The role of integration and clonal expansion in HIV infection: live long and prosper

Integration of viral DNA into the host genome is a central event in the replication cycle and the pathogenesis of retroviruses, including HIV. Although most cells infected with HIV are rapidly eliminated in vivo, HIV also infects long-lived cells that persist during combination antiretroviral therapy (cART). Cells with replication competent HIV proviruses form a reservoir that persists despite cART and such reservoirs are at the center of efforts to eradicate or control infection without cART. The mechanisms of persistence of these chronically infected long-lived cells is uncertain, but recent research has demonstrated that the presence of the HIV provirus has enduring effects on infected cells. Cells with integrated proviruses may persist for many years, undergo clonal expansion, and produce replication competent HIV. Even proviruses with defective genomes can produce HIV RNA and may contribute to ongoing HIV pathogenesis. New analyses of HIV infected cells suggest that over time on cART, there is a shift in the composition of the population of HIV infected cells, with the infected cells that persist over prolonged periods having proviruses integrated in genes associated with regulation of cell growth. In several cases, strong evidence indicates the presence of the provirus in specific genes may determine persistence, proliferation, or both. These data have raised the intriguing possibility that after cART is introduced, a selection process enriches for cells with proviruses integrated in genes associated with cell growth regulation. The dynamic nature of populations of cells infected with HIV during cART is not well understood, but is likely to have a profound influence on the composition of the HIV reservoir with critical consequences for HIV eradication and control strategies. As such, integration studies will shed light on understanding viral persistence and inform eradication and control strategies. Here we review the process of HIV integration, the role that integration plays in persistence, clonal expansion of the HIV reservoir, and highlight current challenges and outstanding questions for future research.     

HIV infected CD4+ T cell clones are more stable than uninfected clones during long-term antiretroviral therapy

Although combination antiretroviral therapy (ART) blocks HIV replication, it is not curative because infected CD4+ T cells that carry intact, infectious proviruses persist. Understanding the behavior of clones of infected T cells is important for understanding the stability of the reservoir; however, the stabilities of clones of infected T cells in persons on long-term ART are not well defined. We determined the relative stabilities of clones of infected and uninfected CD4+ T cells over time intervals of one to four years in three individuals who had been on ART for 9–19 years. The largest clones of uninfected T cells were larger than the largest clones of infected T cells. Clones of infected CD4+ T cells were more stable than clones of uninfected CD4+ T cells of a similar size. Individual clones of CD4+ T cells carrying intact, infectious proviruses can expand, contract, or remain stable over time.     

The integrated genome of murine leukemia virus

The Southern gel filter transfer technique has been used to characterize the integrated genome of Moloney murine leukemia virus (M-MuLV) and the genomes of the endogenous viruses of the mouse. Study of 10 clones of rat cell independently infected by M-MuLV indicates a minimum of 15 integration sites into which the M-MuLV provirus can be inserted. No common integration site is observed among these clones. Clones productively infected by M-MuLV acquire multiple proviruses, whereas infected cells unable to produce virus contain only one M-MuLV provirus. Once established, the integrated genomes are stable for at least two years after initial infection.The use of M-MuLV probe allows detection of a spectrum of Eco RI-cleaved mouse DNA fragments containing endogenous MuLV genomes. DNAs of different inbred laboratory mouse strains yield similar patterns of provirus with each strain showing minor characteristic differences. In some instances, mouse cells infected by M-MuLV reveal additional proviruses beyond those seen in the uninfected cell. DNAs from three different M-MuLV-induced thymomas indicate, as in rat cells, multiple possible integration sites.     

Massive Depletion of Bovine Leukemia Virus Proviral Clones Located in Genomic Transcriptionally Active Sites during Primary Infection

