A JOURNAL OF NEUROLOGY
Corticospinal tract degeneration associated with
TDP-43 type C pathology and semantic dementia
Keith A. Josephs,1,2Jennifer L. Whitwell,3Melissa E. Murray,4Joseph E. Parisi,5
Neill R. Graff-Radford,6David S. Knopman,1Bradley F. Boeve,1Matthew L. Senjem,7
Rosa Rademakers,8Clifford R. Jack Jr,3Ronald C. Petersen1and Dennis W. Dickson4
1 Behavioural Neurology, Department of Neurology, Mayo Clinic, Rochester, MN 55905, USA
2 Movement Disorders, Department of Neurology, Mayo Clinic, Rochester, MN 55905, USA
3 Department of Radiology, Mayo Clinic, Rochester, MN 55905, USA
4 Neuropathology, Department of Neuroscience, Mayo Clinic, Jacksonville, FL 32224, USA
5 Department of Laboratory of Medicine and Pathology, Mayo Clinic, Rochester, MN 55905, USA
6 Department of Neurology, Mayo Clinic, Jacksonville, FL 32224, USA
7 Department of Information Technology, Mayo Clinic, Rochester, MN 55905, USA
8 Molecular Genetics, Department of Neuroscience, Mayo Clinic, Jacksonville, FL 32224, USA
Correspondence to: Keith A. Josephs, MD, MST, MSc,
Professor of Neurology,
Four subtypes of frontotemporal lobar degeneration with TDP-43 immunoreactive inclusions have been described (types A–D).
Of these four subtypes, motor neuron disease is more commonly associated with type B pathology, but has also been reported
with type A pathology. We have noted, however, the unusual occurrence of cases of type C pathology having corticospinal tract
degeneration. We aimed to assess the severity of corticospinal tract degeneration in a large cohort of cases with type C (n = 31).
Pathological analysis included semi-quantitation of myelin loss of fibres of the corticospinal tract and associated macrophage
burden, as well as axonal loss, at the level of the medullary pyramids. We also assessed for motor cortex degeneration and fibre
loss of the medial lemniscus/olivocerebellar tract. All cases were subdivided into three groups based on the degree of corti-
cospinal tract degeneration: (i) no corticospinal tract degeneration; (ii) equivocal corticospinal tract degeneration; and (iii)
moderate to very severe corticospinal tract degeneration. Clinical, genetic, pathological and imaging comparisons were per-
formed across groups. Eight cases had no corticospinal tract degeneration, and 14 cases had equivocal to mild corticospinal tract
degeneration. Nine cases, however, had moderate to very severe corticospinal tract degeneration with myelin and axonal loss. In
these nine cases, there was degeneration of the motor cortex without lower motor neuron degeneration or involvement of other
brainstem tracts. These cases most commonly presented as semantic dementia, and they had longer disease duration (mean:
15.3 years) compared with the other two groups (10.8 and 9.9 years; P = 0.03). After adjusting for disease duration, severity of
corticospinal tract degeneration remained significantly different across groups. Only one case, without corticospinal tract de-
generation, was found to have a hexanucleotide repeat expansion in the C9ORF72 gene. All three groups were associated with
anterior temporal lobe atrophy on MRI; however, the cases with moderate to severe corticospinal tract degeneration showed
right-sided temporal lobe asymmetry and greater involvement of the right temporal lobe and superior motor cortices than the
other groups. In contrast, the cases with no or equivocal corticospinal tract degeneration were more likely to show left-sided
temporal lobe asymmetry. For comparison, the corticospinal tract was assessed in 86 type A and B cases, and only two cases
doi:10.1093/brain/aws324Brain 2013: Page 1 of 16 |
Received May 14, 2012. Revised September 19, 2012. Accepted October 11, 2012
? The Author (2013). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved.
For Permissions, please email: firstname.lastname@example.org
Brain Advance Access published January 28, 2013
at Mayo Clinic Library on January 29, 2013
showed evidence of corticospinal tract degeneration without lower motor neuron degeneration. These findings confirm that
there exists a unique association between frontotemporal lobar degeneration with type C pathology and corticospinal tract
degeneration, with this entity showing a predilection to involve the right temporal lobe.
Keywords: TDP-43 type C; corticospinal tract; MRI; semantic dementia; right temporal lobe
Abbreviations: CSTD = corticospinal tract degeneration; FTD = frontotemporal dementia; FTLD = frontotemporal lobar degener-
ation; MND = motor neuron degeneration
The frontotemporal lobar degenerations (FTLD) are a group of het-
erogeneous pathological disorders that are associated with the fron-
totemporal dementia (FTD) clinical syndromes (Neary et al., 1998;
McKhann et al., 2001; Josephs, 2008; Josephs et al., 2011). Three
main proteins have been associated with FTLD, including microtu-
bule-associated protein tau (Hutton et al., 1998), transactive
response-DNA–binding protein of 43kDa (TDP-43) (Arai et al.,
2006; Neumann et al., 2006) and fused in sarcoma protein
(Neumann et al., 2009). Pathological disorders associated with
TDP-43 are subsumed under the umbrella term FTLD-TDP
(Mackenzie et al., 2009; Josephs et al., 2011) and account for
the majority of cases of FTLD (Josephs et al., 2004; Lipton et al.,
FTLD-TDP can be further subclassified by assessing the distribu-
tion and morphological features of the TDP-43 immunoreactive
inclusions that define FTLD-TDP, in frontotemporal neocortex
and in hippocampal dentate granular cells (Mackenzie et al.,
2006; Sampathu et al., 2006). Initially, two different classification
schemes were published (Mackenzie et al., 2006; Sampathu et al.,
2006), although a recent harmonized classification scheme has
linked the two initial schemes (Mackenzie et al., 2011). Based
on the harmonized scheme, four subtypes of FTLD-TDP (types
A–D) are recognized (Mackenzie et al., 2011). FTLD-TDP type A
(MacKenzie type 1, Sampathu type 3) is characterized by a mix-
ture of neuronal cytoplasmic inclusions, short dystrophic neurites
and intranuclear inclusions; FTLD-TDP type B (MacKenzie type 3,
Sampathu type 2) by a predominance of neuronal cytoplasmic
inclusions; FTLD-TDP type C (MacKenzie type 2, Sampathu 1)
by a predominance of long thick dystrophic neurites in the neo-
cortex and Pick body-like inclusions in the dentate granular cells of
the hippocampus and FTLD-TDP type D by a predominance of
intranuclear inclusions. In addition, cases of FTLD-TDP can be in-
dependently classified as having motor neuron degeneration
(FTLD-MND) (Josephs et al., 2011). Cases of FTLD-MND have
characteristic histological features of FTLD, as well as characteristic
features of MND, such as Bunina bodies, TDP-43 immunoreactive
inclusions in hypoglossal nucleus or anterior horn cells, skein-like
and Lewy body-like inclusions, anterior horn cell loss and corti-
cospinal tract degeneration (CSTD) (Josephs et al., 2006). Cases of
FTLD-MND are almost always associated with type B pathology
(Mackenzie et al., 2006; Snowden et al., 2007; Geser et al., 2009;
Josephs et al., 2009), although a few cases have been associated
with type A pathology (Cairns et al., 2007; Whitwell et al., 2010),
particularlywhenassociated witha hexanucleotiderepeat
expansion sequence in the chromosome 9 opening reading
frame (C9ORF72) gene (Murray et al., 2011).
For the past 3 years, however, we have observed cases of
FTLD-TDP type C pathology associated with severe corticospinal
tract degeneration (CSTD) but without other features of MND
(Josephs et al., 2009). Such cases have not been previously
emphasized. To better understand this unusual combination of
histological features, we aimed to assess the severity of CSTD in
a large cohort of cases with FTLD-TDP type C, and to determine
what proportion of cases have moderate to severe CSTD. This is
important, as cases with FTLD-TDP type C with moderate to
severe CSTD would be atypical and could represent a distinct
FTLD entity. We aimed to assess clinical, genetic and imaging
features of these cases compared with those with absent or
equivocal to mild CSTD.
