Highlights from the Tenth International Workshop on HIV Persistence during Therapy, December 13-16, 2022, Miami, Florida-USA

Abstract

The International Workshop on HIV Persistence during Therapy provides a forum in which HIV/AIDS researchers gather to share the latest research findings related to viral reservoirs and cure. The Tenth Workshop, which was attended by over 400 delegates, extended over 4 days and comprised eight sessions covering topics from the basic science of viral persistence to therapeutic approaches to HIV cure. Furthermore, satellite sessions on the first day of the Conference featuring cure research endeavours being pursued by the Bill and Melinda Gates Foundation as well as those being coordinated under the National Institutes of Health Martin Delaney Collaboratory program, provided important updates on research advances being made in these initiatives. As with previous conferences, the International Workshop on HIV Persistence during Therapy is primarily abstract-driven with only one invited talk for each of the sessions. This format, therefore, increases the number of presentations from early-stage investigators. Furthermore, presentations by Community representatives illustrated approaches to creating cure research literacy with effective messaging for the Community. The following article offers a synopsis of the meeting sessions. Due to space constraints, some presentations may have only been briefly discussed. Nevertheless, the Workshop abstracts can be found online (https://www/sciencedirect.com/journal/journal-of-virus-eradication/vol/8/suppl/S).

1. Session 1: basic science of HIV persistence

The first session of the Conference was focused on basic science concepts related to the nature of the latent HIV reservoir and the proviral genetic landscape, defining, transcriptional patterns, identifying the location of the reservoirs, and understanding the mechanisms of silencing or reactivating the integrated proviruses .¹

In the first talk of the session, Sarah Palmer (Center for Virus Research, Sydney, Australia) showed the different patterns of viral latency distribution after several cycles of analytical treatment interruptions (ATI).¹ She demonstrated that there are differences in the proviral genetic landscape between the various subsets of memory CD4⁺ T cells. Indeed, genetically-intact proviruses appear to be concentrated in specific memory T cell subsets and also in cells which are more proliferative. During her talk she showed evidence that many proviruses are transcriptionally active. This allows the immune system to target these proviruses. However, rapid cellular turnover rates counteract the host immune pressure. Focusing on plasma virions after reactivation, she demonstrated that not all virions in the plasma are infectious; in fact up to 45% are defective. She suggested that investigating the interplay between the virus and the host immune responses will provide insights as to how some HIV-infected individuals control HIV during an ATI.

Eli Boritz (Vaccine Research Center, NIAID, NIH, Bethesda, MD, USA) presented a new technology based on a custom microfluidic process termed “Focused Interrogation of Cells by Nucleic Acid Detection and Sequencing” (FIND-Seq).² Using this technology, the group was able to select HIV-infected cells, perform whole transcriptome sequencing and compare them with non-infected cells. Gene expression patterns of HIV-infected memory CD4⁺ T cells under ART were distinct from those of uninfected memory CD4⁺ T cells. HIV-infected memory CD4⁺ T cell transcriptomes demonstrated inhibition of cell death and anti-proliferative signaling, but interestingly also expressed transcriptomic signatures of HIV silencing. Altogether these results demonstrating that HIV-infected cells modify the host gene transcriptomic patterns to favor HIV silencing, cell survival, and cell proliferation, has important implications for the development of HIV cure strategies.

Adam Capoferri (National Cancer Institute, Frederick, Maryland, USA) reported on a study comparing proviral HIV reservoirs in viremic controllers (VC) versus viremic non controllers and ART-suppressed participants.³ The group observed how, despite the fact that the number of cells infected were lower in VC, the fraction of infected cells with usRNA did not differ across the 3 groups. This shows that the levels of plasma viremia may be determined by the total number of infected cells and the number of rare cells with high levels of HIV usRNA but not by differences in the fraction of infected cells with HIV usRNA. These results indicate that when HIV proviruses are expressed in VC, the host or other factors limit viral spread (e.g. enhanced (natural killer cells (NK) or cytotoxic T lymphocyte (CTL) activity, potent antibody responses, host cell/viral factors).

Yuyang Tang (UNC HIV Cure Center and IGHID, North Carolina, USA) described an interesting study using samples from the special “Last Gift Study” cohort.⁴ They were able to isolate and measure the reservoir in brain myeloid cells from autopsy donations from four persons with HIV (PWH) on antiretroviral therapy (ART) using the quantitative viral outgrowth assay (qVOA). Viral outgrowth was detected and the viral particles recovered were sequenced demonstrating CCR5 tropism and low CD4 usage, indicative of macrophage or microglia origin. This study provides confirmation of brain HIV reservoirs that can persist under ART. HIV persistence dynamics in tissues such as the brain is still not well understood due the difficulties with obtaining biopsies from viral sanctuaries. The “Last Gift Study" offers a great opportunity to understand the nature of HIV reservoirs beyond the blood.

Luisa Mori (UF Scripps Biomedical Research, Florida, USA) presented results from a study conducted to further define the host factors that regulate HIV transcription.⁵ They performed an interesting RNAi screen consisting of shRNAs embedded in micro-RNA backbones in order to simultaneously probe all host chromatin regulatory factors in the J-Lat 10.6 cell line model of HIV latency. They identified the p400 (Tip60/NuA4) complex as a key regulator of HIV latency. Inhibition of members of the p400 complex resulted in a significant induction of HIV transcription suggesting that this complex may play a role in suppressing HIV transcription. They showed by ChIP-seq and co-IP experiments that the p400 complex interacts with RNA pol II, and that the depletion of p400 resulted in more RNA pol II at the HIV genome level, indicating a previously unidentified direct role of this family of proteins in negatively regulating HIV transcription. The p400 complex may represent a novel target that can be exploited to reinforce HIV latency in the "block-and-lock" approach for an HIV cure or to reactivate HIV from latency in the "kick and kill" approach.

Carine Van Lint (Service of Molecular Virology, University of Brussels, Belgium) discussed results from studies undertaken to further define the mechanisms by which the epigenetic integrator, Ubiquitin-Like With PHD And RING Finger Domain 1 (UHRF1) regulates HIV latency.⁶ Using various molecular biology tools, including ChIP-qPCR, RNAi, mass-spectrometry and co-immunoprecipitation assays, her group showed that UHRF1 controls HIV through similar epigenetic mechanisms in both myeloid and lymphoid cells. UHRF1 activity was tat independent, although the protein co-precipitates with tat. Interestingly, disrupting the catalytic RING domain of UHFR1 strongly reduces the ability of this protein to repress HIV transcription, suggesting an additional, non-epigenetic, ubiquitin ligase role of UHFR1 in HIV transcription repression. UHFR1 therefore represents another protein that could be targeted in HIV cure approaches.

Marion Pardons (HIV Cure Research Center, Ghent University, Belgium) reported on recent studies undertaken to improve HIV latency reversal and define the transcriptomic profile of HIV-producing cells after latency reversal.⁷ They tested the ability of a tat mimetic (Tat#1) combined with different latency reversing agents (LRAs) to induce HIV p24 production in CD4⁺ T cells isolated from suppressed aviremic donors. They observed that the combination of Tat #1 and the HDAC inhibitor panobinostat induced the highest proportion of p24⁺ cells as measured by flow cytometry as well as the highest levels of p24 released into culture supernatant relative to PMA/ionomycin treatment. Interestingly, treatment of CD4⁺ T cells with Tat#1 does not significantly alter the transcriptomic profile of treated cells, allowing for transcriptional discrimination between cells producing p24 and those that do not. They observed significantly higher levels of CCL5, GZMA and the long-non-coding RNA SOD1P3 and lower levels of IL7R and GZMB transcripts in p24⁺ cells compared to p24 negative cells, results which were confirmed by flow cytometry.

