Interleukin-10 inhibits Mycobacterium bovis bacillus Calmette–Guérin (BCG)-induced macrophage cytotoxicity against bladder cancer cells

Authors


Y. Luo, University of Iowa, Department of Urology, 3202 MERF, 375 Newton Road, Iowa City, IA 52242, USA.
E-mail: yi-luo@uiowa.edu

Summary

The mechanisms underlying bacillus Calmette–Guérin (BCG) immunotherapy of bladder cancer currently remain elusive. Previously, we demonstrated that macrophages were cytotoxic to bladder cancer cells upon BCG stimulation in vitro. However, macrophages from C57BL/6 mice were less potent than those from C3H/HeN mice for the killing of bladder cancer cells. This study was to determine whether interleukin (IL)-10 produced by macrophages in response to BCG is a causative factor for the reduced cytotoxicity in BCG-stimulated C57BL/6 macrophages. Thioglycollate-elicited peritoneal macrophages were prepared and analysed for the BCG induction of cytotoxicity, cytokines and nitric oxide (NO) in vitro. Compared to BCG-stimulated C3H/HeN macrophages, BCG-stimulated C57BL/6 macrophages exhibited reduced killing of bladder cancer MBT-2 cells and MB49 cells. Studies demonstrated further that BCG-stimulated C57BL/6 macrophages produced a high level of IL-10, which correlated with reduced production of tumour necrosis factor (TNF)-α, IL-6 and NO. Neutralizing endogenous IL-10 during BCG stimulation increased C57BL/6 macrophage cytotoxicity against MB49 cells by 3·2-fold, along with increased production of TNF-α by 6·4-fold and NO by 3·6-fold, respectively. Macrophages from C57BL/6 IL-10−/− mice also exhibited increased killing of MB49 cells and production of TNF-α and NO upon BCG stimulation. In addition, supplementation of exogenous recombinant IL-10 reduced BCG-induced C3H/HeN macrophage cytotoxicity against both MBT-2 cells and MB49 cells in a dose-dependent manner. These results reveal the inhibitory role of IL-10 in BCG-induced macrophage cytotoxicity, suggesting that blockage of IL-10 may potentially enhance the effect of BCG in the treatment of bladder cancer patients.

Introduction

Intravesical instillation of bacillus Calmette–Guérin (BCG) has been used for the treatment of superficial transitional cell carcinoma (TCC) of the bladder for three decades [1–3]. This therapy has been demonstrated to be more effective than localized chemotherapy and radiotherapy. Although the exact mechanisms through which BCG mediates anti-tumour immunity remain elusive, an induction of localized cellular immune responses reflecting activation of various immune cells has been proposed [3–7]. Several lines of evidence suggest that macrophages actively mediate BCG's anti-bladder cancer activity. Following BCG instillation, increased numbers of macrophages, along with T cells, natural killer (NK) cells and dendritic cells (DC), are observed in bladder cancer infiltrates and the peritumoral bladder wall [4,7–10]. The voided urine of bladder cancer patients after BCG treatment also contains increased numbers of macrophages and other types of immune cells [11–13]. Moreover, a transient secretion of various cytokines and chemokines in patients' urine after intravesical BCG treatment has been reported, including those produced predominantly by macrophages such as interleukin (IL)-6, IL-10, IL-12, IL-18 and tumour necrosis factor (TNF)-α[4,13–19]. Furthermore, both murine and human macrophages have been observed to produce a number of cytokines (e.g. TNF-α, IL-6 and IL-10) in response to BCG stimulation in vitro[20–25]. In addition to their function as antigen-presenting cells, macrophages can also act as cytostatic/cytotoxic effector cells against bladder cancer cells [4,6,24–28]. Macrophages exert their tumoricidal activity through cell–cell contact and/or release of soluble cytotoxic effector molecules such as TNF-α and nitric oxide (NO) [6,24–28]. Both TNF-α and NO are known to induce cell death through the apoptotic pathways [29–31].