Materials and methods
The neuropathological databases at Mayo Clinic Rochester, MN and
Jacksonville, FL, USA were queried to identify all cases with FTLD-TDP
type C pathology. A total of 32 cases were identified.
Standard neuropathological procedure
All patients underwent standardized neuropathological evaluation
using the recommended diagnostic protocol for Alzheimer’s disease
(Mirra et al., 1991). Pathological diagnoses were conducted by one
of our two experienced neuropathologists (J.E.P. or D.W.D.). After
removal, the brain was divided into right and left hemibrains. One
hemibrain was fixed in 10% buffered formaldehyde for 7 to 10
days, and then sectioned. In all cases, we sampled mid-frontal, super-
ior temporal, inferior parietal and occipital cortex, amygdala, hippo-
campus and cingulate gyrus, nucleus basalis, basal ganglia, thalamus,
midbrain, pons, medulla and cerebellum. Samples were processed in
paraffin and stained with haematoxylin and eosin and modified
Bielschowsky or Gallyas silver impregnation, and they were immunos-
tained with antibodies to amyloid-b (clone 6F/3D, 1:10 dilution;
Novocastra Vector Labs), phospo-tau (CP-13, 1:100, Peter Davies,
Albert Einstein College of Medicine or clone AT8, 1:1000 dilution;
Endogen), ?-synuclein [NACP, 1:3000 (Gwinn-Hardy et al., 2000;
Beach et al., 2008) or LB509, 1:200 dilution; Zymed], neurofilament
(SMI-31, 1:20 000, Sternberger–Meyer or clone 2F11, 1:75 dilution;
Dako), ubiquitin (Ubi-1, 1:40 000, EnCor Biotechnology or Dako poly-
clonal, 1:100 dilution) and TDP-43 (rabbit polyclonal, 1:3000;
ProteinTech group or a custom made rabbit polyclonal to C-terminal
epitope; Zhang et al., 2009).
Brain 2013: Page 2 of 16K. A. Josephs et al.
at Mayo Clinic Library on January 29, 2013
All 32 cases were re-stained for this study with an anti-TDP-43
antibody to a C-terminal epitope (Zhang et al., 2009) that permits
visualization of pathological TDP-43 without normal nuclear staining,
and they were re-reviewed by one neuropathologist (D.W.D.) to
ensure that all cases met the most recent published criteria for
FTLD-TDP type C pathology (Mackenzie et al., 2011). Specifically,
there had to be long thick dystrophic neurites throughout laminar I–
VI of the frontotemporal neocortex and Pick body-like neuronal cyto-
plasmic inclusions in dentate granular cells of the hippocampus (Fig. 1)
(Josephs et al., 2009). Features of MND, such as Bunina bodies,
TDP-43 immunoreactive inclusions in hypoglossal nucleus, nucleus
ambiguus or anterior horn cells, skein-like and Lewy body-like inclu-
sions, anterior horn cell loss and CSTD, were assessed. The presence of
neuronal intranuclear inclusions was assessed. Eleven randomly se-
lected cases were independently reviewed by another neuropathologist
(J.E.P.) with 100% concordance for the diagnosis of FTLD-TDP type C.
To assess the specificity of the findings to FTLD-TDP type C path-
ology, we also assessed the corticospinal tract and hypoglossal nucleus
of the medulla in 86 cases of FTLD-TDP type A (n = 61) and B
(n = 25). Type A cases were characterized by the presence of dys-
trophic neurites and neuronal cytoplasmic inclusions typically asso-
ciated with neuronal intranuclear inclusions (Mackenzie et al., 2009,
2011). Type B cases were characterized by the presence of neuronal
(Mackenzie et al., 2009, 2011).
The study was approved by the Mayo Clinic Institutional Review
Board, and in all cases, informed consent was given before death.
absence of dystrophicneurites
Semi-quantitative assessment of
corticospinal tract degeneration
A paraffin section of brainstem medulla was used to generate slides for
histological assessment of the corticospinal tract at the level of the
medullary pyramids. In one case, a medulla section was not available;
hence, this case was excluded from further analysis. Slides of medulla
of the remaining 31 cases were stained with Luxol fast blue to assess
for the degree of myelin loss. Immunohistochemistry with ionized cal-
cium binding adaptor molecule 1 (IBA-1) was performed to assess for
the presence and burden of activated and phagocytic microglia
(Ahmed et al., 2007). A semi-quantitative six-point scale was used
to determine the degree of myelin loss observed with Luxol fast
blue stain as follows: 0 = no myelin loss; 0.5 = slight loss; 1 = mild
loss; 2 = moderate loss; 3 = severe loss; and 4 = very severe loss. A
similar six-point scale was used to semi-quantitate activated/phago-
cytic microglia burden as follows: 0 = no activated/phagocytic micro-
glia observed; 0.5 = scant number of activated/phagocytic microglia;
1 = small number of activated/phagocytic microglia; 2 = a moderate
number of activated/phagocytic microglia; 3 = a severe number of
activated/phagocytic microglia; and 4 = a very severe number of acti-
vated/phagocytic microglia. The concept of including a ‘very severe’
category was based on pathological procedures recommended for as-
sessing severity of cortical ?-synuclein pathology in dementia with
Lewy bodies (McKeith et al., 2005).
Semi-quantitative assessment of motor
Sections of motor cortex were available from 22 of the 31 cases. A
semi-quantitative four-point scale was used to document the degree of
motor cortex degeneration on haematoxylin and eosin, based on the
degree of neuronal loss and gliosis, as follows: 0 = none; 1 = mild
neuronal loss and gliosis; 2 = moderate neuronal loss and gliosis; and
3 = severe neuronal loss and gliosis with evidence of status spongiosis.
We also performed TDP-43 immunohistochemistry and graded the
number of TDP-43 immunoreactive inclusions or dystrophic neurites
that were present in motor cortex, as follows: 0 = no inclusions/neur-
ites identified; 1 = a few inclusions/neurites present; 2 = moderate
number of inclusions/neurites present; and 3 = a striking number of
Additional pathological analyses
Immunohistochemistry with IBA-1 antibody was performed to assess
for the presence and severity of activated and phagocytic microglia
(Ahmed et al., 2007) in motor cortex. The presence of myelin loss in
the medial lemniscus and the olivocerebellar tract was assessed with
Luxol fast blue stain in the same sections used to evaluate CSTD. The
presence of axonal loss of the corticospinal tract fibres at the level of
the pyramids was assessed with antibodies to phosphorylated neuro-
filament (SMI-31). The presence of athero- and arteriosclerotic vascu-
lar disease throughout the brain was assessed on haematoxylin and
eosin stain, Lewy bodies with ?-synuclein immunohistochemistry and
senile plaques and cerebral amyloid angiopathy with amyloid-b immu-
nohistochemistry or thioflavin-S fluorescent microscopy. Braak and
Braak staging (Braak and Braak, 1991) of neurofibrillary pathology
was performed based on silver stain (Bielschowsky) or thioflavin-S
All cases were grouped into one of three pathological groups based on
the severity of the semi-quantitative analyses of the medullary pyra-
mids as follows: Group 1 = cases without any myelin loss or activated/
phagocytic microglia [no corticospinal tract degeneration/CSTD(?)];
Group 2 = cases with slight-mild myelin loss or a scant to small
number of activated/phagocytic microglia [equivocal corticospinal
tract degeneration/CSTD(?)]; and Group 3 = cases with moderate to
very severe myelin loss and a moderate to striking number of acti-
vated/phagocytic microglia [definite corticospinal tract degeneration/
The medical records of all 31 FTLD-TDP type C cases were reviewed
for the abstraction of clinical information, including demographics, pre-
senting signs and symptoms, signs and symptoms that developed later
in the disease course and family history of any neurodegenerative
Figure 1 Histological findings included long thick dystrophic
neurites in neocortex (A) and Pick body-like inclusions in hip-
pocampal dentate granular cells (B). High power ?400
FTLD-CSTDBrain 2013: Page 3 of 16 |
at Mayo Clinic Library on January 29, 2013
disease. Disease duration was calculated as the difference in time from
death to time from first symptom onset as reported by the patient/
carer and documented in the medical records. Presence or absence of
the following additional signs and symptoms that have been reported
in FTLD-TDP was abstracted: prosopagnosia (a report of patient not
recognizing family members in the medical records) (Snowden et al.,
2004; Josephs et al., 2008b), topographagnosia (difficulty navigating
in familial areas/surroundings) and psychosis (hallucinations and delu-
sions) (Claassen et al., 2008; Lillo et al., 2010; Snowden et al., 2012).
