- Email: [email protected]
White matter connectivity reflects clinical and cognitive status in Huntington's disease
Neurobiology of Disease 65 (2014) 180–187
Contents lists available at ScienceDirect
Neurobiology of Disease journal homepage: www.elsevier.com/locate/ynbdi
White matter connectivity reflects clinical and cognitive status in Huntington's disease☆,☆☆,★ Govinda R. Poudel a,b,e, Julie C. Stout a, Juan F. Domínguez D a, Louisa Salmon a, Andrew Churchyard c, Phyllis Chua a, Nellie Georgiou-Karistianis a,⁎, Gary F. Egan a,b,d,e a
School of Psychological Sciences, Monash University, Clayton, Victoria, Australia Monash Biomedical Imaging (MBI), Monash University, Melbourne, VIC, Australia Department of Neurology, Monash Medical Centre, Clayton, Victoria, Australia d Centre for Neuroscience, University of Melbourne, Parkville, Victoria, Australia e VLSCI Life Sciences Computation Centre, Melbourne, Victoria, Australia b c
a r t i c l e
i n f o
Article history: Received 25 September 2013 Revised 18 December 2013 Accepted 19 January 2014 Available online 28 January 2014 Keywords: fMRI Huntington's disease Connectivity Tractography Diffusion tensor imaging
a b s t r a c t Objective: To investigate structural connectivity and the relationship between axonal microstructure and clinical, cognitive, and motor functions in premanifest (pre-HD) and symptomatic (symp-HD) Huntington's disease. Method: Diffusion tensor imaging (DTI) data were acquired from 35 pre-HD, 36 symp-HD, and 35 controls. Structural connectivity was mapped between 40 brain regions of interest using tractography. Between-group differences in structural connectivity were identified using network based statistics. Radial diffusivity (RD) and fractional anisotropy (FA) were compared in the white matter tracts from aberrant networks. RD values in aberrant tracts were correlated with clinical severity, and cognitive and motor performance. Results: A network connecting putamen with prefrontal and motor cortex demonstrated significantly reduced tractography streamlines in pre-HD. Symp-HD individuals showed reduced streamlines in a network connecting prefrontal, motor, and parietal cortices with both caudate and putamen. The symp-HD group, compared to controls and pre-HD, showed both increased RD and decreased FA in the fronto-parietal and caudate-paracentral tracts and increased RD in the putamen-prefrontal and putamen-motor tracts. The pre-HDclose, compared to controls, showed increased RD in the putamen-prefrontal and fronto-parietal tracts. In the pre-HD group, significant negative correlations were observed between SDMT and Stroop performance and RD in the bilateral putamen-prefrontal tract. In the symp-HD group, RD in the fronto-parietal tract was significantly positively correlated with UHDRS motor scores and significantly negatively correlated with performance on SDMT and Stroop tasks. Conclusions: We have provided evidence of aberrant connectivity and microstructural integrity in white matter networks in HD. Microstructural changes in the cortico-striatal fibers were associated with cognitive and motor performance in pre-HD, suggesting that changes in axonal integrity provide an early marker for clinically relevant impairment in HD. © 2014 Elsevier Inc. All rights reserved.
Introduction Neuropathological processes in Huntington's disease (HD) primarily target medium spiny neurons of the striatum (Graveland
et al., 1985). Neurodegeneration is also seen in pyramidal projection neurons in the motor and prefrontal cortices, and cingulate and angular gyri (Macdonald and Halliday, 2002; Thu et al., 2010). Together, these neurodegenerative changes are considered to be the cause of onset of
☆ Statistical analysis was performed by Dr Govinda R. Poudel School of Psychological Sciences; Monash Biomedical Imaging (MBI), Monash University, VIC, Australia. ☆☆ Author contributions: Govinda Poudel was involved in drafting/revising the manuscript for content; study concept or design; analysis or interpretation of data and statistical analysis. Julie C. Stout was involved in drafting/revising the manuscript for content; study concept or design; analysis or interpretation of data and obtainment of funding. Juan F. Domínguez D was involved in drafting/revising the manuscript for content; analysis or interpretation of data and acquisition of data. Louisa Salmon was involved in drafting/revising the manuscript for content; analysis or interpretation of data and acquisition of data. Andrew Churchyard was involved in the study concept or design; obtainment of funding and acquisition of data. Phyllis Chua was involved in the study concept or design; obtainment of funding and acquisition of data. Nellie Georgiou-Karistianis was involved in drafting/revising the manuscript for content; study concept or design; analysis or interpretation of data; obtainment of funding and study supervision or coordination. Gary F. Egan was involved in drafting/revising the manuscript for content; study concept or design; analysis or interpretation of data; obtainment of funding and study supervision or coordination. ★ Author Disclosures: All authors have no relevant biomedical, financial or potential conflicts of interest to declare. ⁎ 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). 0969-9961/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.nbd.2014.01.013
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
clinical symptoms in HD (Thu et al., 2010). In contrast, the mechanisms underlying cognitive changes, which often appear several years before clinical diagnosis, remain largely unknown. Disruption of structural connectivity in specific neural circuits has been proposed to be one of the possible mechanisms leading to early cognitive and motor changes in premanifest Huntington's disease (pre-HD) (Li and Conforti, 2013). Structural connectivity can deteriorate in HD due to axonal dysfunction and degeneration associated with huntingtin aggregates which can appear early in HD (reviewed by Li and Conforti, 2013). Hence, white matter atrophy is evident in T1-weighted neuroimaging studies of HD (Tabrizi et al., 2009; Thieben et al., 2002), with posterior-frontal whitematter degeneration apparent even in individuals far from onset (Tabrizi et al., 2009). Diffusion tensor imaging (DTI) studies of HD have also suggested selective microstructural changes in white matter encompassing cortico-striatal motor circuit, corpus callosum, periventricular region, corona radiata, and prefrontal cortex (Bohanna et al., 2011; Dumas et al., 2012; Rosas et al., 2006, 2010; Weaver et al., 2009). DTI also enables investigation of axonal fibers between gray matter structures using tractography methods (Bohanna et al., 2011; Jones, 2008). DTI tractography has been used in HD to isolate structural connections in specific neuroanatomical circuits including the motor loop and fronto-striatal circuit (Bohanna et al., 2011; Dumas et al., 2012; Kloppel et al., 2008), providing evidence for circuit specific alterations in white matter microstructure in HD. For example, microstructural damage in the striatal nodes of the motor loop has been shown to be associated with motor dysfunction in HD (Bohanna et al., 2011). Structural connectivity changes of the fronto-caudal tracts have been shown not only to reflect years to onset, but also to be associated with oculomotor function in pre-HD (Kloppel et al., 2008). Moreover, reduced fiber connectivity between the prefrontal cortex and the caudate has been shown to reflect symptomatology in pre-HD (Kloppel et al., 2008). DTI-tractography has also been used to determine pairwise connections between gray matter structures in the brain enabling calculation of a structural connectivity matrix for individual subjects (Zalesky et al., 2010). Network-based statistical methods can be used to isolate network connections that are altered in disease from the connectivity matrix (Zalesky et al., 2010, 2011). The identification of structural networks in pre-HD and symp-HD may provide insight into early markers of disease progression in HD. The aims of the current study were to identify cortico-striatal networks affected in pre-HD and symp-HD, determine the microstructural alterations in the axonal fibers connecting these pathways, and investigate the relationship between axonal microstructural changes and clinical, cognitive and motor functions in pre-HD and symp-HD. We hypothesized that structural connectivity in neural circuits connecting motor and prefrontal cortices with the caudate and putamen would be affected in both pre-HD and symp-HD. Microstructural white matter degeneration in symp-HD has been reported in the body of the corpus callosum, which structurally connects frontal and parietal areas (Rosas et al., 2010). We hypothesized that symp-HD individuals would show further white matter structural disconnectivity in the fronto-parietal network. Fronto-striatal neural circuits are crucial for cognitive control (Liston et al., 2006). We hypothesized that the microstructural integrity of the fronto-striatal tracts would be associated with cognitive dysfunction in both pre-HD and symp-HD. To test these hypotheses, tractography was used to identify the extent of axonal connectivity between 40 neocortico and striatal brain regions. A network-based statistical method (Zalesky et al., 2010) was used to isolate neuroanatomical networks that showed connectivity differences between the groups. DTI-based measures of radial diffusivity (RD) are thought to be sensitive to demyelinative processes (Song et al., 2005). We measured RD values from the tracts identified as aberrant in pre-HD and symp-HD, and investigated the relationship of RD changes with clinical severity, and cognitive and motor performance.
181
Methods Participants Thirty-five pre-HD, 36 symp-HD, and 35 healthy control volunteers were included in this investigation, all recruited as part of the Australian-based IMAGE-HD study (Georgiou-Karistianis et al., 2013, in press; Gray et al., 2013). Recruitment procedures and inclusion criteria have been published previously (Georgiou-Karistianis et al., 2013). Controls were matched to pre-HD participants for age, gender and IQ [National Adult Reading Test 2nd edition, NART-2 (Nelson et al., 1992)], an estimate of premorbid intelligence. One-way ANOVAs revealed no significant differences in age or IQ scores between the pre-HD group and controls, but significant differences in age between controls and symp-HD, and between pre-HD and symp-HD, respectively (p b 0.05). CAG repeat lengths in the expanded alleles of the participants ranged from 39 to 50 (42 ± 2 for pre-HD; 43 ± 2 for symp-HD). Similar to Tabrizi et al. (2009), inclusion in the pre-HD group was based on UHDRS total motor score ≤ 5. The average years to clinical onset for the pre-HD group was 15 ± 8 years, as determined by a formula based on age, and the number of CAG repeats (Langbehn et al., 2004). The average disease burden score (DBS; Penney et al., 1997) was 270 ± 53. Years since the diagnosis of symptom onset (ascertained by the clinician A.C.) in the symp-HD group ranged from 0 to 5 years with a mean DBS of 379 ± 70 for the group. Demographics, clinical information, and neurocognitive measures of interest are provided in Table 1. The complete battery of neurocognitive data collected for all participants as part of the IMAGE-HD study has been described in previous publications (Georgiou-Karistianis et al., 2013, in press; Gray et al., 2013). The IMAGE-HD study was approved by the Monash University and Melbourne Health Human Research Ethics Committees. Written informed consent was obtained from each participant in accordance with the Helsinki Declaration. MRI data acquisition Structural and functional MR images were acquired with the Siemens Magnetom Tim Trio 3 T MRI scanner (Siemens AG, Erlangen, Germany) and a 32-channel head coil at the Murdoch Children's Research Institute (Royal Children's Hospital, Victoria, Australia). High-resolution T1-weighted images were acquired (192 slices, slice thickness of 0.9 mm, 0.8 mm 0.8 mm in-plane resolution, 320 × 320 matrix, TI = 900 ms, TE = 2.59 ms, TR = 1900 ms, flip angle = 9°). Diffusion weighted data were acquired using a double spin echo diffusion weighted EPI sequence (TR = 8200, TE = 89 ms, flip = 90°, 64 contiguous slices with 2 mm isotropic voxels, acquisition matrix 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) were acquired. MRI data 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. Diffusion tensors were estimated and the reconstructed tensor matrices were diagonalized to obtain eigen values (λ1, λ2, λ3) for estimation of the radial diffusivity (RD) and fractional anisotropy (FA) of each voxel. For spatial normalization of DTI data to a common standard space, an iterative normalization procedure was used (Muller et al., 2011). The FA image from each subject was first normalized to the Montreal Neurological Institute (MNI) 1 mm3 resolution FA Map using non-linear registration (FNIRT). A study specific FA template was created by arithmetically averaging the normalized FA maps for participants from each group.
