NEW YORK - When Robert Mach started looking for ways to image PARP-1 more than a decade ago, cancer was far from his mind.
The University of Pennsylvania's Perelman School of Medicine radiology researcher was, instead, developing tools to image a form of cell death called necrosis with positron emission tomography (PET).
Along the way, he hit on the idea of developing a PET tracer targeting PARP-1, a DNA repair enzyme in the nucleus that can also trigger necrosis. After some digging, Mach narrowed in on a compound capable of binding PARP-1 with high affinity and started labeling its analogues with a favored PET isotope, fluorine-18 (F18).
That compound was rucaparib (now marketed as Rubraca by Clovis Oncology).
These days, Mach's PARP-targeting PET tracer — known as [18F]FluorThanatrace, or [18F]FTT — is making its way into the clinical realm, appearing in clinical trials underway at the University of Pennsylvania, MD Anderson Cancer Center, and Washington University. At those centers, investigators are using the tracer to label and PET image PARP-1 in individuals with breast, ovarian, pancreatic, or prostate cancer who are receiving PARP inhibitor monotherapy.
So far, PET images produced with [18F]FTT are not being used to influence whether PARP inhibitors should be given to patients, explained radiation oncologist Lilie Lin, who is leading an ongoing [18F]FTT trial at MD Anderson. Rather, the investigators want to see if PARP-1 expression in tumors coincides with PARP inhibitor treatment responses, survival outcomes, or other clinical features.
If so, it might be possible, further down the road, to incorporate PARP-1 PET imaging into treatment protocols as a method for non-invasively predicting PARP inhibitor response.
"The whole idea with imaging biomarkers is that they are non-invasive, and they can image the entire tumor repeatedly over the course of therapy," Lin explained. In contrast, testing for many other proposed biomarkers are not only invasive — relying on blood or tumor sampling — but also "only truly give a small snapshot," she said, meaning they "can't necessarily capture the heterogeneity of a tumor."
"If we could use a radiotracer non-invasively to tease out which patients can respond and which can't, there would be several uses for that, including not pursuing what eventually will be futile therapy," added Austin Pantel, a radiologist and clinical instructor in radiology at UPenn.
Though PARP inhibitors have taken a somewhat circuitous path to the clinic, several recent trials have reported very promising results in subsets of patients with ovarian, pancreatic, and other cancers.
The drugs are expected to be most effective in patients with BRCA1/2 mutations or other homologous recombination repair defects. Because those cells are already missing a DNA repair arm, the reasoning goes, it should be possible to kill them by throwing a wrench in the PARP-mediated DNA repair gears. That rationale has held up, to some extent, in PARP inhibitor trials, though response is still variable and does not always line up neatly with BRCA mutation status.
At the American Society for Clinical Oncology annual meeting in Chicago earlier this year, for example, University of Chicago gastrointestinal oncology program medical director Hedy Kindler presented results from the POLO trial, which compared post-chemotherapy maintenance treatment using the PARP inhibitor olaparib (Lynparza from AstraZeneca and Merck) to placebo in patients with metastatic pancreatic cancer who had BRCA1/2 germline mutations.
While the progression-free survival times in the PARP inhibitor maintenance arm were stunning for some patients with metastatic pancreatic cancer, not everyone responded to the maintenance drug. The average progression-free survival time was 7.4 months on maintenance olaparib compared to 3.8 months in the placebo arm, with just over 23 percent of those in the olaparib maintenance group showing an objective response to that drug.
"There are assays under development to try to figure out which tumors are most likely to respond [to PARP inhibitors]. But having a marker of activity — kind of a biological assay — would be very valuable," noted Susan Domcheck, director of the MacDonald Women's Cancer Risk Evaluation Center and at Penn Medicine, and executive director of its Basser Center for BRCA, who is not directly involved in the PET tracer project.
If enhanced PARP-1 expression does track with treatment response, Mach's [F18]FTT tracer and others like it — most notably, PET tracers and other PARP-1 imaging approaches based on olaparib and olaparib analogues that have been developed in the US and the UK — may well hold possibilities as a novel way to predict therapeutic benefit. But there are several open questions that still need to be answered.
Lin cautioned that predictive markers have a high bar to clear for regulatory approval and for acceptance into clinical practice, meaning a great deal of patient data is needed from ongoing and future clinical trials.
Even so, there are early hints that PARP-1 expression varies from one tumor to the next — and not always in cancer types or subtypes where such PARP-1 patchiness is anticipated.