Bedside cognitive testing scores (Mini-Mental State Examination;
Folstein et al., 1975) were recorded, and the motor examinations
were specifically reviewed for any evidence of MND, including dys-
arthria, hyper-reflexia, increased muscle tone, Babinski sign, ankle
clonus, fasciculation or muscle weakness. A retrospective clinical diag-
nosis was rendered, if possible, using the international criteria for FTD
(Neary et al., 1998), by an expert behavioural neurologist (K.A.J.)
blinded to pathological findings. MRI reports were available in 25 of
the 31 cases. Of these 25 cases, 13 had an MRI from our institution,
whereas 12 had an MRI at another institution. The findings that had
been documented by the radiologists in all 25 cases, based on visual
assessment of the pattern of atrophy at the time of clinical assessment,
All 31 cases were screened for genetic mutations in the progranulin
gene (GRN) as previously described (Baker et al., 2006) and for the
newly described hexanucleotide repeat expansion sequence in the
C9ORF72 gene (DeJesus-Hernandez et al., 2011). In brief, for
C9ORF72, PCR was performed with 100ng/ml of genomic DNA in
the presence of 1M betaine (Sigma), 5% dimethyl sulphoxide
(Sigma), 5mM of deoxycytidine triphosphate, deoxyadenosine tripho-
sphate, deoxythymidine triphosphate (Promega) and 7-deaza-2-deoxy-
guanosine triphosphate (Roche) in substitution for deoxyguanosine tri-
phosphate. The cycling programme included an initial denaturation at
98?C for 10min followed by 10 cycles of denaturation at 97?C for 35s,
annealing at 64?C for 2min and extension at 68?C for 8min, followed
by 25 cycles of denaturation at 97?C for 35s, annealing at 64?C for
2min and extension at 68?C for 8min plus an additional 20s for
each cycle (DeJesus-Hernandez et al., 2011). These two genes were
selected, as they are associated with FTLD-TDP pathology.
Quantitative magnetic resonance
A total of 12 cases had undergone an ante-mortem standardized volu-
metric head MRI protocol at 1.5 T that included a T1-weighted 3D
spoiled gradient echo sequence; hence, they were available for quan-
titative morphometric analysis (age at scan = 67 ? 8 years; 58%
female). The first available MRI was used in each case. Each of
these subjects was matched by age and gender to two healthy control
subjects, resulting in a matched cohort of 24 control subjects (age at
scan = 67 ? 7 years, 58% female). All control subjects had undergone
the same standardized MRI protocol at 1.5 T.
All images underwent preprocessing that included corrections for
gradient non-linearity (Jovicich et al., 2006) and intensity inhomogen-
eity using both the N3 bias correction (Sled et al., 1998) followed by
the Statistical Parametric Mapping (SPM5)-based bias correction.
Patterns of grey matter atrophy were assessed using the automated
and unbiased technique of voxel-based morphometry (Ashburner and
Friston, 2000), implemented using SPM5 (http://www.fil.ion.ucl.ac.
uk/spm). Briefly, all images were normalized to a customized template
and segmented using unified segmentation (Ashburner and Friston,
2005), followed by the hidden Markov random field clean-up step
(Zhang et al., 2001). All grey matter images were modulated and
smoothed with an 8-mm full-width at half-maximum smoothing
kernel. A full factorial model was used to assess patterns of grey
matter loss in each of the FTLD-TDP type C groups compared with
control subjects. Results were assessed before and after correction for
multiple comparisons using the false discovery rate correction at
P50.01. As the CSTD(+) group consisted of cases considered to
be novel, a direct comparison was also performed comparing this
group with a group consisting of all the other FTLD-TDP type C
cases [all CSTD(?) and CSTD(?) cases]; assessed uncorrected at
In addition, atlas-based parcellation was applied using SPM5 and an
in-house modified version of the automated anatomic labelling atlas
(Tzourio-Mazoyer et al., 2002), as previously described (Whitwell
et al., 2009), to generate grey matter volumes for specific regions;
selected based on the voxel-based morphometry results. First, grey
matter volume was calculated for the left and right temporal lobes.
An asymmetry score was also calculated as follows: (left temporal lobe
volume ? right temporal lobe volume) ? 2/(left temporal lobe vol-
ume + right temporal lobevolume),
(Whitwell et al., 2010). Second, grey matter volume was calculated
for the left and right motor cortex. All regional volumes were divided
by total intracranial volume to correct for differences in head size.
Statistical analyses were performed using the JMP computer software
(JMP Software, version 8.0.0; SAS Institute Inc) with ? set at 0.05
two-tailed. All binary data were compared across groups with
Pearson’s ?2test; Fisher’s exact test for any analysis with small num-
bers. Kruskal–Wallis test was performed across all three groups for
analysis of continuous data and, if significant, was followed by
Mann–Whitney U-test comparisons across two groups. Spearman
Rank order correlation was used to correlate degree of myelin loss
and activated/phagocytic microglia burden and to correlate myelin
loss and activated/phagocytic microglia burden to disease duration.
Logistic regression analysis was used to determine whether MRI
volumes predicted the CSTD(+) group, after adjusting for disease
duration. Linear regression analysis was used to determine whether
there was a difference in Luxol fast blue/IBA-1 severity across
groups after adjusting for disease duration.
A total of 31 cases of FTLD-TDP type C were analysed for this
study. On gross brain examination, there was often striking atro-
phy of the anterior medial and inferior temporal lobe with relative
sparing of the hippocampus (Fig. 2). Of the 31 cases, 74%
showed evidence of some corticospinal tract pathology, ranging
from mild to very severe. Eight cases were determined to not have
any corticospinal tract pathology and were, therefore, classified as
CSTD(?). Fourteen cases showed equivocal CSTD and were clas-
sified as CSTD(?). The remaining nine cases showed moderate to
very severe myelin loss and a moderate to striking number of
Brain 2013: Page 4 of 16 K. A. Josephs et al.
at Mayo Clinic Library on January 29, 2013
lipid-laden macrophages; hence, they had definite CSTD and were
classified as CSTD(+).
Corticospinal tract findings
Semi-quantitative analyses revealed that the degree of corticosp-
inal tract degeneration was significantly different across the three
groups (Table 1) without any difference between the CSTD(?)
and CSTD(?) groups. A representative case from each of the
three groups showing CSTD findings is shown in Fig. 3. There
was no overlap in the average Luxol fast blue and IBA-1 scores
between the nine CSTD(+) cases and the remaining 22 cases (Fig.
4). As expected, there was a significant correlation between
myelin loss and activated/phagocytic microglia severity (r = 0.9,
P50.0001). The average myelin loss/activated/phagocytic micro-
glia severity did correlate to disease duration across all 31 cases
(r = 0.6, P = 0.004). However, when both disease duration and
group was put in a linear regression model as predictors and aver-
age myelin loss/activated/phagocytic microglia severity as the out-
come, group was a much stronger predictor (P50.0001)
compared with disease duration (P = 0.04). In other words, after
adjusting for disease duration, there remained a strong association
between group and average Luxol fast blue/IBA-1 severity.