IL-10 is a type 2 cytokine secreted by numerous cell types including T helper type 2 (Th2) cells [32,33], B cells [34] and monocytes/macrophages [35]. IL-10 has been shown to inhibit the development of cellular immune responses via a number of mechanisms [36]. IL-10 can block the accumulation of both macrophages and DCs at the tumour site [37,38] and down-regulate the expression of major histocompatibility complex (MHC) class II and co-stimulatory molecules [e.g. B7 and intercellular adhesion molecule-1 (ICAM-1)] on these cells [39–41], thus preventing the induction of specific anti-tumour immune responses. IL-10 can also reduce DCs' capacity to stimulate T cells, leading to the induction of antigen-specific anergy of T cells [42]. It has also been reported that CD4+ T cells in the presence of IL-10 during activation can differentiate into T regulatory cells 1 (Tr1) that are responsible for peripheral immune tolerance induced by IL-10 [43]. In addition, IL-10 can also prevent release of cytokines (e.g. TNF-α) and reactive nitrogen/oxygen intermediates (e.g. NO) by macrophages [35,44–46], thus inhibiting inflammatory, microbicidal and tumoricidal activities of these cells.

Both MB49 cells and MBT-2 cells are widely used murine bladder cancer cell lines for anti-bladder cancer studies. MB49 cells were derived from 7,12-dimethylbenz(a)anthracene-induced C57BL/6 bladder epithelial cells [47], whereas MBT-2 cells were derived from N-[4-(5-nitro-2-furyl)-2-thiazolyl] formamide (FANFT)-induced C3H/He bladder tumours [48]. Although both MB49 cell- and MBT-2 cell-derived bladder cancers in syngeneic mice have been found to be similarly responsive to intravesical BCG treatment, the detailed mechanisms through which BCG induces anti-tumour immunity in these two distinctive genetic backgrounds of mice may not be identical. For example, due to BCG induction of poor cytotoxicity in C57BL/6 macrophages, all reported studies on the potential of BCG-induced macrophage cytotoxicity against bladder cancer cells have been performed exclusively in C3H/HeN macrophages against MBT-2 cells [24–28]. Previously, we observed that BCG could induce potent cytotoxicity in C3H/HeN macrophages [24,25] but reduced cytotoxicity in C57BL/6 macrophages. In this study, we sought to determine what macrophage-derived effector molecule(s) might be responsible for the induction of reduced cytotoxicity in C57BL/6 macrophages. We have observed that IL-10 plays an inhibitory role in the induction of macrophage cytotoxicity by BCG, as increased cytotoxicity can be induced through neutralization of endogenous IL-10 in C57BL/6 mice and in genetically modified C57BL/6 mice lacking IL-10 (IL-10−/−). Our results suggest that blockage of IL-10 may potentially enhance intravesical BCG immunotherapy of bladder cancer, particularly for BCG non-responders who often develop high IL-10 during BCG treatment [16–18].

Materials and methods

Mice

Both C3H/HeN mice and C57BL/6 mice (female, 6–8 weeks old) were purchased from the National Cancer Institute (NCI) and allowed free access to food and water. All animal studies were approved by the University of Iowa Animal Care and Use Committee and were in compliance with the Guide for the Care and Use of Laboratory Animals prepared by the National Institutes of Health.

Cell culture

Both murine bladder cancer MBT-2 cells and MB49 cells were obtained from Dr Timothy Ratliff at Purdue Cancer Center and grown in RPMI-1640 medium containing 10% fetal bovine serum (FBS), 100 units/ml of penicillin and 100 µg/ml of streptomycin at 37°C in a humidified 5% CO2 incubator. Cells were free of mycoplasma as determined by the polymerase chain reaction (PCR)-based analysis using Stratagene Mycoplasma PlusTM PCR Primer Set (La Jolla, CA, USA).