182
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
Table 1 Demographic, clinical and neurocognitive data for participants included in the current investigation. Mean ± SD
N (sample sizes) Gender (M:F) Age (years) UHDRS CAG repeats Disease burden score (DBS) Estimated YtO Duration of illness (years) SDMT Stroop Word Speeded tapping (ITI, ms)
Controls
Pre-HD
Symp-HD
35 11:24 42 ± 14 (24–73) – – –
35 14:21 42 ± 10 (24–65) 1 ± 1 (0–4) 42 ± 2 270 ± 53
36 21:15 52 ± 9⁎⁎++ (37–71) 19 ± 12+++ (6–60) 43 ± 2 379 ± 70+++
– – 57 ± 10 109 ± 16 221 ± 38
15 ± 8 – 51 ± 9⁎
– 2.0 ± 1.5 36 ± 12⁎⁎⁎++ 83 ± 21⁎⁎ 367 ± 164⁎⁎⁎+++
104 ± 17 245 ± 45⁎
SD, standard deviation; UHDRS, motor subscale score, Unified Huntington's Disease Rating Scale (pre-HD, UHDRS b 5; symp-HD, UHDRS ≥ 5); CAG, cytosine–adenine–guanine (number of repeats N40 is full penetrance); disease burden score (CAG-35.5) ∗ age; YTO, years to onset; SDMT, symbol digit modalities test; Stroop Word, Stroop speeded word reading task (number of correct words); ITI, inter-tap interval. 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.
Non-linear normalization of subject-specific FA volumes to the study specific template was iterated until the correlation was N 0.8 between the individual FA-maps and the FA-template. The non-linear transformation matrices resulting from the registration process were saved for later use in tractography registration.
DTI-based tractography Deterministic tractography was carried in each subject's native space using the Diffusion toolkit 0.6.0 (http://trackvis.org/). The interpolated streamline tracking method (Conturo et al., 1999) with fixed increments of 1 mm was used to generate streamlines which followed the direction of the principal eigenvector. Tracking propagation was terminated when a minimum angle of the principal eigenvector direction between adjacent voxels exceeded a threshold of 35° or when the streamlines exceeded the participant's brain mask. Brain masks were generated for each participant by identifying upper and lower thresholds corresponding to minima in the histogram of the diffusion-weighted volumes. Streamlines were initiated from the center of each voxel and streamlines shorter than 10 mm were discarded. For each subject, the tractography generated tens of thousands of streamlines representing structural networks in the cerebral cortex.
Statistical analysis Network Based Statistics (NBS) (Zalesky et al., 2010) was used to non-parametrically compare the group connectivity matrices (Zalesky et al., 2010). NBS has been validated and previously used to map aberrant structural networks in psychiatric and neurological conditions (Bai et al., 2012; Zalesky et al., 2011). A general linear model was used to model group differences with age differences covaried as microstructural changes have been reported in fronto-parietal white matter tracts (Davis et al., 2009). Permutation based unpaired t-tests using 5000 permutations were used to identify networks which differed between pre-HD and controls and symp-HD and controls. NBS identified anatomical networks which show significant differences in structural connectivity between groups (t N 2.6) using a networklevel family-wise-error correction method (p b 0.05). For networks that showed differences for the cortico-striatal brain region pairs between pre-HD and controls, the fiber tracts were visualized using trackvis software. Group-level masks of the cortico-striatal tracts of interest were generated by averaging the tracts from all participants (pre-HD, symp-HD and controls). For each subject, RD and FA values were calculated for the tracts of interest using group-level masks as a measure of microstructural integrity of axonal fiber tracts. For comparison of tract-specific RD between groups, we further divided the pre-HD group into two sub-groups: pre-HD far from onset (pre-HDfar) and pre-HD close to onset (pre-HD close ). This subgrouping was achieved using a median split of the pre-HD group according to their years to onset. Estimated years to onset for preHD far (n = 17) was 21 ± 6 years and for pre-HD close (n = 18) was 10 ± 3 years. Independent t-tests were conducted to compare RD and FA values of specific tracts between pre-HD far and preHDclose, pre-HDfar and controls, pre-HDclose and controls, overall pre-HD group and controls, and symp-HD and controls. Significant differences corrected for family-wise error rate using Holm–Bonferroni method (p b 0.05) are reported. To investigate the relationship between microstructural changes and clinical severity and cognitive and motor performance, partial correlation analyses (with age as a covariate) were performed using average RD in the tracts of interest. We selected SDMT (symbol digit modalities test), Stroop Word, and speeded tapping tasks from our neurocognitive battery as they have been shown to be sensitive to clinically relevant impairment in HD (Hart et al., 2011; Rowe et al., 2010; Tabrizi et al., 2009). SDMT measures visual attention and information processing speed and Stroop measures cognitive control, allowing us to tap into aspects of executive functions dependent on the integrity of fronto-striatal circuits (Liston et al., 2006). Speeded tapping performance (i.e., inter-tapping interval) provided a measure of motor function, which is dependent on integrity of motor loop. For the symp-HD participants, UHDRS motor scores were used as a measure of clinical severity and were correlated with RD values from the tracts disrupted in symp-HD. Significant correlations corrected for family-wise error rate using Holm–Bonferroni method (p b 0.05) are reported.
Structural network mapping Results To isolate neural circuits connecting frontal and parietal cortices with striatum, we generated a structural connectivity matrix for 40 regions of interest in the bilateral frontal and parietal cortices, striatum, and thalamus (Supplementary Fig. 1a). Brain regions of interest were defined using an anatomical labeling (AAL) atlas (Tzourio-Mazoyer et al., 2002). The AAL atlas was transformed from MNI space to each of the subject's DTI native space by applying the previously generated non-linear transformation matrices, and the discrete anatomical labels were preserved using a nearest neighbor interpolation method. For each subject, a 40 × 40 connectivity matrix representing the number of streamlines between pairs of brain regions was calculated so that each cell (i, j) contained the number of fiber streamlines passing through the brain regions i and j (Supplementary Fig. 1c).
A total of (17 ± 1) × 104 streamlines were generated on average for each participant. There were no significant differences in total number of streamlines generated between groups (pre-HD: (17 ± 1) × 104; symp-HD: (17 ± 2) × 104; controls: (17 ± 2) × 104; pre-HD versus controls: t = 1.6, p = 0.12; symp-HD versus controls: t = 1.4, p = 0.17). Network-level differences in structural connectivity Compared with controls, the pre-HD group showed significantly fewer streamlines in right and left hemispheric networks connecting the putamen with prefrontal and primary motor cortices (Figs. 1a, 2a, b). The symp-HD group, compared with controls, showed significantly
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
fewer streamlines in a wide-spread network connecting the frontal and parietal cortices, and the caudate and putamen with prefrontal and parietal cortices (Figs. 1b, 2a,b,c,d). There were no networks that showed significantly more streamlines in symp-HD or pre-HD compared to controls. Tract-specific microstructural differences Compared with controls, both the pre-HD close and symp-HD groups showed significantly greater RD in the bilateral putamenprefrontal tract (pre-HD close versus controls: t = 4.9, p b 0.001; symp-HD versus controls: t = 8.5, p b 0.001, Fig. 3). Both the preHDclose and symp-HD groups increased RD (pre-HDclose versus controls: t = 3.1, p b 0.05, symp-HD versus controls: t = 6.9, p b 0.001) and reduced FA (pre-HDclose versus controls: t = 4.0, p b 0.01; symp-HD versus controls: t = 6.1, p b 0.001) in the fronto-parietal tract connecting frontal and parietal cortices via the corpus callosum (Fig. 3). Symp-HD participants also showed significantly greater RD in the putamenprimary motor tract (t = 5.7, p b 0.001) and left caudate-paracentral tract (t = 7.8, p b 0.001), and reduced FA in left caudate-paracentral tract (t = 6.0, p b 0.001), compared with controls (Fig. 3). The overall pre-HD group, compared to controls, showed a significant increase in
183
RD only in the putamen-prefrontal (t = 4.0, p b 0.01) tract. There was a significant difference in RD between the symp-HD and overall pre-HD groups in the tracts of interest (putamen-motor: t = 5.2; putamenprefrontal: t = 5.0; frontoparietal t = 4.9; caudate-paracentral: t = 5.1; all p b 0.001). Whereas, FA values differed between pre-HD and sympHD only in the frontoparietal (4.2, p b 0.01) and caudate-paracentral (t = 5.0, p b 0.001) tracts. There was a trend (uncorrected p b 0.05) towards an increase in RD in the putamen-primary motor tract and left caudate-paracentral tract in the pre-HDclose group compared with the controls, and in both the putamen-prefrontal and the putamenprimary motor tracts for the pre-HDfar group compared with the controls. The pre-HDclose group also showed only a trend (uncorrected p b 0.05) towards increase in FA in the fronto-parietal and caudateparacentral tracts. Tract-specific comparisons between groups are detailed in the Supplementary Tables 1 and 2. Relationship between microstructure and cognitive and motor performance, and clinical severity Based on network-level differences observed between groups, we selected the putamen-prefrontal and right putamen-motor tracts in pre-HD and symp-HD, and the fronto-parietal and left caudate-paracentral tracts
Fig. 1. White matter networks aberrant in pre-HD and symp-HD relative to controls. (a) The network which showed a significantly (Z N 2.6, p b 0.05, network-level corrected) reduced connectivity in pre-HD compared to controls. (b) The network which showed a significantly (Z N 2.6, p b 0.05, network-level corrected) reduced connectivity in symp-HD compared to controls. The blue and green spheres represent approximate centers of the gray matter nodes which were included in the connectivity analysis. The green spheres represent nodes which were part of the altered networks. The lines represent white matter tracts connecting the green nodes.
