Structural Variants in SNCA Gene and the Implication to Synucleinopathies

Abstract

Synucleinopathies are a group of neurodegenerative diseases that share a common pathological lesion of intracellular protein inclusions largely composed of aggregates of alpha-synuclein protein. Accumulating evidence, including genome-wide association studies, has implicated the alpha-synuclein (SNCA) gene in the etiology of synucleinopathies and it has been suggested that SNCA expression levels are critical for the development of these diseases. This review focuses on genetic variants from the class of structural variants (SVs), including multiplication of large genomic segments and short (<50bp) genomic variants such as simple sequence repeats (SSRs), within the SNCA locus. We provide evidence that SNCA-SVs play a key role in the pathogenesis of synucleinopathies via their effects on gene expression and on regulatory mechanisms including transcription and splicing.

Introduction

Structural Variants (SVs) are genomic variants other than single nucleotide variant (SNVs), are often multiallelic and include deletions, insertions, microsatellites or simple sequence repeats/short tandem repeats (SSRs/STRs), insertion/deletions (indels), copy number variation (CNV), block substitutions, and inversions¹. Recently, there has been increased support for the idea that SVs affect many human phenotypes and complex traits^2–6. We are interested in the functional consequences of SVs and their causative role in neurodegenerative diseases of aging.

Our research strategy expands on the concept that changes, even subtle, in expression levels of normal (wild-type) proteins in the brain can lead to neurodegenerative diseases of the aging brain. The genetic contribution of SNCA (SNCA, MIM #163890) to synucleinopathies is a prominent example. Synucleinopathies are a group of neurodegenerative disorders that share a common pathological lesion composed of protein inclusions in the cytoplasm of select populations of neurons and glia (i.e, oligodendrocytes), known as Lewy bodies (LBs) and Lewy neurites, and glial cytoplasmic inclusions (GCIs), respectively^7–11. Aggregates of the insoluble α-syn protein are the major component of LBs¹², Lewy neurites, and GCIs. This group of disorders include Parkinson’s Disease (PD), dementia with Lewy bodies (DLB), Lewy body variants of Alzheimer disease (LBV/AD), Neurodegeneration with Brain Iron Accumulation (NBIA) type I, pure autonomic failure (PAF), and multiple system atrophy (MSA). Over the last decade, genome wide association studies (GWAS) and candidate gene-based approaches¹³ have implicated SNCA as a highly significant genetic risk factor for synucleinopathies including sporadic PD^14–27, DLB²⁸, MSA^29,30 and LBV/AD^31,32. In addition, accumulating evidence from in vitro systems and in vivo models, suggesting that the α-syn expression levels are critical for the development of synucleinopathies (reviewed in³³). In this review, we will describe the contribution of both long and short SVs in the SNCA locus to the regulation of SNCA gene expression profiles, mRNA levels and splicing, in relation to their possible role in the etiology of synucleinopathies.

Long Structural Variants (SV)- Copy number variations of the SNCA locus

Long SV are defined as DNA variants ≥50 bp. A recent study based on the analysis of an integrated SV map of 2504 human genomes showed that SVs are enriched in haplotypes identified by genome wide association studies (GWAS) and exhibit enrichment for expression quantitative trait loci (eQTLs)⁵.

A sub-class of the long SV are the copy number variations (CNV), segmental multiplication of genomic sequences larger than 1kbp. Two CNVs in the SNCA gene, triplacation and duplication, have been identified in only a few families with an early onset, autosomal dominant form of PD^34–38. Genomic triplication of the SNCA – containing region results in four fully functional copies of SNCA, and 2-fold over-expression of SNCA mRNA and protein. The triplication leads to high penetrance of an early-onset PD phenotype with cognitive impairment and autonomic dysfunction^34,39,40. Duplications of the wild-type SNCA gene result in a 1.5-fold elevation of SNCA expression and, compared with the triplication, a slightly later onset of heritable PD that is characterized by a lower penetrance and a ‘milder’ phenotype with slower progression^35–38, demonstrating the dose-dependent effect of SNCA on disease etiology.

