Background: The evolutionary pressure of endemic malaria has produced highly variable genes encoding erythrocyte proteins and enzymes, many of which modulate cellular structure and function. We hypothesized that a high throughput screen of red blood cell susceptibility to osmotic hemolysis, coupled with a GWAS study, would identify rare and common genetic variants associated with cellular structure and function during red blood cell storage and in hemolytic diseases like sickle cell and malaria.

Methods: The REDS III RED Blood Cell Omics (RBC-Omics) study enrolled 13,403 successful blood donors at 4 major hubs. Two labs were used for red blood cell phenotyping of leukoreduced RBC unit-derived samples stored for ~42 days under blood bank conditions. Red blood cell osmotic fragility was assayed in 12,352 donors, including 1,483 African American, 1,477 Asian and 960 Hispanic donors. DNA was isolated and genotyped using a custom Affymetrix Axiom TM-Array, which contained approximately 875,000 SNPs enriched for blood and transfusion related polymorphisms.

Results: Genome wide association analysis in the complete RBC-omics data set revealed 18 genes/loci to be associated with osmotic hemolysis at a genome wide significance threshold of P<5x10-8. The genes included candidates with clear internal validity such as HBB (specifically rs344 the HbS variant) and promotor regions of HBA2 (thalassemic variants), as well as SNPs in logical candidates, such as Ankyrin, Spectrin, and Aquaporin 1, mutations in which are responsible for spherocytosis and red cell membrane water transport defects, among others. To test the potential relevance of these finding to human disease, we evaluated the top 100 regions in the RBC-Omics studies in the 852 pediatric and adult patients with sickle cell disease enrolled in the PUSH and Walk-PHASST screening studies with genome SNP data, and analyzed the association of the SNPs with the intensity of steady state hemolysis, as measured by a principle component analysis of reticulocyte count, bilirubin, lactate dehydrogenase and aspartate amino transferase. After multiple testing correction, numerous genes were associated with the intensity of hemolysis in both studies, including SNPs in red cell metabolism enzymes genes (HK1, TKTL1 MTOR signaling), ion channels (PIEZO1), as well as regulatory gene (ZNF573) and less obvious genes (FLN, SBF2, GAB3, and TMEM116).

Conclusion: An analysis of the genetic variability underlying normal human red cell susceptibility to osmotic stress hemolysis identified new candidate genes that are also associated with the severity of red blood cell hemolysis during the disease stress of human SCD. We hypothesize that many of these candidate SNPs may modulate the severity of red cell hemolysis in other diseases like malaria and potential outcomes after routine red blood cell storage and transfusion. These studies highlight the discovery of a number of new metabolic and ion/water transport variants that are associated with the severity of steady state hemolysis and under disease stress.


Mast: Novo Nordisk: Research Funding. Gordeuk: Emmaus Life Sciences: Consultancy.

Author notes


Asterisk with author names denotes non-ASH members.