Adeno-associated virus-vectored delivery of HIV biologics: the promise of a “single-shot” functional cure for HIV infection

Abstract

The ability of immunoglobulin-based HIV biologics (Ig-HIV), including broadly neutralizing antibodies, to suppress viral replication in pre-clinical and clinical studies illustrates how these molecules can serve as alternatives or adjuncts to antiretroviral therapy for treating HIV infection. However, the current paradigm for delivering Ig-HIVs requires repeated passive infusions, which faces both logistical and economic challenges to broad-scale implementation. One promising way to overcome these obstacles and achieve sustained expression of Ig-HIVs in vivo involves the transfer of Ig-HIV genes to host cells utilizing adeno-associated virus (AAV) vectors. Because AAV vectors are non-pathogenic and their genomes persist in the cell nucleus as episomes, transgene expression can last for as long as the AAV-transduced cell lives. Given the long lifespan of myocytes, skeletal muscle is a preferred tissue for AAV-based immunotherapies aimed at achieving persistent delivery of Ig-HIVs. Consistent with this idea, recent studies suggest that lifelong immunity against HIV can be achieved from a one-time intramuscular dose of AAV/Ig-HIV vectors. However, realizing the promise of this approach faces significant hurdles, including the potential of AAV-delivered Ig-HIVs to induce anti-drug antibodies and the high AAV seroprevalence in the human population. Here we describe how these host immune responses can hinder AAV/Ig-HIV therapies and review current strategies for overcoming these barriers. Given the potential of AAV/Ig-HIV therapy to maintain ART-free virologic suppression and prevent HIV reinfection in people living with HIV, optimizing this strategy should become a greater priority in HIV/AIDS research.

1. Introduction

The need for an HIV cure is as urgent as ever. Despite substantial global investment, nearly one third of the 38 million people currently living with HIV remain untreated.¹ While antiretroviral therapy (ART) can effectively suppress viral replication, a cure for all has remained elusive. Giving hope to the possibility of a cure for HIV infection, virologic remission has been observed in the “Berlin patient,” the “London patient,” and recently the “New York patient”.2, 3, 4, 5 The successful eradication of HIV in the first two cases involved transplantation of blood stem cells and bone marrow progenitors, respectively, both with a 32-bp deletion in the ccr 5 gene (i.e., CCR5-Δ32) that abrogates CCR5 expression. The New York patient was also transplanted with CCR5-Δ32-bearing cells, but they were of umbilical stem cell origin. Although these cases provide examples of a potential sterilizing cure, stem cell transplantation poses significant health risks and is not a scalable intervention for people living with HIV (PLWH). It is, therefore, important that cure interventions are developed with ease and accessibility in mind so they can be deployable to nations with suboptimal healthcare infrastructures. Although ART is now relatively simple and affordable, the burden of supplying it lifelong to all who need it, especially in low-income countries, is formidable in terms of cost and accessibility. To make matters worse, the COVID-19 pandemic has exposed critical flaws in the HIV continuum of care in many countries.⁶ For example, lockdowns and disruptions in public transportation have created insurmountable barriers to many PLWH who must travel to obtain their medications.^7,8 Additionally, an underappreciated fact is that over 80% of the global supply of ART drugs is produced by only eight Indian companies.⁹ Relying on so few manufacturers for nearly all antiretrovirals used worldwide is problematic because shortfalls in supplies and workforce can threaten the lifeline of millions of PLWH. Thus, the push for moving beyond daily ART as the paradigm for HIV treatment is greater than ever.

Although antiretrovirals can suppress viral replication and prolong the survival of PLWH, ,they are not wholly benign. As these drugs enter most cells and tissues in the body, side-effects will be inevitable after decades of use. In contrast, a few immunoglobulin-based HIV biologics (Ig-HIVs), continuously produced by host cells and functioning outside cells, may be less harmful over a lifetime. A permanent “drug-free holiday” would also help reduce the stigma associated with daily or even monthly ART, which would be highly attractive to many PLWH, especially those in demographics with high incidence of drug non-compliance. Given these limitations of ART, there is great interest in developing Ig-HIVs as alternatives or adjuncts to ART.

An impressive array of monoclonal HIV-specific broadly (b) neutralizing (n) antibodies (Abs) has been isolated in recent years.¹⁰ BnAbs can target various regions of the HIV envelope glycoprotein trimer (Env), including the base of the V3 stem, the apex of V2, the CD4-binding site, the gp120/gp41 interface, the membrane proximal external region (MPER) of the gp41 subunit, and the HIV fusion peptide.^11,12 Following passive infusion, anti-HIV bnAbs can suppress viral replication to varying degrees not only in simian-HIV (SHIV)-infected rhesus macaques (RMs),13, 14, 15 but in PLWH as well.15, 16, 17, 18, 19, 20, 21, 22 Importantly, in contrast to ART, bnAbs can also eliminate infected cells through Fcγ receptor-mediated effector functions, such as Ab-dependent cellular cytotoxicity (ADCC).23, 24, 25 While one bnAb is unlikely to successfully neutralize the vast diversity of circulating HIV isolates, a cocktail of bnAbs may provide the necessary coverage to suppress viral replication and block viral escape.^26,27 However, it is unclear how many bnAbs or Ig-HIVs should be included in such cocktails. Recent clinical trials used passive infusions of 3BNC117 (CD4-binding site) and 10–1074 (V3 glycan) in viremic PLWH whose ART was interrupted prior to bnAb infusions¹⁹ or in the presence or absence of ART.²¹ Gaebler et al. reported that without prior screening for bnAb sensitivity, 13 out of 17 patients maintained virologic suppression off ART until serum bnAb concentrations waned. AAV vectored delivery with optimized transgene expression could potentially solve the issue of maintaining concentrations of circulating bnAbs and therefore prevent viral rebound and the development of resistant escape variants. Julg et al. recently treated viremic PLWH who were not taking ART with three bnAbs: PGDM1400 (V2 glycan), PGT121 (V3 glycan), and VRC07-523-LS (CD4-binding site).²⁰ Although all patients experienced significant reductions in viremia, their viral loads rebounded rapidly, even before the infused bnAbs were cleared from circulation. The recrudescent virus in each person demonstrated partial to complete resistance to PGDM1400 and PGT121 in vitro, whereas susceptibility to VRC07-523-LS was largely preserved. Curiously, virus rebound occurred even though the mean serum concentration of VRC07-523-LS in the cohort was 93 μg/ml. Together, these results underscore the challenges faced by Ig-HIVs to deal with the diversity of circulating HIV isolates and call for more work to identify optimal combinations of bnAbs capable of achieving and maintaining HIV suppression.

