Monocytes/macrophages play a critical role in the initiation and progression of a variety of glomerulonephritides. We sought to define the interactions between physiologically activated human monocytes and glomerular mesangial cells (MC) by employing a cell culture system that permits the accurate assessment of the contribution of soluble factors and cell-to-cell contact. Human peripheral blood monocytes, primed with IFN-γ and GM-CSF, were activated with CD40 ligand (CD40L) or TNF-α and cocultured with MC. CD40L-activated monocytes induced higher levels of IL-6, monocyte chemoattractant protein-1 (MCP-1) and ICAM-1 synthesis by MC. Separation of CD40L-activated monocytes from MC by a porous membrane decreased the mesangial synthesis of IL-6 by 80% and ICAM-1 by 45%, but had no effect on MCP-1. Neutralizing Abs against the β2 integrins, LFA-1 and Mac-1, decreased IL-6 production by 40 and 50%, respectively. Ligation of mesangial surface ICAM-1 directly enhanced IL-6, but not MCP-1, production. Simultaneous neutralization of soluble TNF-α and IL-1β decreased MCP-1 production by 55% in membrane-separated cocultures of MC/CD40L-activated monocytes. Paraformaldehyde-fixed CD40L-activated monocytes (to preserve membrane integrity but prevent secretory activity), cocultured with MC at various ratios, induced IL-6, MCP-1, and ICAM-1 synthesis by MC. Plasma membrane preparations from activated monocytes also induced mesangial IL-6 and MCP-1 synthesis. The addition of plasma membrane enhanced TNF-α-induced mesangial IL-6 production by ∼4-fold. Together, these data suggest that the CD40/CD40L is essential for optimal effector function of monocytes, that CD40L-activated monocytes stimulate MC through both soluble factors and cell-to-cell contact mediated pathways, and that both pathways are essential for maximum stimulation of MC.
Glomerular and interstitial macrophage accumulation is a feature of all glomerulonephritides (GN)2 (1). Macrophages accumulate in the kidney from the peripheral blood in response to a variety of chemokines and adhesion molecules (2, 3). Local macrophage proliferation, presumed to be driven by local growth factors, has also been reported in areas of tissue damage among the more aggressive forms of human GN (4).
Evidence implicating macrophages directly in renal pathology arise from animal studies whereby depletion of macrophages inhibits the induction or progression of renal injury (5). Studies in humans have suggested a correlation of the extent of glomerular macrophage infiltration with glomerular hypercellularity, histologic damage, proteinuria, and loss of renal function (6). Macrophages are also thought to play a vital role in crescent formation and tissue injury in crescentic GN (7).
Macrophages exhibit critical regulatory, as well as effector, functions during inflammatory responses. They communicate with adjacent inflammatory and mesenchymal cells through secretion of soluble factors or by direct cell-to-cell contact (8, 9). In vitro studies have shown that macrophage-derived IL-6, platelet-derived growth factor, and TGF-β promote mesangial expansion (10, 11). Macrophage-derived proinflammatory cytokines such as TNF-α and IL-1β induce mesangial cells (MC) to secrete chemokines, such as monocyte chemoattractant protein-1 (MCP-1), and express adhesion molecules, such as ICAM-1, which facilitate further infiltration and accumulation of macrophages into glomeruli (12, 13, 14).
In contrast with soluble factors, the role of direct cell-to-cell contact in mediating tissue injury by monocytes has received little attention. MC are known to interact with extracellular matrix through a variety of integrins (15, 16, 17), but whether they also interact with local macrophages has not been adequately investigated. This is particularly important in view of data suggesting that direct cell-to-cell interactions mediate important biologic effects, which, in some cases, are distinct from those mediated by soluble factors, and that these interactions may have a pathogenic role in a variety of inflammatory diseases (18, 19, 20, 21).
