Abstract

Progressive nonfluent aphasia (PNFA) is an early stage of frontotemporal degeneration. We identified a novel Cys521Tyr progranulin gene variant in a PNFA family that potentially disrupts disulphide bridging causing protein misfolding. To identify early neurodegeneration changes, we performed neuropsychological and neuroimaging studies in 6 family members (MRI [magnetic resonance imaging], fMRI [functional MRI], and 18f-fluorodeoxygenlucose positron emission tomography, including 4 mutation carriers, and in 9 unrelated controls. Voxel-based morphometry (VBM) of the carriers compared with controls showed significant cortical atrophy in language areas. Grey matter loss was distributed mainly in frontal lobes, being more prominent on the left. Clusters were located in the superior frontal gyri, left inferior frontal gyrus, left middle frontal gyrus, left middle temporal gyri and left posterior parietal areas, concordant with 18FDG-PET hypometabolic areas. fMRI during semantic and phonemic covert word generation (CWGTs) and word listening tasks (WLTs) showed recruitment of attentional and working memory networks in the carriers indicative of functional reorganization. During CWGTs, activation in left prefrontal cortex and bilateral anterior insulae was present whereas WLT recruited mesial prefrontal and anterior temporal cortex. These findings suggest that Cys521Tyr could be associated with early brain impairment not limited to language areas and compensated by recruitment of bilateral auxiliary cortical areas.

Introduction

Primary progressive aphasia (PPA) is an early stage of frontotemporal lobar degeneration (FTLD) characterized by predominant language deficit associated with damage to language areas in the dominant hemisphere (Mesulam 1982). The neuropathology of PPA is heterogeneous, and it is common to that of frontotemporal dementia (FTD). FTLD with ubiquitin-positive, tau-negative inclusions (FTLD-U) (McKhann et al. 2001), dementia lacking distinctive pathology, Pick's disease, and Alzheimer's disease are the most frequent causes of PPA (Mesulam 2001).

Progressive nonfluent aphasia (PNFA) is a PPA variant characterized by nonfluent language, spared language semantics and impaired phonology and syntax. An early PNFA finding is anomia that includes impaired object naming and word-finding difficulties. Global aphasia develops in 77% of PPA individuals (Mendez and Zander 1991; Josephs et al. 2006) and additional cognitive and noncognitive areas are affected over time (Josephs et al. 2006).

Although PPA is considered a sporadic disorder, up to 42% of individuals with PPA have a first-degree affected relative (Neary et al. 1993). Pedigrees reported show an autosomal dominant inheritance (Krefft et al. 2003). No data have been reported on early brain PPA changes or which brain regions appear to show early neuronal loss. Neuroimaging techniques may be used as potential markers for early diagnosis and disease staging of familial frontotemporal degeneration. Structural magnetic resonance imaging could clarify which structures are initially affected by measuring atrophy through morphometry. 18f-fluorodeoxyglucose positron emission tomography (18FDG-PET) may show the hypometabolism of affected frontotemporal areas. The extent of dysfunction can be tested by recording changes in brain blood oxygen level–dependent (BOLD) signals while the subject performs language tasks and can be correlated with the clinical symptoms. This strategy will allow us to identify specific disease patterns of language functionality and the brain regions that are most vulnerable to neurodegeneration.

Two genes in chromosome 17q21 are involved in familial FTLD. Microtubule-associated protein tau (MAPT, OMIM 157140) gene was associated with familial FTD and brain tau deposits (Hutton et al. 1998). Recently, several FTLD-U families linked to chromosome 17q21 have been found to carry GRN gene mutations (GRN, OMIM 138945) (Baker et al. 2006; Cruts et al. 2006; Mukherjee et al. 2006). GRN null mutations cause a premature mRNA stop and loss of one GRN copy, probably degraded by nonsense-mediated mRNA decay (Baker et al. 2006; Cruts et al. 2006). There are missense GRN mutations such as Ala9Asp located in the signal peptide that could alter GRN postraductional process (Gass et al. 2006; Mukherjee et al. 2006) and other missense variants that segregate with FTD which probably alter GRN protein function (Mesulam et al. 2007; van der Zee et al. 2007).

Recently, frameshift and premature stop GRN mutations have been described in 3 families with PPA (Snowden et al. 2006; Mesulam et al. 2007).

GRN encodes for a 68.5-kDa glycoprotein with abundant cysteines that can be cleaved into 8 peptides. Six peptides contain the full granulin domain; the other 2 peptides contain an imperfect and half a granulin domain (Bhandari et al. 1992). Granulin is a secreted growth factor involved in development, inflammation, tissue repair and tumorigenesis. However, its function in the brain is unknown.

We report detailed neuropsychological, neuroimaging (brain MRI, fMRI, and PET) and molecular findings of PPA in a Spanish family.

Methods

Subjects

PPA4 family was native to Spain and presented a positive history of dementia in 3 individuals include one with FTD. An additional sibling presented with language problems. Information was available from 2 generations and suggested an autosomal dominant inheritance. DNA was available for genetic analysis in 10 family members. Seven PPA4 family members underwent a medical history and a neurological exam at the Complexo Hospitalario de Pontevedra or at the Clínica Universitaria de Navarra by an experienced clinician (M.S. and P.P.). Neuropsychological assessment (Supplementary Material), diagnoses and neuroimaging studies were performed blind to genotype status. Specific interviews were carried out on language and memory complaints. Written informed consent was obtained from all individuals. All procedures were approved by the University of Navarra School of Medicine Human Studies Committee.

Sequencing Analysis of GRN Gene

Genomic DNA was isolated from whole blood according to standard procedures. All the GRN exons and their flanking regions were amplified using either previously published methods (Baker et al. 2006; Cruts et al. 2006) or by custom primers designed with the Primer3 software (http://frodo.wi.mit.edu/) (primer sequences and PCR conditions summarized in Supplementary Table 1) and sequenced in both directions on an ABI3130xl automated sequencer (Applied Biosystems, Foster City, CA). We sequenced the entire coding region of GRN in 46 independent individuals with FTD and in 150 Spanish healthy controls coming from the same regions. This sequencing included the proband of the PPA4 family who carried a novel Cys521Tyr GRN mutation (see Results).

Ten PPA4 family members with DNA available and 569 neurologically healthy independent Spanish controls were screened for the Cys521Tyr GRN mutation by using a custom TaqMan probe (TMP) (Assay Information File: AbD_SNP_2249572_446261).

Analysis of Presenilin-1 and MAPT Genes

The coding region and exon–intron boundaries of MAPT and presenilin-1 (PSEN1, OMIM 104311) genes that have been associated with familial dementia (Sherrington et al. 1995; Cruts et al. 1998; Hutton et al. 1998) were amplified using the primers previously reported (Cruts et al. 1998; Hutton et al. 1998) and sequenced as described above.

APOE Genotyping

Fluorogenic allele-specific TMPs were used for APOE isoform genotyping as described (Koch et al. 2002). Allele calling was carried out using the allelic discrimination analysis module of the ABI Sequence Detection Software (Applied Biosystems).

