Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Inflammation of the lung may be triggered, and in some cases perpetuated, by stimuli at the airway epithelial surface. Among immunoinflammatory cells in the airways, resident mononuclear cells and activated T lymphocytes play a major part. Lung mononuclear phagocytes consist of at least three populations, functionally and geographically distinct: dendritic cells, lung tissue macrophages (LTM) located in the interstitium, and alveolar macrophages (AM) located on or near the alveolar surface (1). LTM and AM share several functions: defense against pathogens, ingestion and destruction of potential allergens, clearing of particulates, and presentation of antigens to T cells (2). Several investigators have previously reported the regulated production by AM of interstitial collagenase (MMP-1), stromelysin (MMP-3), 92-kD gelatinase (MMP-9), matrilysin (MMP-7), metalloelastase (MMP-12), and tissue inhibitor of metalloproteinases (TIMP)-1, all of which presumably contribute to the physiologic as well as pathologic remodeling of the extracellular matrix (ECM) (3). To date little is known, however, regarding the ability of LTM to produce matrix metalloproteinases (MMPs) and TIMPs. It has been established that during the process of mononuclear phagocyte differentiation from blood monocytes to AM the profile of MMP and TIMP production varies, an increased expression correlating with a more mature cellular phenotype (6). Indeed, immature promonocytic or monocytic cells contain mostly intracellular serine proteinases and produce MMPs in very small quantities (7).

T lymphocytes, present in small numbers in the normal lung interstitium and actively recruited during inflammatory processes of the airways, secrete various cytokines that affect the functions of other infiltrating inflammatory cells and resident tissue cells. We and others have previously observed that cytokines produced by activated T cells can modulate the expression of MMPs and TIMP-1 by monocyte/macrophages and fibroblasts (8). In addition to the release of soluble factors, interest has also focused on the potential role of direct cell-cell contact in mediating biological events. Thus, we reported previously that direct contact between activated human T lymphocytes and monocytes markedly increased the release of cytokines such as interleukin (IL)-1 and tumor necrosis factor (TNF)- (11, 12), as well as metalloproteinases MMP-1 and MMP-9 and the protease inhibitor TIMP-1 by the latter cells (13). T-cell subsets seem to have opposite effects in that T helper (Th) 1 cells preferentially induce IL-1 and Th2 induce the production of IL-1 receptor antagonist (IL-1Ra) upon cell-cell contact with monocytes (14). In murine models, activated Th2 lymphocytes trigger signals that induce the activation of interferon (IFN)--primed macrophages (15, 16). Stout and coworkers (17) also reported subsequently that CD40-CD40 ligand interaction is necessary for T-cell activation of macrophages. Likewise, direct contact between T cells and monocyte/macrophages via CD40-CD40 ligand mediate, at least in part, the induction of MMPs by mononuclear phagocytes (18).

The potential of direct cell-cell interaction to amplify and perpetuate the inflammatory process prompted us to study the influence of activated T lymphocytes on two distinct mononuclear cell populations in the lungs. We found that T cell-macrophage contact failed to induce AM to increase their production of MMPs or TIMP. In contrast, when T lymphocytes were in direct contact with LTM, their expresssion of MMPs and TNF- was massively induced, and part of the induction was dependent on IL-1 and TNF- bound to activated T-cell membranes.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Reagents

Phorbol myristate acetate (PMA) was purchased from Sigma Chemical Co. (St. Louis, MO), dissolved in dimethylsulfoxide and stored in aliquots of 1 mg/ml at 70°C. Purified phytohemagglutinin (PHA), obtained from E.Y. Laboratories Inc. (San Mateo, CA) and dissolved in phosphate-buffered saline (PBS) free of Ca²⁺ and Mg²⁺, was stored in aliquots of 1 mg/ml at 70°C. Bovine serum albumin, phenylmethylsulfonyl fluoride, ethylenediaminetetraacetic acid, and Triton X-100 were obtained from Sigma Chemical Co. RPMI-1640, PBS, fetal calf serum (FCS), penicillin, streptomycin, and L-glutamine were purchased from GIBCO BRL (Paisley, UK).

