Dr. Ezra Cohen discusses Abstract 103 during Oral Abstract Session C.
“Prostate cancer is the second most common cancer in men,” Eileen Parkes, MD, PhD, of Queen’s University Belfast, in Northern Ireland, who presented results of the new research on January 27, said. She noted that there are more than 300,000 deaths from the disease each year, and there are limited early-stage treatments. Even as checkpoint inhibitor approvals have expanded through a number of malignancies, they have shown minimal impact in studies of prostate cancer. Dr. Parkes noted that the tumor mutational burden is quite low compared to that of cancers where immunotherapies have been found to be more effective.
“All [available] treatments have a limited effect,” she said. “There is a need for new, active treatments for prostate cancer, where we still have a chance to cure our patients.”
The new research Dr. Parkes presented aimed at identifying subsets of patients with prostate who may derive more benefit from immunotherapy than an unselected population does. Her group previously developed and validated a 70-gene signature (PCM assay) that can identify a molecular subgroup of patients with primary prostate cancer using biology that is similar to metastatic disease.
Patients who are PCM-positive have an increased risk of developing metastatic disease, as well as for biochemical recurrence. The assay was further validated using prostate biopsy samples. “These are patients with aggressive prostate cancer that relapses quickly, who should receive more treatment upfront,” Dr. Parkes said.
The researchers conducted an integrative clustering analysis within a cohort of primary prostate cancers from The Cancer Genome Atlas, including 488 matched samples. This was done using RNA-seq, and examined promoter methylation, somatic mutations, and copy number variation data. Samples were scored using the PCM assay.
The clustering analysis divided the samples into four subgroups (C1, C2, C3, and C4). C1 and C4 samples had high PCM scores, and these represented approximately 42% of all of the samples. The clusters also had marked promoter hypermethylation, and somatic mutations of genes known to be involved in aggressive prostate cancer.
The C4 cluster was also marked by overexpression of immune genes, Dr. Parkes said. This group represented approximately 17% of all prostate cancers.
When cells are DNA damage repair deficient, she added, the damaged DNA is sequestered into the cytoplasm instead of being repaired. This results in activation of the cGAS–STING pathway, and subsequently in the activation of immune genes and upregulation of immune checkpoints.
A second assay, an immune-based DNA damage repair deficiency assay (DDRD), was also applied to the cohort. C4 samples showed both high PCM and DDRD scores. These were genomically unstable, with high 8q gains, and 38% of the C4 samples have TP53 mutations whereas only 7% of C1 samples did.
Importantly, the analysis revealed an overexpression of immune response genes in C4 compared to C1. C4 samples showed elevated leukocyte fraction, and expression of immune checkpoint genes, at levels that were significantly higher than C1 (p < 2x10-6).
The C4 subgroup, or the “metastatic immune” subgroup of patients, was associated with a poor prognosis. Compared to all other clusters, patients in the C4 subgroup had a hazard ratio for mortality of 2.586 (95% CI [1.392, 4.803]; p = 0.003). Other factors such as Gleason score were not associated with the differences in outcome.
An analysis of a series of immune checkpoints showed consistent, significant upregulation of these targets in the C4 patients. This included PD-L1, CTLA4, STING, IDO1, TIM3, and others.
“We have demonstrated the identification of a group of patients with early prostate cancer who are at increased risk of developing metastatic disease and present with targetable immune biology,” Dr. Parkes said. “This may represent a viable target population for immune checkpoint and DNA damaging therapies in prostate cancer.”
Ezra Cohen, MD, PhD, of the University of California, San Diego, was the discussant for the abstract. “It does appear that DDRD […] defined a subset that may be responsive to checkpoint inhibition,” he said, although he noted that this may not necessarily be anti–PD-1 therapy, since there are other emerging data suggesting other checkpoints may be more effective in the prostate cancer setting.
He also pointed out that there could be other explanations for the lack of response to immunotherapy in prostate cancer, including possible epigenetic changes to T cells themselves that drive the lack of responsiveness to checkpoint inhibition. Dr. Cohen said this adds to a growing understanding that a more personalized, precision approach is needed in immunotherapy.
“We cannot apply the same tool to every patient just because their histology appears to be similar under the microscope,” he said.