Enhanced Pulmonary and Systemic Response to Endotoxin in Transgenic Sickle Mice J. David Holtzclaw, Daniel Jack, Samuel M. Aguayo, James R. Eckman, Jesse Roman, and Lewis L. Hsu Division of Hematology/Oncology and Bone Marrow Transplantation, Department of Pediatrics, Emory University School of Medicine; Division of Medicine, Atlanta Veterans Affairs Medical Center, Morehouse School of Medicine; Winship Cancer Institute, Hematology/Oncology Department, Emory University School of Medicine; and Division of Pulmonary, Allergy, and Critical Care Medicine, Emory University School of Medicine, Atlanta, Georgia
Some suggest that sickle cell disease (SCD) is associated with a “proinflammatory state” that predisposes patients to acute chest syndrome in the setting of triggering factors. Conflicting data emerged when inflammation markers in SCD were compared with healthy individuals. Therefore, we examined transgenic sickle and control mice at baseline and with endotoxin (LPS) intraperitoneal injection to determine whether a proinflammatory state truly exists. At baseline, sickle mice had elevated levels of circulating leukocytes and soluble vascular cell adhesion molecule 1 (sVCAM-1). No other differences were observed at baseline or in response to saline. However, LPS challenge was associated with significant increases in mortality (p ⬍ 0.05), airway tone (p ⬍ 0.03), serum and bronchoalveolar lavage levels of cytokines tumor necrosis factor-␣ (p ⬍ 0.03), interleukin-1 (p ⬍ 0.02), and sVCAM-1 (p ⬍ 0.01) in sickle mice compared with control subjects. Furthermore, 4 hours after LPS, microarray analysis identified 413 genes differentially expressed in the sickle mice (n ⫽ 5) compared with only 7 in the control subjects (n ⫽ 5). No difference in lung parenchyma was observed by light microscopy. This enhanced response to LPS suggests that the sickle red blood cell confers a subclinical “proinflammatory state.” This enhanced response to inflammatory insult, particularly by adhesion molecules such as sVCAM-1, could play a role in the increased susceptibility to pulmonary dysfunction that has been observed clinically in SCD. Keywords: inflammation; cytokines; adhesion molecules; plethysmography; cDNA microarray
A leading cause of mortality in sickle cell disease (SCD) in the United States is acute chest syndrome (ACS) (1). ACS is a complex and unpredictable lung condition manifested by new infiltrate on the chest radiograph of a sickle cell patient and a constellation of other possible symptoms (e.g., fever, chest pain, cough, hypoxemia, and respiratory distress) (2). Multiple etiologies appear to contribute to the pathophysiology of ACS (3): (1 ) infection, (2 ) marrow fat embolism or other obstruction of blood flow, (3 ) sickle cell entrapment (sequestration) within the microcirculation of the lungs, (4 ) hypoxia, (5 ) atelectasis caused by pain in the chest wall, and (6 ) asthma or abnormal ventilation patterns (4). Histopathology in autopsy specimens supports the multifactorial nature of severe ACS (5–7). ACS continues to be
a therapeutic challenge mainly because it has an unpredictable course, can be triggered by a number of acquired conditions, and because the mechanisms and factors involved in the development of the syndrome are unknown. Several researchers have suggested that a “proinflammatory state” exists in SCD that predisposes patients to ACS or vasoocclusion (VOC) in response to one or more triggering factors (i.e., infection, embolism) (3, 8–13). Evidence supporting a proinflammatory state include elevated serum levels of inflammatory mediators tumor necrosis factor (TNF)-␣ (14–18), interleukin (IL)-1 (19), IL-6 (20), and IL-8 (21) at baseline or during ACS (5, 22–24) suggesting a systemic inflammatory response. However, similar studies failed to find elevated TNF-␣ (17, 25, 26), IL-1 (19, 26), IL-6 (17, 19, 25–28), or IL-8 (26). Elevated steadystate white blood cell (WBC) counts appear to be a significant predictor of adverse outcomes such as ACS in pediatric sickle cell patients (29), and conversely, the landmark hydroxyurea study found that a therapeutic response of sickle cell patients to hydroxyurea was associated more strongly with a decrease in WBC counts than with the rise in fetal hemoglobin (30, 31). Several investigators interpret abnormal endothelial adhesion as evidence of a proinflammatory state (12, 32). Great interest has been focused on vascular cell adhesion molecule-1 (VCAM-1) (33, 34), shown to be elevated at baseline (8, 27, 28, 31, 32, 35) with additional elevations during VOC and ACS (8, 27), followed by a decrease with hydroxyurea or transfusion therapy (29, 35). Several speculate (8, 12) that inflammatory activation of adhesion via VCAM-1 is a key early event in VOC or ACS, and intravital microscopy supports this pathophysiology in -thalassemic transgenic sickle mice (36, 37) and perfused rat retina (38). Overall, the evidence for a “proinflammatory state” remains conflicting with the sole exception of sVCAM-1. In this study, we examined transgenic sickle mice at baseline and in response to endotoxin or saline and discovered that transgenic sickle mice had an enhanced response to inflammatory challenge compared with control subjects. Some results have been previously reported in abstract form (39, 40).
METHODS (Received in original form February 16, 2003; accepted in final form December 11, 2003) Supported by National Institutes of Health grants K08 HL03811 (L.L.H.), by an IRACDA grant 3K12 GM 00680-03S1 to Emory University School of Medicine, by a Glaxo Research Gift to the Atlanta Research and Education Foundation (S.M.A.), and by a Merit Review Award at the Atlanta Veterans Affairs Medical Center (S.M.A.). Correspondence and requests for reprints should be addressed to Lewis L. Hsu, M.D., Ph.D., Pediatric Hematology, St. Christopher’s Hospital for Children, Drexel University College of Medicine, Erie Ave. at Front St., Philadelphia, PA 19134. E-mail:
[email protected] This article has an online supplement, which is accessible from this issue’s table of contents online at www.atsjournals.org Am J Respir Crit Care Med Vol 169. pp 687–695, 2004 Originally Published in Press as DOI: 10.1164/rccm.200302-224OC on December 18, 2003 Internet address: www.atsjournals.org
Buffers, Reagents, and Treatments Mice received an intraperitoneal injection of the endotoxin lipopolysaccharide (LPS, 10 g/g body weight) or saline (5 l/g) as control. LPS (E. coli O111:B4; Sigma, St. Louis, MO) was suspended in 2 ml of sterile phosphate-buffered saline (PBS) and filtered (0.22-m filter; Whatman, Clifton, NJ). To focus on the early phase of LPS response, all pulmonary measurements and tissue samples were collected 4 hours after intraperitoneal injection. PBS and ethylenediaminetetraacetic acid–saline were obtained from Fisher Scientific (Pittsburgh, PA).