In the present study, we sought to critically evaluate potential mechanisms whereby macrophages might contribute to renal pathology through their interactions with resident renal cells. We employed an in vitro experimental system with several unique features. First, we used cytokines of likely importance in the pathogenesis of GN (GM-CSF, IFN-γ) to prime peripheral blood human monocytes (22, 23, 24, 25). Second, to activate primed monocytes, we used soluble CD40 ligand (CD40L). CD40L is a membrane glycoprotein that is transiently expressed on the surface of activated CD4+ T cells. Interaction of CD40L with its counterreceptor CD40 (which is constitutively expressed by macrophages) is essential for the development of the effector function of macrophages and has been implicated in the pathogenesis of autoimmune GN, such as lupus nephritis (26, 27, 28, 29). Third, we employed coculture experiments that used a semipermeable membrane to physically separate cells and thereby delineate the relative contribution of cell-to-cell contact. These experiments were extended using paraformaldehyde-fixed monocytes (which preserves membrane integrity but prevents soluble factor secretion), together with purified plasma membrane preparations, to further characterize the relative contributions of direct cell-to-cell contact and soluble factors. Our data suggest that CD40L-activated monocytes amplify the glomerular inflammatory response through both soluble and cell contact-dependent mechanisms and identify key molecules involved in these interactions.
Materials and Methods
Abs, cytokines, and CD40L trimer
The following Abs, recombinant cytokines, and proteins were used in these experiments: monoclonal mouse anti-human TNF-α (R&D Systems, Minneapolis, MN); monoclonal mouse anti-human IL-1β (Genzyme, Cambridge, MA); monoclonal mouse anti-human CD11a/LFA-1 (PharMingen, San Diego, CA); monoclonal mouse anti-human CD11b/Mac-1 (PharMingen); monoclonal mouse anti-human CD49d/very late activation Ag-4 (VLA-4) (Immunotech, Miami, FL); monoclonal mouse anti-human ICAM-1 (Dako, Carpinteria, CA); R-PE-conjugated F(ab)2 fraction of goat anti-mouse Igs (Dako); recombinant human IFN-γ (Genzyme); recombinant human GM-CSF (R&D Systems); recombinant human TNF-α (R&D Systems); and trimeric human CD40L/leucine-zipper fusion protein (Immunex, Seattle, WA).
Human MC were established and characterized as reported previously (30), and cultured in Waymouth’s medium (Life Technologies, Grand Island, NY) supplemented with 17% heat-inactivated FBS (HyClone Laboratories, Logan, UT), 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mM l-glutamine, 2 mM sodium pyruvate, 1% (v/v) nonessential amino acids, and 26 μg/ml of bovine insulin (all Life Technologies). Four independent cell lines were employed in passages 5–10.
Monocyte purification and culture
Purified monocytes were isolated using counterflow centrifugal elutriation from normal human leukocyte concentrates obtained from the Department of Transfusion Medicine, National Institutes of Health by automated apheresis. Monocyte purity was assessed by FACS analysis using fluoresceinated mAb (Becton Dickinson, Mountain View, CA) with specificity for the following cell surface markers: CD14 (monocytes), CD3 (T cells), and CD19 (B cells). The monocyte fraction routinely consisted of >80% CD14-expressing cells, <0.5% CD19-expressing cells, and <3% CD3-expressing cells. Purified monocytes were cultured in RPMI 1640 (Life Technologies) supplemented with 10% heat-inactivated FBS, 2 mM l-glutamine, 100 IU/ml penicillin, and 100 μg/ml streptomycin.
Monocytes were cultured at 2 215’5f 106/ml for 24 h in 10% FBS/RPMI 1640 medium in the presence or absence of IFN-γ (500 U/ml) and GM-CSF (10 ng/ml) in a 24-well or 6-well plate. Preliminary experiments established these concentrations of cytokines as optimal. Subsequently cells were cultured with or without 3 μg/ml of trimeric CD40L or 10 ng/ml of TNF-α for 24 h. Supernatants were collected and stored for TNF-α, IL-1β, and IL-12 synthesis determination, while activated monocytes were washed three times and resuspended to a final concentration of 1, 5, and 20 215’5f 105/ml in Waymouth’s medium containing 17% FBS. In some experiments, monocytes were fixed for 2 h in 1% paraformaldehyde (Electron Microscopy Sciences, Washington, PA) in PBS at 4°C with gentle agitation; this preserves membrane integrity but prevents secretory activity. MC were harvested at 80% confluence and cultured in 24-well plates at 6 215’5f 104 cell/0.6 ml/well in triplicate for 16 h. Then, 0.6 ml of prepared activated monocytes were added to MC. In some experiments, identical parallel cultures were established in which monocytes were separated from MC by a 0.4-μm pore semipermeable membrane (Biocoat, Falcon, Becton Dickinson Labware, Bedford, MA), while sharing the same medium. After 24 h, supernatants were harvested and stored at –20°C for IL-6 and MCP-1 determination, and cells were harvested and subjected to ICAM-1 expression estimation by FACS analysis.
