Soluble SORL1 in cerebrospinal fluid as a marker for functional impact of rare SORL1 variants.
Background: As part of the retromer, sortilin-related receptor (SORL1) sorts cargo proteins away from the endosome to the trans-Golgi network or to the cell surface, where SORL1 is cleaved into sSORL1 and shed into the interstitial fluid and CSF. SORL1's cargo includes APP and A{beta}, which may explain why protein-truncating genetic variants (PTVs) in SORL1 are observed almost exclusively in Alzheimer's Disease (AD) patients, and that rare, predicted pathogenic missense variants have been associated with a 10-fold increased risk of AD. However, functional evidence supporting variant pathogenicity is warranted. Here we investigated whether soluble SORL1 (sSORL1) concentrations in cerebrospinal fluid (CSF) offer a potential in vivo biomarker to support impaired SORL1 function in variant carriers. Methods: Using an ELISA assay for SORL1 (ABCAM), we determined sSORL1 concentrations in CSF from 218 participants of the Alzheimer Dementia Cohort (ADC) (54% females). We compared sSORLA in CSF derived from 90 carriers of diverse SORL1 variants with concentrations observed in 78 SORL1-WT AD patients, and 50 SORL1-WT controls for whom CSF-pTau-181, CSF-tTau, CSF-A{beta}42 concentrations were available. In a subset of 36 individuals, we used Western blotting (WB) to validate sSORL1 concentrations as determined by ELISA. Results: CSF-sSORL1 concentrations did not differ between SORL1-WT AD patients and controls. While CSF-sSORL1 did not correlate with CSF-A{beta}42 concentrations in SORL1-WT AD patients (p=0.62), it correlated with sCSF-ptau-181 (p=9.7x10-6). The mean CSF-sSORL1 concentration in the SORL1 WT AD cases and controls was 466 pg/ml, with wide variance (SD = 133). Comparatively, concentrations were significantly lower in PTV carriers (260 pg/ml, p=3.2x10-7) and in carriers of predicted damaging SORL1 missense variants (323 pg/ml, p=2.4x10-7). CSF-sSORL1 concentrations measured by ELISA correlated strongly with concentrations estimated by WB ({rho} =0.552; p=5.0x10-4). Conclusion: CSF-sSORL1 increases with CSF-ptau, suggesting that increased SORL1-retromer activity may serve to rescue cellular stress associated with AD-related processes. However, impairing SORL1 genetic variants may preclude SORL1 trafficking to the cell surface, as supported by lower CSF-sSORL1 protein concentrations in carriers. While further refinements are necessary, we present first evidence for the applicability of ELISA-based quantification of sSORL1 in CSF to evaluate the functional impact of rare SORL1 variants.