The list of rare and common genetic variants that sway a person’s risk of Alzheimer’s disease just got substantially longer, as researchers expanded both the breadth and depth of genomic studies aimed at uncovering the etiology of the disease. At this week’s virtual Alzheimer’s Association International Conference, researchers described how harmonizing whole-exome-sequencing data from two massive consortiums in the United States and Europe unearthed rare variants that boost the risk of AD. In addition, geneticists extracted new information from existing genome-wide association studies by flexing their computational muscles of imputation, a feat that uncovered dozens of new AD risk variants, both common and rare. Other geneticists pulled out genetic variants that enhance cognitive resilience against AD. At this point, these lists look like little more than acronym salad, but scientists are quickly trying to find out what these genes are doing normally and in disease.
- Massive exome-sequencing study found new AD risk variants in ATP8B4.
- Variants in its cousin, ATP8B1, seem tied to resilience.
- Imputation of massive GWAS brought up dozens of new AD variants, including ATP8B4, a lipid transporter.
The AD field’s current genome-wide association studies (GWAS) have size on their side, but that’s not enough. They still lack the fine detail needed to capture the rarest variants. Sequencing studies are better suited to fish out these needles in a haystack, but most are limited by sample size. Researchers in Europe merged their sequencing data with data generated by researchers from the United States to address this issue. In the largest whole-exome sequencing study to hunt for AD variants to date, researchers led by Henne Holstege of the VU University in Amsterdam, Gaël Nicolas of Normandie University in Rouen, France, and Jean-Charles Lambert of the Institut Pasteur de Lille in France integrated exome sequences from the Alzheimer Disease European Sequencing (ADES) consortium with data from the American Alzheimer’s Disease Sequencing Project (ADSP), making for a total of 25,982 samples. The findings were posted in a manuscript in medRxiv on July 24, and presented at AAIC by co-first authors Holstege and Marc Hulsman, also at VU University Medical Center in Amsterdam.
Merging these massive datasets was quite a computational undertaking in itself, Holstege told Alzforum. In an effort led by Hulsman, the scientists combed through the data to weed out myriad batch effects from different sources of data. After extensive quality control, the dataset included exome sequences from 12,625 AD cases and 8,693 controls. While the harmonized dataset was generated and stored on the Dutch supercomputer facility in Amsterdam, the subsequent analyses were the result of a close collaboration by the Amsterdam group with the Lambert and Nicolas’ groups in France.
The researchers set out to discover rare genetic variants associated with AD risk. Specifically, they searched for individual variants that occurred in fewer than 1 percent of the samples, but lay within genes in which at least 10 people carried a variant. Out of a total of more than 7 million variants, the researchers detected more than 400,000 that were predicted to either wipe out expression of a gene or mess with its function. They separated them into four categories of badness, ranging from outright loss of function (LOF) to missense mutations with three degrees of predicted damage.
In the end, rare variants in 11 genes appeared linked to AD risk. Three of the 11—SORL1, TREM2, and ABCA7— had come up in previous studies. The others, in order of significance, were ATP8B4, ADAM10, ABCA1, ORC6, CBX3, PRSS3, B3GNT4, and SRC. Applying more stringent statistical techniques weakened some of these finds, but ATP8B4, along with the three known variant genes, remained significantly associated with AD risk. Variants in two of the suggestively associated genes—CBX3 and PRSS3—were less common in cases than controls, casting them as potential protective variants.
The average age at onset across all cases in the dataset was 73. For most of the variants tied to increased AD risk, carriers were younger at onset than noncarriers. Age at onset also decreased with increasing deleteriousness of a given variant; for example, symptoms emerged at an average of 60 years in carriers of LOF variants in SORL1, and not until 68 in carriers of milder missense mutations in the same gene. Variants in ATP8B4 did not appear to associate with age at onset.
A large proportion of the disease-associated signal from SORL1, ADAM10, ABCA1, ORC6, B3GNT4, and SRC came from singletons, that is, variants that were only detected in a single person. For TREM2, ABCA7, ATP8B4, CBX3, and PRSS3, the majority of the signal was derived from variants that were identified in multiple people, but were still very rare.
The findings presented in the medRxiv preprint need to be replicated in an independent set of sequence data from AD patients and non-demented individuals, Holstege noted. This is currently ongoing.