Magnetic Resonance Imaging

MRI studies were performed on a 1.5T Symphony (Siemens Erlangen, Germany) using the 8-channel head receiver array. Coronal T1-weighted structural MRI data was acquired using the 3D MP–RAGE standard sequence (flip angle = 15°, repetition time [RT] = 1900 ms, echo time [ET] = 3.93 ms, TI = 1100 ms, BW = 130 Hz/pixel, FOV = 250 × 188 × 216 mm3, voxel size = 0.98 × 0.98 × 1.5 mm3, matrix = 256 × 192 × 144). Functional MRI data were acquired using a BOLD sensitive 2D gradient echo echo-planar imaging (GE-EPI) sequence with standard parameters (resolution = 3 mm isotropic, FOV = 192 × 192 mm2, matrix = 64 × 64, 36 axial slices with a 0.75 mm gap, TR = 3.7 s, TE = 50 ms, BW = 2170 Hz/pixel, flip angle = 90°). Subjects performed 3 language tasks: 2 covert word generation tasks (CWGTs) that included a category-based semantic fluency task in which subjects silently generated words by category (animals, plants and objects); a letter-based phonemic fluency task in which the starting letter (P, M, and R) was provided; and a word listening task (WLT) in which they were asked to listen to a list of 2-syllable words and sounds. Sounds were generated from the same word list randomizing their frequency components obtained by spectrographic analysis. Tasks were administered in 6 blocks of alternating task-rest periods of 37 s (final run duration = 3.7 min per task). Subjects’ performance was evaluated following the fMRI scans by asking them to repeat the words generated during the scanning session.

18FDG-PET Imaging

Data were collected in 6 family members using the PET (ECAT Exact HR+; Siemens). 18FDG uptake was measured in resting condition. Following 68Ga/68Ge transmission scans, 370 MBq of 18FDG were injected as a bolus at time 0 and at 20 min. PET data acquisition started 40-min postinjection. Sixty-three planes were acquired with septa out, using a voxel size of 2.02 mm × 2.02 mm × 2.43 mm. Venous samples were obtained for glucose and radioactivity measurements.

Image Processing and Data Analysis: VBM

Structural MRI data were analyzed by statisticae parametric mapping 5 (SPM5) unified segmentation procedure (Wellcome Dept. of Imaging Neuroscience, London; http://www.fil.ion.ucl.ac.uk/spm) (Ashburner and Friston 2005) using a customized template from the entire subject sample (n = 15). Briefly, the procedure included: customized template creation (from the whole brain and of the gray matter [GM], white matter, and cerebrospinal fluid sets) from the MRI data of the entire patient and control samples; normalization and segmentation of the original scans using these customized priors; and modulation and smoothing of the resultant GM partitions, using a 8-mm Gaussian filter. The outcome variable was GM volume. Differences between the Cys521Tyr GRN mutation carriers (n = 4) and noncarriers (including the 2 noncarrier family members and 9 controls, n = 11) were assessed using the General Linear Model (GLM). GLM was implemented with a multiple regression model set up with the 2 following regressors: Cys521Tyr carrier/noncarrier and relatedness to the PPA4 family (coded as a binary trait: pedigree or nonpedigree). Each regressor modeled a different explanatory variable. The regressor that defined family relation was introduced to reduce neuroanatomical differences between family members and nonmembers independently of mutation status. The GRN mutation regressor divided the subjects in 2 groups: carriers and noncarriers. Intracranial volume was considered as a covariate of noninterest that was used to eliminate confounding effects due to total brain volume. Atrophic areas of Cys521Tyr mutation carriers were identified with a t-contrast on the GRN mutation regressor, for an uncorrected P value lower than 0.01, and for the cluster size larger than 20 voxels. Atrophic clusters were defined as the regions of interest (ROIs) for the subsequent PET ROI analysis. In order to examine the asymptomatic carriers, we repeated all neuroimaging analyses excluding the only symptomatic individual with PPA (individual III.6) (Supplementary Table 2).

18FDG-PET Data Analysis

Cerebral metabolic rate for glucose (CMRglu) was calculated from the image and blood data using the autoradiographic technique (Huang et al. 1980).

Using SPM5, individual PET data were coregistered onto the respective skull stripped structural MRI. Subsequently, images were normalized into the customized template used for MRI data normalization and, finally, modulated to compensate for the spatial normalization effect (Good et al. 2001).

CMRglu was measured in the ROIs that showed significant atrophy in the carriers according to VBM results. Regional measurements were scaled by global CMRglu. Metabolism differences between Cys521Tyr carriers (n = 4) and noncarriers (n = 2) were assessed in these ROIs, by an unpaired 2-sample t-test. We focused on comparisons between groups in predefined ROIs (not with whole brain volume) to avoid multiple comparison problems seen in VBM voxel-by-voxel volume analyses. Thus, P values lower than 0.05 were acceptable as statistically significant.

Functional Magnetic Resonance Imaging Analysis

Functional images were analyzed using SPM2 (Wellcome Dept. of Imaging Neuroscience, London; http://www.fil.ion.ucl.ac.uk/spm). Spatial preprocessing steps included realignment to correct head motion, coregistration to the anatomical image, and normalization to the custom template brain and spatial smoothing using a 10 mm isotropic full width at half maximum Gaussian kernel. Subsequently, the time series was low-pass filtered and convolved with the standard SPM hemodynamic response function. An individual voxel-wise GLM was conducted. Contrast maps of parameters (task-rest) and t-statistics maps were obtained from the 3 tasks. Random-effects group t-maps were generated by applying one sample t-test on the individual contrast images, separately for the control group and the group of noncarriers. Similar activation maps were found after analyzing semantic and syntactic word generation tasks. Therefore, data from both CWGTs were combined in subsequent analysis to maximize the statistical power. An analysis of covariance (ANCOVA) model with 7 groups (the group of controls and one group per sibling) compared brain activation in each sibling against the control group. The ANCOVA model, which merges ANOVA and regression for continuous variables, allowed introducing the number of words generated after the scanning session as covariate. Mutation effect on activation maps was assessed using a t-contrast to compare the controls with Cys521Tyr GRN mutation carriers, for an uncorrected P value < 0.001 (T > 4.79), and a cluster size larger than 30 voxels.

Results

Clinical and Neuropsychological Assessment of the PPA4 Family

The demographic and main clinical features of PPA4 family members are shown in Table 1. PPA4 relatives reported that individuals II.2 and II.3 were demented but no medical information was available. The proband (individual III.2) was diagnosed with FTD. In addition, individual III.6 complained about word-finding difficulties and was diagnosed with PPA (Supplementary Fig. 1). During the interview he showed slight language fluency impairment and occasional phonemic paraphasia. However, he had no impairment of daily living activities. Memory complaints were reported by 2 other siblings (III.3 and III.5) (Fig. 1 and Table 1). Clinical and neuroimaging studies of the PPA4 affected individuals are displayed in Supplementary Material.

