Matrix Degradation in Human Immunodeficiency Virus Type 1-Associated Tuberculosis and Tuberculosis Immune Reconstitution Inflammatory Syndrome: A Prospective Observational Study

Abstract

Background: Extensive immunopathology occurs in human immunodeficiency virus (HIV)/tuberculosis (TB) coinfection, but the underlying molecular mechanisms are not well-defined. Excessive matrix metalloproteinase (MMP) activity is emerging as a key process but has not been systematically studied in HIV-associated TB.

Methods: We performed a cross-sectional study of matrix turnover in HIV type 1 (HIV-1)-infected and -uninfected TB patients and controls, and a prospective cohort study of HIV-1-infected TB patients at risk of TB immune reconstitution inflammatory syndrome (TB-IRIS), in Cape Town, South Africa. Sputum and plasma MMP concentrations were quantified by Luminex, plasma procollagen III N-terminal propeptide (PIIINP) by enzyme-linked immunosorbent assay, and urinary lipoarabinomannan (LAM) by Alere Determine TB LAM assay. Peripheral blood mononuclear cells from healthy donors were cultured with Mycobacterium tuberculosis and extracellular matrix in a 3D model of TB granuloma formation.

Results: MMP activity differed between HIV-1-infected and -uninfected TB patients and corresponded with specific TB clinical phenotypes. HIV-1-infected TB patients had reduced pulmonary MMP concentrations, associated with reduced cavitation, but increased plasma PIIINP, compared to HIV-1-uninfected TB patients. Elevated extrapulmonary extracellular matrix turnover was associated with TB-IRIS, both before and during TB-IRIS onset. The predominant collagenase was MMP-8, which was likely neutrophil derived and M. tuberculosis-antigen driven. Mycobacterium tuberculosis-induced matrix degradation was suppressed by the MMP inhibitor doxycycline in vitro.

Conclusions: MMP activity in TB differs by HIV-1 status and compartment, and releases matrix degradation products. Matrix turnover in HIV-1-infected patients is increased before and during TB-IRIS, informing novel diagnostic strategies. MMP inhibition is a potential host-directed therapy strategy for prevention and treatment of TB-IRIS.

Keywords: HIV-1; immune reconstitution inflammatory syndrome; matrix metalloproteinase; procollagen III N-terminal propeptide; tuberculosis.

Figure 1.

Pulmonary matrix metalloproteinase (MMP) concentrations are increased in tuberculosis (TB) and differ by human immunodeficiency virus (HIV) serostatus. Pulmonary TB is associated with increased MMP-1, -2, -3, -7, -8, -9, and -10 concentrations in sputum in comparison to respiratory symptomatic (RS) and healthy controls (HC) (A–G). Comparison of TB (HIV⁻) with TB (HIV⁺) demonstrated lower median sputum MMP-1 (A), -2 (B), -3 (C), and -9 (F) concentrations in TB (HIV⁺). Median MMP-8 was also reduced in TB (HIV⁺) compared to TB (HIV⁻), although to a lesser extent (E). TB (HIV⁻) patients with bilateral radiographic abnormalities had elevated sputum MMP-1, compared with TB (HIV⁻) patients with unilateral abnormalities (H). However, in TB (HIV⁺), MMP-1 was similar in patients with bilateral and unilateral chest radiographic involvement (H). In A–G, boxes represent the first and third quartiles and horizontal bars within indicate median values; whiskers indicate minimum and maximum values. In (H), triangles represent TB (HIV⁻) and circles represent TB (HIV⁺). Horizontal bars between the datasets indicate Mann-Whitney U test comparisons; in A–G comparisons between TB (HIV⁻) and HC (HIV⁻), TB (HIV⁺) and HC (HIV⁺), and TB (HIV⁺) and TB (HIV⁻) are shown. *P < .05, **P < .01, ****P < .0001.

Figure 2.

Tuberculosis (TB) increases systemic extracellular matrix turnover, which is further augmented by human immunodeficiency virus (HIV) coinfection. Plasma procollagen III N-terminal propeptide (PIIINP) concentration was elevated in TB patients compared to respiratory symptomatic (RS) and healthy controls (HC) (A). TB (HIV⁺) and TB (HIV⁻) patients had higher plasma PIIINP concentrations than corresponding controls (B). In contrast to sputum matrix metalloproteinases (MMPs), plasma PIIINP was further elevated in TB (HIV⁺) compared with TB (HIV⁻). In HIV⁺ patients, plasma PIIINP concentration and peripheral blood CD4 cell count negatively correlated (C), whereas HIV-1 viral load positively correlated with PIIINP concentration (D). Plasma PIIINP was elevated in patients with extrapulmonary TB (EPTB) compared to those without EPTB (E). Plasma MMP-1 (F) and plasma MMP-8 (G) were elevated in TB (HIV⁻) and TB (HIV⁺) compared to respective RS and HC, and did not differ by HIV serostatus in TB patients. Boxes represent the first and third quartiles and horizontal bars within indicate median values; whiskers indicate minimum and maximum values. Horizontal bars between the datasets indicate Mann-Whitney U test comparisons. *P < .05, **P < .01, ***P < .001, ****P < .0001. Correlations were performed using Spearman rank-order correlation coefficient.

