Cerebrospinal Fluid Biomarker Profile in TDP-43-Related Genetic Frontotemporal Dementia

Cerebrospinal fluid (CSF) biomarkers, namely total tau, phospho-tau and amyloid beta peptides, have received much attention specifically regarding Alzheimer’s disease (AD), since they can detect the biochemical fingerprint of AD and serve as a diagnostic tool for accurate and early diagnosis during life. In the same way, biomarkers for other neurodegenerative disease pathologies are also needed. We present a case series of six patients with genetic frontotemporal dementia (FTD), with TDP-43 underlying proteinopathy, in an attempt to assess TDP-43 as a novel biomarker alone and in combination with established AD biomarkers for this specific patient group, based on the principles of personalized and precision medicine. Our results indicate that genetic TDP-43-FTD is characterized by increased CSF TPD-43 and increased TDP-43 × τΤ/τP-181 combination. Hence, TDP-43 combined with tau proteins could be a useful tool for the diagnosis of genetic FTD with TDP-43 underling histopathology, supplementing clinical, neuropsychological and imaging data.


Introduction
A common neuropathological feature of neurodegenerative diseases is cytoplasmic protein aggregates. Cytoplasmic mislocalization and aggregation of TAR-DNA binding protein 43 (TDP-43) is found in the vast majority of patients with amyotrophic lateral sclerosis (ALS), and in approximately half of patients with frontotemporal dementia (FTD) with ubiquitin inclusions [1].
Recently, great progress in the field of neurogenetics has resulted in the recognition of several gene mutations related to the FTD-ALS spectrum, especially with ubiquitin-TDP-43 predominant histopathology, thus highlighting a new brain proteinopathy. The most recent, but more common discovery, is a genetic mutation in the C9orf72 gene involving a hexanucleotide repeat expansion (GGGGCC) in the first non-coding intron of the gene, responsible for both familial and sporadic FTD and ALS, but more frequently for the combined FTD-ALS phenotype [2][3][4][5].
Heterozygous loss of function mutations in the progranulin gene (GRN) can also lead to FTD with TDP-43 inclusions [6]. GRN mutations have been linked to FTD phenotypes, particularly those of the primary progressive aphasia (PPA) subtype. Mutations in the

Materials and Methods
A total of six patients were included in this case series study. All patients were hospitalized at the 1st Department of Neurology of the National and Kapodistrian University of Athens at Eginition Hospital during the years 2011-2018. The study was performed in accordance with the ethical guidelines of the Declaration of Helsinki, and had the approval of the local Ethical and Deontology committee of our hospital (244/5-7-2010). All study subjects and/or their relatives gave informed consent for inclusion in the study.
Patients underwent detailed clinical, neuropsychological, biochemical and neuroimaging examination, to exclude secondary causes of dementia and establish an FTD diagnosis, based on the criteria of either Rascovsky et al. [24] for probable bvFTD or Gorno-Tempini et al. [25] for PPA. The Mini-Mental State examination (MMSE) and the Frontal Assessment Battery (FAB) were used as crude estimates of dementia severity [26,27].
The CSF levels of τ T , τ P-181 and Aβ 42 and Aβ 40 were measured in duplicate by commercially available enzyme-linked immunosorbent assay (ELISA) kits ("Innotest hTau antigen", "phospho-tau181", "β-amyloid1-42" and "β-amyloid1-40", respectively, Fujirebio, Gent, Belgium) according to manufacturer's instructions, and blind to clinical diagnosis. The above biomarkers are measured routinely in the Neurochemistry and Biomarker Unit of our department, as part of an everyday diagnostic approach, as described elsewhere [23]. TDP-43 was measured in triplicate by double sandwich ELISA, using a commercially available kit (Human TAR DNA-binding protein 43 ELISA kit, Cusabio Biotech Co., Ltd., Wuhan, China) according to manufacturer's instructions, as described elsewhere [18]. The diagnostic performance of TDP-43 was assessed as a biomarker with molecular specificity in patients with an individualized FTD diagnosis, alone and combined with established AD biomarkers (mainly for excluding AD-related dementia). Cut-offs for normal values are shown in Table 1. A 61-year-old female patient was admitted to our department for progressive language impairment over the past 1-2 years. She had word-finding difficulty, with intact comprehension, an absence of phonological speech or grammar errors, no other significant cognitive/behavioral disturbances, and preserved activities of daily living. Neuropsychological assessment revealed mild cognitive impairment, with an MMSE score of 25/30 and a FAB score of 12/18. Her writing skills and phrase and sentence repetition were preserved, too. On magnetic resonance imaging (MRI), moderate cortical atrophy in the frontal and temporal lobes, with left predominance and relative preservation of the hippocampus, was apparent ( Figure 1). CSF biomarker analysis revealed increased τ T = 511.6 pg/mL, normal τ P-181 = 28.7 pg/mL, decreased Aβ 42 = 504.2 pg/mL, but normal Aβ 42 /Aβ 40 = 0.1 and decreased τ P-181 /τ T = 0.056 ratio. TDP-43 = 4.46 pg/mL and the formula TDP-43 × τ T /τ P-181 = 79.5 were increased. Genetic testing revealed a c.463-2A > G splice site variant-heterozygous mutation in the GRN gene. The patient also had more than a 2-fold decrease in plasma progranulin levels, as described elsewhere [28].
AD biomarkers (mainly for excluding AD-related dementia). Cut-offs for normal values are shown in Table 1.

