Recently, we reported oligoadenylate synthetase 1 (OAS1) contributed to the risk of Alzheimer’s disease, by its enrichment in transcriptional networks expressed by microglia. However, the function of OAS1 within microglia was not known. Using genotyping from 1313 individuals with sporadic Alzheimer’s disease and 1234 control individuals, we confirm the OAS1 variant, rs1131454, is associated with increased risk for Alzheimer’s disease. The same OAS1 locus has been recently associated with severe coronavirus disease 2019 (COVID-19) outcomes, linking risk for both diseases. The single nucleotide polymorphisms rs1131454(A) and rs4766676(T) are associated with Alzheimer’s disease, and rs10735079(A) and rs6489867(T) are associated with severe COVID-19, where the risk alleles are linked with decreased OAS1 expression. Analysing single-cell RNA-sequencing data of myeloid cells from Alzheimer’s disease and COVID-19 patients, we identify co-expression networks containing interferon (IFN)-responsive genes, including OAS1, which are significantly upregulated with age and both diseases. In human induced pluripotent stem cell-derived microglia with lowered OAS1 expression, we show exaggerated production of TNF-α with IFN-γ stimulation, indicating OAS1 is required to limit the pro-inflammatory response of myeloid cells. Collectively, our data support a link between genetic risk for Alzheimer’s disease and susceptibility to critical illness with COVID-19 centred on OAS1, a finding with potential implications for future treatments of Alzheimer’s disease and COVID-19, and development of biomarkers to track disease progression.

Genome-wide association studies of late-onset Alzheimer’s disease (AD) have highlighted the importance of variants associated with genes expressed by the innate immune system in determining risk for AD. Recently, we and others have shown that genes associated with variants that confer risk for AD are significantly enriched in transcriptional networks expressed by amyloid-responsive microglia. This allowed us to predict new risk genes for AD, including the interferon-responsive oligoadenylate synthetase 1 (OAS1). However, the function of OAS1 within microglia and its genetic pathway are not known. Using genotyping from 1,313 individuals with sporadic AD and 1,234 control individuals, we confirm that the OAS1 variant, rs1131454, is associated with increased risk for AD and decreased OAS1 expression. Moreover, we note that the same locus was recently associated with critical illness in response to COVID-19, linking variants that are associated with AD and a severe response to COVID-19. By analysing single-cell RNA-sequencing (scRNA-seq) data of isolated microglia from APPNL-G-F knock-in and wild-type C57BL/6J mice, we identify a transcriptional network that is significantly upregulated with age and amyloid deposition, and contains the mouse orthologue Oas1a, providing evidence that Oas1a plays an age-dependent function in the innate immune system. We identify a similar interferon-related transcriptional network containing OAS1 by analysing scRNA-seq data from human microglia isolated from individuals with AD. Finally, using human iPSC-derived microglial cells (h-iPSC-Mg), we see that OAS1 is required to limit the pro-inflammatory response of microglia. When stimulated with interferon-gamma (IFN-γ), we note that cells with lower OAS1 expression show an exaggerated pro-inflammatory response, with increased expression and secretion of TNF-α. Collectively, our data support a link between genetic risk for AD and susceptibility to critical illness with COVID-19 centred on OAS1 and interferon signalling, a finding with potential implications for future treatments of both AD and COVID-19, and the development of biomarkers to track disease progression.

