AJRCCM Articles in Press. Published on September 1, 2005 as doi:10.1164/rccm.200505-753OC
Mycobacterium tuberculosis but not vaccine BCG specifically
up-regulates matrix metalloproteinase-1
Paul T G Elkington,1* Robert K Nuttall,2* Joseph J Boyle,3 Cecilia M O’Kane,1
Donna E Horncastle,3 Dylan R Edwards2 and Jon S Friedland1
1Department of Infectious Diseases and 3Histopathology, Imperial College,
Hammersmith Campus, London, UK; 2School of Biological Sciences, University of
East Anglia, Norwich, UK
* These authors contributed equally to this work
Requests for reprints:
Professor J S Friedland
Department of Infectious Diseases
Du Cane Road
London W12 0NN
Professor J S Friedland
Fax 00 44 20 8383 3394
Tel 00 44 20 8383 8521
Copyright (C) 2005 by the American Thoracic Society.
This work was supported by a Wellcome Trust Clinical Research Training Fellowship
(PE), the Hammersmith Hospitals Trustees Research Committee, and a grant from the
European Union Framework Programme 6 (LSHC-CT-2003-503297).
Running head: Metalloproteinases in tuberculosis
Descriptor number: 122
Total word count: 3,334
This article has an online data supplement, which is accessible from this issue's table
of content online at www.atsjournals.org
Rationale: Pulmonary cavitation is fundamental to the global success of Mycobacterium
tuberculosis. However, the mechanisms of this lung destruction are poorly understood. The
biochemistry of lung matrix predicts matrix metalloproteinase involvement in
immunopathology. Methods: We investigated gene expression of all matrix
metalloproteinases, ADAMs (a disintegrin and metalloproteinase) and tissue inhibitors of
metalloproteinases in M. tuberculosis-infected human macrophages by Real Time PCR.
Matrix metalloproteinase secretion was measured by zymography and Western analysis, and
expression in patients with pulmonary tuberculosis was localized by immunohistochemistry.
Results: MMP-1 and MMP-7 gene expression and secretion are potently up-regulated by M.
tuberculosis, and no increase in tissue inhibitor of metalloproteinase expression occurs to
oppose their activity. Dexamethasone completely suppresses MMP-1 but not MMP-7 gene
expression and secretion. In patients with active tuberculosis, macrophages express MMP-1
and MMP-7 adjacent to areas of tissue destruction. MMP-1 but not MMP-7 expression and
secretion is relatively M. tuberculosis-specific, is not up-regulated by tuberculosis-associated
cytokines and is prostaglandin dependent. In contrast, the vaccine M.bovis BCG does not
stimulate MMP-1 secretion from human macrophages, although M. tuberculosis and BCG do
up-regulate MMP-7 equally. BCG-infected macrophages secrete reduced prostaglandin E2
concentrations compared to M. tuberculosis -infected macrophages, and prostaglandin
pathway supplementation augments MMP-1 secretion from BCG-infected cells.
Conclusions: M. tuberculosis specifically up-regulates MMP-1 in a cellular model of human
infection and in patients with tuberculosis. In contrast, vaccine BCG, which does not cause
lung cavitation, does not up-regulate prostaglandin E2-dependent MMP-1 secretion.
Abstract word count: 239
Mycobacterium tuberculosis (MTb) infects approximately a third of the world’s population
and kills 2-3 million people each year (1, 2). Pulmonary cavitation is vital to the transmission
of MTb and favors bacterial persistence (3). The cavity is an immuno-privileged site with
MTb replicating freely in the cavity wall (3) After treatment of tuberculosis, opportunistic
pathogens may colonize the residual cavity (4). The ability to cause pulmonary cavitation in
the immunocompetent host differentiates MTb from the vaccine strain, M. bovis BCG (BCG).
The standard animal model of tuberculosis (Tb) is the mouse, which is valuable in studying
immunity to MTb (5), but the pathology of murine Tb is divergent to human Tb, being purely
fibrotic without cavitation (5, 6). Guinea pigs develop granulomas similar to humans but
rarely cavitate, while rabbits develop cavitation with BCG, which is not a pulmonary
pathogen in man (5). Therefore, mechanisms by which MTb causes tissue destruction is best
investigated in primary human cells and by clinical studies.
