Verified by 
I was interested to read of research at Penn State University (soon to be published in the journal Biological Psychiatry) that uses a new strain of laboratory mice that possess a very specific genetic defect to model forms of treatment-resistent depression in humans. This work demonstrates both the promise and the peril of animal models of depression.
I see promise in using animal models because I believe that mood has been selected for in a variety of organisms, and certainly in mammals. Thus, I accept that one can identify phenotypes (outward manifestations) of low mood or depression in a mouse. Indeed, the researchers in this case claim and adduce support for the idea that their genetically defective mouse has similar behavioral, hormonal, and neurochemical characteristics to human patients with depression. Perhaps most important, I accept the assays that can be used as a test of whether an animal is exhibiting depression behaviorally, such as reduced exploration of a novel enviornment, which then can in turn be used as a probe of the efficacy of a new antidepressant compound. Indeed, supporting the plausibility of animal models, animal models like the rat swim test have been used in early phases of drug testing and design for decades.
Where I think this work goes wrong is the attempt to model complex disorders like depression in other species by tinkering with single genes. In this case, the genetic defect in the depressed mice has to do with the function of specific receptor in the brain called the GABA-A receptor (this controls the response to the neurotransmitter gamma-aminobutryic acid). More broadly, the effort to find the single genes responsible for any major illness has been an abject failure. Depression candidate genes have come and gone and there is very little reason to hope or expect that we will find a single gene that explains a large percentage of cases of depression in humans. At best, the case is far more complicated and unwieldy: the genetic component of depression risk is explained by a large number of genes that interact with one another and with environmental stress. Thus, a single gene by itself means very little.
The problem here is that the promise of work in animal models has been tethered to this perilous situation in genetics, where studies utilize "knockout" mice, which are engineered to be missing a gene, or to hold a single genetic defect, as in the current paper. To the authors' credit, they are aware that the single gene by itself is insufficient as an account of depression and they are doing studies to integrate the role of developmental stress into their model. Neverthless, the problem remains that this field has become a mess; even if some of the risk genes can be identified (this has proven far more difficult than originally hoped), there are too many of them, and the genes are interacting with one another and with the environment in complex and unpredictable ways. Sadly, we cannot count on unmighty mouse to save the day.
