Unlocking the genetics of sickle cell disease Free

Introduction Sickle cell disease (SCD) is associated with wide variability in clinical disease. Factors like β-globin genotype, co-inherited α-thalassemia, and variants affecting fetal hemoglobin (HbF) levels influence clinical presentation but do not explain all heterogeneity.

Other genetic characteristics may contribute to clinical diversity in SCD. Candidate gene studies are limited to known loci and many traits remain unexplored. Genome-wide association studies (GWAS) offer a broader method to identify novel genetic markers linked to complications. We summarize GWAS findings from a large SCD cohort.

Methods The REDS-III Brazil SCD cohort study enrolled 2,793 participants who in the first phase were followed from 2013-2016. Clinical outcomes were defined according to standardized criteria. Single nucleotide polymorphism (SNP) typing was performed using a custom ~800K SNP transfusion-focused array. The array targeted SNPs enriched for blood- and transfusion-related variants. GWAS were conducted for key SCD complications: priapism, acute chest syndrome (ACS), bacterial infection, avascular necrosis (AVN), ever smoker versus never smoker status, white blood cell (WBC) count, frequent vaso-occlusive (VOC) hospitalizations, cholecystectomy, cholecystitis, cholelithiasis, indirect bilirubin levels and frequent hospitalizations. Additional phenotype data from phase 2 (2021-2024) as part of REDS-IV-P have not yet been included in our analyses.

Initial GWASs used genotype data from the array and Generalized Linear Mixed Model Association Tests (GMMAT). Whole-genome sequencing (WGS) was later performed by the NHLBI Trans-Omics for Precision Medicine program (TOPMed), and Scalable and Accurate Implementation of GEneralized (SAGE) mixed modelling was used. All GWAS used logistic or linear mixed models accounting for kinship and the first 10 principal components, with additional phenotype-specific covariates. Genome-wide significance (GWS) was defined as p≤5×10⁻⁸.

Results For SNP array data, only priapism reached GWS, although other SNPs showed suggestive associations (p≈10⁻⁷). The GWAS for priapism replicated an association with TGFBR3 and identified two novel loci: LINC02537, implicated in prostate cancer, and NAALADL2, previously associated with venous thromboembolism and prostate carcinoma. The ACS analysis, limited to children, suggested an association with ADIPOR2, a gene involved in revascularization following chronic ischemia and in inflammation regulation. Two variants in OPCML, previously linked to venous thromboembolism in African Americans, and MBLAC2, involved in penicillin metabolism and inhibited by sulbactam, were linked to bacterial infections. Smoker status SNPs were associated with AL110114.1, a poorly characterized locus, and NCOR2, previously associated with cocaine dependence. AVN analysis showed associations with two SNPs in DMRT2, a gene implicated in other bone complications.

Additional GWAS following WGS identified associations of ACKR1 (Duffy-null phenotype), TERT (previously linked to bone marrow failure), and FAM3C (implicated in bone mineral density and fractures) with WBC count. Eight SNPs in CTNNA2, a gene involved in cell-cell adhesion and previously associated with multisite chronic pain, and one in METTL4, linked to post-transcriptional regulation under hypoxic conditions, reached GWS for frequent VOC hospitalizations. Analyses of cholecystectomy, cholecystitis, cholelithiasis, and indirect bilirubin replicated known associations, with UGT1A1 variants linked to bilirubin levels and cholecystectomy. Novel loci were found in FER1L6, LRFN5, and SDK2. Frequent hospitalizations (≥3 in 12 months), as a proxy for disease severity, were associated with variants in UHRF2, a potential biomarker for elevated TRV in SCD.

Discussion The REDS-III Brazil SCD cohort collected extensive clinical and laboratory data related to key SCD outcomes. Some well-established genotype-phenotype associations were replicated, and new findings have advanced our understanding of genetic contributors to SCD clinical heterogeneity.

Conclusion The REDS Brazil SCD cohort has significantly advanced the understanding of SCD. Replication of known genetic associations across diverse phenotypes underscores the high quality of the data. Novel findings reveal genes and pathways linked to specific clinical outcomes and shed light on the genetic drivers of disease progression in SCD.