Mutation Analysis of the Genes Associated with Parkinson’s Disease in a Finnish Cohort of Early-Onset Dementia

INTRODUCTION

Clinical characteristics of neurodegenerative diseases, such as Alzheimer’s disease (AD), frontotemporal lobar degeneration (FTLD), Parkinson’s disease (PD), and dementia with Lewy bodies (DLB), often overlap. Each of these diseases have their own typical and core diagnostic symptoms and criteria; however, cognitive decline, neuropsychiatric symptoms, and parkinsonism may be present in all of them. In addition to similarities in the clinical characteristics, the neuropathological features behind these diseases may also overlap. All these neurodegenerative diseases present with certain pathological protein aggregations in the brain, however, mixed neuropathology is also common [1]. Lewy body accumulation is the main neuropathological hallmark in PD, Parkinson’s disease dementia (PDD), DLB, and DLB with AD, which are considered as subtypes of α-synuclein-associated disease spectrum [1, 2]. Nevertheless, AD pathology, including amyloid-β plaques and neurofibrillary tangles formed by hyperphosphorylated tau protein, are also common neuropathological findings in PD, PDD, and DLB [1, 2]. Moreover, Lewy bodies are present in over 40% of the AD patients [3].

The genetics behind these neurodegenerative dementias also overlap. Pathogenic PARK2 gene mutations that cause early-onset PD have previously been found also in sporadic early-onset AD cases. However, it is still unknown if these variants have the ability to affect the susceptibility to AD [4]. There are also a few reports of AD and FTLD patients carrying PINK1, PARK2, or LRRK2 gene variants that have been suggested to contribute to the pathogenesis of these diseases [5–8]. Furthermore, heterozygous GBA mutations have been found to be major genetic risk factors for PD, DLB, and PDD [9–11]. However, the role of the GBA variants in AD and FTLD is unknown.

Regardless of extensive research, there are still unknown genetic causes for early-onset dementias. Approximately10–15% of the early-onset AD (EOAD) patients, who have at least one affected first-degree relative, are estimated to have autosomal dominant inheritance of AD [12]. The most common mutations have been found in the genes coding for amyloid precursor protein (APP) and presenilin 1 and 2 (PSEN1/2) [12]. The C9orf72 hexanucleotide repeat expansion, microtubule-associated protein tau (MAPT), and progranulin (GRN) gene mutations explain the majority of the familial FTLD. In Finland, the prevalence of the C9orf72 repeat expansion is exceptionally high [13], while MAPT and GRN mutations are rare [14–16]. Nevertheless, there are still familial AD and FTLD patients with no genetic etiology identified so far, indicating that there are other still unknown causal genes to be identified. The genome-wide association studies and the next generation sequencing studies have showed that several genes are associated with multiple neurodegenerative disorders. Rare variants widespread in the genome could explain the missing genetic components for complex neurodegenerative diseases by modulating the clinical phenotype, for example the severity and earlier age at onset [17]. Based on the clinical, pathological, and genetic overlap between the neurodegenerative dementias, we aimed to evaluate here the role of the genes associated with PD in a well-defined cohort of early- onset dementia (EOD).

METHODS

The study population consisted of dementia patients who were diagnosed at two memory outpatient clinics in Finland (Oulu University and Kuopio University Hospitals). Patients were examined by experienced neurologists specialized in memory disorders and diagnoses were made using current diagnostic criteria for AD, FTLD, PDD, and DLB [18–22]. All the patients underwent a battery of examinations, including clinical and neurological examinations, routine screening laboratory tests, a neuropsychological examination, and magnetic resonance imaging (MRI) of the brain or rarely computed tomography of the brain (CT). When needed, cerebrospinal fluid analyses of the biomarkers Aβ₄₂, tau, and phospho-tau and/or functional neuroimaging by fluorodeoxyglucose positron emission tomography (FDG-PET) or 123iodine single photon emission computed tomography (FP-CIT SPECT) were performed on the patients to confirm the diagnosis.

