Article Text


Pancreatic cancer regression by intratumoural injection of live Streptococcus pyogenes in a syngeneic mouse model
  1. C Maletzki1,
  2. M Linnebacher2,
  3. B Kreikemeyer3,
  4. J Emmrich1
  1. 1
    Division of Gastroenterology, Department of Internal Medicine, University of Rostock, Rostock, Germany
  2. 2
    Department of General, Thoracic, Vascular and Transplantation Surgery, University of Rostock, Rostock, Germany
  3. 3
    Department of Medical Microbiology and Hospital Hygiene, Institute of Medical Microbiology, Virology and Hygiene, University of Rostock, Rostock, Germany
  1. C Maletzki, Division of Gastroenterology, Department of Internal Medicine, University of Rostock, Ernst-Heydemann-Str. 6, D-18055 Rostock, Germany; Claudia.Maletzki{at}


Background: This study addressed the potential of bacteriolytic therapy using Streptococcus pyogenes in a syngeneic pancreatic carcinoma mouse model.

Methods: Panc02 tumours were either infected with S pyogenes or were treated with the equivalent volume of vehicle. In addition to assessment of tumour histology and immunohistochemistry, isolated splenocytes were analysed by flow cytometry. Interferon (IFN) γ secretion as a reaction of splenocytes against tumour cells was shown through the ELISpot technique. A cytotoxic effect of lymphocytes against tumour targets was detected by lactate dehydrogenase (LDH) release. Cytokine levels in serum were measured.

Results: A single application of live bacteria into established Panc02 tumours resulted in complete tumour regression. This antitumoural effect was accompanied by massive leucocyte infiltration into the tumours as well as a significant and sustained elevation of systemic levels of the proinflammatory cytokines IFNγ, tumour necrosis factor α and interleukin 6. Lymphocytes obtained from treated mice specifically recognised syngeneic tumour cells in IFNγ-ELISpot, and most importantly in cellular cytotoxicity assays, indicating a tumour-specific immune response.

Conclusions: We provide data that both the direct lytic activity of S pyogenes towards tumour cells and the infection-driven infiltration of tumours by cells of the innate immune system lead to damage of tumour cells followed by a dissemination of tumour components. This last outcome allows for the activation of tumour-specific effector cells, most probably in draining lymph nodes, promoted by the proinflammatory context. Taken together, these data indicate that the application of live S pyogenes may be a promising new treatment strategy for advanced pancreatic cancer patients that warrants further investigation.

View Full Text

Statistics from

The aim of cancer immunotherapy is stimulation of the immune system to restore its ability to recognise and respond to residual tumour cells, and it may be a useful complement to conventional anticancer strategies.1 Immunotherapy began over 100 years ago when New York surgeon William B Coley observed that malignant tumours, particularly sarcoma, regressed in patients who developed concurrent bacterial infection after surgical resection.2 Thus, the inoculation of Coley’s toxin, a bacterial vaccine of Streptococcus pyogenes and Serratia marcescens, became the cornerstone in the development of immunotherapy for cancers. Patients who experienced infectious disease processes, including high fever and chills, also experienced a beneficial effect on malignancy and were finally cured of their tumours by the development of a potent immune response.2

Basing on this initial work, researchers have developed many different forms of immunotherapy, including the transfer of immune-active cells—for example, lymphokine-activated killer cells, the application of monoclonal antibodies (mAbs), the activation of specific effector cells by tumour cell lysates or extracts, and the activation of non-specific effector cells (macrophages, natural killer (NK) cells) by microbial or chemical adjuvants.1 Interestingly, instillation of viable attenuated mycobacteria—specifically BCG—is still successfully used in clinical treatment of superficial bladder cancer.3 Today, it is well known that both direct tumouricidal effects and activation of the immune system mediate these antitumour activities of bacteria or their components.46 Recently, several groups have developed novel antitumoural approaches using tumour-therapeutic bacterial strains. Yazawa et al have described the selective growth potential of the anaerobic Bifidobacterium longum in the hypoxic regions of several types of solid tumours.7 8 Similarly, Vogelstein and co-workers developed an efficient antitumoural therapy by combining the intravenous administration of lethal toxin-free spores of Clostridium novyi with standard chemotherapy.9 10 Quite recently, Zhao et al created an attenuated strain of Salmonella typhimurium that selectively grows in and kills prostate tumour cells in a nude mouse model.11