Case 1
A 61-year-old female patient was admitted to our department for progressive language impairment over the past 1-2 years. She had word-finding difficulty, with intact comprehension, an absence of phonological speech or grammar errors, no other significant cognitive/behavioral disturbances, and preserved activities of daily living. Neuropsychological assessment revealed mild cognitive impairment, with an MMSE score of 25/30 and a FAB score of 12/18. Her writing skills and phrase and sentence repetition were preserved, too. On magnetic resonance imaging (MRI), moderate cortical atrophy in the frontal and temporal lobes, with left predominance and relative preservation of the hippocampus, was apparent ( Figure 1). CSF biomarker analysis revealed increased τT = 511.6 pg/mL, normal τP-181 = 28.7 pg/mL, decreased Aβ42 = 504.2 pg/mL, but normal Aβ42/Aβ40 = 0.1 and decreased τP-181/τT = 0.056 ratio. TDP-43 = 4.46 pg/mL and the formula TDP-43 × τΤ/τP-181 = 79.5 were increased. Genetic testing revealed a c.463-2A > G splice site variantheterozygous mutation in the GRN gene. The patient also had more than a 2-fold decrease in plasma progranulin levels, as described elsewhere [28].

Case 2
A 60-year-old male patient, with a family history of dementia on his mother and maternal grandmother's side, presented with language impairment over the last 10 years. Initially, he had difficulty naming objects and writing. Three years later he developed effort-ful speech with sound errors, decreased fluency, and memory complaints and disorientation. At the age of 55, a dysexecutive syndrome as well as behavioral disturbances (overeating and irritability) appeared. Gradually, he became almost mute, while bradykinesia and tremor on the right hand appeared. At the time of his admission, he was completely mute, and unable to eat. Neurological examination revealed diffuse muscle atrophies without fasciculations, bilateral Parkinsonian syndrome (more severe in the right limbs), pyramidal syndrome with spasticity, hyperreflexia, and Babinski sign bilaterally. Neuropsychological assessment was impossible to perform. MRI performed at the age of 57 revealed significant cortical atrophy in the frontal and temporal lobes with left predominance (Figure 2A-C), with further deterioration three years later, as shown in (Figure 2D-F). CSF biomarker analysis revealed increased τ T = 565.12 pg/mL, normal τ P-181 = 41.9 pg/mL, Aβ 42 = 640 pg/m and Aβ 42 /Aβ 40 = 0.11, while τ P-181 /τ T = 0.074, was decreased. TDP-43 = 5.92 pg/mL and formula TDP-43 × τ T /τ P-181 = 79.9 were increased. Genetic testing revealed a c.934-1G > A splice site variant-heterozygous mutation on the GRN gene, and similarly, a decrease in plasma progranulin levels, as in case 1.

