The rs SNP of brain-derived neurotrophic factor (BDNF) is associated with visual cognitive processing in multiple sclerosis

Transcription

1 Pathophysiology 18 (2011) The rs SNP of brain-derived neurotrophic factor (BDNF) is associated with visual cognitive processing in multiple sclerosis Bianca Weinstock-Guttman a, Ralph H.B. Benedict a,b, Miriam Tamaño-Blanco c, Deepa Preeti Ramasamy d, Milena Stosic d, Jennifer Polito c, Robert Zivadinov a,c, Murali Ramanathan a,c, a Jacobs Neurological Institute, Department of Neurology, Buffalo General Hospital, Buffalo, NY 14203, United States b Department of Psychiatry, State University of New York at Buffalo, Buffalo, NY 14260, United States c Department of Pharmaceutical Sciences, State University of New York at Buffalo, Buffalo, NY 14260, United States d Buffalo Neuroimaging Analysis Center, Jacobs Neurological Institute, State University of New York at Buffalo, NY 14203, United States Received 13 October 2009; received in revised form 14 March 2010; accepted 8 April 2010 Abstract Purpose: To investigate the associations between the rs SNP of brain-derived neurotrophic factor (BDNF) and neuropsychological (NP) test measures in multiple sclerosis (MS) patients. Background: BDNF regulates the survival of neuronal and non-neuronal cells and plays a critical role in neurochemical processes underlying learning and memory. Methods: A total of 209 MS patients (161 females; 48 males) underwent brain MRI and genotyping for BDNF rs The NP testing (n = 108) assessed processing speed, working memory, new learning and executive control. The MRI measurements included T1 and T2 lesion volume, whole brain, white and gray matter volumes, magnetization transfer imaging and regional subcortical brain volumes. Results: The T/T rs genotype group performed poorly on the Brief Visuospatial Memory Test-Revised (p = 0.031) and the Symbol Digit Modalities Test (p = 0.045) compared to the C/C genotype group. Because these NP tests both involve visual processing, the relationship with the volume of the thalamus was assessed. The BDNF rs genotype was associated with the volume of the left thalamus (p = 0.036). There were no significant associations with whole brain lesional and atrophy MRI measures. Conclusions: The C allele of BDNF rs is associated with protection against visual cognitive processing deficits via mechanisms that appear associated with the volume of the thalamus Published by Elsevier Ireland Ltd. Keywords: Neurotrophic factors; Multiple sclerosis; Memory; Vision; Thalamus Abbreviations: BDNF, brain-derived neurotrophic factor; BVMTR, Brief Visuospatial Memory Test-Revised; CSF, cerebrospinal fluid; CVLT2, California Verbal Learning Test, second edition; DKEFS, Delis Kaplan Executive Function System; DWI, diffusion-weighted imaging; EDSS, expanded disability status scale; FLAIR, fast, attenuated inversion recovery; FOV, field of view; GA, glatiramer acetate; GM, gray matter; GMV, gray matter volume; IFN-, interferon- ; MD, mean diffusivity; Met, methionine; MPD, mean parenchymal diffusivity; MRI, magnetic resonance imaging; MS, multiple sclerosis; MSFC, MS functional composite; NA, normal appearing; NMDA, N-methyl-d-aspartate; NP, neuropsychological; PASAT, Paced Auditory Serial Addition Test; PBMC, peripheral blood mononuclear cells (PBMC); RR, relapsing remitting; SDMT, Symbol Digits Memory Test; SE, spin echo; T1-LV, lesion volume in T1-weighted imaging; T2-LV, lesion volume in T2-weighted imaging; Val, valine; WM, white matter; WMV, white matter volume. Corresponding author at: Department of Pharmaceutical Sciences, 543 Cooke, Buffalo, NY 14260, United States. Tel.: ; fax: address: murali@buffalo.edu (M. Ramanathan) /$ see front matter 2010 Published by Elsevier Ireland Ltd. doi: /j.pathophys Introduction Cognitive dysfunction, which is estimated to occur in approximately 50 60% of multiple sclerosis (MS) patients [1 5] is a leading cause of vocational disability, disruption of quality of life and well being [5 7]. That cognitive function accounts for vocational disability is not surprising, considering the success of efforts to accommodate physical disability in the work place in recent years. Processing speed and episodic memory are two general domains of cognitive function that are most commonly affected in MS: [8]. Why some patients retain more cognitive capacity than others in the face of cerebral disease is an important area of inquiry, which could lead to a better understanding of brain reserve, plasticity and functional adaptation in MS. Research shows that a wide range of MRI measures

