2005 FASEB

The FASEB Journal express article 10.1096/fj.04-3218fje. Published online March 22, 2005.

Endothelial cell NADPH oxidase mediates the cerebral microvascular dysfunction in sickle cell transgenic mice
Katherine C. Wood,* Robert P. Hebbel, and D. Neil Granger* *Department of Molecular and Cellular Physiology, Louisiana State University Health Sciences Center, Shreveport, LA 71130; and Vascular Biology Center and Division of HematologyOncology-Transplantation, University of Minnesota Medical School, Minneapolis, MN 55455 Corresponding author: D. Neil Granger, Department of Molecular and Cellular Physiology, LSU Health Sciences Center, 1501 Kings Highway Shreveport, LA 71130-3932. E-mail: dgrang@lsuhsc.edu ABSTRACT Although blood cell-endothelial cell adhesion and oxidative stress have been implicated in the pathogenesis of sickle cell disease (SCD), the nature of the linkage between these vascular responses in SCD remains unclear. The objective of this study was to determine whether superoxide derived from endothelial cell-associated NADPH oxidase mediates the leukocyteendothelial (L/E) and platelet-endothelial cell (P/E) adhesion that is observed in the cerebral microvasculature of sickle cell transgenic (S) mice. Intravital fluorescence microscopy was used to monitor L/E and P/E adhesion in brain postcapillary venules of wild-type (WT), SOD1 transgenic (SOD1-TgN), and gp91phox (NADPH oxidase)-deficient mice that were transplanted with bone marrow from S mice. Hypoxia/reoxygenation (H/R) yielded intense P/E and L/E adhesion responses in cerebral venules of S/WT chimeras that were significantly attenuated in both S/SOD1-TgN, and S/gp91phox/ chimeras. Pretreatment of S/WT chimeras with the iron-chelator desferroxamine blunted the blood cell-endothelial cell adhesion responses to H/R, whereas pretreatment with the xanthine oxidase inhibitor allopurinol had no effect. These findings suggest that superoxide derived from endothelial cell NADPH-oxidase and catalytically active iron contribute to the proinflammatory and prothrombogenic responses associated with sickle cell disease. Key words: inflammation thrombogenesis hypoxia/reoxygenation leukocyte-endothelial cell adhesion ickle cell disease (SCD) is an inherited disorder that affects the vasculature of a variety of organ systems (1). In SCD patients, hypoxic insults elicit the release of cytokines that activate both blood and vascular cells to engage in cell-cell adhesion (24). Similarly, in transgenic mouse models of SCD the postcapillary venules assume a proinflammatory and prothrombogenic phenotype, wherein there is pronounced adhesion of both leukocytes and platelets to venular endothelium (57). The few mechanistic studies that have addressed the blood cell-endothelial cell interactions associated with SCD suggest that endothelial cell activation is a critical component of the microvascular responses that accompany this disease.

Page 1 of 19 (page number not for citation purposes)

The endothelial cell activation is manifested as an increased expression of adhesion molecules and an enhanced production of reactive oxygen species (oxidative stress) (5). P-selectin, an adhesion molecule expressed on the surface of activated endothelial cells and platelets, appears to play a major role in the recruitment of adherent leukocytes and platelets in the microvasculature of sickle cell transgenic mice. Data derived from bone marrow chimera mice produced by transplanting bone marrow from sickle cell transgenic mice into mice that are genetically deficient in either P-selectin (7) or both P- and E-selectin (6) indicate that endothelial cell-associated-P-selectin plays a more important role than platelet P-selectin in mediating the blood cell-endothelial cell interactions in SCD. Although much progress has been made toward defining the contributions of specific endothelial cell adhesion molecules to the recruitment of leukocytes and platelets in microvessels of sickle cell transgenic mice, relatively little is known about the enzymatic and cellular sources of the reactive oxygen species (ROS) that are produced by endothelial cells in SCD. Furthermore, the dependence of SCD-induced blood cell-endothelial cell interactions on the accelerated ROS production by microvascular endothelial cells remains unclear. However, there is evidence that supports a potential role for ROS in experimental SCD, including reports describing an enhanced oxidative stress detected by oxidant-sensitive fluorochromes (5, 8) and the ability of exogenous superoxide dismutase (SOD) to attenuate the leukocyte/endothelial cell adhesion observed in cremasteric vessels of sickle cell transgenic mice (9). Although xanthine oxidase has been implicated in SCD, the possible involvement of other sources of ROS has not been previously considered. The overall objective of this study was to further explore the potential contributions of superoxide and ROS to the proinflammatory and prothrombogenic phenotype that is assumed by postcapillary venules of sickle cell transgenic mice. One objective was to determine whether the genetic overexpression of cytosolic (CuZn) SOD or deficiency of NADPH oxidase in vascular endothelial cells alters the leukocyte and platelet responses to SCD. The contribution of xanthine oxidase was assessed using allopurinol treatment. Since previous studies have implicated a role for catalytically active iron in the pathology of SCD (10, 11), we also evaluated the effects of the iron chelator desferroxamine (DFO) on the blood cell adhesion responses in SCD. Our results provide novel insights into the contribution and cellular localization of ROS production in the microvascular responses to SCD. MATERIALS AND METHODS Animals Wild-type (WT) C57BL/6 mice, CD45 congenic B6.SJL-PTPRCPEP/BOY mice (which express CD45.1), and B6.129S6-CYBBTML DIN (gp91phox/; backcrossed on a C57Bl/6J background for 12 generations) mice on a C57BL/6 background were obtained from Jackson Laboratories (Bar Harbor, ME). Homozygous transgenic (TgN) mice overexpressing human CuZn-SOD (TgN[SOD1]3Cje) at approximately 5 times the WT level (12) were also used. Breeder stocks for homozygous C57BL6-TgN(SOD1)3Cje from a backcross generation >9 were obtained from Jackson Laboratories. The SODTgN mice were identified by qualitative demonstration of CuZnSOD activity using nondenaturing gel electrophoresis followed by nitroblue tetrazolium staining. Sickle cell (S) transgenic mice and their negative controls (C57Bl/6) were generated and