Table 1

Demographic and clinical data from the PPA4 kindred

Individual/gender Age Age at onset Clinical diagnosis APOE GRN Cys521Tyr MAPT; IVS10-58A > G Neuropsychological assessment MRI/fMRI 18FDG-PET 
II.1/M Died at 89 — Normala — — — — — — 
II.2/F Died at 73 60 Dementiaa — — — — — — 
II.3/F 90 80 Dementiaa 2/3 Yes No — — — 
III.1/M 70 — Normal 2/4 Yes Yes 
III.2/M 67 52 FTD 2/4 Yes Yes — — — 
III.3/F 65 — Normal 3/4 Yes Yes 
III.4/F 63 — Normala 3/4 Yes Yes — — — 
III.5/F 60 — Normal 3/4 No No 
III.6/M 57 50 PPA 2/4 Yes Yes 
III.7/M 55 — Normal 2/4 No No 
III.8/F Died at 50 — Normala 3/4 No Yes — — — 
III.9/M 49 — Normal 2/4 Yes Yes 
Individual/gender Age Age at onset Clinical diagnosis APOE GRN Cys521Tyr MAPT; IVS10-58A > G Neuropsychological assessment MRI/fMRI 18FDG-PET 
II.1/M Died at 89 — Normala — — — — — — 
II.2/F Died at 73 60 Dementiaa — — — — — — 
II.3/F 90 80 Dementiaa 2/3 Yes No — — — 
III.1/M 70 — Normal 2/4 Yes Yes 
III.2/M 67 52 FTD 2/4 Yes Yes — — — 
III.3/F 65 — Normal 3/4 Yes Yes 
III.4/F 63 — Normala 3/4 Yes Yes — — — 
III.5/F 60 — Normal 3/4 No No 
III.6/M 57 50 PPA 2/4 Yes Yes 
III.7/M 55 — Normal 2/4 No No 
III.8/F Died at 50 — Normala 3/4 No Yes — — — 
III.9/M 49 — Normal 2/4 Yes Yes 

Note: M = male; F = female.

a

Data obtained from family information; + = test performed; — = not available.

Figure 1.

Pedigree of the PPA4 family. Proband is highlighted with an arrow.

Figure 1.

Pedigree of the PPA4 family. Proband is highlighted with an arrow.

In order to detect presymptomatic language impairment, a neuropsychological assessment battery was administered by a certified psychometrician to 6 PPA4 family members and to 9 age, gender and education matched controls to determine the standard normal scores. They were Spanish native speakers and right-handed. There was no difference in years of education across the sample. The MMSE scores were similar across the sample (Supplementary Table 3).

Cys521Tyr GRN Mutation in the PPA4 Family

GRN sequencing of 46 FTD subjects revealed a G-to-A heterozygous nucleotide substitution at the position c.1562 in the proband (III.2) of the PPA4 family, creating a novel missense Cys521Tyr variant in exon 11 (c.1562G > A). The Cys521Tyr mutation is located at the first cysteine residue of the GrnE peptide, and modifies the consensus sequence of this granulin domain (Fig. 2A). The Cys521 GRN amino acid is conserved among the vertebrate species, indicating that it has a significant role in granulin function (Fig. 2B). The heterozygous Cys521Tyr variant was present in 2 individuals with dementia, in one individual with PPA and in the 4 asymptomatic individuals of the PPA4 kindred (Table 1). The Cys521Tyr GRN mutation was absent in 569 healthy control individuals from the Spanish population, indicating that the potential disease variant occurs at a higher rate in the DFT group than in controls (Fisher test P = 0.012). We did not find any missense variants in the GRN coding region among the 150 controls sequenced except for the already described synonymous variant rs25647 (Supplementary Table 4). The lack of variants akin to Cys521Tyr (i.e., change of a conserved residue in granulin domains or a different cysteine) in the Spanish population supports that the Cys521Tyr GRN change is a potential disease-related variant.

Figure 2.

(A) Location of the Cys521Tyr GRN mutation at genomic and mRNA levels. Letters on the squares indicate the granulin peptides. (B) Consensus sequence of the granulin domains. Granulin E peptide conservation among different species. Granulin E sequences were obtained from ExPASy (http://www.expasy.org/). Conserved amino acids are highlighted in gray. (C) Chromatograms showing the wild-type and the mutant GRN DNA and cDNA.

Figure 2.

(A) Location of the Cys521Tyr GRN mutation at genomic and mRNA levels. Letters on the squares indicate the granulin peptides. (B) Consensus sequence of the granulin domains. Granulin E peptide conservation among different species. Granulin E sequences were obtained from ExPASy (http://www.expasy.org/). Conserved amino acids are highlighted in gray. (C) Chromatograms showing the wild-type and the mutant GRN DNA and cDNA.

Scores on executive functions, working memory, attention, visuospatial function, immediate recall, and verbal comprehension were normal among the Cys521Tyr GRN mutation carriers. However, language subtests measuring verbal fluency, spontaneous language (tested by narrative function), and word learning showed lower significant scores in the Cys521Tyr GRN mutation carrier group than in controls, indicating a prodromal impairment of language areas (Supplementary Table 3). The individual with slight language fluency impairment (III.6) was found to have no differences in the language subtests when compared with the other 3 Cys521Tyr GRN mutation carriers examined, which may indicate that he was still in an early stage of the disease (data not shown).

Sequencing of the MAPT gene showed a new intronic variant IVS10-58A>G in some family members. The IVS10-58A>G MAPT change was not present in one of the demented family members and Simwalk2 v2.82 haplotype analysis (Sobel and Lange 1996) inferred the IVS10-58A>G MAPT variant to one paternal chromosome, contrasting the Cys521Tyr GRN variant that was inherited from the mother with dementia. The father did not present any neurological symptoms (data not shown; Table 1 and Fig. 1). In addition, theoretical splice site analyses of the genomic sequence covering MAPT exons 9–10, containing either the A or G nucleotide at IVS10-58 position, was carried out through 3 different online sources (https://splice.cmh.edu/, http://www.fruitfly.org/seq_tools/splice.html, http://www.cbs.dtu.dk/services/NetGene2/). The IVS10-58A>G MAPT variant did not show any significant variation in the MAPT splicing process (Table 1). Sequencing of the entire coding PSEN1 gene region in the proband did not show any pathogenic mutation. APOE genotyping showed no APOE ϵ4/4 carrier, most of the family members being carriers of a single APOE ϵ4 allele (Table 1).

Neuroimaging Studies

MRI and fMRI were performed on 6 PPA4 siblings, 4 of whom were GRN mutation carriers (III.1, III.3, III.6, and III.9) and 2 noncarriers (III.5 and III.9), (mean age = 59.7 ± 7.4 years; range: 49–71) (Table 1), and on 9 healthy controls (mean age = 58.9 ± 8.1 years; range: 48–72). The 6 PPA4 siblings also underwent a 18FDG-PET scan. The effect of the GRN mutation was assessed by VBM and fMRI comparing the group of carriers (n = 4) to the control group (2 noncarrier siblings plus 9 healthy controls) in order to increase the statistical power. For the PET analysis, the carrier group was compared with the group of noncarrier siblings (n = 2).

Voxel-Based Morphometry

The effect of the GNR mutation on GM atrophy was assessed by a whole brain volume segmentation analysis using t-contrast on the mutation regressor (see Methods section for details). This contrast yielded a map of significant atrophy in the carriers compared with control group (Fig. 3A). These uncorrected threshold levels (P < 0.01, uncorrected, k > 20) were selected to obtain atrophic clusters with enough voxels to perform the subsequent 18FDG-PET ROI analysis.

Figure 3.

