Introduction: Themes in Basic Virology of HIV
An overarching theme across the basic science sessions was the central role of the HIV capsid as a dynamic regulatory structure that coordinates various steps of viral replication. Once thought to function primarily as a structural shell protecting the viral genome, the capsid is now recognized as a molecular hub controlling reverse transcription, nuclear trafficking, immune evasion, and integration targeting. Several presentations examined the architecture of the capsid lattice and the mechanisms with which viral proteins and host cofactors interact with the interior and exterior surfaces of the capsid. These studies reveal that the capsid not only organizes viral replication complexes but also helps shield viral nucleic acids from innate immune sensors. Detailed insight was provided on viral RNA organization and packaging inside the capsid. New structural and biochemical studies demonstrated that viral integrase plays an unexpected role in tethering genomic RNA to the interior of the capsid lattice. These findings suggest that the architecture of the viral core is coordinated through intricate interactions between capsid protein, viral RNA, and enzymatic components of the replication complex. Understanding how ribonucleoprotein complexes are assembled within the capsid provides important insight into the earliest stages of infection and may reveal new targets for antiviral drug development.
Important research revealed the interaction between HIV replication and innate immune sensing pathways. Several studies explored how viral RNA and reverse transcription products are detected by host-pattern recognition receptors such as melanoma differentiation-associated protein 5 (MDA5) and cyclic GMP-AMP synthase (cGAS). These findings illustrate how subtle changes in viral RNA structure, capsid stability, or nuclear trafficking can determine whether HIV replication proceeds undetected or triggers potent antiviral responses. Increasing evidence suggests that the virus has evolved sophisticated mechanisms to manipulate these pathways, allowing productive replication and minimizing immune detection.
Several presentations highlighted the power of genomewide functional screens to identify host factors that regulate HIV infection. Clustered regularly interspaced short palindromic repeats (CRISPR)-based screening approaches now allow investigators to systematically identify antiviral restriction factors and host proteins that facilitate viral replication. These studies are beginning to map the full landscape of HIV-host interactions in primary CD4+ T cells, revealing unexpected cellular pathways that influence viral entry, replication, and immune evasion. Together, the presentations reviewed here highlight the remarkable progress being made in dissecting the molecular and cellular mechanisms that govern HIV replication and persistence. The following sections review selected abstracts that exemplify these themes.
HIV Replication and Persistence
The HIV capsid plays a central role in viral replication by forming a conical core that houses the viral genome and replication enzymes. During infection, the capsid coordinates many early steps of the viral life cycle, including reverse transcription, nuclear trafficking, and integration. Because of these essential functions, the capsid has emerged as a promising target for antiviral drug development. Sundquist (Abstract 14) provided a comprehensive overview of the structural biology of the HIV capsid and described the decades-long research effort that ultimately led to the development of lenacapavir, the first US Food and Drug Administration-approved capsid inhibitor. Structural studies have revealed that the mature HIV capsid is composed of approximately 250 hexamers and 12 pentamers of the capsid protein, arranged in a fullerene-like lattice that forms the characteristic conical core. Highresolution cryoelectron microscopy has provided detailed insight into the interfaces that stabilize this lattice and enable capsid assembly.
Biochemical analyses demonstrate that capsid interactions with host proteins such as cyclophilin A, nucleoporins, and cleavage and polyadenylation specificity factor 6 (CPSF6) regulate capsid stability and intracellular trafficking. These interactions help guide the viral replication complex toward the nucleus and protect viral nucleic acids from innate immune sensors.
Building on this structural understanding, researchers identified small molecules capable of binding specific pockets within the capsid lattice. Optimization of these compounds ultimately led to the development of lenacapavir, a long-acting capsid inhibitor that disrupts numerous stages of the viral replication cycle. Lenacapavir binds to the capsid protein at a site involved in host factor interactions, thereby altering capsid stability and interfering with nuclear import.
Clinical trials have demonstrated that lenacapavir provides potent antiviral activity and can be administered as a twice-yearly subcutaneous injection, representing a major advance in long-acting HIV therapy and prevention. The development of lenacapavir represents a remarkable example of translational research bridging basic structural biology and clinical drug development. Nearly 3 decades of collaborative work between academic laboratories and industry partners helped to fully elucidate capsid structure and exploit it as a therapeutic target.
Importantly, capsid inhibitors may provide new opportunities for long-acting treatment and prevention strategies. By targeting a highly conserved structural component of the virus, these drugs may also retain activity against viruses resistant to other antiretroviral classes. The HIV capsid encloses the viral RNA genome and replication enzymes within a conical core. However, because the mature capsid occupies only a portion of the virion interior, the mechanism by which viral RNA remains confined within the capsid during maturation has been unclear. Previous work suggested that viral integrase might interact with RNA, but the structural basis of this interaction had not been defined. Using cryoelectron microscopy, Cherepanov (Abstract 15) described studies to determine how viral RNA is sequestered within the viral capsid. These studies revealed filamentous assemblies composed of integrase octamers, each formed by 2 asymmetric tetramers associated with RNA molecules. The spacing and orientation of these integrase assemblies closely matched the geometry of the capsid lattice. Imaging of purified HIV cores demonstrated that integrase filaments attach to the luminal surface of the capsid shell, forming a tethering network that anchors viral RNA within the core. Mutations disrupting integrase-RNA interactions produced aberrant virions in which viral RNA was mislocalized outside the capsid, resulting in loss of infectivity. These findings reveal an unexpected structural relationship between integrase and the capsid lattice.
Integrase appears to function as a molecular tether that secures viral RNA to the interior of the capsid, ensuring proper genome encapsidation during viral maturation. This finding also provides a potential explanation for the antiviral effects of allosteric integrase inhibitors, which induce eccentric virion formation by disrupting integrase-RNA interactions. Future work will explore whether targeting this structural interaction could provide new strategies for antiviral therapy.