BCG culture

Pasteur strain BCGs containing the kanamycin resistant gene were grown in Middlebrook 7H9 Bacto broth (Difco, Detroit, MI, USA) supplemented with 10% albumin dextrose catalase (ADC; 5% bovine serum albumin fraction V, 2% dextrose and 0·85% NaCl), 0·05% Tween 80 (Sigma, St Louis, MO, USA) and 30 µg/ml kanamycin, as described previously [49]. Log phase cultures of viable BCG were quantified using the absorbance at 600 nm [1 optical density (OD600) unit = 2·5 × 107 colony-forming unit (CFU)/ml] and used for macrophage stimulation.

Macrophage cytotoxicity assay

Macrophages were prepared from mice based on the previously described method with slight modifications [24,25]. Briefly, 3 days after intraperitoneal injection of 2 ml of 3% thioglycollate (Sigma), thioglycollate-elicited peritoneal exudates cells (PECs) were harvested by washing out the peritoneal cavity, suspended in RPMI-1640 – 10% FBS medium, and incubated for 3 h at 37°C in a humidified 5% CO2 incubator. Non-adherent cells were then removed by washing three times. The remaining cells were 96% F4/80-positive cells as determined by flow cytometry. The prepared peritoneal macrophages were stimulated with BCG (0·005 OD/well) or lipopolysaccharide (LPS; Sigma; 10 µg/ml) for 24 h and then incubated with 51Cr-labelled MBT-2 cells or MB49 cells at the indicated effector : target cell (E : T) ratios for 20 h. To determine the role of IL-10 in the BCG induction of macrophage cytotoxicity, a goat anti-IL-10 neutralizing antibody (R&D Systems, Minneapolis, MN, USA) or mouse rIL-10 (Genzyme, Cambridge, MA, USA) was added at the indicated concentrations during BCG stimulation of macrophages. Goat immunoglobulin G (IgG) (Sigma) was used as a control for the anti-IL-10 antibody. Spontaneous 51Cr release was determined by incubating radiolabelled target cells in the absence of macrophages. Maximal 51Cr release was determined by incubating the target cells in 2% Triton X-100. Cytotoxicity was calculated as the mean of triplicate wells according to the formula: % lysis = [(experimental cpm – spontaneous cpm)/(maximal cpm – spontaneous cpm)] × 100.

MTT (thiazolyl blue tetrazolium bromide) assay

To assess the killing of MBT-2 cells and MB49 cells by macrophage-derived cytotoxic effector molecules, cells were incubated with various doses of rTNF-α (BD PharMingen, San Diego, CA, USA) or S-nitroso-N-acetylpencillamine (SNAP; Sigma), a NO donor, in a 96-well flat-bottomed plate at 1 × 105 cells in 200 µl of RPMI-1640 – 10% FBS medium per well for 24 h. After incubation, 20 µl of MTT (5 mg/ml; Sigma) was added to each well and the incubation continued for 4 h. The medium overlying cells was then aspirated and cells were solubilized with 200 µl of dimethylsulphoxide (DMSO). The OD was read at 570 nm. Percentage of cell death was calculated and presented as mean ± standard deviation (s.d.) of triplicate wells referring to non-treated cells (100% viability).

Cytokine enzyme-linked immunosorbent assay (ELISA)

A sandwich format ELISA was used to evaluate IL-10, TNF-α and IL-6 production in the conditioned medium as described previously [50]. The paired capture and detecting antibodies were obtained from BD PharMingen for TNF-α (clones MP6–XT22 and MP6–XT3), IL-10 (clones JES5–2A5 and SXC-1) and IL-6 (clones MP5–20F3 and MP5–32C11). Cytokine concentration was calculated in standard mass/volume format using the standard curve derived from purified rTNF-α (Genzyme), rIL-10 (Genzyme) or IL-6 (R&D Systems) accordingly and presented as mean ± s.d. of duplicate wells.