(A) VBM analysis. Rendered and superior view of the lateral surface of the left hemisphere, showing regions with gray matter loss in the carriers compared with the controls. Coronal and sagittal sections on the customized MR T1 template, showing anatomical details for the clusters found in the superior and middle frontal gyri, left inferior parietal lobule, left precuneus, left middle temporal gyrus and left Broca. Labels indicate anatomical localization, BAs and Talairach coordinates (x, y, z) of the voxel of maximum atrophy. t-map was thresholded at uncorrected P < 0.01, corresponding to a threshold t-value of 2.72 and a minimum cluster size k = 20. (B, C) Functional MRI analysis. Activation observed in the carriers group compared with controls, depicted in a glass brain and rendered onto sagittal and coronal sections of the customized template. (B) CWGT, with main clusters located bilaterally in anterior aspects of the insula (BA 47 and 13), left inferior frontal gyrus (BA 9), left middle frontal gyrus (BA 46) and left cingulate gyrus (BA 24). (C) WLT, with main clusters located in the mesial frontal cortex: superior frontal gyrus (BA 6), AC (BA 24 and 32), DLPFC (BA 10), left superior temporal gyrus (BA 21 and 38), and right lingual gyrus. The labels indicate anatomical localization and Talairach coordinates (x, y, z) of the voxel of maximum activity. t-maps were thresholded at P < 0.001 uncorrected (T > 4.79) and k = 30. Pre-SMA: presupplementary motor area.

Figure 3.

(A) VBM analysis. Rendered and superior view of the lateral surface of the left hemisphere, showing regions with gray matter loss in the carriers compared with the controls. Coronal and sagittal sections on the customized MR T1 template, showing anatomical details for the clusters found in the superior and middle frontal gyri, left inferior parietal lobule, left precuneus, left middle temporal gyrus and left Broca. Labels indicate anatomical localization, BAs and Talairach coordinates (x, y, z) of the voxel of maximum atrophy. t-map was thresholded at uncorrected P < 0.01, corresponding to a threshold t-value of 2.72 and a minimum cluster size k = 20. (B, C) Functional MRI analysis. Activation observed in the carriers group compared with controls, depicted in a glass brain and rendered onto sagittal and coronal sections of the customized template. (B) CWGT, with main clusters located bilaterally in anterior aspects of the insula (BA 47 and 13), left inferior frontal gyrus (BA 9), left middle frontal gyrus (BA 46) and left cingulate gyrus (BA 24). (C) WLT, with main clusters located in the mesial frontal cortex: superior frontal gyrus (BA 6), AC (BA 24 and 32), DLPFC (BA 10), left superior temporal gyrus (BA 21 and 38), and right lingual gyrus. The labels indicate anatomical localization and Talairach coordinates (x, y, z) of the voxel of maximum activity. t-maps were thresholded at P < 0.001 uncorrected (T > 4.79) and k = 30. Pre-SMA: presupplementary motor area.

GM loss in the carrier group was located predominantly in the left hemisphere and the frontal lobes, specifically in bilateral superior frontal gyrus [Zpeak = 3.78(R)/3.60(L), P < 0.0005, k = 156(R), 84(L)], left postcentral gyrus (Brodmann Area [BA] 6 and 8; Zpeak = 2.86, P < 0.002, k = 29), and left inferior frontal gyrus (BA 44; Zpeak = 3.69, P < 0.0005, k = 36). Additional atrophic regions were located in temporal and parietal lobes, especially in left middle temporal gyrus (BA 21 and 22; Zpeak = 3.63, P < 0.0005, k = 55), left inferior parietal lobule (BA 40; Zpeak = 3.60, P < 0.0005, k = 86), left superior parietal lobule and precuneus (BA 7; Zpeak = 3.0, P < 0.001, k = 31). The pattern of GM loss was mostly unchanged after excluding subject III.6. In this analysis only the left postcentral gyrus cluster became nonsignificant. The main clusters varied slightly. Their statistical Z scores and clusters volumes were: Bilateral superior frontal gyrus (Zpeak = 3.50(R)/3.57(L), P < 0.0005, k = 96(R), 67(L)); left inferior frontal gyrus (BA 44; Zpeak = 3.36, P < 0.0005, k = 21), left middle temporal gyrus (BA 21 and 22; Zpeak = 3.54, P < 0.0005, k = 48); left inferior parietal lobule (BA 40; Zpeak = 3.33, P < 0.0005, k = 46); left superior parietal lobule and precuneus (BA 7; Zpeak = 2.78, P < 0.003, k = 21). Peak Z scores in the atrophic ROIs ranged from 2.86 to 3.78 (corresponding to P < 0.002 and P < 0.0005, respectively).

18FDG-PET

The same 6 PPA4 siblings who underwent the MRI studies were subjected to 18FDG-PET scans. However, 18FDG-PET was not carried out for the healthy control subjects. Therefore, the PET data analysis was limited to comparisons between mutation carriers (n = 4) and their noncarriers siblings (n = 2). As a result of the analysis, mutation carriers showed significant hypometabolism in frontal areas, including Broca's area, left middle temporal gyrus and left middle frontal gyrus (Table 2). These areas still appear hypometabolic after excluding subject III.6 (Supplementary Table 2).

Table 2

Comparison of CMRglu between carriers (n = 4) and noncarriers (n = 2) measured in areas of atrophy obtained from VBM

Cluster Number of voxels Noncarriers Carriers P value (one-tail t-test) 
Left Broca 36 1.37 (0.06) 1.14 (0.04) 0.015* 
Left middle temporal gyrus 55 1.41 (0.07) 1.07 (0.05) 0.010* 
Left middle frontal gyrus. BA 6 84 1.22 (0.05) 1.07 (0.04) 0.037* 
Right superior frontal gyrus 156 1.08 (0.16) 0.91 (0.11) 0.204 
Left parietal lobe 29 0.67 (0.06) 0.63 (0.04) 0.291 
Left superior parietal lobule. BA 7 31 1.03 (0.09) 0.96 (0.07) 0.270 
Left inferior parietal lobule. BA 40 86 1.22 (0.13) 1.01 (0.09) 0.133 
Cluster Number of voxels Noncarriers Carriers P value (one-tail t-test) 
Left Broca 36 1.37 (0.06) 1.14 (0.04) 0.015* 
Left middle temporal gyrus 55 1.41 (0.07) 1.07 (0.05) 0.010* 
Left middle frontal gyrus. BA 6 84 1.22 (0.05) 1.07 (0.04) 0.037* 
Right superior frontal gyrus 156 1.08 (0.16) 0.91 (0.11) 0.204 
Left parietal lobe 29 0.67 (0.06) 0.63 (0.04) 0.291 
Left superior parietal lobule. BA 7 31 1.03 (0.09) 0.96 (0.07) 0.270 
Left inferior parietal lobule. BA 40 86 1.22 (0.13) 1.01 (0.09) 0.133 

Note: Data are shown as mean (standard error). Numbers in bold indicate significant differences between the carrier and control groups; *P < 0.05. CMRglu data are expressed relative to global brain metabolism.

Functional Magnetic Resonance Imaging Tests

After analyzing semantic and syntactic word generation tasks, similar activated networks were found in the mutation carrier group. Therefore, data from both tasks were combined to maximize the statistical power.

