Sickle cell anemia (SCA, HBB glu6val) is characterized by a high degree of phenotypic variability: some subjects live relatively normal lives punctuated by sporadic pain crises and acute hospitalizations, while other subjects experience more severe clinical courses with multiple complications. To assess the overall severity of disease, we developed a Bayesian-network based severity score incorporating information about patients' steady-state laboratory data and medical histories. Using this as a phenotype in a genome-wide association study (GWAS) of 1265 subjects from the Cooperative Study of Sickle Cell Disease, we found evidence of genetic association between disease severity and SNPs in ARFGEF2 (p < 5.5 × 10−3 using a novel SNP set enrichment analysis). ARFGEF2 is an important regulator of tumor necrosis factor-α (TNF-α) receptor-1 (TNFR-R1) release from endothelial cells into the plasma, where it functions as a competitive inhibitor of the endothelial membrane-bound TNF-R1 receptor by binding to and sequestering circulating TNF-α. To follow-up this genetic association, we assayed plasma levels of six biomarkers of TNF-α activity, including TNF-α, TNF-R1, P- and E-selectin, vascular cell adhesion molecule-1 (VCAM-1), and intracellular adhesion molecule-1 (ICAM-1), for associations with disease severity in 24 adults with either severe (n=12) or mild (n=12) SCA based on their severity score. Two biomarkers, TNF-R1 and VCAM-1 showed significant increases in severe compared with mild SCA (p < 0.0015 and p < 0.0019, respectively). Interestingly, plasma levels of TNF-R1 and VCAM-1 showed only ambiguous evidence of correlation (p > 0.05 by Spearman correlation, but p ∼ 0.036 by Pearson correlation of log-transformed variables). To evaluate potential sources of confounding, we tested for associations between TNF-R1 or VCAM-1 levels and glomerular filtration rate (GFR), gender, age, and hydroxyurea treatment status. We also examined association with nitric oxide (NO) metabolites (NO2 + NO3), since endothelial cell surface expression of leukocyte adhesion proteins is affected by NO bioavailability. TNF-R1 levels were found to be associated with gender (p < 0.048), but were not significantly associated with GFR, hydroxyurea use, NO2 + NO3 levels, or with age. VCAM-1 levels were not significantly associated with any of these factors. In multivariate models involving SCA severity and gender, the relationship between TNF-R1 and SCA severity remained significant (p < 1.4 × 10-3). The associations of disease severity with both TNF-R1 and VCAM-1 levels suggest that there may be multiple elements of the TNF-α pathway that are important in SCA severity. To look for other severity-related genetic variation in the TNF-α axis, we selected from the GWAS data 47 single nucleotide polymorphisms (SNPs) in the 6 genes coding for the plasma biomarkers (ICAM1, SELE, SELP, TNF, TNFRSF1A, and VCAM1). Of these SNPs, only one (rs 1041163, located in the promoter of VCAM1) was significantly associated with SCA severity (Bayes Factor ∼ 3.5). We also examined 92 SNPs in 9 genes other than ARFGEF2 known to be involved either in the release of these biomarkers or in NO synthesis, including ADAM17, ARF1, ARF3, ERAP-1, IL1B, NOS1, NOS2A, NOS3, and NUCB2. Among these SNPs, the most significant association with SCA severity was in rs1143634, located in the gene IL1B (Bayes Factor > 30). IL1B codes for the cytokine IL-1b, which has a role in TNF-R1 release. Our results suggest that genetic variation in the TNF-α axis may modulate the severity of SCA, and provide biochemical evidence for the validity of the Bayesian-network based severity score as a useful tool for assessing SCA severity.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.