Eat your heart out, eHarmony.com.
While the popular Internet dating site (and others, like Chemistry.com) claims to use science to introduce compatible singles, real laboratory researchers have created a tool that makes even more extraordinary matches: those between diseases and likely new treatments.
In a paper published in the Sept. 29 issue of Science, researchers announced the creation of the “Connectivity Map,” a searchable online database that allows scientists to pinpoint potential new therapies for almost any disease. Researchers looking for a therapy to counteract a certain illness can search the map for drugs, which have desired effects on genes involved in triggering that condition. This process will help scientists select candidate therapies to go through clinical trials.
“[The project] grew out of a frustration that lots of people in biomedical research have of how difficult it is to satisfy the basic requirement of biomedical research: That, I think, is the ability to connect human diseases with small molecules that make treatment,” said lead author Justin Lamb, a senior scientist at the Broad Institute in Cambridge, Mass. “It’s extremely frustrating and time consuming and difficult to make those connections between diseases and drugs. So our primary motivation was to develop a generic solution to that problem.”
The researchers first translated diseases and chemicals into the common language of gene expression. Each of the 22,000 genes in the human genome codes for a protein. Creating these proteins begins with a process that produces molecules known as messenger RNA, which are responsible for carrying genetic information to the sites of protein synthesis. By measuring the cellular concentration of different messenger RNAs, scientists can determine the level of expression of each gene.
The researchers gathered data on how each disease and each drug affects gene expression. They then created a search function allowing biomedical researchers to type in the genetic signature of a given disease—a list of genes affected by the disease, each of which is accompanied by a sign indicating whether expression of that gene increases or decreases. When queried with a disease signature, the Connectivity Map spits out a rank-ordered list of pharmaceutical compounds that either closely match the query or are nearly the opposite of it.
“If you imagine, for example, a comparison of a disease state against a normal state, maybe 50 genes would be increased in their level, and 50 genes would be decreased in their level,” Lamb said. “So that would be the query. And then the database search would look for drugs who make those 50 genes that go up in the disease state go down, and the 50 genes that go down go up. So you would imagine that that would be [essentially] reversing the disease state, if you like.”
The authors have already used their own tool to find successful disease treatments.
In one instance, they targeted the problem of drug resistance in children with acute lymphoblastic leukemia (ALL). They were able to derive a gene-expression signature for response to the leukemia drug by comparing cells from drug-sensitive patients to those of drug-resistant patients. By plugging this signature into the Connectivity Map, they found a corresponding compound, which, when tested on a cell line, turned the formerly resistant cells into cells that were susceptible to the drug.
Lamb noted that since the FDA already improved the compound for another purpose, clinical trials to determine whether it can be used to treat drug-resistance in ALL are likely to proceed quickly. For this reason, the researchers plan to analyze 1,300 compounds that have already been approved by the FDA in hopes of discovering new potential uses for them.
“This is a terrific body of preliminary work demonstrating the potential feasibility of a large scale connectivity map to facilitate the drug discovery process, and enhance understanding of the biology of diseases and their perturbation by medications,” said William Evans, the director of St. Jude Children’s Research Hospital in Memphis, via email.
Evans, whose research focuses on ALL, said the mechanism the authors used to treat drug resistance—lowering one gene’s expression level—is consistent with previous experiments indicating that over-expression of that gene resulted in resistant cell-lines.
Originally published October 5, 2006