The human immune system is a finely honed defense mechanism that quickly detects and destroys bacterial and viral invaders. But so-called “good” bacteria—which perform useful physiological roles—are a puzzling exception: they somehow slip under the immunological radar and form flourishing colonies on or in our bodies. Such “commensal microflora” are particularly abundant in the colon, which teems with some 10 trillion bacteria that help to metabolize nutrients and guide normal tissue development.

The scientific consensus has been that the immune system overlooks the colon’s commensal microbes because they remain “sequestered” within a layer of epithelial cells. However, the immune system does sometimes attack the colon’s commensal bacteria, causing inflammatory bowel diseases such as Crohn’s disease and ulcerative colitis.

In 1997, in collaboration with the late Charles A. Janeway Jr., M.D., Ruslan M. Medzhitov, Ph.D., professor of immunobiology, discovered toll-like receptors (TLRs), a new class of molecules in the innate immune system [“The Toll Road,” Spring 2002], and he suspected that they might be involved in inflammatory bowel diseases. To find out, he and his colleagues injected DSS, a substance that kills colonic epithelial cells, into both normal mice and mice with nonfunctional TLRs. Because of the TLRs’ assumed role in inflammatory bowel diseases as attackers of good bacteria, the researchers fully expected that there would be no inflammation in the mice with the disabled TLRs.

But to Medzhitov’s amazement, the opposite occurred. As reported last July in the journal Cell, even the smallest doses of DSS had no effect on normal mice, but in the mice with compromised TLRs they caused marked weight loss, severe colonic bleeding and death. Moreover, tissue samples from the colons of these mice showed that the normal cell cycle was profoundly disrupted. In short, TLRs looked more like defenders than attackers.

These startling findings led Medzhitov to question the long-held view that commensal microbes are fully shielded from the immune system. Instead he surmised that commensal bacteria may be only partially sequestered, and that recognition of these bacteria by TLRs triggers the production of molecules that protect the colon. In the case of acute epithelial injury, a TLR response to exposed microflora might efficiently recruit and direct healing molecules.

To test these wholly new ideas, Medzhitov and his colleagues turned their original procedure on its head. Instead of knocking out TLRs genetically, they used antibiotics to eliminate good bacteria from mice with normal TLR activity. Doses of DSS caused the same profuse intestinal bleeding and high mortality rate found in mice with faulty TLRs. But DSS caused no ill effects in mice that had received bacterial fragments known to activate TLRs.

“Because this all was so unexpected, it brought up a lot of questions that we’re following up now and finding a lot of interesting and surprising results,” said Medzhitov. But he believes that the studies have one immediate implication for clinical practice. Patients at risk for opportunistic infections, including those undergoing treatment with radiation or chemotherapy, routinely receive powerful antibiotics that kill off commensal intestinal flora. By combining antibiotics with TLR-activating compounds such as the bacterial fragments used in his mouse studies, Medzhitov says, it may be possible to prevent infection while triggering tissue protective responses needed to repair damage induced by radiation and chemotherapy.

“This idea could be directly relevant, and could be tested in clinical trials and used in patients—if not the precise regimen, the basic approach,” Medzhitov said. “That would be very exciting.”