Amygdala subnuclei are differentially affected in the different genetic and pathological forms of frontotemporal dementia

Introduction Frontotemporal dementia (FTD) is a heterogeneous neurodegenerative disorder with multiple genetic and pathological causes. It is characterized by both cortical and subcortical atrophies, with previous studies showing early involvement of the amygdala. However, no prior study has specifically investigated the atrophy of different subnuclei of the amygdala. Methods Using an automated segmentation tool for T1-weighted volumetric magnetic resonance imaging, we investigated amygdalar subnuclei (AS) involvement in a cohort of 132 patients with genetic or pathologically confirmed FTD (age: mean = 61 years (standard deviation = 8); disease duration: 5 (3) years) compared with 107 age-matched controls. Results AS were affected in all genetic and pathological forms of FTD. MAPT mutations/FTDP-17, Pick's disease, and transactive response DNA binding protein 43 kDa type C were the forms with the smallest amygdala (35%–50% smaller than controls in the most affected hemisphere, P < .0005). In most FTD groups, medial subnuclei (particularly the superficial, accessory basal and basal/paralaminar subnuclei) tended to be affected more than the lateral subnuclei, except for the progressive supranuclear palsy group, in which the corticoamygdaloid transition area was the least-affected area. Discussion Differential involvement of the AS was seen in the different genetic and pathological forms of FTD. In general, the most affected subnuclei were the superficial, accessory basal and basal/paralaminar subnuclei, which form part of a network of regions that control reward and emotion regulation, functions known to be particularly affected in FTD.


Introduction
Frontotemporal dementia (FTD) is a heterogeneous neurodegenerative disorder. Genetically, around a third of patients with FTD have an autosomal dominant mutation in microtubule-associated protein tau (MAPT), progranulin (GRN), and chromosome 9 open reading frame 72 (C9orf72) [1]. Neuropathologically, the three most common abnormalities in the brain are tau, transactive response DNA binding protein 43 kDa (TDP-43), and fused in sarcoma inclusions [2,3]. Atrophy in the medial temporal lobe is a common feature in cases of FTD, and the amygdala is often affected in the early stages of the illness, particularly in carriers of mutations in the MAPT gene [4,5] where volume loss has been seen on magnetic resonance imaging 10-15 years before the expected onset [6]. Neuropathological investigations have also shown amygdalar involvement, e.g., one study described severe volume loss of 52% in FTD [7], where Pick's bodies were found in half of the sample, whereas another study showed TDP-43 inclusions in the basolateral nucleus of the amygdala in the earliest stage of the disease [8].
The amygdala is composed of different subnuclei with connections to the limbic system, as well as to the rest of the brain [9][10][11][12]. These are involved in reward learning, motivation, emotion, and several cognitive functions such as attention, perception, and explicit memory [10] (Fig. 1). Owing to recent advances in parcellation methods, it is possible to measure these amygdalar subnuclei (AS) in vivo on magnetic resonance scans [15]. We aimed to investigate the pattern of atrophy of the AS in a cohort with a genetic or pathologically confirmed diagnosis of FTD to clarify whether and to what extent the AS are impaired across these different forms of FTD.

Methods
We reviewed the UCL Dementia Research Center FTD magnetic resonance imaging (MRI) database to identify patients with a genetic or pathologically confirmed diagnosis of FTD and a usable T1-weighted magnetic resonance scan. A total of 132 patients were identified (Table 1). Seventy-five patients were carriers of a mutation in one of the FTD-associated genes: 27 with a mutation in MAPT [16,17], 18 in GRN [18,19], 29 in chromosome 9 open reading frame 72 (C9orf72) [20,21], and one with a dual mutation in both GRN and C9orf72, who was excluded from the genetic analysis. For 79 patients, postmortem confirmation of the underlying neuropathology was available: fused in sarcoma (n 5 3), TDP-43 type A (n 5 16), TDP-43 type B (n 5 3), TDP-43 type C (n 5 20); Pick's disease (n 5 17), progressive supranuclear palsy (n 5 4), corticobasal degeneration (n 5 9), and frontotemporal dementia with parkinsonism linked to chromosome 17 (FTDP-17) (n 5 7). There was an overlap of 21 cases between the genetic groups and the pathological groups: 5 patients with GRN mutation had TDP-43 type A, 7 with C9orf72 mutation had TDP-43 type A and 2 had TDP-43 type B, and 7 with MAPT mutation had FTDP-17; the patient with dual GRN/C9orf72 mutation had TDP-43 type A. A total of 107 cognitively normal subjects, with a similar age to the patients and a usable volumetric T1-weighted MRI, were identified as controls. The study was approved by the local ethics committee, and written informed consent was obtained from all participants. No significant age difference was seen between the FTD groups and controls (Table 1). Among the genetic groups, the GRN group had a shorter disease duration (3.2 [2.8] years) than the C9orf72 (5.5 [3.2], P 5 .040) and MAPT (5.7 [3.2], P 5 .002) groups, but no difference for disease duration was found among the pathological groups (P 5 .722, analysis of variance). No difference was found Volumetric MRI scans were first bias field corrected and whole-brain parcellated using the geodesic information flow (GIF) algorithm [22], which is based on atlas propagation and label fusion. Volumes of the whole amygdala and of 9 AS were subsequently segmented using a customized version of the module available in FreeSurfer 6.0 [15] to adapt the output of GIF to the FreeSurfer format. Based on anatomical subdivision [9], we combined the nine original subnuclei and focused the analysis on the following five regions: (1) lateral nucleus; (2) basal and paralaminar nucleus; (3) accessory basal nucleus (AB); (4) corticoamygdaloid transition area; and (5) the superficial nuclei (central nucleus, cortical nucleus, medial nucleus, anterior amygdaloid area) (Fig. 1). Based on the volumes of the left and right hemispheres extracted from the GIF, we defined the most severely affected cerebral hemisphere in each patient. This allows avoidance of the difficulties that commonly arise in imaging analyses of FTD in which accurate outcomes of volumetric analyses can be obfuscated by combining people with predominantly right-or left-hemisphere atrophy in the same analysis. We then compared AS volumes between groups within the most and least severely affected hemispheres rather than within right and left sides. However, we also investigated asymmetry by calculating an asymmetry index (AI), defined as the absolute difference between the left and right total amygdalar volumes in relation to the total bilateral volume: j(Left 2 Right)j/(Left 1 Right).
AS volumes were expressed as a percentage of the total intracranial volume, computed using SPM12 v6470 (Statistical Parametric Mapping; Wellcome Trust Center for Neuroimaging, London, UK) running under Matlab R2014b (Math Works, Natick, MA) [23]. All segmentations were visually checked for quality.
Statistical analyses were performed on AS volumes and AI using the SPSS software (SPSS Inc., Chicago, IL) v22.0, between control and FTD groups, using an analysis of variance test adjusting for the scanner type, total intracranial volume, gender, and age. Results were corrected for multiple comparisons (Bonferroni's correction) at P , .005 for the genetic and pathological groups.

