The genetic susceptibility to major depression was tested using genotype frequency comparisons between subjects with a history of depression and controls. We detected three polymorphisms (SLC6A4 intron 2 VNTR, 2 COMT polymorphisms) which may contribute to the genetic susceptibility to major depression, however, only one (SLC6A4) which held up to multiple comparison corrections.
The comparison between the non-responders and responders is a method to detect the genetic basis for antidepressant response, in general. Out of the 21 polymorphisms tested, only one (DRD4 exon 3 VNTR) was statistically significant. This is compared to the 3–4 polymorphisms which demonstrate a genetic susceptibility to depression. It was unpredicted that the genetic basis for antidepressant response may differ from the genetic basis of susceptibility to major depression.
We further demonstrated that the number of risk genotypes is not consistent across phenotypes. The non-responders have the most risk genotypes, followed by (in decreasing order) the depressed group as a whole, the responders, and the controls. This supports the notion that Major Depressive Disorder is polygenic, and that the number of risk genotypes may be an indication of the susceptibility to Major Depressive Disorder as well as the severity of the disorder.
In retrospect, we recognize that much of our current understanding of the pathophysiology of depressive disorders has been inferred from the prevailing hypotheses of mechanisms for antidepressant action. Accordingly, candidate genes selected for their function are commonly associated with monoamine function or their putative intracellular responses to neurotransmitter activation. The neurobiology of treatment resistance is not well understood; it may represent an extreme phenotype along a unique pathological continuum of depression. This interpretation may be supported by the fact that the usual monoamine-based treatment interventions, by definition, provide little or no benefit to patients with treatment resistant depression.
The Hamilton Depression Rating Scale has been for decades the gold standard of depression symptoms quantification and has been utilized in >95% of antidepressant treatment studies. There have been lately a number of concerns primarily about the scale's ability to reliably quantify antidepressant responses given the multi-dimensionality of the scale, which may non-specifically determine response in individuals treated with sedating agents such as TCA's. In the other hand, given the number of non-melancholic specific items, it may be less sensitive at detecting an antidepressant response. It is possible that alternative tools such as the Bech Rafelsen or the Montgomery Asberg Depression Rating Scale among others may more accurately reflect antidepressant responses leading to a more reliable detection of therapeutic effects.
This pilot study provides a proof of concept that two depressive phenotypes (antidepressant response and non-response) may be subtypes of Major Depressive Disorder. In terms of medication response, these patients are clearly distinct from each other, thus it is logical to propose their genetic constitution may be distinct. The polymorphisms analyzed in this study is by no means an exhaustive list of polymorphisms which play a role in Major Depressive Disorder, however it is a fairly representative list for the monoamine related polymorphisms thought to confer susceptibility to major depression. It is true that the mean age of the controls is younger compared to the depressed groups, thus they have not necessarily passed the age of onset typical of major depressive disorder. As a result, we may be committing a type I error in the analyses, and be missing or underestimating genetic differences that exist between the controls and depressed subjects, however, we are confident in the effect that we did detect. A large number of comparisons between groups for each polymorphism were initially performed in this study. To correct for false discovery, a q-value was determined, which provided an indication of statistical significance. We accepted a q-value less than 15%, which indicates that 15% of the time we considered out findings to be significant, we would be incorrect. Given that this is a pilot study with relatively small sample size, our single-polymorphism findings may be false positives, and thus necessitate replication in an independent and larger sample. Polymorphisms in alternative neurotransmitter systems, neurotrophic pathways, neurosteroids and antidepressant metabolic pathways (e.g. cytochrome p450 factors) should be the focus of further research.