March 16, 2010 | As Emanuel Petricoin III tells it, “the planets aligned”: a long-standing partnership yields access to a unique sample set, an in-house technology goes from bench to bedside, and suddenly one of the best-selling leukemia drugs looks like a good fit for metastatic colorectal cancer. Petricoin and his colleagues have a model of what personalized medicine could look like.
The planetary alignment started with a close partnership between Istituto Superiore di Sanita (ISS) in Rome and the Center for Applied Proteomics and Molecular Medicine (CAPMM) at George Mason University, of which Petricoin is co-director. The two groups partnered on a study of colorectal cancer in which ISS sent CAPMM a sample set from 35 patients: tumor samples from both the primary colorectal lesion and the metastatic lesion (all from the liver) for each patient, giving the GMU team an opportunity for a side-by-side comparison.
“Oftentimes, therapy is selected by looking at… the primary tumor… but the primary tumor isn’t what kills the patient, it’s the metastatic tumor. And if the metastatic tumor looks different molecularly than the primary, that would be pretty important to know,” explains Petricoin. “So we asked a very simple question of the study set: Is the metastasis the same or different than the primary?”
To answer the question, Petricoin is adamant about a proteomics approach. “We’re using a novel protein microarray technology. There’s no genomics being done in this study,” he says. “The genes aren’t the drug targets. It’s a couple of degrees of separation removed from the real thing,” he says. “Forget the genomic analysis. Look directly at the business. Is the drug target activated?”
The team used reverse phase protein microarray (RPMA), a technology Petricoin and CAPMM’s other co-director, Lance Liotta, invented in 2000 (see, “Portraits in Proteomics,” Bio•IT World, April 2004). Until recently, the technology has been “research only” Petricoin says. But in the past three years, the team has made technical improvements to increase sensitivity and decrease variability and transition the technology to a robust calibrated assay format. “We also developed and verified reliability of the RPMA under CLIA guidelines, developed controls and population data. Basically we improved the performance characteristics and analytical methods to transition the RPMA from research grade to clinical grade,” he explains. RPMA reveals target activation. The updated technology has been vetted on different types of tumors, and other human and animal tissue taken from obesity and cardiovascular disease studies, but this was “the first study that it’s percolated all the way to bedside.”
The GMU team uses laser capture micro-dissection to pluck the tumor cells out of the biopsy because their past work showed the necessity of the procedure in obtaining accurate molecular signaling portraits of the disease cells in question. It then uses the RPMA technology to map the circuitry of the tumor cell. The RPMA allows researchers “to quantitatively measure the circuitry, the protein activation network of hundreds of drug targets from a tiny biopsy specimen. Because that’s how personalized medicine’s going to be done,” says Petricoin. “You’re going to have these tiny biopsy specimens and you’re going to have to hopefully tell a lot of molecular information about what pathways are on from a tiny bit of material.”
The result was that the metastasis looks entirely different from the primary tumor, but what was really intriguing is that in 35 patients, the metastatic tumor all showed activation of the targets of Gleevec, the Novartis blockbuster cancer drug.
“This is a true kind of proteomics-driven molecular stratification that’s very unique,” says Petricoin. “Most people … if they’re doing molecular profiling, they’re doing gene analysis or gene mutation analysis. We’re looking directly at the drug targets, and I think that’s the value of proteomics. Our opinion is, the genes aren’t the drug targets. The proteins are the drug targets.”
Gleevec is indicated for chronic myelogenous leukemia and gastrointestinal stromal tumor, not colorectal cancer. In fact, Gleevec’s targets (the BCR-ABL oncoprotein, cellular protein c-kit, and platelet-derived growth factor receptor (PDGFR)) don’t show up in primary colorectal tumors. But things change as the cancer metastasizes to the liver, and “one of the pathways that lit up like a string of lights was the Gleevec drug target family,” says Petricoin.
