Multi-ancestry Mendelian randomization of omics traits revealing drug targets of COVID-19 severity

Single-cell transcriptome-wide Mendelian randomization and colocalization reveals immune-mediated regulatory mechanisms and drug targets for COVID-19

BACKGROUND

COVID-19 continues to show long-term impacts on our health. Limited effective immune-mediated antiviral drugs have been launched.

METHODS

We conducted a Mendelian randomization (MR) and colocalization analysis using 26,597 single-cell expression quantitative trait loci (sc-eQTL) to proxy effects of expressions of 16,597 genes in 14 peripheral blood immune cells and tested them against four COVID-19 outcomes from COVID-19 Genetic Housing Initiative GWAS meta-analysis Round 7. We also carried out additional validations including colocalization, linkage disequilibrium check and host-pathogen interactome predictions. We integrated MR findings with clinical trial evidence from several drug gene related databases to identify drugs with repurposing potential. Finally, we developed a tier system and identified immune-cell-based prioritized drug targets for COVID-19.

FINDINGS

We identified 132 putative causal genes in 14 immune cells (343 MR associations) for COVID-19, with 58 genes that were not reported previously. 145 (73%) gene-COVID-19 pairs showed effects on COVID-19 in only one immune cell type, which implied widespread immune-cell specific effects. For pathway analyses, we found the putative causal genes were enriched in natural killer (NK) recruiting cells but de-enriched in NK cells. Using a deep learning model, we found 107 (81%) of the putative causal genes (41 novel genes) were predicted to interact with SARS-COV-2 proteins. Integrating the above evidence with drug trial information, we developed a tier system and prioritized 37 drug targets for COVID-19.

INTERPRETATION

Our study showcased the central role of immune-mediated regulatory mechanisms for COVID-19 and prioritized drug targets that might inform interventions for viral infectious diseases.

FUNDING

This work was supported by grants from the National Key Research and Development Program of China (2022YFC2505203).

PLASMA PROTEOME, METABOLOME MENDELIAN RANDOMIZATION IDENTIFIES SEPSIS THERAPEUTIC TARGETS

ABSTRACT

Background : The interrelation between the plasma proteome and plasma metabolome with sepsis presents a multifaceted dynamic that necessitates further research to elucidate the underlying causal mechanisms. Methods : Our investigation used public genome-wide association study data to explore the relationships among the plasma proteome, metabolome, and sepsis, considering different sepsis subgroup. Initially, two-sample Mendelian randomization established causal connections between the plasma proteome and metabolome with sepsis. Subsequently, multivariate and iterative Mendelian randomization analyses were performed to understand the complex interactions in plasma during sepsis. The validity of these findings was supported by thorough sensitivity analyses. Result : The study identified 25 plasma proteins that enhance risk and 34 that act as protective agents in sepsis. After P value adjustment (0.05/1306), ICAM5 emerged with a positive correlation to sepsis susceptibility ( P value = 2.14E-05, OR = 1.10, 95% CI = 1.05-1.15), with this significance preserved across three sepsis subgroup examined. Additionally, 29 plasma metabolites were recognized as risk factors, and 15 as protective factors for sepsis outcomes. After P value adjustment (0.05/997), elevated levels of 1,2,3-benzenetriol sulfate (2) was significantly associated with increased sepsis risk ( P value = 3.37E-05, OR = 1.18, 95% CI = 1.09-1.28). Further scrutiny revealed that this plasma metabolite notably augments the abundance of ICAM5 protein ( P value = 3.52E-04, OR = 1.11, 95% CI = 1.04-1.17), devoid of any detected heterogeneity, pleiotropy, or reverse causality. Mediated Mendelian randomization revealed ICAM5 mediated 11.9% of 1,2,3-benzenetriol sulfate (2)'s total effect on sepsis progression. Conclusion : This study details the causal link between the plasma proteome and metabolome with sepsis, highlighting the roles of ICAM5 and 1,2,3-benzenetriol sulfate (2) in sepsis progression, both independently and through crosstalk.

