Abstract
BACKGROUND: Transfusion of stored red blood cells (RBCs) induces a variable amount of erythrophagocytosis depending on the extent of RBC damage suffered from the storage lesion. Both murine and canine studies show that transfusion of RBCs after prolonged refrigerated storage leads to inflammation, as evidenced by increased post-transfusion levels of circulating pro-inflammatory cytokines. An analogous cytokine response was not observed in healthy humans volunteers. However, underlying inflammation may synergize with the cytokine response to stored RBCs; thus, results with healthy human volunteers may not reflect those with ill patients. Finally, plasma cytokine profiles may not accurately reflect local tissue responses at sites of erythrophagocytosis, such as the liver and spleen.
AIMS: A mouse model was used to identify splenic transcriptional programs underlying the inflammatory response to stored RBC transfusions. Furthermore, mice were pretreated with a sub-lethal dose of lipopolysaccharide (LPS) to test the effect of underlying inflammation on the response to transfusion.
METHODS: RBCs from 8-12 week old wild-type C57BL/6 male mice donors were collected, leukoreduced, and stored in citrate-phosphate-dextrose-adenine (CPDA)-1. Cohorts of syngeneic mice (n=3 per group) were infused with saline or transfused with one human equivalent unit of fresh RBCs (<24 hours old) or stored RBCs (11 days old). In parallel, cohorts of mice were pretreated with 1 μg of LPS (E. coli 0111:B4) and similarly transfused. Two hours after transfusion, blood and organs were collected and plasma cytokine levels were measured by a multiplex assay. Total RNA from each spleen was processed for hybridization to Affymetrix WT Gene microarrays. Whole genome expression profiles were analyzed using publicly available NIA Array and Gene Set Enrichment Analysis (GSEA) software.
RESULTS: The 2-hour post-transfusion RBC recovery averaged 75% for mice transfused older RBCs. Non-LPS treated mice transfused with stored RBCs showed no statistically significant change in any plasma cytokine post-transfusion, however several cytokines were increased by stored RBC transfusion in LPS pretreated mice. Principle Component Analysis revealed a distinct splenic gene expression program associated with stored RBC transfusion in both LPS treated and non-LPS treated mice. GSEA showed significant enrichment of macrophage activation gene sets in spleens from non-LPS treated mice transfused with stored RBCs (see Figure). LPS pre-treatment increased baseline expression of macrophage activation genes, and these genes were further increased by stored RBC transfusion. Specifically, transfusion of stored RBCs acted in synergy with LPS to induce the cytokine genes (Il6 and Il1a), chemokine genes (Ccl2, Ccl3, Cxcl1, Cxcl2), the immediate early gene Fos, and the gene encoding COX-2 (Ptgs2; see Figure). Finally, a set of genes activated by stored blood transfusion independent of inflammation was identified. These genes included oxidative stress inducible genes (Hmox1 and Osgin) as well as GDF-15, which has previously been associated with states of hemolysis or ineffective erythropoieisis such as sickle cell disease and thalassemia.
CONCLUSIONS: Taken together, these results suggest that although transfusion of one unit of stored RBCs in mice does not lead to a significant difference in circulating cytokine levels at two-hours post-transfusion, it does lead to activation of a distinct inflammatory transcriptional program in spleen. Moreover, sublethal endotoxemia synergizes with stored RBC transfusions to enhance expression of a specific subset of these inflammatory genes. Thus, stored RBC transfusions may be well tolerated by healthy human volunteers, but have a more detrimental impact on hospitalized patients, particularly those with sepsis and low-grade endotoxinemia.
No relevant conflicts of interest to declare.
This feature is available to Subscribers Only
Sign In or Create an Account Close Modal