Cell Preparation and Culture

Human AM were obtained from the uninvolved lung tissue of patients undergoing surgical resection for pulmonary carcinoma. Segments or lobes were lavaged and processed as described previously (8). The adherent macrophages were cultured at a concentration of 1 × 10⁶ cells/ml in the presence of RPMI-1640 medium containing 5% FCS, penicillin (100 U/ml), streptomycin (100 ng/ml), and 2 mM L-glutamine in 24-well cluster plates (catalog no. 3524; Costar Corp., Cambridge, MA).

LTM were obtained by proteolytic dispersion of the uninvolved lung tissue of individuals undergoing surgical resection (19). After 24 h of culture under the same conditions as described previously, supernatant was removed and the adherent cells, which represented LTM, were removed from the plate with a rubber policeman, washed three times with PBS, counted, and plated at a concentration of 1 × 10⁶ cells/ml in the conditions described previously. More than 95% of these cells consisted of mature macrophages with their characteristic appearance being revealed by May-Grünwald-Giemsa and inverted macroscopy. Most contaminating cells such as fibroblasts, lymphocytes, and neutrophils were removed by three washes the day before and three further washes the day they were removed from plates by gentle scraping.

AM were obtained from peripheral lung tissue that was unaffected by lung carcinoma. Although most of our patients were current smokers, 30% of the patients were nonsmokers or former smokers. As far as the latter are concerned, no significant changes in the response of AM or LTM were found.

Preparation of Membranes from Human Peripheral Blood T Lymphocytes and T-cell Lines

T lymphocytes from human peripheral blood (PBTL) were harvested from the buffy coat of healthy donors and processed as previously described (11). The T cells obtained were stimulated, their membranes isolated and stored as described (11).

Membranes were prepared from the HUT 78 T-cell line, a human cutaneous T-cell lymphoma line (American Type Culture Collection, Rockville, MD), by the method of Aderem and colleagues (20). Briefly, HUT 78 cells were incubated at 1 × 10⁶ cells/ml in medium with or without PMA (5 ng/ml) and PHA (1 µg/ml) for 24 h. The cells were then processed as described previously (13). The final membrane preparation obtained was resuspended at 50 × 10⁶ cells/ml of RPMI medium and frozen at 70°C until later use.

Culture of AM and LTM with T-Cell Membranes

AM or LTM were dispensed onto 24-well culture plates at 1 × 10⁶ cells/well in 1 ml of medium. Afterwards, the membranes from the unstimulated membrane of T cells (T_umb) or stimulated membranes of T cells (T_smb) were resuspended in culture medium and added to mononuclear cells to obtain a final ratio equivalent to eight T lymphocytes per one mononuclear cell in each well, a ratio found to be optimal for cocultures in previous studies (11, 13). Controls consisted of medium alone or PMA (5 ng/ml). After 48 h of incubation at 37°C, the conditioned media were collected and stored at 20°C until further analysis. For blockade experiments, IL-1Ra was added to lung macrophages (AM or LTM) in culture medium for 30 min at 37°C, while T cell membranes or HUT 78 cells were incubated with TNF binding protein (TNF-bp) or medium alone for 30 min at 4°C. Preincubated membranes were then added to preincubated macrophages for 48 h so that neither membranes nor macrophages were washed before culture.

MMP and TIMP-1 Determination

Samples of conditioned media were subjected to enzyme-linked immunosorbent assay (ELISA) for the determination of MMP-1, MMP-9, and TIMP-1 as described previously (5, 21, 22). The sensitivity for all protein assays was 10 ng/ml.

Flow Cytometric Analysis

The following antibodies were used for cytometry: anti-CD14 (clone MY4) was purchased from Coulter Clone (Hialeah, FL), anti-human leukocyte-associated antigen-DR (HLA-DR) from Dako A/F (Glostrup, Denmark), anti-B7.1 (CD80) (clone L307.4) from Becton-Dickinson (San Jose, CA), anti-B7.2 (CD86) (clone IT2.2) from Pharmingen (San Diego, CA), anti-CD40 (clone MAB89) and control mouse immunoglobulin (Ig)G were obtained from Immunotech (Marseille, France). The expression of surface antigens by LTM and AM was compared. After isolation of the macrophages as described previously, cells were incubated with the different monoclonal antibodies for 45 min at 4°C. Cells were stained by using a fluorescein isothiocyanate (FITC) isotype-specific goat antimouse IgG (Immunotech). The samples were analyzed on a FACScan (Coulter, EPICS XL-MCL; Becton-Dickinson). Dead cells were gated out based on their light scatter properties. Results are expressed in mean ± standard error of the mean (SEM).