Page 2 of 19 (page number not for citation purposes)

characterized in the colony at the University of Minnesota. The S mice used in these studies are the progeny of homozygous sickle (SS) males (hS; m/m/) bred with hemizygous females (hS; m/m+/) and have one normal mouse -globin gene and one human HbS transgene (m/hS). They are developed on a mixed genetic background (FVB/N, 129, DBA/2, C57BL/6, Black Swiss) (13), are homozygous for knockout of murine -globin and heterozygous for knockout of murine -globin, and have one copy of the linked transgenes for human - and human S-globins. The animals were 68 months old and housed under specific pathogen-free conditions. The mice were maintained on standard laboratory chow and fed ad libitum until immediately before the experiments. Chimeras Four combinations of bone marrow chimeras were produced. WT/WT chimeras were WT animals (CD45.2 positive leukocytes) that received bone marrow cells from CD45 congenic mice (CD45.1 positive leukocytes). This resulted in a significant increase of leukocytes expressing CD45.1 (of donor origin), from <5% in WT to >90% in the WT/WT chimeras, allowing verification of proper chimera reconstitution, as described previously (14). S/WT chimeras were produced by transplanting bone marrow from S mice (CD45.1 positive leukocytes) into WT mice (CD45.2 positive leukocytes). Previous studies have shown that S/WT chimeras adopt the donor phenotype a few weeks after transplantation of the S bone marrow (6, 7). Likewise, WT/S chimeras assume a normal (WT) phenotype a few weeks after transplantation of WT marrow into S mice (7). In some experiments, S bone marrow was transferred into gp91phox/ mice (CD45.2 positive leukocytes), yielding S/gp91phox/ chimeras with intact gp91phox on S blood cells and a gp91phox-deficient vessel wall. In other experiments, S bone marrow was transferred into SOD1-TgN mice (CD45.2 positive leukocytes), yielding S/SOD1-TgN chimeras with normal SOD1 expression by S blood cells and an overexpression of SOD1 by the vessel wall. Bone marrow cells were isolated from the femurs and tibias of donor mice and resuspended at 4 107 cells/ml PBS. Recipient mice were irradiated with two doses of 500525 rad, 3 h apart, after which 8 106 donor bone marrow cells in 200 L of PBS were administered via the femoral vein. The chimeras were kept in autoclaved cages, with 0.2% neomycin sulfate (Sigma) in the drinking water for 2 weeks, after which normal drinking water was used. Seven weeks after reconstitution, only chimeric mice demonstrating 90% or better conversion of leukocyte antigen to donor phenotype were subjected to 2 h hypoxia, followed by 4 h reoxygenation, and prepared for intravital microscopy using the closed cranial window preparation. Flow cytometry was used to verify chimera reconstitution in all animals by staining for CD45.1 expression on leukocytes with a FITC-labeled anti-CD45.1 antibody (PharMingen Inc.). Experimental protocols All experimental protocols were applied to WT (C57BL/6) or S chimeras developed from WT or mutant mice with a C57BL/6 background, 68 months of age. Five or more mice were included in each group. Two groups of S/WT chimeric mice were treated with 50 mg/kg body weight allopurinol (Sigma Chemical) using different routes of administration: one group was treated orally (15) for 2 days before surgery and another group was treated intraperitoneally (9,

Page 3 of 19 (page number not for citation purposes)

