Measuring Human Immunodeficiency Virus Reservoirs: Do We Need to Choose Between Quantity and Quality?

Abstract

The persistence of latent viral genomes in people receiving antiretroviral therapy (ART) is the main obstacle to a cure for human immunodeficiency virus (HIV) infection. Viral reservoirs can be defined as cells harboring HIV genomes that have the ability to produce infectious virions. Precise quantification of the cellular reservoirs of HIV is challenging because these cells are rare, heterogeneous, and outnumbered by a larger number of cells carrying defective genomes. In addition, measuring the inducibility of these proviruses requires functional assays and remains technically difficult. The recent development of single-cell and single-viral genome approaches revealed additional layers of complexity: the cell subsets that harbor proviruses are heterogeneous and their ability to be induced is variable. A substantial fraction of intact HIV genomes may be permanently silenced after years of ART, revealing the underappreciated importance of induction assays. As such, a simple approach that would assess simultaneously the genetic intactness and the inducibility of the reservoir is still lacking. In this study, we review recent advances in the development of methods to quantify and characterize persistently infected cells, and we discuss how these findings can inform the design of future assays aimed at measuring the size of the intact and inducible HIV reservoir.

This review describes recent advances in the development of methods to quantify the HIV reservoir in people on ART and how these findings can inform the design of future assays aimed at measuring the size of the intact and inducible HIV reservoir.

Antiretroviral therapy (ART) has revolutionized the treatment of human immunodeficiency virus (HIV) infection by drastically reducing the death rate of people with the virus. Although progress remains to be made to make ART accessible to all and to reduce its cost and associated toxicities, modern therapies have converted a deadly disease into a manageable chronic infection that can be controlled therapeutically and which is no longer transmissible [1]. Despite this formidable success, ART is not a cure: HIV persists in long-lived reservoirs that are predicted to persist for a lifetime [2–5]. As such, viruses hidden in these reservoirs typically reignite infection within a few weeks when the therapeutic pressure is removed, ie, upon ART cessation.

To date, only a few people are considered cured of HIV. These include a few individuals who were transplanted with stem cells naturally resistant to HIV infection [6–9] but also rare cases of spontaneous cure [10, 11]. In all these instances, the demonstration of a cure was made by the interruption of ART, which did not lead to viral rebound. It is interesting to note that traces of HIV genetic material (deoxyribonucleic acid [DNA] or ribonucleic acid [RNA]) could be detected using highly sensitive assays in most cases [8, 9, 12], suggesting that low amounts of viral products may persist even after remission. This reinforces the need to identify a virological marker that could be used to document HIV cure unequivocally.

For many years, the size of the HIV reservoir has been estimated by 2 types of assays. Although HIV DNA measures by polymerase chain reaction (PCR) are easy and cost effective, they largely overestimate the size of the reservoir by capturing genomes harboring genetic defects that would not allow them to produce infectious viral particles [13]. At the other end of the spectrum, viral outgrowth methods only capture inducible and replication competent proviruses but are limited in their use by their cost, their laborious nature, and the fact that they miss a significant fraction of the replication competent reservoir [14].

During the past 10 years, novel generation assays have been developed to measure and study the reservoir. Those can be divided into 2 main categories: (1) quantitative assays measure the frequency of reservoir cells and include novel generations of PCR-based assays and (2) quantitative induction assays that detect viral transcripts or proteins as well as improved viral outgrowth methods. Qualitative approaches allow the characterization of these cells and their associated proviruses but are rarely used to measure the size of the reservoir. In this review, we summarize these novel approaches and propose novel combinatory assays that may provide a better estimate of the viral reservoirs that cause rebound when ART is interrupted (Figure 1).

Figure 1.

