Is HIV Hitching a Ride Inside Cells?

Despite great progress in understanding HIV transmission, researchers still don’t know if it is primarily the result of free virus particles or HIV-infected cells

By Andreas von Bubnoff

More than 25 years after the discovery of HIV, researchers have made great strides in understanding many aspects of HIV transmission. However, one fundamental question remains unanswered: What role, if any, do HIV-infected cells play in HIV transmission?

“This is an understudied topic in the field,” Deborah Anderson, a professor of obstetrics/gynecology and microbiology at Boston University School of Medicine, recently wrote in a review article (AIDS 24, 163, 2010). Anderson became a champion for the importance of understanding cell-associated HIV transmission over 25 years ago, when she first coined the term “Trojan Horse leukocytes,” to illustrate the possibility that the infectious agent later identified as HIV could enter a person’s body hidden inside a white blood cell, in which case it would be shielded from many of the body’s antiviral defense mechanisms (N. Engl. J. Med. 309, 984, 1983).

Anderson has done many studies to characterize HIV-infected cells in semen and cervicovaginal secretions, showing, she says, that they are often present in high numbers and are highly infectious. She says that even though HIV-infected cells in genital secretions—especially infected macrophages and CD4+ T cells—may play an important role in the sexual transmission of HIV, they have been largely overlooked in recent studies of the mechanisms of HIV transmission, and in the design and testing of HIV vaccine and microbicide candidates. In addition, she says, most nonhuman primate (NHP) challenge models use cell-free virus stocks to test vaccine and microbicide candidates, which is why the candidates may not protect against cell-associated viral transmission. This, she adds, may explain the failure of several vaccine and microbicide candidates in recent clinical trials. “It’s really interesting that there is this blind spot,” she says. “People are so invested with their models now that are based on the cell-free system that they are, I think, reticent to branch out.”

Others agree that knowing the role of HIV-infected cells in HIV transmission is important. “It is a basic question, and it is rather amazing that we haven’t answered it yet,” says Grace Aldrovandi, a professor of pediatrics at the Children’s Hospital in Los Angeles. “We do kind of need to know what we are targeting,” says Julie Overbaugh, a member at the Fred Hutchinson Cancer Research Center. “If it’s really cell-associated virus that’s important, then maybe we want immune factors that can target and lyse the infected cell, whereas if it’s cell-free virus, then maybe we want things that are more effective in cell-free virus.” One concern is that cell-associated transmission, if it plays a role, might be harder to prevent. “[For] cell-associated virus [it’s going to be] harder to be able to intervene with things like antibodies and things that would require being able to actually have access to the virus,” says Jairam Lingappa, an associate professor of global health and medicine and an adjunct associate professor of pediatrics at the University of Washington.

One major reason the role of cell-associated transmission is still unclear is that it is difficult to study. Researchers have analyzed genital fluids and breast milk for the presence of HIV-infected cells, and have studied the role of these cells in sexual or mother-to-child transmission (MTCT) of HIV. But so far, no clear consensus has emerged from these studies about the role of cell-associated HIV in transmission.

Understanding sexual transmission

Most HIV infections are the result of sexual transmission of the virus. Semen can contain free HIV particles, as well as white blood cells such as CD4+ T cells, which can be HIV infected and therefore may contribute to transmission. But so far, only a few studies have addressed the relative contribution of HIV in the cell-free and cellular portions of the semen to HIV transmission. Such studies typically separate the semen into the cellular and the cell-free fractions. They then analyze the HIV DNA sequences of the HIV-infected cells (such as HIV provirus integrated into the genome of the infected cell) in the cellular fraction, and the HIV RNA sequences of the free HIV particles in the cell-free fraction.

In 1996, David Ho and colleagues were the first to look at this issue in five men who have sex with men (MSM) transmission pairs (J. Virol. 70, 3098, 1996). They compared the HIV sequences of the cellular and cell-free fractions of the donor’s semen with the sequences in the blood plasma of the recipient. They found that in three cases, the cellular fraction of the donor’s semen was more similar, and in one case the cell-free part was. This was consistent with both cell-associated and cell-free HIV playing a role in transmission, Ho says. However, he adds, the study was of limited value because the samples were taken one to two months after transmission, which means that the HIV sequences in the donor could have changed in the meantime. “One cannot draw definitive conclusions,” says Ho.