Motor cortex findings
Gross atrophy of the motor cortex was observed in some
CSTD(+) cases (Fig. 5). Similar to the findings in medullary pyra-
mids, there was a significant difference in neuronal loss and gliosis
in motor cortex across all three groups (Table 1), without any
difference between the CSTD(?) and CSTD(?) groups. Sections
of motor cortex were available in seven of the nine CSTD(+)
cases. All seven showed subjective decrease in density of Betz
cells in lamina V with neuronophagia in some cases and neuronal
pyknosis in others (Fig. 6); one case had eosinophilic cytoplasmic
inclusions that were not consistent with Bunina bodies. There was
also degeneration of the molecular layer (Fig. 6). Sections of
motor cortex were available in five CSTD(?) cases, and none
had evidence of the aforementioned pathologies. Of the 10
CSTD(?) cases with sections of motor cortex available for ana-
lysis, only two showed mild neuronal loss and neuronal pyknosis,
whereas the other eight did not have any of the aforementioned
pathologies. TDP-43 pathology in the motor cortex was also
significantly greater in the CSTD(+) group compared with the
other two groups, but it was also greater in the CSTD(?) group
compared with the CSTD(?) group (P = 0.01).
Lower motor neuron findings
None of the 31 cases had evidence of lower motor neuron de-
body-like inclusions, TDP-43 immunoreactive inclusions in cranial
nerve XII or anterior horn cells or anterior horn cell loss for those
cases in which spinal cord was available for study (n = 2).
Additional pathological findings
Assessment for other pathologies, including two other brainstem
tracts that are in close proximity to the corticospinal tract, the
medial lemniscus and olivocerebellar tract, revealed an absence
of pathology in all cases across the three groups (Fig. 3A, D and
G). Immunohistochemistry to neurofilament revealed a striking loss
of axons of corticospinal tract fibres in the CSTD(+) cases (Fig. 5).
One CSTD(?) case was found to have Lewy bodies with density
and distribution consistent with incidental Lewy bodies (Gibb and
Lees, 1988). Non-neuritic senile plaques were identified in ap-
proximately one-third of the cases, equally divided between the
three groups. One CSTD(?) case showed cerebral amyloid angio-
pathy. The presence of vascular pathologies, including athero- and
arteriosclerotic vascular disease, was identified in one CSTD(+)
case and three CSTD(?) cases. After excluding all cases with
other degenerative pathologies that could be associated with
microglia activation in motor cortex, only a handful of cases re-
mained. Of the four remaining cases, activated microglial burden
was greatest in the two CSTD(+) cases compared with the other
cases, one each from the CSTD(?) and CSTD(?) groups. Braak
and Braak staging was low across all three groups [CSTD(+) =
1.0 ? 0.9; CSTD(?) = 1.4 ? 1.0; CSTD(?) = 1.3 ? 0.9] and was
not statistically different (P = 0.73).
Corticospinal tract findings in types A
Of the 61 FTLD-TDP type A cases, two had moderate to severe
degeneration of the corticospinal tract and motor cortex, without
involvement of the hypoglossal nucleus; hence, there was no
Table 1 Semi-quantitative findings in the 31 cases of
FTLD-TDP type C by group
(n = 9)
(n = 14)
(n = 8)
Luxol fast blue
2.4 ? 0.7
2.9 ? 0.6
0.7 ? 0.5
1.0 ? 0.5
0.0 ? 0.0
0.0 ? 0.0
1.9 ? 0.70.2 ? 0.40.0 ? 0.0
3.4 ? 0.52.1 ? 0.61.2 ? 0.4 0.0004
Data shown as means ? standard deviation.
Figure 2 On gross brain inspection, there was evidence of an-
terior medial and inferior temporal atrophy, with enlargement of
the temporal horn.
FTLD-CSTDBrain 2013: Page 5 of 16 |
at Mayo Clinic Library on January 29, 2013
evidence for lower motor neuron degeneration. None of the 25
FTLD-TDP type B cases had evidence of degeneration of the corti-
cospinal tract and motor cortex in the absence of lower motor
neuron disease, that is, all type B cases with CSTD also showed
neuropathology affecting the hypoglossal nucleus. Therefore, com-
pared with 9/31 (29%) FTLD-TDP type C cases with CSTD(+)
pathology, only 2/86 (2%) of the other FTLD-TDP types had
CSTD(+) pathology (P50.0001 on ?2testing).
There were no significant differences across the three groups, with
the exception of total disease duration (Table 2). The CSTD(+)
group had a significantly longer disease duration than both the
CSTD(?) (P = 0.01) and CSTD(?) groups (P = 0.03). There was
no difference in disease duration between the CSTD(?) and
CSTD(?) groups (P = 0.41).
The frequency of examination features suggestive of MND did not
differ across the three groups (Table 3), although 42% of the
CSTD(+) group had at least one feature suggestive of corticosp-
inal tract degeneration. Examination features suggestive of CSTD
CSTD(+) group, three cases were unable to ambulate later in
the disease course. No cases in the other two groups had difficulty
ambulating. In no case were features of lower motor neuron dis-
ease reported, such as fasciculations or muscle atrophy. Only one
case, in the CSTD(?) group, had a positive family history in which
the case’s aunt and uncle both had dementia with onset age 565
years. Prosopagnosia was noted to be present across all three
groups. Topographagnosia was only identified in the CSTD(+)
group. Psychosis was extremely rare.
In the CSTD(+) group, the most common presenting symptom
was behavioural change, followed by difficulty naming objects,
loss of word meaning and difficulty recognizing faces. In the
CSTD(?) group, the most common presenting symptoms were
similar to the CSTD(+) group, except that behavioural changes
and prosopagnosia as a presenting symptom were much less fre-
quent. In the CSTD(?) group, difficulty naming and finding words
were the presenting symptoms in the majority of cases, with just
one subject showing personality change.
muscle tone.In the
Figure 3 Histological findings from a representative case from each of the three groups. A CSTD(?) case shows normal myelination of
medial lemniscus and olivocerebellar fibres (A), normal myelination of corticospinal tract fibres (B), and only resting microglia, but no
activated or phagocytic microglia (C). A CSTD(?) case shows normal myelination of medial lemniscus and olivocerebellar fibres (D), subtle
evidence of myelin loss (E) with rare activated and phagocytic microglia (F). A CSTD(+) case shows normal myelination of medial
lemniscus and olivocerebellar fibres (G), severe myelin loss (H) with numerous activated and phagocytic microglia (I). Luxol fast blue stain
(A, B, D, E, G and H) and actin binding protein IBA-1 antibody (C, F and I). Sections taken at the level of the medullary pyramids.
Magnification ?200 for all images.
Brain 2013: Page 6 of 16 K. A. Josephs et al.
at Mayo Clinic Library on January 29, 2013
The most common expert diagnosis rendered across all 31 cases
was semantic dementia (Table 3). Of those diagnosed with se-
mantic dementia, the CSTD(+) group had a higher proportion
of cases with right-sided dominant clinical features, such as be-
havioural changes, topographagnosia and prosopagnosia. In keep-
ing with these clinical findings, MRI reports by the evaluating
radiologists based on visual assessment showed a significatly
higher proportion of cases with right more than left focal anterior
temporal atrophy in the CSTD(+) group (Table 3 and Fig. 7A).
The majority of cases in the other two groups showed left4right
focal anterior temporal lobe atrophy (Table 3 and Fig. 7A). One
CSTD(+) case had multiple MRI with progressive temporal atro-
phy observed (Fig. 7B).
The two cases of FTLD-TDP type A that had CSTD were male with
onset age 56 and 66 years and disease duration of 7 years and 9
months, respectively. Both patients presented with falling and diffi-
culty swallowing, with neurological examination findings of promin-
ent extrapyramidal features. The first case also had striking upper
motor neuron signs and was wheelchair bound before death. Both
patients had been diagnosed with progressive supranuclear palsy.
MRI and PET scans were reported to be normal in both cases.
In no case was a GRN mutation identified. One case, in the
CSTD(?) group, was found to have a hexanucleotide repeat ex-
pansion in the C9ORF72 gene.
