Abstract In the antibody mediated prevention (AMP) trials, the broadly neutralizing antibody (bNAb) VRC01 demonstrated protective efficacy against susceptible HIV strains. To understand how VRC01 shaped breakthrough infections, deep sequencing was performed on 172 participants (>100,000 gag-Δpol and rev-env-Δnef sequences), at diagnosis and over time, in the placebo and treatment arms of the African (HVTN703/HPTN081; NCT02568215) and Americas/Europe (HVTN704/HPTN085; NCT02716675) cohorts. A high frequency of multilineage infections was detected (38%), including co-infection with both VRC01 sensitive and resistant viruses. This high frequency is largely accounted for by low-abundance lineages. Although VRC01 does not significantly affect the genetic transmission bottleneck compared to placebo, higher VRC01 doses trend towards greater VRC01 neutralization differences among co-infecting lineages. Two-thirds of multilineage infections showed evidence of recombination at the diagnostic timepoint. In the treatment group there is evidence of recombinant viruses preferentially inheriting resistance-associated mutations. This study provides critical insights into viral genetic and antigenic diversity that needs to be targeted to achieve protection, and highlights the role of recombination in facilitating escape.

Precise and scalable quantification of the intact HIV reservoir is critical for advancing curative strategies. Current reservoir assays, such as the intact proviral DNA assay (IPDA), are limited by quantification failures or misclassification of defective proviruses due to HIV sequence heterogeneity. Q4ddPCR is a modular, droplet digital PCR simultaneously targeting four conserved regions in the HIV genome to improve specificity, reduce quantification gaps, and provide multi-layered readouts. It comprises two configurations: one fully based on Q4PCR primer/probes and one combining IPDA with gag and pol primer/probes from Q4PCR. We benchmark Q4ddPCR against 3650 near full-length proviral sequences from 13 virally suppressed people with HIV (PWH) generated by Q4PCR. Q4ddPCR closely matches sequence-confirmed reservoir measurements, and multi-probe readouts reveal clonal reservoir dynamics not detectable by IPDA. Q4ddPCR enables intact reservoir quantification in 95% of samples across four independent cohorts and in 16 PWH, strongly correlates with viral outgrowth. In longitudinal samples from 42 participants over the first 4.5 years on antiretroviral therapy (ART), Q4ddPCR reports lower proviral frequencies and a steeper decline in intact proviral DNA compared to IPDA. Collectively, our findings confirm key predictions from mathematical modeling, demonstrating that multi-target assays improve specificity and more accurately capture intact reservoir dynamics.

Broadly neutralizing antibodies (bnAbs) are a promising intervention for HIV prevention, therapy, and cure. bnAb optimization requires precise quantification of in vivo functions, many of which cannot be directly measured in humans. We therefore performed a mathematical modeling meta-analysis which integrated four clinical trials and reproduced serial bnAb concentrations, viral loads, and bnAb sensitivities (IC50) in 43 viremic trial participants who received an infusion of VRC01, VRC01LS, VRC07-523LS, 3BNC117 or 10-1074. We compared >300 mathematical models for their ability to recapitulate multi-strain HIV dynamics following bnAb infusion. For each bnAb, our best model identified a scaling factor of 36-462 to project in vivo activity from in vitro IC50, quantified Fc-mediated infected cell killing in humans over time, and projected the timing of bnAb-resistant strain emergence. Using this holistic profile, VRC07-523-LS was generally optimal.

ABSTRACT Antiviral clinical trial simulation (CTS) is a type of mathematical modeling that couples viral- immune dynamics (VID) unique to each human viral pathogen, with mechanistic, pharmacokinetic (PK), and pharmacodynamic (PD) drug characteristics. Validation is achieved by matching model output to detailed viral load trajectories from trials. Antiviral CTS can be applied at all stages of drug development to viruses with distinct shedding patterns. Models can capture the activity of small molecules, neutralizing antibodies, and cellular therapies, as well as combination strategies to enhance potency and avoid drug resistance. Several principles are observed across antiviral CTS models. First, PK and PD models that recapitulate drug levels and concentration-dependent antiviral activity are often necessary, but never sufficient to predict trial results. VID equations are also required to guide optimal treatment timing because expanding immune responses synergistically eliminate infection but are deleterious if too sustained or intense. Therefore, equivalent antiviral doses may have different efficacy if given during different infection stages. Second, antiviral CTS models identify effective plasma drug concentrations in humans, which are often poorly predicted by in vitro assays. Finally, models that do not consider drug mechanisms lead to incorrect efficacy estimates. Data-validated CTS is increasingly used to inform drug dose and dosing interval, treatment timing and duration, virologic endpoint selection, and sample size, particularly when applied to detailed phase 1 and 2 trial data. Given the high expense of antiviral licensure trials, CTS models are vital to optimize trial efficacy and de-risk the drug development process.