2 44 B. Weinstock-Guttman et al. / Pathophysiology 18 (2011) in MS are correlated with standardized measures of cognitive dysfunction. Recent MRI studies have revealed moderate to large effects when cognitive testing is correlated with ventricle enlargement [9 11], thalamic volume [12] and cortical volume [13,14]. Linear regression models from these studies have accounted for up to half of the variance in MS associated cognitive dysfunction. However, much of the variance remains unexplained. MS is also associated with a significant neurodegenerative component and progressive neuronal loss secondary to the initial inflammatory process [15 17]. BDNF, a neurotrophic factor abundantly expressed in the adult brain and also produced by immune cells, is an attractive candidate mediator that may partly explain the inter-individual differences in clinical outcomes and also the underlying neurochemical and immunological mechanisms. In the brain, BDNF is released by neurons and plays key roles in synaptic plasticity. BDNF and its receptor, TRK B, have been found in active MS lesions. BDNF is also a critical player in long-term potentiation (LTP), a neurochemical process that mediates learning and memory [18,19]. Glutamate receptors are critical for synaptic plasticity and BDNF facilitates glutamatergic synaptic transmission [20]. BDNF signaling via TrkB intersects directly with synaptic plasticity mechanisms involving the N-methyl-d-aspartate (NMDA) receptors [21]. The genetic variation that is the focus of this paper is a single nucleotide polymorphism (SNP), a T-to-C substitution (dbsnp identifier: rs ) in the BDNF gene. The rs SNP has not been investigated in MS and is distally located approximately 47 kb from the well-studied rs6265 SNP, which is associated with higher gray matter volume in MS [22]. We therefore investigated the relationship of rs BDNF genotype to the neurocognitive assessments and MRI parameters in MS patients. 2. Materials and methods 2.1. Study population This is a cross-sectional study of a large consecutive cohort of MS patients enrolled in an ongoing, prospective natural history study evaluating clinical, MRI, neurocognitive and genetic information in MS patients followed at the Baird MS Center, Jacobs Neurological Institute, Buffalo, NY. With informed consent, anti-coagulated peripheral blood was obtained by venipuncture from 209 consecutive patients (Table 1) with MS according to the McDonald criteria [23]. Of the 209 patients, 167 patients (79.9%) had relapsingremitting MS, 40 (19.1%) had secondary progressive MS and 2 (1%) had primary progressive MS. Patients had MRI as part of their routine clinical follow-up at our Center annually. All patients were on disease-modifying therapies (treatment duration 5.2 ± 3.4 years). Of the 209 patients, 174 (83.3%) were on interferon-, 22 (10.5%) were on glatiramer acetate and 12 were on other therapy (e.g., mitoxantrone, azathioprine, etc.). Patients who had exacerbations or received corticosteroid treatments in the preceding 30 days were excluded BDNF genotyping DNA was obtained from peripheral blood mononuclear cells (PBMC) preserved in TRI reagent (Molecular Research Center Inc., Cincinnati, OH) using the manufacturer s instructions [24,25]. The rs single nucleotide polymorphism in BDNF was characterized using the Assays-on-Demand genotyping kit (Applied Biosystems, Redwood City, CA). The fluorescent TaqMan oligonucleotide probes in the Assays-on- Demand genotyping kit specifically discriminate between the T and the C variants of BDNF rs Genotyping was performed according to the manufacturer s instructions on a MX4000 (Stratagene) real-time thermal cycler and the fluorescence outputs were analyzed using the MX4000 software. Non-template controls produced negligible background signals and excellent amplification and accurate genotyping calls were obtained using this method. The allele discrimination assay was crosschecked for a subset of nine patient samples with a restriction fragment length polymorphism assay. A 204 base pair region around the BDNF rs SNP was amplified using PCR (forward primer: TCCAAACATCACACAGCCTAA and reverse primer: GTGGTCAAAAGGGATGTGAGA). The PCR product was digested with the restriction enzyme AclI (New England BioLabs Inc., Ipswich, MA) at 37 C overnight. The restriction enzyme digest was separated on a 2.5% methaphor agarose gel (CAMBREX Rockland, ME) and the genotypes were identified based on the differential band patterns. The allele discrimination and restriction fragment length polymorphism assay were in complete agreement Neuropsychological testing protocol Neuropsychological (NP) testing was conducted in accordance with consensus standards for evaluation of MS patients [26]. It is well established that defects in processing speed, working memory, new learning and executive control are most common in MS, whereas intelligence and language are more often preserved [5,27]. For this reason, our analysis of cognitive capacity was restricted to tests measuring the aforementioned domains. The NP and MRI assessments were obtained in the course of the routine clinical care of the MS patients of the cohort. At our Center, MS patients receive MRIs annually and NP assessments every other year. We used the subset of patients with NP assessments and MRI data within ±3 months of each other. NP data were available for a subset of 108 patients (Table 1).