The ethics committees of the Northern Ostrobothnia Hospital district and Northern Savo Hospital district approved the study. The study was performed in accordance with the Declaration of Helsinki. Written informed consent was obtained from all the patients or their caregivers.

The first cohort consisted of 37 EOD patients (21 men). The inclusion criteria were onset of the disease before 65 years and the patients had to have at least one relative with dementia in the family or the course of the disease was rapidly progressive or presenting with atypical features (e.g., myoclonia or spasticity). Altogether, 27 patients were diagnosed with AD, nine with FTLD, and one with progressive supranuclear paresis. Three of the FTLD patients also had motor neuron disease. The mean age at onset (AAO) was 55.3 years (range 41–65) and the mean age at diagnosis was 58.1 years (range 41–67). At least one first-degree or one second-degree relative with dementia or cognitive decline were identified in 43% (n = 16) and in 51% (n = 19) of the patients, respectively.

DNA was purified with QiAmp DNA Blood Mini Kit (Qiagen, Hilden, Germany) according to the original protocol. The APOE genotypes were determined by HhaI digestion [23]. The repeat-primed polymerase chain reaction assay (RP-PCR) was utilized to exclude the C9orf72 repeat expansion [24]. Targeted gene exon libraries were prepared using Nimble Gen Seq Cap EZ Library kit (Roche, Basel, Switzerland) with customized Seq Cap EZ Neurology Panel Design (Roche, Basel, Switzerland), covering all exons and exon-intron boundaries of 258 genes associated with 87 neurological diseases and disorders (Supplementary Tables 1 and 2). Furthermore, the panel included 22 additional genes. Samples were sequenced using Illumina MiSeq sequencer (Illumina, San Diego, CA, USA).

Pathogenic exonic and exon-intron boundary mutations in APP, PSEN1, PSEN2, MAPT, GRN, VCP, CHMP2B, FUS, TARDBP, TREM2, TMEM106B, UBQLN2, SOD1, UBQLN1, PRNP, and BIN1 as well as the C9orf72 repeat expansion were excluded from the primary cohort of 37 patients. Altogether eighteen patients of this cohort were carriers of the APOEɛ4 allele. Of these, 12 had the ɛ3/ɛ4, five the ɛ4/ɛ4, and one the ɛ2/ɛ4APOE genotype. Based on previous data about known mutations associated with PD, ten genes: SNCA, LRRK2, PINK1, PARK2, PARK7, SNCAIP, GBA, ATP13A2, UCHL1, and HTRA2 were selected for thorough investigation from the targeted panel of 258 genes and 22 additional genes. An in-house developed analysis pipeline was used for the evaluation of raw fastq files generated by the MiSeq sequencer. The data of the cohort of the 37 patients were manually screened to find rare (MAF < 0.01), exonic non-synonymous and other pathogenic mutations. Low-frequency GBA variants (MAF < 0.05) were also screened. The identified variations were confirmed by Sanger sequencing with ABI3500xL Genetic Analyzer (Thermo Fisher Scientific, Waltham, MA, USA).

The variations identified in the cohort of 37 patients were subsequently screened in a second cohort that consisted of 279 EOD patients. In this second cohort the mean age at onset was 57.2 years (range 36–65, 159 men). This cohort included patients with AD (n = 160), mixed dementia with AD and vascular dementia (n = 16), FTLD (n = 85), FTLD with motor neuron disease (n = 7), DLB (n = 5), PDD (n = 3), and progressive supranuclear paresis (n = 2), and multisystem atrophy (n = 1). In the second cohort, there were 50 patients carrying the ɛ4/ɛ4 (n = 50) APOE genotype, ɛ3/ɛ4 (n = 104), ɛ3/ɛ3 (n = 110), ɛ2/ɛ4 (n = 4), ɛ2/ɛ3 (n = 8), and ɛ2/ɛ2 (n = 3) APOE genotypes. The characteristics of both cohorts are described in Table 1. The variations in PARK2, PARK7, LRRK2, and SNCAIP were confirmed by Sanger sequencing. Hardy-Weinberg equilibrium and differences in the allele frequencies were calculated using Arlequin 3.5.1.2 Software [25]. Correction for multiple testing was calculated using Bonferroni correction [26].

