Problems associated with conventional cancer therapy are numerous, but two stand out: lack of specificity and the inability to totally eradicate all cancerous cells. The former leads to severe, sometimes intolerable, adverse effects, whereas the latter contributes to the dismal prognosis for a variety of cancer types.
Effective cancer therapy should consist of components targeting both normoxic and hypoxic tumor tissues.
The progression of solid tumors requires a sufficient blood supply to deliver both nutrients and oxygen. However, the growth rate of malignant tumors often outpaces angiogenesis. In addition, the tumor vasculature is often poorly organized and leaky in nature. Consequently, the oxygenation of tumor tissues is not uniform. Well-oxygenated tumor cells lie close to vascular elements, necrotic cells are far away, and a gradient of hypoxia lies in between (Figure 1a). Confounding this complexity is the fact that the perfusion of tumors is dynamic, resulting in transiently hypoxic areas. Tumor hypoxia poses major problems for conventional cancer therapies, as radiation therapy and chemotherapy both rely on molecular oxygen and/or active replication of the target cells for their cytotoxic effects. Cancer cells in hypoxic areas are quiescent, making them resistant to such therapies. These cells can escape from the initial assault and start proliferating again once perfusion improves at a subsequent time, leading to relapse.
Anaerobic bacteria are attractive adjuncts to conventional therapies because anaerobes can selectively germinate and proliferate in the anoxic or hypoxic environments that are most resistant to drugs and radiation. The first anaerobic bacteriolytic cancer therapy was introduced more than half a century ago with a Clostridial species, C. histolyticus (Parker et al., 1947). To date, bacteria across four genera have been used, including Clostridium, Bifidobacterium, Salmonella, and Corynebacterium (Jain et al., 2001). Early studies involving two Clostridial species, C. histolyticus and C. butyricum (aka C. oncolyticum), showed promising results with respect to bacteria-mediated oncolysis (Parker et al., 1947; Mose et al., 1963). However, the majority of the animals died as the result of the infection. The surviving animals usually suffered from tumor recurrence at the same locations, indicating incomplete oncolysis. Later, clinical trials involving a limited number of advanced cancer patients were conducted with a C. butyricum strain (Carey et al., 1967). Again, bacterial germination in tumor masses and oncolysis were observed, but no cures resulted. These studies suggested that anaerobic bacteria alone are insufficient to eradicate cancers and/or that the bacterial strains used are unable to spread throughout the entire tumor.
In one of the more recent studies aimed at overcoming these problems, Clostridium novyi was shown to be capable of spreading more efficiently throughout the tumor (Dang et al., 2001). Moreover, these bacteria are exquisitely sensitive to oxygen, making it impossible for them to infect normal tissues. Infection of tumors with wild type C. novyi, however, proved lethal to animals because of the secretion of the systemically active alpha-toxin from bacteria that proliferated within the tumors. The C. novyi alpha-toxin gene is located on a phage episome (pseudolysogenous). This episome was eliminated and a clone devoid of the alpha-toxin gene, named C. novyi-NT (NT for non-toxic), was isolated. Clones of this strain proved to be safe when injected systemically into both healthy and tumor-bearing experimental animals. C. novyi-NT alone, however, did not usually cure animals of cancer, often sparing viable rims of well-perfused tumor tissue. To effectively eliminate tumor cells in these regions, cytotoxic therapeutics were employed, resulting in a therapeutic strategy that was named COBALT for Combination Bacteriolytic Therapy (Figure 1b).