Results
Stratifying by genetics, the MAPT group had the smallest AS in both hemispheres (37%-43%, P , .0005), whereas C9orf72 and GRN had similar volume differences in the most affected hemisphere, with the greatest involvement in the superficial group and AB (24%-29%, P , .0005) and the lowest in the lateral nucleus (14%-18%, P , .0005) ( Fig. 2A and Supplementary Table 1). All AS were significantly smaller in MAPT than in C9orf72 and GRN (P , .0005; Supplementary Table 2).
For most groups, no differences were seen between the right and left amygdalae, except for the GRN (0.067

Discussion
In a cohort of patients with genetic and pathologically confirmed FTD, we have shown that AS are the most affected in MAPT/FTDP-17, Pick's disease, and TDP-43 type C groups, with the most affected subnuclei across most groups being the superficial, AB, and basal/paralaminar ones.
Our results showing amygdalar involvement across all forms of FTD are consistent with and extend the work of previous imaging and neuropathological studies [7,8,24,25]. Similarly, our work is consistent with prior studies showing particular involvement of the amygdala in patients with MAPT mutations/FTDP-17 pathology [4][5][6].
However, in this study, we were able to determine in vivo localization of the volume difference within the amygdala, where the most affected nuclei were the superficial, AB, and basal/paralaminar nuclei, which form part of the reward system, and these were also found to be connected to key limbic brain regions (Fig. 1). Impairment of the reward system is regularly seen in patients with FTD but is not universal to all people; previous studies have shown problems with changes in appetite and sexual behavior in patients with FTD due to MAPT mutations, Pick's disease, and TDP-43 type C [26], which were the most frequently diagnosed condition as reported in this study. AS are also involved in emotional processing, which is known to be impaired in both behavioral variant of frontotemporal dementia and semantic variant of primary progressive aphasia (which form the majority of the TDP-43 type C cases) [27]. Consistent with our findings, one previous functional magnetic resonance imaging study [28] has shown that specific AS associated with difficulty in recognizing The reason for differential AS involvement in the different genetic and pathological forms of FTD and how this may lead to different symptoms is currently unclear. It is likely that different neuronal networks have specific vulnerability to particular pathological proteins, and our work here suggests that specific AS may be vulnerable in different proteinopathies [29] rather than the entire amygdala. Future research, brain imaging, both on a macroscopic level and at the cellular level, is needed to investigate this further.
Limitations of this study include the use of different scanners (three manufacturers, two different magnetic fields: 1.5T and 3T) with slightly different MRI sequence types. Even if we correct for the scanner type and gender in the statistical model, we cannot completely remove some of the intrinsic heterogeneity due to these variables. We used an automated method to extract the AS volumes, which is not as accurate as their manual segmentation on dedicated MRIs or on brain tissue postmortem, but we combined the smallest AS into larger groups of nuclei to remove the effect of their less reliable delineation on T1 MRIs and manual segmentation is extremely time-consuming and labor-intensive in such a large cohort.
Future studies of functional and diffusion MRI will be helpful to investigate, in detail, the different connections of the AS in each form of FTD. Furthermore, longitudinal studies, potentially including mutation carriers in their presymptomatic stage, will help to understand the differential involvement of AS over the course of the disease.