Regulatory Genetic Variation at the S100B Gene Associates with Vaso-Occlusive Manifestations in Sickle Cell Disease

In sickle cell disease (SCD) polymerization of hemoglobin S under deoxygenated conditions causes vaso-occlusion, which can manifest as acute pain crisis and progressive bone/organ damage. Molecular studies have attributed vaso-occlusion to elevated vascular adhesion and inflammatory responses, whereas the genetic regulation has only recently been assessed.

Genomic DNA isolated from peripheral blood mononuclear cells (PBMCs) was hybridized to Illumina Human 610-Quad SNP array for the PUSH and Walk-PHaSST cohorts and to Affrymetrix SNP 6.0 array for the Howard SCD expression cohort. Single nucleotide polymorphisms (SNPs) for 381 PUSH, 525 Walk-PHaSST, and 55 Howard patients were imputed to 1000 genomes project phase 3 data. Messenger RNA from PBMCs was profiled using Affymetrix Human Exon 1.0 ST Array for the Howard expression cohort and Affymetrix Human gene 2.0 ST array for the UIC expression cohort.

Patients within the PUSH and Walk-PHaSST cohorts were classified to four groups according to a cumulative pain score, calculated based on pain frequency and questionnaire description of pain intensity. Pain grouping was examined for correlation with other SCD complications using Cochran Armitage test. History of acute chest syndrome (ACS, PUSH P=3.8×10^-9, Walk-PHaSST P=2.4×10^-5) and avascular necrosis (AVN, PUSH P=4.1×10^-4, Walk-PHaSST P=3.7×10^-5) were the most significant clinical manifestations that consistently associated with pain in the two cohorts.

To investigate the genetic control of vaso-occlusive manifestations with appropriate power, we leveraged genetic association of pain, ACS, and AVN with genetic regulation of disease-specific gene expression. We mapped expression quantitative trait loci (eQTL) in the Howard expression cohort for SNPs<1 Mb away from gene ends per expression trait. At a permutation based false discovery rate of 5%, 1004 independent eQTL (linkage disequilibrium r² ≤0.3 per trait) were identified for 880 genes. Focusing on 129 genes whose expression was altered in PBMCs in sickle cell anemia by at least 1.5-fold [1], we identified six eQTL for five differential genes (up-regulated: OSBP2, SLC14A1, RNF182, CCRL2; down-regulated: S100B).

The six eQTL were assessed for association with pain, ACS, and AVN, using the Walk-PHaSST cohort for discovery and the PUSH cohort for validation. At a significance of Bonferroni corrected P=0.05 (nominal P=0.0083), an eQTL of S100B (rs2154586) significantly associated with AVN in the Walk-PHaSST cohort (OR=1.8, P=0.00061) and the association was replicated in the PUSH cohort (OR=2.7, P=0.0052). The A allele of the eQTL (frequency=0.18) associated with increased risk for AVN and increased expression level of S100B in the Howard expression cohort (β=0.40, P=1.6 ×10^-6). In an additional 64 sickle cell anemia patients without hydroxyurea treatment from the UIC expression cohort, expression levels of S100B were significantly elevated in the individuals with AVN (β=0.28, P=0.029). The 24 SNPs in linkage disequilibrium with the eQTL (r² >0.7) constituted the third most significant peak in a meta-analysis of genome-wide association of AVN in the PUSH and Walk-PHaSST cohorts.

To test the hypothesis that genes involved in vaso-occlusion in SCD may affect thrombotic risk in non SCD individuals, we examined the association of the locus with venous thromboembolism (VTE) in the ARIC, JHS and CHS cohorts from dbGaP. The locus was imputed in African Americans and VTE was defined as being told by a doctor to have a blood clot in the leg or lung as answered in questionnaires during medical exams. The SNPs were associated with VTE using logistic linear regression adjusting for age, gender, enrollment site, and the first 15 principal components per cohort. The risk allele of the leading SNP for AVN consistently associated with increased risk of VTE across the cohorts, with a combined P=0.0041 and OR=1.4.

S100B encodes a calcium sensor that appears to intervene in a variety of biological functions. S100B can mediate the inflammatory effects of damage-associated molecular pattern molecules (DAMPs) produced by erythrocyte hemolysis [2, 3]. Serum concentration of S100B correlates with LDH and with TCD-determined peak velocity of the left middle cerebral artery in thalassemia patients[4]. Polymorphisms of S100B that lead to increased serum levels are associated with increased risk of ischemic stroke in the Chinese population[5].