Like the army of White Walkers who march haltingly across the tundra in the HBO series “Game of Thrones,” cancer cells trudge along in a menacing, but hobbled, state. The same genetic defects that help them proliferate also render them vulnerable to attack. Find the right molecule for the right weak point, and it’s like charging the White Walkers with knives made of dragonglass: They’re done for.

For several decades, small-molecule cancer drug researchers have dug for dragonglass among the kinases. By designing small molecules that can “turn off” kinase enzymes stuck in the “on” position, they have managed to ward off cancer’s attack. More recently, scientists have focused on ways to help the body’s own immune cells seek and destroy cancer cells.

But those approaches haven’t cured all cancers. Immunotherapy, while fantastically effective for some people, doesn’t work for everyone. And many of the best-known causes of cancer remain “undruggable”; chemists can’t come up with molecules that safely disrupt their activity.

Take gastric cancer, for example. “Immunotherapy doesn’t work for the majority of patients. There are no active targeted therapies, and chemotherapy is only modestly effective,” points out Barbara Weber, chief executive officer of the oncology-focused start-up Tango Therapeutics. Worse, most cancers feature at least one important gene that has lost its function. “The problem is that when gene function is lost or the gene is gone entirely, you can’t target them,” Weber adds.

Enter the concept known as synthetic lethality. Researchers have long known that a cancer cell hobbling along with one broken gene is vulnerable. Knock out another key gene, they have discovered, and the cell will topple.

The notion of killing cancer cells by damaging their already compromised DNA should sound familiar. Chemotherapy agents do just that, but by using a sledgehammer to smash away at DNA. Because drugs that exploit synthetic lethality act with precision, they promise to be able to discriminate between healthy and diseased cells.

Companies have already developed a class of drugs called PARP inhibitors that take advantage of a synthetic lethal interaction to treat certain types of ovarian cancer.

Now, motivated in part by the real-world proof that the concept can work, a new wave of biotech firms is digging for dragonglass here. They are counting on the potential for the powerful gene-editing tool CRISPR to crack open the next generation of oncology drug targets.

Sidetracked

Geneticist Calvin Bridges first wrote about synthetic lethality nearly a century ago, after noticing that fruit flies with certain combinations of genetic mutations died, while their parents, which harbored just one of the defects, survived. A colleague coined the term, using the original Greek meaning of the word “synthetic”: combining two things to form something new.

In the intervening years, other researchers explored the concept and tried to exploit it in drug discovery. But many decades passed before cancer drug developers had a good test case.

Along came PARP, or poly(ADP-ribose) polymerase, a family of enzymes that repair single-strand breaks in DNA. Drug developers first tried combining PARP inhibitors, which precisely inhibit DNA repair, with chemotherapy, the DNA sledgehammer. The combination proved too toxic.

But the drug class found new life in 2005, when researchers showed that mutations in BRCA genes decreased cells’ ability to undergo homologous repair, a way of patching up double-strand DNA breaks. The defect caused cells to rely on PARP to fix damaged DNA. The scientists reasoned that putting those already stressed cells under more pressure by preventing them from repairing single-strand DNA breaks would be enough to kill them.

That led oncologists to think that PARP could work as a synthetic lethal drug target and that PARP inhibitors might be effective on their own. Indeed, early studies showed promise in using PARP inhibitors to treat ovarian and breast cancers marked by BRCA mutations.

Enthusiasm evaporated in 2011 when Sanofi’s PARP inhibitor iniparib failed to show benefit in a Phase III study of women with triple-negative breast cancer. The result didn’t drag down just the compound; the synthetic lethality concept as a whole was tarnished.

There was a problem, though: Iniparib turned out to not be a true PARP inhibitor. A few companies stuck with the drug class, culminating in three PARP inhibitors winning approval to treat ovarian cancer: AstraZeneca’s Lynparza (olaparib), Clovis Oncology’s Rubraca (rucaparib), and Tesaro’s Zejula (niraparib).