Most of the variants had some functional ties to AD pathophysiology, including APP processing and microglial function. The new risk gene ATP8B4, or ATPase Phospholipid Transporting 8B4, encodes a cation transport ATPase that ferries phospholipids across the cell membrane. In the brain, ATP8B4 is predominantly expressed in microglia, and rare variants in the gene have been associated with systemic sclerosis, an autoimmune disorder (Gao et al., 2016). The present study spotted 74 deleterious ATP8B4 variants in 767 people. Four percent of AD cases, compared with 2.5 percent of controls, carried one of the variants. Three missense variants drove the majority of the effect.
Holstege’s study was not the only one to present an ATP8B gene at AAIC. Another member of the family, ATP8B1, popped up in a genetic study of cognitive resilience. Timothy Hohman of Vanderbilt University, Nashville, Tennessee, reported results of a GWAS for variants associated with cognitive resilience to Alzheimer’s pathology. Essentially, Hohman looked for variants tied to superior cognitive performance in the face of a given amount of Aβ plaques. To assess resilience, Hohman integrated amyloid PET, postmortem neuropathology, and longitudinal cognitive performance data from more than 5,000 participants across four cohorts: Adult Changes in Thought (ACT), the Religious Orders Study and Memory and Aging Project (ROSMAP), the Alzheimer’s Disease Neuroimaging Initiative (ADNI), and the Anti-Amyloid Treatment in Asymptomatic Alzheimer’s (A4) study. For each participant, he charted Aβ burden against cognitive decline, developing a continuous variable of resilience. It ranged from those who performed worse than expected based on their pathological burden, to those who were sharper than expected. Every participant was genotyped.
Only one locus, by the ATP8B1 gene, associated with cognitive resilience at the genome-wide level. Three variants were detected near this gene. Further data will be needed to confirm that ATP8B1 is the gene causing the association, but Hohman said that variant carriers had reduced methylation of the gene, suggesting the variants may have increased its expression.
Hohman told Alzforum that the gene is highly homologous with ATP8B4 detected in the Holstege study, but its expression pattern is different. In the brain, ATP8B1 is expressed predominantly by neurons and by endothelial cells, which form the blood-brain barrier.
In his talk, Hohman focused on the liver, however, where ATP8B1 is known to play a role in regulating the homeostasis of bile acids. Variants in the gene have been tied to liver diseases. Interestingly, Hohman noted that data from ADNI indicate that the composition of bile acids change in AD. Hohman said future studies will need to unravel how ATP8B family genes influence AD risk and cognitive resilience. He noted that the ATP8B ATPases interact with ABC transporters, such as ABCA7, which are known to play a role in lipid metabolism and have been tied to AD risk.
Hohman also reported that the genetic architecture of cognitive resilience is distinct from that of Alzheimer’s disease. In other words, resilience is not simply the opposite of susceptibility. For example, Hohman reported that the genes associated with resilience in his study overlapped significantly with those tied to cognitive performance and education, and modestly with vascular and psychiatric phenotypes. Genetic pathways underlying AD risk were distinct from those tied to resilience.
Growing Larger and Denser, Monster GWASs Spew More Variants
Sequencing data may be the gold standard for pinpointing rare variants, but genome-wide association studies have the edge on size. They are massive, and growing larger still as consortiums pool their data.
Another way to get more variant bang per GWAS buck is to include more single-nucleotide polymorphisms in the analysis. To do this, geneticists have honed imputation methods. Essentially, they use haplotype reference catalogues to infer a genotype at one SNP position based on the genotype at another one nearby. Such catalogues have burgeoned in recent years, enabling the imputation of more SNPs than ever before, said Adam Naj of the University of Pennsylvania in his AAIC presentation. Naj used data from the Haplotype Reference Consortium, which includes nearly 65,000 haplotypes and 40 million SNPs, to impute data from the International Genomics of Alzheimer’s Disease Project (IGAP). Based on more than 25,000 AD cases and 40,000 controls, this GWAS had previously helped identify 25 risk loci (Mar 2019 news on Kunkle et al., 2019).
With deeper imputation, Naj pulled out 14 new common variants plus eight rare ones. Of the 14, only two—HBEGF and BZRAP1—reached genome-wide significance, while a dozen others tickled the significance threshold. They are EIF4G3, EPHA5, ELL, LOC102723838, MTSS2/1L, CLNK, LOC728084/LINC02458, SPPL2A, ANKRD33, LUZP2, GUCY2EP/TSKU, and MC4R. EPHA5 belongs to the ephrin receptor family, of which other members have been implicated in AD. SPPL2A is involved in innate immunity. MC4R encodes a melanocortin receptor that has been tied to obesity, and MC4R agonists reportedly boost neurogenesis and cognition in animal models of AD.
