Motor, cognitive and behavioural profiles of C9orf72 expansion-related amyotrophic lateral sclerosis

Introduction Amyotrophic lateral sclerosis (ALS) individuals carrying the hexanucleotide repeat expansion (HRE) in the C9orf72 gene (C9Pos) have been described as presenting distinct features compared to the general ALS population (C9Neg). We aim to identify the phenotypic traits more closely associated with the HRE and analyse the role of the repeat length as a modifier factor. Methods We studied a cohort of 960 ALS patients (101 familial and 859 sporadic cases). Motor phenotype was determined using the MRC scale, the lower motor neuron score (LMNS) and the Penn upper motor neuron score (PUMNS). Neuropsychological profile was studied using the Italian version of the Edinburgh Cognitive and Behavioral ALS Screen (ECAS), the Frontal Behavioral Inventory (FBI), the Beck Depression Inventory-II (BDI-II) and the State-Trait Anxiety Inventory (STAI). A two-step PCR protocol and Southern blotting were performed to determine the presence and the size of C9orf72 HRE, respectively. Results C9orf72 HRE was detected in 55/960 ALS patients. C9Pos patients showed a younger onset, higher odds of bulbar onset, increased burden of UMN signs, reduced survival and higher frequency of concurrent dementia. We found an inverse correlation between the HRE length and the performance at ECAS ALS-specific tasks (P = 0.031). Patients also showed higher burden of behavioural disinhibition (P = 1.6 × 10–4), lower degrees of depression (P = 0.015) and anxiety (P = 0.008) compared to C9Neg cases. Conclusions Our study provides an extensive characterization of motor, cognitive and behavioural features of C9orf72-related ALS, indicating that the C9orf72 HRE size may represent a modifier of the cognitive phenotype. Supplementary Information The online version contains supplementary material available at 10.1007/s00415-022-11433-z.


Introduction
Amyotrophic lateral sclerosis (ALS) is a fatal neurodegenerative disorder characterized by a progressive loss of upper (UMN) and lower motor neurons (LMN). Approximately 90% of ALS cases are sporadic (SALS), while the remaining 10% are familial (FALS). Mutations in four main genes (C9orf72, SOD1, TARDBP and FUS) are responsible for up to 75% of FALS cases, with variants in > 25 other genes being relatively uncommon [1].
A (G 4 C 2 ) n hexanucleotide repeat expansion (HRE) in the C9orf72 gene accounts for 30-50% of FALS, as well as 5-10% of SALS cases [2], and represents the most common genetic defect in ALS and in frontotemporal dementia (FTD) [3,4]. 1 3 ALS individuals carrying the C9orf72 HRE (C9Pos) often present different phenotypic traits compared to the remaining patients (C9Neg), showing a more aggressive form of the disease characterized by a higher prevalence of bulbar onset, earlier age at onset, and reduced survival [2,[4][5][6][7][8]. C9Pos cases also have an increased frequency of comorbid frontotemporal dementia (ALS-FTD) and a higher prevalence of a family history of ALS and/or other neurodegenerative diseases [2,5,8,9]. Intriguingly, the same HRE length may cause ALS, FTD or mixed ALS-FTD manifestations with no clear genotype-phenotype correlations [10,11]. However, the current molecular diagnosis of C9orf72 HRE by repeat-primed PCR (RP-PCR) gives no information on HRE length above a certain threshold (50-150 repeats), and the difficulty in performing and including Southern blot assay in a clinical setting has made accurate genotype-phenotype correlations inconclusive. Moreover, in contrast to other repeat expansion disorders, C9Pos patients show neither consistent association between HRE size and disease severity nor clear genetic anticipation [12].
The recognition of C9Pos individuals is a key issue, both for stratification in randomized clinical trials and because the development of antisense oligonucleotide-based therapies is already underway [13]. In fact, the presence of C9orf72 HRE is an independent predictor of shorter survival in ALS, and it has been included in prediction models [14]. Here we studied a large cohort of Italian ALS patients with the aim of identifying the phenotypic traits more closely associated with C9orf72 HRE and to assess any possible correlations of the HRE length with motor, cognitive and behavioural features.

Patients and methods
Patients A cohort of 960 Italian patients diagnosed with ALS and other motor neuron diseases (primary lateral sclerosis [PLS] and progressive muscular atrophy [PMA]) according to El Escorial revised criteria [15] was consecutively recruited at a tertiary ALS Centre (IRCCS Istituto Auxologico Italiano, Milan, Italy) between 2008 and 2021. The following data were collected: sex; age at onset; family history of ALS, dementia and/or parkinsonism; site of onset; phenotype; ALSFRS-R score at evaluation; progression rate; clinical stages according to the King and Milano-Torino (MITOS) staging systems [16]; survival.