Cytokine Inhibitors

Human recombinant IL-1Ra was a gift from Dr. R. C. Thompson (Synergen, Boulder, CO) and pegylated soluble tumor necrosis factor receptor type I (PEG sTNF-R1), also named pegylated TNF binding protein (PEG TNFbp), a gift from Dr. C. Edwards III (Amgen, Boulder, CO). LTM and AM were incubated in 96 wells/plate with medium alone or IL-1Ra for 30 min at 37°C. Meanwhile, unstimulated or stimulated PBTL or HUT 78 cells were preincubated with medium alone or TNF-bp for 30 min on ice before their addition to the macrophages. After 48 h of culture, supernatants were collected and analyzed for their contents in MMP-9 and TIMP-1. Final concentrations of IL-1Ra and TNF-bp were 1 µg/ml and 10⁸ M, respectively.

Metabolic Labeling and Immunoprecipitation Studies

All samples subjected to immunoprecipitation were conditioned for 24 h in the presence of [³⁵S]methionine. To study T-cell effects, mononuclear cells were exposed to T_umb or T_smb at the ratio described previously or to PMA (5 ng/ml) for a period of 48 h in RPMI-1640 culture medium. Media were then replaced with fresh media free of methionine but containing [³⁵S]methionine for 24 h (8). For immunoprecipitation, polyclonal antisera to human MMP-1, MMP-7, MMP-9, and TIMP-1 were used as reported previously (21, 22). Processed samples were applied to 12% polyacrylamide slab gels and electrophoresis was performed (23). The gels were exposed to Hyperfilm (CEA AB; Amersham, Solna, Sweden) for 24 to 48 h at 70°C.

RNA Preparation and Northern Blot Hybridization

Total cellular RNA from 1 × 10⁷ LTM or AM, cultured with T_umb, HUT-78_umb, T_smb, HUT-78_smb, or PMA, was extracted by the guanidine isothiocyanate method and purified by cesium chloride density gradient centrifugation (24). Total RNA (5 µg) was processed for Northern blot analysis as described previously (13). Membranes were sequentially hybridized at 42°C overnight with ³²P-labeled complementary DNA (cDNA) probes for MMP-1 (25), MMP-9 (8), and TIMP-1 (kindly provided by D. Carmichael, Synergen, Boulder, CO). glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA, provided by R. Pierce (Washington University School of Medicine, St. Louis, MO), was used as control probe. Filters were then exposed to Hyperfilm for 24 to 48 h at 70°C.

By using a densitometer equipped with ImageQuant software (Molecular Dynamics, Sunnyvale, CA), autoradiographs were scanned and quantified, and values were normalized to GAPDH values.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Difference in Response of LTM and AM to Direct Cell-Cell Contact with Activated T Lymphocytes

Under basal conditions, LTM and AM released small but quantifiable amounts of MMP-1, MMP-9, TIMP-1, and TNF- (Table 1). Values are the results of four different experiments involving four individuals, LTM and AM being derived from the same individual in each experiment from whom the PBTL were obtained. Contact between LTM and T_umb for 48 h did not result in a significant change in the basal production of MMP-1, MMP-9, and TIMP-1 (1.2-fold, 1.3-fold, and 1.6-fold, respectively). However, TNF- was significantly induced 34-fold (Table 1, P < 0.01), but when LTM were cultured in the presence of T_smb, a marked increase in production of MMP-1, MMP-9, TIMP-1, and TNF- was observed (5-fold, 3.5-fold, 2.7-fold, and 250-fold, respectively). In the presence of AM, T_umb yielded a moderate release of MMP-9 (2.2-fold), TIMP-1 (3.8-fold; Table 1, *P < 0.05), and TNF- (20-fold; Table 1, P < 0.01) but failed to affect the production of MMP-1. The moderate increase in MMP-9 and TIMP-1 on AM could be due to the effect of phagocytosis of the membranes by macrophages or to stimulating molecules present at low levels in unstimulated cells. In contrast to LTM, AM exposed to T_smb showed no significant stimulation of MMP or TIMP-1 expression (1.5-fold for MMP-1, 2.7-fold for MMP-9, and 2.7-fold for TIMP-1, respectively) compared with cells incubated with T_umb (see P values in Table 1). Nevertheless, TNF- production was strongly induced (100-fold) in AM cultured with T_smb as compared with T_umb.