Given the dismal prognosis for patients with ductal adenocarcinoma of the pancreas,1 12 novel immunotherapeutic intervention possibilities must be developed to treat this malignancy.4 1316 Streptococcus pyogenes is an important facultative anaerobic pathogen that can cause infections of varying severity in humans, ranging from simple, non-suppurative infections of the pharynx and skin to life-threatening invasive infections.17 In the present study, we sought to address the question of whether bacteriolytic therapy using S pyogenes is applicable for pancreatic carcinoma. To achieve this goal, we analysed the impact of a single intratumoural injection of S pyogenes on established murine pancreatic tumours in a syngeneic mouse model. This single application of S pyogenes resulted in complete tumour regression within 4 weeks. A massive activation of immune response mechanisms secondary to infection accompanied the regression and contributed to eradication of the tumours.

In summary, S pyogenes may be an excellent candidate for the evaluation of an active antitumour therapy. These findings may be of special clinical interest for additive treatment of patients with pancreatic tumours.


Cell lines and mice

Panc02 cells (kindly provided by Dr Arne Scholz, Charite-University Medicine Berlin, Berlin, Germany) were originally established by Corbett and co-workers in 1984 by 3-methylcholanthrene treatment of pancreata of female C57Bl/6 mice.18 EL4 lymphoma cells and the rectal polyploid carcinoma cells CMT-93 were provided by the American Type Culture Collection (Manassas, VA). Fibroblast-like MC3T3-E1 cells were purchased from the Deutsche Sammlung von Mikroorganismen und Zellkulturen (Braunschweig, Germany). All cells were maintained in Dulbecco’s modified Eagle’s medium (DMEM; Biochrom AG, Berlin, Germany), supplemented with 10% fetal calf serum (PAA, Cölbe, Germany) and 1% penicillin/streptomycin (Invitrogen, Karlsruhe, Germany), and incubated at 37°C in a humidified atmosphere of 5% CO2. Female 8–10-week-old C57Bl/6N mice were purchased from Charles River Inc. (Sulzfeld, Germany). Animals were exposed to cycles of 12 h light/12 h darkness and received standard food and water ad libitum. Upon approval by the local animal welfare committee, experiments were performed in accordance with the German legislation on protection of animals and the “Guide for the Care and Use of Laboratory Animals”.

Bacterial strain and culture conditions

A single group A streptococcal isolate, S pyogenes serotype M49 (strain 591), cultured in Todd–Hewitt (TH) broth or on TH agar (Oxoid Unipath, Wesel, Germany), both supplemented with 0.5% yeast extract (THY), was used in this study. For intratumoural infection, S pyogenes was grown in TH broth overnight to mid-log phase and adjusted to a concentration of approximately 1×106 colony-forming units (cfu)/ml on the basis of an optical density reading at 600 nm and on plating analysis.

Animal tumour model and bacterial application

Under brief ether anaesthesia, female C57Bl/6N mice were injected subcutaneously into the right hind flank with 1×106 Panc02 cells. Volumes of outgrowing tumours were evaluated according to the following formula: width2×length×0.52.9 When tumours reached a size of approximately 60 mm3, about 9–12 days post-tumour cell inoculation, one group was infected intratumourally with 1×106 cfu/ml S pyogenes (50 μl volume, dissolved in phosphate-buffered saline (PBS), n = 28). Animals that received equivalent volumes of solvent alone served as vehicle-treated controls (vehicle, n = 6). A group without intratumoural injection served as the tumour-carrying control (tumour, n = 6). An additional group of tumour-free animals was treated with 1×106 cfu/ml S pyogenes intracutaneously into the right hind flank (M49-C, n = 6). Physiological control values were obtained from animals without any intervention (control, n = 3). Mice were sacrificed 7, 14, 21 and 28 days after bacterial infection or when they became moribund before the tumour volume reached 1500 mm3. At the end of each experiment, blood samples, tumour, spleen and mesenteric lymph nodes were removed from the animals of all groups for further analysis. To estimate the frequency of systemic spread of S pyogenes, blood samples from all infected animals were taken at the given time points (day 7, 14, 21 and 28) and plated on sheep blood agar.