Standard protocol approvals and patients' consent
Informed consent for using pseudo-anonymized data for research purposes was obtained from all patients or their authorized legal representatives. This study was approved by the Ethics Committee of Istituto Auxologico Italiano IRCCS (project DAMARE 2021_05_18) and conducted according to the principles expressed in the Declaration of Helsinki.

Evaluation of motor phenotype
The burden of UMN and LMN involvement was determined using three different semi-quantitative scales, namely the MRC (Medical Research Council) scale for muscle strength, the LMN score (LMNS) [17], and the Penn UMN score (PUMNS) [18]. For MRC the strength of three different muscle groups for each limb (shoulder abductors, elbow flexors, wrist dorsiflexors, hip flexors, knee extensors and ankle dorsiflexors) was assessed. The LMNS defines the burden of LMN signs in each limb (0-3), with higher scores corresponding to greater impairment [17]. The score was modified to consider the presence of LMN signs also in the thoracic and bulbar regions, assigning 1 point each, for a total value ranging from 0 to 14. Lastly, PUMNS was used to define the severity of UMN impairment in the bulbar region (score 0-4) and each limb (score 0-7), for a total score of 0-32.

Neuropsychological testing
The cognitive profile was investigated using the Italian version of the Edinburgh Cognitive and Behavioural ALS Screen (ECAS) [19] and the Frontal Behavioural Inventory (FBI) [20]. Comorbid FTD was determined according to the criteria proposed by Neary [21] and Rascovsky [22]. Patients were subsequently classified according to the Strong revised criteria [unimpaired, behavioural (ALSbi), cognitive (ALSci), cognitive and behavioural impairment (ALScbi), ALSFTD] [23]. For cognitive status assessment, the following scores were considered: ECAS total, ALS-Specific, ALS-Nonspecific, as well as the single ECAS subdomains (executive, verbal fluency, language, memory, visuospatial) scores. The behavioural profile was evaluated considering the symptoms detected at the ECAS Carer Interview (disinhibition, apathy/inertia, loss of sympathy/empathy, perseverative/stereotyped/compulsive/ritualistic behaviour and hyperorality/altered food preferences), as well as the scores at FBI-A (negative behaviours) and FBI-B (positive/disinhibited behaviours). Psychological status was explored with the Beck Depression Inventory-II (BDI-II; cognitive-affective and somatic symptoms) [24] and the State-Trait Anxiety Inventory (STAI; Y1 state anxiety, Y2 trait anxiety) [25] C9orf72 genetic analysis Genetic analysis of C9orf72 was performed on DNA extracted from peripheral blood using a two-step PCR protocol, including a first fluorescent amplicon-length analysis with primers designed in the unique sequence flanking the hexanucleotide repeat (FAM-labelled forward 5′-TGT AAA ACG ACG GCC AGT CAA GGA GGG AAA CAA CCG CAGCC-3′ and reverse 5′-GCA GGC ACC GCA ACC GCA G-3′) and, only for samples showing one fluorescent peak, the Repeat-PCR (RP-PCR) was carried out, as previously described [10]. For both amplicon length analysis and RP-PCR, amplicons were run on ABI Prism 3500 Genetic Analyzer (Applied Biosystems) and visualized using Gene Mapper v.4 software (Applied Biosystems). A cut-off value of > 30 repeats was used to define the pathogenic threshold in RP-PCR assay.

Southern blotting
Genomic DNA (12 μg) was digested with XbaI restriction enzyme and run on 0.7% agarose gel. Southern blot was performed using a unique sequence probe, mapping within C9orf72 first intron and obtained by PCR amplification with the following primers: forward 5′-CTT TCT CCA GAT CCA GCA GCC TCC -3′ and reverse 5′-CTG AGT TCC AGA GCT TGC TACAG-3′. Upon probe labelling with dCTP 32 , hybridized membranes were resolved by autoradiography as previously described [10]. For each sample, the HRE size was calculated as mode of the observed smear range [26].

