Killing Cancer With Synthetic Lethality

By Deborah Borfitz 

December 13, 2022 | A key member of the team that discovered the BRCA2 breast cancer susceptibility gene in 1995 is now enamored of the idea of taking on cancer with synthetic lethal therapies, a nontraditional approach that exploits the weaknesses of cancer cells. The promise of the treatment concept was the topic of the Heath Memorial Award keynote address delivered at the recent Leading Edge of Cancer Research Symposium of the University of Texas MD Anderson Cancer Center by Alan Ashworth, Ph.D., president of the University of California, San Francisco (UCSF) Helen Diller Family Comprehensive Cancer Center. 

Ashworth holds patents jointly with AstraZeneca on the use of poly (ADP‐ribose) polymerase (PARP) inhibitors, which effectively kill tumors defective in the BRCA1 or BRCA2 genes through synthetic lethality. PARP inhibitors are now approved by the U.S. Food and Drug Administration (FDA) for the treatment of ovarian, breast, pancreatic, and prostate cancer.  

The backstory here is that the BRCA2 gene discovery led Ashworth down a research path focused on DNA repair and, subsequently, the finding that the gene was critical for the repair of double-strand breaks in breast cancer, he says. “That turned out to be critical for understanding synthetic lethal approaches.” 

It was well appreciated early on that BRCA-mutant cells may have vulnerabilities, for example cross-linking of DNA that likely underlies the sensitivity of ovarian cancer to platinum-based therapies, says Ashworth. In a clinical trial in mutated and triple-negative breast cancer later published in Nature Medicine (DOI: 10.1038/s41591-018-0009-7), platinum-based chemotherapy drug carboplatin proved quite effective relative to the standard of care. 

DNA being a fragile molecule, it gets damaged by chemicals, random breaks, and ionizing radiation, Ashworth notes. Different kinds of DNA breaks get repaired by different machinery that work together to keep the DNA molecule healthy and free of cancer-causing mutations. 

BRCA1 and BRCA2 are critical for a subset of the double-strand break repair process, one known as homologous recombination and the other non-homologous end joining, he continues. “Homologous recombination is the preferred way to repair a double-strand break, but it... is a kind of last-ditch attempt to fix things.” 

Non-homologous end joining, where the break ends are directly ligated, can sometimes be mutagenic, Ashworth says. “Loss of BRCA1 and BRCA2 enhances use of nonhomologous end joining and this is effectively why the cancer risk is increased in individuals carrying BRCA1 and BRCA2 mutations.” 

In identifying the BRCA2 gene, researchers were faced with the intrinsic problem of how to target a deficiency, says Ashworth. BRCA1 and BRCA2 are tumor suppressor genes and affected individuals can bring one mutant copy and one wild-type copy of the genes. The wild-type copy can be inactivated to leave only the mutant copy present—a deficiency famously seen with a BRCA2 mutation common in the Ashkenazi Jewish population. 

The question for many years was what to target when the target is absent, Ashworth says. Synthetic lethality, a longstanding genetic principle, provides one avenue. 

To understand synthetic lethality, he suggests, imagine an organism where the removal of either gene A or gene B has little or no effect but removing them together is unexpectedly deadly. “You wouldn’t predict that from the phenotype of the mutations in each individual gene on its own.” 

The phenomenon is “incredibly familiar to everybody who works on model organisms—yeast and flies and worms—because it has been used for a very long time to understand genetic interactions,” says Ashworth. 

PARP Experiments 

Ashworth and his collaborators decided to try the age-old approach to target DNA repair deficiency in mutant tumors—specifically, by blocking a different DNA repair pathway. If there was a genetic interaction between those two pathways, they reasoned, perhaps it would result in synthetic lethality. 

They knew that normal tissues in a BRCA mutation area had normal homologous recombination activity whereas the tumor tissues did not, Ashworth explains. So, they blocked another excision-repair pathway in hopes of getting a “double-whammy effect” specific to the tumor cells—and it worked. 

It is but one way to think about synthetic lethality, Ashworth quickly points out. He and hi

The enabler was a potent PARP1 inhibitor of AstraZeneca, he says. “The original premise was that it would be used in combination with chemotherapy... and, while that remains an interesting idea, it has never really played out because of combinatorial toxicity.” 

Critically, years earlier Ashworth and colleagues conducted an experiment where they made isogenic cells that lacked BRCA2 completely and then compared them to wild-type BRCA2 cells. The rationale for using normal cells was to have a “clean genetic background” free of other mutations. 

Their experiment showed the BRCA cells were sensitive to “almost homeopathic concentrations” of the PARP inhibitors after being exposed to the drug for about 14 days, say Ashworth. Cancer cells nearly dissolved when treated. It was subsequently learned that inhibiting PARP leads to double-strand breaks, or replication fork collapse, which are much harder to deal with when cells are BRCA-deficient, he adds. 

