Neurobiology of Disease 74 (2015) 406–412

Contents lists available at ScienceDirect

Neurobiology of Disease journal homepage: www.elsevier.com/locate/ynbdi

Longitudinal change in white matter microstructure in Huntington's disease: The IMAGE-HD study Govinda R. Poudel a,b,f, Julie C. Stout a, Juan F. Domínguez D. a, Andrew Churchyard a,d, Phyllis Chua c,d, Gary F. Egan a,b,e,f, Nellie Georgiou-Karistianis a,⁎ a

School of Psychological Sciences, Monash University, Clayton, Victoria, Australia Monash Biomedical Imaging (MBI), Monash University, Melbourne, Victoria, Australia Department of Psychiatry, School of Clinical Sciences, Monash University, Victoria, Australia d Calvary Health Care Bethlehem Hospital, Caulfield, Victoria, Australia e ARC Centre of Excellence for Integrative Brain Function, Monash University, Clayton, Victoria, Australia f VLSCI Life Sciences Computation Centre, Melbourne, Victoria, Australia b c

a r t i c l e

i n f o

Article history: Received 20 August 2014 Revised 14 November 2014 Accepted 8 December 2014 Available online 12 December 2014 Keywords: Longitudinal DTI Huntington's disease TBSS Diffusion tensor imaging

a b s t r a c t Objective: To quantify 18-month changes in white matter microstructure in premanifest (pre-HD) and symptomatic Huntington's disease (symp-HD). To investigate baseline clinical, cognitive and motor symptoms that are predictive of white matter microstructural change over 18 months. Method: Diffusion tensor imaging (DTI) data were analyzed for 28 pre-HD, 25 symp-HD, and 27 controls scanned at baseline and after 18 months. Unbiased tract-based spatial statistics (TBSS) methods were used to identify longitudinal changes in fractional anisotropy (FA), radial diffusivity (RD), and axial diffusivity (AD) of white matter. Stepwise linear regression models were used to identify baseline clinical, cognitive, and motor measures that are predictive of longitudinal diffusion changes. Results: Symp-HD compared to controls showed 18-month reductions in FA in the corpus callosum and cingulum white matter. Symp-HD compared to pre-HD showed increased RD in the corpus callosum and striatal projection pathways. FA in the body, genu, and splenium of the corpus callosum was significantly associated with a baseline clinical motor measure (Unified Huntington's Disease Rating Scale: total motor scores: UHDRS–TMS) across both HD groups. This measure was also the only independent predictor of longitudinal decline in FA in all parts of the corpus callosum across both HD groups. Conclusions: We provide direct evidence of longitudinal decline in white matter microstructure in symp-HD. Although pre-HD did not show longitudinal change, clinical symptoms and motor function predicted white matter microstructural changes for all gene positive subjects. These findings suggest that loss of axonal integrity is an early hallmark of neurodegenerative changes which are clinically relevant. © 2014 Elsevier Inc. All rights reserved.

Huntington's disease (HD) is a neurodegenerative disorder characterized by pathology that spreads beyond its well-known targets in the basal ganglia (Cudkowicz and Kowall, 1990; Li and Conforti, 2012). In particular, structural magnetic resonance imaging (MRI) studies have documented atrophy in the caudate and putamen, as well as in wide spread cortical regions (i.e., frontal and parietal lobes) (Dominguez et al., 2013; Georgiou-Karistianis et al., 2013a; Rosas et al., 2002; Rosas et al., 2011; Tabrizi et al., 2009). Importantly, gray matter atrophy has been shown to parallel localized abnormalities in white matter pathways, including striatal projection fibers and corpus callosum (Dumas et al., 2012; Kloppel et al., 2008; Poudel et al., 2014; ⁎ Corresponding author at: School of Psychological Sciences, Monash University, Clayton 3800, Victoria, Australia. Fax: +61 3 9905 3948. E-mail address: [email protected] (N. Georgiou-Karistianis). Available online on ScienceDirect (www.sciencedirect.com).

http://dx.doi.org/10.1016/j.nbd.2014.12.009 0969-9961/© 2014 Elsevier Inc. All rights reserved.