3 B. Weinstock-Guttman et al. / Pathophysiology 18 (2011) Table 1 Clinical and demographic characteristics of the cohort. Data are mean ± SD, except for EDSS, which is expressed as median (25% quartile 75% quartile range). Characteristic MRI MTR subset FreeSurfer subset NP subset n Females:males 161:48 (77%) 113:32 (78%) 31:12 (72%) 83:25 (77%) Age, years 45.2 ± ± ± ± 7.7 Age of onset, years 31.7 ± ± ± ± 9.5 Disease duration, years * 13.4 ± ± ± ± 8.9 EDSS 2.5 ( ) 2.5 ( ) 2.5 ( ) 2.5 ( ) Disease-modifying therapy, years 5.2 ± ± ± ± 3.2 Education, years 14.4 ± ± ± ± 2.3 * From onset of symptoms. All patients undergoing cognitive testing were screened for impaired, corrected near visual acuity defect using a Snellen near vision chart. Patients with corrected, bilateral near visual acuity poorer than 20/70 were excluded. We did not acquire NP data on patients without adequate vision to enable testing. Processing speed and working memory were evaluated using modified versions of the Symbol Digit Modalities Test (SDMT) [28] and the Paced Auditory Serial Addition Test (PASAT) [29]. Episodic memory was assessed with the California Verbal Learning Test, second edition (CVLT2) [30] and the Brief Visuospatial Memory Test-Revised (BVMTR) [31]. The Delis Kaplan Executive Function System (DKEFS) Sorting Test was used as a measure of higher executive function [32]. The methodology for these tests has been previously described [22]. Data for all neuropsychological tests were transformed to Z-scores using demographically matched controls from Benedict et al. [14] MRI acquisition and analysis Image acquisition Quantitative MRI analysis was available for all 209 patients analyzed for the BDNF rs polymorphism. Patients underwent brain MRI using a 1.5-T General Electric Signa 4x/Lx, scanner. T2-weighted image (WI), diffusionweighted imaging (DWI), 3D-spoiled-gradient recalled (SPGR) T1-WI, spin echo (SE) T1-WI with and without gadolinium (Gd) contrast, fast, attenuated inversion recovery (FLAIR), proton density (PD), and PD with magnetization transfer (MT) pulse images were obtained. The pulse sequences for MRI acquisition have been previously described [22] Image analysis Lesion measures The number of brain T1 Gd positive lesions was based on manual tracing on the digital films [33]. The T2-, T1- and Gd lesion volumes (LVs) were measured using a semi-automated edge detection contouring-thresholding technique that was manually corrected for the region of interest, as previously described [34] Global and tissue-specific atrophy measures For brain extraction and tissue segmentation, we utilized the SIENAX cross-sectional brain atrophy analysis tool [35,36]. Compartment-specific absolute volumes were then quantified and the normalized volumes of whole brain (NBV), GM (NGMV) and white matter (WM) (NWMV) were obtained, as reported previously [27] Subcortical segmentation The regional subcortical brain volumes were available for a subset of 43 patients (Table 1). The volumebased subcortical segmentation and surface based cortical reconstruction on 3D T1-weighted SPGR images were completed using FreeSurfer software ( The volume-based stream is an automated process that reslices 3D T1-weighted SPGR images to approximately 1mm 3 resolutions for whole brain tissue segmentation, subcortical parcellation and quantification of specific subcortical region tissue volumes. The stream consists of five different stages [37]. Initially, the MRI volumes are registered to the Talairach space and the output images are intensity normalized. At the next stage, the skull is automatically stripped off the 3D anatomical data set by using a hybrid method that uses both watershed algorithms and deformable surface models. At this stage, manual intervention is needed to visualize and edit areas of skull and the areas of cortex or cerebellum that should be corrected. After skull stripping, the output brain mask is labeled using a probabilistic atlas where each voxel in the normalized brain mask volume is assigned one of the following labels: cerebral white matter, cerebral cortex, lateral ventricle, inferior lateral ventricle, cerebellum white matter, cerebellum cortex, thalamus, caudate, putamen, pallidum, hippocampus, amygdala, accumbens area, third ventricle, fourth ventricle, brain stem, and cerebrospinal fluid [38] Diffusion-weighted measures The details of the DWI method have been described elsewhere [27,39]. The mean parenchymal diffusivity (MPD)

4 46 B. Weinstock-Guttman et al. / Pathophysiology 18 (2011) values for the brain parenchyma were obtained from the analysis Magnetization transfer measures The MT post-processing was completely automated [22] and a detailed description is provided in [22]. The MTR of T2 and T1-LVs, whole brain (WB) MTR, normal appearing (NA) brain tissue (NABT) MTR, NAWM MTR and NAGM MTR were obtained. Given the superior reliability of mean MTR, and in order to minimize the number of multiple statistical tests, the analyses emphasized the mean MTR measure. The MTR measures were available for a subset of 145 patients (Table 1) Data analysis The analysis plan was developed to search for linear relationships between the rs genotype and neuropsychological tests with established sensitivity in MS and key brain MRI variables. First, in preliminary analyses, descriptive statistics were derived and the relationship between EDSS and rs genotype was tested. Second, we assessed the relationship between eight neuropsychological tests and the rs genotype in separate linear regression models. Finally, the same approach was employed for 12 brain MRI dependent variables. The SPSS (SPSS Inc., Chicago, IL) statistical program was used for all statistical analyses. The cube root transformation was applied to T2-LV and T1-LV prior to statistical analysis [40]. The multivariate linear regression analysis of MRI variables included gender, presence of progressive MS and BDNF rs genotype as factors and age, disease duration, treatment duration as covariates. For the NP variables, the number of years of education was included as an additional covariate. 