Received Date: September 04, 2012; Accepted Date: November 17, 2012; Published Date: November 23, 2012
Citation: Krol M, Balcerowski A, Luczynska M, Szkudlarek U, Nowak D (2012) Increased Exhaled Hydrogen Peroxide in Human Immunodeficiency Virus-Infected Patients without Clinical Signs and Symptoms of Opportunistic Lung Disease. J AIDS Clinic Res 3:183. doi:10.4172/2155-6113.1000183
Copyright: © 2012 Krol M, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Visit for more related articles at Journal of AIDS & Clinical Research
Background: HIV-infected subjects present with decreased antioxidant defense and increased activation of inflammatory cells which may lead to overproduction of oxidants. This study determined whether HIV-infected patients without clinical signs and symptoms of opportunistic lung disease (OLD-negative) exhaled more H2O2 than healthy controls and whether there was association between the exhalation of H2O2 and whole blood chemiluminescence (CL) and clinical variables. Methods: A cross-sectional study was conducted. H2O2 in exhaled breath condensate and CL, resting and agonist-induced with N-formyl-methionyl-leucyl-phenylalanine (fMLP) were measured in 36 OLD-negative patients and 14 healthy controls. Univariate linear regression was used to summarize the average relationship and quantile regression analyzed the relationship at different points of the exhaled H2O2 distribution. Multivariate analyses were carried out using multiple linear regressions. Results: The fold increase of the geometric mean exhaled H2O2 against healthy controls was 3.76-times higher in OLD-negative patients than in controls (95% CI: 2.65-5.33, p<0.001), whereas that of either resting or fMLPinduced CL was 1.46 or 1.63, respectively (95%: 1.17-1.83 and 1.27-2.08, p<0.01). Exhaled H2O2 was not associated with CL, either resting or fMLP-induced. Linear regression detected positive relationship between the exhalation of H2O2 and viral load (R-squared 0.23, p<0.05). The effects of viral load were best revealed at a higher exhalation of H2O2 (quantiles 0.6 and 0.7; both Pseudo R-squared 0.21, p<0.05). In a multivariate model, the main independent contributors to the exhalation of H2O2 were viral load and highly active antiretroviral therapy (HAART), which together accounted for 35% of the variance in exhaled H2O2. If the analysis was limited exclusively to HAART-treated, a better model fit was obtained (R-squared 0.79), confirming that viral load is the main contributor to the exhaled H2O2. Conclusion: Inordinate increase in exhaled H2O2 may reflect airway oxidative stress in HIV-1 infection which may be related to viral load.
Exhaled hydrogen peroxide; HIV-1; Oxidative stress; Whole blood chemiluminescence
Since oxidative stress is implicated in both human immunodeficiency virus, type 1 (HIV-1) expression [1,2] and the pathogenesis of AIDS , this pro-oxidant and antioxidant imbalance has been widely described among HIV-infected patients [4-7]. Overproduction of Reactive Oxygen Species (ROS) is thought to be the result of various cell activation and altered redox status, mostly phagocytic , inappropriately compensated by antioxidants or antioxidant enzymes  and is also augmented by highly active antiretroviral therapy (HAART) [7,9]. Exhaled H2O2 belongs to noninvasive markers of ROS production in the airways [10,11]. Since H2O2 is a volatile compound, it is easily determined by breath analysis. Exhalation of H2O2 is elevated in respiratory tract disorders accompanied by an influx of activated inflammatory cells e.g. bronchial asthma [12,13], COPD [14,15] and pneumonia . To date, this has not been described in HIV-infected patients. They typically show a marked decrease in the concentration of reduced glutathione (GSH) [17,18] and their alveolar macrophages spontaneously produce more superoxide anion  which undergo dismutation to H2O2. Moreover, progression of HIV-1 infection results in a decreased erythrocyte glutathione peroxidase (GSH-Px) activity and suppression of GSH plasma levels . Since the GSH-GSH-Px system is involved in the decomposition of H2O2 one may suspect that HIV-infected subjects have increased H2O2 levels in the airways resulting in increased exhalation of H2O2. Nevertheless, any concomitant Opportunistic Lung Disease (OLD) can influence such measurements .
The mechanisms underlying changes in exhalation of H2O2 are likely multifactorial in nature. Exhaled H2O2 represents a pool of ROS derived from the NADPH-oxidase system and the mitochondrial chain that avoids decomposition by antioxidant systems, subsequently diffusing into the airway surface which is then blown out as vapor and or aerosolized respiratory fluid droplets released from the respiratory epithelial lining fluid finally collected as exhaled breathe condensate (EBC) . Therefore,changes in antioxidant defense in the airway or modifications in the number and activity of pulmonary inflammatory cells may alter the concentration of H2O2 in EBC. In addition, possibly there is a relationship between exhaled H2O2 and the blood oxidant and antioxidant status as measured by luminol enhanced whole blood chemiluminescence. Szkudlarek et al.  found an association between increased exhalation of H2O2 and a greater light emission of whole blood in a cross-sectional study of 41 healthy subjects. This associative finding in HIV-infected patients would provide evidence of a shared mechanism in the oxidative response in the blood and the airways. Furthermore, since nucleoside reverse transcriptase inhibitors (NRTIs) have been shown to increase intracellular H2O2, treatment with NRTIs may also contribute to exhalation of H2O2 .