Last year, Davey Smith, an associate professor at the University of California in San Diego, and colleagues did sequence analysis of six MSM rectal transmission pairs a few months after infection. They reported that the HIV sequences in the cell-free part of the donor’s semen were consistently more similar to the HIV in the recipient’s blood than the sequences in the cell-associated part of the donor’s semen. This suggested that cell-free and not cell-associated HIV is involved in sexual transmission through semen (Sci. Transl. Med. 2, 18re1, 2010).

But then James Mullins, a professor of microbiology and medicine at the University of Washington, and colleagues re-analyzed the data and concluded that Smith and colleagues had included contaminated samples in their study because the HIV sequences in the cellular part of the donor’s semen were too different to have come from the same donor as the HIV sequences in the cell-free part of the semen (Sci. Transl. Med. 2, 50le1, 2010). “What we found, and it’s really unimpeachable, is that the cells were actually from different people,” Mullins says. “That’s why they weren’t closely related [to the recipient’s HIV].” Mullins says Smith and colleagues might have mixed things up in the laboratory, or samples coming into their laboratory were mislabeled. “They didn’t figure it out by a long shot,” adds Mullins. “In fact they have only made the field more confused.”

Smith says he went back and did not find evidence of contamination, mixed up samples, or mislabeling, but acknowledges that he can’t completely rule them out (Sci. Transl. Med. 2, 50lr1, 2010). He says his finding should be validated, which is why he is now trying to better characterize the cellular HIV sequences in the semen of additional transmission pairs. In addition, he says, not enough is known about sequences of HIV found in infected cells in semen to judge what degree of sequence diversity one would expect. “It’s not that easy to find cellular HIV sequences, and very few people have actually done it,” Smith says. “We really don’t know how diverse that population actually is and we can’t really just dismiss everything there just because we haven’t seen it before.”

But for now, the concerns about the study remain. “The problem is there are going to be a lot of people who are just not going to look at this as being very useful information because of the contamination issue,” Lingappa says.

Mullins wonders if it will ever be possible to determine to what extent cell-associated HIV contributes to sexual transmission through semen. “I am not very optimistic,” he says. One challenge, Mullins and others say, is to find transmission pairs where the partners are willing to give blood or genital fluid samples at a time close enough to transmission to be able to get meaningful results when comparing sequences. “It’s hard to get the fluid, and especially hard to get the fluid at the right time,” Mullins says. Even samples collected just a week after transmission might already have changed from the semen that caused the transmission, he adds.

It is also not always possible to find enough cells in the ejaculates of HIV-infected men that would allow for isolation and sequencing of the cell-associated HIV in the donors, says Ron Swanstrom, director of the University of North Carolina Center for AIDS Research. In fact, Swanstrom says, there are often so few cells that it seems questionable whether there are enough cells to mediate transmission. “We do find CD4+ T cells, but not at significant levels that would suggest enough infected cells to contribute to transmission,” he says. Anderson says she has also found that HIV-infected cells such as CD4+ T cells are not always detectable in semen from HIV-infected men, but adds that there is great variation. For example, many more HIV target cells can often be found in the semen of HIV-infected individuals with inflammation in the genital tract from other sexually transmitted diseases.

For all these reasons, the studies of the role of cell-associated virus in HIV transmission in semen have so far involved too few transmission pairs to yield meaningful results, Mullins says. “We really need to do large-scale analyses and do them carefully,” he says.

In what may be the largest effort so far to analyze semen in HIV transmission, Lingappa and colleagues, in collaboration with Mullins and Lisa Frenkel at the University of Washington, hope to analyze semen samples that were taken within six months of infection from HIV-infected men in up to 20 heterosexual transmission pairs. In all of the samples, sequencing has already confirmed that transmission occurred between the partners. The samples come from a study of 3,408 discordant couples from Africa that showed that acyclovir treatment of the HIV-infected partner did not reduce HIV transmission (N. Engl. J. Med. 362, 427, 2010). While the timing between transmission and sampling in the study won’t be very close in most cases, Mullins hopes that the larger number of cases compared with previous studies will provide new insights. “We might be able to identify consistencies not evident previously,” he says.