Rosas et al., 2010; Tabrizi et al., 2009). However, these white matter microstructural changes, measured using diffusion tensor imaging (DTI), have only been cross-sectional in design (Rosas et al., 2010; Rosas et al., 2006) and thus provide limited insights on the development and progression of such changes, or how they may relate to functional outcomes. The only previous longitudinal study of white matter microstructure was not sensitive enough to differentiate HD individuals from healthy controls, possibly due to limitations in methodology and very small sample size (N = 7) (Weaver et al., 2009). Hence, a longitudinal investigation is necessary to better characterize changes in white matter microstructure over time, identify factors that may predict these changes, and provide potential new biomarkers to track the effectiveness of potential therapeutics. In this study we investigated whether: i) pre-HD and symp-HD individuals, compared to healthy controls, show 18-month longitudinal change in whole brain white matter fractional anisotropy (FA), radial

G.R. Poudel et al. / Neurobiology of Disease 74 (2015) 406–412

diffusivity (RD) and axial diffusivity (AD); and, ii) baseline clinical symptoms, cognitive and motor functions in pre-HD and symp-HD are predictive of such changes. We applied an unbiased tensor based registration approach (Keihaninejad et al., 2013), followed by a tract-based spatial statistics (TBSS) analysis of whole brain white matter. Based on previous findings (Rosas et al., 2010; Rosas et al., 2006) we hypothesized that both pre-HD and symp-HD groups would show longitudinal change in white matter microstructure in the corpus callosum, cingulum, and striatal projection white matter pathways, and, that baseline clinical symptoms, cognitive and motor functions would be predictive of such changes. Methods Subjects Eighty-six participants (29 pre-HD, 29 symp-HD, and 28 controls), recruited as part of the Australian-based IMAGE-HD study (Dominguez et al., 2013; Georgiou-Karistianis et al., 2013a; Poudel et al., 2014), had longitudinal DTI data from both baseline and 18months. Six participants were removed from the current analyses (four symp-HD, one pre-HD, and one control) (see MRI preprocessing section), leaving 28 pre-HD, 25 symp-HD, and 27 controls for further investigation. Recruitment procedures and inclusion criteria have been described previously (Dominguez et al., 2013; Georgiou-Karistianis et al., 2013a). Controls were matched to preHD for age, gender and estimated premorbid intelligence [National Adult Reading Test 2nd edition, NART-2 (Nelson et al., 1992)]. Symp-HD participants were significantly older than controls. To account for the age difference between symp-HD and controls, statistical results are reported after covarying out the effects of age. All participants were assessed at baseline and after 18 months. Standard protocol approvals, registrations, and patient consents A written informed consent was obtained from each participant in accordance with the Helsinki Declaration. The study was approved by the Monash University and Melbourne Health Human Research Ethics Committee. Diagnostic criteria and assessment CAG repeat length in the expanded allele ranged from 39 to 50 in both HD groups. The Unified Huntington's Disease Rating Scale motor rating scale was administered by a neurologist (AC) for diagnosis and assessment. Similar to Tabrizi et al. (2009), inclusion in the pre-HD group was based on a UHDRS–TMS b5 and for symp-HD N5. The preHD group was split into pre-HDfar and pre-HDclose groups at the median years to onset, which we estimated using the Langbehn method (Langbehn et al., 2004). The mean diagnostic confidence score for symp-HD was 2.9 (±1.24) at baseline and 3.33 (±1.30) at 18 months. For symp-HD, years since diagnosis ranged from 0 to 4 years. A comprehensive neuropsychological testing battery [described previously (Georgiou-Karistianis et al., 2013a; Georgiou-Karistianis et al., 2014; Gray et al., 2013)] was used to assess participants' cognitive, psychiatric, and motor functions. Table 1 summarizes demographics, clinical information, and neuropsychological measures of interest used in this investigation.

407

Table 1 Demographic, clinical and neurocognitive data for participants included in the current investigation. Controls

Pre-HD

Symp-HD

N (sample sizes) Gender (M:F) Age (years) UHDRS–TMS

28 9:19 42 ± 13 –

Disease Burden Score (DBS) Estimated YTO Duration of illness (years) SDMT Stroop word Tapping performance

– – – 57 ± 11 111 ± 17 30 ± 7

27 10:17 41 ± 10 1±1 (0–4) 265 ± 51 16 ± 7 – 53 ± 10 103 ± 19 25 ± 9⁎

25 15:10 54 ± 10⁎⁎⁎+++ 18 ± 10+++ (6–41) 368 ± 73+++ – 2±1 36 ± 12⁎⁎⁎+++ 85 ± 21⁎⁎⁎+++ 11 ± 6⁎⁎⁎+++

