Effect of Functional BDNF and COMT Polymorphisms on Symptoms and Regional Brain Volume in Frontotemporal Dementia and Corticobasal Syndrome

A range of phenotypes, from behavior and language symptoms to motor neuron disease, can be observed in patients with frontotemporal lobar degeneration (FTLD), even across patients with the same pathogenic mutation (1). There is accumulating evidence that functional polymorphisms, including in transmembrane protein 106b (TMEM106B) (2), catechol-O-methyltransferase (COMT) (3), and the dopamine receptor D4 (DRD4) (4), affect phenotype in FTLD-spectrum disorders. In this study, we examined the effects of common functional polymorphisms in two genes—brain-derived neurotrophic factor (BDNF) and COMT—on clinical and MRI findings in patients with frontotemporal dementia (FTD) and corticobasal syndrome.

BDNF is a member of the neurotrophin family of neuron growth factors (5). BDNF contains a functional polymorphism (rs6265) that results in a valine versus methionine substitution at codon 66, termed the Val66Met polymorphism. The Met allele is associated with decreased BDNF activity (6). BDNF is involved in hippocampus-mediated learning and long-term potentiation (7–10). BDNF appears to be involved in the pathogenesis of Huntington’s disease (11), major depressive disorder (12, 13), and the recovery of executive function after brain injury (14). BDNF genotype appears to affect the volume of several brain regions in healthy subjects, including the hippocampus, lateral temporal lobe, prefrontal cortex, and cingulate cortex (15, 16).

COMT regulates dopamine in the prefrontal cortex by converting released dopamine to 3-methoxytyramine (17–19). COMT has a common functional methionine (Met) for valine (Val) substitution at codon 158, referred to as the Val158Met polymorphism (rs4680). The enzyme in individuals with the Met/Met genotype has three to four times lower activity than the enzyme in individuals with the Val/Val genotype (20). Thus, individuals with the Met allele have greater amounts of available dopamine. Individuals with the Met allele demonstrate superior performance on tests of frontal cognitive function in an allele dose-dependent fashion (21–29). The COMT genotype appears to affect the symptom presentation of Alzheimer’s disease (30), schizophrenia (26, 31, 32), attention deficit hyperactivity disorder (33), and brain injury (34). The COMT genotype has also been shown to affect gray matter volume in dopamine-innervated brain regions in patients with dementia, including FTLD-spectrum illnesses (3). Studies have shown that functional polymorphisms in the DRD₄ gene that reduce dopamine (DA) tone are associated with greater degeneration of brain regions with high DA receptor density and increased apathy in patients with FTD (4), suggesting that changes in DA may affect both neurodegeneration and symptoms in FTLD syndromes.

In the present study, we assessed the effects of the BDNF Val66Met and COMT Val158Met polymorphisms on cognitive, neuropsychiatric, and motor symptoms and regional brain volumes on MRI in 174 patients with FTLD-spectrum illness. Based on previous findings for healthy control subjects and patients with dementia, we hypothesized that the BDNF Met allele would be associated with memory dysfunction and that the COMT Val allele would be associated with executive and motor dysfunction.

Methods

Participants

A total of 174 patients were genotyped as follows: behavioral variant FTD (bvFTD), N=82; nonfluent variant primary progressive aphasia, N=19; semantic variant primary progressive aphasia (PPA), N=4; FTD-motor neuron disease, N=5; and corticobasal syndrome, N=64. Participants were assessed as part of an ongoing research study on FTLD-spectrum illnesses in the Cognitive Neuroscience Section of the National Institute of Neurological Disorders and Stroke (NINDS) of the National Institutes of Health (NIH). They were either self-referred or referred by outside clinicians. Patients arrived at NIH with a caregiver and were diagnosed by standard clinical criteria in an evaluation by a neurologist (E.M.W.), psychiatrist (E.D.H), and neuropsychologist (J.G.) (35–37). Patients also met updated criteria for probable bvFTD (38). The updated terminology is used in this study. Patients, accompanied by their caregivers, underwent approximately 1 week of extensive neuropsychological and imaging studies. Approximately 6% of this patient cohort had an autosomal dominant mutation known to result in FTD, including mutations in progranulin, microtubule-associated protein tau, and C9orf72 (39). The remainder of patients had sporadic illness. While most of the patients have not yet come to autopsy, those patients with a clinical diagnosis of corticobasal syndrome who later died and underwent autopsy (N=15) demonstrated roughly equal pathological diagnoses for Alzheimer’s disease (N=7) and CBD (N=8).

