NEW YORK – Artificial intelligence-powered biotech Berg said on Monday that the US Food and Drug Administration has given the company the go ahead to advance its investigational agent BPM31510 into a Phase II clinical trial as a neoadjuvant treatment for glioblastoma multiforme (GBM).
In the Phase II trial, which will enroll roughly 50 patients with newly diagnosed GBM, Berg will evaluate the survival benefit of BPM31510, a molecule designed to correct cancer cell metabolism, thereby restoring cell death. In the study, patients will receive BPM31510 in combination with vitamin K1 before consecutive treatment with radiation and temozolomide. The company expects to dose the first patient with BPM31510 in the coming weeks. The study's primary endpoint is overall survival, and its secondary endpoints are progression-free survival and tolerability.
In the process of evaluating these outcomes, Berg will also use its AI platform to home in on biomarkers that may predict which patients derive the most benefit from the treatment. According to Berg CEO Niven Narain, the company will further develop the findings within the Phase II trial and apply a biomarker to select patients in Phase III trials.
Framingham, Massachusetts-based Berg plans to conduct multi-omics analysis and use its artificial intelligence platform, dubbed Interrogative Biology, to search for predictive biomarkers. During the course of the trial, which will take place at Stanford University, Yale University, Mount Sinai School of Medicine, and Cedars-Sinai, among other medical centers, Berg will collect samples from patients at various timepoints during their course of treatment. Narain explained that these samples will be shipped to Berg's in-house multi-omics profiling laboratory, which includes a mass spectrometry facility with the instrumentation to conduct genomic, proteomic, lipidomic, and metabolomic analysis.
Berg will pair this multi-omics information with patients' clinical outcome data and build what Narain called an "AI map" for every patient in the trial. Within these maps, Narain explained, researchers will identify the differing biological elements between patients who responded particularly well to BPM31510 and those who didn't and identify potential biomarkers of response.
"We'll be able to see what was different or overexpressed or just expressed in the subgroup of patients who did really well and ask, 'What are those markers?'" Narain said. "And then we can substratify it and say, 'Ok, these are the biological pivot points of responses in patients who do well.'" According to Narain, this AI-enabled process will be useful for not only identifying response biomarkers, but also biology predictive of treatment toxicity.
"Berg really prides itself on using AI and biomarkers and omics to help guide the decision points of how we develop the drugs, who we give the drugs to, and how we make decisions commercially," Narain said. "It really is a robust use of the combination of biology and AI."
Berg is also evaluating BPM31510 in solid tumors beyond GBM and has received orphan drug designation for the agent as a treatment for pancreatic cancer, for which the drug is in Phase II trials. GBM and pancreatic cancers, Narain pointed out, are tumor types that historically have not benefitted from precision medicine approaches as much as some other cancer types have, such as non-small cell lung cancer.
"Let's not parse our words here … pancreatic cancer and brain cancer are pharmaceutical graveyards," Narain said. "The drug development in these two areas are just that much more inefficient." These tumor types also tend to be aggressive, a feature that Narain said is closely tied to cancer cell metabolism, making them potentially strong targets for BPM31510.
"Cancer cells love low-oxygen environments," he explained. "They want to operate in anaerobic respiration, but in return, the BCL2 protein family, which enables a cell to undergo apoptosis, gets turned off, and that's what allows these tumor cells to grow so quickly."
This process, known as the Warburg effect, is disrupted by BPM31510. "The more efficient a cell is in getting to a Warburg state, the more aggressive that tumor will be," Narain continued. "What BPM31510 does is … causes the mitochondria to be shifted back on … It's flipping the cancer cells' metabolism, thereby having an effect on their ability to become sensitized to apoptosis again."
Given this mechanism of action — described by Narain and co-authors in a Scientifics Reports paper last year — it would make sense that BPM31510 would have a greater chance of success treating aggressive, highly metabolic cancers like GMB and pancreatic cancer, as opposed to slower-growing cancers. But Berg didn't have proof of this back when it began evaluating the agent in a tumor-agnostic Phase I basket trial.
Applying the same AI-based approach to the basket trial results, Berg selected the cancer types in which to further develop BPM31510. "That's why we pivoted from our Phase I study that comprised all-comer solid tumors," Narain said. "As we did that analysis on … samples and markers from patients, we realized that the drug worked better in cancers that are highly aggressive. We enriched our strategy."
Now, in the Phase II trial, Narain is hoping that Berg's AI-enabled mapping strategy will identify a biomarker-enriched development path for the drug. In parallel, the company plans to build out companion diagnostic assays to test patients for eligibility.
The company will also conduct validation studies for the AI computational maps that come out of the trials. Narain said Berg has already made progress in this regard within BPM31510's pancreatic cancer program, for which two protein biomarkers with the potential to predict for pancreatic cancer patients' responses to BPM31510 have been identified using the Interrogative Biology platform. These data were presented during the 2020 European Society for Medical Oncology (ESMO) conference. "Because of what we've done with the Phase II study in pancreatic cancer … that's why I'm so excited about brain cancer," Narain said. "These cancers are unmet needs."