Abstract
Background:
Mutations in the GRN gene are causative for an autosomal dominant form of frontotemporal dementia.
Objective/Methods:
The objective of the present study is to describe clinical and molecular features of three siblings harboring the GRN deletion NM_002087.3:c.295_308delTGCCCACGGGGCTT, p.(Cys99Profs*15) identified with next generation sequencing.
Results:
Our patients demonstrated heterogeneous clinical phenotypes, such as progressive supranuclear palsy-like in the proband and the behavioral variant of frontotemporal dementia in the two affected siblings. Progranulin haploinsufficiency was revealed by both gene expression and protein analyses.
Conclusion:
The pathogenicity of the novel GRN deletion c.295_308del TGCCCACGGGGCTT is confirmed by both functional analysis and segregation in three affected siblings.
INTRODUCTION
Single allele mutations in the gene encoding progranulin (GRN) account for approximately 26% of familial frontotemporal dementia (FTD) cases [1]. More than 200 pathogenic GRN variants have been described so far (http://www.hgmd.cf.ac.uk/ac/all.php). The pattern of inheritance is autosomal dominant with onset in late/middle age (up to 90% at age of 70 years) [2]. Less frequently, GRN variants have been associated with both progressive supranuclear palsy (PSP)- and corticobasal syndrome (CBS)-like phenotypes as well as with signs of involvement of first and second motor neuron [3]. Subjects with GRN variants show in 40–60% of cases a mild, akinetic-rigid extrapyramidal syndrome, usually poor responsive to levodopa [2]. In line with the high intra-familial variability of clinical features, studies looking at genotype-phenotype correlations failed to find remarkable results [4].
Irrespective of the specific mutation, GRN heterozygous variants that result in a premature stop codon trigger non sense-mediated under-expression of mRNA and subsequent loss of 50% progranulin levels leading to haploinsufficiency [5].
As for neuropathology, GRN is associated with frontotemporal lobar degeneration with ubiquitinated fragments of TDP-43 (FTLD-TDP) [5]. Animal models suggest that neurodegeneration associated with progranulin deficiency may reflect abnormal lysosomal function and exaggerated microglial inflammation [5].
Herein, we describe the clinical and molecular characterization of a family carrying a GRN 14 bp deletion, NM_002087.3:c.295_308delTGCCCACGGGGCTT, p.(Cys99Profs*15).
MATERIAL AND METHODS
Patients
The proband of the family described in the present study (H311, Fig. 1A) has been identified within the ANAMNESIS (An origiNal Approach to study faMiliarity in NEurodegenerative SYndromeS) study conducted at the Center for Neurodegenerative Diseases (CEMAND), University of Salerno, Italy. In brief, eighty-five patients affected by extrapyramidal syndrome and cognitive or behavioral disturbances with a positive familial history for either a parkinsonian or dementia syndrome underwent a genetic panel analysis with next generation sequencing (NGS).
Pedigree of the family. A) I-1: Dead at 82 years old. Long-lasting drug-resistant depression starting in her 30 s, followed by tremor and dementia; I-3: Dementia; I-4: Motor slowness. II-1 (H312) affected by frontotemporal dementia; II-2 (H311): the proband presenting by progressive supranuclear palsy-like phenotype; II-3: Dead at 8 months for infection; II-4 (H313) affected by frontotemporal dementia. B) Proband’s brain MRI showing mild cortical atrophy and lack of midbrain atrophy. C-E) Sanger sequencing electropherograms. Sanger sequencing electropherograms of the exon 4 GRN gene-mutated region in subjects H311, H312 and H313 showing deletion (C) and SNP rs9897526 located in intron 3 (D). Patients are heterozygous for both mutations. Analyses were done by Sanger-based method sequencing. E) Exon 4 of the GRN gene was investigated (NC_000017.11). Primers forward and reverse were located in the exon2-intron2 boundary and in exon 4 respectively. Primers were designed using Primer3 tool and PCR amplifications were performed on SureCycler 8800 Thermal Cycler (Agilent). The arrows show the position of the deletion of 14 bp NM_002087C.295_308DELTGCCCACGGGGCTT (Homo sapiens chromosome 17 : 44349299-44349810, GRCh38.p12 Primary Assembly) and rs9897526 SNP (chr17 : 44349572).
All patients provided written informed consent, the project was approved by the local Ethics Committee and conducted in compliance with guidelines on human experimentation.