Determination of NO production

NO in the conditioned medium was assessed by measuring the nitrite (NO2-) and nitrate (NO3-) concentrations by the Criess reaction, using the Cayman NO2-/NO3- colorimetric assay kit (Alexis Corporation, San Diego, CA, USA), according to the manufacturer's instructions. The NO2- plus NO3- concentrations were calculated from a graph of absorbance at 540 nm as a function of NO3-concentration using the NO3- standard provided in the assay kit and presented as mean ± s.d. of triplicate wells.

Statistical analysis

All determinations were made in multiple experiments. Statistical significance was determined by Student's t-test using spss version 11·0 software (Chicago, IL, USA) for macrophage effector molecule production. A P-value of < 0·05 was considered significant.

Results

C57BL/6 macrophages exhibit reduced cytotoxicity in response to BCG stimulation compared to C3H/HeN macrophages

To evaluate the effect of BCG on the induction of macrophage cytotoxicity against bladder cancer cells, we prepared peritoneal macrophages from both C3H/HeN mice and C57BL/6 mice and used them as effector cells against MBT-2 cells or MB49 cells in 51Cr release assays. We observed that BCG induced effective killing of MBT-2 cells by macrophages from both murine strains (Fig. 1a and c), although C57BL/6 macrophages were less potent than C3H/HeN macrophages. At an E : T ratio of 30 : 1, BCG-stimulated C3H/HeN macrophages killed 65% of MBT-2 cells whereas BCG-stimulated C57BL/6 macrophages killed 42% of MBT-2 cells. Even at a low E : T ratio of 1 : 1, BCG-stimulated macrophages still killed 36% of MBT-2 cells for C3H/HeN macrophages and 21% of MBT-2 cells for C57BL/6 macrophages, respectively. When MB49 cells were used as targets, BCG-stimulated C3H/HeN macrophages also effectively killed these cells (Fig. 1b), although killing of these cells was less effective than killing of MBT-2 cells. Compared to the killing of MB49 cells by BCG-stimulated C3H/HeN macrophages, BCG-stimulated C57BL/6 macrophages exhibited substantially reduced cytotoxicity against MB49 cells (Fig. 1d). At an E : T ratio of 30 : 1, BCG-stimulated C57BL/6 macrophages killed only 25% of MB49 cells. No killing of MB49 cells was observed at an E : T ratio of 3 : 1. Compared to BCG-induced macrophage cytotoxicity, LPS also induced similar, although low, killing of both MBT-2 cells and MB49 cells by macrophages of both murine strains. These results indicate that macrophages are cytotoxic to bladder cancer cells upon stimulation with BCG or LPS. However, C57BL/6 macrophages are less potent than C3H/HeN macrophages for the killing of bladder cancer cells in response to BCG or LPS stimulation.

Figure 1.

C57BL/6 macrophages exhibit reduced cytotoxicity in response to bacillus Calmette–Guérin (BCG) stimulation compared to C3H/HeN macrophages. Peritoneal macrophages of C3H/HeN mice (left panel) and C57BL/6 mice (right panel) were stimulated with BCG (0·005 optical density) or lipopolysaccharide (LPS) (10 µg/ml) for 24 h. Cytotoxicity was evaluated by further co-culturing BCG- or LPS-stimulated macrophages with 51Cr-labelled murine bladder cancer MBT-2 cells (a,c) or MB49 cells (b,d) at the indicated effector : target ratios for 20 h. Supernatants were then harvested, released 51Cr counted, and percentage of cell lysis calculated and presented as the mean of triplicate incubations.