Animals Used in the Study Transgenic sickle mice (100% sickle hemoglobin [HbS]) expressing exclusively human ␣- and sickle-globin genes (Tg[Hu-miniLCR␣1G␥A␥␦s]
688
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 169 2004
Hba0//Hba0 Hbb0//Hbb0) were provided by the Lawrence Berkley National Laboratory (41). Age-matched colony control subjects were hemizygous mice expressing approximately 20% sickle and 80% mouse single globin. For some studies, age-matched C57BL/6 mice were used as additional control subjects. Additional colony details are in an online supplement. All procedures followed Association for Assessment and Accreditation of Laboratory Animal Care (AALAC) guidelines and were approved by the Institutional Animal Care and Use Committee of Emory University.
Plethysmograph Pulmonary function was assessed using noninvasive, whole-body barometric plethysmography (40, 42–44). Unrestrained and unsedated mice were allowed 10 minutes for acclimation in the plethysmograph and then Penh (enhanced pause) was measured using commercial software (Buxco Electronics, Sharon, CT). Results 4 hours after injection were normalized to Penh values collected before injection. Additional details are in the online supplement.
DNA Microarray Minced lung samples were disrupted and homogenized (Mixer Mill MM 300; Qiagen). Total RNA was isolated using Qiagen RNeasy Minikit (Qiagen), converted to cRNA, and hybridized to the Affymetrix murine U74A version 2 GeneChip (Affymetrix, San Jose, CA) and visualized (Microarray Suite 5.0 software) using manufacturer protocols (46). Normalization and statistical analyses of the data were performed using GeneSpring (Silicon Genetics, Redwood City, CA).
Statistical Methods Data shown represent mean ⫾ SEM and compared by analysis of variance or the Student’s t test unless otherwise denoted. A p of less than 0.05 was considered to be significant. We performed data management and statistical analyses using Excel (Microsoft Corporation, Redmond, WA) unless denoted.
RESULTS Sickle Mice Show Increased Mortality in Response to LPS
Tissue Sampling Lung and blood samples were obtained at euthanasia 4 hours after injection by cervical dislocation under isoflurane anesthesia. Blood samples were obtained immediately by terminal blood draw into heparinized microtubes, centrifuged, and serum frozen for batch analysis. Lung samples for RNA analysis were obtained by immediate necropsy, minced, and promptly placed in RNALater (Qiagen, Valencia, CA). For histologic samples, each lung was inflation-fixed by intratracheal buffered formalin under 5 cm H2O pressure.
Hematologic and Cytokine Analysis Small blood samples (n ⫽ 4 for each group) were obtained retroorbitally before euthanasia for complete blood counts performed on a Hemavet 1500R blood analyzer (CDC Technologies, Oxford, CT) following the manufacturer’s protocol. Serum (n ⭓ 4 for each group) and bronchoalveolar lavage (BAL) (n ⫽ 4 for 20% HbS and n ⫽ 4 for 100% HbS) fluid levels of TNF-␣, IL-1, IL-6, and soluble VCAM-1 were assayed in duplicate using commercially available ELISA kits (R&D Systems, Minneapolis, MN).
BAL BAL was performed using 1.5 cc twice of ethylenediaminetetraacetic acid–saline after right atrial perfusion of crystalloid solution and isolation of the lungs with the tracheotomy tube still in place. BAL fluid (BALF) was immediately centrifuged, and supernatant samples were frozen until assayed. To account for the dilution of BALF, inflammatory markers were normalized by IgA levels, also assayed in duplicate by ELISA (Bethyl Laboratories, Montgomery, TX) (45).
Adult sickle and hemizygote mice received an intraperitoneal injection of LPS (10 g/g weight). In preliminary studies, several single doses of LPS (0–25, g/g weight) were examined in several types of mice from this colony (data not shown). Initially, we wanted to use the maximum sublethal dose, which would have been approximately 8 g/g weight for a normal hemoglobin (Hb) mouse of the same mixed background as the sickle mouse. However, we settled on the 90% survival dose (10 g/g) because it slightly traversed the line between acute disorder and mortality, mimicking the clinical scenario of ACS, which also traverses the line between acute disorder and mortality (ACS is the leading cause of both mortality and morbidity in SCD) (1, 47). The Kaplan-Meier Survival Plot shown in Figure 1 depicts the effect of LPS on mouse phenotype survival. Within 2 days, only 1 of the 15 normal (C57BL/6) and none of the hemizygote mice were lost. However, during the same time period, 60% of the sickle mice died (p ⬍ 0.03 by log-rank test) (48). These data suggest that the sickle mice were less resilient to inflammatory challenge than the hemizygous or normal control subjects. However, the mechanism of death was not determined. Therefore, we continued our study shifting our focus from the organism level to the system level specifically examining pulmonary function and immune system response. Sickle Mice Show Increased Airway Tone in Response to LPS
We measured Penh at baseline and again 4 hours after intraperitoneal injection of LPS or PBS. Longitudinal comparison of Penh
Figure 1. LPS mortality is higher in the 100% sickle mouse. This Kaplan-Meier survival plot depicts the effect of LPS on mouse phenotype survival. Survival was measured after an intraperitoneal injection of LPS (10 g/gm weight). Within 2 days, only 1 of 21 normal subjects or heterozygotes was lost. However, 60% of the sickle mice died as a result of the inflammatory challenge, which was statistically different (p ⬍ 0.03 by log-rank test) than the normal or 20% sickle hemoglobin (HbS) mice.
Holtzclaw, Jack, Aguayo, et al.: A Proinflammatory State in Sickle Mice
689
for each individual mouse was calculated: The postinjection mean Penh value was divided by the baseline mean Penh value to calculate the percentage change in Penh. Figure 2 demonstrates that only the sickle mice given LPS had more than a 10% change in Penh from baseline values. Furthermore, the sickle mice given LPS experienced a statistically significant increase in Penh when compared with either the sickle mice given saline (p ⬍ 0.0001) or the hemizygote mouse also given LPS (p ⬍ 0.005). Although LPS is known to cause pulmonary inflammation (49), it only had a large effect on Penh lung function in the sickle mice and not in the 20% HbS. If lung histology did not change at this early time point (4 hours) after LPS, we could infer that the Penh rise indicates higher airway tone in the sickle mice. Indeed, this was the case, as no differences in lung histology between mouse phenotypes were observed.
lungs was consistent with the known inhibitory effect of mouse hemoglobin on sickle hemoglobin polymerization.