2. Session 2: in vitro and animal model studies of HIV persistence

The session was chaired by Ann Chahroudi from Emory University, School of Medicine Atlanta, USA and Afam Okoye from the Oregon Health and Science University, Beverton, Oregon, USA began with a presentation from Jonah Sacha (Oregon Health and Science University) who discussed the role of CCR5 in HIV prevention and cure.⁸ CCR5 is the major co-receptor for HIV/SIV entry and reports of HIV cure in people with HIV who had received hematopoietic stem cell transplantation (HSCT) from CCR5delta32 donors to treat malignancies suggest that targeting CCR5 may be an effective strategy to eliminate HIV persistence. In dissecting the mechanism of cure following transplant, Sacha reported that graft versus host disease alone was not sufficient, as virus persisted in a Mauritian cynomolgus macaque HSCT model with evidence of full donor chimerism with wild type CCR5. He also showed that virus can accumulate in donor cells as they engraft, highlighting the importance of shielding donor cells from infection. In this context, Sacha highlighted the potential use of the anti-CCR5 humanized monoclonal antibody leronlimab which binds to the CCR5 receptor and competitively inhibits the chemokine CCL5. In a pre-exposure prophylaxis study (PrEP), leronlimab protected rhesus macaques from CCR5-tropic SHIV infection. Sacha's presentation also demonstrated the potential for AAV-vectored delivery of leronlimab for use as PrEP and potentially for facilitating long-term virus suppression.

Ramon Lorenzo-Redondo (Northwestern University, Illinois, USA) next described the use of spatial transcriptomics to characterize SIV reservoirs in macaque tissues during ART and after ART cessation.⁹ Specifically, the investigators used PET-CT-guided tissue collection at necropsy to apply novel 10X genomics Visium Spatial Gene Expression Technology that combines RNA sequencing with classical histology and antibody staining. A series of optimization steps were described that allowed Lorenzo-Redondo and colleagues to identify distinct transcriptomic signatures associated with tissue regions containing SIVgag⁺ cells. These included certain interferon-stimulated genes that were found to be upregulated in the transverse colon early after ART interruption. Furthermore, an association between different expression clusters and specific tissue structures was found. This exciting resource has the capacity to greatly expand our understanding of the features that define tissue reservoirs.

Resistance of HIV-infected cells to their elimination by CTLs is a potential barrier to immune-mediated control of HIV. Andrea Gramatica (Weill Cornell Medicine, New York, USA) reported on the research to generate transcriptomic profiles of HIV-infected CD4⁺ T cells co-cultured with HIV-specific CTLs. Gramatica found the enhancer of the zeste homolog 2 (EZH2) to be significantly upregulated in cells that survive CTL killing.¹⁰ EZH2 is a histone-methyltransferase that silences gene expression by methylating lysine 27 on histone 3 and is involved in the regulation of MHC-I expression. Gramatica suggested that EZH2 may collaborate with HIV-nef to evade CTL recognition. He described how the FDA-approved EZH2 inhibitor tazemetostat (EPZ-6438) was able to enhance the elimination of HIV-infected CD4⁺ T cells in vitro through increased expression of MHC-I. Finally, Gramatica showed a marked decrease in viral loads in HIV-infected humanized mice treated with tazemetostat at a dose of 500 mg/kg.

Gregory Del Prete (Fredrick National Laboratory for Cancer Research, Maryland, USA) next discussed the unresolved question of whether there is ongoing virus replication during ART.¹¹ To tackle this question, Del Prete and colleagues challenged 10 rhesus macaques with the barcoded SIVmac239 M2 and then started them on either a 3- versus 4-drug ART regimen. The 3-drug regimen consisted of the reverse transcriptase inhibitors tenofovir disoproxil fumarate and emtricitabine and the integrase inhibitor dolutegravir; the 4-drug regimen added the protease inhibitor darunavir. ART intensification did not impact plasma viral load decay kinetics or the magnitude of virologic suppression. There was no evidence of viral sequence evolution suggesting that residual plasma viremia in SIV-infected rhesus macaques during ART is likely due to virus production from previously infected cells rather than ongoing virus replication.

Benjamin Varco-Merth (Oregon Health and Science University, Portland, USA) investigated whether alemtuzumab, a licensed pan-lymphocyte-depleting monoclonal antibody targeting CD52⁺ cells, can be used to deplete SIV reservoirs either at the time of ART initiation or during full ART suppression in SIV-infected rhesus macaques.¹² He showed that alemtuzumab can induce a substantial depletion of circulating CD4⁺ T cells in peripheral blood and lymph nodes. However, depletion was followed by CD4⁺ memory T cell proliferation and reconstitution of cells in blood. After ART cessation, most animals rebounded, with no difference in time to rebound between alemtuzumab-treated animals and controls. However, 3 animals with smaller viral reservoirs did not rebound after ART release or subsequent CD8⍺ cell depletion, suggesting smaller reservoirs may be more susceptible to disruption.

Anais Chapel (Institut Pasteur, Paris, France) then described how the constitutive NKG2A expression and timing of ART initiation impact the role of NK cells following ART interruption, providing insights into the potential mechanism of post-treatment viral control.¹³ Cynomolgus macaques were infected with SIVmac251, started on ART either 4- or 24-weeks post-infection and were treated for 2 years prior to ART cessation. Early ART favored post-treatment control as 82% of animals in the early treatment group maintained plasma viral loads below 400 HIV-1 copies/ml compared to the late ART group where 25% were categorized as post-treatment controllers. Interestingly, Chapel also found that the timing of ART impacted the phenotype of NK cells. Indeed, late ART resulted in a higher frequency of NK cells that expressed the activation markers NKp46 and NKp30, which were associated with higher viral loads post-ART interruption.

The final abstract presentation of the session was from Dennis Copertino (Weill Cornell Medicine, New York, USA). He described how the LRA 3-hydroxy-1,2,3-benzotriazin-4(3H)-one (HODHBt) synergizes with IL-15 to enhance the cytotoxic function of HIV-specific CD8⁺ T cells.¹⁴ HODHBt enhances IL-15-mediated reactivation of HIV by increasing STAT5 occupancy of the HIV-LTR. STATs also play an important role in the functional activity of CD8⁺ T cells and the combination of IL-15 and HODHBt increased MHC-I expression and enhanced functional recognition of targets by HIV-specific CTL in vitro. HODHBt plus IL-15 also increased STAT5 phosphorylation and the expression of the cytolytic molecules granzyme A and perforin. In terms of impact on HIV persistence, Copertino showed that in vitro treatment with HODHBt plus IL-15 reduced the level of intact HIV in PBMCs from PWH on ART, however the magnitude of the reduction was variable. Reservoir reduction was not observed in CD8-depleted conditions, suggesting that the decrease in intact HIV was associated with enhancement of CTL function.

3. Session 3: virology of HIV persistence

Nancie Archin¹⁵ (University of North Caroline, USA) opened the session with a discussion on the failure to include equal numbers of women in HIV-1 cure studies. She showed that, although over 50% of people living with HIV-1 are women, only 11.1% of participants in cure-related studies have been women. Studies of HIV-1 viral load, pathogenesis, persistence and immunological responses have suggested that sex-related differences affect clinical outcomes and may influence the establishment and maintenance of the viral reservoir. For example, HIV reservoirs may increase in size in post-menopausal women rather than decrease, as seen in other cohorts. However, more studies are needed to determine the effect of estrogen and estrogen receptor-1 expression on HIV persistence. Women in pre-, peri-, and post-menopause should be included in studies investigating approaches towards achieving HIV remission without ART.

Francesco Simonetti et al.¹⁶ (John Hopkins University, USA) showed that defective proviruses can contribute to non-suppressible viremia on ART, defined as multiple consecutive RNA levels >75 copies/mL of plasma during ART, without drug resistance. Their studies identified several donors with non-suppressible viremia resulting from proviruses with 5’ leader defects, including deletions or mutations in the major splice donor site. Proviruses were present in clones of effector memory T cells that produced non-infectious virus particles with low levels of env on the cell surface. Future studies will investigate how often residual viremia is due to the production of non-infectious virus particles.