Properties of human immunodeficiency virus (HIV) reservoir assays. Each method is classified depending on its ability to quantify reservoirs, and to inform on cell phenotype, on inducibility and/or intactness of the provirus, or on the replication competence of the provirus. quantitative viral outgrowth assay (QVOA) is the only method that allows assessment of the replication competence of proviruses, but it does not allow for the phenotyping of the cells carrying these proviruses. ASAP-Seq, ATAC with select antigen profiling; ECCITE-Seq, expanded CRISPR compatible CITE-Seq; FIND-Seq, focused interrogation of cells by nucleic acid detection and sequencing; FLIPS, Full-Length Individual Proviral Sequencing; IPDA, intact proviral DNA assay; LURE, latent cell capture; NFL, near-full length; PCR, polymerase chain reaction; Phep-Seq, phenotypic and proviral sequencing; PRIP-Seq, parallel HIV-1 RNA, integration site and proviral sequencing; STIP-Seq, simultaneous TCR, integration site and provirus sequencing; TILDA, Tat/rev Induced Limiting Dilution Assay.

HUMAN IMMUNODEFICIENCY VIRUS GENOMES CAN BE EASILY QUANTIFIED BUT ARE NOT EQUALLY IMPORTANT

The simplest and most cost-effective way to measure the frequency of HIV-infected cells is to use PCR-based assays that quantify HIV genomes in a given number of cells, usually peripheral blood mononuclear cells or isolated CD4 T cells. Several quantitative PCR-based approaches have been developed to quantify a variety of HIV DNA forms including integrated HIV DNA [15–17], 1- and 2-long terminal repeat (LTR) circles [18] and unintegrated linear HIV DNA [19]. Quantification of “total HIV DNA” (which includes all forms) is typically performed by amplifying a short and highly conserved region of the HIV genome (usually LTR or gag) and has been extensively used to easily estimate the magnitude of the HIV reservoir. Because they are fast, simple, and target the most abundant form of HIV persistence, HIV DNA quantified PCR (qPCR) can be used on a small number of cells and have been the standard in large clinical studies for years [20–23].

Most HIV genomes in people on long-term ART are integrated [24], and measures of integrated proviruses may represent a better surrogate of the HIV reservoir, particularly in people taking integrase inhibitors or on relatively short-term ART in whom unintegrated genomes are still detectable at substantially high levels [20]. Quantification of integrated HIV DNA can also be achieved by PCR, by combining an HIV-specific primer with primers specific for Alu regions that are distributed throughout the entire human genome [15–17]. However, a major limitation of both total and integrated HIV DNA qPCR-based approaches is that they greatly overestimate the frequency of cells harboring genetically intact HIV proviruses, because they amplify only a small region of the viral genome that was primarily selected for its limited genetic diversity among different variants [16]. This issue was partially resolved with the development of the intact proviral DNA assay (IPDA), a digital droplet (ddPCR)-based assay that uses strategically positioned primer sets in the env and the Ψ regions to discriminate between intact and defective proviruses. The IPDA provides a much more accurate estimate of the frequency of cells harboring potentially intact proviruses than regular PCR assays. As was verified by sequencing, 70% of the proviruses considered as intact by the IPDA were shown to indeed lack defects, whereas this value is less than 10% for regular PCR-based assays [25].

Although IPDA allows a more precise quantification of intact HIV genomes, some proviruses are misclassified as intact because they contain defects outside the 2 targeted regions. Such a misclassification could lead to underestimating the efficacy of cure strategies that target intact proviruses, because defective proviruses incorrectly considered as intact would not be eliminated [26]. Multiprobe assays targeting additional regions of the viral genome can be used to reduce the proportion of misclassified defective proviruses. For instance, the 5T-IPDA, which estimates the number of proviruses containing 5 specific primer-probe regions [27], results in intact provirus counts 10 times above the results of the quantitative viral outgrowth assay (QVOA) and 10 times below the results of the original IPDA. Similarly, a triplex digital PCR assay using a combination of IPDA and Q4PCR demonstrated an increased sensitivity for intact proviruses [28]. Nevertheless, the only way to ascertain the intactness of a provirus is to obtain its entire sequence. For this purpose, long-range PCRs of HIV genomes followed by sequencing, or near-full genome sequencing (nFGS) [14, 29] including the Full-Length Individual Proviral Sequencing (FLIPS) assay [30] and the Q4PCR approach [31], were developed. These assays are typically performed on end-point diluted samples to amplify a single genome per reaction using primer sets located in highly conserved regions of the 5′ and 3′ LTRs. Several rounds of amplifications by nested PCR lead to a near full-length amplicon that is sequenced by next-generation sequencing. The frequency of cells harboring intact proviruses can be calculated from the limiting dilution based on the assumption that all proviruses are quantitatively captured. However, this quantitative aspect has been recently challenged because viral genomes with large deletions may be preferentially amplified [32]. Novel enzymes with greater processivity may resolve this issue [33]. More importantly, the clonality of the HIV reservoir can also be estimated using this approach, assuming that multiple copies of the same genome are the product of cellular mitosis, as strongly suggested in multiple recent studies [34, 35].