We required all participants to have an assigned durable power of attorney before enrollment in the study, and the assigned individuals provided written, informed surrogate consent for the patients to enter the study, and all patients provided assent. All aspects of the study and the consent procedure were approved by the NINDS Institutional Review Board. Demographic and clinical characteristics of the study participants are presented in Table 1.

Genotyping

Peripheral blood samples were collected from patients in 5-ml EDTA tubes. Genomic DNA was extracted from peripheral blood as recommended by the manufacturer (Wizard Genomic DNA Purification Kit, Promega, Madison, Wis.). After quality assessment and quantification using NanoDrop ND-1000 (Thermo Scientific, Waltham, Mass.), 10 ng of DNA samples were genotyped in a 5-µl TaqMan genotyping assay as recommended by the manufacturer (Applied Biosystems, Waltham, Mass.). Primers and probes for the single-nucleotide polymorphism (SNP) rs4680 in the COMT gene and the SNP rs6265 in the BDNF gene were obtained from Applied Biosystems. The assays were run on an ABI PRISM 7900 Sequence Detection System instrument (Applied Biosystems). The genotyping results were analyzed using SDS software version 2.3 (Applied Biosystems). For quality control, 10% of assays were repeated for the genotyping assay, and the results were 100% concordant.

Cognitive, Behavioral, and Motor Analyses

We assessed a broad range of cognition, including all of the neuropsychological tests previously examined by Huey et al. (40): the Delis-Kaplan Executive Function System (D-KEFS) (41), the Mattis Dementia Rating Scale (MDRS-2) (42), and the Wechsler Memory Scale (WMS-III) (43). We assessed neuropsychiatric symptoms with the Neuropsychiatric Inventory (NPI) (44). The same subtests were analyzed in the present study as in Huey et al. (40). In pilot testing, many patients experienced difficulty understanding some of the more challenging D-KEFS tests. Therefore, of the nine D-KEFS tests, only five were administered and analyzed: Trail-Making Test (including Parts A and B), Verbal Fluency, Sorting, Twenty Questions, and the Tower test. All primary scaled measures of the D-KEFS tests (excluding measures that are derived from primary measures) were used in the factor analysis. The sample size and ratio of participants to variables satisfies guidelines for principle components analysis (45). To assess motor function, we analyzed the mean number of taps on the Finger Tapping Test (46) and number of pegs placed on the Grooved Pegboard Test for dominant and nondominant hands and the Composite Standard Score on the Test of Oral and Limb Apraxia (TOLA) (47).

Statistical Analyses

To explore the relationship between allele dosage and symptoms, we used an analysis of variance (ANOVA) to compare the scores on our cognitive and behavioral symptoms between patients with and without a BDNF Met or COMT Val allele dosage. For primary analyses to test our hypotheses that COMT Val alleles would be associated with worsened executive and motor function and BDNF Met alleles would be associated with worsened memory, composite scores for memory, executive and motor function were compared, followed by subsequent test of the subtests that comprised the composite score. If the ANOVA for the composite measure was not significant when corrected for multiple comparisons of the composite measures, we did not continue to subtest analyses. The composite scores were calculated as follows: the mean of the D-KEFS factor scores (executive), the mean of the WMS-III standardized scores (memory), and the mean of z-scores of the Finger Tapping, Grooved Pegboard, and TOLA scores (Table 1). To obtain summary scores for the D-KEFS, we performed the same factor analysis (a principal components analysis using Varimax with Kaiser normalization) on the D-KEFS subscores, as specified in Gorno-Tempini et al. (37). To explore components of the composite scores, we performed subsequent correlations between the number of COMT Val and BDNF Met alleles (termed the allele dosage: 0, 1, or 2 per participant) and measures of cognitive, behavioral, and motor symptoms.

