{"paper_id":"c0c1336d-bf60-42d3-b8ee-ebdcbaf636c6","body_text":"1 \n \nMultiomic Analyses Direct Hypotheses for Creutzfeldt-Jakob Disease Risk Genes \n \nFahri Küçükali1,2,a, Elizabeth Hill 3,a, Tijs Watzeels 1,2, Holger Hummerich 3, Tracy \nCampbell3, Lee Darwent3, Steven Collins4, Christiane Stehmann4, Gabor G Kovacs5, Michael \nD Geschwind 6, Karl Frontzek 7, Herbert Budka 8, Ellen Gelpi 8, Adriano Aguzzi 7, Sven J van \nder Lee 9,10,11, Cornelia M van Duijn 12,13, Pawel P Liberski 14, Miguel Calero 15, Pascual \nSanchez-Juan16, Elodie Bouaziz-Amar17, Jean-Louis Laplanche17, Stéphane Haïk 18,19, Jean-\nPhillipe Brandel 18,19, Angela Mammana 20, Sabina Capellari 21, Anna Poleggi 22, Anna \nLadogana22, Maurizio Pocchiari 22, Saima Zafar 23,24, Stephanie Booth 25, Gerard H \nJansen26, Aušrinė  Areškevi č iū tė 27, Eva Løbner Lund 27,28, Katie Glisic 29, Piero \nParchi,20,21, Peter Hermann 23,30, Inga Zerr 23,30, Jiri Safar 29, Pierluigi Gambetti 29, Brian S \nAppleby29, John Collinge3, Kristel Sleegers1,2, Simon Mead3* \n \nAffiliations \n1Complex Genetics of Alzheimer’s Disease group, VIB Center for Molecular Neurology, VIB, \nAntwerp, Belgium \n2Department of Biomedical Sciences, University of Antwerp, Antwerp, Belgium \n3Medical Research Council Prion Unit, University College London Institute of Prion \nDiseases, London, UK \n4Australian National Creutzfeldt-Jakob Disease Registry, The Florey and Department of \nMedicine (RMH), The University of Melbourne, Victoria, 3010, Australia. \n5Department of Laboratory Medicine and Pathobiology and Tanz Centre for Research in \nNeurodegenerative Disease, University of Toronto, and Laboratory Medicine Program & \nKrembil Brain Institute, University Health Network, Toronto, Ontario, Canada \n6UCSF Memory and Aging Center, Department of Neurology, University of California, San \nFrancisco, USA. \n7Institute of Neuropathology, University of Zürich, Zürich, Switzerland. \n8 Austrian Reference Centre for Human Prion Diseases, Division of Neuropathology and \nNeurochemistry, Department of Neurology, Medical University Vienna, Austria. \n9Genomics of Neurodegenerative Diseases and Aging, Human Genetics, Vrije Universiteit \nAmsterdam, Amsterdam UMC, location VUmc, Amsterdam, The Netherlands \n10Alzheimer Center Amsterdam, Neurology, Vrije Universiteit Amsterdam, Amsterdam UMC \nlocation VUmc, Amsterdam, The Netherlands \n11Amsterdam Neuroscience, Neurodegeneration, Amsterdam, The Netherlands \n12Department of Epidemiology, Erasmus Medical Centre, Rotterdam, The Netherlands  \n13Nuffield Department of Population Health, University of Oxford, UK. \n14Department of Molecular Pathology and Neuropathology, Medical University of Lodz, Lodz, \nPoland \n15Chronic Disease Programme (UFIEC-CROSADIS) and Network Center for Biomedical \nResearch in Neurodegenerative Diseases (CIBERNED), Instituto de Salud Carlos III, \nMadrid, Spain \n16Neurology Service, University Hospital Marqués de Valdecilla (University of Cantabria, \nCIBERNED and IDIVAL), Santander, Spain. \n17Department of Biochemistry and Molecular Biology, Lariboisière Hospital, GHU AP-HP \n.Nord, University of Paris Cité, France \n18Sorbonne Université, INSERM, CNRS UMR 7225, Institut du Cerveau et de la Moelle \népinière, ICM, Paris, France \n19Cellule nationale de référence des maladies de Creutzfeldt-Jakob, AP-HP, University \nHospital Pitié-Salpêtrière, Paris, France \n20IRCCS, Istituto delle Scienze Neurologiche di Bologna, Bologna, Italy. \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \nNOTE: This preprint reports new research that has not been certified by peer review and should not be used to guide clinical practice.\n\n2 \n \n21Department of Biomedical and Neuromotor Sciences, University of Bologna, Bologna, Italy \n22 Department of Neuroscience, Istituto Superiore di Sanità, Roma, Italy. \n23Department of Neurology, Clinical Dementia Center and National Reference Center for \nCJD Surveillance, University Medical School, Göttingen, Germany \n24Biomedical Engineering and Sciences Department, School of Mechanical and \nManufacturing Engineering, National University of Sciences and Technology, Islamabad, \nPakistan \n25Prion Disease Program, National Microbiology Laboratory, Public Health Agency of \nCanada, Winnipeg, Canada \n26Department of Pathology and Laboratory Medicine, University of Ottawa, Ottawa, Canada. \n27Danish Reference Center for Prion Diseases, Department of Pathology, Copenhagen \nUniversity Hospital, Rigshospitalet, Copenhagen 2100, Denmark \n28Department of Clinical Medicine, University of Copenhagen, Copenhagen, Denmark \n29National Prion Disease Pathology Surveillance Center, Case Western Reserve University, \nCleveland, OH, USA \n30German Center for Neurodegenerative Diseases (DZNE), Göttingen, Germany \n \n \n \n*corresponding author \ns.mead@prion.ucl.ac.uk \naThese authors contributed equally to this work. \n  \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n3 \n \nAbstract \nPrions are assemblies of misfolded prion prot ein that cause several fatal and transmissible \nneurodegenerative diseases, with the most common phenotype in humans being sporadic \nCreutzfeldt-Jakob disease (sCJD). Aside from vari ation of the prion protein itself, molecular \nrisk factors are not well understood. Prion and prion-like mechanisms are thought to \nunderpin common neurodegenerative disorders meani ng that the elucidation of mechanisms \ncould have broad relevance. Herein we sought to further develop our understanding of the \nfactors that confer risk of sCJD using a systematic gene prioritization and functional \ninterpretation pipeline based on multiomic integrative analyses. We integrated the published \nsCJD genome-wide association study (GWAS) summary statistics with publicly available \nbulk brain and brain cell type gene and protein expression datasets. We performed multiple \ntranscriptome and proteome-wide association studies (TWAS & PWAS) and Bayesian \ngenetic colocalization analyses between sCJD risk association signals and multiple brain \nmolecular quantitative trait loci signals. We then applied our systematic gene prioritization \npipeline on the obtained results and nominated prioritized sCJD risk genes with risk-\nassociated molecular mechanisms in a tran scriptome and proteome-wide manner. Genetic \nupregulation of both gene and protein expression of syntaxin-6 ( STX6) in the brain was \nassociated with sCJD risk in multiple datasets, with a risk-associated gene expression \nregulation specific to oligodendrocytes. Simila rly, increased gene and protein expression of \nprotein disulfide isomerase family A member 4 ( PDIA4), involved in the unfolded protein \nresponse, was linked to increased disease risk , particularly in excitatory neurons. Protein \nexpression of mesencephalic astrocyte derived neurotrophic factor ( MANF), involved in \nprotection against endoplasmic reticulum stress and sulfatide binding (linking to the enzyme \nin the final step of sulfatide synthesis, encoded by sCJD risk gene GAL3ST1), was identified \nas protective against sCJD. In total 32 genes were  prioritized into two tiers based on level of \nevidence and confidence for further studies. This study provides insights into the genetically-\nassociated molecular mechanisms underlying sCJD susceptibility and prioritizes several \nspecific hypotheses for exploration beyond the prion protein itself and beyond the previously \nhighlighted sCJD risk loci through the newly prioritized sCJD risk genes and mechanisms. \nThese findings highlight the importance of glia l cells, sulfatides and the excitatory neuron \nunfolded protein response in sCJD pathogenesis. \nKey Words: Sporadic Creutzfeldt-Jakob disease (sCJD), Multiomics, Neurodegeneration, \ntranscriptome-wide association studies (TWAS), proteome-wide association studies (PWAS)\n \n  \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n4 \n \nIntroduction \nPrions are infectious, proteinaceous pathogens composed of fibrillar assemblies of misfolded \nforms of host-encoded prion protein (PrP) 1. Prions replicate by templated misfolding leading \nto fibril growth and fission 2. Prion propagation leads to the generation of neurotoxic species \nand neurodegeneration. This underlying molecular mechanism is at the core of a multitude \nof human and animal prion diseases, and several aspects of the mechanism (so-called \n“prion-like”) are shared with the more common neurodegenerative disorders\n2. \n \nHuman prion diseases are unusual amongst neurodegenerative diseases in having three \ndifferent types of aetiology: as well as arising due to rare pathogenic mutations in PRNP \nencoding PrP\nC (inherited prion disease accounting for ~10-15% cases) and spontaneously \n(sporadic prion disease accounting for ~85% cases), the disease can also be acquired \nthrough transmission between humans or zoonotically(<1% cases)\n3-5. Sporadic Creutzfeldt-\nJakob disease (sCJD) is the most common human prion disease, which has a lifetime risk of \n~1:5000\n6, and typically presents as a rapidl y progressing dementia. There are no \nestablished disease-modifying treatments for sCJD although treatments targeting PrP using \ndifferent therapeutic modalities such as employing PrP-targeting monoclonal antibodies have \nbeen reported \n7 and PRNP-targeting ASOs (Phase 1/2a trial employing ION717, \nNCT06153966) are under investigation. Currently how ever the diseases are universally fatal \nand, for optimal disease mitigation, new t herapeutic targets may be required beyond PrP \nitself.  \n \nIn 2020, a collaborative genome-wide association study (GWAS) was conducted in sCJD, \nwhich identified novel risk loci for sCJD susceptibility 8. In addition to the well-known variants \nin the PRNP gene, this study independently replicated findings at two further novel loci, at or \nwithin STX6 and GAL3ST1,  to be associated with sCJD risk. STX6 encodes syntaxin-6, a \nSNARE protein predominantly involved in retrogr ade trafficking from early endosomes to the \ntrans-Golgi network9,10, implicating intracellular trafficking as a causal molecular pathway in \nsCJD. GAL3ST1 encodes galactose-3-O-sulfotrans ferase 1 predominantly in \noligodendrocytes, the exclusive enzyme involved in the final step of sulfatide synthesis, \nwhich is a key constituent of the myelin sheath 11. Two other genes were implicated in sCJD \nrisk by tests that summarise evidence for asso ciation across the entire gene locus, including \nPDIA4, and variants in and near to a further gene, BMERB1 , which were very close to \ngenome-wide thresholds of association8. \nWe aimed to harness transcriptomic and proteomi c datasets to provide further insight into \nsCJD risk in studies such as transcriptome-wide association studies (TWAS) and proteome-\nwide association studies (PWAS), respectively when integrated with the genetic datasets. \nHerein, the latest sCJD GWAS summary statistics 8 were integrated with functional \nannotations (expression quantitative trait loci [eQTL] and protein expression QTL [pQTL]) to \ninfer genetic up- and down-regulation of genes and/or protein expression in brain regions \nand associated with sCJD susceptibility. As the approach in TWAS/PWAS combines \nassociations across variants, thus reducing t he multiple testing burden, these analyses offer \na powerful, complementary approach to c onventional GWAS to develop supporting or \nnegating evidence for loci that were subthreshold ( PDIA4, BMERB1) or loci that did not \nreach the genome-wide significant threshold in the previous GWAS 8. Furthermore, it allows \nexploration of expression-related genetic mechanisms underlying the GWAS association \nsignals already identified ( PRNP, STX6, GAL3ST1) uncovering further mechanistic insights \ninto sCJD risk loci, in addition to nominating new TWAS/PWAS significant prioritized risk \ngenes within subthreshold loci for generat ing novel disease-relevant hypotheses. \nImportantly, there are precedents of similarly designed studies achieving these goals in other \nneurological diseases12-18. \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n5 \n \nThis work provides compelling evidence for risk variants in and around the STX6 locus \ndriving increased transcript and protein expr ession in the brain and consequently disease \nrisk, which intriguingly and unexpectedly predominates in oligodendrocytes. This study also \nprioritizes the previous subthreshold GWAS hit, PDIA4, which is involved in the unfolded \nprotein response (UPR), as being implicated in sCJD susceptibility, driven by PDIA4 \nupregulation. Interestingly, this effect seem ed to localise to excitatory neurons with \ninteractions with the PWAS hit, MANF, providing an intriguing link to sulfatide metabolism \nand GAL3ST1. Several other subthreshold hits were also identified with potential relevance \nto prion disease mechanisms, including the previously identified subthreshold GWAS hit, \nBMERB1. \nTaken together, this study prioritized a num ber of candidate genes, both novel hits and \nrefining existing GWAS hits, at sCJD-associated loci aiding the identification of causal risk \ngenes at GWAS signals by combining results from complementary eQTL and pQTL-based \nstudies. \n  \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n6 \n \n \nMethods \nsCJD GWAS summary statistics \nWe used the summary statistics of the latest and the largest sCJD GWAS available from the \nGWAS Catalogue (GCST90001389)8. The discovery stage of this GWAS was performed on \n17,679 samples (4,110 cases and 13,569 controls), and the summary statistics contained \ninformation on 6,314,492 high-quality imputed single-nucleotide polymorphisms (SNPs) \nacross the autosomes 8. As the original sCJD GWAS summary statistics were in GRCh37 \nhuman reference genome assembly and the molecular QTL catalogues and TWAS/PWAS \npanels used were in GRCh38 assembly, we first lifted over the variant positions from the \nGRCh37 to the GRCh38 genome build by using Picard (v2.22.10) LiftOver tool with \n“RECOVER_SWAPPED_REF_ALT=true” parameter. The SNPs that could not be lifted over \nto the GRCh38 genome build (7,052 SNPs; corresponding to 0.1% of total) were excluded \nfrom this study, and the remaining variants were reannotated with dbSNPv151 (GRCh38) \nusing BCFtools annotate function. The resulting file was used in downstream molecular \nQTL-based analyses (e/pQTL-GWAS coloc and TWAS/PWAS) for the systematic gene \nprioritization pipeline. \nGene prioritization and functional interpretation analyses \nFor the systematic gene prioritization pipel ine we considered three domain-specific \nanalyses, namely variant annotation, eQTL-GWAS integration, and pQTL-GWAS integration \ndomains, for which detailed information is provided below. \nVariant annotation \nWe considered the index variants in each locus described in the sCJD GWAS publication\n8, \nnamely rs3747957 in STX6 locus, rs1799990 in PRNP  locus, rs2267161 in GAL3ST1 locus, \nrs9065 in PDIA4 locus, and rs6498552 BMERB1 locus for three specific criteria. First, we \ninvestigated the nearest protein-coding genes with respect to the genomic position of these \nlead SNPs; then we queried whether they are rare (MAF < 1% in gnomAD v4.1 non-Finnish \nEuropean [NFE] samples) and/or protein-altering (missense or predicted loss-of-function) \ngenetic variants for the nearest protein-codi ng genes they might reside in. Detailed \ninformation on these SNPs can be found in Supplementary Table 1.  \neQTL-GWAS integrative analyses \nFor the eQTL-GWAS integrative analyses, we  processed and used publicly available bulk \nbrain and brain cell-type-specific cis-eQTL catalogues and TWAS reference panels from \ndifferent cohort and datasets. These included 6 bulk brain region datasets (as reanalyzed \nand described in detail in Bellenguez et al. 15) of 3 AMP-AD cohorts; namely, the Mayo \nRNAseq Study (MayoRNAseq 19) temporal cortex (TCX), the Religious Orders Study and \nMemory and Aging Project (ROSMAP 20,21) dorsolateral prefrontal cortex (DLPFC), and The \nMount Sinai Brain Bank study (MSBB22) Brodmann areas (BA) 10, 22, 36, and 44. Moreover, \nthe following 4 additional bulk brain region datasets of GTEx v8 cohorts 23 were used for \neQTL-based analyses: hippocampus, frontal cortex, cortex (right cerebral frontal pole), and \nBA24. Furthermore, we leveraged the inform ation cell-type-specific eQTLs (ct-eQTL) \nmapped in eight major brain cell types (exci tatory neurons, inhibito ry neurons, astrocytes, \noligodendrocytes, microglia, oligodendrocyte precursor cells/committed oligodendrocyte \nprecursors [OPCs/COPs], pericytes, and endothelial cells) from Bryois et al. 24 and in primary \nmicroglia from Young et al. 25 and from the Microglia Genomics Atlas (MiGA) study 26 (medial \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n7 \n \nfrontal gyrus, superior temporal gyrus, subvent ricular zone, thalamus, and meta-analysis of \nfour brain regions). Further information on each cohort and dataset can be found in \nrespective publications cited and in Supplementary Table 2. \nTo investigate the potential genetic colocalization between sCJD risk association signals and \neQTL/ct-eQTL signals controlling cis gene expression of nearby (1 Mb) genes in bulk brain \nand in brain cell types, we performed Bayesian colocalization analyses using coloc (v5.2.2; \n“coloc.abf” function with default priors) 27 for each tested gene within above mentioned 24 \ndistinct eQTL/ct-eQTL catalogues. The coloc analyses outputs for posterior probabilities \n(PPs) for five following hypotheses regarding tw o signals compared: H0 (no causal variant \nfor both traits), H1 (causal va riant only for sCJD GWAS), H2 (causal variant only for eQTL), \nH3 (two different causal variants) and H4 (common causal variant shared between sCJD \nGWAS and eQTL). We defined a eQTL signal as colocalized with sCJD GWAS if coloc PP4 \n(the posterior probability for H4) was ≥ 70%. Furthermore, we investigated the association \nbetween genetically regulated predicted gene expression and sCJD risk by performing \nTWAS in 10 bulk brain gene expression reference panels for each heritable gene expression \nfeature. We used FUSION28 pipeline (using “FUSION.assoc_test.R” with default parameters) \nto run TWAS on 6 bulk brain custom gene expression reference panels from AMP-AD \ncohorts together with a custom linkage disequi librium (LD) reference data derived from 1000 \nGenomes (1KG) project unrelated non-Finnish European samples (as described in detail in \nBellenguez et al.