Genetic analysis
The patient underwent an NGS-based screening of a panel of 32 genes related to neurological syndromes (see Supplementary Material for detailed methods and list of genes tested). To confirm results of the NGS panel in the proband and verify the presence of the same mutations in her affected siblings (H312, H313, Fig. 1A), Sanger sequencing analysis was performed.
RNA isolation, retrotranscription, and quantitative real-time PCR (qPCR)
Functional studies were performed to confirm pathogenicity of GRN deletion. Total RNA was isolated from leukocytes samples using TRI Reagent® (Zymo Research, Irvine, CA, USA). The concentration and purity of the RNA samples were assessed using NanoDrop® 1000 (Thermo Fisher, Milan, Italy), while integrity was checked by agarose gel electrophoresis. One μg of RNA was retrotranscribed into cDNA using SuperScript™ VILO™ cDNA Synthesis Kit (Thermo Fisher, Vilnius, Lithuania) in a final volume of 20μL according to manufacturer’s instructions.
qPCR reactions were carried out as reported elsewhere [6]. Relative gene expression levels were normalized to the reference gene GAPDH (forward primer: AAAATCAAGTGGGGCGATGC; reverse primer: GGCAGAGATGATGACCCTTT), and calculated by the 2– ΔΔ Ct method. Two combinations of primers were used to detect GRN transcript level variations. Specifically, for both GRN wild type and p.(Cys99Profs*15) mutation, the same reverse primer was used, while, to distinguish the two GRN transcripts, two different forward primers were designed on the deletion regions.
Protein analysis by western blotting
Progranulin analysis was performed directly on plasma samples of the proband and both her siblings. Protein concentration was determined by the Qubit® (Invitrogen) fluorometer, 5μg of plasma proteins were subjected to an 8% SDS-PAGE. Two SDS-PAGEs were performed simultaneously, one was analyzed to blue Coomassie staining, the other one was electroblotted on PVDF membrane, which was blocked at room temperature for 1 h in blocking solution (5% milk in TBS-T), incubated overnight at 4°C with anti-GRN antibody (1 : 1000; Thermo Scientific). Membrane was incubated with peroxidase-conjugated anti-rabbit antibody (1 : 10,000; Amersham Biosciences). Proteins were revealed by an ECL kit (Amersham). The intensity of the bands was quantified by scanning densitometry using Scion Image version 4.5 software. Normalization was performed against the signal of Albumin (66.5 KDa), the major component of the total plasma proteins, obtained by blue Coomassie staining of the gel.
Statistical analysis
Both plasmatic progranulin and GRN mRNA levels from the three siblings were compared with an age- and sex-matched group of three healthy controls from the same geographical location (regione Campania, South of Italy) with the two-tailed t-test (Microsoft Excel software). Results were considered significant for p < 0.05.
RESULTS
Clinical characterization
The proband is a 63-year-old right-handed woman evaluated for progressive slowness of movements and camptocormia over the previous 4 years (H311, Fig. 1A). Difficulties with problem-solving and concentration manifested 2 years after motor onset. Family history was relevant for both extrapyramidal and dementia syndrome with an autosomal dominant inheritance pattern affecting other five relatives (Fig. 1A). On mental status examination, she exhibited a Mini-Mental State Examination (MMSE) score of 20 with impaired performances in memory, executive functions, attention tasks, and language and significant impact on activities of daily living (see Supplementary Material for detailed cognitive examination). On neurological examination, she presented a mild, rigid-akinetic, symmetric parkinsonism with good response to levodopa [Unified Parkinson’s disease rating scale part III (UPDRSIII) off levodopa: 35; UPDRSIII on levodopa 300 mg/die: 24], myoclonus, postural instability with tendency to fall and vertical supranuclear gaze limitation with slowing of saccades, suggesting a PSP-like phenotype. After 3 years since onset, overt motor fluctuations appeared with wearing off and pick levodopa-induced dyskinesias prevalent to lower limbs (Video 1). Brain MRI disclosed unremarkable findings but cortical atrophy (Fig. 1B).
The eldest sibling (H312, Fig. 1A) presented cognitive decline and behavioral disturbances (irritability and change in personality) since the age of 60. After 5 years since onset, he showed extrapyramidal syndrome and overt dementia (MMSE not feasible) (Video 2). Brain MRI only disclosed cortical atrophy.