Bladder cancer MBT-2 cells and MB49 cells are susceptible to the killing by macrophage-derived cytotoxic effector molecules TNF-α and NO

We observed that MB49 cells were less susceptible than MBT-2 cells to BCG- and LPS-induced macrophage cytotoxicity (Fig. 1). To exclude the possibility that the observed reduced killing of MB49 cells was due to the inherent resistance of these cells to macrophage-derived cytotoxic effector molecules, we treated MB49 cells with four different concentrations of rTNF-α (range 0·1–100 ng/ml) or SNAP (a NO donor; range 0·01–10 mM) for 24 h, followed by analysis of cell viability. MBT-2 cells were treated with rTNF-α or SNAP in parallel for comparison. Treatment with rTNF-α (Fig. 2a) or SNAP (Fig. 2b) induced cell death for both MB49 cells and MBT-2 cells in a dose-dependent manner. MB49 cells appeared more sensitive than MBT-2 cells to rTNF-α (at concentrations of 10 and 100 ng/ml) and slightly less sensitive than MBT-2 cells to SNAP (at concentrations of 0·1 and 1 mM). These results suggest that MB49 cells are not naturally resistant to macrophage cytotoxicity and can be killed by macrophage-derived cytotoxic effector molecules such as TNF-α and NO.

Figure 2.

Murine bladder cancer MBT-2 cells and MB49 cells are susceptible to killing by the macrophage cytotoxic effector molecules tumour necrosis factor (TNF)-α and nitric oxide (NO). Both MB49 cells and MBT-2 cells were treated with the indicated concentrations of rTNF-α (a) or S-nitroso-N-acetylpencillamine (SNAP) (b), a NO donor, in culture for 24 h. Cell viability was then assessed by thiazolyl blue tetrazolium bromide (MTT) assay. The percentage of cell death was calculated and presented as the mean ± standard deviation of triplicate incubations referring to non-treated cells (100% viability).

High IL-10 production correlates with low production of effector molecules in BCG-stimulated C57BL/6 macrophages

To investigate whether there was a difference in effector molecule production between C57BL/6 macrophages and C3H/HeN macrophages upon BCG or LPS stimulation, which might explain the observed reduced killing of bladder cancer cells by the former macrophages, we stimulated macrophages of both murine strains with BCG or LPS for 48 h, followed by analysis of IL-10 (Fig. 3a,e), TNF-α (Fig. 3b,f), IL-6 (Fig. 3c,g) and NO (Fig. 3d,h) in the conditioned media. No IFN-γ, IL-2 and IL-12 were detected (data not shown). Compared to BCG-stimulated C3H/HeN macrophages (left panel), BCG-stimulated C57BL/6 macrophages (right panel) produced increased IL-10 by 4·5-fold. This high IL-10 production correlated with decreased production of TNF-α by 4·4-fold, IL-6 by 2·4-fold and NO by 2·3-fold in BCG-stimulated C57BL/6 macrophages. Similarly, stimulation with LPS induced IL-10 production by macrophages of both murine strains, along with reduced production of the effector molecules tested (particularly in LPS-stimulated C3H/HeN macrophages). These results suggest that a high level of IL-10 could inhibit macrophage production of cytotoxic effector molecules and thus lead to reduced macrophage cytotoxicity against bladder cancer cells.

Figure 3.

High interleukin (IL)-10 production correlates with low production of tumour necrosis factor (TNF)-α, IL-6 and nitric oxide (NO) in bacillus Calmette–Guérin (BCG)-stimulated C57BL/6 macrophages. Peritoneal macrophages (1 × 105 cells/well) of C3H/HeN mice (left panel) and C57BL/6 mice (right panel) were cultured in the absence or presence of BCG (0·005 optical density) or lipopolysaccharide (LPS) (10 µg/ml) for 48 h. Production of IL-10 (a,e), TNF-α (b,f) and IL-6 (c,g) in culture supernatants were then evaluated by enzyme-linked immunosorbent assay. Data represent the mean ± standard deviation of duplicate incubations. Production of NO (d,h) in culture supernatants was evaluated using the Cayman NO2-/NO3- colorimetric assay kit. Data represent the mean ± standard deviation of triplicate incubations. Mac, macrophages.