All of these PCR-based assays, from the simplest “total HIV DNA” PCR to the more sophisticated IPDA or nFGS, use oligonucleotides that may introduce important variations in assay efficacy between samples due to the large genetic diversity of HIV. Whereas qPCR for LTR-gag can detect multiple clades [16], IPDA and nFGS primers were primarily designed to amplify clade B proviruses, and distinct primer sets are required for other clades. Another important limitation of these approaches is that none of them inform on the inducibility or the replication competence of the provirus. This aspect is particularly critical given the emerging data demonstrating that a substantial fraction of intact genomes is integrated in regions of the chromatin that are not permissive to gene expression [11, 36], particularly in people on ART for many years. Therefore, it is possible, in theory, that an individual with a large frequency of cells carrying genetically intact HIV genomes may not experience HIV rebound upon ATI if these are deeply latent or integrated in regions that are poorly permissive to viral gene expression. Functional measures of the reservoir assessing the inducibility of latent HIV genomes are needed to circumvent this issue.

MOST INDUCTION ASSAYS DO NOT MEASURE REPLICATION COMPETENCE

Most HIV genomes persisting in people on long-term ART are transcriptionally and translationally silent. In vivo, low amounts of elongated or spliced transcripts can be detected in blood and tissues from most people on ART [37–40], but these are often produced by defective genomes [41, 42]. Similarly, low amounts of HIV proteins produced by a minute fraction of cells can be detected during ART [43, 44]. Ex vivo stimulation of CD4⁺ T cells isolated from people on ART results in the rapid induction of viral transcription and translation in much more cells, a small proportion having the ability to produce virions. Sophisticated assays have been developed to measure the frequency of cells that can produce these viral products upon stimulation. The Tat/rev Induced Limiting Dilution Assay (TILDA) uses reverse-transcription PCR to measure the frequency of cells that produce multispliced HIV RNA upon stimulation and therefore assesses the transcription competence of these cells [45]. Although TILDA captures cells that can produce multiply spliced and therefore fully elongated transcripts, it does not ensure that they can produce infectious virions. Similarly, RNA-Flow FISH approaches measure the frequency of cells able to produce different forms of viral transcripts upon stimulation [46, 47]. The HIV^RNA/Gag assay combines RNA fluorescent in situ hybridization for HIV gag and pol mRNAs with p24 protein detection (mRNA Flow-FISH) after stimulation [48]. Although these approaches have been instrumental to identify the magnitude or the transcription-competent reservoir and the nature of the cells that carry it, the detection of HIV RNA by FISH requires multiple steps and 2–3 days, which is not adapted to large cohort studies. Alternatively, the frequency of cells that can produce HIV proteins can be measured using the HIV-Flow assay, a simple approach to measure the translation competent reservoir [49]. After stimulation, cells are stained with 2 antibodies, each targeting a specific epitope of the p24 protein for increased specificity. Unlike the HIV^RNA/Gag assay, which is clade dependent because it requires the hybridization of gag- and pol-specific probes, HIV-Flow can be used with viruses from multiple clades [35], because p24 is well conserved at the protein level.

