Genetic and Neuroimaging Analysis of SIGMAR1 for Frontotemporal Dementia

INTRODUCTION

Frontotemporal dementia (FTD, OMIM 600274) is the second most important cause of dementia and accounted for approximately 10% of all pathologically diagnosed dementias who developed disease before the age of 65 [1]. It represents a spectrum of clinical presentations characterized by insidiously progressive deterioration in behavior and executive and language abilities [2–4]. The clinical presentations of FTD are heterogeneous mainly with two clinical variants: behavioral variant [5] and language variants (progressive nonfluent aphasia and semantic dementia) [6]. Progressive nonfluent aphasia typically results in disruption of phonological and syntactic components of language [7], and by contrast in semantic dementia there is progressive loss of the knowledge base underlying language, results in impaired comprehension with preservation of conversational fluency [8]. FTD has an important genetic component since approximately 40% patients have at least one first degree relative with a disease in the FTD spectrum, and 10–15% of patients have a family history of FTD [9].

The majority of the heritability FTD is accounted by autosomal dominant mutations in three genes [10]: microtubule-associated protein tau (MAPT) identified in 1998 [11], progranulin (GRN) reported in 2006 [12, 13], and chromosome 9 open reading frame 72 (C9orf72) identified in 2011 [14, 15]. Each genetic group causes between  5 and 10% of all FTD, with geographical variability in different case series [16]. In recent years, mutations in an increasing number of genes have been associated with FTD. Recently, genetic studies have identified Sigma nonopioid intracellular receptor 1 (SIGMAR1, NM 005866.2) plays an important role in familial FTLD-MND with genome-wide significant linkage to chromosome 9p [17]. However, it was later reported that the family Aus-14 with a 3’UTR variant (c.672*51 G>T) also has a repeat expansion mutation in C9orf72 (NM_001256054.1), and did not detect C9orf72 expansion mutations in the other family Aus-47 with SIGMAR1 variant c.672*26 C>T [17, 18]. Hence, it is important to verify and validate the contribution of SIGMAR1 independent of C9orf72 gene status to FTD spectrum syndromes. In the present study, first, we performed SIGMAR1 mutation analysis in FTD cases absent of C9orf72 pathogenic repeat expansions. Then, we also present data showing longitudinal neuroimaging changes in SIGMAR1 rare variant case.

MATERIALS AND METHODS

RESULTS

In this study of 82 patients, no abnormal hexanucleotide repeat expansions in C9orf72 gene were identified. Furthermore, we screened all exons and a minimum of 50 bp from each of the flanking introns in SIGMAR1 gene using sanger sequencing with six sets of primers in 82 sporadic FTD patients. We identified a rare variant (rs192856872) (MAF = 0.0002 in 1000Genomes) in a male with semantic dementia (Fig. 1), who had an onset age of 70 and MMSE at the examination was 20. This variant was not found in 417 controls. There are no mutations in MAPT, GRN, or CHCHD10 genes, and the hexanucleotide repeat number of C9orf72 gene in the rs192856872 variant carrier is 2 units.

OPEN IN VIEWER

Fig. 1

Sanger sequencing showed heterozygous mutation of A/C (FWD) at rs192856872 in a patient with semantic dementia.

For in silico analysis, ESEFinder showed that rs192856872 variant creates two new sites, SRSF1(SF2/ASF) binding site (CAGGGGG) with a score of 2.10338 (threshold 1.956), and SRSF1-IgM-BRCA1binding site (CAGGGGG) with a score of 2.65659 (threshold 1.867), so this change is predicted to affect normal splicing (Fig. 2).

OPEN IN VIEWER

Fig. 2

ESEFinder analysis showed the new binding sites of SRSF1 and SRSF1-IgM-BRCA1 binding sites. Region of variation in mutant sequence was encircled.

Regarding the phenotype of the patient, in early-stage, the patient carrying rs192856872 variant show typically asymmetrical atrophy of the anterior temporal lobes (Fig. 3). On coronal MRI, the hallmark is the “knife-edge” type atrophy of the anterior temporal lobes of the left side. The rs192856872 carrier with language deficits underwent routine clinical assessments including domains of general cognitive function (MMSE 20/30 & Memory and Executive Screening 41/100), language (Boston naming test 7/30& Animal Fluency Test 7 number), attention (Symbol Digit Modalities Test 22 number), executive function (Trail Making Test-A 80 seconds, Trail Making Test-B 175 seconds), visuospatial skills (Rey-Osterrich Complex Figure Test: copy 33/35, recall 3/36), and social cognitive function (Reading the Mind in the Eyes Test 16/36). Over 6 years of follow-up, as the disease progresses, the degeneration extends caudally into the left posterior temporal lobes in the majority, rostrally into the frontal lobes (Fig. 3). Furthermore, the disease progress to the right side after six years.

