July | August 2006 | SAN FRANCISCO — Beyond Genome 2006 — one of Cambridge Healthtech Institute’s flagship conferences — traces its origins to 1991, when the Human Genome Project was just getting off the ground. But this year’s meeting* marked the first time it could truly be said to be beyond the human genome — the sequencing of the last human chromosome was published shortly before the meeting.
About 700 attendees gathered in San Francisco to discuss the major technology trends in postgenomic biology, including systems biology, RNAi, proteomics, and personalized medicine.
The meeting was kicked off by Pearl Huang, GlaxoSmithKline’s director of oncology. She described progress in several targeted cancer therapeutic programs, including an inhibitor of BRAF, dubbed 590885, which is mutated in melanoma, colon and other cancers. Another drug, Tykerb, is an ErbB inhibitor, but unlike Tarceva and Iressa, hits ErbB1 and B2 equally. Huang says Tykerb is a promising therapeutic for breast cancer and, as it fortuitously crosses the blood-brain barrier, could be particularly effective against brain metastases.
Perhaps the most popular speaker — judging by the crowd that enveloped her afterward her plenary talk — was Ellen Feigal, who outlined the mission of the Critical Path Institute (C-Path). This is a new nonprofit, based in Arizona, which Feigal says is calling for “collaborations to find innovations for preclinical and clinical testing of small-molecule drugs, biologics, medical devices, and tools.” Areas of particular interest include toxicogenomics, bioinformatics, and clinical bioterrorism.
In the Predictive Safety Test Consortium, for example, Feigal said members share data and check biomarkers to lower the cost of preclinical drug safety assessment. The consortium includes many big pharmas, the FDA, and SRI (Strategic Research International). Additional projects include a search for cardiovascular disease biomarkers, a pilot early-alert system to identify post-approval safety issues earlier than present, and a program to accelerate drug development for rare diseases.
Stephanie Van Bebber (UCSF) discussed the economic value for pharmacogenomics. The much-lauded Herceptin diagnostic has been successful despite initial concerns, with the HER2 diagnostic test serving as a “gatekeeper to accessing the drug.” By contrast, Iressa has been less successful because the drug’s target was only identified after the drug was essentially pulled from the market. Roche’s P450 Amplichip has also experienced “slow adoption despite potentially wide impact,” said Van Bebber. The test is relevant to $12 billion in prescription sales annually.
Brian Van Ness (University of Minnesota) also lamented the lack of clinical utility for cytochrome P450 testing, which he said might have prevented the tragic death of an infant with fetal alchohol syndrome who suffered a Prozac overdose. Many tables of recommended changes in drug dosing depending on CYP2D6 genotypes. Van Ness and colleagues have developed a custom SNP chip of some 3,400 SNPs in candidate cancer genes.
At Celera Genomics, James Devlin detailed a major diagnostics project in coronary heart disease (CHD), in collaboration with the Cleveland Clinic. Many existing risk factors — hypertension, smoking, diabetes, and age — show a relative risk (RR) of about 2. “Most genetic variants don’t exceed an RR of 2,” said Devlin.
In an effort to bring SNP genotyping technology to the clinic by calculating a simple Genetic Risk Score, Celera’s Atherosclerosis Risk in Communities (ARIC) project has identified 11 SNPs in 10 genes that predict CHD risk. By producing a cumulative risk score — awarding +1 for a risk-raising homozygote (two copies), 0 for a heterozygote, and -1 for a risk-lowering homozygote — Devlin said the highest-scoring individuals had an RR of 1.7 — thus likely to be deemed significant by physicians.
Mark McCamish, Perlegen Sciences’ chief medical officer, says his company wants to “add science to the ‘Art of Medicine,’” although he added, “not all drugs can be personalized using pharmacogenomics.” The FDA label for Meridia advises patients to monitor their weight loss for four weeks to indicate whether they have a significant chance to lose weight. “A pharmacogenomics test could give this probability assessment,” said McCamish.
In one case study, Perlegen is collaborating with a biopharma to halve the number of adverse events associated with a specific drug. Using a pooled SNP genotyping approach, Perlegen identified a group of 20 SNPs that were predictive. Half of the at-risk patients could be identified with this SNP panel, while the remaining patients could be offered alternative therapies.
Francois Iris was one of the first scientists at Millennium Pharmaceuticals in the early 1990s, before leaving to establish Biomodeling Systems in Paris. This integrative modeling company has developed 10 disease models in five medical areas. Iris treats data with the “Genghis Kahn procedure — it’s brutal, you keep the top and throw out the rest.” Data, Iris stressed, must be placed in its relative context. Using an iterative modeling process, Iris’ team has identified key proteins in the brain pathology of Creutzfeldt-Jakob disease.
*Beyond Genome 2006; Fairmont Hotel, San Francisco; June 19-21, 2006.