A common limitation to assays focused on detecting viral RNA and/or proteins is that transcription and translation do not necessarily lead to the production of infectious viral particles, so neither HIV^RNA/Gag assay nor HIV-Flow can inform on the replication competence of the proviruses [41]. Nonetheless, assays focusing on the translation competent reservoir dramatically enrich in full-length genomes because they do not capture the pool of cells harboring heavily deleted proviruses [33].

To measure the frequency of proviruses that can produce virions, Cillo et al [50] plated resting CD4⁺ T cells from people on ART in serial dilutions, stimulating them, and measuring virion-associated RNA in the supernatant by RT-qPCR. This approach showed that only 1.5% of proviruses in resting CD4⁺ T cells were reactivated to produce virions. However, this method did not allow for the assessment of the capacity of these virions to fully replicate. Indeed, the only existing methods to quantify cells harboring replication competent proviruses are the QVOA [3, 51] and its derivatives [52, 53]. These techniques are based on the in vitro activation of total or resting CD4⁺ T cells from people on ART to induce viral production. These cells are plated in limiting dilution and cultured with uninfected target cells that will amplify the virus, should the activated cell have produced infectious virions. The capacity for viral production is assessed by dosing the viral p24 protein in the culture supernatant by enzyme-linked immunosorbent assay (ELISA) and is used to measure the frequency of infected cells that harbor replication competent proviruses. The large number of cells and extensive time required to perform it can be an obstacle to the use of the QVOA; therefore, some modifications have been made to simplify the experimental process. For example, the use of cell lines permissive to replication instead of primary CD4⁺ T cells to amplify the virus requires less cells from donors and reduces the length of the experiment. Reverse-transcription PCR can be used to detect viral genomic RNA in the supernatant earlier than p24 by ELISA and is more sensitive [52]. Massanella et al [54] used an improved variation of these inducible assays to demonstrate that the frequency of cells producing inducible cell-free RNA correlated strongly with the QVOA.

The major limitation of all induction assays, including the QVOA, is that they underestimate the frequency of cells with inducible proviruses. Indeed, not all inducible proviruses will be expressed after a single round of activation, or even after several rounds [55]. To improve the efficiency of the activation, Wonderlich et al [53] developed the differentiation QVOA (dQVOA) and showed that inducing the differentiation of CD4⁺ T cells into effector memory cells significantly increases the frequency of cells producing replication competent viruses. This differentiation approach could be applied to all induction assays to enhance the frequency of cells producing viral transcripts, proteins, or virions.

NOVEL GENERATION ASSAYS TO UNDERSTAND THE RESERVOIR

These quantification techniques, although important to measure the size of the reservoir, do not provide qualitative information on the cells that carry these proviruses, nor on their chromatin or epigenetic environment, which are all impacting HIV latency reversal efficacy. During the past 5 years, a plethora of assays have been developed to address these gaps in knowledge. Just as for the aforementioned techniques, these methods are based on the detection of HIV genomes, transcripts, or proteins. However, they allow for a deeper understanding of the reservoir cells by using this detection not just to quantify them, but also to isolate and study them at a single-cell or single HIV genome level.

