Arsenic Counteracts Human Immunodeficiency Virus Type 1 Restriction by Various TRIM5 Orthologues in a Cell Type-Dependent Manner

Abstract

Arsenic trioxide (As₂O₃) increased human immunodeficiency virus type 1 (HIV-1) infectivity when particular Homo sapiens and Cercopithecus aethiops cell lines were used as targets. Knockdown of human TRIM5α by RNA interference eliminated the As₂O₃ effect, demonstrating that the drug acts by modulating the activity of this retroviral restriction factor. In contrast, HIV-1 infectivity in target cell lines from other primate species (Cercopithecus tantalus, Macaca mulatta, and Aotus trivirgatus) was not increased by As₂O₃, despite the potent TRIM5-dependent HIV-1 restriction activity that these cells exhibit. To determine if As₂O₃ responsiveness is characteristic of particular TRIM5 orthologues and not others, TRIM5 cDNAs from these five primate species were transduced into cat fibroblasts, which lack endogenous HIV-1 restriction activity and, therefore, responsiveness to As₂O₃. In this context, the HIV-1 restriction activity conferred by all TRIM5 orthologues was largely eliminated by As₂O₃. The effect of As₂O₃ on HIV-1 restriction is thus shared by different TRIM5 orthologues but dependent on factors specific to the cell line in which TRIM5 is expressed.

Human immunodeficiency virus type 1 (HIV-1) infectivity is inhibited in cells from Old World monkeys, such as rhesus macaques and African green monkeys, and in New World owl monkeys (7). The block to replication occurs early, at a point after virus entry but before integration of the viral DNA (5, 9). The cellular factor responsible for retroviral restriction in Old World monkeys is TRIM5α (18), a member of the tripartite motif family of proteins (14). In owl monkeys, characterization of the HIV-1 CA-binding protein cyclophilin A (CypA) led to the discovery of TRIMCyp (16), an HIV-1-specific restriction factor that was generated by cyclophilin A cDNA retrotransposition into the TRIM5 locus (11, 16). Though HIV-1 replicates well enough in humans to cause AIDS, human TRIM5α possesses a modest anti-HIV-1 restriction activity (6, 8, 13, 18, 21) that perhaps contributes to the long clinical latency observed in most HIV-1-infected people (12).

When added to target cells at the time of infection, arsenic trioxide (As₂O₃) stimulates retroviral infectivity (3, 19). The As₂O₃ effect has only been observed in the context of target cells bearing TRIM5α-mediated retrovirus restriction activity (2, 3). Consistent with the hypothesis that As₂O₃ counteracts TRIM5α-mediated restriction, the transfer of human TRIM5α cDNA to nonrestrictive mouse fibroblasts rendered viral titers on these cells responsive to As₂O₃ (8). Conversely, selection of a human cell line for loss of retrovirus restriction activity resulted in cells that were As₂O₃ unresponsive (15).

As₂O₃ counteracts HIV-1 restriction in some cell lines.

We previously reported that HIV-1 titers on certain restrictive target cell lines were As₂O₃ responsive but that the drug had no effect on HIV-1 titers in other, equally restrictive cell lines (2). To examine the effect of As₂O₃ in more detail, a panel of five primate cell lines that restrict HIV-1 to various extents was tested for the potential of As₂O₃ to relieve HIV-1 restriction. As previously described (2), cells were exposed to vesicular stomatitis virus glycoprotein (VSV-G)-pseudotyped HIV-1_NL4-3 in which the envelope was deleted and green fluorescent protein (GFP) reporter gene replaced nef (HIV-1_GFP). Each cell line was challenged with a constant amount of virus such that a few percent of cells were transduced and rendered GFP positive. As₂O₃ was prepared as described previously (3) and added to media for the first 12 h of infection at the concentrations indicated in Fig. 1.

FIG. 1.

As₂O₃ counteracts HIV-1 restriction in Vero and TE671 cells but not in CV1, OMK, or FRhK4 cells. Cells were infected with VSV-G-pseudotyped HIV-1_NL4-3 bearing an env deletion and GFP in place of nef (HIV-1_GFP), such that <5% of target cells were infected. The indicated concentrations of As₂O₃ were maintained for the first 12 h that cells were exposed to the virus, at which time the medium was replaced without the drug. The percentage of GFP-positive (infected) cells was determined at 48 h postinfection by flow cytometry.