The eight new rare variants were in the genes, NOL10/ODC, LINC00052, LOC101927787, CCDC102B, NCK2, RORA, BSGALT6, and EPHA3, yet another ephrin receptor. Naj said that NCK2 reportedly interacts with presenilin. RORA encodes a nuclear receptor that regulates BDNF and insulin peptides in hippocampus. Finally, B4GALT6 plays a role in neural differentiation.
The imputed GWAS data need replication. Another consortium is already on it. At AAIC, Céline Bellenguez from Institut Pasteur de Lille in France reported preliminary results from a GWAS including data from the European Alzheimer’s Data Bank (EADB). When combined with data from seven smaller consortia, this GWAS analyzed more than 36,000 AD cases and 63,000 controls. Samples from some of the smaller consortia were also included in the IGAP GWAS; however, the majority of samples for IGAP came from the Alzheimer’s Disease Genetics Consortium, which were not included in the EADB GWAS. Bellenguez imputed this GWAS with the Trans-Omics for Precision Medicine (TOPMed) reference panel, comprising more than 125,000 haplotypes.
Bellenguez reported 44 loci tied to AD risk. Fourteen were new: COX7C-RASA1, RASGEF1C, UMAD1, CTSB, ZFPM2, TRIB1, C1S, TBX6, FAM157C, LILRA2, LILRB5-LILRB2, IDUA, GRN and—lo and behold—ATP8B4, the gene found in the WES study. LILRB genes are expressed in microglia and bind Aβ peptides. IDUA and GRN have been implicated in Parkinson’s disease and frontotemporal dementia, respectively.
Among the genome-wide significant hits, none popped up in both the EADB and IGAP imputed GWASs; but Naj noted that hits below the significance threshold may show some overlap. Plans are in the works to combine data from IGAP and EADB GWAS, which will further grow the sample. In addition, Bellenguez said they’ll add samples from UK Biobank soon (Apr 2018 news).—Jessica Shugart
- Paper Alerts: Massive GWAS Studies Published
- GWAS, GWAX: bioRχiv Hosts Bonanza of Alzheimer’s Genetics
- Gao L, Emond MJ, Louie T, Cheadle C, Berger AE, Rafaels N, Vergara C, Kim Y, Taub MA, Ruczinski I, Mathai SC, Rich SS, Nickerson DA, Hummers LK, Bamshad MJ, Hassoun PM, Mathias RA, National Heart, Lung, and Blood Institute GO Exome Sequencing Project, Barnes KC. Identification of Rare Variants in ATP8B4 as a Risk Factor for Systemic Sclerosis by Whole-Exome Sequencing. Arthritis Rheumatol. 2016 Jan;68(1):191-200. PubMed.
- Kunkle BW, Grenier-Boley B, Sims R, Bis JC, Damotte V, Naj AC, Boland A, Vronskaya M, van der Lee SJ, Amlie-Wolf A, Bellenguez C, Frizatti A, Chouraki V, Martin ER, Sleegers K, Badarinarayan N, Jakobsdottir J, Hamilton-Nelson KL, Moreno-Grau S, Olaso R, Raybould R, Chen Y, Kuzma AB, Hiltunen M, Morgan T, Ahmad S, Vardarajan BN, Epelbaum J, Hoffmann P, Boada M, Beecham GW, Garnier JG, Harold D, Fitzpatrick AL, Valladares O, Moutet ML, Gerrish A, Smith AV, Qu L, Bacq D, Denning N, Jian X, Zhao Y, Del Zompo M, Fox NC, Choi SH, Mateo I, Hughes JT, Adams HH, Malamon J, Sanchez-Garcia F, Patel Y, Brody JA, Dombroski BA, Naranjo MC, Daniilidou M, Eiriksdottir G, Mukherjee S, Wallon D, Uphill J, Aspelund T, Cantwell LB, Garzia F, Galimberti D, Hofer E, Butkiewicz M, Fin