Plasma heme oxygenase-1 levels distinguish latent or successfully treated human tuberculosis from active disease

Abstract

Background: Tuberculosis (TB) is associated with oxidative stress and the induction of host anti-oxidants to counteract this response. Heme oxygenase-1 (HO-1) is a critical promoter of cytoprotection in diverse disease models including mycobacterial infection. Nevertheless, the pattern of expression of HO-1 in human tuberculosis has not been studied. Here, we examine expression of HO-1 in M. tuberculosis-exposed and -infected individuals and test its ability to distinguish active from latent and successfully treated TB cases. In addition, we assess correlations between plasma levels of HO-1 and cytokines closely associated with the immunopathogenesis of TB.

Methods: Cross-sectional and longitudinal analyses of levels of HO-1, acute phase proteins and pro-inflammatory cytokines were performed in plasma samples from individuals with active pulmonary, extra-pulmonary or latent TB infection and healthy controls as part of a prospective cohort study in South India.

Results: Systemic levels of HO-1 were dramatically increased in individuals with active pulmonary and extra-pulmonary tuberculosis and particularly those with bilateral lung lesions and elevated bacillary loads in sputum. HO-1 levels effectively discriminated active from latent tuberculosis with higher predictive values than either C-reactive protein or serum amyloid protein. Moreover, there was a marked reduction in HO-1 levels in active TB cases following anti-tuberculous therapy but not in those who failed treatment. Pulmonary TB patients displaying the highest concentrations of HO-1 in plasma exhibited significantly elevated plasma levels of interleukin (IL)-10, interferon (IFN)-γ and IL-17 and diminished levels of tumor necrosis factor (TNF)-α.

Conclusion: These findings establish HO-1 levels as a potentially useful parameter for distinguishing active from latent or treated pulmonary tuberculosis, that is superior in this respect to the measurement of other acute inflammatory proteins.

Figure 1. HO-1 levels associate with disease severity and sputum positivity in tuberculosis.

(A) HO-1 levels were evaluated in plasma samples from 97 individuals with active pulmonary tuberculosis (PTB), 35 with active extra pulmonary tuberculosis (EPTB), 39 with latent M. tuberculosis infection (LTBI) and 40 healthy donors (HD). The samples were collected prior to initiation of anti-tuberculosis treatment (ATT). Kruskal-Wallis test and Dunn’s multiple comparisons were used to evaluate statistical differences between the groups. (B) HO-1 levels were compared between 40 PTB patients with a diagnosis of unilateral lung lesions and 37 patients with bilateral lesions identified by chest radiography. (C) PTB cases were classified according to the presence or absence of the acid fast staining bacilli (AFB) in sputum samples. In (B) and (C), the Mann Whitney test was used for statistical comparisons. (D) PTB cases were stratified according to the quantitative bacillary sputum grade determined by AFB staining. Data was analyzed using the Kruskal-Wallis test with linear trend post-test (left panel) or by Spearman correlation (right panel). In (A), (B), (C) and (D), bars represent median values. (E) Comparison of HO-1 levels pre and post ATT in a subset of patients with PTB for which plasma samples were available (n = 33). Data from 5 patients that failed treatment are also shown. The Wilcoxon matched pairs test was used to evaluate the significance of the effect of ATT on HO-1 values. Tx, treatment. *P<0.05; **P<0.01; ***P<0.001.

Figure 2. HO-1 as a candidate marker of active pulmonary tuberculosis.

Levels of HO-1, C-reactive protein (CRP) and Serum Amyloid Protein-A (SAA) were compared between individuals with LTBI and active TB (A) or between patients with active EPTB and those with PTB (C) using the Mann Whitney test. The power of each of the three different biomarkers to discriminate the above clinical outcomes as analyzed by ROC curves is illustrated in (B) and (D). Symbols represent individual patients and bars represent median values. In (A) and (C), dotted lines represent median values from healthy donors. AUC, area under the curve; 95%CI, 95% confidence interval.

Figure 3. Associations between HO-1 and cytokine levels in active pulmonary tuberculosis.

(A) Correlations between plasma HO-1 levels and systemic concentrations of IL-10, TNF-α, IFN-γ and IL-17 in individuals with active pulmonary tuberculosis (n = 97). Dotted lines on the X-axis represent the median value of HO-1 within the group of patients with pulmonary disease, while dotted lines on each Y-axis indicate median values for each cytokine. In (B), the correlations were stratified according to the bacillary load in sputum smears. Gray areas designate the quadrants that include the individuals simultaneously displaying values of HO-1 and cytokines above the medians or in the case of TNF-α below the median. Percentage of individuals within in the gray areas was compared between the groups with different bacillary sputum grades using a chi-square analysis and correlations evaluated using Spearman test.