Two papers in the July 25 Nature Genetics yield new insight into genetic factors underlying amyotrophic lateral sclerosis (ALS). In one, scientists led by Ammar Al-Chalabi, King’s College London, and Leonard van den Berg and Jan Veldink of the University Medical Center Utrecht, the Netherlands, describe how they combined forces with other ALS researchers to conduct the largest genome-wide association study of the disease to date. C21orf2, a gene predicted to be important in mitochondrial function, emerged as the strongest hit. The second paper comes from geneticists van den Berg, Veldink, Vincenzo Silani of the IRCCS Istituto Auxologico Italiano in Milan, Christopher Shaw at King’s College London, and John Landers at the University of Massachusetts Medical School, Worcester. In it, they detail loss-of-function mutations in the NEK1 gene that predispose carriers to ALS. The NEK1 protein has a range of functions, many of which fit with what scientists already suspect about the etiology of disease.
“A triumph of these two papers is bringing these large, well-defined patient collections together and applying new genetic methods to understand what’s going on in ALS,” said Elizabeth Fisher, University College London, who was not involved in either study. She noted how scientists conducting ALS genetic studies are hampered because ALS is such a rare disease and patient populations are small.
Since numerous genes contribute to this heterogeneous disease, scientists need many patient volunteers to have enough statistical power to uncover genetic associations. In the past nine years, the number of ALS patients genotyped for GWAS grew from just a few hundred to several thousand, leading to the discovery of new ALS genes, including UNC13A and SARM1 (van Es et al., 2009; Fogh et al., 2014). Studies on ALS families have yielded other genes, including SOD1 and FUS. However, because ALS strikes late in life, the number of large affected families available for linkage analyses are limited. Much of the heritability of ALS has yet to be explained.
Working with Al-Chalabi, co-first authors Wouter van Rheenen at University Medical Center Utrecht and Aleksey Shatunov of King’s College London used arrays of common single-nucleotide polymorphisms (SNPs) to genotype 12,577 sporadic ALS cases and 23,475 controls from more than a dozen countries. To maximize the number of SNPs included in their analysis, they used a technique called imputation, which allows geneticists to predict the wider set of SNPs that a person carries even if a particular SNP array did not test for them. For imputation, they sequenced the whole genomes of 1,246 patients with ALS and 615 controls from the Netherlands. This revealed common haplotypes—blocks of DNA that are inherited together—that allowed van Rheenen and colleagues to predict genetic variants inherited by others in the cohort. In the end, the authors tested more than 18 million SNPs for association with ALS. They first examined cases and controls within the same geographical location, and then combined results from various regions.
The authors first confirmed that variants in the previously reported C9ORF72, UNC13A, and SARM1 genes associated with the disease. They also hit upon a new candidate, C21orf2. This codes for a protein found in the mitochondria of immune cells. “That’s important because while the mitochondria have always been thought to be relevant to ALS, we’ve never had direct genetic evidence,” said Al-Chalabi.
They then took a different tack to analyze the data. Rather than combining country-specific data, they used a linear mixed-model approach. This allowed them to examine the whole cohort as one, yet automatically take into account ancestry and how everyone in the sample was related (for a review, see Yang et al., 2014). The result is a more powerful study with a cleaner signal, said Al-Chalabi. The analysis confirmed all four genes identified by the standard GWAS, plus three more—MOBP, SCFD1, and a long non-coding RNA on chromosome 8p23.2. MOBP has been associated with progressive supranuclear palsy and length of survival in frontotemporal degeneration, hinting at a common neurodegenerative pathway among all three diseases, Al-Chalabi said. SCFD1 is involved in vesicle transport, similar to another ALS gene, VCP. The role of the long coding RNA remains a mystery. The authors then replicated the association of the C21orf2, MOBP, and SCFD1 loci in nine independent genetic cohorts totaling 2,579 cases and 2,767 controls. At some point, a larger sample size might also confirm the 8p23.2 variant as an ALS gene, said Al-Chalabi.
Through their analysis, the authors estimated that common genetic variants that can be detected by GWAS explain only a tiny fraction of the total estimated heritability of ALS. Most heritability derives from rare variations that GWAS studies cannot detect, they conclude. ALS is shaping up to have a “polygenic rare variant architecture,” where a handful of rare polymorphisms each have a comparatively large effect on disease, said Al-Chalabi. By contrast, diseases such as schizophrenia appear to be due to many common variants, each with a small effect. “For ALS, methods that go after rare variants may be more productive than GWAS,” said Al-Chalabi.