184
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
Fig. 2. Visualization of key fiber streamlines disrupted in pre-HD and symp-HD groups. DTI-based deterministic tractography was used to track fiber streamlines disrupted in pre-HD and symp-HD and visualized in travkvis software. Streamlines connecting (a) bilateral putamen with prefrontal cortex and (b) right putamen with primary motor cortex were disrupted in both pre-HD and symp-HD. Streamlines connecting left caudate with (c) paracentral cortex and (d) frontal and parietal cortex via cingulum were disrupted in symp-HD only. Color coding is based on the direction of fiber tracts, as displayed on the color coding arrows provided in the figure. Red represents x-direction (left-right), blue represents z-direction (top-down), and green represents y-direction (posterior-anterior) in MNI coordinates.
in symp-HD for partial correlation analysis (Supplementary Table 3). For the pre-HD group, significant negative correlations were observed between average RD in the bilateral putamen-frontal tract and SDMT (r = − 0.44, p b 0.05) and Stroop Word reading (r = − 0.48, p b 0.05) performance (Fig. 4A). For the symp-HD group, RD in the fronto-parietal tract correlated with SDMT and Stroop Word performance and with UHDRS scores (SDMT: r = −0.51, p b 0.01, Stroop r = −0.55, p b 0.01, UHDRS: r = 0.48, p b 0.05) (Fig. 4B). RD in the putamenprefrontal tract correlated positively with UHDRS motor scores (r = 53, p b 0.05), and RD in the putamen-primary motor tract correlated positively with speeded tapping (r = 0.51, p b 0.05) in the symp-HD group. There was a trend (uncorrected p b 0.05) toward correlations between speeded tapping performance (Inter-tap Interval) and RD in the putamen-primary motor tract in pre-HD, and between SDMT and Stroop performance and RD in putamen-prefrontal tract in symp-HD. Discussion This study investigated white matter connectivity between frontal, parietal, and striatal brain regions in Huntington's disease using DTI fiber deterministic tractography and network based statistical analysis. A neuroanatomical network connecting putamen with prefrontal and motor cortices showed reduced white matter connectivity and microstructural changes in pre-HD compared with healthy controls. SympHD showed impairment in a network connecting frontal and parietal cortices, and striatum. RD in white matter tracts reflected clinical severity in symp-HD, and cognitive and motor functions in both pre-HD and
symp-HD. Our findings support the hypothesis that HD is associated with aberrant changes in specific neuroanatomical circuits, which reflects clinical status and cognitive function. Neuroimaging studies have consistently reported gray and white matter degeneration in cortical and sub-cortical structures many years prior to expected diagnosis, and in the absence of overt motor signs of HD (Dumas et al., 2012; Kloppel et al., 2008; Tabrizi et al., 2009). In the current study, impaired structural connectivity in premanifest HD was localized in a neuroanatomical circuit connecting three regionally distinct nodes: putamen, orbitoprefrontal, and precentral cortices. Fewer streamlines in pre-HD between the nodes which are involved in motor and cognitive functions (Liston et al., 2006), indicate early changes in both motor and cognitive parts of cortico-striatal neural circuit in pre-HD. The putamen, prefrontal, and motor cortices have been implicated in several voxel-wise investigations of both gray matter and white matter in pre-HD (Delmaire et al., 2013; Rosas et al., 2006; Tabrizi et al., 2009). Furthermore, segregated microstructural alterations in white matters of cingulum, longitudinal fasciculi, corona radiata, internal capsule, and cerebral peduncles have been previously reported in pre-HD (Stoffers et al., 2010). Consistent with previous studies, we also found increase in RD, suggesting greater demyelination, in the pre-HD close to onset in the tract connecting putamen with prefrontal cortex. The integrity of the putamen-prefrontal tract is critical for efficient recruitment of cognitive control during executive function (Liston et al., 2006). In the pre-HD group, a significant relationship was observed between increase in RD in the prefrontal-putamen tract and
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
185
Fig. 3. Tract-specific microstructral differences in specific tracts from the network showing group differences. In the putamen-motor tract, the symp-HD showed significantly greater RD compared with the control and pre-HD groups. In the putamen-prefrontal tract both the pre-HD close to onset (pre-HDclose) and symp-HD showed significantly greater RD compared with controls, symp-HD showed greater RD compared to the pre-HD group, and pre-HDclose showed greater RD compared to controls In the fronto-parietal tract both the pre-HD close to onset (pre-HDclose) and symp-HD showed significant increase in RD and reduced FA compared with controls, and the symp-HD group showed significantly increased RD and decreased FA compared to the pre-HD group. In the caudate-paracentral tract the symp-HD showed significantly greater RD and reduced FA compared with the controls and pre-HD groups. In the inset above the bar diagram for each tract, group-level masks for each tract (blue), connecting the regions (red and green), are visualized in an example 3D brain. * b 0.05, ** b 0.001. Vertical bars represent standard error.
poorer performance on SDMT and Stroop tasks. Although a previous study has reported association between motor function and integrity of white matter in the sensorimotor cortex in pre-HD (Dumas et al., 2012), we provide fresh evidence for relationship between cognitive function and integrity of a putamen-prefrontal tract in HD. The finding further supports that cognitive function in pre-HD may be associated with damage to specific cortico-striatal neural circuits (Kloppel et al., 2008; Tabrizi et al., 2009). In the symp-HD group, the aberrant network comprised of corticocortical tracts connecting the frontal and parietal cortices via corpus callosum, and the cortico-striatal tracts connecting the left caudate with posterior parietal, and the putamen with the primary motor and prefrontal cortices. The more wide-spread structural network disruption in symp-HD is consistent with previous studies reporting abnormal white matter morphology and diffusion properties in HD (Rosas et al., 2006, 2010; Tabrizi et al., 2009). Structural connectivity of motor circuits and its microstructure have been shown previously to be particularly vulnerable in symp-HD, reflecting both motor function and clinical severity (Bohanna et al., 2011). Here, we also found increased RD and decreased FA in cortico-cortical tracts in symp-HD. Aberrant RD was associated with clinical severity (UHDRS) and tests of cognitive function (SDMT and Stroop). Previous studies have also reported aberrant microstructure in cortico-striatal circuits associated with symptomatology in symp-HD (Bohanna et al., 2011; Kloppel et al., 2008). However, Rosas et al. (2010) recently provided evidence for increased RD and decreased FA in all parts of the corpus callosum and suggested that this may drive the reduced cortico-cortical connectivity in symp-HD. Our findings indicate that structural connectivity of
the cortico-cortical network is not only disrupted in symp-HD but is also associated with clinical signs in motor and cognitive functions. Network based analysis of structural connectivity has been used previously in several case–control studies (Zalesky et al., 2011), and is particularly suitable for detecting subtle changes in neuroanatomical circuits (Zalesky et al., 2010). In HD, early changes during disease progression have been localized in smaller white and gray matter areas (Rosas et al., 2006, 2010; Tabrizi et al., 2009). However, previous reports of damage in white matter in HD did not model pair-wise connectivity disturbances between multiple cortico-striatal brain areas. The use of a network approach in the current study enabled us to isolate a network of cortico-striatal regions as a key macro-circuit affected early in Huntington's disease. A number of cellular and neuroimaging studies, when interpreted together, may help to explain the structural connectivity and microstructure deficits in HD. Although cortico-striatal atrophy is a hallmark feature of HD (Douaud et al., 2006; Fennema-Notestine et al., 2004; Tabrizi et al., 2009), neuropathological studies have provided evidence for striato-cortical axonal degeneration, characterized by disruption of organelle and myelin sheath fragmentation in HD mouse models even when neuronal loss is largely absent (Yu et al., 2003). During all stages of HD, dendritic trees of surviving medium spiny neurons show abnormal growth, divergent changes in spine density, and recurving of dendritic branches (Ferrante et al., 1991; Vonsattel et al., 1985). There is also evidence to suggest aberrant enkephalin immunoreactivity in the substantia nigra in the absence of cell loss (Waters et al., 1988), and reduced density of enkephalin immunoreactive fibers in the external globus pallidus in pre-HD
186
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
Fig. 4. Relationship between tract-specific microstructral changes and clinical and cognitive measures. Scatterplot showing the relationship between average RD in the bilateral putamenfrontal tract and Stroop performance (a) and SDMT performance (b) in the pre-HD group. Scatterplot showing the relationship between RD in the fronto-parietal tract and UHDRS motor scores (c) and SDMT performance in the symp-HD group (d). The r values represent partial correlation coefficient (controlling for effect of age). All correlations were significant with p b 0.05 (corrected).
(Albin et al., 1990). The striatum, which shows very early dysfunction of axonal fibers, is both functionally (Di Martino et al., 2008) and structurally (Wiesendanger et al., 2004) connected to the fronto-parietal cortex. RD increases with the degree of demyelination in WM tracts (Song et al., 2005). Our finding, of increased RD in specific cortico-striatal tracts in pre-HD, when taken together with cellular and new imaging findings, indicates very early network-wide axonal dysfunction in HD. In considering limitations of the current study, it is important to note that tractography was used to create the connectivity matrix between brain regions of interest. Accuracy of the connectivity matrix is dependent on reliability of the fiber tracking method used. While a methodology similar to that used in the current study has been used in several other tractography studies (Bai et al., 2012; Zalesky et al., 2011), some fiber tracts may have been missed due to problems with fitting the diffusion tensor model in the regions of complex fiber geometry (e.g., crossing fibers) and partial volume effects. Crossing fiber model (Tournier et al., 2008) has been developed to better characterize fiber tracts between brain regions, which was not used in the current study. Furthermore, the controls were matched for age with only the pre-HD group, raising the possibility that findings in symp-HD are due to effects of aging. However, the difference between symp-HD and the controls remained significant when age was included as nuisance regressor, arguing against effect of aging on structural connectivity changes in symp-HD.
