TCS Daily

Vaccinating Against AIDS

By Derek Lowe - April 23, 2002 12:00 AM

In my last column, I made the case that a vaccine is the only thing that can stop the catastrophic HIV epidemic, at least in the hardest-hit regions. But that statement raises key questions: Is it even possible to develop a vaccine? How effective could it be? How long is it going to take? And why isn't there one already?

First off, the good news is that there's nothing inherently impossible about an HIV vaccine. Almost everyone develops an immune response when exposed to the virus; it's just that the infection gradually wears down the very T-cells that are on the front lines. (In fact, anti-retroviral vaccines have already been developed, such as the one against the feline leukemia virus.) And there's clinical evidence that the immune system can (in some cases) handle HIV: a subset of infected patients have maintained very low viral counts, with no disease symptoms, after years of multiple exposure. Their immune systems seem to have features that have successfully fought off the infection -- and as you can imagine, these "long-term non-responders" have been exhaustively studied.

But there's nothing inherently easy about developing a vaccine, either. If it were straightforward, we'd have one by now. The standard methods just haven't worked. For example, one of the original vaccination approaches is the use of live attenuated viruses. In the early 1990s, the DNA of the related monkey virus SIV was modified to inactivate a key life-cycle protein. Immunization with this live virus seemed to protect monkeys from infection with real SIV, and hopes rose. But later work showed that even after what seemed to be crippling DNA deletions, the SIV virus could actually repair its genome and make the needed protein again. Then natural mutations of HIV itself were found with partially inactivated forms of the same protein -- but they were still deadly. After these discoveries, no one had the nerve to try a live-virus approach again.

Using killed viruses is the next classic technique. A form of HIV with much of its outer proteins stripped off was evaluated during the 1990s, in work that got press attention because of Jonas Salk's involvement. The immune responses raised have been disappointing, though, and no serious effort seems to be behind the idea now.

The next known step past these whole-virus methods is to use a pure viral protein, usually one found on the surface of the infectious agent. More than a dozen variations of this approach have been tried in humans since 1987. But HIV mutates many of its surface proteins very quickly - its whole life cycle is predicated on avoiding this sort of attack. Immunity can show up to one strain, only to be rendered useless as the virus changes coats. The best shot remaining with this approach comes from Genentech's offshoot, VaxGen, using mixtures of proteins from different HIV strains.

If whole proteins don't work, what about a mix of shorter peptides? That makes it easier to produce such multi-strain mixtures. Unfortunately, this hasn't worked out as well in practice as in theory. Getting the right peptides is the problem here, since it's very difficult to figure out which ones will induce the broadest immune response. The idea's still alive, but as a component of some larger strategy rather than as a stand-alone therapy.

Beyond proteins, there's another idea that's been floating around for years: so-called "naked DNA" vaccination. The HIV situation has everyone pulling out all the stops, though, and this high-risk high-return strategy has now reached the clinic. You directly inject DNA plasmids, which code for viral proteins. The DNA is taken up by various cells around the body, some of which then begin expressing those viral proteins on their own. This has the potential to stimulate just about every kind of immune response there is (T-cells, antibodies, the works). It also has the potential to set off an immune response against your own DNA, among other unwelcome effects. This hasn't happened so far, fortunately, in trials by Merck, Wyeth, and others. If this technique works, it could be a useful one for other diseases.

A similar, more conventional approach uses some other virus (or even a bacterium) as a delivery system for the viral DNA. Getting this to work is complex, because you can vary both the genes chosen and the new virus used to deliver them. The first large-scale attempt at this technique, an Aventis Pasteur canarypox vaccine, delivered unexpectedly poor results in February. The response was weak enough that NIH pulled funding for a larger trial, but the U.S. military is planning to test a variant of the same vaccine. They'll combine it with the Vaxgen protein vaccine mentioned above. No one is really sure yet what factors will be most important in preventing infection, so this study could provide some valuable clues.

That trial illustrates the current state of the art: try everything at once (which is reminiscent of the small-molecule pharmaceutical HIV regimens.) Merck has a promising vaccine system, for example, that uses naked DNA in the first stage, followed by a virus vector in the second. It seems very likely that any successful vaccine will be a multistage protocol; a single-shot approach is probably too much to hope for.

So, what can we hope for? The gold standard would be a vaccine that induces a vigorous, broad response against both the virus and against any cells containing it -- no matter where they're located, no matter where along the HIV infection cycle they might be. Such an agent would protect the unexposed population, would either halt or reverse the disease in those already exposed, and further prevent them from infecting anyone else.

The good news is that, as far as we know, this vaccine is possible. But the road to it has been very rough, and it's going to stay rough. We still don't really know what kinds of immune response best correlate with any of those desired qualities, and without that we're still driving with the windshield painted over. The predictive ability of animal models, even those using SIV, is still questionable, so vaccine candidates are going to have to go into people as quickly as possible. And there just aren't enough patients in the developed world to test them on the scale that's needed. Trials already have to take place in the most stricken parts of Africa and Asia, under very difficult conditions, and this will continue. Some of the difficulties will translate into clinical use of any successful vaccine, as well, such as the logistics of timed, multi-step immunizations.

The best guesses have the first effective vaccine coming out in about 10 years, with a liberal definition of "effective." I hope they're wrong. The immunologists freely admit that they could be mistaken, but in either direction. No matter what, this year and the next will bring some major trial results. We've had enough negative surprises; it's time, more than time, for a positive one.

Derek Lowe is a medicinal chemist with 12 years of experience at major pharmaceutical companies. He conducted PhD and postdoctoral work on natural products synthesis and free radical chemistry. Lowe has worked on drug discovery projects targeted against schizophrenia, Alzheimer's disease, diabetes, osteoporosis and other diseases. He edits the Lagniappe website.

TCS Daily Archives