The largest study to date on hypermutated gliomas has delivered new insights into their origin, as well as their response to several different treatments.
Specifically, even though they are hypermutated, such tumors are unlikely to respond to PD-1 blockers.
“Gliomas, for whatever reason, don’t have a high indel” – insertions and deletions – “burden, even when they have a high point mutation burden,” Keith Ligon told BioWorld.
“Many point mutations, which are single base changes, are useless” for providing neoantigens, he added. “What you need is bigger protein changes.”
And it is indels that are most likely to change protein structure in a way that allows the immune system to recognize them as neoantigens, providing targets for T cells.
Ligon is chief of neuropathology at Brigham and Women’s Hospital and the senior author of the paper reporting the findings, which appeared in the April 16, 2020, issue of Nature.
Understanding why some tumors respond to immunotherapy and others don’t is among the most pressing questions in immune oncology.
A high tumor mutational burden (TMB) was one of the first biomarkers associated with patients who benefit from checkpoint blockade. In 2017, Keytruda (pembrolizumab, Merck & Co Inc.) became the first cancer drug approved for treatment of tumors with deficiencies in mismatch repair and high microsatellite instability (MMRd/MSI-H), which lead to high TMB, regardless of the anatomical site of the tumor.
Even tumors that have high TMB, though, have a less than even chance of responding to checkpoint blockade. And experience with other oncogenes, like BRCA, shows that because different tissues differ in their overall gene expression, the same mutations can have different effects, and different treatment sensitivity, depending on anatomical site.
Keytruda’s tumor agnostic approval, which was granted on the basis of a basket trial with 149 patients, did not include any glioma patients, and anecdotal evidence has suggested that, unlike some other hypermutated tumors, hypermutated gliomas typically do not respond to checkpoint blockade.
First authors Mehdi Touat “started the project, and was excited about it, partly because he knew there was a disconnect” between the fact that temozolomide treatment can lead to hypermutated tumors and the anecdotal evidence that checkpoint blockade is not particularly effective in treating them.
Treatment with the chemotherapy temozolomide can pave the way to recurrence of tumors that are both hypermutated and treatment-resistant.
However, it has been unclear how temozolomide treatment leads to hypermutated tumors, which can result from deficiencies in several different pathways.
Also unknown was whether those hypermutated tumors would become sensitive to checkpoint blockade.
“There’s a lot of excellent prior publications on this,” Ligon said. But hypermutation in response to temozolomide is a rare event in a tumor type that is rare to begin with, and “none of them were of the proper size… to make any sound conclusions.”
In the work now published in Nature, which looked at 500 hypermutated patients and more than 10,000 total tumors, “for the first time, we had enough genomics on lots of patients” to draw such sound conclusions.
“It’s not just an observation,” he said. “It’s a statistical answer.”
In their experiments, the team showed that temozolomide treatment exerted selective pressure that gave a leg up to cells with deficiencies in mismatch repair (MMR) genes.
MMRd “is an acquired resistance mutation, like in targeted therapy with kinase inhibitors,” Ligon said. “There are even hotspot mutations.”
The work shows the importance of repeat biopsies over the course of treatment. Hypermutated, temozolomide-resistant tumors retained their sensitivity to the chemotherapy drug lomustine (CCNU), suggesting that the onset of hypermutations and temozolomide resistance should prompt a switch in treatment strategies, and that combining temozolomide and CCNU might be able to prevent the emergence of hypermutation in the first place.
The team also did a retrospective analysis of hypermutated glioma patients that were treated with checkpoint blockers at the Dana-Farber Cancer Institute and other sites, and found that treatment with checkpoint blockers did not improve their overall survival.
Retrospective analyses, although they are an improvement over case studies (sometimes derided as “anecdata”), cannot replace clinical trials. And Ligon predicted that “there probably will be responders [to checkpoint blockade], but we don’t know how to identify them yet.”
Given the potential side effects and the high cost of checkpoint blockers, however, the results argue against their default use in hypermutated glioma patients.