18 Jul 2023

Make way, APP, PS1, PS2. A fourth gene, SORL1, is joining you on the podium of infamy. For years, researchers have steadily built the case that SORL1 is the fourth familial, sometimes even autosomal-dominant, Alzheimer’s disease gene. Now, a culmination of genetic, epidemiological, and functional data may provide the necessary burden of proof to persuade the field at large. Together, a trio of new studies suggest that variants that derail expression, or proper trafficking, of this endosomal receptor not only cause AD, but indeed spur more cases of the disease than the other three familial AD genes combined.

Coinciding with a flurry of SORL1 studies is the release of Alzforum’s latest—and largest—addition to its Mutations database. Published on Alzforum today, the new SORL1 collection is organized around an interactive diagram of the behemoth protein, with its 2,214 residues carefully arranged into its many domains. The curious reader can find published data about hundreds of variants, as well as their predicted pathogenicity, by clicking on a given residue.

The data-packed diagram is the product of years of painstaking work by Olav Andersen of Aarhus University in Denmark and colleagues. They integrated data about the structure and pathogenicity of proteins with similar domains to SORL1, to estimate the pathogenicity of hundreds of SORL1 variants distributed throughout the gene. Working with Andersen and Henne Holstege at Amsterdam University Medical Center, the Alzforum database team rendered this opus into a visual, interactive form.

Behold the Beast. SORL1 consists of myriad components, including 10 VPS10p domains (green), six YWTD domains (gray), a single EGF domain (orange), 11 CR domains (turquoise), six 3Fn domains (blue), a transmembrane domain, and a cytoplasmic tail. [Courtesy of Andersen et al., bioRxiv, 2023.]

A handful of new studies, three posted on medRxiv/bioRxiv and two coming soon, lend real-life support to these informed predictions of pathogenicity. One, led by Holstege, assessed the AD risk imparted by each of more than 600 SORL1 variants detected in a large sequencing cohort. In agreement with Andersen’s variant prioritization scheme, it found that mutations predicted to derail SORL1 folding, trafficking, or dimerization within the endosome raise the odds of AD immensely for carriers. Two other studies, one led by Andersen and the other by Jessica Young at the University of Washington in Seattle, zeroed in on neuronal mayhem imparted by individual SORL1 variants that tracked with AD in families. They build the case that missense mutations that mess with SORL1’s endosomal trafficking spell trouble for carriers. Additional studies with similar findings await in the wings.

Late (Onset) Bloomer
Unlike variants in APP, PS1, and PS2, which were initially found to segregate with disease in multigenerational families, SORL1 made its debut as a genetic risk factor for sporadic AD. Indeed, noncoding variants around the gene remain at the top of the growing heap of AD risk variants (Jan 2007 news; Feb 2021 news). Later, exome sequencing studies unearthed rare coding variants in SORL1 that appeared to pack a wallop. Mutations predicted to snuff out expression of the gene, or truncate the protein, occurred almost exclusively among people with cognitive impairment or AD (Holstege et al., 2017; Jun 2018 news).

Rare or common, the impact adds up: Scientists estimate that potentially damaging SORL1 variants affect as many as 2.75 percent of all unrelated people with early onset AD, and 1.5 percent of those with late-onset AD (Nov 2022 news).

That SORL1 was first pegged as a risk factor for LOAD may have overshadowed subsequent discoveries of rare coding variants that appeared highly detrimental, Andersen told Alzforum. This sequence of events may partly explain why the field has been slow to warm up to the idea of SORL1 as a causal gene, he added.

“In the case of SORL1, the causal argument is like a jigsaw,” is how Scott Small of Columbia University in New York put it. It's been built a piece at a time. Complicating matters is that SORL1 associates with late-onset forms of the disease, which are inherently more complex.

What’s more, besides the clearly deleterious variants that wipe out expression of one copy of SORL1, 90 percent of SORL1 coding variants are missense mutations with unclear pathogenicity. Though some rare variants have been found almost exclusively among AD cases in large exome-sequencing studies, hinting at their hazardous nature, geneticists have few multigenerational families to draw upon for segregation analysis. For hundreds of variants, Andersen and other researchers therefore had to use different tactics to estimate pathogenicity, and then test out their predictions in large case control studies.

Andersen leveraged SORL1’s homology with other proteins to align the structure of its many domains, and to infer the potential damage wrought by mutations at different residues within each. SORL1 is considered a mosaic protein. It comprises functional domains that hail from the low-density lipoprotein receptor family (LDLR) and from the vacuolar protein sorting-10 family; about a third of its domains have no family ties. Therefore, Andersen and colleagues were able to align many SORL1 domains with those from homologous proteins, and pinpoint conserved residues that were likely critical for that domain’s folding and function. The scientists also considered disease-causing variants in the other protein domains—for example, a variant in the YWTD domain of LDLR that leads to familial hypercholesterolemia—to infer the importance of the corresponding residue in SORL1.

All in the Families. SORL1’s dual membership in the LDLR and VPS10p protein families means that it shares many homologous domains with other proteins. By aligning them, pinpointing conserved residues needed for function, and noting residues that cause other diseases when mutated (listed on bottom), scientists gauged the potential AD risk associated with SORL1 variants. [Courtesy of Andersen et al., bioRxiv, 2023.]