Rechallenge experiment

Four weeks after immunotherapy with S pyogenes, sucessfully treated and M49-C animals (each n = 6) received a tumourigenic dose of Panc02 tumour cells (1×106) into the flank, contralateral to the first tumour application side. Tumour volumes were measured twice a week as described above for 4 weeks before the animals were euthanised.

Apoptosis assay

For detection of apoptosis, cells were trypsinised and subsequently lysed. For quantification of caspase 3 activity, the BD ApoAlert Caspase Assay plate system was used according to the manufacturer’s instructions (BD Biosciences, Heidelberg, Germany). Caspase 3 activity was determined by fluorometric detection on a Cytofluor 2300 (Millipore, Schwalbach, Germany, ex/em: 380/460 nm).19 Additionally, apoptosis was quantified by flow cytometric sub-G1 peak analysis as described before.20

Histology and immunohistochemistry

Leucocytes in resected tumour tissues were stained by the CAE (chloroacetate esterase) technique. The numbers of leucocytes/mm2 were determined in blinded counts by positive staining and morphology in 50 consecutive high power fields (HPFs). To confirm leucocyte cell infiltration into the tumour mass, tissue sections were incubated with goat anti-mouse CD4 (1:100, Santa Cruz Biotechnology, Heidelberg, Germany), rat anti-mouse CD8 (1:200, Santa Cruz Biotechnology) and rabbit anti-mouse CD88 (1:200, Santa Cruz) antibodies for 1 h at room temperature, followed by an alkaline phosphatase-conjugated mouse antibody (Santa Cruz). The sites of phosphatase binding were detected using fuchsin as the chromogen. The sections were counterstained with haematoxylin and examined by light microscopy (Axioskop 40, Zeiss, Oberkochen, Germany).

Flow cytometry

Splenocytes were isolated at different time points from infected and non-infected animals using Ficoll-Paque PLUS (Amersham Biosciences, Uppsala, Sweden) gradient centrifugation. Subsequently, cells were labelled with the following fluorescein isothiocyanate (FITC)-conjugated rat anti-mouse mAbs: CD4, CD8, CD11b, Gr1, CD62L (1:20, Miltenyi Biotec, Bergisch-Gladbach, Germany), CD25 and CD71 (1:20, BD Pharmingen). Samples were analysed on a FACSCalibur Cytometer (BD Pharmingen) using the CellQuest software.

ELISpot assay for interferon γ-secreting lymphocytes

Interferon γ (IFNγ)-specific mAb- (Mabtech, Hamburg, Germany) coated, 96-well microtitre plates were filled with 1×104 target cells/well (Panc02, EL4, CMT-93, MC3T3-E1 and peripheral blood mononuclear cells (PBMCs)) and incubated for 2 h. Splenocytes (105) were given to the targets and co-cultured overnight. Finally, bound antibody was visualised by 5-bromo-4-chloro-3-indolyl phosphate/nitro blue tetrazolium (BCIP/NBT; KPL, Gaithersburg, MD), and spots were counted using a dissection microscope (Zeiss, Oberkochen, Germany). Presented are the numbers of IFNγ-secreting cells per 105 effector cells corrected for background levels counted in the absence of target cells, which usually was between 10 and 20 spots/105 cells. Target cells without effector cells showed no background level.