We conducted a cross-sectional study of HIV-infected men and women without clinical signs and symptoms of concomitant OLD (OLD-negative), accompanied by clinical signs and symptoms of concomitant respiratory tract infection (RTI) without a definite diagnosis of OLD (RTI-positive) and healthy control subjects to determine: 1) the respective amount of exhaled H2O2, 2) luminol enhanced whole blood chemiluminescence, either resting or agonist-induced, 3) whether exhaled H2O2 is associated with whole blood chemiluminescence and finally 4) whether exhaled H2O2 is associated with selected clinical variables, including HIV-infection duration, detectable viral load, treatment with HAART and HAART duration or a history of AIDS.
A total of 310 HIV-infected patients from the Acquired Immune Disorders Outpatient Clinic in Lodz, Poland were screened. The study included 36 OLD-negative subjects, 28 RTI-positive patients and 14 healthy controls. Each patient enrolled had to meet an inclusion criteria: age ≥ 18 and ≤ 60 years, HIV-1 seropositivity, a chest X-ray performed within 10 days prior to the enrollment and a written informed consent. The exclusion criteria included: any episode of alcohol or illicit drug abuse within the last 2 to 6 months before the study, respectively, any history of bronchial asthma, COPD, bronchiectasis, cystic fibrosis, tuberculosis, malignancies, renal or liver damage, pregnancy or breast feeding, pharmacological treatment other than HAART within the last 2 months and regular ingestion of supplements with known antioxidant properties (e.g. vitamin C) within the last 3 weeks before the study. At admission, patients were screened for clinical signs and symptoms of OLD, including: cough, purulent sputum, dyspnea, chest X-ray findings and a history of past respiratory tract infections within the last 3 months. Ongoing HAART was administered to 23 of OLD-negative patients and 7 of RTI-positive subjects. Patients were treated with either lopinavir 600 mg+ritonavir 150 mg twice-daily associated with stavudine 40 mg twice-daily and didanosine 400 mg once daily or lamivudine 150 mg+zidovudine 300 mg twice-daily in combination with efavirenz 600 mg once daily. Except HIV- 1 seropositivity, healthy control subjects also had to meet all inclusion and exclusion criteria. The study was approved by the Ethics Committee of the Medical University of Lodz, Poland (RNN/216/03/KE).
All subjects enrolled (HIV-1 infected patients and healthy controls) were asked to come to the laboratory between the hours of 8 am to 10 pm for EBC collection. Subsequently, 9 ml blood samples were drawn into EDTAK3 Vacuette tubes (Greiner Labor Technik, Austria) for whole blood chemiluminescence assay, blood cell count and HIV-1 RNA copy number.
Exhaled breathe condensate sampling
2-3 mL of exhaled breath condensate was sampled during 15 min of spontaneous tidal volume breathing (respiratory rate ranged 14-20 bpm), using EcoScreen-1 (Erich Jaeger GmbH, Hoechberg, Germany), with saliva trap. Subjects wore a noseclip and rinsed their mouth with distilled water just before and after 7 min of collection . Immediately after the procedure, EBC specimens were stored at -80ºC [23,24], no longer than 7 days until H2O2 measurement. No amylase activity was detected in EBC specimens (control of salivary contamination) . Subjects who were current smokers refrained from cigarette smoking 12 hrs preceding EBC collection.
Measurement of H2O2
HVA method was used to assess the concentration of H2O2 in EBC , as previously described [23,26]. The detection limit of the H2O2 assay was 0.05 μmol/L. The intra-assay variability did not exceed 2.5% for the standard 1 μmol/L of the H2O2 solution. The addition of catalase (30 U) to the EBC specimens of HIV-infected patients (n=4) and healthy subjects (n=3), which previously revealed that detectable exhaled H2O2 levels completely abolished HVA oxidation, demonstrates that the H2O2 assay is specific and other reactive compounds or oxygen species did not contribute to H2O2 readings. Individual results were means from duplicate measurements.
Whole blood chemiluminescence assay
The resting and fMLP-induced luminol enhanced whole blood chemiluminescence (CL) were measured as previously described [21,27]. Two CL parameters were assessed: resting CL prior to the addition of fMLP and peak light emission after the addition of an agonist to a final concentration of 20 μmol/L (fMLP-induced peak CL). Resting CL and fMLP-induced peak CL were expressed as mV per 104 phagocytes in the assayed blood sample. Individual results were obtained as a mean from triplicate measurements.
