Cancer Immunology, Immunotherapy

, Volume 53, Issue 12, pp 1127–1134

Modulation of monocyte–tumour cell interactions by Mycobacterium vaccae

Authors

  • Jarosław Baran
    • Department of Clinical Immunology, Polish-American Institute of PaediatricsJagiellonian University Medical College
  • Monika Baj-Krzyworzeka
    • Department of Clinical Immunology, Polish-American Institute of PaediatricsJagiellonian University Medical College
  • Kazimierz Węglarczyk
    • Department of Clinical Immunology, Polish-American Institute of PaediatricsJagiellonian University Medical College
  • Irena Ruggiero
    • Department of Clinical Immunology, Polish-American Institute of PaediatricsJagiellonian University Medical College
    • Department of Clinical Immunology, Polish-American Institute of PaediatricsJagiellonian University Medical College
Original Article

DOI: 10.1007/s00262-004-0552-6

Cite this article as:
Baran, J., Baj-Krzyworzeka, M., Węglarczyk, K. et al. Cancer Immunol Immunother (2004) 53: 1127. doi:10.1007/s00262-004-0552-6

Abstract

Immunotherapy with Mycobacterium vaccae as an adjuvant to chemotherapy has recently been applied to treatment of patients with cancer. One of the mechanisms of antitumour activity of Mycobacterium bovis bacillus Calmette-Guérin (BCG), the prototype immunomodulator, is associated with activation of monocytes/macrophages. These studies were undertaken to determine how M. vaccae affects monocyte–tumour cell interactions and, in particular, whether it can prevent or reverse deactivation of monocytes that occurrs following their contact with tumour cells during coculture in vitro. Deactivation is characterised by the impaired ability of monocytes to produce tumour necrosis factor α (TNF-α), interleukin 12 (IL-12), and enhanced IL-10 secretion following their restimulation with tumour cells. To see whether deactivation of monocytes can be either prevented or reversed, three different strains of M. vaccae—B 3805, MB 3683, and SN 920—and BCG were used to stimulate monocytes before or after exposure to tumour cells. Pretreatment of monocytes with M. vaccae MB 3683, SN 920 and BCG before coculture resulted in increased TNF-α and decreased IL-10 production. All strains of M. vaccae and BCG used for treatment of deactivated monocytes enhanced depressed TNF-α secretion. Strain SN 920 and BCG increased IL-12 release but only BCG treatment inhibited an enhanced IL-10 production by deactivated monocytes. Thus, although some strains of M. vaccae may either prevent or reverse tumour-induced monocyte deactivation, none of them appears to be more effective than BCG.

Keywords

Cancer cellsCytokinesDeactivationMonocytesMycobacterium vaccae

Introduction

Immunotherapy with the use of Mycobacterium bovis bacillus Calmette-Guérin (BCG) has been applied with limited success for cancer treatment apart from superficial bladder cancer [9]. One of the mechanisms of BCG antitumour activity is due to activation of monocytes/macrophages and enhancement of their cytotoxicity against tumour cells [23]. Mycobacterium vaccae (SRL172) is an environmental saprophyte that has been studied for a possible role in the management of tuberculosis [27]. More recently, M. vaccae immunotherapy combined with chemotherapy is undergoing phase II clinical trials in cancer patients, and there is a trend for improved response rate and survival [1, 14, 19, 20]. It has been shown that M. vaccae possesses T helper-1 (Th1) adjuvanticity [26], expresses heat shock proteins and antineoplastic cell surface proteoglycans [28]. A shift in cytokine responses toward a Th2 pattern, favouring humoral immunity and suppressing cellular immune responses is observed in advanced cancer [3]. M. vaccae down-regulates interleukin 4 (IL-4) and up-regulates IL-2 production, switching toward a Th1 response [12] and activates natural killer cells [14]. To our knowledge no effect of M. vaccae on monocyte/macrophage antitumour activity has been determined.