3. Results 3.1. Patient characteristics The demographic and clinical characteristics of our patient population are summarized in Table 1. The MRI, NP and FreeSurfer analysis cohorts were comparable across the range of demographic and clinical characteristics The BDNF rs genotype distribution in MS patients The T to C SNP variation at rs of BDNF was genotyped in 209 patients with clinically definite MS. As summarized in Table 2, 55 (26.3%) patients were T/T, 99 (47.4%) were T/C, and 55 (26.3%) were C/C. The allele frequency of the T allele and C alleles were both calculated to be 50%. In addition, the genotype and allele frequencies obtained in MS patients were compared separately to Table 2 Distribution of the BDNF rs genotypes. rs genotype Number Percent T/T % T/C % C/C % the frequencies in the cohort of healthy controls reported by the HapMap and SNP Consortium projects and found to be similar. In regression analysis with age of onset as the dependent variable and correcting for gender (p = 0.54), the age of onset was independent of BDNF rs genotype (p = 0.99). In regression analyses with Kurtzke EDSS [41] as the dependent variable (F = 16.8, p < 0.001, adjusted R 2 = 0.38) correcting for gender (p = 0.95), age (p = 0.83), presence of progressive MS (standardized β = 0.488, p < 0.001), disease duration (standardized β = 0.156, p = 0.049) and treatment duration (p = 0.057), the EDSS was not associated with the rs genotype (p = 0.85). These analyses suggest that the BDNF rs genotype does not determine susceptibility to MS and is not associated with the physical disability as measured with the EDSS, a measure of physical disability in MS that does not encompass cognitive status Associations with neuropsychological measures The NP parameters for the rs genotypes are summarized in Table 3. To delineate the factors associated with NP parameters, we conducted regression analysis with the NP variable of interest as the dependent variable, the BDNF rs SNP genotype as independent variable and corrected for age, gender, disease duration and treatment duration (Table 4). The BDNF rs SNP genotype was significantly associated with the BVMTR (standardized β = 0.233, p = 0.031) and SDMT (standardized β = 0.215, p = 0.045). The standardized β values indicate that the presence of the C allele of BDNF rs is associated with better performance on both tests The BDNF rs SNP is not associated with whole brain lesion, atrophy or MTR MRI measures The MRI characteristics of the T/T, C/T and C/C genotypes groups are summarized in Table 5. A detailed summary of the regression analysis results for each MRI parameter is presented in Table 6. No significant associations were found between the BDNF rs genotype and the MRI parameters in Table 6. To assess whether the BDNF rs genotype was associated with differences in the extent of microscopic damage, the MTR values for T2 and T1 lesions, whole brain, normal appearing white matter, normal appearing gray matter and normal appearing brain tissue were also

5 B. Weinstock-Guttman et al. / Pathophysiology 18 (2011) Table 3 Summary of NP characteristics with genotypes of BDNF rs The values for each test are expressed as mean Z-scores ± SD. Test Z-score All T/T homozygous T/C heterozygous C/C homozygous CVLT2 total 0.70 ± ± ± ± 1.11 CVLT2 delay 0.86 ± ± ± ± 1.35 BVMTR total 1.31 ± ± ± ± 1.49 BVMTR delay 1.33 ± ± ± ± 1.41 PASAT 0.59 ± ± ± ± 1.18 SDMT 1.58 ± ± ± ± 1.27 DKEFS correct sorts 0.86 ± ± ± ± 1.21 DKEFS description score 0.90 ± ± ± ± 1.22 Table 4 Regression results for NP characteristics with BDNF rs genotypes. Significant variables are underlined. Test Z-score rs genotype Male gender Age (years) Education (years) Progressive MS Disease duration Treatment duration Std β p Std β p Std β p Std β p Std β p Std β p Std β p CVLT2 total CVLT2 delay BVMTR total BVMTR delay PASAT SDMT DKEFS Correct sorts DKEFS Description score Table 5 MRI characteristics for the cohort of patients genotyped for the BDNF rs SNP and treated with disease-modifying therapy. Data represents mean ± SD. MRI parameter All patients rs genotype TT TC CC T1 lesion volume (ml) 2.38 ± ± ± ± 4.94 T2 lesion volume (ml) 12.4 ± ± ± ± 15.1 Gray matter volume (ml) 752 ± ± ± ± 67 White matter volume (ml) 730 ± ± ± ± 51 Normalized brain volume (ml) 1482 ± ± ± ± 81 Mean diffusivity 1152 ± ± ± ± 97 Mean MTR for T1 lesions 30.1 ± ± ± ± 3.89 Mean MTR for T2 lesions 33.4 ± ± ± ± 3.48 Mean MTR for whole brain 35.2 ± ± ± ± 2.78 Mean MTR for NAWM 38.8 ± ± ± ± 2.64 Mean MTR for NAGM 32.2 ± ± ± ± 3.06 Mean MTR for NA brain tissue 35.3 ± ± ± ± 2.77 Table 6 Regression analysis of MRI characteristics with BDNF rs genotype. Significant variables are underlined. Variable rs genotype Male gender Age (years) Progressive MS Disease duration (years) Treatment duration (months) Std β p Std β p Std β p Std β p Std β p Std β p T1-LV a T2-LV a GMV < WMV NBV < a Cube root transformed.