Lactate dehydrogenase cytotoxicity assay

The colorimetric CytoTox-One Homogeneous Membrane Integrity Assay (Promega, Madison, WI) was used evaluating lactate dehydrogenase (LDH) release from lysed cells. Experiments were performed according to the manufacturer’s protocol. Lytic activity of lymphocytes from mesenteric lymph nodes against target cells (Panc02, EL4, CMT-93, MC3T3-E1 and PBMCs) was measured after overnight incubation at an effector:target cell ratio of 30:1. Maximum LDH release was achieved by incubation of the target cells with the supplied lysis buffer. Target cells without effector cells were used as negative controls. Fluorescence was measured on a Cytofluor 2300 (Millipore, ex/em: 530/620 nm). Cytotoxicity was calculated according to the following formula: % cytotoxicity  = 100× (experimental release – spontaneous release of effector cells – spontaneous release of target cells)/(maximum LDH release – spontaneous release (negative control) of target cells).21 22

Bio-Plex Protein Array system

A panel of serum cytokines was measured in duplicate using the Bio-Plex Protein Array system (BioRad, Munich, Germany), according to the manufacturer’s instructions. With the Bio-Plex cytokine assay kit in combination with the Bio-Plex Manager Software, serum IFNγ, tumour necrosis factor α (TNFα), interleukin (IL) 3, IL5, IL6, granulocyte–macrophage colony-stimulating factor (GM-CSF) and granulocyte colony-stimulating factor (G-CSF) were assessed throughout the time course of the infection. Values of the respective serum cytokine levels of untreated control mice were set as 1, and all other data were given as an x-fold increase.


All values are expressed as mean (SEM). Because no significant difference was measured in any of the evaluated parameters between the vehicle-treated animals and the tumour-carrying and M49-infected tumour-free mice at different time points, only data from one time point (day 28, n = 6) were taken for graphic representation and statistical evaluation. Data from untreated control animals (n = 3) served as physiological values and are displayed in the graphs; however, they were not considered for statistical analysis between groups. Similarly, tumour-free S pyogenes-infected animals served only as internal control for bacterial infection. After proving the assumption of normality and equal variance, comparisons among groups were made using analysis of variance (ANOVA) followed by the appropriate post hoc comparison test (SigmaStat, Jandel Corporation, San Rafael, CA). The criterion for significance was taken to be p<0.05.


Choice of bacterial strain and dose of infection

First, we investigated the effect of different S pyogenes strains on Panc02 pancreatic carcinoma cells in vitro (multiplicity of infection (MOI) = 25)). Most of the strains tested did not mediate significant growth inhibition or killing activity. However, caspase 3 activity tests and sub-G1 peak analysis revealed that S pyogenes serotype M49 strain 591 exerted severe cytotoxic reactivity that is most probably due to the induction of apoptosis towards pancreatic carcinoma cells of murine (Panc02) and human (Panc1, BxPC3, Capan2) origin (fig 1 and data not shown). Consequently, S pyogenes serotype M49 strain 591 was chosen for the evaluation of bacteriolytic antitumoural therapy in vivo. Bacteria were applied intratumourally to maximise local antitumoural impact and minimise potential systemic cytotoxicity.

Figure 1 Quantitative analysis of apoptosis. (A) Increased caspase 3 activity in Panc02 cells after infection with S pyogenes M49 wild type (wt), indicative for induction of apoptosis, which was confirmed by (B) flow cytometric sub-G1 peak analysis. Cells with DNA content lower than the G1 peak were considered to be apoptotic. The results show data from three separate experiments. (C) Panc02 cells show characteristic sub-G1 peaks at 4 h postinfection (representative results out of three independent experiments). Mean (SEM); *p<0.05 vs control; **p<0.001 vs control as calculated by analysis of variance.

Next, we sought to evaluate the optimal therapeutic, non-toxic infection number. Increasing amounts of bacteria were administered (105, 106 and 107 cfu/ml). Animals of the group treated with the highest bacterial dose rapidly developed severe signs of systemic infection and died (fig 2). In contrast, only two animals (15%) died from administration of S pyogenes at the middle (106 cfu/ml) and the low (105 cfu/ml) bacterial doses, respectively, even 4 weeks postinfection. For in vivo application, the highest tolerable dose was chosen (106 cfu/ml).

Figure 2 Survival curve of C57Bl/6N mice with heterotopic syngeneic pancreatic carcinoma. After tumour establishment, mice received a single intratumoural injection of S pyogenes and were monitored for 28 days with adherence to ethical requirements. For evaluation of the optimal therapeutic dose, bacteria were injected in three different doses. Experiments included at least five mice per group.