In an attempt to determine the best bnAb combinations, Wagh et al. assessed the neutralizing activity of 15 bnAbs targeting four distinct epitopes of Env, including the V3-glycan region, the V1/V2-glycan region, the CD4-binding site, and the gp41membrane proximal external region (MPER), against a panel of 200 acute/early clade C HIV-1 Env pseudoviruses.²⁷ They found that a combination of three or four bnAbs proved the most efficacious, with CAP256-VRC26.25, VRC07-523, and 10–1074V as the preferred combination. It would cover various regions of Env with predicted resistant populations of each bnAb displaying little overlap. Ultimately, a combination of Ig-HIVs covering a broad range of Env binding and providing extensive breadth at a sufficient serum concentration would be ideal for prevention and treatment of HIV infection.²⁸

Since each bnAb binds to a single epitope on the Env trimer and displays limited breadth, several strategies have been devised to combine multiple bnAbs in the same molecule in order to increase neutralization coverage. For example, Asokan et al. have constructed a bispecific bnAb consisting of a common Fc region attached to the Fab fragments of the VRC07 (CD4-binding site) and PG9-16 (V1V2 apex) bnAbs.²⁹ Trispecific bnAbs displaying greater breadth and potency than single bnAbs have also been developed.^30,31 A clinical trial of the trispecific bnAb SAR441236, currently in the recruiting phase, will test the safety of this modality and its ability to suppress viremia in PLWH (ClinicalTrials.gov, Identifier: NCT03705169). SAR441236 combines the specificities of the VRC01-LS, PGDM1400, and 10E8v4 bnAbs into one molecule. Finally, utilizing multiple Fab fragments on a human apoferritin nanocage, Rujas et al. have developed the multiaffinity antibody, or “multabody,” platform.³² The combination of N49P7, 10E8v4, and PGDM1400 fragments in multabody T-01 MB.v2 displayed 100% neutralization coverage over a panel of 118 HIV-1 pseudovirus isolates.

Another promising strategy for augmenting the neutralization breadth of Ig-HIVs is the inclusion of the receptor and coreceptor sequences of HIV in the same immunoadhesin. One such Ig-HIV is the chimeric molecule eCD4-Ig, which comprises the ectodomain of CD4, an IgG Fc portion, and a coreceptor-mimetic peptide at its carboxyl terminus.³³ Because eCD4-Ig binds only to functionally constrained regions of Env, it outperforms most single bnAbs in its neutralization coverage.³³ Additionally, HIV variants that are completely resistant to eCD4-Ig have not yet been described.33, 34, 35

A nanobody consists of the variable heavy chain domain of an Ab.³⁶ Because of their small size, nanobodies are less prone to steric hindrance or site inaccessibility issues that can limit the utility of whole bnAbs. Although nanobodies suffer from short serum half-lives and rapid clearance, these caveats can be overcome through PEGylation and the attachment of an IgG Fc domain.³⁶ Schriek et al. recently developed multivalent versions of anti-HIV nanobodies that displayed increased binding to the HIV-1 Env protein and improved neutralization potency compared to monovalent variants.³⁷ Fusion of an IgG1 Fc domain to these multivalent nanobodies also allowed them to direct Fcγ receptor-mediated effector functions.³⁷

Despite the promise of the aforementioned Ig-HIVs to combat HIV/AIDS, the current paradigm for delivering them relies on repeated dosing, which constitutes a major caveat for mass production and deployment to resource-poor regions with suboptimal healthcare infrastructures. However, as described below, this caveat can be addressed through adeno-associated virus (AAV)-mediated gene therapy.

1.1. The promise of AAV-vectored delivery of Ig-HIVs for controlling HIV replication

AAV was originally discovered in the 1960s as a cell culture contaminant in laboratory adenovirus preparations.³⁸ AAV virions have a T = 1 icosahedral capsid consisting of 60 copies of the VP1, VP2, and VP3 proteins at a 1:1:10 ratio.³⁹ AAV is a member of the Parvoviridae family and the genus Dependoparvovirus. The AAV life cycle depends on the presence of helper viruses, such as adenoviruses and herpesviruses to replicate the AAV genome, which contains two genes: rep and cap. In recombinant AAV vector genomes, the rep and cap genes are removed and replaced by the transgene expression cassette, which cannot exceed ∼4.7 kb in single stranded (ss)AAV vectors. This insert capacity is further reduced to ∼2.3 kb in self-complementary (sc)AAV vectors, which include both the sense and anti-sense DNA strands in a single DNA molecule. Although scAAV vectors avoid the rate-limiting step of second strand synthesis and, consequently, display improved transduction efficiency compared to ssAAV constructs,⁴⁰ their double-stranded DNA genome is more prone to triggering the activation of cytosolic DNA sensors, like toll-like receptor (TLR)-9.⁴¹ In keeping with the higher pro-inflammatory potential of scAAV versus ssAAV vectors, scAAV-mediated gene transfer has been associated with the induction of more potent host immune responses against the transgene product.⁴²

AAV vectors are well suited for gene therapy because of their non-pathogenicity and ability to efficiently transduce a wide range of cells.³⁹ The genome of these constructs persists in the nucleus of transduced cells predominantly as episomes and the only protein expressed from AAV vectors is the transgene product; as long as it is non-toxic and does not trigger host immunity, transgene expression can persist for the entire lifespan of transduced cells.³⁹ Because of the long lifespan of muscle cells,⁴³ skeletal muscle is a preferred tissue for AAV/Ig-HIV therapies.^44,45 Hence, one can envision a scenario where a one-time intramuscular (IM) dose of AAV/Ig-HIV vectors would confer passive HIV immunity for years, possibly decades, thereby allowing PLWH to stop ART without experiencing virus rebound. Realizing this vision would not only simplify the treatment of HIV infection, but would also avoid the stigma associated with daily ART..