Neutralizing endogenous IL-10 enhances BCG-induced C57BL/6 macrophage cytotoxicity against MB49 cells

To determine the role of IL-10 in the BCG induction of C57BL/6 macrophage cytotoxicity, we blocked endogenous IL-10 produced by macrophages during cytotoxicity assay using a goat anti-IL-10 neutralizing antibody. Blockage of IL-10 increased the killing of MB49 cells by 3·2-fold by BCG-stimulated C57BL/6 macrophages at an E : T ratio of 10 : 1 (Fig. 4a). The blockage of endogenous IL-10 also increased macrophage production of TNF-α by 6·4-fold (Fig. 4b; P < 0·05) and NO by 3·6-fold (Fig. 4c; P < 0·05), respectively. Similar studies were also performed in genetically modified C57BL/6 mice lacking IL-10 (IL-10−/−). As expected, BCG induced enhanced cytotoxicity against MB49 cells in the IL-10 null macrophages (Fig. 5a; right panel). Accordingly, BCG-stimulated C57BL/6 IL-10−/− macrophages also produced increased TNF-α (Fig. 5b; P < 0·05) and NO (Fig. 5c; P < 0·05) compared to BCG-stimulated C57BL/6 macrophages. These results indicate that IL-10 acts as a strong inhibitor for the BCG induction of macrophage cytotoxicity and that blockage of IL-10 may reverse this inhibition, leading to an enhanced induction of macrophage cytotoxicity by BCG.

Figure 4.

Neutralizing endogenous interleukin (IL)-10 enhances killing of murine bladder cancer (MB49) cells and production of tumour necrosis factor (TNF)-α and nitric oxide (NO) by bacillus Calmette–Guérin (BCG)-stimulated C57BL/6 macrophages. Peritoneal macrophages (1 × 105 cells/well) of C57BL/6 mice were cultured in the absence or presence of BCG (0·005 optical density) or BCG plus goat anti-IL-10 neutralizing antibody (4 µg/ml) for 24 h. Control goat immunoglobulin G (4 µg/ml) was used in parallel. 51Cr-labelled MB49 cells (1 × 104 cells/well) were then added and the incubation continued for 20 h. Cytotoxicity (a) was then assessed by measuring 51Cr release in culture supernatants and percentage of cell lysis calculated and presented as the mean of triplicate incubations. Production of TNF-α (b) and NO (c) was assessed using enzyme-linked immunosorbent assay and the Cayman NO2-/NO3- colorimetric assay kit, respectively. Data represent the mean ± standard deviation of duplicate (for TNF-α) or triplicate (for NO) incubations. Numerical values represent the fold increase referring to the basal levels treated with BCG alone. *P < 0·05 (compared to BCG-treated macrophages). Mac, macrophages.

Figure 5.

Macrophages from C57BL/6 IL-10−/− mice exhibit potent cytotoxicity against murine bladder cancer (MB49) cells and produce increased tumour necrosis factor (TNF)-α and nitric oxide (NO) in response to bacillus Calmette–Guérin (BCG) stimulation. (a) Peritoneal macrophages of both C57BL/6 mice and C57BL/6 IL-10−/−mice were stimulated with BCG [0·005 optical density (OD)] for 24 h. Cytotoxicity was evaluated by further co-culturing BCG-stimulated macrophages with 51Cr-labelled MB49 cells at the indicated effector : target ratios for 20 h. Supernatants were then harvested, released 51Cr counted, and percentage of cell lysis calculated and presented as the mean of triplicate incubations. (b,c) Peritoneal macrophages (1 × 105 cells/well) of both C57BL/6 mice and C57BL/6 IL-10−/−mice were cultured in the absence or presence of BCG (0·005 OD) for 48 h. Production of TNF-α and NO was assessed by enzyme-linked immunosorbent assay and the Cayman NO2-/NO3- colorimetric assay kit, respectively. Data represent the mean ± standard deviation of duplicate (for TNF-α) or triplicate (for NO) incubations. *P < 0·05 (compared to BCG-treated C57BL/6 macrophages). Mac, macrophages.