Gross findings after infection with S pyogenes

The intratumoural administration of S pyogenes did not affect pancreatic carcinoma growth within the first 4 days. Palpable tumours continued to grow and reached an average size of 128.6 (11.8) mm3), which was comparable with the tumour sizes of vehicle-treated animals (151.9 (11.8) mm3). Thereafter, about 6–8 days after infection, tumours in the infection group became noticeably and quantitatively smaller than those of the control groups (p<0.05 vs control and vehicle; fig 3) and were frequently ulcerous and apparently necrotised 10–14 days after infection (fig 3). This finally resulted in nearly complete regression within 4 weeks.

Figure 3 Growth kinetics of tumours in animals at 7, 14, 21 and 28 days after intratumoural bacterial infection and after rechallenge. Animals received a tumourigenic dose of Panc02 cells. Naïve M49-C animals were used as controls. Mice that had undergone bacteriolytic therapy were partially protected from re-exposure to Panc02 cells. Mean (SEM); *p<0.05 vs tumour; †p<0.05 vs vehicle; §p<0.05 vs M49-C rechallenge as calculated by analysis of variance.

To analyse the frequency of systemic spread of S pyogenes, we plated blood samples from all infected animals at the given time points on sheep blood agar, identifying a systemic bacterial infection in 2 of the 28 tumour-carrying mice and in none of the 12 tumour-free, M49 control animals. The two systemically infected animals displayed anorexia, weight loss and ataxia. Because of the severity of infection, these two animals were euthanised and related data excluded.

Following the observation periods (1, 2, 3 and 4 weeks), animals were sacrificed. Macroscopic examination of organs revealed no side effects such as necrosis or haemorrhage, but there was massive splenomegaly only in tumour-carrying mice that had undergone bacteriolytic therapy (an approximately seven- to eightfold increase in size at all time points; data not shown). Interestingly, spleens of tumour-free mice infected with S pyogenes were macroscopically not enlarged.

Similarly, on day 14 postinfection, animals in the treatment group exclusively exhibited a transient leucocytosis (p<0.05 vs control and vehicle groups, respectively) as well as a sustained fall in haematocrit levels (p<0.05 vs control and vehicle) (table 1). Sustained low haematocrit levels are negatively correlated with spleen sizes.23 In this case, a decreasing haematocrit could indicate leukaemic volume expansion.

Table 1 Systemic blood cell count and haematocrit in control groups (control, M49-C, tumour, vehicle) and 7, 14, 21 and 28 days after S pyogenes infection

Tumour rechallenge demonstrates partial protection

In order to test to what extent the bacteriolytic therapy not only led to tumour regression but also induced protective immunological memory towards a re-exposure to Panc02 tumour cells, a tumour rechallenge experiment was performed. Groups consisted of animals that were tumour-free post-treatment and of naïve M49-C mice. Both received a tumourigenic dose of Panc02 cells into the contralateral flank. Four weeks after tumour inoculation, the M49-C mice displayed growing tumours (889.5 (245.7) mm3) (fig 3). In the rechallenge group, tumour development was significantly restrained (226.6 (71.3) mm3; p<0.05) (fig 3), with one animal remaining tumour-free until the end of the experiment. Taken together, this indicates, that bacteriolytic therapy using M49 mediates potent antitumoural reactivity in vivo. However, only one of six animals was completely protected from re-exposure with Panc02 tumour cells.

Infiltration of leucocytes into tumours following intratumoural injection of S pyogenes

To assess the degree of leucocytic infiltration as a measure of local antitumoural response after infection with S pyogenes, tumour tissues were first studied by CAE histology (fig 4). These experiments showed, in accordance with systemic leucocytosis, positive staining of individual leucocytes, with the highest values at days 7 and 14 after bacterial infection (p<0.05 vs control and vehicle, fig 4C), but the values remained at least three times higher than in control animals.