Statistical analysis
Descriptive statistics are reported as numbers and percentages for categorical variables or mean, median and standard deviation for continuous variables. Statistical analysis was carried out with IBM Statistical Package for Social Science (SPSS) version 26. Univariate survival analysis was performed with Kaplan-Meier curves followed by log-rank test. Multivariate survival analysis was performed with Cox regression, using C9orf72 status, age at onset, site of onset and gender as covariates. Differences in C9orf72 status for categorical variables were assessed with Pearson's χ 2 test or Fisher exact test when appropriate. Mann-Whitney U test was used for continuous variables. Pearson bivariate correlation was performed to evaluate the relationship between HRE size and continuous variables. Two-tailed P values < 0.05 were considered significant. Listwise deletion was used to exclude cases missing C9orf72 HRE status. Pairwise deletion was used to handle missing data for correlation with other phenotypic traits.

Data availability
Pseudo-anonymized data are archived on Zenodo and will be disclosed upon reasonable request (doi:10.5281/ zenodo.6245606).
To better assess genotype-phenotype correlations in ALS patients carrying the C9orf72 gene mutation, we determined HRE size by Southern blot in 46/55 C9Pos blood samples. All the C9Pos samples identified by RP-PCR were confirmed as carrying a HRE by Southern blot assay. Our analysis revealed high HRE mosaicism, with G 4 C 2 repeats ranging from about 600 to 7000 units (average length ~ 2150 units) (Fig. 2). When we correlated HRE size with age at onset, survival and other clinical parameters, no significant associations were identified (data not shown).   Fig. 1d) and LMNS scores ( Supplementary Fig. 1e

Assessment of cognitive and behavioural phenotype
In our cohort, C9orf72 HRE was significantly associated to comorbid behavioural variant FTD (bvFTD) [ Similarly, we did not find any difference in ECAS total, ALS-specific, ALS-nonspecific, and individual subdomain scores, either considering the raw scores or the normality cut-offs for the Italian population (Table 3). Since Southern blot data were available for 17/20 C9Pos patients that underwent ECAS, the possible association between HRE size and cognitive features was tested. Interestingly, we found a modest, but significant inverse correlation between C9orf72 HRE length and the ECAS total score (R 2 = 0.261, P = 0.036). This phenomenon appeared to be mostly driven by worse performances at the ALS-specific tasks (R 2 = 0.275, P = 0.031) and specifically at the executive subdomain (R 2 = 0.320, P = 0.018). Within this subdomain, dysfunction at sentence completion (R 2 = 0.404, P = 0.006) and social cognition (R 2 = 0.250, P = 0.041) tasks seemed to account for the greater part of the impairment. These associations were still significant when accounting for age at DNA collection as a covariate. In contrast, no association was found between C9orf72 HRE length and ALS-nonspecific cognitive domains (Fig. 3). We obtained a complete ECAS Carer Interview for 184/211 (87.2%) patients that underwent ECAS, including 17 C9Pos and 167 C9Neg cases. In this subgroup, the most common behavioural symptom was  Fig. 2a-c). The presence of psychotic symptoms was also similar in both groups (data not shown). We also analysed the presence and the severity of depression and anxiety symptoms using the BDI-II and STAI questionnaires. Interestingly, HRE was significantly associated with lower scores at STAI-Y1 (45.2 ± 6.9 vs 50.8 ± 9.9, P = 0.008), STAI-Y2 (45.1 ± 8.4 vs 50.2 ± 10.3, P = 0.033), BDI-II cognitiveaffective (3.7 ± 3.1 vs 6.6 ± 5.8, P = 0.018), BDI-II somatic (5.5 ± 3.1 vs 7.6 ± 4.4, P = 0.042), and BDI-II total scores (9.2 ± 5.6 vs 14.2 ± 9.0, P = 0.015), indicating lower levels of self-reported depression and anxiety in C9Pos patients ( Table 3, Supplementary Fig. 2d-h). However, no significant association was present between HRE size and the manifestation of behavioural, depression and anxiety symptoms.