SD, standard deviation; UHDRS, total motor score, Unified Huntington's Disease Rating Scale (pre-HD, UHDRS b 5; symp-HD, UHDRS ≥ 5); Disease Burden Score (CAG − 35.5) ⁎ age; YTO, years to onset; SDMT, Symbol Digit Modalities Test (number correct); Stroop word, speeded word reading task (number correct); symp-HD or pre-HD versus controls: ⁎p ≤ 0.05; ⁎⁎p ≤ 0.01; ⁎⁎⁎p ≤ 0.001; symp-HD versus pre-HD: ++p ≤ 0.01; +++p ≤ 0.001.

Flip = 90°, 64 contiguous slices with 2 mm isotropic voxels, acquisition matrix of 128 × 128). Diffusion-sensitizing encoding gradients were applied in 60 directions using a b value of 1200 s/mm2, and 10 images without diffusion weighting (b = 0 s/mm2). MRI pre-processing FMRIB's Diffusion Toolbox (FDT) (http://fsl.fmrib.ox.ac.uk/fsl/ fslwiki/FDT) was used for eddy current correction and elimination of motion artifacts. For the TBSS analysis, we estimated rigid registration parameters between each gradient and baseline image of the original diffusion weighted images and removed participants with large motion spikes. Based on this method six participants were removed from subsequent analyses (four symp-HD, one pre-HD, and one control). This step reduces measurement noise and improve reproducibility of microstructure in white matter tracts. Diffusion tensor images were then generated for each participant using the diffusion tensor imaging toolkit (DTI-TK). An unbiased approach recently described and used (Keihaninejad et al., 2013) was chosen for the registration of the longitudinal DTI data, to develop an initial within-subject template from baseline and 18 months of diffusion tensors, and then to iteratively refine, first with affine and then with non-linear registrations. The within-subject templates were then used to create a study specific group-level template using iterative affine and non-linear registrations (Keihaninejad et al., 2013). The group-level population specific diffusion template was registered to an ICBM-152 space enhanced diffusion tensor template (Zhang et al., 2011) using a sequence of rigid, affine, and non-linear registrations. This method of registration of withinsubject data takes into account possible white matter atrophy over time HD. Furthermore, generation of inter-subject group-wise atlas eliminated the potential for bias in estimation of longitudinal changes. Finally, the composition of the mappings from subject space to within-subject template, within-subject template to population template, and population template to ICBM-152 standard template was used to normalize each subject's diffusion tensor data to the ICBM152 template. For each participant, FA, RD, and AD data were estimated using the normalised tensor images. Tract based spatial statistic analysis

MRI data acquisition Diffusion weighted images were acquired on a Siemens Magnetom Tim Trio 3 Tesla MRI (Siemens AG, Erlangen, Germany) using a 32channel head coil at the Murdoch Children's Research Institute (Royal Children's Hospital, Victoria, Australia). Data were acquired using a double spin echo diffusion weighted EPI sequence (TR = 8200, TE = 89 ms,

FA, RD, and AD data from each participant were further processed and analyzed using the tract based spatial statistics (TBSS) analysis tool available in the FSL toolbox (Smith et al., 2006). A skeleton of white matter was generated by thresholding the mean FA map (FA N 0.2) and representing a single line running down the centers of all the common white matter fibers (Smith et al., 2006). Each