Innate immune sensors detect viral nucleic acids and trigger antiviral responses. However, HIV replication often proceeds without strong activation of innate immunity. This suggests that the virus has evolved mechanisms to shield its nucleic acids from immune detection. Kutluay (Abstract 16) described studies highlighting that capsid lattice stability is a crucial determinant of innate immune sensing. Destabilizing capsid mutations increased detection of reverse transcription products by the cGAS-stimulator of interferon genes (STING) pathway. Surprisingly, hyperstable capsids also enhanced immune sensing due to altered nuclear trafficking and reverse transcription dynamics. These results indicate that HIV replication requires a precise balance between capsid stability, reverse transcription, and nuclear import. Disrupting this balance exposes viral DNA to innate immune sensors. A central question in HIV biology is how the components of the incoming viral core within the nucleus shape the final integration event. Although HIV has long been known to prefer integration into transcriptionally active, gene-rich chromatin, more recent work has shifted attention from integration as a purely biochemical process to integration as a highly spatial one. Intact or largely intact HIV cores can enter the nucleus, interact with host nuclear factors, and traffic to specific subnuclear compartments before uncoating. Nuclear speckles, which are membrane-free compartments enriched in splicing and RNA-processing proteins, have emerged as favored sites for incoming HIV complexes. Because speckle-associated chromatin domains are generally transcriptionally active, they offer an attractive explanation for HIV's integration bias. However, it has remained unclear whether the site of uncoating itself directly determines where integration occurs, or whether uncoating and integration are merely correlated downstream events.
Pathak (Abstract 17) described experiments to assess the position of nuclear HIV cores, the timing and location of uncoating, and the final sites of proviral integration. The investigation further examined whether capsid inhibitors such as PF74 and lenacapavir could perturb this spatial choreography after nuclear entry. Using fluorescently labeled HIV-1 cores together with nuclear speckle markers, Pathak showed that most nuclear viral cores localized to nuclear speckles and that uncoating occurred predominantly within these structures. This finding strengthens the emerging model that uncoating is not a diffuse nuclear event but instead happens in a specialized nuclear microenvironment. The work further showed that binding of CPSF6 to capsid is necessary to retain HIV cores within nuclear speckles. When infected cells were treated with PF74 or lenacapavir after nuclear entry, viral cores rapidly lost their interactions with CPSF6 and exited nuclear speckles. These data connect capsid integrity, CPSF6-dependent speckle retention, uncoating position, and integration site selection into a unified pathway, and suggest that the interval between uncoating and integration is very short. Once the core was displaced from speckles and disrupted, integration followed in the altered local environment rather than after extensive diffusion of the viral DNA. CPSF6-mediated retention of viral cores in speckles appears essential for normal integration targeting, whereas postentry treatment with capsid inhibitors disrupts this localization, shifts uncoating outside speckles, and redirects integration toward less transcriptionally active chromatin regions. Future studies should determine whether different cell types, particularly macrophages versus activated CD4+ T cells, display the same speckle-centered uncoating pathway. It will also be important to define how altered integration targeting affects long-term proviral expression, latency establishment, and immune visibility.
The mature HIV-1 capsid core is a sophisticated macromolecular assembly. How viral ribonucleoprotein complexes are physically encapsulated within the mature capsid during maturation is an important but not fully understood question. The mature core contains the viral RNA genome, nucleocapsid, reverse transcriptase, and integrase, all arranged inside the capsid shell in a way that preserves infectivity. Since the conical core is assembled during proteolytic maturation and occupies a defined internal volume, there must be an ordered mechanism that couples capsid formation to genome retention. Prior work suggested that integrase plays a noncatalytic role in proper virion morphogenesis and that disruption of integrase-RNA interactions leads to eccentric, noninfectious particles. However, the structural basis for that process and the role of capsid itself in binding viral RNA were incompletely understood.
Briganti (Abstract 100) described attempts to define the architecture of the HIV-1 core and to identify the molecular determinants that allow viral ribonucleoprotein complexes to be packaged inside it. By combining single-particle cryoelectron microscopy, cross-linking and immunoprecipitation (CLIP) assays, and innovative biochemical assays using capsid-like particles, the investigators aimed to establish a mechanistic model of encapsidation. The structural analysis of native HIV-1 cores identified bridging nucleoprotein interactions involving capsid hexamers, viral RNA, and integrase tetramers on the luminal side of the capsid. Biochemical studies further showed that capsid hexamers, but not capsid monomers, specifically bind viral RNA segments. This indicates that RNA recognition is an emergent property of assembled capsid lattice units rather than of isolated capsid protein, reinforcing the idea that mature capsid architecture generates new functional surfaces.
CLIP experiments confirmed that capsid binds viral RNA within mature virions, and substitutions in capsid residues that impaired interactions with viral RNA produced aberrant noninfectious virions in which viral ribonucleoprotein complexes were mislocalized outside the translucent capsid shell. Thus, capsid-RNA interaction is not incidental but functionally required for proper core organization. These results advance a structural mechanism for one of the least understood steps in HIV maturation: the encapsidation of the viral ribonucleoprotein complex. The finding that capsid hexamers bind viral RNA inside mature virions adds a new functional dimension to the capsid lattice. Rather than acting solely as a shell around preorganized contents, the capsid appears to participate directly in constructing and stabilizing the architecture of the core interior. These results also suggest that noncatalytic integrase functions, especially those related to RNA tethering and encapsidation, are equally fundamental to infectivity and may be pharmacologically targetable.
