Like many biotech entrepreneurs in the late 1990s, Randy Scott, then at Incyte Genomics, and George Poste of SmithKline Beecham believed that sequencing the human genome held great promise for developing new medicines. But a faster route to translating genomic data into profits, they thought, would be to focus initially on finding biomarkers predictive of disease. So they founded DiaDexus in September 1997, and with $25 million in start-up funds from Incyte and SB, Scott and Poste set out to prove it.
Six years after coming into existence, DiaDexus put its first FDA-approved, genomics-derived diagnostics test on the market — showing there is truth to the maxim that diagnostics are a more manageable mechanism for converting discoveries from genomics into products. Furthermore, DiaDexus says its approach has proved significantly faster than traditional diagnostics discovery.
“Given the fact that everyone working in this arena is pioneering, including the FDA, I think six years compared to the evolution of many diagnostic markers has been remarkably quick,” says Poste, who serves on the company’s board of directors and as CEO of Health Technology Networks, a consultancy in Scottsdale, Ariz.
But the path to genomics-based diagnostics is not without serious obstacles. Observers note that DiaDexus’ biomarker pipeline is a work in progress and diagnostics development presents unique challenges. Aside from validating the clinical utility of a marker — a monumental task in itself — companies hoping to turn genomics knowledge into diagnostics products must satisfy the technical demands of reference laboratories, as well as navigate the complexities of healthcare reimbursement. They could learn from DiaDexus’ experience.
Companies as varied as industry powerhouse Roche Diagnostics to upstarts such as Celera Diagnostics are looking to feed tests into the reference labs that physicians rely on to learn about their patients’ problems. DiaDexus’ attempt to apply genomics to diagnostics has yet to pay off in profits — sales of its first test are still at a relative trickle — but its path is illustrative of the challenges facing the industry as a whole.
It’s not hard to see why pursuing the diagnostics market would be a good idea. Industry-wide revenue from molecular diagnostics is today valued around $1.2 billion, and growing at rates between 15 and 25 percent per year, according to figures attributed to Kathy Ordoñez, president of Celera Diagnostics. And according to Heino von Prondzynski, head of Roche Diagnostics, the market for genomics-derived diagnostics will reach $5 billion by 2013.
The money being thrown at developing new diagnostics is also impressive. Although DiaDexus CEO Patrick Plewman declined to disclose the size of the company’s annual R&D budget, DiaDexus raised nearly $103 million in a private equity offering in 2000, and has two years’ worth of cash left, he says. Von Prondzynski at Roche Diagnostics says his company is spending between 25 and 30 percent of its $550 million annual R&D budget on discovering new biomarkers via genomics and proteomics. Celera, in conjunction with sister company Applied Biosystems, has spent $100 million just sequencing 39 human and chimp genomes, in part to identify new markers for disease.
Step 1: Construct a Pipeline
How is DiaDexus trying to turn genomics and proteomics research into new candidate biomarkers? It should suffice to say the company isn’t relying on just one approach. As the cutting edge of genomics technology has evolved, so has the nature of DiaDexus’ collaborations.
At its founding, DiaDexus became the sixth licensee of Incyte’s Lifeseq databases, and since then has employed a team of bioinformaticians to pore through Lifeseq and other databases to identify genes potentially useful as disease markers. In 2002 DiaDexus began working with Compugen of Tel Aviv to predict patterns of alternative splicing in genes that could be used as diagnostic markers, and it is employing DNA microarrays from Agilent Technologies, says CEO Patrick Plewman. Agilent prints the custom arrays in five designs representing 25,000 cancer-associated ESTs, and DiaDexus researchers use the chips to detect sequences over-expressed in tumor samples, iteratively whittling down the number of viable markers. “It has been a productive collaboration for us,” Plewman says.
More recently, DiaDexus and its 70-odd employees have joined forces with Sam Hanash, a proteomics expert and cancer researcher at the University of Michigan Medical School, and with Beyond Genomics, a Cambridge, Mass., company using mass spectrometry and other techniques to create a systems biology platform for identifying biomarkers and drug targets. Although Plewman says the proteomics partnerships are still in their early stages, “the appeal is significant, because you can shortcut a lot of painful science” by examining the proteins directly.
Not to say that DiaDexus is the only company using genomics and proteomics to discover new biomarkers. Celera is in part relying on sequencing additional human and chimp genomes to generate new SNPs with potential as biomarkers, and Roche Diagnostics even has a protein-chip platform designed for use in clinical reference labs in the works, although it’ll take several years before the system is deployed, says Roche’s von Prondzynski.
But DiaDexus co-founder Poste says the company may have an edge because of its relationship with Compugen. “To the best of my knowledge we probably have one of the most sophisticated insights now into alternative splicing patterns in major malignancies, and that is a very important element of new markers that we wouldn’t have found if we had not undertaken that [collaboration with Compugen],” he says.
Step 2: Get Your Markers into Large Patient Cohort Studies
DiaDexus’ main efforts go toward validating the clinical utility of its most promising candidate markers, which initially requires proof-of-concept studies involving quantitive PCR experiments with tissue samples, and developing immunoassays to determine if the levels of the expressed protein in serum correlate with disease. “You want your work validated by opinion leaders,” says Plewman, “and you need to persuade them to open their serum banks.”
