Research Briefs

SIV May Be Much Younger Than Previously Thought

Simian Immunodeficiency Virus (SIV) may have existed in chimpanzees and sooty mangabeys for just hundreds of years before it jumped to humans, giving rise to the HIV/AIDS pandemic, a study has found (1). This is substantially less time than previously thought—SIV was thought to have coexisted in its natural hosts for perhaps millions of years, long enough to render it nonpathogenic. SIV typically does not cause disease in its natural hosts, including sooty mangabeys.

Joel Wertheim and Michael Worobey of the University of Arizona in Tucson, the authors of the study, estimated the rate of virus evolution from sequences of conserved regions of the gagpol, and env genes from samples collected from humans, chimpanzees, and sooty mangabeys between 1975 and 2005. Based on that rate of viral evolution, they then determined how long ago the common ancestor of the SIVs in chimpanzees or in sooty mangabeys must have existed.

They found that the common ancestor of chimpanzee SIV dates back to about 1492. This is just a little over 400 years before this SIV is thought to have jumped to humans, in 1908, to give rise to HIV-1 group M, which makes up the vast majority of HIV-1 infections. They also found that the common ancestor of sooty mangabey SIV dates back to about 1809, only about 120 years before this virus is thought to have jumped to humans to give rise to HIV-2. The analysis also found that SIV in sooty mangabeys is evolving at the same rate as HIV-2.

“These results were surprising because SIV has long been thought to be millions of years old,” says Wertheim, a doctoral candidate and first author of the study. Still, he notes that it is possible that SIV really coexisted longer in its natural hosts than these estimates suggest. There could be an unknown bias that could mask an older age of SIV, and if so, then that same unknown bias may also affect the estimates as to how long ago HIV-1 and HIV-2 jumped from nonhuman primates to humans. “This is a serious issue that needs to be addressed if these biases exist,” Wertheim says.

David Robertson of the University of Manchester, who was not connected to the study, also says that SIV could have existed in primates longer than was shown in the study. “What we have circulating now are the descendents of some successful virus that existed some number of years ago,” Robertson says. “That’s not necessarily the point when they first entered primates—that’s just the common ancestor of the ones that circulate now.”

However, if the findings from this study are true, it could mean that SIV evolved avirulence or nonpathogenicity in its animal hosts over a much shorter period of time than previously thought, Wertheim says, or that, alternatively, SIV may have been nonpathogenic to begin with. “The disease in humans and chimps [could be] an aberration,” Wertheim says. Chimpanzees have recently been shown to get sick from infection with SIV (http://www.retroconference.org/2009/Abstracts/34339.htm).

Another argument for SIV being old is that closely related SIV strains are found in closely related host species, suggesting that SIV had previously infected the common ancestor of these species, according to Wertheim. But two years ago, Wertheim and Worobey showed evidence that the evolution of SIV in African green monkeys doesn’t exactly mirror the evolution of its host (2), suggesting that SIV might have been transmitted more recently between closely related host species, instead of having infected their common ancestor.

Robertson has also developed a model, which showed that the observation that closely related viruses infect closely related species can be explained by a tendency for SIV to successfully jump between more closely related species. The new study by Wertheim and Worobey “very much confirms that message,” Robertson says. —Andreas von Bubnoff

1. PLoS Comput. Biol. 5, e1000377, 2009  
2.PLoS Pathog. 3, e95, 2007   

New Research Suggests HIV Enters Target Cells by Endocytosis

HIV has long been thought to enter cells by direct fusion with the outer plasma membrane. But a recent study suggests that instead, it enters target cells by endocytosis, and fuses with the target cell membrane only once it is inside the endosome (1). As a result, HIV particles may be harder to reach for antibodies that target HIV while it is fusing with target cells.

“[This] goes against dogma,” says Gregory Melikyan, an associate professor of microbiology and immunology at the University of Maryland in Baltimore, who led the study. “The dogma in the field was that HIV fuses directly with the cell plasma membrane.”

Melikyan and colleagues infected cultured cells with HIV particles that were stained with two different dyes: One for the viral lipid membrane, the other for the content of HIV particles. They found that at the plasma membrane, HIV particles fused only partially, without emptying their contents into the cell. In contrast, once HIV particles were endocytosed, they emptied their contents into the target cell (see Figure 3, below). “We never expected that to happen,” Melikyan says, “and then several thousand particles later we convinced ourselves that the fusion of the plasma membrane simply doesn’t happen.”

Figure 3. Entry Pathways for HIV. By visualizing the mixing of viral lipids (red) and contents (blue) with host cell membranes and cytosol, respectively, Miyauchi et al. (2009) observe three distinct routes for entry of host cells by HIV. These include endocytic events in which two-colored HIV particles are internalized, undergo lipid mixing with the vesicle membrane, and deliver their contents into the cytoplasm (A). There are fusion events that occur at the plasma membrane and proceed at least to the stage of hemifusion (B). These are followed by subsequent endocytosis and content mixing. There are also fusion events at the plasma membrane that do not result in any subsequent content mixing (C). Image and legend courtesy of Cell Press, Elsevier Inc., Uchil and Mothes, 2009. Cell Volume 137 n3, pages 402-404.

