INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: exhaled nitric oxide; sickle cell disease; acute chest syndrome; serum nitrates; human; genetics; NOS I

Acute chest syndrome (ACS) is the second most common cause of hospitalization in subjects with sickle cell disease (SCD) and accounts for 25% of premature deaths (1). Survivors of ACS may be left with incapacitating pulmonary sequelae, and in one series, 43% of the subjects who suffered one episode had at least one other episode of ACS (2). As such, ACS represents a threat with a significant risk of death and disability in subjects with SCD.

Nitric oxide (NO) metabolic derangements are a possible contributing factor in the pathogenesis of ACS (3). Compared with healthy control subjects (HC), subjects with SCD are hypermetabolic (4), have reduced serum concentrations of arginine (5), have increased plasma concentrations (6) of nitrates, and have increased excretion of NO metabolites (7). Recent anecdotal reports have also suggested that inhaled nitric oxide decreases the hypoxemia of ACS (8, 9). These findings, along with other evidence implicating NO abnormalities in the pathogenesis of SCD (6, 7, 10) suggested to us that low levels of FE_NO may contribute to ACS. In the present study, we tested the hypothesis that FE_NO levels are altered in subjects with SCD who have had at least one episode of ACS. We also tested the hypothesis that the number of AAT repeats in intron 20 of the NOS I gene correlates with FE_NO levels. This latter hypothesis is supported by a recent study that reported FE_NO in patients with asthma was inversely related to the number of AAT repeats in intron 20 (11). Herein we report that FE_NO levels in patients who have had ACS are approximately one-third those observed in HC and in patients who have not had ACS, and that FE_NO is inversely related to the allelic sum of AAT repeats.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients with SCD and subjects who served as HC were recruited from the Hematology Clinic and our local community. All study participants were African-Americans between 6 and 18 yr of age, were nonsmokers, and were matched for age and sex in each group. The HC group had no history of any acute or chronic illness at the time of the study. SCD patients with ACS⁺ must have had at least one episode of ACS necessitating hospital admission at least 3 mo before inclusion in the study. ACS was defined by the presence of rapidly progressive multilobar infiltrates, cough, hypoxemia, and dyspnea in children with SCD. Exclusion criteria for subjects included the presence of chronic heart or lung diseases unrelated to SCD, any acute illness, any history of drug intake within the past 3 d, history of cigarette smoking, allergy, or asthma. Sickle cell subjects meeting study criteria underwent a screening visit that included a full medical history and physical examination. All study participants gave written informed consent and the Institutional Review Committee for Research on Human Subjects approved the study.

FE_NO Measurements

FE_NO measurements in parts per billion (ppb) were done online using Sievers 280 NOA analyzer (Sievers, Boulder, CO) as previously described (12, 13). All measurements were done at the same time of the day (0800 h). Subjects were asked to rinse their mouths and gargle with water prior to sampling in order to avoid contamination with the presence of nitrites and nitrates in the saliva. The measurement circuit consists of a mouthpiece connected to a two-way valve through which the seated patient inhales ambient air that has passed through a scrubber (Window-Cator Cannister Scrubber, Type N-B100; Mine Alliances Company, Pittsburgh, PA). The subjects were required to insert the mouthpiece producing a good seal and then inhale via the mouth to total lung capacity. They were then asked to exhale to a constant pressure of 20 mm Hg, which was displayed on a monitor incentive screen. A resistor placed online in series with exhalation circuit resulted in an expiratory flow rate of 50 ml/s. The end point of the measurement was defined as a plateau for at least 4 s. FE_NO measurements were repeated three times to yield values that were within 5% of each other. Each trial was performed after 15 s of rest between trials.

Pulmonary Evaluation

Each subject underwent pulmonary function testing using a Fleish Pneumotach MultiSpiro PC (Medical Equipment Design, Laguna Hills, CA). This was done in order to determine FEV₁ (forced expiratory volume in 1 s) and forced vital capacity (FVC). Three sets of readings were obtained in the standing position. The best effort, defined by the highest FEV₁, was recorded for each patient without the examiner's knowledge of the patient's medical history or group assignment. The best results were compared with normals for age and expressed as percentage predicted by Crapo Standards (14).

Plasma Nitrite and Nitrate Measurements (NO_x)

Plasma nitrite and nitrate concentrations were quantified as described in the Nitric Oxide Analyzer NOA 280 Operation and Service Manual (Sievers Instrument, Inc, Boulder, CO). Blood was collected in purple top (EDTA) vacutainer tubes then separated by centrifuging for 20 min at 3,000 rpm under refrigeration (4° C). The plasma was collected in a labeled cryo-tube and frozen at 80° C until the day of analysis. On the day of analysis, the samples were allowed to thaw at room temperature and deproteinized by treating 0.5 ml of plasma with 1.0 ml of cold ethanol. Samples were then incubated at 4° C for 30 min followed by centrifugation at 14,000 rpm for 5 min. The supernatant was collected for analysis.