Deoxyribonucleic Acid-Based Assays

As discussed above, the advantage of DNA-based assays is that they do not require stimulation and capture the entire reservoir (inducible or not). Einkauf et al [56] developed the matched integration site and proviral sequencing (MIPSeq) to analyze the proviral sequence and its corresponding integration site in the human chromatin. Infected cells are diluted to a single infected cell per well, and whole-genome amplification (WGA) by multiple displacement amplification (MDA) is performed. Proviruses are studied by near full-length sequencing and integration sites by integration site loop amplification (ISLA). This combined approach showed that intact proviruses are enriched in nongenic chromosomal positions, suggesting that transcriptional latency provides a selective advantage for the long-term persistence of intact HIV genomes. This protocol was modified into the Parallel HIV-1 RNA, Integration site and Proviral Sequencing (PRIPSeq) by adding a step to assess viral transcription [57]. Biotinylated primers specific for HIV RNA are used to isolate and quantify viral transcripts. This demonstrated that after prolonged ART, the integration site profile of intact HIV proviruses is heavily biased toward heterochromatin locations and that these proviruses are less transcriptionally active, suggesting a selection of reservoir cells during therapy. The Phenotypic and Proviral Sequencing (PHEP-Seq) assay combines assessment of cell phenotype and proviral sequences [58]. Cells are labeled with oligonucleotide-tagged antibodies directed to cellular markers of interest, encapsulated in water-in-oil droplets, and singularly barcoded. Multiplex PCRs are then performed to amplify the oligonucleotides tags and the HIV DNA, which are subsequently sequenced. Hence, this approach can be used to identify cellular markers that are preferentially expressed by cells carrying putatively intact genomes and revealed that these cells express higher levels of immune checkpoint molecules and their ligands, as well as cell surface receptors associated with T-cell survival (CD28, IL-7R etc). In addition, using HIV DNA detection to study reservoir cells, the Focused Interrogation of cells by Nucleic acid Detection and Sequencing (FIND-Seq) assay enables the study of whole transcriptome of infected cells [59]. Single cells are separated by encapsulation within water-in-oil droplets and singularly barcoded, after which the polyadenylated RNAs are reverse transcribed into cDNA, generating single cells that retain both the genomic DNA and the linked whole transcriptome. Infected cells are identified using gag PCR and sorted separately from their uninfected counterparts and used for Whole Transcriptome Amplification (WTA). Thereby, Clark et al [59] identified a distinctive host gene expression pattern in HIV DNA + cells in people on long-term ART associated with HIV silencing, cell survival, and cell proliferation. The ATAC with Select Antigen Profiling (ASAP-Seq) assay allows us to identify HIV-infected cells using a modified Assay for Transposase Accessible Chromatin at the single-cell level (scATAC-Seq) in combination with the detection of cell surface markers using oligo-conjugated antibodies [60]. This approach gave access to the epigenetic environment of proviruses integrated in open chromatin regions in association with the phenotype of the infected cell in unperturbed samples. This study elegantly confirmed the large phenotypic heterogeneity of the HIV reservoir.

Induction Assays

By isolating infected cells based on the detection of viral DNA, the aforementioned methods can be performed ex vivo, without activating the cells. However, this implies that they cannot attest to the inducibility of the proviruses. The following methods are based on the detection of viral transcripts or proteins upon cell stimulation and therefore only quantify and study inducible reservoirs. The HIV-1 SortSeq consists in the hybridization of HIV RNA-specific probes to sort stimulated infected cells using flow cytometry [61]. The transcriptome of these cells can then be obtained using scRNA-Seq, which revealed that aberrant host gene transcription occurs in HIV-infected cells. More recently, Collora et al [62] conducted a report on the Expanded CRISPR compatible CITE-Seq (ECCITE-Seq) assay, an approach that concomitantly captures surface protein expression, cellular transcriptome, HIV-RNA, and TCR sequences from individual infected cells. They observed that HIV genomes are enriched in GzmB + cytotoxic effector memory Th1 cells that express Bcl2 and SERPINB9, which could potentially explain their resistance to cell death. Because cells that have the potential to produce viral proteins may better reflect the reservoir that causes viral rebound upon ART interruption, several assays have been developed to characterize the translation-competent reservoir. In the latent cell capture (LURE) assay, cells expressing HIV envelope proteins upon stimulation are labeled with a cocktail of anti-Env broadly neutralizing antibodies and sorted for subsequent scRNA-Seq [63]. Gene-expression analysis revealed that these cells share a transcriptional profile that includes expression of genes implicated in silencing the virus. Near-full length (NFL)-HIV-Flow consists in a single-index cell sorting and NFL sequencing of the proviral genome harbored by cells producing p24 upon stimulation [33]. By phenotypically characterizing cells with inducible and translation-competent reservoir while simultaneously assessing intactness and defects of the corresponding proviruses, Dufour et al [33] identified the integrin α₄β₁ (VLA-4) as a marker of CD4⁺ T cells carrying intact and inducible HIV genomes. Clonality of reservoir cells can also be assessed using this approach. Using TCR sequencing in translation-competent reservoir cells, Gantner et al [35] showed that clonal expansions highly contribute to the persistence of the HIV reservoir and that reservoir cells displaying a differentiated phenotype are the progeny of infected central memory cells undergoing antigen-driven clonal expansion during ART. More recently, STIP-Seq (Simultaneous TCR, Integration site and Provirus sequencing), which adds WGA by MDA of single-sorted cells following HIV-Flow protocol [34], allowed a multimodal assessment of clonality of translation-competent reservoir cells combining integration site amplification by ISLA, NFL proviral sequencing, and TCRβ chain sequencing. Integration site sequencing revealed that cell clones with predicted pathogen specificity can harbor inducible proviruses integrated into cancer-related genes.