\n15), meanwhile MASHR models of remaining 4 GTEx v8 brain region \nreference panels were used with S-PrediXcan 29,30 (with non-default parameters “--\nkeep_non_rsid --model_db_snp_key varID --additional_output –throw”) implemented in \nMetaXcan v 0.6.12 tools 29. We determined the transcriptome-wide significance thresholds \nbased on the Bonferroni correction on transcriptome -wide number of tested features in each \ngene expression reference panel ( Supplementary Table 2 ). Moreover, fine-mapping of \nsignificant TWAS results was performed with Fine-mapping Of CaUsal gene Sets (FOCUS)31 \nv0.803 tool within five distinct genetic regions constructed by 1 Mb extended GWAS index \nvariant coordinates (with “--locations” parameter), where we calculated posterior inclusion \nprobabilities (PIPs) for TWAS associations and used these to define associations within 90% \ncredible sets as fine-mapped TWAS associations. \npQTL-GWAS integrative analyses \nFor the pQTL-GWAS integrative analyses, we accessed the publicly available bulk brain cis-\npQTL datasets from Wingo et al. 16 and reprocessed and reannotated these for pQTL-GWAS \ncoloc and PWAS analyses. First, pQTL-GWAS coloc analyses were performed as described \nabove using coloc pipeline, and by using pQTL catalogue (v2) from ROSMAP DLPFC \ncohort. Second, ROSMAP DLPFC (v2) and Banner  Sun Health Research Institute (Banner) \nDLPFC PWAS reference panels were used using FUSION pipeline described above. \nDetailed information on these datasets and cohorts , including PWAS significance thresholds \nand number of samples, can be found in Supplementary Table 2. \nSystematic gene prioritization \nTo combine evidence for each candidate sCJD  risk gene and nominate prioritized sCJD risk \ngenes and related risk-associated molecular mechanisms, we applied a systematic gene \nprioritization and functional interpretation anal ysis pipeline adapted from Bellenguez et al. \nstudy\n15 for Alzheimer’s disease (AD). We first brought together all evidence for the candidate \nsCJD risk genes as a result of (i) variant annotation, (ii) eQTL-GWAS integration, and (iii) \npQTL-GWAS integration domain analyses, each having various categories and \nsubcategories with predetermined weighting schem e for single hits and replicated hits \n(across different e/pQTL coloc or TWAS/PWAS analyses), all described in detail in \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n8 \n \nSupplementary Table 3 . The weighted sum of the hits in different categories resulted in a \ngene prioritization score (between 0-42) for each candidate gene (i.e, a gene with a hit in at \nleast one subcategory and with a gene prioritization score >0). \nThis was followed by the assignment of each candidate gene based on their genomic \ncoordinates to 3 different types of loci and indexed: (i) the genes within 1 Mb extended \ncoordinates of 3 genome-wide signifi cant (GWS) index variants (with P ≤  5x10-8) from the \nsCJD GWAS assigned to respective 3 GWS loci ( STX6 [G1], PRNP [G2], and GAL3ST1 \n[G3] loci), (ii) the genes within 1 Mb ext ended coordinates of 2 highlighted subthreshold \nindex variants (with P ≤  5x10 -6) from the sCJD GWAS assigned to two subthreshold loci \n(PDIA4 [S1] and BMERB1  [S2] loci), and (iii) the remaining candidate genes were grouped \ntogether if they were positioned together (<1 Mb) and these resulted in an additional 26 \nother loci (indexed as O1-O26). Using the pipeline described in Bellenguez et al. 15, we then \nranked all the protein-coding candidate genes in each locus based on their total weighted \nscores, determined the top-ranked genes, and compared the relative score differences \nbetween the top-ranked genes and the other genes in each locus to classify them as tier 1 \nand tier 2 prioritized risk genes, representing hi gher and lower levels of confidence for being \ntrue risk genes in loci, respectively. Furthermore, using a large publicly available single-\nnucleus RNA sequencing (snRNA-s eq) study of 1.4 M nuclei from 84 human dorsolateral \nprefrontal cortex brain samples (The Seattle Alzheimer’s Disease Cell Atlas [SEA-AD]\n32), we \nfirst estimated average gene expression of each candidate risk gene within annotated major \nbrain cell type clusters and then calcul ated the cell-type-specific gene expression \nproportions across 7 major brain cell types. Fi nally, gene set enrichment and protein-protein \ninteraction analyses for the gene lists of tier 1 and all prioritized risk genes were performed \nusing STRING v1233 with default parameters. \n  \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n9 \n \nResults \nOur systematic gene prioritization pipeline ident ified 17 tier 1 prioritized risk genes and 15 \ntier 2 prioritized risk genes in 30 risk loci ( Fig. 1 and Supplementary Table 4 ). We could \nresolve all 3 GWS and 2 subthreshold sCJD risk loci with nominated tier 1 risk genes. Our \nintegrative multiomic analyses identified candidate ri sk genes in another 26 loci (‘other’ loci), \nof which 12 harboured tier 1 prioritized risk genes.  \nGenome-wide significant loci \nAt the chromosome 1 STX6 locus (G1) we observed 23 GWS SNPs. STX6 was the nearest \ngene to the synonymous index variant rs3747957 and also the tier 1 prioritized gene with the \nhighest score in this study (23), as its prio ritization was supported by replicated hits in \nmultiple subcategories ( Fig. 1-2 and Supplementary Tables  5-9). Remarkably, across 10 \nbulk brain cohorts, we found strong evidence of eQTL-GWAS colocalization (PP4s = 94.3-\n98.0%), in addition to having a ct-eQTL-GWAS colocalization hit specific for \noligodendrocytes (PP4 = 97.7%) and a pQTL-GWAS coloc hit in DLPFC (PP4 = 99.2%) \n(Fig. 2 ). Moreover, the fine-mapped TWAS results showed that genetic upregulation of \nSTX6 was significantly associated with increased sCJD risk in multiple studies (FOCUS PIPs \n= 0.92-1; the most significant being in the brain region BA44; P = 7.92x10 -9, Z-score = \n+5.77), which was also supported by protein expression level with replicated PWAS hits (P = \n1.34x10-8, Z-score = +5.68 and P = 1.25x10-6, Z-score = +4.85 in the DLPFC analyses of the \nROSMAP and Banner cohorts, respectively; Fig. 3). \nAt the chromosome 20 PRNP locus (G2) 16 GWS SNPs were located within PRNP. While \nwe did not detect any coloc or TWAS driven evidence for any gene in this locus, we \nprioritized PRNP as the tier 1 risk gene, because the index variant rs1799990 was a \ncommon (NFE MAF = 34.3%) missense (p.Met129Val; CADD = 17.85) variant ( Fig. 1 and \nSupplementary Table 1). The same variant was also the most significant pQTL at this locus \namong the 183 tested. The major and protective G allele (p.129Val) was nominally \nassociated with increased PrP levels in DLPFC ( P = 8x10\n-3, beta = +0.019). Nevertheless, \npQTL coloc results for PRNP showed limited pQTL-GWAS coloc (PP4 = 41%) due to the \nmodest pQTL signal in the locus. Furthermore, no heritable PWAS models were available for \nPrP protein expression, thus it could not be tested in PWAS. Risk conferred by rs1799990 is \ncomplex, in that it varies between aetiological types of prion disease 34-37, but the \nheterozygous genotype is strongly associated wi th reduced risk of sCJD and more slowly \nprogressive phenotypes relative to both homozygous genotypes 34. The biological relevance \nof this pQTL is therefore unclear. \nAt the chromosome 22 GAL3ST1  locus (G3) we found 2 GWS SNPs centering GAL3ST1 \nas the nearest gene. GAL3ST1 could be prioritized as the tier 1 risk gene, as the index \nvariant rs2267161 (p.Met29Val, CADD score 18.13) was a common (NFE MAF = 31.1%) \nmissense variant, despite the presence of two other candidate genes in the locus: TCN2 (50 \nkb downstream from the index variant with fine-mapped TWAS hit in BA22) and INPP5J \n(>500 kb downstream from the index variant with a borderline eQTL coloc hit in DLPFC in \nthe ROSMAP cohort) ( Fig. 1  and Supplementary Tables 5 and 7 ). Moreover, GAL3ST1 \np.Met29Val index variant has strong associations with sulfatide (SHexCer) blood lipids (five \ndifferent classes and total SHexCer, P=2.5x10\n-15 – 2.7x10 -37) with the sCJD risk allele \nrs2267161-C conferring increased lipid levels 38. \nSubthreshold loci \nAt the chromosome 7 PDIA4 locus (S1), which was a hit in gene-wide analyses in the \nprevious study 8, the GWAS association signal surrounded PDIA4  with a minimum P  of \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n10 \n \n1.66x10-6 for the 3' UTR index variant rs9065.  