Antiretroviral therapy (ART) suppresses HIV replication in people living with HIV (PWH), but a persistent population of reservoir cells prevents cure. Reservoir cells are mostly anatomically dispersed, latently infected CD4+ T cells harboring one copy of chromosomally integrated, replication-competent HIV proviral DNA. Despite their low frequency (0.01%-0.1%) among CD4+ T cells and the quiescence of most genetically intact proviruses, viremia usually recurs within weeks after ART cessation. When PWH are not on ART, the reservoir is sustained through viral infection and infected cell proliferation. During suppressive ART, HIV reservoir cells persist via mechanisms sustaining uninfected CD4+ T cells including antigen-responsive and homeostatic clonal proliferation, programmed cell death, and T cell subset differentiation. Rates of latently infected cell proliferation and death must exist in quasi-equilibrium to explain limited change in reservoir volume over decades of ART, and the rarity of cancers or lymphoproliferative disorders emerging from infected cells. Some reservoir cells are under additional selection forces during ART, illustrated by slightly higher clearance rates of genetically intact versus replication-defective HIV proviral DNA and by a gradual transition to a less transcriptionally active and more clonal reservoir. While a small but meaningful percentage of latently infected cells are negatively selected due to lytic viral replication or elimination by adaptive immune responses, most reservoir cell death occurs independently of harboring intact HIV DNA. Given that HIV is often a passenger in reservoir cells, CD4+ T cell proliferation, targeted death, and subset differentiation may be viable therapeutic targets for curative interventions.

Accurate timing estimates of when participants acquire HIV in HIV prevention trials are necessary for determining antibody levels at acquisition. The Antibody-Mediated Prevention (AMP) Studies showed that a passively administered broadly neutralizing antibody can prevent the acquisition of HIV from a neutralization-sensitive virus. We developed a pipeline for estimating the date of detectable HIV acquisition (DDA) in AMP Study participants using diagnostic and viral sequence data. Using a Bayesian strategy that combines three streams of data (REN [rev/vpu/env/Δnef] sequence, GP [gag/Δpol] sequence, and diagnostic) where their 95% credible intervals overlap based on pre-specified criteria and decision rules. We evaluated the performance of our AMP pipeline using PacBio viral sequence data from 41 participants across two prospective acute HIV acquisition cohort studies, FRESH and RV217, with twice-weekly sampling. These cohort studies enrolled young women in South Africa and men and women in Kenya and Thailand, respectively, with a high likelihood of HIV acquisition. In evaluating performance, "true DDA" was the center of bounds between last-negative and first-positive RNA diagnostic tests (median time 4 days, range 2-7 days); bias was the mean difference between estimated and true DDA. Using diagnostic data alone yielded timing estimates with a bias of 2.4 days and root mean square error (RMSE) of 7.9 days. These results were improved using sequence + diagnostic data (bias 1.5 days, RMSE 6.9 days), as well as by restricting sequence-based estimation to samples from ≤5 weeks post-DDA (bias 0.2 days, RMSE 7.8 days).IMPORTANCEIn HIV prevention trials, accurate timing estimates of when individual participants acquire HIV can be used to estimate antibody levels at the time of acquisition, which is useful for projecting antibody levels needed for prevention. The results we report here suggest that if sequence-based estimation of acquisition timing is used in future clinical trials of combination broadly neutralizing antibody (bnAb) regimens or multispecific bnAbs for HIV prevention, a sampling frequency of at least monthly is needed. Moreover, in the samples analyzed here, we observed less bias in sequence-based timing estimation for samples taken <5 weeks post-DDA. This observation is consistent with the timing of immune-driven selective pressures that may negatively impact the power to detect acquisition sieve effects.

