Neuroimaging and neuropathological studies of bipolar disorder (BD) have provided substantial evidence that brain structures that normally modulate emotional, autonomic and neuroendocrine responses to stress and emotionally provocative stimuli contain abnormal reductions in grey matter volume, synaptic markers and glial cells. The significance of these abnormalities with respect to the pathogenesis and clinical manifestations of BD remain unclear, but we obtained preliminary evidence that BD subjects with the most prominent reductions in grey matter (GM) are more likely to develop psychosis and to follow a chronic illness course without complete recovery between episodes. If confirmed in larger subject samples these preliminary observations would imply that treatments which could prevent or reverse the GM volume changes may prove particularly beneficial in improving the prognosis (long term outcome) in BD. These discoveries have occurred in parallel with the additional discoveries that the mood stabilizing drugs, lithium and divalproex, exert robust neuroprotective and neurotrophic effects in experimental animals. For example, in rodents lithium can reverse the persistent neuronal atrophy induced in the hippocampus and medial prefrontal cortex (PFC) by 2 months of repeated stress; although a variety of other psychotropic drug classes are incapable of reversing this atrophy. Together with Dr. Husseini Manji, we are investigating whether lithium and divalproex can reverse toward normal the abnormal reductions in the volumes of hippocampal and medial PFC structures in humans suffering from BD. In a longitudinal study of lithium treatment in BD subjects the grey matter volume increased (toward normal) in the left ventral anterior cingulate cortex, a medial PFC structure where the GM volume was previously shown to be abnormally reduced in bipolar disorder. This increase in volume was specific to the subjects who showed clinical improvement during lithium treatment. Studies performed in experimental animals or humans have suggested that this brain structure plays a major role in inhibiting or regulating the autonomic and endocrine responses to threat or stress, and the emotional feelings that occur in response to internally generated sad or anxious thoughts and memories. The abnormal reduction in volume may indicate the function of this region is impaired in BD, dysregulating emotional expression and experience. Lithium has been shown to increase the genetic expression of proteins that exert neuroprotective and neurotrophic effects, so correcting this brain abnormality may play an integral role in exerting mood stabilizing effects. Effects of the psychotropic medications pramipexole and divalproex are currently being pursued using MRI measures of regional GM volume. In the previous year we obtained MRI images at higher spatial resolution than has been possible in previous studies by using a 3 Tesla MRI scanner, an MRI pulse sequence optimized for tissue contrast resolution and a 40 minute long scan. These images provide exquisite anatomical detail to increase the sensitivity of the volumetric measures. We obtained these scans in 28 bipolar subjects, of whom one-half have been recently treated with lithium or divalproex, and other one-half have not. In addition, 56 healthy controls have been scanned. During the past year such images also were obtained from 44 patients with unipolar depression and an additional 16 healthy controls. We compared matched groups of patients versus controls on the basis of prior treatment with mood stabilizers or antidepressant drugs. We were unable to establish that in the infralimbic portion of the medial PFC mood stabilizer treatment was associated with larger GM volumes than existed in both healthy controls and unmedicated BD subjects. However, in unipolar depressives, we discovered that the infralimbic portion of the anterior cingulate cortex volume is larger in patients who have been chronically or intermittently ill as compared with cases who can remain in remission for long time periods. We also demonstrated that this region was particularly small in psychotic depressives, and that the magnitude of this effect predicted outcome over the next two years of clinical follow-up with respect to the more severe abnormality, the worse the prognosis. Also in the past year we obtained lower resolution 3T MRI Images from 36 BD subjects and 65 healthy controls in an approach called voxel-based morphometry which could survey larger regions for GM changes in cortical areas that were anatomically related to the medial PFC and hippocampus, namely the orbitofrontal cortex, superior temporal cortex and posterior cingulate cortex. To account for possible neurotrophic and neuroprotective effects of psychotropic medications, BD subjects were divided into recently medicated and unmedicated groups. Images were segmented into tissue compartments, which were examined on a voxel-wise basis to determine the location and extent of morphometric changes. The GM was reduced in the posterior cingulate/retrosplenial cortex and superior temporal gyrus of unmedicated BD subjects relative to medicated BD subjects, and in the lateral orbital cortex of medicated BD subjects relative to controls. The morphometric abnormalities in the posterior cingulate, superior temporal, and lateral orbital cortices in BD support the hypothesis that the extended network of neuroanatomical structures subserving visceromotor regulation contains structural alterations in BD. Additionally, localization of morphometric abnormalities to areas known to exhibit increased metabolism in depression supports the hypothesis that repeated stress and elevated glucocorticoid secretion may result in neuroplastic changes in BD. Finally, we expanded the MRI imaging studies in children and adolescents who had mothers with BD or MDD to determine whether persons at high genetic risk for developing a mood disorder also have a reduction in these regions. In the genual portion of the corpus callossum, which carries fibers from the ventral anterior cingulate, orbitofrontal cortex, and medial PFC areas where we identified structural abnormalities in unipolar and/or bipolar depression, was also shown to be reduced in volume both in females with unipolar depression and in females with BD. In addition, this abnormality was evident in their children and adolescent offspring who had not yet suffered from mood disorders themselves. Male and female offspring of depressed mothers who became depressed after the MRI, but had not been depressed prior to MRI, had smaller genu areas than high-risk offspring who never became depressed either preceding or following MRI (p < .05). The genu and whole corpus callosum areas were negatively related to family stress in female offspring (p < .03), but not in the male offspring or the mothers. These data collectively demonstrated that these abnormalities may be associated with abnormalities of brain development that precede illness and may be associated with the vulnerability to developing depression. In the coming year we plan to initiate a longitudinal study examining the long term relationships between clinical and brain structural effects during years of maintenance mood stabilizing agents. These results were reported at multiple scientific meetings and 3 manuscripts describing these results have been submitted for publication, while three additional manuscripts remain in preparation.
