Deciphering Immunology's Dirty Secret

Indeed, what's old is new again. "We've known for a very long time that the body mounts an intense inflammatory response to microbes," says Bruce Beutler, professor of immunology at the Scripps Research Institute in La Jolla, Calif. Scientists even knew about what caused the response: In the 1890s Robert Koch and Richard Pfeiffer identified what they dubbed endotoxin, the complex molecule that coats Gram-negative bacteria. They found that endotoxin - later pegged specifically as lipopolysaccharide (LPS) - could induce fever and shock in guinea pigs.

Fly geneticists changed the name of the game. In 1996, scientists had learned that a fruit fly development gene called Toll also protected the flies from fungal infections. ² In 1997, Janeway and Ruslan Medzhitov, then a postdoc at Yale, found that a known human gene, similar to the fly's Toll, could activate immune genes that stimulate T cells. ³ The discovery of several more receptors in the Toll family soon followed, and 11 have now been identified and validated in humans.

It was Beutler and his colleagues who ultimately connected bacterial endotoxin - and, on a grander scale, practically any microbial invader - to the innate immune system. Using mice resistant to LPS, they found the toxin's receptor, mapping the mutation to a single gene, TLR4. ⁴ "The findings in the last 10 years have shown that the toll-like receptors are what really drive the innate immune system more than anything else," Beutler says.

The era was marked by an explosion of Toll papers. "There was a general sense of excitement," says Pulendran who was a postdoc at the time, working on dendritic cells at Seattle-based Immunex. "All of us in this field could feel it reaching an event horizon, so to speak. We were all on the cusp of finding new things."

A code of sorts was emerging. Each TLR uniquely recognizes an invader. For example, TLR4 recognizes LPS, while TLR7 and TLR8 sense single-stranded RNA from viruses. TLR5 sees flagellin on mobile bacteria, and TLR9 recognizes CpG DNA from bacteria or viruses.

A dramatic reduction in viral loads for Pulendran's 25 macaques could prove the validity of a TLR-targeting strategy.

Arthur Krieg unwittingly stumbled on that last relationship in 1994, even before the TLRs had officially been discovered. Then a researcher at the University of Iowa, Krieg was working with antisense oligonucleotides. He made a batch of them that were supposed to serve as controls, but instead they caused a massive B cell activation. "I just knew it was the most powerful immune activator that I had ever seen," says Krieg, now chief scientific officer of Coley Pharmaceutical Group in Boston, a company he formed to commercialize his discovery.

The common component he found in those stimulatory oligos was a cytosine-guanine sequence. In human DNA, CGs are normally methylated, but in bacteria they are not. A bit of detective work confirmed that unmethylated CGs somehow trigger an immediate immune response. ⁵ He later worked out a 24-base oligo sequence which contains four CpG dinucleotides. Krieg started Coley to research and commercialize the oligo.

Coley's, CPG 7909, as the oligo is called, is similar to one of the adjuvants that Pulendran gave to some of the monkeys in his trial. In winter 2004, they received the first dose of the vaccine, a prime with a naked DNA vaccine containing three HIV genes. Weeks later, they received a boost with the viral vector MVA, altered to express those same genes. One group of monkeys also received CpG DNA. Another received a small molecule that targets TLR7 and TLR8.

Thanks to the activity of Beutler, Krieg, and others, a handful of companies have developed several synthetic ligands that they're testing alone or in combination with vaccines. Each such adjuvant triggers a unique response, so the challenge is in learning which response is most effective at fighting a pathogen, complementing the immune reaction a vaccine elicits, or toning down a hyperactive immune system.

Some hopeful hints have indicated that these strategies are effective. For example, in October Dynavax Technologies of Berkeley, Calif., reported results from a Phase II clinical trial of its CpG DNA oligonucleotide conjugated to a ragweed antigen. ⁶ A six-week course of weekly vaccinations reduced clinical allergy symptoms over two ragweed seasons in people who received the vaccine compared to those who received a placebo.

"Prior to these toll-receptor ligands, there hadn't been a new vaccine adjuvant licensed in the last 15 to 20 years," says Robert Coffman, vice president and chief scientific officer of Dynavax, which is also testing its oligo in other allergy vaccines. "It was a field that, in my view, was stuck with limited technology. It's given vaccinologists a whole new toolbox of things they can use to really tailor to specific needs."

Pulendran says these tools could enable a process long in the making: rational vaccine design. "Almost all previously developed vaccines were made empirically," he says. From smallpox to measles, most approved vaccines have been developed though trial and error: Substitute, kill, attenuate, fragment, mix, and test for the bullet that works.

Attenuated viruses have been successful generally because they retain enough of their structure to stimulate innate immunity, essentially serving as their own adjuvants. One such vaccine for yellow fever has been around for more than 60 years and has been given to 450 million people worldwide. When Pulendran first came to Emory in 2001, the vaccine center's director, Rafi Ahmed, was working on understanding the nature of T-cell responses stimulated by the yellow fever vaccine. "One injection of the vaccine can induce different types of T-cell responses, and neutralizing antibodies for up to 30 years, but how this vaccine was so effective wasn't known," says Pulendran.

To determine the TLR that yellow fever vaccine activates, Pulendran and his graduate student, Troy Querec, injected the vaccine into knockout mice deficient for various TLRs. They ultimately found that the yellow fever vaccine prompts such a polyvalent immune response because it activates not one but four TLRs on different types of dendritic cells. "If we could only make something that would work as well as this against malaria or TB [tuberculosis], for example, we wouldn't be here," he says. "We'd close up shop and go home."

