Research Briefs

New Insights on Antibody Inhibition of Cell-associated HIV Spread

HIV can spread either as free particles, or from infected cell to uninfected cell. One important mode of cell-to-cell spread occurs through virological synapses (VS), contact zones that form transiently between uninfected and infected cells and enable HIV to pass through. Previous studies found that cell-bound spread is more efficient, although estimates vary as to how much. Also, some previous studies have concluded that HIV transmitted through VS might be harder for drugs or antibodies to reach than free HIV particles, which would be a concern for HIV vaccine and drug design.

A research group led by Quentin Sattentau, a professor of immunology at the University of Oxford, and colleagues has revisited these issues (1). The researchers found that in vitro, cell-associated HIV spread between CD4+ T cells via VS is about ten times more efficient than cell-free spread.

They also found that neutralizing antibodies such as b12, 2F5, and 2G12, as well as other HIV entry inhibitors, can inhibit VS-mediated HIV spread between CD4+ T cells and entry of free HIV particles into such cells equally well. They determined this by mixing uninfected CD4+ T cells with CD4+ T cells infected with R5-tropic HIV, or with free particles of R5-tropic HIV in the presence of the entry inhibitors. Researchers then measured the synthesis of new HIV DNA to indicate infection of the target cells. This is “encouraging for vaccine and drug design,” the authors conclude.

Sattentau and colleagues also found that some inhibitors such as the b12 antibody could destabilize the VS when added after VS formation. Even when added after VS formation, b12 colocalized with HIV and host cell proteins at the VS and could inhibit HIV infection of the target cells as measured by viral HIV DNA synthesis. “We are pretty sure that b12 is getting into the synapse,” Sattentau says. “In effect it’s neutralizing virus within the synapse.” Consistent with this, electron tomograms of VS between uninfected and infected CD4+ T cells showed sufficient space for large molecules such as antibodies to gain access.

“What people were worried about was that if the virus could efficiently spread cell-to-cell across junctions that are really sealed, then antibodies would be useless and perhaps so would some of the other inhibitors,” Sattentau says. “We don’t think it is like that. We think that these synapses are actually rather porous open structures and antibodies can get in.”

This doesn’t mean that VS-mediated HIV spread between cell types other than CD4+ T cells is also easy to inhibit. “This is a T cell-T cell synapse,” Sattentau says, “and you can’t necessarily say the same thing about other synapses. We work also on macrophage-T cell synapses, and they look much tighter.”


Virological Synapse-mediated Spread of HIV  

Surface rendering of membranes in electron tomographic reconstructions reveals the three-dimensional (3D) morphology of a synapse that is formed between an HIV-1 infected (orange) and an uninfected T-cell (grey). HIV-1 virions (red) are seen at the interface of the two cells.

Infected cell samples courtesy of Nicola Martin (Dunn School of Pathology, University of Oxford). Electron tomography and 3D image by Sonja Welsch (Structural and Computational Biology Unit, European Molecular Biology Laboratory). Originally published in Supplemental Movie 3 of J. Virol. 84, 3516, 2010.


The observations of Sattentau and colleagues differ from those of other studies, such as a 2007 study which showed that antibodies such as 2F5 did not block HIV transfer from T cell to T cell through VS (2).

This might be due to different experimental approaches such as the assays used to assess infection, according to Benjamin Chen, an associate professor of medicine at the Mount Sinai School of Medicine, who led the 2007 study. In the 2010 study, Sattentau and colleagues measured inhibition of HIV DNA synthesis as an indicator of target cell infection, whereas Chen and colleagues in the 2007 study measured inhibition of the transfer of HIV Gag proteins from the infected to the uninfected cells, which doesn’t always result in an infection of the target cell. Because 2F5 doesn’t block Gag transfer, but does block fusion of the virus with the target cell membrane, the two approaches could yield different results.

In the most recent study, researchers used cultured cells that have been infected with HIV for about a week, Chen says. His concern is that virus that has been repeatedly propagated in chronically infected cells is more likely to carry deletions in regulatory genes such as nef or vpu. This may make the virus more likely to behave in a manner similar to free virus. In addition, Chen says it will be important to test different sera and different viruses with Envelopes that more closely resemble those circulating in vivo. Last year, Chen and colleagues found that serum isolated from HIV-infected people can neutralize cell-free HIV infection better than cell-associated infection of CD4+ T cells (3). —Andreas von Bubnoff

1. J. Virol. 84, 3516, 2010
2. J. Virol. 81, 12582, 2007
3. Science 323, 1743, 2009

CMV Superinfection No Longer Shrouded in Mystery

Researchers have elucidated an important aspect of the mechanism that enables cytomegalovirus (CMV) to overcome pre-existing immune responses and therefore superinfect rhesus macaques already infected with the virus (1). The research team, led by Louis Picker, a professor of pathology at Oregon Health & Science University (OHSU), and Klaus Früh, a professor of molecular microbiology and immunology at OHSU, found that to superinfect, rhesus CMV needs genes that prevent major histocompatibility complex (MHC) class I of infected host cells from presenting CMV proteins to CD8+ T cells. “The virus prevents the infected cell from putting a big sign up that says, I am infected, kill me!” Picker says of this immune evasion strategy.