No Detectable Difference in Histology between Mouse Phenotypes
Histology of sickle and hemizygote mice lungs, with or without LPS, showed little sign of inflammation by light microscopy (Figure 3). Alveoli contained few red blood cells (RBCs), leukocytes, or exudates; these were patchy and focal and consistent with artifacts of changes at the time of death. Bronchioles did not have increased inflammatory cells. Airway smooth muscle and basement membrane were not thickened, and vessels did not show thrombi. Chronic vascular changes were seen in arteriolar smooth muscle (data not shown here, but described in detail in a separate study by Kean and colleagues) (50). Elongated and sickle erythrocyte morphology were often seen in the vasculature and scattered in the alveoli of sickle mouse lungs, but not in the hemizygote lungs. The presence of elongated RBC in the sickle mouse lungs does not prove that the RBCs were sickled in vivo because both terminal acidosis and hypoxia or formalin fixation can induce RBC sickling postmortem. The absence of elongated sickle RBC in hemizygote
Figure 2. Only 100% HbS mice increase airway obstruction with LPS as measured by Penh. This figure shows the percent change in Penh ⫾ SD, 4 hours after intraperitoneal injection with either LPS (10 g/gm) or saline normalized to the baseline value taken before intraperitoneal injection. On average, both the normal (BL/6) and heterozygotes (20% HbS) experienced less than 10% change from baseline in response to either LPS or saline. However, the homozygous sickle mice (100% HbS) experienced a 35% increase in airway obstruction in response to LPS, significantly different than the results obtained from either the 100% sickle mice given saline or the 20% sickle mice given LPS. PBS ⫽ phosphate-buffered saline.
Sickle Mice Demonstrate Increased Leukocytosis
Average peripheral blood leukocyte counts ⫾ SD for four mice in each treatment group are shown in Figure 4. At baseline (PBS), sickle (100% HbS) mice have significantly higher concentrations of circulating total leukocytes (including lymphocytes, neutrophils, and monocytes) compared with hemizygote (p ⬍ 0.02 by Student’s t test) and normal hemoglobin control subjects (C57BL/6, p ⬍ 0.01). In response to LPS injection, all three mice phenotypes experienced a decrease in leukocytes, neutrophils, lymphocytes, monocytes, and platelets (platelet data not shown). However, the sickle mice experienced the greatest and only statistically significant drop in circulating total WBCs (p ⬍ 0.001), as well as lymphocytes (p ⬍ 0.001), neutrophils (p ⬍ 0.05), and monocytes (p ⬍ 0.05). Hemizygote mice (20% HbS) had a significant decrease only in lymphocytes (p ⬍ 0.05) in response to LPS. To examine this proinflammatory state further, we transitioned from a cellular level to a molecular level and measured protein expression. Sickle Mice Show Increased Cytokine and sVCAM-1 Levels
Various inflammatory cytokines and chemokines from sickle cell patients have been reported to be elevated at baseline (14–16, 18, 51) and during ACS (5, 22, 23), suggesting a systemic inflammatory response. Therefore, serum cytokine levels TNF-␣, IL-1, IL-6, and sVCAM-1 were measured in duplicate by commercially available ELISA kits for normal (BL/6, n ⫽ 4), hemizygote (20% HbS, n ⭓ 4), and sickle (100% HbS, n ⭓ 5) mice and were plotted in Figure 5 (mean values ⫾ SEM). At baseline, only sVCAM-1 (Figure 5C) was detectible in any mouse, and the sickle mice expressed higher levels of sVCAM-1 than either control (p ⬍ 0.03). These findings are consistent with elevated sVCAM-1 in human SCD (8, 27, 29, 32). Saline injection produced results similar to the baseline, with all three groups of mice expressing minimal levels of TNF-␣, IL-1, and IL-6. Sickle mice have higher serum sVCAM-1 levels than control subjects at baseline or after an intraperitoneal injection of saline. In response to LPS, all three types of mice had increased TNF-␣ (p ⬍ 0.01), IL-1 (p ⬍ 0.01), sVCAM-1 (p ⬍ 0.001), and IL-6 (p ⬍ 0.001) by two-factor analysis of variance. The sickle mice showed significantly greater increases in serum TNF-␣, IL-1, and sVCAM-1 compared with either of the control subjects. Thus, although all three groups of mice expressed similar levels at baseline, the sickle mice have an enhanced serum cytokine response to LPS when compared with either the hemizygous colony control subjects or normal (BL/6) control subjects. Similar to the serum protein measurements, BALF from sickle and hemizygote mice was examined for the presence of TNF-␣, sVCAM-1, and IL-6 by ELISA. BALF was collected 4 hours after intraperitoneal injection with either LPS or saline. Protein levels in BALF were normalized by the amount of IgA present in the sample, also determined by ELISA, to adjust for variability in lung fluid volume obtained by lavage (45). BAL results, given in Figure 6 (mean values ⫾ SEM), mirrored the results obtained from serum measurements. BAL at baseline and in response to PBS of both the hemizygote and sickle had nearly identical levels. LPS significantly elevated BAL TNF-␣ levels in both the hemizygote (66% increase, n ⫽ 4) and sickle (860% increase, n ⫽ 5) mice (Figure 6A) by two-factor analysis of variance. Similarly, LPS significantly elevated IL-6 levels in BAL of both types of mice in response to LPS (Figure 6C) by two-factor analysis of variance, but the sickle mice (n ⫽ 5)
690
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 169 2004
Figure 3. No difference in lung histology between mice with and without LPS. For histologic samples, each lung was inflation fixed by intratracheal buffered formalin, under 5 cm H2O pressure, and then the trachea was tied off to maintain inflation while the lungs were allowed to fix. After standard histologic processing and embedding in paraffin, sections (5 m) were stained with hematoxylin and eosin and analyzed by light microscopy (Olympus AX70) using a ⫻68 magnification oil objective. Digital images were captured by CCD imaging using MagnaFire 2.0 software (Optronics, Goleta, CA). Alveoli contained few red blood cells (RBCs) or white blood cells (WBCs). Bronchioles did not have increased inflammatory cells or chronic changes. Airway smooth muscle and basement membrane were not thickened, and vessels did not show thrombi or wall thickening.