Blood cell count was performed using the 5-DIFF LH 750 Hematology Analyzer (Beckman-Coulter, Inc. USA). Blood CD4 count was determined with anti-CD4 monoclonal antibodies (Becton Dickinson, NJ, USA), following flow cytometry (Beckman Coulter Epics XL, USA). Serum anti- HIV-1 antibodies were detected with enzyme linked immunosorbent assay (Bio-Rad, USA). HIV-1 seropositivity was confirmed by Western blot analysis (Calypte Biomedical, USA). HIV-1 RNA copies (viral load) were determined by COBAS Amplicor HIV-1 monitor test (Roche, Branchburg, NJ, USA), with detection limit of 50 copies/mL and expressed as a common logarithm of RNA copies per mL of plasma.
Statistical analysis was carried out using the Stata 12 (Stata Corp., College Station, TX, USA). Normally distributed continuous variables and variables of log 10-transformed toward normality were compared between groups using one-way ANOVA with post-hoc Bonferroni adjustment and unpaired Student’s t-test for equal variances; non-normally distributed data were compared using the Kruskall-Wallis rank test and the Wilcoxon rank-sum test. Categorical variables were compared between groups using the Pearson’s chi-squared test and the Fisher’s exact test. The one-way analysis of covariance (ANCOVA) was used for comparison of variables adjusted for covariate. Correlations between continuous variables were determined nonparametrically using Spearman’s rho. Univariate linear regression analyses were carried out with nonparametric bootstrap and 10 000 replications (to avoid transformation of the dependent variables where appropriate). Linear regression was used to summarize the average relationship while simultaneous quantile regression (sqreg function) was applied to analyze the relationship at different points of the distribution . Multivariate analyses were performed using multiple linear regression with bootstrap estimates of coefficient standard errors and 10 000 replications. Independent variables were manually implemented based on clinical judgment where appropriate. The stepwise estimation technique was used with p to enter and p to leave both equal to 0.15, and unitless standardized coefficients were presented. Statistical significance was set at p<0.05. No formal adjustments for multiple comparisons were made.
A total of 36 OLD-negative adults and 14 healthy control subjects matched for gender, age and smoking habits were studied. Moreover, a group of 28 RTI-positive patients, of whom 15 had a cough, 19 had purulent sputum, 6 were dyspneic, 11 had the chest X-ray findings and 12 with a history of past RTI within the last 3 months, were included in the analysis (Table 1). EBC of H2O2 as well as resting and fMLP-induced whole blood chemiluminescence measurements was observed in all patients. Viral load was determined in 28 of OLD-negative patients and 10 of RTI-positive subjects. Viraemia measurement was solely dependent on test availability. Table 1 shows the clinical and demographic findings of HIV-infected subjects and healthy controls. Evaluation of white blood cell (WBC) count and polymorphonuclear leukocytes (PMNs) count revealed significantly lower values in OLD-negative patients as compared to healthy controls (Table 1). Moreover, WBC and lymphocyte counts were higher, estimated duration of HIV-infection was shorter and the number of HAART-treated was lower in RTI-positive subjects as compared to OLD-negative patients.
|Variable||(1) Control (n=14)||(2) HIV-infected OLD-negative (n=36)||(3) HIV-infected RTI-positive (n=28)||p-value||post-hoc p-value||Difference||95% CI|
|Male gender, n (%)||8 (57%)||20 (55%)||19 (68%)||0.591|
|Age [yrs], median (95% CI, range)||31(25.6 to 36.4, 20-52)||28.5(25.5 to 31.5, 23-54)||29.5(25.9 to 33.1, 20-56)||0.626|
|Smokers, n (%)||7(50%)||23(64%)||23(82%)||0.091|
|Cough/Sputum/Dyspnea/Chest X-ray/Past RTI, n||0 / 0 / 0 / 0 / 0||0 / 0 / 0 / 0 / 0||15 / 19 / 6 / 11 / 12|