Inhibition of cytokine synthesis by Ab
Abs against CD11a/LFA-1, CD11b/Mac-1, CD49d/VLA-4, TNF-α, IL-1β, or IgG1 of irrelevant specificity (Sigma, St. Louis, MO) were incubated at 10 μg/ml with activated monocytes for 30 min at room temperature. Cells were added to MC as described before. Cytokine synthesis after 24 h was determined by ELISA.
Membrane preparation from monocytes
Crude plasma membrane was prepared as described before (19). Briefly, monocytes were disrupted by sonication (five 5-s bursts of 90 W each) in PBS containing 0.68 M sucrose, 200 mM PMSF, and 5 mM EDTA. The lysate was centrifuged for 15 min at 4000 215’5f g to discard nuclei and intact cells. The supernatant was centrifuged for 45 min at 100,000 215’5f g, and the pellet containing the membrane fraction was resuspended at the theoretical concentration of 20 215’5f 106 cell equivalent/ml in Waymouth’s medium.
MC were harvested and cultured in 96-well plates at 1 215’5f 104 cell/0.1 ml/well in triplicate for 16 h. Subsequently, 0.1 ml of prepared medium, including 10 μg/ml of anti-ICAM-1 Ab and 6.7 μg/ml of F(ab′)2 of goat anti-mouse IgG, Fcγ specific, were added to cross-link mesangial ICAM-1. An irrelevant IgG was used as control instead of anti-ICAM-1 Ab. After 24 h incubation, supernatants were harvested and stored at –20°C for IL-6 and MCP-1 determination.
Determination of IL-6 and MCP-1 production and mesangial ICAM-1 expression
IL-12 p70, IL-1β, IL-6, and MCP-1 were measured by ELISA (Endogen, Woburn, MA). ICAM-1 expression was determined by FAC analysis. Briefly, MC were harvested with trypsin/EDTA, and 1 215’5f 105 of cells were labeled by indirect immunofluorescence using anti-human ICAM-1 and R-PE-conjugated F(ab′)2 goat anti-mouse Igs. ICAM-1 expression was analyzed using a Becton Dickinson FACSort running Lysis II. MC were carefully gated using forward scatter and side scatter to minimize the contamination by monocytes. In preliminary experiments, the extent of contamination by monocytes labeled with FITC-conjugated anti-CD14 was consistently <5%. The significant difference in the profile of the cells by forward scatter and sice scatter between MC and monocytes made it possible to distinguish MC from monocytes in subsequent experiments without labeling monocytes.
Statistical significance was determined by Student’s t test. A value of p < 0.05 was considered to represent a statistically significant difference between group means.
The effect of trimeric CD40L on IL-1β, TNF-α, and IL-12 production by primed monocytes
We first optimized the activation of peripheral blood-derived monocytes by measuring cytokine synthesis after various stimuli. Both IFN-γ and GM-CSF are known to activate monocytes and independently up-regulate CD40 expression (31). In preliminary experiments, a combination of IFN-γ and GM-CSF was much more effective in priming monocytes than either protein alone, as judged by IL-12 synthesis (data not shown). Primed monocytes were further incubated in the presence or absence of CD40L or TNF-α. After 24 h, supernatants were harvested to determine IL-1, TNF-α, and IL-12 by ELISA. CD40L induced high levels of IL-1β, TNF-α, and IL-12 p70 (Fig. 1, A–C) production by primed monocytes. Primed monocytes without CD40L produced low levels of IL-1β but failed to produce detectable level of TNF-α and IL-12. Although TNF-α-activated monocytes produced comparable levels of IL-1β with CD40L-activated monocytes, they failed to secrete significant amounts of IL-12 p70. Stimulation of monocytes with IFN-γ, GM-CSF, TNF-α, or CD40L alone failed to induce detectable levels of IL-1β, TNF-α, and IL-12 production. Increasing TNF-α to 50 ng/ml or CD40L to 15 mg/ml did not lead to further synthesis of IL-1β, TNF-α, or IL-12 p70 (data not shown). Together, these data indicate that CD40L activates monocytes to produce inflammatory cytokines such as TNF-α and IL-1β. The effect of CD40L on monocyte activation is distinct from that of TNF-α because only the former induces IL-12 production.