16) one day before surgery. Both groups received an additional dose administered before hypoxia. Another group of S/WT chimeric mice was treated with intraperitoneal injection of 300 mg/kg body weight DFO (Sigma Chemical) (17) for 3 days before intravital microscopy. Surgical protocol All animals were anesthetized intraperitoneally with ketamine hydrochloride (50 mg/kg) and pentobarbitol (50 mg/kg). The left femoral artery was cannulated for monitoring mean arterial blood pressure (BP-1, World Precision Instruments) and sampling arterial blood for blood gas analysis (Omni-Modular System), while the femoral vein was cannulated for intravenous administration of carboxyfluorescein succinimidyl ester (CFSE, Molecular Probes; Eugene, OR) labeled platelets and/or acridine orange (Sigma; St. Louis MO) for in vivo labeling of leukocytes. Arterial blood samples allowed for hemoglobin (g/dl), hematocrit (%), and methemoglobin (%) measurements. Core body temperature was maintained at 35 0.5C. The experimental procedures described above were reviewed and approved by the Louisiana State University Health Sciences Center-Shreveport Institutional Animal Care and Use Committee and performed according to the criteria outlined in the National Institutes of Health guidelines. Experimental preparation Mice were maintained on normal mouse chow until 6 h before the experiment. Mice subjected to hypoxia/reoxygenation, to simulate a hypoxic crisis that promotes red blood cell sickling, were placed individually in a hypoxia chamber (10% O2, 0.05% CO2, balance nitrogen) for 2 h, and then returned to normoxic conditions for the remainder of the experiment. Lidocaine (1%) was used for local anesthesia. All mice were tracheostomized and mechanically ventilated (Harvard Rodent Ventilator Model 683; South Natick MA) with room air. The head of each mouse was fixed in a stereotaxic frame, and the left parietal bone was exposed by a longitudinal midline skin incision. Two PE-50 polyethylene tubes (Clay Adams; Parsippany, NJ) were fixed on the skull with ethyl cyanoacrylate (Elmers Products, Columbus, OH) and a closed cranial window was constructed. A craniectomy was made using a high-speed microdrill (Fine Science Tools; Foster City, CA), the dura was carefully reflected, and a cover glass affixed with a quick self-curing acrylic resin (Biobond, Japan) used to render the window air-tight. Artificial cerebrospinal fluid (in mEq/L): 147.8 Na+, 3.0 K+, 2.3 Mg2+, 2.3 Ca2+, 135.2 Cl, 19.61 HCO3, 1.67 lactate, and (in mmol/L) 1.1 phosphate, and 3.9 glucose filled the space beneath the cranial window. The animals were allowed to stabilize for 30 min before observation of postcapillary venules under an upright Nikon Xanophot microscope (HLX64610). Branches of postcapillary venules that drain blood from the anterior, superior, and middle cerebral arteries were scanned and images of 1-min duration were videotaped using a videocassette recorder (BR-S601MU, JVC; Japan) for each of 5 nonoverlapping 300-m venular segments (2070 m diameter). A time-date recorder (WJ810, Panasonic) displayed a stopwatch function onto the video screen. Circulating cell-endothelial cell interactions Monitoring the interactions of circulating cells with each other and with endothelium in the cerebral microcirculation required fluorescent labeling of cells; leukocytes were labeled in vivo with acridine orange (0.05% in 100 l) and platelets were labeled ex vivo with CFSE (90 M).

Page 4 of 19 (page number not for citation purposes)

The following manipulations, which have been shown to have no significant effect on the activation state or viability of isolated platelets (as assessed by flow cytometry) (18) were used to isolate and label donor platelets from WT mice. Platelets were isolated from whole blood obtained via the carotid artery of donor mice. The blood was collected in a polypropylene tube containing 0.1 ml acid citrate dextrose buffer (Sigma; St. Louis, MO) and then centrifuged twice at 120 g for 8 and 3 min. Between centrifugations, platelet-rich plasma (PRP) was removed by pipette and placed in a clean polypropylene tube. The PRP was centrifuged at 550 g for 10 min to pellet the platelets. The platelet pellet was gently resuspended in 1500 l PBS (pH 7.4) and incubated with the fluorescent dye CFSE for 10 min at room temperature. The platelets were then centrifuged for 10 min at 550 g, resuspended in 500-l PBS, and protected from light until use. Contamination of the platelet suspension by leukocytes was evaluated from a 25-l sample and did not exceed 0.01%. Leukocytes were stained by addition of 465 L 3% citric acid and 10 L 1% crystal violet (Sigma) and quantified with the aid of a hemocytometer and light microscope. Platelets (100106) were infused over 5 min, yielding ~5% of the total platelet count. Infused platelets were allowed to circulate for a period of 5 min before videotaping. Estimates of pseudoshear rate in venules were obtained using measurements of venular diameter (Dv) and the maximal velocity of flowing platelets (Vplt) in venular segments videotaped for offline quantitative analysis of cell adhesion measurements. Pseudoshear rate was determined for each animal using the following formulation: pseudoshear rate = (Vplt/1.6)/Dv 8. (19) Intravital fluorescence microscopy Platelets were visualized, as described previously, (20) with a Nikon Xanophot upright microscope assisted by a silicon-intensified target camera (C-2400, Hamamatsu Photonics). Cerebral venules were epi-illuminated, observed through a 20 long distance objective lens (Leitz Weitzler, Germany), and recorded on videotape using a videocassette recorder (BRS601MU, JVC; Japan). A video time/date generator projected the time, date, and stopwatch function on the monitor (diagonal 50.6 cm, PVM-2030, Sony Trinitron; Japan). CFSE (excitation: 490 nm, emission: 518 nm) and acridine orange (excitation: 500 nm, emission: 526 nm) visualization required a filter block with an excitation filter (B-2A, Nikon; Japan) for 450 490 nm, a dichroic mirror for 510 nm, and a barrier filter of 520 nm. The cerebral surface was scanned for five nonoverlapping nonmeningeal venular segments per animal, and each was recorded for 1 min. Venules with a diameter of 2070 m and identifiable branch extensions to subsurface brain tissue were considered nonmeningeal. Platelet and leukocyte adhesion were quantified separately within cerebral venules. Video analysis Leukocyte and platelet interactions with the endothelium were determined by off-line analysis of videotaped images, as described previously (7). Nonoverlapping venular segments of 300-m length with diameters ranging between 20 and 70 m (39.9 m mean diameter) were chosen for quantification of adherent cells. Platelets and leukocytes were considered adherent if they arrested on the vascular endothelium 2 and 30 s, respectively. Platelet and leukocyte adherence were expressed as the number of cells per square millimeter of venular surface, calculated from diameter and length, assuming cylindrical vessel shape (21).