6 48 B. Weinstock-Guttman et al. / Pathophysiology 18 (2011) analyzed in regression analyses. No associations with BDNF rs were found (data not shown) The BDNF rs SNP is associated with thalamic volumes measures from FreeSurfer analysis Thus, we observed a relationship between NP testing and the BDNF rs polymorphism but there was no relationship with the whole brain lesional, atrophy and MTRbased MRI measures. However, we did note that both of the tests showing significance the BVMTR and SDMT involve visual processing. We hypothesized that the general MRI measures may not have fully captured the regional pathology specifically related to these tests. In previous work from our group, we have shown robust correlation between SDMT and thalamic volume [12]. Our regional MRI analyses focused on the thalamus to explain the deficiencies on the SDMT and the BVMTR because: (i) it is affected in MS [12,42] and, (ii) it contains the pulvinar nuclei that are associated with visual cognitive processing. The pulvinar nuclei represent a substantial portion of thalamic volume and have extensive reciprocal connectivity with the cortex and receive input from the colliculus and retina [43]. We obtained regional analyses of various regions of the brain using the FreeSurfer software in 42 patients to address this hypothesis. The overall disease characteristics of this subset of patients were similar to that of the overall cohort (Table 1). The BDNF rs genotype distribution in this subset consisted of 13 (30.2%) T/T homozygous, 14 (32.6%) T/C heterozygous and 16 (37.2%) C/C homozygous. We conducted regression analysis with the NP measures (BVMTR Z-score or SDMT Z-score) as the dependent variable and with the following as independent variables: age, gender, years of education, and the volume of the brain region of interest to determine whether the brain region volume was associated with the NP measure. We focused on the volume of thalamus (left, right and total). In regression analyses with the BVMTR total Z-score as the dependent variable, a strong association with the volume of the left thalamus (standardized β = 0.409, p = 0.020) and the total volume of the thalami (standardized β = 0.470, p = 0.008) was found (Table 7). In corresponding regression analyses with the SDMT Z-score as the dependent variable, also indicated strong associations with the volume of the left thalamus (standardized β = 0.549, p = 0.002) and the total volume of the thalami (standardized β = 0.471, p = 0.012) was found (Table 8). The volume of the right thalamus was associated with the BVMTR total Z- score (standardized β = 0.504, p = 0.006) but not the SDMT (standardized β = 0.328, p = 0.10). The exact reasons for these left-right differences are not known. In the next analyses, we assessed whether the genotypes of BDNF rs were associated with the volumes of brain regions of interest in MS patients. We conducted regression analysis with the volume of the brain region of interest as the dependent variable and with age, gender, disease duration, treatment duration, presence of progressive MS and BDNF Table 7 Regression results of NP scores with various brain regions as dependent variable with genotypes of rs Significant variables are underlined. Brain region BVMTR Z-score dependent variable SDMT Z-score dependent variable Brain region volume (mm 3 ) Male gender Age (years) Education (years) Brain region volume (mm 3 ) Male gender Age (years) Education (years) Std β p Std β p Std β p Std β p Std β p Std β p Std β p Std β p Left thalamus Right thalamus Both thalami Left hippocampus Right hippocampus Both hippocampi Brain region PASAT Z-score dependent variable CVLT Z-score dependent variable Std β p Std β p Std β p Std β p Std β p Std β p Std β p Std β p Left thalamus Right thalamus Both thalami Left hippocampus Right hippocampus Both hippocampi

7 B. Weinstock-Guttman et al. / Pathophysiology 18 (2011) Table 8 Regression results with various brain regions as dependent variable with BDNF rs genotype. Significant variables are underlined. Brain region volume rs genotype Male gender Age (years) Progressive MS Disease duration Treatment duration Std β p Std β p Std β p Std β p Std β p Std β p Left thalamus Right thalamus Total thalamus Left hippocampus Right hippocampus Total hippocampus rs genotype as independent variables (see Table 8). The BDNF rs genotype was associated with the volume of the left thalamus (standardized β = 0.314, p = 0.036) and the association total volume of thalami approached significance (standardized β = 0.269, p = 0.055). The BDNF rs genotype was not associated with the volume of the hippocampus. We also examined the relationship between the thalamus volumes and the performance on the PASAT and CVLT2 (Table 8). These tests also require processing speed and memory but require auditory and verbal stimuli. The PASAT and CVLT2 Z-scores associations with thalamic and hippocampus volume measures were not significant. These results suggest that the effects of the BDNF rs genotype on the volume of the thalamus can explain the selective effects on the BVMTR and SDMT, which involve visual processing. 4. Discussion This is the first study examining the relationship between BDNF rs and neurological outcomes in MS and we have presented data that demonstrate that BDNF rs genotype is associated with specific neurocognitive deficits in MS patients. These associations between the rs SNP of BDNF and BVMTR and SDMT neuropsychological parameters are promising findings and highlight the BDNF rs polymorphism as a potential genetic variation determining neuropsychological vulnerability in MS patients. Our data suggest that the effects of BDNF rs are not directly linked to several of the currently used whole brain atrophy, lesion burden and microscopic damage MRI parameters. However, the associations between BDNF rs genotype with volume of the thalamus were observed and these findings provide a mechanistic explanation for the visual processing deficits found in our study. As noted previously, the BDNF rs genotype showed significant association in regression analysis with two tests, the BVMTR and SDMT. These tests measure different aspects of cognition, processing speed and episodic memory, respectively. Both require the sensory processing of visual stimuli and visual/spatial processing in order to derive correct responses. We focused on the thalamus because this region of the brain plays a critical role in processing and transmitting sensory input to cortical regions. More importantly, emerging data indicate the thalamus is atrophic in MS [12,42]. Based on prior literature, the BVMTR and SDMT performance were also known to be strongly associated with thalamic volume [12]. These factors provided the rationale for examining the relationships between thalamic volume as obtained in FreeSurfer analyses and BDNF rs genotype when the conventional lesional and whole brain atrophy measures did not provide satisfactory explanations. A potential limitation is that our sample size (n = 43) in the FreeSurfer

