HIV Transmission: The Genetic Bottleneck

Researchers are finally getting a handle on the nature of the virus that is transmitted and establishes infection, offering promising hints for AIDS vaccine development

By Simon Noble

Plenty of scientific quandaries cause AIDS vaccine researchers restless nights, but one overriding challenge has always trumped them all and is a source of nightmares—the astonishing degree of genetic diversity that HIV presents. To put it in perspective, the genetic diversity of HIV within a single infected individual after six years of infection is roughly equivalent to the entire global diversity of influenza A virus in a year. With an estimated 33 million people around the world currently infected with HIV, that makes for a mind-boggling degree of genetic diversity.

Designing and developing an AIDS vaccine to tackle that degree of diversity can seem an overwhelming prospect. So any indication that the virus that is transmitted and establishes a new infection is less diverse would be welcome news. Evidence has been accumulating for the past decade suggesting that the dominant virus during primary infection is relatively homogeneous because of a genetic bottleneck during transmission, which effectively limits the degree of variation. And more recently, technical advances, the formation of large-scale consortia, plus a growing realization that the very earliest events of HIV infection are crucial, have given researchers a better look at the enemy—the virus that a preventive AIDS vaccine would have to vanquish.

Sampling advances

Although HIV was first identified over 25 years ago, researchers are just beginning to unravel the earliest events in HIV infection, the crucial window of opportunity that a vaccine needs to exploit. Within the first two weeks, HIV is disseminated throughout the body and rapidly devastates the reservoir of CD4+ T cells, firmly establishing infection. A detailed molecular understanding of the transmission and the early evolution of HIV, including a precise description of the transmitted or early founder virus, would seem to be critical steps in the development of an effective AIDS vaccine. But one of the major reasons for the lack of insight until recently is pragmatic; HIV is, most often, a sexually transmitted infection, so for obvious reasons it’s extremely difficult to identify and study the actual infectious event.

That’s where improved sampling has paid off. Many efforts to identify newly infected individuals have come about through the various research programs that have been established in recent years. Susan Allen of Emory University and director of the Rwanda Zambia HIV Research Group has established cohorts of discordant couples—in which one partner is HIV infected and the other is not—in Rwanda and Zambia with support from IAVI. The development of these cohorts entails screening huge numbers of couples, identifying those that are HIV discordant, and then persuading those couples to come back on a regular basis to undergo sampling and counseling. The nature of the cohort allows researchers to identify newly infected individuals when they are viral antigen (p24) positive but antibody negative.

Discordant couple cohorts uniquely enable both the study of the virus that establishes infection and the virus population in the chronically infected partner, and in numbers of individuals that would be impossible in other cohorts. “Having the viral quasispecies that the transmitted virus originated from, you can ask was it a dominant or minor variant, was it enriched in the genital compartment, what’s the history of the transmitted virus?” says Cynthia Derdeyn, assistant professor of pathology and laboratory medicine at Emory University, who studies the samples from these cohorts. Traditionally, cohort members have returned to the clinic at three-month intervals, but recently they’ve been asked to return for monthly visits. This allows for more frequent sampling and also helps reinforce counseling messages about condom use among couples.

Other recent efforts by the Center for HIV/AIDS Vaccine Immunology (CHAVI) to identify newly HIV-infected individuals have required keen detective work to track down historical serial plasma specimens.

The genetic bottleneck

When a study appeared in 2004 suggesting that the virus that establishes HIV infection goes through a severe genetic bottleneck, and might be more sensitive to antibody neutralization, it caused quite a stir (1). Eric Hunter, a professor at Emory University, and colleagues, including Derdeyn, studied eight heterosexual transmission pairs from a discordant couple cohort in Zambia, four male-to-female (M-F) and four female-to-male (F-M) subtype C HIV transmissions. Viral env sequences, specifically the region spanning the V1-V4 loops, were studied from peripheral blood mononuclear cells (PBMCs) and plasma. They found that an extreme bottleneck occurred in all the transmissions, which they interpreted as the transmission or outgrowth of a single sequence from the donor quasispecies. They also found that the transmitted virus tended to have shorter V1-V4 regions, which meant it had fewer glycosylation sites than the donor virus. A likely functional consequence of the fewer glycosylation sites is greater exposure of the CD4 binding domain, which often results in an augmented susceptibility to antibody neutralization. Indeed, the authors found that the recipient viruses were up to 10 times more sensitive to neutralization by antibody present in the plasma from the donor.