CD40L-activated monocytes induce IL-6, MCP-1, and ICAM-1 synthesis by MC
We next examined whether CD40L-activated monocytes could, in turn, stimulate MC. IFN-γ/GM-CSF-primed monocytes were stimulated with CD40L or TNF-α for 24 h, then harvested, washed, and cocultured with MC, which had been precultured to facilitate adherence. After 24 h, IL-6 and MCP-1 production was determined by ELISA, and ICAM-1 expression was determined by FACS. In preliminary studies, we established the optimal macrophage:MC ratio and duration for coculture (not shown). Monocytes activated with CD40L consistently induced higher levels of IL-6, MCP-1, and ICAM-1 synthesis by MC than did monocytes activated with other stimuli, including TNF-α (Fig. 2, A–C and Fig. 3). Production of TGF-β, an important growth factor for matrix expansion that is constitutively secreted in vitro by cultured MC (10), was not affected by adding activated monocytes (data not shown). Similarly, platlet-derived growth factor, a potent mitogen for MC (10), or RANTES, a chemokine produced by MC (32), were not induced during coculture (data not shown). Thus, CD40L-activated monocytes enhanced mesangial synthesis of IL-6, MCP-1, and ICAM-1 but not of TGF-α, platlet-derived growth factor, and RANTES. Although in this experiment it is difficult to precisely evaluate the relative contribution of the two cell types to IL-6 and MCP-1 production, we assumed that the major source of cytokines produced are MC because 1) in the absence of MC, monocytes activated maximally by IFN-γ, GM-CSF, and rCD40L failed to induce significant amounts of IL-6 and MCP-1 (Fig. 2, A and B), and 2) as shown in Fig. 2,C, ICAM-1 expression on MC parallels the up-regulation of IL-6 and MCP-1 in Fig. 2, A and B.
Cell-to-cell contact induce IL-6 and ICAM-1 expression by MC
We next investigated the relative contribution of soluble factors and cell-to-cell contact in MC activation by monocytes. Primed monocytes were stimulated with CD40L and cocultured with MC as previously described at a ratio of 1:1. In parallel cultures, activated monocytes were separated from MC by a porous (0.4 μm) membrane in otherwise identical conditions. Twenty-four hours later, supernatants were collected for cytokine determination and cells were harvested for ICAM-1 determination by FACS. Separation of monocytes from MC markedly decreased the synthesis of IL-6 (Fig. 4,A) and ICAM-1 (Fig. 4,C), but had little effect on MCP-1 synthesis (Fig. 4,B). The obligatory contribution of cell-to-cell contact was calculated (Fig. 4 D) and was found to be 80% for IL-6 and 45% for ICAM-1 enhancement. These data clearly indicate that direct cell-to-cell contact is essential for enhancement of mesangial IL-6 production by CD40L-activated monocytes, whereas it is not necessary for MCP-1 production.
β2 integrin/ICAM-1 pathway contributes to enhanced MC IL-6 production
We next sought to identify the cell-surface molecules involved in these interactions between activated monocytes and MC. Previous studies have suggested that β2 integrins such as LFA-1 and Mac-1 may play critical roles in interactions between 1) monocytes and endothelial cells and 2) monocytes and MC (13, 33). CD40L-activated monocytes were preincubated with neutralizing mAbs specific for LFA-1 and Mac-1 or with irrelevant control Ig. After thorough washing, monocytes were added to MC and cultured for 24 h. Blockade of LFA-1 and Mac-1 decreased IL-6 production by 50 and 60%, respectively (Fig. 5 A). No additive effect was observed when LFA-1 and Mac-1 were simultaneously blocked. Inhibition of VLA-4/VCAM-1, another important pathway for monocyte adhesion (34), by neutralizing Ab against VLA-4 did not affect IL-6 synthesis by MC. These data directly implicate the β2 integrin/ICAM-1 pathway in MC/monocyte cell-to-cell interactions.