Case 2
A 60-year-old male patient, with a family history of dementia on his mother and ma ternal grandmother's side, presented with language impairment over the last 10 years Initially, he had difficulty naming objects and writing. Three years later he developed ef fortful speech with sound errors, decreased fluency, and memory complaints and disori entation. At the age of 55, a dysexecutive syndrome as well as behavioral disturbance (overeating and irritability) appeared. Gradually, he became almost mute, while bradyki nesia and tremor on the right hand appeared. At the time of his admission, he was com pletely mute, and unable to eat. Neurological examination revealed diffuse muscle atro phies without fasciculations, bilateral Parkinsonian syndrome (more severe in the righ limbs), pyramidal syndrome with spasticity, hyperreflexia, and Babinski sign bilaterally Neuropsychological assessment was impossible to perform. MRI performed at the age o 57 revealed significant cortical atrophy in the frontal and temporal lobes with left predom inance (

Case 3
A 72-year-old male patient, with a family history of motor neuron disease (brother) gradually developed language (phonological speech errors, impaired comprehension, in ability to write, severe difficulty in naming objects and reduced phonemic and categorica fluency) and memory impairment, bradykinesia and stereotypical movements of the righ hand over the last two years. Three months before admission, his symptoms aggravated Neurological examination revealed bilateral akinetic-rigid Parkinsonian syndrome, mild pyramidal signs (mainly on the right lower limb with hyperreflexia and extensor planta response) and apraxia. Neuropsychological assessment revealed severe diffuse cognitiv deficits (MMSE score 5/30, FAB score 1/18). On MRI, severe cortical atrophy in frontal and temporal lobes was observed ( Figure 3). CSF biomarker analysis revealed normal τT = 316.16 pg/mL and τP-181 = 27.2 pg/mL, decreased Aβ42 = 501.51 pg/mL but normal Αβ42/Αβ4 = 0.14 and low τP-181/τT = 0.086 ratio. TDP-43 levels were 5.74 pg/mL and TDP-43 × τΤ/τP-18

Case 3
A 72-year-old male patient, with a family history of motor neuron disease (brother), gradually developed language (phonological speech errors, impaired comprehension, inability to write, severe difficulty in naming objects and reduced phonemic and categorical fluency) and memory impairment, bradykinesia and stereotypical movements of the right hand over the last two years. Three months before admission, his symptoms aggravated. Neurological examination revealed bilateral akinetic-rigid Parkinsonian syndrome, mild pyramidal signs (mainly on the right lower limb with hyperreflexia and extensor plantar response) and apraxia. Neuropsychological assessment revealed severe diffuse cognitive deficits (MMSE score 5/30, FAB score 1/18). On MRI, severe cortical atrophy in frontal and temporal lobes was observed ( Figure 3). CSF biomarker analysis revealed normal τ T = 316.16 pg/mL and τ P-181 = 27.2 pg/mL, decreased Aβ 42 = 501.51 pg/mL but normal Aβ 42 /Aβ 40 = 0.14 and low τ P-181 /τ T = 0.086 ratio. TDP-43 levels were 5.74 pg/mL and TDP-43 × τ T /τ P-181 = 66.7, both increased. The genetic testing revealed a pathologic hexanucleotide repeat expansion in the C9orf72 gene. = 66.7, both increased. The genetic testing revealed a pathologic hexanucleotide repeat expansion in the C9orf72 gene.

Case 4
A 54-year-old female patient, with a medical history of depression with psychotic features was referred for evaluation. She had a family history of dementia (mother) and symptoms of motor neuron disease (two maternal uncles). The patient presented with apathy, inertia, disinhibition, loss of empathy and sympathy, unusual stereotyped and perseverative behaviors, and hyperorality with overeating of carbohydrates during the last two and a half years. One and a half year before her admission, she also developed inappropriate crying and a language impairment. Neurological examination revealed severely impaired non-fluent speech with single-word answers, severe disorientation in time and space, and frontal gait disorder. Neuropsychological assessment revealed moderate cognitive deficits with an MMSE score of 15/30 and a FAB score of 6/18. On MRI, some T2 hyperintense lesions attributed to chronic small vessel ischemia (Fazekas II), and mild cortical atrophy of the temporal lobes was observed ( Figure 4). CSF biomarker analysis revealed normal τT = 250.3 pg/mL, τP-181 = 35.45 pg/mL and Aβ42 = 935.95 pg/mL, while τP-181/τT ratio was 0.14. The levels of TDP-43 = 2.48 pg/mL and TDP-43 × τΤ/τP-181 = 17.5 were also normal. Genetic testing, however, revealed a hexanucleotide repeat expansion in the C9orf72 gene.