Altered IL-6 signalling and risk of tuberculosis: a multi-ancestry mendelian randomisation study

BACKGROUND

The role of IL-6 responses in human tuberculosis risk is unknown. IL-6 signalling inhibitors, such as tocilizumab, are thought to increase the risk of progression to tuberculosis, and screening for latent Mycobacterium tuberculosis infection before using these drugs is widely recommended. We used single nucleotide polymorphisms (SNPs) in and near the IL-6 receptor gene (IL6R), including the non-synonymous variant, rs2228145, for which the C allele contributes to reduced classic (cis) IL-6 signalling activity, to test the hypothesis that altered IL-6 signalling is causally associated with the risk of developing tuberculosis.

METHODS

We performed a meta-analysis of genome-wide association studies (GWAS) published in English from database inception to Jan 1, 2024. GWAS were identified from the European Bioinformatics Institute, MRC Integrative Epidemiology Unit catalogues, and MEDLINE, selecting publicly available studies for which tuberculosis was an outcome and that included the IL6R rs2228145 SNP. Using each study's population-level summary statistics, effect estimates were extracted for each additional copy of the C allele of rs2228145. We used these estimates to perform multi-ancestry, two-sample mendelian randomisation analyses to estimate the causal effect of reduced IL-6 signalling on tuberculosis. Our primary analyses used rs2228145-C as a genetic instrument, weighted on C-reactive protein (CRP) reduction as a measure of the effect on IL-6 signalling. We also took an alternative, ancestry-specific, multiple SNP approach using IL-6 receptor plasma protein as an exposure. Additionally, we compared the effects of rs2228145 in tuberculosis with those in critical COVID-19, rheumatoid arthritis, Crohn's disease, and coronary artery disease using the summary statistics extracted from GWAS.

FINDINGS

17 GWAS were included, collating data for 19 302 individuals with tuberculosis (cases) and 1 019 821 population controls across multiple ancestries. For each additional rs2228145-C allele, the odds of tuberculosis reduced (odds ratio [OR] 0·94 [95% CI 0·92-0·97]; p=6·8 × 10-6). Multi-ancestry mendelian randomisation analyses supported these findings, with decreased odds of tuberculosis associated with readouts of reduced IL-6 signalling (0·52 [0·39-0·69] for each natural log CRP decrease; p=6·8 × 10-6), with weak evidence of heterogeneity (I2=0·315; p=0·11). Ancestry-specific, multiple SNP mendelian randomisation using increase in IL-6 receptor plasma protein as an exposure revealed a similar reduced risk of tuberculosis (OR 0·94 [95% CI 0·93-0·96]; p=2·4 × 10-10). The protective effects on tuberculosis seen with rs2228145-C were similar in size and direction to those observed in critical COVID-19 (0·66 [0·50-0·86]), Crohn's disease (0·57 [0·44-0·74]), and rheumatoid arthritis (0·45 [0·36-0·58]), all of which benefit from the therapeutic effects of IL-6 antagonism.

INTERPRETATION

Our findings propose a causal relationship between reduced IL-6 signalling and lower risk of tuberculosis, akin to the effect seen in other IL-6 mediated diseases. This study suggests that IL-6 antagonists do not increase the risk of tuberculosis but rather should be investigated as therapeutic adjuncts in its treatment.

FUNDING

UK National Institute for Health and Care Research, Wellcome Trust, EU European Regional Development Fund, the Welsh Government, and UK Research and Innovation.

CCN3/NOV as a potential therapeutic target for diverticular disease: A proteome-wide Mendelian randomization study

