Several research papers published in the last month have reported strong correlations between specific immune responses and protection against HIV infection or its effects. These include a comparison of women in the CAPRISA 004 microbicide trial who became infected and ones who did not, despite high exposure to HIV; an analysis of female sex workers in Kenya who similarly seem to be able to resist HIV infection; and a study of immune responses in a group of monkeys given pre-exposure prophylaxis (PrEP) who all became infected but subsequently developed much lower viral loads.
The hunt for correlates of protection
One of the most critical aspects of developing an effective vaccine against HIV is to define exactly what constitutes a protective immune response to the virus. Such immune responses could either be associated with protection against infection, so-called sterilising immunity, or protection against disease in people who are infected, so-called functional immunity.
Without reliable and consistent ‘correlates of protection’, developers of vaccines must rely on educated guesses and lucky finds in order to progress towards an effective vaccine.
In this hunt they are hampered by several factors. Animal immune responses do not always mimic human ones; resistance to infection or progression does not seem to be characterised by one type of immune response, but rather different ones in different people; there are a huge number of genetic variants that have to be searched through to find protective ones; and even if a strong correlate of protection is found, it is hard to ‘reverse-engineer’ a vaccine that will produce this response in the body.
Infected and non-infected women in CAPRISA 004
One way of looking for correlates of protection is to study so-called ‘highly exposed seronegatives’, people who do not acquire HIV despite frequent contact with it.
In a study that looked at women who took part In the CAPRISA 004 study of a vaginal tenofovir-gel microbicide, the researchers noted that while younger age and genital herpes (HSV-2) infection were both associated with infection with HIV in the study, factors such as the frequency of sex, number of partners and condom use were not, and a large amount of the variation in susceptibility to infection remained unexplained. They therefore compared immune responses to HIV in 44 selected women who became infected (33 on tenofovir gel and 11 on placebo) and 37 women who, despite higher levels of risk behaviour, did not (22 on tenofovir and 15 on placebo).
The infected women were on average younger (mean age 23.3 versus 27.6), but while the non-infected women had sex on average 11 times month, the infected women had it less than six times. These are baseline figures, but did not change during the trial.
The researchers found that women who were infected had higher levels of certain cytokines (immune-modulating chemicals) including TNF-α, IL-2, IL-7 and IL-12. They also found higher platelet counts in the blood of women in the visit immediately before infection compared with non-infected women, even though platelet counts were the same in both groups at baseline. All these factors point to a higher state of immune activation existing in infected women shortly before HIV infection.
When they looked deeper, the researchers found that the cytokine and platelet increases were manifestations of the activation of natural killer (NK) cells, a component of the so-called innate immune system.
Higher animals have three main branches to their immune system. They have the humoral system (antibodies) and the cellular system (T-cells), collectively called the acquired immune system, which are finely tuned to ‘remember’ specific infections and recognise and defeat them again if re-encountered. But we also share with all animals and plants the innate immune system, a less precise but faster response, which recognises general characteristics of infected cells and tissue damage. NK cells are central to this response; they send signals that increase production of broad-spectrum antibodies called immunoglobulins, which recruit acquired immune system cells to sites where the body’s defences are broken. They also stimulate blood platelet production, which prepares the body for wound healing and physically ‘tangles’ bacterial invaders in blood clots.
Women who were infected had higher levels of receptors (cell-surface proteins) called HLA-DR and CD69 on the surface of their NK cells than uninfected women, and lower levels of CD38. In T-cells, high levels of CD38 indicate activation, but in NK cells the reverse applies.
These differences were all significant in themselves, but when all three were lumped together, as composite indicators of innate immune activation or quiescence, they were much more so. In multivariate analysis, women with composite innate immune activation were eleven times more likely to acquire HIV and women with composite immune quiescence 17 times less likely.
In terms of other variables, women on placebo were seven times more likely to acquire HIV than women on tenofovir and HSV-2 positive women 22 times more likely. Women were also 28% less likely to acquire HIV for every year older.
The strength of correlation with particular immune indicators is not only good news in terms of being able to predict who may or may not be vulnerable to HIV infection, but, as the researchers point out, suggests new ways of making new HIV prevention technologies more effective. The researchers point out that the tenofovir-gel microbicide in CAPRISA 004 was only 54% effective even in women who claimed 100% adherence, but if it could be coupled with therapies that dampened down NK-cell activity, many more infections might be prevented.
Antibody-driven responses in non-infected sex workers
Meanwhile, a study presented at the recent 19th International AIDS Conference in Washington, of immune responses in highly exposed but uninfected Kenyan female sex workers, found evidence to corroborate a hypothesis already advanced in another paper published this month (see this report). In this paper, evidence from the trial of the first HIV vaccine to show some efficacy, RV144, show that it may have conferred some protection against disease progression in people who acquired HIV as well as some protection against infection. The RV144 researchers suggest that an antibody called immunoglobulin A (IgA), which is produced in large amounts by mucous membranes in the genitals and gut in response to foreign infections, may be responsible for some of this immunity/disease resistance.