408

G.R. Poudel et al. / Neurobiology of Disease 74 (2015) 406–412

participant's normalized FA image was then projected onto this common skeleton to minimize any residual misalignment of tracts. The skeleton projection was then applied to RD and AD images to create a separate skeleton representing the RD and AD values. To estimate white matter microstructural change over 18 months for each participant, we subtracted baseline data (i.e., FA, RD, and AD skeleton) from 18 month data. The difference images were then used in a general linear model based analysis, which included predictors for four groups (preHDclose, pre-HDfar, all pre-HD, symp-HD, and controls) and three covariates (age, gender, and inter-scan interval). Statistical contrasts were defined for difference in change over time between symp-HD and controls, pre-HDclose and controls, pre-HDfar and controls, and symp-HD and overall pre-HD group. The model was estimated using the “randomize” tool available in FSL, which used non-parametric permutation based inference using 5000 permutations. The results were corrected for multiple comparisons across space using the threshold free cluster enhancement method (p b 0.05). Statistical analysis Statistical analyses of clinical and neurocognitive variables were performed using the MatLab Statistics Toolbox (MathWorks, Natik, MA, USA). Only the white matter regions showing significant changes in FA, RD, or AD over time in symp-HD or pre-HD, compared to controls, were selected for this analysis. White matter regions of interest were delineated using the JHU white matter atlas, by linearly and nonlinearly registering (FSL, FLIRT + FNIRT) average study specific FA map with JHU FA map. Partial correlation analyses [with age and Disease Burden Score (DBS) as covariates] were performed between percent change in FA and baseline UHDRS–TMS in the overall HD group (i.e., pre-HD and symp-HD combined). Disease Burden Score is a function of age and CAG repeat and indicates the pathological load of the mutated HTT protein (Penney et al., 1997). We considered correlations to be significant at p b 0.05 corrected using Bonferroni's method. To investigate whether baseline clinical symptoms, cognitive and motor functions predict the longitudinal change in HD, stepwise linear regressions were performed with percent change in FA as dependent variables against four sets of independent variables: clinical (age, UHDRS–TMS, DBS), cognitive-quantified motor [age, Symbol Digit Modalities Test (SDMT), Stroop word reading, Paced Tapping], and combined (age, UHDRS–TMS, DBS, SDMT, Stroop, Paced Tapping). Separate stepwise regressions were fitted for the clinical, cognitive, and combined models. Results were considered to be significant at p b 0.05. Our previous publications provide a more comprehensive summary of all neurocognitive tests used as part of the IMAGE-HD study (Georgiou-Karistianis et al., 2013a; Georgiou-Karistianis et al., 2013b; Poudel et al., 2013). Results 18 month changes in white matter microstructure The symp-HD group, compared with controls and pre-HD, showed significantly reduced (p b 0.05, TFCE corrected) FA over 18 months in the corpus callosum and middle cingulum white matter (Fig. 1). The symp-HD group, compared to pre-HD, showed significantly increased (p b 0.05, TFCE corrected) RD over 18 months in the corpus callosum, cingulum, internal capsule, and striatal projection white matter (Fig. 2). Fig. 3 shows annualized percent change in FA of the genu, body, and splenium of the corpus callosum, and cingulum white matter regions, which showed significant longitudinal change in symp-HD compared to controls. On average, the symp-HD group showed an annual FA reduction of 1.5% in all parts of the corpus callosum, and a reduction of ~ 3.5% in the cingulum. A post-hoc t-test revealed reduced FA in preHDclose (p b 0.03), compared to controls, only in the genu of the corpus callosum.

Relationship between white matter microstructure and clinical symptoms For a partial correlation analysis with UHDRS–TMS, we selected the genu, body, and splenium of the corpus callosum and cingulum white matter, as they showed a significant change in symp-HD. Scatterplots of partial correlations are shown in Fig. 4. Significant negative correlations were observed between baseline UHDRS–TMS and 18 month change in FA in the genu (r = 0.41, pcorr = 0.01), body (r = − 0.50, pcorr b 0.001) and splenium (r = − 0.44, pcorr = 0.005) (Fig. 3) of the corpus callosum in the combined pre-HD and sympHD groups. No significant correlation was observed with change in FA in the cingulum white matter. Baseline predictors of longitudinal white matter microstructural change For the stepwise regression analysis, we selected change in FA in the genu, body, and splenium of the corpus callosum and cingulum white matter. For change in FA in the genu of the corpus callosum, only the UHDRS–TMS at baseline was an independent predictor in the clinical model (adjusted R2 = 0.22, p b 0.001) and in the combined model (R2 = 0.23, p b 0.001). On using the cognitive–quantitative motor model, Paced Tapping performance was the only independent predictor (adjusted R2 = 0.12, p = 0.008). For the change in FA in the body of the corpus callosum, only the UHDRS–TMS at baseline was an independent predictor in the clinical model (adjusted R2 = 0.36, p b 0.001) and in the combined model (R2 = 0.27, p b 0.001). Paced Tapping was the only independent predictor (adjusted R2 = 0.28, p b 0.001) in the cognitive-quantitative motor model. For the change in FA in the splenium of the corpus callosum, UHDRS– TMS at baseline was the only independent predictor in the clinical model (adjusted R2 = 0.22, p b 0.001) and in the combined model (R2 = 0.24, p b 0.001). Paced Tapping was the only independent predictor (adjusted R2 = 0.13, p = 0.006) in the cognitive–quantitative motor model. For the change in FA in the cingulum, Paced Tapping was the only independent predictor in the cognitive–quantitative motor model (adjusted R2 = 0.10, p = 0.013) and in the combined model (R2 = 0.10, p = 0.013). Discussion This study characterized 18-month changes in white matter microstructure in pre-HD and symp-HD individuals. Symp-HD, compared to controls and pre-HD, showed decreased FA over 18-months in the corpus callosum and middle cingulum white matter. Symp-HD also showed 18-month increased RD in widespread white matter, but this change was only significant compared to the change in pre-HD. Change in FA in all parts of the corpus callosum (genu, body, and splenium) was linearly associated with baseline UHDRS–TMS across both HD groups. UHDRS–TMS was also a significant independent predictor of longitudinal FA change in the body, genu, and splenium of corpus callosum. Baseline tapping performance across both HD groups was an independent predictor of the longitudinal decline in middle cingulum FA. These findings provide the first direct evidence for clinically relevant progressive decline in white matter microstructure in HD. Longitudinal change in white matter FA and RD in symp-HD supports our hypothesis of progressive decline of white matter microstructure in HD. To our knowledge, this is the first evidence of longitudinal change in white matter microstructure in symp-HD, a characteristic that significantly differentiated symp-HD from both pre-HD and control groups. A previously published longitudinal study of white matter microstructure in HD reported only within-group reduction in similar white matter areas in a small group of n = 7 (Weaver et al., 2009). Our findings are also consistent with previous cross-sectional studies demonstrating compromised white matter microstructural integrity of the corpus callosum and cingulum in HD (Dumas et al., 2012; Poudel