express almost double the BALF IL-6 as the hemizygote mice (n ⫽ 4). LPS also significantly increased BALF sVCAM-1 levels in the sickle mice (n ⫽ 4) compared with baseline or saline control (Figure 6B). These BALF findings support the serum findings that the sickle mice have an exaggerated response to challenge compared with hemizygous or normal control subjects. Sickle Mice Show Increased Gene Expression in Response to LPS
Figure 4. Elevation in leukocytes in 100% HbS mice at baseline. WBC counts (n ⫽ 4) were also obtained 4 hours after an intraperitoneal injection by small blood samples obtained retro-orbitally before euthanasia. Blood samples were analyzed by an automated veterinary blood analyzer (MASCOT; CDC Technologies, Inc., Oxford, CT) optimized for murine blood. Mean leukocyte concentrations ⫾ SD were plotted above for each cell type. At baseline, shown in the solid bars, we observed elevated levels of circulating WBCs, including lymphocytes (LY), neutrophils (NE), and monocytes (MO) in 100% HbS sickle mice over the heterozygous or normal control subjects (p ⬍ 0.01). In response to LPS, shown in the hashed bars, we observed a significant decrease in circulating WBCs in 100% sickle mice, but not the hemizygotes.
Briefly reported here are the expression levels of approximately 12,000 genes per sample, analyzed in 20 high-density, oligonucleotide cDNA microarrays (five arrays for each treatment group: 20% HbS PBS, 20% HbS LPS, 100% HbS PBS, and 100% HbS LPS). cDNA microarray analysis was performed with GeneSpring software (Silicon Genetics). Affymetrix data files were imported into GeneSpring and were normalized per gene based on the median gene intensity value. Statistical analysis was performed by the nonparametric, Wilcoxon-Mann-Whitney test, with the p value cutoff set to 0.05, and using the Benjamini and Hochberg False Discovery Rate multiple testing correction (52). Therefore, only 5% of the identified genes would be expected to pass the restriction by chance. Comparing hemizygote mice (20% HbS) given LPS to hemizygote mice given saline (Table 1), only seven genes were identified as statistically different. However, the same comparison in the sickle mice (100% HbS) yielded over 400 genes, nearly 60 times the response observed in the hemizygote mice. Again, these data support our other findings that the sickle mice had an enhanced response to LPS that was
Holtzclaw, Jack, Aguayo, et al.: A Proinflammatory State in Sickle Mice
691
Figure 5. Enhanced response to LPS in 100% HbS mice in serum cytokines. Average serum cytokine levels of tumor necrosis factor-␣ (TNF-␣) (A ), interleukin (IL)-1 (B ), soluble vascular cell adhesion molecule 1 (sVCAM-1) (C ), and IL-6 (D ) were measured by commercially available ELISA kits (R&D Systems) and were plotted ⫹ SEM. In all three types of mice, TNF-␣, IL-1, and IL-6 levels were undectable or minimal at baseline or in response to PBS. However, soluble VCAM-1 baseline levels were significantly different between mice phenotypes (p ⬍ 0.03). *An significant increase in serum protein expression in the LPS treated mice compared with baseline or PBS-treated mice by two-way analysis of variance. Although LPS significantly increased TNF-␣, IL-1, sVCAM-1, and IL-6 in all mice, LPS-treated 100% sickle mice typically demonstrated a twofold or greater increase over LPS treated hemizygotes or normal mice. Similar results were obtained with sVCAM-1.
not present at baseline or in the hemizygote mice. Detailed pathway analyses of this differential gene expression are in progress, together with reverse transcriptase–polymerase chain reaction validation of the results, and will be reported separately.
DISCUSSION We determined baseline, saline, and early LPS-induced pulmonary inflammation in transgenic sickle mice to address the question of whether or not a proinflammatory state exists in SCD. Our analysis was performed at multiple levels, including whole animal (survival), organ function (plethysmograph), cellular (WBC counts), and molecular levels (protein and mRNA expression). LPS was chosen because it is a well-characterized experimental model for acute systemic and pulmonary inflammation, triggering inflammatory pathways mediated by TNF-␣ and IL-1, and activating endothelial cells. Furthermore, infection is one of the triggers for sickle cell ACS and VOC. Although E. coli infection and LPS would be less common in SCD clinically than Gram-positive bacterial (29) and viral infections, the earliest responses to activation of WBC and endothelium by infection can be initially modeled by LPS injection rather than actual infection. In this article, we present a transgenic animal model for sickle acute chest inflammatory response, which has several advantages. First, because of the pathogen-free housing environment, the transgenic mouse model allows us the ability to establish a true baseline condition with no exogenous infections. Second, through noninvasive methods such as whole-body plethysmograph, we can perform serial measurements and observe the progression of the disease or time course of events after inflammatory challenge on the same mouse. Similarly, these mice enable us to perform invasive studies such as histology, cDNA microarrays, and BAL. Furthermore, this transgenic sickle mouse model displayed an enhanced response to inflammation
stimulus and enhanced airway reactivity (Figure 2) similar to the response observed in human patients (3, 5, 22, 23, 53). Although no mouse is a perfect model for SCD, the BerkeleyPaszty mouse closely parallels the pathophysiology of the human disease (54), including severe anemia, red cell morphology at low oxygen tension, extensive organ damage, irreversible sickle cells, and VOC (55). Consistent with our model, a recent study by Belcher and colleagues (56) demonstrated that the BerkeleyPaszty mouse, like human sickle cell patients, has an active inflammatory response at baseline and is a good model to study vascular inflammation and potential antiinflammatory therapies. Limitations of this sickle mouse include the use, for breeding purposes, of a super-enriched feed containing additional vitamins, antioxidants, and l-arginine (which produces free nitric oxide). Furthermore, we recognize that the intraperitoneal injection of LPS inflammatory challenge is not the same as clinical infection. Finally, we have focused on the earliest acute events at 4 hours after LPS and not other subacute or chronic time points; thus, we may have missed some later events like acute respiratory distress syndrome histology or lung fibrosis. At baseline, sickle mice had elevated circulating leukocyte counts (Figure 4) and serum levels of sVCAM-1 (Figure 5) compared with both hemizygous and normal control subjects. These features of the animal model are similar to human SCD. These observations of high WBC counts at baseline in sickle mice match clinical observations (6) and have been well established by clinical studies as a significant predictor of more sickle cell complications (29, 31). Similar to the data presented in Figures 1 and 2 where Sickle mice have increased mortality and airway tone, sickle cell patients with higher WBCs are more likely to develop recurrent ACS and other SCD complications such as frequent pain and stroke (29). More importantly, the leukocyte counts presented in Figure 4 are consistent with an upregulated immune response or proinflammatory state in the sickle mice at
692
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 169 2004
Figure 6. Enhanced response to LPS in 100% HbS mice in BALF cytokines cytokine levels of TNF-␣ (A), sVCAM-1 (B), and IL-6 (C) in bronchoalveolar lavage fluid (BALF) was also measured by commercially available ELISA kits (R&D Systems) and were plotted ⫾ SEM. Mice were euthanized, and blood was cleared from the lungs using right atrial perfusion. Isolated lungs were lavaged twice with 1.5 cc of ethylenediaminetetraacetic acid–saline. BALF was immediately filtered and centrifuged, and ELISAs were performed on the supernatant. To adjust for variation in recovery volumes of BALF, results were normalized to IgA concentrations, which remain constant during inflammation. BAL protein levels were similar to serum. *A significant increase in serum protein expression in the LPStreated mice compared with baseline or PBS-treated mice by two-way analysis of variance. At baseline, BAL results were similar for all types of mice. LPS-treated 100% sickle mice demonstrated a greater response to LPS than LPS-treated hemizygotes. Only the 100% HbS mice saw a significant increased in sVCAM-1 levels in response to LPS compared with the 100% HbS mice treated with PBS.