But much of the excitement that ensued was due to misinterpretation, because even though the founder viruses seemed to be more sensitive to antibody in the donor plasma, there was no statistical difference in their sensitivity to pooled plasma from subtype C infection. “It didn’t appear that the founder viruses were globally more sensitive to neutralization,” says Hunter. “We were at pains to convey that the neutralization sensitivity was not absolute, only in relation to the antibodies present in the donor plasma.” Rather, Hunter says the founder viruses “seemed as if they had lost some of the protection that had developed in the majority of circulating viruses in the chronically infected partner.”

This genetic bottleneck at the point of HIV transmission was recently confirmed by George Shaw, a professor in the department of medicine at the University of Alabama at Birmingham, and colleagues working within CHAVI when they defined the env genes of transmitted subtype B HIV from 102 plasma donors who became newly infected with HIV, a study that Hunter calls “a tour de force in terms of numbers.” Shaw’s group employed some technical and theoretical insights, using a combination of single genome amplification (SGA) and direct sequencing to analyze viral RNA in plasma samples—virus in plasma has an extremely short lifespan and so reflects very recent viral replication. From historical serial samples they were able to work back and identify the sample from each individual plasma donor closest to the infectious event and, using a mathematical model of random viral evolution, infer unambiguously the transmitted founder env sequence in 98 of 102 individuals (2).

In the majority of transmissions (76%) in this cohort, a single virus was responsible for productive clinical infection, with the remainder showing evidence of infection by between two and five viruses. But Shaw is careful to define exactly what he means when he talks about the transmitted virus. “Our inference of the transmitted founder viral sequences obtained near peak viremia is, I think, well accepted in the field right now. If we qualify it by saying that these are transmitted founder sequences that are leading to productive clinical infection, people agree with that,” he says. If, as Shaw’s work suggests, limited viral evolution precedes peak viremia, it suggests that a vaccine would only need to be effective against a small inoculum. “In the first two to six days of infection, the extent of viral diversity is quite low,” he says.

Shaw and colleagues also reported some biological phenotypes of the transmitted founder viruses that are important for vaccine design. Invariably, they were R5 tropic, meaning they used CCR5 as a coreceptor to gain entry into cells. “The virus is not using X4 [CXCR4 as a coreceptor] and then being selected for R5 tropism,” says Shaw, “it’s R5 tropic at the moment of transmission.”

Also, the phenotype of the envelopes is typical for primary virus strains—R5 tropic, CD4 dependent, and both the coreceptor-binding surface of gp120 and the V3 sequences are effectively concealed. “If a vaccine were to be based either on CD4-induced epitopes or V3 epitopes, our data would suggest it’s likely to be ineffective against the transmitted founder virus,” says Shaw. He and his colleagues also determined that the susceptibility of the founder envelopes to broadly neutralizing antibodies was similar to that of primary virus strains. “It’s important to know that it’s not going to be a magic antibody that’s going to block all these viruses, they’re all different,” says Derdeyn. In future studies, Shaw wants to investigate additional biological, immunological, and antigenic properties of transmitted HIV to determine how these viruses behave, their tissue/cell tropism, and to identify a possible Achilles’ heel that could be exploited by vaccine researchers.

Hunter has also confirmed the genetic bottleneck at HIV transmission in additional studies in the discordant couple cohorts, extending it to another subtype of HIV. In a paper currently in press, his group has now looked at a total of 20 transmission pairs—11 subtype C and 9 subtype A infections—and in 90% of transmissions they see a single genetic env variant initiating infection. His group also used the SGA and direct sequencing methodologies since they wanted to determine the frequency with which the genetic variant in the newly infected individual was present within the donor and determine whether it was the most frequent or a rare variant in donor plasma or PBMCs. They frequently see an identical or very closely related variant present in the donor, but in almost every case that variant is a minor species within the donor quasispecies. “We see a very homogeneous virus population early on, and can track that back to what we believe is the founder virus,” says Hunter. “The added value of our study is that we can relate that back to the virus in the donor.”