An intriguing aspect of HIV biology is that viral RNA must retain features needed for genome packaging, translation, dimerization, and export, yet those same unusual features risk recognition by host innate immune sensors. HIV-1 partially solves this by exploiting transcription start site heterogeneity to generate distinct unspliced RNAs with 1 or 3 guanosines at the 5’ end (cap1G RNA and cap3G RNA, respectively). Hughes (Abstract 101) highlighted previous work from her group and others showing that HIV-1 unspliced RNA exported through the Rev-chromosome region maintenance 1 (Rev-CRM1) pathway can activate MDA5-dependent innate immune responses in macrophages. What had not been clear was why HIV universally conserves heterogeneous transcription start site usage and whether different unspliced RNA species differed intrinsically in immunogenicity. In primary monocytederived macrophages, the investigators found that infection with cap1G-only virus induced high levels of type I interferon, whereas infection with cap3G-only virus produced little to no detectable interferon response. The interferon response induced by cap1G virus was blocked by inhibitors of reverse transcription, integrase, and Rev-CRM1-mediated export, indicating that productive early replication and canonical viral RNA export were required.
Knockdown studies showed that MDA5 and mitochondrial antiviral signaling (MAVS) protein, but not retinoic acid-inducible gene-I (RIG-I) protein, were essential for cap1G-driven interferon secretion. Consistent with this, MDA5 immunoprecipitated HIV-1 unspliced RNAs during wild-type and cap1G-only infection, but not during cap3G-only infection. These results provide direct evidence that cap1G RNA is the immunostimulatory viral RNA species recognized by MDA5. The investigators then went further to dissect mechanism. Viruses engineered to express cap1G-like leader structures with exposed caps behaved differently from native cap1G RNAs, showing that cap sequestration versus exposure is a crucial variable. In addition, replacing the Rev-CRM1 export pathway with a constitutive transport element that drives nuclear RNA export factor 1 (NXF1)-dependent export altered immunogenicity, demonstrating that the route of nuclear export is also a determinant of sensing. Together, the data suggest that MDA5 recognition depends on a combination of RNA start site, 5’ leader conformation, and export pathway. This is also consistent with published work from the same group showing that MDA5 sensing is tightly governed by transcription start site and 5’ leader structure. In summary, these studies demonstrate that cap1G HIV-1 unspliced RNA is the principal driver of MDA5-MAVS-dependent type I interferon responses in infected macrophages, whereas cap3G RNA is largely nonimmunogenic. The immunogenicity of HIV unspliced RNA depends on transcription start site choice, 5’ leader structure, cap exposure, and use of the Rev-CRM1 export pathway.
Transmitter-founder and rebound HIV-1 viruses frequently display enhanced replication fitness and relative resistance to type I interferon, yet classical sequence analysis has not explained these phenotypes fully. This has prompted interest in non-sequence-based viral determinants, including epitranscriptomic RNA modifications. N6-methyladenosine, or m6A, is a reversible chemical modification that influences RNA stability, translation, structure, and recognition by host proteins. HIV RNA is known to carry m6A marks, but whether these modifications directly alter viral fitness and interferon sensitivity has remained uncertain. Ye and colleagues (Abstract 102) tested the hypothesis that m6A-modified HIV RNAs evade cellular RNA sensing and thereby enhance replication fitness and interferon resistance, which might point to a mechanism by which viral phenotype could change without fixed coding mutations. The investigators experimentally altered the m6A status of HIV-1 RNA by treating producer cells with inhibitors of the m6A writer machinery, including methyltransferase-like 3 (METTL3)/ METTL14 inhibitors, or with inhibitors of the m6A eraser fat mass and obesity (FTO)-related protein. Virions with reduced or increased m6A modification were then used to infect activated and polarized human T-helper 17 cells in the presence or absence of type I interferon. Viruses with depleted m6A modifications showed lower HIV RNA and p24 expression after infection and were markedly more sensitive to interferon. In contrast, hypermethylated viruses displayed enhanced resistance to type I interferon. These data support the central premise that m6A on viral RNA contributes to replication fitness and interferon escape.
The investigators extended these findings to infected primary cells. In acutely infected Th17 cells and in reactivated memory CD4+ T cells from people with HIV (PWH), type I interferon reduced viral RNA and protein levels and decreased the number of cells containing intact proviruses. These effects were amplified when m6A levels were reduced pharmacologically. Single-cell RNA sequencing and protein analyses showed that lowering m6A enhanced interferon-induced expression of several antiviral genes, including interferon-induced protein with tetratricopeptide repeats 1 (IFIT1), IFIT5, IFITM1, and tripartite motif 13 (TRIM13). The combined treatment also perturbed cell-cycle regulatory pathways, reduced cell viability, and appeared to favor elimination of cells harboring intact proviruses. This study is novel because it reframes interferon resistance as an epitranscriptomic property of HIV, not just a genetic one, and suggests that the virus can gain phenotypic advantages through chemical decoration of its RNA genome, potentially shaping early infection, transmission, rebound, and reservoir persistence. Future studies may evaluate whether m6A-targeting strategies can be used safely in vivo.
A defining feature of HIV pathogenesis is the constant evolutionary contest between viral accessory proteins and host restriction factors. Although many classical restriction factors have already been identified, it is increasingly clear that the known set is incomplete. Standard CRISPR screens have been useful for discovering host factors, but many approaches do not recapitulate the full viral replication cycle or the dynamic selective pressures that shape replication in infected cells. Gosálbez and colleagues (Abstract 104) previously developed an HIV-guided CRISPR platform in which replication-competent HIV itself carries single-guide RNAs targeting host genes. In Cas9-expressing cells, viruses that knock out antiviral genes gain a replication advantage, allowing enriched single-guide RNAs to reveal the targeted restriction factors. Related prior work established the sensitivity of this strategy for discovering physiologically relevant defense factors.