B, Scarpini E, Sarnowski C, Bush WS, Meslage S, Kornhuber J, White CC, Song Y, Barber RC, Engelborghs S, Sordon S, Voijnovic D, Adams PM, Vandenberghe R, Mayhaus M, Cupples LA, Albert MS, De Deyn PP, Gu W, Himali JJ, Beekly D, Squassina A, Hartmann AM, Orellana A, Blacker D, Rodriguez-Rodriguez E, Lovestone S, Garcia ME, Doody RS, Munoz-Fernadez C, Sussams R, Lin H, Fairchild TJ, Benito YA, Holmes C, Karamujić-Čomić H, Frosch MP, Thonberg H, Maier W, Roschupkin G, Ghetti B, Giedraitis V, Kawalia A, Li S, Huebinger RM, Kilander L, Moebus S, Hernández I, Kamboh MI, Brundin R, Turton J, Yang Q, Katz MJ, Concari L, Lord J, Beiser AS, Keene CD, Helisalmi S, Kloszewska I, Kukull WA, Koivisto AM, Lynch A, Tarraga L, Larson EB, Haapasalo A, Lawlor B, Mosley TH, Lipton RB, Solfrizzi V, Gill M, Longstreth WT Jr, Montine TJ, Frisardi V, Diez-Fairen M, Rivadeneira F, Petersen RC, Deramecourt V, Alvarez I, Salani F, Ciaramella A, Boerwinkle E, Reiman EM, Fievet N, Rotter JI, Reisch JS, Hanon O, Cupidi C, Andre Uitterlinden AG, Royall DR, Dufouil C, Maletta RG, de Rojas I, Sano M, Brice A, Cecchetti R, George-Hyslop PS, Ritchie K, Tsolaki M, Tsuang DW, Dubois B, Craig D, Wu CK, Soininen H, Avramidou D, Albin RL, Fratiglioni L, Germanou A, Apostolova LG, Keller L, Koutroumani M, Arnold SE, Panza F, Gkatzima O, Asthana S, Hannequin D, Whitehead P, Atwood CS, Caffarra P, Hampel H, Quintela I, Carracedo Á, Lannfelt L, Rubinsztein DC, Barnes LL, Pasquier F, Frölich L, Barral S, McGuinness B, Beach TG, Johnston JA, Becker JT, Passmore P, Bigio EH, Schott JM, Bird TD, Warren JD, Boeve BF, Lupton MK, Bowen JD, Proitsi P, Boxer A, Powell JF, Burke JR, Kauwe JS, Burns JM, Mancuso M, Buxbaum JD, Bonuccelli U, Cairns NJ, McQuillin A, Cao C, Livingston G, Carlson CS, Bass NJ, Carlsson CM, Hardy J, Carney RM, Bras J, Carrasquillo MM, Guerreiro R, Allen M, Chui HC, Fisher E, Masullo C, Crocco EA, DeCarli C, Bisceglio G, Dick M, Ma L, Duara R, Graff-Radford NR, Evans DA, Hodges A, Faber KM, Scherer M, Fallon KB, Riemenschneider M, Fardo DW, Heun R, Farlow MR, Kölsch H, Ferris S, Leber M, Foroud TM, Heuser I, Galasko DR, Giegling I, Gearing M, Hüll M, Geschwind DH, Gilbert JR, Morris J, Green RC, Mayo K, Growdon JH, Feulner T, Hamilton RL, Harrell LE, Drichel D, Honig LS, Cushion TD, Huentelman MJ, Hollingworth P, Hulette CM, Hyman BT, Marshall R, Jarvik GP, Meggy A, Abner E, Menzies GE, Jin LW, Leonenko G, Real LM, Jun GR, Baldwin CT, Grozeva D, Karydas A, Russo G, Kaye JA, Kim R, Jessen F, Kowall NW, Vellas B, Kramer JH, Vardy E, LaFerla FM, Jöckel KH, Lah JJ, Dichgans M, Leverenz JB, Mann D, Levey AI, Pickering-Brown S, Lieberman AP, Klopp N, Lunetta KL, Wichmann HE, Lyketsos CG, Morgan K, Marson DC, Brown K, Martiniuk F, Medway C, Mash DC, Nöthen MM, Masliah E, Hooper NM, McCormick WC, Daniele A, McCurry SM, Bayer A, McDavid AN, Gallacher J, McKee AC, van den Bussche H, Mesulam M, Brayne C, Miller BL, Riedel-Heller S, Miller CA, Miller JW, Al-Chalabi A, Morris JC, Shaw CE, Myers AJ, Wiltfang J, O'Bryant S, Olichney JM, Alvarez V, Parisi JE, Singleton AB, Paulson HL, Collinge J, Perry WR, Mead S, Peskind E, Cribbs DH, Rossor