Prioritization of therapeutic targets for amyotrophic lateral sclerosis using protein-wide Mendelian randomization analysis

Background and Objectives Amyotrophic lateral sclerosis (ALS) is a rapidly progressing neurodegenerative disease with an increasing global burden. Available treatments for ALS present marginal efficacy. To identify novel candidate therapeutic targets for ALS, we conducted a proteome-wide Mendelian randomization (MR) study.

Methods We leveraged data from the largest summary statistics for ALS to date (27,205 patients with ALS and 110,881 controls). Genetic instruments of more than 4,000 proteins defined by cis-protein quantitative loci (pQTL) genetic instruments on plasma and cerebrospinal fluid (CSF) were obtained from Fenland (discovery, n=10,709), deCODE (replication, n=35,559), and a recently published dataset (replication, n=971). To investigate the causal ALS-associated proteins, proteome-wide Mendelian randomization based on summary-data-based MR (SMR and multi-SNP-based SMR) were performed. Then, two-sample MR analyses using five additional methods were conducted as sensitivity analyses. To further address the linkage disequilibrium bias, heterogeneity in dependent instruments test and colocalization analyses were performed. Steiger filtering and bi-directional MR analyses were conducted to address the potential reverse causality. Four drug target datasets were searched to extract druggability profiles for candidate target proteins. In addition, we carried out a case-control study involving up to 21 patients with ALS and 21 matched controls to assess the protein levels difference in CSF for evidence triangulation.

Results Genetically predicted levels of six circulating proteins were associated with incident ALS in primary SMR analysis. After removing proteins with any linkage disequilibrium bias, SHBG, SIGLEC7, and SIGLEC9 presented consistent associations with ALS risk, supported by medium-to-high colocalization across both plasma pQTL datasets. In CSF, higher level of SHBG was also causally associated with the risk of ALS. There was no reverse causality detected. The case-control study using CSF proteomics conducted in our center observed consistent alteration in the levels of SHBG and SIGLEC7 with MR prediction, further suggesting their functionally relevant to ALS as potential druggable targets.

Discussion Combined with the findings from MR and our observational study, we prioritize SHBG, SIGLEC7, and SIGLEC9 as drug candidate proteins for ALS, and further studies are needed to verify our findings and elucidate the underlying mechanism.