The researchers managed to pinpoint the variant by studying people of different ancestries, a feat they say highlights the importance of conducting clinical trials that include people of diverse descents. The results are published in the journal Nature Genetics.
In addition to old age and certain underlying diseases, genetics can influence whether we become severely affected or only suffer mild illness from COVID-19. Previous studies on mainly people of European ancestry have found that individuals carrying a particular segment of DNA have a 20 percent lower risk of developing a critical COVID-19 infection.
This region of DNA is, however, packed with numerous genetic variants, which makes it challenging to disentangle the exact protective variant that could potentially serve as a target for medical treatment against severe COVID-19 infection.
Studied people of different ancestries
To identify this specific gene variant, researchers in the current study looked for individuals carrying only parts of this DNA segment. Since the Neandertal inheritance occurred after the ancient migration out of Africa, the researchers saw a potential in focusing on individuals with African ancestry who lack heritage from the Neanderthals and therefore also the majority of this DNA segment.
The researchers found that individuals of predominantly African ancestry had the same protection as those of European ancestry, which allowed them to pinpoint a specific gene variant of particular interest.
“The fact that individuals of African descent had the same protection allowed us to identify the unique variant in the DNA that actually protects from COVID-19 infection,” says Jennifer Huffman, the first author of study and a researcher at the VA Boston Healthcare System in the U.S.
The analysis included a total of 2,787 hospitalized COVID-19 patients of African ancestry and 130,997 people in a control group from six cohort studies. Eighty percent of individuals of African ancestry carried the protective variant. The outcome was compared with a previous, larger metastudy of individuals of European heritage.
According to the researchers, the protective gene variant (rs10774671-G) determines the length of the protein encoded by the gene OAS1. Prior studies have shown that the longer variant of the protein is more effective at breaking down SARS-CoV-2, the virus causing the disease COVID-19.
“That we are beginning to understand the genetic risk factors in detail is key to developing new drugs against COVID-19,” says co-author Brent Richards, senior investigator at the Lady Davis Institute of the Jewish General Hospital and professor at McGill University in Canada.
Underscores need for diversity
The COVID-19 pandemic has spurred considerable collaboration among researchers in different parts of the world, which has made it possible to study genetic risk factors in a wider diversity of individuals than in many previous studies. Even so, the majority of all clinical research is still being done on individuals of predominantly European descent.
“This study shows how important it is to include individuals of different ancestries. If we had only studied one group, we would not have been successful in identifying the gene variant in this case,” says the study’s corresponding author Hugo Zeberg, assistant professor at the Department of Neuroscience at Karolinska Institutet.
Interindividual variability in response to SARS-CoV-2 infection resulting in the range from asymptomatic to fatal disease requires an urgent understanding of the underlying molecular mechanisms. Results of our cases-case analyses in individuals of European ancestry confirmed the genetic association within chr12q24.13 locus initially reported comparing patients with COVID-19 to the general population2.
We fine-mapped the risk of hospitalized disease to a 19Kb-haplotype that included 76 OAS1 variants forming a 95% credible set for association in this region. In functional analyses, we showed that alleles of two OAS1 variants – a splicing variant rs10774671 and a missense variant rs1131454 in exon 3 regulate expression levels of OAS1 through NMD.
We demonstrated that the OAS1 haplotype most targeted by NMD was associated with the lowest baseline OAS1 expression, increased risk of severe COVID-19, and impaired spontaneous clearance of SARS-CoV-2, which could be compensated by early treatment with pegIFN-λ1.
A 185Kb-haplotype was reported as introgressed into the genomes of modern humans from the Neandertal and Denisova lineages of non-African ancestors5-7,20. It was hypothesized that the introgression of this haplotype was beneficial, perhaps by protecting humans from select pathogens; several associations with immune-related phenotypes reported for individual variants comprising this haplotype8-16 supported this idea.
