By Jennifer Crebs and Meredith Salisbury
Pharmacoproteomics may not be a buzzword that gets your paper published — at press time, PubMed only returned 15 citations using the term — but don’t think that this will remain the case for long. In fact, the phenomenon of pharmaceutical companies and biotechs teaming up with proteomics expertise from outside companies or academics to find better-targeted therapeutics is becoming more common than ever before.
Taking a stroll through the admittedly skimpy literature record, one can see that the term started cropping up in journals around 2002. But the application of proteomic techniques such as mass spec, SELDI, and protein microarrays to pharmacology, toxicology, and drug development was already in place by the late ’90s. That said, several sources estimate that pharma truly started to embrace the prospect of outsourcing or striking partnerships sometime around 2003. Since then, the trend has only grown. This year, in fact, the third Biologie Perspective meeting in Santorini will feature an entire session devoted to pharmacoproteomics.
Walid Qoronfleh, executive director of Michigan’s Core Technology Alliance, has seen the shift happen from the front lines. “Relative to other sectors, the pharma/biotech industry has been slow to adopt outsourcing,” he says, but he does affirm that the demand for R&D outsourcing by pharmaceutical companies is expected to grow.
Citing data from a 2005 report by the Pharmaceutical Research and Manufacturers of America, Qoronfleh points out that total research and development investments by pharma have quickly outpaced NIH spending on biomedical research. R&D doesn’t come cheap, and cutting-edge technologies such as those used in proteomics involve expensive, labor-intensive processes to speed drug discoveries to the clinic. For this reason, big pharma is turning to the benefits of outsourcing key components of its drug, biomarker, and diagnostic development operations.
Likewise, smaller pharmaceutical or biotechnology companies also have a reason to outsource. Whereas some firms may be able to foot the bill for specialized instruments or a handful of proteomic specialists, the prospect of building a complete team to collect and interpret proteomic data involves tremendous cost. “Outsourcing provides the small pharmaceutical or biotechnology company with the ability to take new drugs through clinical trials by avoiding investment in infrastructure,” Qoronfleh says.
This is just one trend that Qoronfleh has seen take place in recent years. Prior to taking on the post at the CTA, a not-for-profit organization founded by four major research institutions, he gained first-hand experience with diverse players in the pharmacoproteomic nexus. Before heading up his organization’s collaborative network of fee-for-service technology facilities, Qoronfleh worked for SmithKline Beecham, the NCI, and Perbio Sciences, among others.
In addition to the obvious financial benefits of outsourcing proteomics services, Qoronfleh sees a couple of larger issues informing the practice. For instance, he notes that the aging baby boomer generation will demand more therapeutics and provide extra incentive to invest in biomedical research. Add to this the FDA’s increasing pressure to speed the approval process, and you have the perfect conditions for a cottage industry dedicated to specialized services.
But Qoronfleh points out that these relationships won’t necessarily be easy. “This is not a risk-free business, and a pharma or biotech company will not outsource unless they are sure that they have protected their rights.” Fee-for-service providers like the CTA, with their explicit disavowal of any IP claims, may become even more popular with drug developers in the years to come. But proteomic partners will have to really go the distance to secure their positions in this field. Martin LeBlanc, COO of Caprion Pharmaceuticals, says, “The biggest issue with proteomics has been one of credibility,” and he stressed that these firms first have to prove they can deliver in order to gain the trust of pharmas and biotechs.
The profiles that follow show how a handful of companies are addressing the issue of outsourcing proteomics expertise, or even striking up these relationships. Keep reading to get a glimpse of how pharma and proteomics-based companies have arranged these relationships to speed the development of drugs, diagnostics, and biomarkers.
Protein-based diagnostics are heart of Biosite’s biz
Biosite is a diagnostics company that got its start by creating a drug screening test to meet the need of emergency rooms to rapidly triage new patients. Since then, the company has branched out into new disease areas, while remaining focused on the development of diagnostics. It’s a unique business model that Gunars Valkirs, Biosite’s senior vice president of discovery, says benefits both the company and its collaborators, who are primarily interested in validating drug targets.
The company’s initial drug screening product was followed by a cardiac test that could detect markers accepted for the diagnosis of acute myocardial infarction. The test was able to quantify three different proteins in a single device, and return results in less than 15 minutes — crucial for the effective treatment of heart failure. At the time, this was the first protein array that had ever been commercialized, says Valkirs; these days, the test can measure up to 12 markers.