As₂O₃ increased HIV-1 titer 10-fold in Vero cells (Cercopithecus aethiops), with a maximal effect observed at a concentration of 12.5 μM (Fig. 1). HIV-1 titer was increased threefold by As₂O₃ when human TE671 cells were used as targets, with a maximal drug effect at 5 μM (Fig. 1). The decline in HIV-1 titer with higher As₂O₃ concentrations can be attributed to target cell toxicity (3). In contrast, As₂O₃ did not stimulate HIV-1 titer in Macaca mulatta FRhK4 cells, Aotus trivirgatus OMK cells, or Cercopithecus tantalus CV1 cells (Fig. 1), even at drug concentrations high enough to cause grossly evident target cell death in these particular cell lines (Fig. 1 and data not shown). Using the XTT assay (3), no significant differences in the sensitivities of the cell lines to As₂O₃ were detected that would explain the differential drug responsiveness.

TRIM5α is required for the HIV-1 stimulatory effect of As₂O₃.

To determine if the positive effect of As₂O₃ on the HIV-1 titer is related to effects of the drug on TRIM5α, we created a TE671 cell line in which TRIM5 was knocked down by stable transduction of a short hairpin RNA (shRNA), as previously described (18). Successful TRIM5 knockdown in these cells was documented previously by demonstrating complete removal of the potent restriction activity that targets N-tropic murine leukemia virus (MLV) (17). Cells transduced with shRNA targeting luciferase were generated concurrently to serve as a control.

TRIM5 knockdown resulted in a threefold increase in HIV-1 titer (Fig. 2). The same magnitude of increase in HIV-1 infectivity was observed with the control, luciferase knockdown cells when infection was conducted in the presence of 2.5 μM As₂O₃ (Fig. 2). However, As₂O₃ had no effect on HIV-1 titer in TRIM5 knockdown cells (Fig. 2), demonstrating that As₂O₃ counteracts HIV-1 restriction mediated by TRIM5α. Although we knocked down TRIM5 expression in FRhK4 cells (1) and others did the same in a CV1-derived cell line (8), multiple attempts to knock down TRIM5 in Vero cells were unsuccessful, precluding confirmation of these results in another As₂O₃-responsive cell line.

FIG. 2.

TRIM5 knockdown in human cells eliminates the stimulatory effect of As₂O₃ on HIV-1 infectivity. TE671 cells were transduced with pSUPER.retro.puro (Oligoengine), encoding an shRNA against TRIM5 (target sequence, 5′-GCCUUACGAAGUCUGAAAC-3′) or against luciferase as a control (target sequence, 5′-CGUACGCGGAAUACUUCGA-3′). After selection in puromycin, cells were transduced with HIV-1_GFP in the presence or absence of 2.5 μM As₂O₃, and the percentage of GFP-positive (infected) cells was determined at 48 h postinfection. RT units, reverse transcriptase units.

HIV-1 restriction in cat cells transduced with TRIM5 cDNA is suppressed by As₂O₃.

One possible explanation for our findings with the five cell lines is that As₂O₃ suppresses the restriction activity of certain TRIM5 orthologues and not others. To test this hypothesis, TRIM5 cDNAs from the two As₂O₃-responsive cell lines (TE671 and Vero) and the three As₂O₃-unresponsive cell lines (CV-1, FRhK4, and OMK) were transduced into CRFK, a cat fibroblast cell line without apparent retroviral restriction activity (6, 8). This was done by using murine stem cell virus-derived vectors as previously described (16). Each cDNA was expressed as a fusion to a C-terminal FLAG epitope tag. Despite the range in expression of the different orthologues (Fig. 3), all exhibited HIV-1 restriction activity that corresponded in relative magnitude to what is observed in the cell lines that served as the sources for the TRIM5 cDNAs (Fig. 4, left column).

FIG. 3.

Western blot showing synthesis of TRIM5 orthologues in cat fibroblasts. CRFK cells were transduced with retroviral vectors bearing TRIM5α cDNAs from the primate cell lines TE671, FRhK4, CV1, and Vero or TRIMCyp cDNA from OMK cells. Each cDNA encoded a C-terminal FLAG epitope tag. The vector used for transduction was pMIP, which is pMSCV-IRES-GFP (20) with a puromycin resistance cassette in place of the GFP coding sequence. Lysate from each cell population was resolved by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and immunoblotted with anti-FLAG M2 antibody (Sigma).

FIG. 4.

As₂O₃ counteracts the restriction of HIV-1 that is conferred by TRIM5 orthologues from various species. CRFK cells expressing TRIM5α from Homo sapiens (TE671), Macaca mulatta (FRhK4), Cercopithecus aethiops (Vero), or Cercopithecus tantalus (CV-1), or Aotus trivirgatus TRIMCyp (OMK), were infected with HIV-1_GFP in the absence or presence of 15 μM As₂O₃. The drug was present for the first 12 h of infection. The relative quantity of virus used to infect each sample is shown in RT units on the x axis. The percentage of GFP-positive (infected) cells was determined at 48 h postinfection.

