Functional validation of Alzheimer’s disease risk genes using established and pluripotent human cell culture models

Abstract

Alzheimer’s disease is a neurodegenerative disorder that leads to the progressive loss of neurons and affects millions of patients worldwide. The risk of Alzheimer’s disease is mainly conferred through genetic risk factors, some of which are still not identified. The recent advances in sequencing technologies have uncovered many previously unknown and potentially pathogenic single-nucleotide polymorphisms (SNPs) in Alzheimer’s disease patients. Given the large number of these Alzheimer’s disease-associated SNPs, the functional characterization and proof of their relevance for this disease is urgently needed. Here, I functionally analyzed the relevance of coding and non-coding risk SNPs in the SORL1 and SORT1 gene that encode for the VPS10P domain receptors SORLA and sortilin. VPS10P domain receptors are a unique class of neuronal sorting receptors that direct intracellular transport of target proteins between trans-Golgi network (TGN), cell surface and endosomes. Sortilin acts as a neuronal clearance receptor for both APOE (apolipoprotein E), the major genetic risk factor for late-onset Alzheimer’s disease, and progranulin, a protective factor for frontotemporal lobar degeneration. A risk locus of several genetically linked SNPs upstream of the SORT1 gene was associated with SORT1 expression levels in the liver. To test whether this risk locus is indeed a regulator of SORT1 expression in the liver and possibly in the brain, I differentiated human iPSCs carrying minor or major variants of these SNPs into hepatocytes and cortical neurons. I also used genome-editing techniques to specifically exchange the proposed functional SNP rs12740374 from major to minor variant in order to verify the functionality of this SNP in isogenic cell lines. My data demonstrated that the SORT1 risk locus predicts SORT1 expression in iPSC-derived hepatocytes, but not in iPSC-derived neurons, which may be explained by lower expression levels of the transcription factor C/EBPα in neurons. Analysis of isogenic iPSC lines suggested that the proposed functional SNP rs12740374 indeed determines SORT1 expression in hepatocytes. SORLA is a trafficking receptor for both the Aβ peptide as well as its precursor protein (APP) and thus affects Aβ accumulation in the brain, one of the hallmarks of Alzheimer’s disease. Recently, potentially pathogenic coding SORL1 mutations were identified in early-onset Alzheimer’s disease (EOAD) patients. I functionally characterized the EOAD-associated SORL1 mutations by overexpressing mutant receptor variants in an established neuronal cell line (SH-SY5Y). Out of the three analyzed EOAD-associated SORL1 mutations, the N1358S mutation demonstrated an increase in extracellular Aβ levels compared to the wildtype condition. Neither impaired binding of APP nor affected lysosomal targeting of Aβ could explain the increase in Aβ levels in the mutant SORLAN1358S cells. However, the subcellular trafficking of both SORLA and APP was altered in the SORLAN1358S overexpressing cell line, with more SORLA and APP localizing to endosomal instead of Golgi compartments when compared to the SORLAWT overexpressing cell line. I performed an unbiased interactome screen of SORLAWT and SORLAN1358S and identified several interaction partners of SORLA related to endosomal trafficking that may be affected by the N1358S mutation. Furthermore, the interactome screen revealed the exosomal protein MFG-E8 as a previously unknown interaction partner of SORLA, suggesting a novel role of SORLA in sorting MFG-E8, and potentially Aβ, for exosome secretion. Taken together, my data elucidate the cell type specific transcriptional regulation of SORT1 and the pathological mechanism of the EOAD-associated N1358S mutation in SORL1. They also highlight the importance of validating disease-associated genetic variants and substantiate iPSC-derived neurons and hepatocytes for modeling VPS10P domain receptor function.