Time Course of MMP-9 and TIMP-1 Production by LTM versus AM in Response to T-Cell Membranes

To determine to what extent the duration of contact between T_umb and LTM or AM affected the induction of MMPs and TIMP-1, macrophages were exposed to T-cell membranes (Tmb) for various periods of time. LTM in contact with T_smb released significant amounts of MMP-9 after 48 h (Figure 1A). The release of TIMP-1 also took place 48 h after contact with T_smb (Figure 1B). As shown previously (Table 1), AM contrary to LTM failed to exhibit any increased production of MMP-9 (Figure 1C) or TIMP-1 (Figure 1D) at any time point examined, whether in the presence of T_umb or T_smb. The occasional reduction below levels is also seen with T_umb. Kinetics of MMP-1 production were analyzed in only one experiment of three performed due to the small number of cells available. The production of MMP-1 was markedly increased when LTM were in contact with T_smb compared with T_umb, but the enzyme was undetectable in media from AM in contact with either unstimulated or stimulated T cells (data not shown).

View larger version (20K): [in this window] [in a new window]   Figure 1.   Time course of MMP-9 and TIMP-1 production by mononuclear cells exposed to T-cell membranes (Tmb). Either LTM (A and B) or AM (C and D) at 1 × 106 cells/ml were cultured in the presence of Tumb or Tsmb for the times indicated. Culture supernatants were tested by ELISA for contents in secreted MMP-9 (A and C) and TIMP-1 (B and D). MMP-9 levels were 1,497 ± 164 and 585 ± 158 ng/ml on LTM plus PMA and on AM + PMA, respectively. Values are mean ± SEM from three different experiments involving three different individuals.

Effect of Tmb Concentrations on the Capacity to Induce MMP-9 and TIMP-1 in LTM versus AM

Production of MMP-9 and TIMP-1 by each LTM and AM was determined in the presence of increasing amounts of T_umb or T_smb. Tmb were added to mononuclear cells for 48 h, yielding a final ratio equivalent to 2, 4, 8, and 16 T lymphocytes per 1 mononuclear cell/well. Consistent with results shown in Table 1 and Figure 1, LTM production of MMP-9 and TIMP-1 was markedly stimulated by T_smb (Figures 2A and 2B). Whereas the expression of MMP-9 was optimal when using T_smb in amounts equivalent to eight T cells per one LTM (Figure 2A), the stimulation of TIMP-1 required lesser amounts of T_smb, the magnitude of response being somewhat attenuated by higher concentrations of T_smb (Figure 2B). In contrast to LTM, AM did not respond to T_smb at whatever concentration in terms of MMP-9 or TIMP-1 production (Figures 2C and 2D). These findings are further proof of the dichotomous response of LTM and AM to T_smb versus T_umb exposure. Kinetics of MMP-1 production were not analyzed because of the small number of cells available and the low levels of produced enzyme.

View larger version (20K): [in this window] [in a new window]   Figure 2.   Effect of ratio of T-cell membranes to macrophages on the production of MMP-9 and TIMP-1. LTM (A and B) or AM (C and D) were cultured for 48 h at 1 × 106 cells/ml in the presence of increasing concentrations of Tmb. Supernatants were tested by ELISA for contents in secreted MMP-9 (A and C) and TIMP-1 (B and D). MMP-9 levels were 263 and 1,552 ng/ml on LTM plus PMA and AM plus PMA, respectively. Values are mean ± SEM from three different experiments on three different individuals.

