No abstract

contributed to the result interpretation. JLB, RD, KB, JPB, BAB contributed to the genotype and RNA-Seq data generation. JCM and PJP provided the Knight ADRC data. CC, OH, JDD, ZL designed the study. CC supervised the project. All authors read and approved the manuscript.

Expression quantitative trait loci (eQTL) mapping has successfully resolved some genome-wide association study (GWAS) loci for complex traits. However, there is a need for implementing additional "omic" approaches to untangle additional loci and provide a biological context for GWAS signals. We generated a detailed landscape of the genomic architecture of protein levels in multiple neurologically relevant tissues (brain, cerebrospinal fluid (CSF) and plasma), by profiling thousands of proteins in a large and well-characterized cohort. We identified 274, 127 and 32 protein quantitative loci (pQTL) for CSF, plasma and brain respectively. We demonstrated that cis-pQTL are more likely to be shared across tissues but trans-pQTL are tissue-specific. Between 78% to 87% of pQTL are not eQTL, indicating that protein levels have a different genetic architecture than gene expression. By combining our pQTL with Mendelian Randomization approaches we identified potential novel biomarkers and drug targets for neurodegenerative diseases including Alzheimer disease and frontotemporal dementia. In the context of personalized medicine, these results highlight the need for implementing additional functional genomic approaches beyond gene expression in order to understand the biology of complex traits, and to identify novel biomarkers and potential drug targets for those traits.

Alpha-synuclein is the main protein component of Lewy bodies, the pathological hallmark of Parkinson’s disease. However, genetic modifiers of cerebrospinal fluid (CSF) alpha-synuclein levels remain unknown. The use of CSF levels of amyloid beta1–42, total tau, and phosphorylated tau181 as quantitative traits in genetic studies have provided novel insights into Alzheimer’s disease pathophysiology. A systematic study of the genomic architecture of CSF biomarkers in Parkinson’s disease has not yet been conducted. Here, genome-wide association studies of CSF biomarker levels in a cohort of individuals with Parkinson’s disease and controls (N = 1960) were performed. PD cases exhibited significantly lower CSF biomarker levels compared to controls. A SNP, proxy for APOE ε4, was associated with CSF amyloid beta1–42 levels (effect = − 0.5, p = 9.2 × 10−19). No genome-wide loci associated with CSF alpha-synuclein, total tau, or phosphorylated tau181 levels were identified in PD cohorts. Polygenic risk score constructed using the latest Parkinson’s disease risk meta-analysis were associated with Parkinson’s disease status (p = 0.035) and the genomic architecture of CSF amyloid beta1–42 (R2 = 2.29%; p = 2.5 × 10−11). Individuals with higher polygenic risk scores for PD risk presented with lower CSF amyloid beta1–42 levels (p = 7.3 × 10−04). Two-sample Mendelian Randomization revealed that CSF amyloid beta1–42 plays a role in Parkinson’s disease (p = 1.4 × 10−05) and age at onset (p = 7.6 × 10−06), an effect mainly mediated by variants in the APOE locus. In a subset of PD samples, the APOE ε4 allele was associated with significantly lower levels of CSF amyloid beta1–42 (p = 3.8 × 10−06), higher mean cortical binding potentials (p = 5.8 × 10−08), and higher Braak amyloid beta score (p = 4.4 × 10−04). Together these results from high-throughput and hypothesis-free approaches converge on a genetic link between Parkinson’s disease, CSF amyloid beta1–42, and APOE.

The identification of multiple genetic risk factors for Alzheimer Disease (AD) provides evidence to support that many pathways contribute to AD onset and progression. However, the metabolomic and lipidomic profiles in carriers of distinct genetic risk factors are not fully understood. The metabolome can provide a direct image of dysregulated pathways in the brain, including information on treatment targets. In this study, we interrogate the metabolomic and lipidomic signatures in the AD brain, including carriers of pathogenic variants in APP, PSEN1, and PSEN2 (autosomal dominant AD; ADAD), APOE ε4 and TREM2 risk variant carriers, and non-carrier sporadic AD (sAD). We generated metabolomic and lipidomic data from parietal cortical tissue from 366 participants with AD pathology and 26 cognitively unimpaired controls using the Metabolon global metabolomics platform. We identified 133 metabolites associated with disease status (FDR q-value<0.05). In sAD brains these include tryptophan betaine (b=-0.57) and N-acetylputrescine (b=-0.14). Metabolites associated with sAD and ADAD include ergothioneine (b=-0.21 and -0.26 respectively) and serotonin (b=-0.34 and -0.58, respectively). TREM2 and ADAD showed association with α-tocopherol (b=-0.12 and -0.12) and CDP-ethanolamine (b=-0.13 and -0.10). β-citrylglutamate levels are associated with sAD, ADAD, and TREM2 compared to controls (b=-0.15; -0.22; and -0.29, respectively). Additionally, we identified a signature of 16 metabolites that is significantly altered between genetic groups (sAD vs. control p = 1.05×10-7, ADAD vs. sAD p = 3.21×10-5) and is associated with Braak tau stage and disease duration. These data are available to the scientific community through a public web browser (http://ngi.pub/Metabolomics). Our findings were replicated in an independent cohort of 327 individuals.