Gosálbez expanded this concept to a genome-wide scale and incorporated a parallel viral protein U (Vpu)-deficient HIV-1 screen to identify factors specifically antagonized by Vpu. The rationale was to build a more complete, unbiased map of antiviral genes across the viral life cycle and to illuminate the immune evasion pathways governed by Vpu. The investigators generated replication-competent HIV-1 constructs encoding more than 77,000 distinct single-guide RNAs for genome-wide targeting and passaged these libraries in Cas9-expressing CEM-M7 cells with or without interferon-β. This yielded a large-scale selection system in which enrichment of specific single-guide RNAs in viral supernatants marked genes whose loss benefited viral replication. The screen identified numerous candidate antiviral factors. Among the strongest signals was a cluster of genes involved in DNA replication and repair, including deoxyribonuclease 1 (DNaseL1), G protein subunit beta 1 like (GNB1L), MPN domain-containing (MPND), Nth like DNA glycosylase 1 (NTHL1), DNA polymerase epsilon subunit 4 (POLE4), and RecQ-mediated genome instability 2 (RMI2). Validation in primary CD4+ T cells showed that knockout of these genes increased viral yield approximately 2-to 3-fold, supporting genuine antiviral activity. This is particularly interesting because it implicates DNA maintenance and genome integrity pathways in the restriction of HIV, areas not usually considered central in canonical HIV defense models. The parallel Vpu-deficient screen yielded more than 50 factors that were specifically enriched when Vpu was absent, including ADAM metallopeptidase with thrombospondin type 1 motif 16 (ADAMTS16), adenosine diphosphate-ribosylation factor interacting protein 1 (ARFIP1), carcinoembryonic antigen-related cell adhesion molecule 5 (CEACAM5), and RAB7A. Their selective enrichment suggests that these genes encode antiviral activities normally counteracted by Vpu. Validation in primary CD4+ T cells was ongoing, but the design itself provides a powerful route to identify Vpu-sensitive immune evasion targets.
Systematic Discovery of Pro-and Anti-HIV Host Factors in Primary Human CD4+ T Cells
Although HIV research has identified many host proteins required for or restrictive to infection, the landscape remains incomplete, particularly in primary human CD4+ T cells, which are the principal targets of HIV in vivo. Prior work has relied on transformed cell lines or focused on candidate genes, leaving open the possibility that key regulators in authentic primary T cells have been missed. Technical barriers, especially the difficulty of performing genome-wide perturbation screens at scale in primary T cells, have limited progress.
Dugan and colleagues (Abstract 105) addressed that challenge by applying orthogonal genome-wide CRISPR activation and CRISPR knockout screens directly in primary human CD4+ T cells. The investigators also designed the work to dissect whether host factors act at entry, coreceptor expression, T-cell activation, or postentry stages. The screens nominated hundreds of candidate antiviral and proviral factors. Among the most notable antiviral hits identified by CRISPR activation was peptidase inhibitor 16 (PI16). Overexpression of PI16 in CD4+ T cells led to potent inhibition of HIV entry. Mechanistically, PI16 physically interacted with CD4+ and with numerous HIV host factors involved in fusion, indicating that it acts at or near the plasma membrane entry interface.
A second hit was peptidylprolyl isomerase D (PPID), also known as cyclophilin 40 (Cyp40). PPID is a homologue of cyclophilin A, but in contrast with CypA, which helps shield HIV cores from TRIM5α and facilitates early postentry events, PPID acted as a potent antiviral factor. The study showed that PPID inhibited nuclear import of the viral core. Its antiviral activity depended on residues G89 and P90 in precursor 55 Gag (Pr55Gag) for incorporation into virions, indicating that PPID is packaged into viral particles and exerts its effect from within or immediately after virus entry. The investigators further demonstrated that the tetratricopeptide repeat (TPR) motifs in the PPID C-terminal domain are sufficient to convert CypA into an antiviral factor. In addition, a TPR-domain site required for interaction with translocase of outer mitochondrial membrane 70 (TOM70) was essential for PPID-mediated restriction, suggesting that TOM70 recruitment contributes to the antiviral mechanism. The experimental design also included challenges with C-X-C chemokine receptor type 4 (CXCR4)-and C-C chemokine receptor type 5 (CCR5)-tropic viruses, as well as VSV-G pseudotyped viruses, allowing the team to infer whether hits were entry related or post entry in function. Secondary pooled and arrayed validation experiments further linked candidate factors to the regulation of receptor expression and T-cell activation states. Future work should determine how broadly these findings generalize across naive, memory, activated, and tissue-resident CD4+ T-cell subsets. It will also be important to define the biochemical mechanism by which PPID blocks nuclear import and how TOM70 participates in that process. For PI16, the next key question will be whether its antiviral activity can be enhanced pharmacologically or mimicked therapeutically.
Viral Reservoir and HIV Cure Research
The viral reservoir and cure sessions highlighted how the HIV cure field is evolving from descriptive studies of persistence toward increasingly mechanistic, quantitative, and intervention-oriented science. A central message across the presentations was that HIV persistence during suppressive antiretroviral therapy (ART) is not explained by a single process, cell type, or molecular mechanism. Rather, the reservoir is a dynamic and heterogeneous biologic system composed of infected cell populations that vary in proviral integrity, inducibility, proliferative behavior, anatomic location, transcriptional activity, and susceptibility to immune clearance. This complexity has major implications for cure strategies, because no single intervention is likely to be effective across all reservoir states.
A dominant theme was the need to understand reservoir heterogeneity at many levels. At the molecular level, only a minority of proviruses in treated individuals is intact and replication competent, yet defective proviruses are not biologically silent and can still contribute to chronic immune stimulation and antigen expression. At the cellular level, infected cells differ in phenotype, proliferative history, differentiation state, and transcriptional responsiveness. At the clonal level, long-term ART does not simply “freeze” the reservoir in place; instead, individual infected clones may decay, persist, plateau, or expand over time, sometimes over decades. At the tissue and systems level, the reservoir behaves as an ecologic population shaped by cell survival, homeostatic proliferation, immune pressure, and therapeutic intervention.