Last year, Picker led a study that showed that rhesus macaques that were already CMV infected could be superinfected with a CMV vector expressing SIV genes (2; see Research Briefs, IAVI Report, Mar.-Apr. 2009). The mechanism that enabled this superinfection was not known, although in in vitro experiments, Früh had shown that rhesus CMV expresses genes called US2, 3, 6, and 11 that can downregulate MHC class I presentation. The in vivo significance of such genes was unclear, because infecting naive mice with CMV lacking genes with a similar function didn’t have any effect (3).

Then Früh and Picker joined forces and infected rhesus macaques by subcutaneous injection with rhesus CMV with and without these genes. They found that wildtype CMV can superinfect rhesus macaques at much lower, more physiologic doses than previously shown. They also showed that CMV lacking the US2, 3, 6, and 11 genes could infect rhesus macaques that were never infected with CMV. In contrast, CMV lacking these genes could not superinfect rhesus macaques with preexisting immunity to CMV, unless their CD8+ T cells were depleted.

This suggests that CMV superinfection, but not initial CMV infection, requires the US-gene-mediated downregulation of MHC class I presentation of CMV proteins on the surface of infected cells, which normally activates CMV-specific CD8+ T cells. The study also found that this requirement for MHC class I downregulation disappears later, presumably because CMV moves to places where it is hidden from CD8+ T cells.

“This is the first [study] to show the importance of MHC class I immune evasion genes for the ability of CMV to superinfect,” says Ann Hill, a professor of molecular microbiology and immunology at OHSU who led the 2004 study of CMV lacking such genes in naive mice. “This provides a really appealing answer to the puzzle of what these genes do for the virus.”

Although the findings are in rhesus macaques, human CMV has very similar genes, suggesting that it uses the same immune evasion mechanism. This is good news for the development of CMV as a vector for candidate HIV vaccines because it suggests that widespread pre-existing immunity to CMV would not hamper the use of CMV as a vector for HIV vaccine candidates in humans. Last year, Picker and colleagues showed that a CMV vector expressing SIV genes could protect rhesus macaques from systemic infection after low-dose rectal challenge with SIVmac239 (2; see Raft of Results Energizes Researchers, IAVI Report, Sep.-Oct. 2009). They found that CMV, a replicating vector, induced effector memory T cells, which Picker suggests are better at protecting from challenge virus in mucosal tissues than the central memory T cells induced by non-replicating vectors.

Picker says the new study shows that CMV vectors for HIV vaccine candidates need to have the US genes in order to elicit immune responses in people with pre-existing CMV immunity. However, CMV still needs to be attenuated to keep it from causing problems in people with compromised immune systems. “We are working to make the virus safer,” Früh says. “We already have put the first attenuated viruses into monkeys and the results so far look good.”

The new findings on CMV superinfection also suggest that it may be difficult to develop a CMV vaccine that prevents infection in individuals with a compromised immune system, including fetuses whose mothers haven’t yet been exposed to CMV. CMV infection of fetuses is the main cause of non-genetic birth defects such as deafness. The new study suggests CMV would likely be able to overcome any immune response induced by such a vaccine, in the same way it can overcome previous CMV immune responses when it superinfects. “The natural infection doesn’t even prevent superinfection, so having a vaccine to prevent infection is not going to work,” Picker says.

However, while a CMV vaccine probably won’t be able to prevent CMV infection, it should be able to protect against disease by keeping the virus in check, Picker says. “You could have a vaccine [where] the mothers would still get infected, but if they got infected during pregnancy, they wouldn’t transmit to their fetus,” Picker says. “So it’s still possible to make a useful CMV vaccine. It’s just not possible to make a sterilizing CMV vaccine.”

While the study shows that the US2, 3, 6, and 11 genes are required for CMV superinfection, it is still not clear what biological advantage CMV gains from being able to superinfect its hosts. “Why would the virus want to do that?” asks Früh. One possibility is that CMV needs to be able to overcome immune responses to other immunogens that happen to cross-react with CMV even to establish initial infection. Another possibility is that superinfection enables different viral species to get into people, says Picker. “You have something for evolution to operate on.” —Andreas von Bubnoff

1. Science 328, 102, 2010
2. Nat. Med. 15, 293, 2009
3. J. Immunol. 172, 6944, 2004