Caroline Dufour et al.¹⁷ (Montreal University, Canada) reported that a cell-adhesion molecule also known as VLA-4 or α4β1 is expressed on the surface of cells carrying replication-competent HIV proviruses and that this molecule may be targeted to reduce the size of the HIV reservoir. The investigators sequenced 309 HIV genomes from single cells expressing HIV Pp4 collected from 6 donors on ART. Of the 12 proviruses that they found to be sequence-intact, 92% were present in memory T cells expressing higher levels of α4β1 than the total CD4⁺ T cell population (1.62-fold more α4 and 1.2-fold more β1). Sorting for α4β1^high CD4⁺ T cells resulted in cell populations that were 27 times more enriched for replication-competent HIV as measured by viral outgrowth assays. The team is currently exploring whether this finding can be recapitulated in humanized mice to develop a model that can be used to test possible interventions.

Mary Grace Katusiime et al.¹⁸ (NCI, USA) demonstrated that HIV can persist in clones of CD4⁺ naïve T cells in children born with HIV and on long-term ART. Measuring HIV levels in 8 children, they found all to have detectable HIV DNA in the sorted naïve T cell populations at levels about 12-fold lower than the memory T cell populations. In one child with the highest level of HIV DNA, they identified 8 clonal populations of infected naïve T cells using the HIV integration sites assay. At least 30% of the integration sites identified in the naïve cell population were present in T cell clones. Using HIV LTR as a denominator, 2% of the infected naïve cells had proviruses that were predicted to be intact. The investigators concluded that infected naïve T cells may be a renewable source of infected central and effector memory T cells during ART.

Diane Bolton et al.¹⁹ (US MHRP, Bethesda, USA) examined HIV-1 RNA expression in memory CD4⁺ T cells from 24 peripheral blood samples and 4 lymph node samples from the RV254/SEARCH 010 acute infection cohort. They measured cell-associated HIV-1 RNA in limiting dilution to assess the frequency of HIV-1 RNA⁺ cells during acute infection. HIV-1 RNA expression level (including unspliced (gag), spliced (env/vpu/nef and tat/rev), and both (LTR-U3 at the 3’ end of RNA) was compared with 27 inflammatory cytokine levels in plasma in order to identify inflammatory markers correlating with HIV-1 RNA expression. They found that a median of 4% memory CD4⁺ T cells (0.1–30%) were HIV-1 RNA⁺. Only 0.2% of cells expressed spliced HIV-1 RNA, while 5% of cells express any HIV-1 RNA. IL-12p70 was the plasma protein that shows significant and positive correlation with HIV-1 RNA expression (rho = 0.7, P-FDR = 0.01).

Joshua Kufera from Robert Siliciano's group²⁰ (John Hopkins University School of Medicine , USA) presented an updated version of microwell culture to measure the proliferation capacity of cells harboring intact versus defective HIV-1. In the previous study by Brunauer et al. (Nature 2019), cells from PLWH were plated at limiting dilution so that each well contained only one HIV-1-infected cell (along with hundreds or thousands of uninfected cells from the same individual). After weeks of CD3/CD28 stimulation in the presence of ART (to prevent new rounds of infection in vitro), these culture microwells were split into two aliquots for integration site analysis in one aliquot and frequency of intact provirus measurement (by digital-droplet PCR-based method IPDA). Cells harboring intact proviruses would die of viral cytopathic effect with few proliferating cells remaining in the microwell, while cells harboring defective proviruses did not cause significant viral cytopathic effects and proliferated extensively. Extending from this knowledge, Kufera et al. measured the supernatant HIV-1 RNA production in these microculture wells. They confirmed as in their previous study that cells harboring intact proviruses, which may produce more viral particles upon CD3/CD28 stimulation, cannot proliferate because of viral cytopathic effect.

Sarah Gowanlack from Guinevere Lee's group²¹ (Weill-Cornell, New York, USA) presented an updated version of the HIV-1 intact provirus DNA assay (IPDA), allowing cohort and clade-specific measurement of the frequency of intact HIV-1 proviruses to broaden the use of the initial IPDA to clades prevalent in Africa, particularly A1, D, and their recombinants. Briefly, to more accurately capture the HIV-1 genome diversity by digital droplet qPCR probes, specifically for participants recruited at Rakai, Uganda, Gowanlack et al. performed HIV-1 near-full-length proviral sequencing using FLIPseq, calculated the predictive value of new probe designs, and tested on clinical samples. They found that by designing custom digital droplet qPCR probes to better capture the sequence diversity, they could increase the sensitivity for HIV-1 detection and accuracy for measuring the frequency of cells harboring intact HIV-1 proviruses.

4. Session 4: immunology of HIV persistence

The session on the Immunology of HIV Persistence focused on cellular immune responses – primarily CD8⁺ T cell responses, but also NK cells. These arms of the adaptive and innate immune systems – respectively – are known to exert a partial control of HIV replication in the absence of ART and it is suggested that they can be harnessed to reduce or eliminate viral reservoirs. Presentations in this session furthered our understanding of both the requirements for natural immune-mediated control of HIV replication and of the interplay between the immune system and the HIV reservoir on ART.

Decades of research have established some fundamental distinctions between effective versus ineffective CD8⁺ T cell responses in the context of untreated HIV infection, where effective responses tend to: i) target conserved and highly networked epitopes ii) maintain proliferative potential and polyfunctionality, and iii) be restricted by ‘protective’ HLA-B alleles. Nonetheless, our understanding of CD8-based control remains incomplete. In this session, D. R. Collins et al. (USA) focussed on an anatomical dimension of CD8⁺ T cell efficacy by studying their frequencies and functionalities within lymph node follicles – a key sanctuary for HIV. While other studies have established that CD8⁺ T cells in follicles are generally poorly cytotoxic relative to those in blood, Collins et al. showed that, in virological controllers, these exhibit cytotoxic profiles in proportion to both tissue viral burden and proximity of CD8⁺ T cells to infected cells – implicating a role in containing HIV replication within follicles. Keeping with the same anatomical theme, M. Müller-Trutwin et al. (Pasteur Insitut, Paris, France) asked whether NK cells might be taught to target and reduce HIV reservoirs in follicles – perhaps an alternative way of containing virus replication at these sites. In SIV-infected rhesus macaques, they showed that IL-6 promoted NK cell migration into follicles, while administering IL-21 to ART-treated SIV-infected macaques, promoted NK cell maturation and reductions in lymph node viral reservoirs. Together, these two talks highlighted the importance of anatomical viral sanctuaries in the outcome of viral control and persistence and gave insights into how immune effector function at these sites can be enhanced.

Focusing on a second compartment of HIV persistence, M, Wang and al. (Yale University, USA) explored the types of HIV-infected CD4⁺ T cells in cerebral spinal fluid (CSF) and how they related to the peripheral blood. Using single-cell analyses, they found that HIV RNA expressing cells in the CSF were predominantly expanded clones of central memory CD4⁺ T cells. Sequencing revealed increased expansion of T cell clones in the CSF compared to blood, with lower overlap between CSF and peripheral clones in some participants. Together, these results support the compartmentalization of some HIV-infected T cells within the CNS and highlight the need to continue study of this potentially distinct HIV reservoir.

