Environmental Factor, September 2010, National Institute of Environmental Health Sciences

NIEHS-led study advances predictive genomics

By Thaddeus Schug
September 2010

Bushel is head of the Microarray and Genome Informatics Group in the NIEHS Biostatistics Branch. With his colleague, Paules, Bushel has been instrumental in spearheading the Toxicogenomics Integrated with Environmental Science series of international conferences. (Photo courtesy of Steve McCaw)

This spring the Society of Toxciology recognized Paules with the 2010 Leading Edge in Basic Science Award "for his work in the integration of genomics into the investigation of the molecular basis of injury and disease processes," describing him as "a visionary... who has diligently positioned NIEHS at the forefront of the field." (Photo courtesy of Steve McCaw)

In a new study published online July 30, a team of 23 investigators led by NIEHS bioinformatician Pierre Bushel, Ph.D., shows that genomic indicators in the blood can serve as diagnostic biomarkers predictive of drug-induced liver injury (DILI).

The study, which points to potential applications in the clinical setting and in drug discovery, provides a proof of principal for ongoing efforts by the MicroArray Quality Control Phase II (MAQC-II) project led by Leming Shi, Ph.D., of the U.S. Food and Drug Administration (FDA) National Center for Toxicological Research (NCTR) to develop best practices for microarray gene expression utilization in biomarker discovery.

According to the authors of the study(http://www.nature.com/tpj/journal/v10/n4/full/tpj201033a.html) , drug-induced hepatotoxicity results in thousands of emergency room visits each year and is the major cause for a drug to be withdrawn from the market, have its use restricted, or be required to use a warning label. They explain that DILI is difficult to diagnose in its early stages with serum markers currently in use or by liver biopsy.

"There's a huge need to put better information in the hands of clinicians, and I think genomics is an important tool for helping us reach that goal," said one of the coauthors of the study, Rick Paules, Ph.D. Paules is a senior scientist and principal investigator in the Laboratory of Toxicology and Pharmacology and director of the NIEHS Microarray Core facility.

The NIEHS-led study is one of 11 papers published in a special August 2010 issue of The Pharmacogenomics Journal (TPJ) by researchers in the MAQC-II consortium. Bushel, Paules, National Toxicology Program Scientific Administrator Jennifer Fostel, Ph.D., and other NIH members are also co-authors on a summary report(http://www.nature.com/nbt/journal/vaop/ncurrent/abs/nbt.1665.html) of the entire MAQC-II project published online July 30 in Nature Biotechnology.

Genomic indicators predict a phenotype of toxicity

To test the utility of genomic indicators for predicting very early stage DILI, the team used a unique data set from the MAQC-II project that was generated by the NIEHS National Center for Toxicogenomics (NCT) and contributed by Paules, consisting of gene expression data from two tissues in rats - blood and liver. They found that by analyzing genomic indicators in the blood as biomarkers, they could predict liver necrosis or hepatic cell death across a variety of chemical compounds that target the liver, as shown by histology of liver samples after rats had developed frank DILI.

The researchers compared results on two microarray platforms for profiling the liver data and were able to predict DILI with as great as 92.1% accuracy. In an independent validation component of the study, they also showed that these genomic biomarkers predicted acetaminophen-induced liver injury with a high accuracy, as well as predicted liver injury induced by two non-therapeutic chemical compounds with an accuracy that was slightly lower than with acetaminophen.

"Our results strongly support the claim that genomic indicators in the blood can serve as biomarkers of necrosis as a form of a chemically-stressed adverse effect on the rat liver, and give credence to the acquisition of gene expression signatures from minimally invasive biomaterial sources potentially for diagnostic testing of DILI in humans," explained Bushel.

Reflecting on the long history of pioneering research in toxicogenomics at NIEHS, Paules concluded, "While the challenges of developing signatures to utilize as clinical biomarkers of specific adverse effects are great, the prospects have never been brighter. We really hope to provide better science for better treatment of disease."

Figure 4 - Network analysis of the upstream and downstream regulation. The nine genes (marked with solid circles) are direct targets of 10 transcription factors. The downstream genes belong to three processes implicated in liver injury. (Figure courtesy of Nature Publishing Group, Pierre Bushel, and Rick Paules)

The study's findings underscore the potential for using microarray gene expression technology as a means to individually tailor drug treatments to a patient's genome. Use of diagnostic genomic indicators is also likely to speed up development of high-throughput drug screening assays. Ultimately, genomic medicine could reduce the costs and duration associated with clinical trials, which often spell death for promising new drugs.

Citation: Huang J, Shi W, Zhang J, Chou JW, Paules RS, Gerrish K, et al.(http://www.nature.com/tpj/journal/v10/n4/full/tpj201033a.html) 2010. Genomic indicators in the blood predict drug-induced liver injury. Pharmacogenomics J 10(4):267-277.

The Microarray Quality Control Consortium. (http://www.nature.com/nbt/journal/vaop/ncurrent/abs/nbt.1665.html) 2010. The MicroArray Quality Control (MAQC)-II study of common practices for the development and validation of microarray-based predictive models. Nat Biotechnol. Epub ahead of print. doi:10.1038/nbt.1665

(Thaddeus Schug, Ph.D., is a postdoctoral research fellow in the NIEHS Laboratory of Signal Transduction and a regular contributor to the Environmental Factor. He is currently on detail as a program analyst in the NIEHS Division of Extramural Research and Training.)

Moving forward - the SEQC project

Despite the highly successful efforts of the NIEHS-led team, several of the other teams experienced difficulty in deriving classifiers from certain microarray clinical studies for prediction of certain clinical end-points. In an editorial accompanying the 11 papers in TPJ, William Slikker Jr., Ph.D., observed, "Quality control and appropriate bioinformatics processing will remain a challenge for any new high-throughput molecular technology."

These challenges also led the editors of the two Nature-published journals to conclude, "The results from this large-scale project involving participants from academia, industry, and governmental agencies will likely be useful for developing classifiers for data from high-throughput assays other than DNA microarrays."

As phase II of the MAQC project nears its end, FDA plans a third phase, MAQC-III, also called the Sequencing Quality Control (SEQC) project. The SEQC project aims to assess the technical performance of next-generation sequencing platforms by generating benchmark datasets with reference samples. They also plan to evaluate the advantages and limitations of various bioinformatics strategies in RNA and DNA analyses.

NIEHS will continue its productive partnership with the FDA and other agencies to extend the successes of MAQC-II and to look further into the quality and reproducibility of next generation sequencing through SEQC. "This is another example of how NIH and FDA are working together to advance translational and regulatory science," said Linda Birnbaum, Ph.D., director of NIEHS and the National Toxicology Program. "I'm proud of the role NIEHS investigators played in this important undertaking."

"Atrazine causes abnormalities ..." - previous story next story - "Unruly protein may shape learning..."
September 2010 Cover Page