Swiss Time
After Petricoin presented his results to Novartis, the Swiss pharma provided funding for a small pilot clinical trial using Gleevec for metastatic colorectal cancer patients. The trial has just begun accruing patients the first of 20 patients with end-stage colorectal cancer with liver metastases. None would have received Gleevec, because it is simply not the standard of care, Petricoin explains, and there was no molecular reason. “We identified this molecular rationale,” says Petricoin. “So in this clinical trial, patients are truly being stratified for therapy. It’s truly a personalized therapy trial. If they have the pathways activated, they get Gleevec. If they do not, they get put on standard of care. And we follow these patients to see if they respond or not.”
Petricoin is both realistic and compassionate about the patients. “The bar is fairly low here from a clinical outcome standpoint, because unfortunately nearly all of these patients will die of their disease very quickly because they are at end-stage metastatic [disease]. So any stabilization would be considered a success.” Yet he believes these are precisely the patients that can benefit the most from the trial’s potential success. “These patients are going to their doctor’s office and the doctor is basically throwing darts on what to do next because everything else has already failed them,” he says. “Now we’re saying there’s a new type of assay or platform that can maybe rationally help you select drugs you may not have even considered, based on which drug targets are activated in the metastatic tumor itself in that patient.”
“It’s probably the first personalized therapy trial for metastatic colorectal cancer ever,” says Petricoin, and it’s testing a paradigm. “The current paradigm really is that we give therapeutics… based on site of the tumor. Herceptin for breast cancer; Gleevec for leukemia. But the new paradigm we’re trying to test is that we give target inhibitors based on whether or not the tumor has a drug target activated regardless of where it rests. There’s no need to think about tumors according to where they’re localized, but more about their molecular profile.”
For now, all of the testing will be done in CAPMM. Four scientists (two of which are backups) can receive samples and return results within 72 hours. Patients will have biopsies and in three days the team will let the physician know if their patients are showing activation of the Gleevec targets.
“In some ways, the trial is testing the obvious question of whether or not turning the Gleevec drug targets off works clinically,” Petricoin says, “But we’re also testing the analytical capabilities of the technology, of the workflow to actually be done in the first place… Could we exploit this novel proteomics technology? Could we actually do this in a quick turnaround time that you’d actually need to do this in a real clinical world?”
All of this begs the question, “What if it works?”
Although Petricoin is quick to point out that lit up drug targets do not necessarily mean responsive disease, the clinical-grade RPMA technology as well as specific biomarkers for prediction and prognosis have already been licensed to Theranostics Health, a 2007 CAPMM spinoff. Petricoin is a founder along with Liotta and sits on the Scientific Advisory Board of Theranostics and will only say, vaguely, that the company is “working with pharma right now on a variety of different opportunities,” though no assays for physicians to order are being sold quite yet.
Petricoin sees the model as a glimpse into the potential of personalized medicine: companies like Theranostics with “gatekeeper technologies” can help inform physicians. Maybe the drug company sits back and enjoys the increasing revenues, or maybe pharma funds trials and works with a technology company to provide companion diagnostics.
In this case, Petricoin predicts that Novartis will be involved. “If one patient has stable disease from this trial, I think Novartis will be very interested in this. And they’ve told us that if we have some success in this trial, they’d love to see this go to a multi-institute large trial, where you really test this statistically and robustly.”
Petricoin isn’t waiting for the results of the Gleevec tests though. A new personalized therapy trial, sponsored by the Side-Out Foundation for metastatic breast cancer has just begun recruiting patients. Liver, skin, or lung metastases samples will be taken from metastatic breast cancer patients and biopsied. Following micro-dissection and RPMA, the team then sends a report that says what drug targets are activated to a medical treatment decisions committee, which will make a decision. “But rather than just Gleevec, they’ll choose from dozens of therapies the patient would never have been considered for; we can profile for the activation level of all of the FDA approved targeted therapies at once.”
Petricoin is clearly excited about the possibilities. “You can really start to see how this could take shape if things work, especially for metastatic patients that have no other hope. And you can provide the physicians with key information they’ve never had before that they really can use to decide what to do next with these patients.”
This article also appeared in the March-April 2010 issue of Bio-IT World Magazine.
Subscriptions are free for qualifying individuals. Apply today.