baseline. Likewise, elevated sVCAM-1 levels at baseline (8, 27) and further elevation during ACS (8) were previously shown in humans. Decreased plasma sVCAM-1 levels were measured during hydroxyurea treatment (57) or in sickle patients with below-average cell-free plasma hemoglobin (58). Apart from these results, the baseline condition of sickle mice had little apparent difference from control subjects when examined at multiple levels of lung structure and function. The baseline cytokine findings are not typical for a proinflammatory state in terms of multiple elevated serum cytokines yet actually are similar to human SCD at baseline. Literature on cytokines at baseline in sickle patients includes data sets with many subjects whose cytokine levels are no different from control subjects, with only a few patients, possibly outliers, with
cytokine levels increased at baseline (20). These sickle mouse results suggest that there are only markers for very subtle signs of a state of activation at baseline: increased sVCAM-1 and increased WBCs. The rest of the system is compensated and not a typical picture of elevated proinflammatory cytokines at baseline. As expected, the response to injection of saline did not differ from baseline, indicating that minimal response was caused by the stress of intraperitoneal injection with a needle. In an exaggerated response to LPS at several different levels, sickle mice experienced increased mortality, Penh, serum and BALF cytokine levels, decrease WBCs, and greater differences in mRNA expression as measured by cDNA microarray. Although these differences to LPS were measured at multiple levels, no difference
TABLE 1. ENHANCED RESPONSE TO LIPOPOLYSACCHARIDE IN 100% SICKLE HEMOGLOBIN MICE IN MESSENGER RNA EXPRESSION Wilcoxon-Mann-Whitney Comparisons Genes, No. LPS effect on 20% HbS LPS effect on 100% HbS
7 413
Genes of Interest E-selectin,
3 cytokines, and chemokine genes 5 TNF genes, IL-1, IL-6, IFN-␥, IFN-, sVCAM-1, ICAM-1, E-selectin, P-selectin
Definition of abbreviations: HbS ⫽ sickle hemoglobin; ICAM-1 ⫽ intercellular adhesion molecule-1; IL ⫽ interleukin; sVCAM-1 ⫽ soluble vascular cell adhesion molecule 1; TNF ⫽ tumor necrosis factor. Four hours after intraperitoneal injection, mice were killed and one lung lobe was removed for complementary DNA microarray analysis. Microarrays were obtained from Affymetrix and lung tissue complementary RNA was isolated and hybridized following manufacturer’s protocol. Murine Genechip U74A version 2 chips (Affymetrix) were used, which contain 12,488 genes, of which approximately 6,000 are full-length and functionally identified genes. Genes considered significantly different between groups were analyzed using GeneSpring software (Silicon Genetics), using multiple Wilcoxon-Mann-Whitney comparisons, a p value cutoff of 0.05, and the Benjamini and Hochberg false discovery rate multiple testing correction. This table shows the number of genes either upregulated or downregulated in response to LPS for both the 100% HbS and 20% sickle mice. As shown, the 100% sickle mice had 413 genes with differential response to LPS. This was a markedly greater response to LPS than that of the heterozygous control mice. Results for TNF-␣, IL-1, and sVCAM-1 were confirmed by RT-PCR (data not shown). Furthermore, the 100% sickle mice had an upregulation of a number of adhesion molecules such as sVCAM-1, sICAM-1, E-selectin, and P-selectin, as well as an upregulation in cytokines, such as TNF, IL-1, IL-6, and IFN-␥, .