REMAINING CHALLENGES AND FUTURE DIRECTIONS

Which assay should we use? There is no simple answer to this question, and none of the measures described above can be qualified as the “best” (Table 1). The choice of the most adapted reservoir assay heavily depends on the question to be addressed [64]. When assessing the efficacy of therapeutic strategies aimed at reducing the size of the HIV reservoir, it is critical to select assays that are sensitive, reproducible, and that measure a form of viral persistence targeted by the clinical approach evaluated. For instance, PCR-based assays that measure total HIV genomes essentially capture defective genomes that cannot be reactivated to produce viral proteins and that are minimally affected by shock and kill interventions. Similarly, assays that measure intact genomes without assessing their inducibility may not be adapted, particularly in people on long-term ART in whom most intact proviruses may be epigenetically repressed [36] and insensitive to latency reversal. Although defective proviruses cannot cause rebound, some maintain the ability to produce viral transcripts and proteins [33, 42]. Therefore, they can be targeted by specific cure strategies (like shock and kill) and should not be overlooked [65, 66]. Proviruses that can cause viral rebound, which are the ones most investigators are interested in, should be endowed with 2 characteristics: genetic intactness and inducibility (Figure 2). To date, an assay that would measure these rare genomes in a quantitative and high-throughput manner is not available. Although PRIPseq combines near-full genome sequencing (for intactness) and viral RNA measures (for inducibility) [57], this elegant assay has been mostly used to study the spontaneous transcriptional activity of the reservoir and not its inducibility, and it is cumbersome and relatively expensive. In addition, although single cell-based assays can provide critical information on the biology of the reservoir, the efficiency of detection of infected cells using these sophisticated approaches is generally low and the numbers of infected cells studied small. Therefore, additional assays and a larger number of samples are needed to confirm the findings generated using these single-cell methods.

Table 1.