Understanding the tissue-specific genetic architecture of protein levels is instrumental to understand the biology of health and disease. We generated a genomic atlas of protein levels in multiple neurologically relevant tissues (380 brain, 835 cerebrospinal fluid (CSF) and 529 plasma), by profiling thousands of proteins (713 CSF, 931 plasma and 1079 brain) in a large and well-characterized cohort. We identified 274, 127 and 32 protein quantitative loci (pQTL) for CSF, plasma and brain respectively. cis-pQTL were more likely to be shared across tissues but trans-pQTL tend to be tissue-specific. Between 44% to 68.2% of the pQTL do not colocalize with expression, splicing, methylation or histone QTLs, indicating that protein levels have a different genetic architecture to those that regulate gene expression. By combining our pQTL with Mendelian Randomization approaches we identified potential novel biomarkers and drug targets for neurodegenerative diseases including Alzheimer disease and frontotemporal dementia. Here we present the first multi-tissue study yielding hundred of novel pQTLs. This data will be instrumental to identify the functional gene from GWAS signals, identify novel biological protein-protein interactions, identify novel potential biomarkers and drug targets for complex traits.

Alpha-Synuclein (a-Syn) is the main protein component of Lewy bodies (LB), the pathological hallmark of Parkinson's disease (PD). Cerebrospinal fluid (CSF) levels of a-Syn are not currently used as a clinical biomarker but may be a proxy for pathological a-Syn accumulation in the brain. Therefore, identifying genetic modifiers of CSF a-Syn levels could provide insights into the underlying molecular mechanisms leading to PD. However, genetic modifiers of CSF a-Syn levels remain unknown. CSF levels of amyloid beta1-42 (Ab42), total tau (t-tau), and phosphorylated tau181 (p-tau181) are standard biomarkers for the diagnosis of Alzheimer disease (AD); its use as quantitative traits in genetic studies have provided novel insights into AD pathophysiology. A systematic study of the genomic architecture of CSF biomarkers in PD has not been conducted. Here, genome-wide association studies (GWAS) were performed using CSF biomarker levels as quantitative traits in four PD cases and control cohorts (combined N=1,960). CSF biomarker (a-Syn, Ab42, t-tau, and p-tau181) levels were significantly lower in PD cases compared with controls. An SNP, proxy for APOE e4, was associated with CSF Ab42 levels (effect=-0.5, p=9.2x10-19). Several genome-wide suggestive loci associated with CSF a-Syn, t-tau, or p-tau181 were found. Polygenic risk scores (PRS) were constructed using the latest PD risk meta-analysis (49,731 PD cases and 784,343 controls) and the largest CSF biomarkers GWAS (N=3,146). PRS calculated using META-PD were associated with PD status in the four cohorts included in the present study (p= 2.2x10-16). A highly significant correlation (Nagelkerke pseudo-R2 =2.29%; p=2.5x10-11) of the genomic architecture between CSF Ab42 and PD risk was also found. Higher PRS scores were associated with lower CSF Ab42 levels (p=7.3x10-04). Two-sample Mendelian Randomization (MR) approach revealed that CSF Ab42 plays a role in PD risk (p=1.4x10-05) and age at onset (p=7.6x10-06), an effect mainly mediated by variants in the APOE locus. Subsequently, the APOE e4 allele was associated with significantly lower levels of CSF Ab42 (p=3.8x10-06), higher mean cortical binding potentials (cortical binding of Pittsburgh compound B PET) (p=5.8x10-08) and higher Braak Ab; score (p=4.4x10-04) in PD participants. Together these results from high-throughput and hypothesis-free approaches (GWAS, PRS and MR) converge on a genetic link between PD with CSF Ab42 and APOE.

Background Cerebrospinal fluid (CSF) levels of amyloid beta (Aβ), tau and p‐tau are the standard biomarkers for diagnosis and progression of Alzheimer disease (AD). CSF levels of Alpha‐synuclein (α‐Syn) have not been very informative in Parkinson’s disease (PD). The burden of Aβ plaques and Tau tangles inversely correlates with cognitive status in PD cases with dementia. However. Lewy body aggregated correlates with dementia progression in PD. A systematic study of CSF biomarkers in PD is yet to be complete, thus, we aimed to investigate the relationship between dementia biomarkers and PD using polygenic risk scores and Mendelian Randomization. Methods Genome wide association analyses (GWAs) were performed in N = 1,960 individuals to define the genetic architecture of CSF biomarker levels (α‐syn, Aβ, tau and p‐tau). CSF values were normalized using Z‐score; linear regression was corrected by sex, age and population structure. We performed the same analyses in PD affected individuals only (N = 700) to evaluate if the signals were driven by the PD population. PD genetic architecture was obtained from the latest PD risk meta‐analysis (Nalls, et al 2019) summary statistics. . Results PRS analysis showed that CSF Aβ genetic architecture was correlated to that of PD (p = 2.5 × 10−11; R2 = 2.29%.), but not α‐syn, tau or p‐tau. Mendelian randomization methods found that CSF Aβ have a causal role in PD (p = 1.44 × 10−05); an effect mediated by a SNP near the APOE gene (rs769449). Additionally, a trend towards a causal association was found for α‐Syn and PD (p = 0.06). GWAs on PD only population did not reveal any new or known signal for those CSF proteins. Conclusion We demonstrated that Aβ is involved in the causal pathway of PD, even though APOE does not seem to be related to CSF Aβ levels or PD risk in PD cohorts. α‐Syn also seems to have a causal relation with PD, but studies with a larger number of samples are needed.