In addition, the researchers used a small peptide that cannot cross the plasma membrane and inhibits HIV fusion only at the outer plasma membrane, but not once an HIV particle is inside an endosome. The later this fusion inhibitor was added, the more virus was able to fuse with the membrane, presumably because it was inside the endosome and therefore protected from the inhibitor. Low temperature, however, which inhibits all types of fusion, even inside endosomes, inhibited HIV’s fusion to the target cell membrane for much longer. This suggests that the virus was first endocytosed and then fused with the target cell membrane only later, once inside the endosomes, Melikyan says.

The study also found that dynasore, an agent that inhibits dynamin—a protein important for endocytosis—inhibited HIV endocytosis and also its fusion with the membrane inside endosomes. This suggests that HIV may need target cell factors like dynamin to fuse with the membrane inside endosomes, Melikyan says, adding that such host cell factors could be future drug targets.

Dynamin has been suggested before to be required for HIV entry (2), but Melikyan says it wasn’t clear as to whether dynamin was directly involved. He says that while such previous evidence was “not sufficiently strong to convince the majority of scientists that this is serious,” his study now demonstrates that dynamin is required for both endocytosis of HIV as well as the delivery of viral content into the cytoplasm.

Overall, Melikyan says, his findings are the strongest evidence so far that HIV infects cells via endocytosis and not via direct fusion to the plasma membrane. However, he acknowledges that the study used cultured cell lines as target cells for HIV infection, and that the findings should be tested in primary T cells. “This has to be done in natural target cells that are more relevant,” he says. “[In] cell lines we may not get an adequate picture.”

If true, the findings mean that HIV is endocytosed into the target cell before undergoing fusion, making it difficult to inhibit HIV with certain drugs or antibodies that cannot cross plasma membranes and target intermediate conformations of Env that only form while HIV fuses with the target cell.

While Melikyan and colleagues did their study on cell-free HIV, their findings add to recent observations, in living cells, that HIV transmission between cells may also involve the endosomal pathway in the target cell (3; seeResearch Briefs, IAVI Report, March-April 2009). Benjamin Chen, an assistant professor at the Mount Sinai School of Medicine and lead author of that study, says the study by Melikyan and colleagues is timely and related. “We are coming at it from very different directions but we are coming to the same conclusions, so in a way it’s a bit of a convergent discovery.”

In a commentary on the study in the same issue of Cell, Pradeep Uchil and Walther Mothes of Yale University wrote that the study “presents the most comprehensive analysis of HIV entry to date and demonstrates that it does depend on endocytosis.” While not everyone in the field is ready to accept the findings, Mothes adds, many research groups will now take a closer look. “Without the paper being in Cell, people would not address these issues.” —Andreas von Bubnoff

1. Cell 137, 433, 2009
2.J. Virol. 79, 1581, 2005
3.Science 323, 1743, 2009

Identifying Broadly Neutralizing Antibody Activity in HIV-infected Individuals

So far, only a handful of broadly neutralizing monoclonal antibodies against HIV have been isolated from HIV-infected individuals. To identify others, researchers are actively screening HIV-infected individuals throughout the world. In an effort led by IAVI’s Neutralizing Antibody Consortium (NAC), researchers have conducted the largest screening and evaluation, to date, of virus neutralization patterns for sera collected from non-clade B HIV-infected individuals. This study identified some HIV-infected individuals, referred to as “elite neutralizers,” whose sera have broadly neutralizing activity (1). The study’s authors define an elite neutralizer as having neutralizing antibody activity against more than one pseudovirus in a panel of five—representing clades A, B, and C and one circulating recombinant form referred to as CRF01_AE—with a 50% inhibitory concentration neutralization titer of at least 300 within a single clade, as well as across at least four clades. Out of approximately 1,800 individuals screened in this study, 1% of them were classified as elite neutralizers.

Initially, researchers created an algorithm to assess neutralization activity based on 463 sera samples. They then used this algorithm to score and rank the neutralization capabilities of an additional 1,234 sera samples collected from individuals in Ivory Coast, Kenya, South Africa, Thailand, and the US. “Our results suggest that neutralizing activity across multiple geographic regions, with different spectra of circulating HIV-1, can be reliably assessed using a small panel of pseudoviruses,” the study’s authors write. All individuals were HIV infected for at least three years, did not meet a clinical AIDS diagnosis—either according to the World Health Organization’s criteria or by having CD4+ T-cell levels above 200 cells/ml, and had not been taking antiretrovirals in the year prior to sample collection.