We detected multiple lines of evidence \nsupporting PDIA4 (gene prioritization score of 21, the second highest in this study after \nSTX6) as a tier 1 prioritized in this locus ( Fig. 1 and Supplementary Tables 5-9 ). Across 8 \nbulk brain cohorts, we found strong evidence of eQTL and sCJD risk colocalization (PP4s \nbetween 85.1%-96.1%). This appeared to be largely driven by excitatory neurons (PP4 = \n76.2%). PDIA4 was also a fine-mapped TWAS hit (FOCUS PIP = 98.9%, P = 1.02x10\n-6, Z-\nscore = +4.89) and a significant PWAS hit in Banner DLPFC ( P = 1.1x10-5, Z-score = +4.39; \nFig. 3). Genetic upregulation of both transcript and protein expression confer increased risk \nof sCJD. Finally, PDIA4 pQTLs also colocalized with the sCJD GWAS (PP4 = 94.7%). \nAt the chromosome 16 BMERB1 locus (S2), the intronic index variant rs6498552 was close \nto the GWS threshold (rs6498552 P = 5.73x10-8) 8. BMERB1 (formerly known as C16orf45) \nwas the only candidate gene in S2 and we prioritized it as a tier 1 risk gene, as it had \nreplicated fine-mapped TWAS hits in GTEx Frontal Cortex ( P = 4.7x10-6, Z-score = -4.58, \nFOCUS PIP = 96.5%) and GTEx Hippocampus (P = 4.7x10-6, Z-score = -4.58, FOCUS PIP = \n96.9%) analyses where the predicted gene expression was conversely associated with the \nrisk of sCJD (Fig. 1 and Supplementary Table 7).  \nOther loci \nOf the remaining 26 ‘other’ loci, 22 had pr otein-coding genes in which we performed gene \nprioritization analysis. Of note, variant annotation domain does not contribute to gene \nprioritization in these loci because they do not harbour GWAS index variants 8. Nevertheless, \nwe could assign a tier 1 prioritized risk gene in 12 of these 22 loci. Moreover, for the \nremaining 10 risk loci, 8 had a single tier 2 prioritized risk gene and 2 (O7 and O9) had two \ntier 2 prioritized risk genes with similar weighted gene prioritization scores ( Fig. 1  and \nSupplementary Table 4 ). While full results on these prioritized genes are available in \nSupplementary Table 4 , below we highlight 5 of these loci containing the five highest \nscoring candidate genes (gene prioritization scores ≥ 7; all supported by hits in multiple \nsubcategories, see Fig. 1), in addition to SIRPB1 in O25 with considerable GWAS evidence. \nIn locus O10 we identified MANF as tier 1 prioritized risk gene, which was also the highest \nscoring gene (gene prioritization score of 8) among the other loci candidate genes. MANF \nhad a pQTL-GWAS coloc hit (PP4 = 88.1%) and PWAS hit ( P = 1.35x10-6, Z-score = -4.35; \nFig. 3 ) in DLPFC in the ROSMAP cohort, where genetic downregulation of protein \nexpression was associated with increased risk of sCJD. Moreover, LPIN1 (O1) and GSAP \n(O16) tier 1 prioritized risk genes both notabl y exhibited replicated bulk brain eQTL-GWAS \ncoloc hits (in 9 and 7, across 10 analyses, respectively) and also had ct-eQTL-GWAS coloc \nhits in multiple brain cell types (3 and 4, across 8 cell types), showing the importance of \nsCJD risk-associated genetic variation in both loci in terms of gene expression regulation \nacross multiple brain regions and cell types. In locus O14, HBS1L was identified as the tier 1 \nprioritized gene through a pQTL-GWAS coloc hit in DLPFC (PP4 = 86.1%) and borderline \nmicroglia ct-eQTL-GWAS coloc hit in meta- analysis of the MiGA data (PP4 = 70.2%). \nFurthermore, locus O9 had 4 protein-coding c andidate genes (the highest among all loci), in \nwhich two genes were prioritized as tier 2 ri sk genes as the weighted evidence was similar: \nNCKIPSD and INKA1 (formerly known as FAM212A), positioned furthest away from each \nother in the locus (>1.1 Mb), had both eQTL-GWAS coloc and TWAS hits. NCKIPSD scored \none point higher than INKA1 because of having replicated eQTL-GWAS coloc hits (9 out of \n10 analyses), although coloc PP4 for INKA1 in DLPFC in the ROSMAP cohort was higher \n(98.8% vs 87.8%). Finally, in locus O25, located >3 Mb upstream of PRNP, SIRPB1 was \nprioritized as a tier 2 risk gene as a result of an eQTL-GWAS coloc hit in BA10 (PP4 = \n86.1%). Of note, SIRPB1 had the second most significant GWAS P evidence among other \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n11 \n \nloci candidate genes after the genes within locus O9, as the GWAS P for its 3’UTR variant \nrs2422615 was 5.26x10-6 (Fig. 1 and Supplementary Tables 5-9).  \nGene set enrichment and interaction analyses \nUsing STRING, we performed gene set enrichment and protein-protein interaction analyses \nfor the gene lists of tier 1 and all prioritized risk genes. While no significant pathways (FDR < \n0.05) were found to be enriched when corrected for multiple comparisons, we detected a \nstrong protein-protein interaction relationship between PDIA4 and MANF  on the basis of \nexperimental/biochemical data, co-expression, and mentions of both genes in abstracts in \nthe literature (see Discussion). There was also some suggestive evidence for protein-protein \ninteractions between PRNP and SIRPB1, TRANK1 and DCLK3, as well as LPIN1 and \nACAA1, although none of these were based on human experimental/biochemical data. \n  \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n12 \n \nDiscussion \nTranscriptome and proteome-wide association studies (TWAS and PWAS) and molecular \nQTL-GWAS colocalization analyses can contribute to a better understanding of genetic risks \nfor diseases through refining hypotheses about implicated genes, direction of effects, cell \ntypes and pathways using GWS and subthreshold fi ndings. Human prion diseases have not \npreviously been studied in this way and, beyond t he prion protein locus itself, suffer from a \npaucity of genetically validated targets for therapeutic development. In 2020, a large GWAS \nstudy in the prion disease field led to the di scovery of three proposed genetic loci associated \nwith sCJD risk\n8 in or near to PRNP, STX6  and GAL3ST1, and we highlighted two \nsubthreshold loci (PDIA4 and BMERB1). We aimed to harness transcriptomic and proteomic \ndatasets to provide further insight into sCJD risk. Herein we report considerable molecular \nQTL-based evidence that supports  a causal role for genetically upregulated syntaxin-6 gene \nand protein expression in risk of sCJD relative to other genes at the locus, and a cell-type-\nspecific relevance of the GWAS signal in regulating STX6 gene expression in \noligodendrocytes but not in other brain cell types. Furthermore, both subthreshold hits we \npreviously highlighted, PDIA4 and BMERB1 , also show significant associations between \ntheir genetically regulated expression and sCJD risk. We also found that reduced protein \nexpression of a further gene product, pr eviously unconnected to prion diseases, MANF , was \nassociated with increased sCJD risk in PWAS. Interestingly, the sCJD proposed risk gene \nGAL3ST1 encodes an enzyme involved in the synthesis  of sulfatides, which are a major lipid \ncomponent of the myelin sheath and are known to have experimental links with both MANF \nand ER stress\n39, providing an indirect link to PDIA4. This work therefore refines and \nproposes new hypotheses about mechanisms of risk in human prion diseases. \nVariants in and near to the syntaxin-6 (STX6) gene are genetic risk factors for sCJD8 and the \nmost common primary tauopathy, progressive supranuclear palsy (PSP) 40-44. Syntaxin-6 is a \nmember of the SNARE protein family 10, which mediate the final step of membrane fusion \nduring vesicle transport, and thus its identification in GWAS implicated intracellular trafficking \nas a causal disease mechanism. However, although STX6 appears to modify disease \nsusceptibility8, in more recent work we have shown there is no association with age of onset \nor disease progression 45, and knockout of Stx6 expression in mouse has no, or modest \neffects, on prion disease incubation time 46. In this work, we show increased STX6 \nexpression was significantly linked to risk of sCJD across multiple reference panels both for \nTWAS and PWAS, along with e/pQTL-GWAS colocalization, whereas evidence was limited \nfor other genes (including KIAA1614) at the same locus. These findings are concordant with \nprevious studies in tauopathies correlating genetic risk loci with transcriptomic and proteomic \ndata. Indeed, using reference data from the GTEx Consortium, a PSP TWAS study identified \nthat the STX6 risk haplotype was associated with differential expression of the gene\n41. \nFurthermore, a recent frontal cortex ca se–control EWAS meta-analysis identified STX6 as \nbeing hypomethylated at CpG sites in PSP compared to controls 47. Interestingly, STX6 has \nalso been identified as conferring Alzheimer’s disease (AD) risk in a recent AD PWAS study, \nwith increased syntaxin-6 protein levels in the brain being causally associated with the \ndisease48. We conclude that syntaxin-6 has pleiotropic risk effects in neurodegenerative \ndiseases, which are driven by a common genetic mechanism of increased protein \nexpression. \nAs expected, PRNP  and GAL3ST1 were not identified as PWAS or TWAS hits, which is in \nkeeping with the candidate mechanisms of these genes being driven by common missense \nvariants. At PRNP, the p.Met129Val polymorphism is known to be a strong modifier of prion \ndisease determining predisposition to sCJD 49 and iatrogenic CJD (iCJD) 50, as well as \ninfluencing age of disease onset and/or disease progression in kuru 37 and some inherited \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n13 \n \nprion diseases 36, where in general the heterozygous genotype is protective compared to \nboth homozygous genotypes. It is important to note that most molecular QTL studies, \nincluding the ones used in our study, are based on additive models (where the effect of \nincreasing number of alleles are tested against the molecular phenotype outcome), therefore \nthis can be one of the limiting factors for finding significant downstream effects of this genetic \nvariant on PrP expression. Moreover, codon 1 29 has complex effects, exemplified by \nsusceptibility to variant CJD (vCJD), the human form of bovine spongiform encephalopathy, \nwith all but one definite case being homozygous for methionine at codon 129\n51. These \nhuman associations correlate well with modelling of the codon 129 genotype in mouse 52 and \nare in keeping with a mechanism of codon 129 genotypic risk that involves the selection of \nprion strains and dominant negative effects.  