Page 5 of 19 (page number not for citation purposes)

Statistical analysis All values are reported as means SEM. ANOVA with the Scheffes post hoc test was used to compare groups, with statistical significance set at P < 0.05. RESULTS Comparisons of venular pseudoshear rate (55621 vs. 46239 s1, P=0.11), venular diameter (38.41.2 vs. 38.90.7 m), and arterial blood pressure (Table 1) between WT/WT and S/WT chimeric groups revealed no statistically significant differences. Both S mice and S/WT chimeras exhibited hemoglobin concentration (Hb), as well as hematocrit (Hct) and methemoglobin (metHb) values (Table 2) that were different from their WT counterparts. One objective of this study was to determine whether genetic overexpression of cytosolic (CuZn) superoxide dismutase (SOD) in vascular endothelial cells alters the leukocyte and platelet adhesion responses to SCD. The role of endothelial cell-associated superoxide in mediating the leukocyte and platelet adhesion responses to 2 h of whole body hypoxia (10% O2, 0.05% CO2, balance nitrogen) followed by 4 h reoxygenation (room air) in S mice was addressed in bone marrow chimeric mice (Fig. 1). Transfer of marrow from S mice into WT mice (S/WT, n=9) yielded significantly higher H/R-induced leukocyte (10118 vs. 4410 per mm2, P=0.01) and platelet (1104203 vs. 411163 per mm2, P=0.029) adhesion responses, compared with WT mice receiving bone marrow from WT mice (WT/WT, n=7). However, the transfer of bone marrow from S mice into SOD-1 TgN mice (S/SOD-1TgN; n=5) yielded leukocyte (226 vs. 4410 per mm2, P=0.47) and platelet (21863 vs. 411163 per mm2, P=0.77) adhesion responses to H/R that were no different from those noted in WT/WT chimeras and significantly reduced compared with leukocyte (226 per mm2, P=0.0009) and platelet (21863 per mm2, P=0.008) adhesion responses observed in S/WT chimeras. The blunted adhesion responses of platelets and leukocytes in S/SOD-1TgN chimeric mice, with normal SOD expression by circulating S blood cells, but overexpression of SOD by the vessel wall, strongly implicates endothelial cell-associated superoxide in these adhesion responses. A second objective of this study was to determine the relative contributions of endothelial cell NADPH oxidase and xanthine oxidase to the superoxide-dependent recruitment of platelets and leukocytes in venules of sickle cell transgenic mice. The involvement of endothelial cell NADPH oxidase (Fig. 2) was addressed in gp91phox/ mice that were transplanted with bone marrow derived from S mice (S/gp91phox/, n=5). These chimeras exhibited platelet (28345 per mm2, P=0.90) and leukocyte (223 per mm2, P=0.52) adhesion responses to H/R that were no different from those noted in WT/WT chimeras (n=7) and significantly reduced compared with platelet (P=0.001) and leukocyte (P=0.02) adhesion responses observed in S/WT chimeras (n=9). The attenuated adhesion responses in S/gp91phox/ chimeric mice, with functional NADPH oxidase expression by circulating S blood cells and inactive NADPH oxidase in the vessel wall, strongly implicates endothelial cell-associated NADPH oxidase in these adhesion responses. The contribution of xanthine oxidase to H/R-stimulated leukocyte and platelet adhesion responses in S mice was addressed in S/WT mice treated with the xanthine oxidase inhibitor,

Page 6 of 19 (page number not for citation purposes)