In spite of the ample evidence supporting the use of AAV-expressed Ig-HIVs for HIV prophylaxis,^33,35,46, 47, 48, 49, 50 only two pre-clinical studies have evaluated the ability of this approach to achieve ART-free control of immunodeficiency virus replication. In one study, Horwitz et al. assessed whether gene therapy with AAV8/bnAb vectors could maintain virologic control in humanized (hu)-mice with pharmacologically controlled HIV–1_YU2 infection following ART interruption.⁵¹ Unexpectedly, the ART regimen used in that study (tenofovir disproxil-fumarate, raltegravir, and emtricitabine) appeared to interfere with the efficiency of AAV8/bnAb transduction in vivo. To overcome this limitation, the authors stopped the ART regimen and waited 2 weeks before the hu-mice were treated with an AAV8/10–1074 vector. During this 2-week period, virologic control was maintained by treating the hu-mice with passive infusions of recombinant 10–1074, which ended at the time of the AAV8/10–1074 dosing. The hu-mice were then followed for 67 days, during which time high serum concentrations of AAV8-expressed 10–1074 were detected in all animals. Of the seven AAV8/10-1074-treated hu-mice, only one lost control of virus replication due to the emergence of 10-1074-resistant viral variants. It is noteworthy that hu-mice do not make clinically relevant levels of anti-drug antibodies (ADAs) following AAV/Ig-HIV treatment, so it is unclear if the positive results obtained by Horwitz and colleagues could be reproduced in humans.

In another study, Martinez-Navio et al. have assessed whether AAV-expressed anti-HIV bnAbs could control an established SHIV-AD8 infection in rhesus monkeys (RMs) without prior ART.⁵² In their first study group, four RMs received IM injections of AAV vectors expressing “rhesusized” versions of 10–1074 (V3), 3BNC117 (CD4-binding site), and 10E8 (MPER). Consistent with the high immunogenicity of anti-HIV bnAbs in the context of AAV-mediated gene therapy, most of the AAV-treated animals developed ADAs, resulting in low or undetectable levels of bnAbs in serum and failure to contain the infection. Interestingly, one animal (rh2438; also known as the “Miami Monkey”) was an exception in that it mounted little or no ADAs against 3BNC117 and 10–1074 and, consequently, developed high and persistent levels of these bnAbs.⁵³ The rise in serum concentrations of 3BNC117 and 10–1074 in rh2438 coincided with a sharp decline in viremia, which went from 10,000 vRNA copies/ml at the time of AAV/bnAb dosing (week 86 post infection) to <15 vRNA copies/ml five weeks later. Remarkably, monkey rh2438, which is still alive at the time of this writing, was treated once with AAV/bnAb vectors in August 2015 and has remained aviremic ever since (R. Desrosiers, personal communication). Importantly, two additional animals in that study have also controlled plasma viral loads (PVLs) to below detection limits after receiving two doses of AAV/bnAb vectors. Although those RMs experienced transient episodes of viremia during the first few months after the AAV/bnAb inoculations, they too remain alive and aviremic at the time of writing (R. Desrosiers, personal communication).

1.2. Barriers to realizing the full potential of AAV-mediated gene therapy with Ig-HIVs

Pre-existing immunity to AAV vectors. Natural infection with wild-type AAV is common among humans and RMs. Despite its apparent benign biology, wild-type AAV infection triggers innate immune pathways that culminate in the generation of long-lived adaptive immune responses against the AAV capsid.⁵⁴ Due to amino acid sequence conservation across different serotypes, AAV-specific immune responses tend to be highly cross-reactive, leading to immune recognition of most AAV capsids used for gene transfer. Cellular immune responses induced by natural AAV infection are less frequently observed than humoral responses, at least in peripheral blood.^55,56 However, AAV-specific T cells can undergo significant expansion following parenteral delivery of AAV vectors and thwart the efficacy of liver-targeted AAV-based gene therapies. Indeed, in an early clinical trial of AAV-mediated gene therapy for hemophilia B, patients were treated with an AAV2 vector encoding Factor IX (F.IX) through hepatic artery infusion and developed a stable expression of F.IX until week 4.⁵⁷ However, serum concentrations of F.IX began to decay in subsequent weeks, ultimately dropping below detection limits by week 12. These patients also experienced a spike in liver transaminases, which peaked at week 6 and then gradually returned to normal levels in the next 2 months. Tellingly, a time course analysis of AAV-specific T cell responses in one of the AAV2/F.IX-treated subjects revealed the expansion of AAV-specific T cells, which responded to peptides corresponding to the AAV2 capsid by producing IFN-γ.⁵⁷ This observation, combined with reports of human AAV-specific T cells expanding following AAV-mediated gene therapy,⁵⁵ implicates CD8⁺ T cell-mediated killing of AAV vector-transduced hepatocytes as the cause of the transient transaminitis and the decline in F.IX expression in the aforementioned hemophilia B clinical trial.

Although AAV-specific T cells can limit transgene expression in hepatocytes, the impact of cellular immune responses on muscle gene transfer is less clear, as IM AAV-treated patients can maintain persistent transgene expression despite the presence of anti-AAV capsid T cells at the site of the AAV inoculation.58, 59, 60 In those cases, analysis of muscle biopsies from the AAV injection site revealed T cells displaying regulatory (Foxp3+) and exhausted (PD-1+) phenotypes,58, 59, 60 which might prevent the elimination of AAV-transduced myocytes. Curiously, despite being able to present epitopes through major histocompatibility class-I molecules, recent evidence suggests that myocytes are relatively resistant to cytotoxic T lymphocyte-mediated killing and can serve as an antigen reservoir for viral vector-driven transgene expression.⁶¹ More studies are needed to confirm this possibility. Thus, while pre-existing T cell responses against the AAV capsid can limit liver transduction, cellular immune responses seem unable to affect AAV vector persistence in muscle cells in a meaningful way.

Anti-AAV nAbs pose a key barrier to AAV vector transduction following parenteral injection.62, 63, 64, 65, 66, 67, 68, 69 Accordingly, AAV seropositive (+) RMs that undergo plasmapheresis to remove their anti-AAV nAbs prior to AAV vector dosing develop equivalent levels of transgene expression in muscle cells as non-pheresed AAV seronegative (−) animals that received the same AAV vector in parallel.⁷⁰ However, plasmapheresis is not supposed to affect memory T cells. Consistent with IgG being the predominant class of anti-AAV nAbs, enzymatic cleavage of serum IgG prior to AAV vector dosing leads to a significant drop in anti-AAV nAb titers and improved transgene expression in AAV (+) mice and monkeys.^71,72 Although even low titers of anti-AAV nAbs can completely abrogate liver transduction following intravenous (IV) AAV administration,⁶⁶ this nAb susceptibility diminishes when AAV vectors are given intramuscularly. Indeed, Greig et al. have shown that pre-existing midpoint anti-AAV nAb titers of 160 or less had no impact on AAV-driven mAb expression in RMs following IM delivery of an AAV/mAb vector.⁷³ However, medium or high titers of anti-AAV nAbs at the time of IM AAV vector dosing can limit or even abrogate transgene expression.^48,73,74 Thus, PLWH who are AAV (+) might still benefit from AAV/Ig-HIV therapy as long as their anti-AAV nAbs do not inactivate the AAV/Ig-HIV vector or appropriate interventions are used to lower anti-AAV nAb titers sufficiently to allow AAV transduction of myocytes.