Another major theme was the increasing importance of single-cell and systems approaches for dissecting latency and reservoir behavior. Genome-wide CRISPR screening, perturbation sequencing (Perturb-seq), single-cell RNA sequencing (scRNAseq), and computational modeling are now allowing investigators to move beyond bulk measurements of reservoir size and toward the identification of the regulatory circuits that govern whether a provirus remains silent or becomes transcriptionally active. These approaches are powerful because they capture the fact that genetically similar or even clonally related cells can respond differently to the same latency-reversing perturbation. This has shifted the field away from viewing latency reversal as a simple pharmacologic trigger and toward understanding it as a context-dependent outcome shaped by cell state. The sessions also emphasized that long-term ART profoundly remodels the host immune system, with important implications for cure strategies. Conventional thinking has often assumed that HIV-specific immunity in people treated during chronic infection is irreversibly exhausted or senescent, particularly after decades of infection and aging. However, emerging data suggest that this view may be overly simplistic. Under prolonged viral suppression, HIV-specific CD8+ T-cell populations may undergo clonal renewal and regain features of functional competence, including proliferative potential and cytolytic capacity. These observations are highly relevant because most cure strategies ultimately depend, directly or indirectly, on immune-mediated control or clearance of infected cells. A further theme was the increasing convergence between HIV cure research and lessons from cancer immunotherapy. The field is paying growing attention to principles established in oncology, including the importance of durable expansion, persistence, trafficking, and fitness of cytotoxic effector cells. Just as successful cancer immunotherapies require not only target recognition but also robust effector cell persistence and function, effective HIV remission strategies will likely require immune effectors capable of recognizing rare, infected cells across diverse anatomic and transcriptional contexts. Broadly neutralizing antibodies (bNAbs), therapeutic vaccines, engineered T cells, and combinatorial immunotherapies are therefore being evaluated not in isolation but as part of integrated strategies to reshape the reservoir and the host response. Taken together, the cure-related presentations reinforced several key ideas. First, latency is not a uniform, binary state but a spectrum of proviral and cellular phenotypes. Second, the reservoir is maintained largely through infected cell survival and clonal expansion rather than ongoing rounds of productive replication. Third, immune competence under long-term ART may be more recoverable than previously appreciated. Finally, successful cure strategies will likely need to integrate mechanistic understanding of latency, longitudinal reservoir dynamics, and immune reconstitution into tailored combination approaches. The following abstracts illustrate these emerging themes particularly well.
Spatial Transcriptomic and Proteomic Profiling of HIV Reservoirs in Lymphoid Tissues
Despite the remarkable success of combination antiretroviral therapy, HIV persists in long-lived cellular reservoirs that prevent eradication of infection. Lymphoid tissues, particularly lymph nodes, represent a major anatomical reservoir where infected cells can evade immune surveillance. Previous studies have shown that HIV-infected cells often localize within B-cell follicles where cytotoxic T lymphocytes have limited access. However, the precise cellular and spatial organization of these reservoir niches remains incompletely understood.
Ndhlovu (Presentation 2) reviewed efforts to address this knowledge gap by applying integrated spatial transcriptomics and proteomic imaging approaches to lymph node tissues from individuals receiving suppressive ART. By combining spatial gene expression profiling with highdimensional histocytometry, the investigators aimed to characterize the follicular immune microenvironment that permits persistence of HIV-infected cells. Understanding the structural and cellular factors that limit immune clearance is essential for developing strategies aimed at reservoir elimination. Using spatial transcriptomics platforms, the investigators generated high-resolution maps of gene expression within lymph node sections obtained from ART-suppressed individuals. These analyses were integrated with multiplex protein imaging and histocytometric analyses that allowed the identification of specific immune cell subsets within the tissue microenvironment. Several important observations emerged.
First, HIV RNA-positive cells were found to be highly enriched within germinal center and follicular regions of lymph nodes. These regions exhibited unique transcriptional signatures characterized by elevated expression of genes associated with T follicular helper (Tfh) cells and follicular dendritic cells. Importantly, and consistent with previously published work,1 cytotoxic CD8+ T cells were relatively excluded from these regions, suggesting that spatial segregation contributes to immune evasion. Second, the investigators identified distinct cellular networks surrounding HIV-infected cells, including Tfh cells, regulatory T cells, and follicular dendritic cells that collectively formed a specialized microenvironment. These niches exhibited elevated expression of immune checkpoint molecules and antiinflammatory signaling pathways, which may dampen antiviral immune responses. Third, high-dimensional histocytometry revealed that infected cells frequently interacted with follicular helper T cells and dendritic cell populations that produce survival and activation signals. These interactions may support long-term persistence of infected cells despite suppressive therapy. These findings provide further evidence that HIV reservoirs persist within highly structured immunological niches in lymphoid tissues. The spatial segregation of infected cells from cytotoxic effector cells may represent a key barrier to reservoir elimination. Moreover, the presence of regulatory immune networks within these regions suggests that local immunosuppressive signals further protect infected cells from immune clearance.
Latent HIV Reversal Via Functional Genomics and Single-Cell Perturbation Modeling
A major barrier to HIV cure is the persistence of latently infected cells harboring replication-competent proviruses that remain transcriptionally silent during suppressive ART. These latent proviruses evade antiviral drugs, which act only on actively replicating virus, and immune-mediated clearance, which generally depends on antigen expression. For many years, efforts to reverse latency have focused on identifying individual latency-reversing agents capable of broadly inducing proviral transcription. However, clinical and preclinical studies have repeatedly shown that no single perturbation is sufficient to reactivate the full latent reservoir. This has pointed to a deeper biologic problem: latency is not governed by a single dominant block, but by a multilayered network of host pathways, chromatin states, transcription factor availability, signaling thresholds, and cell-state features.