We have previously shown that human monocytes respond to tumour cells in vitro by production of proinflammatory cytokines and reactive oxygen intermediates (ROI) [17, 32]. However, after the first contact with tumour cells, monocytes become deactivated (unresponsive) to the in vitro rechallenge with tumour cells and are polarised toward a M2 (TNF-α, IL-12, IL-10+) phenotype [18]. This phenomenon is identical to tumour-induced dysfunction of tumour-infiltrating macrophages (TIMs) [11]. In the present study, we asked whether treatment of monocytes with M. vaccae or with BCG, which was used for comparison as the prototype of monocyte/macrophage activator, may enhance their response to tumour cells and prevent or reverse their unresponsiveness. The in vitro model of macrophage-tumour interactions was used, in which monocytes are cocultured with tumour cells and their antitumour potential is studied [16, 31, 32]. In this condition, monocytes respond to stimulation with tumour cells by production of tumour necrosis factor α (TNF-α), IL-10 and IL-12 [18, 32]. However, monocytes isolated from the coculture are unresponsive to restimulation by tumour cells, as defined by decreased production of TNF-α and IL-12, and increased secretion of IL-10, but respond normally to lipopolysaccharide (LPS)—a process called deactivation [18].

This study shows that although all strains of M. vaccae used (B 3805, MB 3683 and SN 920) enhanced tumour cell–induced production of TNF-α, but not IL-10 and IL-12, only some (MB 3805 and SN 920) were able to prevent or reverse tumour cell–induced monocyte deactivation. However, none of M. vaccae strains used was more effective than BCG.

Materials and methods

Mycobacteria

Mycobacterium vaccae strains: ATCC SN 920, NRRL MB 3683 and NRRL B 3805 were kindly provided by Professor Leon Sedlaczek (Centre for Microbiology and Virology, Polish Academy of Sciences, Łódź, Poland). SN 920 is the reference strain of M. vaccae (Reference Laboratory for the genus Mycobacterium, Forschungszentrum Borstel, Borstel, Germany), B 3805 and MB 3683 are the mutants of the same parental strain [24]. Bacteria were grown in NB medium pH 6.0–6.2 containing 8.0 g/l of nutrient broth (Difco; BD Diagnostic Systems, Sparks, MD, USA) and 10.0 g/l glucose; supplemented with 0.2% (v/v) of Tween 80 (Aldrich, Sigma-Aldrich Corporation, St Louis, MO, USA), for 120 h at 35°C and washed twice with a large volume of saline. Cell number was calculated using the standard curve based on colony-forming unit (CFU) counts and adjusted to the final concentration of 5×107/ml. Lyophilized BCG, Moreau strain (OnkoBCG 50—Vaccinum BCG; Biomed, Lublin, Poland), was dissolved in 5 ml of saline to the final concentration of 5×107/ml of viable bacilli.

Isolation of cell populations

Human peripheral blood mononuclear cells (PBMCs) were isolated from EDTA blood of healthy donors by the standard Ficoll-Hypaque density gradient centrifugation (Pharmacia, Uppsala, Sweden). Monocytes were separated from mononuclear cells by counter-flow centrifugal elutriation with a JE-6B elutriation system equipped with a 5-ml Sanderson separation chamber (Beckman-Coulter, Palo Alto, CA, USA) as previously described [17]. The cells were suspended in RPMI 1640 medium (Biochrom, Berlin, Germany) with gentamycin (25 μg/ml; Biochrom), glutamine (2 mM; Gibco, Paisley, UK) and 10% foetal bovine serum (FBS; Biochrom). Monocytes were 90–96% pure as judged by FACS analysis using anti-CD14 monoclonal antibody (mAb; BD Biosciences Pharmingen, San Diego, CA, USA).

Cell lines

The following human cell lines were used: HPC-4 (pancreatic adenocarcinoma) and DeTa (colorectal adenocarcinoma), as previously described [17]. Cells were cultured by biweekly passages in RPMI 1640 with 5% FBS. Cell lines were regularly tested for Mycoplasma sp. contamination using a PCR-ELISA kit according to the manufacturer’s instructions (Roche, Mannheim, Germany).

Cell cultures

Isolated monocytes were preincubated for 2 h with Mycobacteria at a ratio of 1:10 and washed. Then HPC-4 cells were added and cultured for a further 2 h. Monocytes were washed three times and, following incubation with anti-CD14, conjugated to MicroBeads (Miltenyi Biotec, Bergisch Gladbach, Germany). CD14+ cells were separated using miniMACS and MS columns (Miltenyi Biotec). Isolated CD14+ monocytes were cultured either alone or with tumour cells at a ratio of 1:0.3 (coculture) in flat-bottom 96-well microtiter plates (Nunc, Roskilde, Denmark) for 18 h in a humidified 5% CO2 atmosphere. Then supernatants were collected and tested for cytokine content. Alternatively, CD14+ cells after separation from the coculture with tumour cells were incubated with Mycobacteria for 3 h, washed three times and restimulated with tumour cells for 18 h. Then supernatants were collected and tested for cytokine content.

