Genes Have a Lot to Say About Depression Treatment Results

Genetic variations may account for differences in treatment outcomes, according to the results of a large antidepressant trial. “This is an encouraging finding that is not quite ready for use, a proof of concept, but it's a step on the way to improving treatment for individual patients,” said study coauthor Husseini Manji, M.D., chief of the Laboratory of Molecular Pathophysiology and director of the Mood and Anxiety Program at the National Institute of Mental Health (NIMH). “It does show that genetic differences play an important role in treatment response.”

Manji and colleagues compared DNA data with patient outcomes of the Sequenced Treatment Alternatives for Depression (STAR^*D) trial, which treated patients with the SSRI citalopram (Psychiatric News, January 20, April 21). DNA samples were taken from 1,953 of the 4,041 eligible participants in the trial. Sixty-eight candidate genes were chosen by an expert panel based on existing evidence of association with antidepressant outcomes, association with mood disorders, and known functional variants. Subjects were divided by whether they remitted or responded to treatment with citalopram. The study results were published in the May American Journal of Human Genetics.

Researchers compared allele and genotype frequencies between subjects in both groups. They tested 768 single nucleotide proteins (SNPs) for association with response and remission in a discovery sample of 1,380 subjects. Only one SNP, rs7997012, reached significance for the same allele and phenotype in the discovery sample and in the replication sample (n=438). That SNP resides on the gene HTR2A, which encodes the serotonin receptor. They found a significant association between the A allele of rs7997012 and treatment response, as measured by the QIDS-C score. Patients homozygous for the A allele showed an 18 percent lower risk of not responding to treatment.

“Eighteen percent won't change clinical practice, but in combination with other genes it could become a practical clinical tool by telling us who is likely to respond to a drug and what side effects they might experience,” said Manji in an interview.

However, response was not even across all subgroups. Patients in the trial had identified themselves regarding race, and the researchers grouped them into “white” (78.2 percent), “black” (16.0 percent) and “other/mixed” (5.8 percent). Dividing the sample into“ white” and “black” strata revealed that the A allele associated with better treatment outcome was found six times more often among white than black participants. The allele frequency in whites was 0.42, in blacks, 0.06. In fact, said the researchers, “[T]he association between HTR2A and treatment outcome was largely confined to the white participants.”

Differences in response to antidepressants tied to race are often attributed to psychosocial causes, noted the authors, but this study suggested that genetic variants ought to be considered as well.

“Further genetic discoveries like this could take some of the guesswork out of treatment,” said Manji. “Perhaps using a panel of simple genetic tests in the doctor's office would give us an 80 percent chance of starting a patient on the right antidepressant, rather than trying two or three different drugs over a period of months to find the right one.”

However, both researchers and clinicians must be cautious about using racial categories expressed by patients in the study, Alan Guttmacher, M.D., deputy director of the National Human Genome Research Institute (NHGRI), told Psychiatric News. “Genes vary by the history and geography of one's ancestors, but are not linked to racial definitions.”

The NIMH researchers, concerned about that issue, compared several genetic markers known to be more prevalent in African populations with what patients said about themselves. “We were a little surprised that it matched up as well as it did with self-identification,” said Manji.

Use of the candidate-gene method probably represents an imperfect middle step until a whole genome approach becomes more practical for researchers, said both Manji and Guttmacher. The NIMH researchers took a good guess at a group of genes and found one that worked, but some other gene could turn out to be even more important, said Guttmacher.

Using the whole human genome would have been impractical when the STAR^*D research began. In 2002, scanning the entire DNA sequence of 1,000 people with a disease and another 1,000 controls without it would have cost $10 billion, said Guttmacher. Today, following research advances from the NHGRI's International Haplotype Map (HapMap) project, the price for such a project has fallen to $2 million to $3 million. Further development should lower costs into a range that could be considered by many researchers.

“The technology is there,” said Manji. “Discussions are ongoing to drive down the price.”

The NIMH study will have to be replicated and then tried with other genes. Plans are under way to accelerate that process, said Guttmacher. The Genetic Association Information Network (GAIN), a collaboration of the National Institutes of Health, the Foundation for National Institutes of Health, and private industry, has been set up to use the whole genome approach to link potential genetic patterns with a disease. Once those patterns are identified, it should be possible to pinpoint more easily specific genes involved in the disease process than has been possible so far. Mental health research will have a literal seat at the table, since the GAIN executive committee includes NIMH Director Thomas Insel, M.D.

An abstract of “Variation in the Gene Encoding the Serotonin 2A Receptor Is Associated With Outcome of Antidepressant Treatment” is posted at<www.journals.uchicago.edu/AJHG/journal/issues/v78n5/43135/brief/43135.abstract.html>.▪