click to enlarge
A lab technician deploys an update to the fully automated Complete Genomics sequencing system.  

Drug development is an extremely costly endeavor. Estimates of the total expense of advancing a new drug from the chemistry stage to the market are as high as $2 billion. Much of that cost is attributable to drug failures late in development, after huge investments have been made. Drugs are equally likely to fail at that stage for safety reasons, as for a lack of efficacy, which is often well-established by the time large trials are launched.

Great strides have been made in identifying new drug targets and understanding mechanisms of disease. However, the science of drug safety lags behind, and the rare toxic reactions that torpedo otherwise promising drugs are often never fully explained.

Genomic sequencing tools are now becoming available that may close the gap between the science of drug efficacy and the science of drug safety.

Genome-wide association studies (GWAS)—scanning the genome for single nucleotide polymorphisms—have been the backbone of genomic studies since they were first introduced in the mid-2000’s. The Hamner-UNC Institute for Drug Safety, at the University of North Carolina, is leading a consortium of investigators to develop new tools for predicting and characterizing drug-induced liver injury (DILI).

DILI is a greater contributor to failed drug approvals, market withdrawals, usage restrictions, and black box warnings than any other toxicology event. The industry lacks predictive tools that can identify potential for DILI early in the drug development process.

It was once hoped that science could identify a human phenotype that was most susceptible to DILI, but that effort has come to nothing. Instead, it’s becoming apparent that DILI is a unique event that is specific to individual drugs. That means rather than identifying one or a few genes that generally contribute to DILI, each drug requires its own screening process.

If liver toxicity problems can be pinpointed and prevented, some drugs that have been withdrawn may even be able to re-enter the market with a companion diagnostic test to screen for toxicity.

The best example of an effort in that direction so far is a companion diagnostic developed by Novartis for lumiracoxib. The drug is a COX-2 inhibitor that was available in Europe in 2006 and 2007. It was carefully formulated and tested to avoid the cardiovascular side effects that had caused the failure of Vioxx. In 2007, however, lumiracoxib was withdrawn from the market because of liver-related adverse events. Using genome-wide association studies, Novartis has developed a genetic test that it plans to market as a companion diagnostic for lumiracoxib, to screen out patients who are susceptible to liver injury.

At the Hamner Institute, GWAS are carried out using Illumina chips with a million or more SNPs that are found at a frequency in the population of 5% or higher. That means anything that turns up on a GWAS will be relatively common.

Researchers at Hamner have identified over a thousand patients who have had liver reactions to over 200 different drugs. "What we found was that there are no common susceptibilities to having liver injury from multiple drugs. It appears the susceptibilities are going to be drug or drug-class specific," says Paul Watkins, MD, chair of Steering and Genetics for the National DILI Network.


click to enlarge

Complete Genomics’ data storage facility.  

According to Watkins, the safety bar is rising for drug approvals. Many newer cancer drugs, for example, don’t have the harsh toxic side effects of the standard chemotherapy regimens. "When you develop a new drug," says Watkins, "it has to be very special indeed if it has serious risks to it."

Although GWAS has been been very productive for these types of studies, the costs of whole genome sequencing and exome sequencing have fallen to the point that GWAS is starting to fall by the wayside.

Complete Genomics offers an outsourced whole human genome sequencing and analysis service that provides genomic data to biopharmaceutical researchers at an unprecedentedly low cost. Its sequencing facility in Mountain View, Calif., has the capacity to sequence and analyze over 600 complete human genomes per month, and the company is planning to increase that capacity to as much as 1,200 genomes per month.

In addition to drug safety studies, Complete Genomics serves researchers sequencing needs to solve a variety of problems. In one perplexing case from the University of Texas Southwestern, Complete Genomics helped to sequence the genome of an 11-month old child with very high levels of LDL cholesterol. The standard test for the genetic defect failed to detect a missing protein, because the child was receiving it from his mother’s milk, and it was present in his blood. Genomic sequencing confirmed that he was missing that gene.

At the Scripps Translational Science Institute in La Jolla, Calif., researchers are leveraging the power of whole genome sequencing for a number of projects, including one called idiopathic diseases of man, or IDIOM. Taking a similar approach as drug safety screening, IDIOM seeks to identify the genetic origins of idiopathic conditions. These conditions have origins that are unknown, and patients with them defy straightforward diagnosis.

"If you identify a mutation that disrupts functioning of a particular gene, that might shed light on the pathophysiology or pathobiology of the disorder. Maybe a therapy could be crafted based on that insight," says Nicholas Schork, Director of Biostatistics and Bioinformatics for the institute.

Schork says that sequencing needs to be highly accurate, and even with the most accurate technology, some errors will creep in to the sequence. "It’s very rare to get a perfect sequence," says Schork. In addition, genome sequencing requires powerful computational and data analysis tools to analyze the terabytes of data produced by human genome sequences.

One human genome creates approximately one quarter to a whole terabyte of data to be stored and processed.

Complete also provided sequencing services to Genentech to characterize mutations in a non-small cell lung cancer genome. The sequence revealed 50,000 mutations, roughly one mutation for every three cigarettes smoked by the subject.

"We take a factory approach to genome sequencing," says Jill Hagenkord, MD, Complete Genomics’s chief medical officer. "And, as such, benefit from economies of scale."

Exome sequencing has been another popular cost-saving alternative for genomic studies. It only targets about 1% of the genome, whereas whole genome sequencing addresses a majority of the genome, including coding and non-coding regions. Like GWAS, however, exome sequencing may be heading rapidly toward obsolescence.

"For approximately $4,000, almost the same as sequencing an exome, Complete Genomics can now sequence a whole human genome and produce high quality variant data in an easy-to-understand format. One has to ask: If the cost of these two approaches are comparable, and the data is similarly accessible, why would a researcher settle for incomplete data?" says Hagenkord.

About the Author
Catherine Shaffer is a freelance science writer specializing in biotechnology and related disciplines with a background in laboratory research in the pharmaceutical industry.