Proteases from inflammatory cells have been hypothesized to cause tissue destruction in Tb
since the era of Koch and Virkow (7). For cavitation to occur, the extracellular matrix of the
lung must be degraded. Fibrillar type I collagen provides the tensile strength of the matrix
and elastin allows distensibility (8). Type I collagen is highly resistant to enzymatic
degradation and only matrix metalloproteinases (MMPs) are able to cleave it at neutral pH
(9). Excess MMP activity in the lung causes tissue destruction. For example, mice
constitutively over-expressing MMP-1 (interstitial collagenase) develop emphysema (10) and
MMPs are implicated in other animal models of emphysema (11-13).
MMPs are a family of zinc-binding proteases that collectively degrade all components of the
extracellular matrix (14) and are from the metalloprotease class that includes the ADAM (a
disintergrin and metalloproteinase) family. MMP-1, -8 and -13 are the primary enzymes
responsible for degradation of type I collagen (15). Multiple MMPs can degrade elastin,
including MMP-2, -7, -9, and -12 (15). MMP-7 (matrilysin) is the most potent elastase
produced by macrophages (16). In addition to matrix remodelling, MMPs have important
roles in immunity, such as the proteolytic processing of chemokines and cytokines (17, 18).
MMP activity is regulated by multiple mechanisms, including transcriptional control,
secretion as pro-enzymes that require proteolytic activation, compartmentalization and
secretion of specific inhibitors, TIMPs (17).
MMP expression alters with cellular differentiation: macrophages secrete a broader spectrum
and greater quantities of these enzymes than monocytes (19, 20). Macrophage MMP
production is primarily regulated through a prostaglandin E2 – cAMP dependent pathway (21-
23). Macrophages are a principal source of MMPs at inflammatory loci (24). In Tb,
macrophages form the central core of the granuloma and are the key effector of the host
immune response to MTb (25). In pulmonary MTb infection, excessive proteolytic activity
causes matrix breakdown, caseation of the granuloma, liquefaction and finally cavitation (26).
We hypothezise that MMPs are critical to this pulmonary destruction in Tb.
This work has been previously presented in part in the form of abstracts (27, 28).
Methods word count: 500
Additional details of the following methods are provided in the online supplemental material
MTb and BCG culture
M. tuberculosis H37Rv and M. bovis BCG were cultured in supplemented Middlebrook 7H9
medium (BD Biosciences, Oxford, UK). For infection experiments, mycobacteria were used
at mid-logarithmic growth at optical density of 0.60 (Biowave Cell Density Meter, WPA,
Monocyte purification and maturation
Monocytes were isolated from single donor buffy coats (National Blood Transfusion Service,
UK) by density centrifugation and adhesion purification. Monocytes were matured to
monocyte-derived macrophages (macrophages) for 5 days. Medium was then changed to
RPMI with 2mM glutamine and 10µg/ml ampicillin and macrophages were infected with
MTb or BCG at a multiplicity of infection of 1.
Gene expression analysis by Real Time Real Time Polymerase Chain
Macrophages were lysed using Tri-Reagent (Sigma, Poole, UK) and total RNA extracted.
1µg RNA was reverse transcribed using 2µg random hexamers (Amersham Biosciences, Little
Chalfont, UK) and 200 units of Superscript II reverse transcriptase (Invitrogen, Paisley, UK),
according to the supplier’s instructions. qPCR reactions were done on the ABI Prism 7700
(Applied Biosystems, Warrington, UK) according to previously described methods (29). The
cycle threshold (CT) at which amplification entered the exponential phase was determined and
this number was used to indicate the amount of target RNA in each sample.
Casein and gelatin zymography
For assessment of MMP-1 and MMP-7 activity, samples were analyzed on 12% casein gels
(Invitrogen) and incubated for 40h in collagenase buffer at 37oC. Caseinolytic activity was
revealed by Coomassie blue staining (Pharmacia, Sweden). All experimental samples were
run in parallel to a lane containing 5ng recombinant MMP-1 (Oncogene, Nottingham, UK) to
standardize between gels. Densitometric analysis of zymography gels was performed using
NIH Image version 1.61. Gelatin zymography to detect MMP-9 was performed as previously
Western blotting for MMP-1 and MMP-7
Western blotting was performed using anti-human MMP-1 Ab (The Binding Site,
Birmingham, UK) and MMP-7 Ab-3 (Oncogene) detected using chemiluminence.
