Project 4. Epigenetic Mechanisms of Depression in Human Limbic Circuits (Carol Tamminga, UT Southwestern)
The timely translation of molecular hypotheses derived from animal models to the human syndrome of depression is one of the primary goals of this Center. We focus on two of the brain’s limbic regions, PFC and NAc: both are implicated in human depression, and the direct stimulation of each exerts antidepressant effects in clinical trials (see above). The most direct and feasible methodology with which to query these limbic circuits for epigenetic changes in depression is the analysis of postmortem brain. This is because virtually all of the chromatin regulatory proteins implicated in animal models cannot yet be investigated by brain imaging, given the lack of suitable radioligands. In contrast, analysis of these proteins and their encoding genes and mRNAs is straightforward in human postmortem brain - assuming high quality tissue, detailed clinical information, and customized, accurate brain dissections. Project 4 is based on the hypothesis that many of the molecular alterations observed in the PFC and NAc of rodent depression models, derived from Projects 1-3, will show similar abnormalities in these regions of depressed humans. Indeed, we now have substantial evidence to support this hypothesis as detailed in several recent publications. Such changes include altered levels of H3K9me2, G9a, L3MBTL1, H3.3, and several other epigenetic endpoints that form the basis of all three of our preclinical Projects (see Figure). Conversely, we expect that such analyses of human tissue will reveal novel modes of regulation in depressed humans and thereby generate new hypotheses to be tested functionally in Projects 1-3. The proposed studies include genome-wide ChIP-Seq analyses of both standard histone modifications, and the more advanced chromatin endpoints of interest to Projects 1-3, on postmortem tissue. We focus on several regions of PFC (e.g., sgACC) that are homologous with regions of PFC (e.g., medial or mPFC) in mice, as best as such homology can be established. In addition to analyzing whole extracts of PFC and NAc, we are examining isolated neuronal nuclei from PFC. Our Center has unique expertise in being able to accomplish this complex task. In this Project, we not only evaluate molecular hypotheses of depression, developed from the Center's uniquely broad animal models, in unipolar depression per se; we also test these hypotheses in symptomatic depression, across several diagnoses, including bipolar disorder and schizophrenia. This work will help us define the association of a chromatin abnormality with symptoms of depression or with a specific diagnostic category. Likewise, we are relating various chromatin abnormalities not with a diagnosis, but with exposure to early life stress, which is feasible because of our extensive clinical phenotyping of the brain samples in our collections. Careful attention is given to possible medication effects, both by examining brains of patients with and without a history of antidepressant treatment and by comparing differences observed in depressed humans to patterns of chromatin regulation seen in mice (stressed and unstressed) treated with antidepressants. Confidence that the results we obtain truly represent the in vivo state of these molecular candidates rests on the quality of the tissue used and the rigor with which we evaluate results for postmortem artifact. Together, we believe that these proposed studies will provide a uniquely powerful analysis of epigenetic and related pathologies in the human PFC and NAc associated with depression and antidepressant treatment.
FIGURE. Abnormal H3.3 expression in depression. Levels of H3.3A and H3.3B are reduced in the sgACC of depressed humans, an effect not seen in individuals treated with antidepressants (ADT). In contrast, H3.3B levels are increased in the NAc in individuals off ADT, but H3.3A and H3.3B are decreased dramatically in individuals on ADT. These data are based on a small number of subjects and will now be repeated. Data are mean ± sem (N=6/group). *p<0.05 by ANOVAs and t-tests.