In conclusion, our findings provide evidence for network-specific decline in structural connectivity and microstructural integrity of white matter tracts in both pre-HD and symp-HD. Aberrant structural connectivity in pre-HD possibly represents early axonal dysfunction which is also related to subtle deficits cognitive and motor functions. Proliferation of disruption in structural connectivity may subsequently lead to the more advanced clinical severity profile observed in sympHD. Our findings suggest that DTI-based tractography has the potential to serve as a sensitive method for the assessment of cerebral integrity to future treatments. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.nbd.2014.01.013. Acknowledgments We would like to acknowledge the contribution of all 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, in collaboration with Melbourne, Monash and La Trobe Universities and this research is also an initiative of the Victorian Government of 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.
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
References Albin, R.L., Reiner, A., Anderson, K.D., Penney, J.B., Young, A.B., 1990. Striatal and nigral neuron subpopulations in rigid Huntington's disease: implications for the functional anatomy of chorea and rigidity-akinesia. Ann. Neurol. 27, 357–365. Bai, F., Shu, N., Yuan, Y., Shi, Y., Yu, H., Wu, D., et al., 2012. Topologically convergent and divergent structural connectivity patterns between patients with remitted geriatric depression and amnestic mild cognitive impairment. J. Neurosci. 32, 4307–4318. Bohanna, I., Georgiou-Karistianis, N., Egan, G.F., 2011. Connectivity-based segmentation of the striatum in Huntington's disease: vulnerability of motor pathways. Neurobiol. Dis. 42, 475–481. Conturo, T.E., Lori, N.F., Cull, T.S., Akbudak, E., Snyder, A.Z., Shimony, J.S., et al., 1999. Tracking neuronal fiber pathways in the living human brain. Proc. Natl. Acad. Sci. U. S. A. 96, 10422–10427. Davis, S.W., Dennis, N.A., Buchler, N.G., White, L.E., Madden, D.J., Cabeza, R., 2009. Assessing the effects of age on long white matter tracts using diffusion tensor tractography. NeuroImage 46, 530–541. Delmaire, C., Dumas, E.M., Sharman, M.A., Van den Bogaard, S.J., Valabregue, R., Jauffret, C., et al., 2013. The structural correlates of functional deficits in early Huntington's disease. Hum. Brain Mapp. 34, 2141–2153. Di Martino, A., Scheres, A., Margulies, D.S., Kelly, A.M., Uddin, L.Q., Shehzad, Z., et al., 2008. Functional connectivity of human striatum: a resting state FMRI study. Cereb. Cortex 18, 2735–2747. Douaud, G., Gaura, V., Ribeiro, M.J., Lethimonnier, F., Maroy, R., Verny, C., et al., 2006. Distribution of grey matter atrophy in Huntington's disease patients: a combined ROI-based and voxel-based morphometric study. NeuroImage 32, 1562–1575. Dumas, E.M., van den Bogaard, S.J., Ruber, M.E., Reilman, R.R., Stout, J.C., Craufurd, D., et al., 2012. Early changes in white matter pathways of the sensorimotor cortex in premanifest Huntington's disease. Hum. Brain Mapp. 33, 203–212. Fennema-Notestine, C., Archibald, S.L., Jacobson, M.W., Corey-Bloom, J., Paulsen, J.S., Peavy, G.M., et al., 2004. In vivo evidence of cerebellar atrophy and cerebral white matter loss in Huntington disease. Neurology 63, 989–995. Ferrante, R.J., Kowall, N.W., Richardson Jr., E.P., 1991. Proliferative and degenerative changes in striatal spiny neurons in Huntington's disease: a combined study using the section-Golgi method and calbindin D28k immunocytochemistry. J. Neurosci. 11, 3877–3887. Georgiou-Karistianis, N., Gray, M.A., Dominguez, D.J., Dymowski, A.R., Bohanna, I., Johnston, L.A., et al., 2013. 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., Stout, J.C., Dominguez, D.J., Poudel, G.R., Churchyard, A., Chua, P., et al., 2014. Functional magnetic resonance imaging of working memory in Huntington's disease: IMAGE-HD cross-sectional analysis, Human Brain Mapping. Hum. Brain Mapp. (in press). Graveland, G.A., Williams, R.S., DiFiglia, M., 1985. Evidence for degenerative and regenerative changes in neostriatal spiny neurons in Huntington's disease. Science 227, 770–773. Gray, M.A., Egan, G.F., Ando, A., Churchyard, A., Chua, P., Stout, J.C., et al., 2013. Prefrontal activity in Huntington's disease reflects cognitive and neuropsychiatric disturbances: the IMAGE-HD study. Exp. Neurol. 239, 218–228. Hart, E., Middelkoop, H., Jurgens, C.K., Witjes-Ane, M.N., Roos, R.A., 2011. Seven-year clinical follow-up of premanifest carriers of Huntington's disease. PLoS Curr. 3, RRN1288. Jones, D.K., 2008. Studying connections in the living human brain with diffusion MRI. Cortex 44, 936–952. Kloppel, S., Draganski, B., Golding, C.V., Chu, C., Nagy, Z., Cook, P.A., 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., Brinkman, R.R., Falush, D., Paulsen, J.S., Hayden, M.R., 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.
187
Li, J.Y., Conforti, L., 2013. Axonopathy in Huntington's disease. Exp. Neurol. 246, 62–71. Liston, C., Watts, R., Tottenham, N., Davidson, M.C., Niogi, S., Ulug, A.M., et al., 2006. Frontostriatal microstructure modulates efficient recruitment of cognitive control. Cereb. Cortex 16, 553–560. Macdonald, V., Halliday, G., 2002. Pyramidal cell loss in motor cortices in Huntington's disease. Neurobiol. Dis. 10, 378–386. Muller, H.P., Glauche, V., Novak, M.J., Nguyen-Thanh, T., Unrath, A., Lahiri, N., et al., 2011. Stability of white matter changes related to Huntington's disease in the presence of imaging noise: a DTI study. PLoS Curr. 3, RRN1232. Nelson, H.E., Willison, J., Owen, A.M., 1992. National adult reading test, 2nd edition. Int. J. Geriatr. Psychiatry 7, 533. Penney Jr., J.B., Vonsattel, J.P., MacDonald, M.E., Gusella, J.F., Myers, R.H., 1997. CAG repeat number governs the development rate of pathology in Huntington's disease. Ann. Neurol. 41, 689–692. Rosas, H.D., Tuch, D.S., Hevelone, N.D., Zaleta, A.K., Vangel, M., Hersch, S.M., 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., Lee, S.Y., Bender, A.C., Zaleta, A.K., Vangel, M., Yu, P., et al., 2010. Altered white matter microstructure in the corpus callosum in Huntington's disease: implications for cortical “disconnection”. NeuroImage 49, 2995–3004. Rowe, K.C., Paulsen, J.S., Langbehn, D.R., Duff, K., Beglinger, L.J., Wang, C., et al., 2010. Selfpaced timing detects and tracks change in prodromal Huntington disease. Neuropsychology 24, 435–442. Song, S.K., Yoshino, J., Le, T.Q., Lin, S.J., Sun, S.W., Cross, A.H., et al., 2005. Demyelination increases radial diffusivity in corpus callosum of mouse brain. NeuroImage 26, 132–140. Stoffers, D., Sheldon, S., Kuperman, J.M., Goldstein, J., Corey-Bloom, J., Aron, A.R., 2010. Contrasting gray and white matter changes in preclinical Huntington disease: an MRI study. Neurology 74, 1208–1216. Tabrizi, S.J., Langbehn, D.R., Leavitt, B.R., Roos, R.A., Durr, A., Craufurd, D., et al., 2009. Biological and clinical manifestations of Huntington's disease in the longitudinal TRACKHD study: cross-sectional analysis of baseline data. Lancet Neurol. 8, 791–801. Thieben, M.J., Duggins, A.J., Good, C.D., Gomes, L., Mahant, N., Richards, F., et al., 2002. The distribution of structural neuropathology in pre-clinical Huntington's disease. Brain 125, 1815–1828. Thu, D.C., Oorschot, D.E., Tippett, L.J., Nana, A.L., Hogg, V.M., Synek, B.J., et al., 2010. Cell loss in the motor and cingulate cortex correlates with symptomatology in Huntington's disease. Brain 133, 1094–1110. Tournier, J.D., Yeh, C.H., Calamante, F., Cho, K.H., Connelly, A., Lin, C.P., 2008. Resolving crossing fibres using constrained spherical deconvolution: validation using diffusion-weighted imaging phantom data. NeuroImage 42, 617–625. Tzourio-Mazoyer, N., Landeau, B., Papathanassiou, D., Crivello, F., Etard, O., Delcroix, N., et al., 2002. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. NeuroImage 15, 273–289. Vonsattel, J.P., Myers, R.H., Stevens, T.J., Ferrante, R.J., Bird, E.D., P., RJE, 1985. Neuropathological classification of Huntington's disease. J. Neuropathol. Exp. Neurol. 44, 559. Waters, C.M., Peck, R., Rossor, M., Reynolds, G.P., Hunt, S.P., 1988. Immunocytochemical studies on the basal ganglia and substantia nigra in Parkinson's disease and Huntington's chorea. Neuroscience 25, 419–438. Weaver, K.E., Richards, T.L., Liang, O., Laurino, M.Y., Samii, A., Aylward, E.H., 2009. Longitudinal diffusion tensor imaging in Huntington's Disease. Exp. Neurol. 216, 525–529. Wiesendanger, E., Clarke, S., Kraftsik, R., Tardif, E., 2004. Topography of cortico-striatal connections in man: anatomical evidence for parallel organization. Eur. J. Neurosci. 20, 1915–1922. Yu, Z.X., Li, S.H., Evans, J., Pillarisetti, A., Li, H., Li, X.J., 2003. Mutant huntingtin causes context-dependent neurodegeneration in mice with Huntington's disease. J. Neurosci. 23, 2193–2202. Zalesky, A., Fornito, A., Bullmore, E.T., 2010. Network-based statistic: identifying differences in brain networks. NeuroImage 53, 1197–1207. Zalesky, A., Fornito, A., Seal, M.L., Cocchi, L., Westin, C.F., Bullmore, E.T., et al., 2011. Disrupted axonal fiber connectivity in schizophrenia. Biol. Psychiatry 69, 80–89.