HIV cure is exceptionally rare, with only six cases documented among the estimated 88 million individuals who have acquired HIV since the onset of the epidemic1-6. Successful cures, including that of the pioneering individual known as the Berlin patient, are limited to those who received allogeneic stem cell transplants (allo-SCTs) for haematological cancers. HIV resistance from stem cell donors with the rare homozygous CCR5Δ32 mutation was long considered the main mechanism for HIV remission without antiretroviral therapy. However, recent reports have highlighted CCR5-independent mechanisms as important contributors to HIV cure6-8. Here we provide new evidence for this conceptual shift, whereby long, treatment-free HIV remission was achieved after allo-SCT with functionally active CCR5. A man with heterozygous CCR5 wild-type/Δ32 living with HIV received allo-SCT from a HLA-matched unrelated heterozygous CCR5 wild-type/Δ32 donor as treatment for acute myeloid leukaemia. Three years after allo-SCT, the patient discontinued antiretroviral therapy. So far, HIV remission has been sustained for more than 6 years with undetectable plasma HIV RNA. Reservoir analysis revealed intact proviral HIV before transplantation, but no replication-competent virus in blood or intestinal tissues after allo-SCT. Declining or absent HIV-specific antibody and T cell responses support the absence of viral activity. High antibody-dependent cellular cytotoxicity activity at the time of transplantation may have contributed to HIV reservoir clearance. These results demonstrate that CCR5Δ32-mediated HIV resistance is not essential for durable remission, which underscores the importance of effective viral reservoir reductions in HIV cure strategies.

To inform cure in children living with HIV (CWH), we elucidated the dynamics and mechanisms underlying HIV persistence during antiretroviral therapy (ART). In 120 Kenyan CWH who initiated ART between 1-12 months of age, 55 had durable viral load suppression, and 65 experienced ART interruptions. We measured plasma HIV RNA levels, CD4+ T cell count, and levels of intact and defective HIV DNA proviruses via the cross-subtype intact proviral DNA assay (CS-IPDA). By modeling data from the durably suppressed subset, we found that during early ART (year 0-1 on ART), plasma RNA levels decayed rapidly and biphasically and intact and defective HIV DNA decayed with mean 3 and 9 month half-lives, respectively. After viral suppression was achieved (years 1-8 on ART), intact HIV DNA decay slowed to a mean 22 month half-life, whilst defective HIV DNA no longer decayed. In five CWH, we found individual CD4+ TCRβ clones wax and wane, but average kinetics resembled those of defective DNA and CD4 count, suggesting that differential decay of intact HIV DNA arises from selective pressures overlaying normal CD4+ T cell kinetics. Finally, by modeling HIV RNA and DNA in CWH with treatment interruptions, we linked temporary viremia to transient rises in HIV DNA, but long-term intact reservoirs were not strongly influenced, suggesting brief treatment interruptions may not significantly increase HIV reservoirs in children.

Fitting mathematical models of viral dynamics to serial, quantitative viral load data provides inferences on the mechanisms in virus infection. This process can reveal the speed and magnitude of viral replication, cell proliferation and death, immune responses, and/or treatment efficacy. Viral dynamics modeling involves developing conceptual models, translating them into equations, and applying the appropriate statistical tools to determine the optimal parameters such that the model recapitulates observations from human and animal infections. In this review, we outline the theoretical foundations needed to understand model fitting, parameter estimation, and what it means to achieve a good fit. We provide examples and explain the strengths and limitations of three commonly used model fitting approaches: individual fitting, population mixed effects fitting, and feature fitting. We briefly review fitting algorithms and highlight powerful available computer software packages that can be used for fitting and parameter estimation. We discuss different model types, parameter identifiability, and how future modeling efforts can leverage advances in multi-dimensional data. Finally, we conclude with simple guidelines for choosing the best approach based on available data and scientific questions.

Broadly neutralizing antibodies (bnAbs) show promise in HIV prevention, yet viral escape remains a challenge. In the Antibody Mediated Prevention (AMP) trials, the CD4 binding site (CD4bs) bNAb VRC01 blocked acquisition by VRC01-sensitive strains. However, its influence on viral evolution post-acquisition is not fully understood. Here we analyzed >12,000 HIV env sequences from 47 participants from the AMP trials, identifying VRC01-mediated de novo escape mutations in 8 of 26 VRC01-treated participants but none in 21 placebo participants. These mutations were found at very low frequency (<1%) in global viruses. Escape mutations, primarily located in the Loop-D and β23/V5 regions of Env, conferred cross-resistance to several CD4bs bnAbs, while more potent CD4bs bnAbs like N6 and 1-18 largely retained their activity. Our findings demonstrate that prophylactic VRC01 can select for viral escape after infection, underscoring the need for next-generation bnAbs with improved breadth and potency to enhance durability and efficacy of antibody-based HIV prevention.