Random-effects group analysis of the CWGTs in the controls showed the maximum activation in left frontal gyrus, as expected. The carrier group showed higher bilateral insular activation than controls (Left: Z = 4.97; P < 0.0001; k = 131; Right: Z = 4.03; P < 0.0001; k = 64). The largest cluster was located in the left hemisphere covering the anterior aspect of the insula (Fig. 3B). Other activation clusters were found in left inferior frontal gyrus, middle frontal gyrus and left cingulate gyrus (Fig. 3B). Analyzed individually, the youngest carrier (individual III.9) showed greater activation than controls only in left inferior frontal gyrus (BA 47 and 13) and left middle frontal gyrus (BA 46) (−48, 26, 28) (Supplementary Fig. 3A). However, the oldest mutation carrier (individual III.1) showed a significant activation not only in left anterior insular cortex and left inferior frontal gyrus (BA 45 and 46) (−50, 26, 18), but also in the parahippocampal gyrus (BA 36) (Supplementary Fig. 3B). Frontal motor and somatosensory cortical activation was not observed in any subject, indicating absence of subvocalization during CWGTs. After excluding subject III.6 only the left cingulate gyrus cluster dropped under the set threshold for significance (T > 4.79; P < 0.001; k > 30).

Interestingly, the number of words generated after the scanning session correlated linearly with left inferior frontal gyrus activation (BA 47 and 13) in the whole study group, independently of GRN mutation status (x, y, z = −34… −34, 16, −6; Z = 4.44; P < 0.001), supporting the specificity of CWGTs for measuring language generation brain activity.

During WLT, second level analysis of random effects in the control group showed strong bilateral activation in middle temporal gyrus, predominantly on the left, as well as in left inferior and middle frontal gyrus. The carrier group showed higher activation than controls in mesial bilateral superior frontal gyrus, anterior cingulated (AC) and left superior frontal gyrus. Other clusters appeared in left superior temporal gyrus and right lingual gyrus (Fig. 3C). Again, the activation pattern was essentially unmodified after excluding the subject with word production difficulties (individual III.6). Only the cluster located in the right lingual gyrus did not reach significance at the threshold of T > 4.79, P < 0.001, k > 30.

For both fMRI whole brain volume activations during language tests, in order to generalize the obtained fMRI results to the population level, we performed random-effects analysis. Although the threshold was set at an uncorrected P value of 0.001, the maximum voxels of each cluster reached Z-values ranging from 3.35 to 4.97.

Discussion

We have characterized clinically and genetically a Spanish family with PNFA. The variable clinical expressivity among the siblings led us to investigate whether early subclinical brain changes are present among the Cys521Tyr carriers. We identified a pattern of cortical atrophy and functional language reorganization among the asymptomatic and mildly affected Cys521Tyr carriers, showing the brain degeneration map that could be associated with the novel GRN variant.

The initial aphasic disturbance of the PPA4 proband resembles Broca's aphasia with agrammatism and phonemic paraphasias. The nonverbal memory, learning and language comprehension were spared at early stages of the disease. There was neither impairment of other cognitive functions nor motor neuron signs. A brother of the proband showed minor word-finding difficulties and phonemic paraphasias.

Cys521Tyr GRN Variant in PPA4

Screening of the GRN, PSEN1, and MAPT genes in the proband (III.2) revealed a novel Cys521Tyr missense mutation in the exon 11 of the GRN gene, which was present in 7 out of 10 tested family members 2 with dementia, one with PPA and in 4 nonaffected members, but was absent in the control population. Spontaneous language, word generation, and word learning tests showed lower scores in the carrier group than in noncarriers; the differences were not significant, probably owing to the small sample size. The carrier group showed significantly lower scores than controls in word generation and word learning tests, indicating that they could be at a prodromal stage of the disease (Supplementary Table 3). Premorbid executive function alterations has been found in young MAPT mutation carriers, a highly penetrant, dominantly inherited dementing condition, as an example of regional asymmetric neurodegeneration (Geschwind and Miller 2001; Geschwind et al. 2001). Most of the Cys521Tyr carriers were nondemented, but indeed they showed neuropsychological low scores in language tests. The fact that Cys521Tyr carriers are older than MAPT mutation carriers with premorbid executive function (Geschwind et al. 2001) could be due to a reduction of neuronal vulnerability by the Cys521Tyr GRN mutation. The Cys521Tyr GRN mutation could be associated with a different regional vulnerability that modifies premorbid brain architecture, visualized by structural and functional neuroimaging.

These results indicate that the Cys521Tyr GRN mutation behaves not only with reduced penetrance, as is the case with other GRN mutations (Gass et al. 2006), but also with variable clinical expressivity. The PPA4 affected family members have a long delay from the onset of language fluency impairment to the development of global aphasia or dementia. In fact, the proband developed language problems 10 years before he showed a full clinical FTD. Frameshift and nonsense granulin mutations resulting into a loss of one GRN gene copy have been associated with PPA in 3 families recently characterized by progressive severe loss of fluency, which progressed into global aphasia in 3–6 years (Snowden et al. 2006; Mesulam et al. 2007). The PPA4 family proband showed a pattern of similar language impairment with word-finding difficulties and reduced fluency but the duration before he developed global aphasia was longer than in the other PPA families associated with GRN mutations. In addition, we observed high phenotypical variability within the PPA4 family in which some apparently healthy mutation carriers showed subclinical and neuroimaging alterations, indicating that the mutated GRN allele may still be functional to a certain degree.

Other data also support the role of Cys521Tyr GRN mutation as a probable disease variant of PPA in this family. First, the novel Cys521Tyr mutation changes a highly conserved cysteine of the GrnE peptide, modifying the consensus sequence of GrnE perfect domain and disrupting a disulphide bond necessary to maintain GrnE quaternary structure and its function (Tolkatchev et al. 2000). Next, the absence of Cys521Tyr in 569 healthy controls from the same population supports its pathogenic role. In addition, Sorting Intolerant From Tolerant (SIFT v.2) analysis (Ng and Henikoff 2003) predicted that the Cys521Tyr mutation could alter protein function (P < 0.01). This hypothesis is supported by the fact that most of the Nocth3 gene pathogenic mutations alter the location of cysteines at epidermal growth factor-like repeats altering the disulphide bridging (Federico et al. 2005; van der Zee et al. 2007). Finally, we have not found any missense variant in 150 healthy subjects apart from some intronic or synonymous single nucleotide polymorphisms already reported indicating the importance of conservation of cysteines at the granulin domains for protein function.

A missense GRN mutation (Ala9Asp) located in the signal peptide could alter the GRN postraductional process (Gass et al. 2006; Mukherjee et al. 2006); other missense variants present in familial FTD alter GRN protein production and secretion (Gass et al. 2006; Mesulam et al. 2007; Spina et al. 2007; van der Zee et al. 2007; Shankaran et al. 2008). Arg433Cys GRN mutation modifies the number of cysteines in the full granulin protein and could have a similar pathogenic mechanism to the Cys521Tyr GRN mutation (van der Zee et al. 2007). Sequencing of a large number of cases and controls in North American and European populations led to the identification of 2 missense cysteine variants whose implication in GRN function is uncertain. The Cys105Arg located in the granulin G domain is absent in controls but does not segregate with the disease (Gass et al. 2006; Spina et al. 2007). The Cys158Tyr variant is located in the granulin F domain and is present in 15% of controls (Federico et al. 2005; Gass et al. 2006; van der Zee et al. 2007). The lack of pathogenicity of these cysteine variants could be explained because the crystal structure for the 31 first amino acids of the GRN A (1g26.pdb) indicate that, as in the case of the Cys521, the first cysteines of the GRN domains are implicated in a disulfide bond; in contrast, other cysteines, such as cysteine number 11 (Cys105) and cysteine number 7 (Cys158), have an unknown role in the formation of disulfide bonds necessary to maintain the quaternary structure of granulin domains. These data indicate that variation effects could depend on their impact on GRN protein structure and stability. In addition, it has been described that different GRN peptides can display different functions (Plowman et al. 1992) indicating that the cysteine variants at granulins F, G, and E in brain could have different effects depending on their location across the gene (Federico et al. 2005; van der Zee et al. 2007). Subsequent studies about the specific function of the different granulins in brain and the modification of the quaternary structure of the mutated granulins will help to understand the pathogenic mechanism of the GRN missense mutations. These data support the role of GRN gene in PPA4 family.