These studies are still ongoing but so far Hunter and colleagues have not yet been able to answer a perennial question in HIV research—whether the initial infection was caused by cell-associated virus or cell-free virus. “I think that’s still an open question in the field,” he says.


To get an even clearer picture of the actual transmission events, Hunter’s group is now characterizing the virus in the genital fluids rather than the peripheral blood of the donor. In a presentation at the recent AIDS Vaccine 2008 Conference in Cape Town, Debi Boeras, an affiliate scientist in Hunter’s group, presented data indicating similar evidence of the genetic bottleneck with viruses that are present in the genital compartment of the donor. They studied five F-M transmissions and three M-F transmissions, and while there is compartmentalization of the virus in the donor genital fluids, “the variant that is most closely related to the founder virus in the recipient is not in those enriched populations at the time we’re looking,” says Hunter. It is not yet clear whether this is a real finding or another consequence of the practical barriers hindering the study of HIV transmission. “We obviously can’t be there at the time that transmission actually occurs, and we’re now trying to determine just how much virus turnover there is in these genital populations,” says Hunter. “It may be that the virus that establishes infection in the genital mucosa and then becomes systemic may have to have specific properties that enrichment in the genital fluid doesn’t provide.”

Multiple founders

Given the mounting evidence that suggests HIV passes through a genetic bottleneck during sexual transmission, the frequency with which two or more viruses establish infection actually occurs more often than would be expected—in 24% of cases in Shaw’s study. “That’s much higher than you’d expect by chance,” he says, “something else must be going on.” Hunter and colleagues found that in individuals who became HIV infected by somebody other than their spouse—termed epidemiologically unlinked transmissions—the frequency with which more than one virus established infection increased dramatically. They also found a statistically significant association between the presence of either a chronic ulcerative disease or an inflammatory genital infection and multiple genetic variants establishing infection in the recipient. “It seems that the genetic bottleneck can be modulated by infections or inflammatory reactions, either by compromising the mucosal barrier or providing an enriched population of target cells for new infection,” says Hunter.

Vaccine implications

These findings of a single founder virus and limited viral evolution may have important implications for vaccine design. “The good news is that in the majority of cases, very early on a single defined virus is initiating infection,” says Hunter. “That suggests that the frequency with which a virus that has the capacity to breach the mucosa and become systemic is quite low. It gives us hope that if you can contain the newly infecting virus for long enough, with neutralizing or even binding antibody, to allow the CTL response to be triggered, you may be able to confer some protection with a vaccine and stop virus from becoming systemic.”

Shaw believes his data corroborates what the clinical data shows—HIV transmission is an uncommon event, perhaps as low as 1:1000 sexual acts under some circumstances. So it fits that infection is probably going to be due to only one or two viruses. “A vaccine that has breadth and potency will probably be effective because these are uncommon events,” says Shaw.

But John Moore, a professor of microbiology and immunology at Weill Cornell Medical College, advises caution. He thinks the perception that “because there’s only one virus expanding in the new host, you only have to block the transmission of one virus” is dangerous. “That’s like saying that to stop pregnancy a contraceptive only needs to stop one sperm from among the millions present,” adds Moore. “The analogy is not exact, but I’m concerned that some people in the vaccine and microbicide fields might take home the wrong message and misunderstand what’s really going on.”