Immunohistochemistry for MMP-1 and MMP-7 was performed on paraffin embedded lung
biopsies from 6 patients with culture-proven MTb infection and 6 non-infected controls.
Sections were probed with primary antibodies (MMP-1 Ab-1, Oncogene and MMP-7 Ab-3,
Oncogene and isotype controls) and antibody was detected with the Menarini non-
biotinylated kit according to manufacturer’s instructions. Ethical consent was obtained from
the Hammersmith Hospitals Research Ethics Committee for the use of archived lung biopsies.
Prostaglandin E2 and TIMP-1 ELISAs
Prostaglandin E2 levels in cell culture medium were determined by competitive binding
immunoassay (R&D, Abingdon, UK) according to manufacturer’s instructions. The lower
level of sensitivity was <15.9pg/ml. TIMP-1 levels in cell culture medium were measured by
ELISA (R&D) according to the manufacturer’s instructions. The lower level of sensitivity
To analyze the effect of multiple treatments (LPS and MTb) at different times two-way
ANOVA was used, followed by Tukey’s multiple comparison. To analyze the MMP-1 and
MMP-7 activity of MTb infected macrophages compared to control the student’s t-test was
used. A p value of <0.05 was taken as statistically significant.
MMP, ADAM and TIMP gene expression in MTb-infected human
Gene expression of all 23 human MMPs, 9 ADAMs and 4 TIMPs was analyzed by RT-PCR
in unstimulated macrophages and in lipopolysaccharide (LPS)- and MTb-stimulated
macrophages at 6 and 24 hours (Figure 1). MTb increased MMP-1 gene expression at 24h to
levels 220-fold greater than control, MMP-3 expression 547-fold control, MMP-7 expression
10-fold control and MMP-10 expression 114-fold control (all p< 0.05, summarised in Table
1). No significant change in MMP-2, -8, -9, -11, -14, -15 or -17 expression was detected.
LPS caused a similar 8-fold increase in MMP-7 expression compared to MTb but only
increased MMP-1 expression to 21% of the level stimulated by MTb. LPS also increased
MMP-12 expression 15-fold (p< 0.05). MMP-19 to MMP-28 expression was not
significantly changed by either stimulus (Figure E1). MMP-13, -16, -20, and -26 mRNA were
In contrast, MTb did not significantly increase TIMP expression. MTb stimulated a non-
significant 2.8 fold increase in TIMP-1 expression, while TIMP-2, -3 and -4 gene expression
were suppressed by 2-, 20- and 2-fold respectively (Figure 1). Gene expression analysis of the
9 ADAMs with a putative protease domain and of the cell-bound MMP inhibitor reversion
including cysteine-rich protein with kazal motifs (RECK) did not demonstrate any significant
changes following MTb or LPS stimulation (Online supplemental material, Figure E1 and
E2). ADAM33 was not detected.
The data in figure 1 present relative mRNA levels normalized to 18S rRNA, indicating the
proportional, not absolute, MMP expression level. The cycle threshold (CT) where PCR
amplification enters the logarithmic phase indicates levels of gene expression; a low threshold
reflects high gene expression. Therefore, the CT was used to assess the expression level of
each MMP (Figure 2A). Of the MMPs up-regulated by MTb, MMP-1 and MMP-7 were the
most highly expressed, with mean threshold in MTb-infected cells of 24.1 and 18.0
respectively, while MMP-3 and MMP-10 mRNA levels only reached moderate levels with
mean thresholds greater than 25 (Figure 2A). Taken together, the data in Figure 1 & 2A
demonstrate that the principal MMPs up-regulated by MTb are MMP-1 and MMP-7.
MTb infection drives macrophage MMP-1 and MMP-7 secretion
We next analyzed MMP secretion to investigate whether increased gene expression results in
greater enzyme activity. Casein and gelatin zymography demonstrate that MMP-1, -7 and -9