Quantitative magnetic resonance
Of the 12 cases with volumetric imaging, five were classified as
CSTD(?), five were classified as CSTD(?) and two were classified
as CSTD(+). Grey matter loss in all three groups was largely re-
stricted to the temporal lobes (Fig. 8). Only the CSTD(+) cases
showed atrophy of bilateral motor cortices compared with control
subjects (Fig. 8). After correcting for multiple comparisons, both
the CSTD(?) and CSTD(?) groups showed greater involvement
of the left hemisphere, with loss observed in anterior regions of
the left temporal lobe, involving temporal pole, fusiform gyrus,
inferior and middle temporal gyrus, amygdala and hippocampus.
In contrast, the CSTD(+) cases showed more bilateral patterns of
temporal lobe loss, with greater involvement of the right hemi-
sphere. The temporal pole, fusiform gyrus, inferior and middle
temporal gyrus, amygdala and hippocampus were involved in
both hemispheres, although to a greater degree on the right.
The CSTD(+) cases showed significantly greater loss in the right
amygdala, fusiform gyrus, parahippocampal gyrus, superior tem-
poral gyrus and bilateral superior motor cortices, than the group of
subjects that consisted of both the CSTD(?) and CSTD(?) cases,
on direct comparison (Fig. 9). The CSTD(?) and CSTD(?) group
did not show any regions of greater loss than the CSTD(+) cases.
Eleven of the 12 cases had a temporal lobe asymmetry score
greater than control subjects, demonstrating the presence of
asymmetry. Seven cases showed greater involvement of the left
hemisphere and four showed greater involvement of the right
hemisphere. The average asymmetry score in the CSTD(+) sub-
jects was 0.24 ? 0.08, with both subjects showing right-sided
asymmetry. In contrast, the average asymmetry score in the
group that consisted of both CSTD(?) and CSTD(?) cases was
?0.20 ? 0.27, with only 20% (2/10) of subjects showing
right-sided asymmetry. Volumes of the right temporal lobe
(P = 0.001) and right motor cortex (P = 0.001) were smaller in
the CSTD(+) group after adjusting for disease duration, compared
with the rest of the cases (Fig. 10A and B). No differences were
observed in the left hemisphere.
We have identified a novel pathological entity characterized by
focal temporal lobe atrophy, FTLD-TDP type C pathology and
moderate to very severe CSTD. This combination of features has
not been previously emphasized, yet the identification of this
entity has important implications for FTLD classification.
Of cases with FTLD-TDP type C pathology, those with moder-
ate to very severe CSTD accounted for almost 30% of cases;
hence, they are not uncommon. Importantly, however, these
cases were not associated with pathological features of lower
motor neuron degeneration, which is the cardinal feature of
FTLD-MND. There is one small case series in which FTLD-TDP
Figure 4 Box-plot showing the average Luxol fast blue/IBA-1
severity in the CSTD(?), CSTD(?) and CSTD(+) cases. Boxes
represent 25th and 75th percentiles with means for all 31 cases.
Whiskers extend to minimum and maximum values. Plots show
no overlap in values between the nine CSTD(+) cases and the
remaining 22 cases. LFB = Luxol fast blue.
FTLD-CSTDBrain 2013: Page 7 of 16 |
at Mayo Clinic Library on January 29, 2013
Figure 5 Figure demonstrating gross and histological features of a typical CSTD(+) case. There is gross atrophy of the motor cortex. Low
power microscopic images (left images = ?100 magnification; right images = ?200 magnification) reveal loss of myelin on Luxol fast blue
(LFB) stain, axonal loss on neurofilament (pNF) stain and the presence of activated microglia with IBA-1 in the medullary pyramids. In
contrast, there is preservation of myelinated fibres in the adjacent medial lemniscus (ML). MTR = motor cortex.
Brain 2013: Page 8 of 16K. A. Josephs et al.
at Mayo Clinic Library on January 29, 2013
type C pathology has also been associated with CSTD, without
pathologicalfindings of lower
(Kobayashi et al., 2010). The authors demonstrate that such
CSTD(+) cases showed moderate to severe CSTD, FTLD-MND
cases had only mild CSTD. In fact, at the level of the midbrain,
some FTLD-MND cases did not have any evidence of CSTD.
This further supports our contention that CSTD(+) cases are dif-
ferent from FTLD-MND. Our CSTD(+) cases also do not meet
criteria for any other neurodegenerative disease. The absence of
amyloid-related entities. The absence of ?-synuclein and tau de-
position eliminates all entities that are defined by the deposition of
these two proteins. The absence of fused in sarcoma protein
deposition excludes neuronal intermediate filament inclusion dis-
ease (Josephs et al., 2003; Cairns et al., 2004), basophilic inclusion
body disease (Munoz et al., 2009) and atypical FTLD with
ubiquiting-positive inclusions (FTLD-U) (Josephs et al., 2008a;
Mackenzie et al., 2008). Finally, given the presence of TDP-43,
our cases do not meet criteria for FTLD-U, FTLD with no inclusions
(FTLD-Ni), FTLD with epitopes of the ubiquiting–proteasome
system (FTLD-UPS) or dementia-lacking distinctive histology
(Knopman et al., 1990; Mackenzie et al., 2009). Therefore, we
are left with a pathological entity characterized by focal temporal
lobe degeneration with striking CSTD, without any evidence of
anterior horn cell disease and TDP-43 immunoreactive inclusions.
There is currently no such recognized neurodegenerative entity
(Dickson, 2003). Furthermore, the demonstration of axonal loss
excludes the likelihood that it is a type of demyelinating disease.
We previously reported two cases of FTLD with CSTD that
predated TDP-43 typing (Josephs and Dickson, 2007b). One of
the observations from that brief report was that the two cases
with CSTD had longer disease duration than typical cases of
FTLD-MND. In the current study, we confirm the long disease
duration in CSTD(+) cases. Compared with FTLD-MND, in
which the mean disease duration is ?2–3 years (range between
6 months and 6 years) (Hodges et al., 2003; Josephs et al., 2005;
Hu et al., 2009; Coon et al., 2011a; Elamin et al., 2011); mean
disease duration in our CSTD(+) cases was ?15 years. This is not
necessarily surprising, as long disease duration is associated with
FTLD-TDP type C pathology (Grossman et al., 2007; Josephs
et al., 2009, 2011). Surprisingly, however, the CSTD(+) cases
had longer disease duration than the other cases with FTLD-TDP
type C pathology. Although it is possible that the changes in the
Figure 6 Motor cortex from the three representative cases that are shown in Fig. 3. In the CSTD(?) case, Betz cells are present (A), there
is preservation of the molecular layer (B), only scant dystrophic neurites (C) and scant activated microglia (D) are observed. In the
CSTD(?) case, Betz cells can be identified, although less prominent compared with the CSTD(?) case (E), there is less preservation of the
molecular layer (F), more dystrophic neurites (G) and more activated microglia (H) can be observed. In the CSTD(+) case, Betz cells are
absent (I), there is spongiosis in the molecular layer (J), there are a moderate number of dystrophic neurites (K) and a frequent number of
activated microglia (L). The inset in I shows neuronophagia in lamina V in the CSTD(+) case. Haematoxylin and eosin low power (?200)
images (A, B, E, F, I and J), TDP-43 (C, G and K) and IBA-1 (D, H and L) high power (?400) images (C, D, G, H, K and L).
FTLD-CSTD Brain 2013: Page 9 of 16 |
at Mayo Clinic Library on January 29, 2013
CSTD(+) group are because of longer survival, our findings do not
support that hypothesis. First, the differences in severity measures
of CSTD across the groups persisted after we accounted for dis-
ease duration, suggesting that disease duration alone cannot ex-
plain the findings. Second, moderate to very severe CSTD in the
absence of lower motor neuron degeneration was not associated
with type A and type B pathology, despite the fact that some
cases with these pathologies do have long disease duration
(Cairns et al., 2007; Josephs et al., 2007a; Whitwell et al.,
2010), suggesting that the association is specific to type C path-
ology. Third, if it was simply a matter of longer disease duration
and, hence, disease progression, we would expect changes to be
observed in other tracts, yet we did not observe any evidence of
degeneration in the olivocerebellar tract or medial lemniscus. The
more likely possibilities, therefore, are that there exists a unique
association between FTLD with type C pathology and CSTD, and
that this is a novel neurodegenerative entity in which long disease
duration is a feature of the disease. The notion that the presence
of CSTD could be causing longer disease duration in FTLD-TDP
type C cases is counter-intuitive and highly unlikely.