In the initial study, microtubule-interacting agents, such as dolastatin-10, were found to be particularly useful in combination with C. novyi-NT (Dang et al., 2001). This beneficial interaction may result from the vascular collapse caused by the microtubule interacting agents, potentiating the hypoxic conditions at the tumor and therefore promotion of bacterial growth. When a single dose of C. novyi-NT spores was administered intravenously in combination with both mitomycin C and dolastatin-10, seven out of eight animals showed dramatic tumor regression and four out of five that survived the therapy became cancer free. This therapeutic regimen, however, was toxic, as three out of eight animals died from the treatment. In an effort to identify chemotherapeutic agents that were effective but less toxic when combined with C. novyi-NT, a panel of microtubule-interacting agents was screened for their ability to synergize with C. novyi-NT in experimental cancer treatment (Dang et al., 2004). Microtubule-interacting agents can be categorized into two classes, microtubule-stabilizers and -destabilizers. Drugs from both categories were found to be efficacious without causing significant toxicity, though different types of tumor responses were noted. When combined with C. novyi-NT, microtubule-destabilizers (HTI-286, vinorelbine) caused massive hemorrhagic necrosis followed by rapid tumor regression, whereas microtubule-stabilizers (docetaxel, MAC-321) caused progressive tumor shrinkage without apparent accompanying hemorrhagic necrosis. Though these combinations caused dramatic tumor regressions in most tumor models tested, the long-term responses varied significantly: while COBALT with HTI-286 or docetaxel cured 40%~100% of the animals bearing HCT116 (colorectal cancer) or HuCC-T1 (cholangiocarcinoma) xenografts, the cure rate dropped to <10% when the same regimens were applied to animals bearing HT29 (colorectal cancer) or CaPan-1 (pancreatic cancer) xenografts. The reasons underlying this heterogeneity are unknown but are likely to be related to the sensitivities of the various tumor models to the therapeutic drug used in combination with C. novyi-NT spores.
The success of radiation therapy is dependent on the level of oxygenation within tumors, as hypoxic cells are three-fold more resistant to ionizing radiation than normoxic cells. To overcome this obstacle, investigators have attempted to increase tumor oxygenation by various means and to use radiosensitizing agents. Because of their ability to destroy the hypoxic components of cancers, anaerobic bacteria provide an additional avenue for radiation enhancement. In a recent study, C. novyi-NT was used in conjunction with external beam radiation (Bettegowda et al., 2003). When fractionated doses of 5×2 Gy were applied, approximately 30% of the animals bearing HCT116 xenografts could be cured, compared to the 0% cure rate when radiation alone was administered. When a higher dose (50 Gy) was administered locally using brachytherapy (direct application of radiation therapy to affected tissue) in combination with a single intravenous injection of C. novyi-NT spores, all mice bearing HCT116 or HuCC-T1 xenografts were cured. Thus, C. novyi-NT can serve as an effective adjunct to radiation.
Converting cancer to a localized infection using engineered anaerobic bacteria
Conventional cancer therapies rely on their potent cytotoxicity for effective tumor control. This cytotoxicity is not particularly specific: they kill normal replicating cells almost as effectively as they kill cancer cells. Genetically engineered anaerobic bacteria have the potential to deliver cytotoxic therapeutics in high concentrations without significant systemic toxicity because they can selectively germinate within solid tumors. In addition to their tumor specificity, anaerobic bacteria have several advantages when used as gene delivery vectors:
(1) Delivery capacity. Because of the size, bacterial genomes are able to accommodate multiple exogenously introduced genes. Cancers are genetically unstable and it is likely that eradication of cancers will require a combination of therapeutic agents targeting various cellular components of tumors, analogous to the strategy used to eliminate genetically unstable HIV viruses.
(2) Widespread applicability. When anaerobes are used as anti-cancer agents, the only requirement for their targeting specificity is necrosis or hypoxia, a phenomenon associated with most solid tumors. This is a significant advantage compared to other gene therapies involving vectors with more stringent tropisms.
(3) Amplification ability. Systemically delivered therapeutics has limited access to poorly perfused regions of a solid tumor. Live bacteria have the capability of self-replication in a supportive environment. Therefore, the efficacy of bacterium-based delivery systems is less sensitive to the initial concentration reached in poorly perfused tumor tissues.