RESULTS

No pathogenic mutations were found in the first cohort of 37 EOD patients. Altogether seven different variants in four different genes were identified. A total of 24.3% of the patients (n = 9) were carriers of at least one of the variants. The identified variants were LRRK2 p.Arg793Met, PARK2 p.Ala82Glu, SNCAIP p.Arg240Gln, SNCAIP p.Phe735Leu, GBA p.Asn409Ser, GBA p.Glu365Lys, and GBA p.Thr408Met (Table 2). At least one of these variants was found in 14.9% of the patients (n = 47) in the total cohort (n = 316). One male FTD patient (AAO 61 years) was homozygous for the GBA p.Glu365Lys variant, while all the other identified variants were heterozygous. Two patients carried two different variants (one patient had GBA p.Glu365Lys and PARK2 p.Ala82Glu variants and another patient LRRK2 p.Arg793Met and SNCAIP p.Arg240Gln variants). All in all, the allele frequencies of the variants were higher in the first selected EOD cohort compared to the frequencies of the total cohort. Except for the LRRK2 p.Arg793Met and GBA p.Thr408Met variants, the allele frequencies of all the other variants were also slightly higher in the total EOD cohort when compared to the frequencies reported in the gnomAD database for Finnish Europeans (gnomAD 03092019) (Table 2). In our EOD cohorts, the allele frequencies of the SNCAIP p.Arg240Gln variant in the total cohort were significantly higher than the reported allele frequencies for Finnish Europeans in gnomAD. However, after multiple testing correction (Bonferroni), the result did not remain statically significant. The phenotypes of the patients carrying these variants are shown in Table 3.

DISCUSSION

The aim of this study was to evaluate the role of different variants and mutations of the genes associated with PD in Finnish cohorts of EOD. No pathogenic mutations were found. However, several possible risk variants in the genes coding for LRRK2, PARK2, SNCAIP, and GBA were identified.

LRRK2 is a large gene in which over 1600 exonic variants have been found (gnom AD 03092019). So far, six autosomal dominant LRRK2 mutations (p.Arg1441Cys/Gly/His, p.Tyr1699Cys, p.Gly2019Ser, p.Ile2020Thr) have been identified in PD patients with incomplete penetrance [27]. In previous studies, the association between LRRK2 variants (p.Asn551Lys, p.Gly2385Arg, p.Arg1628Pro, p.Gly2019Ser, and p.Ile2020Thr) and AD has been investigated, but a recent meta-analysis found no connection between those variants and AD [28]. We identified the LRRK2 p.Arg793Met variant in three of our EOD patients. This variant has been previously been found in several PD patients and most of them have had late onset PD [29–32]. LRRK2 p.Arg793Met has been found also from a patient with non-fluent variant of primary progressive aphasia (nfvPPA) with the age at onset of 66 years [31]. Since this variant has also been detected in healthy controls, it is suspected to be a polymorphism. Nevertheless, the controls of that study were quite young (40–61 years) and therefore the possibility of later onset neurological diseases could not be excluded in their case [29, 34]. The LRRK2 p.Arg793Met variant might be a risk factor for dementia or later onset PD, but more studies are needed to further evaluate this.