Now Biosite is fixed on opportunities for which no diagnostic exists, which is both a departure from the company’s early days and a potential boon to pharma companies downstream. In the late ’90s, around the same time the heart attack test was commercialized, the company developed the capacity to design antibodies with a huge number of targets. By being able to generate reagents for immunoassays along with the ability to actually implement them on devices, Biosite was able to broaden its mandate to build diagnostics for sepsis, stroke, acute kidney injury, and cancer — “areas where we believe that single markers are not going to be adequate for the job of diagnosing complex disease,” Valkirs says.
So far, Biosite has only one publicly identified pharmacoproteomic collaboration, which is with Eli Lilly. Last March, Lilly and Biosite entered into a collaborative partnership to develop an assay that will be used to identify patients who may benefit from taking Xigris, Lilly’s severe sepsis drug. Physicians don’t know to whom Xigris should be administered, Valkirs says, so consequently it’s not as widely used as it could be. Biosite will be providing the drug company with a point-of-care assay based on a platform designed to detect levels of Protein C, a multipurpose biomarker, which may enable physicians to tailor dosage requirements during treatment. Lilly will use the Biosite test to enroll patients in an upcoming Phase IIb clinical trial, slated to take place later this year.
Biosite also has a number of collaborations between biotech and pharma companies, mostly on the antibody development side, Valkirs says. In the case of proteomic biotech companies, the diagnostic company has recently signed a deal with Oxford Genome Sciences. Biosite will use the targets that company has discovered to develop assays for use in the detection of colorectal cancer, both in asymptomatic and relapsing patients.
Regarding these collaborations, Valkirs says that the company tries to identify other players who are interested in the same disease. Because of Biosite’s ability to cost-effectively generate antibodies to a large number of targets, therapeutic-minded collaborators can validate leads, while Biosite benefits by gaining access to proprietary targets relating to disease areas of interest.
“Eventually, we’ll develop diagnostics based on those targets, and our partners will develop therapeutics.” Valkirs says. “We really have no interest in competing with our partners, and that’s a very different model from other companies who may want a slice of the same pie.”
Caprion, MDS lend proteomic expertise to Biomarker Alliance
Pharmacoproteomics is just one component of the Biomarker Alliance, a network of biomarker discovery and validation providers including MDS Pharma Services, Caprion Pharmaceuticals, Gentris, and the Massachusettes General Hospital’s radiology department.
But that’s what makes the network a success, says Martin LeBlanc, chief operating officer of Caprion: pharma clients come to the Biomarker Alliance because it’s a one-stop shop for any kind of biomarker, from pharmacogenomics to SNPs to proteomics to imaging.
Caprion lends most of the proteomics power, especially for the discovery phase, while MDS contributes to validating the biomarkers, according to LeBlanc. “Caprion has built an integrated proteomics technology to discover biomarkers,” he says. “We are able to, in a nutshell, profile thousands of proteins and their changes in expression level in plasma.” LeBlanc says Caprion’s use of sera — mostly blood or urine — is essential in getting a “more direct readout of what’s going on at the protein level” than one would get from looking at a tissue sample.
Caprion’s technique, which includes mass spectrometry and bioinformatics to analyze the proteomic data, renders a dataset documenting “all the proteins that are changing between normal and disease state” or between treated and untreated patients or various dosings of a treatment, LeBlanc says. That dataset is what gets delivered to the pharma client, serving as a list of “biomarkers indicative of disease or response.”
It’s a system that pharmas, for their part, have embraced. The Biomarker Alliance has active collaborations with Pfizer, Wyeth, AztraZeneca, Boehringer Ingelheim, and others. Historically, collaborating with competitors like this might not have been a feasible business model; however, LeBlanc says that biomarkers have proven to be noncompetitive enough to allow this to work. “The industry in many cases is looking to make those biomarkers a bit more of a publicly shared type of information,” he says. Still, he adds, “We certainly have adopted some strict confidentiality procedures.”
OGS (version 2) relies on proteomics database
Christian Rohlff, CEO of Oxford Genome Sciences, has been with his database since it began in the proteomics division of Oxford Glycosciences, where he was director of proteome research collaborations.
The first OGS was purchased by Celltech, but the proteomics database was spun out and morphed into the Oxford Genome Anatomy Project, a cornerstone for Oxford Genome Sciences.
“The database becomes absolutely key in having a plan” for which biomarkers get pushed ahead in the drug discovery and development process compared to those that get put into a holding pattern, says Rohlff. Thanks to the protein database and the proteomic expertise headquartered at his company, he says, “we can develop [a] protein biomarker with a pharmacoproteomic approach in less than three years.”