Contents lists available at ScienceDirect
Neurobiology of Disease journal homepage: www.elsevier.com/locate/ynbdi
White matter connectivity reflects clinical and cognitive status in Huntington's disease☆,☆☆,★ Govinda R. Poudel a,b,e, Julie C. Stout a, Juan F. Domínguez D a, Louisa Salmon a, Andrew Churchyard c, Phyllis Chua a, Nellie Georgiou-Karistianis a,⁎, Gary F. Egan a,b,d,e a
School of Psychological Sciences, Monash University, Clayton, Victoria, Australia Monash Biomedical Imaging (MBI), Monash University, Melbourne, VIC, Australia Department of Neurology, Monash Medical Centre, Clayton, Victoria, Australia d Centre for Neuroscience, University of Melbourne, Parkville, Victoria, Australia e VLSCI Life Sciences Computation Centre, Melbourne, Victoria, Australia b c
a r t i c l e
i n f o
Article history: Received 25 September 2013 Revised 18 December 2013 Accepted 19 January 2014 Available online 28 January 2014 Keywords: fMRI Huntington's disease Connectivity Tractography Diffusion tensor imaging
a b s t r a c t Objective: To investigate structural connectivity and the relationship between axonal microstructure and clinical, cognitive, and motor functions in premanifest (pre-HD) and symptomatic (symp-HD) Huntington's disease. Method: Diffusion tensor imaging (DTI) data were acquired from 35 pre-HD, 36 symp-HD, and 35 controls. Structural connectivity was mapped between 40 brain regions of interest using tractography. Between-group differences in structural connectivity were identified using network based statistics. Radial diffusivity (RD) and fractional anisotropy (FA) were compared in the white matter tracts from aberrant networks. RD values in aberrant tracts were correlated with clinical severity, and cognitive and motor performance. Results: A network connecting putamen with prefrontal and motor cortex demonstrated significantly reduced tractography streamlines in pre-HD. Symp-HD individuals showed reduced streamlines in a network connecting prefrontal, motor, and parietal cortices with both caudate and putamen. The symp-HD group, compared to controls and pre-HD, showed both increased RD and decreased FA in the fronto-parietal and caudate-paracentral tracts and increased RD in the putamen-prefrontal and putamen-motor tracts. The pre-HDclose, compared to controls, showed increased RD in the putamen-prefrontal and fronto-parietal tracts. In the pre-HD group, significant negative correlations were observed between SDMT and Stroop performance and RD in the bilateral putamen-prefrontal tract. In the symp-HD group, RD in the fronto-parietal tract was significantly positively correlated with UHDRS motor scores and significantly negatively correlated with performance on SDMT and Stroop tasks. Conclusions: We have provided evidence of aberrant connectivity and microstructural integrity in white matter networks in HD. Microstructural changes in the cortico-striatal fibers were associated with cognitive and motor performance in pre-HD, suggesting that changes in axonal integrity provide an early marker for clinically relevant impairment in HD. © 2014 Elsevier Inc. All rights reserved.
Introduction Neuropathological processes in Huntington's disease (HD) primarily target medium spiny neurons of the striatum (Graveland
et al., 1985). Neurodegeneration is also seen in pyramidal projection neurons in the motor and prefrontal cortices, and cingulate and angular gyri (Macdonald and Halliday, 2002; Thu et al., 2010). Together, these neurodegenerative changes are considered to be the cause of onset of
☆ Statistical analysis was performed by Dr Govinda R. Poudel School of Psychological Sciences; Monash Biomedical Imaging (MBI), Monash University, VIC, Australia. ☆☆ Author contributions: Govinda Poudel was involved in drafting/revising the manuscript for content; study concept or design; analysis or interpretation of data and statistical analysis. Julie C. Stout was involved in drafting/revising the manuscript for content; study concept or design; analysis or interpretation of data and obtainment of funding. Juan F. Domínguez D was involved in drafting/revising the manuscript for content; analysis or interpretation of data and acquisition of data. Louisa Salmon was involved in drafting/revising the manuscript for content; analysis or interpretation of data and acquisition of data. Andrew Churchyard was involved in the study concept or design; obtainment of funding and acquisition of data. Phyllis Chua was involved in the study concept or design; obtainment of funding and acquisition of data. Nellie Georgiou-Karistianis was involved in drafting/revising the manuscript for content; study concept or design; analysis or interpretation of data; obtainment of funding and study supervision or coordination. Gary F. Egan was involved in drafting/revising the manuscript for content; study concept or design; analysis or interpretation of data; obtainment of funding and study supervision or coordination. ★ Author Disclosures: All authors have no relevant biomedical, financial or potential conflicts of interest to declare. ⁎ 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). 0969-9961/$ – see front matter © 2014 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.nbd.2014.01.013
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
clinical symptoms in HD (Thu et al., 2010). In contrast, the mechanisms underlying cognitive changes, which often appear several years before clinical diagnosis, remain largely unknown. Disruption of structural connectivity in specific neural circuits has been proposed to be one of the possible mechanisms leading to early cognitive and motor changes in premanifest Huntington's disease (pre-HD) (Li and Conforti, 2013). Structural connectivity can deteriorate in HD due to axonal dysfunction and degeneration associated with huntingtin aggregates which can appear early in HD (reviewed by Li and Conforti, 2013). Hence, white matter atrophy is evident in T1-weighted neuroimaging studies of HD (Tabrizi et al., 2009; Thieben et al., 2002), with posterior-frontal whitematter degeneration apparent even in individuals far from onset (Tabrizi et al., 2009). Diffusion tensor imaging (DTI) studies of HD have also suggested selective microstructural changes in white matter encompassing cortico-striatal motor circuit, corpus callosum, periventricular region, corona radiata, and prefrontal cortex (Bohanna et al., 2011; Dumas et al., 2012; Rosas et al., 2006, 2010; Weaver et al., 2009). DTI also enables investigation of axonal fibers between gray matter structures using tractography methods (Bohanna et al., 2011; Jones, 2008). DTI tractography has been used in HD to isolate structural connections in specific neuroanatomical circuits including the motor loop and fronto-striatal circuit (Bohanna et al., 2011; Dumas et al., 2012; Kloppel et al., 2008), providing evidence for circuit specific alterations in white matter microstructure in HD. For example, microstructural damage in the striatal nodes of the motor loop has been shown to be associated with motor dysfunction in HD (Bohanna et al., 2011). Structural connectivity changes of the fronto-caudal tracts have been shown not only to reflect years to onset, but also to be associated with oculomotor function in pre-HD (Kloppel et al., 2008). Moreover, reduced fiber connectivity between the prefrontal cortex and the caudate has been shown to reflect symptomatology in pre-HD (Kloppel et al., 2008). DTI-tractography has also been used to determine pairwise connections between gray matter structures in the brain enabling calculation of a structural connectivity matrix for individual subjects (Zalesky et al., 2010). Network-based statistical methods can be used to isolate network connections that are altered in disease from the connectivity matrix (Zalesky et al., 2010, 2011). The identification of structural networks in pre-HD and symp-HD may provide insight into early markers of disease progression in HD. The aims of the current study were to identify cortico-striatal networks affected in pre-HD and symp-HD, determine the microstructural alterations in the axonal fibers connecting these pathways, and investigate the relationship between axonal microstructural changes and clinical, cognitive and motor functions in pre-HD and symp-HD. We hypothesized that structural connectivity in neural circuits connecting motor and prefrontal cortices with the caudate and putamen would be affected in both pre-HD and symp-HD. Microstructural white matter degeneration in symp-HD has been reported in the body of the corpus callosum, which structurally connects frontal and parietal areas (Rosas et al., 2010). We hypothesized that symp-HD individuals would show further white matter structural disconnectivity in the fronto-parietal network. Fronto-striatal neural circuits are crucial for cognitive control (Liston et al., 2006). We hypothesized that the microstructural integrity of the fronto-striatal tracts would be associated with cognitive dysfunction in both pre-HD and symp-HD. To test these hypotheses, tractography was used to identify the extent of axonal connectivity between 40 neocortico and striatal brain regions. A network-based statistical method (Zalesky et al., 2010) was used to isolate neuroanatomical networks that showed connectivity differences between the groups. DTI-based measures of radial diffusivity (RD) are thought to be sensitive to demyelinative processes (Song et al., 2005). We measured RD values from the tracts identified as aberrant in pre-HD and symp-HD, and investigated the relationship of RD changes with clinical severity, and cognitive and motor performance.
181
Methods Participants Thirty-five pre-HD, 36 symp-HD, and 35 healthy control volunteers were included in this investigation, all recruited as part of the Australian-based IMAGE-HD study (Georgiou-Karistianis et al., 2013, in press; Gray et al., 2013). Recruitment procedures and inclusion criteria have been published previously (Georgiou-Karistianis et al., 2013). Controls were matched to pre-HD participants for age, gender and IQ [National Adult Reading Test 2nd edition, NART-2 (Nelson et al., 1992)], an estimate of premorbid intelligence. One-way ANOVAs revealed no significant differences in age or IQ scores between the pre-HD group and controls, but significant differences in age between controls and symp-HD, and between pre-HD and symp-HD, respectively (p b 0.05). CAG repeat lengths in the expanded alleles of the participants ranged from 39 to 50 (42 ± 2 for pre-HD; 43 ± 2 for symp-HD). Similar to Tabrizi et al. (2009), inclusion in the pre-HD group was based on UHDRS total motor score ≤ 5. The average years to clinical onset for the pre-HD group was 15 ± 8 years, as determined by a formula based on age, and the number of CAG repeats (Langbehn et al., 2004). The average disease burden score (DBS; Penney et al., 1997) was 270 ± 53. Years since the diagnosis of symptom onset (ascertained by the clinician A.C.) in the symp-HD group ranged from 0 to 5 years with a mean DBS of 379 ± 70 for the group. Demographics, clinical information, and neurocognitive measures of interest are provided in Table 1. The complete battery of neurocognitive data collected for all participants as part of the IMAGE-HD study has been described in previous publications (Georgiou-Karistianis et al., 2013, in press; Gray et al., 2013). The IMAGE-HD study was approved by the Monash University and Melbourne Health Human Research Ethics Committees. Written informed consent was obtained from each participant in accordance with the Helsinki Declaration. MRI data acquisition Structural and functional MR images were acquired with the Siemens Magnetom Tim Trio 3 T MRI scanner (Siemens AG, Erlangen, Germany) and a 32-channel head coil at the Murdoch Children's Research Institute (Royal Children's Hospital, Victoria, Australia). High-resolution T1-weighted images were acquired (192 slices, slice thickness of 0.9 mm, 0.8 mm 0.8 mm in-plane resolution, 320 × 320 matrix, TI = 900 ms, TE = 2.59 ms, TR = 1900 ms, flip angle = 9°). Diffusion weighted data were acquired using a double spin echo diffusion weighted EPI sequence (TR = 8200, TE = 89 ms, flip = 90°, 64 contiguous slices with 2 mm isotropic voxels, acquisition matrix 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) were acquired. MRI data 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. Diffusion tensors were estimated and the reconstructed tensor matrices were diagonalized to obtain eigen values (λ1, λ2, λ3) for estimation of the radial diffusivity (RD) and fractional anisotropy (FA) of each voxel. For spatial normalization of DTI data to a common standard space, an iterative normalization procedure was used (Muller et al., 2011). The FA image from each subject was first normalized to the Montreal Neurological Institute (MNI) 1 mm3 resolution FA Map using non-linear registration (FNIRT). A study specific FA template was created by arithmetically averaging the normalized FA maps for participants from each group.
