Matrix metalloproteinase inhibitors enhance the efficacy of frontline drugs against Mycobacterium tuberculosis

Abstract

Mycobacterium tuberculosis (Mtb) remains a grave threat to world health with emerging drug resistant strains. One prominent feature of Mtb infection is the extensive reprogramming of host tissue at the site of infection. Here we report that inhibition of matrix metalloproteinase (MMP) activity by a panel of small molecule inhibitors enhances the in vivo potency of the frontline TB drugs isoniazid (INH) and rifampicin (RIF). Inhibition of MMP activity leads to an increase in pericyte-covered blood vessel numbers and appears to stabilize the integrity of the infected lung tissue. In treated mice, we observe an increased delivery and/or retention of frontline TB drugs in the infected lungs, resulting in enhanced drug efficacy. These findings indicate that targeting Mtb-induced host tissue remodeling can increase therapeutic efficacy and could enhance the effectiveness of current drug regimens.

Fig 1. MMP-9 and MMP-2 level are up-regulated during Mtb infection or TDM induction.

(A): Heatmap showing gene expression profile generated from microarray dataset [34] analysis comparing 5 caseous human pulmonary TB granuloma and 2 normal lung parenchyma (GSE20050). (B): Volcano plot showed 12878 up-regulated genes (red dots, p.value < 0.05, Fold change >2) and 8062 down-regulated genes (blue dots, p.value < 0.05, Fold change <0.5) in TB granuloma. MMP-9 and MMP-2 were labeled among the up-regulated genes. (C): Protein level of MMP-9 and MMP-2 in uninfected and infected mice lung (n = 3). Representative images were showed. (D): Representative images of luminescence from MMP-9 or MMP-2 reporter cell lines within TDM granulomas in C57BL/6J mice. The same mice were imaged at Day 3 and Day 7 (n = 5). Experiment performed twice with similar result. (E): Quantitative analysis of luminescent signal at Day 3 and Day 7 of MMP-2 and MMP-9 reporter cell lines within TDM granuloma (n = 5). Data represented mean ± SD. ****: p < 0.0001, 2-way ANOVA with Šidák multiple comparison test.

Fig 2. MMP inhibitors facilitate INH killing of Mtb.

(A): Experimental setting of Marimastat’s effect in lung infection model (n = 5). (B): CFU count of 2 groups (PBS and Marimastat) at Day 14, and 4 groups (PBS, Marimastat, INH and Mariamstat+INH) at Day 28 post infection. Each dot represented one mouse (n = 5). Experiment was repeated 3 times with similar observation. Data showed results from one representative experiment. (C): H&E stain of the 4 groups (PBS, Marimastat, INH and Mariamstat+INH) at Day 28 infection (n = 5). Scale bar: 1mm. Arrow indicated consolidated region circled by black line. (D): Quantification of inflammatory region percentage within tissue of in all mice from different treatment groups (n = 5). Data represented mean ± SD. **: p < 0.01, One-way ANOVA with Šidák multiple comparison test.

Fig 3. Other frontline TB drug (RIF) and other MMP inhibitors can induce synergistic effect to reduce Mtb burden.

(A): CFU count of lung tissue from Mtb-infected mice treated with Marimastat and RIF. Each dot represented one mouse (n = 4). Data represented mean. ***: p < 0.001, Two-tailed Unpaired Student t test with Welch-correction. (B): CFU count of lung tissue from Mtb-infected mice treated with other MMP inhibitors and INH. Each dot represented one mouse (n = 5). Data represented mean. ***: p < 0.001, One-way ANOVA with Šidák multiple comparison test.

Fig 4. Inhibition of MMP results in an increase in both collagen and mannose binding lectin (MBL) within the infected tissue.

(A): Hydroxyproline concentration in lung tissue from infected mice under different treatments (n = 5). Data represented mean ± SD. **: p < 0.01, ***: p < 0.001, One-way ANOVA with Šidák multiple comparison test. (B): MBL protein level from lung lysate of infected mice with different treatments (n = 5).

Fig 5. CD31 and α-SMA staining in the lung of infected mice treated with Marimastat.

(A): Mice infected with Mtb were treated with PBS (n = 5) or Marimastat (n = 5). Lung tissue from infected mice was stained for CD31 (upper panel) and α-SMA (lower panel). Both PBS and Marimastat groups had positive staining of CD31 and α-SMA. Scale bar: 80μm. (B, C, D): Quantification of CD31 positive blood vessel number (B), percentage of α-SMA positive staining (C), and α-SMA positive blood vessel number (D) in PBS or Marimastat treated mice (n = 5) 4-week post infection. Data represent mean ± SD. *: p < 0.05, **: p < 0.01, Two-tailed Unpaired Student t test with Welch-correction. (E, F): Quantification of percentage of α-SMA positive staining (E), and α-SMA positive blood vessel number (F) in consolidated area (granuloma like area) and surrounding normal tissue from PBS or Marimastat treated mice (n = 5) 4-week post infection. Data represented mean ± SD. *: p < 0.05, **: p < 0.01, One-way ANOVA with Šidák multiple comparison test.

Fig 6. Leakage and delivery of blood vessels measured by fluorescent dextran.

(A): Mice infected with mCherry Mtb bacteria (red) were treated with PBS (n = 3) or Marimastat (n = 3). 70k dextran (green) and 10k dextran (yellow) were injected to mice before euthanization. Lung tissue from infected mice was stained for CD31 (blue) to label the blood vessels. Both PBS and Marimastat groups had positive staining of CD31 and positive signal of 70k and 10k dextrans. Scale bar: 20μm. (B): Quantification of the ratio of 10k dextran outside vs inside of blood vessel (C): Quantification of the ratio of 70k dextran outside vs inside of blood vessel. Data represented mean ± SD. **: p < 0.01, One-way ANOVA with Šidák multiple comparison test.

Fig 7. Delivery and/or retention of Evans blue dye or frontline TB drugs in infected animals treated with Marimastat.

(A): Mtb infected mice treated with or without Marimastat were injected intravenously with Evans blue dye before euthanization. Retention of Evans blue dye within the lung of infected mice treated with or without Marimastat. Data represented mean ± SD. **: p < 0.01, Two-tailed Unpaired Student t test with Welch-correction. Experiment was repeated twice. (B and C): Mtb infected mice treated with or without Marimastat were injected intravenously with RIF and INH before euthanization. At indicated time points, lung and blood samples were collected for drug measurement. The concentration of RIF (B) and INH (C) in the lung were normalized by drug concentration in the plasma. Data represented mean ± SD. *: p < 0.05, ****: p < 0.0001, One-way ANOVA with Šidák multiple comparison test.