The molecular mechanism that mediates the pathogenic effect of SNCA-CNV has been also investigated in vitro using isogenic induced Pluripotent Stem Cells (iPSC)-derived models. Differentiating of iPSCs from a patient with SNCA triplication into dopaminergic neurons confirmed the vulnerability of the dopaminergic neurons to overexpression of SNCA^41,42. The iPSC-derived dopaminergic neurons carrying the SNCA triplication showed an approximately two-fold increase in α-syn levels compared to the control cells and the accumulation of α-syn protein in this iPSC-derived model coincided with increased oxidative stress markers, and conferred increased vulnerability to oxidative stress-induced cell death⁴³. Another study that investigated the SNCA triplication iPSCs, reported that iPSCs-derived Neural Precursors Cells (NPCs) displayed overall normal cellular and mitochondrial morphology, but showed substantial changes in growth, viability, cellular energy metabolism and stress resistance when challenged by starvation or toxicant challenge. These phenotypic changes were reversible upon knockdown of SNCA overexpression⁴⁴. A recent study demonstrated that dopaminergic neurons derived from SNCA triplication iPSCs recapitulate several PD-related phenotypes and exhibit cellular and molecular characteristic of neurodegeneration⁴⁵. Collectively, these studies introduced the iPSCs-derived dopaminergic neurons as a model to investigate the effects of SNCA-CNV mutations on PD related cellular phenotypes, and demonstrated that the phenotypic effect of the SNCA-CNV mutation is mediated by upregulation of SNCA expression levels. Patients carrying SNCA-CNV mutation- triplications and duplications also manifest clinical and pathological features similar to DLB. While the above studies focused on iPSC-derived dopaminergic neurons to model PD, it is also important to gain insight into the effect of SNCA-CNVs in cortical neurons to model DLB. In unpublished work, we established a model system of iPSCs-derived neurons from a normal subject and a patient with SNCA-triplication to evaluate the effect of SNCA-triplication and up regulation of SNCA-mRNA. Those iPSC lines were differentiated into dopaminergic and cholinergic neurons to model PD and DLB, respectively. SNCA-mRNA (Fig. 1A) and protein (Fig. 1B) expression in the SNCA-triplication cells exhibited two-fold increase compared to the cells from a normal control for each neuronal type across differentiation, as shown in Fig. 1 an example for iPSCs and the derived precursor and matured cholinergic neurons. These results recapitulate the in vivo observations in human tissues. Notable, SNCA-mRNA expression in both the SNCA triplication and the control iPSCs lines increased along the maturation process of the cholinergic neurons (Fig. 1A). These observations warrant a deeper characterization of the differentiated neuronal lines derived from each of the isogenic-iPSCs in the context of the implication to susceptibility to PD compared to DLB.

Figure 1. SNCA-mRNA and α-syn protein in iPSC-derived neurons.

(A) Levels of SNCA-mRNA were measured by real-time RT-PCR using TaqMan expression assay and SNCA-mRNA fold levels were calculated relative to the geometric mean of GAPDH-mRNA and PPIA-mRNA reference controls using the 2^−ΔΔCT method. (B) Immunofluorescence (IF) labeling for α-syn protein (green) was performed using rabbit recombinant monoclonal anti-alpha synuclein primary antibody (ab138501 ABCAM) and AlexaFluor 488 goat anti-Rabbit secondary antibody (A11034 ThermoFisher). Signal intensities were analyzed using ImageJ. Example are shown for the iPSC-derived basal forebrain cholinergic neurons (BFCNs) differentiation.

Short Structural Variants (SSV) within noncoding regions of SNCA locus

Short SVs (SSVs) are short genomic variants (<50bp) other than SNVs, and generally present multi alleles. SSVs are thought to affect phenotype by altering the regulation of gene transcription^46–51, splicing⁵², and translation, and it is by these mechanisms that SSVs may play a role in the etiology of human diseases, including complex disorders. A new study identified >2000 expression STRs (eSTRs) in the human genome and found that eSTRs contribute to ~10–15% of the cis heritable variation in gene expression attributed to common variants⁵³. This study also showed that these eSTRs are enriched in genomic regions associated with clinically relevant phenotypes⁵³.