Three inter-related presentations studied the relationship between HIV-specific T cell responses and the HIV reservoir on ART. Such studies were motivated, in part, by recent publications highlighting: i) ongoing HIV expression on ART^22,23 and ii) progressive decay of intact HIV proviruses (but not total HIV DNA).²⁴ Together, these raise the question of whether ongoing CD8⁺ T cell recognition is driving reductions in the ‘active reservoirs’. Two previous studies provided evidence that HIV nef-specific CD8⁺ T cells preferentially recognize HIV antigens in vivo in individuals on long-term ART (median 7 years ART at study entry).^25,27 In this Session, H. Takata et al. (Oregon Health & Science University, USA) studied a subtype A/E infected cohort on relatively short-term ART (>1.5 years) and showed direct correlations between CD8⁺ T cell responses to each of HIV-nef, pol, and gag (but not env) and levels of HIV DNA on ART, implicating ongoing antigenic stimulation. Moreover, drivers of exhaustion – PD-1 and TOX – on HIV-specific CD8⁺ T cells correlated directly with HIV DNA, while drivers of stem-like memory – IL-7R and TCF-1 – showed inverse correlations. In a complementary study, M, Dube et al. (CRCHUM, Montreal, Canada) used a method to detect ‘leaky’ HIV reservoirs (flow cytometric detection of HIV-RNA positive cells by RNAflow-FISH) in individuals on ART for ≥3 years. They observed HIV RNA in 14/18 donors, with a median of 21 HIV-RNA positive cells/106 CD4⁺ T cells. Strikingly, the magnitudes of nef- pol- and env-specific CD4⁺ T cell responses, as well as gag-specific CD8⁺ T cell responses, correlated with frequencies of HIV-RNA + cells, with parallel trends for gag-specific CD4⁺ T cell and gag- pol- nef-specific CD8⁺ T cells. These two studies provide strong evidence that HIV-specific CD8⁺ T cell responses continue to encounter appreciable levels of antigen on ART. It will be interesting for future studies to explore the role of ART duration in these versus published observations, i.e: could the observation that only nef-specific responses correlate with HIV DNA on long-term ART (>7 years)^25,26 reflect the observation that proviruses with intact nef ORFs are preferentially maintained over long-term ART²⁸?

In the third related study, A. Ward and al. (Weill Cornell, USA) asked whether ongoing recognition of ‘active’ or ‘leaky’ HIV reservoirs may drive on ART decay of intact HIV proviruses. As reported by others, they observed a decay in intact HIV proviruses – but not total HIV DNA – over 144 weeks of ART, alongside marked reductions in poly-adenylated unspliced HIV RNA. Contrary to the hypothesis, however, neither of these declines correlated with magnitudes of granzyme-B- and IFN-g-producing T cell responses to any HIV gene product. While this study did not reveal evidence for HIV-specific T cell responses driving on ART reservoir decay, such a role would be revealed by parameters not measured here. Alternatively, factors other than HIV-specific CD8⁺ T-cells may play dominant roles in reducing reservoirs. Taking an unbiased approach to identifying such drivers, M. Peluso et al. (UCSF, USA) assessed correlations between 32 pro-inflammatory or regulatory cytokines and intact provirus decay in 76 individuals on ART for a median of 10.4 years. They observed that higher levels of Galectin-9 most strongly associated with greater levels of subsequent intact HIV decay. Roles of galectin-9 in both HIV expression and cytotoxic NK and CTL effector function were noted as leads for mechanistic studies into this association.

While the other studies in this session generally focused on canonical cytotoxic functions of HIV-specific CD8⁺ T cells, S Mutascio et al. (Emory University, USA) provided novel mechanistic insights into a non-canonical function – the recently reported ability of these cells to promote HIV latency.²⁹ Using an in vitro latency model, they showed that co-culture with CD8⁺ induced reductions in both gag expression and NF-kb activation. These impacts were sustained in CD4⁺ T cells for 72 h after removal of CD8⁺ T cells suggesting a pro-latency effect. The ongoing dissection of the underlying mechanisms may lead to novel approaches to enhancing latency reversal.

In summary, this session on the Immunology of Persistence raised new insights in several areas of relevance to the field. Beginning with HIV persistence in lymph node follicles, authors addressed the evolving role of cytotoxic T cells in HIV containment and the potential for immunotherapy to drive more effective NK responses. In vitro studies supported a pro-latency role for CD8⁺ T cells in HIV persistence, while HIV persistence in the CNS was shown to involve clonal expansion of potentially CNS-specific HIV-infected CD4⁺ T cells. Several studies sought cellular immune correlates of ongoing HIV expression and declining intact virus reservoirs over time, elucidating the complex, likely multidirectional relationship between persistent HIV antigen production and HIV-specific CD8⁺ T cell numbers and function, and identified galectin-9 as a potential marker of HIV decay. Together, these presentations highlighted key areas of current and future study.

5. Session 5: drug discovery development & pharmacology

Session 5 of the Tenth International Workshop on HIV Persistence during Therapy, co-chaired by Romas Geleziunas (Gilead Sciences, Foster City, California, USA) and Jan van Lunzen (Radboud University Medical Center, Nijmegen, The Netherlands), focused on novel approaches to reactivating or silencing the HIV reservoir.

The session began with an overview lecture by Richard Dunham (ViiV Health Care, Raleigh-Durham, North Carolina, USA) on the potential role of inhibitors of apoptosis proteins (IAPis) as LRAs.³⁰ Given the relatively modest activity of LRAs tested in clinical trials to date, more potent agents are needed. One problem with more potent activators of gene expression such as T-cell receptor (TCR) and protein kinase C (PKC) agonists is that they increase expression of approximately 3000 genes (roughly 10% of human genes). Dunham's group sought a more selective NF-κB agonist that would decrease by 10-fold the number of genes activated. They hypothesized that activators of the non-canonical NF-κB pathway would be more effective LRAs with fewer off-target effects. One class of IAPis are Smac mimetic (Smacm) compounds, which bind to the baculovirus IAP repeat (BIR) domains found on X-linked IAP (XIAP) and cellular IAP (cIAP). Binding of Smacm to the BIR domain targets these IAPs for degradation, resulting in accumulation of NF-κB-inducing kinase (NIK) and increased activation of the non-canonical NF-κB pathway. The most potent of these agents, AZD5582, is a bivalent Smacm that activates proviral expression in vitro, in HIV-infected, ART-suppressed humanized mice, and in SIV-infected ART-suppressed macaques. Modest additivity was observed when AZD5582 was combined with an SIV-specific rhesus mAb and N-803 (an IL-15 superagonist), resulting in up to a 0.5-log10 decrease in SIV DNA in various tissues. In vitro synergy has been observed between AZD5582 and inhibitors of bromodomain and extra-terminal motif (BET) proteins such as iBET151 and JQ1. Ciapavir, another bivalent Smac mimetic, also activates HIV expression in the humanized mouse model. This promising data supports further preclinical and early-phase clinical trials of IAPis in combination with other interventions that target the HIV reservoir.

Alberto Bosque (Georgetown University, Washington, DC, USA) presented results of the characterization of a dual PTNP1/PTPN2 inhibitor to target latent HIV reservoirs.³¹ His group previously showed that 3-Hydroxy-1,2,3-Benzotriazin-4(3H)-one (HODHBt) reactivates latent HIV-1 and enhances IL-15-mediated NK cell effector function by reducing STAT5 turnover. Using a thermal shift assay followed by mass spectrometry, they identified the non-receptor phosphatases PTPN1 and PTPN2 as targets of HODHBt in peripheral blood mononuclear cells (PBMCs) and NK cells. The role of these check-point proteins in STAT5 phosphorylation and transcriptional activity was confirmed by CRISPR/CAS9 knock-out of the phosphatases. Current efforts are focused on identifying more potent PTPN1/PTPN2 inhibitors that might target the HIV reservoir alone or in combination with other immune-based strategies.

Mario Manzaneres (Instituto de Salud Carlos III, Madrid, Spain) reported results of a phase 2 exploratory trial of the tyrosine kinase inhibitor (TKI) ponatinib.³² Like other TKIs such as dasatinib, ponatinib inhibits SAMDH1 phosphorylation, reduces IL-7-mediated T cell proliferation and TCR-mediated activation, avoids proviral reactivation, increases CTL antiviral activity and interferes with reservoir reseeding. In this study, 9 participants with chronic myeloid leukemia (CML) and without HIV who had responded to induction therapy with imatinib received 1 year of consolidation treatment with ponatinib. Sustained reduction in SAMHD1 phosphorylation in CD4⁺ T cells was maintained for 1 year after ponatinib. CD4⁺ T cells showed a mean 8.8-fold reduction in infectibility by HIV in vitro following 1 year of ponatinib treatment, an effect that persisted for an additional year among the 5 participants who remained relapse-free. This effect was accompanied by increased antiviral activity of CD8⁺ T cells and increased degranulation capacity of NK and Tγδ cells. Because the experiments were carried out with total PBMCs, it is unclear to what extent the reduction in susceptibility to HIV was mediated by intrinsic resistance of CD4⁺ cells to infection induced by ponatinib versus enhanced antiviral activity of CD8⁺ T cells, NK cells and Tγδ cells. This question could be resolved by quantifying susceptibility to infection of purified CD4⁺ T cells from these patients.