Methods to Quantify and Characterize the HIV Reservoir

Assay	Reference	Main Type of Technology	Target	HIV Genomes Captured	Advantage	Practicality	Intactness/Inducibility
DNA PCR (total or integrated)	O’Doherty et al, 2002 [15]; Vandergeeten et al, 2014 [16]	qPCR or ddPCR	Total or integrated HIV DNA	All	Provides a frequency of infected cells	Fast, easy, requires small number of cells	No/No
IPDA	Bruner et al, 2019 [25]	ddPCR	Intact and defective HIV DNA	Intact	Captures and quantifies intact proviruses	Faster, cheaper and easier than full-length sequencing	Yes/No
FLIPS	Hierner et al, 2017 [30]	long range PCR + sequencing	Intact and defective HIV DNA	Intact	Gives exact sequence of the provirus	Time consuming and expensive	Yes/No
TILDA	Procopio et al, 2015 [45]	Limiting dilution + RT-PCR	tat/rev RNA	Transcription competent	Measures HIV transcription inducibility	Fast, requires small number of cells	No/Yes
QVOA	Finzi et al, 1997 [3]	Cell culture in limiting dilution + ELISA	p24 in supernatant	Replication competent	Only approach that quantifies the replication competent reservoir	Time consuming and expensive	Yes/Yes
HIV GAG/RNA assay	Baxter et al, 2017 [48]	Flow cytometry	viral RNA + p24 protein	Translation competent	Measures HIV translation inducibility	Requires large amount of cells.	No/Yes
HIV-Flow	Pardons et al, 2019 [49]	Flow cytometry	p24 protein	Translation competent	Measures HIV translation inducibility	Requires large amount of cells. Fast and cost-effective	No/Yes
NFL-HIV Flow	Dufour et al, 2023 [33]	Flow cytometry single-cell sorting + long range PCR + sequencing	p24 protein	Intact and translation competent	Assesses translation inducibility, intactness, cell phenotype and clonality	Expensive, time consuming, requires large amount of cells	Yes/Yes
STIP-Seq	Cole et al, 2021 [34]	Flow cytometry single-cell sorting + WGA + sequencing	p24 protein	Intact and transcription competent	Assesses translation inducibility, intactness, cell phenotype, clonality, TCR sequence and integration site	Expensive, time consuming, requires large amount of cells	Yes/Yes
LURE	Cohn et al, 2018 [63]	Flow cytometry single-cell sorting + scRNAseq	Env protein	Intact and translation competent	Assesses translation inducibility, intactness, cell clonality and TCR sequence	Expensive, time consuming, requires large amount of cells	Yes/Yes
SORT-Seq	Liu et al, 2020 [61]	Flow cytometry single-cell sorting + scRNAseq	HIV RNA	Intact and transcription competent	Assesses intactness and cellular and viral transcriptomes	Expensive, time consuming, requires large amount of cells	No/Yes
ECCITE-Seq	Collora et al, 2022 [62]	Single-cell encapsulation in droplet + scRNAseq + antibody-derived Tags	HIV RNA	Transcription competent	Assesses cellular and viral transcriptomes, cell phenotype, TCR sequence and clonality	Expensive, time consuming, requires large amount of cells	No/Yes
PHEP-Seq	Sun et al, 2023 [58]	Single-cell encapsulation in droplet + antibody-derived Tags + multiplex PCR	HIV DNA	Intact	Assesses intactness, cell phenotype and clonality	Expensive, time consuming, requires large amount of cells	Yes/No
FIND-Seq	Clark et al, 2023 [59]	Single-cell encapsulation in droplet + PCR + RNAseq	HIV DNA	All	Assesses cell transcriptome	Expensive, time consuming, requires large amount of cells	No/No
MIP-Seq	Einkauf et al, 2019 [56]	Single-cell dilution + WGA + ISLA	HIV DNA	Intact	Assesses intactness and integration site	Expensive, time consuming, requires large amount of cells	Yes/No
PRIP-Seq	Einkauf et al, 2022 [57]	Single-cell dilution + WGA + ISLA + RT-PCR	HIV DNA + RNA	Intact	Assesses intactness, viral transcriptome, and integration site	Expensive, time consuming, requires large amount of cells	Yes/No
ASAP-Seq	Wu et al, 2023 [60]	Single-cell encapsulation in droplet + ATACseq + antibody-derived Tags	HIV DNA	All	Assesses the epigenetic environment of the provirus and cell phenotype	Expensive, time consuming, requires large amount of cells	No/No