The NOA 280 was set up for liquid measurements using the Radical Purger, which quantifies nitrates and nitrites and expresses the total amount as nitrate. Five milliliters of a saturated solution of 0.56 M VCl₃/0.1 M HCl and 100 µl of antifoaming reagent were added to the purge vessel of the Radical Purger. The system was allowed to equilibrate and a baseline was established using helium as the flow gas. A standard curve was prepared from sodium nitrate (NaNO₃) and nitrate-free deionized water and used to calculate the concentration of nitrates and nitrites in the plasma samples in triplicate.

Arginine and Citrulline Levels

Arginine and citrulline concentrations were measured by an ion- exchange chromatography method using a Beckman 6300 Amino Acid Analyzer (Beckman Coulter, Fullerton, CA) in our College of American Pathologists accredited Biochemical Genetics Laboratory.

Genotyping (NOS I Intron 20 AAT Repeats)

Genomic DNA was extracted from 0.3 ml of whole blood with a DNA isolation kit (Gentra Systems, Inc., Minneapolis, MN) following the manufacturer-suggested protocol. The final DNA was dissolved in 10 mM Tris-HCl, pH 7.5, and diluted to a concentration of 100 ng/µl. Oligonucleotides were synthesized by Operon Technologies, Inc. (Alameda, CA). The polymerase chain reaction (PCR) cocktail consisted of 17 µl of Platinum PCR SuperMix (GIBCO-BRL Life Technologies, Gaithersburg, MD), 1 µl of 10 µM forward primer (5'-CTG GGGGCAATGGTGTGT-3'), 1 µl of 10 µM reverse primer (5'-CAT CATGGAAACTGGCAAACATGAAAATCAAGG-3'), and 1 µl of genomic DNA (100 ng/µl).

The PCR consists of 35 cycles of denaturation at 97° C for 20 s, annealing at 59° C for 30 s, and extension at 72° C for 45 s (GeneAmp PCR System 9600, Perkin Elmer). The sizes of the amplicons were determined by loading 5 µl of the PCR products onto a 10% nondenaturing polyacryamide gel with a 10-bp DNA ladder (GIBCO-BRL Life Technologies) as the molecular size standard.

Statistical Analysis

Data are presented as mean ± standard deviation. A one-way ANOVA with least significant differences test was used with p < 0.05 between groups considered significant. The correlation coefficient of the relationship between FE_NO and the sum of AAT repeats on intron 20 of the NOS I was determined by standard means. The hypothesis that the slope of the line differed from zero was tested as previously described (15).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Eight ACS⁺ children had one episode of ACS, while three had two episodes each. In all cases the most recent episode was greater than 6 mo ago. Table 1 and Figure 1 summarize our data. There was no significant difference between groups with respect to age. Mean ± SD FE_NO (ppb) was approximately one-third lower in ACS⁺ children as compared with ACS children and HC (Table 1 and Figure 1). Significant differences in FE_NO values were found between ACS⁺ and ACS (p = 0.002) and ACS ⁺ and HC (p = 0.0001). There were no differences between ACS and HC.

View larger version (24K): [in this window] [in a new window]   Figure 1.   FENO levels in ACS+, ACS, and HC. **p < 0.001 versus HC; *p < 0.05 versus ACS+.

Arginine, NO, NO_x, and citrulline are substrate and products of the NO metabolic pathway. We measured NO_x in 11 of 13 ACS⁺ subjects, 6 of 7 ACS subjects, and 3 of 6 HC who consented to have blood drawn. Because of insufficient sample volumes, only 8 of 13 ACS⁺ subjects and 4 of 7 ACS subjects had citrulline and arginine levels measured. The results in Table 1 show that there were no significant differences between ACS⁺ and ACS and HC with respect to NO_x levels and that there was no difference between ACS⁺ and ACS groups with respect to arginine or citrulline levels.

Pulmonary function tests were done in order to assess lung function. There were significant differences between ACS⁺ and HC with respect to both FEV₁ (p = 0.014) and FVC (p = 0.023) (see Table 1). There were no significant differences between ACS and HC as well as between ACS⁺ and ACS groups.