Galactose-3-O-sulfotransferase 1 (GAL3ST1) is \nan oligodendrocyte expressed enzyme, which catalyses the sulfation of Golgi-membrane \nsphingolipids to form sulfatides. These are important lipids in the brain and essential \nconstituents of the myelin sheath 11. In the GAL3ST1 gene, a common amino acid variant \n(p.Val29Met) confers increased risk of sCJD. In recent lipidomics GWAS studies the \np.Val29Mel variant was associated with altered concentrations of blood sulfatides \n38,53. \nTherefore, as there is alread y strong evidence for a genetic mec hanism at both of these loci \nindependent from expression change, we would not expect either PRNP nor GAL3ST1 to be \na TWAS/PWAS hit. Of note, TCN2, upstream at the GAL3ST1 locus, was identified as a \nfine-mapped TWAS hit in a single cohort, and is t herefore an alternative albeit lower priority \ncandidate at the locus. \nPreviously, we reported suggestive evidence that the PDIA4 locus was associated with \nsCJD risk by gene-based testing in the discovery stage of GWAS 8. These TWAS and PWAS \nanalyses provide an additional, complementary appr oach to explore the association of the \nPDIA4 locus with sCJD risk. PDIA4 was both TWAS and PWAS significant with a consistent \npositive Z-score suggesting genetic upregulation of this gene increases risk for sCJD, \nsupported also by the replicated e/pQTL-GWAS colocalization. PDIA4  encodes a member of \nthe protein disulphide isomerase (PDI) family  of proteins and is localised to the endoplasmic \nreticulum (ER) where it mediates oxygen-dependent disulphide bond formation and \nconsequently the correct folding of both transmembrane and secreted proteins 54. It has \nbroad brain expression and its function has been linked to the unfolded protein response \n(UPR). Interestingly, PDIA4 has been implicated in prion disease pathogenesis 55 as well as \nindependently emerging as a central, generic pl ayer in other neurodegenerative diseases \n(reviewed in 56) suggesting it may have risk effects across multiple protein misfolding \ndiseases. Specifically, the PDI gene family is upregulated in prion-infected cultured cells as \nwell as in prion-infected hamster brains ea rly in disease pathogenesis, which progressively \nincreases at later stages of the disease 55. This is further supported by two further \nindependent studies showing Pdia4 is upregulated both at the RNA and protein level in mice \ninfected with RML prions57.  \nThe identification of PDIA4 as a TWAS/PWAS hit localising to excitatory neurons (through \nthe ct-eQTL-GWAS coloc analyses) further implicates the UPR in human sCJD. Although \nthe UPR is a physiologically protective cellul ar response, which protects against ER stress \ndriven by the accumulation of misf olded proteins or other stressors 58, dysregulation of the \nUPR across multiple neurodegenerative diseases leads to translational failure ultimately \nculminating in neuronal loss 59-61. This translational failure is driven by the phosphorylation of \nthe α -subunit of eukaryotic translation initiation factor, eIF2 α 62. Importantly, the UPR has \nbeen highlighted as a mechanism in prion disease pathogenesis, with eIF2 α -P driving \npersistent translational repression of global prot ein synthesis in prion-infected mice, leading \nto synaptic failure and neuronal loss 63. In a more recent study it has been shown that the \nprotracted UPR typical of prion diseases also  induces diacylation of a key phosphoinositide \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n14 \n \nkinase, PIKfyve, resulting in its degradation and consequently endolysosomal hypertrophy \nand activation of TFEB-dependent lysosomal enzymes 64. This has been proposed to \nunderpin a defining histopathological trait of sCJD: spongiform degeneration. Therefore, the \nidentification of PDIA4 in this study, and its strong links to the UPR ,are in keeping with the \nemerging theme in the prion disease field that a dysregulated UPR is a driver of \nneurotoxicity. \nContinuing with this theme, Mesencephalic Astrocyte-derived Neur otrophic Factor ( MANF), \nalso implicated in the ER stress response, was a PWAS and pQTL-GWAS coloc hit. \nAlthough it did not surpass the stringent threshold of significance in the Banner DLPFC \nPWAS reference panel ( P = 3x10 -4, Z-score = -3.65; Fig. 3 ), this analysis supported the \nsame direction of effect at a suggestive sign ificance level and its conserved position in the \ntop three most significant hits across panels provides confidence its levels are associated \nwith risk of the disease. Mammalian MANF was first reported to have neurotropic effects on \ndopaminergic neurons65, promoting their survival 66. It has particularly high expression in the \nbrain (reviewed in67) with ER stress promoting its upregulation 68 as well as its secretion into \nthe extracellular environment69,70. MANF has been shown to be an important regulator of the \nUPR68,71, which is further supported with studies using mice with Manf knockout which show \nabnormal activation of the UPR 72. Interestingly, it has recently been shown that human \nMANFs directly bind to sulfatide promoting the cellular uptake of MANF, which alleviates the \nER stress response in cells thereby conferring cytoprotection 39. Its identification in this study \nas a new candidate gene therefore provides potential convergence with another sCJD risk \ngene, GAL3ST1. Additionally, as a secreted factor fr om astrocytes, it provides support for \nthe increasingly accepted notion that the inte rplay between astrocytes and neurons in prion \ndisease is a key pathogenic phenomenon\n73. \nInterestingly SIRPB1, located >3 Mb upstream of PRNP and genetically linked to PRNP , \nwas prioritized as a tier 2 risk gene with  there being suggestive evidence for a protein-\nprotein interaction between PRNP and SIRPB1. SIRPA encodes signal regulatory protein α  \n(SIRPα ), a protein enriched in microglia which plays a key modulatory role of phagocytosis. \nHowever, SIRPα  does not appear to play a role in prion pathogenesis in vivo74. \nAnother fascinating finding that came out of this study comes from analysis of cell-type-\nspecific eQTLs (ct-eQTLs), which revealed stri king cell-type-specific effects in the genetic \ncontrol of STX6 gene expression by risk variants with the STX6 signal specifically \ncolocalizing with oligodendrocyte eQTLs ( Fig. 2 ). This provides suggestive evidence that \nSTX6 may be exerting its risk effects in ol igodendrocytes. Oligodendrocytes are an \nunderstudied cell population in the prion disease fi eld, but one study provided evidence that \noligodendrocytes do not replicate prions and are resistant to prion infection 75. However, it is \npossible that the relationship between neurons, oligodendrocytes and other brain cell types \nis crucial for prion formation, propagation, clearance or neurotoxicity. Indeed, there is \nsuggestive evidence for a role of oligodendrocytes in prion disease through dysregulation of \noligodendrocyte-specific genes in transcriptomic studies\n76-78. Furthermore, a recent study \nshowed that NG2 glia, oligodendrocyte-lineage cells , exert a protective effect against prion-\ninduced neurotoxicity by interacting with microglia and inhibiting critical signalling \npathways\n79. It is also noteworthy that in human patients, oligodendroglial PrP pathology has \nbeen reported in certain histotypes of sCJD 80. Therefore, oligodendrocytes may be \nimplicated in prion pathogenesis, which is further supported by the convergence of the two \nnon-PRNP sCJD risk factors, STX6 and GAL3ST1, in this cell type. \nThis study has also several limitations. Fi rstly, our molecular QTL-based analyses were \nlimited to eQTLs and pQTLs; however, the inclusion of other molecular QTLs such as \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n15 \n \nsplicing QTLs (sQTLs), methylation QTLs (mQT Ls), and histone acetylation QTLs (haQTLs) \nin future studies could provide additional sCJD risk-associated molecular mechanisms, \nwhich can be complementary in terms linking the GWAS signals to similar sets of prioritized \nrisk genes or to other candidates. Secondl y, the molecular QTL-based