Microorganisms and Cancer: Quest for a Therapy

Many microorganisms are known to cause cancer. Some examples are Helicobacter pylori, which causes gastric cancer in humans and animals (6), and Agrobacterium tumefaciens, which causes crown gall in plants (37). The use of microorganisms or their products in the treatment of cancer is, however, less widely known, particularly among the readership of the Journal of Bacteriology, as most of these papers are published in more medically oriented journals and very few relevant papers have been published in the Journal of Bacteriology.

Live bacteria in the treatment of cancer.

The use of bacteria or their extracts in the treatment of cancer goes back more than 100 years. The most cited case is that by the physician and surgeon William B. Coley of the then Memorial Hospital in New York City, now called Memorial Sloan-Kettering Hospital, who observed that many of his patients with various forms of cancer had their tumors regress when they were infected with bacterial pathogens. Treatment to eliminate the infections allowed the cancer to come back (5, 19). He developed a treatment modality by making extracts of some of these bacteria, later described as Coley's toxin, which he used to help shrink the tumors in his patients (5, 19). Subsequently, many bacteria have been used in an effort to reduce the growth rate or size of tumors. The most prominent example would be the use of Mycobacterium bovis BCG, the vaccine strain, in the treatment of bladder cancer. Several well-coordinated studies have shown a clear relationship between the use of M. bovis BCG immunoprophylaxis after surgical removal of the tumor and the decreased recurrence rate or the delayed period during which recurrence could occur (15, 18). The long-term use of high titers of BCG with booster shots poses problems, such as side effects, lack of predictability of its effectiveness, and rare cases of sepsis, leading to the deaths of patients, so new ways of administering BCG have often been discussed (20). The mode of action of BCG to exert its antineoplastic effect is believed to be due to its effect on the immune system, with CD4 and CD8 T lymphocytes playing a major role (27), although the intravesical instillation of BCG is thought to result in nonspecific cystitis, which is likely accompanied by the local production of cytokines and the accumulation of inflammatory cells that are more damaging to the malignant rather than normal cells (2). The requirement of live cells of BCG for its anticancer activity is reflected in the fact that monocytes and T helper type 1 cells are often important for its effectiveness (34) and that large doses of vitamins that presumably enhance BCG survival and proliferation have a positive effect on the treatment of bladder cancer in human clinical trials (16). Similar positive effects have also been reported to be seen with immunoadjuvants such as tumor necrosis factor alpha (12). Using a human in vitro system to analyze the role of NK cells in BCG-induced cellular cytotoxicity, Brandau et al. (3) treated mononuclear cells with BCG for 7 days and demonstrated the ability of the BCG-activated killer cells to significantly destroy bladder tumor cells. Similarly, using C57BL/6 wild-type mice, NK-deficient beige mice, and mice treated with anti-NK1.1 monoclonal antibody, these authors noticed that viable BCG cells significantly prolonged survival in wild-type mice compared with control nontreated mice while BCG therapy was completely ineffective in NK-deficient beige mice or in mice treated with anti-NK1.1 monoclonal antibody (3). These studies demonstrated a key role for NK cells in BCG immunotherapy. BCG cell wall skeleton stimulation of the human innate immune system, as judged by altered levels of interleukin-12 (IL-12), IL-18, IL-10, and gamma interferon in the blood lymphocytes of patients with lung cancer, also demonstrated a role of BCG in the modulation of immune system activation (17). It should be emphasized, however, that the cancer that shows the maximal response to BCG treatment is superficial bladder cancer. Similar effects that are statistically significant have not been observed with other forms of cancer such as lung cancer (30) or melanoma (1). The reasons for such differential effects are not clear, although it is possible that certain forms of malignancies such as bladder cancer are more susceptible to BCG treatment because of cytokine network modulation or of the functional T-lymphocyte receptor repertoire expressed by surrounding tissues, recruited cells, or the malignant cells themselves.