Genome-wide association studies (GWAS) identified over 100 susceptibility loci and candidate causal genes for diverticular disease (DD) at the transcriptional level. However, effective therapeutics or preventions based on underlying disease mechanisms remain to be elucidated. In this study, we explored potential causal genes for DD at the protein level. We used 2 GWAS summary statistics of DD; 1 was obtained from the United Kingdom Biobank (UKBB) with 31,917 cases and 419,135 controls, and the other from the FinnGen consortium with 30,649 cases and 301,931 controls. For the primary analysis, we employed proteome-wide Mendelian randomization (MR) studies using 738 cis-acting protein quantitative trait loci (pQTLs) for 735 plasma proteins from the 5 published studies. For external validation, we conducted 2-sample MR analyses using plasma pQTLs of the screened proteins from another study by deCODE genetics. Moreover, we performed a series of sensitivity analyses including reverse MR and Bayesian colocalization tests. The primary MR identified 4 plasma proteins that were associated with DD risk including CCN3/NOV (odds ratio [OR], 0.98; 95% confidence interval [CI], 0.97-0.99; P = 1.2 × 10-11 for UKBB. OR, 0.73; 95% CI, 0.66-0.81; P = 7.2 × 10-10 for FinnGen). The validation MR well replicated the primary result of CCN3/NOV (OR, 0.95; 95% CI, 0.93-0.96; P = 1.9 × 10-11 for UKBB. OR, 0.43; 95% CI, 0.33-0.56; P = 7.0 × 10-10 for FinnGen). Sensitivity analyses supported the causal association. We prioritized plasma CCN3/NOV protein as a protective factor for DD for follow-up functional studies to elucidate the disease mechanisms and therapeutics.

Identifying genetic determinants of sarcopenia-related traits: a Mendelian randomization study of druggable genes

BACKGROUND

Sarcopenia, characterized by progressive muscle mass and function loss, particularly affects the elderly, and leads to severe consequences such as falls and mortality. Despite its prevalence, targeted pharmacotherapies for sarcopenia are lacking. Utilizing large-sample genome-wide association studies (GWAS) data is crucial for cost-effective drug discovery.

METHODS

Herein, we conducted four studies to understand the putative causal effects of genetic components on muscle mass and function. Study 1 employed a two-sample Mendelian randomization (MR) on 15,944 potential druggable genes, investigating their potential causality with muscle quantity and quality in a European population (N up to 461,089). Study 2 validated MR results through sensitivity analyses and colocalization analyses. Study 3 extended validation across other European cohorts, and study 4 conducted quantitative in vivo verification.

RESULTS

MR analysis revealed significant causality between four genes (BLOC-1 related complex subunit 7, BORCS7; peptidase m20 domain containing 1, PM20D1; nuclear casein kinase and cyclin dependent kinase substrate 1, NUCKS1 and ubiquinol-cytochrome c reductase complex assembly factor 1, UQCC1) and muscle mass and function (p-values range 5.98 × 10-6 to 9.26 × 10-55). To be specific, BORCS7 and UQCC1 negatively regulated muscle quantity and quality, whereas enhancing PM20D1 and NUCKS1 expression showed promise in promoting muscle mass and function. Causal relationships remained robust across sensitivity analyses, with UQCC1 exhibiting notable colocalization effects (PP·H4 93.4 % to 95.8 %). Further validation and in vivo replication verified the potential causality between these genes and muscle mass as well as function.

CONCLUSIONS

Our druggable genome-wide MR analysis identifies BORCS7, PM20D1, NUCKS1, and UQCC1 as causally associated with muscle mass and function. These findings offer insights into the genetic basis of sarcopenia, paving the way for these genes to become promising drug targets in mitigating this debilitating condition.

Exploring the gut microbiota's effect on temporomandibular joint disorder: a two-sample Mendelian randomization analysis

Background

Temporomandibular joint disorders (TMD) are highly prevalent among people. Numerous investigations have revealed the impact of gut microbiota in many diseases. However, the causal relationship between Temporomandibular joint disorders and gut microbiota remains unclear.

Methods

Genome-Wide Association Studies (GWAS) refer to the identification of sequence variations, namely single nucleotide polymorphisms (SNPs), existing across the entire human genome. GWAS data were collected on gut microbiota and TMD. Then, instrumental variables were screened through F-values and removal of linkage disequilibrium. These SNPs underwent mendelian analysis using five mathematical models. Sensitivity analysis was conducted to further verify the stability of the results. Pathogenic factors of TMD mediate the causal relationship between gut microbiota and TMD were explored through a two-step Mendelian randomization analysis. Finally, reverse mendelian analysis was conducted to account for potential reverse effects.