G.R. Poudel et al. / Neurobiology of Disease 74 (2015) 406–412

409

Fig. 1. Longitudinal change in fractional anisotropy in symptomatic Huntington's disease relative to healthy controls and premanifest HD. The study-specific FA skeleton, representing the centers of principal white matter tracts, is displayed in green, overlaid on a study specific averaged fractional anisotropy map. Symp-HD showed significant reduction in FA (p b 0.05, TFCE corrected) compared to controls (A) and pre-HD (B) in the corpus callosum and cingulum white matter. Blue indicates significant reduction. The horizontal line on the sagittal view indicates the axial slices displayed. The vertical line on the coronal view indicates the sagittal slices displayed. The colorbar represents 1-p values. FA = fractional anisotropy, symp-HD = symptomatic Huntington's disease, pre-HD = premanifest Huntington's disease. Slice location in ICBM MNI template is indicated on the top of each slice. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

et al., 2014; Rosas et al., 2010). Some changes in white matter microstructure have also been reported in pre-HD as early as 10 years prior to symptom onset (Rosas et al., 2010). Although the whole-brain TBSS analysis in the present study was not sensitive to longitudinal FA or RD change in pre-HD, a post-hoc comparison between pre-HDclose and controls showed that FA change in the genu of corpus callosum may be sensitive to longitudinal progression as individuals approach onset. In symp-HD, the annual decrease of 1.5% in FA in all sections of the corpus callosum suggests that microstructural decline in this brain region accelerates after symptom onset. In the symp-HD group, 18-month reduction in FA was also observed in the middle cingulum white matter. The cingulum, a collection of white matter tracts that interconnects the parietal, frontal, and motor cortices, is known to be prone to early degeneration in HD (Stoffers et al., 2010). In particular, the middle segment of cingulum, which shows changes in the current study, contains connections with

premotor and motor cortical regions. Since motor and pre-motor cortical gray matter may represent the first of the cortical structures to show neurodegeneration in HD (Rosas et al., 2002), 18-month reduction in the structural integrity of the white matter connecting these regions is not surprising. However, because measurable longitudinal changes in FA were observed only in symp-HD, and not in pre-HD, this suggests that longitudinal microstructural changes reflective of myelin damage may be sensitive only during the manifest stages of the disease. This finding is somewhat surprising since previous large-scale studies in pre-HD (Aylward et al., 2012; Tabrizi et al., 2013) have demonstrated longitudinal atrophy of white matter. However, it is important to note that microstructural changes during the pre-HD stage may be dynamic, depending on a number of factors including how far along the disease continuum the individual is in terms of estimated years to onset. It is also possible that the diffusivity changes during the premanifest stages are more variable and perhaps more dynamic over time compared with