Almost all CSTD(+) cases, and in fact all FTLD-TDP type C
cases, fulfilled criteria for semantic dementia (Neary et al.,
1998), although in some patients, limited clinical information pre-
vented retrospective diagnosis. The fact that the majority of cases
Table 3 Clinical features
CSTD(+)(n = 7) CSTD(?)(n = 12)
CSTD(?)(n = 8)
Presented with behavioural change or prosopagnosia
Positive family history
Any examination features suggestive of MNDa,b
Increased muscle tone
Unable to ambulatec
Expert clinical diagnosis
Proportion of semantic dementia cases with rSD
MRI findings based on the radiologist report CSTD(+)(n = 6)
a Noted at any time during the disease course.
b Findings are not mutually exclusive (i.e. a patient could have hyper-reflexia and a Babinski sign).
c Feature was noted to occur late in the disease course.
bvFTD = behavioural variant FTD; rSD = right-sided dominant features; L = left; R = Right; NA = Not able to provide retrospective diagnosis according to current clinical
Table 2 Demographics of the 31 cases of FTLD-TDP type C by group
Demographics CSTD(+) (n = 9)CSTD(?) (n = 14)CSTD(?) (n = 8)P-
Age at death, years
Age at onset, years
Total disease duration, years
MMSE at initial evaluation
15.6 ? 3.9 (10–21)
74. 4 ? 8.8 (61–93)
63.3 ? 10.8 (47–87)
10.8 ? 2.7 (6–15)
23.9 ? 3.1 (18–29)
12.8 ? 3.0 (8–16)
71. 5 ? 5.6 (60–78)
61.6 ? 4.6 (53–69)
9.9 ? 2.9 (6–15)
24.0 ? 3.3 (18–27)
15.8 ? 3.3 (12–21)
74.3 ? 8.5 (58–81)
59.0 ? 8.3 (46–70)
15.3 ? 4.4 (11–22)
14.8 ? 11.5 (3–27)
Data shown as means ? standard deviation (range).
MMSE = Mini-Mental State Examination.
Brain 2013: Page 10 of 16K. A. Josephs et al.
at Mayo Clinic Library on January 29, 2013
were diagnosed with semantic dementia was not surprising, as
FTLD-TDP type C has been associated with semantic dementia
(Mackenzie et al., 2006; Grossman et al., 2007; Snowden et al.,
2007; Josephs et al., 2009; Rohrer et al., 2011). Unexpectedly,
however, clinical features of CSTD were not prominent and were
reported in 550% of the CSTD(+) cases. CSTD is typically con-
sidered to be associated with striking clinical features of spastic
dysarthria, dysphagia, upper motor neuron type weakness,
hyper-reflexia, Babinski sign and clonus. In our CSTD(+) cases
of FTLD-TDP type C, some of these features were documented,
but in most cases, these features were not emphasized in the
medical records. Three cases were unable to ambulate later in
the disease course, which could signify upper motor neuron dis-
ease, but this was not clear from the medical records. In one
reported case of semantic dementia that had pathologically con-
firmed FTLD-MND, it was indeed stated that the patient de-
veloped walking difficulties 1 year before death (Davies et al.,
2005). The fact that upper motor neuron features were not
more frequent is somewhat surprising, as one would have pre-
dicted more clinical symptoms suggestive of CSTD in more
cases. Although it is possible that clinical features associated
with CSTD were simply absent to mild in some of our CSTD(+)
cases, it is also possible that some of these features were missed
by the evaluating physicians who may have been more focused on
assessing cognitive and behavioural impairment. The absence of
clinical features suggestive of CSTD has been previously reported
in another degenerative disease, corticobasal degeneration, where
CSTD was pathologically identified (Tsuchiya et al., 2005). Clinical
features suggestive of underlying CSTD were identified in only
60% of the subjects with corticobasal degeneration (Tsuchiya
et al., 2005), suggesting that in some degenerative diseases with
CSTD, clinical signs may not always be observed. However, it is
possible that the clinical manifestations of CSTD may be a late
feature of the disease (Davies et al., 2005). In a case report of
a patient with semantic dementia and CSTD with detailed annual
examination, clinical features in-keeping with CSTD were not
observed until 17 years after disease onset (Yokota et al.,
2006). In fact, in one of our patients that had yearly examinations,
clinical features of CSTD were still not observed even after 9 years
(Supplementary material). Findings from this study cannot answer
the question of whether clinical signs of CSTD were absent or
missed by the evaluating physicians; most likely it was because
Figure 7 Coronal volumetric MRI for representative CSTD(?), CSTD(?) and CSTD(+) cases shown at the level of the anterior temporal
lobes (A). The CSTD(+) cases both show right4left temporal atrophy, whereas the other cases show left4right temporal atrophy. One
of the CSTD(+) cases had annual brain MRI scans for a period of 5 years, demonstrating progressive atrophy of both temporal lobes,
although the right temporal lobe remained the more affected at each time point (B).
FTLD-CSTDBrain 2013: Page 11 of 16 |
at Mayo Clinic Library on January 29, 2013
of a combination of both. However, the findings support the im-
portance of performing a thorough neurological examination of
both the upper and lower motor systems in all patients with an
FTD syndrome, especially those with semantic dementia. We could
also speculate that the lack of clinical features of CSTD may have
biological underpinnings. Features of CSTD tend to be prominent
in patients with strokes affecting the corticospinal tract. In such
instances, there is acute damage to the corticospinal tract fibres.
Features of CSTD can also be prominent in amyotrophic lateral
sclerosis, in which degeneration is relatively rapid for a period of
Figure 8 Patterns of grey matter loss in the CSTD(?), CSTD(?) and CSTD(+) groups compared with control subjects. Results are shown
on 3D renderings uncorrected for multiple comparisons. Findings in the motor cortex are circled. L = left; R = right.
Figure 9 Regions that showed greater grey matter loss in the CSTD(+) subjects compared with the rest of the FTLD-TDP type C cases
[CSTD(?) and CSTD(?)]. Results are shown on 3D renderings and on representative coronal and axial slices, uncorrected for multiple
comparisons. L = left; R = right.
Brain 2013: Page 12 of 16K. A. Josephs et al.
at Mayo Clinic Library on January 29, 2013
months to a few years. On the contrary, in our CSTD(+) cases, it
is more likely that degeneration occurred much more gradually for
a period of many years. It has been suggested that in degenerative
diseases, there is more involvement of the caudal corticospinal
tract than the rostral corticospinal tract, for example, more
severe degeneration at the level of the pyramids than at the
level of the cerebral peduncles, and, hence, a dying back phenom-
enon (Kato, 2008; Kobayashi et al., 2010). Although we found
more TDP-43 immunoreactive inclusions in the motor cortex and
changes in the molecular layer in our CSTD(+) cases compared
with the other two groups, suggesting that there is involvement of
the corticospinal tract beyond the level of the pyramids, we did
not assess the corticospinal tract at the level of the cerebral ped-
uncles or pons; hence, we cannot refute or support the suggestion
that this may be a dying back phenomenon.
In all CSTD(+) cases that could be diagnosed, there was evi-
dence of significant anomia, as well as varying combinations of
behavioural and personality change, prosopagnosia and topogra-
phagnosia. The presence of behavioural and personality changes,
topographagnosia and prosopagnosia is suggestive of involvement
of the right temporal lobe (Thompson et al., 2003; Seeley et al.,
2005; Josephs et al., 2008b; Chan et al., 2009), and indeed, visual
assessment of the MRIs did find a predominance of right temporal
lobe atrophy. In contrast, clinical and visual MRI findings in the
other two groups were more consistent with left temporal lobe
involvement. However, prosopagnosia was observed in the
CSTD(?) and CSTD(?) groups, likely reflecting the fact that
the right temporal lobe was involved to some degree, although
not dominant, in some of these cases. Psychosis was rare in all
FTLD-TDP type C groups, which contrasts with previous reports of
psychosis being associated with FTLD-MND (Claassen et al., 2008;
Lillo et al., 2010); further evidence that our CSTD(+) cases differ
from FTLD-MND. The fact that only one case had a repeat ex-
pansion in the C9ORF72 gene, and none had mutations in GRN,
concords with the lack of family history and the lack of psychosis,
as the latter has been associated with the C9ORF72 repeat ex-
pansions (Snowden et al., 2012).