Results

The analysis of the data in this article suggests that some gut microbiota, including Coprobacter, Ruminococcus torques group, Catenibacterium, Lachnospiraceae, Turicibacter, Victivallis, MollicutesRF9, Methanobacteriales, Methanobacteriaceae, FamilyXI, Methanobacteria were identified as risk factors, while Peptococcaceae provides protection for TMD.

Conclusion

The research reveals the relation of gut microbiota in TMD. These findings provide insights into the underlying mechanisms and suggest potential therapeutic strategy.

Circulating Proteins and IgA Nephropathy: A Multiancestry Proteome-Wide Mendelian Randomization Study

Key Points

Background

The therapeutic options for IgA nephropathy are rapidly evolving, but early diagnosis and targeted treatment remain challenging. We aimed to identify circulating plasma proteins associated with IgA nephropathy by proteome-wide Mendelian randomization studies across multiple ancestry populations.

Methods

In this study, we applied Mendelian randomization and colocalization analyses to estimate the putative causal effects of 2615 proteins on IgA nephropathy in Europeans and 235 proteins in East Asians. Following two-stage network Mendelian randomization, multitrait colocalization analysis and protein-altering variant annotation were performed to strengthen the reliability of the results. A protein–protein interaction network was constructed to investigate the interactions between the identified proteins and the targets of existing medications.

Results

Putative causal effects of 184 and 13 protein–disease pairs in European and East Asian ancestries were identified, respectively. Two protein–disease pairs showed shared causal effects across them (CFHR1 and FCRL2). Supported by the evidence from colocalization analysis, potential therapeutic targets were prioritized and four drug-repurposing opportunities were suggested. The protein–protein interaction network further provided strong evidence for existing medications and pathways that are known to be therapeutically important.

Conclusions

Our study identified a number of circulating proteins associated with IgA nephropathy and prioritized several potential drug targets that require further investigation.

A robust cis-Mendelian randomization method with application to drug target discovery

Mendelian randomization (MR) uses genetic variants as instrumental variables (IVs) to investigate causal relationships between traits. Unlike conventional MR, cis-MR focuses on a single genomic region using only cis-SNPs. For example, using cis-pQTLs for a protein as exposure for a disease opens a cost-effective path for drug target discovery. However, few methods effectively handle pleiotropy and linkage disequilibrium (LD) of cis-SNPs. Here, we propose cisMR-cML, a method based on constrained maximum likelihood, robust to IV assumption violations with strong theoretical support. We further clarify the severe but largely neglected consequences of the current practice of modeling marginal, instead of conditional genetic effects, and only using exposure-associated SNPs in cis-MR analysis. Numerical studies demonstrated our method's superiority over other existing methods. In a drug-target analysis for coronary artery disease (CAD), including a proteome-wide application, we identified three potential drug targets, PCSK9, COLEC11 and FGFR1 for CAD.

Using genetics and proteomics data to identify proteins causally related to COVID-19, healthspan and lifespan: a Mendelian randomization study

BACKGROUND

COVID-19 pandemic poses a heavy burden on public health and accounts for substantial mortality and morbidity. Proteins are building blocks of life, but specific proteins causally related to COVID-19, healthspan and lifespan have not been systematically examined.

METHODS

We conducted a Mendelian randomization study to assess the effects of 1,361 plasma proteins on COVID-19, healthspan and lifespan, using large GWAS of severe COVID-19 (up to 13,769 cases and 1,072,442 controls), COVID-19 hospitalization (32,519 cases and 2,062,805 controls) and SARS-COV2 infection (122,616 cases and 2,475,240 controls), healthspan (n = 300,477) and parental lifespan (~0.8 million of European ancestry).