While M. bovis BCG has been used to bolster the immune system against specific cancers such as bladder cancer, attenuated bacterial vaccine vectors such as Listeria monocytogenes and Salmonella enterica serovar Typhimurium that target the antigen-presenting cells or are powerful inducers of an innate immune response and immune mediators such as IL-12 have also been recommended for use in cancer prevention and therapy (21, 22). However, facultative anaerobic bacteria such as Salmonella are also known to target tumor cells for growth and proliferation (10) and attenuated Salmonella mutants, when injected intraperitoneally in plasmacytoma-bearing mice, were reported to allow tumor regression and prolonged survival of the mice (8). Similar use of Shigella (31) and Clostridia (9) species for the delivery of therapeutic agents to target cells,including tumor cells, has been reported. S. enterica serovar Typhimurium by far has attracted considerable attention because it is a facultative anaerobe capable of growing under both aerobic and anaerobic conditions and is highly manipulable genetically. Salmonella also targets tumors specifically for growth, with titers in tumors far exceeding those in the liver, the normal site of Salmonella replication in non-tumor-bearing mice. With attenuated Salmonella, the tumor-liver titers were shown to range between 250:1 and 9,000:1 (25). Such a propensity of Salmonella to target tumor cells for proliferation was shown to allow tumor regression, prolonging the survival of melanoma-tumor-harboring mice when injected intraperitoneally into the mice (25, 33). To investigate the role of Salmonella pathogenicity islands in tumor regression, Pawelek et al. (24) examined the effect of various Salmonella mutants defective in pathogenicity island genes on systemic invasion and survival in various cell types in mice. They demonstrated that Salmonella pathogenicity island 2 is essential for Salmonella antitumor effects, perhaps by aiding bacterial amplification within the tumors. The oral route of vaccination via Salmonella as the vector was used by Weth et al. (36) to compare a cytotoxic-T-lymphocyte- versus T-helper-cell-directed vaccination approach. Mice were vaccinated either orally with attenuated S. enterica serovar Typhimurium or subcutaneously with dendritic cells loaded with gp100 peptides predicted to bind to H2-kb or H2-Db molecules. Salmonella cells were transformed with murine gp100 cDNA (SL-gp100) or with a fusion construct of gp100 and a fragment of invariant chain cDNA (SL-gp00/Ii). Retardation of murine B16F1 melanoma growth was more efficiently achieved by vaccination with SL-gp100 than with the dendritic cells. Indeed, the efficacy of vaccination was limited by tumor-induced immunosuppression.

The ability of Salmonella, Clostridia, and other anaerobic bacteria to target tumors for their preferential replication, leading to tumor regression, was taken advantage of in a very interesting study reported by Dang et al. (7). It was previously demonstrated that a protozoan parasite, Toxoplasma gondii, when injected into melanoma-bearing mice, caused tumor regression by blocking angiogenesis (formation of blood vessels). It was thought that inhibition of angiogenesis in the tumor was due to the production of infection-induced antiangiogenic soluble factor(s) that created hypoxic conditions in the tumors, leading to their necrosis (13). Since various anaerobic bacteria have been reported to allow tumor regression (7, 14) and since such bacteria proliferate mainly in the anaerobic core of the tumors, Dang et al. (7) used the chemotherapeutic agent mitomycin C and the antivascular agent dolastatin-10 in combination with the spores of an attenuated anaerobic bacterium, Clostridium novyi, to treat colorectal cancer cells. The rationale behind this combined therapy (called combination bacteriolytic therapy or COBALT) was that, while the anaerobic bacteria grew in the anaerobic zone of the core of the tumors, the antivascular agent would create more extensive hypoxic areas for bacterial growth and starve the tumors of oxygen and essential nutrients while the chemotherapeutic agent attacked the tumor cells in the well-perfused, nonnecrotic outside cells of the tumors, leading to their total destruction (7, 14). The results of such a study were highly impressive. In the absence of the bacteria, but in the presence of mitomycin C and dolastatin-10, the tumors persisted for a much longer time and showed limited regression, while inclusion of the bacteria led to extensive disappearance of the tumors within a short period of time and, in some cases, complete dissolution of the tumors, leaving the animals tumor free. Similar results were shown with a melanoma tumor cell line (7). A downside of this combination therapy was the high level of toxicity, where 15 to 45% of the mice died within a few days after the treatment started, presumably due to the release of the highly toxic metabolic products of the disintegrating tumors (7).

Bacterial products in the treatment of cancer.

