By Kevin Davies
PharmaWeek |Dec 3| Pediatric oncologist Alan d’Andrea has enjoyed a distinguished academic career over the past two decades. As a postdoc at the Whitehead Institute in the late 80s, he almost single-handedly cloned the erythropoietin receptor. At the Dana Farber Cancer Institute, his interests pivoted toward Fanconi anemia (FA), a rare genetic cancer disorder where patients are prone to solid tumors, leukemias, and bone marrow failure. d’Andrea’s group has helped define some of the disease’s 13 known mutated genes and the underlying biochemical pathways. Now, as co-founder of The DNA Repair Company (DNAR), d’Andrea is hoping that his insights into the mechanisms of DNA repair’s six major pathways will help pharma identify patient responders and improve drug safety.
PharmaWeek recently caught up with d’Andrea in his Dana-Farber digs to learn more about the commercial prospects of analyzing DNA repair.
Alan, what was your motivation behind The DNA Repair Company (DNAR)?
I can give you a couple of different renditions. Conventional chemotherapy and radiation kills cancer by causing DNA damage – that’s how most cancer chemotherapy works. Even in the post-Gleevec, post-targeted therapy days, still, 18 of the 20 major anti-cancer drugs cause DNA damage. Cancer cells become resistant to chemotherapy and radiation by amplifying their DNA repair pathways. So the idea here [of DNAR] is, if we could get a fresh tumor sample and profile its DNA repair – which pathways are on or off – we might be able to predict the best drug(s) or radiation treatment for that patient. It’s another example of personalized medicine.
At the same time, other starting companies were saying, ‘We don’t know what’s important, so let’s just transcriptionally profile everything.’ But we think DNA repair is important. So that was the impetus behind the company, to pull together people I viewed as the five or six major leaders in DNA repair and ask them: Do you have biomarkers? Do you have antibodies that would be a sign as to whether your pathway is working? I co-founded DNAR with Michael Yaffe at MIT, who has intellectual property (IP) in DNA damage response pathways, while most of my work has been on FA and homologous recombination. So we said, let’s put all this together.
Our long-term goal is to screen for small-molecule inhibitors of these pathways. We’re interested in Chk1, PARP inhibitors, etc. The original intent was to develop a drug screening company, and eventually, things may move in that direction. But to get started, it made more sense to say, there’re a lot of conventional drugs out there already. Can we better define which patients will respond?
The personal story is a fun one – for the last 15 to 20 years, I’ve been working on this rare genetic disease, FA. Children with FA are born with a DNA repair defect. I wrote a tiny review for Nature Genetics in the early days on forging a novel pathway when the second FA gene was cloned. Well, there are now 13 FA genes! It’s astounding, and it’s been a fun ride for us. Studying this disease entailed working with pediatricians, getting samples – we have a clinical lab here running the diagnostic test for FA, and we’ve been a repository for collecting a lot of FA cell lines. So I was able to establish IP in the context of the FA pathway, which I tried to license to Big Pharma, but nobody was all that interested because it’s such a rare disease. That’s when the idea of starting our own company came about, after we reached a critical mass of IP here at Dana Farber.
How did former Human Genome Sciences CEO Bill Haseltine [Chairman of the Board] get involved?
Ah, that’s an interesting story. Thirty years ago, as an undergraduate at Harvard, I wrote my thesis while working at the [then] Sydney Farber Cancer Institute with Bill, who was a new assistant professor. He’d been a postdoc with [Nobel laureate] David Baltimore, working on retroviruses. Bill had two or three people in his lab; he was just getting started. He got very interested in DNA damage and repair – the Maxam-Gilbert protocol was, you take a piece of DNA and you damage it with alkylating agents. Bill always had a passion for DNA sequencing, and we got the protocol from Allan Maxam before it was published – I was probably one of the first people in the world sequencing DNA with this method. I kept up with Bill as a friend over the years; he worked on HIV and ultimately started Human Genome Sciences. Three or four years ago, Mike [Yaffe] and I visited him – he was very interested in the idea [of DNAR] and agreed to help.
What’s the status of the company today?
We have venture backing from Mohr-Davidow Ventures; they have particular expertise in diagnostics and personalized medicine. We have about 12 employees and research space in the Photonics Center at Boston University, but we’re moving to a bigger space in Cambridge. Also, we’re currently raising a B round of financing.