Deltaretroviruses such as human T-lymphotropic virus type 1 (HTLV-1) and bovine leukemia virus (BLV) induce a persistent infection that remains generally asymptomatic but can also lead to leukemia or lymphoma. These viruses replicate by infecting new lymphocytes (i.e. the infectious cycle) or via clonal expansion of the infected cells (mitotic cycle). The relative importance of these two cycles in viral replication varies during infection. The majority of infected clones are created early before the onset of an efficient immune response. Later on, the main replication route is mitotic expansion of pre-existing infected clones. Due to the paucity of available samples and for ethical reasons, only scarce data is available on early infection by HTLV-1. Therefore, we addressed this question in a comparative BLV model. We used high-throughput sequencing to map and quantify the insertion sites of the provirus in order to monitor the clonality of the BLV-infected cells population (i.e. the number of distinct clones and abundance of each clone). We found that BLV propagation shifts from cell neoinfection to clonal proliferation in about 2 months from inoculation. Initially, BLV proviral integration significantly favors transcribed regions of the genome. Negative selection then eliminates 97% of the clones detected at seroconversion and disfavors BLV-infected cells carrying a provirus located close to a promoter or a gene. Nevertheless, among the surviving proviruses, clone abundance positively correlates with proximity of the provirus to a transcribed region. Two opposite forces thus operate during primary infection and dictate the fate of long term clonal composition: (1) initial integration inside genes or promoters and (2) host negative selection disfavoring proviruses located next to transcribed regions. The result of this initial response will contribute to the proviral load set point value as clonal abundance will benefit from carrying a provirus in transcribed regions.     

Insertion elements and transitions in cloned mouse mammary tumour virus DNA: further delineation of the poison sequences.

The provirus of mouse mammary tumour virus (MMTV) is reputed to contain sequences within the viral gag gene that prevent or inhibit its propagation as a recombinant DNA clone in Escherichia coli. Here we report the successful isolation of several lambda and plasmid clones comprising the 5' virus-host DNA junction fragments from integrated MMTV proviruses in BR6 mice. Although the lambda clones appeared intact, almost all of the plasmids were found to contain the bacterial insertion sequences IS1 or IS2 within a small region of the gag gene. One nondisrupted clone was recovered which had undergone multiple G to A transitions, some of which created stop codons in gag. These results have provided more precise information as to the location of the poison sequences and are discussed in relation to possible explanations for the phenomenon.     

Chromosomal position or virus mutation permits retrovirus expression in embryonal carcinoma cells

Retrovirus expression is restricted in embryonal carcinoma (EC) cells. To study how a virus can overcome this block, we selected and analyzed rare proviruses that are expressed in F9 cells. Our results indicate that provirus expression occurs by two different mechanisms: one provirus acquired a single base pair mutation in the retrovirus tRNA primer binding site, permitting provirus expression; expression of three proviruses was mediated by 5'-flanking DNA sequences. Surprisingly, five proviruses in 17 selected cell lines integrated into the same two distinct chromosomal regions, suggesting that the number of chromosomal positions in the cellular genome that allows virus expression is very limited. Our results suggest that genomic sequences that are actively transcribed in EC cells can be isolated by selection for retrovirus expression.     

Identification of a new common provirus integration site in gross passage A murine leukemia virus-induced mouse thymoma DNA.

The Gross passage A murine leukemia virus (MuLV) induced T-cell leukemia of clonal (or oligoclonal) origin in inoculated mice. To study the role of the integrated proviruses in these tumor cells, we cloned several newly integrated proviruses (with their flanking cellular sequences) from a single tumor in procaryotic vectors. With each of the five clones obtained, a probe was prepared from the cellular sequences flanking the provirus. With one such probe (SS8), we screened several Gross passage A MuLV-induced SIM.S mouse tumor DNAs and found that, in 11 of 40 tumors, a provirus was integrated into a common region designated Gin-1. A 26-kilobase-pair sequence of Gin-1 was cloned from two lambda libraries, and a restriction map was derived. All proviruses were integrated as a cluster in the same orientation within a 5-kilobase-pair region of Gin-1, and most of them had a recombinant structure of the mink cell focus-forming virus type. The frequency of Gin-1 occupancy by provirus was much lower in thymoma induced by other strains of MuLV in other mouse strains. Using somatic-cell hybrid DNAs, we mapped Gin-1 on mouse chromosome 19. Gin-1 was not homologous to 16 known oncogenes and was distinct from the other common regions for provirus integration previously described. Therefore, Gin-1 appears to represent a new common provirus integration region. The integration of a provirus within Gin-1 might be an important event leading to T-cell transformation, and the Gin-1 region might harbor sequences which are involved in tumor development.     

Biological characterization of human immunodeficiency virus type 1 clones derived from different organs of an AIDS patient by long-range PCR.