Modulation by Tmb of De Novo Synthesis of MMPs and TIMP-1 in LTM but not AM

To gain insight into the intracellular processes governing the production of MMPs and TIMP-1 during cell contact with Tmb, LTM or AM were metabolically labeled with [³⁵S]methionine. Labeled proteins were immunoprecipitated with antisera specific for MMP-1, MMP-9, MMP-7, and TIMP-1. Data on MMP-7 (matrilysin) were obtained only by this method because ELISA was not available. T_smb stimulated LTM for the biosynthesis of MMP-1, MMP-9, and MMP-7 as well as TIMP-1 (Figure 3A). T_umb were only slightly stimulatory. In contrast, when AM were tested for de novo protein synthesis, production of MMP-1, MMP-9, and MMP-7 was not stimulated to a greater extent by T_smb than by T_umb (Figure 3B). PMA was used as positive control in these experiments. It is noteworthy that in this experiment, addition of T_smb to AM actually tended to decrease the synthesis of TIMP-1 as compared with T_umb (Figure 3B). This observation was confirmed in another similar experiment, and this phenomenon will be the subject of further studies. Taken as a whole, these data largely mirror protein release in culture supernatants (Table 1).

View larger version (45K): [in this window] [in a new window]   Figure 3.   Effect of cell-cell contact on de novo biosynthesis of MMP-1, MMP-9, MMP-7, and TIMP-1. Either LTM (A) or AM (B) were cultured in medium alone or were exposed to Tumb, Tsmb, or PMA for 48 h. Cells were metabolically labeled with [35S]methionine for the last 24 h of incubation as described in MATERIALS AND METHODS. Labeled proteins were immunoprecipitated with antiserum to MMP-1, MMP-9, MMP-7, and TIMP-1 as indicated.

Pretranslational Control of Macrophage Expression of MMP-9 and TIMP-1 after Contact with Tmb

To further assess at which level T cells exert control on phagocytic cells, Northern blot analysis was performed. LTM and AM were cultured for 24 h in the presence of T_umb, T_smb, or PMA before harvesting total RNA. Both LTM and AM expressed basal messenger RNA (mRNA) levels of MMP-1, MMP-9, and TIMP-1 (Figure 4), and addition of T_umb slightly stimulated the expression of MMP and TIMP-1 mRNA. However, after exposure to T_smb, both MMP and TIMP-1 mRNA were increased in LTM while remaining partly unmodified in AM, when compared with the effect of T_umb. PMA used as positive control does not stimulate MMP and TIMP-1 expression in AM. Another experiment was performed and confirmed that on LTM, PMA stimulates both MMP-1 and MMP-9 mRNA expression while slightly decreasing TIMP-1 mRNA levels on AM (data not shown). Once again, neither PMA, T_smb, or T_umb stimulated markedly MMP-9 expression in AM. These data confirm those obtained previously for protein secretion and protein biosynthesis, and indicate that the regulatory effects induced on LTM by T_smb occur at a pretranslational level.

View larger version (32K): [in this window] [in a new window]   Figure 4.   Pretranslational regulation of MMP-1, MMP-9, and TIMP-1. Human LTM (A and B) and AM (C and D) from a single donor were cultured for 24 h alone or in the presence of T-cell membranes derived from HUT 78 cells (Tumb, Tsmb) or PMA (5 ng/ml). Total RNA was harvested as described in MATERIALS AND METHODS, and subjected to Northern blot analysis using cDNA probes for MMP-1, MMP-9, TIMP-1, and GAPDH. For each condition, 107 cells were used, and each lane was loaded with 5 µg of total RNA. Results are from a single experiment representative of one other experiment yielding similar results. (B and D) The rate of MMPs and TIMP-1 normalized to GAPDH after densitometric scanning quantification. A.U. = arbitrary units.

Analysis of Cell-Surface Molecules on Macrophages

Because LTM and AM are macrophages obtained from two locations in the lung tissue, we analyzed their expression of cell-surface molecules in order to define phenotypic differences between these two macrophage populations. After isolation of both cell types as described in MATERIALS AND METHODS, cells were incubated with various markers, revealed with FITC-conjugated antibody, and fluorescence-activated cell sorter analysis was performed. Among the various cell-surface molecules studied, only the expression of human leukocyte-associated antigen-DR (HLA-DR) molecules appears to be significantly different in LTM and AM (Table 2). In Figure 5, the small difference in mean fluorescence intensity (delta means) for CD14, CD40, CD80, or CD86 on AM or LTM compared with controls is in the same order of magnitude as the SEM of the mean fluorescence of the controls (Table 2) with the exception of HLA-DR, where delta means are increased by several orders of magnitude. These data suggest that both LTM and AM have many phenotypic characteristics in common and that the only difference demonstrated so far is their surface expression of HLA-DR molecules, which could be consistent with a better function as antigen-presenting cells for LTM.