8 50 B. Weinstock-Guttman et al. / Pathophysiology 18 (2011) analysis was limited; however, the genotype distributions in the limited sample were favorable (13 or more patients in each genotype group) and not excessively skewed. Because MS can cause visual impairment, all patients were screened for visual acuity before NP testing to avoid confounding of the NP measurements. Patients with the corrected visual acuity of worse than 20/70 were excluded. There is evidence that selective atrophy of thalamus, hippocampus, caudate and putamen occurs in MS patients and the atrophy of deep gray matter structures can occur at rates that are two to three times faster than that in global and cortical regions [44]. There is also evidence for the disproportionate vulnerability for thalamic atrophy relative to whole brain atrophy [12]. Selective atrophy of the left thalamus has been reported in SP-MS [45]. Genetic factors could also alter the baseline values of regional volumes [46,47]. However, the cross-sectional design of our study precludes complete assessments of the relative contributions of disease and baseline differences [47]. The rs SNP is in a non-coding region of the BDNF gene and its genotypes were not associated with immune cell secretion of BDNF (data not shown). The biological mechanisms of the rs effects are not fully known but could potentially involve yet uncharacterized effects on BDNF function in neurons as genetic associations involving rs in haplotypes have been reported. There is now considerable evidence that the regulation of BDNF is complex and involves a range of posttranscriptional mechanisms. Furthermore, there are at least three studies that indicate that BDNF rs containing haplotypes are associated with susceptibility in other conditions. Beuten et al. conducted haplotype analysis of four BDNF SNPs, rs rs rs rs , and observed that the major T-C-T-G haplotype was significantly associated with nicotine dependence measures in the European-American male sub-group [48]. Qian et al. conducted haplotype analysis of three BDNF SNPs and the (GT) n dinucleotide repeat, rs6265-(gt) n -rs rs [49]. They observed that the A-274-C-T haplotype at these respective loci was protective against schizophrenia in Chinese subjects. Interestingly, a common theme vis-à-vis BDNF rs that can be inferred from combining both reports is that the T allele may be generally associated with adverse outcomes. Interestingly in this report, we too found that the group with the C allele had better BVMTR total and SDMT Z-scores. In contrast to the relative dearth of information on the functional effects of BDNF rs , the BDNF rs6265 SNP, which results in a substitution of a methionine for a valine in the BDNF pro-protein, has been widely investigated [46,50]. The activity dependent secretion of BDNF protein is impaired with the methionine substitution and the volume of the hippocampus is reduced in healthy subjects with the allele coding for methionine [46,47,50]. Wehave investigated the effect of the BDNF rs6265 on MRI and NP parameters in MS patients and found an association with the gray matter volume and an association trend with the PASAT Z-score [22]. The strong correlations between the rs genotype and visual cognitive processing and memory should provide the rationale for the in vitro and in vivo characterization of its effects on BDNF function in neuronal and non-neuronal cells and during brain development. It is becoming increasingly clear from large case control, hypothesis-generating, genome-wide association studies that the contributions of individual genetic polymorphisms to MS susceptibility is relatively modest. In a recently published study, polymorphisms in the IL7RA and IL2RA genes were identified as strongly linked to MS (p < ) [51]. Despite the low p-values, these polymorphisms accounted for less than 0.2% of the variance in risk [51]. No clear genetic determinants for MS disease course and clinical characteristics have emerged but the majority of these studies have focused on measures such as age of onset or the Kurtzke EDSS scores that are considered to be relatively coarse instruments. The APOE4 polymorphism has been investigated using clinical, MRI and NP approaches in MS [52 54] but the results have been mixed [55]. The p-values for significantly associated NP components when found have been in the range of [56,57]. We are currently genotyping the BDNF gene comprehensively in large cohort of MS patients with longitudinal MRI that are designed to provide independent validation of the findings of this paper. The effects of individual genes and genetic variations (including BDNF) on complex, critical functions in the brain are expected to be in the small to moderate range and we are also conducting power and sample size calculations for several study designs for MRI and NP parameters, which are quantitative traits. Indeed, it should be acknowledged that the genetics, environmental epidemiology, immunology and neurobiology underlying the MS disease process and its effects on cognition are in themselves complex and therefore, identifying the individual contributing factors is challenging. Nonetheless, investigation of promising genes such as BDNF and polymorphisms such as rs in patient populations is a starting point that could provide valuable insights particularly when combined with quantitative techniques such as MRI and NP to measure MS disease effects. Conflict of interest Dr. Zivadinov received personal compensation from Teva Neuroscience, Biogen Idec, Aspreva, Pfizer and Serono for speaking and consultant fees. These were unrelated to the research in this manuscript. Dr. Zivadinov received financial support for research activities from National Institute of Health, National Multiple Sclerosis Society, National Science Foundation, Biogen Idec, Teva Neuroscience, Aspreva and Jog for the Jake Foundation. Dr. Bianca Weinstock-Guttman received honoraria and compensation from Teva Neuroscience, Biogen Idec, Berlex/Bayer and Serono for speaking and consultant fees.

