A step forward for AAV-mediated delivery of bNAbs

Richard Jefferys, TAG

Many years ago, Phil Johnson from the Children’s Hospital of Philadelphia began pursuing a novel workaround to overcome the challenge of inducing broadly neutralising antibodies (bNAbs) against HIV.

The idea borrows from gene therapy, using adeno-associated virus (AAV) as a vector to deliver the genetic code for a bNAb into the body. The aim is for the AAV to persist inside cells and act as a factory for churning out the bNAb into systemic circulation.

As previously covered on the blog, promising results were obtained in macaques over a decade ago, but the first human trial (conducted in HIV negative volunteers) did not achieve detectable levels of the bNAb PG9 – likely due to the induction of anti-PG9 antibodies. [1, 2]

At CROI 2020, Joseph P. Casazza from the Vaccine Research Center (VRC) at the National Institutes of Health presented evidence that the approach may yet have promise. [3]

Casazza described preliminary results from an ongoing trial of the bNAb VRC07 delivered via an AAV serotype 8 (AAV8) vector. [4] The vector was designed by David Baltimore and Alejandro Balazs, and is different to the AAV serotype 1 vector used in the previous human trial. The study is recruiting people with HIV on ART with undetectable viral loads, and Casazza presented data from eight participants (six men and two women, with five African American and three Caucasian).

Three escalating doses are being evaluated, and Casazza’s presentation included data from three recipients of the lowest dose, two recipients of the intermediate dose and three recipients of the highest dose. Follow up ranged from five months to a little over two years.

In contrast to the prior trial with AAV1, the bNAb VRC07 became persistently detectable in a majority of participants (six out of eight). In most cases, a pattern was observed in which an initial peak in levels occurred 2-4 weeks after injection, followed by a decline and then a rebound toward steady state levels after around 14-16 weeks.

Three participants developed anti-VRC07 antibodies, which has been the Achilles heel of the AAV-based delivery approach. In two of these individuals VRC07 levels became undetectable after the initial peak, and in the third the secondary peak was blunted. Casazza noted that anti-VRC07 antibodies did not appear to be the only factor influencing VRC07 levels, because one participant in the high dose group experienced a significant decline between weeks 22 and 40 which has yet to be explained.

Adverse events were minimal, with mild pain and tenderness at the injection sites reported in the high dose group and one case of muscle pain in the intermediate dose group—all resolved within seven days.

Casazza did not discuss the potential reasons for the greater success compared to the results achieved with AAV1-based delivery but, in addition to differences in the technical aspects of vector design, AAV8 is liver tropic and it’s been suggested that this is more likely to lead to immunological tolerance of the delivered bNAb. [5]

Notably, the maximum VRC07 concentration achieved was a little over one ug/mL in the high dose group, which is low compared to the levels that are obtained by intravenous infusion of bNAbs. For example, levels of well over 50 ug/mL have been reported in studies involving IV administration of the bNAb VRC01. [6] In a recent trial that administered a combination of the bNAbs 3BNC117 and 10-1074 and then interrupted antiretroviral therapy (ART), an undetectable viral load was maintained for as long as the concentration of both antibodies was above 10 ug/mL. [7]

Casazza’s results offer some hope that the use of AAV vectors to deliver bNAbs (or other therapeutic proteins) will turn out to be feasible. However, work remains to improve both the consistency and magnitude of bNAb production. In a preclinical macaque study in which an AAV8 vector was used to deliver VRC07, the addition of transient immune suppression with cyclosporine was able to increase average peak bNAb levels to ~40 ug/mL (compared to ~5 ug/mL without), however it’s unclear if this approach could be practically adapted for human use. [8]

Casazza and his colleagues at the VRC are not the only research group still pursuing the idea. Michael Farzan from the Scripps Institute gave a talk at CROI describing the latest results obtained with AAV-based delivery of eCD4-Ig, a protein that inhibits HIV replication by binding the virus envelope at sites that attach to CD4 and CCR5 receptors on CD4 T cells. [9]

In a therapeutic experiment involving six macaques infected with the SIV/HIV hybrid virus SHIV-AD08, eCD4-Ig delivered by AAV (two injections, the first using an AAV8 vector and the second using an AAV1 vector) was able to maintain viral load control to varying degrees after an ART interruption. At the most recent timepoint after 80-90 weeks off ART, five animals have viral loads ≤15 copies/mL and the sixth has a viral load of 25 copies/mL. Concentrations of eCD4-Ig ranged from 4.7 to 10.3 ug/mL, and Farzan explained that they hope to further refine delivery to achieve higher levels.

The laboratory of Ron Desrosiers is also working on AAV delivery, focusing on payloads of combination bNAbs. Farzan cited the best known example of this work: the “Miami monkey,” a recipient of AAV-delivered 3BNC117 and 10-1074 that has exhibited prolonged containment of a SHIV AD8 challenge at such vanishingly low levels that the virus has been extremely challenging to even detect (at one point it was thought virus eradication may have occurred). [5]

On 17 March 2020, Desrosiers and colleagues provided an update on another individual macaque that was originally part of a prevention study published in 2015. [10] This animal has now maintained high levels (240-350 ug/mL) of the anti-SIV antibody 5L7 for over six years. The researchers are now working to create this type of response more reliably and they conclude: “If satisfactory delivery methods are found, it becomes possible to envision long-term control of viral replication in the absence of antiretroviral treatment by delivering a combination of antibodies in people, and long-lasting protection when this approach is used in a prophylactic setting.”

It may still be a big “if,” but the data clearly justify efforts to solve the technological challenge.


  1. Jefferys R. 2G12: The sweetest broadly neutralizing antibody. (18 May 2009).
  2. Priddy FH et al. Adeno-associated virus vectored immunoprophylaxis to prevent HIV in healthy adults: a phase 1. Lancet HIV, 6(4)pe230-e329. (15 March 2019). DOI: 10.1016/S2352-3018(19)30003-7.
  3. Casazza JP. Durable HIV-1 antibody production in humans after AAV8-mediated gene transfer. CROI 2020, Oral abstract 41. (abstract and webcast)
  4. VRC 603: A phase 1 dose-escalation study of the safety of AAV8-VRC07 (VRC-HIVAAV070-00-GT) recombinant AAV vector expressing VRC07 HIV-1 neutralizing antibody in antiretroviral -treated, HIV-1 infected adults with controlled viremia.
  5. Jefferys R. Update on AAV vectors as delivery vehicles for broadly neutralizing antibodies: Ron Desrosiers and the Miami macaque (14 December 2017).
  6. Bar KJ. Effect of HIV Antibody VRC01 on Viral Rebound after Treatment Interruption. N Engl J Med 2016; 375:2037-2050. DOI: 10.1056/NEJMoa1608243. (24 November 2016).
  7. Mendoza P et al. Combination therapy with anti-HIV-1 antibodies maintains viral suppression. Nature. 2018 Sep; 561(7724): 479–484. doi: 10.1038/s41586-018-0531-2.
  8. Saunders KO et al. Broadly neutralizing HIV Type 1 antibody gene transfer protects nonhuman primates from mucosal simian-HIV. J Virology. DOI: 10.1128/JVI.00908-15. (3 june 2015).
  9. Farzan M. Toward durable control of HIV-1 with eCD4-Ig. CROI 2020. Oral abstract 51. (abstract and webcast)
  10. Martinez-Navio JM et al. Long-Term Delivery of an Anti-SIV Monoclonal Antibody With AAV. Front. Immunol. doi: 10.3389/fimmu.2020.00449. (17 March 2020).

Links to other websites are current at date of posting but not maintained.