View larger version (19K): [in this window] [in a new window]   Figure 5.   Determination of antigen expression on the surface of AM and LTM. AM and LTM were obtained and purified as described in MATERIALS AND METHODS, and expression of surface antigen by AM and LTM was compared. Cells were incubated with the different monoclonal antibodies for 45 min at 4°C, then stained by using an FITC isotype-specific goat antimouse IgG, and finally analyzed on a FACScan. This histogram is representative of four experiments performed on different individuals (see Table 2). For each experiment, AM and LTM were obtained from the same individual. MF means the difference of the mean fluorescence intensity of the whole cell population stained either with the control antibody or a given marker expressed for each analysis.

Involvement of Cell-Surface-Associated Cytokines on Contact Between T Cells and Lung Macrophages

To assess the role of T-cell membrane-bound cytokines in the induction of MMPs during T cell-macrophage contact, specific inhibitors of IL-1 and TNF- were used. Tmb (T_u and T_s) and lung macrophages (AM and LTM) were exposed to TNF-bp and IL-1Ra, respectively, before coincubation. From Figure 6, it is evident that on contact with membrane from stimulated PBTL (Figure 6A), the production of MMP-9 by LTM is markedly decreased by the addition of TNF-bp (44% of inhibition), whereas this cytokine had no effect on the level of MMP-9 produced by LTM in contact with stimulated HUT 78 cells (Figure 6B). When used alone, IL-1Ra did not show any significant inhibitory effect but added to TNF-bp, the inhibition of MMP-9 was marked upon contact with stimulated PBTL and HUT 78 cells (60 and 66%, respectively). Of note, TNFbp added to IL-1Ra exerted an inhibitory effect on MMP-9 production due to unstimulated T cells (60% of inhibition on LTM + PBTL contact, and 58% on LTM + HUT 78 cell contact) (Figures 6A and 6B). However, TNF-bp alone only had an inhibitory effect on LTM + HUT 78 cell contact (28%), but very little on LTM + PBTL contact (6%). As pointed out in this report, stimulated T cells have no effect on the production of MMP-9 by AM, making it difficult to analyze the potential inhibitory effect of TNF-bp or IL-1Ra (Figures 6C and 6D). When TNF-bp or TNF-bp and IL-1Ra were added to AM, the production of MMP-9 decreased by 25 to 45% in all culture conditions. However, as concentration of MMP-9 is low (< 100 ng/ml), small variations could also be due to the decreased sensitivity of the ELISA at such low values. Both cytokine inhibitors do not have any effect on the production of TIMP-1 during PBTL-LTM contact or T cell- AM contact. Only when added to HUT 78 cells was TNF-bp able to inhibit 25% of the production of TIMP-1 by LTM (data not shown). Moreover, when TNF- production was measured by ELISA in this blockade experiment, IL-1Ra alone or added to TNF-bp did not decrease the production of TNF- in LTM, whereas IL-1Ra alone or added to TNF-bp had a strong inhibitory effect (95%) on the production of TNF- by AM (data not shown). These data suggest that LTM and AM use different signaling pathways to produce cytokines when in contact with activated T cells.

View larger version (22K): [in this window] [in a new window]   Figure 6.   Effect of cytokine inhibitors on T cell-lung macrophage contact. PBTL and HUT 78 cells were stimulated and processed as described in MATERIALS AND METHODS. IL-1Ra was added to the lung macrophages in culture medium 30 min before addition of T-cell membranes. TNF-bp was added to the membrane preparation of T cells for 30 min at 4°C, then the reaction mixture was added to macrophages. Final reaction was incubated for 48 h. MMP-9 levels (ng/ml) on LTM alone were 216: on LTM + PMA, 1,383; on AM alone, 95; and on AM + PMA, 147.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