Mesangial MCP-1 synthesis is reduced by TNF-α and IL-1β neutralization
TNF-α and IL-1β are critical proinflammatory monokines for the induction of glomerular injury in experimental GN (35). To identify the role of soluble factors released by monocytes, we examined the effects of TNF-α and IL-1β neutralization in membrane-separated cocultures of monocyte/MC on MCP-1 production. We used MCP-1 production as a read-out because its production was more soluble factor dependent than IL-6 and ICAM-1. Neutralizing cytokine-specific or irrelevant control Abs were used during monocyte preincubation and active coculture. Simultaneous blockade of TNF-α and IL-1β led to significant inhibition of mesangial production of MCP-1 (up to 60%; Fig. 5 B), whereas neutralization of either TNF-α or IL-1β alone did not significantly affect its production. These data indicate that activated monocytes stimulate MC, at least in part, through soluble TNF-α and IL-1β. Moreover, functional redundancy likely exists, as single cytokine targeting was insufficient.
Cell-to-cell contact alone is sufficient to induce IL-6, MCP-1, and ICAM-1 synthesis by MC
Although these data established the importance of cell-to-cell contact between activated monocytes and MC, soluble factors were present in all cultures. Whether cell-to-cell contact-mediated signals alone might be sufficient to induce inflammatory responses remained unclear. Therefore, activated monocytes were fixed with 1% paraformaldehyde to preserve membrane integrity but prevent soluble factor secretion. Fixed monocytes were cultured with MC either in contact with or separated from MC at variable ratios. Fixed CD40L-activated monocytes induced IL-6, MCP-1, and ICAM-1 synthesis by MC in a dose-dependent manner, which was abrogated by membrane separation (Fig. 6). Plasma membrane preparation from CD40L-activated monocytes also induced IL-6 and MCP-1 production by MC (Fig. 7,A). As before, CD40L-activated monocytes induced higher responses than TNF-α-activated monocytes (data not shown). Compared with previous live cocultures, the monocyte:MC ratio was increased 20-fold to induce comparable mesangial responses. In addition, membrane preparations from CD40L-activated monocytes induced an ∼5-fold increase in TNF-α-induced IL-6 production by MC (Fig. 7 B), suggesting that cell contact-mediated signals operate synergistically with soluble factors. Intriguingly, MCP-1 synthesis was induced by cell-to-cell contact and was not blocked by addition of TNF-α- and IL-1β-specific mAbs (data not shown), indicating that membrane-bound cytokine may not replace soluble cytokine in this system and that proinflammatory monokine “leakage” from fixed cells is unlikely to explain our observations. Together, these data clearly demonstrate that monocyte cell-surface molecules are potent inducers of inflammatory mediator production by MC in the absence of soluble factors. This activity is significantly enhanced in CD40L-activated, compared with TNF-α-activated, monocytes, suggesting that the former activation pathway may be of prime importance in vivo.
Ligation of ICAM-1 induces IL-6, but not MCP-1, production by MC
There is increasing evidence that ICAM-1 acts not only as an adhesion molecule, but also has signaling functions (36, 37). The finding that the blockade of the β2 integrin/ICAM-1 pathway reduced MC IL-6 synthesis (Fig. 5,A), and that cell-to-cell contact alone was sufficient to induce it, suggested that ICAM-1 ligation might directly activate MC. Therefore, we investigated whether stimulation of MC through ICAM-1 induced proinflammatory responses. ICAM-1 on the surface of MC were cross-linked by simultaneously adding mouse monoclonal anti-ICAM-1 Ab and the F(ab′)2 of goat anti-mouse IgG Fcγ fragment. After 24 h, IL-6 and MCP-1 synthesis by MC were determined by ELISA. ICAM-1 ligation induced IL-6 production, at a level comparable to that induced by 10 ng/ml of TNF-α (Fig. 8). Of interest, ICAM-1 ligation failed to induce MCP-1 production by MC. These data clearly demonstrate that the β2 integrin/ICAM-1 pathway plays a critical role, not only for monocyte/MC adhesion, but also for direct mesangial activation. These data may also partly explain the differential contribution of cell-to-cell contact to IL-6 and MCP-1 synthesis by MC (Fig. 4).