Holtzclaw, Jack, Aguayo, et al.: A Proinflammatory State in Sickle Mice
693
in alveoli airway or bronchiole histology was observed by light microscopy. The relationship between sickle cell vascular problems and pulmonary function remains unclear, but the clinical picture of SCD includes increased airway hyperreactivity at baseline (59, 60). At the 4-hour time point, we are already seeing an increase in airway tone due to LPS in the sickle mice that does not appear in the hemizygote mice (Figure 2) with the same insult. These data, along with the survival data presented in Figure 1, suggested a pathophysiologic response to inflammatory challenge in the sickle mice that is not present in the hemizygote mice or normal control subjects. These animal plethysmographic data are consistent with results from a large multicenter clinical study that demonstrated a high incidence of airway obstruction (3, 53) in sickle cell patients. Clinical lung complications in SCD become severe when an ordinary community-acquired viral or bacterial pneumonia can trigger ACS, a far more severe illness than their siblings with sickle trait would have in response to the same inflammation. Respiratory deterioration in sickle cell patients can lead to regional hypoxia, ventilation–perfusion mismatch, pulmonary edema, and then acute lung injury and acute respiratory distress syndrome (7). The increase in mortality in sickle mice compared with control subjects (Figure 1) was quite striking. LPS levels that were tolerated by the normal or hemizygous mice are fatal to the sickle mice. These sickle mouse findings with inflammatory challenge are similar to human sickle ACS, especially the increase in Penh. The sickle mice had a 40% increase in sVCAM-1, a 400% increase in TNF-␣, and a 150% increase in IL-1 when compared with the hemizygote mice. This enhanced response by TNF-␣ and IL-1, both initiators of the inflammatory system, combined with the elevated levels of sVCAM-1 and leukocytes at baseline could explain the increase in airway tone and mortality, similar to a “two-hit” paradigm. Whole-body plethysmography has many advantages: (1 ) serial observations on the same animal over time, (2 ) avoiding invasive procedures or physical restraint that may alter the airway physiology, and (3 ) avoiding anesthetics that may alter breathing pattern and airways reactivity. The major output of this method is the “Enhanced Pause” (Penh), a dimensionless parameter that incorporates ratios of exhalation time to relaxation time and peak inspiratory flow to peak expiratory flow (see additional online supplement for more details). Proper interpretation of the enhanced pause (Penh) output parameter is complex, and many pitfalls exist (61, 62). However, nearly all pitfalls such as body temperature and humidity can be avoided by longitudinal comparisons (normalizing by baseline conditions as done in Figure 2), showing no histologic changes (as shown in Figure 3), measuring residual capacity and tidal volume independently (61, 63) or demonstration that changes in Penh are reversible in the same individual mouse (42). Increased airway tone occurred without any histologic change but was associated with increased cytokine expression, suggesting that airway obstruction in SCD is not typical asthma pathophysiology. Recently, asthma has been found to be associated with both risk and frequency of ACS (64). Interestingly, we have been unable to detect a histologic correlate so far at these early time points. Speculatively, the pathophysiologic problem in the sickle mouse model under normoxic condition is either abnormal endothelial adhesion, or abnormal oxidative free radicals, which consume nitric oxide via xanthine oxidase activity (65), or increased cell-free hemoglobin, which decreases nitric oxide bioavailability (58), or some other “hit” initiating an acute event. This area is rich for future studies. Further studies may characterize whether there is abnormal endothelial adhesion in lung, parallel to cremaster microcirculation (36, 66).
Lung problems in SCD may have conceptual similarities to the susceptibility of those with alcoholism to acute respiratory distress syndrome with community-acquired pneumonia (67) where the patient is chronically compensated for “one-hit” pathophysiology, but has no reserve, and is easily tipped over into acute catastrophic pathophysiology with another “hit.” In alcoholism, the first hit appears to be a diminished antioxidant reserve of hepatic glutathione from chronic alcohol exposure, setting the stage for the second hit (pneumonia) to cause acute lung injury and acute respiratory distress syndrome. An alcoholic rat animal model given experimental inflammation by LPS demonstrates increased acute lung injury (67). Even though the mechanism of the first hit to the lungs may be different in SCD than the oxidative hit in alcoholism, the concept that a chronic disease causes decreased reserve for coping with an inflammatory second hit and allows acute lung injury or acute respiratory distress syndrome to develop from community-acquired pneumonia is strikingly similar to the clinical picture of sickle ACS. In summary, the results of this study support a subtler concept of a “proinflammatory state,” one that is sufficiently compensated at baseline to differ from control subjects only in elevated WBCs and sVCAM-1 but ready to decompensate with LPS challenge. Furthermore, both endothelium and leukocytes contribute to the response to inflammation in this animal model of SCD, including adhesion molecules and cytokines, to produce the exaggerated response even without hypoxic exposure. These findings support recent emphasis (8, 66) that the pathophysiology of SCD involves more than RBC sickling and VOC but may help identify the mechanisms connecting RBC to pulmonary endothelium and leukocytes (8, 12, 32). Conflict of Interest Statement : J.D.H. has no declared conflict of interest; D.J. has no declared conflict of interest; S.M.A. has no declared conflict of interest; J.R.E. has no declared conflict of interest; J.R. has no declared conflict of interest; L.L.H. has no declared conflict of interest. Acknowledgment : The authors appreciate Jennifer Perry (Emory University) for her expertise with maintaining the sickle mouse colony, Trudy McDermott (Atlanta Veterans Affairs Medical Center), Giri Polavarapu, Melissa Bowman, Lou Ann Brown, and Frank Harris (all of Emory University) for their technical assistance.
References 1. Platt OS, Brambilla DJ, Rosse WF, Milner PF, Castro O, Steinberg MH, Klug PP. Mortality in sickle cell disease: life expectancy and risk factors for early death. N Engl J Med 1994;330:1639–1644. 2. Castro O, Brambilla DJ, Thorington B, Reindorf CA, Scott RB, Gillette P, Vera JC, Levy PS. The acute chest syndrome in sickle cell disease: incidence and risk factors: the Cooperative Study of Sickle Cell Disease. Blood 1994;84:643–649. 3. Vichinsky EP, Neumayr LD, Earles AN, Williams R, Lennette ET, Dean D, Nickerson B, Orringer E, McKie V, Bellevue R, et al. Causes and outcomes of the acute chest syndrome in sickle cell disease: National Acute Chest Syndrome Study Group. N Engl J Med 2000;342:1855– 1865. 4. Bellet PS, Kalinyak KA, Shukla R, Gelfand MJ, Rucknagel DL. Incentive spirometry to prevent acute pulmonary complications in sickle cell diseases. N Engl J Med 1995;333:699–703. 5. Weil JV, Castro O, Malik AB, Rodgers G, Bonds DR, Jacobs TP. NHLBI Workshop Summary: pathogenesis of lung disease in sickle hemoglobinopathies. Am Rev Respir Dis 1993;148:249–256. 6. Davies SC, Luce PJ, Win AA, Riordan JF, Brozovic M. Acute chest syndrome in sickle-cell disease. Lancet 1984;1:36–38. 7. Manci E, Haynes JHJ, Voelkel NF. Pulmonary complications in sickle cell disease. In: Embury S, et al., eds. Basic Principles and Clinical Practice. New York: Raven Press; 1994. 8. Stuart MJ, Setty BN. Sickle cell acute chest syndrome: pathogenesis and rationale for treatment. Blood 1999;94:1555–1560. 9. Stuart MJ, Setty BN. Acute chest syndrome of sickle cell disease: new light on an old problem. Curr Opin Hematol 2001;8:111–122. 10. Rucknagel DL. Progress and prospects for the acute chest syndrome of sickle cell anemia. J Pediatr 2001;138:160–162.
