On February 28 in Nature Genetics online, geneticists from the International Genomics of Alzheimer’s Project published the largest genome-wide association meta-analysis to date. Brian Kunkle, University of Miami Miller School of Medicine, Benjamin Grenier-Boley, INSERM, Lille, France, and colleagues, including five research conglomerates, collaborated to analyze data from 35,274 people who had been diagnosed with late-onset Alzheimer’s disease. They identified five new loci with genome-wide significance. They fell near the IQCK, ACE, ADAM10, ADAMTS1, and WWOX genes. The last encodes a protein believed to interact with tau.

The paper follows the January 7 publication in Nature Genetics of a massive GWAS/GWAX study. In GWAX, researchers study AD by proxy, using not diagnosed AD cases, but people with a family history of the disease—the idea being that they carry a higher genetic risk for the disorder. Iris Jansen and Jeanne Savage, VU University, Amsterdam, and colleagues reported nine new risk loci for AD among a total of 71,880 diagnosed or proxy cases. This analysis also turned up ADAM10. The other eight hits lay near the ADAMTS4, HESX1, CLNK, CNTNAP2, APH1B, KAT8, ALPK2, and AC074212.3 genes.

These two papers follow the publication in May 2018 of another GWAX. Riccardo Marioni and colleagues at the University of Edinburgh analyzed IGAP data to identify ADAM10, ACE, and KAT8 as AD loci (Marioni et al., 2018). Alzforum covered these papers when they were uploaded to the BioRχiv preprint server (Apr 2018 news). 

“These three papers mark steady progress in the identification of new loci for AD in Caucasian populations,” noted John Hardy, University College London. He was one of 474 co-authors on the Kunkle paper. “As more loci are identified, nearly all of them fit into categories that are already established—lipid metabolism, microglial activation, and APP processing.”

Since Kunkle and colleagues uploaded their paper to BioRχiv, they have analyzed roughly an additional 4,600 case/control samples in a replication analysis. “That led to the additional genome-wide locus at WWOX, which we did not have in the BioRχiv,” said Kunkle.

How might these loci contribute to AD risk? The functional variants are yet to be identified. Meanwhile, Kunkle and colleagues ranked genes near the loci for their likely involvement in AD pathogenesis based on eight criteria: deleterious coding, loss of function, or splicing variant in the gene; significance in gene-based testing; expression in a tissue relevant to AD, including astrocytes, neurons, microglia/macrophages, and oligodendrocytes; enrichment in a human microglial (HuMi) database (Olah et al., 2018); correlation with expression quantitative trait loci (eQTL) in any tissue, an AD-relevant tissue, or co-localization with an eQTL; being involved in a biological pathway enriched in AD; expression being correlated with BRAAK stage; and evidence of differential expression in more than one AD study (see figure below). 

Top Hits. A biological ranking system prioritizes genes near known and newly discovered GWAS hits. [Courtesy of Kunkle et al., Nature Genetics 2019.]

By this ranking, ADAM10, which encodes α-secretase, was the top hit near that locus. The secretase is well known for non-amyloid processing of amyloid-β precursor protein (APP) and for shedding the ectodomain of TREM2, the microglial receptor that carries AD risk variants in its own right. While APP lies near the ADAMTS1 locus, the researchers believe ADAMTS1 itself the likely gene at that locus, though they do not rule out it regulating APP somehow. ADAMTS1 encodes ADAM metalloproteinase with thrombospondin type 1 motif, a potentially neuroprotective factor that is induced by interleukin-1β (Kuno et al., 1997). ICQX was also the top rank near its locus. This function of this gene is unclear, but it has been linked to obesity. Ranking high near the ACE locus were immune response genes, including PSMC5, MAP3K3, and CD79B. ACE itself might be a candidate since it has been previously linked to LOAD and Aβ levels (Baranello et al., 2015; Kauwe et al., 2014). 

Near the WWOX locus the WW-domain containing oxidoreductase gene itself ranked highly, but so did MAF, which encodes a transcription factor expressed in microglia and macrophages. “Studies have shown that WWOX may control neuronal survival and block neurodegeneration via direct binding of tau or interactions with tau-phosphorylating enzymes,” noted Kunkle (Sze et al., 2004). WWOX dysfunction might also cause aggregation of tau and Aβ (Chang and Chang, 2015). “It is also possible that WWOX or a gene in this locus influences risk of Alzheimer’s through other mechanisms. For instance, this locus has been associated with obesity, HDL cholesterol, and triglyceride levels, all of which may influence Alzheimer’s risk,” wrote Kunkle.