Macrophages occupy a central position in many T cell-mediated inflammatory processes. They promote the inflammatory response and function as effector cells that mediate tissue damage. Furthermore, they may process and present Ags to T cells. Monocyte/macrophages may be preactivated by T cells via processes that are in part CD40 dependent. CD40-CD40L interactions up-regulate the production of proinflammatory cytokines and chemokines, such as IL-1α, IL-1β, TNF-α, IL-6, and IL-8, and promote monocyte rescue from apoptotic death at inflammatory sites (26). Macrophages and MC are vital to the initiation and progression of renal injury in a variety of GN models (6, 33, 38, 39). However, less is known about those pathways through which macrophages might regulate the proinflammatory activity of MC. The present study sought to define such interactions between “physiologically activated” macrophages and MC in modifying the biologic response of the latter.
To study these interactions, we used an in vitro system to quantify the relative contribution of soluble and cell-to-cell contact-dependent pathways, achieved either through cell separation using semipermeable membranes or by effector cell fixation and utilization of isolated plasma membranes. The latter methods facilitate the assessment of cell contact-dependent pathways, independent of secreted bioactivities. Similar in vitro approaches have yielded useful information about the pathogenesis of several other human autoimmune diseases (18, 19, 20, 21, 22). We have shown that activated monocytes amplify glomerular responses by promoting proinflammatory activity in MC through discrete but nonexclusive pathways. That these included elaboration of soluble factors, such as TNFα and IL-1β, was expected. However, our data provide strong evidence that the regulatory activities of macrophages in the glomerulus extend beyond monokine secretion to include significant contributions by cell-to-cell contact. Previous reports have demonstrated that β2 integrins, such as LFA-1 and Mac-1, are involved in cell-to-cell interactions, which contribute to the glomerular accumulation of inflammatory cells and subsequent injury. LFA-1 forms an adhesion bridge with ICAM-1 on endothelial cells and MC, and the ICAM-1/LFA-1 pathway is important for various homotypic as well as heterotypic cell-to-cell interactions. Abs to ICAM-1 or LFA-1 prevent glomerular injury in a variety of experimental models of GN (40, 41, 42). Our findings suggest that such effects might be mediated through blockade of critical macrophage/MC interactions.
In the present study, nonprimed peripheral blood monocytes failed to induce any significant response in MC, allowing us to define physiologically relevant priming requirements before coculture. Our data illustrate the importance of CD40 ligation in assuring maximum effector functions in monocytes. Thus, in contrast to IL-1β, stimulation via CD40 was essential for the production of IL-12, a cytokine essential for functional Th1 responses (Fig. 1). It is of interest that most T cell clones isolated from lupus GN, in which high levels of CD40L are expressed, are predominantly of the Th1 phenotype (43). CD40L-activated monocytes also resulted in higher biologic responses from MC when compared with those stimulated by TNF-α. MC activation was assessed by measuring the synthesis of molecules with a reported pathogenic role in GN. Thus, IL-6 is a sensitive indicator of mesangial activation and has been implicated in the proliferation of MC (11). MCP-1, a chemotactic cytokine that is highly specific for lymphocytes and monocytes, is excreted in excessive amounts in the urine of patients with GN (44, 45, 46). Animal data suggest that MCP-1 is involved in glomerular crescent formation and interstitial fibrosis (47, 48, 49). ICAM-1 enhances monocyte adhesion and retention within the mesangium and facilitates cellular interaction with infiltrating mononuclear cells (13).