The quantitative MRI findings of predominant focal anterior
temporal atrophy across all FTLD-TDP type C groups were con-
sistent with previous reports in FTLD-TDP type C (Rohrer et al.,
2010; Whitwell et al., 2010) and the clinical diagnosis of semantic
dementia (Mummery et al., 2000; Chan et al., 2001; Galton et al.,
2001; Gorno-Tempini et al., 2004; Josephs et al., 2008b). In the
majority of cases, the temporal lobe atrophy was asymmetric.
However, the CSTD(+) cases showed right-sided asymmetry,
whereas left-sided asymmetry was more common in the other
cases. Although there was an association between right-sided
temporal atrophy and CSTD, we also had right-sided cases that
did not have moderate to very severe CSTD. This disassociation
suggests that not all patients with right-sided dominant semantic
dementia will have CSTD, although right-sided patients may be
more likely to have CSTD than patients with left-sided semantic
dementia. On the other hand, one case has been previously re-
ported with semantic dementia and CSTD with left4right
Figure 10 Box-plots showing right temporal (A) and motor cortex (B) residual volumes after regressing out the effect of disease duration
for CSTD(?), CSTD(?) and CSTD(+) cases. Residuals rather than the observed volumes were plotted to graphically illustrate groupwise
differences after accounting for disease duration. Residual volumes were obtained by fitting a linear regression model of regional volume
(y-axis) versus disease duration (x-axis) and then calculating the raw volume minus the predicted volume.
FTLD-CSTDBrain 2013: Page 13 of 16 |
at Mayo Clinic Library on January 29, 2013
temporal atrophy (Yokota et al., 2006); therefore, cases with
left-sided semantic dementia could also develop CSTD. Further
complicating the issue of asymmetry is that patients with right-
left temporal atrophy can also be associated with FTLD-MND
(Coon et al., 2011b). Such cases, however, typically have clinical
features of lower motor neuron disease and have been reported to
have FTLD-TDP type B pathology (Coon et al., 2011b). These
cases should be differentiated from our CSTD(+) cases that
have FTLD-TDP type C pathology and no clinical features of
lower motor neuron disease.
In addition to differences within the temporal lobe, the CSTD(+)
group also showed greater atrophy in the motor cortex compared
with the other FTLD-TDP type C groups. The region-of-interest ana-
lysis revealed no overlap in motor cortex volumes, even when ac-
counting for disease duration. Consistent with motor cortex atrophy
observed on imaging was the fact that pathological analysis revealed
motor cortex atrophy on gross examination, as well as motor cortex
degeneration in the CSTD(+) cases.
Even after excluding the CSTD(+) cases, a high proportion of
FTLD-TDP type C cases did have some evidence of CSTD (64%).
However, the data did not support separating the CSTD(?) group
from the CSTD(?) group. In fact, we observed no clinical, genetic
or imaging differences between these two groups. Subtle involve-
ment of the corticospinal tract, therefore, seems to be a feature of
FTLD-TDP type C. However, clinical and imaging data clearly
show that it is important to segregate the CSTD(+) group,
which is characterized by moderate to very severe CSTD, from
those cases with only subtle features of involvement of the corti-
cospinal tract. The finding of moderate to very severe CSTD in the
absence of lower motor neuron degeneration seems to be specific
to FTLD-TDP type C pathology. We only observed two cases with
CSTD and absent lower motor neuron degeneration out of 86
cases of FTLD-TDP type A or B. These two cases were clearly
clinically different from our CSTD(+) type C cases. Both had strik-
ing extrapyramidal and pyramidal features, short disease duration
Furthermore, neither case had focal temporal lobe atrophy.
There have been a few reports of semantic dementia associated
with motor neuron degeneration, although these cases differ from
our CSTD(+) cases. In one series of 18 pathologically confirmed
cases with semantic dementia, one case was reported to have
features of motor neuron degeneration (Davies et al., 2005).
This case showed degeneration of medullary motor nuclei, that
is, lower motor neuron degeneration; hence, it differed from our
CSTD(+) cases. Another case has been reported with clinical
features of semantic dementia and amyotrophic lateral sclerosis
without autopsy confirmation (Kim et al., 2009). Given the neuro-
physiological confirmation of lower motor neuron disease in that
case, one would predict FTLD-TDP type B pathology and the pres-
ence of lower motor neuron degeneration. However, features of
motor neuron degeneration, including CSTD, have not been
emphasized in other large clinicopathological series of semantic
dementia (Godbolt et al., 2005; Hodges et al., 2010), and one
study assessing motor neuron dysfunction in FTD variants did not
identify corticospinal tract dysfunction in those with semantic de-
mentia (Burrell et al., 2011). There are two possible explanations:
(i) CSTD occurs late in those with semantic dementia or (ii)
semantic dementia with CSTD may not have been represented
in that series. Therefore, it seems that there are at least two vari-
ants of FTLD in which CSTD can be identified: (i) those with
pathological findings of lower motor neuron degeneration, ab-
sent–moderate CSTD, FTLD-TDP type B pathology and relatively
short disease duration (6 months to 6 years) and (ii) those without
lower motor neuron degeneration yet moderate to very severe
CSTD, FTLD-TDP type C pathology and long disease duration
(10–20 years). The former is FTLD-MND and is predominantly
associated with behavioural variant FTD or agrammatic aphasia
and features of lower motor neuron disease, whereas the latter
could be referred to as FTLD-CSTD and is predominantly asso-
ciated with semantic dementia and no features of lower motor
In summary, we have identified a group of cases with FTLD-
TDP type C pathology with moderate to very severe CSTD, that is,
FTLD-CSTD, presenting as semantic dementia predominantly with
right-sided features (Box 1). This entity seems to be associated
with right predominant temporal lobe atrophy and evidence of
motor cortex grey matter atrophy. Given that our quantitative
imaging analysis was limited by a small number of cases in some
categories, confirmation of these imaging findings will be needed
in larger cohorts. This distinct clinicopathological entity represents
a relatively high proportion of FTLD-TDP type C cases and, there-
fore, should not be ignored. In fact, it might be important to
separate out such cases in future genetic studies of FTLD-TDP.
The authors acknowledge the histological assistance of Virginia
Phillips and Linda Rousseau and immunohistochemistry support
Box 1 Features suggestive of FTLD-TDP type C pathology
with moderate to very severe CSTD
(1) Long disease duration
(1) Prosopagnosia and topographagnosia
(2) Behavioural and personality change
(3) Anomia and loss of word meaning
(2) Babinski sign
(4) Upper motor neuron pattern weakness
(1) Right4left anterior temporal lobe atrophy
(2) Atrophy of the motor cortex
Brain 2013: Page 14 of 16 K. A. Josephs et al.
at Mayo Clinic Library on January 29, 2013
of Monica Casey-Castanedes. They thank Dr Leonard Petrucelli for
generous donation of C-terminal TDP-43 antibody used for
Funding for this study was received from the National Institutes of
Health (NIH) (R01 AG037491, P50 AG16574, P01 AG03949 and
Supplementary material is available at Brain online.
Ahmed Z, Shaw G, Sharma VP, Yang C, McGowan E, Dickson DW.
Actin-binding proteins coronin-1a and IBA-1 are effective microglial
markers for immunohistochemistry. J Histochem Cytochem 2007; 55:
Arai T, Hasegawa M, Akiyama H, Ikeda K, Nonaka T, Mori H, et al.
TDP-43 is a component of ubiquitin-positive tau-negative inclusions in
frontotemporal lobar degeneration and amyotrophic lateral sclerosis.
