Systematic review of genetic association studies in people with Lewy body dementia

Lewy body dementia (LBD) causes more morbidity, disability, and earlier mortality than Alzheimer disease. Molecular mechanisms underlying neurodegeneration in LBD are poorly understood. We aimed to do a systematic review of all genetic association studies that investigated people with LBD for improving our understanding of LBD molecular genetics and for facilitating discovery of novel biomarkers and therapeutic targets for LBD.

prognosis than AD, with accelerated cognitive decline and a greater negative impact on quality of life. 5 The search for disease modifying drugs and reliable peripheral biomarkers for LBD is still ongoing. 2 Despite the public health importance of LBD, very little is known about the molecular pathology underlying neurodegeneration in LBD.
Systematic research on the genetics of LBD remains sparse. While most cases of LBD appear sporadic, several studies have reported familial aggregation of LBD and its core features such as visual hallucinations and cognitive fluctuations. 6,7 Siblings of probands with DLB have been reported to have significantly higher risk of developing DLB than the siblings of probands with AD. 7 Research in these families has supported a role for genes implicated in both AD (APP, PSEN1, PSEN2, PGRN, and PRNP) and Parkinson disease (PD) (SNCA, SNCB, LRRK2, and GBA) with the development of DLB and PDD. 8 However, autosomal dominant inheritance mutations in SNCA and LRRK2 in people with LBD have been reported. PDD has been associated with variants in PARK1, PARK4, GBA, MAPT, LRRK2, and APOE. 9 While the variants in APOE and GBA have been associated with both DLB and PDD, the associations are stronger for DLB over PDD. 10 Most of the candidate gene association studies that investigated the genetics of LBD were small, and their findings were poorly replicated. The first genome-wide association study (GWAS) investigating DLB was published in January 2018. 11 It estimated 36% heritability of DLB, and it confirmed the associations between DLB and variants in APOE, SNCA, and GBA. 11 Further imputation and genome-wide complex trait analysis of the GWAS data have updated the heritability of DLB as 59.9% and have indicated that the genetic risk factors for DLB are likely to be independent from known AD and PD risk variants. 12 Identification of genetic variants associated with LBD will improve our understanding of neurodegeneration in LBD and its molecular pathogenesis. Identifying a unique genetic profile will help in distinguishing LBD from AD and defining the nosological boundaries between DLB and PDD. This can facilitate discovery of reliable diagnostic biomarkers for LBD and of novel targets for future therapeutic approaches. In order to provide a comprehensive summary of all available evidence on the genetic associations of LBD, we aimed to conduct the first systematic review of all genetic association studies that investigated people with LBD.

| Study design
The protocol for this systematic review has been registered in the international prospective register of systematic reviews (PROSPERO protocol CRD42018087114; available at http://www.crd.york.ac.uk/ PROSPERO/display_record.php?ID=CRD42018087114). OR "variant*" OR "polymorphism*")) OR (Genome AND association*)

| Search strategy
OR "mutation*" OR "SNP" OR "CNV" OR "copy number variant*" OR "rare variant*" OR "microsatellite*" OR "chromosome*." The searches were limited to 3 February 2018 and to English. Reference lists of the studies included in the review were explored for identifying other potentially eligible studies.

| Eligibility criteria
We included all genetic association studies that satisfied the following inclusion criteria: (a) they were human studies. Studies on animals or cell lines were not included: (b) they presented original research data; (c) participants in at least one study group were clinically diagnosed to have DLB or PDD or LBD; (d) there was a control group in which LBD was clinically ruled out. The controls were either older people without cognitive impairment or those with other neurodegenerative disorders excluding LBD. We excluded studies that were not published in English.

| Study selection
We screened for all eligible candidate gene association studies and GWAS investigating the genetic associations of LBD. We merged our search results and removed duplicates. We excluded the abstracts that did not mention investigating the genetic association(s) between LBD and one or more genetic variants. We attempted retrieving full texts of all potentially eligible abstracts, and a three-member review team assessed the eligibility of the full-text papers. When a conference abstract was not accompanied by its full presentation, we requested further details from the corresponding author if the contact information was provided. If the corresponding author did not respond to our request within 14 days, we excluded that abstract.