However, all these studies explored only one or a few variants, not providing sufficient clarity on the extent of the protective part of the Neandertal haplotype or the plausible molecular mechanisms underlying its protective effect. The analysis of COVID-19-related outcomes affords a unique opportunity to address these questions.
Based on the COVID-19 GWAS meta-analysis, a 97Kb block of linked variants (r2>0.8) associated with hospitalized disease compared to the general population was nominated as a potentially protective part of the Neandertal haplotype20. Based on association and credible set analyses, we further narrowed this fragment to a 19Kb-region, which included 76 OAS1 variants in r2>0.9 with the lead variant rs10774671 and comparably associated with the hospitalized disease.
Because rs10774671-G allele, associated with protection from severe COVID-191,2, creates the OAS1-p46 protein isoform (vs. OAS1-p42 and some other isoforms), this marker can be considered the lead functional variant in the OAS1 credible set8,9. Previous reports21 and our results showed that OAS1-p46 expression is enriched in trans-Golgi compartments, while OAS1-p42 is expressed in the cytosol.
Targeting OAS1 to endomembrane structures may benefit response to pathogens that hide from cytosolic pattern recognition receptors31. Virus-induced formation of complex membrane rearrangements originating from the endoplasmic reticulum was demonstrated as the mechanism used by plus-strand RNA flaviviruses (such as West Nile, hepatitis C, Dengue, Yellow Fever, and Zika viruses) to evade sensing and elimination31. Trans-Golgi localization of OAS1-p46 would support enhanced sensing and clearance of flaviviruses, contributing to the strong differences in antiviral activities observed between OAS1-p42 and p46 isoforms21.
It was hypothesized that coronaviruses and SARS-CoV-2 specifically might use a similar mechanism to evade the immune response32. In embryonic kidney cells stably expressing ACE2 (HEK293-ACE2), SARS-CoV-2 was cleared in the presence of both OAS1 protein isoforms but more efficiently in the presence of OAS1-p46 than OAS1-p4221. We tested the functional effects of rs10774671 and several linked missense OAS1 variants by creating corresponding OAS1 plasmids. In A549-ACE2 cells used as our experimental model, all OAS1 plasmids provided a similar anti-SARS-CoV-2 response.
The presence of the C-terminal OAS1-p46 CaaX motif, which is responsible for trans-Golgi targeting, was insufficient to explain differences in antiviral activities of OAS1-p42 and OAS1-p4621. Furthermore, OAS1-p42 and OAS1-p46 were similarly functional as enzymes activating the antiviral RNaseL pathway21. These protein isoforms overexpressed in A549 cells similarly blocked in-vitro replication of Dengue virus via an RNaseL-dependent mechanism33. Overall, previous and our results suggest that both OAS1-p42 and OAS1-p46 have anti-SARS-CoV-2 activity, which is likely to be context-dependent, such as determined by the requirement of localization to specific cellular compartments for viral sensing, and expression levels of each isoform.
Our expression analyses showed that in all datasets explored, mRNA expression of OAS1-p46 transcript was on average 3.9-fold higher than OAS1-p42. We did not find evidence for transcriptional regulation of OAS1 expression but demonstrated that OAS1 expression is regulated by NMD differentially affecting OAS1 isoforms. By creating an NMD-resistant OAS1-p46 transcript, the non-risk rs10774671-G allele plays the major role in preserving the OAS1 expression. Additionally, we identified rs1131454 within exon 3 of OAS1 as the second variant contributing to the regulation of OAS1 NMD.
Although rs1131454 is a missense variant in exon 3 (Gly162Ser), it did not appear to have a functional impact on the enzymatic activity of recombinant OAS1 proteins34 and in our anti-SARS-CoV-2 assays. Instead, the non-risk rs1131454-G allele of this variant creates an ESE/ESS to include the Short vs. Long forms of exon 3. Although NMD targets all OAS1-p42 transcripts, transcripts with rs1131454-G allele are partially rescued from NMD. Because the OAS1-p46 transcripts always carry the rs1131454-G allele, they are most NMD-resistant. Trans-Golgi accumulation of OAS1-p46 coupled with its increased expression due to the combined effects of non-risk rs10774671-G and rs1131454-G alleles may significantly enhance the antiviral activity of OAS1, offering a functional mechanism underlying the association of the Neandertal haplotype with COVID-19 outcomes.