9 B. Weinstock-Guttman et al. / Pathophysiology 18 (2011) These were unrelated to the research in this manuscript. Dr. Bianca Weinstock received financial support for research activities from National Institutes of Health, National Multiple Sclerosis Society, National Science Foundation, Biogen Idec, Teva Neuroscience, Aspreva, EMD Serono and Jog for the Jake Foundation. Dr. Ralph Benedict received consulting and research fees from Biogen Idec and Cognition Pharmaceuticals. These were unrelated to the research in this manuscript. Dr. Ralph Benedict received financial support for research activities the National Multiple Sclerosis Society. Dr. Murali Ramanathan financial support for research activities from the National Institutes of Health, National Multiple Sclerosis Society, National Science Foundation, Kapoor Foundation, Pfizer Inc., Novartis Inc., EMD Serono and Jog for the Jake Foundation. This work was funded by research grant from the National Multiple Sclerosis Society. The other funding was not related to this grant. All other authors do not have conflicts of interest. References [1] M.P. Amato, G. Ponziani, G. Pracucci, L. Bracco, G. Siracusa, L. Amaducci, Cognitive impairment in early-onset multiple sclerosis. Pattern, predictors, and impact on everyday life in a 4-year follow-up, Arch. Neurol. 52 (1995) [2] O. Lyon-Caen, R. Jouvent, S. Hauser, et al., Cognitive function in recent-onset demyelinating diseases, Arch. Neurol. 43 (1986) [3] J.M. Peyser, S.M. Rao, N.G. LaRocca, E. Kaplan, Guidelines for neuropsychological research in multiple sclerosis, Arch. Neurol. 47 (1990) [4] M. Prosiegel, C. Michael, Neuropsychology and multiple sclerosis: diagnostic and rehabilitative approaches, J. Neurol. Sci. 115 (Suppl.) (1993) S51 S54. [5] S.M. Rao, G.J. Leo, L. Bernardin, F. Unverzagt, Cognitive dysfunction in multiple sclerosis. I. Frequency, patterns, and prediction, Neurology 41 (1991) [6] W.W. Beatty, Cognitive and emotional disturbances in multiple sclerosis, Neurol. Clin. 11 (1993) [7] K.V. Wild, M.D. Lezak, R.H. Whitman, D.N. Bourdette, Psychosocial impact of cognitive impairment in the multiple sclerosis patient, J. Clin. Exp. Neuropsychol. 13 (1991) 74. [8] R.H.B. Benedict, J.H. Bobholz, Multiple sclerosis, Semin. Neurol. 27 (2007) [9] R.H.B. Benedict, B. Weinstock-Guttman, I. Fishman, J. Sharma, C.W. Tjoa, R. Bakshi, Prediction of neuropsychological impairment in multiple sclerosis: comparison of conventional magnetic resonance imaging measures of atrophy and lesion burden, Arch. Neurol. 61 (2004) [10] C. Christodoulou, L.B. Krupp, Z. Liang, et al., Cognitive performance and MR markers of cerebral injury in cognitively impaired MS patients, Neurology 60 (2003) [11] A. Tekok-Kilic, R.H.B. Benedict, B. Weinstock-Guttman, et al., Independent contributions of cortical gray matter atrophy and ventricle enlargement for predicting neuropsychological impairment in multiple sclerosis, Neuroimage 36 (2007) [12] M.K. Houtchens, R.H. Benedict, R. Killiany, et al., Thalamic atrophy and cognition in multiple sclerosis, Neurology 69 (2007) [13] M.P. Amato, M.L. Bartolozzi, V. Zipoli, et al., Neocortical volume decrease in relapsing-remitting MS patients with mild cognitive impairment, Neurology 63 (2004) [14] R.H.B. Benedict, J.M. Bruce, M.G. Dwyer, et al., Neocortical atrophy, third ventricular width, and cognitive dysfunction in multiple sclerosis, Arch. Neurol. 63 (2006) [15] M.H. Barnett, J.W. Prineas, Relapsing and remitting multiple sclerosis: pathology of the newly forming lesion, Ann. Neurol. 55 (2004) [16] C. Lucchinetti, W. Bruck, J. Parisi, B. Scheithauer, M. Rodriguez, H. Lassmann, Heterogeneity of multiple sclerosis lesions: implications for the pathogenesis of demyelination, Ann. Neurol. 47 (2000) [17] B.D. Trapp, J. Peterson, R.M. Ransohoff, R. Rudick, S. Mork, L. Bo, Axonal transection in the lesions of multiple sclerosis, N. Engl. J. Med. 338 (1998) [18] B. Lu, W. Gottschalk, Modulation of hippocampal synaptic transmission and plasticity by neurotrophins, Prog. Brain Res. 128 (2000) [19] M.M. Poo, Neurotrophins as synaptic modulators, Nat. Rev. Neurosci. 2 (2001) [20] K. Yamada, T. Nabeshima, Brain-derived neurotrophic factor/trkb signaling in memory processes, J. Pharmacol. Sci. 91 (2003) [21] I.B. Black, Trophic regulation of synaptic plasticity, J. Neurobiol. 41 (1999) [22] R. Zivadinov, B. Weinstock-Guttman, R.H.B. Benedict, et al., Preservation of gray matter volume in multiple sclerosis patients with the Met allele of the rs6265 (Val66Met) SNP of brain-derived neurotrophic factor, Hum. Mol. Genet. 16 (2007) [23] W.I. McDonald, A. Compston, G. Edan, et al., Recommended diagnostic criteria for multiple sclerosis: guidelines from the International Panel on the diagnosis of multiple sclerosis, Ann. Neurol. 50 (2001) [24] P. Chomczynski, A reagent for the single-step simultaneous isolation of RNA, DNA and proteins from cell and tissue samples, Biotechniques 15 (1993) , [25] P. Chomczynski, K. Mackey, Short technical reports. Modification of the TRI reagent procedure for isolation of RNA from polysaccharideand proteoglycan-rich sources, Biotechniques 19 (1995) [26] R.H. Benedict, J.S. Fischer, C.J. Archibald, et al., Minimal neuropsychological assessment of MS patients: a consensus approach, Clin. Neuropsychol. 