- Top of page
- Materials and methods
Neuroimaging studies have linked the methionine (Met) allele of the brain-derived neurotrophic factor (BDNF) gene to abnormal regional brain volumes in several psychiatric and neurodegenerative diseases. However, no neuroimaging studies assessed the effects of this allele on brain morphology in alcohol use disorders and its demonstrated change during abstinence from alcohol. Here we assessed the effects of the BDNF Val66Met (rs6265) polymorphism on regional brain tissue volumes and their recovery during short-term abstinence in treatment-seeking alcohol-dependent individuals. 3D T1 weighted magnetic resonance images from 62 individuals were acquired at 1.5 T at one week of abstinence from alcohol; 41 of the participants were rescanned at 5 weeks of abstinence. The images were segmented into gray matter (GM), white matter (WM) and cerebrospinal fluid and parcellated into regional volumes. The BDNF genotype was determined from blood samples using the TaqMan technique. Alcohol-dependent Val (Valine)/Met heterozygotes and Val homozygotes had similar regional brain volumes at either time point. However, Val homozygotes had significant GM volume increases, while Val/Met heterozygotes increased predominantly in WM volumes over the scan interval. Longitudinal increases in GM but not WM volumes were related to improvements in neurocognitive measures during abstinence. The findings suggest that functionally significant brain tissue volume recovery during abstinence from alcohol is influenced by BDNF genotype.
Brain-derived neurotrophic factor (BDNF) is a neurotrophin that is widely expressed in the mammalian brain with particularly high levels in the hippocampus and the cerebral cortex (Hofer et al. 1990; Phillips et al. 1990; Wetmore et al. 1990). BDNF plays an important role in the long-term survival, differentiation, growth and maintenance of neurons (Ernfors et al. 1995; Murer et al. 2001). It also promotes myelination in both the peripheral and central nervous systems (Cellerino et al. 1997; Vondran et al. 2010; Xiao et al. 2010). The methionine (Met) allele of the BDNF Val66Met (rs6265) polymorphism is associated with impaired intracellular trafficking and activity-dependent secretion of BDNF (Chen et al. 2004; Egan et al. 2003). The Met allele is also associated with poorer verbal memory (Egan et al. 2003; Hariri et al. 2003), processing speed (Miyajima et al. 2008; Raz et al. 2009) and general intelligence (Tsai et al. 2004) in controls and individuals with various neuropsychiatric conditions.
In voxel-based morphometric magnetic resonance imaging (MRI) studies of healthy controls, BDNF Met allele carriers showed smaller hippocampi, prefrontal, parietal and occipital lobes, as well as smaller frontal and temporal gyri compared with valine (Val) homozygotes (Eker et al. 2005; Pezawas et al. 2004). The Met carriers compared with Val homozygotes also had higher rates of gray matter (GM) volume reductions in the dorsolateral prefrontal cortex and the dorsal premotor cortex as a function of normal aging (Nemoto et al. 2006); BDNF Met carriers also showed smaller volumes of specific brain regions in schizophrenia (Takahashi et al. 2008), major depression (Frodl et al. 2007), Alzheimer's disease (Hashimoto et al. 2009) and mild cognitive impairment (Forlenza et al. 2010); and Met carriers with schizophrenia had larger time-related reductions of frontal GM volumes than Val homozygotes (Ho et al. 2007). Furthermore, hypo-myelination of developing axons of the spinal cord, the optic nerve and the corpus callosum was reported in BDNF knock-out/deficient mice (Cellerino et al. 1997; Vondran et al. 2010; Xiao et al. 2010). Together, these findings indicate that the BDNF Met allele is related to abnormalities in brain morphology and function in normal and in various neuropsychiatric conditions. Similar to psychiatric and neurodegenerative diseases, alcohol use disorders [i.e. abuse or dependence according to the diagnostic and statistical manual of mental disorders, fourth edition (DSM-IV) criteria] are also associated with abnormalities in brain structure and function. Neuroimaging studies of those with an alcohol use disorder have consistently showed GM and/or white matter (WM) reductions (particularly in the frontal and temporal regions) and enlarged ventricles and sulci (Bühler and Mann 2011), which at least partially normalize within the first month of sustained abstinence from alcohol (Gazdzinski et al. 2005; Pfefferbaum et al. 1995). Although the Val66Met BDNF polymorphism has been associated with brain morphological abnormalities in individuals with neurodegenerative or psychiatric disorders, no MRI studies have assessed the potential effect of this polymorphism on brain morphology in those with an alcohol use disorder. Therefore, the primary goal of this study was to assess the influence of the BDNF Val66Met polymorphism on MRI-derived regional brain tissue volumes and their longitudinal changes in alcohol-dependent individuals (ALC) during the first few weeks of abstinence from alcohol. We also assessed the functional relevance of any longitudinal brain volume changes. Specifically, we hypothesized that:
- At 1 week and 5 weeks of abstinence, BDNF Val homozygotes have significantly larger GM and WM volumes in the frontal, parietal and temporal lobes than Val/Met heterozygotes.
