February 15, 2017 (Vol. 37, No. 4)
Lisa Heiden Ph.D. Director of Business Development MyBioSource
Diverse Immuno-Oncology Approaches Call for Diverse Biomarkers
Tumor biomarkers act as important contextual cues or signals of cancer status. “If there are T cells infiltrating your cancer, you are going to do better than if you don’t have them.
For almost every cancer, immune infiltrates are prognostic markers for better overall outcome,” says Bernard A. Fox, Ph.D., CEO of UbiVac and Harder Chair for Cancer Research, Earle A. Chiles Research Institute.
Predictive biomarkers are indicators of whether a patient should get a specific therapy. The use of such biomarkers has been given various names, notes Robert Anders, M.D., Ph.D., associate professor of pathology and assistant professor of oncology at Johns Hopkins University. These names include personalized medicine or individualized medicine.
DNA tumor mutation burden is a predictive biomarker, although RNA alterations must be considered, too. “The androgen receptor (AR) splice variant AR-V7 is an RNA biomarker predicting drug resistance to anti-AR agents. Such agents, which include enzalutamide and abiraterone, represent standard-of-care therapy in castration-resistant prostate cancer,” informs Shidong Jia, M.D., Ph.D., founder and CEO, Predicine.
“Immuno-oncology aims to overcome cancer’s trick, or invisible cloak, by enabling the body’s immune system’s T cells to attack cancer cells,” remarks David Duffy, Ph.D., CTO and vp of research, Quanterix. “Knowing if therapeutic strategies have worked will likely require measurement of immune-response molecules, rather than genetic analysis of tumor cells.”
“After over a decade of preclinical work, researchers uncovered the potential of immune checkpoint blockade by targeting CTLA4 and the PD-1/PD-L1 axis in human tumors,” adds Daniel E. Carvajal-Hausdorf, M.D., postdoctoral associate, Yale School of Medicine.
Each of the preceding comments is offered in anticipation of a conference, CHI Biomarkers for Cancer Immunotherapy. At this conference, which is scheduled to take place February 23–24 in San Francisco, Drs. Fox, Anders, Jia, Duffy, and Carvajal-Hausdorf will elaborate on their insights. In this article, these scientists summarize their most relevant findings.
Extending Gene RADAR’s Range
DNA-based molecular tests have entered the mainstream of cancer diagnostics, but even the most informative DNA-only tests may fall short as guides to immuno-oncology. “In many diseases, progression happens at both the DNA level and the RNA level, and sometimes at the RNA transcription level alone,” explains Dr. Jia.
Predicine’s Gene RADAR (ctRNA and ctDNA single molecule digital Reading technology) measures RNA and DNA concurrently, and it works with liquid biopsies or tissue samples. The beauty of Gene RADAR is its ability to pick up both cancer genetic blueprints (DNA) and functional biology (RNA) in a single test. “The combined DNA-plus-RNA test further boosts detection sensitivity and specificity for genes of interest,” asserts Dr. Jia.
For example, the BRAF inhibitor Zelboraf has therapeutic responses in a majority of melanoma patients carrying BRAF V600 DNA mutations. The DNA-based mutation may be silent in the nonresponders and never transcribed into RNA or mutant protein.
“RNA-based biomarker profiling at the functional level could help authenticate DNA-positive patients, thus increasing the probability of success in biomarker-driven clinical trials,” explains Dr. Jia.
Predicine, which aims to accelerate drug discovery, diagnostic test development, and therapeutic applications, maintains mirrored operations in California and Shanghai. Both operations interface with the Predicine Cloud, an integrated biomarker and cloud-computing platform.
To develop this platform, Predicine drew on the contributions of its team members. Hailing from Guardant Health, Harvard, Illumina, and other leading institutions, Predicine scientists possess expertise in noninvasive diagnostics, precision medicine, cancer immunotherapy, clinical drug development, diagnostic assay development, next-generation sequencing, computational biology, and big data analysis.
RADAR encompasses a wide range of testing panels and is in clinical trials globally. “The PrediSeq-Cancer Immunotherapy panel reflects our scientific expertise in cancer immunotherapy, drug development, clinical trials, bioinformatics, and big data,” states Dr. Jia.