Imaging

A 1.5-T GE MRI scanner (GE Medical Systems, Milwaukee, Wis.) and standard quadrature head coil were used to obtain all images. A T1-weighted spoiled gradient echo sequence was used to generate 124 contiguous 1.5-mm-thick axial slices (repetition time=6.1 msec; echo time=min full; flip angle=20°; field of view=240 mm; 124 slices, slice thickness=1.5 mm; matrix size=256×256×124). Voxel-based morphometry analysis of the data was performed with SPM8 and followed the principles outlined by Ridgway et al. (48). Except as noted below, all default SPM options were used. Images were segmented into gray matter, white matter, and CSF. Spatial normalization, segmentation, and modulation were processed using a unified segmentation algorithm (49). This algorithm simultaneously calculated image registration, tissue classification, and bias correction using our participants’ structural MR images combined with the tissue probability maps provided in SPM. The segmented and modulated normalized gray matter images were smoothed with an 8-mm full width at half-maximum Gaussian kernel. An explicit mask encompassing the entire brain was used in the analyses to control for background signal outside the brain. This mask was downloaded from the SPM Anatomical Automatic Labeling toolbox. An explicit absolute threshold of p<0.05 for masking was used in the SPM second-level model interface (50). Total intracranial volume was calculated in SPM8 from the unsmoothed, modulated gray matter, white matter, and CSF images from each patient and used as a nuisance variable to account for the possible effect of varying brain volumes.

Two sets of imaging analyses were performed. First, two whole-brain ANOVAs were performed in SPM comparing two to zero copies of the COMT Val and BDNF Met alleles on gray matter volumes. Clusters surviving an uncorrected p-value threshold <0.001 and a cluster size of 30 voxels were considered significant (3). Second, the MarsBar toolbox was used to create regions of interest from clusters found to be significant on the whole brain analysis, as in Gennatas et al. (3). Mean gray matter intensity values were extracted from these regions of interest and used for correlations with our cognitive and behavioral measures.

Results

Demographic Characteristics

We examined a total of 174 participants with FTLD-spectrum illnesses. Table 1 summarizes the results of our analysis of the relationship between COMT Val and BDNF Met allele dosage and composite cognitive, behavioral, and motor measures in patients with FTLD as well as demographic data for the cohort. There were no statistically significant differences in the demographic characteristics of the FTD and PPA versus corticobasal syndrome groups. Therefore, in analyzing the effect of BDNF and COMT polymorphisms, these groups were combined. There was no effect of BDNF or COMT polymorphism on age at onset or other demographic variables (Table 1).

Cognitive, Behavioral, and Motor Analyses

In the principal components analysis, the measures of the correlation between the variables were good, indicating that a principle components analysis was appropriate (the Kaiser-Meyer-Olkin measure was 0.83; Barlett’s Test of Sphericity: χ²=1337.71, df=190, p<0.001). Five components had an eigenvalue >1.0, accounting for 75.5% of the total value. Each of the five components corresponded to a single D-KEFS subtest (i.e., each of the subtests was significantly loaded; absolute value >0.70 on the rotated component matrix). There was a significant effect of the COMT Val allele on our composite executive function measure (F=6.14, df=1, 76, p=0.015) but no significant effect of the COMT Val allele on the memory or motor composite measures. Correlations between the D-KEFS subtests and allele dosage revealed that the significant effect of COMT on executive function was due to its effect on the Sorting test (r=0.22, two-tailed p=0.049, corresponding to a Cohen’s d of 0.45 and r² of 0.05). None of the other correlations between the D-KEFS subtest factor scores and COMT or BDNF allele dosage were significant. There was a significant effect, surviving correction for multiple comparisons, of the BDNF Met allele only on depressive symptoms as measured by the NPI (F=9.50, df=1, 136, p=0.002) (Table 1). The BDNF Met allele did not have a significant effect on the composite motor measure.

Imaging

Two versus zero BDNF alleles did not have a significant effect on gray matter volume in the whole brain analysis. A COMT Val dosage of 2 was significantly associated with reduced gray matter volume in the bilateral caudate nuclei compared with a dose of 0 (Figure 1). Within the areas found to have significantly reduced gray matter volume, gray matter density was negatively correlated with verbal fluency and sorting subtests of the D-KEFS, with the conceptualization subscale of the MDRS-2, and visual memory on the WMS-III. Gray matter density in the right and left caudate was also negatively correlated with the total NPI and several of the NPI subscales, including agitation, euphoria, apathy, disinhibition, and motor behavior (Table 2).