Enzyme Trapping 

The biochemical effect of inhibiting PARP is through a “trapping” mechanism, preventing the enzyme from coming off DNA and blocking its replication, continues Ashworth. In fact, the selectivity of different PARP inhibitors has been tried clinically for BRCA-mutant cells based on their trapping potency. 

Talazoparib (Pfizer) is the “most trapping” of the PARP inhibitors, and therefore the most potent on BRCA-mutant cells, while Veliparib (AbbVie) is much less selective, he says. “The truth is we don’t really fully understand what the properties are that cause the selectivity, but we think the original premise is [substantially correct].” 

Clinical efficacy testing proceeded without delay because PARP agents were already on track for the clinic, says Ashworth. “We were able to divert them away from being used in all comers to being used in a phase 1b setting as an expansion of a phase 1 early optimal dose trial in BRCA mutation carriers.” 

In this setting, more than half of patients with the BRCA mutations responded quickly to the PARP inhibitors, he reports. And this was despite their resistance to multiple lines of established therapeutics. 

However, over the course of eight or nine years several pharmaceutical companies lost confidence in the PARP inhibitors after some concerning data came out about an agent that turned out to not even be a PARP inhibitor, Ashworth says. Even AstraZeneca dropped its development program for a couple of years. 

But in the end, work was done to establish that AstraZeneca’s PARP inhibitor, olaparib, was effective in the BRCA mutant setting, he adds. The original indication was approved at the end of 2014 for the treatment of advanced ovarian cancer. Since then, three other PARP inhibitors have been approved— rucaparib (Clovis Oncology), talazoparib, and niraparib (GlaxoSmithKline)—for various cancers and “there has been some to-ing and fro-ing about whether that goes beyond the BRCA group.” 

For the three being used for heavily pretreated ovarian cancer patients (olaparib, rucaparib, and niraparib), the FDA recently withdrew its approval for this indication based on survival data. But overall, the evidence supports the use of PARP inhibitors in patients with the BRCA mutations and homologous recombination-deficient tumors, says Ashworth. 

Clinical Data 

Some of the most exciting clinical data on olaparib emerged from the Solo-1 Trial of PARP inhibitor maintenance therapy in BRCA ovarian cancer (New England Journal of Medicine, DOI:  10.1056/NEJMoa1810858). Compared with placebo, opalarib given after platinum-based chemotherapy provided substantial benefit in terms of progression-free survival among women with newly diagnosed advanced ovarian cancer and a BRCA1/2 mutation.  

“The really quite striking observation is that in the placebo-controlled arm, 50% of patients relapsed in about a year,” says Ashworth. Among the olaparib-treated group, half of patients weren’t relapsing until year four.  

More recently, at a meeting of the American Society of Clinical Oncology, researchers shared results of the OlympiA trial of adjuvant olaparib in patients with HER2-negative, high-risk ­early-stage breast cancer and BRCA1 and BRCA2 mutations. The heroic trial, enrolling 2,000 patients at 200 centers, demonstrated a significant overall survival benefit; specifically, the drug reduced the risk of death over placebo by 32% and yielded an absolute improvement of 3.8% at three years. This established the evidence allowing the FDA to approve olaparib as a treatment for adjuvant breast cancer—as Ashworth points out, 27 years after the discovery of the BRCA genes.   

Ashworth believes that PARP inhibitors target the underlying DNA repair defect, homologous recombination, rather than the BRCA mutants specifically. Back in 2004—well before the discovery of modern chemotherapeutic agents and PARP inhibitors—he helped coin the term “BRCAness” describing the phenotypic traits that some ovarian tumors share with tumors in BRCA1/2 germline mutation carriers causing a defect in how genetic information gets exchanged. 

Good evidence now exists that “at least some gene defects in homologous recombination do in fact respond to PARP inhibitors,” Ashworth says. But they likely target a network of genes rather than the individual BRCA1 and BRCA2 genetic mutations.  

Resistance Puzzle 

“This story has not been without its challenges, drug resistance being one of them,” says Ashworth. “When we started this, we always imagined there would be some mechanism of resistance, but it was kind of puzzling to imagine what it would look like [for] the synthetic lethal approach.” 

Early on, Ashworth and team developed cells from a pancreatic cell line carrying the 6174delT mutation (Capan-1) that were highly resistant to a PARP inhibitor, as with the Ashkenazi Jewish population. The mutation causes truncation in the BRCA2-associated protein since it lacks a thymine residue in the DNA, he explains, and the back end of the protein is missing. 

The resistant cell lines were all reconstituted to the “zombie” form of the protein, recreating enough activity of homologous recombination to make the cells resistant to a PARP inhibitor, says Ashworth. While concerning, this proved that the research team was targeting synthetic lethality because an enzyme was being inhibited and resistance had nothing to do with the PARP inhibitor but rather the synthetic lethal target of the BRCA2 gene. 