|Hemoglobin [g/dL], mean (95% CI, range)||14.5(13.8 to 15.3, 12.3-16.5)||14.3(13.9 to 14.7, 11.4-16.7)||14.9(14.4 to 15.5, 11.0-17.2)||0.183|
|White blood cells count [×103 cells/µL], geometric mean (95% CI, range)||6.61 (5.82 to 7.50, 5.10-11.0)||4.75 (4.25 to 5.30, 2.01-8.72)||5.83 (5.24 to 6.49, 4.03-11.40)||<0.0013||0.0024 1 vs. 2 0.0204 2 vs. 3||0.72 1.23||(0.58 to 0.89) (1.03 to 1.47)|
|PMNs count [×103 cells/µL], median (95% CI, range)||4.30(3.65 to 4.95, 3.80-8.90)||3.14(2.53 to 3.75, 1.06-7.66)||3.58(2.76 to 4.40, 1.91-8.98)||0.0036||0.0017 1 vs. 2||-1.45||(-2.24 to -0.59)|
|Lymphocyte count [×103 cells/µL], mean (95% CI, range)||1.79(1.55 to 2.04, 1.30-2.30)||1.49(1.30 to 1.68, 0.31-2.46)||1.86(1.60 to 2.12, 0.26-3.30)||0.0393||0.0434 2 vs. 3||0.37||(0.02 to 0.72)|
|CD4 count [cells/µL], mean (95% CI, range)||N/A||361.6(287.1 to 436.0, 6-1063)||391.7(320.8 to 462.6, 90-822)||0.565|
|HIV-infection duration [yrs], median (95% CI, range)||N/A||4.2(1.9 to 6.5, 0.2-12.0)||2.1(0.7 to 3.6, 0.1-9.2)||0.0038||-1.8||(-3.3 to -0.7)|
|Viral load assays, n (%)||N/A||28(78%)||10(36%)||0.0012|
|Detectable viral loada, n(%)||N/A||19(68%)||6(60%)||0.712|
|Viral load [×103 RNA copies/mL], geometric mean (95% CI, range)||N/A||1.08(0.26 to 4.42, 0-935)||1.33(0.07 to 23.71, 0-41.9)||0.885|
|Treatment with HAART, n (%)||N/A||23(64%)||7(25%)||0.0032||-39%|
|Treatment duration [mths], geometric mean (95% CI, range)||N/A||17.4(10.9 to 28.0, 1-84)||7.9(1.7 to 36.0, 1-62.5)||0.155|
|History of AIDSb, n (%)||N/A||14(39%)||5(18%)||0.102|
p-values: 1Pearson’s chi-squared test; 2Fisher’s exact test; 3one-way ANOVA; 4post hoc ANOVA test with Bonferroni adjustment; 5unpaired Student’s t-test (equal variances); 6Kruskal-Wallis rank test; 7Wilcoxon rank-sum test with Bonferroni adjustment; 8Wilcoxon rank-sum test.
HAART, highly active anti-retroviral therapy; PMNs, polymorphonuclear leukocytes.
aDetectable viral load defined as ≥50 copies of HIV-1 RNA/mL. bAIDS defined as a history of CD4 count <200 cells/μL or AIDS-defining illness.
Table 1: Comparison of demographic and clinical variables between healthy controls vs. HIV-infected patients without clinical signs and symptoms of opportunistic lung disease (OLD-negative) or with clinical signs and symptoms of concomitant respiratory tract infection without a definite diagnosis of OLD (RTI-positive).
Exhaled H2O2 and whole blood chemiluminescence
Old-negative vs. RTI-positive subjects: Increased oxidative status defined as elevated exhalation of H2O2 and enhanced whole blood chemiluminescence in comparison to healthy controls was a consistent feature of HIV infection, regardless of concomitant RTI (Table 2, Figures 1A and 1B). The highest significant difference was seen in exhaled H2O2 (Table 2, Figure 1B). Fold increase of the geometric mean exhaled H2O2 against healthy controls was about 2 to 3-times higher than that of either resting or fMLP-induced CL, respectively (post hoc ANOVA against healthy controls; all p<0.05) (Table 2, Figure 1B). Moreover, there were no significant differences between OLD-negative patients and RTI-positive subjects (post hoc ANOVA against OLD-negative; all p>0.05) (Table 1, Figure 1A). Adjustment of CL variables for the significant difference on PMNs count with ANCOVA changed the conclusion concerning the significant difference in resting and fMLP-induced CL (Table 2, Figure 1B). While controlling for the effect of PMNs count in OLD-negative patients, there were no significant differences in the PMNs-adjusted means of CL variables (ANCOVA testing against healthy controls with PMNs as covariate; all p>0.05 in OLD-negative patients). For the exhaled H2O2 the assumptions for ANCOVA with PMNs as a covariate were not met. Regardless of that, the strong increase in exhaled H2O2 was also highly significant in terms of 95% CI (the entire 95% confidence interval for the ratio of geometric means was well over 2-times that of HIV-infected to control geometric means ratio) (Figure 1B).