Li's presentation (Presentation 4) was framed around the idea that successful latency reversal must be understood as a systems problem rather than a single-pathway problem. Genome-wide CRISPR knockout screening can identify host regulators that constrain or enable HIV transcription, and single-cell perturbation approaches such as Perturb-seq can determine how individual cells respond to those perturbations. The rationale for combining these methods is particularly strong, because bulk reactivation assays often obscure the underlying heterogeneity of response. A treatment that appears modestly effective overall may strongly reactivate a small subset of cells, leaving most of the reservoir unaffected. Thus, to design rational cure strategies, it is necessary not only to identify the pathways controlling latency, but also to understand which reservoir cell states are poised to respond. Using a latent HIV cell-line model, the investigators performed genome-scale CRISPR knockout screens and identified a complex regulatory landscape controlling proviral silencing and reactivation. Rather than pointing to a single dominant suppressive pathway, the screen revealed numerous host networks wherein perturbation affected latency. The results suggested that combinations of targeted interventions, rather than single gene disruptions or single pharmacologic classes, had the greatest potential to induce robust reactivation. This is an important observation because it supports the growing view that latency is maintained redundantly through intersecting pathways involving chromatin repression, transcriptional elongation blocks, signaling constraints, and cell-state determinants.
The investigators then applied Perturb-seq and computational modeling to dissect how individual cells responded to specific perturbations. These analyses showed that genetically identical cells can follow divergent transcriptional trajectories after the same intervention. In other words, the effect of a latency-reversing strategy was shaped by cell-to-cell heterogeneity. Only particular subpopulations appeared primed for productive reactivation, whereas others remained silent or followed incomplete or abortive transcriptional responses. This may explain why even strongly rational interventions may fail to induce broad reservoir expression when assessed across a heterogeneous cell population. Future studies should extend these findings into primary-cell and ex vivo reservoir models from people on long-term ART, where latency is likely even more heterogeneous than in cell-line systems. It will also be important to identify reproducible biomarkers of the cell states most amenable to reactivation so that interventions can be designed or stratified accordingly.
The HIV Reservoir: Persistence and Expansion Under Long-Term ART
Although ART effectively suppresses plasma viremia, it does not eliminate the reservoir of HIV-infected cells that gives rise to rebound when treatment is stopped. Over the past decade, a broad consensus has emerged that the principal reservoir consists of CD4+ T cells carrying integrated, inducible, replication-competent proviruses, and the much larger pool of defective proviruses may still contribute to immune activation and viral antigen production. At the same time, the field has increasingly recognized that the reservoir is maintained predominantly through clonal expansion of infected cells, not through sustained cycles of ongoing replication in the face of effective ART. However, many crucial questions remain unresolved: how does the reservoir evolve over decades of suppressive therapy, why do some individuals exhibit continued reservoir decay while others plateau or expand, and which biologic features of infected clones predict long-term persistence?
Mellors (Presentation 26) addressed these fundamental questions by providing an overview of emerging data on the long-term dynamics of the HIV reservoir and linking them to therapeutic efforts aimed at achieving durable remission. He summarized several key issues that now define the modern view of the HIV reservoir. First, intact proviruses account for only a small fraction of infected cells, whereas most proviruses are defective. Nonetheless, it is the intact, inducible proviruses that principally determine rebound risk after ART interruption. Second, the reservoir is maintained mainly through clonal expansion of infected cells rather than continuous de novo infection. Third, although proviral integration sites do not appear to be the dominant driver of clonal expansion in most cases, intact proviruses are disproportionately found in Krüppel-associated box zinc finger (KRAB-ZNF) genes. This suggests that genomic context may influence persistence, perhaps by favoring low-level expression, immune invisibility, or stable maintenance, although the mechanisms remain unclear. Fourth, proviral expression is highly variable. Even on ART, most reservoir proviruses are latent, but a meaningful degree of expression persists across some clones and some individuals, sufficient in many cases to generate chronic low-level viremia. Whether this residual viremia is infectious or materially contributes to rebound remains uncertain, but it indicates that reservoir silence is not absolute. Longitudinal studies over decades of ART have revealed at least 3 patterns of intact proviral dynamics, ie, continuous decay, decay to a stable plateau, and expansion during the second decade of therapy. Expansion appears least frequent but can be dramatic, including cases in which a single infected clone undergoes pronounced outgrowth. These observations indicate that long-term reservoir behavior is not uniform across individuals and that the reservoir remains biologically active in population terms even when virologically suppressed.
Mellors also reviewed the current therapeutic landscape, including latency reversal or silencing approaches, proviral knockout strategies, therapeutic vaccination, cellular therapies, bNAbs, and combination regimens. Several approaches, especially those involving bNAbs, have shown promising signals, but the field remains limited by small studies and insufficient statistical power. Drawing from cancer immunotherapy, Mellors emphasized that durable remission likely requires highly functional cytotoxic effector cells capable of expansion and persistence, as illustrated by the success of chimeric antigen receptor-and T-cell receptor-engineered approaches in oncology. Future work should define the biologic mechanisms that distinguish reservoirs that continue to decay from those that plateau or expand during long-term ART. Particular attention should be paid to the role of antigen-driven proliferation, homeostatic signals, integration-site context, and immune pressure in shaping these trajectories. It will also be important to determine whether residual proviral expression promotes immune recognition, fuels inflammation, or contributes directly to rebound competence. On the therapeutic side, more rigorously powered studies are needed to evaluate bNAbs, engineered T-cell therapies, and combination approaches. A particularly important future direction will be to integrate longitudinal reservoir measurements with immune fitness profiling so that remission strategies can be matched to the biology of each individual reservoir.
