Science magazine, 22 December 2006
Toll-like receptors (TLRs), once an obscure class of cell-surface proteins, have become hot targets for drugs and vaccines. And many immunologists have come to regard these key microbial sensors as the long-sought link between the immune system’s initial inflammatory response and its more tailored and enduring antibody and cellular response (Science, 14 April, p. 184).
Now, however, a new study casts doubt on the importance of TLRs as general-purpose immune stimulators, particularly for the type of long-lasting immunity required by vaccines. On page 1936, immunologist Amanda Gavin at the Scripps Research Institute in San Diego, California, and her colleagues report that mice lacking the ability to respond to all TLR signals can nevertheless mount impressive antibody responses to four different vaccine adjuvants, two of which were thought to work through TLRs.
Although the new work doesn’t contest the role of TLRs in innate immunity, the early containment of microbial infection, it does challenge the notion that TLRs are necessary for turning on the adaptive immune response, in which T and B cells become armed against a specific microbe and remember it to deploy defenses against subsequent attacks. “This paper shows that TLRs are not the essential link between the innate and adaptive immune systems where classical adjuvants are concerned,” says Scripps immunologist Bruce Beutler, a co-author of the paper. “A large body of literature has to be reexamined.”
The work “is wonderfully provocative,” says Thomas Hawn, an infectious-disease specialist at the University of Washington, Seattle. “It reminds the field that there are alternative pathways for the innate immune response to influence an adaptive immune response.” Some of those pathways might provide superior targets for vaccine adjuvants, he adds; Hawn and others worry that stimulating TLRs could lead to serious side effects.
Yet some immunologists dismiss the new study, citing its use of an “artificial” antigen as one of several flaws in its methodology. Others, such as Arthur Krieg, chief scientific officer at the TLR firm Coley Pharmaceutical Group in Wellesley, Massachusetts, point out that immune stimulators thought to work through TLRs appear safe so far in human trials and have shown preliminary signs of efficacy.
The controversial Scripps work began some 20 months ago when Gavin, along with graduate student Bao Duong and Scripps immunologist David Nemazee, wanted to test whether TLRs were required for antibody-making B cells to respond to synthetic molecules belonging to a narrow class known as T-cell independent antigens. They injected a large molecule made of linked sugars into a strain of mice whose TLR signaling is defective. These mice generated just as many antibodies to the antigen as did ordinary mice.
The researchers then tested both the TLR-disabled and normal rodents’ immune responses to a protein antigen. This time, Gavin added an adjuvant called alum, because proteins typically are weakly immunogenic. Her team saw the same strong antibody response in both kinds of mice. Just as they got this result last year, Yale University immunologists Ruslan Medzhitov and Chandrashekhar Pasare reported in Nature that work with a different strain of TLR-disabled mice, including cell-transfer studies, led them to conclude that TLRs on B cells as well as dendritic cells are required for optimal antibody responses. In a letter to Nature, the Scripps team disputed that conclusion, supplying some of their then-unpublished data. In a reply, Medzhitov and Pasare argued, among other points, that the Scripps team would have obtained different results with other vaccines or adjuvants.
Responding to that challenge, the Scripps team injected both TLR-disabled and normal mice with a chemically modified protein antigen and other immune boosters, including Freund’s complete adjuvant (FCA), an oily microbial mixture that includes TLR ligands, and Ribi adjuvant, a TLR4 activator used in a hepatitis B vaccine. They saw robust antibody responses to the antigen for all the adjuvants in both types of mice. “We were surprised,” Gavin says. “We too had been sucked into the misconception that TLRs are the only road there is” to a strong antibody response.
Medzhitov believes the study is fatally flawed, however. The robust B-cell responses in the TLR-signaling mutants, he claims, result from the use of a chemically modified protein. If Gavin’s team were to use a regular antigen, he predicts, they would see a big difference between the mice. “TLRs are not the only possible target for vaccines,” Medzhitov maintains, “but as far as we know, most of the major adjuvants work through TLRs.”
Whether TLR stimulants are safe and effective adjuvants should be resolved as large-scale human trials come to a close in the next several years. But if Gavin and her colleagues are correct, biotech firms may want to shift gears. “TLRs are moving rapidly in the clinic,” Krieg says. “But could there be something better in the future? Absolutely. Clearly, you can generate strong immune responses without TLRs.”