|Variable||(1) Control (n=14)||(2) HIV-infected OLD-negative (n=36)||(3) HIV-infected RTI-positive (n=28)||p-value||post-hoc p-value||Difference||95% CI|
|Exhaled breath H2O2 [µmol/L], geometric mean (95% CI, range)||0.21 (0.15 to 0.31, 0.05-0.63)||0.80 (0.67 to 0.95, 0.30-2.20)||0.90 (0.75 to 1.09, 0.31-2.20)||<0.0011||<0.0012 1 vs. 2 <0.0012 1 vs. 3 1.002 2 vs. 3||3.76 4.24||(2.53 to 5.58) (2.81 to 6.40)|
|Resting CL [mV/104 cells], geometric mean (95% CI, range)||0.59 (0.50 to 0.70, 0.28-0.84)||0.84 (0.74 to 0.95, 0.48-1.35)||0.71 (0.61 to 0.83, 0.35-1.28)||0.0081||0.0062 1 vs. 2 0.0422 1 vs. 3 1.002 2 vs. 3||1.46 1.36||(1.10 to 1.94) (1.02 to 1.83)|
|fMLP-induced peak CL [mV/104 cells], geometric mean (95% CI, range)||0.92 (0.74 to 1.13, 0.42-2.24)||1.49 (1.30 to 1.71, 0.62-2.97)||1.67 (1.34 to 2.09, 0.63-4.71)||<0.0011||0.0042 1 vs. 2 <0.0012 1 vs. 3 0.982 2 vs. 3||1.63 1.83||(1.20 to 2.19) (1.27 to 2.63)|
|Adjusted variables for significant difference on PMNs count:|
|Exhaled breath H2O2 [µmol/L], geometric mean (95% CI)||N/A||N/A||N/A||N/A3|
|Resting CL [mV/104 cells], geometric mean (95% CI)||0.71(0.63 to 0.81)||0.75(0.70 to 0.81)||0.81(0.75 to 0.88)||0.184|
|peak CL [mV/104 cells], geometric mean (95% CI)||1.03(0.82 to 1.31)||1.40(1.21 to 1.62)||1.71(1.45 to 2.00)||0.0044||0.0035 1 vs. 3||1.65||(1.17 to 2.33)|
p-values: 1one-way ANOVA; 2post hoc ANOVA test with Bonferroni adjustment; 3one-way ANCOVA (assumptions not met); 4one-way ANCOVA (assumptions met); 5post
hoc ANCOVA test with Bonferroni adjustment.
CL, whole blood chemiluminescence; fMLP, N-formyl-methionyl-leucyl-phenylalanine; N/A, not applicable; PMNs, polymorphonuclear leukocytes.
Table 2: Comparison of exhaled H2O2 and whole blood chemiluminescence between healthy controls vs. HIV-infected patients without clinical signs and symptoms of opportunistic lung disease (OLD-negative) or with clinical signs and symptoms of concomitant respiratory tract infection without a definite diagnosis of OLD (RTI-positive).
Figure 1: Exhaled H2O2 as marker of oxidative stress in HIV-infected patients without signs and symptoms of opportunistic lung disease (OLD-negative) and with signs and symptoms of concomitant respiratory tract infection without definite diagnosis of OLD (RTI-positive). (A - top) Dot plot of individual results in healthy controls and HIV-1 infected patients. Box plot shows the geometric mean as a solid line and 95% confidence intervals as a rectangle. (B - bottom) Fold increase of geometric mean exhaled H2O2, resting and fMLP-induced peak whole blood chemiluminescence (CL) in OLD-negative patients against healthy controls. Raw data vs. adjusted for differences in polymorphonuclear leukocytes count (PMNs) among groups with ANCOVA (not applicable for exhaled H2O2 - assumptions not met).
To recapitulate, there exist a highly significant difference in the exhalation of H2O2 between HIV-infected patients and healthy controls; what is more, exhaled H2O2 was the most prominent marker of oxidative stress in HIV-infected individuals, regardless of concomitant RTI.
HAART-naive vs. HAART-treated subjects: A total of 23 OLDnegative patients commenced aggressive antiretroviral treatment regimens with HAART and 13 OLD-negative subjects remained off therapy until clinically indicated (Table 3). Viral load assays confirmed a significant decrease in the number of HIV-1 RNA copies associated with antiretroviral treatment (p<0.05) (Table 3). Along with suppression of viral load no further differences occurred between HAART-naive and HAART-treated arms (post hoc ANOVA against HAART-naive; all p>0.05) (Table 3, Figure 2A). The analysis confirmed an elevated exhalation of H2O2 in comparison to healthy controls, regardless of HAART and no significant difference in resting and fMLP-induced CL after adjustment of CL variables for a difference in PMNs with ANCOVA (Table 3, Figures 2A and 2B).
