By David N. Leff

Behind the scenes at major airports, trained dogs sniff out baggage suspected of containing illicit drugs. Their canine sense of smell has been conditioned to target molecules emanated by narcotics.

For decades, immunologists have labored mightily to unmask the molecules that presumably emanate exclusively from cancerous tumors. When one scientist in the 1970s announced discovery of what he confidently called "tumor-specific antigens," his holy grail turned out to be somewhat holey.

Since then, the cancer-targeting hunters prudently term their discoveries "tumor-associated," rather than "specific."

Now, taking a different tack, microbiologists at Yale University and a spin-off company, Vion Pharmaceuticals Inc., both of New Haven, Conn., report having mutated a species of infective bacteria that seem to sniff out mammalian tumors, while leaving non-malignant tissues strictly alone.

Their paper, in the January 1999 issue of Nature Biotechnology, bears the title "Lipid A mutant Salmonella with suppressed virulence and TNFa induction retain tumor-targeting in vivo." The article's senior author is microbiologist David Bermudes, associate director of biology at Vion.

Salmonella's sinister reputation in man and beast rests mainly on two species, S. choleraesuis, which gives pigs hog cholera, and S. typhi, inflicter of typhoid fever on humans - and humans only. That's one reason why Bermudes and his co-authors enlisted a third version, S. typhimurium, which primarily causes food poisoning in all warm-blooded mammals, plus snakes.

"What we did," he told BioWorld Today, "was genetically engineer a mutation in the outer coat of that bacteria's cell wall. It contains lipopolysaccharide, also called lipid A. Lipid A," he continued, "is the main stimulator of the immune system, which causes septic shock." (See BioWorld Today, Dec. 21, 1998, p. 1.)

"That mutation," Bermudes explained, "attenuated the chosen Salmonella's wild-type virulence. A second mutation made the bacteria auxotrophic for purines - nucleotide bases, which are building blocks for DNA. Those auxotrophic mutants could no longer make their own purine metabolites to survive and grow."

He and his co-authors determined that "the bacteria find the purine in abundance in the tumor environment. So, at the expense of the tumor, they replicate to very high numbers - 10 to the ninth - one billion per gram of tissue."

Tumor cells, which by definition divide rapidly and endlessly, need a lot of purines, primarily adenines and guanines, to fuel their replication. So converting the virulent wild-type, S. typhimurium bacteria from pit bulls into pussycat mutants that need exogenous purines just to stay alive and multiply, gave the Vion/Yale team the tumor-targeting vehicle they were looking for.

"The unmodified, wild-type Gram-negative Salmonella," Bermudes said, "is extremely virulent for mice. As few as 10 or 20 organisms will kill a mouse. We injected a couple of million mutant bacteria, with no toxic consequences to the animal. So that's our basis for claiming that it's 100,000-fold safer."

Technology Passes Blindfold Tumor-Spotting Test

That phenomenon, Bermudes went on, "is the basic premise of Vion's current flagship technology, called TAPET [Tumor Amplified Protein Expression Therapy]. What's generally remarkable about TAPET, compared to any viral gene therapy approach, is that our bacteria can be administered systemically, intravenously. And they locate the tumors without knowing ahead of time where they are. They have a natural ability to survive in the bloodstream, find their way to the tumor site, and when they get there, preferentially replicate."

He contrasted this strategy with "the various viral treatments that are directly injected into the tumor in order to be effective. And there are a lot of Salmonella and Listeria vaccines out there, but those are all for oral delivery."

Dispelling Catch-22 Cloud On TAPET's Horizon

When Salmonella's lipid A coat protein overstimulates the immune system to take action, one of its first actions is to liberate the two-faced cytokine, tumor necrosis factor-alpha (TNF-a) "What that does," Bermudes explained, "is to release TNF-a at such high levels that it initiates the cytokine cascade that triggers septic shock and organ failure. But the mutation we made in lipid A drastically reduces the amount of TNF-a released from macrophages by wild-type lipid A.

"Mice and men are different," he noted. "One of the major differences is that mice are not susceptible to septic shock. In other words, just because you can make a bacterial therapy work in mice, and it doesn't cause septic shock, doesn't mean it would work that way in other animals, including humans."

Pigs are something else.

When the co-authors injected their mutated Salmonella into mice and swine, neither animal died. But wild-type bacteria brought death to all of the mice within four days, and 90 percent of the swine within five.

"Pigs make a good animal model for septic shock," Bermudes pointed out. "In swine, the bacteria reduced septic-shock potential. We didn't screw them up just by making these genetic modifications. So, it shows that you can make something safe for an animal that is in some regards similar to humans.

"Probably, the main therapeutic scenario for this, as for most cancer therapies, is that they're not curative by themselves," he said. "That's as true for our Salmonella as for most of the other treatments. The best new anticancer therapies are combinations with other standard therapeutics. And we don't have anything yet to say what that combination is."

On that score, Bermudes added that "the bacteria we are developing as anticancer vectors remain fully sensitive to a wide range of antibiotics."

Phase I trials of Vion's Salmonella construct are planned for the first quarter of 1999, Bermudes said. "Melanoma will certainly be one of the primary tumor-type targets. I think we've shown efficacy in mice with the bug we are using against melanoma, colon, lung and breast tumors. So, the justification for expanding a trial to include all of those indications is there.

"No one's injecting Gram-negative bacteria into humans for treating cancer," he concluded. "That part of it is totally novel." *