From Ed Yong of Not Exactly Rocket Science (edited for length):

Nitya Venkataraman from the Univeristy of Central Florida has managed to reawaken a guardian gene that has been lying dormant in our genomes for 7 million years --- retrocyclins. They are known to protect monkeys from HIV-like viruses. The hope is that by rousing these genes from their slumber in humans, they could do the same for us.

Retrocyclins are the only circular proteins in our bodies and belong to a group of proteins called defensins, which as they suggest defend against bacteria, viruses, fungi, and other foreign invaders. They have only been found in the white blood cells of macaques, baboons, and orangutans. They have proven to be remarkably good at protecting cells from HIV infections....but in humans the genes don't work anymore. Somewhere in the course of evolution these genes developed a mutation that resulted in a useless retrocyclin.

Here's whats promising:

Despite this lone crippling mutation, the genes are intact (in all of us) and 90% identical to the mokey versions (which have been studied). Now, Venkatarman's group has reawakened them. She found two ways to fix the fault in human white blood cells --- one involving gene transfer and the other using a simple antibiotic. Either way she restored the cell's ability to manufacture the protective proteins. AND, the resurrected human proteins did their job well --- they stopped HIV from infecting a variety of human immune cells (up to 80% of the cells) and reduced levels of virus in cells already infected.

Gene transfer is expensive and an unlikely help in rolling out any future cures, treatments or preventions tools in resource-poor countries. But Venkatarman has also discovered something very cool --- the effect of a certain type of antibiotics, called aminoglycosides.

In bacteria, these drugs work by blocking them from creating proteins. But in the more complex cells of animals, they do something different - they react with the protein-making machinery of our cells so that they make slightly more mistakes than usual. Normally, that would be a bad thing but for retrocyclins, it's an unexpected boon. It means that the machinery barrels straight through the mutation that causes retrocyclins to be built half-finished. It doesn't stop prematurely, and produces a full-length protein.

Venkataraman found that one of these drugs, tobramycin, was especially good at restoring retrocyclins, and did so in both white blood cells and actual vaginal tissue. The drug slashed the rate of HIV infection by about 50% - a respectable figure but clearly a smaller one compared to the sizeable benefits bestowed by the gene transfer method. On the plus side, the technique didn't seem to harm the cells in any way.

These results are promising ones indeed, and Venkataraman thinks that with more work, aminoglycoside-based creams [a form of microbicide] could be used to prevent HIV infections in the real world.

HIV kills by infecting the very cells that are meant to defend us from infections and destroying them. But retrocyclins are something it hasn't encountered before. Humans lost the ability to create these guardians millions of years ago and by reawakening them, we could have a new but ancient weapon against this sneakiest of foes.