Methods to avoid pre-existing immunity. In order for AAV-mediated transfer of Ig-HIV genes to be a viable treatment option for PLWH regardless of their AAV serostatus, immune recognition of the AAV vector must be addressed. One approach to evade recognition of the AAV vector capsid and improve transduction of specific target cells is the development of synthetic capsids that are not recognized by natural anti-AAV humoral responses. Such artificial capsids have been engineered either by disruption of immunogenic epitopes, by site-directed mutagenesis,⁷⁵ capsid chimeras such as AAV2.5,⁷⁶ or through directed evolution,⁷⁷ as done in the development of AAV-KP1.⁷⁸ Evasion of pre-existing anti-AAV nAbs remains difficult due to the diversity of neutralizing epitopes and the fact that changes in immunodominant regions can hinder vector transduction.⁷⁹

Alternatively, it is also possible to transiently lower host anti-AAV nAb titers prior to AAV vector dosing. One strategy to accomplish this is enzymatic cleavage of host anti-AAV nAbs prior to AAV/bnAb vector treatment using bacterial IgG endopeptidases.⁷¹ The Streptococcus pyogenes-derived endopeptidase IdeS, for example, cleaves all IgG subclasses at their hinge region, yielding two F (ab’)2 and Fc fragments (Fig. 1).⁸⁰ Without the IgG Fc region, F (ab’)2 fragments display a markedly reduced serum half-life of only 12–20 hours.⁸¹ Because of its efficient IgG-depleting activity, IdeS is being considered for treating conditions involving pathogenic IgG Abs.82, 83, 84, 85 In work by Leborgne et al., AAV (+) cynomolgus macaques were treated with IdeS prior to IV administration of AAV8.⁷² While IdeS successfully depleted pre-existing nAbs, the macaques in that study had very low pre-existing titers of anti-AAV8 nAbs, so it is unclear whether pre-dosing with IdeS alone would be able to lower clinically relevant anti-AAV nAbs sufficiently to enable AAV vector transduction. Should IdeS be considered as an adjunct for AAV/Ig-HIV-based cure strategies in PLWH on ART, attention must be paid to the timing of analytic treatment interruption (ATI) since residual levels of IdeS could, in theory, cleave the AAV-encoded Ig-HIV in vivo. However, this effect would only be relevant post ATI. Given the short half-life of IdeS in serum (i.e., ∼27 hours),⁷² IdeS-mediated cleavage of Ig-HIVs during their therapeutic window could be avoided by delaying ATI by a few weeks after AAV inoculation, thereby providing enough time for the endopeptidase to be cleared from serum and for AAV-driven expression Ig-HIVs to peak and stabilize.

Fig. 1.

The IgG endopeptidase IdeS cleaves IgG molecules at the hinge region, accelerating their decay in circulation. IdeS cleaves IgG molecules at the hinge region, yielding two fragments: one F (ab)₂ and one Fc. F (ab)₂ fragments have markedly reduced half-lives in serum than intact IgG molecules.

Host anti-AAV nAb titers may be transiently reduced for a brief period in vivo through blockade of the neonatal Fc receptor (FcRn), which regulates IgG transport and catabolism.⁸⁶ At steady state, serum IgG Abs are continuously internalized by myeloid and endothelial cells where the IgG molecules bind to FcRn in the acidic environment of early endosomes. FcRn-bound IgG is then sorted into recycling endosomes that shuttle IgG back to the cell surface. The neutral pH of the extracellular milieu dissipates the FcRn-IgG interaction, resulting in the release of IgG from the cells (Fig. 2A). This recycling process accounts for the long half-life (>20 days) of circulating IgG. In contrast, IgG molecules that do not bind to FcRn end up degraded in lysosomes (Fig. 2B). Consistent with the role of FcRn in rescuing IgG from catabolic degradation, inhibiting this receptor can markedly shorten the half-life of serum IgG molecules. Indeed, FcRn blockade is being investigated as a potential treatment for IgG-mediated autoimmunity and graft rejection.87, 88, 89, 90 A single dose of the FcRn-blocking mAb Rozanolixizumab, for example, can safely and transiently lower total IgG levels in mice, nonhuman primates (NHPs), and humans.^86,90, 91, 92, 93 In NHPs, IgG concentrations dropped by 49%–90% from baseline, depending on the dosing schedule.^88,90 Maximal effects were achieved at 4–25 days post treatment and IgG levels generally returned to baseline by week 8.^88,90 Therefore, pre-treatment with a FcRn-blocking mAb prior to AAV/Ig-HIV administration may improve vector transduction and allow for higher levels of therapeutic transgene expression (Fig. 2C and D).

Fig. 2.

Monoclonal antibody-based neonatal Fc receptor blockade as a strategy for evading pre-existing anti-AAVneutralizing antibodies (nAbs)prior to intramuscular AAV/Ig-HIV vector dosing. A) Serum IgG antibodies (Abs) are continuously internalized by endothelial cells and bind to the neonatal Fc receptor (FcRn) within the acidic environment of the early endosome. FcRn-bound IgG molecules are then sorted into recycling endosomes that traffic the IgG back to the cell surface, where the neutral pH of the extracellular environment dissociates the FcRn/IgG interaction. IgG molecules that are not bound to FcRn end up degraded in lysosomes. B) Anti-FcRn monoclonal (m)Ab therapy accelerates the catabolic degradation of serum IgG molecules by blocking their interaction with FcRn. C-D) AAV/Ig-HIV vector dosing of AAV (+) individuals harboring medium to high titers of anti-AAV nAbs results in little or no expression of Ig-HIVs due to Ab-mediated neutralization of AAV/Ig-HIV vectors (C). However, transiently reducing anti-AAV nAb titers through anti-FcRn mAb therapy prior to the AAV/Ig-HIV inoculation might improve vector transduction, leading to higher levels of Ig-HIV expression (D).