Immunotherapy uses a patient’s own immune system to fight disease, and so immunotherapy outcomes may be predicted by immune markers such as T cells and PD-1/PD-L1 expression. Immunotherapy, however, has another kind of predictive biomarker: DNA tumor mutation burden.
“Mutations must be functional in order to trigger an immune response,” stipulates Dr. Jia. Immune RADAR RNA-based mutation tests, together with Predicine’s immune gene signature panel, are powerful tools to help define the panel content that constitutes a meaningful list for DNA tumor mutation burden.
Documenting Induced Immunity
“Most patients lack an anticancer immune response,” laments Dr. Fox. “And the absence of this response can prevent patients from responding to checkpoint blockade with PD-1/PD-L1.”
“A big question is how to identify combinations of immunotherapy that will increase anticancer immunity and response rates,” he continues. “UbiVac exploits proteasome inhibitor-induced autophagy (cell self-eating) as a strategy for producing vaccines.”
UbiVac is the developer of an immuno-oncology technology called DRibble. This technology packages cancer targets into dendritic cell-targeted micro-vesicles that can educate immune cells to recognize cancer.
DRibble vaccines prime T cells to recognize tumor antigens that otherwise may remain “hidden” biomarkers. DRibbles are autophagosomes containing defective ribosomal products (DRiPs) and short-lived proteins (SLiPs) and other antigenic components that facilitate cross-presentation, including C-type lectin domain family 9A (CLEC9A) ligands, toll-like receptor agonists, damage-associated molecular patterns, and heat-shock protein molecular chaperones. (DRibbles refer to DRiPs- and SLiPs-containing blebs.)
A key aspect of DRibbles is capturing and stabilizing SLiPs as whole proteins. In contrast, endogenous SLiPs are quickly degraded into small peptides and soon become unavailable for cross-presentation by dendritic cells to prime immunity.
The allogeneic, off-the-shelf DPV-001 DRibbles vaccine, in clinical trials, is manufactured from cancer cell lines (UbiLT3 and UbiLT6) induced to undergo autophagy.
“There are over 150 proteins in DPV-001 commonly overexpressed by the average non-small cell lung cancer (NSCLC),” notes Dr. Fox. “Also, there are up to 1,700 neoantigen epitopes or altered peptide ligands.”
Autophagosomes are targeted to CLEC9A+ antigen-presenting dendritic cells, where they are internalized through CLEC9A for antigen processing. Protein arrays and CD4/CD8 T-cell cytokine responses against NSCLC cells are among the assays used to measure DPV-001-induced immune responses.
“Most cancers have many overexpressed genes in common, and data suggests that SLiPs are the dominant epitopes on the surface of cancer cells,” states Dr. Fox. The prevalence of SLiP epitopes may account for the efficacy of the autophagosome strategy in inducing an anticancer immune response in preclinical models and vaccinated patients.
“If you are in a low Immunoscore category, our hypothesis,” says Dr. Fox, “is that vaccines plus stimulatory antibodies against OX40 (CD134) will significantly boost anticancer immunity and lead to tumor regression in patients who don’t respond to anti-PD-1/PD-L1.”
Looking Squarely at Resistance
Hormone-positive breast cancer (60–70% of all breast cancers) has not responded effectively to immunotherapies in the metastatic setting, observes Dr. Carvajal-Hausdorf. The group led by Dr. Carvajal-Hausdorf develops objective assays for immuno-oncology biomarkers. It focuses on ID01, B7-H3, and B7-H4 in anti-PD-1/PD-L1-resistant subtypes of breast cancer that already have indications of response in Phase I studies.
“We start with antibody clones from companies that we trust and subject them to a validation battery,” informs Dr. Carvajal-Hausdorf. “Many of results in the biomarker literature are inaccurate, because antibodies were tested and results extracted without further validation.”
Researchers are responsible for appropriately testing their reagents for the intended use. Dr. Carvajal-Hausdorf’s validation battery for in situ measurement includes positive and negative controls, such as cell lines with inducible or knocked-down expression and tissues with known levels of the target molecule.
Objective, quantitative fluorescence technology developed at Yale measures the signal in a selected area and transforms it into a score to obtain readings on a continuous scale. The measurement is a score in units allowing for a stricter evaluation than classical immunohistochemistry, which has categorical measurements such as 0, 1+, 2+, or 3+. “Nonetheless, visual assessment is still fundamental to ensure validation,” emphasizes Dr. Carvajal-Hausdorf.