Discussion

In patients with FTD and corticobasal syndrome, a greater number of Met alleles at position 66 of BDNF was associated with more depressive symptoms, and more Val alleles at position 158 of COMT was associated with poorer performance on the D-KEFS Sorting task, a test of executive function. We did not find an effect of either polymorphism on motor function. Our findings are consistent with previous studies that have shown associations between the COMT Val158Met and DRD₄ polymorphisms and executive function (15, 29) and the BDNF Val66Met polymorphism and depression (12, 13). However, this is the first study, to our knowledge, to implicate BDNF polymorphisms in the clinical presentation of FTLD illnesses. Unexpectedly, the effect is that a greater number of BDNF Met alleles are associated with decreased depression. This is the opposite direction from what has been observed in most previous studies in neurologically intact participants (51, 52), although it is the same direction we observed in our previous study on recovery of executive function after TBI (14). As can be seen in Table 2, the anatomic association between depression and caudate gray matter density is also in the opposite direction of the associations with most of the other neuropsychiatric symptoms (i.e., greater gray matter density is associated with greater depression). However, these results agree with previous findings that suggest that damage to certain brain structures and regions can be protective against internalizing psychiatric symptoms, including depression and anxiety (53–56). Our subsequent analyses showing an association between our cognitive and behavioral measures and the gray matter density in the caudate nuclei further support the relationship between COMT genotype and phenotypic presentation in FTLD (Table 2).

The effect size of BDNF allele dosage on depression, as well as the COMT allele dosage on the Sorting test in patients with FTLD, is medium (Cohen’s d=0.45), accounting for approximately 5% of the variance. This is similar to the amount of variance accounted for by the BDNF Val66Met polymorphism on executive function in the recovery from acute brain injury (6%) (14). However, it is considerably larger than the effects of both of these polymorphisms in healthy subjects (3, 12, 57). This may suggest that these polymorphisms play a larger role in the setting of brain dysfunction than in cognition in healthy subjects.

In our MRI analysis, we found that having two Val alleles at codon 158 of COMT was associated with decreased gray matter volumes only in the bilateral caudate nucleus (Figure 1) compared with zero Val alleles. Gennatas et al. (2012) (3) also found gray matter volume changes associated with increasing COMT Val dose in DA-innervated brain regions including the striatum in patients with dementia, including FTLD. In both studies, associations were observed between cognitive measures, behavioral measures, and gray matter density within DA-innervated brain regions (Table 2). Thus, we can assert with greater confidence the conclusions that were reached by Gennatas et al. (3): increased synaptic DA catabolism associated with the COMT Val allele appears to promote neurodegeneration within DA-innervated brain areas. However, Gennatas et al. (3) found this in several brain structures, whereas we found a significant association only in the bilateral caudate. This difference could be due in part to differences in our patient groups. Gennatas et al. (3) included patients with Alzheimer’s disease, bvFTD, and semantic dementia, whereas our study was limited to patients with FTLD-spectrum illnesses. In addition, our patient group included individuals with corticobasal syndrome, who have more disease involvement of the basal ganglia than other patients with FTLD (58).

The specificity we found of the volume loss related to COMT genotype for the caudate is striking (Figure 1). The caudate is heavily DA-innervated (59), and the caudate and putamen receive extensive afferent projections from the cerebral cortex (60). In corticobasal syndrome, the caudate and putamen show significant tau deposition, but so do several other structures, including the primary motor cortex and the subthalamic nucleus (61). Because all FTLD disorders involve significant cortical degeneration of areas that project to the basal ganglia, one might expect that the caudate, as a major target of those projections, could be particularly affected. Indeed, there is animal evidence that the caudate is more vulnerable to DA depletion than other structures in the basal ganglia (59).

Alzheimer’s disease appears to selectively target the cholinergic neurotransmitter system, and most available treatments for this illness increase available acetylcholine. The brain areas affected in FTD receive extensive dopaminergic and serotonergic projections (62). However, it is not known whether this is an anatomical coincidence (i.e., the disease process targets the frontal and anterior temporal lobes, and these just happen to receive dopaminergic and serotonergic projections) or whether these neurotransmitter systems influence the tropism and progression of the illness. The present study and related studies (3, 4) suggest that the DA system can affect the presentation of the illness and that DA-innervated areas may be particularly vulnerable.

Together, the findings of this study suggest that COMT and BDNF polymorphisms may explain, in part, the phenotypic variability of FTD and corticobasal syndrome. In addition, our results, along with those of other studies, highlight the importance of DA balance in FTLD and support the trial of medications augmenting dopamine in patients with FTD. Indeed, clinical trials of COMT inhibition in patients with FTLD are under way.