It is estimated that 36.9 million people worldwide are infected with HIV (go.nature.com/2pugvhd). Drug treatment can prevent this infection from developing into AIDS and becoming lethal, by controlling the level of virus in the bloodstream. However, this therapy doesn’t eradicate HIV, and is associated with toxicity and conditions such as metabolic disorders1. As part of ongoing efforts to improve the treatment options available, clinical-trial results are now reported in two papers from the same research group — one in Nature by Mendoza et al.2 and the other by Bar-On et al.3 in Nature Medicine. The studies assess the effects of a treatment for HIV that uses two antibodies to target a viral protein.
The current standard treatment for HIV infection is long-term use of antiretroviral therapy (ART), a daily regime of drugs that block steps needed for viral replication. If ART treatment ceases, virus becomes detectable in the bloodstream within days, a phenomenon termed viral rebound that can lead to progression towards AIDS. ART does not provide a cure because the drugs do not kill infected cells in the viral reservoir4, which is established when virus inserts its genome into the genome of a host immune cell (Fig. 1).
ART drugs can prevent viral production from the part of the reservoir termed the active viral reservoir, in which viral replication occurs. However, these drugs cannot target the viral reservoir that exists in a ‘dormant’ state, termed the latent viral reservoir. A latent viral reservoir can exist in certain tissues5; here, virus production occurs at an extremely low level and might increase in the future. This ‘hidden’ reservoir can evade destruction by the immune system. If these cells could be targeted, more-effective treatments might be possible.
There has been interest in the potential use of HIV-targeting antibodies to control the viral level in the bloodstream. Such antibodies can bind to and block a protein on the viral surface called Env, which is needed for HIV entry into host immune cells called CD4+ T cells. The discovery6 of highly potent antibodies, termed broadly neutralizing antibodies (bNAbs), boosted this idea. These antibodies target regions of the Env molecule that are present in nearly all HIV strains. Individual bNAbs have already been tested6–9 for use as clinical treatments for HIV. They are safe and well tolerated10,11, and, in contrast to the daily dosage necessary for ART, administration is needed only every few weeks to maintain constant bNAb levels in the bloodstream and potentially in tissues.
Previous clinical studies7–9 testing the individual effects of two bNAbs12,13 — either 3BNC117 or VRC01 — found that the level of HIV in the bloodstream was initially suppressed for 6–10 weeks, but then viral rebound occurred, and the studies7–9 reported the presence of antibody-resistant viral variants.
Mendoza et al. and Bar-On et al. carried out phase Ib clinical trials (small-scale trials to test the safety of a treatment) to investigate whether combining two bNAbs (3BNC117 and 10-1074) that target distinct sites on Env might decrease the probability of virus resistance occurring, and might control virus levels in HIV-infected people who did not receive ART during the trial period. Mendoza and colleagues assessed the effect of giving the two antibodies to 11 people who were temporarily stopping ART. Bar-On and colleagues analysed the effect of these two bNAbs on seven people who hadn’t yet received ART.
Both studies reported an impressive reduction in bloodstream HIV levels compared with levels at the start of the trial. Viral rebound took many weeks to occur, and when it did, there was little or no evidence that viruses were resistant to both bNAbs. This suggests that using two antibodies might offer an alternative treatment option to ART, pending further studies.
Mendoza and colleagues observed that resistance to 10-1074 developed more rapidly than that to 3BNC117. This could be explained if there was a more rapid natural decline in the blood levels of 3BNC117, which had a shorter half-life than 10-1074. It effectively resulted in periods of treatment with a single type of bNAb rather than with the intended two. If the antibody dosage could be adjusted so that the half-lives of the two bNAbs were more closely matched, the development of treatment-resistant viral variants might be avoided. However, triple antibody combinations might be necessary to tackle the emergence of antibody resistance.
Maintained viral suppression from antibody treatment was evident in Mendoza and colleagues’ study. The authors found that when the bNAbs were introduced 2 days before ART was stopped, and further doses given 3 and 6 weeks later, virus in the bloodstream was suppressed to undetectable levels for a median time of 21 weeks before viral rebound. The rebound timeframe ranged from 5 to 30 weeks, and 9 of the 11 trial participants maintained viral suppression without rebound for more than 15 weeks.
One way to determine whether the characteristics of viral infection change after bNAb therapy is to compare the diversity of the nucleotide sequences that encode the Env protein before and after treatment. Mendoza and colleagues found that the rebounding viruses after antibody treatment had low sequence diversity, whereas a high degree of diversity has been observed in viral rebound after ART treatment14,15. It is possible that this difference arises because the antibodies restrict the growth of viruses from the tissue viral reservoir, in addition to targeting actively replicating viruses in the bloodsteam, leading to a low level of sequence diversity.
How effectively might antibodies control virus levels without pretreatment with ART? In Bar-On and colleagues’ study, the treatment suppressed virus levels in the bloodstream for an average of 86 days, about 60 days longer than previously observed7 after treatment with just one HIV-targeting bNAb. This dual treatment approach in people who already had detectable levels of HIV in their bloodstream clearly limited the emergence of resistant viral variants, providing a notable advance on treatment with a single type of antibody. However, suppression was complete only in people who had low levels of virus in their bloodstream at the beginning of the trial.
The results provide cause for optimism that a cocktail of antibodies might provide substantial and durable viral control in the absence of ART. However, such enthusiasm must be tempered with caution. Bar-On and colleagues initially conducted laboratory tests to determine whether their study participants had a viral strain that would be sensitive to the bNAbs. In three individuals in whom the virus level rebounded most quickly, these antibody-sensitivity tests failed to detect viruses that had partial or complete pre-existing resistance to one or both bNAbs. This suggests that a more-sensitive test than that used by the authors will be needed to assess whether people are likely to respond to a particular bNAb combination.
In the war against HIV, there has been some controversy regarding the use of bNAbs as a treatment approach. Antibodies have been thought of as providing defences equivalent to a naval blockade, stopping HIV from initially infecting or preventing the infection from spreading. But once this barrier is breached and infection takes hold, it is assumed that T cells are the front line of immune defence against infected cells, with antibodies relegated to bystander status. However, models of infection with a simian–human chimaeric virus called SHIV in primates suggest that antibodies might have an active role in tackling infected cells.
In these primate models, human antibodies become widely distributed in tissues16, and could be acting in a manner equivalent to minesweepers, by limiting the spread of infection. Not only are such antibodies distributed efficiently to tissues far from the site of antibody administration, but they can also help to destroy infected cells, at least in the early stages of an infection, and to prevent the establishment of the viral reservoir16,17. Once viral infection is established, the introduction of bNAbs that target Env can suppress the virus to undetectable levels in the bloodstream, as long as the antibodies persist there18,19. The role of these antibodies in tissues is not fully understood. When the antibody levels decay, viral rebound occurs and immune defences are mediated mainly by T cells20.
Given the results in primates, is there any evidence that bNAbs can tackle the viral reservoir in human HIV infection? Mendoza and colleagues used an in vitro quantitative virus-outgrowth test to assess the viral-reservoir size. They found no significant differences in reservoir size before and after bNAb treatment. However, longer and more-effective bNAb treatment and larger study groups might be needed to establish definitively whether bNAbs affect the viral reservoir.
As strategies are sought to target the reservoir, it is encouraging to know that bNAbs might provide a useful weapon in the arsenal of viral-control tools. Moreover, if it turns out that a cocktail of bNAbs could act as a potential temporary replacement for ART, this type of ‘drug holiday’ might give people time to recover from ART-induced toxicity.
Nature 561, 468-470 (2018)