Toxin shows promise in killing off lurking HIV


An antibody and toxin mix has successfully detected and killed HIV-infected cells lurking in the organs and bone marrow of mice that were altered to have a human immune system.

The results, reported Thursday in the online journal PLOS Pathogens, offer conceptual proof that a reservoir of HIV-infected cells in organs can sought out and destroyed, a scenario that would potentially end the stalemate between the virus and antiretroviral drug therapies.

The altered mice, developed about eight years ago, can be infected by the human immunodeficiency virus in an identical manner to humans; they exhibit the same viremia and respond the same way to current antiretroviral drug therapy, but do not come down with AIDS, according to the study.


Assays of the mice organs – a process that can’t readily be done on human subjects – showed that the viral load and the number of infected cells in marrow and organ tissue dropped by an order of 10 to 1,000 in mice treated with the antibody-toxin compound, according to the study. That suggests this mix could be a promising candidate for the “kick and kill” end-game strategy to awaken “silent” HIV so antibodies can find it, dock with the infected cell, and deliver a lethal payload.

Current antiretroviral drug cocktails, which must be taken daily, thwart the virus successfully enough to bring its detectable level in blood close to zero. Achieving that therapeutic upper hand over the virus is considered a major milestone in the battle against AIDS, but it leaves open a path for HIV to evolve defenses against the drugs.

“In the organs, the virus continues to produce RNA and the therapy is doing nothing to it,” said University of North Carolina virologist J. Victor Garcia, lead author of the study. “The idea was: Can we kill those residual cells that are in the tissues, that are maintaining or contributing to the maintenance of the virus in the patient?”

Garcia, part of the team that developed the mouse model in 2006, believes the answer is a qualified yes – it works on partially “humanized” mice, but remains untested on actual humans.

Garcia likened his technique to firing a heat-seeking missile at an obscure target.

“We took an antibody that recognized a little protein of the virus when it’s expressed on the cell surface of an infected cell,” he said. “So, when the cell expresses this envelope protein, this antibody can bind to it. But the antibody on its own can’t kill the cell. It needs to have a payload. And in this particular case, the payload is a highly effective bacterial toxin.”

Although Garcia’s experiment targeted cells that express this protein, and thus are part of the active reservoir of HIV, the technique ultimately is aimed at the so-called silent HIV that can’t be readily detected by antibodies. Researchers are pursuing strategies that would induce this latent reservoir to betray its location and allow antibodies to dock and deliver toxins.

Garcia used a truncated version of a bacterial endotoxin as the payload, carried by a relatively well-tested antibody that targets the GP120 envelope protein that helps HIV enter human immune cells.

Compared with those treated with only the drug cocktail, mice that also received the antibody-toxin compound saw a thousand-fold decrease in the viral RNA in bone marrow. Decreases also were evident in the mouse’s thymic tissue, spleen, lymph nodes, liver, lung, intestines and blood cells. The decrease across all those tissues was about tenfold, according to the study.

There also was a nearly hundred-fold decrease in the number of cells producing viral RNA, according to the study – suggesting that cell death had a direct effect on the viral load.

“Everywhere we look,” said Garcia, “the antibody is able to kill those infected cells.”

Garcia said his experiments offer further support for using the humanized mouse model, known as the BLT mouse (for bone marrow, liver, thymus) to speed development of the lethal element of the three-part assault on HIV.

“We know how to treat the patients - we know very well how to do the therapy,” Garcia said. “We’re learning how to do induction, but we’re a little behind in how to do the kill.

“Now, if something better comes around, we would be able to test it in a relatively quick manner and then hopefully translate it into clinical applications in a shorter time period.”