Determination of cytokines

Appropriate ELISA kits (BD Biosciences Pharmingen, San Diego, CA, USA) were used to measure concentrations of TNF-α, IL-10 and IL-12p40 in the culture supernatants, according to the manufacturer’s instructions. Detection level for TNF-α was 20 pg/ml, and 10 pg/ml for IL-10 and IL-12.

Statistical analysis

Statistical analysis was performed by paired Student’s t-test using Excel software. Differences were considered significant at p<0.05.

Results

Production of cytokine by monocytes stimulated with M. vaccae strains

In the initial experiments, monocytes were cultured with different strains of M. vaccae to determine the induction of cytokine secretion. Monocytes were cultured with M. vaccae for 4 h, washed and then cultured in the medium for an additional 18 h. All three strains of M. vaccae induced secretion of TNF-α and IL-10, but not IL-12, to a comparable extent (Fig. 1). BCG, which was used for comparison, also caused secretion of TNF-α and IL-10, but the levels were lower.
Fig. 1

Production of cytokines by monocytes stimulated with different strains of M. vaccae or BCG. Monocytes were cultured with M. vaccae or BCG for 4 h (ratio 1:10), washed and then cultured in the medium for an additional 18 h. Supernatants were collected and tested for cytokine content. No IL-12 was detected. Means ± SD of four independent experiments are shown. *Significant at p<0.05

Modulatory effect of M. vaccae on tumour cell–induced secretion of cytokines by monocytes

In this set of experiments, we asked whether M. vaccae bacilli can modulate cytokine secretion by human monocytes activated with tumour cells. To answer that question, monocytes were cultured either in the medium alone or with M. vaccae or BCG for 2 h, washed and then stimulated with HPC-4 cancer cells for an additional 18 h. Then supernatants were collected and tested for cytokine content. Monocytes cultured in the presence of tumour cells secreted TNF-α, IL-10 and IL-12. All strains of M. vaccae used for pretreatment of monocytes caused an enhancement of TNF-α, but not IL-10 or IL-12 secretion (Fig. 2). BCG exerted its strongest enhancing effect on TNF-α and also increased IL-12 production. This suggested that M. vaccae facilitated interactions of monocytes with tumour cells, as determined by the increased production of TNF-α.
Fig. 2

Cytokine production by monocytes stimulated with HPC-4 cancer cells in the presence of different strains of M. vaccae or BCG. Monocytes were cultured either in the medium alone or were pretreated either with M. vaccae or BCG for 2 h (ratio 1:10), washed and then stimulated with HPC-4 cancer cells for an additional 18 h. Then supernatants were collected and tested for cytokine content. *Significant at p<0.05

Prevention of tumour cell–induced monocyte deactivation by M. vaccae

We have previously reported that a short contact of monocytes with tumour cells leads to selective monocyte unresponsiveness to the same or different tumour cells, but not to LPS (deactivation), defined by decreased production of proinflammatory cytokines (TNF-α, IL-12) and enhanced release of IL-10 [18]. In these experiments we asked whether pretreatment of monocytes with M. vaccae can prevent their unresponsiveness to the challenge in vitro with tumour cells. Monocytes were cultured for 2 h with M. vaccae, and then HPC-4 cells were added and cocultured for an additional 2 h. CD14+ monocytes were isolated from the coculture by FACS sorting and restimulated with tumour cells for 18 h, when supernatants were collected and tested for cytokine content. Untreated monocytes showed significantly depressed (approximately 30% of control monocyte response) secretion of TNF-α (Fig. 3). Treatment of monocytes before coculture with M. vaccae SN 920 and BCG, and to a lesser extend by M. vaccae MB 3683, significantly increased TNF-α secretion. The enhanced IL-10 release by cocultured and restimulated monocytes was decreased by treatment of monocytes with M. vaccae B 3808 and SN 920, and BCG (Fig. 4). However, strain MB 3683, which enhanced TNF-α secretion, did not caused significant inhibition of IL-10 secretion, hence no clear reciprocal relationship between the effect on TNF-α and IL-10 secretion was observed. These data imply that some strains of M. vaccae can prevent tumour cell–induced monocyte deactivation, as determined by TNF-α production.
Fig. 3

