4.1. Near Full-Length (NFL) Sequencing
The first set of assays, like single-genome/proviral sequencing assays, provide no real estimate of the replication-competent viruses. Therefore, new assays have been developed to look at the intactness of individual proviral genomes, allowing for a good characterization of the genetic composition and thus making a better estimation of the replication-competent fraction of the persisting viral reservoir. In 2013, Ho et al. presented a novel technique to generate NFL proviral sequences [14
]. Starting again with a limiting dilution to ensure the presence of a single proviruses per reaction, an outer PCR reaction was performed ranging from the 5′ LTR to the 3′ LTR. Next, inner PCRs targeting subgenomic regions were performed to verify hypermutations, followed by a set of inner PCRs amplifying overlapping regions of the HIV-1 genome of non-hypermutated HIV-1 positive wells. The resulting products were visualized on an agarose gel and cut out to send for direct Sanger sequencing followed by sequencing analysis.
The initial results gave novel insights in the composition of the HIV-1 latent reservoir, revealing that the frequency of replication-competent proviruses was at least 60-fold higher than expected when measuring via a standard quantitative Virus Outgrowth Assay (qVOA), thus indicating the severe underestimation of this cell culture assay [14
]. In 2016, a follow-up paper using the same technique by Bruner et al. studied proviruses present in non-induced CD4+ T cells from patients on ART [27
]. A vast majority of proviruses were harboring a variety of genetic defects such as large internal deletions or APOBEC3G-mediated mutations. The observed fractions of intact proviruses were very low, 5% in patients who started ART during acute infection and 2% for patients starting ART during chronic infection.
In 2017, two papers by Hiener et al. and Lee et al. presented a modified protocol, this time using a two-step nested PCR that amplifies a 9kb amplicon fragment (~92% of provirus), followed by deep next-generation sequencing (NGS) via Illumina [15
]. For each positive well, de novo assembled proviruses using these short reads were generated and aligned to an HXB2 reference genome for further downstream analysis. Each provirus was assessed on a list of varying criteria such as inversions, large internal deletions, premature stop codons, frameshift mutations, and the presence of mutations or deletions in either the major splice donor (MSD) site or the PSI. This NGS strategy allows for a simplified, faster, and more cost-effective assessment of NFL proviral genomes than using multiple internal primers to sequence via Sanger, while also reducing the chance of erroneously labeling a virus as defective due to primer mismatch failing to amplify that region (however, it is important to note that the primers used in nested NFL PCR reactions might still show mismatches despite their location in a conserved region).
These novel developments allow for an in-depth characterization of the viral reservoir, shedding a new light on the genomic composition and reservoir dynamics. Lee et al. studied several functionally polarized memory CD4+ T cells in five individuals of which Th1-polarized CD4+ T cells harbored the most intact [16
]. Meanwhile, Hiener et al. looked at the proviral composition across different differentiation stages of CD4+ T cells in six participants, identifying effector memory T cells as the subset with highest amount of intact sequences [15
]. Another study employing full-length NGS by Lu et al. sequenced proviral genomes from CD4+ T cells of 12 patients who underwent analytic treatment interruption (ATI) and compared those with results from VOA. They showed a partial overlap of intact sequences picked up from limiting dilution on lysed DNA from CD4+ T cells [28
]. This demonstrates that the intactness of the sequence can be linked to replication competence, however no overlap with rebounding sequences was found [28
]. Recent papers by Pinzone et al. and Lee et al. employed NFL sequencing to longitudinally study the evolution of the HIV-1 reservoir in different setups [26
]. The former, spanning several years, focused on the selective pressures driving the dynamics of the viral reservoir in four individuals on ART and revealed a stronger clearance of intact proviruses due to a negative selection pressure driven by their effective protein expression [26
]. The other paper by Lee et al. gave rare insights in the early stages of HIV-1 infection in four women infected with subtype C, observing the initial composition consisting of almost solely intact, non-identical proviral genomes and the increase in diversity of the proviral landscapes over time due to multiple factors such as truncation and hypermutation [29
]. Additionally, all these studies report the detection of proportions of identical proviral sequences in several patients on ART, suggesting a role for clonal expansion of infected cells in the maintenance of HIV-1 reservoir. Yet, these NGS assays can merely lead to assumptions on clonality, as the integration site to confirm clonal origins is still missing [24