Evasion of pre-existing immunity to AAV may also be achieved through the use of DNA vectors. Since DNA does not induce adaptive immune responses, it could, in theory, be re-administered until therapeutic levels of Ig-HIVs are achieved. Along this line, Weiner and colleagues recently reported that NHPs developed detectable levels of Ig-HIVs in serum following IM delivery of DNA plasmids via electroporation (EP).^94,95 In one study, NHPs were treated with DNA plasmids encoding the bnAbs PGDM1400 and PGT121 and developed peak bnAb serum concentrations of >30 μg/ml on day 14 post dosing. However, bnAb expression levels decreased rapidly after day 20, concomitantly with the emergence of ADA responses.⁹⁴ The fact that the authors delivered unmodified human versions of PGDM1400 and PGT121 likely contributed to their immunogenicity in NHPs. However, even “rhesusized” versions of those bnAbs might have induced ADAs in that study considering the pro-inflammatory responses triggered by IM EP. In a separate study, Xu et al. used IM EP to deliver a DNA plasmid encoding the anti-HIV immunoadhesin eCD4-Ig to immunodeficient mice, which resulted in peak serum concentrations of eCD4-Ig of 100 μg/ml.⁹⁵ Another caveat to delivering Ig-HIVs via IM EP of DNA plasmids is promoter silencing, which can be caused by methylation of the CMV promoter.⁹⁶ The role of CpG motifs on plasmid DNA-driven transgene expression remains uncertain. One study reported that removal of CpG motifs from plasmid DNA constructs improved the degree and duration of luciferase expression in mice following IV administration.⁹⁷ However, a separate study showed that removal of CpGs from the bacterial plasmid backbone did not have a significant effect on transcriptional silencing.⁹⁸ Future studies should attempt to reconcile the impact of promoter silencing on the ability of DNA vectors to promote long-term transgene expression.

Interestingly, injection of linearized plasmid DNA into mouse tail veins has been shown to result in persistent transgene expression⁹⁹ at levels that were nearly 100-fold higher than those achieved in control mice injected with circular DNA constructs.¹⁰⁰ An analysis of the structure of these DNA constructs in the liver from mice in both groups revealed the presence of large unintegrated concatemers in tissue from animals that had received the linearized, but not the circular, DNA plasmids, implicating concatemerization as a potential mechanism for the sustained transgene expression observed in the linearized DNA group. Although the clinical utility of systemic delivery of naked DNA is uncertain, the aforementioned studies illustrate how the molecular structure of DNA vectors can affect transgene expression. Non-plasmid vectors for transgene delivery have also been developed with the use of minicircles.¹⁰¹ Since these lack the origin of replication and antibiotic resistance sequences present in most plasmids, they avoid the issue of innate recognition of bacterial sequences. A major advantage to the use of minicircles is their enhanced long-term transgene expression in comparison to conventional plasmids, which could make them an attractive gene delivery option for the long-term expression of Ig-HIV.101, 102, 103 However, issues with minicircle practicality as a treatment option remain, one of these being scalability. The production and purification processes available for minicircles are lengthy and would likely prove too costly to be deployed as a treatment for the millions of PLWH.¹⁰⁴ Additionally, similar to the issue of empty AAV capsids contaminating the production of AAV/bnAb vectors,¹⁰⁵ minicircle preparations can also contain parental plasmids used in their production.¹⁰⁴ While the above methods could aid in the evasion of pre-existing immunity to AAV, the immunogenicity of the bnAb itself would still remain an issue.

Vermeire et al. recently compared the efficiency of mAb expression by conventional versus minimal plasmid delivery strategies.¹⁰⁶ In vitro experiments, including transfections of Expi293F and C2C12 cells, and in vivo studies utilizing IM EP in mice, have assessed the impact of plasmid backbone size and expression cassette design on the expression of the mAb 4D5. Although in vitro data suggested that decreases in plasmid backbone size would improve 4D5 expression, in vivo results have shown similar levels of expression between conventional plasmids and nanoplasmids. In terms of cassette design, the use of two plasmids, each encoding the heavy or light chain, produced higher levels of 4D5 than a single plasmid encoding both mAb chains, especially in the case of minimal plasmid backbone size. Notably, the authors found that in mice subjected to IM EP delivery of DNA vectors, the use of a muscle-specific promoter outperformed the CAG promoter in driving 4D5 expression. This is an encouraging observation since restricting transgene expression to skeletal muscle cells would not only increase the safety profile of this approach, but it might also reduce the immunogenicity of the transgene product and avoid ADAs.

Immunogenicity of the AAV-encoded Ig-HIV. One of the largest hurdles to AAV-mediated gene therapy with Ig-HIVs is that the AAV-delivered molecule often becomes an immune target, leading to the development of host ADAs that can limit or even abrogate Ig-HIV expression. As expected, the immunogenicity of Ig-based biologics increases proportionately with the degree of species mismatch between the Ig molecule and the host. However, ADAs have been detected even in cases of complete species match between the AAV-delivered Ig-HIV and the host.^47,107,108 In the International AIDS Vaccine Initiative (IAVI)-sponsored clinical trial of an AAV1 vector expressing the bnAb PG9, 10 out of 16 AAV1/PG9-treated participants developed ADAs, even though PG9 could not be detected in any of the individuals.¹⁰⁷ PG9 was expressed using a CMV promoter for the heavy chain and an EF1α promoter for the light chain. Levels of PG9 were undetectable by ELISA at each of the doses ranging from 4.0 × 10¹² vector genomes (vg)/kg to 1.2 × 10¹⁴ vg/kg. AAV1 was chosen as the capsid for this trial due to its efficiency in comparison to other serotypes at transducing skeletal muscle.¹⁰⁹ More recently, in the VRC 603 clinical trial of an AAV8 vector expressing the bnAb VRC07 in PLWH with ART-suppressed viremia, 3 out of 8 study participants developed ADAs.¹⁰⁸ The treatment doses ranged from 5.0 × 10¹⁰ vg/kg to 2.5 × 10¹² vg/kg and used a synthetic “CASI” promoter system to drive transgene expression. VRC07 was expressed as a single ORF containing a self-cleaving F2A peptide between the heavy and light chain coding regions. AAV8 was used in this trial due to its low seroprevalence in humans in comparison to other serotypes.¹¹⁰ Serum expression of VRC07 peaked at 3.3 μg/ml, which is considerably lower than the levels reported in RMs treated with a similar AAV8/VRC07 construct.⁵⁰ It is unclear why VRC07 expression was lower in VRC 603 study participants than in NHPs. It is also worth noting that the VRC07 molecule used in the VRC 603 study lacked FcRn affinity-enhancing mutations. The “LS mutations” (M428L/N434S), for example, increase the affinity of IgG for FcRn and have been shown to extend the serum half-life of the anti-HIV bnAb VRC01 by ∼3-fold in RMs.¹¹¹ It is tempting to speculate that greater bnAb expression levels would have been achieved in the VRC 603 study, had the VRC07 molecule contained the LS mutations and an AAV capsid with greater muscle tropism than AAV8 been used, but it is difficult to predict the magnitude of this improvement. Additional improvements in both trials could have been made in the choice of promoter for the expression cassette. For example, the surge in transgene expression enabled by the CMV promoter may have favored the induction of ADAs in the IAVI trial.¹¹² An alternative promoter, such as the chicken beta-actin (CBA) one, may bypass this issue by resulting in lower but sustained levels of Ig-HIV expression. In addition, the use of a dual promoter system in the IAVI trial, rather than the single promoter plus self-cleaving peptide combination, may have contributed to the low levels of PG9 expression in that study.¹¹³ Additional details of the IAVI and VRC 603 clinical trials are shown in Table 1.