Precise and scalable quantification of the genetically intact HIV reservoir is critical for advancing curative strategies. However, current HIV reservoir assays such as the intact proviral DNA assay (IPDA) are limited by quantification failures or misclassification of defective proviral genomes due to HIV sequence heterogeneity. Q4ddPCR is a modular, droplet digital PCR assay that simultaneously targets four conserved regions in the HIV genome to improve specificity, reduce quantification gaps, and provide multi-layered readouts. We benchmarked Q4ddPCR against 3,650 near full-length proviral sequences from 13 virally suppressed people with HIV (PWH) generated by Q4PCR using the same primer/probe sets. Q4ddPCR enabled intact reservoir quantification in 95% of samples from three independent cohorts and closely matched sequence-confirmed Q4PCR reservoir measurements. In addition, multi-probe readouts revealed clonal intact reservoir dynamics that are not detectable by IPDA. In longitudinal samples from 42 participants over the first 4.5 years on antiretroviral therapy (ART), Q4ddPCR reported lower proviral frequencies and a steeper decline in intact proviral DNA compared to IPDA. Collectively, our findings confirm key predictions from mathematical modeling, demonstrating that multi-target assays provide greater specificity and more accurately capture the dynamics of the intact HIV reservoir.

To determine whether HIV persistence arises from the natural dynamics of memory (m)CD4+ T cells, we compare clonal dynamics of HIV proviruses and mCD4+ T cells from the same people living with HIV (PWH) on antiretroviral therapy and from matched HIV-seronegative people (N = 51). HIV proviruses are more clonal than mCD4+ T cells but similarly clonal to antigen-specific cells. Increasing reservoir clonality over time and differential decay of intact and defective proviruses are not explained by mCD4+ T cell kinetics alone. We develop and validate a stochastic model trained on 10 quantitative data metrics, which shows that negative selection against HIV-infected cells is necessary to explain all metrics. We estimate the strength of negative selection, finding that death of cells harboring intact and defective proviruses is infrequently (∼6% and ∼2% on average) due to HIV-specific factors. Thus, our data indicate that HIV persistence is mostly, but not entirely, driven by natural mCD4+ kinetics.

On July 19th, 2023, the National Institute of Allergy and Infectious Diseases co-organized a workshop with the Society of Mathematical Biology, with the authors of this paper as the organizing committee. The workshop, “Bridging multiscale modeling and practical clinical applications in infectious diseases” sought to create an environment for mathematical modelers, statisticians, and infectious disease researchers and clinicians to exchange ideas and perspectives.

Most proviruses persisting in people living with HIV (PWH) on antiretroviral therapy (ART) are defective. However, rarer intact proviruses almost always reinitiate viral rebound if ART stops. Therefore, assessing therapies to prevent viral rebound hinges on specifically quantifying intact proviruses. We evaluated the same samples from 10 male PWH on ART using the two-probe intact proviral DNA assay (IPDA) and near full length (nfl) Q4PCR. Both assays admitted similar ratios of intact to total HIV DNA, but IPDA found ~40-fold more intact proviruses. Neither assay suggested defective proviruses decay over 10 years. However, the mean intact half-lives were different: 108 months for IPDA and 65 months for Q4PCR. To reconcile this difference, we modeled additional longitudinal IPDA data and showed that decelerating intact decay could arise from very long-lived intact proviruses and/or misclassified defective proviruses: slowly decaying defective proviruses that are intact in IPDA probe locations (estimated up to 5%, in agreement with sequence library based predictions). The model also demonstrates how misclassification can lead to underestimated efficacy of therapies that exclusively reduce intact proviruses. We conclude that sensitive multi-probe assays combined with specific nfl-verified assays would be optimal to document absolute and changing levels of intact HIV proviruses.

Persistence of HIV in people living with HIV (PWH) on suppressive antiretroviral therapy (ART) has been linked to physiological mechanisms of CD4+ T cells. Here, in the same 37 male PWH on ART we measure longitudinal kinetics of HIV DNA and cell turnover rates in five CD4 cell subsets: naïve (T N ), stem-cell- (T SCM ), central- (T CM ), transitional- (T TM ), and effector-memory (T EM ). HIV decreases in T TM and T EM but not in less-differentiated subsets. Cell turnover is ~10 times faster than HIV clearance in memory subsets, implying that cellular proliferation consistently creates HIV DNA. The optimal mathematical model for these integrated data sets posits HIV DNA also passages between CD4 cell subsets via cellular differentiation. Estimates are heterogeneous, but in an average participant’s year ~10 (in T N and T SCM ) and ~10 4 (in T CM , T TM , T EM ) proviruses are generated by proliferation while ~10 3 proviruses passage via cell differentiation (per million CD4). In simulations, therapies blocking proliferation and/or enhancing differentiation could reduce HIV DNA by 1-2 logs over 3 years. In summary, HIV exploits cellular proliferation and differentiation to persist during ART but clears faster in more proliferative/differentiated CD4 cell subsets and the same physiological mechanisms sustaining HIV might be temporarily modified to reduce it.