182
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
Table 1 Demographic, clinical and neurocognitive data for participants included in the current investigation. Mean ± SD
N (sample sizes) Gender (M:F) Age (years) UHDRS CAG repeats Disease burden score (DBS) Estimated YtO Duration of illness (years) SDMT Stroop Word Speeded tapping (ITI, ms)
Controls
Pre-HD
Symp-HD
35 11:24 42 ± 14 (24–73) – – –
35 14:21 42 ± 10 (24–65) 1 ± 1 (0–4) 42 ± 2 270 ± 53
36 21:15 52 ± 9⁎⁎++ (37–71) 19 ± 12+++ (6–60) 43 ± 2 379 ± 70+++
– – 57 ± 10 109 ± 16 221 ± 38
15 ± 8 – 51 ± 9⁎
– 2.0 ± 1.5 36 ± 12⁎⁎⁎++ 83 ± 21⁎⁎ 367 ± 164⁎⁎⁎+++
104 ± 17 245 ± 45⁎
SD, standard deviation; UHDRS, motor subscale score, Unified Huntington's Disease Rating Scale (pre-HD, UHDRS b 5; symp-HD, UHDRS ≥ 5); CAG, cytosine–adenine–guanine (number of repeats N40 is full penetrance); disease burden score (CAG-35.5) ∗ age; YTO, years to onset; SDMT, symbol digit modalities test; Stroop Word, Stroop speeded word reading task (number of correct words); ITI, inter-tap interval. 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.
Non-linear normalization of subject-specific FA volumes to the study specific template was iterated until the correlation was N 0.8 between the individual FA-maps and the FA-template. The non-linear transformation matrices resulting from the registration process were saved for later use in tractography registration.
DTI-based tractography Deterministic tractography was carried in each subject's native space using the Diffusion toolkit 0.6.0 (http://trackvis.org/). The interpolated streamline tracking method (Conturo et al., 1999) with fixed increments of 1 mm was used to generate streamlines which followed the direction of the principal eigenvector. Tracking propagation was terminated when a minimum angle of the principal eigenvector direction between adjacent voxels exceeded a threshold of 35° or when the streamlines exceeded the participant's brain mask. Brain masks were generated for each participant by identifying upper and lower thresholds corresponding to minima in the histogram of the diffusion-weighted volumes. Streamlines were initiated from the center of each voxel and streamlines shorter than 10 mm were discarded. For each subject, the tractography generated tens of thousands of streamlines representing structural networks in the cerebral cortex.
Statistical analysis Network Based Statistics (NBS) (Zalesky et al., 2010) was used to non-parametrically compare the group connectivity matrices (Zalesky et al., 2010). NBS has been validated and previously used to map aberrant structural networks in psychiatric and neurological conditions (Bai et al., 2012; Zalesky et al., 2011). A general linear model was used to model group differences with age differences covaried as microstructural changes have been reported in fronto-parietal white matter tracts (Davis et al., 2009). Permutation based unpaired t-tests using 5000 permutations were used to identify networks which differed between pre-HD and controls and symp-HD and controls. NBS identified anatomical networks which show significant differences in structural connectivity between groups (t N 2.6) using a networklevel family-wise-error correction method (p b 0.05). For networks that showed differences for the cortico-striatal brain region pairs between pre-HD and controls, the fiber tracts were visualized using trackvis software. Group-level masks of the cortico-striatal tracts of interest were generated by averaging the tracts from all participants (pre-HD, symp-HD and controls). For each subject, RD and FA values were calculated for the tracts of interest using group-level masks as a measure of microstructural integrity of axonal fiber tracts. For comparison of tract-specific RD between groups, we further divided the pre-HD group into two sub-groups: pre-HD far from onset (pre-HDfar) and pre-HD close to onset (pre-HD close ). This subgrouping was achieved using a median split of the pre-HD group according to their years to onset. Estimated years to onset for preHD far (n = 17) was 21 ± 6 years and for pre-HD close (n = 18) was 10 ± 3 years. Independent t-tests were conducted to compare RD and FA values of specific tracts between pre-HD far and preHDclose, pre-HDfar and controls, pre-HDclose and controls, overall pre-HD group and controls, and symp-HD and controls. Significant differences corrected for family-wise error rate using Holm–Bonferroni method (p b 0.05) are reported. To investigate the relationship between microstructural changes and clinical severity and cognitive and motor performance, partial correlation analyses (with age as a covariate) were performed using average RD in the tracts of interest. We selected SDMT (symbol digit modalities test), Stroop Word, and speeded tapping tasks from our neurocognitive battery as they have been shown to be sensitive to clinically relevant impairment in HD (Hart et al., 2011; Rowe et al., 2010; Tabrizi et al., 2009). SDMT measures visual attention and information processing speed and Stroop measures cognitive control, allowing us to tap into aspects of executive functions dependent on the integrity of fronto-striatal circuits (Liston et al., 2006). Speeded tapping performance (i.e., inter-tapping interval) provided a measure of motor function, which is dependent on integrity of motor loop. For the symp-HD participants, UHDRS motor scores were used as a measure of clinical severity and were correlated with RD values from the tracts disrupted in symp-HD. Significant correlations corrected for family-wise error rate using Holm–Bonferroni method (p b 0.05) are reported.
Structural network mapping Results To isolate neural circuits connecting frontal and parietal cortices with striatum, we generated a structural connectivity matrix for 40 regions of interest in the bilateral frontal and parietal cortices, striatum, and thalamus (Supplementary Fig. 1a). Brain regions of interest were defined using an anatomical labeling (AAL) atlas (Tzourio-Mazoyer et al., 2002). The AAL atlas was transformed from MNI space to each of the subject's DTI native space by applying the previously generated non-linear transformation matrices, and the discrete anatomical labels were preserved using a nearest neighbor interpolation method. For each subject, a 40 × 40 connectivity matrix representing the number of streamlines between pairs of brain regions was calculated so that each cell (i, j) contained the number of fiber streamlines passing through the brain regions i and j (Supplementary Fig. 1c).
A total of (17 ± 1) × 104 streamlines were generated on average for each participant. There were no significant differences in total number of streamlines generated between groups (pre-HD: (17 ± 1) × 104; symp-HD: (17 ± 2) × 104; controls: (17 ± 2) × 104; pre-HD versus controls: t = 1.6, p = 0.12; symp-HD versus controls: t = 1.4, p = 0.17). Network-level differences in structural connectivity Compared with controls, the pre-HD group showed significantly fewer streamlines in right and left hemispheric networks connecting the putamen with prefrontal and primary motor cortices (Figs. 1a, 2a, b). The symp-HD group, compared with controls, showed significantly
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
fewer streamlines in a wide-spread network connecting the frontal and parietal cortices, and the caudate and putamen with prefrontal and parietal cortices (Figs. 1b, 2a,b,c,d). There were no networks that showed significantly more streamlines in symp-HD or pre-HD compared to controls. Tract-specific microstructural differences Compared with controls, both the pre-HD close and symp-HD groups showed significantly greater RD in the bilateral putamenprefrontal tract (pre-HD close versus controls: t = 4.9, p b 0.001; symp-HD versus controls: t = 8.5, p b 0.001, Fig. 3). Both the preHDclose and symp-HD groups increased RD (pre-HDclose versus controls: t = 3.1, p b 0.05, symp-HD versus controls: t = 6.9, p b 0.001) and reduced FA (pre-HDclose versus controls: t = 4.0, p b 0.01; symp-HD versus controls: t = 6.1, p b 0.001) in the fronto-parietal tract connecting frontal and parietal cortices via the corpus callosum (Fig. 3). Symp-HD participants also showed significantly greater RD in the putamenprimary motor tract (t = 5.7, p b 0.001) and left caudate-paracentral tract (t = 7.8, p b 0.001), and reduced FA in left caudate-paracentral tract (t = 6.0, p b 0.001), compared with controls (Fig. 3). The overall pre-HD group, compared to controls, showed a significant increase in
183
RD only in the putamen-prefrontal (t = 4.0, p b 0.01) tract. There was a significant difference in RD between the symp-HD and overall pre-HD groups in the tracts of interest (putamen-motor: t = 5.2; putamenprefrontal: t = 5.0; frontoparietal t = 4.9; caudate-paracentral: t = 5.1; all p b 0.001). Whereas, FA values differed between pre-HD and sympHD only in the frontoparietal (4.2, p b 0.01) and caudate-paracentral (t = 5.0, p b 0.001) tracts. There was a trend (uncorrected p b 0.05) towards an increase in RD in the putamen-primary motor tract and left caudate-paracentral tract in the pre-HDclose group compared with the controls, and in both the putamen-prefrontal and the putamenprimary motor tracts for the pre-HDfar group compared with the controls. The pre-HDclose group also showed only a trend (uncorrected p b 0.05) towards increase in FA in the fronto-parietal and caudateparacentral tracts. Tract-specific comparisons between groups are detailed in the Supplementary Tables 1 and 2. Relationship between microstructure and cognitive and motor performance, and clinical severity Based on network-level differences observed between groups, we selected the putamen-prefrontal and right putamen-motor tracts in pre-HD and symp-HD, and the fronto-parietal and left caudate-paracentral tracts
Fig. 1. White matter networks aberrant in pre-HD and symp-HD relative to controls. (a) The network which showed a significantly (Z N 2.6, p b 0.05, network-level corrected) reduced connectivity in pre-HD compared to controls. (b) The network which showed a significantly (Z N 2.6, p b 0.05, network-level corrected) reduced connectivity in symp-HD compared to controls. The blue and green spheres represent approximate centers of the gray matter nodes which were included in the connectivity analysis. The green spheres represent nodes which were part of the altered networks. The lines represent white matter tracts connecting the green nodes.