analyses we used \nwere designed to capture GWAS-relevant regulatory variants for the features in cis (typically \nwithin a window of < 1 Mb from the features), yet GWAS signals could be related to trans-\nQTLs, linking associations to distant  candidate genes. However, generation of trans-eQTL \nand trans-pQTL catalogues have been historically difficult due to multiple problems related to \nsample size and control of confounders 81, although there has been recent progress in large-\nscale brain trans-eQTL catalogues 82, opening up new analysis opportunities in the future for \nrare cases where a GWAS signal is acting through a trans-eQTL signal. Thirdly, despite the \nrecent progress in availability of brain ct-eQTL catalogues 24,83, no such cell-type-specific \npQTL catalogues are available to our knowledge ; but the latest advances in the field for \nsingle-nucleus proteomics84 may lead to brain ct-pQTL datasets in the foreseeable future. \nIn conclusion, our results are compatible with the leading hypotheses for the three known \ngenetic risk factors for sCJD, with there being robust evidence for increases in STX6 \nexpression driving disease risk, but not for PRNP and GAL3ST1, which are thought to be \ndriven by missense SNPs. Furthermore, this functionally-informed analysis of sCJD GWAS \nsummary statistics provides additional suggestive evidence and connections between other \nprioritized genes, including PDIA4, BMERB1 and MANF, and generally, for a role of glial \ncells and the UPR in sCJD aetiology ( Fig. 4 ). Future functional studies may confirm the \ntarget prioritized sCJD risk genes and risk-associated molecular mechanisms highlighted in \nour study, leading to better understanding of  the disease mechanisms and consequently \nproviding new therapeutic opportunities for sCJD, with potential relevance to other \nneurodegenerative diseases. \nData availability and URLs \nThe sCJD GWAS 8 summary statistics is available at  the European Bioinformatics Institute \nGWAS Catalog portal (https://www.ebi.ac.uk/gwas/) under accession no. GCST90001389. \nSEA-AD32 brain single nucleus gene expression matrices \n(https://registry.opendata.aws/allen-sea-ad-atlas/) \nFull e/pQTL-GWAS coloc and TWAS/PWAS results from this study are available at \nhttps://doi.org/10.5281/zenodo.12507355, while significant-only results are shown in \nSupplementary Tables 5-9. \nMolecular eQTL and pQTL related datasets used in this study are publicly available (see \nalso Supplementary Table 2): \neQTLs and TWAS reference panels in AD-relevant bulk brain regions from AMP-AD cohorts, \nas analyzed by Bellenguez et al.15: (https://doi.org/10.5281/zenodo.5745927); \nGTEx v823 eQTL catalogues (https://www.gtexportal.org/); \nGTEx v8 MASHR29,30 expression prediction models for TWAS \n(https://predictdb.org/post/2021/07/21/gtex-v8-models-on-eqtl-and-sqtl/#mashr-based-\nmodels); \nBryois et al.24 ct-eQTL catalogues (https://doi.org/10.5281/zenodo.5543734); \nMiGA eQTL catalogues (https://doi.org/10.5281/zenodo.4118605 and \nhttps://doi.org/10.5281/zenodo.4118676); \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n16 \n \nWingo et al.16 v2 pQTL catalogues & PWAS reference panels \n(https://www.synapse.org/#!Synapse:syn23627957). \n \nFunding and Acknowledgements \nThe work was funded by the Medical Research Council (UK). SM and JC are National \nInstitute for Health Research (NIHR) Senior In vestigators (JC is emeritus). FK receives a \npostdoctoral fellowship (BOF 49758) from the University of Antwerp Research Fund. \nThe data available in the AD Knowledge Portal would not be possible without the \nparticipation of research volunteers and the contribution of data by collaborating \nresearchers. The results published here are in  whole or in part based on data obtained from \nthe AD Knowledge Portal (https://adknowledgeportal.org\n). Data generation was supported by \nthe following NIH grants: P30AG10161, P30AG72975, R01AG15819, R01AG17917, \nR01AG036836, U01AG46152, U01AG61356, U01AG046139, P50 AG016574, R01 \nAG032990, U01AG046139, R01AG018023, U01AG006576, U01AG006786, R01AG025711, \nR01AG017216, R01AG003949, R01NS080820, U24NS072026, P30AG19610, \nU01AG046170, RF1AG057440, and U24AG061340, and the Cure PSP, Mayo and Michael J \nFox foundations, Arizona Department of Health Services and the Arizona Biomedical \nResearch Commission. We thank the participant s of the Religious Order Study and Memory \nand Aging projects for the generous donation, the Sun Health Research Institute Brain and \nBody Donation Program, the Mayo Clinic Brain Bank, and the Mount Sinai/JJ Peters VA \nMedical Center NIH Brain and Tissue Repository. Data and analysis contributing \ninvestigators include Nilüfer Ertekin-Taner, Steven Younkin (Mayo Clinic, Jacksonville, FL), \nTodd Golde (University of Florida), Nathan Pr ice (Institute for Systems Biology), David \nBennett, Christopher Gaiteri (Rush University), Philip De Jager (Columbia University), Bin \nZhang, Eric Schadt, Michelle Ehrlich, Vahram Haroutunian, Sam Gandy (Icahn School of \nMedicine at Mount Sinai), Koichi Iijima (National Center for Geriatrics and Gerontology, \nJapan), Scott Noggle (New York Stem Cell Foundation), Lara Mangravite (Sage \nBionetworks). Study data were generated from postmortem brain tissue obtained from the \nUniversity of Washington BioRepository and Integrated Neuropathology (BRaIN) laboratory \nand Precision Neuropathology Core, which is supported by the NIH grants for the UW \nAlzheimer's Disease Research Center (P50AG005136 and P30AG066509) and the Adult \nChanges in Thought Study (U01AG006781 and U19A G066567). This study is supported by \nNIA grant U19AG060909. \nCompeting Interests \nNo competing interests.\n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n17 \n \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n18 \n \nFigures \nFigure 1. Gene prioritization results for sCJD GWAS. A visual summary of weighted evidence category scores \nfor each prioritized risk gene, together with brain cell-type-specific gene expression proportions. The figure shows \na total of 32 prioritized risk genes (17 tier 1 and 15 tier 2). The leftmost squares indicate the locus indexes where \n“G” is used for the genome-wide significant loci, “S” for the subthreshold loci, and “O” for the remaining other loci. \nThe types of evidence for each category are coloured according to the three different domains to which they \nbelonged. Weighted scores for each evidence category are rescaled to a 0–100 scale based on the maximum \nscore a candidate gene can obtain from a category (see Supplementary Table 3). The darker colours represent \nhigher scores in categories or higher average gene expression proportions in the 7 major brain cell types, while \ntier 1 prioritized genes are displayed in dark green and tier 2 prioritized genes are displayed in light green. Only \ntier 1 and tier 2 genes are shown for each locus, and all candidate genes considered and scored can be found in \nSupplementary Table 4 . CADD (v1.7) PHRED scores for index variants are labelled in white within the \nrespective squares in variant annotation domain. eQTL, expression QTL; pQTL, protein-expression QTL; ct-\neQTL, cell-type-specific eQTL; coloc, colocalization; TWAS, transcriptome-wide association study; PWAS, \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n19 \n \nproteome-wide association study; OPCs, oligodendrocyte precursor cells; COPs, committed oligodendrocyte \nprecursors. \n \nFigure 2. Regulation of STX6 brain gene and protein expression by the sCJD-risk-colocalized eQTLs and \npQTLs within the STX6  locus. The regional plots of (i) sCJD GWAS association signal ( n = 17,679), (ii) STX6 \nbrain pQTL signal in DLPFC (ROSMAP DLPFC pQTL catalogue, n = 376), (iii) STX6  bulk brain eQTL signal in \nDLPFC (ROSMAP DLPFC eQTL catalogue, n = 560), and (iv) STX6  ct-eQTL signal in oligodendrocytes (Bryois \net al.24 ct-eQTL catalogue, n = 192) are shown for 100 kb extended genomic coordinates of the STX6 locus index \nvariant rs3747957 (chr1:180884717-181084717). Boxes in each panel shows QTL-GWAS coloc PP4 values \nbetween the molecular QTL signal and the GWAS signal for all tested variants (see Supplementary Tables 5-6). \nThe index variant is shown in purple, and LD r 2 values (calculated within 1 KG non-Finnish European samples [n  \n= 404] with respect to the index variant) are indicated on a color scale, and variants that are not available in the \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n20 \n \nLD reference panel are shown in grey. y axis,/i1−/i1log10 GWAS or QTL P; x axis, GRCh38 genomic position on \nchromosome 1 together with the annotation for the genomic positions of the protein-coding genes in the locus. \neQTL, expression QTL; pQTL, protein-expression QTL; ct-eQTL, cell-type-specific eQTL; coloc, colocalization. \n \nFigure 3. sCJD brain proteome-wide association study results. sCJD brain proteome-wide association study  \n(PWAS) results are shown proteome-wide for both of the PWAS reference panels with two mirrored Manhattan \nplots on the x-axis; the upper side of the plot displays the results