Fetal hemoglobin (HbF) production induced by hydroxyurea is the mainstay of treatment for sickle cell anemia (SCA). Increased HbF production correlates with a higher number of HbF-containing red blood cells (RBCs) called F-cells. Successful treatment with hydroxyurea is associated with an increased number of F-cells, less hemolysis, improvement of anemia, and decreased frequency of vaso-occlusive crises in SCA patients. Comparison of in vitro sickling among blood specimens from sickle mouse models and from patients with different HbF levels has provided compelling evidence that increasing the percentage of circulating F-cells is associated with improvement of hemolytic biomarkers. While it has been demonstrated that higher HbF content prolongs sickle RBC survival, there is only indirect evidence of the response to hypoxia of F-cells compared to non-F-cells. We investigated the influence of HbF content on sickling through our recently developed Sickle Imaging Flow Cytometry Assay (SIFCA). SIFCA allows simultaneous analysis of both expression of intracellular proteins and morphological features of each cell in a robust, reproducible, operator-independent sickling assay. Peripheral venous blood samples were collected upon written consent from adult SCA patients with a wide range of HbF percentages (HbF range 2.0-26.9%) (n=15, nine on hydroxyurea treatment). RBC pellets were used to prepare 1% suspensions that were subjected to deoxygenation for 2 hours at 2% oxygen. RBCs were then labeled for HbF using a standard protocol for F-cell quantitation and a minimum of 20,000 cells were analyzed by imaging flow cytometry (ImageStreamX Mk II, Amnis Corporation), allowing the correlation between shape change and intensity of HbF expression for each RBC. We confirmed previous observations using conventional flow cytometry that F-cell count percentages significantly correlate with mean HbF determined by HPLC (r2P=0.9700, 95% CI 0.9098-0.9902, P<0.0001). F-cell count by SIFCA correlated highly with conventional F-cell flow cytometry by an independent CLIA-certified facility (r2P =0.9976, 95% CI 0.9861-0.9996, P<0.0001). SIFCA morphological analysis showed that the percentage of non-F-cells sickling upon deoxygenation was significantly higher than among F-cells (17.75% [95% CI 12.5-23.00] vs. 12.41% [95% CI 8.67-16.15], P=0.0015), a 1.498-fold difference (95% CI 1.228-1.768). Image analysis also allowed us to identify the presence of F-cells that still sickle despite their high HbF content, as well as non-F-cells that are resistant to sickling (Figure 1). Transmission electron microscopy of F-cells enriched by fluorescence activated cell sorting confirmed that sickled F-cells contained hemoglobin S polymers. In summary, we have documented for the first time at the individual RBC level that human F-cells are less prone to sickle under hypoxia ex vivo than non-F-cells. This study also illustrates the power of imaging flow cytometry to characterize predisposition to sickling in populations of red blood cells from the same patient, and would be suitable for use as a supportive biomarker assay in clinical trials investigating the efficacy of novel HbF inducers and their anti-sickling effect in a single assay. While the finding that F-cells sickle less than non-F-cells is not unexpected, it seems surprising to us that the difference in hypoxia-induced sickling between F-cells and non-F-cells is so small. This finding emphasizes the need to characterize additional RBC features that render individual cells more susceptible or resistant to sickling. Identification of factors besides HbF that modulate sickle hemoglobin polymerization may help design novel therapies for hydroxyurea-resistant SCA patients.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.
© 2013 by The American Society of Hematology