The Scientist

Photo: James Platt

Friends and colleagues: From left, K. Brooks Low, David Bermudes, John M. Pawelek

When the public hears about Salmonella, it is usually in a warning about food poisoning, but a group of researchers in New Haven, Conn., is using the bacteria to target cancer. It turns out that Salmonella preferentially colonize and multiply within a tumor, thereby inhibiting growth. Vion Pharmaceuticals is taking advantage of this trait by genetically altering Salmonella typhimurium to reduce the serious toxicities associated with wild-type infection and create a novel cancer treatment.

The association between bacterial infection and tumor regression has existed since the turn of the century. William B. Coley, attending bone surgeon at Memorial Hospital, now Memorial Sloan-Kettering Cancer Center, noticed that patients who developed severe infections after surgery for sarcoma fared much better than those who did not develop postoperative infections. Based on these observations, Coley purified a component of the bacterial cell wall for the treatment of cancer.¹ "Coley's toxin" was eventually marketed by Parke-Davis until the 1950s.

Coley's work was a starting point for all modern immunotherapy. It led to the discovery of tumor necrosis factor alpha (TNF-*) in the 1970s.² The main component in his concoction was lipid A, a membrane lipid of gram-negative bacteria that stimulates the immune system to produce TNF-*. The antitumor effect resulting from TNF-* is currently undergoing clinical development.

In the 1940s, researchers found another relationship between bacteria and cancer: Clostridia, strict anaerobic bacteria that cause gangrene and botulism, could infect tumors. Clostridia seemed to like the necrotic and anaerobic environment in large tumors. Nonpathogenic strains showed some success in mice, but only on larger tumors. Further research showed that Clostridia did not work as an antitumor agent in humans.

"We based our research on the hypothesis that Salmonella could have a preference for tumor cells," says David Bermudes, director of microbiology at Vion Pharmaceuticals. Melanoma cells have been described as macrophagelike, and Salmonella have specificity for attacking macrophages. Salmonella genetics is well understood, so manipulations to make it safe would be easier than with other bacteria. They also can grow under aerobic or anaerobic conditions, another advantage because a tumor is usually heterogeneous and has both conditions.

The project is multidisciplinary: Bermudes, a parasitologist, worked with John M. Pawelek, a melanoma biologist at Yale, and K. Brooks Low, a bacterial geneticist also at Yale. "We are able to connect scientifically and creatively in a way that complements our different backgrounds," explains Pawelek. Once the researchers proved that Salmonella targeted tumor cells in vitro, they developed strains better able to invade melanoma cells.³ Salmonella injected into mice found their way to tumors, where they preferentially replicated. This provides an advantage over most viral gene therapy agents, which need to be directly injected into tumors. Unfortunately, many deadly tumors are small and difficult to locate.

The bacteria disseminate throughout the body, but they cannot cross the blood-brain barrier. In a tumor they reach a concentration that is 1,000 times greater than in the liver and spleen, the next-highest concentration. "Compared to other targeting mechanisms such as monoclonal antibodies, the specificity is remarkable," comments Bermudes.

The researchers found invasive strains were better at targeting the tumor. In selecting more invasive Salmonella strains, they also selected bacteria that were auxotrophic for purines. These bacteria can't make the purines they need for growth and depend on the environment for the nutrient. "By serendipity we stumbled on the fact that the tumor environment is rich in purines, so there is no growth limitation for the Salmonella," comments Bermudes. The purine auxotrophy enhances the specificity of Salmonella for tumors and lessens their pathogenicity because purines are limited in the rest of the body.

The second alteration, which makes the bacteria safe for treatment, is a lipid mutation. Though TNF-* stimulation may contribute to the antitumor activity of bacterial infection, it can also cause life-threatening sepsis and irreversible septic shock. "We wanted to find a lipid mutant that would retain tumor-targeting and antitumor properties but lose the septic shock liability," explains Bermudes. A purine auxotroph with a lipid mutation turns out to be a safe and effective combination.⁴

William B. Coley Founder of modern immunotherapy

Administering the attenuated Salmonella to mice dramatically slows, but doesn't completely halt, tumor growth. The mice live twice as long with treatment. At the time of death, the mice still have some Salmonella in the tumors, but the concentration is less than at initial treatment. In monkeys, the bacteria cleared from normal tissues within 41 days. If necessary, the bacteria can be cleared sooner by using antibiotics.

The mechanism of tumor growth retardation is still uncertain. Low speculates that many factors may contribute to the antitumor activity. The bacteria secrete proteins into host cells through a needlelike structure. The proteins may interact with the signal transduction machinery and cause apoptosis. The bacteria may stimulate other arms of the immune system, which attack the infected cancer cells. The competition between cancer cells and the Salmonella for oxygen and nutrients may also contribute to slowing tumor growth.

Pawelek and Low are getting back to the basic biology now. "We are knocking out different genes in the bacteria to see which bacterial functions are important in causing the tumor retardation," explains Low. They have attracted the attention of the Salmonella community. Many geneticists are donating mutant strains so the researchers can see the effect different mutations have on the bacteria's ability to colonize tumors.

"The ultimate goal is to increase the antitumor effect so [the Salmonella treatment] does the entire job and completely kills off the tumor," explains Bermudes. The researchers are incorporating gene-based cancer-fighting drugs into the attenuated bacteria. By bringing the drug factory preferentially into the tumor, cancer therapy could be more concentrated, more effective, and less toxic to normal tissue. For example, six of seven mice treated with Salmonella expressing the TNF gene showed complete tumor regression.⁵

Combination therapy has shown some promise in the fight against cancer, so the researchers are exploring this possibility in mice. They reason that Salmonella is a good candidate for combination therapy because it does not act like standard chemical agents. Says Low, "We are confident that it can be used in combination." Last December the Salmonella treatment started Phase I clinical trials to test safety in humans.

Nadia S. Halim can be contacted at nhalim@the-scientist.com.

References

1. W.B. Coley, "The cancer symposium at Lake Mohonk," American Journal of Surgery, 1:222-5, 1926.

2. E.A. Carswell et al., "An endotoxin-induced serum factor that causes necrosis of tumors," Proceedings of the National Academy of Sciences, 72:3666-70, 1975.

3. J.M. Pawelek et al., "Tumor-targeted Salmonella as a novel anticancer vector," Cancer Research, 57:4537-44, 1997.

4. K.B. Low et al., "Lipid A mutant Salmonella with suppressed virulence and TNF-alpha induction retain tumor-targeting in vivo," Nature Biotechnology, 17:37-41, January 1999.

5. S.L. Lin et al., "Tumor-directed delivery and amplification of tumor-necrosis factor-alpha (TNF) by attenuated Salmonella typhimurium," Clinical Cancer Research, Suppl. 5, 1999.