(4) Dissemination. The spreading of both biological (e.g., viruses) and non-biological agents (e.g., liposomes or polymers) within a solid tumor can be limited by fibrosis and other non-cellular components, thereby attenuating the efficacy. Free-living bacteria, on the other hand, can secrete enzymes capable of digesting the extracellular matrix during invasion, thus liquefying the tumor mass. Furthermore, bacterial strains such as C. novyi-NT can be highly motile, greatly increasing their radius of attack.
(5) Versatility. The efficacy of bacteriolytic therapies is not dependent on the replicating status of cells.
(6) Safety. Bacterial strains can be selected or generated which are widely present in the environment and nonpathogenic to healthy animals. These bacteria can serve as gene delivery vehicles that are safe for patients as well as for the general public. In addition, bacteria are relatively easy to eradicate with antibiotics when the therapy is concluded or if adverse effects are encountered.
Facultative or strictly anaerobic bacteria including C. acetobutylicum (Theys et al., 2001), C. sporogenes (Liu et al., 2002), B. longum (Nakamura et al., 2002), and an attenuated strain of Salmonella typhimurium (TAPET-CD) (Zheng et al., 2000) have been used as gene-delivery systems in different animal tumor models. Efforts are underway to transform C. novyi-NT and thereby convert it into a gene delivery vehicle as well as a direct bacteriolytic agent.
Though several types of anaerobic bacteria appear to be promising anti-tumor agents, major questions remain. How do bacteria interact and kill cancer cells? What kind of a role does the host response play in bacteriolytic therapy and how can it be manipulated to maximize the therapeutic effect? Are the impressive anticancer effects seen in experimental animal models translatable to human clinical trials? How safe is bacteriolytic therapy to human patients? What genes are optimal to employ when bacteria are used as delivery vectors? Hopefully, further studies to address these questions will result in the progress required to apply these fascinating pre-clinical approaches to the clinical level.
Bettegowda C, Dang LH, Abrams R, Huso DL, Dillehay L, Cheong I, Agrawal N, Borzillary S, McCaffery JM, Watson EL, Lin KS, Bunz F, Baidoo K, Pomper MG, Kinzler KW, Vogelstein B, Zhou S.
Overcoming the hypoxic barrier to radiation therapy with anaerobic bacteria.
PNAS 100(25):15083-15088, 2003.
Summary: Radiation therapy is particularly sensitive to hypoxia. C. novyi-NT can eliminate the hypoxic component of the tumor, and thereby has the potential to function as a radiosensitizing agent. In this study, C. novyi-NT was used to treat experimental tumors in conjunction with three forms of radiation therapy - external radiation, brachytherapy, and radioimmunotherapy. Remarkable improvements were observed for all three forms of radiation therapy when C. novyi-NT was included.
Dang LH, Bettegowda C, Huso DL, Kinzler KW Vogelstein B.
Combination bacteriolytic therapy for the treatment of experimental tumors.
PNAS 98(26):15155-15160, 2001.
Summary: This paper introduced the concept of Combination Bacteriolytic Therapy (COBALT) and described generation of C. novyi-NT, a C. novyi strain devoid of the alpha-toxin gene. This strain was then used in combination with chemotherapeutic agents to treat experimental tumors. Despite toxicity observed with this regimen, a significant percentage (80%) of the surviving animals were cured, demonstrating the greatly enhanced efficacy of COBALT.
Dang LH, Bettegowda C, Agrawal N, Cheong I, Huso DL, Kinzler KW, Vogelstein B, Zhou S.
Targeting vascular and avascular compartments of tumors with C. novyi-NT and anti-microtubule agents.
Cancer Biol Ther Mar. 2004.
Summary: Two classes of microtubule-interacting agents, microtubule stabilizers and destabilizers, were investigated in this study for their antitumor effects when used in combination with C. novyi-NT. Xenograft tumors showed similarly impressive long-term responses to selected drugs from both classes, although short-term responses appeared different. Responses varied among individual tumors.