M, Pierce A, Ryan NS, Poon WW, Nacmias B, Potter H, Sorbi S, Quinn JF, Sacchinelli E, Raj A, Spalletta G, Raskind M, Caltagirone C, Bossù P, Orfei MD, Reisberg B, Clarke R, Reitz C, Smith AD, Ringman JM, Warden D, Roberson ED, Wilcock G, Rogaeva E, Bruni AC, Rosen HJ, Gallo M, Rosenberg RN, Ben-Shlomo Y, Sager MA, Mecocci P, Saykin AJ, Pastor P, Cuccaro ML, Vance JM, Schneider JA, Schneider LS, Slifer S, Seeley WW, Smith AG, Sonnen JA, Spina S, Stern RA, Swerdlow RH, Tang M, Tanzi RE, Trojanowski JQ, Troncoso JC, Van Deerlin VM, Van Eldik LJ, Vinters HV, Vonsattel JP, Weintraub S, Welsh-Bohmer KA, Wilhelmsen KC, Williamson J, Wingo TS, Woltjer RL, Wright CB, Yu CE, Yu L, Saba Y, Alzheimer Disease Genetics Consortium (ADGC),, European Alzheimer’s Disease Initiative (EADI),, Cohorts for Heart and Aging Research in Genomic Epidemiology Consortium (CHARGE),, Genetic and Environmental Risk in AD/Defining Genetic, Polygenic and Environmental Risk for Alzheimer’s Disease Consortium (GERAD/PERADES),, Pilotto A, Bullido MJ, Peters O, Crane PK, Bennett D, Bosco P, Coto E, Boccardi V, De Jager PL, Lleo A, Warner N, Lopez OL, Ingelsson M, Deloukas P, Cruchaga C, Graff C, Gwilliam R, Fornage M, Goate AM, Sanchez-Juan P, Kehoe PG, Amin N, Ertekin-Taner N, Berr C, Debette S, Love S, Launer LJ, Younkin SG, Dartigues JF, Corcoran C, Ikram MA, Dickson DW, Nicolas G, Campion D, Tschanz J, Schmidt H, Hakonarson H, Clarimon J, Munger R, Schmidt R, Farrer LA, Van Broeckhoven C, C O'Donovan M, DeStefano AL, Jones L, Haines JL, Deleuze JF, Owen MJ, Gudnason V, Mayeux R, Escott-Price V, Psaty BM, Ramirez A, Wang LS, Ruiz A, van Duijn CM, Holmans PA, Seshadri S, Williams J, Amouyel P, Schellenberg GD, Lambert JC, Pericak-Vance MA. Genetic meta-analysis of diagnosed Alzheimer's disease identifies new risk loci and implicates Aβ, tau, immunity and lipid processing. Nat Genet. 2019 Mar;51(3):414-430. Epub 2019 Feb 28 PubMed.
- Bellenguez C, Grenier-Boley B, Lambert JC. Genetics of Alzheimer's disease: where we are, and where we are going. Curr Opin Neurobiol. 2020 Apr;61:40-48. Epub 2019 Dec 18 PubMed.
- Holstege H, Hulsman M, Charbonnier C, Grenier-Boley C, Quenez O, Grozeva D, van Rooij J, Sims R, Ahmad S, Amin N, Norsworthy P, Dols O, Hummerich H, Kawalia A, Amouyel P, Beecham G, Berr C, Bis J, Boland A, Bossu P, Bouwman F, Campion D, Daniele A, Dartigues JF, Debette S, Deleuze JF, Denning N, Destefano A, Farrer L, Fox N, Glimberti D, Genin E, Haines J, Holmes C, Ikram MA, Ikram M, Jansen I, Kraaij R, Lathrop M, Lemstra A, Lleo A, Luckcuck L, Marschall R, Martin E, Masullo C, Mayeux R, Mecocci P, Meggy A, Mol M, Morgan K, Nacmia B, Naj A, Pastor P, Pericak-Vance M, Raybould R, Redon R, Richard AC, Riedel-Heller S, Rivadeneira F, Rousseau S, Ryan N, Saad S, Sanchez-Juan P, Schellenberg G, Scheltens P, Scott J, Seripa D, Spalletta G, Tijms B, Uitterlinden A, van der Lee S, Wagner M, Wallon D, Wang LS, Zarea A, Reinders M, Clarimon J, van Swieten J, Hardy J, Ramirez A, Mead SH, van der Flier W, van Duijn C, Williams J, Nicolas G, Bellengue C, Lambert JC. Exome sequencing identifies novel AD-associated genes. medRxiv. 2020.07.22.20159251