Table 1.

Two clinical trials of AAV/bnAb vectors have been conducted to date.

	IAVI Trial	VRC 603 Trial
Number of Participants	21	8
HIV Status	Negative	Positive
Country of Conduct	United Kingdom	United States
AAV/bnAb	AAV1-PG9	AAV8-VRC07
BnAb Expression Cassette	CMV and EF1α promoters for the heavy and light chains respectively	CASI promoter expressing a single ORF and a self-cleaving F2A peptide between the heavy and light chain coding regions
Vector Doses	4.0 × 1012 vg 4.0 × 1013 vg 8.0 × 1013 vg 1.2 × 1014 vg	5.0 × 1010 vg/kg 5.0 × 1011 vg/kg 2.5 × 1012 vg/kg
bnAb Expression	No serum expression detected by ELISA	Detectable serum expression in all participants, with 3 participants having >1.0 μg/ml
Rate of ADAs	10/16 participants	3/8 participants

Considering that the AAV vectors evaluated in the IAVI and VRC 603 clinical trials (Table 1) encoded bnAbs that are authentic human IgG molecules, why did those molecules induce ADAs in some of the AAV/bnAb-treated participants? One possibility is that innate sensing of the AAV vectors creates a pro-inflammatory milieu during bnAb expression in vivo, thereby rendering a “self” molecule immunogenic. Indeed, Martino et al. have previously reported that plasmacytoid dendritic cells respond to AAV vectors by producing type I interferons within 2 hours of vector administration and that AAV sensing relied on the TLR-9-MyD88 pathway.⁴¹ The authors have also documented the rapid infiltration of innate immune cells (i.e., macrophages, neutrophils) to the site of the AAV administration in mice and the subsequent induction of adaptive immunity. Although TLR-9 inhibition decreased the immunogenicity of AAV vectors, the AAV capsid can also activate TLR-2,¹¹⁴ indicating that broader inhibition of innate immunity might be needed to decrease the immunogenicity of AAV vectors. Another factor contributing to the immunogenicity of HIV-specific bnAbs, in the context of AAV-mediated gene therapy, is their hypermutated nature and unusual structural features. Even though all therapeutic HIV bnAbs evolved in PLWH, the Env sequences and checkpoint systems that shaped the variable domains of those molecules are highly unique. From this perspective, it seems plausible that monoclonal bnAbs, especially their complementarity determining regions, would be perceived as foreign entities by bnAb-treated individuals who have never been exposed to those particular variable motifs. Consistent with this notion, Martinez-Navio et al. have previously reported a significant positive correlation between the magnitude of ADA responses and the degree of sequence divergence from germline of AAV-delivered bnAbs in RMs.¹¹⁵ As ADAs can inactivate the AAV-delivered bnAb and accelerate its clearance, there is great interest in the development of strategies capable of curbing their production.

Methods to limit ADA production. The mammalian target of rapamycin (mTOR) participates in IL-2 signal transduction, a critical step in the activation of both T and B cells.¹¹⁶ Hence, pre-treatment with rapamycin has been tested for its potential to inhibit ADAs following AAV vector treatment in mice treated with Factor VIII (F.VIII).¹¹⁷ Guided by this rationale, Selecta Biosciences developed ImmTOR®, a biodegradable PLA (poly (D,l-lactide) and PLA-PEG (poly (D,l-lactide)-block-poly (ethylene-glycol)) polymer nanoparticle containing rapamycin.¹¹⁸ Combining nanoparticle methodology with rapamycin, antigen presenting cells are targeted to become tolerogenic, which can then induce tolerance to a specific antigen through activation of antigen-specific regulatory T cells. In addition to preventing ADAs, recent studies conducted in mice have shown that treatment with rapamycin or ImmTOR® prior to AAV vector inoculation can also prevent the generation of anti-AAV nAbs, thereby opening the possibility of vector redosing.^119,120 A recent clinical trial tested the safety of ImmTOR® and its ability to prevent ADAs against pegadricase, a therapeutic uricase enzyme, in patients with hyperuricemia.¹²¹ In a dose-dependent manner, a single IV infusion of recombinant pegadricase and ImmTOR® prevented the production of ADAs and resulted in transient control of serum uric acid levels. Though encouraging, it remains to be determined whether combining ImmTOR® dosing with AAV-mediated gene therapy would prevent induction of ADAs and enable long-term transgene expression.