Abbreviations: ASAP-Seq, ATAC with select antigen profiling; ATAC-Seq, assay for transposase accessible chromatin with high-throughput sequencing; CITE-Seq, cellular indexing of transcriptomes and epitopes by sequencing; ddPCR, droplet digital polymerase chain reaction; DNA, deoxyribonucleic acid; ECCITE-Seq, expanded CRISPR compatible CITE-Seq; ELISA, enzyme-linked immunosorbent assay; FIND-Seq, focused interrogation of cells by nucleic acid detection and sequencing; FLIPS, Full-Length Individual Proviral Sequencing; IPDA, intact proviral DNA assay; HIV, human immunodeficiency virus; ISLA, integration site loop amplification; LURE, latent cell capture; MIP-Seq, matched integration site and proviral sequencing; NFL, near-full length; PHEP-Seq, phenotypic and proviral sequencing; PRIP-Seq, parallel HIV-1 RNA, integration site and proviral sequencing; RNA, ribonucleic acid; RT-PCR, reverse-transcription PCR; STIP-Seq, simultaneous TCR, integration site and provirus sequencing; TCR, T-cell receptor; TILDA, Tat/rev Induced Limiting Dilution Assay; qPCR, quantitative PCR; QVOA, quantitative viral outgrowth assay; WGA, whole-genome amplification.

Figure 2.

Properties of human immunodeficiency virus (HIV) genomes that can cause viral rebound. Viral rebound is caused by the resurgence of viral production from genetically intact and inducible viral genomes. Inducibility of the provirus is regulated by several factors including the sites of integration within the human chromatin, epigenetic regulations, and cell activation status. Although all of these parameters can be measured by different assays, a simple approach that would simultaneously measure the intactness and inducibility of HIV genomes in people on antiretroviral therapy is still lacking.

Novel generation assays that could simultaneously measure the genetic intactness of the provirus and their inducibility, ideally using viral proteins or virions rather than viral transcripts as a readout, are warranted. Indeed, it is well described that viral transcription does not always result in the production of HIV proteins or virions [38]. Even the production of HIV proteins is not a good surrogate of infectious viral particles production: most genomes that produce p24 upon stimulation are genetically defective, particularly in the psi/MSD region [33, 34], a defect that was also recently reported in virions causing nonsuppressive viremia [67, 68]. Of note, although these defective genomes cannot produce infectious viral particles and are therefore irrelevant for cure studies, they most likely contribute to the residual inflammation reported in people on ART and should be considered.

The IPDA [25], which is now widely used in clinical studies, presents several important advantages compared to other approaches that measure the intactness of HIV genomes: it does not require sequencing and can be applied to relatively large numbers of samples in a short time. However, because it interrogates only 2 small regions of the viral genomes, it may misclassify some proviruses with defects in other viral genes or underestimate intact reservoir size because of within-host HIV diversity [69]. In addition, although some efforts have been developed to adapt IPDA to other clades [70], it has been validated for clade B viruses only.

Although near full-length genome amplification followed by sequencing is sometimes presented as the only approach to assess HIV genome intactness in a definitive manner, it presents 2 limitations. First, some long-range PCR may be inefficient and miss up to 70% of full-length proviruses due to amplification failure at the initial PCR step [32], although novel versions of these assays using more processive enzymes may resolve this issue. A second limitation, which is common to all assays that rely on sequencing to predict function, is the fact that defect criteria used in most intactness prediction algorithms are largely subjective. There is a critical need to experimentally verify these criteria by reconstructing more infectious clones from the viral reservoir [14] to improve the accuracy of prediction algorithms.

CONCLUSIONS

The recent development of sophisticated assays that measure not only the amount of proviruses but also their genetic intactness together with parameters that may impact their inducibility has profoundly changed our view of the viral reservoir. Although current scalable measures of HIV persistence do not combine these 2 parameters, the extraordinary progress that has been made in this area during the past 10 years suggests that this should be soon a reality.

Notes

Acknowledgments. We thank Remi Fromentin for critical reading of the manuscript.

Financial support. This work was supported by the Canadian Institutes of Health Research (Operating Grant 451304 and Team Grant HB2-164064, Canadian HIV Cure Enterprise [CanCURE]; to NC), the National Institute of Allergy and Infectious Diseases (Grant UM1AI164560, Delaney AIDS Research Enterprise [DARE] to Cure HIV), and the Réseau Sida et Maladies Infectieuses du Fonds de Recherche du Québec - Santé (FRQ-S).