Extensive analysis of other genes involved in familial dementia such as MAPT and PSEN1 genes (Cruts et al. 1998; Hutton et al. 1998) excluded other coding pathogenic variants associated with the PPA4 family. However, we found a nucleotide substitution in MAPT (IVS10-58A>G), which is probably a rare benign variant because of its lack of segregation with the disease.

GRN gene mutations have been associated with a wide range of disease onset (average range: 48–83 years) (Gass et al. 2006) as occurs in the PPA4 family, pointing to additional genetic or environmental factors that may be involved in the disease's presentation and severity. The absence of APOE ϵ4/4 genotype carriers suggests that APOE gene is not a major modifying genetic risk factor. However, the presence of an APOE ϵ4 allele in most of the family members could still represent a precipitating factor for dementia in the PPA4 family.

Cortical Atrophy and Cerebral Hypometabolism in Cys521Tyr Carriers

One of the main objectives of this study was to detect early specific atrophic or dysfunctional cortical brain areas associated with the Cys521Tyr GRN mutation in order to evaluate the mutation effects. Because neuropsychological test scores were similar across the Cys521Tyr mutation carriers, the scores were grouped together to be compared against the controls. Atrophic regions identified by VBM in the carrier group were specifically located in primary language centers and included the cerebral network related to language processing (Indefrey and Levelt 2004) (Fig. 3A). Besides the areas classically involved in word generation, the carrier group showed atrophic changes in temporal and parietal areas previously identified as the phonological memory store and correlated with word learning impairment (Paulesu et al. 1993). GM loss in the PPA4 family was located in the depth of the sulci suggesting where the degeneration process associated with the Cys521Tyr mutation is triggered. This GM loss pattern is different from aging brain atrophy, characterized by an increase in sulcal width and a decrease in sulcal depth resulting in cortical shrinkage and decreased brain volume (Kochunov et al. 2005). VBM studies on PNFA have identified a neurodegeneration pattern in inferior left frontal regions including the insula, followed by left anterior and ventral temporal cortex (Rosen et al. 2002; Sonty et al. 2003). Most studies included PPA patients with heterogeneous etiology and in advanced stages of disease, which introduced a high variability in their analyses and meant that early atrophic changes as well as specific etiologic patterns were missing. However, in our study we have potentially detected presymptomatic atrophic changes in a genetically and environmentally homogeneous sample with better preservation of language.

In the carrier group, the main areas with significant GM loss previously identified in the VBM study showed significant hypometabolism in 18FDG-PET ROI analysis (Table 2). Hypometabolic regions such as left inferior frontal and middle temporal areas corroborate the findings of other 18FDG-PET studies (Soriani-Lefevre et al. 2003). Hypometabolism extended to nonlanguage areas responsible for executive functions has previously been interpreted as a sign of advanced disease (Perneczky et al. 2007). Our study identifies hypometabolic and atrophic cortical areas outside the language network in Cys521Tyr GRN carriers that could instead be related to prodromal stages of PPA.

Language Network Reorganization in Cys521Tyr Carriers

Functional MRI strategies to perform basic language functions in carriers consisted of extra activation and recruitment of additional neural populations compared with age and education matched controls. During fMRI CWGT, activation clusters were found in both anterior insulae with a greater activation on the left insula considered the final common pathway in language articulation. This region has been associated with abnormal motor planning of speech in patients with focal lesions (Dronkers et al. 2004). The activation of both insulae in Cys521Tyr carriers indicates a supplementary effort of word articulation planning. Recruitment of contralateral areas has been found to be a plasticity mechanism in chronic aphasic stroke patients associated with language comprehension (Crinion and Price 2005). Additionally, the dorsolateral prefrontal cortex (DLPFC) activation might reflect an increased working memory load to control word generation lists. The additional activation in the left AC probably reflects attentional demands of the task or error monitoring (Pardo et al. 1990; Phelps et al. 1997). Interestingly, the overall number of words generated by the carriers was lower than the number generated by controls during the same amount of time, even though they recruited cognitive network areas to perform the task (Fig. 3B,C). Connectivity analyses during language tasks in advanced disease PPA series have shown a functional disruption between superior temporal and inferior frontal areas despite apparently normal levels of BOLD activation (Sonty et al. 2007). Our fMRI findings suggest that increased BOLD signal and recruitment of new areas precedes uncoupling of language areas in PPA.

During WLT controls showed extensive activation not only in the middle temporal gyri but also in left frontal opercular including Broca's region, as described before (Hickok and Poeppel 2000; Indefrey and Levelt 2004). We measured the early language processing through WLT; this requires both passive listening and discrimination between sounds and words, but not specific demands of phonemic discrimination or word learning.

Cys521Tyr carriers showed a significantly higher synaptic demand by activating left prefrontal areas such as the presupplementary motor area, AC and DLPFC. These areas are associated with task-relevant contents of memory (MacDonald and Joordens 2000), and on-line monitoring, error detection, and response execution tasks (Botvinick et al. 2001). In addition, the carriers recruited the anterior superior temporal gyrus, which is a multimodal area related to phonetic information (Fig. 3C) (Spitsyna et al. 2006). The recruitment of right lingual gyrus suggests a multimodal mechanism to improve word discrimination by adding word shape processing to the auditory perception (Fink et al. 1996).

Our functional studies of word generation and listening suggest that the brain could compensate for the dysfunction of language areas with bilateral recruitment of working memory and attention networks. This compensatory mechanism could explain the absence of noticeable language impairment in some of the Cys521Tyr GRN mutation carriers.

Limitations of the Study

Neuroimaging techniques could potentially be sensitive and specific to detect early brain dysfunction in late-onset inherited neurodegenerative diseases; however, limitations of our study arise from the lack of a large sample size for the imaging segment of the study and from the lack of additional DFT families with the Cys521Tyr variant. These findings must be interpreted with great caution and should be investigated in other PPA kindreds with GRN gene variants. Detection of early brain changes associated with the new GRN mutation will facilitate early application of targeted disease therapies that could prevent neurodegeneration. Follow-up of the PPA4 family will determine the long-term effects of the Cys521Tyr variant.

Supplementary Material

Supplementary material can be found at: http://www.cercor.oxfordjournals.org/

Funding

Government of Navarra (“Ayudas para la Realización de Proyectos de Investigación” 2006–2007) to P.P.; and Government of Navarra Postdoctoral Fellow (2005–2007) to C.C.