But iniparib’s failure continued to weigh on the field of synthetic lethality. “It’s really very easy for a whole field to almost get stalled because of what is effectively an erroneous result,” says Christopher Lord, leader of the gene function team at the Institute of Cancer Research in London.

Despite the approval of that first PARP inhibitor, “the pharmaceutical community has just sort of held its breath a bit and not really pursued DNA repair,” says Niall Martin, CEO of Artios Pharma, which is pursuing the concept of DNA damage response. Martin led the team that developed Lynparza at KuDOS Pharmaceuticals, acquired in 2005 by Astra­Zeneca.

Even without the iniparib red herring, the field had other problems. Notably, the process of finding synthetic lethal gene pairs that could kill cancer cells was arduous and riddled with technical challenges.

Until recently, big drug firms and academic researchers spent much time and energy screening for novel cancer targets using RNA interference (RNAi) technology. In order to generate clues about synthetic lethal drug targets, they filed through giant libraries containing either small interfering RNA (siRNA), which are short oligonucleotides that can directly silence genes, or short hairpin RNA, which use a viral vector to dampen gene expression.

The RNAi technology was alluring but unreliable. Off-target effects were rampant, and validating targets was daunting. Big pharma researchers who lived through that period—many of whom now populate the biotech firms working on synthetic lethality—look back in frustration.

“When you ran a screen, you came out with a massive number of false positives,” recalls Michael Zinda, who previously led oncology bioscience at AstraZeneca’s Boston research site and now heads R&D at synthetic lethality-focused Repare Therapeutics. “It was inefficient to dig through the gene set to find one or two real hits.”

Weber, who before taking the helm at Tango was head of oncology translational medicine at Novartis, agrees. Both the publicly available academic screens and the ones going on inside companies yielded long lists of potential synthetic lethal gene pairs that ran the gamut from strong possibilities to ones with limited effect, she says.

Even the seemingly strong hits were problematic. “You’d burn through postdocs very rapidly because the targets would not validate,” Weber adds. “There was no real threshold to separate the true positives from the noise.”

New generation

The new biotech companies have been motivated largely by a possible solution to the signal-versus-noise problem: CRISPR, the powerful gene-editing technology that in just a few years has become a workhorse of industrial and academic labs.

Named for the “clustered regularly interspaced short palindromic repeats” seen in the bacterial immune system from which the technology was discovered, CRISPR allows genes to be turned on and off with exquisite precision.

Meanwhile, new information about cancer genetics has proliferated, as have cancer cell lines for testing that information. “In the last five years, tens of thousands of genomes from cancer patients have been sequenced, so that now we have a much better understanding of what the mutational spectrum looks like,” Repare’s Zinda says.

In addition, academic researchers worked out how to use CRISPR RNA libraries to find new synthetic lethal interactions, and these discoveries have become the basis for several well-financed companies.

The $68 million in first-round funding for Repare was the biotech industry’s biggest early-stage financing in the first half of 2017. That figure was surpassed this month by the fundraising for KSQ Therapeutics, which got $76 million to build a CRISPR-based discovery platform that will be used in part to pursue synthetic lethal drug targets.

Academic and industry scientists are taking varied approaches to their screens. One method uses CRISPR guide RNAs to screen two cell lines that are identical except for the knockout of one key gene, and look for places where the cells are perturbed differently than others.

Another strategy is to screen a CRISPR library against a panel of cancer cell lines that includes cells with and without a mutation of interest—BRCA, for example. Good targets should cause a perturbation in only the mutated cell line.

Both methods produce a more refined list of potential targets than what was found with earlier technology. “With RNAi screens, you would have lists with hundreds of genes that looked like potential hits,” many of which might not be legitimate, Zinda says. A CRISPR screen yields a more focused list of genes, the vast majority of which are “real,” he adds.

“Five years ago, I was lucky to have RNAi,” says Frank Stegmeier, who previously led oncology target discovery at Novartis. Now, as chief scientific officer of KSQ, he’s amazed at how precisely he can interrogate gene functions.

