|Send to printer »|
Feature Articles : Jan 1, 2013 ( )
Genomic Analysis Drives Biomarker Discovery
The application of genomic technology and R&D strategies to the development of personalized medicine and the enhancement of global food production were among the topics discussed at “ICG Americas” held recently in Philadelphia.
The meeting was co-organized by Shenzhen, China-based BGI and The Children’s Hospital of Philadelphia (CHOP).
Genomics offers the possibility of accelerating the identification and assessment of new therapeutic strategies, Huanming Yang, Ph.D., chairman and professor at BGI, explained. When genomic analysis reveals a mutation affecting a biochemical pathway or gene network, a safe drug able to intervene in that pathway may already exist, he added, providing a shortcut to later-stage clinical testing in a targeted patient population.
Jun Wang, Ph.D., executive director of BGI, gave the participants a sense of how quickly the field of genomics is advancing and a glimpse into the future. BGI has sequenced 639 different plant and animal reference genomes, with the oyster genome the most recent to be published. To annotate a genome “is now routine,” Dr. Wang said. The company is currently focusing on completing sets of genomes, such as the 50 bird species included in its avian phylogenomic project.
In agriculture, BGI is utilizing genomics to select for desired traits in crop plants. Essentially, said Dr. Wang, the goal is to predict phenotype based on whole genome sequence data—to do “breeding in the computer, then the field.” This “digital revolution of agriculture” will begin in maize.
BGI recently announced a collaboration with the Bill & Melinda Gates Foundation that will involve both global health and agricultural development.
Cancer Genomics Project
The Cancer Genomics and Personalized Medicine project under way at BGI involves building a cancer mutation map and determining the frequency of recurrent mutations and identifying cancer driver genes. Much progress has been made to date in surveying the landscape of somatic mutations and genomic variation in cancer and how they relate to tumor biology. Presenters at the meeting described work under way to identify biomarkers to support oncology drug development and drive the discovery of personalized cancer therapies and methods for predicting tumor response to treatment.
James Watters, Ph.D., head of applied genomics at Sanofi Oncology, suggested that it can hopefully help improve the field’s overall track record. About 1 in 20 oncology drugs in development succeed in becoming commercialized, with many failing in late-stage trails; approximately half of the compounds that make it to Phase III studies fail.
Describing next-generation sequencing (NGS) technology as “quite transformative” for translational medicine, Dr. Watters outlined four main approaches for how NGS is being applied to study human tumors: PCR amplicon sequencing, targeted gene panel hybrid capture, whole exome sequencing/hybrid capture, and RNA-Seq. Each strategy has advantages and disadvantages, and when a novel somatic mutation is identified, one of the main challenges is to determine whether it has functional consequences and to understand what the biological and phenotypic effects might be.
Assessing structural changes and complex rearrangements, detecting copy number variation (CNV) with high sensitivity, and performing NGS at the single-cell level represent additional ongoing challenges.
Funda Meric-Bernstam, M.D., medical director and professor, Institute for Personalized Cancer Therapy (IPCT), University of Texas, MD Anderson Cancer Center, described the evolution of genomic biomarker-selected clinical trials at her institution. When possible, patients are allocated to Phase II trials based on a molecular profile composed of somatic mutations linked to tumor types and predicted drug response.
Dr. Meric-Bernstam explained that nearly 600 patients are enrolled in the IPCT Clearinghouse Protocol, which is aimed at establishing a database of frequent somatic mutations in various patient populations and tumor types to assist in the delivery of personalized cancer therapy.
Mao Mao, Ph.D., research fellow at Pfizer and president of the Asian Cancer Research Group, jointly established by Eli Lilly, Merck, and Pfizer, emphasized the importance of identifying actionable mutations as the basis for developing targeted cancer therapeutics. To search for drivers of oncogenetic changes at the cellular and subcellular level, the group screens whole tumor genomes to identify potential drivers and their prevalence in well-annotated cohorts. The initial findings of NGS have shown that “30X whole genome sequencing is cost effective,” said Dr. Mao.
Dr. Mao explained that about half of all hepatocellular carcinomas were related to hepatitis B virus (HBV) infection. HBV is a double-stranded virus capable of integrating into the human genome. Dr. Mao’s group is taking a whole genome sequencing approach, using paired positive/negative tumor samples to study the hypothesis that HBV integration is associated with cancer.