A transient course of immunosuppressants before and after AAV/Ig-HIV therapy has been shown to suppress ADA responses. This approach was employed by Saunders et al., who have reported that four RMs treated IM with an AAV8 vector encoding a simianized version of VRC07 developed peak plasma bnAb concentrations in the 2.5–7.7 μg/ml range between week 2 and 4 post treatment.⁵⁰ However, bnAb expression fell sharply in the ensuing weeks, concomitantly with the emergence of ADAs, which drove VRC07 levels to below detection limits by weeks 6–9 in all four RMs. In an attempt to suppress these anti-VRC07 antibody responses, a separate cohort of RMs was treated with cyclosporine (CsA) beginning 9 days prior to the AAV8/VRC07 administration and lasting until week 4 post vector injection. A clear enhancement in VRC07 expression was observed in those animals. Plasma bnAb concentrations peaked at an average of 38 μg/ml over the first 3 weeks, with one monkey making up to 66 μg/ml of VRC07 during this period. Of the six AAV8/VRC07-treated RMs that had received CsA, three developed ADAs that abrogated bnAb expression around the end of the CsA regimen. The remaining three RMs developed persistent VRC07 expression, with plasma concentrations eventually plateauing at 1–10 μg/ml and no apparent induction of ADAs. These results are encouraging since a short course of CsA increased the proportion of AAV8/VRC07-treated animals with persistent levels of bnAb in circulation from zero to 50%. Although CsA has been approved for human use for decades and can be taken orally, prescribing an immunosuppressant as an adjunct to AAV/Ig-HIV therapy for PLWH might be problematic given the immune deficit already incurred by HIV infection. Nevertheless, if transient immunosuppression provides a viable path to achieving virologic remission following AAV/Ig-HIV therapy, this strategy might seem appealing to many PLWH willing to try it as an alternative to lifelong ART. Thus, a risk-benefit analysis of combining AAV/Ig-HIV vectors with transient immunosuppression seems warranted.

Another approach to inhibit ADAs involves tolerizing the host to the AAV-delivered Ig-HIV. One way to achieve this is by directing transgene expression to the liver to leverage the tolerogenic properties of that organ.^54,122, 123, 124, 125, 126, 127, 128, 129, 130 Guided by this rationale, Fuchs et al. have used AAV expression cassettes encoding the rhesus monkey SIV-specific mAb 4L6 to evaluate different promoters and delivery schemes in these animals.⁷⁴ Overall, dosing with the liver-tropic AAV8 capsid first, followed by a second inoculation with AAV1 resulted in the highest serum concentrations of 4L6. However, the magnitude of ADA responses inversely correlated with the extent to which 4L6 expression was restricted to the liver following the AAV8 administration. For example, when 4L6 expression was under the control of the ubiquitous cytomegalovirus (CMV) promoter and both vectors were given IM, RMs developed low to modest levels of ADAs following the AAV8-CMV-4L6 inoculation. These ADAs decayed over time, but they still decreased, albeit partially, 4L6 production. Interestingly, when those animals were boosted with the AAV1-CMV-4L6 vector 50 weeks later, there was a marked increase in mAb expression, but no measurable rise in ADAs. Even when AAV vectors are administered IM, they can still leak into the circulation and reach other organs, including the liver.¹³¹ In a subsequent experiment, Fuchs et al. have attempted to restrict the initial AAV8-driven mAb expression to hepatocytes by placing the 4L6 gene under the control of the liver-specific thyroxine-binding globulin (TBG) promoter and delivering the AAV8-TBG-4L6 vector intravenously. This strategy resulted in very low levels 4L6 expression after the IV AAV8-TBG-4L6 inoculation, but none of the RMs mounted detectable ADAs.⁷² Notably, when those animals were boosted intramuscularly 14 weeks later with the AAV1-CMV-4L6 construct, a sharp rise in 4L6 levels ensued, with only a modest and transient increase in ADAs. Although the study by Fuchs et al. is limited by small group sizes, the results are consistent with the ability of AAV8 vectors to induce tolerance to transgene products, possibly because of its tropism for hepatocytes. If liver-directed AAV-driven expression of Ig-HIVs can indeed help limit ADAs, this strategy should be considered in designing future trials of AAV/Ig-HIV-based therapies.

Tolerance to the AAV-encoded Ig-HIV may also be achieved through the long recognized^132,133 but poorly understood phenomenon of high-zone tolerance (HZT).134, 135, 136 HZT refers to the ability of injections of large quantities of a soluble non-aggregated and non-adjuvanted antigen to induce a state of immune unresponsiveness specific to the delivered antigen. HZT has been leveraged for the treatment of hemophilia A, an X-linked recessive disease characterized by the lack of endogenous production of wild-type full length F.VIII.¹³⁷ As a result, patients with this condition require periodic administrations of F.VIII to prevent life-threatening bleeding episodes. Importantly, because these patients do not produce a functional form of F.VIII, the passively infused therapeutic protein can be perceived as “foreign” in 20%–30% of cases, leading to the induction of anti-F.VIII nAbs that inhibit clotting.¹³⁸ While pharmacological immunosuppression can be used to manage ADAs in this setting, repeated infusions of large doses of F.VIII have also been used to counteract these responses through HZT.^132,137,138 Early attempts to generate mAbs against subdominant cellular antigens also benefitted from subtractive immunization protocols based on HZT.¹³⁹ Those cases involved treating mice with high doses of purified soluble cell lysates containing undesirable immunodominant epitopes but lacking the B-cell determinants of interest. One week later, the same animals would be given cells containing both the immunodominant and the subdominant epitopes. Due to the HZT created by the first immunization, B cells targeting the immunodominant epitopes would fail to respond after the second immunization, whereas those specific for the subdominant ones would reach greater frequencies, thereby facilitating their isolation. HZT might also help to explain the inverse relationship between serum concentrations of anti-TNF-α mAbs and incidence of ADAs observed in some patients undergoing treatment for rheumatoid arthritis.^135,140,141 Furthermore, two recent studies suggest that it might be possible to leverage HZT to lower the immunogenicity of anti-HIV bnAbs.^142,143 For example, Shapiro et al. have monitored the pharmacokinetics of human versions of the bnAbs VRC07-523 and PGT121 in infant RMs following two dosing regimens.¹⁴² The first one consisted of four subcutaneous (SC) injections of 5 mg/kg of each bnAb given every 2–3 days, while the second regimen comprised a single SC dose of 20 mg/kg. Although the study was not originally designed to compare bnAb and ADA levels between the two groups, both the incidence and magnitude of ADA responses were lower in the latter group, raising the possibility that the 20-mg/kg infusion was more effective in inducing tolerance to the bnAbs than the low-dose regimen. Chang et al. have made similar observations about the immunogenicity of the anti-CCR5 humanized IgG4 mAb Leronlimab in RMs.¹⁴³ While weekly SC injections of 10 mg/kg of this mAb were immunogenic in a subset of animals, increasing the dose of Leronlimab to 50 mg/kg and delivering it every other week did not induce ADAs. Based on the aforementioned data, it might be possible to prevent or at least reduce the incidence of ADAs following gene therapy with AAV/bnAb vectors by pre-treating recipients with large doses of the same bnAbs being delivered by AAV-mediated gene transfer. Future studies should address this possibility, considering the excellent safety profile of passive bnAb infusions in clinical trials.