Case 4
A 54-year-old female patient, with a medical history of depression with psychotic features was referred for evaluation. She had a family history of dementia (mother) and symptoms of motor neuron disease (two maternal uncles). The patient presented with apathy, inertia, disinhibition, loss of empathy and sympathy, unusual stereotyped and perseverative behaviors, and hyperorality with overeating of carbohydrates during the last two and a half years. One and a half year before her admission, she also developed inappropriate crying and a language impairment. Neurological examination revealed severely impaired non-fluent speech with single-word answers, severe disorientation in time and space, and frontal gait disorder. Neuropsychological assessment revealed moderate cognitive deficits with an MMSE score of 15/30 and a FAB score of 6/18. On MRI, some T2 hyperintense lesions attributed to chronic small vessel ischemia (Fazekas II), and mild cortical atrophy of the temporal lobes was observed ( Figure 4). CSF biomarker analysis revealed normal τ T = 250.3 pg/mL, τ P-181 = 35.45 pg/mL and Aβ 42 = 935.95 pg/mL, while τ P-181 /τ T ratio was 0.14. The levels of TDP-43 = 2.48 pg/mL and TDP-43 × τ T /τ P-181 = 17.5 were also normal. Genetic testing, however, revealed a hexanucleotide repeat expansion in the C9orf72 gene.
= 66.7, both increased. The genetic testing revealed a pathologic hexanucleotide repeat expansion in the C9orf72 gene.

Case 4
A 54-year-old female patient, with a medical history of depression with psychotic features was referred for evaluation. She had a family history of dementia (mother) and symptoms of motor neuron disease (two maternal uncles). The patient presented with apathy, inertia, disinhibition, loss of empathy and sympathy, unusual stereotyped and perseverative behaviors, and hyperorality with overeating of carbohydrates during the last two and a half years. One and a half year before her admission, she also developed inappropriate crying and a language impairment. Neurological examination revealed severely impaired non-fluent speech with single-word answers, severe disorientation in time and space, and frontal gait disorder. Neuropsychological assessment revealed moderate cognitive deficits with an MMSE score of 15/30 and a FAB score of 6/18. On MRI, some T2 hyperintense lesions attributed to chronic small vessel ischemia (Fazekas II), and mild cortical atrophy of the temporal lobes was observed ( Figure 4). CSF biomarker analysis revealed normal τT = 250.3 pg/mL, τP-181 = 35.45 pg/mL and Aβ42 = 935.95 pg/mL, while τP-181/τT ratio was 0.14. The levels of TDP-43 = 2.48 pg/mL and TDP-43 × τΤ/τP-181 = 17.5 were also normal. Genetic testing, however, revealed a hexanucleotide repeat expansion in the C9orf72 gene.

Case 5
A 62-year-old male patient, with a medical history of arterial hypertension and an operation for a herniated lumbar disc at the L4-L5 level, was referred for evaluation of neuropsychiatric symptoms. He had a family history of "presenile dementia" in his mother, ALS of his older brother, who died at the age of 53, and behavioral symptoms of another brother. The patient presented with progressively deteriorating steppage gait 20 years ago. A year before admission, the patient was unable to walk independently. Additionally, he developed obsessive behavior, aggression, palilalia and apathy. His neuropsychological evaluation revealed severe cognitive impairment, suggestive of predominant frontal lobe dysfunction, MMSE: 2 and FAB: 0 (not applicable). Brain MRI showed mild frontal lobe atrophy ( Figure 5). EMG showed diffuse myopathic findings and mild spontaneous activity. Due to the non-AD biomarker profile, the family history of dementia and ALS in members of consecutive generations, which is indicative of a possible autosomal dominant inheritance pattern and the clinical and laboratory findings suggesting myopathy, a high suspicion of a possible VCP-associated syndrome was raised. Genetic testing (DNA isolated from skeletal muscle) revealed a pathogenic heterozygous missense mutation p.R159H (c.476G > A) in the VCP gene, confirming the diagnosis of IBMPFD [29].

