acterial DNA triggers a rapid response by the so-called innate immune system of mammals. But how does the innate immune system—dismissed for decades by serious immunologists as "unsophisticated, unintelligent, indiscreet, and obsolescent,"according to Douglas Fearon, a researcher at Cambridge University—make such a fine distinction? DNA is DNA is DNA, regardless of its source. What could detect that a sequence comes from an invader? The answer appears at #10 in the latest list of Hot Papers.

Innate immunity, the instant response to general signals of invasion by pathogens, is evolutionarily ancient, with some elements shared among plants, insects, and mammals. But immunologists tended to ignore it as little more than a holding action while the crack troops of acquired immunity assembled themselves. That view started to change about five years ago, when Charles Janeway, a Howard Hughes Medical Institute investigator at Yale University, identified the human equivalent of a Drosophila gene called Toll. In the fruit fly, Toll not only signals one of the key axes of development, it also helps the fly fight off fungal infections.

In humans a succession of Toll-like receptor genes has been identified, from TLR1 to TLR10. Many have been shown to respond to very specific pathogenic signatures. TLR2 responds to the peptidoglycans that characterize the cell walls of Gram-positive bacteria. TLR4 is triggered by the lipopolysaccharides of Gram-negative bacteria. And TLR5 responds exclusively to flagellin, a protein found only in the whippy flagellae that some bacteria use to move themselves about. These findings, with their enormous implications for therapy, have turned the innate immune system into a hot topic for investigators (see, for example, Science Watch,11[4]:8, July/August 2000, and 11[5]:8, September/October 2000). The new TLR announced in #10 confirms the interest in the area. Shizuo Akira and his colleagues at Osaka University in Japan identified TLR9 and showed that it responds very specifically to DNA sequences that distinguish bacteria from mammals.

Akira’s group started off with a mouse DNA sequence that showed a high similarity to previously identified TLRs. That sequence enabled them to fish for the complete sequence in a mouse DNA library and to find its human equivalent. The gene’s sequence revealed that it was indeed a new TLR. To find out what it did the group generated knockout mice that lacked the gene, the same technique they previously used to examine TLR4. The key discovery was that these TLR⁹-/- mice did not respond to a specific bacterial DNA sequence.

The sequence in question consists of an unmethylated pair of nucleotides, cytosine and guanine (CpG) and flanking regions. This type of fragment is 20 times more common in bacteria than in mammals. CpG pairs in mammals are much more likely to be methylated. So TLR9 is much more likely to be triggered by bacterial DNA than by the host’s own DNA. Details are still being worked out, but in one crucial respect TLR9 seems to differ from the other TLRs. They respond to bacterial signals that arrive at the cell surface. But TLR9 seems to require CpG sequences to be scooped up and taken into the cell as endocytotic vesicles.

One reason Akira’s paper is being highly cited is that it may shed new light onto DNA vaccines. Traditional vaccines use protein antigens to induce acquired immunity. DNA-based vaccines use a sequence that codes for an antigen that will stimulate the proliferation of T cells, but they also contain CpG elements that activate the innate immune system. Those in turn boost T cell responses, rather like the adjuvants in conventional vaccines, making the DNA-based vaccine more effective. Akira’s group shows clearly that activation of TLR9 by CpG sequences results in the release of IL-12, which does stimulate T-helper-1 cells. That could account for the adjuvant effect of CpG sequences.

So far, DNA-based vaccines have been more effective in mice than in humans. The CpG sequence that best activates mouse innate immunity differs by one amino acid from the sequence that activates human innate immunity: GACGTT as opposed to GTCGTT. Akira’s paper suggests that differences between mouse and human TLR9s may underly differences in the way they respond to particular CpG sequences. On the horizon may be more effective DNA-based vaccines for people, trials of which are already under way.