allopurinol. Intraperitoneal (n=5) and oral allopurinol (n=6) pretreated S/WT chimeras did not significantly differ from one another in terms of platelet (1036368 vs. 1066324, P=0.95) and leukocyte (8310 vs. 5611, P=0.10) adhesion responses. Intraperitoneal allopurinol pretreatment (1036368 per mm2, P=0.86) (Fig. 3) and oral allopurinol pretreatment (1066324 per mm2, P=0.92) yielded platelet adhesion responses to H/R that did not differ from those noted in S/WT chimeras (n=9). Although intraperitoneal allopurinol pretreatment had no effect on leukocyte adhesion responses in S/WT chimeras (8310 per mm2, P=0.48), oral allopurinol pretreatment did have an attenuating effect, although the responses were not statistically significant (5611 per mm2, P=0.08). The role of catalytically active iron in mediating the leukocyte and platelet adhesion responses to H/R in S mice was addressed in S/WT mice treated with the iron chelator desferroxamine (DFO). Intraperitoneal DFO pretreatment (Fig. 4) of S/WT mice (n = 7) yielded platelet (520136 per mm2, P=0.62) and leukocyte (343 per mm2, P=0.27) adhesion responses to H/R that were no different from those noted in WT/WT chimeras (n=7) and significantly reduced compared with platelet (1104203 per mm2) and leukocyte (10118 per mm2) adhesion responses observed in S/WT chimeras (n=9). DISCUSSION Oxidative stress has been proposed to be a major factor in the pathogenesis of SCD. This contention is supported by several lines of evidence, including the demonstration of accelerated ROS generation in postcapillary venules (5), an attenuating effect of exogenous SOD on some of the microvascular alterations induced by SCD (9), and the ability of xanthine oxidase inhibitors to attenuate the SCD-associated impairment of endothelial cell-dependent arteriolar dilation (8). Previous reports from our laboratory and by others indicate that SCD also induces a proinflammatory (57) and prothrombogenic (7) phenotype in postcapillary venules in different tissues, including skeletal muscle and brain. The objectives of this study were to determine whether ROS contribute to these phenotypic changes in the microvasculature during SCD and to identify the potential enzymatic and cellular sources of these ROS. Because superoxide has been implicated in the defective arteriolar and venular responses in mutant mice with SCD, we studied the potential involvement of this reactive oxygen species in our model using SOD1-TgN mice. SOD1-TgN mice typically overexpress the cytosolic form (CuZn) of SOD (up to 600% in the brain), depending on the organ (12). We focused on cytosolic SOD because this form of the enzyme accounts for 5080% of total vascular SOD activity (22). Our findings indicate that overexpression of this cytosolic antioxidant largely abolishes both the proinflammatory and prothrombogenic phenotype that is normally assumed by SCD mice. Our results on leukocyte adhesion are consistent with a recent report by Kaul et al. that demonstrates complete inhibition of hypoxia/reoxygenation-induced leukocyte adhesion in postcapillary venules of the cremaster of a transgenic model of SCD treated with exogenous SOD (9). The fact that an increased extracellular concentration of SOD is as effective as overexpression of intracellular vascular SOD in preventing SOD-induced leukocyte adhesion suggests that the source of superoxide lies at or near the endothelial cell surface. Irrespective of the location of this enzyme, all of the available evidence strongly implicates superoxide in the inflammatory process that is associated with SCD.

Page 7 of 19 (page number not for citation purposes)

There are several potential sources of the superoxide that is generated in postcapillary venules of S mice. These include xanthine oxidase, auto-oxidation of hemoglobin (Hb), and NADPH oxidase. Two previous reports have implicated xanthine oxidase as a major source of ROS produced in arterioles (8) and venules (9) of SCD mice. Both of these studies used the xanthine oxidase inhibitor allopurinol. In the present study, allopurinol administration at inhibitory doses either orally or intraperitoneally had no effect on the proinflammatory and prothrombogenic responses seen in cerebral venules of S mice. These findings suggest that xanthine oxidase is not an important source of the superoxide generated in the cerebral microvessels of SCD mice. This apparent discrepancy between our findings in the brain and previous reports on cremaster microcirculation (9) likely reflects differences in xanthine oxidase activity between skeletal muscle and brain microcirculation. Indeed, it has been shown that xanthine oxidase is absent from bovine brain capillary endothelium but present in skeletal muscle endothelium. (23) Another enzymatic source of superoxide that has been implicated in several disease processes, including hypertension (2426), atherosclerosis (27), and stroke (20, 28) is NADPH oxidase. This enzyme has been localized to vascular endothelial cells (29) and comprises different protein subunits, including gp91phox. The enzyme is also found in other cell types, including leukocytes and platelets. The results of the present study strongly support the involvement of NADPH oxidase in SCD-induced blood cell-vessel wall interactions. Our findings, which are based on mice that are incapable of producing superoxide through NADPH oxidase due to a genetic deficiency in the gp91phox subunit of this enzyme, indicate that NADPH oxidase inactivation is as effective as SOD overexpression in abolishing both the leukocyte and platelet adhesion responses that are normally observed in SCD mice. This study therefore represents the first demonstration of NADPH oxidase involvement in the blood cell-vessel wall interactions elicited by SCD and suggests that this enzyme likely represents the major source of superoxide in this disease model. Sickle hemoglobin (HbS) auto-oxidation represents a potentially important nonenzymatic source of ROS in SCD (30). Iron decompartmentalization within sickle erythrocytes, possibly reflective of HbS instability, contributes to membrane-associated free iron that is redox active (11, 31). This redox active iron is capable of catalyzing intraerythrocytic ROS formation (32), as noted by a near-twofold increase in spontaneous superoxide production by sickle erythrocytes (32). The elevated baseline and post-H/R methemoglobin (metHb) levels (a measure of autooxidation) detected in our S mice compare quantitatively with the accelerated rates of metHb formation observed in hemoglobin (Hb) isolates from erythrocytes of sickle patients (30). Enhanced intraerythrocytic ROS formation, in the context of decreased catalase and glutathione peroxidase stores within sickle erythrocytes (33, 34), may contribute to the multiple membrane defects and enhanced susceptibility to hemolysis previously reported for sickle cells. Chelation of catalytically active iron has been shown to reduce erythrocyte membrane peroxidation and the concomitant hemoglobin denaturation (35). To address the contribution of plasma catalytically active iron in promoting the proinflammatory and prothrombogenic phenotype observed in SCD, we treated S mice with the iron chelator desferroxamine (DFO). The results of experiments reveal that DFO pretreatment significantly attenuates the adhesion of both leukocytes and platelets in cerebral venules of S mice. Because our DFO treatment regimen has been reported to produce a preconditioned tolerance effect in a mouse model of focal cerebral ischemia (17), it is conceivable that some of the protective effects that we noted are also related to an increased

