HIV vaccine update: the “Miami macaque” as proof-of-concept breakthrough?
Richard Jefferys, TAG
The idea of using adeno-associated virus (AAV) as a vehicle to deliver genes encoding anti-HIV broadly neutralising antibodies (bNAbs) has been around for some time, and has been covered a number of times previously on the blog (see posts from May 2009 and January 2013). [1, 2]
The approach was first developed by Phil Johnson as a possible way of circumventing the challenges associated with inducing broadly neutralising antibodies using traditional vaccines. AAVs primarily take up residence in muscle tissue and can act as a factory for producing proteins encoded by genes inserted into the AAV genome. At the opening lecture of the 8th HIV Persistence Workshop in Miami, the researcher Ron Desrosiers presented an update on efforts to deliver bNAbs with AAV, including the intriguing tale of a macaque in which the method appeared to have a profound therapeutic effect. 
Many years ago Desrosiers, then at the New England Primate Research Center, collaborated with Phil Johnson on the first study to demonstrate that anti-SIV antibodies delivered by AAV could protect macaques against a highly pathogenic SIV challenge.  Desrosiers has since moved to the University of Miami and is now exploring the potential of AAV to deliver the more recently discovered potent bNAbs.
At the workshop, Desrosiers described a preliminary study in which four macaques were infected with a SHIV AD8 challenge virus and, 86 weeks later, given three AAV vectors encoding the bNAbs 10E8, 3BNC117 and 10-1074, respectively. Because they originated in humans, the bNAbs were modified to make them compatible with rhesus macaque antibodies (rhesusised, to use Desrosiers term). No antiretroviral therapies were employed in the experiment.
Evaluations of bNAb levels after AAV administration produced disappointing results: 10E8 was very low or undetectable in all cases, 3BNC117 was delivered successfully in just one out of four animals and 10-1074 achieved significant levels in three out of four. As Desrosiers and colleagues explained in a paper published in Molecular Therapy last year, the problem was caused by the generation of antibodies against the AAV-encoded bNAbs (anti-antibody responses). This problem was also seen in SIV prevention experiments. 
There was a more encouraging finding, however. One macaque developed sustained levels of both 3BNC117 and 10-1074, and this was associated with a persistent decline in SHIV AD8 viral load to undetectable levels – below 15 copies/mL in 26 samples taken over a 24-month period (viral load analyses were conducted by Jeff Lifson at the National Cancer Institute). Virus reservoirs also became undetectable: 62 weeks after AAV administration, no SHIV AD8 could be recovered from 180 million peripheral blood mononuclear cells (PBMC) using a quantitative virus outgrowth assay (QVOA).
Several additional experiments were conducted in an attempt to ascertain if a cure of the challenge virus might have been achieved. Cells from an entire lymph node sampled after 74 weeks were transferred into an uninfected macaque, without causing SHIV AD8 infection. A follow up five weeks later, in which approximately 140 million cells derived from an extracted cluster of lymph nodes were transferred, also failed to establish SHIV AD8 infection. Desrosiers presented these results earlier this year in a talk that is available on Youtube (thanks to @DanWilliamsVisa for sharing this). 
In the workshop lecture, Desrosiers reported that 87 weeks after AAV was given, SHIV AD8 was finally recovered at very low levels on three occasions by QVOA, with the frequency of infected cells estimated to be ~1 in 50 million PBMC (depleted of CD8 cells). While this ends hopes that the virus may have been entirely eradicated, Desrosiers noted that the animal—dubbed the Miami macaque by some of his colleagues—might legitimately be described as functionally cured, although he acknowledged the major caveat: it is just one case.
Desrosiers’s research group is now focused on circumventing the problem of anti-bNAb antibodies. The AAV variant used in studies to date has been AAV1, and he showed evidence that AAV8 appears less prone to inducing anti-antibodies. This may be because AAV8 is tropic for the liver, a site in the body where immune responses against vector-encoded bNAbs are less likely to be induced (a phenomenon known as the “liver tolerance effect”). 
Experiments in which AAV8 administration was followed later by AAV1 have suggested the combination might further reduce anti-antibody levels; in essence a type of “prime-boost” in which the booster is enhancing immunological tolerance rather than enhancing immune responses. An additional strategy involves including a piece of genetic code in the AAV vector that is designed to shut down antigen presentation by cells (e.g. dendritic cells) that might otherwise promote the development of anti-antibodies.
Plans are now underway to conduct a study in which 12 macaques infected with SHIV AD8 will be divided into two groups and receive either AAV8 vectors encoding 3BNC117 and 10-1074 (possibly with an AAV1 boost) or antiretroviral therapy.
There is one ongoing trial of an AAV1 vector encoding a bNAb (PG9) in humans, launched several years ago by a collaboration involving Phil Johnson and the International AIDS Vaccine Initiative (IAVI). The trial population is HIV-negative men. Results were initially due in January 2016 and have been eagerly anticipated. In his presentation at the Persistence Workshop, Desrosiers stated that he recently attended a meeting at the Gates Foundation at which preliminary data were presented and obtained permission to disclose them. 
According to his description, what the results revealed is that the anti-antibody problem is not limited to macaques: nine trial participants that were analysed did not have detectable PG9 levels, but seven out of the nine had readily detectable anti-PG9 antibodies. There may be subtleties to the findings that could not be conveyed in a brief aside, but it sounds like the promise of the AAV approach – both in the preventive and therapeutic contexts – will only be realised if a means to avoid anti-antibodies can be developed.
Jefferys R. Update on AAV vectors as delivery vehicles for broadly neutralising antibodies: Ron Desrosiers and the Miami macaque. TAG Basics Science Blog. (14 December 2017).
- Jefferys R. Genetically engineered immunity. TAG basic science blog. (18 May 2009).
- Jefferys R. David Baltimore and vectored immunoprophylaxis. TAG basic science blog. (18 January 2013).
- Desrosiers RC. Long-term virological suppression mediated by AAV-delivered antibodies. 8th International Workshop on HIV Persistence during Therapy, 12-15 December 2017, Miami. Opening lecture.
- Johnson PR et al. Vector-mediated gene transfer engenders long-lived neutralizing activity and protection against SIV infection in monkeys. Nat Med. 2009 Aug; 15(8): 901–906. doi: 10.1038/nm.1967.
- Martinez-Navio J et al. Host anti-antibody responses following adeno-associated virus–mediated delivery of antibodies against HIV and SIV in rhesus monkeys. Molecular Therapy (January 2016), 24(1): 76-86. doi: 10.1038/mt.2015.191.
- Desrosiers RC. Long-term delivery of anti-HIV antibodies for treatment and prevention. (7 April 2017). (YouTube video)
- Tiegs G et al. Immune tolerance: What is unique about the liver. Journal of Autoimmunity 34(1):1-6. (February 2010). doi: 10.1016/j.jaut.2009.08.008.
- https://www.sciencedirect.com/science/article/pii/S089684110900105Xclinicaltrials.org. A phase 1, randomized, blinded, dose-escalation study of rAAV1-PG9DP recombinant AAV Vector coding for PG9 antibody in healthy male adults.