Monkey transmission

Both Hunter and Shaw are now using the simian immunodeficiency virus (SIV)/rhesus macaque model to complement their clinical studies and gain further insight into acute infection. The general aim of these SIV studies is to elucidate the early replication events in the eclipse phase, the period between virus transmission and the broader dissemination of the virus, when a vaccine might have its best chance of containing or eliminating an infection. “If we can understand the kinetics and the events that occur in this eclipse phase and then look at the effect of candidate vaccines on that eclipse phase, it could be a powerful tool,” says Shaw. That applies equally to clinical studies of vaccine candidates; characterizing founder viruses that establish infection after breakthrough infection in vaccinees (or SIV challenge in macaques) allows researchers to study routes of immune escape. “If we vaccinate with an immunogen that raises responses to certain T-cell epitopes, we can look very, very early at the virus that leads to breakthrough infection in those vaccinees and determine whether or not there is strong early selection pressure at the relevant epitopes,” says Shaw.

The research groups of Norman Letvin, professor of medicine at Harvard Medical School, and Gary Nabel, director of the Vaccine Research Center at the National Institute of Allergy and Infectious Diseases, have developed a low-dose, atraumatic (non-abrasive) mucosal challenge macaque model that uses heterogeneous inoculums. Shaw’s group has studied such animals and finds that “the virus that is transmitted is literally the same virus, nucleotide for nucleotide throughout the entire env gene, to a virus that is present in the inoculum,” says Shaw. They have also seen the same phenomenon in a clinical transmission pair—in an acutely infected individual with very low virus diversity and in a second individual who contracted HIV infection from that person, the transmitted virus is identical to a virus identified in the donor—not a single nucleotide has changed between the moment of transmission and about three to four weeks later at peak viremia.

Hunter is working in collaboration with David Evans, a researcher at the New England Primate Center at Harvard University, to set up a similar multiple low-dose rectal challenge in macaques. So far he finds a similar genetic bottleneck to that seen in humans. In four of six macaques challenged with SIV, a single genetic variant from the pool of challenge viruses established infection; the other two animals’ infection arose from two viruses.

Moore thinks there is an important message to the field in these and other SIV studies of the transmitted virus, and hopes it will remove some of the prejudice against animal models, at least from the perspective of the viral dose. “It appears that it doesn’t really matter which nonhuman primate model is used, low- or high-dose challenge, rectal or vaginal, the data is very similar to naturally, mucosally infected humans,” he says. “That undermines the argument that animal models are misleading due to the dose of virus being too high, or that the model is too stringent. The data coming out supports the contention that animal models are reasonable mimics of the human infection, at least from the perspective of the challenge dose.”

Benefits of viral diversity?

Some studies indicate that the genetic diversity of HIV within an infected individual may actually be good news for vaccine researchers. In chronic infection, HIV is continually being selected by the host immune response, both neutralizing antibody and cytotoxic T-lymphocyte responses, and this immunological pressure forces the virus to mutate and generate escape mutants. The virus that gets transmitted, then, has been selected for survival in one immunogenetic environment. Hunter and his colleagues investigated how that viral quasispecies copes when it enters a new environment, how rapidly it escapes, and what happens to those escape mutations that were selected for in the chronically infected partner to see whether these escape mutations confer a fitness defect on the transmitted virus before it has chance to revert. Also, on a population basis, they looked at whether escape mutations in particular genes affect subsequent viral load.

In a study of 114 discordant couple transmission pairs, he and his colleagues found that multiple mutations in thenef gene didn’t seem to impair the transmitted virus. In contrast there was a progressive effect of escape mutations in gag, particularly in p24, such that when five or more escape mutations were present, the viral load in the newly infected partner was significantly lower (3). This suggests that the virus is placed at a disadvantage by having had to escape the immune response in the infecting partner. Those partners whose immune systems are most effective at targeting Gag may actually transfer viruses that are least fit, and this may have a positive long-term effect for the newly infected partner because peak viral loads and the attendant destruction of mucosal tissue might be reduced in those individuals. “It might also be telling us that if we can use a vaccine to target cellular immune responses to as many epitopes as possible in Gag,” says Hunter, “then the virus is really fighting an uphill battle because it’s trying to escape the immune response, but at the same time it’s decreasing its replicative capacity and committing harakiri.”

1. Science 303, 2019, 2004
2. Proc. Natl. Acad. Sci. 105, 7552, 2008
3. J. Exp. Med. 205, 1009, 2008