- Val homozygotes during abstinence from alcohol recover faster from lobar GM and WM volume reductions than Val/Met heterozygotes, particularly in the frontal lobes.
- Longitudinal increases of tissue volumes relate positively to improvements on measures of working memory, visuospatial scanning speed and incidental learning, auditory-verbal and visuospatial learning.
- Top of page
- Materials and methods
In this quantitative MRI study assessing the effects of the BDNF Val66Met (rs6265) polymorphism on brain tissue volume changes in abstinent ALC, Val homozygotes and Val/Met heterozygotes differed on the degree of regional GM and WM volume recovery over 1 month of abstinence from alcohol. Total cortical GM volume increased significantly in Val homozygotes, whereas total lobar WM volume increased significantly in Val/Met heterozygotes. Specifically, Val homozygotes showed significant increases in GM volumes of all but the occipital lobes, while showing no longitudinal changes in any lobar WM volume. Val/Met heterozygotes showed significant increase of frontal WM and trends for increased WM in the parietal and temporal lobes. For the subcortical volumes, thalamic GM increased only in Val homozygotes, whereas total volumes of the cerebellum and the brainstem increased only in Val/Met heterozygotes. Temporal GM volume increased significantly in both genotypes, the lenticular nuclei showed no significant change in either genotype, and the caudate showed a significant decrease in Val homozygotes.
Overall, our results indicate two disparate patterns of brain tissue volume recovery for the BDNF Val homozygotes and Val/Met heterozygotes in this alcohol-dependent cohort: Val/Val was associated with significant volume increases restricted to cortical GM, while Val/Met was primarily related to volume increases in lobar WM. Despite the genotype-specific longitudinal changes, there were no significant cross-sectional regional volume differences between Val homozygotes and Val/Met heterozygotes at baseline or follow-up. However, the alcohol-dependent Val homozygotes had significantly larger ICV compared with their Val/Met heterozygous counterparts. Cole et al. (2011) reported smaller ICV for Met carriers compared with Val homozygotes with major depression, but this difference was not apparent in healthy controls (Bueller et al. 2006; Cole et al. 2011), despite high heritability of ICV (Glahn et al. 2007). The lack of BDNF genotyping for our control participants in this study and the small control sample did not allow assessing the effects of BDNF Val66Met polymorphism on ICV in controls. Our light-drinking controls showed no regional GM or WM volume changes over 7 months, and none of our alcohol consumption measures were associated with brain tissue volume change over the 4-week-interval in ALC; together, this data suggest that the observed ALC volume changes were indeed related to structural changes over approximately 1 month of abstinence, and several of these regional changes were mediated by BDNF genotype.
Given that BDNF is associated with neuronal survival, neuronal growth and synaptic plasticity in the adult brain (Lu 2003), our observation of significant longitudinal recovery of cortical and subcortical GM in only Val homozygous ALC over 1 month of abstinence is congruent with the action of BDNF binding with tyrosine kinase (Trk) B receptors in the cortex. Both BDNF and TrkB show high expression in the cortex, and BDNF-mediated activation of TrkB is associated with dendritic growth, axonal outgrowth and synaptic plasticity (Bath and Lee 2006). As the Met allele is associated with suppressed secretion of BDNF and its extracellular trafficking (Egan et al. 2003; Chen et al. 2004), the reportedly higher concentrations of BDNF in blood of Val homozygotes relative to Met allele carriers (Egan et al. 2003; Duncan et al. 2009) appears consistent with our observation of greater cortical GM recovery in Val homozygotes. In mammals, optimal neuronal growth and survival depends on the concentration of growth and neurotrophic factors, such as BDNF (Davies 2000; Hetman and Gozdz 2004).