Jansen and Savage took a similar approach, using three mapping strategies to link the genetic loci to potential functional genes. They identified 99 genes that lay within 10 kilobases of a locus, matched loci to 168 genes via eQTL analysis, and linked loci to 21 genes based on tertiary structure of chromatin. In other words, they tried to find distant DNA regions that might contort to juxtapose the AD loci when the DNA is wound (see image below).

Any two methods together identified 80 genes; all three methods identified 16. Notable hits were CLU and PTK2B on chromosome 8, which may interact or confer risk independently. ADMTS4 itself might be a risk factor since it has been implicated in AD previously. Ditto for ADAM10 and APH1B, which encodes a component of γ-secretase. The researchers noted that expression and methylation of KAT8, a lysine acetyltransferase, are regulated by numerous variants near that locus, making it a likely functional candidate. It is also regulated by KANSL1, a component of a histone acetylation complex that the authors say has been associated with AD in the absence of an ApoE4 allele.

Mapping Function. Positional, eQTL, and chromatin mapping links genome-wide significant single nucleotide polymorphisms to 192 genes. Sixteen were identified by all three maps, 80 by at least two. [Courtesy Jansen et al., 2019, and Nature Genetics.]

Kunkle and colleagues also gauged how loci might relate to other co-morbidities. To do so, they correlated LOAD genetics with 792 other human diseases, traits, and behaviors. Family history of AD, fewer years of education, and indications of cardiovascular disease positively correlated with LOAD, while intelligence and more years of education correlated negatively. Some correlations were complex—many individual measures of cardiovascular disease and diabetes, such as family history of high blood pressure and fasting insulin, negatively correlated with AD, which the authors suggest is evidence that treating these disorders may be protective.

Where do GWAS studies go from here? Kunkle and colleagues noted loci that deserve further study because they almost reached genome-wide significance, including those near miR142/TSPOAP1-AS1, NDUFAF6, NME8, and MEF2C genes. Still, Hardy thinks the AD field is coming to the end of GWAS in Caucasians. “For related phenotypes, such as dementia with Lewy bodies and progressive supranuclear palsy, the current GWAS are small and more are needed. For AD, we clearly need GWAS in Asian and African populations,” he wrote.

Senior authors on the Kunkle paper were Agustin Ruiz, Universitat Internacional de Catalunya, Barcelona, Spain; Cornelia M. van Duijn, Medical Center, Rotterdam, the Netherlands; Peter A. Holmans and Julie Williams, Cardiff University, U.K.; Sudha Seshadri, Boston University; Phillippe Amouyel and Jean-Charles Lambert, University of Lille; Gerard D. Schellenberg, University of Pennsylvania Perelman School of Medicine, Philadelphia; and Margaret Pericak-Vance from the University of Miami.  

Stephan Ripke, Charité–Universitätsmedizin, Berlin, Ole Andreassen, University of Oslo, and Danielle Posthuma at VU University were co-senior authors on the Jansen paper.—Tom Fagan


No Available Comments

Make a Comment

To make a comment you must login or register.


News Citations

  1. GWAS, GWAX: BioRxiv Hosts Bonanza of Alzheimer’s Genetics

Paper Citations

  1. . GWAS on family history of Alzheimer's disease. Transl Psychiatry. 2018 May 18;8(1):99. PubMed.
  2. . A transcriptomic atlas of aged human microglia. Nat Commun. 2018 Feb 7;9(1):539. PubMed.
  3. . Molecular cloning of a gene encoding a new type of metalloproteinase-disintegrin family protein with thrombospondin motifs as an inflammation associated gene. J Biol Chem. 1997 Jan 3;272(1):556-62. PubMed.
  4. . Amyloid-beta protein clearance and degradation (ABCD) pathways and their role in Alzheimer's disease. Curr Alzheimer Res. 2015;12(1):32-46. PubMed.
  5. . Genome-wide association study of CSF levels of 59 alzheimer's disease candidate proteins: significant associations with proteins involved in amyloid processing and inflammation. PLoS Genet. 2014 Oct;10(10):e1004758. Epub 2014 Oct 23 PubMed.
  6. . Down-regulation of WW domain-containing oxidoreductase induces Tau phosphorylation in vitro. A potential role in Alzheimer's disease. J Biol Chem. 2004 Jul 16;279(29):30498-506. PubMed.
  7. . WWOX dysfunction induces sequential aggregation of TRAPPC6AΔ, TIAF1, tau and amyloid β, and causes apoptosis. Cell Death Discov. 2015;1:15003. Epub 2015 Aug 3 PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . 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.
  2. . Genome-wide meta-analysis identifies new loci and functional pathways influencing Alzheimer's disease risk. Nat Genet. 2019 Mar;51(3):404-413. Epub 2019 Jan 7 PubMed.