Figure 3.

Paradoxical tuberculosis immune reconstitution inflammatory syndrome (TB-IRIS) is characterized by systemic inflammation at TB diagnosis and during TB-IRIS. We studied 47 antiretroviral therapy (ART)–naive TB patients with advanced human immunodeficiency virus (HIV; CD4 count <200 cells/µL) at enrollment, who underwent clinical observation at TB diagnosis (TB0) and biweekly for the first 4 weeks of ART (ARV0, ARV2, ARV4). TB-IRIS patients were characterized by elevated heart rate (A) and elevated respiratory rate (B), compared to non-IRIS controls. CD4 count increased in both TB-IRIS patients and non-IRIS controls following ART initiation (C) and concurrently HIV-1 viral load reduced (D), although median HIV-1 viral load was higher in TB-IRIS patients than in non-IRIS patients at TB diagnosis and at ARV2. Elevated C-reactive protein was a feature of TB-IRIS onset (E). Median lymphocyte counts were lower in TB-IRIS patients at all timepoints (F), whereas neutrophil counts (G) and monocyte counts (H) were increased at ARV2 and to a lesser extent at ARV4. In A, B, F, G, and H, boxes represent the first and third quartiles, horizontal bars within the median values, and whiskers the minimum and maximum values. In C and D, data are median values (TB-IRIS, filled circles; non-IRIS, open squares) linked by horizontal lines, and interquartile ranges are shown by vertical bars. In E, individual data points are shown, including results for unscheduled visits (ARV1, 3, 5, and 6 representing visits at 1, 3, 5 and 6 weeks of ART, respectively), and horizontal lines represent the median. In all panels, asterisks indicate Mann-Whitney U test comparisons; summary P values: *<.05 and **<.01.

Figure 4.

Immunopathology in paradoxical tuberculosis immune reconstitution inflammatory syndrome (TB-IRIS) is associated with increased matrix metalloproteinase (MMP) activity. Plasma procollagen III N-terminal propeptide (PIIINP) was elevated in TB-IRIS patients compared to non-IRIS control patients at the time of TB diagnosis and also at IRIS onset (A). Plasma MMPs were elevated in TB-IRIS compared to non-IRIS controls, including MMP-1 (B), MMP-3 (C), and, most consistently, MMP-8 (D). Plasma MMP-8 positively correlated with plasma PIIINP (E) and also with neutrophil count (F). In A–D, boxes represent the first and third quartiles; horizontal bars within median values and whiskers minimum and maximum values. Comparisons are by Mann-Whitney U test. *P < .05, **P < .01, ***P < .001. In E and F, individual data points are plotted by filled circles. Spearman rank-order correlation coefficient r and P values are reported for correlations.

Figure 5.

Elevated matrix metalloproteinases (MMPs) associate with increased tuberculosis (TB) antigen load, and Mycobacterium tuberculosis (Mtb)–driven MMP activity is inhibited by doxycycline (Doxy/Dox). Tuberculosis immune reconstitution inflammatory syndrome (TB-IRIS) patients with positive urinary lipoarabinomannan (LAM) had increased plasma MMP-3, MMP-7, and MMP-8 compared with TB-IRIS patients who were LAM negative (A). Plasma MMP-3 was higher at TB diagnosis (TB0) but not at 2 weeks of antiretroviral therapy (ARV2), whereas MMP-7 and MMP-8 were most significantly increased at ARV2. MMP-8 concentrations were measured in culture supernatants of peripheral blood mononuclear cells (PBMCs) stimulated with heat-killed H37Rv Mtb in a cohort of 22 TB-IRIS patients and 22 non-IRIS controls (B). After stimulation, MMP-8 secretion was greater from TB-IRIS PBMCs than non-IRIS controls. In a 3D cell culture model of TB, microspheres were impregnated with ultraviolet-killed Mtb-stimulated PBMCs and either DQ-gelatin or DQ-collagen (DQC), which increase in fluorescence when cleaved. Mtb stimulation increased total gelatin degradation within microspheres compared to control PBMCs (C). Addition of doxycycline to the surrounding cell culture media inhibited extracellular matrix breakdown (D). Similarly, doxycycline suppressed Mtb-driven collagen degradation in a dose-dependent manner (E). In A and B, horizontal lines indicate medians, and comparisons between groups are by Mann-Whitney U test analysis; in C–E, means and standard error of the mean are shown and analyses are by 2-way repeated measures analysis of variance, with Tukey posttest comparison. Comparisons shown are for gelatin + PBMC + Mtb compared to gelatin + PBMC (C) and doxycycline 20 μg/mL compared to Mtb alone. Summary P values: *<.05, **<.01, ****<.0001.