This study demonstrates that the direct contact between T lymphocytes and lung macrophages may play a major part in the pathogenesis of lung inflammation by mediating the induction of both proinflammatory cytokines and MMPs. Among several potential T-cell surface factors involved in the stimulation of lung macrophages, both membrane- associated IL-1 and TNF- are potent triggers of the stimulation observed upon cell-cell contact. Direct contact was studied by using plasma membrane preparations of T cells to avoid contamination of cytosolic or intracellular molecules due to leakage. The striking observation is the existence of a fundamental difference in the capacity of macrophages originating from interstitial tissue versus alveolar space to produce MMPs on contact with activated T cells, presumably reflecting a distinct pattern of responsiveness and biological significance. Only LTM are stimulated by membranes of activated T cells to express MMPs. Thus, LTM may play a pivotal role in tissue remodeling because they can mediate tissue destruction directly by secreting their own MMPs or indirectly by releasing cytokines such as TNF- that in turn induce the production of MMPs by fibroblasts in the vicinity (26, 27).

In contrast, AM, which are located mostly in the alveolar spaces of the lung, do not secrete more MMPs after direct contact with activated T cells than with nonactivated T cells; however, TNF- is increased in these conditions. Of importance, on comparing the effects of T_smb and T_umb, the production of MMP-1, MMP-9, TIMP-1, and TNF- was significantly increased on LTM, whereas on AM only TNF- was increased. This suggests that AM play a more important role in maintaining the inflammatory state rather than in directing the destructive and remodeling process when in direct contact with activated T cells. Even if phenotypic analysis with the reagents used did not reveal any difference between these two cell populations, the differential responses in MMP production of LTM and AM emphasize that these are two distinct functional types of lung macrophages whose variant states of differentiation could result in specific patterns of responsiveness and biological behavior (6).

As AM have previously been shown to produce MMPs and inhibitors through the action of various stimuli (6), it seems reasonable to postulate that the difference in the capacity of the two cell populations to respond to T cells lies mainly in the manner in which they interact with T cells. In previous reports, we (8) and others (9, 10) demonstrated that AM are able to respond to soluble products expressed by T cells (i.e., soluble cytokines), whereas in the present study we show that AM are not able to produce MMPs and TIMP-1 when stimulated by direct T-cell contact. The decrease in TIMP-1 expression by stimulated T cells on AM has not been observed on other monocytic cells (13). According to a recent report published by our group (28), the expression of TIMP-1 by dermal fibroblasts or synoviocytes varies depending on the time of T-cell stimulation. This suggests that T cells stimulated for short periods of time (1 to 4 h) are able to induce the production of TIMP-1 in fibroblasts or synoviocytes, whereas T cells stimulated for long periods of time (24 to 48 h) lose the capacity to modify TIMP-1 production. These effects are probably due to the distinct modifications in the expression of surface factors depending on the time of stimulation. Because T cells were stimulated for long periods of time (48 h) in our study, they probably lost the capacity to modify TIMP-1 production by AM, which is in keeping with the report by Burger and coworkers (28). Owing to the difficulty in obtaining sufficient amounts of cells from human sources we were not able to perform rigorous time course experiments. In the past (11), we observed that T lymphocytes stimulate monocytes through the action of multiple cell-surface factors: these studies revealed that the IL-1 production by the monocytic cell line THP-1 was partially inhibited by antibodies to CD11b and to CD11c (59 and 50% inhibition, respectively) (11) and by antibodies to CD69 (33% inhibition) (12; data not shown). Interestingly, in the same experiments, none of these antibodies was able to inhibit the production of MMP by T cell-activated THP-1 cells. The presence of cell-associated cytokines, such as TNF- (29) or IL-1 (30), has been described and may represent important stimulatory signals. TNF- expressed on membranes of activated T cells has been shown to induce production of IL-10 by monocytes (31) or the expression of integrins and cytokines by endothelial cells (32). Moreover, detectable levels of membrane-associated IL-1 and TNF- have been measured on membrane preparation of PHA/PMA-activated PBTL (28). Recently, two of our reports (28, 33) have pointed out that those cell-associated cytokines play a major part during T cell and fibroblast interaction. The first report suggests that the production of MMP-1 and prostaglandin E₂ by synoviocytes or fibroblasts during direct contact with T cells is mainly dependent on membrane-associated TNF- and IL-1 (28). In contrast, during T-cell and monocyte interaction, neither the antagonist to TNF nor that to IL-1 had significant effects (12, 33). According to the previous report (28), the association of IFN-, TNF-, and IL-1 on membranes of T cells results in the potent inhibition of collagen I production by fibroblasts (34). We also analyzed the involvement of membrane-associated TNF- and IL-1 on activation of macrophages during interaction with T cells. The blockade of TNF-, particularly in association with IL-1Ra inhibitor, decreased by > 50% the production of MMP-9 by LTM, without specific activity on AM. However, as the inhibition is not complete, unknown cell-surface factorsnot characterized yetmay be required for T cells to stimulate the production of MMPs by mononuclear phagocytes. A possibility not to be ruled out is that some of the inhibitory effects of cytokine antagonists may not exclusively be due to the effect on membrane-associated cytokines but also on cytokines produced in an autocrine fashion by macrophages. However, this possibility does not seem too likely because in the presence of T_smb, TNF- is produced to an equal extent by AM and LTM (Table 1). Others observed in a murine model that soluble products of T lymphocytes (i.e., IFN-) are important for macrophage activation (35), and TNF- or CD40 ligand expressed on T-cell membrane contact contributes to stimulating macrophage functions upon direct cell-cell (15, 35).