184
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
Fig. 2. Visualization of key fiber streamlines disrupted in pre-HD and symp-HD groups. DTI-based deterministic tractography was used to track fiber streamlines disrupted in pre-HD and symp-HD and visualized in travkvis software. Streamlines connecting (a) bilateral putamen with prefrontal cortex and (b) right putamen with primary motor cortex were disrupted in both pre-HD and symp-HD. Streamlines connecting left caudate with (c) paracentral cortex and (d) frontal and parietal cortex via cingulum were disrupted in symp-HD only. Color coding is based on the direction of fiber tracts, as displayed on the color coding arrows provided in the figure. Red represents x-direction (left-right), blue represents z-direction (top-down), and green represents y-direction (posterior-anterior) in MNI coordinates.
in symp-HD for partial correlation analysis (Supplementary Table 3). For the pre-HD group, significant negative correlations were observed between average RD in the bilateral putamen-frontal tract and SDMT (r = − 0.44, p b 0.05) and Stroop Word reading (r = − 0.48, p b 0.05) performance (Fig. 4A). For the symp-HD group, RD in the fronto-parietal tract correlated with SDMT and Stroop Word performance and with UHDRS scores (SDMT: r = −0.51, p b 0.01, Stroop r = −0.55, p b 0.01, UHDRS: r = 0.48, p b 0.05) (Fig. 4B). RD in the putamenprefrontal tract correlated positively with UHDRS motor scores (r = 53, p b 0.05), and RD in the putamen-primary motor tract correlated positively with speeded tapping (r = 0.51, p b 0.05) in the symp-HD group. There was a trend (uncorrected p b 0.05) toward correlations between speeded tapping performance (Inter-tap Interval) and RD in the putamen-primary motor tract in pre-HD, and between SDMT and Stroop performance and RD in putamen-prefrontal tract in symp-HD. Discussion This study investigated white matter connectivity between frontal, parietal, and striatal brain regions in Huntington's disease using DTI fiber deterministic tractography and network based statistical analysis. A neuroanatomical network connecting putamen with prefrontal and motor cortices showed reduced white matter connectivity and microstructural changes in pre-HD compared with healthy controls. SympHD showed impairment in a network connecting frontal and parietal cortices, and striatum. RD in white matter tracts reflected clinical severity in symp-HD, and cognitive and motor functions in both pre-HD and
symp-HD. Our findings support the hypothesis that HD is associated with aberrant changes in specific neuroanatomical circuits, which reflects clinical status and cognitive function. Neuroimaging studies have consistently reported gray and white matter degeneration in cortical and sub-cortical structures many years prior to expected diagnosis, and in the absence of overt motor signs of HD (Dumas et al., 2012; Kloppel et al., 2008; Tabrizi et al., 2009). In the current study, impaired structural connectivity in premanifest HD was localized in a neuroanatomical circuit connecting three regionally distinct nodes: putamen, orbitoprefrontal, and precentral cortices. Fewer streamlines in pre-HD between the nodes which are involved in motor and cognitive functions (Liston et al., 2006), indicate early changes in both motor and cognitive parts of cortico-striatal neural circuit in pre-HD. The putamen, prefrontal, and motor cortices have been implicated in several voxel-wise investigations of both gray matter and white matter in pre-HD (Delmaire et al., 2013; Rosas et al., 2006; Tabrizi et al., 2009). Furthermore, segregated microstructural alterations in white matters of cingulum, longitudinal fasciculi, corona radiata, internal capsule, and cerebral peduncles have been previously reported in pre-HD (Stoffers et al., 2010). Consistent with previous studies, we also found increase in RD, suggesting greater demyelination, in the pre-HD close to onset in the tract connecting putamen with prefrontal cortex. The integrity of the putamen-prefrontal tract is critical for efficient recruitment of cognitive control during executive function (Liston et al., 2006). In the pre-HD group, a significant relationship was observed between increase in RD in the prefrontal-putamen tract and
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
185
Fig. 3. Tract-specific microstructral differences in specific tracts from the network showing group differences. In the putamen-motor tract, the symp-HD showed significantly greater RD compared with the control and pre-HD groups. In the putamen-prefrontal tract both the pre-HD close to onset (pre-HDclose) and symp-HD showed significantly greater RD compared with controls, symp-HD showed greater RD compared to the pre-HD group, and pre-HDclose showed greater RD compared to controls In the fronto-parietal tract both the pre-HD close to onset (pre-HDclose) and symp-HD showed significant increase in RD and reduced FA compared with controls, and the symp-HD group showed significantly increased RD and decreased FA compared to the pre-HD group. In the caudate-paracentral tract the symp-HD showed significantly greater RD and reduced FA compared with the controls and pre-HD groups. In the inset above the bar diagram for each tract, group-level masks for each tract (blue), connecting the regions (red and green), are visualized in an example 3D brain. * b 0.05, ** b 0.001. Vertical bars represent standard error.
poorer performance on SDMT and Stroop tasks. Although a previous study has reported association between motor function and integrity of white matter in the sensorimotor cortex in pre-HD (Dumas et al., 2012), we provide fresh evidence for relationship between cognitive function and integrity of a putamen-prefrontal tract in HD. The finding further supports that cognitive function in pre-HD may be associated with damage to specific cortico-striatal neural circuits (Kloppel et al., 2008; Tabrizi et al., 2009). In the symp-HD group, the aberrant network comprised of corticocortical tracts connecting the frontal and parietal cortices via corpus callosum, and the cortico-striatal tracts connecting the left caudate with posterior parietal, and the putamen with the primary motor and prefrontal cortices. The more wide-spread structural network disruption in symp-HD is consistent with previous studies reporting abnormal white matter morphology and diffusion properties in HD (Rosas et al., 2006, 2010; Tabrizi et al., 2009). Structural connectivity of motor circuits and its microstructure have been shown previously to be particularly vulnerable in symp-HD, reflecting both motor function and clinical severity (Bohanna et al., 2011). Here, we also found increased RD and decreased FA in cortico-cortical tracts in symp-HD. Aberrant RD was associated with clinical severity (UHDRS) and tests of cognitive function (SDMT and Stroop). Previous studies have also reported aberrant microstructure in cortico-striatal circuits associated with symptomatology in symp-HD (Bohanna et al., 2011; Kloppel et al., 2008). However, Rosas et al. (2010) recently provided evidence for increased RD and decreased FA in all parts of the corpus callosum and suggested that this may drive the reduced cortico-cortical connectivity in symp-HD. Our findings indicate that structural connectivity of
the cortico-cortical network is not only disrupted in symp-HD but is also associated with clinical signs in motor and cognitive functions. Network based analysis of structural connectivity has been used previously in several case–control studies (Zalesky et al., 2011), and is particularly suitable for detecting subtle changes in neuroanatomical circuits (Zalesky et al., 2010). In HD, early changes during disease progression have been localized in smaller white and gray matter areas (Rosas et al., 2006, 2010; Tabrizi et al., 2009). However, previous reports of damage in white matter in HD did not model pair-wise connectivity disturbances between multiple cortico-striatal brain areas. The use of a network approach in the current study enabled us to isolate a network of cortico-striatal regions as a key macro-circuit affected early in Huntington's disease. A number of cellular and neuroimaging studies, when interpreted together, may help to explain the structural connectivity and microstructure deficits in HD. Although cortico-striatal atrophy is a hallmark feature of HD (Douaud et al., 2006; Fennema-Notestine et al., 2004; Tabrizi et al., 2009), neuropathological studies have provided evidence for striato-cortical axonal degeneration, characterized by disruption of organelle and myelin sheath fragmentation in HD mouse models even when neuronal loss is largely absent (Yu et al., 2003). During all stages of HD, dendritic trees of surviving medium spiny neurons show abnormal growth, divergent changes in spine density, and recurving of dendritic branches (Ferrante et al., 1991; Vonsattel et al., 1985). There is also evidence to suggest aberrant enkephalin immunoreactivity in the substantia nigra in the absence of cell loss (Waters et al., 1988), and reduced density of enkephalin immunoreactive fibers in the external globus pallidus in pre-HD
186
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
Fig. 4. Relationship between tract-specific microstructral changes and clinical and cognitive measures. Scatterplot showing the relationship between average RD in the bilateral putamenfrontal tract and Stroop performance (a) and SDMT performance (b) in the pre-HD group. Scatterplot showing the relationship between RD in the fronto-parietal tract and UHDRS motor scores (c) and SDMT performance in the symp-HD group (d). The r values represent partial correlation coefficient (controlling for effect of age). All correlations were significant with p b 0.05 (corrected).
(Albin et al., 1990). The striatum, which shows very early dysfunction of axonal fibers, is both functionally (Di Martino et al., 2008) and structurally (Wiesendanger et al., 2004) connected to the fronto-parietal cortex. RD increases with the degree of demyelination in WM tracts (Song et al., 2005). Our finding, of increased RD in specific cortico-striatal tracts in pre-HD, when taken together with cellular and new imaging findings, indicates very early network-wide axonal dysfunction in HD. In considering limitations of the current study, it is important to note that tractography was used to create the connectivity matrix between brain regions of interest. Accuracy of the connectivity matrix is dependent on reliability of the fiber tracking method used. While a methodology similar to that used in the current study has been used in several other tractography studies (Bai et al., 2012; Zalesky et al., 2011), some fiber tracts may have been missed due to problems with fitting the diffusion tensor model in the regions of complex fiber geometry (e.g., crossing fibers) and partial volume effects. Crossing fiber model (Tournier et al., 2008) has been developed to better characterize fiber tracts between brain regions, which was not used in the current study. Furthermore, the controls were matched for age with only the pre-HD group, raising the possibility that findings in symp-HD are due to effects of aging. However, the difference between symp-HD and the controls remained significant when age was included as nuisance regressor, arguing against effect of aging on structural connectivity changes in symp-HD.