Sonia Mediouni Jablonski (UF Scripps Biomedical Research, Jupiter, Florida, USA) described the identification and characterization of novel HIV tat inhibitors.³³ Previous efforts in the Valente lab had identified didehydro-Cortistatin A as a tat inhibitor that blocked HIV transcription with nanomolar potency. Unfortunately synthesis required 13 steps and the molecule was considered undruggable. The laboratory therefore performed high throughput screening of a library of ∼580.000 small molecules to identify novel Tat inhibitors. This screen identified 3 candidate molecules with activity in the low micromolar range and 50% cytotoxicity concentrations (CC50) >50 μM. These molecules, which can be synthesized in 10 steps or less, reduced HIV expression in OM 10.1, JLAT and ACH2 cells. All 3 molecules induced tat degradation but had no effect on tat mRNA. The proposed mechanism of action is that these novel compounds act as “molecular glues” that promote tat recruitment to the Cullin4A/B-Cereblon E3 Ubiquitin ligase complex and subsequent degradation by the 26 S proteasomal pathway. Next steps include improving the potency and pharmacological properties of the lead compounds through medicinal chemistry. If these compounds advance into clinical trials it will be important to determine whether limited courses lead to sustained HIV inhibition or whether chronic administration will be required.

Guiomar Casado-Fernández (Instituto de Salud Carlos III, Madrid, Spain) reported the results of studies on the effect of the PKI dasatinib on the metabolic activity of CD4⁺ T cells.³⁴ As previously noted, TKIs can interfere with the maintenance of the HIV reservoir by a number of mechanisms (see Manzaneres, above). Because HIV selectively infects CD4⁺ T cells with enhanced glycolysis and oxidative phosphorylation, the investigators sought to determine whether dasatinib induces metabolic reprogramming of CD4⁺ T effector memory (TEM) and TEM-CD45RA+ (TEMRA) cells, resulting in reduced glycolytic activity and thereby interfering with the replenishment of the HIV reservoir. The phosphoproteome of CD4⁺ T cells from HIV-negative donors as well as from PWH with CML who received ART and dasatinib, was analyzed by LC-MS/MS. Dasatinib modified phosphorylation of more than 130 proteins involved in metabolic pathways, reduced mitochondrial production of ATP and inhibited glucose uptake in TEM and TEMRA cells. Whether these metabolic effects are mechanistically related to the inhibition of HIV reactivation observed in previous studies of dasatinib or is simply a correlate of reduced T cell activation remains to be determined.

Matthew Marsden (University of California, Irvine, California, USA) described efforts to identify better and safer PKC agonists to serve as LRAs.³⁵ Prototype drugs, including bryostatin 1, prostratin and ingenane are potent activators of HIV in vitro but are considered too toxic for use in PWH. These drugs have two domains, one that binds to PKC and another involved in membrane interaction and downstream signaling. Prodrugs of these PKC agonists demonstrated delayed activity in vitro and in vivo (in HIV-infected humanized mice), and were better tolerated over a wide range of doses in animal studies. Treatment of ART-suppressed HIV-infected humanized mice with the bryostatin-1 analog SUW133 resulted in significantly delayed viral rebound after ART interruption and reduced diversity of the rebounding virus. Additional preclinical safety studies are needed to determine whether these PKC prodrugs are suitable for testing as LRAs in PWH.

Youry Kim (Peter Doherty Institute, Melbourne, Australia) presented data from in vitro studies on the effect of the combination of romidepsin and the BCL-2 antagonist venetoclax on the HIV reservoir.³⁶ Venetoclax is approved for the treatment of chronic lymphocytic leukemia and sensitizes malignant cells to apoptosis. In the experiments reported here, PBMCs from PWH on suppressive ART were harvested, treated with venetoclax for 24 h and then with either romidepsin or PMA/PHA. Venetoclax alone induced high levels of cell death within T cell subsets, particularly naïve and effector memory T cells. The combination of romidepsin plus venetoclax resulted in significantly greater reductions in integrated HIV DNA compared to either drug alone. A similar pattern of decline in intact proviral DNA, assessed by the intact proviral DNA assay was observed. Whether a drug such as venetoclax can safely be used in otherwise healthy PWH is an open question and may limit the applicability of these results.

6. Session 6: cell and gene therapies

The sixth oral session of the Workshop focused on cell and gene therapies and was chaired by Tricia Burdo of Lewis Katz School of Medicine at Temple University, Philadelphia, Pennsylvania, and Javier Martinez-Picado of IrsiCaixa AIDS Research Institute, Barcelona, Spain.

The session began with a presentation by Matthew Gardner (Emory National Primate Research Center), who discussed the use of broadly neutralizing antibodies (bNAbs) expressed through adeno-associated virus (AAV) vectors to suppress viremia after antiretroviral therapy (ART) removal.³⁷ His preliminary showed eCD4-Ig, an antibody-like HIV entry inhibitor, can suppress SIV viremia despite low levels of expression. In a pilot study of SIV239-infected macaques treated with AAV8-eCD4 followed by an AAV1 boost 12 weeks later, one animal did not rebound, and one had partial control. With AAV9-eCD4, one macaque had partial control. Since some animals that received AAV9-eCD4 vectors resulted in uncontrolled viral rebound, Gardner discussed ways to improve the consistency of gene therapy strategy. They engineered new eCD4 expression cassettes for improved expression using a P2A peptide to express bNAbs. He evaluated five different AAV capsids in macaques and showed AAV9 had the highest expression, was well-tolerated with no anti-drug antibodies (ADA). Thus, AAV9 is suitable replacement for AAV1/8 for intramuscular inoculations, especially as AAV9 negative (42%) macaques are easier to find than AAV1 (8%) or AAV8 (16%) negative animals.³⁷

The next speaker was Priti Kumar from Yale University, who showed administration of anti-CD7 resulted in the labelling of human T, NK, monocytes, but not CD7-non-expressing B cells in humanized mice.³⁸ Her first experiments used anti-CD7 coated integrating lentivirus to package an expression vector for Cas9 and guide RNAs targeting CCR5. In vitro, 50% of T cells were transduced. In vivo, CCR5 editing efficiencies of about 60% were achieved in T cells of NSG mice. Dr. Kumar's humanized mice studies showed mice treated with CD7-lentivirus controlled HIV infection and maintained CCR5-edited cells. Next, they generated DNA-free virus-like particles (VLP) surface-decorated with anti-CD7 for delivery of CRISPR-Cas9 ribonucleoprotein complexes (RNPs). In mice treated with the CD7-targeted VLP, CCR5-edited CD4⁺ T cells (30% efficiency) expanded after ART interruption, leading to stabilization of blood CD4⁺ T cells, control of plasma viremia, and ART-free remission with no viral outgrowth from splenocytes.³⁸

Gabriela Webb from the Oregon Health and Science University next discussed CCR5 blockade as a scalable non-transplantation approach for long-term ART-free HIV remission.³⁹ She sought to mimic the CCR5Δ32/Δ32 phenotype by using the CCR5-blocking IgG4 monoclonal antibody leronlimab and test if AAV9 could induce long-term expression in SHIV-infected macaques. Webb discussed two proof-of-concept cases of long-term transgene expression and suppressed viremia. In case #1 the macaque received AAV9 human leronlimab, dexamethasone and tacrolimus to limit immune activation. This animal reached 100% CCR5 receptor occupancy on CD4⁺T cells, possessed detectable plasma leronlimab with no antidrug antibodies (ADA). SHIV viremia became undetectable at 4 weeks until 33 weeks. In case #2, the macaque received AAV9 macaque leronlimab with no immune suppression. 100% CCR5 receptor occupancy was reached and plasma leronlimab was detected within 2 weeks. ADA developed at 4 weeks post-AAV, but then resolved. SHIV viremia became undetectable shortly after resolution of ADA.³⁹ This demonstrates the potential of AAV vectors for sustained antibody-based CCR5 blockade as a gene therapy approach for long-term ART-free HIV remission. Future studies will aim to reduce ADA, increase leronlimab expression, and test alternate AAV capsids.

