherapy for patients with advanced melanoma. Multiple biomarkers from large numbers of patient samples generate large datasets, which makes analysis of the data challenging. As a unique way to analyze large datasets, cancer researchers are learning from astronomers. The work done by astronomers to map stars and galaxies in the universe also generates large datasets. Astronomers use sophisticated methods to store and analyze data about galaxies and stars. In work supported by MRA, dermatology and pathology professor Dr. Janis Taube and her colleagues at Johns Hopkins University are working with an astronomer named Dr. Alexander Szalay. Drs. Taube and Szalay and their team are using the methods and lessons learned from astronomy to analyze these large melanoma biomarker datasets with the goal of developing a test that incorporates multiple biomarkers to predict a response to PD-1/PD-L1 inhibitors.⁴

Another important aspect of biomarker research is the fact that tumor cells, immune cells, and the molecules they make interact with one another in 3D space. Drs. Taube and Szalay and their team are also learning how to analyze these 3D relationships and incorporate this information during development of a test to predict who will respond to PD-1/PD-L1 inhibitors.⁴ The platform the team is developing for this purpose is called Astropath. New potential biomarkers can be added to the system as they are discovered.⁷ Spatial representations of the interactions between tumor cells and immune cells may serve as a biomarker for immunotherapy.⁵ For example, the density of CD8-positive, FoxP3-positive immune cells is related to a good response to treatment with PD-1 inhibitors.⁴ Another example is that the presence of CD163-positive, PD-L1-negative immune cells along with a high density of PD-L1-negative tumor cells identifies patients with a poor long-term outcome.⁴ A better understanding of the spatial organization of multiple immune factors in melanoma will allow for improved patient selection for PD-1 inhibitors, the rational combination of PD-1 inhibitor drugs with other therapies, and more precise selection of immunotherapy for an individual patient (that is, precision immune-oncology).⁴

Dr. Taube stated that her team is partnering with industry to move the diagnostic test into clinical care to benefit patients. Because it is not practical to perform testing for dozens of biomarkers for every patient, they have narrowed down a short list of biomarkers that is most informative and actionable. Their research has shown that approximately six biomarkers provide the best information. The team is also looking at the value of incorporating the status of the microbiome into the information used to select treatment.

The team is currently engaged in MRA-funded efforts to validate the six biomarkers, establish a threshold for the levels of each biomarker that distinguishes response from non-response, ensure that results are reproducible across institutions and over time, and assess other quality control measures. Dr. Taube anticipates beginning a clinical trial to study the test they are developing in approximately 3 years.

Dr. Taube would like patients to know that this test is coming. Not only do they anticipate that the test will determine who is likely to respond to PD-1/PD-L1 inhibitor treatment, but the test may also help select the best PD-1 inhibitor and who is likely to respond to combinations of immunotherapy drugs. Dr. Taube stated “we appreciate the patients who have consented to allowing us to use their tumor samples for research.” These studies are expected to benefit many patients and help match the right patient with the right treatment.

‍