Biochem Biophys Res Commun 2006; 351: 602–11.
Ashburner J, Friston KJ. Voxel-based morphometry—the methods.
Neuroimage 2000; 11: 805–21.
Ashburner J, Friston KJ. Unified segmentation. Neuroimage 2005; 26:
Baker M, Mackenzie IR, Pickering-Brown SM, Gass J, Rademakers R,
Lindholm C, et al. Mutations in progranulin cause tau-negative fron-
totemporal dementia linked to chromosome 17. Nature 2006; 442:
Beach TG, White CL, Hamilton RL, Duda JE, Iwatsubo T, Dickson DW,
et al. Evaluation of alpha-synuclein immunohistochemical methods
used by invited experts. Acta Neuropathol 2008; 116: 277–88.
Braak H, Braak E. Neuropathological stageing of Alzheimer-related
changes. Acta Neuropathol 1991; 82: 239–59.
Burrell JR, Kiernan MC, Vucic S, Hodges JR. Motor neuron dysfunction in
frontotemporal dementia. Brain 2011; 134: 2582–94.
Cairns NJ, Grossman M, Arnold SE, Burn DJ, Jaros E, Perry RH, et al.
Clinical and neuropathologic variation in neuronal intermediate fila-
ment inclusion disease. Neurology 2004; 63: 1376–84.
Cairns NJ, Neumann M, Bigio EH, Holm IE, Troost D, Hatanpaa KJ, et al.
TDP-43 in familial and sporadic frontotemporal lobar degeneration
with ubiquitin inclusions. Am J Pathol 2007; 171: 227–40.
Chan D, Anderson V, Pijnenburg Y, Whitwell J, Barnes J, Scahill R, et al.
The clinical profile of right temporal lobe atrophy. Brain 2009; 132:
Chan D, Fox NC, Scahill RI, Crum WR, Whitwell JL, Leschziner G, et al.
Patterns of temporal lobe atrophy in semantic dementia and
Alzheimer’s disease. Ann Neurol 2001; 49: 433–42.
Claassen DO, Parisi JE, Giannini C, Boeve BF, Dickson DW, Josephs KA.
Frontotemporal dementia mimicking dementia with Lewy bodies.
Cogn Behav Neurol 2008; 21: 157–63.
Coon EA, Sorenson EJ, Whitwell JL, Knopman DS, Josephs KA. Predicting
survival in frontotemporal dementia with motor neuron disease.
Neurology 2011a; 76: 1886–93.
Coon EA, Whitwell JL, Parisi JE, Dickson DW, Josephs KA. Right tem-
poral variant frontotemporal dementia with motor neuron disease.
J Clin Neurosci 2011b; 19: 85–91.
Davies RR, Hodges JR, Kril JJ, Patterson K, Halliday GM, Xuereb JH. The
pathological basis of semantic dementia. Brain 2005; 128: 1984–95.
DeJesus-Hernandez M, Mackenzie IR, Boeve BF, Boxer AL, Baker M,
Rutherford NJ, et al. Expanded GGGGCC hexanucleotide repeat in
noncoding region of C9ORF72 causes chromosome 9p-linked FTD
and ALS. Neuron 2011; 72: 245–56.
Dickson D, editor. Neurodegeneration: the molecular pathology of de-
mentia and movement disorders. Basel: ISN Neuropath Press; 2003.
Elamin M, Phukan J, Bede P, Jordan N, Byrne S, Pender N, et al.
Executive dysfunction is a negative prognostic indicator in patients
with ALS without dementia. Neurology 2011; 76: 1263–9.
Folstein MF, Folstein SE, McHugh PR. “Mini-mental state”. A practical
method for grading the cognitive state of patients for the clinician. J
Psychiatr Res 1975; 12: 189–98.
Galton CJ, Patterson K, Graham K, Lambon-Ralph MA, Williams G,
Antoun N, et al. Differing patterns of temporal atrophy in Alzheimer’s
disease and semantic dementia. Neurology 2001; 57: 216–25.
Geser F, Martinez-Lage M, Robinson J, Uryu K, Neumann M,
Brandmeir NJ, et al. Clinical and pathological continuum of multisys-
tem TDP-43 proteinopathies. Arch Neurol 2009; 66: 180–9.
Gibb WR, Lees AJ. The relevance of the Lewy body to the pathogenesis
of idiopathic Parkinson’s disease. J Neurol Neurosurg Psychiatry 1988;
Godbolt AK, Josephs KA, Revesz T, Warrington EK, Lantos P, King A, et al.
Sporadic and familial dementia with ubiquitin-positive tau-negative in-
clusions:clinicalfeaturesof onehistopathological abnormality underlying
frontotemporal lobar degeneration. Arch Neurol 2005; 62: 1097–101.
Gorno-Tempini ML, Dronkers NF, Rankin KP, Ogar JM, Phengrasamy L,
Rosen HJ, et al. Cognition and anatomy in three variants of primary
progressive aphasia. Ann Neurol 2004; 55: 335–46.
Grossman M, Wood EM, Moore P, Neumann M, Kwong L, Forman MS,
et al. TDP-43 pathologic lesions and clinical phenotype in frontotem-
poral lobar degeneration with ubiquitin-positive inclusions. Arch Neurol
2007; 64: 1449–54.
Gwinn-Hardy K, Mehta ND, Farrer M, Maraganore D, Muenter M,
Yen SH, et al. Distinctive neuropathology revealed by alpha-synuclein
antibodies in hereditary parkinsonism and dementia linked to chromo-
some 4p. Acta Neuropathol 2000; 99: 663–72.
Hodges JR, Davies R, Xuereb J, Kril J, Halliday G. Survival in frontotem-
poral dementia. Neurology 2003; 61: 349–54.
Hodges JR,Mitchell J, DawsonK,Spillantini MG,Xuereb JH, McMonagle P,
et al. Semantic dementia: demography, familial factors and survival in a
consecutive series of 100 cases. Brain 2010; 133: 300–6.
Hu WT, Seelaar H, Josephs KA, Knopman DS, Boeve BF, Sorenson EJ,
et al. Survival profiles of patients with frontotemporal dementia and
motor neuron disease. Arch Neurol 2009; 66: 1359–64.
Hutton M, Lendon CL, Rizzu P, Baker M, Froelich S, Houlden H, et al.
Association of missense and 5’-splice-site mutations in tau with the
inherited dementia FTDP-17. Nature 1998; 393: 702–5.
Josephs KA. Frontotemporal dementia and related disorders: deciphering
the enigma. Ann Neurol 2008; 64: 4–14.
Josephs KA, Ahmed Z, Katsuse O, Parisi JF, Boeve BF, Knopman DS,
et al. Neuropathologic features of frontotemporal lobar degeneration
with ubiquitin-positive inclusions with progranulin gene (PGRN) muta-
tions. J Neuropathol Exp Neurol 2007a; 66: 142–51.
Josephs KA, Dickson DW. Frontotemporal lobar degeneration with upper
motor neuron disease/primary lateral sclerosis. Neurology 2007b; 69:
Josephs KA, Hodges JR, Snowden JS, Mackenzie IR, Neumann M,
ity in frontotemporal dementia. Acta Neuropathol 2011; 122: 137–53.
Josephs KA, Holton JL, Rossor MN, Braendgaard H, Ozawa T, Fox NC,
et al. Neurofilament inclusion body disease: a new proteinopathy?
Brain 2003; 126: 2291–303.
Josephs KA, Holton JL, Rossor MN, Godbolt AK, Ozawa T, Strand K,
et al. Frontotemporal lobar degeneration and ubiquitin immunohisto-
chemistry. Neuropathol Appl Neurobiol 2004; 30: 369–73.
Josephs KA, Knopman DS, Whitwell JL, Boeve BF, Parisi JE, Petersen RC,
et al. Survival in two variants of tau-negative frontotemporal lobar
degeneration: FTLD-U vs FTLD-MND. Neurology 2005; 65: 645–7.
FTLD-CSTDBrain 2013: Page 15 of 16 |
at Mayo Clinic Library on January 29, 2013