Secretion of TNF-α by monocytes treated with different strains of M. vaccae or BCG before coculture with tumour cells. Monocytes were cultured for 2 h with M. vaccae, and then HPC-4 cells were added and cocultured for an additional 2 h. CD14+ monocytes were isolated from the coculture by FACS sorting and restimulated with tumour cells for 18 h, when supernatants were collected and tested for cytokine content. The release of TNF-α by control monocytes stimulated with the bacteria was 2,687 ± 1,128 pg/ml; by control monocytes stimulated with tumour cells, 4,668 ± 1,388 pg/ml; and by cocultured monocytes (medium), 1,160 ± 1,241 pg/ml. Results are expressed as the percentage of response of control monocytes stimulated with tumour cells. Means ± SD of four independent experiments are shown. *Significant at p<0.05

Fig. 4

Secretion of IL-10 by monocytes treated with different strains of M. vaccae or BCG before coculture with tumour cells. Monocytes were treated as described for Fig. 3. The release of IL-10 by control monocytes stimulated with the bacteria was 1,709 ± 908 pg/ml; by control monocytes stimulated with tumour cells, 2,434 ± 916 pg/ml; and by cocultured monocytes (medium), 4,014 ± 1,045 pg/ml. Results are expressed as the percentage of the response of control monocytes stimulated with tumour cells. *Significant at p<0.05

Effect of treatment of deactivated monocytes with M. vaccae

To determine whether M. vaccae can reverse deactivation of monocytes, monocytes isolated from the coculture were treated with M. vaccae strains for 3 h, washed and then restimulated with tumour cells. Depressed TNF-α secretion was significantly enhanced by treatment with different strains of M. vaccae or BCG (Fig. 5), but only SN 920 and BCG increased IL-12 release (Fig. 6). None of M. vaccae strains, but BCG, decreased enhanced production of IL-10 (Fig. 7). It was concluded that tumour cell–induced monocyte unresponsiveness can only be efficiently reversed by treatment with M. vaccae SN 920 and BCG, as judged on their effect on TNF-α and IL-12 production.
Fig. 5

Secretion of TNF-α by monocytes treated with different strains of M. vaccae after coculture with tumour cells. Monocytes isolated from the coculture were treated either with M. vaccae strains or BCG for 3 h, washed and then restimulated with tumour cells for 18 h. Then supernatants were collected and tested for cytokine content. The TNF-α release by control monocytes stimulated with tumour cells was 3,262 ± 1,334 pg/ml, and by cocultured and restimulated monocytes (medium), 2,217 ± 1,067 pg/ml

Fig. 6

Secretion of IL-12 by monocytes treated with different strains of M. vaccae or BCG after coculture with tumour cells. Monocytes were treated as described for Fig. 5. The IL-12 release by control monocytes stimulated with tumour cells was 937 ± 536 pg/ml, and by cocultured and restimulated monocytes (medium), 200 ± 100 pg/ml

Fig. 7

Secretion of IL-10 by monocytes treated with different strains of M. vaccae after coculture with tumour cells. Monocytes were treated as described for Fig. 5. The IL-10 release by control monocytes stimulated with tumour cells was 2,500 ± 1,257 pg/ml, and by cocultured and restimulated monocytes (medium), 3,800 ± 1,706 pg/ml

Discussion

Clinical trials using BCG as an adjunct to chemotherapy in different types of cancer have yielded conflicting results. However, in superficial bladder cancer, BCG is an effective immunostimulator which provides superior treatment results to those from chemotherapy [9]. In vitro BCG enhances cytotoxic activity of monocytes [23], induces production of TNF-α [8] and, in patients with gastric cancer receiving BCG immunotherapy, enhances tumour cell–induced TNF-α production by mononuclear cells [31]. Therefore, BCG was used in the present study as a prototype activator of monocytes.

Patients with metastatic cancer exhibit a shift in the immune response toward a Th2 pattern, favouring humoral immunity and suppressing cellular immune response [3]. Work in both cancer and allergy has demonstrated the ability of M. vaccae to down-regulate IL-4 and up-regulate IL-2 production, with a switch toward a Th1 pattern of response [12, 31]. M. vaccae has been recently undergoing clinical trials as a potential immunological adjuvant in patients with cancer. The phase I/II trials have suggested the clinical efficacy of M. vaccae SRL172 as an anticancer agent in hormone-refractory prostate cancer [5], malignant melanoma, non–small cell lung cancer and malignant mesothelioma [1, 14, 19, 20]. This prompted the present study designed to determine how M. vaccae affects monocyte–tumour cell interactions in vitro, and in particular to define its effect on tumour cell–induced monocyte deactivation [18]. Three different strains of M. vaccae were used and their effect was compared with that of BCG (Moreau strain). This strain exhibited strong immunotherapeutic activity in experimental tumours [21] and showed improved survival of stage III gastric cancer patients when added to chemotherapy [22].