The youngest sibling (H313, Fig. 1A) showed behavioral disturbances (irritability, apathy, obsessive thinking) since the age of 57. After 2 years since onset, he presented a symmetric, rigid-akinetic extrapyramidal syndrome, myoclonus and dementia (Video 3). Brain MRI was not available. Both siblings carried a diagnosis of behavioral variant FTD (bvFTD) and were treated with levodopa (300–400 mg/die) but reported no motor improvement nor levodopa-induced dyskinesia.
Genetic analysis
NGS identified one proband (H311, Fig. 1A) with a novel GRN 14bp deletion, NM_002087.3:c.295_308delTGCCCACGGGGCTT, p.(Cys99Profs*15). Varsome classified such novel variant as likely pathogenic, in accordance with the American College of Medical Genetics and Genomics (ACMG) classification criteria. NGS disclosed also the pathogenic SQSTM1 missense variant p.(Ale33Val) in the same proband [7]. No other variants were found in the other genes screened in the panel (see Supplementary Material).
Sanger analysis confirmed the presence of such novel GRN deletion in the proband and both her siblings. Furthermore, the Sanger sequencing in the region identified the SNP c.264+21G>C (rs9897526) associated with the GRN deletion in all three siblings (Fig. 1C–E). Finally, the Sanger sequencing confirmed the presence of the SQSTM1 missense variant NM_003900.5:c.98C>T;p.(Ala33Val) in the proband (H311) and her eldest sibling (H312). No other genes were tested in the proband’s siblings.
Functional analysis
Western blotting demonstrated lower progranulin plasmatic levels in the proband and both her siblings compared to age-matched healthy controls (p < 0.05) (Fig. 2A, B). Moreover, GRN gene expression was also analyzed by quantitative real-time PCR (qPCR) of total RNA extracted from white blood cells with a primer set corresponding to sequences within ex4 and ex5 (Fig. 1C, 2C). To discriminate between wild type and deletion sequence (c.295_308del TGCCCACGGGGCTT mutation), we designed a forward primer encompassing the deletion mutation and able to perfectly match only with the wild type sequence. The qPCR assay showed a range of cycle threshold (Ct) from 19.31 to 19.69 for the healthy controls, while a Ct range from 21.09 to 21.53 for the three siblings, supporting a lower level of transcript in the subjects harboring the GRN deletion as already evidenced by the fold change calculated using one of the controls as calibrator (Fig. 5). To further corroborate such results, we analyzed mRNA expression using a different forward primer GRN Ex4M –Ex5, able to perfectly match only with the mutated sequence of the three siblings within the ex4 c.295_308del TGCCCACGGGGCTT mutation. In this case, the range of Ct was lower in the three siblings compared to the healthy controls (from 24.22 to 24.95 for the three siblings and from 29.94 to 31.60 for healthy controls), demonstrating that only for the three siblings the primer specific for the deletion mutation was able to efficiently work in the qPCR assay (data not shown).
A) Western blotting analysis of plasmatic progranulin in three healthy controls (lanes 1, 2, 3) and three subjects carrying the GRN deletion (lanes 4, 5, 6 corresponds to H311, H312, H313), with fetal bovine serum as a negative control. B) Densitometric analysis of the signal intensity from all 3 assays normalized against albumin (66.5 KDa), the major component of the total plasma protein signal obtained by Coomassie blue staining of the gels. Results are represented as means±SD (standard deviation) of the three experiments. *p < 0.05 versus healthy patients. C) Expression level of wild type GRN mRNA in WBC samples. The expression level was normalized to the reference transcript GAPDH and calculated with the 2– ΔΔ Ct method. The CTRL1 sample was used as a calibrator. The results from two technical replicates are expressed as the mean of fold change±SD (standard deviation). Difference between the three subjects harboring the GRN deletion and healthy controls is significant at p > 0.01 (**). CTRL1, CTRL2, and CTRL3 are samples from healthy controls. Ct (Cycle threshold).
DISCUSSION
Herein, we describe for the first time the clinical and molecular characterization of three siblings harboring the 14 bp GRN deletion NM_002087.3, c.295_308delTGCCCACGGGGCTT, p.(Cys99Profs *15) in exon 4. Such deletion has been recently found in a single patient in a large, international genome-wide association study, suggesting this is a rare GRN mutation [8].