On the other hand, the significant lobar WM volume increases observed in Val/Met heterozygotes seem counterintuitive, because BDNF is associated with promotion of central nervous system (CNS) myelination. Specifically, BDNF knock-out mice (a single Val allele deleted) exhibited hypo-myelination in the optic nerve, hippocampus, cerebellum and cortex after 21 days of delivery (Cellerino et al. 1997), suggesting the necessity of BDNF for normal CNS myelination. This observation was confirmed by two other studies in the developing brain of mice models (Vondran et al. 2010; Xiao et al. 2010). Given that BDNF has been described as a promyelination neurotrophin in the developing brain (Cellerino et al. 1997; Vondran et al. 2010; Xiao et al. 2010), we predicted our Val homozygous ALC would show increases in regional WM volumes; instead, the Val/Met heterozygotes showed greater WM volume increases during short-term abstinence than the Val homozygotes. These discrepancies may be related to differences in BDNF activity during brain development and brain repair of the adult brain after removal of a chronic insult (here alcohol).
In the combined ALC group (i.e. Val homozygotes + Val/Met heterozygotes), longitudinal changes in regional cortical and subcortical GM volumes during abstinence were positively associated with improvement in neurocognition. Most notably, increases in all cortical and subcortical GM regions were related to improving visuospatial learning. Additionally, changes in measures of working memory related positively to frontal and parietal GM volume increases; and improvement in measures of processing speed related positively to parietal GM volume increases. On the other hand, longitudinal increases in WM volumes, which were observed mainly in the Val/Met heterozygotes, were not associated with improvement in any neurocognitive measure. This may suggest that the WM increases observed primarily in Val/Met heterozygotes were of little functional significance in this cohort. However, the complexity of the cortical–cortical and cortical–subcortical WM interconnectivity is not well captured by our lobar WM volumetric measures (Pfefferbaum and Sullivan 2005); volumetrics cannot determine if remyelination, a likely contributor to the WM volume increases (Harper 2009; Sullivan and Pfefferbaum 2005), is the primary contributing factor to the WM increases demonstrated here by Val/Met heterozygotes. More regionally specific measures of WM microstructural integrity provided by diffusion weighted imaging measures (i.e. fractional anisotropy, radial and longitudinal diffusivity) may serve as better proxies of neurocognitive function (Chiang et al. 2011; Pfefferbaum et al. 2000; Pfefferbaum et al. 2005; Pfefferbaum and Sullivan 2005) and may assist in determining if the WM volume increases are indeed related to myelination.
We expected that the observed differential recovery rates of regional GM and WM volumes between the two genotypes (which were statistically equivalent at TP1) would result in cross-sectional group differences at 1 month of abstinence. However, we observed statistically similar regional brain volumes between genotypic groups at the 1-month follow-up (i.e. TP2). The 1-month scan interval may not have been of sufficient duration to allow the observed differences in recovery rates between the groups to result in significant volume differences at TP2. If Val homozygotes and Val/Met heterozygotes continue to show different rates of change between regional GM and WM tissue volumes, significant cross-sectional differences may become apparent with extended abstinence.
This study has limitations which may affect the generalizability of our findings. The sample, although of adequate size for neuroimaging studies, was relatively small for genetic research. However, our analysis focused on only one specific common polymorphism with a demonstrated influence on human brain morphology. The majority of the participants were males, which precluded assessment of potential sex effects on regional tissue volumes. Also, we did not have a control sample with BDNF genotypes for comparisons of genotype-specific regional brain volumes. Furthermore, we did not evaluate our alcohol-dependent participants for DSM-IV Axis II disorders, such as antisocial personality disorders (Grant et al. 2004; Pridmore et al. 2005), and we did not measure potential group differences in nutrition, exercise and other genetic predispositions.
In summary, the BDNF Val66Met (rs6265) polymorphism was significantly related to the recovery of regional GM and WM tissue volumes within the first 5 weeks of sobriety, suggesting genetic influences on brain tissue changes during abstinence from alcohol in this alcohol-dependent cohort. Significant associations of cortical and sub-cortical GM volume increases (observed predominantly in Val homozygotes) with improved performance on multiple neurocognitive abilities during short-term abstinence highlight the functional relevance of genotype-specific brain tissue recovery during abstinence from alcohol. To better understand the mechanisms contributing to long-term morphological changes in alcohol use disorders, it is necessary to determine the influence of the BDNF Val66Met polymorphism, (and other polymorphisms associated brain morphology (e.g. superoxide dismutase; Srivastava et al. 2010) on changes in regional brain volumes and other morphometrics (e.g. cortical thickness) with extended abstinence from alcohol. Although the mechanisms by which BDNF genotype and protein mediate brain tissue recovery in abstinent ALC are unknown, these neuroimaging findings highlight the need to consider genetic factors such as BDNF in the treatment of alcohol use disorders, as gene variants may influence the pattern of brain tissue recovery and associated neurocognitive improvements.