These results confirm previous observations, indicating that chronically HIV-infected individuals have broadly neutralizing antibody activity, according to the study’s authors. Studying elite neutralizers may lead to the identification of additional broadly neutralizing antibodies against HIV, creating more targets for AIDS vaccine researchers. —Kristen Jill Kresge

1. J. Virol. 83, 7337, 2009

Broadly Neutralizing Antibodies Bind HIV Mostly With Just One Arm

The broadly neutralizing HIV antibodies b12 and 4E10 appear to mostly bind the Env spike with only one of their two antigen-binding arms at a time, a recent study suggests (1). The reason could be that HIV has only about 14 spikes on its surface. These spikes are too few and often too far apart from each other for both antibody arms to reach two spikes at the same time (2; see Figure 4, below). By comparison, the flu virus, while of a similar size as HIV, has about 450 spikes on its surface, making it more likely for both arms to be able to bind two spikes at the same time.

Figure 4. Env Spikes on SIV and HIV Virions. SIVmac239 virion (A) and three representative HIV-1 virions (B-D) showing the "front" (Top panel) and "back" (bottom panel) of the virions, and the distribution of the Env spikes (white). The images come from an analysis of individual virions by cryoelectron microscopy that showed that SIVmac239 virions have about 73 Env spikes per particle, and HIV-1 virions have about 14 spikes per particle. A version of the image originally appeared in Nature 441, 847, 2006. Reprinted with permission.

These findings could explain why HIV can easily escape from antibody neutralization by accumulating a few mutations on the Env protein, says Joshua Klein, the first author of the study who was a doctoral student at the California Institute of Technology (Caltech) at the time the study was published. “If only one arm is able to bind,” Klein says, “then it becomes much easier [for HIV] to acquire mutations to render those antibodies harmless.”

That’s because mutations that result in less efficient binding will have much less of an effect if both arms of the antibody can bind. “When HIV mutates, it really matters because you don’t have the buffering effect of [crosslinking],” Klein says.

The results of this study also give researchers clues about how to engineer larger versions of antibodies that can bind with both arms and could be administered in the form of gene therapy. “Our results suggest that the traditional vaccine approach—i.e., injecting an antigen in order to elicit an immune response to a virus—may never produce effective anti-HIV antibodies due to the inability of most anti-HIV antibodies to bind bivalently to the virus,” says Pamela Bjorkman, a professor at Caltech who led the study.

In the study, Klein and colleagues broke two broadly neutralizing antibodies, b12 and 4E10, down into their component parts, and tested how these different antibody parts could bind and neutralize HIV in vitro. As expected, the two-armed immunoglobulin G (IgG) versions of both antibodies could better neutralize HIV than the one-armed Fab versions. However, the improvement of two-armed over one-armed versions was much smaller than for antibodies to other pathogens like the flu virus, Klein says. Two-armed antibody neutralization was 17-fold better for b12 and 4.4-fold better for 4E10, compared to one-armed neutralization. By comparison, others have shown that two-armed IgG antibodies to the flu virus neutralize the virus about 1,000 times better than one-armed versions, Klein says.

Klein and colleagues also observed that two-armed antibodies with a shorter reach had less of a neutralization advantage compared with their one-armed versions than antibodies with a longer reach compared with their one-armed versions. This suggests that if one arm of an antibody was bound to a spike, then the second arm was less likely to bind to another spike if the arms were shorter than if the arms were longer. The researchers concluded, based on this, that the spikes on the surface of HIV probably don’t move freely.

Another finding of the study was that smaller versions of 4E10—which bind to the inner part of the Env spike called gp41—could neutralize HIV better than larger versions. This suggests that the site on the Env spike where the 4E10 antibody binds is not easy for the antibody to reach. “The bottom line is smaller is better,” Klein says, adding that it could also explain why previous studies have found that 4E10 is generally not a very potent antibody, and less potent at neutralizing different HIV strains than b12 (3).

Although an earlier study showed that larger IgM versions of 4E10—a ring of five IgGs—are weaker than IgG versions (4), Klein says the current study is “by far the most thorough analysis” to date, to show evidence for a steric occlusion of 4E10.

Pascal Poignard, an adjunct professor at the Scripps Research Institute and a principal investigator at IAVI’s Neutralizing Antibody Center who was not connected to the study, says he is not too surprised by the study’s findings since previous studies have shown similar improvements between the neutralization abilities of the two-armed IgG versus the one armed Fab versions of b12 and 4E10.

Next, Klein plans to use the insights from this study to engineer antibodies that can be used in gene therapy experiments, in a collaborative project with Bjorkman’s group and David Baltimore’s group at Caltech (seeEngineering Immunity, IAVI Report Jul.-Aug. 2008). “Now that I’ve figured out what this thing needs to look like to get bivalent binding,” Klein says, “my job is to try to make that molecule.”

Bjorkman says such an engineered antibody would have to be able to bind two sites on the same spike or two spikes at the same time. “By increasing the distance between their [binding] sites, it might be possible to create anti-HIV reagents that can take advantage of avidity effects,” she says. —Andreas von Bubnoff

1. Proc. Natl. Acad. Sci. 106, 7385, 2009
2. Nature 441, 847, 2006
3. J. Virol. 78, 13232, 2004
4. AIDS Res. Hum. Retroviruses 20, 755, 2004