RESULTS

We identified 35, 43, and 63 proteins for severe COVID, COVID-19 hospitalization, and SARS-COV2 infection, and 4, 32, and 19 proteins for healthspan, father's attained age, and mother's attained age. In addition to some proteins reported previously, such as SFTPD related to severe COVID-19, we identified novel proteins involved in inflammation and immunity (such as ICAM-2 and ICAM-5 which affect COVID-19 risk, CXCL9, HLA-DRA and LILRB4 for healthspan and lifespan), apoptosis (such as FGFR2 and ERBB4 which affect COVID-19 risk and FOXO3 which affect lifespan) and metabolism (such as PCSK9 which lowers lifespan). We found 2, 2 and 3 proteins shared between COVID-19 and healthspan/lifespan, such as CXADR and LEFTY2, shared between severe COVID-19 and healthspan/lifespan. Three proteins affecting COVID-19 and seven proteins affecting healthspan/lifespan are targeted by existing drugs.

CONCLUSIONS

Our study provided novel insights into protein targets affecting COVID-19, healthspan and lifespan, with implications for developing new treatment and drug repurposing.

Integration of human organoids single-cell transcriptomic profiles and human genetics repurposes critical cell type-specific drug targets for severe COVID-19

Human organoids recapitulate the cell type diversity and function of their primary organs holding tremendous potentials for basic and translational research. Advances in single-cell RNA sequencing (scRNA-seq) technology and genome-wide association study (GWAS) have accelerated the biological and therapeutic interpretation of trait-relevant cell types or states. Here, we constructed a computational framework to integrate atlas-level organoid scRNA-seq data, GWAS summary statistics, expression quantitative trait loci, and gene-drug interaction data for distinguishing critical cell populations and drug targets relevant to coronavirus disease 2019 (COVID-19) severity. We found that 39 cell types across eight kinds of organoids were significantly associated with COVID-19 outcomes. Notably, subset of lung mesenchymal stem cells increased proximity with fibroblasts predisposed to repair COVID-19-damaged lung tissue. Brain endothelial cell subset exhibited significant associations with severe COVID-19, and this cell subset showed a notable increase in cell-to-cell interactions with other brain cell types, including microglia. We repurposed 33 druggable genes, including IFNAR2, TYK2, and VIPR2, and their interacting drugs for COVID-19 in a cell-type-specific manner. Overall, our results showcase that host genetic determinants have cellular-specific contribution to COVID-19 severity, and identification of cell type-specific drug targets may facilitate to develop effective therapeutics for treating severe COVID-19 and its complications.

Sex-stratified genome-wide association and transcriptome-wide Mendelian randomization studies reveal drug targets of heart failure

The prevalence of heart failure (HF) subtypes, which are classified by left ventricular ejection fraction (LVEF), demonstrate significant sex differences. Here, we perform sex-stratified genome-wide association studies (GWASs) on LVEF and transcriptome-wide Mendelian randomization (MR) on LVEF, all-cause HF, HF with reduced ejection fraction (HFrEF), and HF with preserved ejection fraction (HFpEF). The sex-stratified GWASs of LVEF identified three sex-specific loci that were exclusively detected in the sex-stratified GWASs. Three drug target genes show sex-differential effects on HF/HFrEF via influencing LVEF, with NPR2 as the target gene for the HF drug Cenderitide under phase 2 clinical trial. Our study highlights the importance of considering sex-differential genetic effects in sex-balanced diseases such as HF and emphasizes the value of sex-stratified GWASs and MR in identifying putative genetic variants, causal genes, and candidate drug targets for HF, which is not identifiable using a sex-combined strategy.

Plasma Proteome-Wide Mendelian Randomization Analysis Reveals Biomarkers and Therapeutic Targets for Different Stages of COVID-19