|Variable||(1) Control (n=14)||(2) HIV-infected OLD-negative HAART-naive (n=13)||(3) HIV-infected OLD-negative HAART-treated (n=23)||p-value||post-hoc p-value||Difference||95% CI|
|Viral load assays, n (%)||N/A||5(39%)||23(100%)||<0.0011|
|Detectable viral loada, n(%)||N/A||5(100%)||14(61%)||0.141|
|Viral load [x103 RNA copies/mL], geometric mean (95% CI, range)||N/A||13.65(0.43 to 436.13, 0.84-935)||0.53(0.11 to 2.47, 0-625)||0.0472||0.04||(0.002 to 0.95)|
|Exhaled breath H2O2 [µmol/L], geometric mean (95% CI, range)||0.21 (0.15 to 0.31, 0.05-0.63)||0.77 (0.54 to 1.10, 0.42-2.20)||0.81 (0.66 to 1.00, 0.30-1.80)||<0.0013||<0.0014 1 vs. 2 <0.0014 1 vs. 3 1.004 2 vs. 3|| 3.64
|(2.18 to 6.07) (2.44 to 6.01)|
|Resting CL [mV/104 cells], geometric mean (95% CI, range)||0.59 (0.50 to 0.70, 0.28-0.84)||0.79 (0.66 to 0.94, 0.47-1.22)||0.87 (0.73 to 1.04, 0.43-2.14)||0.0043||0.0704 1 vs. 2 0.0044 1 vs. 3 1.004 2 vs. 3||1.51||(1.14 to 2.01)|
|fMLP-induced peak CL [mV/104 cells], geometric mean (95% CI, range)||0.92 (0.74 to 1.13, 0.42-2.24)||1.61 (1.29 to 2.02, 0.90-2.97)||1.42 (1.19 to 1.70, 0.62-2.87)||<0.0013||0.0014 1 vs. 2 0.0054 1 vs. 3 1.004 2 vs. 3|| 1.76
|(1.23 to 2.52) (1.13 to 2.13)|
|Adjusted variables for significant difference on PMNs count:|
|Exhaled breath H2O2 [µmol/L] geometric mean (95% CI)||N/A||N/A||N/A||N/A5|
|Resting CL [mV/104 cells], geometric mean (95% CI)||0.75(0.68 to 0.83)||0.73(0.67 to 0.80)||0.77(0.72 to 0.83)||0.636|
|fMLP-induced peak CL [mV/104 cells], geometric mean (95% CI)||1.13(0.95 to 1.35)||1.52(1.28 to 1.79)||1.30(1.14 to 1.47)||0.0786|
p-values: 1Fisher’s exact test; 2unpaired Student’s t-test (equal variances); 3one-way ANOVA; 4post hoc ANOVA test with Bonferroni adjustment; 5one-way ANCOVA (assumptions
not met); 6one-way ANCOVA (assumptions met).
CL, whole blood chemiluminescence; fMLP, N-formyl-methionyl-leucyl-phenylalanine; HAART, highly active anti-retroviral therapy; N/A, not applicable; PMNs, polymorphonuclear
Table 3: Comparison of viral load, exhaled H2O2 and whole blood chemiluminescence between healthy controls vs. HIV-infected patients without clinical signs and symptoms of opportunistic lung disease (OLD-negative) and either HAART-naive or HAART-treated.
Figure 2: Exhaled H2O2 as marker of oxidative stress in HIV-infected patients without signs and symptoms of opportunistic lung disease (OLD-negative) and either HAART-naive or HAART-treated. (A - top) Dot plot of individual results in healthy controls and HIV-1 infected patients. Box plot shows the geometric mean as a solid line and 95% confidence intervals as a rectangle. (B - bottom) Fold increase of geometric mean exhaled H2O2 (H2O2), resting (rCL) and fMLP-induced peak whole blood chemiluminescence (pCL) in OLD-negative patients either HAART-naive or HAART-treated against healthy controls. Raw data vs. adjusted (adj.) for differences in polymorphonuclear leukocytes count (PMNs) among groups with ANCOVA (not applicable for exhaled H2O2 - assumptions not met).
Relationship between exhaled H2O2 and whole blood chemiluminescence: Spearman’s rank correlations were calculated between exhaled H2O2 and CL variables. Exhaled H2O2 did not significantly correlate with resting CL and fMLP-induced peak CL in OLD-negative patients (all Spearman’s rho p>0.05; detailed data not shown). This was in agreement with no significant associations between exhaled H2O2 and either resting CL or fMLP-induced peak CL in healthy controls (all Spearman’s rho p>0.05; detailed data not shown).
Factors determining exhaled H2O2 in OLD-negative subjects
Univariate analyses: When exhaled H2O2 was established as dependent variable linear regression by nonparametric bootstrap did not find any evidence of significant association with any demographic or clinical variables except a detectable viral load being revealed as a significant and positive predictor for exhaled H2O2 (R-squared=0.23, p=0.014) (Table 4, Figure 3A). On the contrary, the linear regression by nonparametric bootstrap did not show any significant relations between CL variables, either resting CL or fMLP-induced peak CL and viral load (all p>0.05, detailed data not shown).