Figure 4 Representative images of leucocytic infiltration of pancreatic tissue after vehicle treatment (A) or at day 7 (B) after S pyogenes infection. (C) Quantitative analysis of chloroacetate esterase (CAE)-positive leucocytes infiltrating the tumour tissue. Data from control groups are representative for other time points (original magnification, ×40). Mean (SEM); *p<0.05 vs tumour; †p<0.05 vs vehicle as calculated by analysis of variance.

As shown in fig 5, the majority of these infiltrating cells were granulocytes (CD88-positive) and CD4+ T cells, while cytotoxic T cells (CD8-positive) were found markedly within necrotic regions of the tumour (fig 5). Moreover, inflammatory cells were only rarely detected in control tumours and, among them, there were no CD4+ or CD8+ T cells or CD88+ granulocytes.

Figure 5 Representative images of CD4, CD8 and CD88 immunohistochemical staining of pancreatic tumour tissue after vehicle treatment (A, D, G) or at day 7 (B, E, H) and day 21 (C, F, I) after S pyogenes infection (106 cfu/ml, n = 7 per time point). Note the massive accumulation of CD4-positive T-helper cells and granulocytes within bacterially infected tumour tissue (original magnification, ×40).

Identification of splenocyte subsets by flow cytometric analysis

Next, we analysed the composition of splenic immune cells (table 2). We observed an elevation of macrophages and monocytes (CD11b+) as well as mature granulocytes (Gr1+) within the first 2 weeks of infection (table 2). Numbers of transferrin receptor-positive (CD71) cells were significantly raised, whereas expression of L-selectin (CD62L) was gradually downregulated. Levels of CD4+ (T helper), CD8+ (cytotoxic) and CD25+ T cells were slightly elevated during the first 3 weeks postinfection (table 2). However, the dramatic decrease of CD4+ as well as CD8+ T cells at day 28 possibly reflected migration of immunological effector cells more into lymphoid tissues than into the spleen.

Table 2 Analysis of splenocyte subsets in control groups (control, M49-C, tumour, vehicle) and 7, 14, 21 and 28 days after S pyogenes infection (%)

Cytokine level changes as a consequence of infection

The data described above indicated an inflammatory reaction in the tumour tissue of M49-treated animals. Therefore, we wondered whether a systemic effect could be measured in addition to these local phenomena. Serum cytokine levels were determined from infected and non-infected animals. Analysis of neutrophil chemotaxis polypeptide G-CSF showed a 158.5 (33.8) and a 21.1 (4.6) increase at days 3 and 7, respectively, after S pyogenes treatment when compared with control values (p<0.001, fig 6). In addition, we found significant increases for IFNγ and IL3 at day 7, with return to near-physiological values of all three Th1 cytokines at later time points (table 3), indicative of a strong proinflammatory stimulus, especially during the early phases of S pyogenes infection. A similar pattern was seen for TNFα (table 3). On day 14, S pyogenes infection caused a fivefold increase in IL5 and a 15-fold increase in IL6, followed by a threefold rise of GM-CSF levels on day 21 (table 3).

Figure 6 Kinetics of the cytokine granulocyte colony-stimulating factor (G-CSF) during the experimental time course. Animals that were infected intratumourally with S pyogenes showed a massive increase of G-CSF at day 3 after treatment, while no significant differences were measured in control animals. Results show data from three mice per time point. Mean (SEM); *p<0.05 vs tumour; †p<0.05 vs vehicle as calculated by analysis of variance.
Table 3 Serum cytokine activities in control groups (control, M49-C, tumour, vehicle) and 7, 14, 21 and 28 days after S pyogenes infection

Lymphocytes of infected animals specifically recognise tumour cells

We next examined whether immune cells respond to stimulation from tumour cells. First, the reaction of splenocytes co-incubated overnight with different target cell populations was measured in IFNγ-ELISpot assays. Panc02 tumour cells triggered IFNγ release exclusively in splenocytes obtained from M49-treated animals. However, this reaction was statistically significant only at day 14 postinfection (p<0.05 vs control and vehicle; fig 7A). Interestingly, similar results were obtained with other syngeneic tumour cell lines, such as the lymphoma cell line EL4 and the colorectal carcinoma cells CMT-93 (fig 7A). We next determined the reactivity of splenocytes to non-cancerous control cell lines to exclude an unspecific response. Regarding the IFNγ release, none of the leucocytes showed reactivity against syngeneic MC3T3-E1 fibroblasts or PBMCs (fig 7B).