Why would an antibody that has no direct effect against HIV produce immunity to it? The answer is that IgA and similar antibodies, like the NK cells that guide their production, congregate wherever there are cells that look as if they have been infected and attach themselves. These attached antibodies then act as ‘flags’ to alert the slower but more potent and precise CD8 cells and HIV-specific antibodies of the acquired immune system to come to the site and either destroy infected cells or neutralise free virus. This process is called antibody-driven cellular cytotoxicity (ADCC) and may be responsible for the somewhat unexpected success of the RV144 vaccine.
In the Kenyan sex workers’ study, researchers from Rush Medical Centre in Chicago studied the ADCC activity, against cells displaying the HIV gp120 envelope protein, of cells taken from cervico-vaginal fluid in ten highly exposed HIV-negative women and two who had acquired HIV .
Cells from the non-infected women showed consistently high levels of IgA-driven ADCC activity. In contrast, the cells from the two infected women showed no or low levels of response involving IgA. They did show high levels of response of ADCC driven by another broad spectrum immunoglobulin, IgG, which only appeared several years after infection, though the researchers did not comment on whether this was associated with any kind of viral control.
Immune responses in monkeys infected on PrEP
Finally, a study of pre-exposure prophylaxis (PrEP) in monkeys, in which the PrEP regimen failed to prevent infection, nonetheless found that monkeys who were infected on PrEP developed viral loads two logs (one hundred times) lower than monkeys infected while not taking PrEP.
The monkeys infected while on PrEP had low levels of inflammatory cytokines (immune-modulating chemicals) around the time of infection, because their bodies were responding to a much lower peak level of virus in their blood. They also had 100 times the level of CD4 cells in their blood at this point (because there was less HIV destroying cells) and also – in a finding that brings us back to the findings in the CAPRISA study – about half as much interleukin-15 (IL-15), a cytokine that stimulates the activity of natural killer cells.
They also showed no signs of the decimation of gut-associated lymphocytes (white immune cells of all types) that is normally seen around the time of peak viral load in animals and humans who contract HIV, indicating that infection had happened without the long-term immune dysregulation that is normally a legacy of HIV infection.
A series of tests showed that the CD4 and CD8 T-cells in the monkeys given PrEP responded to a broader range of HIV proteins during the period of highest viral load, though this alertness to HIV was not maintained over time – a good thing if the idea is to avoid chronic immune over-activation. The monkeys given PrEP also had fewer chronically activated memory cells and relatively more of the quiescent central-memory cells than those not given PrEP.
Finally, the viral loads in the monkeys given PrEP were no longer suppressed and went up two logs if their CD8 cells were artificially depleted, showing that HIV-specific CD8 cells were playing an important part in viral control.
It is important to note that the blunted viral load seen in monkeys who are infected on PrEP is not a new phenomenon – researchers first saw this in studies published in 2008. So far, however, a similar study of ‘PrEP failures’ in humans had not been published. The kind of virus used in this study also does not exactly mimic HIV, as it is not pathogenic and the infection is eventually contained naturally, even in monkeys not on PrEP. Further studies using viruses more analogous to human ones are taking place.
However this study does clarify the sequence of immune events in animals that are receiving ARVs at the time of infection – as is inevitably likely to happen in some people taking PrEP. The researchers also tried to produce HIV drug resistance in the monkeys by continuing PrEP for several weeks past the date of infection, but no monkey developed drug resistance.
What should we make of these varied results? The studies show that the body’s immune response to HIV consists of a complex series of events, and that high levels of activation against HIV may be helpful at one point but harmful at another. It shows that activity in one part of the immune system may need to be coupled with quiescence in another part, if HIV is to be contained. And it shows that immune responses that later become harmful when they become systemic, such as the general level of inflammation provoked by the innate immune system, may in some ways be useful earlier on in infection when concentrated at the specific infection sites.
In general, they show that immune responses in the mucous membranes need to be much more carefully studied, as there is a lot we still do not know about how the body repels or, conversely, embraces HIV.
Naranbhai et al. Innate Immune Activation Enhances HIV Acquisition in Women, Diminishing the Effectiveness of Tenofovir Microbicide Gel. Journal of Infectious Diseases August 2012. Epub ahead of print – see abstract here.
Aziz M et al. (presenter Baum L) Comparative activity of IgA mediated antibody dependent cell-mediated cytotoxicity (ADCC) in the genital mucosa of HIV seroconverters and highly exposed seronegative women. Nineteenth international AIDS conference, Washington DC. Abstract MOLBA04. 2012. See abstract here.
Kersh EN et al. Reduced inflammation and CD4 loss in acute SHIV infection during oral pre-exposure prophylaxis. Journal of Infectious Diseases 206(5):770-9. 2012. See abstract here.