Page 8 of 19 (page number not for citation purposes)

HIF-1 and enhanced production of the potent neuroprotectant erythropoietin. However, the protection that we observed with DFO, when considered in the context of our findings with SOD1-TgN overexpression and gp91phox (NADPH oxidase) deficiency, is consistent with a mechanism wherein NADPH oxidase-derived superoxide leads to the formation of irondependent radicals, which ultimately mediate the microvascular responses observed in S mice. Therefore, chelation of the catalytically active iron that is free in hemolyzed blood blunts the microvascular responses observed in our model. Leukocytes, platelets, and endothelial cells are all capable of producing superoxide via an NADPH oxidase-dependent mechanism (27). This raises the issue about the likely cellular source of NADPH oxidase-derived superoxide in our experiments. Because we employed bone marrow chimeras, wherein S marrow was transferred to either gp91phox/ or SOD1-TgN recipients, the mutants created possessed typical S levels of gp91phox and CuZnSOD in all circulating blood cells, while the vessel wall (endothelial cells) was either deficient in gp91phox or overexpressed CuZnSOD. Therefore, our findings in these mutants suggest that the vessel wall (endothelial cells) is the major source of NADPH oxidase-derived superoxide rather than leukocytes or platelets. While the allopurinol and DFO treatment groups do not provide the cell specificity that is afforded by the bone marrow chimeras, the results obtained with these pharmacological agents are nonetheless consistent with the possibility that the vessel wall is the dominant source of ROS in the brain of S mice. Our findings invoke a novel role for endothelial cell NADPH oxidase in the pathophysiology of SCD. The experimental strategies used in this study reveal a strong dependence of the blood cell interactions in SCD on NADPH oxidase-derived superoxide and catalytically active iron (Fig. 5). Furthermore, our findings point to endothelial cells as a potentially important source of the ROS that contribute to the microvascular dysfunction elicited by SCD. The relative unimportance of xanthine oxidase in mediating the SCD-induced microvascular alterations in the brain, compared with the previously reported involvement of this enzyme in skeletal muscle venules (9), emphasizes the need for caution when extrapolating the responses of one regional vascular bed to another, and it underscores the need to consider multiple therapeutic targets for treatment of the microvascular complications of SCD. ACKNOWLEDGMENT This study was supported by grants from the National Institutes of Health (P01-DK43785) and (P01-HL55552). REFERENCES 1. 2. Ballas, S. K. (1998) Sickle cell disease: clinical management. Baillieres Clin. Haematol. 11, 185214 Gurkan, E., Tanriverdi, K., and Baslamisli, F. (2005) Clinical relevance of vascular endothelial growth factor levels in sickle cell disease. Ann Hematol. 84, 7175

Page 9 of 19 (page number not for citation purposes)

3.

Wun, T., Cordoba, M., Rangaswami, A., Cheung, A. W., and Paglieroni, T. (2002) Activated monocytes and platelet-monocyte aggregates in patients with sickle cell disease. Clin. Lab. Haematol. 24, 8188 Inwald, D. P., Kirkham, F. J., Peters, M. J., Lane, R., Wade, A., Evans, J. P., and Klein, N. J. (2000) Platelet and leucocyte activation in childhood sickle cell disease: association with nocturnal hypoxaemia. Br. J. Haematol. 111, 474481 Kaul, D. K., and Hebbel, R. P. (2000) Hypoxia/reoxygenation causes inflammatory response in transgenic sickle mice but not in normal mice. J. Clin. Invest. 106, 411420 Turhan, A., Weiss, L. A., Mohandas, N., Coller, B. S., and Frenette, P. S. (2002) Primary role for adherent leukocytes in sickle cell vascular occlusion: a new paradigm. Proc. Natl. Acad. Sci. USA 99, 30473051 Wood, K. C., Hebbel, R. P., and Granger, D. N. (2004) Endothelial cell P-selectin mediates a proinflammatory and prothrombogenic phenotype in cerebral venules of sickle cell transgenic mice. Am. J. Physiol. Heart Circ. Physiol. 286, H1608H1614 Aslan, M., Ryan, T. M., Adler, B., Townes, T. M., Parks, D. A., Thompson, J. A., Tousson, A., Gladwin, M. T., Patel, R. P., Tarpey, M. M., et al. (2001) Oxygen radical inhibition of nitric oxide-dependent vascular function in sickle cell disease. Proc. Natl. Acad. Sci. USA 98, 15,21515,220 Kaul, D. K., Liu, X., Choong, S., Belcher, J. D., Vercelotti, G. M., and Hebbel, R. P. (2004) Anti-inflamamtory therapy ameliorates leukocyte adhesion and microvascular flow abnormalities in transgenic sickle mice. Am. J. Physiol. Heart Circ. Physiol. 287, H293 H301

4.