In conclusion, our findings provide evidence for network-specific decline in structural connectivity and microstructural integrity of white matter tracts in both pre-HD and symp-HD. Aberrant structural connectivity in pre-HD possibly represents early axonal dysfunction which is also related to subtle deficits cognitive and motor functions. Proliferation of disruption in structural connectivity may subsequently lead to the more advanced clinical severity profile observed in sympHD. Our findings suggest that DTI-based tractography has the potential to serve as a sensitive method for the assessment of cerebral integrity to future treatments. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.nbd.2014.01.013. Acknowledgments We would like to acknowledge the contribution of all 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, in collaboration with Melbourne, Monash and La Trobe Universities and this research is also an initiative of the Victorian Government of 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.
G.R. Poudel et al. / Neurobiology of Disease 65 (2014) 180–187
References Albin, R.L., Reiner, A., Anderson, K.D., Penney, J.B., Young, A.B., 1990. Striatal and nigral neuron subpopulations in rigid Huntington's disease: implications for the functional anatomy of chorea and rigidity-akinesia. Ann. Neurol. 27, 357–365. Bai, F., Shu, N., Yuan, Y., Shi, Y., Yu, H., Wu, D., et al., 2012. Topologically convergent and divergent structural connectivity patterns between patients with remitted geriatric depression and amnestic mild cognitive impairment. J. Neurosci. 32, 4307–4318. Bohanna, I., Georgiou-Karistianis, N., Egan, G.F., 2011. Connectivity-based segmentation of the striatum in Huntington's disease: vulnerability of motor pathways. Neurobiol. Dis. 42, 475–481. Conturo, T.E., Lori, N.F., Cull, T.S., Akbudak, E., Snyder, A.Z., Shimony, J.S., et al., 1999. Tracking neuronal fiber pathways in the living human brain. Proc. Natl. Acad. Sci. U. S. A. 96, 10422–10427. Davis, S.W., Dennis, N.A., Buchler, N.G., White, L.E., Madden, D.J., Cabeza, R., 2009. Assessing the effects of age on long white matter tracts using diffusion tensor tractography. NeuroImage 46, 530–541. Delmaire, C., Dumas, E.M., Sharman, M.A., Van den Bogaard, S.J., Valabregue, R., Jauffret, C., et al., 2013. The structural correlates of functional deficits in early Huntington's disease. Hum. Brain Mapp. 34, 2141–2153. Di Martino, A., Scheres, A., Margulies, D.S., Kelly, A.M., Uddin, L.Q., Shehzad, Z., et al., 2008. Functional connectivity of human striatum: a resting state FMRI study. Cereb. Cortex 18, 2735–2747. Douaud, G., Gaura, V., Ribeiro, M.J., Lethimonnier, F., Maroy, R., Verny, C., et al., 2006. Distribution of grey matter atrophy in Huntington's disease patients: a combined ROI-based and voxel-based morphometric study. NeuroImage 32, 1562–1575. Dumas, E.M., van den Bogaard, S.J., Ruber, M.E., Reilman, R.R., Stout, J.C., Craufurd, D., et al., 2012. Early changes in white matter pathways of the sensorimotor cortex in premanifest Huntington's disease. Hum. Brain Mapp. 33, 203–212. Fennema-Notestine, C., Archibald, S.L., Jacobson, M.W., Corey-Bloom, J., Paulsen, J.S., Peavy, G.M., et al., 2004. In vivo evidence of cerebellar atrophy and cerebral white matter loss in Huntington disease. Neurology 63, 989–995. Ferrante, R.J., Kowall, N.W., Richardson Jr., E.P., 1991. Proliferative and degenerative changes in striatal spiny neurons in Huntington's disease: a combined study using the section-Golgi method and calbindin D28k immunocytochemistry. J. Neurosci. 11, 3877–3887. Georgiou-Karistianis, N., Gray, M.A., Dominguez, D.J., Dymowski, A.R., Bohanna, I., Johnston, L.A., et al., 2013. 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., Stout, J.C., Dominguez, D.J., Poudel, G.R., Churchyard, A., Chua, P., et al., 2014. Functional magnetic resonance imaging of working memory in Huntington's disease: IMAGE-HD cross-sectional analysis, Human Brain Mapping. Hum. Brain Mapp. (in press). Graveland, G.A., Williams, R.S., DiFiglia, M., 1985. Evidence for degenerative and regenerative changes in neostriatal spiny neurons in Huntington's disease. Science 227, 770–773. Gray, M.A., Egan, G.F., Ando, A., Churchyard, A., Chua, P., Stout, J.C., et al., 2013. Prefrontal activity in Huntington's disease reflects cognitive and neuropsychiatric disturbances: the IMAGE-HD study. Exp. Neurol. 239, 218–228. Hart, E., Middelkoop, H., Jurgens, C.K., Witjes-Ane, M.N., Roos, R.A., 2011. Seven-year clinical follow-up of premanifest carriers of Huntington's disease. PLoS Curr. 3, RRN1288. Jones, D.K., 2008. Studying connections in the living human brain with diffusion MRI. Cortex 44, 936–952. Kloppel, S., Draganski, B., Golding, C.V., Chu, C., Nagy, Z., Cook, P.A., 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., Brinkman, R.R., Falush, D., Paulsen, J.S., Hayden, M.R., 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.
187
Li, J.Y., Conforti, L., 2013. Axonopathy in Huntington's disease. Exp. Neurol. 246, 62–71. Liston, C., Watts, R., Tottenham, N., Davidson, M.C., Niogi, S., Ulug, A.M., et al., 2006. Frontostriatal microstructure modulates efficient recruitment of cognitive control. Cereb. Cortex 16, 553–560. Macdonald, V., Halliday, G., 2002. Pyramidal cell loss in motor cortices in Huntington's disease. Neurobiol. Dis. 10, 378–386. Muller, H.P., Glauche, V., Novak, M.J., Nguyen-Thanh, T., Unrath, A., Lahiri, N., et al., 2011. Stability of white matter changes related to Huntington's disease in the presence of imaging noise: a DTI study. PLoS Curr. 3, RRN1232. Nelson, H.E., Willison, J., Owen, A.M., 1992. National adult reading test, 2nd edition. Int. J. Geriatr. Psychiatry 7, 533. Penney Jr., J.B., Vonsattel, J.P., MacDonald, M.E., Gusella, J.F., Myers, R.H., 1997. CAG repeat number governs the development rate of pathology in Huntington's disease. Ann. Neurol. 41, 689–692. Rosas, H.D., Tuch, D.S., Hevelone, N.D., Zaleta, A.K., Vangel, M., Hersch, S.M., 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., Lee, S.Y., Bender, A.C., Zaleta, A.K., Vangel, M., Yu, P., et al., 2010. Altered white matter microstructure in the corpus callosum in Huntington's disease: implications for cortical “disconnection”. NeuroImage 49, 2995–3004. Rowe, K.C., Paulsen, J.S., Langbehn, D.R., Duff, K., Beglinger, L.J., Wang, C., et al., 2010. Selfpaced timing detects and tracks change in prodromal Huntington disease. Neuropsychology 24, 435–442. Song, S.K., Yoshino, J., Le, T.Q., Lin, S.J., Sun, S.W., Cross, A.H., et al., 2005. Demyelination increases radial diffusivity in corpus callosum of mouse brain. NeuroImage 26, 132–140. Stoffers, D., Sheldon, S., Kuperman, J.M., Goldstein, J., Corey-Bloom, J., Aron, A.R., 2010. Contrasting gray and white matter changes in preclinical Huntington disease: an MRI study. Neurology 74, 1208–1216. Tabrizi, S.J., Langbehn, D.R., Leavitt, B.R., Roos, R.A., Durr, A., Craufurd, D., et al., 2009. Biological and clinical manifestations of Huntington's disease in the longitudinal TRACKHD study: cross-sectional analysis of baseline data. Lancet Neurol. 8, 791–801. Thieben, M.J., Duggins, A.J., Good, C.D., Gomes, L., Mahant, N., Richards, F., et al., 2002. The distribution of structural neuropathology in pre-clinical Huntington's disease. Brain 125, 1815–1828. Thu, D.C., Oorschot, D.E., Tippett, L.J., Nana, A.L., Hogg, V.M., Synek, B.J., et al., 2010. Cell loss in the motor and cingulate cortex correlates with symptomatology in Huntington's disease. Brain 133, 1094–1110. Tournier, J.D., Yeh, C.H., Calamante, F., Cho, K.H., Connelly, A., Lin, C.P., 2008. Resolving crossing fibres using constrained spherical deconvolution: validation using diffusion-weighted imaging phantom data. NeuroImage 42, 617–625. Tzourio-Mazoyer, N., Landeau, B., Papathanassiou, D., Crivello, F., Etard, O., Delcroix, N., et al., 2002. Automated anatomical labeling of activations in SPM using a macroscopic anatomical parcellation of the MNI MRI single-subject brain. NeuroImage 15, 273–289. Vonsattel, J.P., Myers, R.H., Stevens, T.J., Ferrante, R.J., Bird, E.D., P., RJE, 1985. Neuropathological classification of Huntington's disease. J. Neuropathol. Exp. Neurol. 44, 559. Waters, C.M., Peck, R., Rossor, M., Reynolds, G.P., Hunt, S.P., 1988. Immunocytochemical studies on the basal ganglia and substantia nigra in Parkinson's disease and Huntington's chorea. Neuroscience 25, 419–438. Weaver, K.E., Richards, T.L., Liang, O., Laurino, M.Y., Samii, A., Aylward, E.H., 2009. Longitudinal diffusion tensor imaging in Huntington's Disease. Exp. Neurol. 216, 525–529. Wiesendanger, E., Clarke, S., Kraftsik, R., Tardif, E., 2004. Topography of cortico-striatal connections in man: anatomical evidence for parallel organization. Eur. J. Neurosci. 20, 1915–1922. Yu, Z.X., Li, S.H., Evans, J., Pillarisetti, A., Li, H., Li, X.J., 2003. Mutant huntingtin causes context-dependent neurodegeneration in mice with Huntington's disease. J. Neurosci. 23, 2193–2202. Zalesky, A., Fornito, A., Bullmore, E.T., 2010. Network-based statistic: identifying differences in brain networks. NeuroImage 53, 1197–1207. Zalesky, A., Fornito, A., Seal, M.L., Cocchi, L., Westin, C.F., Bullmore, E.T., et al., 2011. Disrupted axonal fiber connectivity in schizophrenia. Biol. Psychiatry 69, 80–89.