Cortical volumes and atrophy rates in FTD-3 CHMP2B mutation carriers and related non-carriers

Abstract

Frontotemporal dementia constitutes the third most prevalent neurodegenerative disease with dementia. We compared cortical structural changes in nine presymptomatic CHMP2B frontotemporal dementia mutation positive individuals with seven mutation negative family members. Using serial MRI scans with a mean interval of 16 months and surface based cortical segmentation we measured cortical thickness and volume, and quantified atrophy rates. Cortical thickness and atrophy rates were averaged within major lobes and focal effects were determined by parametric statistical maps.

The volumetric atrophy rates in the presymptomatic CHMP2B mutation carriers were statistically significant, though of a lower magnitude than those previously reported in patients of other types of frontotemporal dementia. Cortical thickness measurements revealed cortical thinning in mutation carriers bilaterally in the frontal and occipital lobes, and in the left temporal lobe. Results indicated that cortical thickness has a higher sensitivity for detecting small changes than whole-brain volumetric measures. Comparing mutation carriers with non-carriers revealed increased atrophy rates in mutation carriers bilaterally in the inferio-temporal cortex, the superior frontal cortex, and the insular cortex. These findings indicated impairment of regions involved in both behaviour and language. The symptoms previously reported in clinical CHMP2B frontotemporal dementia patients are associated with the anatomically affected regions here found in the presymptomatic mutation carriers.

Introduction

Frontotemporal dementia (FTD) is a syndromic clinical variant of frontotemporal lobar degeneration (FTLD) which constitutes the third most prevalent group of neurodegenerative diseases with cognitive impairment (Ratnavalli et al., 2002, Harvey et al., 2003). Within recent years the clinical, molecular genetic and pathological classifications of FTD have evolved (Cairns et al., 2007, Davion et al., 2007). Up to 40% of FTD cases are considered autosomal dominantly inherited. One of the rarer causes of familial FTD is CHMP2B-mutation related FTD with a pathogenic G-to-C transition in the acceptor splice site of CHMP2B exon 6 (c.532-1G > C) on chromosome 3 (FTD-3) (Skibinski et al., 2005). The CHMP2B protein is a part of the Escort-3 complex involved in trafficking proteins destined for degradation in the Golgi apparatus. The molecular disease mechanism is not yet fully known. The disease was primarily described in a large Danish family (Gydesen et al., 2002, Lindquist et al., 2008), but a novel nonsense mutation in the CHMP2B gene was recently identified in a Belgian familial FTD patient further supporting the gene to be involved in FTD (van der Zee et al., 2008).

The Danish FTD-3 family is very large with 33 identified patients and another 250 at risk for developing the disease within the next 60 years. The average FTD-3 clinical onset is 57 years with a broad range from 43 to 65. As the symptom onset is insidious the exact time of onset can be difficult to determine. Patients present with primarily a clinical syndrome of frontotemporal dementia with behavioural changes, apathy, sometimes aggression and/or changed eating behaviour. During the early course they rarely have language disturbances, but when neuropsychologically tested they often have impairment of more posteriorly cortically located functions such as memory and visuospatial problems. Urinary incontinence and gait disturbances are normally late features although they sometimes can be seen during the early years of the disease. Disease duration from diagnosis to death ranges from 2 to over 20 years.

In FTD cortical structural changes are per se primarily found in the frontal and temporal lobes (Broe et al., 2003). However, studies have also reported changes in the parietal lobes (Grossman et al., 2004, Avants et al., 2005, Whitwell et al., 2007c). The aim of the present study is to assess cortical structural changes in preclinical FTD-3 CHMP2B mutation positive cases compared to mutation negative family members, and furthermore to assess a possible progression of cortical changes. A secondary aim has been to try to identify possible preclinical focal cortical abnormalities.

In vivo investigation of brain cortical structural changes using magnetic resonance imaging (MRI) has primarily employed manual or semi-automatic tracing of tissue boundaries to quantify anatomical structures (Jack et al., 1999, Rankin et al., 2004). Such approaches are time consuming and subject to inter-rater variability. Therefore, automatic unbiased computational approaches have gained popularity when studying cohorts of subjects.

A variety of studies (Whitwell and Jack, 2005) have used voxel-based morphometry (VBM) to detect brain changes in diseases with FTLD and differences between disease FTLD sub-types and healthy controls since the introduction of the method (Wright et al., 1995, Ashburner and Friston, 2000). VBM performs voxel-wise comparisons between spatially aligned MRI scans of subject groups enabling identification of tissue growth and tissue loss throughout the entire brain. A related method is tensor-based morphometry (TBM) (Ashburner et al., 1998), analyzing the deformation field involved in non-linear mapping of images, such as mapping of intra-subject serial scans and mapping of subject to group average. This way, local expansions and contractions can be identified, and the tensor maps can be used to quantify longitudinal effects and differences between subjects and groups. TBM has been used in different areas, such as studying the developing human brain (Chung et al., 2001) and measuring degeneration in Alzheimer's disease (Fox et al., 2001). TBM has been used less extensively than VBM within the field of FTLD, but recently more studies using TBM have been reported (Studholme et al., 2004, Avants et al., 2005, Cardenas et al., 2007).

A third type of method for measuring cortical changes is the explicit segmentation of the cerebral cortex for measurements of cortical thickness using parametric or geometric deformable models (Dale et al., 1999, Zeng et al., 1999, Fischl and Dale, 2000, Jones et al., 2000, Kriegeskorte and Goebel, 2001, Han et al., 2004, June et al., 2005, Xue et al., 2007). The cortex is explicitly or implicitly represented as surfaces of the white matter/gray matter boundary and the gray matter/cerebrospinal fluid boundary fitted to the images with subvoxel precision. This enables measurements of cortical thickness throughout the entire cortex with the advantage of standardized thickness measures, which is unavailable through VBM or TBM. In addition, VBM does not consider the cortical geometry, and cannot differentiate the cortical thickness of opposing walls in sulci. The drawback of surface based methods is the lack of quantification of subcortical regions, such as the thalamus, and basal ganglia. However, in studies where cortical structures are the objective and subcortical structures are less relevant, surface based methods are preferable. Surface based methods have been used to quantify changes in a variety of diseases, such as schizophrenia (Kuperberg et al., 2003, Narr et al., 2005, Yoon et al., 2007), obsessive–compulsive disorder (Shin et al., 2007), and Alzheimer's disease (Lerch et al., 2008). In diseases with FTLD surface based methods are apt, as degeneration is expected in the cortical lobes.

In this study we applied a surface based cortical segmentation method to serial MRI scans of preclinical individuals with CHMP2B-mutation related FTD and individuals without the mutation from the same family. Global volume measurements and local cortical thicknesses were determined from the cortical surfaces. Longitudinal and cross-sectional differences in cortical thickness were evaluated by lobe averages and by construction of statistical parametric maps.