As described in the second paper, geneticists led by Landers took a stab at searching for those rare variants in patients with familial forms of disease. Co-first authors Kevin Kenna at UMass and Nicola Ticozzi in Milan developed a machine-learning protocol to filter for rare variants predicted to have functional effects. The scientists previously used this exome-wide rare variant analysis to identify TUBA4A as an ALS gene (see Oct 2014 news). Here, they optimized this algorithm by searching for variants in 10 previously described genes associated with ALS in the whole exomes of 1,022 people with FALS and 7,315 controls. They settled on the algorithm that was most sensitive for finding all 10.
The authors then used the algorithm in the same sample set to find new ALS genes. Out popped NEK1, which codes for the serine/threonine kinase NIMA (never in mitosis gene-A)-related kinase. NEK1 has diverse functions in cell-cycle progression, mitosis, microtubule stability, and the formation of primary cilia that are crucial for sensing mechanical and chemical stimuli. It also regulates the permeability of the mitochondrial membrane and helps with DNA repair, processes that are impaired in SOD1 and FUS mutants (Chen et al., 2009; Pelegrini et al., 2010; Sama et al., 2014; Tafuri et al., 2015).
On average, 10 NEK1 variants in 12 FALS patients conferred an 8.2 odds ratio of having the disease, with those that cause a loss of function being the strongest driver of this signal. These variants appeared scattered throughout the sequences encoding the protein kinase domain and protein-interacting regions. NEK1 had been identified previously as a candidate gene in ALS studies with smaller samples (Cirulli et al., 2015; Brenner et al., 2016). Intriguingly, the protein interacts with C21orf2 (Fang et al., 2015).
Unbeknownst to Landers and colleagues at the time, co-first authors Perry van Doormaal and Annelot Dekker from Veldink’s lab had gone in search of new ALS genes in still another way. They sequenced the genomes of four patients from a genetically isolated population in the Netherlands, looking for homozygous variants that could explain the disease. Two copies of a rare allele seldom turn up in the same genome, but this occurs more often in communities with low genetic variation. They came across a missense Arg261His variant in NEK1, which was homozygous in two of the patients and heterozygous in the others. They found the same change in 40 out of 6,172 sporadic ALS patients compared to 14 of 4,417 matched controls, giving an odds ratio of 2.4 for ALS risk.
When Landers and Veldink learned that they had uncovered the same gene by two different methods, they teamed up to confirm their finding. Landers found that the Arg261His variant turned up more often in patients than controls in the FALS cohort. Sequencing the gene in 2,303 more SALS cases yielded 10 loss-of-function polymorphisms in NEK1 in 23 patients, compared to zero in 1,059 controls. The average odds ratio for the combined familial discovery and sporadic replication cohorts came to 8.8.
Altogether, these analyses uncovered 120 NEK1 variants. Those that caused a loss of function appeared in 1.2 and 1.0 percent of familial and sporadic patients, respectively, compared with 0.17 percent of controls. Some of the latter could reflect low penetrance or a subclass of variants restricted to the C-terminus of the protein, which are likely to have weaker effects on function, noted Landers.
The findings make sense in terms of what is already known about the genes associated with ALS, Fisher told Alzforum. “This study helps us fit more pieces into the puzzle to give a better understanding of what makes people more susceptible to motor neuron degeneration,” she said. Hande Ozdinler, Northwestern University, Chicago, agreed, saying the real promise of these genetics studies will be to gather as many gene candidates as possible and sort them into underlying compromised canonical pathways. “The genes are just the tip of the iceberg,” she told Alzforum. When researchers can pinpoint cellular pathways that are defective, which may differ depending on the group of patients, they may design therapeutics to compensate, she said. Together, ALS genes are converging on molecular pathways involved in mitochondrial dysfunction, DNA repair, and microtubule stability that could hint at therapeutic targets. Fisher added that the evidence in these papers is convincing enough to start investigating the effects of related mutations in mice and human induced pluripotent stem cells.
Given that most of the missing heritability in ALS appears to be due to rare alleles, sequencing techniques may offer real promise, said Matthew Harms, Columbia University, New York. “The take-home message is that our inclination to use whole-genome sequencing in large groups of patients is right on and will yield additional genetic risk variants.” Techniques that search for rare variants are already being used in other neurodegenerative diseases, such as Parkinson’s, he said. He said scientists should also examine epigenetics as a potential player and that very large GWAS will still be useful in ALS. “As additional cases are sequenced and incorporated, we will continue to chip away at the missing heritability,” he added.
