Stroke is a major cause of morbidity and mortality among children with sickle cell anemia (SCA). Children with SCA at risk for stroke can be identified by transcranial Doppler (TCD) ultrasound screening for abnormally high cerebral artery blood flow velocity and strokes can be prevented by chronic packed red blood cell (RBC) transfusion. However, the mechanisms that lead to cerebral vasculopathy and stroke in SCA and that explain the beneficial effects of chronic RBC transfusions in stroke prevention are poorly understood. We have previously shown that pre-treatment serum levels of brain derived neurotropic factor (BDNF) and platelet derived growth factor (PDGF) subtypes, biomarkers of cerebral ischemia and arterial remodeling, were associated with both high TCD velocity and development of stroke. We hypothesized that frequency of RBC transfusion would be associated with altered serum levels of neurodegenerative, inflammatory and angiogenic markers in SCA children with high TCD velocity and tested this hypothesis by assaying levels of these markers in post-STOP serum samples.
Frozen serum samples drawn one year after subject's exit from the STOP clinical trial phase were utilized. Given the positive trial results, all but 9 subjects had been on chronic transfusion regimen for at least one year at the time of sample collection. Eighty samples were assayed using multiplex antibody immobilized beads (Millipore Corp, Billerica, MA). The mean fluorescent intensity was measured using the Milliplex xMAP system powered by Luminex (Bio-Rad, Hercules, CA). Ten biologically related neurodegenerative, inflammatory and angiogenic biomarkers were tested. The total number (frequency) of RBC transfusions recorded over the study period (4 years) for each participant was categorized into High (≥ 40), Moderate (20 – 39) or Low (< 20) frequency of transfusion. Median distribution with 10 – 90th percentile of the levels of biomarkers and TCD velocity were expressed using box-plots and the differences in median distribution between groups based on frequency of transfusion was estimated using Kruskal-Wallis test. A principal component analysis (PCA) loading plot was used to demonstrate the biological relationships between the biomarkers, taking into consideration linear correlations and spatial relationships between them.
There were no significant differences in the hematological and anthropometric measures between groups. Overall, our result showed that low transfusion frequency was associated with high serum levels of biomarkers and vice versa, despite no significant difference in hemoglobin level between groups. The high frequency transfusion group had lower serum levels of BDNF (p = 0.02), sVCAM-1 (p < 0.001), PDGF-AA (p < 0.001), CCL5 (p < 0.01), tPAI-1 (p < 0.01) and NCAM (p < 0.01) levels compared with the low frequency transfusion group (figure 1 a – e). Although not shown in the figures, the same pattern was observed with TCD velocity which was lower (160, 115.7 – 204.9 cm/s) in the low compared with the high (195, 154 – 272 cm/s) frequency transfusion group. In addition, the medium frequency transfusion group had significantly lower serum sVCAM-1 (p < 0.01) compared with the low frequency transfusion group and higher PDGF-AA (p < 0.01) compared with the high frequency transfusion group. A PCA loading plot (figure 2) shows clustering of the biomarkers that are most closely biologically related, these are also the biomarkers that were significantly affected by the frequency of transfusion.
Red blood cell transfusions in the STOP study were associated with reduced serum levels of biomarkers of angiogenesis (PDGF-AA and sVCAM-1), cerebral ischemia/neuronal survival adaptation (BDNF and NCAM) and inflammation (RANTES/CCL5), and this effect was most pronounced in the group with the highest frequency of transfusions (equivalent to most chronic transfusion regimen). This suggests that the protective effects of chronic RBC transfusions on stroke development in children with SCA may be attributable to improved cerebral perfusion, reduced inflammation and down-regulation of hypoxia-induced angiogenic responses that promote arterial remodeling. One or more in this group of biologically-related and relevant markers may be useful for monitoring children with SCA receiving stroke prevention therapies and for designing treatment targets.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.