The next speaker was Dan Claiborne from Wistar who described CCR5-based genome editing using a CRISPR-Cas9/RNP approach targeting human hematopoietic stem cells (HSCs).⁴⁰ He used in silica prediction software and in vitro screening methods for gRNA. HSCs (80–97% CCR5 disruption) were engrafted into 6–8-week-old NSG mice. CCR5-edited HSCs displayed slightly delayed, but otherwise normal hematopoiesis. High frequency of CCR5 editing was detected in the myeloid (84–94%) and lymphoid lineages (95–97%), and the frequency of CCR5⁺ T cells was reduced >100-fold. Mice engrafted with CCR5-edited HSCs were refractory to multiple challenges with CCR5-tropic HIV. Claiborne discussed the level of CCR5 ablation needed for HIV cure as a threshold of 72% CCR5 ablation.⁴⁰

Jose Martinez-Navio (Miller School of Medicine, University of Miami, USA) showed data on three SHIV-infected monkeys who, after receiving AAVs encoding a cocktail of broadly neutralising antibodies (bNAbs) have shown suppressed viral loads for years and appear functionally cured.⁴¹ Further attempts to create more such functional cures have been hampered by the generation of ADA responses. He showed that ADA levels correlated with antibody mutations, so the more divergent from germline, the more immunogenic. Four SHIV-infected macaques received recombinant AAVs expressing three bNAb (DH270, PCIN63 and DH511) that were naturally closer to the germline. High levels of two AAV-delivered antibodies were obtained in three of the four macaques. Sustained viral load suppression was achieved in one of those three monkeys with a second one suppressed for 20 weeks. The third monkey showed transient effects on viral load levels. The fourth monkey had low antibody levels due to ADA and little or no virologic suppression. Overall, the speaker concluded that the use of closer-to-germline bNAbs may be a viable strategy for avoiding ADA.⁴¹ Minimizing ADA responses will be crucial to make the AAV-delivery of antibodies a consistent and reliable approach against HIV.

The final talk of the session was from Helen Wu of the Oregon Health and Science University, who examined the impact of CCR5-wildtype allogeneic haematopoietic cell transplantation (HCT) on the SIV reservoir in ART-suppressed Mauritian cynomolgus macaque recipients.⁴² Her focus was the impact of graft-versus-host (GVH) immunity on the viral reservoir. Grafts collected from MHC-matched, SIV-naive donors were transplanted into four SIV+, ART-suppressed recipients following reduced intensity conditioning. Engraftment differed with some quickly reaching full donor chimerism in blood T cells and some with mixed chimerism. CD4⁺ T cell-associated SIV DNA levels decreased 10-fold within 30 days in all recipients. A strong inverse correlation between SIV DNA and donor chimerism was observed. GVH disease manifested in two animals. Allogeneic HCT-mediated viral reservoir clearance occurred concomitantly with full CD4⁺ T cell donor chimerism and was associated with GVHD, suggesting that alloreactivity may drive viral reservoir clearance in HIV cure.⁴²

7. Session 7: human studies

The seventh oral session of the 10th Workshop on HIV Persistence During Therapy focused on human studies. In the opening plenary, “Challenges and Advances in Identification and Immune Targeting of HIV-Infected Cells: Implications for Cure Studies,” Tim Henrich of the UCSF Division of Experimental Medicine discussed work towards these major challenges he defined as needed to induce immune control of HIV infection in the absence of ART. A major need was to define measurable biological markers in a person that would predict the ability to control HIV viremia without ART.

He provided an overview of previously published findings that correlate parameters such as lower CD4⁺ T cell nadir, higher pre-ART viral load, or earlier ART initiation with time to rebound after ART interruption, while CD4⁺ and CD8⁺ T cell responses, protective or neutral HLA type, and HLA-associated gag polymorphisms were associated with a lower viral setpoint after ATI. However, these associations were of varied strength in different studies and in different cohorts. No definitive markers or predictors of the goal of HIV RNA <200 copies/ml for an extended period of time (presumably years) have yet been identified.

Henrich noted challenges in the execution of clinical trials in this space. The Intact Proviral DNA Assay (IPDA), that has shown promise as a simplified measure of persistent infection, but cannot yet assess non-B clade infection, and may be confounded if lentiretroviral gene therapy are used as part of a cure or control strategy. The speaker pointed to the many challenges of studies where success is measured by an ATI, some of which might be addressed by the ongoing development of sensitive at-home assays to monitor HIV plasma viremia.

Henrich then turned to focus on HIV in tissue, and efforts to develop non-invasive imaging assays to measure the extent of HIV infection, immune response, and the distribution of therapeutics in tissues. Much of this is under study using 89Zr-labeled antibodies such as VRC-01 by PET scan. He shared images of people during an ATI where uptake of PET signal was seen in some tissues such as nasal-associated lymphoid tissue (NALT), gut, lymph nodes, spleen, bone marrow, and perhaps CNS prior to detectable viremia. In a primate study, a latency reversal strategy induced new PET signal in imaging.

Finally, he raised a number of current questions around the capacity to clear infected cells, such as whether there is sufficient HIV-1 env expression on infected cells under the cover of ART. Or if viral proteins being produced remain in close anatomical proximity to infected cells, or if HIV-1 env is processed and expressed as antigen on CD4⁺ T or other infected immune cells. He proposed that these and other questions around the mechanics of clearance of infected cells, at different stages of viral expression, might be addressed by some of the same imaging techniques.

Ming Lee, from the Fidler group at Imperial College London, presented very early results for 4 participants in the RIO trial (NCT04319367). In this randomized study,⁴³ eligible participants initiated ART in early HIV infection (<3 months after diagnosis) with no evidence of bNAb resistance by in-silico sensitivity testing. They received a single dose of 2 long-acting bNAbs, 3BNC117LS and 10-1074LS. There was an ATI, and after rebound, ART was restarted. ART restart criteria were six consecutive viral load measurements >1000 HIV-1 copies/mL or two measurements >100.000 HIV-1 copies/m BNAb serum concentrations were measured by validated anti-idiotype ELISA.

This report focused on 4 individuals. For these 4 participants, rapid viral rebound was observed during the first ATI (median 8.5 weeks after ATI, range 4–10). These participants were MSM, with a mean age of 39 years (range 26–52). When rebound after first ATI occured, they elected to restart ART and undergo bNAb re-administration (dose: 10 mg/kg (10-1074LS) and 30 mg/kg (3BNC117LS)). Here the median time to HIV viral load resuppression (<20 HIV-1 copies/mL) was 4.5 weeks (range 2–8 weeks). Then a second ATI was initiated six months after bNAb dosing. Two participants restarted ART after 26 and 22 weeks. Two participants have not yet met ART-restart criteria after 32 and 8 weeks. The second ATI exhibited lower peak rebound levels compared to the first (median peak viremia 3.8 log (10) vs 6.0 log 10 copies/ml, p=0.0025). Serum bNAb concentrations appeared to be adequate. The presenter suggested that there might have been a beneficial immunological effect and improved control of viremia after ATI, perhaps due to a vaccinal effect of these bnAbs. However, questioners raised the issue that prior studies of serial ATIs have shown some improved control following repeated ATIs, simply from more rapid recall of prior immune responses. Further studies are underway.