We observed a decrease in SARS-CoV-2 expression after treating cells with interferons (either IFNβ or IFN-λ cocktail) before or after infection, suggesting that interferons can overcome insufficient viral clearance. Indeed, in a clinical trial with pegIFN-λ1, we observed that spontaneous but not the interferon-induced clearance of SARS-CoV-2 was associated with OAS1 haplotypes.
Thus, our results suggest that spontaneous clearance of SARS-CoV-2 impaired in the presence of specific OAS1 variants can be compensated for by treatment with interferons. Although this treatment accelerated viral clearance in all patients, patients with the risk OAS1 haplotype (AAA for rs1131454-A, rs10774671-A, and rs2660-A) would benefit from this treatment the most because of their impaired ability to clear the virus without treatment.
In our clinical trial, patients were treated with a single subcutaneous injection of pegIFN-λ118. Due to the restricted expression of receptors, IFN-λ1, a type III interferon, is well tolerated without causing systemic side effects often associated with administration of type I interferons35. Inhaled nebulized type I interferons, IFNβ-1a36 and IFN-α2b37,38, are also tested as an early treatment for SARS-CoV-2 infection, with promising results.
The strengths of our study include genetic analyses evaluating outcomes in individuals with laboratory-confirmed COVID-19 and integrated analyses of multiple genomic datasets (e.g., RNA-seq, ATAC-seq, ChIP-seq, and Hi-C) supported by experimental testing of our hypotheses. In addition, we analyzed multi-variant haplotypes in association studies with COVID-19 severity and a clinical trial with pegIFN-λ1.
The extensive multi-method investigation provides strong plausibility for our findings. Of multiple associated genetic variants, we identified rs10774671 and rs1131454 as most functional within OAS1 for clearance of SARS-CoV-2 and COVID-19 severity. However, we cannot exclude additional functional variants, especially in non-European populations, in which we had low statistical power for genetic analyses. Some of the associations with severe COVID-19 might be related to the function of OAS3, which we did not explore.
Specifically, in Europeans, we identified two missense OAS3 variants that might be independently contributing to the risk of severe disease. The activity of OAS1 and OAS3 enzymes might be synergistic but cell type and condition-dependent, which further genetic and functional studies should explore. Our study did not explore the genetics of immune response to SARS-CoV-2 variants.
Overall, we propose that non-risk alleles of two variants (rs10774671-G and rs1131454-G) protect OAS1 transcripts from NMD (Figure 8). The rs10774671-G allele has the major role by generating the OAS1-p46 isoform, while rs1131454-G additionally and independently contributes by creating an ESE that increases inclusion of Long exon 3, thus protecting both OAS1-p46 and OAS1-p42 from elimination by NMD. The non-risk G alleles of both variants create the most abundant and NMD-resistant OAS1 isoform (OAS1-p46-Long).
In contrast, the risk A alleles of both variants create the NMD-vulnerable and low-expressed isoform (OAS1-p42), while the haplotype with rs10774671-A but rs1131454-G allele creates the OAS1-p42 isoform with an intermediate NMD resistance and expression levels (Figure 8). Deficient spontaneous clearance in individuals carrying these risk OAS1 haplotypes can be compensated by early treatment with interferons, which should be further explored in clinical trials.
reference link :https://www.medrxiv.org/content/10.1101/2021.07.09.21260221v1.full
More information: Hugo Zeberg, Multi-ancestry fine mapping implicates OAS1 splicing in risk of severe COVID-19, Nature Genetics (2022). DOI: 10.1038/s41588-021-00996-8. www.nature.com/articles/s41588-021-00996-8