694
AMERICAN JOURNAL OF RESPIRATORY AND CRITICAL CARE MEDICINE VOL 169 2004
11. Wun T. The Role of Inflammation and leukocytes in the pathogenesis of sickle cell disease: haemoglobinopathy. Hematology 2001;5:403–412. 12. Platt OS. The acute chest syndrome of sickle cell disease. N Engl J Med 2000;342:1904–1907. 13. Chies JA, Nardi NB. Sickle cell disease: a chronic inflammatory condition. Med Hypotheses 2001;57:46–50. 14. Francis RB Jr, Haywood LJ. Elevated immunoreactive tumor necrosis factor and interleukin-1 in sickle cell disease. J Natl Med Assoc 1992; 84:611–615. 15. Malave I, Perdomo Y, Escalona E, Rodriguez E, Anchustegui M, Malave H, Arends T. Levels of tumor necrosis factor alpha/cachectin (TNF alpha) in sera from patients with sickle cell disease. Acta Haematol 1993;90:172–176. 16. Kuvibidila S, Gardner R, Ode D, Yu L, Lane G, Warrier RP. Tumor necrosis factor alpha in children with sickle cell disease in stable condition. J Natl Med Assoc 1997;89:609–615. 17. Bourantas KL, Dalekos GN, Makis A, Chaidos A, Tsiara S, Mavridis A. Acute phase proteins and interleukins in steady state sickle cell disease. Eur J Haematol 1998;61:49–54. 18. Wun T, Paglieroni T, Field CL, Welborn J, Cheung A, Walker NJ, Tablin F. Platelet-erythrocyte adhesion in sickle cell disease. J Investig Med 1999;47:121–127. 19. Croizat H. Circulating cytokines in sickle cell patients during steady state. Br J Haematol 1994;87:592–597. 20. Taylor SC, Shacks SJ, Qu Z, Wiley P. Type 2 cytokine serum levels in healthy sickle cell disease patients. J Natl Med Assoc 1997;89:753–757. 21. Duits AJ, Schnog JB, Lard LR, Saleh AW, Rojer RA. Elevated IL-8 levels during sickle cell crisis. Eur J Haematol 1998;61:302–305. 22. Abboud MR, Taylor EC, Habib D, Dantzler-Johnson T, Jackson SM, Xu F, Laver J, Ballas SK. Elevated serum and bronchoalveolar lavage fluid levels of interleukin 8 and granulocyte colony-stimulating factor associated with the acute chest syndrome in patients with sickle cell disease. Br J Haematol 2000;111:482–490. 23. Little F, Hess B, Moore RB. Changes in inflammatory cytokines and alpha1-acid glycoprotein in sickle cell subjects with acute chest syndrome [abstract]. Blood 1996;88:12a. 24. Naumov A, Hall GA. Serum cytokine and circulating adhesion molecule levels in patients with sickle cell disease: relationships to markers of disease severity and therapy with HbF-inducing agents [abstract]. Blood 1996;88:10a. 25. Fadlon E, Vordermeier S, Pearson TC, Mire-Sluis AR, Dumonde DC, Phillips J, Fishlock K, Brown KA. Blood polymorphonuclear leukocytes from the majority of sickle cell patients in the crisis phase of the disease show enhanced adhesion to vascular endothelium and increased expression of CD64. Blood 1998;91:266–274. 26. Graido-Gonzalez E, Doherty JC, Bergreen EW, Organ G, Telfer M, McMillen MA. Plasma endothelin-1, cytokine, and prostaglandin E2 levels in sickle cell disease and acute vaso-occlusive sickle crisis. Blood 1998;92:2551–2555. 27. Duits AJ, Pieters RC, Saleh AW, van Rosmalen E, Katerberg H, Berend K, Rojer RA. Enhanced levels of soluble VCAM-1 in sickle cell patients and their specific increment during vasoocclusive crisis. Clin Immunol Immunopathol 1996;81:96–98. 28. Saleh AW, Duits AJ, Gerbers A, de Vries C, Hillen HF. Cytokines and soluble adhesion molecules in sickle cell anemia patients during hydroxyurea therapy. Acta Haematol 1998;100:26–31. 29. Miller ST, Sleeper LA, Pegelow CH, Enos LE, Wang WC, Weiner SJ, Wethers DL, Smith J, Kinney TR. Prediction of adverse outcomes in children with sickle cell disease. N Engl J Med 2000;342:83–89. 30. Charache S. Mechanism of action of hydroxyurea in the management of sickle cell anemia in adults. Semin Hematol 1997;34(3 Suppl 3):15–21. 31. Steinberg MH, Lu ZH, Barton FB, Terrin ML, Charache S, Dover GJ. Fetal hemoglobin in sickle cell anemia: determinants of response to hydroxyurea: Multicenter Study of Hydroxyurea. Blood 1997;89:1078– 1088. 32. Hebbel RP. Blockade of adhesion of sickle cells to endothelium by monoclonal antibodies. N Engl J Med 2000;342:1910–1912. 33. Swerlick RA, Eckman JR, Kumar A, Jeitler M, Wick TM. Alpha 4 beta 1-integrin expression on sickle reticulocytes: vascular cell adhesion molecule-1-dependent binding to endothelium. Blood 1993;82:1891– 1899. 34. Gee BE, Platt OS. Sickle reticulocytes adhere to VCAM-1. Blood 1995; 85:268–274. 35. Saleh AW, Hillen HF, Duits AJ. Levels of endothelial, neutrophil and platelet-specific factors in sickle cell anemia patients during hydroxyurea therapy. Acta Haematol 1999;102:31–37. 36. Liu XW, Pierangeli SS, Barker J, Wick TM, Hsu LL. RBC adhesion to
37.
38.
39.
40.
41.
42.
43.
44.
45.
46. 47.
48. 49.
50.
51.
52.
53.
54. 55.
56.
57.
