Changes in prefrontal cortical regions

Pezawas et al provided the first report of an effect of the BDNF Val66Met polymorphism in the caudal middle frontal cortex, revealing reduced grey-matter volume in healthy individuals who were Met carriers compared with those homozygous for Val. ^{Reference Pezawas, Verchinski, Mattay, Callicott and Kolachana7} The potential contribution of cortical thinning to the volumetric reductions has recently been added, ^{Reference Yang, Liu, Sun, Wang, Zeng and Yuan20} and reduced middle frontal activity has also been observed in a healthy Met carrier group. ^{Reference Schofield, Williams, Paul, Gatt, Brown and Luty9} In our study we found a main effect of Met carrier status on cortical thinning in the caudal middle frontal cortex in both healthy individuals and in people with depression. Our study extends both the observation by Yang et al in healthy Chinese adults, ^{Reference Yang, Liu, Sun, Wang, Zeng and Yuan20} and the original finding, ^{Reference Pezawas, Verchinski, Mattay, Callicott and Kolachana7} by localising the contribution of cortical thinning to the reductions in caudal middle frontal volume. Moreover, this is the first report of the extent that Met carrier status leads to middle frontal cortical thinning, as the effect of the Met allele appears to supersede the disease process effects of MDD on cortical thickness. Grey-matter reductions in the caudal middle frontal region have been frequently observed in MDD, with further reductions in recurrent MDD. ^{Reference Bora, Fornito, Pantelis and Yucel28} Our findings indicate that the impact of the BDNF Val66Met polymorphism on cortical thinning in the caudal middle frontal region is even greater than that of recurrent MDD. It is possible that the influence of the Met allele in this region may not necessarily be causal in itself for MDD, but may be a predisposing, contributory factor in association with other neurogenic effects. Stress is linked to an increase in cortisol, which in turn causes a reduction in BDNF. ^{Reference Castren, Voikar and Rantamaki29} Individuals who are Met carriers are less able to compensate for this BDNF reduction owing to deficient transport of the BDNF preprotein which causes clumping around the nucleus, perhaps leading to neuronal atrophy in response to reduced BDNF levels. ^{Reference McAllister, Katz and Lo4,Reference Duman and Monteggia5}

We also observed a significant interaction effect in surface area in the caudal anterior cingulate and rostral middle frontal cortices. The anterior cingulate and middle frontal cortices are key regions in the neurocircuitry of mood disorders, and the anterior cingulate is a well-replicated predictive marker of clinical response in MDD, ^{Reference Fu, Steiner and Costafreda30} which is evident at the individual level. ^{Reference Costafreda, Chu, Ashburner and Fu31} Surface area and cortical thickness have independent genetic and developmental origins. ^{Reference Panizzon, Fennema-Notestine, Eyler, Jernigan, Prom-Wormley and Neale19} The radial unit hypothesis proposes that cortical thickness is determined by the number of cells within a neuronal column, and cortical surface area is determined by the number of neuronal columns. ^{Reference Rakic32} There is support for a general regional expansion of surface area from childhood into adolescence, particularly in boys, ^{Reference Koolschijn and Crone33} followed by subsequent decreases in adulthood with increasing age, ^{Reference Hogstrom, Westlye, Walhovd and Fjell34} whereas cortical thinning is a pronounced feature of adolescence that continues into adulthood. ^{Reference Koolschijn and Crone33} There is, however, some notable variability in the regional changes, ^{Reference Koolschijn and Crone33,Reference Hogstrom, Westlye, Walhovd and Fjell34} for example the surface area of the anterior cingulate cortex may show relatively fewer changes in adulthood. ^{Reference Koolschijn and Crone33} Modulation of grey-matter density by the BDNF Val66Met polymorphism in both the anterior cingulate and middle frontal cortices has been found in bipolar disorder, ^{Reference Matsuo, Walss-Bass, Nery, Nicoletti, Hatch and Frey35} and in the anterior cingulate in healthy individuals with a history of childhood abuse, ^{Reference Gerritsen, Tendolkar, Franke, Vasquez, Kooijman and Buitelaar12} whereas no prefrontal regional effect has been reported in schizophrenia. ^{Reference Ho, Milev, O'Leary, Librant, Andreasen and Wassink36} Our study localises the genetic influence to cortical surface area in its contribution to grey-matter volume, which had not been examined in previous studies in patient populations. ^{Reference Matsuo, Walss-Bass, Nery, Nicoletti, Hatch and Frey35,Reference Ho, Milev, O'Leary, Librant, Andreasen and Wassink36} It is possible that the surface area reduction we observed is an endophenotype for MDD, or more generally for mood disorders including bipolar disorder, or it may also be a feature in schizophrenia that has not been well captured by morphometric studies to date. The lack of an association of BDNF with schizophrenia, ^{Reference Kawashima, Ikeda, Kishi, Kitajima, Yamanouchi and Kinoshita37} however, suggests that the effect may be more strongly expressed in mood disorders.

