Understanding the complex genetic architecture connecting rheumatoid arthritis, osteoporosis and inflammation: discovering causal pathways
Kasher M., Williams FMK., Freidin MB., Malkin I., Cherny SS., Livshits G., Benjamin E., Chasman DI., Dehghan A., Ahluwalia TS., Meigs J., Tracy R., Alizadeh BZ., Ligthart S., Bis J., Eiriksdottir G., Pankratz N., Gross M., Rainer A., Snieder H., Wilson JG., Psaty BM., Dupuis J., Prins B., Vaso U., Stathopoulou M., Franke L., Lehtimaki T., Koenig W., Jamshidi Y., Siest S., Abbasi A., Uitterlinden AG., Abdollahi M., Schnabel R., Schick UM., Nolte IM., Kraja A., Hsu YH., Tylee DS., Zwicker A., Uher R., Davey-Smith G., Morrison AC., Hicks A., Van Duijn CM., Ward-Caviness C., Boerwinkle E., Rotter J., Rice K., Lange L., Perola M., De Geus E., Morris AP., Makela KM., Stacey D., Eriksson J., Frayling TM., Slagboom EP.
Rheumatoid arthritis (RA) and osteoporosis (OP) are two comorbid complex inflammatory conditions with evidence of shared genetic background and causal relationships. We aimed to clarify the genetic architecture underlying RA and various OP phenotypes while additionally considering an inflammatory component, C-reactive protein (CRP). Genome-wide association study summary statistics were acquired from the GEnetic Factors for OSteoporosis Consortium, Cohorts for Heart and Aging Research Consortium and UK Biobank. Mendelian randomization (MR) was used to detect the presence of causal relationships. Colocalization analysis was performed to determine shared genetic variants between CRP and OP phenotypes. Analysis of pleiotropy between traits owing to shared causal single nucleotide polymorphisms (SNPs) was performed using PL eiotropic A nalysis under CO mposite null hypothesis (PLACO). MR analysis was suggestive of horizontal pleiotropy between RA and OP traits. RA was a significant causal risk factor for CRP (β = 0.027, 95% confidence interval = 0.016-0.038). There was no evidence of CRP→OP causal relationship, but horizontal pleiotropy was apparent. Colocalization established shared genomic regions between CRP and OP, including GCKR and SERPINA1 genes. Pleiotropy arising from shared causal SNPs revealed through the colocalization analysis was all confirmed by PLACO. These genes were found to be involved in the same molecular function 'protein binding' (GO:0005515) associated with RA, OP and CRP. We identified three major components explaining the epidemiological relationship among RA, OP and inflammation: (1) Pleiotropy explains a portion of the shared genetic relationship between RA and OP, albeit polygenically; (2) RA contributes to CRP elevation and (3) CRP, which is influenced by RA, demonstrated pleiotropy with OP.