DiaDexus followed this route successfully to validate a cardiac disease marker that the FDA approved for marketing last July. The marker, lipoprotein-associated phospholipase A2 (Lp-PLA2), was first identified at SmithKline Beecham through its collaboration with Human Genome Sciences and later passed on to DiaDexus as part of a four-year agreement designed to give DiaDexus the right to license drug targets from SB (now part of GlaxoSmithKline) for diagnostic purposes. That agreement ended in 2001.
When DiaDexus researchers first found evidence that Lp-PLA2 could be effective as a predictor of coronary heart disease in serum, they, along with scientists at SB and the University of Glasgow, published their findings in The New England Journal of Medicine in 2000. This work paved the way for access to a broader NHLBI-sponsored study of 13,000 patients, the Artherosclerosis Risk in Communities Study, from which DiaDexus presented positive results at an American College of Cardiology conference last April.
By July 2003, DiaDexus had received FDA marketing clearance for the test — “a surprisingly short amount of time,” according to Dawn Caruso, a consultant to the diagnostics industry who worked with DiaDexus in a previous position at Quest Diagnostics. The speedy approval could have been due to a number of factors, including DiaDexus’ close relationship with FDA during the approval process. But Caruso praises the company’s diligence in validating the marker’s utility in three well-respected clinical studies, efforts that the FDA undoubtedly looked upon with favor. And although FDA approval is not a prerequisite for a diagnostic test to enter the market — many bypass the agency altogether — gaining marketing approval immensely improves DiaDexus’ ability to advertise the effectiveness of its test to physicians and medical societies.
DiaDexus is taking a similar approach to working up its candidate markers for cancer. Of the 25,000 ESTs the company initially identified, it has filed patents on 6,000 and taken 1,000 marker sequences through quantitative PCR experiments in tissue samples. At the moment, says Plewman, company researchers are expressing proteins corresponding to 100 of the targets, and producing antibodies for use in immunoassays as part of proof-of-concept studies.
Step 3: Convince Doctors and Patients to Order Your Test
Publishing favorable results in respected journals and presenting data at medical conferences goes a long way toward informing physicians and medical societies of the utility of a particular test. Gaining marketing approval from the FDA allows a compay to appeal directly to physicians. There are, however, additional steps a company can take to smooth a new diagnostic’s entry into the marketplace.
The challenge with diagnostics, says David Anderson, a consultant with Boston Healthcare Associates, is to make a test that provides physicians with additional options for treatment. It’s debatable whether an early warning test truly helps a patient if a doctor has no way of mitigating that threat. Conversely, “What benefit to the physician does a test for cardiovascular disease provide if the patient already weighs 400 pounds?” asks Anderson.
DiaDexus’ close relationship with GlaxoSmithKline may help the company overcome these obstacles. The company’s Lp-PLA2 test, marketed as the PLAC test, owes a great deal of its promise to GSK’s simultaneous program to develop a therapeutic designed to inhibit Lp-PLA2, an enzyme involved in atherosclerosis. If and when FDA approves GSK’s therapeutic, currently in phase II clinical trials, physicians would have another option for reducing patients’ susceptibility to heart disease. “If the drug is accepted, that would work well for [DiaDexus],” says Jorge Leon, a former Quest Diagnostics executive now consulting independently through his company Leomics.
DiaDexus has also tailored its product development strategy to the needs of the large clinical reference laboratories such as LabCorp and Quest Diagnostics. Instead of designing its PLAC test around new instrumentation that reference labs would have to acquire and DiaDexus would have to build, the company has made the test compatible with the ultra-high-throughput automated immunoassay systems already in place, manufactured by the likes of Roche Diagnostics and Abbott Laboratories. “Developing a platform requires an engineering feat, as well as the ability to get it distributed,” says Sharon Tetlow, DiaDexus’ vice president for finance and CFO. “We looked at that early on, and found it would distract us from making markers.”
Launching a diagnostic test also requires negotiating with the American Medical Association, the College of American Pathologists, and the Centers for Medicare and Medicaid Services to set the rate at which insurance companies will reimburse a physician for the cost of ordering a diagnostic test, Plewman says. For a generic enzyme test, the standard reimbursement rate is set at $17 a test, but often companies try to show that their proprietary markers are worth more.
In the case of PLAC, DiaDexus has plans to apply for a reimbursement rate early next year, Plewman says, and will attempt to prove via a cost-benefit analysis that the PLAC test is worth reimbursing at a higher rate than $17. “The toughest thing in any diagnostic company has got to be reimbursement,” says Anderson at Boston Healthcare Associates. “It’s very different from the pharmaceutical world, where everyone is focused on the FDA. In the diagnostics world it’s the opposite. You don’t even need FDA approval to launch a test, but reimbursement is a major factor.”
Thus DiaDexus has its hopes pinned on physicians ordering more of its PLAC tests, while shuttling its next biomarker candidates into clinical studies. The company currently offers its PLAC test in a semi-automated, 96-well plate format through Mayo Medical Laboratories, but physicians have yet to order the test often enough to convince the manufacturers of the automated immunoassay systems to create their own versions of the test. “There’s no indication that this is going to be a large product yet,” says consultant Leon. “We have to see how much success it will have.”
Poste and Plewman have confidence in their pipeline, particularly in the markers they hope to validate for early prediction of certain types of cancers. Poste says the company’s head start and relationship with GSK give it a lead its competitors will find hard to overcome. “It’s pleasing to see that what we said in terms of the impact that diagnostics would have as a new element of the value chain in therapeutics is now beginning to gain in momentum,” Poste says. “We were ahead of the curve on that.”