In a bid to identify biomarkers that could predict a patient's response to asthma therapies and to provide insights that could lead to better treatments for the disease, GNS Healthcare is lending its bioinformatics expertise and Reverse Engineering Forward Simulation platform to a project that aims to develop a predictive model for asthma based on data collected during a clinical trial.
Partnering with GNS in this endeavor is Harvard Medical School’s Brigham and Women’s Hospital. During the two-year collaboration, the groups will perform “iterations of model building along with validation of our results," Iya Khalil, executive vice president and co-founder of GNS Healthcare, told BioInform. "We also plan to look at other genomic data types … and as we generate more data, we will include [that] in the model building and come up with updated results."
BWH and GNS will use de-identified patient data collected from more than 700 participants in the Childhood Asthma Management Program, a five-year clinical trial that compared inhaled corticosteroids to placebo in patients who were 5-12 years old when enrolled in the study. To build the models, the partners will use genotype and gene expression data; clinical and environmental measures; and clinical outcomes such as hospitalizations, requirements for fast-acting asthma drugs, and reduced lung growth.
Financial terms of the agreement were not disclosed.
Primarily, the models could prove useful for pharmaceutical companies looking to develop new therapies for asthma and for clinicians who treat asthma patients, Khalil said, because they will provide information about the “mechanisms of these drugs within the context of childhood asthma — how they work and also why the drugs work for some people and not for others.”
In addition to better treatments, she noted that there is the potential for intellectual property around any new markers discovered during the course of the study.
This collaboration marks GNS’ first project in the life sciences space under new leadership and an updated business model.
Last month, several founders of the company formerly known as Gene Network Sciences launched a new firm called Via Sciences, which now serves as a holding company for the renamed GNS Healthcare and for Fina Technologies, a GNS spinoff that is applying the REFS technology to the financial industry (BI 11/05/2010).
For this project, GNS researchers will work with Scott Weiss, a professor of medicine at Harvard and an associate physician at Brigham and Women's who was involved in the CAMP trial.
Weiss’ team will provide the genetic, gene expression, and clinical data that will be used to develop the models as well as validate the predictions generated by the platform.
"We will use the variation in genetic and molecular profiles and clinical response across the 700 patients to [create] a model of the disease and drug response," Khalil said.
She explained that CAMP tried to discover "whether anti-inflammatory drugs would change asthmatic outcomes with the primary endpoint being change in lung function based on forced expiratory volume after the administration of a bronchodilator."
The CAMP trial involved 1,041 children with mild to moderate asthma. Of those, 311 children received 200 micrograms of budesonide, a corticosteroid that is distributed by AstraZeneca; 312 participants received 8 milligrams of nedocromil, a noncorticosteroid that was distributed by King Pharmaceuticals until April 2008 when the drug was discontinued in the US; and 418 children received a placebo twice daily.
So far, the researchers have concluded that in children with mild-to-moderate asthma, neither budesonide nor nedocromil is better than placebo in terms of lung function, but inhaled budesonide improves airway responsiveness and provides better control of asthma than placebo or nedocromil. The side effects of budesonide are limited to a small, transient reduction in growth velocity, Khalil said.
The trial, which is sponsored by the National, Heart, Lung, and Blood Institute, is now in its third phase and has been extended through June 2011. This is to allow the researchers to determine the impact of anti-inflammatory therapy on lung function, study the history of asthma through age 26, and to "define patterns of reduced lung function growth and early decline of lung function in young adults" according to the trial description.
The GNS and BWH team will use the REFS platform to build computer models from the trial data that will connect genetics to clinical outcomes. These computer models will be used to discover the key genetic and molecular drivers of both childhood asthma disease progression and patient response to inhaled corticosteroids.
Khalil said that in the first year of the project, the partners plan to produce predictions that can be validated via testing in the lab and then generate more data that can be used to update the models and make new predictions in the second year.
REFS, which is based on a hypothesis-free approach that uses machine learning algorithms, will “learn a model of the system” that identifies connections between the different variables — for example, “how genetic variation can influence changes in gene expression profile that lead to changes in ways that your clinical phenotype can be influenced when you are on the different drugs,” Khalil explained.
During the simulation step, Khalil said, the team will be able to query the model by selecting specific patient genotypes and phenotypes, for example, and then predicting different scenarios such as whether the patient would respond to a particular drug or the likelihood that a patient would benefit from taking inhaled therapies on a daily basis.
Furthermore, she said, “you can … come up with new predictions about new biology … and now that you have learnt a model, you can actually do those predictions and simulations.”
Meanwhile, Khalil said that the partners would work together to select appropriate validation experiments, which could involve analyzing additional patient data from other studies or from the same study to identify new biomarkers. The predictions could also be validated by testing them on immortalized cell lines.
The partners plan to publish their findings as the models are validated.
Khalil also noted that GNS Healthcare hopes to establish additional partnerships with research groups that work particularly in inflammatory disease areas as well as central nervous system disorders.
Past GNS partnerships in these areas include a project with Biogen Idec, which used the REFS platform to identify new drug targets for rheumatoid arthritis patients using genome-wide SNP data, expression data, and clinical data (BI 04/16/2010).
Have topics you'd like to see covered in BioInform? Contact the editor at uthomas [at] genomeweb [.] com.