58.
cremaster endothelium in mice with abnormal hemoglobin is increased by topical endotoxin. Ann N Y Acad Sci 1998;850:391–393. Hsu LL, Liu XW, Pierangeli S, Eckman JR, Jack D, Aguayo SM, Wick TM. Microcirculatory effects of blocking cell adhesion molecules in transgenic sickle mice [abstract]. Blood 2000;96:528a. Lutty GA, Taomoto M, Cao J, McLeod DS, Vanderslice P, McIntyre BW, Fabry ME, Nagel RL. Inhibition of TNF-alpha-induced sickle RBC retention in retina by a VLA-4 antagonist. Invest Ophthalmol Vis Sci 2001;42:1349–1355. Hsu LL, Aguayo SM, Eckman JR, Jack D, Roser S, Holtzclaw JD, Moore M, Roman JA. A survey of cytokine expression in the lungs of transgenic sickle mice: is there a pro-inflammatory state in sickle cell disease [abstract]? Am J Respir Crit Care Med 2002;165:C47. Holtzclaw JD, Jack D, Aguayo SM, Roman J, Hsu LL. Cytokine and adhesion molecule expression in transgenic sickle mice lung tissue: evidence supporting a pulmonary proinflammatory state in sickle cell disease [abstract]. Blood 2002;100:449a. Paszty C, Brion CM, Manci E, Witkowska HE, Stevens ME, Mohandas N, Rubin EM. Transgenic knockout mice with exclusively human sickle hemoglobin and sickle cell disease. Science 1997;278:876–878. Hamelmann E, Schwarze J, Takeda K, Oshiba A, Larsen GL, Irvin CG, Gelfand EW. Noninvasive measurement of airway responsiveness in allergic mice using barometric plethysmography. Am J Respir Crit Care Med 1997;156:766–775. Hamada K, Goldsmith CA, Goldman A, Kobzik L. Resistance of very young mice to inhaled allergen sensitization is overcome by coexposure to an air-pollutant aerosol. Am J Respir Crit Care Med 2000;161:1285– 1293. Shore SA, Schwartzman IN, Le Blanc B, Murthy GG, Doerschuk CM. Tumor necrosis factor receptor 2 contributes to ozone-induced airway hyperresponsiveness in mice. Am J Respir Crit Care Med 2001;164: 602–607. Brown LA, Harris FL, Guidot DM. Chronic ethanol ingestion potentiates TNF-alpha-mediated oxidative stress and apoptosis in rat type II cells. Am J Physiol Lung Cell Mol Physiol 2001;281:L377–L386. Affymetrix. GeneChip eukaryotic expression analysis technical manual. Santa Clara, CA: Affymetrix Incorporated; 2001. Vichinsky EP, Styles LA, Colangelo LH, Wright EC, Castro O, Nickerson B. Acute chest syndrome in sickle cell disease: clinical presentation and course: cooperative study of sickle cell disease. Blood 1997;89: 1787–1792. Altman DG. Practical statistics for medical research. New York: Chapman & Hall; 1991. p. 532–534. de Rochemonteix-Galve B, Marchat-Amoruso B, Dayer JM, Rylander R. Tumor necrosis factor and interleukin-1 activities in free lung cells after single and repeated inhalation of bacterial endotoxin. Infect Immun 1991;59:3646–3650. Kean LS, Manci EA, Perry J, Balkan C, Coley S, Holtzclaw D, Adams AB, Larsen CP, Hsu LL, Archer DR. Chimerism and cure: hematologic and pathologic correction of murine sickle cell disease. Blood 2003;102:4582–4593. Makis AC, Hatzimichael EC, Mavridis A, Bourantas KL. Alpha-2-macroglobulin and interleukin-6 levels in steady-state sickle cell disease patients. Acta Haematol 2000;104:164–168. Benjamini Y, Hochberg Y. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J Royal Stat Soc 1995;57: 289–300. Nickerson B, Browning I, Vichinsky E. Cooperative study of sickle cell disease, (CSSCD): pulmonary function in children with sickle cell disease [abstract]. Am J Respir Crit Care Med 1994;149:A374. Nagel RL. A knockout of a transgenic mouse–animal models of sickle cell anemia. N Engl J Med 1998;339:194–195. de Jong K, Emerson RK, Butler J, Bastacky J, Mohandas N, Kuypers FA. Short survival of phosphatidylserine-exposing red blood cells in murine sickle cell anemia. Blood 2001;98:1577–1584. Belcher JD, Bryant CJ, Nguyen J, Bowlin PR, Kielbik MC, Bischof JC, Hebbel RP, Vercellotti GM. Transgenic sickle mice have vascular inflammation. Blood 2003;101:3953–3959. Gladwin MT, Shelhamer JH, Ognibene FP, Pease-Fye ME, Nichols JS, Link B, Patel DB, Jankowski MA, Pannell LK, Schechter AN, et al. Nitric oxide donor properties of hydroxyurea in patients with sickle cell disease. Br J Haematol 2002;116:436–444. Reiter CD, Wang X, Tanus-Santos JE, Hogg N, Cannon 3rd RO, Schechter AN, Gladwin MT. Cell-free hemoglobin limits nitric oxide bioavailability in sickle-cell disease. Nat Med 2002;8:1383–1389.
Holtzclaw, Jack, Aguayo, et al.: A Proinflammatory State in Sickle Mice
695
59. Leong MA, Dampier C, Varlotta L, Allen JL. Airway hyperreactivity in children with sickle cell disease. J Pediatr 1997;131:278–283. 60. Elion J, Belloy M, Grossi Y. Airway obstruction in children with sickle cell disease [abstract]. In: National Sickle Cell Program. Bethesda, MD: National Institutes of Health; 1997. Abstract 352. 61. Lundblad LK, Irvin CG, Adler A, Bates JH. A reevaluation of the validity of unrestrained plethysmography in mice. J Appl Physiol 2002; 93:1198–1207. 62. Mitzner W, Tankersley C. Interpreting Penh in mice. J Appl Physiol 2003; 94:828–832. 63. DeLorme MP, Moss OR. Pulmonary function assessment by whole-body plethysmography in restrained versus unrestrained mice. J Pharmacol Toxicol Methods 2002;47:1–10. 64. Henderson JN, Moinuddin A, DeBaun MR. Asthma increases the risk
acute chest syndrome in children with sickle cell disease [abstract]. Blood 2002;100:453a. 65. Aslan M, Ryan TM, Adler B, Townes TM, Parks DA, Thompson JA, Tousson A, Gladwin MT, Patel RP, Tarpey MM, Batinic-Haberle I, White CR, Freeman BA. Oxygen radical inhibition of nitric oxidedependent vascular function in sickle cell disease. Proc Natl Acad Sci USA 2001;98:15215–15220. 66. Turhan A, Weiss LA, Mohandas N, Coller BS, Frenette PS. Primary role for adherent leukocytes in sickle cell vascular occlusion: a new paradigm. Proc Natl Acad Sci USA 2002;99:3047–3051. 67. Holguin F, Moss I, Brown LA, Guidot DM. Chronic ethanol ingestion impairs alveolar type II cell glutathione homeostasis and function and predisposes to endotoxin-mediated acute edematous lung injury in rats. J Clin Invest 1998;101:761–768.