Supplementation of exogenous rIL-10 inhibits BCG-induced C3H/HeN macrophage cytotoxicity against MBT-2 cells and MB49 cells

To confirm the inhibitory role of IL-10 in BCG-induced macrophage cytotoxicity, we added rIL-10 at four different concentrations (range 2–16 ng/ml) during BCG stimulation of C3H/HeN macrophages, followed by analysis of macrophage cytotoxicity against MBT-2 and MB49 cells. Supplementation of rIL-10 reduced BCG-induced C3H/HeN macrophage cytotoxicity in a dose-dependent manner. Up to 91% inhibition (at the rIL-10 concentration of 16 ng/ml) on the killing of MBT-2 cells at an E : T ratio of 10 : 1 was observed (Fig. 6a). Similarly, up to 92% inhibition (at the rIL-10 concentration of 8 ng/ml) on the killing of MB49 cells at an E : T ratio of 30 : 1 was observed (Fig. 6b). These results support the inhibitory role of IL-10 in the BCG induction of macrophage cytotoxicity against bladder cancer cells such as that observed for the BCG-induced C57BL/6 macrophage cytotoxicity.

Figure 6.

Supplementation of exogenous recombinant interleukin (rIL)-10 inhibits bacillus Calmette–Guérin (BCG)-induced C3H/HeN macrophage cytotoxicity against murine bladder cancer MBT-2 cells and MB49 cells. Peritoneal macrophages of C3H/HeN mice were cultured in the absence or presence of BCG (0·005 optical density) or BCG plus the indicated doses of rIL-10 for 24 h. 51Cr-labelled MBT-2 cells (a) or MB49 cells (b) were then added and the incubation continued for 20 h. Cytotoxicity was assessed by measuring 51Cr release in culture supernatants and percentage of cell lysis calculated and presented as the mean of triplicate incubations. Numerical values represent the fractions of reduction referring to the basal lysis treated with BCG alone. Mac, macrophages.

Discussion

To study the mechanisms underlying BCG treatment of bladder cancer, we investigated the BCG induction of macrophage cytotoxicity towards bladder cancer MBT-2 cells and MB49 cells in vitro. We observed that BCG induced clear killing of bladder cancer cells by both C3H/HeN macrophages and C57BL/6 macrophages. However, the latter macrophages were less potent than the former macrophages. To determine which factor(s) might be responsible for the reduced cytotoxicity in BCG-stimulated C57BL/6 macrophages, we analysed macrophage-derived effector molecules produced in response to BCG stimulation and compared them between C57BL/6 macrophages and C3H/HeN macrophages. We have found several lines of evidence supporting that IL-10 plays an inhibitory role in the BCG induction of macrophage cytotoxicity and is responsible for the reduced cytotoxicity in BCG-stimulated C57BL/6 macrophages. First, production of a high level of IL-10 was observed in BCG-stimulated C57BL/6 macrophages, which correlated with low cytotoxicity and low production of TNF-α and NO in these cells. Second, neutralization of endogenous IL-10 during BCG stimulation led to increased C57BL/6 macrophage cytotoxicity against MB49 cells along with increased production of TNF-α and NO. Third, macrophages from C57BL/6 IL-10−/− mice exhibited increased cytotoxicity against MB49 cells and produced increased TNF-α and NO in response to BCG stimulation. In addition, we observed that supplementation of exogenous rIL-10 reduced BCG-induced C3H/HeN macrophage cytotoxicity against both MBT-2 cells and MB49 cells in a dose-dependent manner.