Rohlff identifies two main activities where OGS (the second OGS, that is) works with pharmas. The first, in preclinical development, is all about “early biological annotation of molecules,” he says, and profiling proteins to try to predict which will fare better in later stages of the process. The main consideration at this phase, he says, is cost effectiveness. In the second activity, clinical development, he says, “there are more stringent criteria” than simply cost. For this work, OGS works with pharmas to try to understand which patients are responding to certain treatments. Rohlff says a protein blood-based assay is one example of how this could play out, showing exactly how and whether a patient is reacting to therapy. “It’s clear that because of the urgency of the clinical need but also because of the associated costs that having these kinds of tools will be very useful,” he says.
Recently, Rohlff says, pharma companies have grown more willing to team up with external companies for assistance in developing and validating biomarkers. “The most strenuous and the most costly phase is the validation of these biomarkers,” Rohlff says — which has given OGS plenty of business. Collaborators include Bayer Healthcare and Bayer Diagnostics, as well as Medarex and the University of Oxford.
Pharmas are also looking at companies like OGS for help developing companion protein-based diagnostics for therapies, Rohlff says. “The majority of drug companies don’t have their own diagnostic capability [so] they want somebody who can help them and present it on the diagnostic platform.”
The rise of PPx: reverse phase protein arrays head to the clinic
For a paper he recently published in Nature, Emanuel Petricoin spent three weeks going back and forth with journal editors about the definition of the term “pharmacoproteomics.” Indeed, even though the field is gaining ground, there are only about a dozen references that come up when you use it in a keyword search on PubMed.
“It’s a nascent field,” says Petricoin, co-director of the Center for Applied Proteomics and Molecular Medicine at George Mason University, speaking of this concept of using proteomic technologies and protein endpoints that enable “the physician or the pharmaceutical company to tailor medicine to the right patient populations.” But if his predictions prove correct, pharmacoproteomics will continue to become more common — eventually, he says, even eclipsing its sister field of pharmacogenomics.
Part of the philosophy behind that prediction stems from the fact that most targeted therapies act on a protein rather than a gene, Petricoin says. “We have from day one felt that the ability to truly stratify and target medicine — especially the newer classes of molecular targeted therapeutics — will require one to actually look at the cellular circuitry, the protein pathways in patient material.” To that end, he and longtime collaborator Lance Liotta have developed the reverse phase protein microarray, which allows scientists to study hundreds of phosphorylation events in parallel using just one antibody per analyte with minimal sample from a patient. That’s a significant improvement over the traditional sandwich assay requiring two antibodies, Petricoin says. “People don’t realize what a huge roadblock that is [to find two antibodies]. There’s a huge attrition rate to get to two that don’t interfere with each other.”
The arrays currently look at 300 phosphorylation endpoints. Functionally speaking, Petricoin says, everything that happens winds up going through the human kinome — which has much more tractable numbers than the genome. “The beauty is we don’t have to measure 30,000 genes,” he says. “We may only have to measure 300 phosphorylation events.”
The reverse phase protein arrays have been under development for the past six years, and now they’ve reached the point in terms of reproducibility, precision, and accuracy that Petricoin and his team are beginning to use them in more clinical settings. For the time being, he says, their use in clinical trials will serve to generate data showing whether the arrays are actually measuring the correct endpoints. Once that is proven, he says, they will be able to introduce the tool into clinical trials with the purpose of stratifying patients into populations to recommend treatments or dosages.
That’s where pharmacoproteomics will show its advantage over pharmacogenomics, Petricoin contends. “You’re going to have the ability to stratify responders and nonresponders … using a pharmacogenomics approach. That’s great if you’re in the responder category,” he says. “But [for] the nonresponder category, genomics does not tell you what to do with that population. Proteomics, because they’re measuring the drug targets, will give you the sweet spot of theranostics.” Pharmacoproteomics will not only be able to distinguish responders from nonresponders, Petricoin says, but will also be able to indicate, for example, a particular signaling pathway in the nonresponder group that can be targeted to convert the nonresponders into responders.
Petricoin, who is talking to a number of pharmaceutical companies about the reverse phase array technology, says pharmacoproteomics has an advantage even at the most practical level: cost. Because they’re so much less finicky, protein tests are significantly cheaper than their gene-based counterparts, which can require cleanrooms and highly trained technicians. “Protein tests … will underprice the gene-based test out of existence,” Petricoin says.