OPEN IN VIEWER

Fig. 3

Coronal MRI of the rs192856872 patient showed striking asymmetric (left > right) temporal lobe atrophy. Top: in the early stage; Bottom: six years later.

DISCUSSION

To our knowledge, this is the first genetic and neuroimaging investigation on SIGMAR1 gene from China. Before this study, we reported two mutations of MAPT gene, one variant of GRN gene and fifteen variants of CHCHD10 gene in 18 FTD patients previously [21]. However, no abnormal hexanucleotide repeat expansions in C9orf72 gene were identified in 82 patients for this study. Although the frequency of the SIGMAR1 gene variants in this Chinese FTD cohort is very rare (0.007), it is seven times higher compared to general East Asian population, which is 0.001 from the database of 1000 Genomes. The rare variant rs192856872 (MAF = 0.0002 in 1000Genomes) is located in 3’UTR of SIGMAR1, similar to other FTD causative SIGMAR1 variants: c.672*26 C>T/43 G>T/47 G>A/ 51 G>T /58T>C [17, 22, 23]. It postulate that the 3’UTR nucleotide substitution is a mutation that alters transcript stability and gene expression [17]. There is increasing evidence that 3’UTRs contain regulatory elements, which have an important role in posttranslational control of gene expression [24], could bind to either proteins or small noncoding regulatory RNAs [25]. Further, ESEFinder analysis showed that rs192856872 variant creates two new sites, and this change in SRSF1 binding site is predicted to affect normal splicing. Therefore, rs192856872 variant with uncertain significance need more research in the future.

More recent evidence from clinical, pathological, and genetic studies has emphasized the multisystem nature of amyotrophic lateral sclerosis (ALS) and FTD with overlapping symptoms and causes [26]. Approximatively, 10–15% of FTD patients display features of motor neuron disease, while around 50% of ALS cases show cognitive and behavioral impairment of which 10–15% reach diagnostic criteria for FTD [27]. Pathologically, the TDP-43 protein was found to be an important constituent of aggregates in neurons of MND– FTLD patients [28]. In genetic study, SIGMAR1 variants were reported to segregate in FTLD– MND pedigrees [17], leading to significant advances in understanding of the etiology and molecular mechanisms underlying ALS and FTD. SIGMAR1 is highly expressed in motor neurons [29],and dysregulated in tissues from ALS patients [30]. Impairment of SIGMAR1 leads to axonal and motor neuron degeneration both in vitro and in vivo [31], and the Sigmar1 knock-out mouse presents motor disabilities [32]. ALS-FTD was linked to chromosome 9p13.2-21.3 [33], while SIGMAR1 located on chromosome 9p13.3 and C9orf72 located on chromosome 9p21. It is unknown whether changes in SIGMAR1 gene could lead to genetic or epigenetic changes in its downstream genes including C9orf72. In the present study, we found mutations in SIGMAR1 could present in FTD patients independent of C9orf72 gene status.

Previously, the majority of studies have demonstrated cross-sectional associations between genetic findings and cognitive impairment. However, few have reported genetic and neuroimaging relationships. As well known, MRI studies demonstrate that frontal atrophy worse in the bvFTD cases while the temporal worse in semantic dementia, and this feature can distinguish between FTD and other dementing conditions [34]. MRI scans were abnormal in the majority of FTD patients (75%), with focal atrophy present in 100% of semantic dementia patients [35]. Lateralization of atrophy was a feature of semantic dementia as with the majority showing left-sided atrophy [35]. Previous cohort studies of MRI and FTD have focused primarily on findings in cross-sectional associations, the strengths of our study had 6 years follow-up periods. In rs192856872 carrier, the degeneration extends mainly caudally in the left posterior temporal lobes, then progress to the right side. A hallmark of the semantic dementia is the presence of semantic errors (e.g., ‘banana’ for apple; ‘dog’ for sheep), which is consistent with the gradual erosion of the capacity for discrimination between related concepts [36]. In the traditional neuropsychological tests, the Boston Naming Test and Animal Fluency Test scores reflect semantic memory, which were significantly decreased in rs192856872 carrier. Therefore, there is growing evidence that left anterior temporal lobe conceptual hub binds information from modality-specific systems and stores it in a modal format [37].

In summary, we identified a heterozygous variant of SIGMAR1 in a sporadic FTD patient without any mutation of MAPT and GRN genes, and without pathogenic repeat expansions in C9orf72. ESEFinder analysis showed the variant creates two new SRSF1 binding sites, which is predicted to affect normal splicing. The follow-up MR imaging showed asymmetrical atrophy in the anterior temporal lobes. Generally, our study provides genetic and neuroimaging support for the validity that SIGMAR1 variant plays an important role in FTD.