The Antibody Mediated Prevention (AMP) trials (NCT02716675 and NCT02568215) demonstrated that passive administration of the broadly neutralizing monoclonal antibody VRC01 could prevent some HIV-1 acquisition events. Here, we use mathematical modeling in a post hoc analysis to demonstrate that VRC01 influenced viral loads in AMP participants who acquired HIV. Instantaneous inhibitory potential (IIP), which integrates VRC01 serum concentration and VRC01 sensitivity of acquired viruses in terms of both IC50 and IC80, follows a dose-response relationship with first positive viral load ( p = 0.03), which is particularly strong above a threshold of IIP = 1.6 ( r = -0.6, p = 2e-4). Mathematical modeling reveals that VRC01 activity predicted from in vitro IC80s and serum VRC01 concentrations overestimates in vivo neutralization by 600-fold (95% CI: 300–1200). The trained model projects that even if future therapeutic HIV trials of combination monoclonal antibodies do not always prevent acquisition, reductions in viremia and reservoir size could be expected.

Magnetic particle spectroscopy (MPS) in the Brownian relaxation regime, also termed magnetic spectroscopy of Brownian motion (MSB), can detect and quantitate very low, sub-nanomolar concentrations of molecular biomarkers. MPS/MSB uses the harmonics of the magnetization induced by a small, low-frequency oscillating magnetic field to provide quantitative information about the magnetic nanoparticles’ (mNPs’) microenvironment. A key application uses antibody-coated mNPs to produce biomarker-mediated aggregation that can be detected using MPS/MSB. However, relaxation changes can also be caused by viscosity changes. To address this challenge, we propose a metric that can distinguish between aggregation and viscosity. Viscosity changes scale the MPS/MSB harmonic ratios with a constant multiplier across all applied field frequencies. The change in viscosity is exactly equal to the multiplier with generality, avoiding the need to understand the signal explicitly. This simple scaling relationship is violated when particles aggregate. Instead, a separate multiplier must be used for each frequency. The standard deviation of the multipliers over frequency defines a metric isolating viscosity (zero standard deviation) from aggregation (non-zero standard deviation). It increases monotonically with biomarker concentration. We modeled aggregation and simulated the MPS/MSB signal changes resulting from aggregation and viscosity changes. MPS/MSB signal changes were also measured experimentally using 100 nm iron-oxide mNPs in solutions with different viscosities (modulated by glycerol concentration) and with different levels of aggregation (modulated by concanavalin A linker concentrations). Experimental and simulation results confirmed that viscosity changes produced small changes in the standard deviation and aggregation produced larger values of standard deviation. This work overcomes a key barrier to using MPS/MSB to detect biomarkers in vivo with variable tissue viscosity.

Modern HIV research depends crucially on both viral sequencing and population measurements. To directly link mechanistic biological processes and evolutionary dynamics during HIV infection, we developed multiple within-host phylodynamic models of HIV primary infection for comparative validation against viral load and evolutionary dynamics data. The optimal model of primary infection required no positive selection, suggesting that the host adaptive immune system reduces viral load but surprisingly does not drive observed viral evolution. Rather, the fitness (infectivity) of mutant variants is drawn from an exponential distribution in which most variants are slightly less infectious than their parents (nearly neutral evolution). This distribution was not largely different from either in vivo fitness distributions recorded beyond primary infection or in vitro distributions that are observed without adaptive immunity, suggesting the intrinsic viral fitness distribution may drive evolution. Simulated phylogenetic trees also agree with independent data and illuminate how phylogenetic inference must consider viral and immune-cell population dynamics to gain accurate mechanistic insights.

The emergence of new SARS-CoV-2 variants of concern (VOC) has hampered international efforts to contain the COVID-19 pandemic. VOCs have been characterized to varying degrees by higher transmissibility, worse infection outcomes and evasion of vaccine and infection-induced immunologic memory. VOCs are hypothesized to have originated from animal reservoirs, communities in regions with low surveillance and/or single individuals with poor immunologic control of the virus. Yet, the factors dictating which variants ultimately predominate remain incompletely characterized. Here we present a multi-scale model of SARS-CoV-2 dynamics that describes population spread through individuals whose viral loads and numbers of contacts (drawn from an over-dispersed distribution) are both time-varying. This framework allows us to explore how super-spreader events (SSE) (defined as greater than five secondary infections per day) contribute to variant emergence. We find stochasticity remains a powerful determinant of predominance. Variants that predominate are more likely to be associated with higher infectiousness, an SSE early after variant emergence and ongoing decline of the current dominant variant. Additionally, our simulations reveal that most new highly infectious variants that infect one or a few individuals do not achieve permanence in the population. Consequently, interventions that reduce super-spreading may delay or mitigate emergence of VOCs.