We thank the PPA4 family members and the controls for their collaboration. We thank Dr Pablo Martinez-Lage (Fundacion ACE, Barcelona, Spain) for providing many samples for the genetic analysis, Dr Francisca Lahortiga for her neuropsychological advice and Dr Secundino Fernández for his work on language spectrographic analysis and auditory stimuli processing. Conflict of Interest: None declared.

References

Ashburner
J
Friston
KJ
Unified segmentation
Neuroimage.
 , 
2005
, vol. 
26
 (pg. 
839
-
851
)
Baker
M
Mackenzie
IR
Pickering-Brown
SM
Gass
J
Rademakers
R
Lindholm
C
Snowden
J
Adamson
J
Sadovnick
AD
Rollinson
S
, et al.  . 
Mutations in progranulin cause tau-negative frontotemporal dementia linked to chromosome 17
Nature.
 , 
2006
, vol. 
442
 (pg. 
916
-
919
)
Bhandari
V
Palfree
RG
Bateman
A
Isolation and sequence of the granulin precursor cDNA from human bone marrow reveals tandem cysteine-rich granulin domains
Proc Natl Acad Sci USA.
 , 
1992
, vol. 
89
 (pg. 
1715
-
1719
)
Botvinick
MM
Braver
TS
Barch
DM
Carter
CS
Cohen
JD
Conflict monitoring and cognitive control
Psychol Rev.
 , 
2001
, vol. 
108
 (pg. 
624
-
652
)
Crinion
J
Price
CJ
Right anterior superior temporal activation predicts auditory sentence comprehension following aphasic stroke
Brain.
 , 
2005
, vol. 
128
 (pg. 
2858
-
2871
)
Cruts
M
Gijselinck
I
van der Zee
J
Engelborghs
S
Wils
H
Pirici
D
Rademakers
R
Vandenberghe
R
Dermaut
B
Martin
JJ
, et al.  . 
Null mutations in progranulin cause ubiquitin-positive frontotemporal dementia linked to chromosome 17q21
Nature.
 , 
2006
, vol. 
442
 (pg. 
920
-
924
)
Cruts
M
van Duijn
CM
Backhovens
H
Van den Broeck
M
Wehnert
A
Serneels
S
Sherrington
R
Hutton
M
Hardy
J
St George-Hyslop
PH
, et al.  . 
Estimation of the genetic contribution of presenilin-1 and -2 mutations in a population-based study of presenile Alzheimer disease
Hum Mol Genet.
 , 
1998
, vol. 
7
 (pg. 
43
-
51
)
Dronkers
NF
Wilkins
DP
Van Valin
RD
Jr
Redfern
BB
Jaeger
JJ
Lesion analysis of the brain areas involved in language comprehension
Cognition.
 , 
2004
, vol. 
92
 (pg. 
145
-
177
)
Federico
A
Bianchi
S
Dotti
MT
The spectrum of mutations for CADASIL diagnosis
Neurol Sci.
 , 
2005
, vol. 
26
 (pg. 
117
-
124
)
Fink
GR
Halligan
PW
Marshall
JC
Frith
CD
Frackowiak
RS
Dolan
RJ
Where in the brain does visual attention select the forest and the trees?
Nature.
 , 
1996
, vol. 
382
 (pg. 
626
-
628
)
Gass
J
Cannon
A
Mackenzie
IR
Boeve
B
Baker
M
Adamson
J
Crook
R
Melquist
S
Kuntz
K
Petersen
R
, et al.  . 
Mutations in progranulin are a major cause of ubiquitin-positive frontotemporal lobar degeneration
Hum Mol Genet.
 , 
2006
, vol. 
15
 (pg. 
2988
-
3001
)
Geschwind
DH
Miller
BL
Molecular approaches to cerebral laterality: development and neurodegeneration
Am J Med Genet.
 , 
2001
, vol. 
101
 (pg. 
370
-
381
)
Geschwind
DH
Robidoux
J
Alarcon
M
Miller
BL
Wilhelmsen
KC
Cummings
JL
Nasreddine
ZS
Dementia and neurodevelopmental predisposition: cognitive dysfunction in presymptomatic subjects precedes dementia by decades in frontotemporal dementia
Ann Neurol.
 , 
2001
, vol. 
50
 (pg. 
741
-
746
)
Good
CD
Johnsrude
IS
Ashburner
J
Henson
RN
Friston
KJ
Frackowiak
RS
A voxel-based morphometric study of ageing in 465 normal adult human brains
Neuroimage.
 , 
2001
, vol. 
14
 (pg. 
21
-
36
)
Hickok
G
Poeppel
D
Towards a functional neuroanatomy of speech perception
Trends Cogn Sci.
 , 
2000
, vol. 
4
 (pg. 
131
-
138
)
Huang
SC
Phelps
ME
Hoffman
EJ
Sideris
K
Selin
CJ
Kuhl
DE
Noninvasive determination of local cerebral metabolic rate of glucose in man
Am J Physiol.
 , 
1980
, vol. 
238
 (pg. 
E69
-
82
)
Hutton
M
Lendon
CL
Rizzu
P
Baker
M
Froelich
S
Houlden
H
Pickering-Brown
S
Chakraverty
S
Isaacs
A
Grover
A
, et al.  . 
Association of missense and 5′-splice-site mutations in tau with the inherited dementia FTDP-17
Nature.
 , 
1998
, vol. 
393
 (pg. 
702
-
705
)
Indefrey
P
Levelt
WJ
The spatial and temporal signatures of word production components
Cognition.
 , 
2004
, vol. 
92
 (pg. 
101
-
144
)
Josephs
KA
Duffy
JR
Strand
EA
Whitwell
JL
Layton
KF
Parisi
JE
Hauser
MF
Witte
RJ
Boeve
BF
Knopman
DS
, et al.  . 
Clinicopathological and imaging correlates of progressive aphasia and apraxia of speech
Brain.
 , 
2006
, vol. 
129
 (pg. 
1385
-
1398
)
Koch
W
Ehrenhaft
A
Griesser
K
Pfeufer
A
Muller
J
Schomig
A
Kastrati
A
TaqMan systems for genotyping of disease-related polymorphisms present in the gene encoding apolipoprotein E
Clin Chem Lab Med.
 , 
2002
, vol. 
40
 (pg. 
1123
-
1131
)
Kochunov
P
Mangin
JF
Coyle
T
Lancaster
J
Thompson
P
Riviere
D
Cointepas
Y
Regis
J
Schlosser
A
Royall
DR
, et al.  . 
Age-related morphology trends of cortical sulci
Hum Brain Mapp.
 , 
2005
, vol. 
26
 (pg. 
210
-
220
)
Krefft
TA
Graff-Radford
NR
Dickson
DW
Baker
M
Castellani
RJ
Familial primary progressive aphasia
Alzheimer Dis Assoc Disord.
 , 
2003
, vol. 
17
 (pg. 
106
-
112
)
MacDonald
PA
Joordens
S
Investigating a memory-based account of negative priming: support for selection-feature mismatch
J Exp Psychol Hum Percept Perform.
 , 
2000
, vol. 
26
 (pg. 
1478
-
1496
)
McKhann
GM
Albert
MS
Grossman
M
Miller
B
Dickson
D
Trojanowski
