Multinucleate Giant Cells Release Functionally Unopposed Matrix Metalloproteinase-9 In Vitro and In Vivo

Multinucleated giant cells (MGCs) are characteristically formed during inflammatory responses leading to granuloma formation. MGCs occur in granulomas in many infections, including tuberculosis (TB), brucellosis, and leishmaniasis, as well as in inflammatory conditions such as rheumatoid arthritis and sarcoidosis. The absence of MGCs may indicate a relatively ineffective host immune response, as occurs in some patients with AIDS who have TB [1]. Although most MGCs arise from fusion of mononuclear phagocytes [2], some, such osteoclasts and HIV-1-induced T cell-derived syncytia, are not monocyte derived—and, paradoxically, the majority of functional data concerning MGCs relate to such specialized cells, rather than to the more common monocyte-derived MGCs.

Matrix metalloproteinases (MMPs) are zinc-dependent proteases that degrade extracellular matrix and are involved in tissue remodeling and cellular migration [3]. The proteolytic activity of MMPs is inhibited by specific tissue inhibitors of metalloproteinase (TIMPs), with which they form 1:1 complexes. The MMP:TIMP ratio is critical in regulating the proteolysis of connective tissues and in controlling tissue damage [4]. Modification of MMP activity is of potential therapeutic importance in inflammatory lung disease [5]. Several types of MMPs have been found in granulomatous inflammatory lesions, but the gelatinases (i.e., MMP-2 and -9) are most frequently reported in association with MGCs [6, 7].

There are no data on MGC MMP secretion in response to infection, and the relative contribution that monocytes, macrophages, and MGCs make to MMP activity is unknown. MMP-9 is quantitatively the most important of several MMPs secreted by monocytic cells. Active and latent MMP-9 is inhibited by TIMP-1, the major TIMP secreted by mononuclear phagocytes. We first performed a comparative study investigating MMP-9 and TIMP-1 secretion during and after human MGC formation in vitro, comparing this secretion with that by monocytes and macrophages. During formation, MGCs secrete more MMP-9 and less TIMP-1 than do macrophages and, once formed, spontaneously release higher MMP-9 concentrations than do monocytes or macrophages. This suggests that they might play a role in tissue destruction in infectious granulomatous disease. Therefore, we investigated MMP-9 and TIMP-1 secretion in granulomas from patients with TB, a disease that causes ∼2 million deaths per year and that, because of its increasing drug resistance, is likely to require new therapeutic approaches [8]. We demonstrate that high-level expression of MMP-9, but of not TIMP-1, is associated with MGCs and is adjacent to areas of caseous necrosis in human lymph nodes.

Materials and methods. Peripheral blood monocytes were isolated by density-gradient centrifugation over Ficoll-Paque and adhesion purification, on plastic for 2 h, of leukocyteenriched buffy coats obtained from the blood-transfusion service. We have reported elsewhere that the highest MGC fusion rates (>80%) are achieved by culturing fresh, undifferentiated monocytes at a density of 5 × 10⁵/mL in Dulbecco's modified Eagle medium for 3 days, in the presence of 5 µg/mL concanavalin A (ConA), 1000 IU/mL interferon (IFN)–γ, and 10% human serum [9]; monocyte-derived macrophages were obtained by culturing monocytes for 3 days in the absence of ConA and IFN-γ. After formation of MGCs or macrophages, cells were washed before further experiments. At the start of each experiment, ∼3 × 10⁶ cells were placed in standard 6-well tissue-culture plates

MMP-9 was detected by zymography using standard methodology [10]. In brief, cell-culture supernatants were mixed with 5 × loading buffer (50 mmol/L Tris HCl [pH 7.6], 10% glycerol, 1% SDS, and 0.01% bromophenol blue) and were resolved on 11% SDS gel impregnated with 0.12 mg/mL gelatin. Gels were run at 180 V, washed for 1 h in 2.5% Triton X-100, and incubated, overnight at 37°C, in collagenase buffer (50 mmol/L Tris HCl [pH 7.6], 0.2 mol/L NaCl, and 5 mmol/L CaCl2). After a Coomassie blue-based staining, proteolytic activity was revealed as white bands on a dark background. Gel images were then digitized, and densitometric analysis was performed by use of the Image 1.61 analysis program (National Institutes of Health). A linear range for MMP-9 quantitation (6–170 ng/mL) was determined on the basis of standard curves obtained by use of known quantities of recombinant MMP-9 (32–8000 pg). Samples with MMP-9 activity above the upper limit of this range were diluted as necessary.