1. van Rheenen W, Shatunov A, Dekker AM, McLaughlin RL, Diekstra FP, Pulit SL, van der Spek RA, Võsa U, de Jong S, Robinson MR, Yang J, Fogh I, van Doormaal PT, Tazelaar GH, Koppers M, Blokhuis AM, Sproviero W, Jones AR, Kenna KP, van Eijk KR, Harschnitz O, Schellevis RD, Brands WJ, Medic J, Menelaou A, Vajda A, Ticozzi N, Lin K, Rogelj B, Vrabec K, Ravnik-Glavač M, Koritnik B, Zidar J, Leonardis L, Grošelj LD, Millecamps S, Salachas F, Meininger V, de Carvalho M, Pinto S, Mora JS, Rojas-García R, Polak M, Chandran S, Colville S, Swingler R, Morrison KE, Shaw PJ, Hardy J, Orrell RW, Pittman A, Sidle K, Fratta P, Malaspina A, Topp S, Petri S, Abdulla S, Drepper C, Sendtner M, Meyer T, Ophoff RA, Staats KA, Wiedau-Pazos M, Lomen-Hoerth C, Van Deerlin VM, Trojanowski JQ, Elman L, McCluskey L, Basak AN, Tunca C, Hamzeiy H, Parman Y, Meitinger T, Lichtner P, Radivojkov-Blagojevic M, Andres CR, Maurel C, Bensimon G, Landwehrmeyer B, Brice A, Payan CA, Saker-Delye S, Dürr A, Wood NW, Tittmann L, Lieb W, Franke A, Rietschel M, Cichon S, Nöthen MM, Amouyel P, Tzourio C, Dartigues JF, Uitterlinden AG, Rivadeneira F, Estrada K, Hofman A, Curtis C, Blauw HM, van der Kooi AJ, de Visser M, Goris A, Weber M, Shaw CE, Smith BN, Pansarasa O, Cereda C, Del Bo R, Comi GP, D’Alfonso S, Bertolin C, Sorarù G, Mazzini L, Pensato V, Gellera C, Tiloca C, Ratti A, Calvo A, Moglia C, Brunetti M, Arcuti S, Capozzo R, Zecca C, Lunetta C, Penco S, Riva N, Padovani A, Filosto M, Muller B, Stuit RJ; PARALS Registry; SLALOM Group; SLAP Registry; FALS Sequencing Consortium; SLAGEN Consortium; NNIPPS Study Group, Blair I, Zhang K, McCann EP, Fifita JA, Nicholson GA, Rowe DB, Pamphlett R, Kiernan MC, Grosskreutz J, Witte OW, Ringer T, Prell T, Stubendorff B, Kurth I, Hübner CA, Leigh PN, Casale F, Chio A, Beghi E, Pupillo E, Tortelli R, Logroscino G, Powell J, Ludolph AC, Weishaupt JH, Robberecht W, Van Damme P, Franke L, Pers TH, Brown RH, Glass JD, Landers JE, Hardiman O, Andersen PM, Corcia P, Vourc’h P, Silani V, Wray NR, Visscher PM, de Bakker PI, van Es MA, Pasterkamp RJ, Lewis CM, Breen G, Al-Chalabi A, van den Berg LH, Veldink JH. Genome-wide association analyses identify new risk variants and the genetic architecture of amyotrophic lateral sclerosis. Nat Genet. 2016 Jul 25. [Pubmed].
2. Kenna KP, van Doormaal PT, Dekker AM, Ticozzi N, Kenna BJ, Diekstra FP, van Rheenen W, van Eijk KR, Jones AR, Keagle P, Shatunov A, Sproviero W, Smith BN, van Es MA, Topp SD, Kenna A, Miller JW, Fallini C, Tiloca C, McLaughlin RL, Vance C, Troakes C, Colombrita C, Mora G, Calvo A, Verde F, Al-Sarraj S, King A, Calini D, de Belleroche J, Baas F, van der Kooi AJ, de Visser M, Ten Asbroek AL, Sapp PC, McKenna-Yasek D, Polak M, Asress S, Muñoz-Blanco JL, Strom TM, Meitinger T, Morrison KE; SLAGEN Consortium, Lauria G, Williams KL, Leigh PN, Nicholson GA, Blair IP, Leblond CS, Dion PA, Rouleau GA, Pall H, Shaw PJ, Turner MR, Talbot K, Taroni F, Boylan KB, Van Blitterswijk M, Rademakers R, Esteban-Pérez J, García-Redondo A, Van Damme P, Robberecht W, Chio A, Gellera C, Drepper C, Sendtner M, Ratti A, Glass JD, Mora JS, Basak NA, Hardiman O, Ludolph AC, Andersen PM, Weishaupt JH, Brown RH Jr, Al-Chalabi A, Silani V, Shaw CE, van den Berg LH, Veldink JH, Landers JE. NEK1 variants confer susceptibility to amyotrophic lateral sclerosis. Nat Genet. 2016 Jul 25. [Pubmed].
To view commentaries, primary articles and linked stories, go to the original posting on Alzforum.org here.
Copyright 1996–2016 Biomedical Research Forum, LLC. All Rights Reserved.