To ascertain whether low FE_NO associates with a genetic polymorphism in the NOS I gene (11), we explored the relationship between FE_NO and the number of AAT repeats in intron 20 of the NOS I gene in our small patient population. The results shown in Figure 2 reveal that there is a significant inverse relationship between FE_NO and the sum of AAT repeats of both NOS I alleles. The correlation coefficient (r) was 0.62 (p < 0.005).

View larger version (14K): [in this window] [in a new window]   Figure 2.   Correlation between FENO and number of AAT repeats on intron 20 of NOS I genes in study subjects. Open circles indicate ACS+, closed circles indicate ACS, and closed triangles indicate HC.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

To our knowledge, this is the first report that establishes a link between reduced FE_NO, genetic variation, and the risk of subjects with SCD to develop ACS. We have shown that subjects with SCD who are ACS⁺ have markedly lower FE_NO levels than HC or ACS subjects, and that high numbers of AAT repeats in intron 20 of the NOS I gene are inversely correlated with FE_NO levels. SCD is accompanied by a hypermetabolic state during which arginine stores are depleted as a consequence of increased demands for NO (4, 6, 16). It is known that factors can trigger ACS in 10% to 20% of subjects with SCD, suggesting a genetic predisposition. Our data suggest that the number of AAT repeats is one genetic variation that is associated with the risk of patients with SCD to develop ACS. We speculate that subjects with SCD develop ACS when exposed to an environmental trigger if they have a high number of AAT repeats in intron 20 of the NOS I gene because subjects with this polymorphism may not be able to produce sufficient NO to meet the demand. Therefore this genetic variation may be a true risk factor for SCD patients to develop ACS.

NO plays a key role in the regulation of pulmonary vascular tone and blood flow (17, 18). The vasodilating properties of NO reduces vascular resistance in the pulmonary circulation by inhibiting smooth muscle contraction (19, 20). Moreover, NO also reduces platelet aggregation. NO is therefore a very important molecule in preventing sludging in pulmonary vessels that may occur as a result of pulmonary vascular compromise and platelet aggregation, which are inciting factors for ACS (3). Several respiratory diseases have been associated with low levels of NO and FE_NO with clinical similarities to the sequelae of ACS such as cor pulmonale (21) and pulmonary hypertension (22, 23). Our data indicate that ACS is another disease that is characterized by low FE_NO levels.

The decreased levels of FE_NO cannot be explained by acute hypoxia. Dweik and colleagues found that FE_NO levels correlated directly with inspired oxygen (FI_O2) in the hypoxic range (0.05 to 0.15 FI_O2) (24). This mechanism is unlikely to be responsible for our findings because our subjects were breathing room air and had normal exercise tolerance, respiratory rate, and work of breathing. Although small differences in PO₂ may dramatically affect NO delivery by hemoglobin, differences in PO₂ will affect FE_NO levels only if hypoxia is extreme (24). Therefore, other mechanisms to explain the low FE_NO values must be considered.

To unravel the most likely explanation requires a thorough understanding of airway NO origin. There is no consensus as to which cell types and NOS isoforms contribute to FE_NO production. Likewise, the physical location of origin of FE_NO is controversial. A study with perfused pig lungs suggests that FE_NO originates in the lung itself and is not carried from other tissue beds via the bloodstream (25). Moreover, this study reported that a portion of the FE_NO is derived from the pulmonary endothelial cells (25) and may be derived from NOS III expressed in the pulmonary endothelium (26). Other studies have not supported the theory that pulmonary vascular NO production or delivery is mirrored in the FE_NO (27, 28). Evidence strongly suggests that FE_NO represents only airway epithelial cell production. Indeed, mathematical modeling and sequential bronchoscopic airway sampling have indicated that the majority of pulmonary NO emanates from larger central airways as opposed to terminal gas exchanging units (24, 29, 30) where hemoglobin may avidly trap NO.

There are several possible explanations other than a polymorphic variant of NOS for the low FE_NO levels in subjects with previous episodes of ACS. First, ACS may have altered the lung architecture by deposition of iron with widening and or inflammation in the interstitium. Second, low FE_NO may reflect an abnormality in transfer of NO from deoxyhemoglobin. Third, substrate limitations and alternate metabolic fates of airway NO are possibilities.

The possibility that the difference in FE_NO values may be due to fibrotic or inflammatory changes in lung tissues is reasonable and warrants consideration. In our study, the ACS⁺ group had a significantly lower FEV₁ and FVC as compared with HC but was similar to the ACS group. However, inflammatory changes are unlikely to cause low FE_NO because these changes upregulate inducible NOS (NOS II) activity and hence may increase FE_NO as reported in subjects with asthma and lung transplants (31, 32). Pulmonary fibrosis may cause restrictive lung disease in subjects with SCD who have suffered repeated episodes of lung injury (33, 34), and this is possible in our ACS⁺ group. Because the ACS⁺ and ACS groups had significantly different FE_NO values but similar FEV₁ and FVC, altered lung architecture does not offer a satisfactory explanation for the low FE_NO observed. Moreover, iron deposition and widening of the interstitium are an unlikely explanation because FE_NO originates mostly from the large central airways (31, 32).