First, we studied the ability of M. vaccae to induce cytokine production. Data show that all strains of M. vaccae stimulated production of TNF-α and IL-10 to a comparable level that was substantially higher than that obtained with BCG. Induction of IL-12 was not detected. To our knowledge this is the first demonstration of the capacity of M. vaccae to induce cytokine production by human monocytes. Then the immunomodulatory effect on monocyte–tumour cell interactions was determined. Tumour cell–induced production of TNF-α was significantly enhanced by preactivation of monocytes with strain B 3805 (levels comparable to BCG), and to a lesser extent by preactivation with two other strains of M. vaccae. No changes in the secretion of IL-10 were observed. Furthermore, although BCG enhanced IL-12 production by monocytes stimulated with tumour cells, M. vaccae was without effect. These data indicated a somewhat preferential immunomodulatory effect of M. vaccae on TNF-α, but not other cytokines, secreted by monocytes stimulated with tumour cells. Although it is unknown what constituents of M. vaccae may be responsible for stimulation of monocytes, the differential effect of M. vaccae on TNF-α and IL-10 production may be related to the different regulation of their gene expression via different transcriptional factors [2, 10].

Substantial data demonstrate that tumour growth adversely affects macrophage function [4]. Our recent observations indicate that contact of monocytes with tumour cells during coculture leads to monocyte unresponsiveness to in vitro challenge with tumour cells. Such monocytes following restimulation show significantly compromised ability to produce TNF-α and IL-12, and enhanced secretion of IL-10. This phenomenon is called deactivation [18]. Otherwise, contact with tumour cells causes polarisation of monocytes toward a M2 (TNF-α, IL-12, IL-10+) phenotype [11]. With this in mind, in the next set of experiments, we asked whether treatment of monocytes with M. vaccae can prevent or reverse monocyte deactivation occurring following coculture with tumour cells. Treatment of monocytes with BCG, M. vaccae strains SN 920 and MB 3683, but not B 3805, enhanced depressed secretion of TNF-α, despite the fact that all stains of M. vaccae had a similar capacity to induce TNF-α production by control monocytes. BCG, and SN 920 and B 3805 strains, inhibited enhanced IL-10 production. Hence, there was no clear-cut reciprocal effect of M. vaccae B 3805 on IL-10 and TNF-α secretion. In this context, it is interesting that IL-12 production by BCG–cell wall skeleton (BCG-CWS)-stimulated monocytes from patients with lung cancer was up-regulated, while IL-10 was inhibited [13]. This is also in keeping with other findings indicating that BCG infection of monocytes is accompanied by impairment of IL-10 and enhancement of TNF-α secretion [7].

Treatment of monocytes after coculture (deactivated) with BCG and M. vaccae led to enhancement of depressed TNF-α secretion to a comparable extent by all strains used, although enhanced IL-10 production was inhibited by BCG only. Interestingly enough, only BCG and SN 920 significantly reversed the down-regulation of IL-12 secretion. This data again suggest that TNF-α and IL-10 production is differently regulated [2, 10]. Furthermore, different surface molecules may be responsible for activation of monocytes, e.g. for TNF-α production by tumour cells and Mycobacteria. While BCG-CWS or lipoarabinomannan activate Toll-like receptors 2 and 4 [6, 29], CD44 and MHC class II are involved in the interactions of monocytes with tumour cells [32].

In summary, the results of this study indicate that treatment of monocytes with some strains of M. vaccae before or after exposure to tumour cells can prevent or reverse their deactivation, but none of the strains show superior effect in comparison to BCG. It would be interesting to compare other M. vaccae strains (e.g. SRL 172) which showed promising preliminary results in recent clinical trials [1, 5, 12, 14].

TIMs are polarised toward M2 macrophages [11, 25]. If these observations were valid in vivo, treatment with M. vaccae might prevent or partly reverse this polarisation of TIMs.

Acknowledgements

This study was supported by the National Committee for Scientific Research (grant No. 6 PO5A 096 20). We wish to thank Prof. Leon Sedlaczek (Centre for Microbiology and Virology, Polish Academy of Sciences, Łódź, Poland) for kind donation of M. vaccae strains. We also thank Ms Barbara Hajto and Mariola Ożóg for skillful technical assistance.

Copyright information

© Springer-Verlag 2004