Background. The COVID-19 pandemic caused by the SARS-CoV-2 virus has resulted in a global health crisis with significant morbidity and mortality. While effective vaccinations have been developed, drug treatments for the disease are still required, particularly for different stages of the disease and to combat evolving variants. Identifying reliable biomarkers and potential therapeutic targets for the different stages of COVID-19 is crucial. Methods. Mendelian randomization using the largest publicly available datasets was conducted to identify potential causal plasma proteins for severe COVID-19, hospitalized COVID-19, and SARS-CoV-2 infection. Independent, and strongly associated cis- or pan-pQTLs were used as instrumental variables for each protein. The FDR q-value was used to correct for multiple testing followed by sensitivity analyses, reverse MR and genetic colocalization to ensure the robustness of the results. Results. We identified proteins with strong evidence of causal association with different stages of COVID-19. Some of these proteins were identified previously, such as BGAT and BCAT2, but we also identified the novel proteins, such as KLC1, MRVI1, CACO2, and PCNP. Conclusion. These proteins provide valuable insights into the underlying mechanisms of COVID-19. The identification of these proteins offers new opportunities for developing potential therapeutic targets or biomarkers for the treatment and prevention of COVID-19.

Integrating population-level and cell-based signatures for drug repositioning

Drug repositioning presents a streamlined and cost-efficient way to expand the range of therapeutic possibilities. Furthermore, drugs with genetic evidence are more likely to progress successfully through clinical trials towards FDA approval. Exploiting these developments, single gene-based drug repositioning methods have been implemented, but approaches leveraging the entire spectrum of molecular signatures are critically underexplored. Most multi-gene-based approaches rely on differential gene expression (DGE) analysis, which is prone to identify the molecular consequence of disease and renders causal inference challenging. We propose a framework TReD (Transcriptome-informed Reversal Distance) that integrates population-level disease signatures robust to reverse causality and cell-based drug-induced transcriptome response profiles. TReD embeds the disease signature and drug profile in a high-dimensional normed space, quantifying the reversal potential of candidate drugs in a disease-related cell screen assay. The robustness is ensured by evaluation in additional cell screens. For an application, we implement the framework to identify potential drugs against COVID-19. Taking transcriptome-wide association study (TWAS) results from four relevant tissues and three DGE results as disease features, we identify 37 drugs showing potential reversal roles in at least four of the seven disease signatures. Notably, over 70% (27/37) of the drugs have been linked to COVID-19 from other studies, and among them, eight drugs are supported by ongoing/completed clinical trials. For example, TReD identifies the well-studied JAK1/JAK2 inhibitor baricitinib, the first FDA-approved immunomodulatory treatment for COVID-19. Novel potential candidates, including enzastaurin, a selective inhibitor of PKC-beta which can be activated by SARS-CoV-2, are also identified. In summary, we propose a comprehensive genetics-anchored framework integrating population-level signatures and cell-based screens that can accelerate the search for new therapeutic strategies.

Genetic insight into the putative causal proteins and druggable targets of osteoporosis: a large-scale proteome-wide mendelian randomization study

Background: Osteoporosis is a major causative factor of the global burden of disease and disability, characterized by low bone mineral density (BMD) and high risks of fracture. We aimed to identify putative causal proteins and druggable targets of osteoporosis. Methods: This study utilized the largest GWAS summary statistics on plasma proteins and estimated heel BMD (eBMD) to identify causal proteins of osteoporosis by mendelian randomization (MR) analysis. Different GWAS datasets were used to validate the results. Multiple sensitivity analyses were conducted to evaluate the robustness of primary MR findings. We have also performed an enrichment analysis for the identified causal proteins and evaluated their druggability. Results: After Bonferroni correction, 67 proteins were identified to be causally associated with estimated BMD (eBMD) (p < 4 × 10^-5). We further replicated 38 of the 67 proteins to be associated with total body BMD, lumbar spine BMD, femoral neck BMD as well as fractures, such as RSPO3, IDUA, SMOC2, and LRP4. The findings were supported by sensitivity analyses. Enrichment analysis identified multiple Gene Ontology items, including collagen-containing extracellular matrix (GO:0062023, p = 1.6 × 10^-10), collagen binding (GO:0005518, p = 8.6 × 10^-5), and extracellular matrix structural constituent (GO:0005201, p = 2.7 × 10^-5). Conclusion: The study identified novel putative causal proteins for osteoporosis which may serve as potential early screening biomarkers and druggable targets. Furthermore, the role of plasma proteins involved in collagen binding and extracellular matrix in the development of osteoporosis was highlighted. Further studies are warranted to validate our findings and investigate the underlying mechanism.