PARK2 mutations have been reported to cause early-onset autosomal recessive PD. So far, the role of heterozygous PARK2 variations has been controversial. However, PD patients with heterozygous PARK2 variations have been shown to have more dementia than the non-carriers [35]. Compound heterozygous mutations have been reported in early-onset PD patients as well as in one bvFTD patient [36]. We found heterozygous PARK2 p.Ala82Glu variant in six EOD patients. This variant has been considered to be a rare polymorphism as it has been found both in PD patients and in disease-free control subjects [37–39]. Previously, a possible pathogenic role for this variant has been suggested in patients carrying a heterozygous PARK2 p.Ala82Glu variant in combination with PINK1 and DJ-1 sequence alterations [40]. In addition, PARK2 variations have been identified in a cohort of sporadic EOAD (n = 408), in which two confirmed heterozygous mutation carriers were found, one patient with heterozygous PARK2 p.Thr240Met and one patient with heterozygous PARK2 p.Gln34fs. [4]. However, these patients were also homozygous APOE4 allele carriers, which may also partly explain the early-onset of the disease [4]. In conclusion, the PARK2 variations may predispose to neurodegeneration, but larger studies are needed to confirm this.

SNCAIP encodes a protein called synphilin-1, which interacts with alpha-synuclein. It might have a role in Lewy body formation, but its association to PD is uncertain [41, 42]. We found two different variants in SNCAIP (p.Arg240Gln and p.Phe369Leu) from two EOAD patients. Previously SNCAIP p.Arg240Gln variant has been detected in late-onset sporadic PD patients [43]. SNCAIP has been a candidate gene for PD but the association studies have failed to show any connection among idiopathic and familial patients [42, 44–46].

GBA gene encodes a lysosomal protein called glucocerebrosidase, which hydrolyses glucosylceramide to ceramide and glucose. Some homozygous mutations in GBA lead to Gaucher disease. GBA mutations are known to be the single largest risk factor for the development of PD [10, 11]. We identified three different GBA variants (p.Glu365Lys, p.Thr408Met, p.Asn409Ser), which all have been associated with PD in previous studies [47–49]. However, the allele frequencies of these variants did not differ significantly from the gnomAD Finnish control population. The GBA p.Glu365Lys variant was found in 10.4% of our cases and, interestingly, one patient was homozygous for this variation. The homozygous GBA p.Glu365Lys carrier had a rapidly progressing FTD (AAO 61 years) with occasional visual hallucinations, mild hypokinesia, and hypomimia, but no rigidity or resting tremor. Later on, the patient developed symptomatic epilepsy. This variant has been found to increase the risk for PD, PDD, and DLB [9, 47]. It has also been associated with cognitive impairment in PD [50]. The GBAp.Thr408Met variant was identified from three patients of our cohort. Previously, carriers of this variant have been reported to have a slight elevation in the risk for developing PD and a faster progression of motor symptoms in PD [48, 50]. Furthermore, two patients of our EOD cohort carried the GBA p.Asn409Ser variant. This variant has been strongly associated with PD and increased risk for dementia, faster development of motor symptoms, dyskinesia, daytime sleepiness, and wearing off in PD patients [49, 52].

Parkinsonism was not emphasized in our patients carrying the identified variants in the PD genes. Over half of the now characterized patients with rare variants did not have any gait disturbances, resting tremor, rigidity, or hallucinations, and the patients who had these symptoms had only one or two of the symptoms. It is known, that at least 30% of AD patients have gait disturbances, tremor, rigidity, and hallucinations in the advanced stages of the disease and extrapyramidal symptoms and signs are detected in over 20% of FTD patients [53–55]. However, in most cases, we had the opportunity to follow the patients of our cohorts for only a few years after the diagnosis, indicating that the clinical characteristics described here represent the early stages of the diseases.

The strengths of our study are the well-characterized cohorts and the fact that the clinical diagnoses have been made by experienced neurologists specialized in memory disorders. However, due to rather small number of patients in our cohorts, our findings may be indicative and future studies with larger cohorts of well-characterized patients are needed to evaluate the impact of the PD gene mutations in EOD. Considering the rareness of EOD, international collaboration is needed to obtain larger cohorts. A limitation in the present study is that neuropathological confirmation of the diagnosis was lacking in the majority of the patients.