A Step Toward Personalized Medicine in Depression

Over the past decade, considerable efforts have been expended at developing possible genetic markers for predicting responses to antidepressant agents. Results to date have largely indicated that some genetic variants may be particularly helpful in predicting risk for side effects. These have included the study of pharmacokinetic-related genes, such as P450-2D6, as well as pharmacodynamics markers involving genetic variation for specific reuptake sites or receptors.

We recently reported in the American Journal of Psychiatry (1) that a common variation for a single nucleotide polymorphism (SNP) for the ABCB-1 gene—which encodes for a protein pump that controls efflux across the blood brain barrier of a number of commonly used antidepressants out of brain—could predict both high rates of response and lower side effects to specific antidepressants.

This observation follows on another publication (in Pharmacogenomics Journal) by our group this year of another, independent SNP that similarly predicted better response and lower peripheral side-effect burden stemming from drugs that are substrates to the transport pump such as escitalopram (2).

This predictive combination may seem counterintuitive, since some may expect that higher brain levels of antidepressants that are capable of activating the transport pump would be associated with greater antidepressant effect and higher side-effect burden. A possible explanation might be that patients who carry specific alleles rapidly achieve high brain levels and respond quickly, but in other organs in which the specific pump is not active, higher concentrations of drugs are not particularly achieved, and thus patients actually experience lower peripheral side effects.

In a commentary in the same issue of our recent AJPmanuscript, Francis McMahon, M.D., a senior investigator at the National Institute of Mental Health, indicated our results were encouraging for developing biomarkers to guide treatment (3).

How would a marker that controls efflux of certain antidepressants be used in clinical practice? Aspects of this have been explored by investigators at the Max Planck Psychiatric Institute in Munich, where variants for the gene were originally demonstrated to be associated with significantly higher rates of response (4-6). They have noted that knowing if a patient carried a favorable allele that resulted in less efflux of a drug from the brain resulted in higher remission rates than if the information was not known. Patients who are carriers of such an allele can be given a drug that is actively transported by the pump, ensuring maximum benefits of treatment. Conversely, patients who are not carriers would likely need higher doses to achieve therapeutic brain concentrations. This might need to be done gradually.

What needs to be done next? The SNP we reported on recently needs replication, and a sample has already been collected. Those other ABCB-1 SNPs that have had multiple replications in relatively smaller studies (2, 4, 5) would benefit from a larger scale cost-benefit study to determine the cost-benefit of the test.

In the end, we are getting closer to personalized approaches to treatment. As McMahon pointed out in AJP, genetic markers might also be combined with other markers—imaging, cognitive measures, etc.—to establish a truly effective basis for selection of treatment. ■