Andersen told Alzforum that he spent the better part of the COVID pandemic glued to a chair, aligning these protein domains. What came of this immersion? A compendium of SORL1 variants, in which each residue was categorized as a high or moderate priority risk for developing AD (Andersen et al., 2023). Meanwhile, Alzforum database curators were there every residue of the way, cataloguing information about each variant, and using Andersen’s structural alignments to chart each of SORL1’s domains (image below).

SORL1 Sorted. Screenshot of Alzforum’s new SORL1 page. In the database, readers can click on any residue with known variants to access requisite findings. Noncoding variants, not pictured here, are also listed on the page. Loss-of-function variants are denoted in red, missense mutations in yellow.

The resulting SORL1 page on Alzforum allows users to view the whole protein at the level of single residues, each nestled within its domain. SORL1’s byzantine ectodomain features a 10-folded ring of the VPS10p domain, ending with a 10CC-region, a 6-folded ring of the YWTD domain, followed by a single epidermal growth factor (EGF) domain, 11 complement-type repeat (CR) domains, and six fibronectin-type III (3Fn) domains.

The VPS10p and CR domains are known to bind substrates, including Aβ and APP, respectively. However, a recent study found that once in the endosome, SORL1 dimerizes via its VPS10p and 3Fn domains, enabling it to associate with the retromer complex via its cytoplasmic tail (Feb 2023 news). This retromer association is key to SORL1's function, Small emphasized.

Alzforum users can click on each SORL1 residue for which variants have been identified to access the published findings on it. Overlaid on the diagram are Andersen’s priority rankings, and users can apply a filter to highlight only high- or moderate-priority variants. These classifications are not specific to each individual variant that has been identified at that residue, but rather to the importance of the residue itself for domain folding and function. The diagram is also conveniently searchable, by residue number or specific variant.

Andersen hopes that researchers around the world will use this resource to pick variants for deeper functional studies, and that clinical geneticists may reference it to give prognostic insight to patients and their families.

That said, much of the prioritization for missense variants is based on alignment with other proteins and predictions about how each residue influences domain structure. Thus far, few mutation carriers are available to help scientists nail down their potency in real life. For examples of  how reliably this protein structure/function-based classification system estimates the odds of AD, see Part 2 of this series.—Jessica Shugart

Comments

No Available Comments

Make a Comment

To make a comment you must login or register.

References

Mutation Interactive Images Citations

1. SORL1 18 Jul 2023

News Citations

1. SORLA Soars—Large Study Links Gene to Late-onset AD 15 Jan 2007
2. Massive GWAS Meta-Analysis Digs Up Trove of Alzheimer’s Genes 24 Feb 2021
3. Gaining Notoriety, SORL1 Claims Spot Among Top Alzheimer’s Genes 7 Jun 2018
4. Rare Variants in Lipid Transporter Genes Increase Risk for Alzheimer’s Disease 23 Nov 2022
5. When SORL1 Dimerizes in Endosomes, Retromers Recycle APP Faster 1 Feb 2023
6. When Missense Variants Derail SORL1 Traffic, Destination Is Dementia 18 Jul 2023

Paper Citations

1. Holstege H, van der Lee SJ, Hulsman M, Wong TH, van Rooij JG, Weiss M, Louwersheimer E, Wolters FJ, Amin N, Uitterlinden AG, Hofman A, Ikram MA, van Swieten JC, Meijers-Heijboer H, van der Flier WM, Reinders MJ, van Duijn CM, Scheltens P. Characterization of pathogenic SORL1 genetic variants for association with Alzheimer's disease: a clinical interpretation strategy. Eur J Hum Genet. 2017 Aug;25(8):973-981. Epub 2017 May 24 PubMed.
2. Andersen OM, Monti G, Jensen AM, deWaal M, Hulsman M, Olsen JG, Holstege H. Relying on the relationship with known disease-causing variants in homologous proteins to predict pathogenicity of SORL1 variants in Alzheimer's disease. 2023 Feb 27 10.1101/2023.02.27.524103 (version 1) bioRxiv.

Other Citations

1. Mutations

Further Reading

Papers

1. Small SA, Kent K, Pierce A, Leung C, Kang MS, Okada H, Honig L, Vonsattel JP, Kim TW. Model-guided microarray implicates the retromer complex in Alzheimer's disease. Ann Neurol. 2005 Dec;58(6):909-19. PubMed.
2. Vardarajan BN, Zhang Y, Lee JH, Cheng R, Bohm C, Ghani M, Reitz C, Reyes-Dumeyer D, Shen Y, Rogaeva E, St George-Hyslop P, Mayeux R. Coding mutations in SORL1 and Alzheimer disease. Ann Neurol. 2015 Feb;77(2):215-27. PubMed.
3. Campion D, Charbonnier C, Nicolas G. SORL1 genetic variants and Alzheimer disease risk: a literature review and meta-analysis of sequencing data. Acta Neuropathol. 2019 Aug;138(2):173-186. Epub 2019 Mar 25 PubMed.

News

1. Voilà SorLA! Sorting Receptor’s Structure Solved 6 Feb 2015
2. In Mini-Pigs Missing SORL1, Endosomes Swell and Aβ Peptides Rise in CSF 28 Sep 2022