JQ
Clinical and pathological diagnosis of frontotemporal dementia: report of the Work Group on Frontotemporal Dementia and Pick's Disease
Arch Neurol.
 , 
2001
, vol. 
58
 (pg. 
1803
-
1809
)
Mendez
MF
Zander
BA
Dementia presenting with aphasia: clinical characteristics
J Neurol Neurosurg Psychiatry.
 , 
1991
, vol. 
54
 (pg. 
542
-
545
)
Mesulam
M
Johnson
N
Krefft
TA
Gass
JM
Cannon
AD
Adamson
JL
Bigio
EH
Weintraub
S
Dickson
DW
Hutton
ML
, et al.  . 
Progranulin mutations in primary progressive aphasia: the PPA1 and PPA3 families
Arch Neurol.
 , 
2007
, vol. 
64
 (pg. 
43
-
47
)
Mesulam
MM
Slowly progressive aphasia without generalized dementia
Ann Neurol.
 , 
1982
, vol. 
11
 (pg. 
592
-
598
)
Mesulam
MM
Primary progressive aphasia
Ann Neurol.
 , 
2001
, vol. 
49
 (pg. 
425
-
432
)
Mukherjee
O
Pastor
P
Cairns
NJ
Chakraverty
S
Kauwe
JS
Shears
S
Behrens
MI
Budde
J
Hinrichs
AL
Norton
J
, et al.  . 
HDDD2 is a familial frontotemporal lobar degeneration with ubiquitin-positive, tau-negative inclusions caused by a missense mutation in the signal peptide of progranulin
Ann Neurol.
 , 
2006
, vol. 
60
 (pg. 
314
-
322
)
Neary
D
Snowden
JS
Mann
DM
Familial progressive aphasia: its relationship to other forms of lobar atrophy
J Neurol Neurosurg Psychiatry.
 , 
1993
, vol. 
56
 (pg. 
1122
-
1125
)
Ng
PC
Henikoff
S
SIFT: Predicting amino acid changes that affect protein function
Nucleic Acids Res.
 , 
2003
, vol. 
31
 (pg. 
3812
-
3814
)
Pardo
JV
Pardo
PJ
Janer
KW
Raichle
ME
The anterior cingulate cortex mediates processing selection in the Stroop attentional conflict paradigm
Proc Natl Acad Sci USA.
 , 
1990
, vol. 
87
 (pg. 
256
-
259
)
Paulesu
E
Frith
CD
Frackowiak
RS
The neural correlates of the verbal component of working memory
Nature.
 , 
1993
, vol. 
362
 (pg. 
342
-
345
)
Perneczky
R
Diehl-Schmid
J
Pohl
C
Drzezga
A
Kurz
A
Non-fluent progressive aphasia: cerebral metabolic patterns and brain reserve
Brain Res.
 , 
2007
, vol. 
1133
 (pg. 
178
-
185
)
Phelps
EA
Hyder
F
Blamire
AM
Shulman
RG
FMRI of the prefrontal cortex during overt verbal fluency
Neuroreport.
 , 
1997
, vol. 
8
 (pg. 
561
-
565
)
Plowman
GD
Green
JM
Neubauer
MG
Buckley
SD
McDonald
VL
Todaro
GJ
Shoyab
M
The epithelin precursor encodes two proteins with opposing activities on epithelial cell growth
J Biol Chem.
 , 
1992
, vol. 
267
 (pg. 
13073
-
13078
)
Rosen
HJ
Gorno-Tempini
ML
Goldman
WP
Perry
RJ
Schuff
N
Weiner
M
Feiwell
R
Kramer
JH
Miller
BL
Patterns of brain atrophy in frontotemporal dementia and semantic dementia
Neurology.
 , 
2002
, vol. 
58
 (pg. 
198
-
208
)
Shankaran
SS
Capell
A
Hruscha
AT
Fellerer
K
Neumann
M
Schmid
B
Haass
C
Missense mutations in the progranulin gene linked to frontotemporal lobar degeneration with ubiquitin-immunoreactive inclusions reduce progranulin production and secretion
J Biol Chem.
 , 
2008
, vol. 
283
 (pg. 
1744
-
1753
)
Sherrington
R
Rogaev
EI
Liang
Y
Rogaeva
EA
Levesque
G
Ikeda
M
Chi
H
Lin
C
Li
G
Holman
K
, et al.  . 
Cloning of a gene bearing missense mutations in early-onset familial Alzheimer's disease
Nature.
 , 
1995
, vol. 
375
 (pg. 
754
-
760
)
Snowden
JS
Pickering-Brown
SM
Mackenzie
IR
Richardson
AM
Varma
A
Neary
D
Mann
DM
Progranulin gene mutations associated with frontotemporal dementia and progressive non-fluent aphasia
Brain.
 , 
2006
, vol. 
129
 (pg. 
3091
-
3102
)
Sobel
E
Lange
K
Descent graphs in pedigree analysis: applications to haplotyping, location scores, and marker-sharing statistics
Am J Hum Genet.
 , 
1996
, vol. 
58
 (pg. 
1323
-
1337
)
Sonty
SP
Mesulam
MM
Thompson
CK
Johnson
NA
Weintraub
S
Parrish
TB
Gitelman
DR
Primary progressive aphasia: PPA and the language network
Ann Neurol.
 , 
2003
, vol. 
53
 (pg. 
35
-
49
)
Sonty
SP
Mesulam
MM
Weintraub
S
Johnson
NA
Parrish
TB
Gitelman
DR
Altered effective connectivity within the language network in primary progressive aphasia
J Neurosci.
 , 
2007
, vol. 
27
 (pg. 
1334
-
1345
)
Soriani-Lefevre
MH
Hannequin
D
Bakchine
S
Menard
JF
Manrique
A
Hitzel
A
Kotzki
PO
Boudousq
V
Vera
P
Evidence of bilateral temporal lobe involvement in primary progressive aphasia: a SPECT study
J Nucl Med.
 , 
2003
, vol. 
44
 (pg. 
1013
-
1022
)
Spina
S
Murrell
JR
Huey
ED
Wassermann
EM
Pietrini
P
Baraibar
MA
Barbeito
AG
Troncoso
JC
Vidal
R
Ghetti
B
, et al.  . 
Clinicopathologic features of frontotemporal dementia with progranulin sequence variation
Neurology.
 , 
2007
Spitsyna
G
Warren
JE
Scott
SK
Turkheimer
FE
Wise
RJ
Converging language streams in the human temporal lobe
J Neurosci.
 , 
2006
, vol. 
26
 (pg. 
7328
-
7336
)
Tolkatchev
D
Ng
A
Vranken
W
Ni
F
Design and solution structure of a well-folded stack of two beta-hairpins based on the amino-terminal fragment of human granulin A
Biochemistry.
 , 
2000
, vol. 
39
 (pg. 
2878
-
2886
)
van der Zee
J
Le Ber
I
Maurer-Stroh
S
Engelborghs
S
Gijselinck
I
Camuzat
A
Brouwers
N
Vandenberghe
R
Sleegers
K
Hannequin
D
, et al.  . 
Mutations other than null mutations producing a pathogenic loss of progranulin in frontotemporal dementia
Hum Mutat.
 , 
2007
, vol. 
28
 pg. 
416
 

Author notes

Dr. Maria A. Pastor and Dr. Pau Pastor contributed equally to this article.