For example, Lin was senior author on a paper published in the Journal of Clinical Investigation last year, in which she and colleagues from UPenn used the [F18]FTT tracer to profile ovarian cancer cell lines, mouse xenograft models, and 20 ovarian cancer patients, including cell lines in which PARP-1 was lopped out with CRISPR-Cas9 editing.
That study suggested that PET imaging with [F18]FTT lined up with PARP-1 levels measured by immunohistochemistry, for example. But it also revealed a range of [F18]FTT uptake in the ovarian cancer tumors, shoring up interest in the kind of clinical trials that are now underway to explore PET-based PARP-1 expression profiles in the treatment response setting.
Likewise, in a study involving 30 breast cancer patients with a range of tumor subtypes, University of Pennsylvania Perelman School of Medicine radiology researcher Elizabeth McDonald and colleagues used the [18F]FTT PET tracer to uncover enhanced PARP-1 expression in some tumors types where it was not expected, such as estrogen receptor-positive tumors, as well as in tumor types, such as triple-negative cancers, that are thought to be more responsive to PARP inhibition.
"The most significant and striking finding from that first, focused trial in breast cancer was that PARP-1 expression was independent of breast cancer immunohistochemistry subtype," explained McDonald, who presented the abstract for that research at the San Antonio Breast Cancer Symposium in late 2018.
The imaging data also revealed variable PARP-1 levels in triple-negative breast tumors, prompting speculation that such variability might contribute to the heterogeneous PARP inhibitor responses that have been documented in these cancer cases.
In their JCI study, Lin and her colleagues found that PARP-1 must be present for cells to respond to olaparib and other PARP inhibitors, though it remains to be seen whether higher-than-usual PARP-1 expression boosts sensitivity to such drugs.
See It, Treat It?
A missing drug target is one thing. But does it mean we will see enhanced PARP inhibitor response in tumors with more of that PARP-1 target?
That remains to be seen, according to Ruth Plummer, an oncologist and experimental cancer medicine researcher at Newcastle University.
"What we've never known is, if you've got a tumor which has got PARP up-regulated, are you more likely to respond to a PARP inhibitor?" cautioned Plummer, who was involved in the earliest human trials using rucaparib but is not part of the [F18]FTT tracer team.
On the contrary, she pointed to past research in colon cancers and surrounding tissue samples that saw elevated PARP-1 expression in tumors compared to the unaffected tissues neighboring them.
Moreover, Plummer noted that any tumor that responds to a PARP inhibitor must be reliant on PARP-1 for one reason or another, even if it is not due to BRCA1/2 mutations. That may well be the case for some patients, she said, but, "I just think we haven't got that clinical tie in yet."
To explore such issues, Lin is leading a clinical trial at MD Anderson Cancer Center that will use the [18F]FTT tracer-based PET imaging in three distinct cohorts: patients with primary or recurrent ovarian, fallopian tube, or primary peritoneal cancer; breast cancer patients; and individuals with non-ovarian, non-breast solid cancers.
That team is currently recruiting patients and hopes to enroll 120 participants to look at PARP-1 levels in relation to [F18]FTT uptake while further evaluating the safety of the tracer and tracking its potential relationship to BRCA mutation status and treatment response.
At UPenn's Abramson Cancer Center, meanwhile, Pantel is leading a pancreatic cancer trial that began in 2018 and aims to enroll 30 individuals with advanced pancreatic ductal adenocarcinoma. For that trial, investigators will use the [F18]FTT tracer and PET imaging to take a look at PARP-1 levels in tumors from individuals with prior platinum-based chemotherapy who have gone on to receive a PARP inhibitor for another clinical trial or as part of their standard clinical care.
Pantel and other members of the UPenn team are also participating in several more [F18]FTT trials underway in breast, prostate, and other cancer types at UPenn, MD Anderson, and/or Washington University, which share an exploratory investigational new drug application for the compound from the US Food and Drug Administration for the tracer.
For the moment, investigators are focused on oncology applications for the tracer. Even so, in their initial preclinical tests on [F18]FTT, investigators included individuals with and without cancer in an effort to start profiling broad applications of the tracer, explained Delphine Chen, who was senior author on the first in-human study of [F18]FTT, published in Radiology in early 2017.
Chen, who recently relocated from Washington University to the University of Washington has started talking to oncologists there about potential trials through the Seattle Cancer Care Alliance, and members of the broader collaboration are interested in moving to multi-center trials in the future.