In order to characterize the biological properties of human immunodeficiency virus type 1 (HIV-1) variants from different tissues (peripheral blood mononuclear cells [PBMC], lymph node, spleen, brain, and lung) of one patient, we have chosen long-range PCR to amplify virtually full-length HIV proviruses and to construct replication-competent viruses by adding a patient-specific 5' long terminal repeat. To avoid selection during propagation in CD4+ target cells, we transfected 293 cells and used the supernatants from these cells as challenge viruses for tropism studies after titration on human PBMC. Despite differences in the V3 loop of the major variants found in brain and lung compared to lymphoid tissues all recombinant HIV clones obtained showed identical cell tropism and replicative kinetics. After infection of human PBMC these viruses replicated with similar kinetics, with a slow/low-titer, non-syncytium-inducing phenotype. In contrast to the prediction of macrophage tropism, drawn from the V3 loop sequence, none of these viruses infected monocyte-derived macrophages. The challenge of blood dendritic cells by these recombinant viruses in the presence of tumor necrosis factor alpha, granulocyte-macrophage colony-stimulating factor, and interleukin-4 resulted in a productive infection only after adding stimulated CD4+ T lymphocytes. Therefore, the biological properties of the HIV-1 variants derived from nonlymphoid tissue of this patient did not differ from those of HIV-1 variants from lymphoid tissue with respect to tropism for primary cells such as PBMC, macrophages, and blood dendritic cells.     

Construction of a full-length human T cell leukemia virus type I genome from MT-2 cells containing multiple defective proviruses using overlapping polymerase chain reaction

Human T cell leukemia virus type I (HTLV-I), the etiological agent of adult T cell leukemia, integrates into the host genome as a provirus. Multiple defective copies of the integrated provirus are often present in the host genome. For this reason it is difficult to clone the intact provirus from HTLV-I-infected cells using conventional techniques. Here, we used overlapping polymerase chain reaction (PCR) to construct a full-length provirus of HTLV-I directly from an HTLV-I-transformed cell line, MT-2, which contains multiple defective proviruses. First, four overlapping proviral HTLV-I fragments (1.4-3.9 kb each) were constructed from genomic MT-2 DNA using PCR. Next, the complete HTLV-I proviral DNA (9 kb) was generated from these fragments using asymmetric PCR and cloned into a plasmid vector. 293 T cells transfected with this plasmid produced virus-like particles, and we show that these particles are capable of infecting a human T cell line. We propose that this cloning technique constitutes a powerful tool for constructing infectious molecular clones from cells of patients infected with HTLV-I or other viruses.     

Human T-cell leukemia virus type 1 (HTLV-1) infection of mice: proliferation of cell clones with integrated HTLV-1 provirus in lymphoid organs

Human T-cell leukemia virus type 1 (HTLV-1) is suggested to cause adult T-cell leukemia after 40 to 50 years of latency in a small percentage of carriers. However, little is known about the pathophysiology of the latent period and the reservoir organs where polyclonal proliferation of cells harboring integrated provirus occurs. The availability of animal models would be useful to analyze the latent period of HTLV-1 infection. At 18 months after HTLV-1 infection of C3H/HeJ mice inoculated with the MT-2 cell line, which is an HTLV-1-producing human T-cell line, HTLV-1 provirus was detected in spleen DNA from eight of nine mice. No more than around 100 proviruses were found per 10(5) spleen cells. Cellular sequences flanking the 3' long terminal repeat (LTR) and the clonalities of the cells which harbor integrated HTLV-1 provirus were analyzed by linker-mediated PCR. The results showed that the flanking sequences are of mouse genome origin and that polyclonal proliferation of the spleen cells harboring integrated HTLV-1 provirus had occurred in three mice. A sequence flanking the 5' LTR was isolated from one of the mice and revealed the presence of a 6-nucleotide duplication of cellular sequences, consistent with typical retroviral integration. Moreover, PCR was performed on DNA from infected tissues, with LTR primers and primers derived from seven novel flanking sequences of the three mice. Data revealed that the expected PCR products were found from lymphatic tissues of the same mouse, suggesting that the lymphatic tissues were the reservoir organs for the infected and proliferating cell clones. The mouse model described here should be useful for analysis of the carrier state of HTLV-1 infection in humans.     

Host sequences flanking the HIV provirus.