That low FE_NO is a marker rather than a result of ACS is an intriguing possibility. If this is the case, values may reflect an abnormality in transfer of NO from deoxyhemoglobin. This abnormality may be in the red cell membrane anion exchanger. Indeed, Pawloski and colleagues demonstrated the presence of a membrane-bound protein in erythrocytes that is suggested to be responsible for translocation of NO bioactivity from the red cell to the tissues concomitantly with oxygen unloading (35). This occurs at the cytosol-membrane interface of the erythrocyte in the presence of the anion exchanger AE1 (35). It is conceivable that abnormalities of cell-bound proteins responsible for the movement of NO in the endothelium, erythrocyte, or airway epithelial cells are present in ACS⁺ subjects. This is an interesting possibility that deserves consideration.

Substrate limitation is also unlikely to explain differences in FE_NO because arginine levels were similar in the ACS⁺ and ACS groups. Morris and colleagues reported substrate depletion during acute crises in children with SCD (6). Stuart and Setty also reported that plasma nitric oxide metabolites were decreased during an attack of ACS (36). Their findings differ from ours, which show similar NO_x levels between ACS and ACS⁺ subjects. This apparent discrepancy may be explained by different experimental conditions. Our sampling was done during healthy periods (at least 3 mo after an episode of ACS), whereas the samplings of both Morris and Stuart were done during acute illness. It is possible that if their sampling was extended, levels of arginine may be higher.

Another possible explanation for low FE_NO levels in ACS⁺ children may be the rapid conversion to S-nitrosothiols. Nitric oxide in the biological systems is quickly metabolized to other compounds (37, 38). In oxygenated environments, potential fates for NO include conversion to nitrite (NO₂) and nitrate (NO₃), reaction with O₂ to form peroxynitrite anion (OONO-) (37), reaction with transition metal ions that are components of heme and nonheme proteins, and reactions with nitrogen and sulfur-containing compounds resulting in nitrosoamines (RNH-NO) and S-nitrosothiols (RS-NO) (37). It is postulated that the S-nitrosothiol compounds subserve NO bioactivity and are instrumental in matching ventilation and perfusion (38). Finally, it has been convincingly demonstrated that NO in the presence of hemoglobin reacts to form S-nitrosohemoglobin (SNO-Hb). SNO-Hb is believed to represent a vast biological sink for NO and is actually the bioactive form of NO that is transported to tissues (39, 40). We did not measure nitrosothiols in our study and hence cannot discount this explanation for low FE_NO levels.

Lastly, if we accept the idea that ACS subjects have an FE_NO value that is chronically lower than normal or SCD subjects, then it is reasonable to hypothesize a genetic basis. Thus, our data showing the association of the sum of the number of AAT repeats in intron 20 of the two NOS I alleles are both consistent with this hypothesis and other published work (11). Whether the increased number of AAT repeats results in less or unstable NOS I mRNA and subsequently lower protein production is unknown. It is possible that this polymorphism is in linkage disequilibria with an SNP in NOS I or an SNP in a nearby gene whose product impacts on FE_NO production. These questions must be addressed experimentally. Second, it is important to remember that this variant of NOS I is clinically important only as a modifier of SCD in the ^s genetic background. Other genetic loci that operate on NO metabolism may in turn modulate this genotype. Conversely, variants in genes involved in other biochemical pathways also may be relevant in precipitating ACS. Most importantly, however, should these hypotheses be upheld in subsequent studies, it would have diagnostic value in risk assessment and it would lead to the design and testing of nutritional and pharmacogenomic strategies to increase NO production.

In summary, we have shown that FE_NO levels are low in ACS⁺ subjects as compared with ACS subjects and HC. The exact reason for this finding is unclear but does not appear to be related to the degree of lung injury, abnormality in transfer of NO from deoxyhemoglobin, or substrate deficiency. The differences may be due alternative metabolic pathways for the breakdown of NO, but our study design did not allow us to draw conclusions on this possibility. Our data suggest, however, that there may be a genetic basis to our findings as FE_NO levels were associated with an NOS I genetic variant. Whether low FE_NO or genetic analysis can prospectively identify those prone to ACS before the first episode, and whether therapy to increase NO may prevent further episodes of ACS, are unknown, but are intriguing possibilities.