Each biotech firm is trying to add its own twist to the screening process. Tango, for example, identifies a subset of people with cancer it would like to treat and then uses cell lines that match their genetic profiles. And rather than a full CRISPR library, Tango interrogates the cell lines using a tailored “druggable” CRISPR library so that it can create small molecules against the hits that come out of its screens.

KSQ, on the other hand, is going broad and deep. The biotech has already generated massive data sets by knocking out, one gene at a time, the entire genome of nearly 600 cell lines. The firm says the vast screening effort will help it understand whether targets can be safely inhibited with drugs.

Dose of reality

Although companies are enthusiastically forging ahead with target validation using CRISPR, not everyone sees the technology as a panacea. University of Sussex structural biologist and DNA damage expert Laurence Pearl likes to point out that screens are still just cells, not humans. Using siRNA to knock down expression of a protein in a cell or knocking out its gene by CRISPR/Cas9, is a whole lot different from using a small molecule to block the function of a protein in the body.

“In experiments, you have a perfectly happy cell, you remove one gene and challenge it with one stress, and it dies. You think, ‘Yeah, I’ve got a target now!’ ” says Pearl, whose lab is working on drugs against synthetic lethal targets. “But that’s not a tumor.”

Simon Powell, chair of the department of radiation oncology at Memorial Sloan Kettering Cancer Center, agrees. “Right now, the field is okay at the cellular level,” says Powell, who is on the advisory board of Artios. “But finding drugs and making sure they actually work in a real tumor is where they need to go.”

Biotech companies counter that the technology is also helping them pressure test their targets. “There’s a convergence of understanding patient tumors through sequencing, new technology like CRISPR, as well as a proliferation of novel models allowing us to ask really much more informed questions and identify patient populations who would best respond to our drugs,” Repare’s Zinda says.

The next phase for these young biotech firms will be proving that this convergence does indeed lead to cancer-killing drugs. Companies are already well into winnowing the hits emerging from CRISPR screens using factors including whether drugs already exist against those targets and how easily the hits can be broached with small molecules.

Ironically, today’s most prominent synthetic lethal drug target is one identified before the CRISPR craze. DNA polymerase θ, encoded by a gene called POLQ, is a key player in a method for repairing DNA breaks that becomes important when more prominent repair mechanisms, including the homologous recombination used by BRCA, are lost. Both Repare and Artios are pursuing Polθ inhibitors.

The bigger picture

As companies that are focused on synthetic lethality proliferate, they are also contemplating how their drug candidates will fit into the quickly evolving oncology landscape. Cancer immunotherapies have become a prominent part of the drug pipeline, and every new firm has to consider how its DNA repair inhibitors might combine with already approved immunotherapies, like checkpoint inhibitors and the new immuno-oncology drugs.

Although pairing DNA repair inhibitors with immunotherapies makes sense on paper, studying combinations in the lab is a challenge. “At the moment, we’re suffering from a dearth of data,” says the Institute of Cancer Research’s Lord. Because the immune system is difficult to model in cells and animals, “there’s a big research gap about whether DNA repair and immunotherapy work together.”

Still, one appeal of combining checkpoint and DNA repair inhibitors is the perceived safety of the approach. Unlike adding kinase inhibitors to immunotherapy, which has proved toxic in some cases, early clinical data suggest side effects aren’t exacerbated by PARP inhibitor-immunotherapy combinations.

Even as they contemplate blue-sky questions like how to meld theoretical DNA repair drugs with an evolving immuno-oncology pipeline, executives from recently established biotech companies remain focused on the major task ahead: getting cancer drugs to the people who need them. Although Tango, Repare, and KSQ have secured significant funding and quickly grown to several dozen employees, they still have a lot more work to do.

All of them have selected their initial drug targets, but none of them expects to have a compound in the clinic before 2019.

That’s not to say this new generation of biotech firms doesn’t have grand ambitions. Oncology drug developers have never been able to get at the genetic changes caused by a loss in gene function. Synthetic lethality is “the only way” to do that, Tango’s Weber says. “What I hope, and what our experiments are now showing, is that if you could get that piece, you open up a whole new target space in oncology.”