Risk factor SORL1: from genetic association to functional validation in Alzheimer's disease

Abstract

Alzheimer's disease (AD) represents one of the most dramatic threats to healthy aging and devising effective treatments for this devastating condition remains a major challenge in biomedical research. Much has been learned about the molecular concepts that govern proteolytic processing of the amyloid precursor protein to amyloid-β peptides (Aβ), and how accelerated accumulation of neurotoxic Aβ peptides underlies neuronal cell death in rare familial but also common sporadic forms of this disease. Out of a plethora of proposed modulators of amyloidogenic processing, one protein emerged as a key factor in AD pathology, a neuronal sorting receptor termed SORLA. Independent approaches using human genetics, clinical pathology, or exploratory studies in animal models all converge on this receptor that is now considered a central player in AD-related processes by many. This review will provide a comprehensive overview of the evidence implicating SORLA-mediated protein sorting in neurodegenerative processes, and how receptor gene variants in the human population impair functional receptor expression in sporadic but possibly also in autosomal-dominant forms of AD.

Fig. 1

SORLA, a member of the VPS10P domain receptor gene family of neuronal sorting receptors. Sorting-related receptor with A-type repeats (SORLA) is member of the vacuolar protein sorting (VPS10P) domain receptor gene family, a group of five related type-1 transmembrane proteins found in mammalian cell types [101]. Other family members are sortilin as well as sorting receptor CNS expressed (SORCS) 1, SORCS2, and SORCS3. All receptors share an extracellular VPS10P domain, a single transmembrane domain, and a short cytoplasmic tail. The receptors are produced as precursor proteins containing a short pro-peptide at the amino terminus that blocks ligand binding in the VPS10P domain. Proteolytic processing of the pro-peptide by convertases in the Golgi is a precondition for activating the ligand-binding capability of the receptors [41]. SORLA is unique among the members of the VPS10P domain receptor gene family as it contains additional functional modules not shared by the other receptors including domains for protein–protein interaction (fibronectin-type III domains, complement-type repeats) or for pH-dependent release of ligands in endosomes (6-bladed β-propeller). Complement-type repeats and the β-propeller are functional elements also found in lipoprotein receptors, such as the low-density lipoprotein receptor, suggesting the possibility of SORLA to act in cellular lipoprotein transport [80]

Fig. 2

Single nucleotide polymorphisms in SORL1 associated with sporadic AD. Selected coding (red) and non-coding (black) single nucleotide polymorphisms (SNPs) in human SORL1 associated with sporadic AD are indicated. The numbering of the SNPs follows the nomenclature introduced by Rogaeva and colleagues [76]. Most SNPs cluster in two haplotype blocks in the 5′ and 3′ region of the gene locus. SNPs rs11218343 and rs3781834 are associated with sporadic AD at a genome-wide level in individuals of Caucasian and Asian origin [51, 62]. The structure of SORL1 on chromosome 11q23-q24 is indicated below with exons represented by vertical lines. For coding variants, the change in amino acid sequence is given in brackets [12]

Fig. 3

SORLA-dependent sorting of APP and Aβ. a Nascent APP molecules move through the trans-Golgi-network (TGN) to the cell surface where they are subject to non-amyloidogenic processing initiated by α-secretase (α) cleavage. APP molecules not cleaved by α-secretase at the cell surface undergo endocytosis and move to endosomes where they are processed by β- and γ-secretases (β, γ) producing Aβ. Aβ peptides are released from cells through various exocytic mechanisms. b SORLA acts as a sorting receptor for APP causing retrograde endosome to TGN retrieval and slowing down exit from the Golgi to reduce the number of APP molecules subjected to amyloidogenic processing. SORLA also acts as anterograde sorting factor that directs newly produced Aβ molecules to lysosomal degradation, further decreasing the amyloidogenic burden. Anterograde sorting also results in lysosomal breakdown of SORLA, negatively regulating receptor levels [23]. Some cytosolic adaptors required for shuttling of SORLA between TGN and endosomes are indicated and include GGAs (Golgi-localizing, y-adaptin ear homology domain, ARF-interacting proteins) for anterograde as well as retromer and PACS1 (phosphofurin acidic cluster sorting protein 1) for retrograde sorting. Figures adapted from [100]

Fig. 4

Protein interactions at the cytoplasmic domain of SORLA. The amino acid sequence of the cytoplasmic domain of human SORLA (Q92673, Uniprot) is shown. Binding sites for GGAs, PACS1, AP1 and 2 as well as for the VPS26 subunit of the retromer complex are color coded in green. Binding sites for SPAK, ROCK2, and PKC in the receptor tail are currently unknown. Amino acid residue serine 2206 is subject to phosphorylation by ROCK2

Fig. 5

Functional modules in SORLA targeted by coding disease gene variants. Structural organization of the mature SORLA polypeptide indicating proposed functions for the various protein modules (to the left). Selected sequence variations identified in late- (red) or early onset (black) cases of AD are shown to the right. Mutations p.E270K, p.A528T, and p.T947M cause missorting of the receptor and APP ligand, increasing the extent of amyloidogenic processing [93]. Mutation p.G511R disrupts the binding site for Aβ, impairing SORLA-dependent lysosomal catabolism of this peptide [13]. Other domains in SORLA harboring AD-associated sequence alterations are the complement-type repeats, the fibronectin-type III domain, and the cytoplasmic tail as exemplified by the indicated coding variants [65, 95]. Whether these sequence variations also alter receptor functions is unresolved so far