Multiple effector mechanisms are involved in the killing of bladder cancer cells by BCG-induced macrophage cytotoxicity. Previously, we observed that both direct effector : target cell contact and release of soluble cytotoxic factors (such as TNF-α) by macrophages contributed to the killing of MBT-2 cells by BCG-stimulated C3H/HeN macrophages [25]. In this study, we observed further that BCG-stimulated C57BL/6 macrophages could also produce TNF-α. In addition, we also observed that both BCG-stimulated C3H/HeN macrophages and C57BL/6 macrophages produced NO. These observations were consistent with previous observations that macrophages are capable of producing TNF-α and NO in response to BCG stimulation [6,20–25], which were tumoricidal for human and murine bladder cancer cells [6,24,25,51,52]. TNF-α is known to act as an autocrine cytokine essential for macrophage activation and NO production [22,53,54], whereas NO is known to be required for macrophage activation and TNF-α production [53]. Thus, production of these two effector molecules provided a positive feedback loop for macrophage activation and cytotoxicity induction. Indeed, we have observed that an increase in BCG-induced C57BL/6 macrophage cytotoxicity (by neutralizing endogenous IL-10) correlated with increased production of both TNF-α and NO. Because macrophages exert multiple pivotal functions in host immune responses and serve as a first line of defence against BCG infection [55,56] it is likely that, in the case of BCG intravesical treatment of bladder cancer, macrophages initiate BCG-induced anti-bladder cancer responses and at the same time also act as tumoricidal cells against bladder cancer cells. As observed in our and others' studies [24–28], such BCG-induced macrophage cytotoxicity can be very potent for both C3H/HeN macrophages and C57BL/6 macrophages (if endogenous IL-10 is blocked) against bladder cancer cells.

Although IL-10 is known to have pleiotropic inhibitory functions, its role in regulating the BCG induction of macrophage tumoricidal activity has not well been studied. Previously, we observed that BCG could induce enhanced delayed-type hypersensitivity (DTH) in C57BL/6 mice treated with neutralizing anti-IL-10 antibody or genetically deficient in IL-10 (IL-10−/−), resulting in enhanced inhibition on MB49 orthotopic tumour growth [18]. In this study, we observed further that high IL-10 production was associated with low killing of MB49 cells by BCG-induced C57BL/6 macrophage cytotoxicity in vitro. In addition, we also observed that blockage of endogenous IL-10 could enhance this BCG-induced macrophage cytotoxicity against MB49 cells, which correlated with increased production of TNF-α and NO. As supplementation of exogenous rTNF-α failed to increase the BCG-induced C57BL/6 macrophage cytotoxicity (data not shown), IL-10 appeared to play a predominant role in controlling the BCG induction of C57BL/6 macrophage cytotoxicity. To support the regulatory role of IL-10, we also observed that BCG induced increased cytotoxicity against MB49 cells in C57BL/6 IL-10−/− macrophages and that supplementation of exogenous rIL-10 inhibited BCG-induced C3H/HeN macrophage cytotoxicity against both MBT-2 cells and MB49 cells in a dose-dependent manner. These results suggest that blockage of IL-10 may potentially enhance the effect of BCG in the treatment of bladder cancer patients.

We observed that MB49 cells were less susceptible than MBT-2 cells to the BCG-induced macrophage cytotoxicity. Because both bladder cancer cell lines were found to be similar in the killing by macrophage-derived cytotoxic effector molecules TNF-α and NO (simulated by rTNF-α and SNAP), the reduced killing of MB49 cells suggests the involvement of other factors in the BCG-induced macrophage cytotoxicity against this cell line. Indeed, it has been reported that MB49 cells are capable of inducing IL-10 production by other immune cells [57,58]. In this study, we also observed that MB49 cells could induce IL-10 production by C57BL/6 macrophages (data not shown). Moreover, differential expression of IL-10 receptors on tumour cells or macrophages from different genetic backgrounds may also play a role in susceptibility of the tumour cells to BCG-induced macrophage cytotoxicity. However, these assumptions need to be validated.

This study reveals that IL-10 is an inhibitory factor for the BCG induction of macrophage cytotoxicity. Our results provide the rationale for further studies on IL-10 blockage in BCG immunotherapy of bladder cancer, particularly for BCG non-responders who are often associated with high IL-10 production in response to BCG treatment.

Acknowledgements

We thank Dr Timothy Ratliff for providing the murine bladder cancer cell lines and Ms Kris Greiner for editorial review of this manuscript.

Disclosure

The authors have no conflict of interest.

Ancillary