Additional Clues  

One hint that HIV-infected cells might play a role in heterosexual HIV transmission is that as many as 20% of heterosexual transmissions involve more than one transmitted founder virus, says Ron Swanstrom, director of the University of North Carolina Center for AIDS Research. That’s a much higher percentage than what would be expected if the infections were due to cell-free virus because the probability of infection with cell-free virus is thought to be 1% or lower, and so the combined probability of independent infections with more than one free virus would be expected to be even lower than that. An alternative explanation is that the infection occurs through cells that carry several proviral DNA copies of HIV, especially if additional sexually transmitted infections in the donor result in more HIV-infected cells that could be transmitted. However, it is also possible that infections increase the number of HIV target cells in the recipient, or that circumstances such as a break in the skin of the recipient increase the probability of transmission of multiple viruses. —AvB

Mother-to-child transmission

HIV-infected cells also seem to play a role in MTCT of HIV. A 2004 study of almost 300 antiretroviral (ARV) untreated mother-infant pairs in Africa led by Overbaugh showed that cell-associated virus levels in the mother were a more important predictor of transmission risk to the child than cell-free virus levels (J. Infect. Dis. 190, 1880, 2004). This finding suggests that the children might get infected by ingesting HIV-infected cells from the mother, Overbaugh says.

Aldrovandi and her colleagues are planning an even bigger study of this type, using samples from about 600 mother-child transmission pairs from the Zambia Exclusive Breastfeeding Study (N. Engl. J. Med. 359, 130, 2008). The researchers will correlate the levels of cell-free and cell-associated virus in the mothers’ breast milk with the transmission risk to the infants. One advantage is that the researchers know quite well when the infants were infected, Aldrovandi says. She and her colleagues may also try to determine HIV sequences from some of the mother-child transmission pairs to see more directly whether the cell-associated or the cell-free HIV caused the infection in the infants.

Overbaugh’s study used samples that were collected before mothers were advised to take ARVs during delivery or breastfeeding, which is now becoming more common practice because it has been shown to dramatically reduce rates of MTCT. However, Overbaugh and others have found that ARV treatment of the mother mostly reduces the mother’s free virus level, not the reservoir of HIV-infected cells in the breast milk (AIDS 22, 1475, 2008). That raises a question, Overbaugh says: If, as her 2004 study suggests, cell-associated virus plays an important role in MTCT, how is it possible that ARV therapy dramatically reduces the MTCT risk if it mostly reduces cell-free, but not cell-associated virus levels in the mothers? One possibility, she says, is that the ARVs are actually working as prophylaxis in the infant. Consistent with this, Overbaugh says, a recent study showed that infant ARV prophylaxis during breastfeeding can reduce MTCT even if the mother is not receiving treatment (N. Engl. J. Med. 362, 2271, 2010).

Cell-bound virus challenge models

While most challenge models in NHPs currently use cell-free challenge stocks, in a handful of studies researchers have tried to infect NHPs with cell-associated virus. Studies in female chimpanzees and cynomolgus macaques suggested that infection with cell-associated HIV or SIV respectively is possible, while an attempt to infect female rhesus macaques with SIV-infected cells failed (AIDS Res. Hum. Retroviruses 14 Suppl 1, S119, 1998).

But a study last year (J. Infect. Dis. 202, 337, 2010) that was led by Roger Le Grand, head of the division of immunovirology at the Institute for Emerging Diseases and Innovative Therapies at the Atomic Energy Commission in France, was an important advance towards the development of a challenge model for cell-associated transmission in macaques, says Anderson, who wrote a commentary on the study (J. Infect. Dis. 202, 333, 2010).

