Cell Cycle Arrest by Vpr in HIV-1 Virions and Insensitivity to Antiretroviral Agents

See allHide authors and affiliations

Science  10 Jul 1998:
Vol. 281, Issue 5374, pp. 266-269
DOI: 10.1126/science.281.5374.266


Expression of human immunodeficiency virus–type 1 (HIV-1) Vpr after productive infection of T cells induces cell cycle arrest in the G2 phase of the cell cycle. In the absence of de novo expression, HIV-1 Vpr packaged into virions still induced cell cycle arrest. Naturally noninfectious virus or virus rendered defective for infection by reverse transcriptase or protease inhibitors were capable of inducing Vpr-mediated cell cycle arrest. These results suggest a model whereby both infectious and noninfectious virions in vivo, such as those surrounding follicular dendritic cells, participate in immune suppression.

The HIV-1 vpr gene encodes a 14-kD nuclear protein (Vpr) that is expressed within infected cells and is packaged into virions (1). Vpr is nonessential for viral replication in T cell lines and activated peripheral blood lymphocytes (PBLs) in vitro but it is necessary for efficient infection of nondividing cells such as macrophages (2, 3). Simian immunodeficiency virus–induced disease progression in macaques was attenuated when either vpr alone or both vpr and the related gene vpx were mutated, indicating that Vpr plays an important role in viral pathogenesis (4). Functions ascribed to Vpr include transport of the viral core into the nucleus of nondividing cells (3) and up-regulation of viral gene expression (5). One particularly intriguing function of Vpr is the ability to induce cell cycle arrest at the G2checkpoint in a variety of mammalian cells, including human PBLs (6–8). This cell cycle arrest is characterized by alterations in the activation and phosphorylation state of Cdc2 kinase (7, 8) and resembles the G2checkpoint induced by genotoxic agents (9). Over time, virally infected cells arrested in the G2 phase by Vpr die by apoptosis (10).

In previous studies of Vpr-mediated cell cycle arrest, we used HIV-1 bearing the murine Thy 1.2 reporter gene to allow concurrent visualization of infected cells by cell surface staining for Thy 1.2 and cell cycle status by staining with propidium iodide (9,10). In some experiments, there was significant cell cycle arrest even in cells that were not productively infected (Thy 1.2), particularly at early time points before detectable Thy 1.2 expression (11). We tested the possibility that virion-associated Vpr can mediate cell cycle arrest by generating a vesicular stomatitis virus G protein (VSV-G) pseudo-typed HIV-1 that carried Vpr but was incapable of synthesizing it because of a frameshift mutation within vpr[HIV-1NL4-3-thy env(-)-vprX]. Epitope-tagged Vpr was supplied to the virion by cotransfection, resulting in trans-complementation [HIV-1NL4-3-thy env(-)vprX + BSVpr] (Fig. 1A) (12). Virion particles formed in this fashion contained Vpr in amounts comparable to wild-type virus [HIV-1NL4-3-thy env(-)]. Infection with virions complemented in trans with Vpr (VprX mutant + Vpr in trans) resulted in cell cycle arrest in the Thy 1.2+population (G1/G2 = 0.26) compared with the Thy 1.2 population (G1/G2= 1.8) and with infection by the vprX mutant (G1/G2 = 1.9) (Fig. 1B). The level of arrest induced by virion-associated Vpr alone was typically less than that observed with wild-type virus capable of de novo Vpr production (Fig. 1B), probably because of higher levels of Vpr expressed in cells de novo. Similar results were observed after infection of primary human PBLs and SupT1 T cells with an HIV-1 vector that packages Vpr but whose genome is defective for de novo expression of all HIV-1 genes (13). Virions pseudotyped with a CXCR-4 tropic HIV-1 envelope also induced cell cycle arrest (14) (Fig. 2B). The extent of G2 arrest induced by virion-associated Vpr varied depending on the virion preparation and cell type. However, HIV-1 particles prepared without envelope, but still containing Vpr within the viral core, did not induce G2 arrest (15). Thus, cell cycle arrest required both Vpr and viral entry into cells and was not the result of contaminating soluble Vpr in our virion preparations. Significantly, the cell cycle arrest observed with virion-associated Vpr was observed at low multiplicities of infection (MOI = 0.15; Fig. 1B) (16).