A conserved property of retroviral proviruses is the presence of a direct repeat in the host DNA immediately flanking the viral sequence; each virus generates a repeat with a characteristic length. By sequencing the viral/host DNA junctions from five HIV-1 proviral clones, we have confirmed that integration of HIV results in the generation of a five basepair direct repeat. A target sequence in uninfected host DNA was analyzed to establish that the five basepair sequence flanking the provirus was present only once prior to integration. Of the five proviruses examined, two were found to have integrated in known repetitive sequence elements of the human genome; one in a Line-1 element and a second in satellite DNA.     

A flexible model of HIV-1 latency permitting evaluation of many primary CD4 T-cell reservoirs

Latently infected cells form the major obstacle to HIV eradication. Studies of HIV latency have been generally hindered by the lack of a robust and rapidly deployable cell model that involves primary human CD4 T lymphocytes. Latently infected cell lines have proven useful, but it is unclear how closely these proliferating cells recapitulate the conditions of viral latency in non-dividing CD4 T lymphocytes in vivo. Current primary lymphocyte models more closely reflect the in vivo state of HIV latency, but they are limited by protracted culture periods and often low cell yields. Additionally, these models are always established in a single latently infected cell type that may not reflect the heterogeneous nature of the latent reservoir. Here we describe a rapid, sensitive, and quantitative primary cell model of HIV-1 latency with replication competent proviruses and multiple reporters to enhance the flexibility of the system. In this model, post-integration HIV-1 latency can be established in all populations of CD4 T cells, and reactivation of latent provirus assessed within 7 days. The kinetics and magnitude of reactivation were evaluated after stimulation with various cytokines, small molecules, and T-cell receptor agonists. Reactivation of latent HIV proviruses was readily detected in the presence of strong activators of NF-κB. Latently infected transitional memory CD4 T cells proved more responsive to these T-cell activators than latently infected central memory cells. These findings reveal potentially important biological differences within the latently infected pool of memory CD4 T cells and describe a flexible primary CD4 T-cell system to evaluate novel antagonists of HIV latency.     

DNA methylation affecting the expression of murine leukemia proviruses.

The endogenous, vertically transmitted proviral DNAs of the ecotropic murine leukemia virus in AKR embryo fibroblasts were found to be hypermethylated relative to exogenous AKR murine leukemia virus proviral DNAs acquired by infection of the same cells. The hypermethylated state of the endogenous AKR murine leukemia virus proviruses in these cells correlated with the failure to express AKR murine leukemia virus and the lack of infectivity of cellular DNA. Induction of the endogenous AKR murine leukemia virus proviruses with the methylation antagonist 5-azacytidine suggested a causal connection between DNA methylation and provirus expression. Also found to be relatively hypermethylated and noninfectious were three of six Moloney murine leukemia virus proviral DNAs in an unusual clone of infected rat cells. Recombinant DNA clones which derived from a methylated, noninfectious Moloney provirus of this cell line were found to be highly active upon transfection, suggesting that a potentially active proviral genome can be rendered inactive by cellular DNA methylation. In contrast, in vitro methylation with the bacterial methylases MHpaII and MHhaI only slightly reduced the infectivity of the biologically active cloned proviral DNA. Recombinant DNA clones which derived from a second Moloney provirus of this cell line were noninfectious. An in vitro recombination method was utilized in mapping studies to show that this lack of infectivity was governed by mechanisms other than methylation.     

Quantification of unintegrated HIV-1 DNA at the single cell level in vivo

In the nucleus of HIV-1 infected cells, unintegrated HIV-1 DNA molecules exist in the form of one and two LTR circles and linear molecules with degraded extremities. In tissue culture they are invariably more numerous than the provirus, the relative proportion of integrated to unintegrated forms varies widely from ～1:1 to 1:10 and even over 1:100. In vivo, this ratio is unknown. To determine it, single nuclei from two infected patients with a known provirus copy number were microdissected, HIV DNA was amplified by nested PCR, cloned and individual clones sequenced. Given the extraordinary sequence complexity, we made the assumption that the total number of distinct sequences approximated to real number of amplifiable HIV-1 DNA templates in the nucleus. We found that the number of unintegrated DNA molecules increased linearly with the proviral copy number there being on average 86 unintegrated molecules per provirus.