HIV-Specific T-Cell Responses in PWH on Long-Term ART
CD8+ T cells are central to immune control of HIV infection and are a key component of nearly all immune-based cure strategies. However, in people who initiate ART during chronic HIV infection, HIV-specific CD8+ T cells have traditionally been viewed as dysfunctional, exhausted, and potentially limited in their capacity to mediate durable viral control if treatment is stopped. This concern is further compounded by the fact that many PWH are now aging on long-term ART, raising the possibility that immune senescence and thymic involution could further compromise the quality of HIV-specific cellular immunity. These issues are highly relevant to cure research because strategies such as therapeutic vaccination, latency reversal plus immune clearance, bNAb combinations, and engineered cellular therapies all depend, directly or indirectly, on a host immune system capable of exerting control over infected cells.
Latency is maintained redundantly through intersecting pathways involving chromatin repression, transcriptional elongation blocks, signaling constraints, and cell-state determinants. Appay (Presentation 27) reviewed whether prolonged rates of viral suppression and immune reconstitution under ART might allow the emergence of renewed or replacement clones with improved functional properties. Using flow cytometry and single-cell RNA sequencing in PWH on very long-term ART, including individuals with 25 to 30 years of clinical follow-up, the study aimed to define the phenotypic and transcriptomic characteristics of persistent HIV-specific CD8+ T-cell responses. The key finding of the study was that instead of being dominated by exhaustion and senescence markers, HIV-specific CD8+ T cells in PWH undergoing long-term treatment displayed features of early differentiation and stemness. These cells appeared to retain proliferative and cytolytic potential, suggesting a more rejuvenated and functionally competent phenotype than many would have predicted in people treated during chronic infection and followed for decades.
Single-cell and phenotypic analyses supported a model of clonal succession. According to this model, older, more exhausted clones are progressively replaced over time by newly generated clones with improved functional properties. This concept is important because it suggests that successful ART does not simply preserve a damaged
HIV-specific repertoire; rather, it may permit ongoing immune remodeling and selective renewal of antiviral T-cell populations. This observation is noteworthy given the expected effects of aging and reduced thymic output, implying that the immune system in PWH retains more regenerative capacity than previously assumed. The presentation also linked these findings to cure relevance by noting that similarly stem-like, less terminally differentiated HIV-specific CD8+ T cells have recently been associated with posttreatment control. This raises the possibility that the quality of the persisting HIV-specific T-cell pool under long-term ART may be more favorable for remission strategies than older models of irreversible dysfunction would suggest. Latency is maintained redundantly through intersecting pathways involving chromatin repression, transcriptional elongation blocks, signaling constraints, and cell-state determinants.
In addition to the biologic findings, the presentation emphasized a translational point: immune interventions should be aligned with the physiologic and clinical context of the target population. The optimal strategy for inducing effective CD8+ T-cell responses may differ depending on age, immune reconstitution status, treatment history, and baseline T-cell composition. The proposed model of clonal succession suggests that prolonged ART may gradually reshape the HIV-specific repertoire, enabling the replacement of older, exhausted clones with functionally fitter ones. If correct, this has implications for the timing and design of cure interventions. Individuals on stable longterm therapy may in some cases be better candidates for immune-based remission strategies than previously appreciated, provided their HIV-specific responses can be appropriately boosted or redirected.
Future studies should determine how broadly this phenomenon extends across different age groups, treatment histories, and timing of ART initiation, including individuals treated during acute versus chronic infection. It will also be important to establish whether these rejuvenated HIV-specific CD8+ T-cell populations can directly mediate reservoir control in ex vivo or in vivo models and whether their presence predicts outcomes in analytic treatment interruption studies.
Characterization of HIV Proviruses in Brain Microglia in PWH on ART
An unresolved question in HIV cure research is whether the central nervous system (CNS) contains a true, biologically meaningful reservoir that is distinct from the better-characterized reservoir in CD4+ T cells. Among candidate CNS reservoir cells, brain microglia have long attracted particular interest because they are long-lived, self-renewing myeloid cells resident in the brain, and are capable of surviving for prolonged periods in an inflammatory environment. If infected microglia harbor intact proviruses, they could represent a uniquely challenging reservoir because of their anatomic location, limited accessibility to immune effectors, and distinct cellular biology compared with circulating T cells. At the same time, interpretation of HIV nucleic acid detected in brain tissues has remained difficult, because signals attributed to microglia may in some cases reflect infiltrating infected T cells, adherent material, or phagocytosed HIV-containing debris rather than authentic myeloid infection.
Thomas and colleagues (Abstract 110) addressed these issues directly using autopsy tissues from PWH on suppressive ART enrolled in the Last Gift rapid autopsy program. The investigators asked 3 central questions: how often does the brain microglia harbor HIV proviruses, what fraction of those proviruses are potentially intact, and where are those proviruses integrated in the human genome? The investigators isolated DNA from several autopsy-derived tissue compartments, including brain microglia, mixed brain tissue, brain T cells, spleen bulk cells, and spleen T cells. Microglial isolates were defined as CD11b+ cells with high transmembrane 119 (TMEM119) positivity, providing strong support for microglial enrichment. DNA was diluted to the proviral endpoint and subjected to multiple displacement amplification, followed by screening for potentially intact proviruses using polymerase chain reaction detection of the long terminal repeat (LTR), packaging signal (Psi), and Rev response element (RRE) regions, which are commonly lost in defective genomes. Integration site sequencing was then used to define host genomic locations and assess clonality. HIV proviruses were detected in brain and spleen compartments, and as expected, the majority were defective. However, in 1 donor, a substantial fraction of proviruses met criteria for being potentially intact, including 27% of microglia-derived proviruses and 48% of spleen-derived proviruses.