Leila Giron, of the Abdel-Mohsen group at the Wistar Institute presented work⁴⁴ that evolved from the ACTG study 5345, an observational study of ATI. Forty-five participants who had initiated ART during chronic (n = 33) or early (n = 12) infection underwent ATI. Viral rebound (≥1000 HIV-1 copies/ml) was seen at a median of 3 weeks. The group has for some time being studying the hypothesis that plasma metabolic and glycomic analytes reflected ongoing immune biology and responses during ART, and may be used as predictors of viral control. Pre-ATI plasma glycans and metabolites were measured by lectin microarray and mass spectrometry. Correlations were examined between pre-ATI glycans or metabolites and time to viral rebound.

As in prior cohorts the plasma n-glycan FA2G2S1 and the binding glycan ACG/rACG were found to correlate with shorter and faster time to rebound, respectively. Overall, multiple sialylated glycans correlated with slow time-to-viral-rebound, whereas high levels of the plasma markers of tryptophan catabolism were associated with fast time-to-viral-rebound. The anti-oxidant l-Ergothioneine was associated with slow time-to-viral-rebound. Markers of rapid rebound also correlated in a consistent direction with higher CD8⁺ T cell activation (CD38⁺ HLADR⁺ CD8⁺ T cells), fewer effector CD8⁺ T cells, more TIM3+ T cells, and higher plasma sCD163 levels. Eventually such plasma markers might help predict the effect of therapies to improve viral control and prolong aviremia after ATI.

The meeting was entranced by the presentation of Josephine Nabukenya of the MUJHU Research Collaboration, Uganda⁴⁵ (Fig. 1). She described her project's effort to simplify HIV cure research terms and findings into easily accessible community language. The team recruited adolescents and youth peers from 10 health facilities for an initial workshop to identify scientific terms used in HIV cure science, and adopt a community lay language. During COVID, further work was done via a WhatsApp group, and dissemination material develop in the form of illustrations and animated cartoon-like characters. These were shared via Facebook, Instagram and YouTube targeting adolescents and young people. The resulting platform known as #seriesofjojo is now available to disseminate HIV cure information via youth-friendly animations and illustrations.

Fig. 1.

A way to disseminate HIV Cure information in a community language.

Maria Puertas of the Martinez-Picado group at the AIDS Research Institute IrsiCaixa, Barcelona, Spain, reported observations made using the VIPSPOT assay within the eCLEAR trial.⁴⁶ It is (NCT03041012) a phase 1 b/2a, open-label, multicenter, randomized controlled trial among people starting ART randomly allocated to one of 4 treatment groups: ART only, ART and the bnAb 3BNC117, ART and the LRA romidepsin, or ART plus bnAB plus romidepsin. The bnAb was given after 1 and 3 weeks of ART; Romidepsin was given after 10, 17, and 14 days of ART. Thus far in previously published results, there was a somewhat accelerated viral decay with the bnAb, and a somewhat longer time to rebound in participants with a bnAb sensitive virus receiving the bnAb, however the contribution of romidepsin was unclear. Here, the investigators assessed the ability of the VIP-SPOT assay to detect changes in the productive HIV reservoir. This was of interest because it could be that the bnAb had an effect on productively infected cells in this study, without clearance of the truly latent reservoir.

Cryopreserved PBMCs from most of the study participants (50 or 60) were used for the VIP-SPOT assay. VIP-SPOT data was compared to plasma viremia, HIV viral clade from sequencing, proviral HIV DNA, and intact provirus measured by an IPDA-like duplexed ddPCR (3dPCR) assay. The VIP-SPOT assay captures cells on a plate coated with a mix of antibodies that activated cells (anti CD3/CD28) and capture viral capsid antigen (anti-p24). Spots of antigen, presumably marking single infected cells, are detected as spots in the assay with another luminescent anti-p24 Ab.

Puertas found that the frequency of cells able to reactivate and produce viral protein detected in VIP-SPOT is 1000-fold lower than total HIV DNA, and that even most intact proviruses are not able to reactivate productively in this assay. However, there was a correlation between the VIP-SPOT assay and other virological parameters at baseline. About half of the participants had non-B clade infection, but the VIP-SPOT appeared to detect productively infected cells regardless of clade. Overall, when assessed after one year of ART, there was substantial, and apparently similar, decline in VIP-SPOT positivity across all the study groups, regardless of intervention. It is possible that there was a small decay in the durably latent reservoir that was not substantial enough, or uniform enough in this study, to be detected by the VIP-SPOT, especially in the background of substantial ART-driven decay in the first year of therapy. There was a non-significant trend towards greater undetectability of VIP-SPOT in participants receiving 3BNC117, especially in individuals whose pre-ART plasma viruses were sensitive to the antibody.

Liliana Perez led work in the Bortiz lab at the Vaccine Research Centre, USA that described HIV clonal dynamics⁴⁷ in ART-treated participants who received the PD-1 inhibitor pembrolizumab (clinical trial NCT03239899). Six people with chronic HIV infection, on ART and with CD4 T cell counts >350 cells/μL received one infusion of 200 mg pembrolizumab. PBMCs were collected at baseline, week 3 and 24 post-infusion. CD4⁺ T cells were FACS-sorted into naïve, central/transitional memory, and effector memory (EM) subsets. HIV DNA was quantified by limiting dilution PCR of env and Sanger sequencing. HIV RNAs were quantified by ddPCR. Changes in gene expression patterns and T-cell receptor (TCR) repertoires were evaluated. Bulk integration site analysis (ISA) in CD4⁺ T cells was used to assess similarity and clonality of all integration sites (IS) within each person across timepoints. Intact HIV proviruses in CD4⁺ T cells were enumerated with an intact proviral DNA assay.

Overall, the investigators found that there was increased cell proliferation in effector memory CD4⁺ T cells at week 3 following PD-1 blockade, with a shift in distribution of HIV DNA postivie cells to the EM pool. PD-1 blockade induced HIV transcriptional initiation, possibly due to latency reversal or proliferation of cells that are “initiation-competent.” A reduced diversity of cell-associated HIV DNA in EM CD4⁺ T cells, could be consistent with the proliferation of some clones and clearance of others. In some cases, changes in the population of unique integration sites of HIV-infected CD4⁺ T cells was seen, consistent with clonal expansion and after PD-1 blockade. There was, however, no clear change in the number of intact proviruses after anti-PD-1 treatment.

Finally, Tine Struyve of the Vandekerckhove group in Ghent University Belgium, presented an in-depth assessment of the total and inducible viral reservoir after 1 year of ART.⁴⁸ Near full-length (NFL) proviral sequences and integration sites were analyzed in 9 men with subtype B HIV infection treated during acute infection (Fiebig II-III: n=7; Fiebig IV: n=1; Fiebig V: n=1) and who had received ART for 0.49–1.93 years, using matched integration site and proviral sequencing (MIP-Seq), and near-full length sequencing (NFL). Multiplex digital PCR was used to assess total and intact HIV proviral DNA (IPDA). The translation-competent reservoir was assessed by HIV-Flow, and finally an as-yet unpublished multiplex IPDA-type technique called the Rainbow assay (which amplifies provirus using primers of 5 regions, the RU5, the PSI and env region from IPDA and 2 regions in gag and pol from the Q4PCR assay was used to quantitate total and intact HIV DNA.

Not unexpectedly, despite early ART initiation, total HIV DNA remained detectable, but was significantly lower in individuals who initiated therapy during acute infection compared to chronic infection. The IPDA and Rainbow assays gave similar results, with the number of intact proviruses 4 times lower than total HIV DNA. In a combined assessment of all amplified proviral genomes, after a median of 0.96 years of ART, 9% of proviruses were intact. The fraction of intact and hypermutated proviruses was higher for early treated individuals than a comparable chronically infected cohort.