Case 5
A 62-year-old male patient, with a medical history of arterial hypertension and an operation for a herniated lumbar disc at the L4-L5 level, was referred for evaluation of neuropsychiatric symptoms. He had a family history of "presenile dementia" in his mother, ALS of his older brother, who died at the age of 53, and behavioral symptoms of another brother. The patient presented with progressively deteriorating steppage gait 20 years ago. A year before admission, the patient was unable to walk independently. Additionally, he developed obsessive behavior, aggression, palilalia and apathy. His neuropsychological evaluation revealed severe cognitive impairment, suggestive of predominant frontal lobe dysfunction, MMSE: 2 and FAB: 0 (not applicable). Brain MRI showed mild frontal lobe atrophy ( Figure 5). EMG showed diffuse myopathic findings and mild spontaneous activity. Due to the non-AD biomarker profile, the family history of dementia and ALS in members of consecutive generations, which is indicative of a possible autosomal dominant inheritance pattern and the clinical and laboratory findings suggesting myopathy, a high suspicion of a possible VCP-associated syndrome was raised. Genetic testing (DNA isolated from skeletal muscle) revealed a pathogenic heterozygous missense mutation p.R159H (c.476G > A) in the VCP gene, confirming the diagnosis of IBMPFD [29]. Lumbar puncture and CSF biomarker analysie were performed one year later and revealed normal τT = 356.75 pg/mL, τP-181 = 32.77 pg/mL, low Aβ42 = 431 pg/mL, but normal Αβ42/Αβ40 = 0.1 and reduced ratio τP-181/τT = 0.091, while TDP-43 levels = 5.27 pg/mL, and TDP-43 × τΤ/τP-181 = 57.3 were increased.

Case 6
A 41-year-old male patient, with a medical history of dyslipidemia was admitted to the neurology department for delusions of religious content over the past 12 years. Moreover, apathy, a change in eating habits with carbohydrate craving and weight gain gradually developed, while dysarthria, executive deficits, stereotypical behaviors and inappropriate laughter appeared over the last two years. Finally, difficulty in swallowing liquids was noticed six months before admission. Neurological examination revealed pseudobulbar dysarthria, an absence of pharyngeal reflex, bilateral pyramidal syndrome and fasciculations in tongue and back muscles. Neuropsychological assessment revealed mild to moderate frontal cognitive deficits with an MMSE score 28/30, and a FAB score 15/18. Neurophysiological examination could not confirm lower motor neuron involvement. On MRI, significant cortical atrophy in frontal, parietal and temporal lobes was observed (Figure 6). CSF biomarker analysis revealed normal τT = 328.05 pg/mL, τP-181 = 32.07 pg/mL, decreased Aβ42 = 397.1 pg/mL, with normal ratio Αβ42/Αβ40 = 0.17 and decreased τP-181/τT = Lumbar puncture and CSF biomarker analysie were performed one year later and revealed normal τ T = 356.75 pg/mL, τ P-181 = 32.77 pg/mL, low Aβ 42 = 431 pg/mL, but normal Aβ 42 /Aβ 40 = 0.1 and reduced ratio τ P-181 /τ T = 0.091, while TDP-43 levels = 5.27 pg/mL, and TDP-43 × τ T /τ P-181 = 57.3 were increased.

Case 6
A 41-year-old male patient, with a medical history of dyslipidemia was admitted to the neurology department for delusions of religious content over the past 12 years. Moreover, apathy, a change in eating habits with carbohydrate craving and weight gain gradually developed, while dysarthria, executive deficits, stereotypical behaviors and inappropriate laughter appeared over the last two years. Finally, difficulty in swallowing liquids was noticed six months before admission. Neurological examination revealed pseudobulbar dysarthria, an absence of pharyngeal reflex, bilateral pyramidal syndrome and fasciculations in tongue and back muscles. Neuropsychological assessment revealed mild to moderate frontal cognitive deficits with an MMSE score 28/30, and a FAB score 15/18. Neurophysiological examination could not confirm lower motor neuron involvement. On MRI, significant cortical atrophy in frontal, parietal and temporal lobes was observed ( Figure 6). CSF biomarker analysis revealed normal τ T = 328.05 pg/mL, τ P-181 = 32.07 pg/mL, decreased Aβ 42 = 397.1 pg/mL, with normal ratio Aβ 42 /Aβ 40 = 0.17 and decreased τ P-181 /τ T = 0.098, while TDP-43 levels 5.74 pg/mL and the formula TDP-43 × τ T /τ P-181 = 57.66 were increased. Genetic testing finally revealed a hexanucleotide repeat expansion in the C9orf72 gene.