TIMP-1 concentrations were measured by ELISA [11]. In brief, standard 96-well plates were coated, overnight at 4°C, with monoclonal antibody, 5 µg/mL RRU-T5, in PBS (antibody was provided by Prof. T. Cawston, University of Newcastle, United Kingdom). Plates were blocked with 10 mg/mL bovine serum albumin (BSA) in PBS before TIMP-1 standards and samples were added. After incubation overnight at 4°C, 25 µg/ mL biotinylated polyclonal antibody, B-anti CL1 (in 0.5 mg/ mL BSA in 0.1% Tween 20 in PBS), was added for 2 h at room temperature. Binding was detected by addition of streptavidin horseradish peroxidase (1:1000), with O-phenylenediamine as a substrate. The reaction was stopped by addition of 0.5 mol/ L H₂SO₄, and absorbency at 492 nm was measured. The lower limit of sensitivity of this assay is 5 ng/mL. Data for MMP-9 and TIMP-1 are presented as mean ± SE. Student's t test was used for intergroup comparisons, and P < .05 was considered to be significant.

Immunohistochemical analysis of MMP-9 and TIMP-1 was performed on paraffin-embedded sections that, because of clinical indications, had been obtained from patients with tuberculous lymphadenitis. The study received approval from the local ethical committees of AB Prisma (Lima, Peru) and Hammersmith Hospital (London, United Kingdom). Sections were washed in xylene on 3 occasions, and endogenous peroxidase activity was blocked with 30% hydrogen peroxide. Sections for TIMP-1 staining, but not those for MMP-9 staining, were heated in 500 mL of citrate buffer for 18 min. Sections were blocked and then incubated, overnight at room temperature, with either anti-MMP-9 (VIIC2; Neomarkers) or anti-TIMP- 1 (102D1; Neomarkers) murine monoclonal antibodies. Slides were washed 3 times in 0.1% Tween 20 in PBS before being exposed first to biotinylated goat anti-mouse IgG and then to peroxidase-labeled avidin. After further washing, slides were incubated with 3,3′-diaminobenzidine substrate, according to the manufacturer's instructions (Ultravision Detection System; Labvision), and subsequently were dehydrated and mounted. The same procedures were also performed with appropriate control antibodies

Results and discussion. First, a comparative investigation of MMP-9 and TIMP-1 secretion was performed in cells of the monocyte lineage. MMP-9 secretion is up-regulated more during MGC differentiation than during macrophage differentiation (figure 1A and B). At 72 h, monocyte-derived MGCs secreted 153 ± 2.5 ng/mL MMP-9, compared with 115 ± 3.8 ng/mL MMP-9 during macrophage differentiation (P < .05). Data on MMP-9 secretion were obtained from experiments performed at least 5 times. Such MMP-9 secretion might be attributed to the fact that, to generate MGCs, monocytes must be stimulated with ConA and IFN-γ (reviewed in our description of MGC generation, published elsewhere [9]). It is probable that IFN-g is involved in MGC generation in vivo, particularly in host responses to intracellular pathogens, responses that usually involve IFN-γ secretion. In experiments using mycobacteria with culture supernatants from transformed T cell clones to drive MGC formation, IFN-γ was found to be important [12]. IFN-γ is often considered a proinflammatory cytokine. However, MMP-9 secretion by lipopolysaccharideand Mycobacterium tuberculosis-stimulated human monocytes is down-regulated by IFN-g [13]. Thus, it is unlikely thatMMP- 9 secretion by MGCs is caused by IFN-γ stimulation.

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 1.