As for the clinical phenotype, the proband presented an uncommon onset of disease characterized by parkinsonism followed by cognitive impairment. Indeed, both vertical supranuclear gaze limitation and early-onset postural instability pointed towards a PSP-like phenotype (Video 1). However, the positive familial history with autosomal dominant inheritance pattern as well as the good response to levodopa with development of overt motor fluctuations and dyskinesia, the stooped posture, and the absence of a severe progression of disease, all represented exclusion criteria for a diagnosis of idiopathic PSP according to recent criteria [9]. In addition, brain MRI did not show midbrain atrophy (Fig. 1B). Conversely, both proband’s siblings presented a classical bvFTD with the only red flag for an underlying genetic cause being the positive familial history [10].
As for the functional impact, such GRN deletion was already classified as likely pathogenic in accordance with ACMG classification criteria and, thus, highly suggestive for a whole- or partial-gene deletion. In keeping with this hypothesis, segregation in other affected siblings as well as functional analysis with both quantitative real-time PCR of the gene and protein analyses further confirmed its pathogenicity, suggesting such deletion generates a frameshift defect in GRN gene sequence leading to haploinsufficiency [11]. In detail, gene expression analysis by qPCR assay, using primers designed to amplify both the GRN ex4 wild type sequence and GRN Ex 4 mutated in the second analysis, showed that the quantity of mRNA molecules containing wild type GRN Ex4 is reduced in the three siblings compared to healthy controls (Fig. 2C). Such results parallel the reduced protein levels observed in the western blotting analysis in the same patients (Fig. 2A, B) and clearly point to progranulin haploinsufficiency associated to loss-of-function heterozygous deletion in the GRN gene.
On a clinical ground, our report has several practical consequences. First, patients with PSP-like phenotype and positive familial history for dementia should be screened for GRN. We also highlight the difficulty in classifying genetic patients according to clinical diagnostic criteria. As a matter of fact, our proband was labeled as a PSP-like phenotype due to several atypical findings excluding idiopathic PSP according to current clinical criteria [12]. Finally, we report for the first time the presence of a good response to levodopa as well as the development of overt motor fluctuations and levodopa-induced dyskinesia in a subject harboring a pathogenic GRN deletion and SQSTM1 missense variation p.(Ala33Val). No final conclusions can be drawn on the role of SQSTM1 missense variant p.(Ala22Val) in determining the clinical phenotype. As a matter of fact, both the proband and her eldest sibling shared such variant, irrespective of the different clinical diagnosis and response to levodopa.
On a functional ground, we describe the molecular effect of the GRN 14 bp deletion c.295_308delTGCCCACGGGGCTT, p.(Cys99Profs*15) segregating in all affected members of the pedigree and associated with under-expression of plasmatic progranulin levels. Our data would further support the usefulness of progranulin plasma dosage assessments (e.g., ELISA) for pre-screening of mutation carriers and for validation of GRN mutations.
The disease spectrum associated with GRN variants is wide with high inter- and intra-familial variability. bvFTD, agrammatic variant of primary progressive aphasia, corticobasal syndrome, and Alzheimer’s disease (AD)-like phenotype are the most common presentations [13]. Our case is of special interest as PSP-like presentation is very rare. Furthermore, levodopa response with development of overt dyskinesias are unique features. As such, although the proband did not present any other non-motor symptom suggestive of alpha-synucleinopathy co-pathology, we cannot exclude the presence of concomitant Parkinson’s disease. Another unique clinical feature of our report is the presence myoclonus in all three siblings. Although NGS panel excluded mutations in other genes associated with AD pathology, we failed to perform cerebrospinal fluid analysis or nuclear imaging to verify the presence of markers of AD pathology. Thus, we cannot exclude myoclonus is related to AD pathology in our cases.
We also failed to perform a formal electrophysiologic examination in our proband. Although, none of the patients presented overt motor neuron loss signs, we cannot exclude a subclinic motor neuron involvement.
In conclusion, although we acknowledge the lack of cellular functional model, our results confirm the pathogenicity of the GRN deletion of 14 bp in exon 4 (c.295_308del TGCCCACGGGGCTT) and its association to the development of both PSP-like and FTD phenotypes in different siblings from the same pedigree. The significance of levodopa-induced dyskinesia in our proband as well as the role of concomitant SQSTM1 missense variant p.(Ala33Val) in determining the clinical phenotype need to be explored in further studies.