Discussion
In our single-centre analysis of 960 Italian ALS patients, we compared motor, cognitive and behavioural profiles between C9Pos and C9Neg individuals to define if these two groups have different phenotypes. Our results largely confirmed previous evidence from the literature. Indeed, our C9Pos patients showed a younger onset, higher odds of bulbar onset and a reduced survival [2,[4][5][6][7][8]. As expected, they also showed a higher prevalence of family history of ALS and dementia. On the contrary, no family history was observed for parkinsonian syndromes confirming the hypothesis that, although C9orf72 HRE can be observed in ALS-plus phenotypes, the mutation is strictly related to TDP-43 proteinopathies [27,28]. In our analysis, C9Pos patients presented with a more severe UMN involvement in all body regions and specifically in the bulbar segment. While different studies have reported the association between C9orf72 HRE and bulbar onset [4][5][6]8], ours is the first study to show a major burden of UMN involvement at the global and regional level in C9Pos patients. As for the cognitive phenotype, C9Pos patients more frequently presented a phenotype within the FTD spectrum disorders as already described [2,5,8,9]. The analysis of individual ECAS domains, however, did not show any qualitative difference between mutated and non-mutated patients, a result apparently in contrast with previous studies indicating a more severe impairment in executive functions and verbal memory in the former group [29]. However, when analysing the subgroup of patients for which Southern Blot was available, we found a statistically significant inverse correlation between HRE length and performance at the ECAS. This correlation was specifically relevant for ALS-specific cognitive functions, particularly for the executive function subdomain (sentence completion and social cognition tasks), thus suggesting a specific impairment in inhibitory cognitive components. Indeed, C9Pos patients with larger expansions had a more severe cognitive impairment, thereby supporting the emerging hypothesis that C9orf72 HRE size may be a modifier factor of phenotype along the ALS-FTD clinical spectrum [30]. Given the difficulty in obtaining informative Southern blots for C9orf72 HRE size and the high tissue mosaicism of the pathological expansion, conclusive evidence on the effect of HRE length on clinical phenotype, age of onset or disease duration is still lacking. Some studies describe significantly larger pathogenic HRE size in ALS patients compared to FTD cases [31,32], which was not confirmed by other reports [33]. We too did not find any association with HRE length and disease manifestation in an Italian C9Pos pedigree presenting high intrafamilial variability [10]. In contrast to other repeat expansion disorders, the phenomenon of anticipation is still disputable since both expansions and contractions of the G 4 C 2 repeat number have been described transgenerationally in C9Pos families [11,[34][35][36]. Also, with regard to disease duration, data are contrasting: some reports did not find any correlation with HRE size [31], while others observed that a longer expansion size was associated with shorter disease duration [32].
To our knowledge, very few reports have analyzed the differences in behavioral symptoms, degree of depression and anxiety between C9Pos and C9Neg patients. One study found that behavior was similarly impaired in both cohorts [29], while another report found increased apathy levels in C9Pos compared to C9Neg individuals [9]. Conversely, we did not appreciate any significant difference in the levels of apathy and other negative behavioural symptoms in our cohort. Remarkably, however, we found that behavioural disinhibition was significantly more present in C9Pos patients. This finding is further supported by the direct association between a more significant UMN impairment and the presence of behavioural symptoms [37], similarly to what observed in our cohort of C9Pos patients when compared to C9Neg ones. A previous study reported a higher presence of anxiety and depression in C9Neg vs C9Pos patients, even if the difference was not significant [29]. In a similar way, in our cohort C9Pos individuals showed significantly lower degrees of depression and anxiety when compared to C9Neg cases. This result could be interpreted as a loss of insight and a higher level of anosognosia in these patients. Loss of insight still represents an important under-recognized issue that should be measured within ALS patients' current assessment [38]. Future studies on awareness and emotional processing in C9orf72 patients will help to better clarify if the emotional response in these subjects is affected.
A limitation of our study is that ECAS was systematically performed in a sub-population of the whole cohort after the validation of the Italian version [19]. Despite this, no differences in demographic, clinical and motor features were observed between patients evaluated prior or after this timepoint (data not shown), suggesting that any bias due to missing data is limited. In fact, our study reports a consistent clinical description of a large cohort of Italian ALS patients in which C9Pos individuals were extensively characterized regarding the motor phenotype with the use of semiquantitative scales to assess the burden of UMN and LMN impairment. We also show for the first time that C9orf72 HRE size may represent a modifier of cognitive phenotype along the ALS-FTD spectrum. It must be noted that our observations are based on blood DNA samples and given the occurrence of HRE mosaicism across tissues, they might be not fully representative of the HRE in affected cerebral tissues. Other possible limitations of our study are the absence of longitudinal data, the lack of correlation with neuroimaging markers and the analysis of a centre-based cohort rather than a population registry.
In summary, our results showed a distinct motor, cognitive and behavioural profile in ALS patients carrying the C9orf72 HRE. We observed a significant association with the burden of UMN signs, the severity of ALS-specific cognitive impairment, as well as with the presence of positive behavioural alterations, such as disinhibition. Finally, anxiety and depression were uncommon in our C9Pos patients, and it is likely that the lower frequency of such features is related to reduced insight. A detailed motor, cognitive and behavioural assessment, including awareness and insight features, should thus represent a fundamental aspect of the clinical evaluation of ALS patients, specifically those carrying C9orf72 HRE. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.