for ROSMAP DLPFC PWAS, while the lower \nside of the plot displays the results for Banner DLPFC PWAS. Proteome-wide significance thresholds (0.05 \ndivided by number of tested associations; see Supplementary Table 2 ) for both analyses are indicated with red \ndashed lines and suggestive significance thresholds (1 divided by number of tested associations) with a blue \ndotted line, and all the genes whose protein products are passing these thresholds are labeled, and colored \nbased on their significance (red: proteome-wide significant, blue: suggestive significant). The directionality of Z-\nscores of each PWAS association are represented with up-pointing triangles (positive Z-score) and down-pointing \ntriangles (negative Z-score). y axis,/i1−/i1log10 PWAS P; x axis, GRCh38 chromosomal positions. DLPFC, \ndorsolateral prefrontal cortex. \n \n \n \n \n \n \n \n \n \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n \n \n \nFigure 4. Speculative Model of the Cell Types and the Potential Relationship Between Prioritized Risk \nGenes and their Mechanisms . A common amino acid variant (p.Val29Met) in the GAL3ST1 gene, encoding \ngalactosylceramide sulfotransferase, increases sulfatide production predominantly in oligodendrocytes, \nconferring increased risk of sCJD ( 1). Sulfatide may act as a cofactor in PrP C conversion or prion propagation, \nwhich may be intercepted by the astrocyte-secreted factor, MANF, which binds to sulfatide extracellularly ( 2). \nSulfatide may additionally promote the cellular uptake of MANF allowing it to work in concert with PDIA4 to \nprotect against the adverse effects of ER stress and the sustained unfolded protein response characteristic of \nprion infection ( 3). Increased syntaxin-6 expression predominantly in oligodendrocytes may be altering the \ntrafficking of either PrP C or prions with implications on prion formation, propagation and/or clearance ( 4). Figure \ncreated on Biorender. \n  \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n22 \n \nReferences \n1. Manka SW, Wenborn A, Betts J, et al. A structural basis for prion strain diversity. Nat \nChem Biol. 2023;19(5):607-613. \n2. Collinge J. Mammalian prions and their wider relevance in neurodegenerative \ndiseases. Nature. 2016;539(7628):217-226. \n3. Ladogana A, Puopolo M, Croes EA, et al. Mo rtality from Creutzfeldt-Jakob disease \nand related disorders in Europe, Australia, and Canada. Neurology. 2005;64(9):1586-\n1591. \n4. NCJDRSU. Annual Report. 2021; \nhttps://www.cjd.ed.ac.uk/sites/default/files/report29.pdf. \n5. Collinge J. Prion diseases of humans and animals: their causes and molecular basis. \nAnnu Rev Neurosci. 2001;24:519-550. \n6. Genevieve MJAK, Handan W, Marion S, et al. Intensity of human prion disease \nsurveillance predicts observed disease incidence. Journal of Neurology, \nNeurosurgery &amp;amp; Psychiatry. 2013;84(12):1372. \n7. Mead S, Khalili-Shirazi A, Potter C, et al. Prion protein monoclonal antibody \n(PRN100) therapy for Creutzfeldt-Jakob disease: evaluation of a first-in-human \ntreatment programme. The Lancet Neurology. 2022;21(4):342-354. \n8. Jones E, Hummerich H, Vire E, et al. Identification of novel risk loci and causal \ninsights for sporadic Creutzfeldt-Jakob disease: a genome-wide association study. \nLancet Neurol. 2020;19(10):840-848. \n9. Bock JB, Klumperman J, Davanger S, Scheller RH. Syntaxin 6 functions in trans-\nGolgi network vesicle trafficking. Molecular biology of the cell. 1997;8(7):1261-1271. \n10. Bock JB, Lin RC, Scheller RH. A new syntaxin family member implicated in targeting \nof intracellular transport vesicles. The Journal of biological chemistry. \n1996;271(30):17961-17965. \n11. Takahashi T, Suzuki T. Role of sulfatide in normal and pathological cells and tissues. \nJ Lipid Res. 2012;53(8):1437-1450. \n12. Gusev A, Mancuso N, Won H, et al. Transcriptome-wide association study of \nschizophrenia and chromatin activity yields mechanistic disease insights. Nat Genet. \n2018;50(4):538-548. \n13. Pain O, Jones A, Al Khleifat A, et al. Harnessing Transcriptomic Signals for \nAmyotrophic Lateral Sclerosis to Identify Novel Drugs and Enhance Risk Prediction. \nmedRxiv. 2023. \n14. Dall'Aglio L, Lewis CM, Pain O. Delineating the Genetic Component of Gene \nExpression in Major Depression. Biol Psychiatry. 2021;89(6):627-636. \n15. Bellenguez C, Kucukali F, Jansen IE, et al. New insights into the genetic etiology of \nAlzheimer's disease and related dementias. Nature Genetics. 2022;54(4):412-+. \n16. Wingo AP, Liu Y, Gerasimov ES, et al. Integrating human brain proteomes with \ngenome-wide association data implicates new proteins in Alzheimer’s disease \npathogenesis. Nature Genetics. 2021;53(2):143-146. \n17. Wingo TS, Gerasimov ES, Liu Y, et al. Integrating human brain proteomes with \ngenome-wide association data implicates novel proteins in post-traumatic stress \ndisorder. Mol Psychiatry. 2022;27(7):3075-3084. \n18. Wingo TS, Liu Y, Gerasimov ES, et al. Shared mechanisms across the major \npsychiatric and neurodegenerative diseases. Nat Commun. 2022;13(1):4314. \n19. Allen M, Carrasquillo MM, Funk C, et al. Human whole genome genotype and \ntranscriptome data for Alzheimer’s and other neurodegenerative diseases. Scientific \nData. 2016;3(1):160089. \n20. Bennett DA, Buchman AS, Boyle PA, Barnes LL, Wilson RS, Schneider JA. Religious \nOrders Study and Rush Memory and Aging Project. J Alzheimers Dis. \n2018;64(s1):S161-s189. \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n23 \n \n21. Mostafavi S, Gaiteri C, Sullivan SE, et al. A molecular network of the aging human \nbrain provides insights into the pathology and cognitive decline of Alzheimer’s \ndisease. Nature Neuroscience. 2018;21(6):811-819. \n22. Wang M, Beckmann ND, Roussos P, et al. The Mount Sinai cohort of large-scale \ngenomic, transcriptomic and proteomic data in Alzheimer's disease. Scientific Data. \n2018;5(1):180185. \n23. The GC, Aguet F, Anand S, et al. The GTEx Consortium atlas of genetic regulatory \neffects across human tissues. Science (New York, NY). 2020;369(6509):1318-1330. \n24. Bryois J, Calini D, Macnair W, et al. Cell-type-specific cis-eQTLs in eight human brain \ncell types identify novel risk genes for psychiatric and neurological disorders. Nature \nNeuroscience. 2022;25(8):1104-1112. \n25. Young AMH, Kumasaka N, Calvert F, et al. A map of transcriptional heterogeneity \nand regulatory variation in human microglia. Nature Genetics. 2021;53(6):861-868. \n26. Lopes KP, Snijders GJL, Humphrey J, et al. Genetic analysis of the human microglial \ntranscriptome across brain regions, aging and disease pathologies. Nat Genet. \n2022;54(1):4-17. \n27. Giambartolomei C, Vukcevic D, Schadt EE, et al. Bayesian test for colocalisation \nbetween pairs of genetic association studies using summary statistics. PLoS Genet. \n2014;10(5):e1004383. \n28. Gusev A, Ko A, Shi H, et al. Integrat ive approaches for large-scale transcriptome-\nwide association studies. Nature Genetics. 2016;48(3):245-252. \n29. Barbeira AN, Dickinson SP, Bonazzola  R, et al. Exploring the phenotypic \nconsequences of tissue specific gene expression variation inferred from GWAS \nsummary statistics. Nature Communications. 2018;9(1):1825. \n30. Gamazon ER, Wheeler HE, Shah KP, et al. A gene-based association method for \nmapping traits using reference transcriptome data. Nature Genetics. \n2015;47(9):1091-1098. \n31. Mancuso N, Freund MK, Johnson R, et al. Probabilistic fine-mapping of \ntranscriptome-wide association studies. Nat Genet. 2019;51(4):675-682. \n32. Gabitto MI, Travaglini KJ, Rachleff VM, et al. Integrated multimodal cell atlas of \nAlzheimer’s disease. bioRxiv. 2024:2023.2005.2008.539485. \n33. Szklarczyk D, Kirsch R, Koutrouli M,  et al. The STRING database in 2023: protein-\nprotein association networks and functional enrichment analyses for any sequenced \ngenome of interest. Nucleic Acids Res. 2023;51(D1):D638-d646. \n34. Palmer MS, Dryden AJ, Hughes JT, Collinge J. Homozygous prion protein genotype \npredisposes to sporadic Creutzfeldt–Jakob disease. Nature. 1991;352(6333):340-\n342. \n35. Collinge J, Palmer MS, Dryden AJ. Genet ic predisposition to iatrogenic Creutzfeldt-\nJakob disease. The Lancet. 1991;337(8755):1441-1442. \n36. Mead S, Poulter M, Beck J, et al. Inherited prion disease with six octapeptide repeat \ninsertional mutation--molecular analysis of phenotypic heterogeneity. Brain. \n2006;129(Pt 9):2297-2317. \n37. Collinge J, Whitfield J, McKintosh E, et  al. Kuru in the 21st century--an acquired \nhuman prion disease with very long incubation periods. Lancet. \n2006;367(9528):2068-2074. \n38. Cadby G, Giles C, Melton PE, et al. Comprehensive genetic analysis of the human \nlipidome identifies loci associated with lipid homeostasis with links to coronary artery \ndisease. Nat Commun. 2022;13(1):3124. \n39. Bai M, Vozdek R, Hnizda A, et al. Conserved roles of C. elegans and human MANFs \nin sulfatide binding and cytoprotection. Nat Commun. 