With commercial ligands, new knowledge about the innate immune system, and lessons learned, Pulendran says he hopes vaccinology can tackle the big dragons - HIV, TB, and malaria - that have largely eluded vaccinologists.

It is perhaps surprising that Beutler, a pioneer in the field, is one of a few to doubt that adjuvants targeting the TLRs will play much role in this fight. "I think people still don't understand fully how the adaptive immune system gets activated. They know the TLRs are one way to do it. Are they the best way? I'm quite skeptical about that," he says. Beutler's work has shown that TLR-independent mechanisms exist to activate the adaptive immune response.

Pulendran is undeterred. "I think the jury is out because, really, no one has done the experiment," he says. "That's why I'm so excited. If we can show that there's a dramatic reduction in viral loads [in the current HIV trial], it will then prove that the TLR strategy might work."

HIV has proven especially recalcitrant. But, "HIV is not different than any other viral infection," says Robert Seder, chief of the cellular immunology section of the Vaccine Research Center at the National Institute of Allergy and Infectious Diseases. "It requires a neutralizing antibody to protect you from getting infection or to severely limit the infection. The $800 billion question is, what's the neutralizing antibody."

HIV's structure has made it impossible, so far, to stop with a neutralizing antibody. Instead, Seder and others are turning their attention to specific T-cell types that can effectively kill infected cells. Malaria, TB, and HIV all require T cells at some point to keep infections under control. "What's holding up vaccines for malaria ... are ways to generate the appropriate magnitude and quality of responses that are durable," says Seder.

For HIV, Seder and his colleagues have found that CD8+ T cells that can produce multiple cytokines may be the most important for killing virus-infected cells and controlling viral load in HIV. In other words, an effective vaccine needs to generate not only many T cells, but also the right kind of T cells.

If that's true, then Pulendran's HIV trial is on the right track. After administering the complete course of the vaccine, which takes about four months, they found that certain groups of monkeys - even those that received the vaccine but no TLR adjuvant - had high counts of T cells that made three different cytokines: interferon gamma, interleukin 2, and tumor necrosis factor. "The immune responses are looking terrific postchallenge," Pulendran says. They still don't know, however, exactly how the adjuvants add to or complement that response, and ultimately it will be the viral-load data that prove whether those high T-cell counts were meaningful.

While they wait for those final results, Pulendran and his colleagues have already moved on to see how they can improve their vaccines. He acknowledges that the strategy that Seder's group developed - conjugating a protein-based vaccine to a TLR7/8 agonist - seems to provoke more of an immune response than his own strategy, in which vaccine and the adjuvants were simply mixed together. Moreover, the companies that have developed specific TLR ligands are doing much work on their own to see what types and combinations of responses their products can generate.

Pulendran's group is now exploring the possibility of delivering combinations of different TLR ligands with an antigen encapsulated in nanoparticles. He wants these nanoparticles to include multiple TLR ligands, much like the yellow fever vaccine. They have also learned from the yellow fever vaccine that when vaccines activate multiple TLRs, the whole isn't always the sum of its parts. While different TLRs can synergize with each other, in some instances they antagonize each other's effects. "It's a double-edged sword when you stimulate the immune system," Pulendran says. "It's got to be not too hot, not too cold, but just right."

Pulendran suspects that this is part of the issue in fighting HIV. People infected with the virus might be immunocompromised in some way, so it's not enough to stimulate an immune response. "It's like trying to push down the accelerator on a car, but the handbrake is on," Pulendran explains. "Even though there's noise and smoke, the car isn't moving forward. The challenge is to relieve the immune brake while pressing down the accelerator."

The TLRs generate incredibly potent inflammatory responses, so researchers have to gauge the safety of each synthetic ligand. As TLR-ligand combinations become more popular, researchers need to understand not only the unique cascade that each one triggers, but also their composite outcomes. Timing and delivery of a vaccine matters, as does whether the antigen is conjugated to the TLR ligand or simply mixed in with it. Nanoparticles have already been scrutinized in bioengineering circles for their possible toxic effects in the body, and nanovaccines will be subject to similar scrutiny.

"Everything is a question of therapeutic index in this business," Beutler says. Someone with cancer might be able to tolerate a little bit of toxicity, but a purely preventive vaccine must have almost no side effects.

Beyond safety concerns, scientists must also determine the exact tweaks that will result in the correct immune response. Which ligand? Which TLR? What T cell? What dendritic cell? How many cytokines? "The challenge is figuring out ... what immune response you want to end up with, and then trying to figure out how to apply what you know about innate immunity to make that response," Coffman says. With HIV in its sights, the Bill and Melinda Gates Foundation recently funded a consortium headed up by Julie McElrath at the Fred Hutchinson Cancer Research Center in Seattle. Pulendran and colleague Ahmed are on that team, and they will contribute their understanding of innate immunity to the overall picture of making a vaccine against HIV.

The work dovetails nicely with Pulendran's HIV trial, even as he awaits the final word on how successful their experimental strategies actually were. By early December, Pulendran had received the initial results on the monkeys' viral load figures three to four weeks after the SIV challenge. "It looks encouraging," he says. Any further detail, though, he's keeping to himself until he gets results from 10-to-12-week data, which he says is more indicative of whether the vaccine has actually worked. Should the results not pan out as he hopes they will, even that information will be valuable, he says. "We'll have to go back to the drawing board and figure out why - despite such a huge expansion of dendritic cells and great activation of the innate immune system and very good T-cells responses - why didn't we get better control? ... It would say to us that it doesn't matter if you have such a huge immune response [if] we're not looking at the right type of response."