Key Points
• Genetic associations between DLB and genetic variants in APOE, GBA, and SNCA have been replicated by at least two good quality studies.
• Genetic associations of PDD with variants in APOE and GBA have been replicated.
• Our meta-analyses confirm the associations of APOE-ε4 with DLB and PDD.

| Data extraction
We extracted the following data: (a) population characteristics including their mean age and ethnicity, (b) sample size in each study group, (c) definition of the phenotype, (d) investigated genetic variant(s), (e) genotyping method, (f) study findings with effect size and P values, (g) statistical correction for multiple testing, and (h) statistical analyses addressing the effects of potential confounders.

| Quality assessment
We assessed the quality of eligible studies using the quality of genetic association studies tool (Q-Genie). 13 The Q-Genie assesses the follow-

| Data reporting
When the studies included in this systematic review have reported the dbSNP identifiers (rs IDs) of their investigated genetic variants, we have extracted the information and have reported them in this review.
When the included studies have not reported the dbSNP identifiers, we searched the dbSNP database (https://www.ncbi.nlm.nih.gov/snp) with the reported names of the variants. When our search could not establish a unique dbSNP identifier, we have reported the variant name as it was reported by the original study authors. We report the results of included studies using the descriptors "positive" for statistically significant associations with P values less than .05 (after multiple testing correction, if available) and "negative" for the lack of statistically significant (P ≥ .05) associations.

| Data synthesis
A descriptive synthesis was carried out using the extracted data and major findings of each included study. We have synthesised the data by listing the genetic associations of investigated variants with a specific outcome variable (LBD/DLB/PDD). If three or more studies investigated the genetic association between a single genetic variant and a specific outcome variable, we conducted either fixed or random-effects meta-analyses using the STATA 15.1 software (StataCorp LLC, TX, USA) and its "metan" command. Later, we grouped these genetic associations by their potential functional links to the complex aetiopathogenesis of LBD. We have discussed the potential functional implications of the reported genetic associations within the context of available literature.

| RESULTS
We identified and screened 5125 papers after removing the duplicates and found 75 papers eligible to be included in this systematic review.

Figure 1 presents the study selection process in the Preferred Reporting
Items for Systematic Reviews and Meta-Analyses (PRISMA) format. 15 Our quality assessment using the Q-Genie rated 31 (41.3%) included studies as poor, 27 (36.0%) studies as moderate quality, and 17 (22.7%) studies as good (see Table S1). Statistically significant genetic associations have been reported between LBD and the genetic variants in in APOE and GBA have been replicated by two or more studies. There has been only one GWAS investigating DLB that has reported genomewide significant associations between DLB and rs429358 (APOE), rs7681440 (SNCA), rs35749011 (GBA), rs897984 (BCL7C/ STX1B), and rs1426210 (GABRB3) in its discovery stage. 11 There has not been any GWAS investigating people with PDD so far.

| APOE
The APOE variants, especially its ε4 allele, are the most studied among all genetic variants in people with LBD.  Apolipoprotein E (APOE) is involved in cholesterol mobilisation and redistribution during neuronal growth and injury, 16 and it may promote β-amyloid aggregation. 16 Among the 25 studies that investigated the association of APOE-ε4 with DLB, 21 have demonstrated a statistically significant association between APOE-ε4 and DLB ( Table 1). The 21 positive studies included six good quality, eight moderate quality, and seven poor quality studies, and the negative studies included one good, one moderate, and two poor quality studies. We conducted a meta-analysis including data from 17 positive studies and all four negative studies ( Figure 2).
We did not include four positive studies, because they did not provide allele frequency data or used the same participants as another study.
Moreover, statistically significant associations with probable reduced risk of DLB 22 have been reported between APOE-ε2 and DLB. [21][22][23]29 Similarly, one of the studies investigating the association between APOE-ε3 and DLB has reported statistically significant reduced risk for DLB. 29 Furthermore, two moderate quality studies 32,42 and one poor quality study including PD controls without dementia (PDND) 43 have reported statistically significant associations between APOE-ε4 and PDD. Another moderate quality study has reported significantly increased frequency of APOE-ε4 in a LBD group including people with DLB and PDD compared to PDND controls. 44 However, eight studies including two good, four moderate, and two poor quality studies did not find statistically significant association between APOE-ε4 and PDD. 19,30,[45][46][47][48][49][50] We conducted a meta-analysis including data from the four positive studies and eight negative studies ( Figure 3). There was significant heterogeneity (χ 2 = 29.22, df = 11; P = .002) among the studies, and the random-effects meta-analysis confirmed that the APOE-ε4 is significantly associated with increased risk of PDD (POR = 1.60; 95% CI, 1.21-2.11; P = .001).

| SNCA
α-synuclein encoding SNCA variants rs974711, 19 rs1348224, 19 and rs7681440 11 have been associated with DLB. However, studies that investigated SNCA triplication, 53 SNCA variant rs104893877, 54 and variants in α-synuclein interacting protein encoding SNCAIP 55 did not find statistically significant associations with DLB. Moreover, SNCA variants rs10018362 and rs7689942 19 were significantly increased in PDD, and SNCA variants rs1372525, rs2583988, rs2619364, rs2619363, and rs2301135 were not significantly associated with PDD. 56 Apart from the replicated association between DLB and rs7681440, 11 the reported genetic association findings between LBD and other SNCA variants have not been replicated so far. haplotype and DLB, but two good quality studies have not replicated this finding. 11,37 Moreover, a moderate quality study has reported associations between PDD and MAPT H1 haplotype and another probably protective variant rs1467967. 60 Tables 2 and 3 present other reported genetic associations of DLB and PDD, respectively. The GWAS has replicated the association between DLB and rs7314908 of CNTN1. 11 The reported genetic associations between DLB and rs897984 (BCL7C), 11    APOE-ε4 variant has the largest body of evidence in this topic.  Two GWAS have confirmed the genome-wide significant association between APOE-ε4 and DLB. Our meta-analyses have confirmed this association, and the genetic association between APOE-ε4 and PDD. Similar to this genetic association, the molecular genetics of LBD has been hypothesised to overlap with known genetic associations of AD and PD. 8 However, the genetic overlap is limited to only a few genes including APOE, ESR1, MAPT, PSEN1, TFAM, and TREM2