The majority of integration sites were unique, with small clones retrieved in 2 out of the 8 individuals. Finally, to assess the inducibility of latent proviruses, using a flow cytometry-based method, enriched CD4⁺ T cells were stimulated for 24 h, and surface memory markers and intracellular p24 were assessed. While cells were minimally reactive to PMA/ionomycin, a combination of PMA and an HIV tat mimetic protein induced detectable p24 expression in 0.4–20 cells/million. The positive p24 cells tended to be enriched in the TEM subset. When comparing the phenotype of p24⁺ cells between the acute and chronic cohorts, a higher frequency of p24⁺ cells were in the naïve population in early treated individuals compared to the chronically treated individuals.

Struyve concluded that, as expected, early ART initiation did not prevent the establishment of the intact viral reservoir and in this cohort, after 1 year of treatment, the contribution of clonal expansion to the persistence of the viral reservoir was minimal. This conclusion will have to be re-examined over longer periods of ART.

8. Session 8: antibody & immune based therapies

In this session presenters shared results from clinical and non-clinical studies of immune-based modalities and discussed implications of presented data to future cure studies.

Marina Caskey from the Rockfeller University, NY, USA, gave an overview of studies with bNAbs.⁴⁹ She discussed studies with 3BNC117 (CD4bs) and 10–1074 (V3 loop) where bNAbs transiently reduced viremia and maintained viral suppression after ART interruption in PWH harboring sensitive viruses.⁵⁰ She reviewed available data on potential bNAb effects on immune responses and reservoir. Nonclinical studies showed that bNAbs accelerate clearance of HIV-infected cells in a hu-mouse model, and induce virologic control in a subset of SHIV-infected macaques that is at least in part mediated by CD8⁺ T cells,51, 52, 53 two clinical studies showed that bNAbs enhance gag-specific T cell responses when administered during ART interruption or during viremia at ART initiation.^53,54 Lastly, 3BNC117 plus 10–1074 immunotherapy was associated with changes in intact proviral reservoir size, although magnitude of changes varied and were not associated with virological control.⁵⁵ This data suggests that bNAbs can modulate immune responses and may impact HIV persistence but additional studies are needed to confirm these initial observations.

Michael Freeman from Case Western University, USA discussed results from a study of interleukin-2 as a LRA.⁵⁶ Nine PWH on ART received one or two 4-day cycles of IL-2. IL-2 led to T and NK cell activation, increased cell cycling, loss of Bcl-2 expression, and enhanced cytolytic function; it also promoted CD4⁺ Treg phenotype. It led to measurable plasma viremia, and small reductions in proviral reservoir by IPDA. Despite the immunologic and virologic effects, the study was terminated because of toxicity in three participants: one with capillary leak syndrome and two with hypothyroidism. These results show the potential of IL-2 as a LRA, but strategies to deliver IL-2 directly to target cells are likely needed to circumvent toxicity risks.

Beatriz Mothe from Hospital GermansvTrias I Pujol, Badalona, Spain, discussed results from the AELIX-002 therapeutic vaccine trial, which evaluated the safety, immunogenicity and impact on viral rebound of a combination of DNA.HTI, MVA.HTI and ChAdOx1.HTI vaccines in 45 early-treated PWH.^57,58 The vaccine regimen was well tolerated and highly immunogenic, inducing high frequency of HTI-specific polyfunctional CD8⁺ and CD4⁺ T cells. During ATI, viral rebound occurred in all 41 participants. In a post-hoc analysis, participants without beneficial HLA class I alleles (32 of 41), 1 (8%) placebo and 8 (40%) vaccine recipients remained off ART for 22 weeks; with 1 placebo and 5 vaccinees maintaining viral load <2000 HIV-1 copies/ml. Importantly, time off ART and viral load level at the end of ATI were associated with the frequency of CD8⁺ T cells expressing GzmB, while reservoir measurements were not associated with ATI outcomes.

H. King, from MHRP, USA presented results from a study in SIVmac251-infected and ART-treated rhesus macaques that evaluated a combination of TLR agonists and SIV mAbs.⁵⁹ Animals (8 per group) received: (1) 2BXy (TLR7/8); (2) CpG (TLR9); (3) LPS (TLR4); or (4) intravenous BCG (TLR2/4/9) in combination with two anti-SIV mAbs (ITS103, I109) during suppressive ART; (5) SIV mAbs only; or (6) ART only. ART was discontinued 24 weeks later. T and NK cell immune activation occurred, however viral blips were not observed during TLR dosing and about 1/3 of animals in the TLR groups developed anti-drug antibody responses. No changes in SIV-specific T cell responses were observed. During ATI, all animals experienced return of viremia within 4 weeks, without differences in time to rebound or setpoint viremia across different groups. These results are in contrast with previous studies with HIV mAbs in a SHIV model, however the mechanisms underlying the difference in outcomes are unclear, which highlights the need to further investigate immunologic and virologic features associated with post-ATI control.

Elena Martinelli, Northwestern University, USA presented results from a study in non-human primates that evaluated TGF-beta blockade as a mechanism to reactivate HIV from latency.⁶⁰ Seven SIVmac239M2-infected macaques, ART-treated for 7 months received four 2-week cycles of galunisertib (Gal), a TGF-beta receptor inhibitor. Gal induced viremia in 5 out of 7 animals. PET/CT imaging showed SIV reactivation in gut tissues. Gal stimulated an effector phenotype in T cells, and these changes correlated with decreases in cell-associated viral DNA in PBMCs and gut and lymphoid tissues by the end of the 4th cycle (IPDA measurements are underway). The presented results suggest that TGF-beta blockade is a potential strategy to perturb HIV persistence in vivo.

Katharine Bar University of Pennsylvania, shared results from a study in SHIV-D-infected rhesus macaques evaluating autologous neutralizing antibody responses (anAbs) kinetics during ART and bNAb therapy.⁶¹ Eighteen SHIV-D-infected macaques were started on ART 3 months post-infection and discontinued ART 6 months later; 9 animals received VRC07523.LS at the start of ATI. Control animals were treated with ART after rebound and underwent a second ATI in the presence of VRC01 or VRC07523LS; viral rebound peaks during the second ATI were lower. Prior to ART start, 56% of macaques had anAb responses against SHIV.D; plasma and purified IgG responses were tightly correlated and remaining stable during ART. In contrast, 8 weeks post-rebound, plasma anAb responses rose ∼1000- fold against both SHIV.D and rebound Envs; with greater magnitude against SHIV.D, indicating antibody imprinting. Rebound virus escaped baseline anAb responses, suggesting ongoing humoral immune pressure at ATI. This study shows the utility of SHIV infection models to study antibody and viral kinetics, and the contribution of anAb responses to viral restriction during bNAb therapy.

Derek O'Hagan Uf Scripps Biomedical Research, USA showed results of a study of AAV delivery of eCD4-Ig, a potent HIV entry inhibitor, in SHIV-AD8 infected macaques.⁶² Ten to 14 weeks after onset of SHIV-AD8 viremia, 5 animals were administered AAV8- and AAV1-eCD4-Ig. Two of 5 treated animals developed persistent serum levels of eCD4-Ig (70–100 mcg/ml) but experienced only partial or transient viremia reductions. The R315G, A436 T, G471E mutations were selected, conferred resistance to neutralization and were associated with fitness cost. These appear to confer eCD4-Ig resistance in part by taking advantage of a single amino acid difference between rhesus CD4 and eCD4-Ig (I39 N) and by modulating CCR5 engagement, potentially allowing the use of alternate coreceptors. These results provide information on evolutionary pathways that remain open to env in the presence of bNAbs mimicking major entry receptors.

Declaration of competing interest

D. Kuritzkes has received research support and/or consulting honoraria from AbbVie, Excision Biotherapeutics, Gilead, GlaxoSmithKline, Janssen, Merck, Roche and ViiV, and speaker fees from Gilead, Janssen, Novartis and Merck. The other authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.