Figure 7 Quantitative analysis of interferon (IFN)γ-ELISpot assay. The reactivity of splenocytes against tumour targets was tested after co-incubation overnight at an effector:target cell ratio of 30:1. Splenocytes from infected animals at 7, 14, 21 and 28 days after intratumoural bacterial infection showed a significantly higher reactivity against target cells than those from control animals (day 28). Data from control groups are representative for other time points. Surprisingly, the reactivity of splenocytes against Panc02, EL4 and CMT-93 cells (A) decreased from days 7 to 28 after infection. Furthermore, lymphocytes showed no reactivity against MC3T3-E1 fibroblasts or peripheral blood mononuclear cells (PBMCs) (B) during the time period of infection, underscoring a tumour cell-specific activity. The results show data from three separate experiments. Mean (SEM); *p<0.05 vs tumour; †p<0.05 vs vehicle as calculated by analysis of variance.

In more functional cytotoxicity assays, we used lymphocytes obtained from lymph nodes as effector cells. Again, lymphocytes of S pyogenes-treated mice recognised Panc02 tumour cells, but lymphocytes from control animals did not show lytic activity. It is notable that the killing activity of lymphocytes from the tumour-infected group continuously increased postinfection (p<0.05 vs control and vehicle; fig 8A). Moreover, other syngeneic tumour cell lines (EL4 and CMT-93) were also effectively lysed (fig 8A). Tumour specificity could be confirmed by absent or marginal cytotoxicity against non-cancerous control cells (MC3T3-E1 and PBMCs) (fig 8B). Therefore, the recognition of tumour cells in the previous experiments is specific and suggests the induction of a specific immune response that contributes to the observed tumour regression in S pyogenes-infected animals.

Figure 8 Quantitative analysis of cytotoxicity using the lactate dehydrogenase release assay. Lymphocytes were isolated from mesenteric lymph nodes and co-cultured with targets for 24 h at an effector:target cell ratio of 30:1. Lymphocytes from infected animals at 7, 14, 21 and 28 days after intratumoural bacterial infection showed a significantly higher lytic activity against target cells than those from control animals (day 28). Cytotoxicity of lymphocytes against (A) Panc02, EL4 and CMT-93 cells increased from days 7 to 28 after infection, up to >50%. In contrast, no lytic activity of lymphocytes against (B) non-cancerous MC3T3-E1 cells or peripheral blood mononuclear cells (PBMCs) was seen. Data from control groups are representative for other time points. The results show data from three separate experiments. Mean (SEM); *p<0.05 vs tumour; †p<0.05 vs vehicle as calculated by analysis of variance.


Given the poor prognosis of patients with advanced pancreatic cancer and the unsatisfactory results of standard therapeutic interventions, evolving new treatment strategies is imperative. Therefore, we explored the antitumoural potential of a bacteriolytic therapy based on facultative anaerobic bacteria. We chose S pyogenes as a model bacterial strain because it can exert direct cytolytic effects on eukaryotic cells,24 25 mediated by a variety of cytotoxic enzymes and pore-forming toxins.26 27 Among these, the haemolysins streptolysin S and streptolysin O seem to play a pivotal role.25 28 29 In preliminary in vitro investigations, we found that S pyogenes indeed mediates severe pancreatic carcinoma cell injury (unpublished data). Another reason for using S pyogenes is the exceptional sensitivity of streptococci to penicillin, which allows for easy eradication following therapeutic application.17 As a model of pancreatic carcinoma, we chose the aggressively growing, poorly immunogenic Panc02 mouse model.18 These tumours show intrinsic insensitivity to anticancer agents.18 3032

There are important and relevant features common to infection with S pyogenes in mice and human infection: (1) rapid cell damage after infection, accompanied by ulceration and necrosis of the infected tissue; (2) significant and sustained elevation of proinflammatory cytokines (TNFα, IFNγ) and chemokines (G-CSF); (3) high influx of inflammatory cells of the innate immune system (eg, neutrophils, macrophages and NK cells) into the focus of infection; and (4) activation of an adaptive, cell-mediated immune response.16 3335 Moreover, intravenous administration of live bacteria causes severe toxicity.25 36 Given these facts and the previously mentioned requirement for direct bacterial and tumour cell contact, we focused on local, intratumoural injection, rather than on systemic application.