16 (2002) [27] R.H. Benedict, D. Cookfair, R. Gavett, M. Gunther, F. Munschauer, N. Garg, B. Weinstock-Guttman, Validity of the minimal assessment of cognitive function in multiple sclerosis (MACFIMS), J. Int. Neuropsychol. Soc. 12 (2006) [28] A. Smith, Symbol Digit Modalities Test: Manual, Western Psychological Services, Los Angeles, [29] D.M. Gronwall, Paced auditory serial-addition task: a measure of recovery from concussion, Percept. Motor Skills 44 (1977) [30] D.C. Delis, J.H. Kramer, E. Kaplan, B.A. Ober, The California Verbal Learning Test Manual, second edition, The Psychological Corporation, San Antonio, [31] R.H.B. Benedict, D. Schretlen, L. Groninger, M. Dobraski, Revision of the Brief Visuospatial Memory Test: studies of normal performance, reliability, and validity, Psychol. Assess.: J. Consult. Clin. Psychol. 8 (1996) [32] D.C. Delis, E. Kaplan, J.H. Kramer, Delis Kaplan Executive Function System TM (D-KEFS TM ), Harcourt Assessment, Inc., San Antonio, TX, [33] R. Zivadinov, M. Dwyer, K.L. Watt, Measurement of cerebral grey and white matter atrophy measurement from various MRI pulse sequences using different segmentation algorithms, J. Neurol. 251 (2004) S89. [34] R. Zivadinov, R.A. Rudick, R. De Masi, et al., Effects of IV methylprednisolone on brain atrophy in relapsing-remitting MS, Neurology 57 (2001) [35] S.M. Smith, Y. Zhang, M. Jenkinson, et al., Accurate, robust, and automated longitudinal and cross-sectional brain change analysis, Neuroimage 17 (2002)

10 52 B. Weinstock-Guttman et al. / Pathophysiology 18 (2011) [36] S.M. Smith, N. De Stefano, M. Jenkinson, P.M. Matthews, Normalized accurate measurement of longitudinal brain change, J. Comput. Assist. Tomogr. 25 (2001) [37] B. Fischl, D.H. Salat, E. Busa, et al., Whole brain segmentation: automated labeling of neuroanatomical structures in the human brain, Neuron 33 (2002) [38] D. Ramasamy, D. Fritz, J.L. Cox, et al., Extent of deep gray matter atrophy in patients with multiple sclerosis. A case control study, J. Neurol. 254 (2007), O65:14. [39] E. Tavazzi, M.G. Dwyer, B. Weinstock-Guttman, et al., Quantitative diffusion weighted imaging measures in patients with multiple sclerosis, Neuroimage (2007). [40] D.K. Li, U. Held, J. Petkau, et al., MRI T2 lesion burden in multiple sclerosis: a plateauing relationship with clinical disability, Neurology 66 (2006) [41] J.F. Kurtzke, Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS), Neurology 33 (1983) [42] A. Cifelli, M. Arridge, P. Jezzard, M.M. Esiri, J. Palace, P.M. Matthews, Thalamic neurodegeneration in multiple sclerosis, Ann. Neurol. 52 (2002) [43] K.L. Grieve, C. Acuna, J. Cudeiro, The primate pulvinar nuclei: vision and action, Trends Neurosci. 23 (2000) [44] R. Zivadinov, D. Ramasamy, E. Havrdova, et al., A longitudinal study of deep gray matter atrophy in patients with multiple sclerosis. A case control study, Mult. Scler. (2007), P606:S182. [45] E. Pagani, M.A. Rocca, A. Gallo, et al., Regional brain atrophy evolves differently in patients with multiple sclerosis according to clinical phenotype, AJNR Am. J. Neuroradiol. 26 (2005) [46] M.F. Egan, M. Kojima, J.H. Callicott, et al., The BDNF val66met polymorphism affects activity-dependent secretion of BDNF and human memory and hippocampal function, Cell 112 (2003) [47] L. Pezawas, B.A. Verchinski, V.S. Mattay, et al., The brain-derived neurotrophic factor val66met polymorphism and variation in human cortical morphology, J. Neurosci. 24 (2004) [48] J. Beuten, J.Z. Ma, T.J. Payne, et al., Significant association of BDNF haplotypes in European-American male smokers but not in European- American female or African-American smokers, Am. J. Med. Genet. B Neuropsychiatr. Genet. 139 (2005) [49] L. Qian, J. Zhao, Y. Shi, et al., Brain-derived neurotrophic factor and risk of schizophrenia: an association study and meta-analysis, Biochem. Biophys. Res. Commun. 353 (2007) [50] A.R. Hariri, T.E. Goldberg, V.S. Mattay, et al., Brain-derived neurotrophic factor val66met polymorphism affects human memoryrelated hippocampal activity and predicts memory performance, J. Neurosci. 23 (2003) [51] D.A. Hafler, A. Compston, S. Sawcer, et al., Risk alleles for multiple sclerosis identified by a genomewide study, N. Engl. J. Med. 357 (2007) [52] C. Enzinger, S. Ropele, S. Smith, et al., Accelerated evolution of brain atrophy and black holes in MS patients with APOE-epsilon 4, Ann. Neurol. 55 (2004) [53] F. Fazekas, S. Strasser-Fuchs, H. Kollegger, et al., Apolipoprotein E epsilon 4 is associated with rapid progression of multiple sclerosis, Neurology 57 (2001) [54] T. Masterman, Z. Zhang, D. Hellgren, et al., APOE genotypes and disease severity in multiple sclerosis, Mult. Scler. 8 (2002) [55] R.M. Burwick, P.P. Ramsay, J.L. Haines, et al., APOE epsilon variation in multiple sclerosis susceptibility and disease severity: some answers, Neurology 66 (2006) [56] B.A. Parmenter, D.R. Denney, S.G. Lynch, L.S. Middleton, L.M. Harlan, Cognitive impairment in patients with multiple sclerosis: association with the APOE gene and promoter polymorphisms, Mult. Scler. 13 (2007) [57] J. Shi, C.B. Zhao, T.L. Vollmer, T.M. Tyry, S.M. Kuniyoshi, APOE epsilon 4 allele is associated with cognitive impairment in patients with multiple sclerosis, Neurology 70 (2008)