Secretion of matrix metalloproteinase (MMP)-9 and tissue inhibitor of metalloproteinase (TIMP)-1. Multinucleated giant cells (MGCs) secrete higher concentrations of MMP-9, but not of TIMP-1, than do macrophages or monocytes, both during and after differentiation. A, Representative zymogram showing MMP-9 secretion during differentiation of monocytes to macrophages and MGCs (in the presence of 5 µg/mL concanavalin A, 1000 IU/mL interferon-γ, and 10% human serum). B, MMP-9 secretion during cell differentiation. Densitometric analysis of MMP-9 secretion measured by zymography is greater during MGC differentiation (black bars) than during macrophage differentiation (gray-shaded bars). C, TIMP-1 secretion during cell differentiation. TIMP-1 secretion measured by ELISA is significantly less during MGC differentiation than during macrophage differentiation. D and E, Representative zymogram (D) and densitometric analysis (E) of MMP-9 secretion by monocytes (white bars), differentiated macrophages (grayshaded bars), and differentiated MGCs (black bars). F, Similarity of TIMP-1 secretion by monocytes (white bars),, macrophages (gray-shaded bars), and MGCs (black bars). All data are the mean of at least 5 independent experiments. *P < .05.

In contrast to MMP-9 secretion, TIMP-1 secretion was decreased during MGC formation, compared with that during macrophage formation (P < .05) (figure 1C). Such a decrease will result in relatively unopposed proteolytic activity and is consistent with data on patients with granulomatous tuberculous meningitis, in whom increased MMP-9 activity has been found to be relatively unopposed by TIMP-1 [14].

After formation, MGCs continue to secrete higher MMP-9 concentrations than do macrophages (at 48 h, 141 ± 6 ng/mL vs. 105 ± 7 ng/mL, resepctively; P < .05) and 3-fold more MMP-9 than do monocytes (P < .05) (figure 1D and E). Absolute MMP-9 concentrations are greater after 48-h culturing of mature MGCs than after 48-h culturing during MGC formation (P < .05) (figure 1B and E). In contrast, TIMP-1 secretion is similar in all 3 cell types (figure 1F). Thus, compared with other monocyte-derived cells, MGCs continue to release increased MMP-9 concentrations, an effect that is not opposed by increased TIMP-1 and that would have the potential to be important in tissue destruction in granuloma.

In view of these observations, we investigated tissue expression of MMP-9 and TIMP-1 in MGCs from lymph nodes from patients with TB. Cells mainly of the monocyte lineage that were adjacent to caseous necrosis stained positive for MMP-9, and MGCs expressed particularly high MMP-9 levels (figure 2). In contrast, TIMP-1 staining was not found in MGCs, although it was identifiable in positive-control breast-cancer biopsy specimens and can be seen at low levels within granuloma [14]. We have previously shown that scattered stromal cells within tuberculous granuloma stain positive for MMP-9 [14]. These TB data are consistent with the results of studies of inflammatory granuloma such as those found in noninfectious giant-cell arteritis [15].

View larger version:

• In this window
• In a new window

• Download as PowerPoint Slide

Figure 2.

Matrix metalloproteinase (MMP)-9 secretion by multinucleated giant cells (MGC) adjacent to caseous necrosis in a tuberculous granuloma. Lymph-node biopsy specimens that, because of clinical indications, had been obtained from patients with tuberculosis were dewaxed, blocked, incubated, and then either (A) specifically stained for MMP-9 or (B) exposed to control antibody, before being visualized by use of the 3,3′-diaminobenzidine Ultravision system (for details, see the text).

The fact that extensive tissue destruction was not evident is consistent with the fact that MMP-9 is involved in other aspects of the immune response to TB, including control of leukocyte migration [3]. MMP-associated tissue damage may be more evident in disseminated or advanced pulmonary TB than in lymph-node disease. It is not possible, on the basis of examination of clinical biopsy specimens taken at a single time point during a chronic infection, to state whether monocytes and nonfused monocyte-derived cells, although they are present in greater numbers, secrete more MMP-9 than do MGCs in granuloma. The data do allow the conclusion that one function of MGCs is MMP-9 secretion and that this occurs at a relatively high level. MGCs may contribute to the generation of a matrixdegrading phenotype in which MMP activity is up-regulated more than TIMP activity, leading to relatively unopposed enzymatic activity. In summary, the present study demonstrates that MGCs may be an important source of MMP-9, although not of TIMP-1, both during and after differentiation. Furthermore, MMP-9 is expressed at high levels in MGCs associated with inflammatory necrosis in human TB.

• Received January 22, 2007.
• Accepted April 16, 2007.

• © 2007 by the Infectious Diseases Society of America