5. 6.

7.

8.

9.

10. Repka, T., and Hebbel, R. P. (1991) Hydroxyl radical formation by sickle erythrocyte membranes: role of pathologic iron deposits and cytoplasmic reducing agents. Blood 78, 27532758 11. Browne, P., Shalev, O., and Hebbel, R. P. (1998) The molecular pathobiology of cell membrane iron: the sickle red cell as a model. Free Radic. Biol. Med. 24, 10401048 12. Epstein, C. J., Avraham, K. B., Lovett, M., Smith, S., Elroy-Stein, O., Rotman, G., Bry, C., and Groner, Y. (1987) Transgenic mice with increased Cu/Zn-superoxide dismutase activity: animal model of dosage effects in Down syndrome. Proc. Natl. Acad. Sci. USA 84, 8044 8048 13. Noguchi, C. T., Gladwin, M., Diwan, B., Merciris, P., Smith, R., Yu, X., Buzard, G., Fitzhugh, A., Keefer, L. K., Schechter, A. N., et al. (2001) Pathophysiology of a sickle cell trait mouse model: human alpha(beta)(S) transgenes with one mouse beta-globin allele. Blood Cells Mol. Dis. 27, 971977

Page 10 of 19 (page number not for citation purposes)

14. Tomita, Y., Sachs, D. H., and Sykes, M. (1994) Myelosuppressive conditioning is required to achieve engraftment of pluripotent stem cells contained in moderate doses of syngeneic bone marrow. Blood 83, 939948 15. Akgur, F. M., Brown, M. F., Zibari, G. B., McDonald, J. C., Epstein, C. J., Ross, C. R., and Granger, D. N. (2000) Role of superoxide in hemorrhagic shock-induced P-selectin expression. Am. J. Physiol. Heart Circ. Physiol. 279, H791H797 16. Jeon, B. R., Yeom, D. H., and Lee, S. M. (2001) Protective effect of allopurinol on hepatic energy metabolism in ischemic and reperfused rat liver. Shock 15, 112117 17. Prass, K., Ruscher, K., Karsch, M., Isaev, N., Megow, D., Priller, J., Scharff, A., Dirnagl, U., and Meisel, A. (2002) Desferrioxamine induces delayed tolerance against cerebral ischemia in vivo and in vitro. J. Cereb. Blood Flow Metab. 22, 520525 18. Tailor, A., and Granger, D. N. (2003) Hypercholesterolemia promotes P-selectin-dependent platelet-endothelial cell adhesion in postcapillary venules. Arterioscler. Thromb. Vasc. Biol. 23, 675680 19. Tsujikawa, A., Kiryu, J., Yamashiro, K., Nonaka, A., Nishijima, K., Honda, Y., and Ogura, Y. (2000) Interactions between blood cells and retinal endothelium in endotoxic sepsis. Hypertension 36, 250258 20. Ishikawa, M., Stokes, K. Y., Zhang, J. H., Nanda, A., and Granger, D. N. (2004) Cerebral microvascular responses to hypercholesterolemia: roles of NADPH oxidase and P-selectin. Circ. Res. 94, 239244 21. Massberg, S., Enders, G., Leiderer, R., Eisenmenger, S., Vestweber, D., Krombach, F., and Messmer, K. (1998) Platelet-endothelial cell interactions during ischemia/reperfusion: the role of P-selectin. Blood 92, 507515 22. Faraci, F. M., and Didion, S. P. (2004) Vascular protection: Superoxide dismutase isoforms in the vessel wall. Arterioscler. Thromb. Vasc. Biol. 24, 13671373 23. Jarasch, E. D., Bruder, G., and Heid, H. W. (1986) Significance of xanthine oxidase in capillary endothelial cells. Acta Physiol. Scand. Suppl. 548, 3946 24. Rey, F. E., Li, X. C., Carretero, O. A., Garvin, J. L., and Pagano, P. J. (2002) Perivascular superoxide anion contributes to impairment of endothelium-dependent relaxation: role of gp91(phox). Circulation 106, 24972502 25. Fukui, T., Ishizaka, N., Rajagopalan, S., Laursen, J. B., Capers, Q. T., Taylor, W. R., Harrison, D. G., de Leon, H., Wilcox, J. N., and Griendling, K. K. (1997) p22phox mRNA expression and NADPH oxidase activity are increased in aortas from hypertensive rats. Circ. Res. 80, 4551

Page 11 of 19 (page number not for citation purposes)