410

G.R. Poudel et al. / Neurobiology of Disease 74 (2015) 406–412

Fig. 2. Longitudinal change in radial diffusivity in symptomatic Huntington's disease relative to healthy controls and premanifest HD. The study-specific FA skeleton, representing the centers of principal white matter tracts, is displayed in green, overlaid on a study specific averaged fractional anisotropy map. (A) Symp-HD showed no significant change in RD (p b 0.05, TFCE corrected) when compared to controls. (B) Symp-HD showed increase in RD (p b 0.05, TFCE corrected) when compared to pre-HD in widespread white matter pathways. The colorbar represents 1-p values. Red indicates significant increase. RD = radial diffusivity, symp-HD = symptomatic Huntington's disease, pre-HD = premanifest Huntington's disease. Slice location in ICBM MNI template is indicated on the top of each slice. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3. Annualized percent change in FA of the corpus callosum and cingulum in pre-HD, symp-HD, and controls. Significant difference observed in whole-brain TBSS analysis is denoted by * and the significant difference observed in post-hoc exploratory analysis is denoted by ^.

G.R. Poudel et al. / Neurobiology of Disease 74 (2015) 406–412

411

Fig. 4. Partial correlation scatterplot showing relationship between tract-specific FA changes and baseline UHDRS–TMS in HD. Residual UHDRS and Residual FA change (after correcting for age and DBS) are used in the scatterplot to show linear relationship between UHDRS and average change in FA in the genu (A), body (B), splenium (C), and cingulum (D). The r values represent partial correlation coefficient. All correlations were significant with p b 0.05 (corrected).

the diffusivity changes during the early manifest stages. Interestingly, we observed trends towards reduced RD over time in pre-HD (see Supplementary Table 1), which may be reflective of compensatory remyelination (Song et al., 2005). This may in-part explain the lack of a significant reduction in FA in the pre-HD group, which may be associated with myelin damage (See Supplementary Table 2). The clinical relevance of the longitudinal findings is highlighted by the significant association between corpus callosum FA change and baseline UHDRS–TMS. This finding was further strengthened by stepwise regression modeling, which revealed UHDRS–TMS to be the most predictive marker of longitudinal decline in all parts of the corpus callosum. The UHDRS–TMS measures motor function and taps into symptomatology that is critical in HD diagnosis. The association between UHDRS and longitudinal change in corpus callosum microstructure may therefore reflect its sensitivity to disruption in interhemispheric structural connectivity in HD. The ensuing rapid loss of inter-hemispheric communication, due to deterioration of corpus callosum microstructure, may account for some of the symptoms evident in symptomatic HD that are not entirely explained by corticostriatal dysfunction. Although the specific cellular mechanism underlying changes in DTI measures remains unclear, our findings of decreased FA and increased RD over time in symp-HD may be explained by either loss of axonal fibers in HD, or HD-related demyelination, or both. In support of the first possibility, studies in animal models of HD have shown axonal degeneration associated with huntingtin aggregates. Postmortem studies in HD have also found decreases in pyramidal neurons and their axonal connections (Cudkowicz and Kowall, 1990). Support for the second

explanation comes from post-mortem studies of pre-HD brains, which show significantly elevated densities of oligodendrocytes, perhaps due to a physiological attempt to repair and remyelinate (Myers et al., 1991). DTI studies have also shown that myelin pathology is reflected by an increase in RD, and not AD (Song et al., 2005). In the current study, decreased FA in the corpus callosum and cingulum white matter, accompanied by increased RD but no change in AD, is suggestive of progressive myelin pathology in HD. In this study we used an unbiased tensor-based registration framework for investigating changes in whole brain white matter microstructure over time. Previous studies have primarily relied on scalar maps (e.g., FA) derived from the diffusion tensor to perform independent registrations to a standard template (Weaver et al., 2009), and have generally failed to differentiate longitudinal changes in HD from controls (Sritharan et al., 2010; Weaver et al., 2009). Registration of DTI data using full tensor information improves both specificity and precision for detecting longitudinal changes compared to standard scalar approaches (Keihaninejad et al., 2013). Registration of within-subject data in this study, using a non-linear method, also takes possible white matter atrophy over time into account, which is consistently reported in HD (Dominguez et al., 2013; Poudel et al., 2014; Tabrizi et al., 2013). Furthermore, generation of an inter-subject group-wise atlas eliminated the potential for bias in estimation of longitudinal changes. Hence, the significant findings from the current study represent biomarkers that are unbiased and are sensitive to neuropathology in HD. In conclusion, this study has shown that symp-HD individuals are particularly vulnerable to a rapid change in white matter