|Type||Model||Dependent Variable||Independent Variable||Regression Type||R-squared||Coefficient (95% CI)||Standardized Coefficient||p-value|
|Univariate||1||Exhaled H2O2 [µmol/L] (n=19) (detectable viral load subgroup)||Log10 (detectable viral load) [RNA copies/mL]||Linear||0.23||0.180 (0.037 to 0.323)||0.014|
|Multivariate||2||Exhaled H2O2 [µmol/L] (n=19) (detectable viral load subgroup: HAART-naive and HAART-treated)||Log10 (detectable viral load) [RNA copies/mL] Treatment with HAARTConstant||Linear, stepwise||0.35||0.229 (0.080 to 0.377) 0.401 (0.010 to 0.791) -0.189 (-0.818 to 0.440)||0.6082 0.365||0.003 0.044 0.556|
|3||Exhaled H2O2 [µmol/L] (n=14) (detectable viral load subgroup: HAART-treated only)||Log10 (detectable viral load) [RNA copies/mL]CD4 count [cells/µL] HIV-infection duration [yrs] HAART duration [mths]Constant||Linear, stepwise||0.79||0.464 (0.270 to 0.659) 0.0017 (0.0003 to 0.0031) 0.155(0.011 to 0.298) -0.015 (-0.035 to 0.003) -1.50 (-2.56 to -0.44)||1.162 0.938 0.932 -0.944||<0.001 0.015 0.035 0.110 0.005|
HAART, highly active anti-retroviral therapy
Table 4: Summary of regression models in HIV-infected patients without clinical signs and symptoms of opportunistic lung disease (OLD-negative).
Figure 3: Regression models in HIV-infected patients without signs and symptoms of opportunistic lung disease (OLD-negative). Linear regression (A - top) to estimate changes in exhaled H2O2 (y) as a function of log10- transformed detectable HIV-1 viral load (x). Solid line: slope (β1); long dashed lines: 95% confidence intervals. Quantile regression (B-middle) to estimate relationships between exhaled H2O2 and log10-transformed detectable HIV-1 viral load for the exhaled H2O2 distribution from 0.2 to 0.8 quantiles. Solid line: slopes for quantiles (b1), connected with white circles (non-significant) or black circles (significant at p<0.05) and with error bars corresponding to 95% confidence intervals (bolded if p<0.05). Linear regression slope β1 is shown as solid line with long dashed lines. corresponding to 95% confidence intervals. Multiple linear regression (C - bottom) in HAART-treated subgroup only, with exhaled H2O2 as the dependent variable and log10-transformed HIV-1 detectable viral load (A), CD4 count (B), HIV-infection duration (C) and HAART duration (D) as the independent variables. Predicted values of exhaled H2O2 from the multiple regression equation are graphed on the X-axis and the observed values of exhaled H2O2 are plotted on the Y-axis.
Moreover, quantile regression was employed to estimate the relationships between exhaled H2O2 and viral load for a large part of the exhaled H2O2 distribution. We present results by simultaneous bootstrap analysis narrowed to a range from 0.2 to 0.8 quantiles, as justified by the small sample (n=19) and large sampling variation for upper quantiles (Figure 3B). Quantile regression estimates indicated some significant and positive relations between exhaled H2O2 and viral load. The effects of viral load on exhaled H2O2 were best revealed at higher H2O2 exhalation as shown for quantiles from 0.6 to 0.7 (Figure 3B). There was a significant increase in exhaled H2O2 in response to viral load at quantile 0.6 (Pseudo R-squared=0.21, p=0.043) and at quantile 0.7 (Pseudo R-squared=0.21, p=0.042).
Insofar as it can be ascertained, the estimated effects of viral load were well represented by changes in exhaled H2O2.
Multivariate analysis: In order to determine the factors contributing to exhaled H2O2 (to generate hypotheses regarding the causes in variation of the exhalation of H2O2), a multivariate analysis was carried out with exhaled H2O2 as the dependent variable together with smoking habits, duration of HIV-infection (in years), a detectable viral load (in log10 of RNA copies/mL), treatment with HAART and a history of AIDS as possible explanatory factors. In this model, the main contributors to exhaled H2O2 as described by a multiple linear regression equation were viral load and treatment with HAART. The standardized coefficients indicated that viral load contributed most considerably to the model, followed by treatment with HAART. Together, they accounted for 35% of the variance in exhaled H2O2 (Table 4). If the analysis was limited to HAART-treated only and CD4 count (in cells/mL) was added as a possible predictor, then a better model fit was obtained (R-squared=0.79). The analysis confirmed that viral load is the main contributor to the exhaled H2O2; it also suggests that CD4 count has slightly more influence than the duration of HIV-infection in the HAARTtreated subgroup. Figure 3C shows a scatter graph with the predicted values of exhaled H2O2 on the X-axis from the multiple regression equation and the observed values of exhaled H2O2 on the Y-axis. Since the points fall close to the diagonal line, this illustrates the fit of the multiple regression model for prediction of H2O2 exhalation.