Effects on amygdala volume

Contrary to our hypothesis, no significant interaction effect was found in the grey-matter volume of the amygdala. Our observations are consistent with findings in healthy individuals carrying the BDNF Met allele with a history of childhood adversity and in depression, ^{Reference Frodl, Schule, Schmitt, Born, Baghai and Zill10,Reference Gerritsen, Tendolkar, Franke, Vasquez, Kooijman and Buitelaar12} but there have also been reports of reduced amygdala volume in healthy people with and without a history of stressful events. ^{Reference Montag, Weber, Fliessbach, Elger and Reuter8,Reference Costafreda, Chu, Ashburner and Fu31} Similarly, studies of amygdala responsivity have been mixed, with reports of significant effects, ^{Reference Montag, Reuter, Newport, Elger and Weber38,Reference Gasic, Smoller, Perlis, Sun, Lee and Kim39} but more frequently of no impact, ^{Reference Schofield, Williams, Paul, Gatt, Brown and Luty9,Reference Hariri, Goldberg, Mattay, Kolachana, Callicott and Egan40,Reference Hashimoto, Moriguchi, Yamashita, Mori, Nemoto and Okada41} of the BDNF polymorphism. The literature on amygdala volume in depression, however, is inconsistent, with some suggestion that amygdala volume is reduced in more chronic forms of depression. ^{Reference Drevets, Price and Furey42} In our study, participants had a history of recurrent depression characterised by discrete acute depressive episodes with periods of euthymia rather than a more chronic, treatment-resistant type of depression. Our findings are most comparable with those of Frodl et al, who similarly did not observe any effect of diagnosis or BDNF genotype on amygdala volume. ^{Reference Frodl, Schule, Schmitt, Born, Baghai and Zill10}

Study limitations

We limited our analysis to a priori defined regions in the prefrontal cortex and the amygdala within the neurocircuitry of depression. However, our sample size was relatively modest and replication in an independent sample is required. Cortical thinning in the medial orbitofrontal region has been reported in first-episode depression, ^{Reference Grieve, Korgaonkar, Koslow, Gordon and Williams43} and in a younger cohort of patients with MDD (mean age 34 years) than in our study. ^{Reference Li, Zhang, Wang, Chen, Fan and Guan44} Although there was evidence of left medial orbitofrontal cortical thinning in MDD patients relative to healthy controls in our sample, this difference did not survive correction for multiple comparisons. Furthermore, BDNF Val66Met effects have also been observed in the temporal and parietal cortices in healthy Chinese adults. ^{Reference Yang, Liu, Sun, Wang, Zeng and Yuan20} There may also be a modulatory effect of early life stress, as healthy individuals carrying the Met allele with a history of greater stressful events have shown reduced grey-matter volume in the amygdala and hippocampus, ^{Reference Gatt, Nemeroff, Dobson-Stone, Paul, Bryant and Schofield45} and also in the anterior cingulate cortex, ^{Reference Gerritsen, Tendolkar, Franke, Vasquez, Kooijman and Buitelaar12} compared with a group homozygous for Val. Another limitation of our study is the absence of data on the history of possible childhood trauma in our participants, although it is unclear whether these effects would persist alongside the pathophysiological effects of the illness. Moreover, most of the participants with MDD were taking antidepressant medication, which may have had an effect on BDNF levels, ^{Reference Duman and Monteggia5} and in turn potentially on neural volumes. ^{Reference Schmidt and Duman46} The anterior cingulate region has been consistently identified as a predictive marker of clinical response, ^{Reference Fu, Steiner and Costafreda30} and the BDNF polymorphism has shown an association with treatment response. ^{Reference Licinio, Dong and Wong47} Furthermore, an interaction effect of BDNF and its high-affinity receptor, neurotrophic tyrosine kinase receptor 2 (NTRK2), gene polymorphisms has been associated with the development of treatment-resistant depression. ^{Reference Van Eijndhoven, van Wingen, Katzenbauer, Groen, Tepest and Fernandez48} How these potentially complementary markers may interact at a neural level in predicting clinical response requires further investigation. Neural correlates of BDNF associations with clinical response in a sample of patients who were medication-free and perhaps in their first episode of depression would elucidate the effects of medication, recurrent episodes of depression and prediction of clinical response.

Concluding remarks

In summary, we demonstrated sustained effects of the BDNF Val66Met polymorphism on distinct subregions in the prefrontal cortex in depression. The effects in the caudal middle frontal regions exceeded those of the illness as Met carrier status was associated with greater cortical thinning in both MDD and control groups. Effects in the anterior cingulate and rostral middle frontal regions revealed an interaction with BDNF Val66Met genotype, in which the MDD Met carrier group showed the greatest reduction in surface area. Our findings specify the anterior cingulate and middle frontal regions as key regions within the neurotrophin hypothesis of depression.