412

G.R. Poudel et al. / Neurobiology of Disease 74 (2015) 406–412

microstructure, which can be observed over periods as short as 18 months. Although no significant change in microstructure was detected during the pre-HD stage, baseline clinical symptoms (UHDRS–TMS) and Paced Tapping were predictive of change in white matter microstructure in overall pre-HD and symp-HD over 18 months. Thus, change in white matter microstructure may be a candidate biomarker for consideration in future clinical trials. Author contribution Dr Govinda R. Poudel conceived the study, analyzed the data, wrote the article, reviewed the article, and approved the final version for publication. He has received funding from the CHDI Foundation Inc. (grant number A-3433) and, the National Health and Medical Research Council (NHMRC) (grant number 606650). Dr. Poudel reports no disclosures. Prof. Julie C Stout conceived the study, wrote the article, reviewed the article, and approved the final version for publication. She has received funding from the CHDI Foundation Inc. (grant number A-3433) and, the National Health and Medical Research Council (NHMRC) (grant number 606650). Prof. Stout reports no disclosures. Dr Juan F. Domínguez D collected the data, wrote the article, reviewed the article, and approved the final version for publication. He has received funding from the CHDI Foundation Inc. (grant number A-3433) and, the National Health and Medical Research Council (NHMRC) (grant number 606650). Dr Domínguez D reports no disclosures. Dr Andrew Churchyard conceived the study, collected the data, reviewed the article, and approved the final version for publication. He has received funding from the CHDI Foundation Inc. (grant number A-3433) and, the National Health and Medical Research Council (NHMRC) (grant number 606650). Dr Churchyard reports no disclosures. Prof. Gary F. Egan conceived the study, wrote the article, reviewed the article, and approved the final version for publication. He has received funding from the CHDI Foundation Inc. (grant number A-3433) and, the National Health and Medical Research Council (NHMRC) (grant number 606650). Prof Egan reports no disclosures. Dr Phyllis Chua conceived the study, collected the data, reviewed the article, and approved the final version for publication. She has received funding from the CHDI Foundation Inc. (grant number A-3433) and, the National Health and Medical Research Council (NHMRC) (grant number 606650). Dr Chua reports no disclosures. Prof. Nellie Georgiou-Karistianis conceived the study, analyzed the data, wrote the article, reviewed the article, and approved the final version for publication. She has received funding from the CHDI Foundation Inc. (grant number A-3433) and, the National Health and Medical Research Council (NHMRC) (grant number 606650). Prof GeorgiouKaristianis reports no disclosures. Study funding CHDI Foundation Inc. (grant number A-3433) and, National Health and Medical Research Council (NHMRC) (grant number 606650). Acknowledgments We would like to acknowledge the contribution of the participants who took part in this study. We are also grateful to the CHDI Foundation Inc. (grant number A-3433), New York (USA), and to the National Health and Medical Research Council (NHMRC) (grant number 606650) for their support in funding this research. This research was supported by the VLSCI's Life Sciences Computation Centre, a collaboration between Melbourne, Monash and La Trobe Universities and an

initiative of the Victorian Government, Australia. We also thank the Royal Children's Hospital for the use of their 3 T MR scanner. GFE is a Principal NHMRC Research Fellow.

Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.nbd.2014.12.009.