Previous investigation of macrophage/MC interactions, which used U-937 myelomonocytic leukemia cells, suggested that the latter may cause cytotoxic damage to and subsequent proliferation of MC (50). Although ICAM-1, LFA-1, and VLA-4 were implicated in this system, the relative contribution of secreted and membrane-bound factors was not defined, nor were differential effects on MC function determined, as was the case in the present study. Moreover, the use of leukemic cells limits the generalization of these observations to physiologic systems. In contrast, we endeavored to employ primary monocytes, priming protocols, and measures of MC activation with direct physiologic relevance to human GN.
The differential effects of cell-to-cell contact, as compared with those of soluble factors, were striking. Whereas IL-6 production appeared largely cell contact dependent, MCP-1 production was unaffected by cell separation. Functional redundancy was evident in the latter cultures because optimal neutralization of neither TNFα nor IL-1β alone was sufficient to modify total MCP-1 output. Moreover, when the ratio of monocytes to MC was increased from 1:1 to 5:1, MCP-1 production was clearly induced by a cell contact-dependent pathway. This effect was not evident in vitro until fixed and live macrophage/MC cocultures were compared. Similarly, ICAM-1 expression appeared dependent on both pathways. Stimulation of MC by activated macrophages enhanced ICAM-1 expression. This, in turn, can increase adhesion of macrophages to MC, thus creating a positive feedback loop. Our experimental system does not exactly mimic the kinetics of in vivo responses, and it is probable that soluble and contact factors may contribute differentially through the evolution of an inflammatory lesion. Thus, soluble factors might up-regulate ICAM-1 expression, which thereby facilitates enhanced cell contact-dependent effector function. Certainly, when cells were fixed, we observed a significant increase (>5-fold) in the macrophage cell number required to activate MC, suggesting that optimal activation of MC likely required both soluble and cell contact-dependent factors. Thus, cell contact may increase the efficiency of cytokine-mediated bystander activation by facilitating paracrine mechanisms, perhaps through modification of cytokine receptor expression. However, Abs to TNF-α did not modify cell contact-dependent function in our culture system, which argues against a role at least for membrane-bound TNF-α (9, 51). The differential secretion of cytokines and chemokines that we observed further suggests that cell-to-cell contact uses different intracellular pathways than those activated by soluble factors in MC. This has potential relevance in therapeutic targeting of signaling pathways.
Up-regulation of CD40 in resident renal cells and CD40L expression in infiltrating mononuclear cells have been observed in renal biopsies of patients with proliferative, but not membranous, lupus nephritis (26). In animal models of lupus nephritis, anti-CD40L Abs ameliorate nephritis even when administered after disease onset (52). CD40-mediated signals induce secretion of chemokines by resident renal cells, such as IL-8, MCP-1, and RANTES, and promote tissue damage and fibrosis (53). To this end, we envision the following scenario in vivo (Fig. 9). T cells activated by specific Ags up-regulate CD40L expression, which in turn activate macrophages either in the peripheral blood, in lymph nodes, or in situ in the kidney. Activated peripheral blood macrophages enter the glomerulus either through the mesangial area (or though holes in the Bowmanns’ capsule in the case of crescentic GN) where they activate MC and epithelial cells. Under certain conditions, activated macrophages may also express CD40L (54), raising the possibility of local fraternal activation of macrophages in the glomerulus. It is conceivable that in animal models, interruption of the CD40/CD40L pathways, in addition to preventing autoantibody production, inhibits macrophage activation and resultant renal injury. Thus, CD40L-targeted approaches may exert salutary effects on immune-mediated GN above and beyond its effects on autoantibody production
In summary, we have shown the essential role of CD40/CD40L interactions for optimal effector function of monocytes. We also have shown that CD40L-activated monocytes mediate a variety of effects on cultured human MC, which may lead to amplification of the inflammatory response. Our data demonstrate the importance of cell-to-cell interactions in this response and reiterate the role of cellular immunity in the pathogenesis of GN.
We thank Drs. R. L. Wilder, J. Rivera, J. Kopp, and J. E. Balow for critical review of the manuscript.
Abbreviations used in this paper: GN, glomerulonephritides; MCP-1, monocyte chemoattractant protein-1; CD40L, CD40 ligand; VLA-4, very late activation Ag-4; MC, mesangial cells.