Figure 1

Virus complemented with Vpr in trans can induce cell cycle arrest. (A) Western blot analysis of HIV-1NL4-3-thy env(−) vprX (lane 1), HIV-1NL4-3-thy env(−) (lane 2), and HIV-1NL4-3-thy env(−) vprX + BSVpr (lane 3) for the presence of Vpr in virions (28). Molecular size standards are indicated on the left. Upper arrow indicates Vpr fused at the NH2-terminus to the influenza hemagglutinin nonapeptide expressed by BSVpr (12) and has appeared as a doublet in multiple experiments. Lower arrow indicates wild-type Vpr expressed from HIV-1NL4-3-thy env(−). (B) Cell cycle profile, 24 hours after infection, of HeLa cells infected with HIV-1NL4-3-thy env(−) vprX (vprX mutant), HIV-1NL4-3-thy env(−) (wild type), or HIV-1NL4-3-thy env(−) vprX + BSVpr (vprX mutant + Vpr in trans) (29).

Figure 2

Vpr can induce cell cycle arrest in the presence of RTI. (A) HeLa cells were infected with wild-type virus pseudotyped with VSV-G. One set of samples was treated with AZT and nevirapine (+RTI) (18). At the indicated times after infection, cells were analyzed for Thy 1.2 expression and DNA content (29). Data shown are representative of three separate experiments. (B) HeLa CD4 cells were infected with wild-type virus pseudotyped with HIV-1LAI envelope (18) and treated as in (A) except cells were synchronized in the G1 phase by a double thymidine block at the time of infection (9). Experiment shown is representative of three separate experiments.

We tested whether infection with viruses blocked at reverse transcription by zidovudine (AZT) and nevirapine, which are nucleoside and nonnucleoside reverse transcriptase inhibitors (RTIs) (17), respectively, would still result in cell cycle arrest. Infection with wild-type virus in the absence of RTI produced Thy 1.2+ cells that were arrested in the G2 phase (Fig. 2A). Most of the Thy 1.2 cells were also arrested at the 18-hour time point, suggesting early arrest before expression of viral genes. Infection with wild-type virus in the presence of RTI resulted in few Thy 1.2+ cells, which indicates that these drugs were effective in preventing productive infection (18) (Fig. 2A). Nevertheless, there was a significant increase in the G2 population evident 18 hours after infection and continuing to 43 hours after infection. At the later time point, the level of arrest was lower than that observed in non-RTI–treated infected cells: G1/G2 ratios of 0.35 and. 0.21, respectively. This difference was consistent in three separate experiments and is likely due to higher levels of Vpr synthesized in the non-RTI–treated infected cells. Similar results were observed when HeLa CD4 cells were infected with wild-type virus pseudotyped with the CXCR-4 tropic HIV-1LAI envelope (Fig. 2B).

Protease inhibitors (PIs) are anti-HIV-1 drugs that prevent viral replication by blocking cleavage of the Gag precursor, resulting in production of immature noninfectious particles (19). Compared with untreated virus preparations, wild-type virus or thevprX mutant produced in the presence of the PI indinavir contained reduced amounts of processed Gag proteins (20) (Fig. 3A). Wild-type virus, but not thevprX mutant, packaged Vpr within the virions in the presence or absence of indinavir (Fig. 3B). HeLa cells infected with wild-type virus produced in the presence of indinavir resulted in few Thy 1.2+ cells (Fig. 3C), demonstrating the relative inability of this virus to establish a productive infection. Nevertheless, these noninfectious viral particles were still capable of inducing cell cycle arrest, as indicated by the decreased G1/G2ratio relative to mock-infected or vprX mutant-infected cells. Similar results were obtained with the use of another PI, nelfinavir (11). In several experiments, the level of cell cycle arrest induced by wild-type virus treated with indinavir was generally comparable to the G2 arrest mediated by wild-type virus treated with RTI [G1/G2 ratios of 0.58 (Fig. 3C) and. 0.35 (Fig. 2A), respectively].