The HIV-1 titer on cells transduced with empty vector was not increased by any concentration of As₂O₃ that was applied to target cells (Fig. 4 and data not shown). This is consistent with our other data showing a requirement for TRIM5-mediated restriction in order to see the effect of As₂O₃ (Fig. 2) and with the apparent lack of HIV-1 restriction activity in cat cells (6, 8). As₂O₃ treatment of cat cells transduced with any of the TRIM5 orthologues, however, resulted in a potent increase in the HIV-1 titer (Fig. 4). A 15 μM As₂O₃ concentration was found empirically to be the drug concentration with the largest effect on HIV-1 titers in TRIM5-transduced CRFK cells (data not shown). The magnitude of the increase in titer due to As₂O₃ correlated with the magnitude of HIV-1 restriction that a given cell line possessed. For example, the biggest effect of As₂O₃ was found with cells transduced with Macaca mulatta TRIM5α or Aotus trivirgatus TRIMCyp, the two cDNAs that conferred the greatest HIV-1 restriction activities on cat cells (Fig. 4). Of the TRIM5 cDNAs, human TRIM5α cDNA conferred the least HIV-1 restriction activity and, correspondingly, the smallest response to As₂O₃ (Fig. 4). Thus, HIV-1 restriction activity attributable to transduction of any of the TRIM5 cDNAs into cat cells was inhibited by As₂O₃.

As₂O₃ overcomes TRIM5α-mediated restriction of SIV_MAC239.

To determine if the findings with HIV-1 might be extended to other retroviruses, the effect of As₂O₃ on SIV_MAC239 infectivity was tested.Of the five primate cell lines examined here, only CV1 cells exhibit significant restriction activity against SIV_MAC239 (4-6, 10). The SIV_MAC239 titer in CV1 cells did not respond to As₂O₃ treatment (Fig. 5A), though, confirming our previous findings for this cell line with HIV-1 (Fig. 1). As₂O₃ had no effect on SIV_MAC239 titer in any of the other four cell lines (data not shown), as expected since these cells do not possess significant restriction activity against SIV_MAC239. The cat cell line transduced with the TRIM5α cDNA derived from CV1 cells exhibited a very potent restriction activity against SIV_MAC239 which was effectively suppressed by As₂O₃ (Fig. 5B). Cat cell lines transduced with the other TRIM5 orthologue cDNAs exhibited more modest SIV_MAC239 restriction activity and, correspondingly, smaller responses to As₂O₃ (data not shown). Similarly, As₂O₃ stimulated N-tropic MLV infectivity, but not B-MLV infectivity, in those TRIM5-transduced cat cell lines that possessed N-MLV-specific restriction activity (data not shown). Thus, the cell-specific suppression of TRIM5α restriction activity by As₂O₃ is not unique to HIV-1.

FIG. 5.

As₂O₃ does not increase the titer of SIV_MAC239 on CV1 cells but counteracts the SIV_MAC239 restriction that is conferred on cat cells by CV1 TRIM5α cDNA. (A) CV1 cells were infected with VSV-G-pseudotyped, GFP-encoding SIV_MAC239 (SIV_GFP) while being treated with the indicated amounts of As₂O₃. Infectivity was measured as described in the legend to Fig. 1. (B) Experiments were performed as described in the legend to Fig. 4, using SIV_GFP instead of HIV-1_GFP.

Conclusions.

Here we have directly shown that the stimulatory effect of As₂O₃ on HIV-1 titers involves suppression of TRIM5-mediated restriction activity. It has been suggested that the variable As₂O₃ responsiveness observed with different cell lines is a property of the particular TRIM5α orthologue expressed by those cells, but this conclusion was reached with single data points for two orthologues that differed in restriction activity by almost 2 orders of magnitude (8). The results presented here clearly demonstrate that all TRIM5 orthologues that exhibit retroviral restriction activity are likely to respond to As₂O₃ when they are expressed in the proper cell type. The particular property of a cell which determines whether TRIM5-mediated restriction will be As₂O₃ responsive remains to be determined. Given that human cells, monkey cells, and even cat cells can be As₂O₃ responsive, it seems unlikely that this is a species-specific property. Another possibility to consider is that some cells may be As₂O₃ unresponsive because they possess a retroviral restriction factor other than TRIM5 which is not sensitive to inhibition by As₂O₃. The molecular basis for the As₂O₃ effect on TRIM5-mediated restriction is unknown. As₂O₃ had no detectable effect on the half-life or subcellular localization of TRIM5 proteins in cat cells (data not shown). Ultimately, studies with As₂O₃ may aid attempts to elucidate the mechanism of TRIM5-mediated restriction.