Discussion
Herein, we present six patients with clinical, neuropsychological and imaging data fulfilling the criteria of Rascovsky et al. [24] and Gorno-Tempini et al. [25] for probable FTD, in which, due to the genetic findings, patients were classified as definite FTLD with TDP-43 histopathology. Three of our cases had a pathogenic hexanucleotide repeat expansion in the C9orf72 gene; two had a novel mutation in the GRN gene which led to more than a 2-fold decrease of plasma progranulin levels, as expected [28,30]; meanwhile, one patient harbored a pathogenic mutation in the VCP gene. Although small in number, this group of patients is indicative of the most common mutations found in the Greek population [28,31].  Table 2 summarizes the CSF biomarker profile for all cases. All cases had normal τ P-181 and only two out of six cases had increased τ T . Four out of six cases had decreased Aβ 42 ; however, all had a normal Aβ 42 /Aβ 40 ratio. Five out of six patients had increased TDP-43 and formula (TDP-43 × τ T /τ P-181 ) values, while τ P-181 /τ T was abnormal in all.

Discussion
Herein, we present six patients with clinical, neuropsychological and imaging data fulfilling the criteria of Rascovsky et al. [24] and Gorno-Tempini et al. [25] for probable FTD, in which, due to the genetic findings, patients were classified as definite FTLD with TDP-43 histopathology. Three of our cases had a pathogenic hexanucleotide repeat expansion in the C9orf72 gene; two had a novel mutation in the GRN gene which led to more than a 2-fold decrease of plasma progranulin levels, as expected [28,30]; meanwhile, one patient harbored a pathogenic mutation in the VCP gene. Although small in number, this group of patients is indicative of the most common mutations found in the Greek population [28,31].
According to the AT(N) system proposed by the National Institute of Aging and Alzheimer's Association (NIA-AA) Research Framework group, AD biomarkers are divided into those with molecular specificity for AD, namely τ P-181 , Aβ 42 and Aβ 42 /Aβ 40 ratio, and those with neurodegeneration, τ T in the present case [15]. All presented cases had a biomarker profile of either A-T-(N+) or A-T-(N−) according to CSF data, compatible with a "non-Alzheimer's pathological change" profile. Only one patient (case 4) had a "normal" biomarker profile.
In four out of the six presented cases, Aβ 42 was reduced, but the use of Aβ 42 /Aβ 40 ratio restored normality, ruling out amyloidopathy (A). The Aβ 42 /Aβ 40 ratio has been suggested to better reflect the presence or absence of amyloid pathology than Aβ 42 alone. Furthermore, the Aβ 42 /Aβ 40 ratio correlates better with amyloid load in PET (positron emission tomography) [32], as it has been found to be reduced in some cases of FTD and vascular dementia (VD), but most commonly in dementia with Lewy bodies (DLB) [33][34][35][36].
The main findings of our study, however, were the increase of CSF TDP-43 levels (higher than the above cut-off value) in five out of six cases, and the increased values of the TDP-43 × τ T /τ P-181 formula. TDP-43 is a protein, which may be suggestive and specific of TDP-43 proteinopathy. Additionally, the τ P-181 /τ T ratio was decreased, which in non-AD (FTD) patients could be considered as an indirect indication for a non-tauopathic pathology [22,23].
TDP-43 aggregation appears in approximately 50% of all FTLD cases (familial and sporadic) and in 97% of all ALS cases [37]. Mutations associated with TDP-43 histopathology have been documented for C9orf72, GRN, VCP, and TARDBP genes for FTLD and the FTD-ALS spectrum [38]. TDP-43 was initially considered as an intracellular/intranuclear protein; however, it is now recognized as an important protein for the existence and wellbeing of cells, through its various functions in RNA metabolism and homeostasis. Apart from FTLD, abnormalities in this protein's homeostasis are associated with other severe neurodegenerative disorders [39]. Abnormal TDP-43 seems to cause neurotoxicity by various pathogenic processes, although the exact pathophysiological mechanisms that eventually lead to neurodegeneration remain unclear [1].
Although thought to exist as an intracellular/intranuclear protein, recent studies using ELISA and Western blot techniques suggest that TDP-43 can also be detected in extracellular fluids such as plasma and CSF [40].
