Sickle cell disease (SCD) is characterized by chronic inflammation, continuous oxidative stress and severe hemolysis. Activation of endothelial cells, adhesion of sickle erythrocytes and polymorphonuclear neutrophils (PMNs) have been reported to mediate SCD vaso-occlusion. PMNs from SCD patients produce significantly higher basal levels of reactive oxygen species (ROS) and contain less intracellular ROS scavengers than normal individuals. The formation of neutrophil extracellular traps (NETs), which requires the upregulation of intracellular ROS, has been described as a critical innate mechanism to contain infectious insults and to participate in thromboinflammatory injury. In addition, histones and granular proteins released during NET formation can directly damage lung microvascular endothelial cells. Based on these results, we explored the possibility that neutrophils contribute to SCD pathophysiology through NETs. To investigate the presence of NETs in vivo, we used a humanized mouse model of SCD that expresses human sickle hemoglobin exclusively and shares many features of human SCD. Immunofluorescence images of lung samples revealed a significantly higher number of NET DNA fibers in SCD mice (22.6 ± 2.8 NETs/field of view) compared to control mice (5.96 ± 0.86 NETs/field of view, p<0.0001) following proinflammatory cytokine (TNF-α) challenge. We confirmed that these DNA fibers were not resulting from section artifacts or remnants of apoptotic or necrotic cells by co-staining these fibers with antibodies against citrullinated histone H3 and neutrophil elastase (NE), both of which have been identified as markers of NETs. Citrullinated histone H3+ NE+ NETs were identified in the lungs of SCD mice and were found within blood vessels. Concurrently, the soluble NET component plasma DNA was significantly higher in the peripheral blood of SCD mice compared to control mice (∼524 versus ∼48 ng/ml, p<0.001). Further, plasma nucleosomes were detectable only in SCD mice (∼0.11 U/ml). We also found that SCD mice experienced body temperature decline (a reduction of 5.3 ± 1.0°C) following TNF-α challenge whereas the body temperature of control mice was unaffected. A significant positive correlation was found between numbers of NETs and reduction of body temperature in these mice (r=0.72, p<0.0001). DNase I treatment reduced NETs by 38% in SCD mice (12.2 ± 0.9 versus 7.5 ± 0.7 NETs/field of view, for vehicle- and DNase I-treated, respectively, p<0.001), protected these mice from NET-associated hypothermia and significantly prolonged their survival (p<0.05) compared to vehicle-treated SCD mice. As a result of chronic hemolysis, SCD is associated with increased plasma heme. Like SCD patients, SCD mice have significantly higher plasma heme concentration compared to control mice (50.8 ± 5.2 versus 23.8 ± 2.8 μM, p<0.001). We hypothesized that heme is the plasma factor that stimulates PMNs to produce NETs in SCD mice. In vitro study showed that heme activated PMNs to produce citrullinated histone H3+ NE+ NETs in a ROS- and heme-iron dependent manner. Heme injection significantly increased NET formation in control mice (4.5 ± 0.7 versus 9.6 ± 1.7 NETs/field of view for vehicle- and heme-treated, respectively, p<0.01) and accounted for the acute body temperature decline in these mice (3.1 ± 0.8°C, p<0.01). On the other hand, hemopexin administration, which decreased total plasma heme concentration in SCD mice, reduced NET formation by 41% (11.6 ± 0.8 versus 6.9 ± 0.7 NETs/field of view, for vehicle- and hemopexin-treated, respectively, p<0.0001) and protected treated mice from NET-associated hypothermia. Our study thus demonstrates that NETs may contribute to the pathogenesis of SCD and that heme released during hemolysis promotes NET generation in vivo. Targeting heme-mediated NET formation may represent a useful strategy in managing SCD and other pathological states involving severe hemolysis.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.