Figure 3

Ability of virions containing Vpr to cause cell cycle arrest is not affected by treatment with PI. VprX mutant or wild-type virus was produced in the absence (−) or presence (+) of indinavir (20). (A) The presence of Gag in virions was analyzed by Western blot analysis (28). (B) The presence of Vpr in virions was analyzed by Western blot analysis (28). (C) Cell cycle profile, 48 hours after infection, of HeLa cells that were mock infected or infected with vprX mutant or wild-type virus produced in the absence (−PI) or presence (+PI) of indinavir (29).

Most viral particles contained within HIV-1 preparations are noninfectious (21). In addition, a high proportion of noninfectious HIV-1 particles are found in infected patients in vivo (22), which raises the possibility that naturally occurring noninfectious particles could contribute to cell cycle arrest. We examined the potential of these particles to induce G2 arrest by examining the degree of cell cycle arrest in the Thy 1.2 population in an HIV-1–infected culture over time. Mock-infected HeLa cell cultures contained 35 to 37% of cells in the G2 phase at any one time (Table 1). At 18 hours after infection, 98% of the Thy 1.2+ population from HIV-1–infected cultures were in the G2 phase. Similarly, 85% of the Thy 1.2 cells were in the G2 phase. The increase over mock-infected cells in the proportion of Thy 1.2 cells in G2 was maintained at 42 hours after infection. It is unlikely that arrest of the Thy 1.2 population was due to productively infected cells that had yet to express Thy 1.2 on their cell surface because the percentage of Thy 1.2+ cells remained constant in this experiment at 49% at the two time points. Therefore, our interpretation of these experiments is that naturally existing noninfectious virus is capable of causing cell cycle arrest. In three independent experiments, we calculated that there were approximately equivalent numbers of G2-arrested Thy 1.2 cells and G2-arrested Thy 1.2+ cells, which suggests that in these virus preparations there were approximately equal numbers of noninfectious particles and infectious particles capable of causing cell cycle arrest. We note that our estimate of noninfectious particles that can induce arrest is less than the generally accepted estimate of noninfectious particles present in virus preparations (21). This is likely due to multiple factors responsible for loss of infectivity (21), some of which might preclude entry of Vpr into cells.

Table 1

Noninfectious HIV-1 virions can induce cell cycle arrest. HeLa cells were infected with wild-type virus sufficient to productively infect 49% of the cells. At the indicated times after infection, cells were analyzed for Thy 1.2 expression and DNA content (29). Results are expressed as percent of cells in G1 and G2. Similar results were obtained in three independent experiments.

View this table:

One consequence of Vpr activity may be the suppression of effective immune responses through, as we proposed (8, 10), inhibition of clonal expansion of T cells in response to antigen, or transcriptional dysregulation (23). By blocking key early steps in the immune response, Vpr would contribute to immune dysfunction and serve an important function for the virus by facilitating the persistence of virus-producing cells. Our results raise the possibility that high proportions of both infectious and noninfectious HIV-1 particles found in the blood and lymph nodes of infected individuals (22)—for example, those surrounding antigen-presenting follicular dendritic cells (24)—could arrest CD4+ T cells and contribute to the CD4+ immune dysfunction observed throughout the course of HIV-1 disease (25). Importantly, the cell cycle arrest observed with virion-associated Vpr is achievable at low MOIs, which indicates that the arrest of CD4+ T cells in vivo could be highly efficient. In addition, our results indicate that the current anti–HIV-1 drugs, including nucleoside and nonnucleoside RTIs and PIs, effectively block infectivity but do not affect this virion-associated function of Vpr. Thus, in patients undergoing HIV-1 antiviral therapy, any HIV-1 particles produced from residual infected cells at reservoir sites (26) would likely still have a T cell suppressive effect. This could contribute to the slow recovery of T cell function observed after effective antiviral suppression (27).


View Abstract

Stay Connected to Science

Editor's Blog

Navigate This Article