Ageing is the greatest global healthcare challenge, as it underlies age-related functional decline and is the primary risk factor for a range of common diseases, including neurodegenerative conditions such as Alzheimer's disease (AD). However, the molecular mechanisms defining chronological age versus biological age, and how these underlie AD pathogenesis, are not well understood. The objective of this study was to integrate common human genetic variation associated with human lifespan or AD from Genome-Wide Association Studies (GWAS) with co-expression networks altered with age in the central nervous system, to gain insights into the biological processes which connect ageing with AD and lifespan. Initially, we identified common genetic variation in the human population associated with lifespan and AD by performing a gene-based association study using GWAS data. We also identified preserved co-expression networks associated with age in the brains of C57BL/6J mice from bulk and single-cell RNA-sequencing (RNA-seq) data, and in the brains of humans from bulk RNA-seq data. We then intersected the human gene-level common variation with these co-expression networks, representing the different cell types and processes of the brain. We found that genetic variation associated with AD was enriched in both microglial and oligodendrocytic bulk RNA-seq gene networks, which show increased expression with ageing in the human hippocampus, in contrast to synaptic networks which decreased with age. Further, longevity-associated genetic variation was modestly enriched in a single-cell gene network expressed by homeostatic microglia. Finally, we performed a transcriptome-wide association study (TWAS), to identify and confirm new risk genes associated with ageing that show variant-dependent changes in gene expression. In addition to validating known ageing-related genes such as APOE and FOXO3, we found that Caspase 8 (CASP8) and APOC1 show genetic variation associated with longevity. We observed that variants contributing to ageing and AD balance different aspects of microglial function suggesting that ageing-related processes affect multiple cell types in the brain. Specifically, changes in homeostatic microglia are associated with lifespan, and allele-dependent expression changes in age-related genes control microglial activation and myelination influencing the risk of developing AD. We identified putative molecular drivers of these genetic networks, as well as module genes whose expression in relevant human tissues are significantly associated with AD-risk or longevity, and may drive "inflammageing." Our study also shows allele-dependent expression changes with ageing for genes classically involved in neurodegeneration, including MAPT and HTT, and demonstrates that PSEN1 is a prominent member/hub of an age-dependent expression network. In conclusion, this work provides new insights into cellular processes associated with ageing in the brain, and how these may contribute to the resilience of the brain against ageing or AD-risk. Our findings have important implications for developing markers indicating the physiological age and pre-pathological state of the brain, and provide new targets for therapeutic intervention.

Ageing is the greatest global healthcare challenge, as it underlies age-related functional decline and is the primary risk factor for a range of common diseases, including neurodegenerative conditions such as Alzheimer’s disease (AD). However, the molecular mechanisms defining chronological age versus biological age, and how these underlie AD pathogenesis, are not well understood. The objective of this study was to integrate common human genetic variation associated with human lifespan or AD from Genome-Wide Association Studies (GWAS) with co-expression networks altered with age in the central nervous system, to gain insights into the biological processes which connect ageing with AD and lifespan. Initially, we identified common genetic variation in the human population associated with lifespan and AD by performing a gene-based association study using GWAS data. We also identified preserved co-expression networks associated with age in the brains of C57BL/6J mice from bulk and single-cell RNA-sequencing (RNA-seq) data, and in the brains of humans from bulk RNA-seq data. We then intersected the human gene-level common variation with these co-expression networks, representing the different cell types and processes of the brain. We found that genetic variation associated with AD was enriched in both microglial and oligodendrocytic bulk RNA-seq gene networks, which show increased expression with ageing in the human hippocampus, in contrast to synaptic networks which decreased with age. Further, longevity-associated genetic variation was modestly enriched in a single-cell gene network expressed by homeostatic microglia. Finally, we performed a transcriptome-wide association study (TWAS), to identify and confirm new risk genes associated with ageing that show variant-dependent changes in gene expression. In addition to validating known ageing-related genes such as APOE and FOXO3, we found that Caspase 8 (CASP8) and APOC1 show genetic variation associated with longevity. We observed that variants contributing to ageing and AD balance different aspects of microglial function suggesting that ageing-related processes affect multiple cell types in the brain. Specifically, changes in homeostatic microglia are associated with lifespan, and allele-dependent expression changes in age-related genes control microglial activation and myelination influencing the risk of developing AD. We identified putative molecular drivers of these genetic networks, as well as module genes whose expression in relevant human tissues are significantly associated with AD-risk or longevity, and may drive “inflammageing.” Our study also shows allele-dependent expression changes with ageing for genes classically involved in neurodegeneration, including MAPT and HTT, and demonstrates that PSEN1 is a prominent member/hub of an age-dependent expression network. In conclusion, this work provides new insights into cellular processes associated with ageing in the brain, and how these may contribute to the resilience of the brain against ageing or AD-risk. Our findings have important implications for developing markers indicating the physiological age and pre-pathological state of the brain, and provide new targets for therapeutic intervention.