26. Harrison, D. G., Cai, H., Landmesser, U., and Griendling, K. K. (2003) Interactions of angiotensin II with NAD(P)H oxidase, oxidant stress and cardiovascular disease. J. Renin Angiotensin Aldosterone Syst. 4, 5161 27. Griendling, K. K., Sorescu, D., and Ushio-Fukai, M. (2000) NAD(P)H oxidase: role in cardiovascular biology and disease. Circ. Res. 86, 494501 28. Walder, C. E., Green, S. P., Darbonne, W. C., Mathias, J., Rae, J., Dinauer, M. C., Curnutte, J. T., and Thomas, G. R. (1997) Ischemic stroke injury is reduced in mice lacking a functional NADPH oxidase. Stroke 28, 22522258 29. Gorlach, A., Brandes, R. P., Nguyen, K., Amidi, M., Dehghani, F., and Busse, R. (2000) A gp91phox containing NADPH oxidase selectively expressed in endothelial cells is a major source of oxygen radical generation in the arterial wall. Circ. Res. 87, 2632 30. Hebbel, R. P., Morgan, W. T., Eaton, J. W., and Hedlund, B. E. (1988) Accelerated autoxidation and heme loss due to instability of sickle hemoglobin. Proc. Natl. Acad. Sci. USA 85, 237241 31. Hebbel, R. P. (1985) Auto-oxidation and a membrane-associated 'Fenton reagent': a possible explanation for development of membrane lesions in sickle erythrocytes. Clin. Haematol. 14, 129140 32. Hebbel, R. P., Eaton, J. W., Balasingam, M., and Steinberg, M. H. (1982) Spontaneous oxygen radical generation by sickle erythrocytes. J. Clin. Invest. 70, 12531259 33. Saad, S., Salles, S. I., and Velho, P. E. (1991) Decreased reduced glutathione and glutathione reductase activity in subjects with hemoglobin C. Nouv. Rev. Fr. Hematol. 33, 1114 34. Das, S. K., and Nair, R. C. (1980) Superoxide dismutase, glutathione peroxidase, catalase and lipid peroxidation of normal and sickled erythrocytes. Br. J. Haematol. 44, 8792 35. Marva, E., and Hebbel, R. P. (1994) Denaturing interaction between sickle hemoglobin and phosphatidylserine liposomes. Blood 83, 242249
Received October 26, 2004; accepted February 1, 2005.

Page 12 of 19 (page number not for citation purposes)

Table 1 Arterial pressure and venular diameter in the different experimental groups Group WT/WT S / WT S / SOD1-TgN S / gp91phox/ S / WT-allopurinol i.p. S / WT- DFO Arterial pressure, mmHg 69.5 2.6 70.3 4.7 70.0 3.8 68.7 5.4 66.4 4.1 76.6 2.5 Venular diameter, m 38.4 1.2 39.1 1.1 36.3 0.8 38.3 2.0 41.0 2.9 38.1 2.0

Asterisk indicates P < 0.05 versus S / WT chimeric mice.

Page 13 of 19 (page number not for citation purposes)

Table 2 Hemoglobin, hematocrit and methemoglobin values in the different experimental groups Group WT (baseline) S (baseline) WT (H/R) S (H/R) WT/ WT (H/R) S / WT (H/R) Hemoglobin, g/dL 3.1 0.3 11.2 0.5 * 12.2 0.3 9.2 0.4 * 11.7 0.4 8.6 0.4 ** Hematocrit, % 39.2 0.8 33.6 1.8 * 40.6 1.2 35.2 2.2 * 42.0 1.1 34.2 1.1 ** Methemoglobin, % 1.1 0.2 1.4 0.1 * 1.2 0.0 1.9 0.2 * 1.2 0.1 2.0 0.1 **

Five to 16 animals were studied in each group. *P < 0.05 versus WT mice; **P < 0.05 versus WT/ WT chimeric mice.

Page 14 of 19 (page number not for citation purposes)

Fig. 1

Figure 1. Leukocyte and platelet adhesion responses in wild-type (WT) mice receiving bone marrow from either WT
(WT/WT) or S (S/WT) donor mice, and in copper-zinc superoxide dismutase (SOD-1) transgenic (TgN) mice receiving bone marrow from S mice (S/SOD1-TgN). Asterisk indicates P < 0.05 compared with WT/WT; #P < 0.05 relative to S/WT mice.
Page 15 of 19 (page number not for citation purposes)

Fig. 2

Figure 2. Leukocyte and platelet adhesion responses in wild-type (WT) mice receiving bone marrow from either WT

(WT/WT) or S (S/WT) donor mice, and in gp91phox/ mice receiving bone marrow from S mice (S/ gp91phox/). Asterisk indicates P < 0.05 compared with WT/WT; #P < 0.05 relative to S/WT mice.

Page 16 of 19 (page number not for citation purposes)

Fig. 3

Figure 3. Leukocyte and platelet adhesion responses in wild-type (WT) mice receiving bone marrow from either WT
(WT/WT) or S (S/WT) donor mice, and in S/WT mice treated intraperitoneally (ip) with allopurinol. Asterisk indicates P < 0.05 compared with WT/WT.
Page 17 of 19 (page number not for citation purposes)

Fig. 4

Figure 4. Leukocyte and platelet adhesion responses in wild-type (WT) mice receiving bone marrow from either WT
(WT/WT) or S (S/WT) donor mice, and in S/WT mice treated with the iron chelator desferrioxamine (DFO). Asterisk indicates P < 0.05 compared with WT/WT; #P < 0.05 relative to S/WT mice.
Page 18 of 19 (page number not for citation purposes)

Fig. 5

Figure 5. Schematic diagram summarizes the relative importance of NADPH-oxidase-derived superoxide and irondependent radicals in the production of SCD-associated adhesion of leukocytes and platelets in the cerebral microcirculation.

Page 19 of 19 (page number not for citation purposes)