The results of the present study indicate that CSF TDP-43 levels were increased in five out of six patients with TDP-43 histopathology-associated mutations. Case 4, with a C9orf72 pathological repeat expansion, represents an exception. Nonetheless, the proposed mechanisms through which C9orf72 mutations cause neurodegenerative changes are three: C9orf72 loss-of-function through haploinsufficiency, RNA toxic gain of function, and gain of function through the aggregation of toxic dipeptide repeat proteins [41]. The first two mechanisms cause neurodegeneration related to TDP-43 aggregation. The "gain of function" hypothesis due to the formation and accumulation of dipeptide repeats (DPRs) causes neurodegeneration via repeat-associated non-AUG (RAN) translation of the hexanucleotide GGGGCC from both sense and antisense strands [42]. As a result, there is a production of five distinct DPRs from sense {(poly-GA (glycine-alanine), poly-GP (glycine-proline), poly-GR (glycine-arginine) and antisense poly-GP (glycine-proline), as well as from poly-PR (proline-arginine) and poly-PA (proline-alanine)} strands [43,44]. These DPRs are amyloidogenic, and accumulate in different parts of the central nervous system without TDP-43 aggregation [43][44][45]. The pathogenic contribution of DPR-associated toxicity to disease progression remains unknown. As there is no aggregated TDP-43 in neuronal cytoplasmic inclusions in such patients, we can assume that the TDP-43 concentration in their CSF would be normal. This reasoning provides an explanation for the normal values of TDP-43 and formula in case 4.
TDP-43 abnormalities have also been observed in other neurodegenerative diseases as well, mainly in ALS, the other end of the ALS-FTD spectrum, but also in AD, limbicpredominant age-related TDP-43 encephalopathy (LATE), cerebral age-related TDP-43 with sclerosis (CARTS), and some other even rarer conditions [27,46,47]. Reports on CSF TDP-43 levels are mostly derived from patient cohorts of sporadic FTD cases, with inadequate (74%) clinicopathological concordance rates [48], resulting in relatively low sensitivities and specificities vs. controls. Thus, we enrolled genetic cases in order to reach safer conclusions. Sensitivities and specificities compared with other dementing disorders would have to be investigated as well.
There are also other biomarkers that can be used to differentiate FTD patients from patients with other neurodegenerative diseases, such as NfL [49]. However, NfL, total tau protein and the specific patterns of atrophy, are considered biomarkers of neurodegeneration in general [15]. The aim of the current study was to assess the diagnostic value of CSF TDP-43 protein as a biomarker with molecular specificity for FTD. Therefore, we combined TDP-43 with established AD biomarkers only, mainly to exclude AD, which often enters into a FTD differential diagnosis.
The current study has certain limitations. TDP-43 pathology can only be presumed from genetic mutations, since it is an ante-mortem study. Another disadvantage is the small number of patients, coming from a single-center study that included only patients of Greek origin; however, the mutations that the patients harbored are indicative of the most commonly observed in TDP-43-related FTLD found in the Greek population [28,31].
Our study also has some important advantages. We describe patients that have undergone extended clinical, neuropsychological and imaging examinations, fulfilling the criteria of Rascovsky et al. [24] and Gorno-Tempini et al. [25] for probable FTD, while they are classified as definite FTLD with TDP-43 underlying pathology, due to the genetic findings. This homogeneous genetic FTD cohort may serve as a very good example of how relatively safe conclusions may be reached through personalized and precision medicine. Further studies with larger patient cohorts are nevertheless required to corroborate our results.

Conclusions
Based on the results of the present personalized case series study, which is nonetheless indicative of the mutational profile of the population of Greece, the following conclusions can be drawn:

•
Established AD biomarkers are reliable indicators for the exclusion of AD pathology in genetic FTLD. • A decreased τ P-181 /τ T ratio may be useful as an indirect marker for non-tauopathic pathology. • TDP-43 alone and combined with tau proteins, could be a useful tool for the diagnosis of genetic FTD patients with TDP-43 underlying pathology.
Future studies with larger and well characterized patient cohorts are needed to further explore the diagnostic potential of these biomarkers in differentiating TDP-43 proteinopathy from non-TDP-43-related pathologies.  Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee and the Scientific Board of "Eginition" Hospital (protocol code: 244, date of approval: 5-7-2010).
Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.