Furthermore, recent studies proposed a potential mechanism whereby SSRs affect transcription (illustrated in Fig 2). These studies showed that certain repetitive DNA sequences, when present in the flanking regions of specific transcription factor (TF) binding sites, can have a magnitude effect on the intensity of TF-DNA binding, through a mechanism we termed “non-consensus binding”^54,55.

Figure 2. Schematic illustration of a putative mechanism whereby SSRs affect differential transcription.

The overall hypothesis is that differential non-consensus binding of transcription factors (TFs) to the variable alleles at repeat sequence sites is responsible, at least in part, for differential expression of the cis-regulated genes, which can contribute to disease pathogenesis. For alleles of different lengths, some SSRs will have differential non-consensus binding free-energy, and thus differential TF binding strength.

Using our newly developed SSV Evaluation System⁵⁶, we screened for SSVs within a ~215 Kb segment of the SNCA genomic region, +/−50kb (hg19, chr4: 90,595,250–90,809,447), with the goal of identifying candidate causal variants for synucleinopathies. We identified 555 SSVs, of which 75 are from the SSR category. Below we describe a few SSVs that have been studied experimentally in relation to their association with disease phenotypes and mode of function.

(1) SNCA-CT rich haplotype

Combined in silico and wet lab approaches led to the identification of an intronic polymorphic CT-rich region in SNCA that is a candidate regulatory SSV. The SSV evaluation system provided high scores (representative Total Impact Score=42) in a CT-rich, low-complexity region of intron 4 of SNCA. We cloned and sequenced this region in case (LBV/AD) and control (AD) subjects and identified four distinct haplotypes within this region, with specific haplotype-conferred risk to develop LB pathology in AD patients³¹. We further demonstrated that the risk haplotype was significantly associated with elevated levels of SNCA-mRNA in human brain tissues relevant to the LB pathology and suggested that the CT-rich site acts as an enhancer element of SNCA transcription³¹.

(2) SNCA-Rep1

Rep10⁴⁷ is a polymorphic dinucleotide complex repeat site located ~10 kb upstream of the SNCA transcription start site^57,58. The length of Rep1 appears to be associated with increased risk of PD⁵⁹. The overwhelming majority of the reported association studies, including a large meta-analysis, have shown that the extended alleles of SNCA-Rep1 confer increased risk to develop late-onset, ‘idiopathic’ PD, while the shorter allele is protective^24,60–63. We investigated the effect of the Rep1 polymorphism on SNCA expression, and discovered that Rep1 regulates SNCA transcription in human brain tissues⁶⁴. These results have been confirmed using luciferase reporter assay^47,65 and in a humanized mouse model⁶⁶. The PD-risk Rep1 allele led to increased SNCA-mRNA levels, providing further support to the pathogenic effect of SNCA overexpression^47,64,66. In support of this, we identified a factor, poly(ADP-ribose) transferase/polymerase-1 (PARP-1), that binds specifically to Rep1 and modulates SNCA transcription⁶⁷, suggesting that the association of Rep1 alleles with sporadic PD may be mediated, in part, by the effect of PARP-1 on SNCA expression. Retrospectively, the Rep1 site was assessed using the SSV evaluation system and had a high score (Total Impact Score=31) relative to the complete list of SSRs within SNCA genomic regions including +/−50kb flanking regions in SSV evaluation system.

However, these results conflict with the recent findings that showed no effect of Rep1 on SNCA-mRNA level using genome edited iPSC-derived neurons⁶⁸. It is possible that difficulties in quantifying the total SNCA transcripts levels affected the validity of the reported conclusions. In fact, we recognized that the allele-specific assay used to measure SNCA-mRNA was designed to target only the long 3′UTR isoform of SNCA transcript, and therefore the method Soldner et al. used to quantify SNCA-mRNA levels did not capture all SNCA transcript species. It is crucial to note that the long 3′UTR isoform of SNCA is not as abundant as the short 3′UTR isoform and represents only a small fraction of SNCA transcripts. Therefore, we cannot rule out the possibility of a false negative for the SNCA-Rep1 finding reported by Soldner et al⁶⁸.

