Integration of Genomic and Genetic Approaches Highlight a Novel Validated Gene Signature for Pulmonary Hypertension Associated with Sickle Cell Disease

Pulmonary hypertension (PH) is a serious complication of sickle cell disease (SCD) associated with increased mortality. Gene expression profiles of peripheral blood mononuclear cells (PBMC) and genetic strategies have been studied in pulmonary arterial hypertension and in SCD. We hypothesized that a PBMC-derived gene signature in SCD patients may be utilized as a PH biomarker which may be further validated using integrated genomic and genetic strategies. Methods & Results: Twenty-seven patients with homozygous SCD underwent transthoracic echocardiography and PBMC isolation for genome—wide expression profiling. An independent SCD cohort (n=132) was genotyped using Affymetrix 6.0 SNP array and also underwent TTE. PH was defined as estimated right ventricular systolic pressure (RVSP)>30 mmHg with a peak tricuspid regurgitation velocity (TRV)>2.5m/s. Genome-wide mRNA and miRNA expression profiles were correlated against PH severity using RVSP (correlation coefficient ρ_R) and TRV (ρ_T) as surrogates. Utilizing a correlation threshold (ρ_R²>0.15, ρ_T²>0.15) for prioritization yielded 631 transcripts and 12 miRNAs. A support vector machine analysis on the transcripts identified a 10 gene signature which discriminated patients with and without echo-defined PH with 100% accuracy. This gene signature was then validated in an independent cohort of SCD patients with PH confirmed by right heart catheterization (n=10) and without PH (n=10) with 90% accuracy. In silico analyses of the top PH-related miRNAs revealed strong binding predictions of miR-301a to polypeptide N-acetylgalactosaminyltransferase 13 (GALNT13), a PH signature gene, which was further validated by microarray data confirming correlation between miR-301a and GALNT13 expression (p=0.024). Genome-wide association study in 132 adult SCD patients comparing echo-defined PH (n=51) versus no PH (n=81) revealed 12 significant SNPs, which were within or upstream to the PH signature genes (Table 1, P<0.01). Seven of the 12 SNPs were associated with GALNT13, further validating this top candidate PH signature gene. Integrating all available genomic and genetic information from a sub-population of 24 patients with SCD revealed significant expression quantitative trait loci (eQTLs) associated with echo-defined PH. Within our PH signature genes, we found four trans-acting eQTLs, based on a FDR<5% with Bonferroni-Holm correction (p=2.1 e-07 for all four) and 1 cis-acting eQTL (p=0.6e-04) upstream of the adenosine A2B receptor gene (ADORA2B) based on a nominal p value 1e–03. Conclusion: These genomic signatures are potential biomarkers to screen at-risk populations in SCD for the presence of PH. Integrative analyses with genomic and genetic analyses highlight ADORA2B and GALNT13, a glycosyltransferase enzyme, as potential candidate genes in SCD-related PH.

No relevant conflicts of interest to declare.