After stringent evaluations are made, checkpoint markers and inflammatory infiltrates such as CD3, CD4, CD8, and CD20 are used in curated cohorts representing several cancer types.
The goal is to identify biological differences in various cancer subtypes and determine prognostic value. It is a step toward developing assays for prospective trials that could evaluate people that will go into or have already resisted treatment. “If you don’t have a validated assay that can be evaluated in an objective way, your results and conclusions might not be accurate,” concludes Dr. Carvajal-Hausdorf.
Surveying Tumor Microenvironments
PD-L1 expression is a biomarker of response to PD-1 blockade. “But it is not a perfect biomarker,” cautions Dr. Anders. Patients who express the PD-L1 protein have about a 50% chance of responding to PD-1 blockade therapy, whereas patients who don’t express PD-1 protein have about a 15% chance of responding.
“It’s difficult to predict whether an individual patient is going to respond,” remarks Dr. Anders. “Given the effectiveness of anti-PD-1/PD-L1 therapies, individual patients shouldn’t be excluded from a study based on just PD-L1 expression status.”
A better job of predicting response can be done if more pieces of data are used, such as lymphocyte density in the tumor microenvironment, mutational density, and PD-L1 expression, “especially for colon cancer,” informs Dr. Anders.
In colon cancer characterized by mismatch-repair deficiency, there is microsatellite instability. Polymerase errors are not repaired at an efficient rate, producing an abundance of mutations leading to new epitopes or antigens that are potential stimulants for the immune system. When mutations induce a change in a protein that the immune system sees as nonself, the immune system is more likely to attack it.
“Having the patient’s own immune system remove the cancer is one of the principles of immunotherapy,” reminds Dr. Anders.
Patients with microsatellite instability-associated colon cancer that have lots of lymphocytes infiltrating their tumor, lots of mutations, and PD-L1 expression respond well to PD-1 blockade. In a Phase II trial, albeit with a small number of patients, the objective response rate was 62% and the disease control rate was 92%.
“The whole field of immune cancer therapy is very encouraging, and there’s great excitement in continuing to measure which cancers have which checkpoint inhibitors,” reflects Dr. Anders. “There’s probably at least another dozen checkpoints.”
Beneath the Tip of the ELISA Iceberg
The challenge for drug companies is to establish baseline levels of protein biomarkers in each patient’s blood and to monitor changes in those levels to determine if they are responding to the therapy. “Unfortunately, today they often can’t measure them,” laments Dr. Duffy. The level of the target might simply be too low. He estimates up to 90% of the protein targets people want to measure are part of the ELISA iceberg that is below the surface.
Quanterix’s Simoa (Single molecule array) is a digital ELISA-based immunoassay with incredible sensitivity for detecting proteins. It’s about thousandfold more sensitive then traditional ELISA and can be multiplexed. The Simoa antibody-based technology specifically captures, labels, and counts single molecules in traditional sample types, such as blood, plasma, serum, tissues, and cells.
“You can’t get more sensitive than a single molecule,” exclaims Dr. Duffy.
Simoa brings new opportunities to flow cytometry. After sorting cells into different populations and isolating single cells, you might try breaking them open and analyzing what’s inside with Simoa.
The Simoa HD-1 Analyzer™ is a fully automated sample-in to results-out instrument with high precision and lab-to-lab reproducibility. It’s designed to take markers from biomarker discovery through industry validation and drug development to clinical trials. Quanterix partnered with bioMerieux to take the platform toward FDA approval.
“The goal is to expedite FDA approval of biomarkers that are discovered in research,” reminds Dr. Duffy.
Many FDA-approved tests start as laboratory-developed tests (LDTs). “LDTs are where a lot of groundbreaking work takes place, and we have customers that are setting up LDTs around Simoa,” says Dr. Duffy. Customers use Homebrew Kits to develop custom assays with their own antibodies.
In drug development, people who are looking at specific epitopes cannot tolerate interference between the drug and measurement of the protein. “Flexibility in building assays is required, and flexibility is one of the unique features of Simoa technology,” asserts Dr. Duffy. “We provide training and help people develop their assays.”