1.3. Priority research areas

AAV-mediated gene therapy with HIV biologics is safe and can result in broad, potent, and persistent anti-HIV immunity after a single dose. No other HIV immune intervention comes close to achieving this. Despite the promise of this approach and the 20 years that have elapsed since the first use of an AAV vector to promote continuous expression of an anti-HIV mAb in vivo,¹¹³ surprisingly few studies have evaluated AAV-mediated delivery of Ig-HIVs for suppressing HIV replication. Indeed, only two pre-clinical studies have been conducted and shown that AAV/bnAb therapy can achieve ART-free control of immunodeficiency virus infection,^51,52 and the VRC 603 trial remains the only clinical test of an AAV/bnAb vector in PLWH¹⁰⁸ (Table 1). This paucity of research on AAV/Ig-HIV-based cure strategies stands in sharp contrast to the dozens of active clinical trials of HIV cure modalities based on cord blood transplantation, gene-modified stem cells, and checkpoint inhibitors that are currently registered on www.clinicaltrials.gov. Many of those regimens are inherently dangerous and/or require complex treatment schedules. In contrast, persistent control of HIV infection might be attained after a one-time IM administration of AAV/Ig-HIV vectors, which for AAV (−) individuals, could be done after a single visit to an outpatient or a mobile clinic. Because many PLWH are likely AAV (+), these patients may require additional visits to lower their titers of pre-existing nAbs prior to AAV/Ig-HIV vector dosing, unless muscle-tropic AAV capsids capable of evading pre-existing anti-AAV nAbs are developed. More work is also needed for developing strategies to prevent the induction of ADAs. Finally, pre-clinical and clinical studies should focus on improving expression of the delivered Ig-HIV, considering the low serum concentrations of VRC07 reported in the VRC 603 trial.¹⁰⁸ As a combination of Ig-HIVs will likely be necessary to provide sufficient breadth to durably suppress viral replication, the composition of such Ig-HIV cocktail still needs to be determined. High-throughput neutralization screens and machine learning may prove useful to parse the wealth of data available from the numerous Ig-HIVs that are being considered for clinical testing.

Design considerations for the expression cassette include the promoter system as well as modifications in the DNA backbone to avoid innate immune sensing. While the heavy and light chains of mAbs can be expressed from separate plasmids or from bicistronic cassettes, the use of self-cleaving 2A peptides can enable expression of both mAb chains from a single ORF.¹⁴⁴ With various self-cleavage peptide designs for efficient transgene expression available, a systematic comparison of the best 2A peptides for AAV-driven bnAb expression in primates seems warranted.^145,146 The ITRs, which bookend the AAV transgene expression cassette, are necessary for the concatemerization of DNA within the episome and implicated in supporting transgene expression. These regions contain CpG motifs which are recognized by TLR-9. Recent studies have focused on bypassing this recognition system either through removal of CpG motifs¹⁴⁷ or the inclusion of short DNA oligonucleotides to inhibit TLR-9 activation.¹⁴⁸ The removal of all CpG motifs in the ITRs did not impact transgene encapsidation or transgene expression in muscle tissue proceeding IM delivery in mice, but did result in reduced vector production.¹⁴⁷ While TLR-9 activation is likely evaded by the use of CpG-free ITRs, recognition through the innate immune system is still possible through TLR-2.¹¹⁴ Short DNA oligonucleotides designed to inhibit TLR-9 activation incorporated into the AAV vector genome have resulted in reduced innate immune and T cell responses in addition to increased transgene expression.¹⁴⁸ By evading or inhibiting innate immune activation, these strategies may also prove useful in preventing the elicitation of ADAs.

AAV transduction of myocytes can be enhanced through various approaches. This specificity can be engineered by fusion of cell-specific ligands to the AAV capsid, as is the case with eAAV9,¹⁴⁹ or through directed evolution, as done in the development of myoAAV.¹⁵⁰ To enhance transduction efficiency, eAAV9 contains an insulin-mimetic peptide that increases its binding avidity for the highly expressed insulin receptor on the skeletal muscle cell surface.¹⁴⁹ Compared to control AAV9 vectors, eAAV9 was 18-fold more efficient in transducing mouse skeletal muscle. Of note, the authors noticed that the efficiency of eAAV9 transduction was significantly enhanced by fasting prior to the vector inoculation, indicating that circulating insulin levels may negatively impact eAAV9-based gene transfer. MyoAAV was identified by Tabebordbar et al., using directed evolution of AAV capsids based on the levels of transgene RNA produced in vivo.¹⁵⁰ In contrast to other directed evolution strategies that rely on transduction efficiency as the readout for success, the method developed by Tabebordbar and colleagues has utilized successful transgene expression in addition to cellular-target transduction to select leading AAV capsids.¹⁵⁰ The use of myoAAV has resulted in increased muscle transduction compared to the parental AAV9 capsid and, notably, had reduced transduction of liver tissue. As off-target tissue transduction and hepatotoxicity are potential concerns in AAV-mediated gene transfer to skeletal muscle,151, 152, 153, 154, 155, 156 engineered capsids with target-specific enhancements may bypass these issues.

1.4. Concluding remarks

The potential for a single administration of AAV/Ig-HIV vectors to achieve lifelong control of HIV replication would make this treatment modality especially attractive in places with suboptimal healthcare infrastructures. In those settings, recurrent visits for parenteral injections with long-acting antiretrovirals or Ig-HIV cocktails may prove impractical. Despite the promise of AAV/Ig-HIV therapy as a “single-shot” cure for HIV infection, more work is still needed to improve transgene expression and optimize adjunct modalities to counteract host immune responses. Special consideration should be given to the selection of Ig-HIVs with maximal coverage and the design of efficient expression cassettes. With further preclinical and clinical work focused on generating therapeutic levels of Ig-HIV expression, an optimized AAV/Ig-HIV therapy regimen may lead to a functional cure of HIV infection. Finally, in order for AAV/Ig-HIV-based therapies to live up to their promise, PLWH will need to feel comfortable being treated with AAV/Ig-HIV vectors. As ART is already capable of suppressing HIV viremia, it will be crucial for healthcare providers to clearly outline the benefits and risks of AAV/Ig-HIV therapy to PLWH before administration of this permanent intervention.¹⁵⁷

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Dr. Martins has a consulting financial interest in Emmune, Inc., a company that is developing HIV immunotherapies based on the immunoadhesin eCD4-Ig. This potential conflict of interest is being managed by the University of Florida.

Data availability

No data was used for the research described in the article.