Leading genetic mutations and gene-expression profiles linked to rheumatoid arthritis, including the adverse effects of TNF-inhibitors such as the widely prescribed drug Remicade, were presented in a majority of studies reported by Millennium Pharmaceuticals, according to a company official.
Remicade, made by Johnson & Johnson unit Centocor, was recently approved by the US Food and Drug Administration as a first-line therapy for RA. The drug, approved for Crohn’s disease in 2002 and for RA in 1999, has been linked to an increased risk of lymphoma, and the FDA last week asked Centocor to include the risk in its labeling for the drug.
Presented during the 2004 American College of Rheumatology meeting in San Antonio, Tex., this week, the studies relate certain genomics and proteomics data to risk, severity, and response rates for drugs used to treat RA. Millennium analyzed “about half” of its growing RA-patient registry — a collaborative effort with Brigham and Women’s Hospital in Boston — to produce the studies, said Ronnen Roubenoff, Millennium’s senior director of molecular medicine.
The RA registry contains data on more than 885 patients, but will grow to include 1,000, Roubenoff said. It collects demographic, clinical, health outcome, pharmacogenomics, and economic data, as well as cross-sectional data on other therapies given to patients taking TNF inhibitors. Data from the registry will be made available to other companies.
Even though this may be the largest group of patients assembled for the study of RA, more and larger studies may be needed. “Some gene-expression and proteomic signatures are apparently popping up as potentially useful, but these will need to be confirmed on even larger cohorts,” wrote Peter Gregersen in an e-mail. Gregersen is director of the Robert S. Boas Center for Genomics and Human Genetics at the North Shore Long Island Jewish Research Institute and principal investigator of the North American Rheumatoid Arthritis Consortium.
Using registry data, Millennium is looking at two classes of biomarkers: predictors of disease severity and predictors of therapeutic response, said Roubenoff. “The most clinically important [of the studies] are regarding the predictors of methotrexate and anti-TNF response, and we also have predictors of toxicity,” said Roubenoff. We focused on TNF inhibitors, as a class, and methotrexate — these are really the lynchpins of treatment today.”
Methotrexate is a cancer drug used to treat autoimmune diseases such as RA and Crohn’s disease.
None of the research addressed the association of Remicade with lymphoma, said Roubenoff. “The risk of infection is what we went after, and one of our abstracts is about the prediction of ill effects from both methotrexate and anti-TNF therapies as far as predictors of risk,” said Roubenoff.
One study of methotrexate and anti-TNF drugs, conducted by Allessandro Foti and colleagues, incorporated proteomic and genetic markers linked to drug benefit or adverse events. “We found a marker for adverse events which had no overlap with the marker set for effectiveness,” said Roubenoff. Foti’s group found one locus was associated with methotrexate adverse events, and four with loci associated with anti-TNF drugs. However, two loci were linked to severe adverse events for methotrexate, while six loci associated with severe events related to anti-TNF drugs.
Roubenoff said the company has not yet discussed its results with Centocor.
More accurate predictions of RA severity can be gained by first stratifying patients based on SNP genetic data, then evaluating the expression of 28 genes based on patient classification, said Roubenoff. This research, by Anita Singh and colleagues, “extends predictive capability beyond just the shared epitope,” Roubenoff said. Without first stratifying patients before expression profiling, “we don’t find a whole lot,” but stratified by “zero, one, or two shared epitopes, all of a sudden the data become very striking,” he added.
The shared epitope is a series of six genes in the HLA locus that confers a “much higher” RA risk, said Roubenoff. The company is validating the study, he said.
A combination of genetic, proteomic, and clinical biomarkers provides “about an 88 percent discrimination for predicting the presence or absence of joint erosion,” according to a study by Nancy Shadick and colleagues, said Roubenoff. The method incorporates shared epitope, serum levels of MMP3, and rheumatoid nodules as the clinical marker to produce an indicator of RA severity, he said. Erosion is the major predictor of disability in RA.
As far as the FDA is concerned, a disease-modifying drug must be able to stop erosions, and the combined method provides a tool, said Roubenoff. “For any Phase III trials, you need to show that effect on erosion will be successful,” he said.
Indeed, the use of these kinds of tools is not coming quickly enough for some. “The rest of the community, particularly the pharmaceutical community, needs to recognize that these types of analyses need to be routinely incorporated into clinical trials so that the information can be mined and interpreted properly, and then translated into a meaningful source of data for patient management” said the North American Rheumatoid Arthritis Consortium’s Gregersen in an e-mail.
The role of Brigham and Women’s Hospital in creating the registry is to recruit and assess patients and collect blood samples, and send the samples to Millennium, which does the metabolomic, genomic, and proteomic work, said Roubenoff. Millennium funds the collaboration; company officials declined to disclose the amount.
Because of RA’s complexity, “we really can’t anticipate successful use of pharmacogenomics in rheumatoid arthritis unless we have the ability to look across a large number of patients before we get into clinical trials, and that’s why we think the registry is so important.” It gives a lot of “lead time” to understand complicated issues to come up with a marker set to apply to the constrained world of clinical trials.
For its part, Millennium hopes to discover biomarkers to apply in its own Phase II clinical trials to “see how well they will predict the response,” said Roubenoff. “If that works and they are validated, we can use them for real in our Phase III studies.” The whole point of the registry is to give the company a “leg up” — without it, the company would find markers in Phase II and validate them in Phase III, Roubenoff added.