At DNAR, we’re generating small panels of antibodies against different DNA repair proteins and protein modifications. These are useful in screening particular tumor types for DNA repair activity, and thereby predicting drug and radiation sensitivity. So the clinical validation here requires the use of these antibody sets in analyzing archival tumor samples, which we obtain from collaborators around the country. We often get tumor sets from 50-100 patients with a particular tumor type. These are clinically annotated, meaning that we know how they responded to particular drugs. Next, we apply the DNAR antibodies to these samples to see if our antibody profile does correlate with the patient’s response, with the idea of coming up with predictive tests.
Is there a paradigmatic example of this approach that you can give?
There’s actually a practical example in play. Most of the validation for our company has been external. Steve Jackson [Cambridge University] started a company called Kudos, which has a real drug: a PARP inhibitor that blocks base excision repair. A certain subset of breast tumors have a deficiency in one of the six repair pathways – homologous recombination (HR) repair; they’re hyper-dependent on base excision repair instead. So predictably, a PARP inhibitor is an excellent drug here. It’s a classic example of synthetic lethality. The initial thought was that this drug could be used as a sensitizer, in combination with other chemotherapeutic agents. But astoundingly, for this subset of patients where the repair profile is known – deficient in HR, heavily dependent on base excision repair – the PARP inhibitor works as a single agent. So that has bred a fair amount of excitement for our company. Maybe our panels can actually predict small groups of patients with very distinct DNA repair profiles – they might respond to drugs in trials as single agents, allowing us to stratify subsets of patients for new drugs coming along. All drug companies, when they start a drug in the clinic, want to use it as a monotherapy rather than part of a combination, as this might increase the chances of getting positive results. So we’re hopeful that our panels could prove useful to big pharma companies in launching new drug trials.
Could you use your panels as a service like Genomic Health’s Oncotype Dx, where physicians would send clinical samples to you?
We looked at the Genomic Health model very closely. They did a couple of things very well – they set up a centralized lab, rather than send a kit out to pathology labs around the country. They also asked a very discrete clinical question: What’s one of the most vexing problems in breast cancer treatment? In other words, tamoxifen versus no tamoxifen? I think Genomic Health was wise, not necessarily going after the biggest market, but instead going after a problem that was particularly troubling to clinicians managing breast cancer.
Right now at DNAR, we’re going after tumor types – breast, ovarian, colon – and subsets of patients who will respond to conventional therapies. We’re looking for panels of antibodies that aren’t just prognostic, but predictive of drug response. So it’s less of a crapshoot using this drug set. It also says, here’s a subset of patients who will not respond – because it’s as important to disqualify patients from the drug as to qualify them.
DNAR’s other mission is partnering with bigger companies who have PARP inhibitors or DNA repair inhibitors, with the idea that we can find subsets of patients who are most likely to respond to these drugs, and stratify them for clinical trials. Let’s get them into Phase I trials, so the trial isn’t simply a toxicity test; you’re trying to enroll patients who might actually respond. The whole success of Velcade at Millennium was having a few multiple myeloma patients in their Phase I trials. You want to stack the deck.
Do you have any ongoing Big Pharma collaborations?
We do, but I’m not sure I’m allowed to reveal them yet. We’re hopeful that over time, these same pathways and biomarkers will be useful in screening drugs. That’s an important component of this: Why have a company doing immunohistochemistry when there are 5-10 times as many companies doing transcriptional profiling, looking at nucleic acids? I’d argue that the logic of having a pathology-based company is, you’re actually looking at a pathway and if you find a marker, it can also be pharmacodynamically useful for developing drugs around that pathway.
Are you looking for additional partners?
When I go to ASCO or clinical oncology meetings, I talk to small companies that have drugs – Chk1 inhibitors, PARP inhibitors – and who may be modulating DNA repair pathways. I then try to set up relationships with DNAR. In terms of the competitive landscape, there are several other companies coming up in the DNA repair area, but I haven’t seen any as focused as we are.
Who are some of the other key management team players?
Dan Paterson, our president, used to work at Dana Farber; he has a lot of knowledge of clinical trial networks, and he’s a business guy. Our chief scientific officer, David Weaver, is himself a highly regarded expert in DNA repair, while Brian Ward, our executive vice president, previously worked at Myriad and Genomic Health – he’s the development guy, trying to get these ideas into a nice little packaged kit. That’s the race with all this! Some postdocs from my own lab are now working at DNAR. Each scientist has a different tumor assignment; they collaborate with experts in academic medical centers for samples and clinical information. I think it’s working well.