(3) Intron 2 poly-T

The SNCA126 splicing variant may have a protective role because the in-frame deletion of exon 3 leads to the interruption of the N-terminal protein-membrane interaction domain which may lead to less aggregation^69,70. SNCA126 levels were decreased in the prefrontal cortex of DLB patients⁷¹. In contrast, SNCA126 expression was increased in the frontal cortex of PD brains and no significant differences in MSA⁷².

A polyT variant in intron 2 of SNCA gene comprises three alleles (5T, 7T, an12T) and the length of the polyT stretch is directly associated with SNCA126 expression levels in the normal brain, influencing the splicing efficiency of SNCA exon 3. Whereas the shortest 5T allele was associated with lower expression of SNCA126-mRNA, the longest 12T led to the highest SNCA126-mRNA levels⁷³. The same study also reported that the 12T-allele-carrying genotypes accumulated with increasing age in the normal population, while the frequency of 5T-allele-carrying genotypes decreased in successive age groups until reaching zero in the oldest group. Collectively these observations imply that the longest poly-T allele has a protective effect in aging, presumably via its association with higher SNCA126-mRNA levels.

Summary

We have reviewed evidence of the causative role of SNCA SVs in the etiology of synucleinopathies in general, and PD in particular, and described the molecular mechanisms that underlie their pathogenic impact (summarizes in Fig 3). We proposed that a 1.5- to 2-fold increase in expression caused by CNVs can lead to early onset familial disease, while even subtle changes in the SNCA expression are able to trigger the onset of non-Mendelian synucleinopathies. Subtle alterations in the SNCA expression are driven by different mechanisms and factors, and here we focused on noncoding cis-SSV and their corresponding regulatory functions. This implies that common and distinct SSVs and regulatory mechanisms through which they act might be involved in the etiology of the different synucleinopathies. Noteworthy, as summarized in Fig 3, selective vulnerability of each cell types involved in the different synucleinopathies to overexpression and/or particular SNCA-isoform will contribute, among other factors, to the development of a particular disease in the synucleinopathies spectrum disorders.

Figure 3. Schematic representation of the role of SNCA SVs in synucleinopathies.

The effect of SNCA-CNVs is mediated via increases in overall SNCA-mRNA and protein levels. The effects of SNCA-SSVs can be mediated via several RNA-based mechanisms. The cell-type selective vulnerability to overexpression and/or particular SNCA-isoform will contribute, among other factors, to the development of a particular disease in the spectrum of synucleinopathy disorders.

While it is widely accepted that up-regulation of the wild-type SNCA plays a causative role³³, it is still not completely understood if SNCA produces toxic isoforms and, if so, their precise identity. Current technologies and novel model systems, including iPSCs –derived neurons and CRISPR/Cas9 genome editing, will provide the opportunities to look at the functional and phenotypic consequences of each candidate SVs to fill in those gaps. In return, this important knowledge will result in the development of precision medicine including, biomarkers for pre-clinical diagnosis and effective treatment approaches for synucleinopathies in general and for a specific disease in the spectrum.

Highlights.

• Structural Variants (SVs) have a large impact on many human phenotypes including complex diseases such as synucleinopathies, and also contribute significantly to variation in gene expression in human as exemplified here for SNCA gene.

• Copy number variations (CNVs) of the SNCA gene cause familial Parkinson’s disease (PD), where disease severity and the levels of wild-type SNCA mRNA and protein are correlated with SNCA gene dose.

• A specific haplotype at a CT-rich region in SNCA-intron 4 that consists of a cluster of SSVs conferes risk to develop LBV/AD and acts as an enhancer element to increase SNCA-mRNA expression.

• Rep1, a SSV ~10kb upstream of SNCA gene that comprises of a cluster of SSRs, has been associated with PD-risk and has a role in the regulation of SNCA transcription

• A polyT variant in SNCA-intron 2 influences the splicing efficiency of exon 3, deletion of this exon is differently expressed in synucleinopathies and has been suggested to lead to less protein aggregation.