| Other genetic associations
that have been reported to be associated with both LBD and AD. [92][93][94] Despite LBD is an α-synucleinopathy, 95 and its reported genetic associations indicate the importance of autophagy lysosomal pathway (ALP), ubiquitin proteasome system (UPS), oxidative stress, and mitochondrial dysfunction in its complex etiopathology. Aggregation of α-synuclein leads to the formation of Lewy bodies, 96  associated with increased tau fibrilisation and deposition. 59,60,99,100 Potential interactions between these variants may lead to synergistic neurodegenerative effects of amyloid, tau, and α-synuclein deposition. 60,101 Moreover, PSEN1 variants may contribute to neurodegeneration in LBD through increased amyloid deposition. 97 The L435F PSEN1 minor allele reportedly leads to progressive loss of cortical neurons, increased apoptosis, astrogliosis, and microgliosis in PSEN1 knock-in mice. 102 The genetic associations between GBA variants and PD are well known, and the people with PD carrying GBA variants are at higher risk for developing PDD. 103,104 However, little is known about how GBA variants contribute to neurodegeneration in people with LBD. ALP and UPS are important cellular systems responsible for the degradation of misfolded proteins. 105 GBA variants are likely to impair ALP and to cause cytoplasmic accumulation of misfolded proteins. Lysosomal dysfunction coupled with higher misfolded protein burden may overwhelm the UPS and autophagy pathways and may increase α-synuclein aggregation. 51 Functional loss of GBA and consequent impaired lysosomal protein degradation have been reported to cause α-synuclein aggregation and neurotoxicity in stem cell-derived neurons. 106 Such aggregated α-synuclein may set off a self-propagating disease by inhibiting neuronal lysosomal activity. 106 Moreover, SCARB2 gene encodes a lysosomal membrane protein that transports GBA to lysosomes, and its deficiency may lead to reduced GBA activity and α-synuclein accumulation. 107 Furthermore, UCHL1 is essential for reuse of free ubiquitin and hydrolysis of substrates by neuronal UPS, and its loss of function may contribute to the formation of Lewy bodies. 79,108 The mitochondrial cascade hypothesis for AD states that an individual's mtDNA determines baseline mitochondrial function that declines with age and environmental insults resulting in AD pathology. 109 As LBD has been found to be associated with mtDNA haplogroup H, independent of APOE genotype, 25 the similar mitochondrial cascade hypothesis can be considered for LBD. APP has been found to be targeted to the mitochondria, and its progressive accumulation on mitochondrial membrane may cause mitochondrial dysfunction. 110 TFAM encodes mitochondrial transcription factor A (TFAM) that is essential for mitochondrial transcription and mtDNA replication. TFAM variants impairing its function may lead to mitochondrial dysfunction and neurodegeneration. 111 TFAM overexpression has been reported to improve hippocampal long-term potentiation and motor learning memory in mice. 106 It has been found to reduce expression of inflammatory mediators such as interleukin-1β and to reduce mtDNA damage in microglia. 111 33 and it has been associated with reduced tau phosphorylation in people with dementia. 123 The BCHE-K variant has been found to be significantly less common among people with DLB. 33 Additionally, the reported genetic association between CHRFAM7A and LBD may highlight the importance of cholinergic system dysfunction in LBD. 17,124 CHRFAM7A is a duplicated gene complex including CHRNA7 that encodes neuronal acetylcholine receptor subunit α-7 (nAChRα7).
nAChRα7 has been implicated in the pathology of several neuropsychiatric disorders, and it is involved in memory, sensory information processing, and neuronal survival. 125 In comparison with the field of molecular genetics of AD or PD,

CONFLICT OF INTEREST
None declared.

AUTHOR CONTRIBUTIONS
HS and APR conceived the study, and HS wrote the initial study protocol. The systematic review team included HS, RS, and HM. They completed necessary quality assessment and data extraction. APR completed all data analyses. HS wrote the initial manuscript with the supervision of APR. All authors were involved in further critical revisions of the manuscript, and all authors have approved the final version of the manuscript.

DATA AVAILABILITY STATEMENT
Data sharing is not applicable to this article as no new data were created or analyzed in this study.