This is a notable observation because it indicates that at least some HIV DNA associated with brain microglia is not merely fragmented or grossly defective, but may retain the structural hallmarks of replication competence. Integration site analyses identified large clones with identical integration sites present across brain and spleen tissues, a pattern consistent with a T-cell origin. These shared clones were detected in brain microglia, brain T cells, and spleen, suggesting that some of the proviral signal observed in brain-associated microglial fractions may derive from infected T cells that have trafficked into the CNS, adhered to microglia, or been phagocytosed by them. At the same time, the study identified smaller clones observed only in the microglial compartment, some of which were potentially intact. These microglial-only clones provide evidence that at least part of the CNS proviral burden may reside in a more specific myeloid reservoir.
The overall picture is therefore mixed but informative: the brain compartment contains signals consistent with infiltrating or clonally expanded T-cell-derived proviruses as well as signals supporting the existence of infected microglia with potentially intact proviruses. These results build upon earlier published studies indicating that brain microglia harbor inducible proviruses that exhibit a macrophage-tropic phenotype.2
A major challenge in HIV cure research is the inability to noninvasively track where and when viral rebound begins during analytical treatment interruption. Plasma viremia provides only a delayed and indirect signal of recrudescent infection, reflecting viral production after rebound has already propagated beyond its earliest tissue foci. Yet, the biologic events that determine whether and where rebound emerges likely begin much earlier within discrete anatomic compartments such as lymphoid tissue, gut, bone marrow, or the CNS. Understanding these earliest events is important for cure strategies, because it could reveal the active tissue reservoirs responsible for rebound and identify windows of vulnerability in which host immune responses or therapeutic interventions might still contain infection before it becomes systemic.
Henrich and colleagues (Abstract 113) presented a study that was designed to address that gap using multimodal positron emission tomography (PET) imaging during analytic treatment interruption (ATI). The investigators used one tracer, [89Zr]-VRC01, to visualize tissue-associated HIV envelope (Env)/glycoprotein 120 (gp120), and another, [18F] FAraG, to detect activated or cycling T cells. The investigators first performed longitudinal [89Zr]-VRC01 PET imaging in 3 individuals participating in an ATI study. Imaging time points included 2 days prior to first detectable plasma HIV RNA level, during viral rebound, and after ART reinitiation in a person with subsequently controlled virus. The 2 participants imaged in the pre-or perirebound period showed higher levels of tracer uptake in more various tissues than reported in previous studies on PWH on suppressive ART. These included inguinal lymph nodes, colorectal tissue, femoral bone marrow, and perivascular brain parenchyma, with particularly notable increases in the frontal cortex and cerebellum. In contrast, the controller imaged after ART reinitiation showed VRC01 uptake values that were similar to those seen in ART-suppressed individuals.
The investigators then examined T-cell activation dynamics using [18F]-FaraG PET-magnetic resonance imaging before ATI and again a median of 9 days after ART interruption. Repeat scanning was performed between 5 and 35 days before detectable plasma HIV RNA level, allowing the assessment of events in a truly prerebound state. In 3 of 4 participants, the investigators observed up to a 2.3-fold increase in tracer uptake in various inguinal lymph nodes compared with pre-ATI baseline. The highest uptake occurred in a participant who subsequently rebounded 35 days after imaging, suggesting that tissue immune activation may precede measurable plasma viremia by a substantial interval. In addition, 1 individual showed an approximately 1.5-fold increase in tracer uptake in several brain regions, including the frontal, temporal, cerebellar, and pontine areas, 17 days before the first detectable plasma viral load.
Together, these findings indicate that HIV-associated signal and immune activation can be visualized in tissues before virus becomes detectable in plasma. The imaging patterns also suggest that rebound may arise from variable focal sites rather than from simultaneous diffuse activation of a uniform systemic reservoir. The study offers direct evidence that HIV recrudescence during ATI is a tissue-first event that can be detected before standard plasma assays show rebound. The identification of focal activity in lymph nodes, gut, bone marrow, and brain supports the view that rebound emerges from anatomically heterogeneous reservoir sites that differ among individuals. The CNS-related findings are noteworthy. Increased uptake in perivascular brain regions and broader brain activation prior to plasma rebound raises the possibility that the brain may participate more actively in rebound dynamics than often appreciated, at least in some individuals. Although the imaging results do not by themselves prove productive infection in specific brain cell types, they do suggest that the CNS can be involved in early viral-host responses during ATI.
Methodologically, the work points toward a powerful new toolkit for cure studies. Noninvasive whole-body imaging could complement reservoir assays, tissue biopsies, and plasma measurements by revealing where active processes are occurring in vivo. This could be especially valuable in small interventional studies, where identifying the anatomic impact of a therapeutic strategy may be more informative than relying only on time-to-rebound endpoints.
At a broader level, the study underscores that reservoir biology is spatially organized. Cure strategies that succeed in blood may still fail if they do not adequately target anatomically privileged or variably active tissue sites. Future studies should evaluate these imaging approaches in larger ATI cohorts and directly correlate tracer uptake patterns with tissue biopsies, plasma virology, immune phenotyping, and rebound timing. It will be important to determine whether specific anatomic uptake signatures predict rebound kinetics, posttreatment control, or response to cure-directed interventions.
Notes
All abstracts cited in the text appear in the CROI 2026 Abstract eBook, available online at www.CROIconference.org
The IAS-USA has identified and resolved ahead of time any possible conflicts of interest that may influence CME activities with regard to exposition or conclusion. All financial relationships with ineligible companies for the author and planners/reviewers are below.
Financial affiliations in the past 24 months: Dr Stevenson reported no relevant financial relationships with ineligible companies. (Updated March 17, 2026)
Additional References Cited in Text
S Li, JM Folkvord, EG Rakasz, . Simian immunodeficiency virus-producing cells in follicles are partially suppressed by CD8+ cells in vivo. J Virol. 2016;90(24):11168–11180. doi: 10.1128/JVI.01332-16
Y Tang, A Chaillon, S Gianella, . Brain microglia serve as a persistent HIV reservoir despite durable antiretroviral therapy. J Clin Invest. 2023;133(12):e167417. doi: 10.1172/JCI167417