Thus, the important role in mediating contact response of certain receptors and their ligand, as well as membrane-associated cytokines, may explain the differential behavior of LTM and AM when in direct contact with T cells. We suggest that LTM have cell-surface glycoproteins capable of interacting with the glycoproteins of activated T cells, whereas AM do not express such cell-surface glycoproteins, responsible for MMP induction. Other studies have shown that as a result of biological maturation, AM, contrary to their precursor monocytes and LTM, tend to cease to express several surface markers (36). Like other markers, CD14 is considerably reduced or altogether absent on normal AM, unlike monocytes (36). Contrary to normal AM, which express little or no CD80 (B7-1), CD86 (B7-2), or CD40, AM of patients with active sarcoidosis do, as do mature dendritic cells (37). LTM did not differ significantly from AM with regard to these markers, except for the expression of HLA-DR; thus, with the latter exception, they do not express moleculespresent on mature dendritic cellsthat are likely to interact with T cells. It is, however, possible that other glycoproteins are present on LTM, which may explain why the response of LTM and AM, when in contact with T_umb or T_smb, was not identical.

To ascertain the viability of AM and their potential to be activated, each experiment was performed in the presence of PMA, known to induce cytokine and MMP production on macrophages. We consistently observed that PMA stimulated, at various levels, the expression of MMPs and TIMP-1 by AM (Figures 3 and 4), contrary to T_smb cells, which were unable to induce such a response. Further demonstration of AM viability was their capacity for full induction of TNF- expression in the presence of stimulated T cells. Moreover, most of our observations are based on HUT 78 T-cell lines, used as a substitute for lymphocytes derived from peripheral blood or tissue-resident lymphocytes. Because previous studies have confirmed that in direct contact with monocytes or dermal fibroblasts, T cells or T-cell clones have similar biological functions to those of T-cell lines (13, 28, 33), we postulate that the latter have similar biological effects on lung macrophages. So far, we have performed two experiments with freshly isolated peripheral blood T lymphocytes confirming their similar biological effect on lung macrophages (Figure 6).

ECM degradation, which occurs during tissue remodeling in inflammatory processes, is tightly regulated, matrix remodeling being mediated largely by secreted proteins of the MMP and TIMP families. The finding that direct cell-to-cell contact between activated T cells and mononuclear phagocytes induces MMPs, which facilitate matrix breakdown, may be highly relevant to the understanding of immunopathologic processes. The key finding regarding the functional aspect of differential mature macrophage phenotypes is novel and of clinical importance. The biological significance can be that cell-cell interaction is of greater relevance at the tissue level than in the extracellular space (i.e., alveolar space). Moreover, the capacity of restricted, specific populations of lung macrophages to participate in certain aspects of these responses may start to explain the diversity of the underlying pathobiology of lung inflammation, injury, and repair. Interfering in a specific manner at the level of the direct contact between T cells and distinct subsets of mononuclear phagocytes may be a useful addition to the therapeutic arsenal.