Whereas systemic oxidative stress is a common feature of HIV- 1 infection, the lungs are one of the major targets of HIV-1 attack. Accumulation of ROS induces airway inflammation that can be deteriorated by opportunistic lung diseases. Exhaled H2O2 is a known noninvasive inflammatory marker of the respiratory tract which has not been previously reported in HIV-1-infected patients. Moreover, the relationship between exhaled H2O2 and whole blood chemiluminescence has not been established. In this cross-sectional study, we found a high level of exhaled H2O2 among HIV-infected patients as compared to healthy controls, regardless of concomitant respiratory tract infection and despite treatment with HAART. This was accompanied by greater luminol enhanced light emission of the whole blood, either resting or agonist-induced, even though an increase in the exhalation of H2O2 was more evident. Nevertheless, elevated exhalation of H2O2 and appreciable chemiluminescence result from enhanced activity of phagocytes and may be a compensatory mechanism in response to the underlying immunodeficiency . We found that the observed differences in CL variables were explained by an adjustment to a lower whole blood PMNs count. Elbim et al. confirmed that PMNs counts in HIV-infected patients were significantly decreased, though circulating PMNs were activated producing more H2O2 . Interestingly, there was no association between exhaled H2O2 and whole blood chemiluminescence in HIV-infected patients. This associative finding had been previously reported in a study of 41 healthy subjects using the same methods to measure both the exhalation of H2O2 and light emission of blood phagocytes . These results indicate that exhalation of H2O2 in HIV-infected patients does not dependent on ability of blood phagocytes to generate ROS, to a higher extent, this may involve phagocytes within the lungs.
In addition, increased exhalation of H2O2 in OLD-negative patients was associated with detectable viral load. The association was more evident at higher levels of exhaled H2O2 as revealed by quantile regression analysis. The mechanism underlying the observed associations aims to uncover direct casual pathway between viral load and exhaled H2O2. Numerous studies have shown that H2O2 strongly activates HIV long terminal repeat (LTR), containing sequences required for the initiation of HIV-1 transcription via a post-translational control of NF-kappaB [2,31-33]. Moreover, alveolar macrophages are susceptible to HIV-1 virus infection and can be recognized as latent viral reservoir . These cells isolated from asymptomatic HIV-1 positive subjects exhibited a constitutive activation of phosphatidylinositol 3-kinase pathway . The Nef (Negative Regulatory Factor) protein of the HIV-1 virus could be one of the activators in the signal transduction pathway leading to stimulation of the NADPH oxidase complex  and increased oxidants release from macrophages . Additionally, Tat protein has been shown to induce the release of cytokines, thereby enhancing the production of H2O2 in a variety of cells, including macrophages [38,39]. In fact, a study by Buhl showed that alveolar macrophages isolated from the lungs of HIV-infected subjects presented with an increased spontaneous release of oxidants . In all, these can favor conditions for increased H2O2 activity in the airways, rendering an augmentation in viral replication. This is in agreement with the findings of Elbim et al. who reported that basal production of H2O2 in whole blood monocytes is correlated with viral load . These observations point towards the existence of a positive feedback interplay between the production of H2O2 in the airways and HIV-1 viral load. Possibly, this may be a leading mechanism responsible for increased exhaled H2O2 levels in HIV-infected patients.
Moreover, instead of an association with viral load there was a weaker association between increased exhaled H2O2 and treatment with HAART as revealed by a multiple regression analysis. This concurs with study by Mandas et al. showing oxidative imbalance in HIV-1 infected patients treated with antiretroviral therapy , and report by Ngondi et al. demonstrating enhancing (pro-oxidant) effect of HAART on systemic lipid peroxidation . Although, in the latter study, the majority of HIV- 1 infected patients were diagnosed relatively late and presented with an increased severity in clinical status along with active opportunistic infections .
When the analysis was narrowed exclusively to HAART-treated, we found other positive associations with CD4 count and HIV-1-infection duration. Bucy et al. showed evidence that increase in CD4 lymphocytes after HIV antiretroviral therapy reflects redistribution from lymphoid tissues . The link between the increased exhalation of H2O2 and the duration of HIV-infection is likely complex, resulting from disease factors, such as decrease in the GSH concentration over time in the alveolar lining fluid , decreased erythrocyte GSH-Px activity , interactions with HAART  and patient factors including genetically based susceptibility. For example, Delanghe et al. reported that HIV-seropositive patients with the antioxidant protein haptoglobin 2-2 phenotype, known to bind free hemoglobin more slowly, had a higher mortality and worse prognosis than patients with other phenotypes, suggesting enhanced hemoglobin-driven oxidative stress .
Our study has several limitations. Given our small sample size, which necessitates testing in larger groups, we were unable to fully explore all the hypotheses. Secondly, the cross-sectional study design makes the assessment of casual relationships difficult. Finally, we enrolled OLD-negative patients, which included normal chest X-rays, to avoid opportunistic infections, though asymptomatic presentations of OLD could not be excluded. Despite these limitations, the findings in our study should encourage an answer to the question of whether or not the increased exhalation of H2O2 in HIV-1 infected subjects evinces clinical significance. Implications for further studies are HIV associated pulmonary emphysema , since H2O2 is linked to breakdown of elastic fibers  and Kaposi’s sarcoma , as H2O2 mediates herpesvirus reactivation from latency. Therefore, it is possible that determination of exhaled H2O2 can be helpful in the selection of patients with a higher risk of some HIV-1 associated diseases.
This work was supported by Medical University of Lodz Grants: 503/0-079- 01/503-41 and 503/0-079-06/503-01.