Many studies are being published describing various mechanisms of resistance, he continues, and their relative importance remains unknown. “I know some of them are relatively theoretical and less likely to happen in patients. We do know reversal mutations occur quite frequently, [but] how frequently is not clear and it may well differ by the position of the mutation.” 

Some of the anecdotes are quite persuasive, says Ashworth. He points to a study, published in Cancer Discovery (DOI: 10.1158/2159-8290.CD-17-0146), finding BRCA2 reversion mutations in patients with DNA repair-deficient prostate cancer who developed resistance to PARP inhibitors, as well as demonstrating the use of circulating cell-free DNA to monitor the phenomenon.  

In one patient whose PSA level shot up quickly and strongly after a transitory response to olaparib, he notes, the research team found 34 reversion mutations normally detectable prior to administration of a PARP inhibitor. Importantly, these were in-frame mutations that “by chance... should be one in three, so this isn’t just some random mutation that is going on.” More likely, he adds, it is Darwinian selection of preexisting clones that have gone undetected prior to PARP inhibitor treatment. 

“The confounder is they are BRCA deficient, so they are also DNA repair deficient,” Ashworth says. “It may well be that the pool of mutants is higher in that group than it is when you don’t have a BRCA mutation.” 

As reported in Cancer Discovery (DOI: 10.1158/2159-8290.CD-19-1485), more than 300 reversion mutations in BRCA1 and BRCA2 have been catalogued and analyzed. It was the first such attempt to get at the absolute prevalence and context of these mutations, says Ashworth, which remains unknown. 

But the analysis suggests that the position of BRCA2 mutations affects the risk of reversion. Many reversions were also predicted to encode tumor neoantigens, which might provide a route for managing resistance. 

PEGylated Agent 

Several groups are now working on addressing difficulties in combining PARP inhibitors and chemotherapeutic agents to overcome resistance and enhance the efficacy of combination therapies, reports Ashworth. 

AstraZeneca is developing a next-generation PARP1 selective inhibitor that may have properties that will allow it to be combined with chemotherapy, he says. “I’m hopeful, but we’ve been thinking about other potential ways of doing this.” 

UCSF scientists have teamed up with a company called ProLynx, which has a technology that involves using polyethylene glycol (PEG) as the drug carrier, Ashworth says. In one experiment in pediatric solid tumor models, where 15-nanometer particles of PEGylated talazoparib were coupled with an oral chemotherapy drug, the combination increased uptake into tumor tissue while reducing permeability to normal vasculature and thus toxic side effects. Protracted drug exposure was accomplished by a releasable “linker” developed by ProLynx that degrades at a predictable rate. 

The UCSF team has also experimented with the agent in xenograft mouse models using colorectal adenocarcinoma cells engineered to lack BRCA2, he reports. Both the mice treated with talazaparib over 21 days and those given a single dose of the PEGylated version responded minimally, but similarly, to treatment. This suggests the new agent could be used in combination with chemotherapy. 

Viral Connection 

In early 2020, researchers at UCSF brought their individual expertise to bear on the question of how to target the new coronavirus, says Ashworth. “We had been aware for some time that ADP-ribosylation was really quite important in viral infection and, weirdly, three months before the pandemic we actually started to work on... ADP-ribosylation targeting SARS-1.”  

The rationale for the approach is that certain viral macrodomain mutations as well as mono-ADP-ribosylation proteins (triggered by interferon) are known to protect cells against viruses, setting up a “kind of arms race” between the two, he explains. With coronaviruses, the macrodomain serves to remove covalently attached ADP-ribosylation proteins, but with certain mutations the virus can cause weight loss in mice, but they otherwise rapidly recover. 

Researchers produced an in vitro assay for the SARS-Cov-2 macrodomain demonstrating the ability of the SARS-CoV-2 macrodomain to remove mono-ADP-ribosylation proteins. They also worked together to develop inhibitors of the macrodomain. 

One approach, Ashworth says, was to use high-throughput crystallography to expose proteins in 96-well plates to fragments of chemicals to see if any would bond to the three-dimensional structure of the macrodomain. On the first go, the team had 110 hits from 641 fragments. The team subsequently used in vitro docking of 20 million molecules to identify another 60 candidates—20 of which co-crystallized with the macrodomain, as reported in Science Advances (DOI: 10.1126/sciadv.abf8711).   

Currently, UCSF researchers have 400 three-dimensional structures that will bond to compounds of the macrodomain, he continues. Preliminary evidence suggests the compounds could be used as antivirals for SARS-CoV-2 and, potentially, for other pathogens and endemic viruses such as chikungunya and hepatitis E. “Human macrodomains we believe are druggable as well,” says Ashworth, and this potentially extends to cancer. 

Load more comments
comment-avatar