References Aylward, E.H., et al., 2012. Striatal volume contributes to the prediction of onset of Huntington disease in incident cases. Biol. Psychiatry 71, 822–828. Cudkowicz, M., Kowall, N.W., 1990. Degeneration of pyramidal projection neurons in Huntington's disease cortex. Ann. Neurol. 27, 200–204. Dominguez, D.J., et al., 2013. Multi-modal neuroimaging in premanifest and early Huntington's disease: 18 month longitudinal data from the IMAGE-HD study. PLoS One 8, e74131. Dumas, E.M., et al., 2012. Early changes in white matter pathways of the sensorimotor cortex in premanifest Huntington's disease. Hum. Brain Mapp. 33, 203–212. Georgiou-Karistianis, N., et al., 2013a. Automated differentiation of pre-diagnosis Huntington's disease from healthy control individuals based on quadratic discriminant analysis of the basal ganglia: the IMAGE-HD study. Neurobiol. Dis. 51, 82–92. Georgiou-Karistianis, N., et al., 2013b. Functional and connectivity changes during working memory in Huntington's disease: 18 month longitudinal data from the IMAGEHD study. Brain Cogn. 83, 80–91. Georgiou-Karistianis, N., et al., 2014. Functional magnetic resonance imaging of working memory in Huntington's disease: IMAGE-HD cross-sectional analysis. Hum. Brain Mapp. 35, 1847–1864. Gray, M.A., et al., 2013. Prefrontal activity in Huntington's disease reflects cognitive and neuropsychiatric disturbances: the IMAGE-HD study. Exp. Neurol. 239, 218–228. Keihaninejad, S., et al., 2013. An unbiased longitudinal analysis framework for tracking white matter changes using diffusion tensor imaging with application to Alzheimer's disease. Neuroimage 72, 153–163. Kloppel, S., et al., 2008. White matter connections reflect changes in voluntary-guided saccades in pre-symptomatic Huntington's disease. Brain 131, 196–204. Langbehn, D.R., et al., 2004. A new model for prediction of the age of onset and penetrance for Huntington's disease based on CAG length. Clin. Genet. 65, 267–277. Li, J.Y., Conforti, L., 2012. Axonopathy in Huntington's Disease. Exp Neurol. Myers, R.H., et al., 1991. Factors associated with slow progression in Huntington's disease. Arch. Neurol. 48, 800–804. Nelson, H.E., et al., 1992. National adult reading test, 2nd edition. Int. J. Geriatr. Psychiatry 7, 533. Penney Jr., J.B., et al., 1997. CAG repeat number governs the development rate of pathology in Huntington's disease. Ann. Neurol. 41, 689–692. Poudel, G.R., et al., 2013. Functional Changes During Working Memory in Huntington's Disease: 30-month Longitudinal Data From the IMAGE-HD Study. Brain Struct Funct. Poudel, G.R., et al., 2014. White matter connectivity reflects clinical and cognitive status in Huntington's disease. Neurobiol. Dis. 65, 180–187. Rosas, H.D., et al., 2002. Regional and progressive thinning of the cortical ribbon in Huntington's disease. Neurology 58, 695–701. Rosas, H.D., et al., 2006. Diffusion tensor imaging in presymptomatic and early Huntington's disease: selective white matter pathology and its relationship to clinical measures. Mov. Disord. 21, 1317–1325. Rosas, H.D., et al., 2010. Altered white matter microstructure in the corpus callosum in Huntington's disease: implications for cortical “disconnection”. Neuroimage 49, 2995–3004. Rosas, H.D., et al., 2011. A tale of two factors: what determines the rate of progression in Huntington's disease? A longitudinal MRI study. Mov. Disord. 26, 1691–1697. Smith, S.M., et al., 2006. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 31, 1487–1505. Song, S.K., et al., 2005. Demyelination increases radial diffusivity in corpus callosum of mouse brain. Neuroimage 26, 132–140. Sritharan, A., et al., 2010. A longitudinal diffusion tensor imaging study in symptomatic Huntington's disease. J. Neurol. Neurosurg. Psychiatry 81, 257–262. Stoffers, D., et al., 2010. Contrasting gray and white matter changes in preclinical Huntington disease: an MRI study. Neurology 74, 1208–1216. Tabrizi, S.J., et al., 2009. Biological and clinical manifestations of Huntington's disease in the longitudinal TRACK-HD study: cross-sectional analysis of baseline data. Lancet Neurol. 8, 791–801. Tabrizi, S.J., et al., 2013. Predictors of phenotypic progression and disease onset in premanifest and early-stage Huntington's disease in the TRACK-HD study: analysis of 36-month observational data. Lancet Neurol. 12, 637–649. Weaver, K.E., et al., 2009. Longitudinal diffusion tensor imaging in Huntington's disease. Exp. Neurol. 216, 525–529. Zhang, S., et al., 2011. Enhanced ICBM diffusion tensor template of the human brain. Neuroimage 54, 974–984.

Longitudinal change in white matter microstructure in Huntington's disease: The IMAGE-HD study.

To quantify 18-month changes in white matter microstructure in premanifest (pre-HD) and symptomatic Huntington's disease (symp-HD). To investigate bas...
2MB Sizes 2 Downloads 10 Views