The introduction of NFL sequencing provides a new level of understanding of the viral reservoir. Still, the costs associated with NGS due to use of limiting dilution are still high, and might explain why the aforementioned papers included, on average, six individuals, making more informative large-population studies too expensive [15
4.2. Quadruplex qPCR (Q4PCR)
In 2019, Gaebler et al. presented a novel method using a combination of quantitative PCR (qPCR) and NGS to examine the latent HIV-1 reservoir dubbed Q4PCR [17
This approach consists of a nested NFL PCR, but with an additional qPCR step between the outer and inner PCR reaction. The probes of their multiplex qPCR strategy cover four different regions of the HIV-1 region: packaging signal (PS), gag, pol, and env, which are optimized for optimal detection via in silico design using intact proviral sequences from the Los Alamos HIV Database. After the multiplex qPCR, only those reaction-positive for two probes are further selected for NFL PCR. Due to the inclusion of those four conserved regions, they can make a prediction on intactness. By adding this intermediate step, they can improve the efficiency of picking up intact or NFL proviruses by filtering out defective ones while still retaining qualitative data on non-sequenced wells (not observed on agarose gel comparison). The acquired intact sequences from six individuals were positive for any of two of the four probe combinations, as were 99% of the intact subtype B sequences on the Los Alamos HIV database, confirming the robustness of the assay design.
Gaebler et al. present a sensitive, specific, and informative assay, capable of an unbiased characterization of the HIV-1 latent reservoir when compared with regular NFL NGS. It could help to reduce the costs associated with NGS of HIV-1 proviruses in larger clinical studies. Of note, the Q4PCR is not a standalone absolute quantitative assay as compared with the IPDA, as the quantitative nature of the Q4PCR requires a prior limiting dilution gag qPCR coupled with a near full-length PCR protocol.
4.3. Multiple-Displacement Amplification (MDA)-Based Techniques
As pointed out, clonal proliferation is thought to be a major driver of the viral persistence of the HIV-1 reservoir, as suggested by detection of similar integration sites via ISS and clusters of identical sequences via the current full-length assays. In 2019, Einkauf et al. and Patro et al. published similar novel approaches, respectively dubbed matched integration site and proviral sequencing (MIP-seq) and MDA-SGS, to get combined information indisputably linking proviral integrity and integration sites [18
The setup is as follows: the first step includes an MDA at limiting dilution, which produces hundreds to thousands of long templates consisting of the proviral genome and the upstream and/or downstream host genomic sequence. Parts of these templates are subsequently used for viral sequencing by the SGS assay or NFL assay, and parts are used to go after the integration site with an ISS method of choice.
Both papers present the first combined data of proviruses and their chromosomal positions respectively on three and five participants. This offers a big change to earlier studies, which could only rely on assumptions of clonality based on expansion of identical viral sequences [24
]. In addition to sequencing, Einkauf et al. also analyzed the chromatin accessibility and gene expression in proximity to integration sites to get a comprehensive overview of the persisting reservoir. On the basis of their observations, the researchers suggest the occurrence of deep latency, where patients on long-term ART would develop a reservoir consisting of less activatable, yet still intact proviruses due to their location in either non-genic/pseudogenic regions, their reverse orientation to the host gene, or location in less accessible chromatin [18
]. Patro et al. used the new assay to explore clonality of proviral genomes in patients on ART, being either the result of clonal expansion of infected cells or the result of viral genetic bottlenecks during the absence of ART [19
]. Like Einkauf et al., they noted the importance of reverse orientation to the host gene of integrated intact clones. Next, the authors were able to match an intact expanded clone to an NFL provirus picked up via VOA, confirming its replication-competent state.
The arrival of MDA-based techniques will allow for a better characterization of the HIV-1 reservoir. It allows the most informative view of the proviral landscape to date, providing insights in clonality and the replication-competent state of the persistent reservoir of individuals on ART and, more importantly, links these two together. Still, the associated high costs might hamper the wide use of this technique in large-scale studies.