A lack of predictability of efficacy, a strong immune response, and the associated side effects and toxicity have often limited the use of live bacteria in the treatment of cancer. Preparations of microbial products, the lipopolysaccharide (LPS) vaccines in particular, have also been shown to have an anticancer property. As early as 1944, the endotoxin of Serratia marcescens was shown to be the “hemorrhage-producing factor” that promoted tumor regression (29). Many subsequent reports have shown the varying effectiveness of certain detoxified bacterial LPS preparations (23, 28, 32), with or without additional components, including some preparations of the LPS vaccines of Pseudomonas aeruginosa (4) that provided significant prolongation of remission and survival in patients with acute myelogenous leukemia compared to patients not treated with the LPS (4). A couple of macrolides, epothilone A and epothilone B (EpoA and EpoB), from the myxobacterium Sorangium cellulosum and particularly a chemically modified synthetic form of EpoB, desoxyepothilone B, have been shown to have antitumor activity against a range of human tumors (3a).

More recently, purified redox proteins such as azurin have been shown to allow cancer regression in nude mice harboring human melanoma (39). Azurin is a copper-containing oxidoreductase that is normally involved in denitrification in P. aeruginosa. Yamada et al. (39) demonstrated that azurin enters into the cytosol of a human melanoma cell line (UISO-Mel-2), is transported to the nucleus, and forms a complex with the tumor suppressor protein p53, thereby stabilizing it. Stabilization of p53 allows the significant generation of reactive oxygen species, which is a potent inducer of apoptosis (11, 38). Indeed, antioxidants that scavenge reactive oxygen species significantly reduce azurin-mediated cytotoxicity (11, 38). p53 is normally a labile protein with a half-life of a few minutes. It acts primarily as a transcriptional regulator within the cell, although its role in nontranscriptional processes and various instances of cell networking is also well known (35). The two primary functions of p53 within the cell are to allow growth arrest and to induce apoptosis in cells. Stabilization of p53 in azurin-treated cells enhances the level of intracellular p53, which triggers apoptosis in the melanoma cells xenotransplanted in nude mice, leading to their in vivo regression (39).

Concluding remarks.

While the ability of many attenuated bacteria to cause tumor regression has been well known for many years, the exact mode of their action was unknown. It has generally been believed that bacteria activate the immune system upon entry to the host cells and that the activated immune system then allows cancer regression. The experiments with T. gondii, however, indicated that the inhibition of angiogenesis, rather than an activated immune system, may contribute to cancer regression, at least in the case of T. gondii (13). While the proliferation of anaerobic bacteria in the core of the tumors is a well-established fact (7, 25, 33), there is very little understanding if such proliferating bacteria may also secrete soluble factors that could contribute to cancer cell death. The ability of purified azurin to mediate apoptosis in melanoma cells thus begs the question of whether other bacteria similarly secrete other proteins that may have cytotoxicity for cancer cells, somewhat similar to the many antibiotics that bacteria produce to kill other bacteria or fungi. Since azurin appears to act through complex formation with p53 and since 50% of human cancer cases have mutations in the p53 gene (35), it remains to be seen if azurin will have any cytotoxic effects towards these cancer cells as well. Similarly, it will be of great interest to bacteriologists and to microbiologists in general to explore why bacteria secrete redox proteins to kill mammalian cells. It has recently been hypothesized (26) that both mammalian cell mitochondria and present-day bacteria such as P. aeruginosa secrete redox proteins in response to certain signals (death signals for mitochondria; presence of certain eukaryotic proteins in the case of P. aeruginosa) that enter the mammalian cell cytosol, where they normally are not present. Mitochondria are believed to have prokaryotic ancestry, and therefore, the release of such redox proteins to effect eukaryotic cell death is an evolutionarily conserved function. The release of redox proteins, which are normally mitochondrion specific, in the cytosol of the eukaryotic cell signals the cell to an impending energy catastrophe, as mitochondria are the energy storehouse of the cell and any release of mitochondrion-specific proteins in the cytosol signals the loss of mitochondrial structural integrity. Since the redox activity of azurin (11) or other mammalian apoptogenic redox proteins such as AIF and PRG3 is not important for the induction of apoptosis, it has been hypothesized that the physical presence of the redox proteins in the cytosol is the signal for the triggering of apoptosis. The eukaryotic cell has evolved mechanisms that then initiate the apoptotic process through complex formation of the redox proteins with specific cytosolic proteins such as Apaf-1 and p53 (26). It would be of great interest to know if bacterial redox proteins other than azurin will enter the eukaryotic cell cytosol to trigger an apoptotic event in such cells, including human cancer cells.