In the study, Le Grand and colleagues infected cynomolgus macaques with SIVmac251 and isolated SIV-infected CD4+ T cells and macrophages from their spleen at peak viremia. They washed off any free SIV and placed the infected cells onto the vaginal mucosa of uninfected cynomolgus macaques that had been treated with Depo-Provera to facilitate transmission by thinning the vaginal mucosa. Four of the five macaques challenged this way got infected after one challenge. In addition, when the researchers labeled the cells placed on the vaginal mucosa, they found the cells, as well as SIV, in distant lymph nodes as soon as 21 hours after challenge.

Still, it’s hard to really prove that the macaques in this model got infected from the SIV-infected cells and not from free SIV particles, says Ronald Veazey of Tulane University. “SIV/HIV infected cells in fluids of an inoculum are not just quiescent, they are shedding virus like crazy,” Veazey says. “So if you inoculate animals even with highly washed infected cells, it’s still difficult to prove infection didn’t occur due to shed virus from the cells instead of the cells themselves.”

Le Grand says he addressed this issue by washing the SIV-infected cells several times, and by demonstrating that the supernatants from these washes, which should contain the free virus particles, did not infect the macaques. Still, he adds, it is possible that some SIV particles may have stayed attached to the cells, or that the cells may have infected the macaques by shedding virus close to the vaginal epithelium or in contact with cells in the vaginal epithelium. “We are careful by saying that we transmit infection by cell-associated virus and not really by infected cells because we don’t want to exclude the different mechanisms by which this can occur,” Le Grand says.

Nevertheless, he says, the fact that labeled cells were observed in lymph nodes far from the vaginal mucosa just 21 hours after the challenge suggests that the cells did migrate away from the vaginal mucosa. “If this mechanism is true, it would be a very efficient mechanism to transmit infection,” Le Grand says, and strategies designed to block the free viral particles, such as antibodies, may not work against these infected cells. He says he wants to use the challenge model to test if microbicides that have so far shown some efficacy to protect against cell-free challenge can also protect against cell-associated challenge.

The model is an important advance, Anderson says, in part because the dose of cellular HIV copies needed to infect half of the female macaques is similar to the number she and others have found to be present in human ejaculates. This is in contrast to the super-physiological doses needed to achieve infection with cell-free HIV.

However, the reception by the field to the study has been “flat,” Anderson says. Indeed, some researchers are skeptical. Ashley Haase at the University of Minnesota believes that the cell-associated NHP model still has a long way to go until it can be used as a challenge model. He says that while there is a large body of information on what has been learned from the high-dose cell-free model of SIV infection of rhesus macaques, the literature on cell-associated infections is “scanty.”

Many cell-associated infection studies so far, he adds, are also limited by confounding factors that could affect the results, such as the use of Depo-Provera by Le Grand and colleagues to thin the vaginal mucosa. Haase also notes that he and Eva Rakasz of the University of Wisconsin-Madison used macaques with chemically induced ulcers in the vaginal mucosa when they showed, similar to Le Grand, that SIV-infected cells placed on the vaginal mucosa are disseminated throughout the body (J. Virol. 82, 4154, 2008).

Anderson, for her part, is not deterred. “I don’t intend to give up on our cell-associated HIV transmission research anytime soon,” she says. “My colleagues and I that study cell-associated HIV transmission believe that the sidelining of this research area by leaders in the HIV prevention field may have delayed the development of highly effective HIV microbicides and vaccines.”


Figure 1

Potential mechanisms underlying cell-associated HIV transmission. (a) Columnar epithelium: (1) Infected cell migrates between epithelial cells to infect susceptible host cells. (2) HIV transcytosis through epithelial cells to infect susceptible target cells. (b) Stratified squamous epithelium: (3) Transfer of HIV from infected leukocyte to epithelial cell, which transfers virus to target cells through transcytosis or attraction via release of chemokines. (4) Direct cell-to-cell transfer of HIV from infected leukocyte to target cell via viral synapses. (5) Transepithelial migration of infected leukocyte to infect target cells within the epithelium. (6) Transepithelial migration of infected cell to infect target cells in the subepithelium or draining lymph nodes. Originally published in AIDS 24, 163, 2010.