A notable result of the current experiments was the complete tumour regression within 4 weeks subsequent to intratumoural administration of live S pyogenes. This antitumoural effect suggests strong participation of the innate arm of the immune system. In support of this interpretation, it has been found that during the initial phase of S pyogenes infection, innate immune cells—that is, granulocytes, monocytes and macrophages—are the major responding effectors recruited to the site of infection.9 28 33 34 Our own observations strongly argue for an ongoing inflammatory response: these observations include granulocytes (CD88-positive) that massively infiltrate infected tumours; significant increases in splenic granulocytes and in monocytes and macrophages (CD11b-positive); and systemic production of proinflammatory cytokines such as IL3, IFNγ and TNFα. The primary goal of this inflammatory response is the prevention of a systemic spread of bacteria,9 mainly by the very effective phagocytic activity of murine macrophages, which are especially capable of killing streptococci.28 36

The functional in vitro analysis showed that the specific immune response in terms of IFNγ release was restricted to immune cells obtained from M49 bacteria-treated tumour-bearing animals. This finding is additionally supported by the lymphocytes’ potential to recognise Panc02 cells in cytotoxicity assays. Tumour specificity was further emphasised by the lack of reactivity to non-cancerous control cells. Interestingly, we also observed a cross-reactivity of those lymphocytes with cells of other syngeneic tumour entities, possibly attributable to the presence of lymphocytes recognising shared tumour antigens.

To explain our findings, we propose the following model. Intratumourally administered S pyogenes induce an infiltration of inflammatory cells from the peripheral blood into the infected tissue,29 37 where they take up infected cells to clear the bacterial infection. In draining lymph nodes, the cellular debris components are presented to other immune cells. As a consequence of the ongoing inflammation and the activated status of presenting cells, specific immune responses may easily be initiated. It is conceivable that in this context, tumour-specific immune responses are also triggered.37 Consequently, animals are at least partially protected from a tumour rechallenge. However, the partial nature of this protective immunity further hints at the necessity for strong inflammatory stimuli to eradicate the tumour burden completely.

To the best of our knowledge, we are the first to show that S pyogenes infection initially activates the innate immune system and results in generation of tumour-specific effector cells, contributing to pancreatic carcinoma regression. This hypothesis is further supported by findings of other groups that used an inactivated S pyogenes preparation (OK-432) for treatment of several types of experimental tumours. The usefulness of OK-432 for in vivo maturation of dendritic cells leading to induction of oral tumour-specific cytotoxic T lymphocytes when combined with the application of chemotherapeutic agents has been demonstrated.38 Likewise, Ono et al proved the usefulness of OK-432 as a vaccination adjuvant in combination with a tumour-specific T cell epitope.39 Contrary to our findings using a single intratumoural injection of live bacteria, the experimental tumours of the former studies did not completely regress subsequent to therapy. Impressive tumour eradication rates similar to the ones observed by us were, however, reported by Avogadri et al using attenuated Salmonella.37 Interestingly, the highest antitumoural reactivity was strictly dependent on an extended repertoire of Salmonella-specific T cells accomplished by a prevaccination protocol.

Taken together, we have shown that bacteriolytic immunotherapy using facultative anaerobic bacteria such as S pyogenes is a promising new strategy for the treatment of advanced pancreatic carcinoma that warrants further investigation.


We thank Dr B Krueger for providing access to the Cytofluor 2300.


View Abstract


  • Competing interests: None declared.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • Digest
    Robin Spiller Magnus Simren