Carey RW, Holland JF, Whang HY, Neter E, Bryant B.
Clostridial oncolysis in man.
Eur J Cancer 3:37-46, 1967.
Summary: This was one of the early clinical studies employing anaerobic bacteria in cancer therapy. Five patients with advanced malignancies were treated with the C. butyricum strain (M-55) via intravenous infusion of bacterial spores. Oncolysis was observed in three out of five patients and positive C. butyricum identification was made in two. No cures resulted.
Jain RK, Forbes NS.
Can engineered bacteria help control cancer?
PNAS 98(26):14748-14750, 2001.
Summary: This commentary characterized the unique tumor microenvironment and its impact on conventional therapies. It also tabulated the published studies involving bacteriolytic tumor therapies, providing a historical perspective.
Liu SC, Minton NP, Giaccia AJ, Brown JM.
Anticancer efficacy of systemically delivered anaerobic bacteria as gene therapy vectors targeting tumor hypoxia/necrosis.
Gene Ther 9(4):291-296, 2002.
Summary: C. sporogenes was engineered to express the E. coli cytosine deaminase. It was shown that C. sporogenes could target the expression of cytosine deaminase to tumor tissue and convert the nontoxic prodrug 5-fluorocytosine to the anticancer drug 5-fluorouracil, which resulted in significant growth delay of experimental tumors.
Mose JR, Mose G.
Oncolysis by Clostridia. I. Activity of Clostridium butyricum (M-55) and Other Nonpathogenic
Clostridia against the Ehrlich Carcinoma.
Cancer Res 24:212-216, 1963.
Summary: This paper described one of the early studies using Clostridial species to treat experimental tumors. In this study, a C. butyricum strain (M-55) and strains from several other Clostridial species were tested. Impressive lysis of Ehrlich ascites tumors was observed after intravenous injections of spores, accompanied with intolerable toxicity. No long-term cures were reported.
Nakamura T, Sasaki T, Fujimori M, Yazawa K, Kano Y, Amano J, Taniguchi S.
Cloned cytosine deaminase gene expression of Bifidobacterium longum and application to enzyme/pro drug therapy of hypoxic solid tumors.
Biosci Biotechnol Biochem 66(11):2362-2366, 2002.
Summary: B. longum was genetically engineered and used as a vector to deliver cytosine deaminase specifically to experimental tumors.
Parker RC, Plummer HC, Siebenmann CO, Chapman MG.
Effect of histolyticus infection and toxin on transplantable mouse tumors.
Proc So Exp Biol Med 66:461, 1947.
Summary: This is the first published study in which experimental tumors were treated by injection of anaerobic bacteria. Although oncolysis and tumor regression were observed, most animals did not survive the therapy because of the acute toxicity.
Theys J, Landuyt W, Nuyts S, Van Mellaert L, van Oosterom A, Lambin P, Anne J.
Specific targeting of cytosine deaminase to solid tumors by engineered Clostridium acetobutylicum.
Cancer Gene Ther 8(4):294-297, 2001.
Summary: The E. coli cytosine deaminase gene was used to transform C. acetobutylicum and the engineered bacteria were able to express the enzyme activity selectively in the experimental tumors.
Zheng LM, Luo X, Feng M, Li Z, Le T, Ittensohn M, Trailsmith M, Bermudes D, Lin SL, King IC.
Tumor amplified protein expression therapy: Salmonella as a tumor-selective protein delivery vector.
Oncol Res 12(3):127-135, 2000.
Summary: A strain of attenuated S. typhimurium engineered to express either green fluorescent protein (GFP) or cytosine deaminase (CD) was injected into tumor-bearing mice. GFP and CD activities could be detected in tumors, but not in normal tissue such as liver.
[Discovery Medicine, 4(20):33-37, 2004]