2018;9(1):897. \n40. Hoglinger GU, Melhem NM, Dickson DW, et al. Identification of common variants \ninfluencing risk of the tauopathy progressive supranuclear palsy. Nat Genet. \n2011;43(7):699-705. \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n24 \n \n41. Chen JA, Chen Z, Won H, et al. Joint genome-wide association study of progressive \nsupranuclear palsy identifies novel susceptibility loci and genetic correlation to \nneurodegenerative diseases. Mol Neurodegener. 2018;13(1):41. \n42. Chen Z, Chen JA, Shatunov A, et al. Genome-wide survey of copy number variants \nfinds MAPT duplications in progressive supranuclear palsy. Mov Disord. \n2019;34(7):1049-1059. \n43. Ferrari R, Hernandez DG, Nalls MA, et al. Frontotemporal dementia and its subtypes: \na genome-wide association study. Lancet Neurol. 2014;13(7):686-699. \n44. Sanchez-Contreras MY, Kouri N, Cook CN, et al. Replication of progressive \nsupranuclear palsy genome-wide association study identifies SLCO1A2 and DUSP10 \nas new susceptibility loci. Mol Neurodegener. 2018;13(1):37. \n45. Holger H, Helen S, Tracy C, et al. Genome wide association study of clinical duration \nand age at onset of sporadic CJD. medRxiv. 2023:2023.2010.2017.23297050. \n46. Jones E, Hill E, Linehan J, et al. Ch aracterisation and prion transmission study in \nmice with genetic reduction of sporadic Creutzfeldt-Jakob Disease risk gene Stx6. \n2023:2023.2001. 2010.523281. \n47. Fodder K, Murthy M, Rizzu P, et al. Brain DNA methylomic analysis of frontotemporal \nlobar degeneration reveals OTUD4 in shared dysregulated signatures across \npathological subtypes. Acta neuropathologica. 2023;146(1):77-95. \n48. Wingo AP, Liu Y, Gerasimov ES, et al. Integrating human brain proteomes with \ngenome-wide association data implicates new proteins in Alzheimer's disease \npathogenesis. Nat Genet. 2021;53(2):143-146. \n49. Palmer MS, Dryden AJ, Hughes JT, J C. Homozygous prion protein genotype \npredisposes to sporadic Creutzfeldt-Jakob disease. Nature. 1991;352:340-342. \n50. Collinge J, Palmer MS, Dryden AJ. Genet ic predisposition to iatrogenic Creutzfeldt-\nJakob disease. Lancet. 1991;337:1441-1442. \n51. Mok T, Jaunmuktane Z, Joiner S, et al. Variant Creutzfeldt-Jakob Disease in a \nPatient with Heterozygosity at PRNP Codon 129. N Engl J Med. 2017;376(3):292-\n294. \n52. Wadsworth JD, Asante EA, Collinge J. Review: contribution of transgenic models to \nunderstanding human prion disease. Neuropathology and applied neurobiology. \n2010;36(7):576-597. \n53. Harshfield EL, Fauman EB, Stacey D, et al. Genome-wide analysis of blood lipid \nmetabolites in over 5000 South Asians reveals biological insights at cardiometabolic \ndisease loci. BMC Med. 2021;19(1):232. \n54. Ferrari DM, Söling HD. The protein disulphide-isomerase family: unravelling a string \nof folds. The Biochemical journal. 1999;339 ( Pt 1)(Pt 1):1-10. \n55. Wang SB, Shi Q, Xu Y, et al. Protein Disulfide Isomerase Regulates Endoplasmic \nReticulum Stress and the Apoptotic Process during Prion Infection and PrP Mutant-\nInduced Cytotoxicity. PLoS ONE. 2012;7(6):e38221. \n56. Andreu CI, Woehlbier U, Torres M, Hetz C. Protein Disulfide Isomerases in \nNeurodegeneration: from disease mechanisms to biomedical applications. FEBS \nLett. 2012. \n57. Moore RA, Sturdevant DE, Chesebro B, Priola SA. Proteomics analysis of amyloid \nand nonamyloid prion disease phenotypes reveals both common and divergent \nmechanisms of neuropathogenesis. J Proteome Res. 2014;13(11):4620-4634. \n58. Ron D, Walter P. Signal integration in the endoplasmic reticulum unfolded protein \nresponse. Nature reviews Molecular cell biology. 2007;8(7):519-529. \n59. Remondelli P, Renna M. The Endoplasmic Reticulum Unfolded Protein Response in \nNeurodegenerative Disorders and Its Potential Therapeutic Significance. Frontiers in \nMolecular Neuroscience. 2017;10. \n60. Hoozemans JJ, van Haastert ES, Eikelenboom P, de Vos RA, Rozemuller JM, \nScheper W. Activation of the unfolded protein response in Parkinson's disease. \nBiochemical and biophysical research communications. 2007;354(3):707-711. \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n25 \n \n61. Hoozemans JJ, Veerhuis R, Van Haastert ES, et al. The unfolded protein response is \nactivated in Alzheimer's disease. Acta neuropathologica. 2005;110(2):165-172. \n62. Harding HP, Zhang Y, Ron D. Protein translation and folding are coupled by an \nendoplasmic-reticulum-resident kinase. Nature. 1999;397(6716):271-274. \n63. Moreno JA, Radford H, Peretti D, et al. Sustained translational repression by eIF2 \nalpha-P mediates prion neurodegeneration. Nature. 2012;485(7399):507-511. \n64. Lakkaraju AKK, Frontzek K, Lemes E, et al. Loss of PIKfyve drives the spongiform \ndegeneration in prion diseases. EMBO Mol Med. 2021;13(9):e14714. \n65. Petrova P, Raibekas A, Pevsner J, et  al. MANF: a new mesencephalic, astrocyte-\nderived neurotrophic factor with selectivity for dopaminergic neurons. J Mol Neurosci. \n2003;20(2):173-188. \n66. Palgi M, Lindström R, Peränen J, Piepponen TP, Saarma M, Heino TI. Evidence that \nDmMANF is an invertebrate neurotrophic factor supporting dopaminergic neurons. \nProceedings of the National Academy of Sciences of the United States of America. \n2009;106(7):2429-2434. \n67. Kim Y, Park SJ, Chen YM. Mesencephalic astrocyte-derived neurotrophic factor \n(MANF), a new player in endoplasmic reticulum diseases: structure, biology, and \ntherapeutic roles. Transl Res. 2017;188:1-9. \n68. Lindström R, Lindholm P, Kallijärvi J, Palgi M, Saarma M, Heino TI. Exploring the \nConserved Role of MANF in the Unfolded Protein Response in Drosophila \nmelanogaster. PloS one. 2016;11(3):e0151550. \n69. Henderson MJ, Richie CT, Airavaara M, Wang Y, Harvey BK. Mesencephalic \nastrocyte-derived neurotrophic factor (MANF) secretion and cell surface binding are \nmodulated by KDEL receptors. J Biol Chem. 2013;288(6):4209-4225. \n70. Glembotski CC, Thuerauf DJ, Huang C, Vekich JA, Gottlieb RA, Doroudgar S. \nMesencephalic astrocyte-derived neurotrophic factor protects the heart from ischemic \ndamage and is selectively secreted upon sarco/endoplasmic reticulum calcium \ndepletion. J Biol Chem. 2012;287(31):25893-25904. \n71. Palgi M, Greco D, Lindstrom R, Auvinen P, Heino TI. Gene expression analysis of \nDrosophilaa Manf mutants reveals perturbations in membrane traffic and major \nmetabolic changes. BMC Genomics. 2012;13:134. \n72. Pakarinen E, Danilova T, Võikar V, et al. MANF Ablation Causes Prolonged \nActivation of the UPR without Neurodegeneration in the Mouse Midbrain Dopamine \nSystem. eNeuro. 2020;7(1). \n73. Smith HL, Freeman OJ, Butcher AJ, et al. Astrocyte Unfolded Protein Response \nInduces a Specific Reactivity State that Causes Non-Cell-Autonomous Neuronal \nDegeneration. Neuron. 2020;105(5):855-866 e855. \n74. Nuvolone M, Paolucci M, Sorce S, et al. Prion pathogenesis is unaltered in the \nabsence of SIRP\nα -mediated \"don't-eat-me\" signaling. PloS one. \n2017;12(5):e0177876. \n75. Prinz M, Montrasio F, Furukawa H, et al. Intrinsic resistance of oligodendrocytes to \nprion infection. J Neurosci. 2004;24(26):5974-5981. \n76. Andres Benito P, Dominguez Gonzalez M, Ferrer I. Altered gene transcription linked \nto astrocytes and oligodendrocytes in frontal cortex in Creutzfeldt-Jakob disease. \nPrion. 2018;12(3-4):216-225. \n77. Andrés-Benito P, Carmona M, Douet JY, Cassard H, Andreoletti O, Ferrer I. \nDifferential astrocyte and oligodendrocyte vulnerability in murine Creutzfeldt-Jakob \ndisease. Prion. 2021;15(1):112-120. \n78. Slota JA, Sajesh BV, Frost KF, Medina SJ, Booth SA. Dysregulation of \nneuroprotective astrocytes, a spectrum of microglial activation states, and altered \nhippocampal neurogenesis are revealed by single-cell RNA sequencing in prion \ndisease. Acta Neuropathologica Communications. 2022;10(1):161. \n79. Liu Y, Guo J, Matoga M, Korotkova M, Jakobsson P-J, Aguzzi A. NG2 glia protect \nagainst prion neurotoxicity by inhibiting microglia-to-neuron prostaglandin E2 \nsignaling. Nature Neuroscience. 2024. \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint \n\n26 \n \n80. Gelpi E, Klotz S, Vidal-Robau N, et al. Histotype-Dependent Oligodendroglial PrP \nPathology in Sporadic CJD: A Frequent Feature of the M2C \"Strain\". Viruses. \n2021;13(9). \n81. Lappalainen T, Li YI, Ramachandran S, Gusev A. Genetic and molecular architecture \nof complex traits. Cell. 2024;187(5):1059-1075. \n82. de Klein N, Tsai EA, Vochteloo M, et al. Brain expression quantitative trait locus and \nnetwork analyses reveal downstream effects and putative drivers for brain-related \ndiseases. Nature Genetics. 2023;55(3):377-388. \n83. Fujita M, Gao Z, Zeng L, et al. Cell subtype-specific effects of genetic variation in the \nAlzheimer’s disease brain. Nature Genetics. 2024;56(4):605-614. \n84. Derks J, Jonson T, Leduc A, et al. Single -nucleus proteomics identifies regulators of \nprotein transport. bioRxiv. 2024:2024.2006.2017.599449. \n \n . CC-BY 4.0 International licenseIt is made available under a \n is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)\nThe copyright holder for this preprint this version posted July 21, 2024. ; https://doi.org/10.1101/2024.07.19.24310476doi: medRxiv preprint","source_license":"CC-BY-4.0","license_restricted":false}