The group has extracted the HBV integration breakpoints and found that tumor samples tend to have more HBV integrations than nontumor samples. Of 344 HBV integrations in tumor samples, 163 (47.4%) significantly affected the tumor, reported Dr. Mao. Integration breakpoints recurred in six particular genes in more than one sample, suggesting that HBV has preferred integration sites.
Predicting Therapeutic Response
Merck Research Laboratories is using molecular profiling to evaluate disease subpopulation and cancer drivers, but one “cannot just rely on [the study of] tumors,” said Andrey Loboda, Ph.D., director and head of data analysis, informatics & analysis, who described Merck’s cell line biomarker discovery project.
Compound libraries are screened against large panels of cell lines. The cell lines are randomly divided into two sets, a test set and a training set, and this strategy is used to identify biomarkers that can be used to predict the response to a compound. The drug-response profiles developed from cell-line screening well represent treatment responses seen in the clinic, reported Dr. Loboda.
As work progresses to identify genomic biomarkers for molecular subtypes of tumors and to identify drivers of oncogenesis, and as the sample size and database of tumor profiles continue to grow, these efforts will have increasing implications for clinical trial design and for preselective biomarker-driven enrollment, Dr. Loboda predicted.
Noting the publication of five major papers on the breast cancer genome within the past few months, Joe Gray, Ph.D., director of the Spatial Systems Biomedicine Center and chair of biomedical engineering, Oregon Health & Science University, said, “I think we pretty much know chapter and verse the [genomic] landscape of untreated breast cancers.” He described 45 regions of recurrent amplification or deletion and approximately 10 distinct subtypes associated with outcomes based on expression and copy number.
Furthermore, he said, “the spectrum of aberrations differs according to subtype.” Dr. Gray reported a somatic mutation rate of about 1/Mb (3,000 per tumor) in breast cancer, composed of a few recurrent and many rare mutations.
From a clinical perspective, this information can be used to develop a genomic decision tree, mapping information on treatment response onto the arms of this tree. In this way it might be possible to develop a model of a tumor’s intrinsic genomic diversity and to link the genomic aberrations to therapeutic response.
Ultimately, the goal is to discover predictive markers of therapeutic response, and Dr. Gray’s group is pursuing a quantitative approach based on screening a range of doses of more than 140 therapeutic compounds in replicates of 54 cell lines. The researchers then use this data and machine-learning methods to develop a predictive algorithm that can select the best drug for an individual patient.
They have developed a methodology for studying pathways in tumors and cell lines aimed at identifying activity signatures and demonstrating subtype specificity for pathway activities. This involves the use of an siRNA interrogation platform for gene knockdown in cells spotted on a culture surface and stained for a specific phenotypic characteristic related to cancer.
Using this approach the group has been able to develop imaging assays for most of the phenotypic characteristics of tumors, reported Dr. Gray. They are now working to improve this model by controlling the microenvironment around the printed cell spots.
BGI and CHOP Launch Clinical NGS Services
The Children’s Hospital of Philadelphia (CHOP) and BGI formed the [email protected] Partnership and established the Joint Genome Center (JGC) to conduct large-scale human genome sequencing and bioinformatics analysis focused on the discovery of genes linked to rare and common pediatric diseases. The JGC recently announced that it will begin to offer clinical NGS services at CHOP through the hospital’s department of pathology and laboratory medicine in a CAP/CLIA-compliant environment.
The Center’s services will include exome sequencing for the diagnosis of specific diseases including heritable disorders and pediatric cancers. The JGC will also extend the research services it has been providing to support new drug development and clinical trials in compliance with CAP and CLIA standards.
The JGC has five high-throughput sequencers, with plans to scale up to 20 sequencers. It can perform human whole exome sequencing, targeted sequencing, whole genome re-sequencing, ChIP-Seq and RNA-Seq, and NGS data analysis. CHOP maintains a pediatric biobank, with many samples linked to clinical data using an electronic medical record system. The robot-controlled, -20°C DNA sample storage unit can accommodate two million bar-coded samples stored in 96-well microtiter plates or 2 mL tube racks.
© 2013 Genetic Engineering & Biotechnology News, All Rights Reserved