FoxP3 Enhances HIV-1 Gene Expression by Modulating NFκB Occupancy at the Long Terminal Repeat in Human T Cells^*,S

Plasmids, Antibodies, and Cell Lines

FoxP3 was cloned into the HSPG retrovirus vector (47). LTR linker scanning mutants (48), LTR-luc, LTR deletion or point mutants, minimal promoter with LTR 3xκB-luc (41) were previously reported. Mis-sense mutations identified from IPEX patients (8) (I363V, F371C, A384T, and R397W) were introduced in the FoxP3 cDNA by site-directed mutagenesis. Deletion mutants in FoxP3 were generated by PCR amplification to specifically remove the N terminus (aa 1–189) as in FoxP3-ZFLZFKH, the N terminus, including the zinc finger (aa 1–241) as in FoxP3-LZFKH, or the N terminus, including all domains except Forkhead (aa 333–429) as in FoxP3-FKH. MHC-3XκB-luc and IκB-SR plasmids were provided by A. Baldwin (University of North Carolina-Chapel Hill). Jurkat cells were maintained in RPMI 1640 (Invitrogen), 293T cells were maintained in Dulbecco’s modified Eagle’s medium, and primary human T cells were maintained in Iscove’s minimal essential medium, supplemented with 10% fetal bovine serum (Sigma), 2 mM L-glutamine, 100 units/ml penicillin, and 100 mg/ml streptomycin. For T-cell activation, mouse anti-CD3 and -CD28 (BD Biosciences) and goat anti-mouse IgG (Caltag) antibodies were used. To determine T-cell purity, anti-human CD25-PE (Miltenyi) and CD4-FITC (BD Biosciences) were used. 7AAD (7-actinomycin D) (Molecular Probes) was used to determine cell viability.

Retrovirus Production and Transduction

293T cells were co-transfected with retroviral vector-, VSV-G-, and gag/pol-containing plasmids as previously described (47). For production of single cycle HIV-Luc reporter virus, 293T cells were transfected with VSV-G and HIV-Luc plasmids by Effectene (Qiagen). For retroviral transduction, Jurkat cells were spin-inoculated with HSPG control or FoxP3 retrovirus to >95% efficiency as measured by GFP expression (47).

Cell Culture and T-cell Purification and Transfection

CD4+ primary T cells were purified from Ficoll-separated PBMC (peripheral blood mononuclear cell) by MACS human CD4+ T-cell isolation kit (>95% purity, Miltenyi-Biotech, Auburn, CA). CD4+ T cells were activated by anti-CD3/CD28 monoclonal antibody cross-linked by plate-bound anti-mouse IgG as previously described (49). Briefly, purified cells are stained with anti-CD3 (0.5 μg/ml) and anti-CD28 (1 μg/ml) and incubated on plated-bound goat anti-mouse IgG. Purified primary CD4+ T cells were transfected by using the Amaxa nucleofector kit (Amaxa Biosystems, Gaithersburg, MD). Briefly, 5 × 10⁶ unstimulated CD4+ T cells were transfected with 5 μg of plasmid DNA (>40% efficiency by GFP expression), and cultured for 48 h before infection with the VSV-G pseudotyped HIV-luciferase reporter virus and activation by CD3 and CD28 cross-linked with plate-bound IgG. Cells were activated for 48 h, and luciferase was measure by luciferase assay. Alternatively, T cells were transduced with vector or FoxP3 retrovirus (>60% efficiency by GFP) and infected with VSV-G pseudotyped HIV-luciferase virus.

IL-2 Measurement by Real-time qPCR and Enzyme-linked Immunosorbent Assay

Amaxa-transfected primary CD4+ T cells were sorted for GFP expression and activated with anti-CD3/CD28 as above. IL-2 mRNA levels were determined following activation. 5 × 10³ cells were lysed in TRIzol reagent (Invitrogen) and column-purified (RNeasy, Qiagen). cDNA was generated using Cells-to-cDNA II kit (Ambion, Austin, TX), and quantitative PCR was performed on an ABI Prism 7000 (Applied Biosystems) using TaqMan IL-2 probe/primer set (Applied Biosystems). Each sample was normalized to 18 s. To measure IL-2 production in Jurkat T cells, transduced Jurkat cells were stimulated with PHA (phytohemagglutinin) and ionomycin (1 μM) for 12 h (41), and supernatants were collected. Secreted IL-2 was quantitated by the Human IL-2 enzyme-linked immunosorbent assay kit (BD Biosciences).

Transfection of Jurkat T Cells and Luciferase Assay

Jurkat cells were transfected using Geneporter transfection reagent (Genlantis, San Diego, CA), and transfection efficiency was monitored by GFP expression or co-transfection with pAX-βgal reporter plasmid. Cells were lysed, and luciferase expression was determined by Luciferase Assay System (Promega). Experiments were done in triplicates and repeated at least three times.

ChIP Assays

Jurkat T cells with a stably integrated HIV-1 LTR were transduced with control or FoxP3-expressing vectors. Sonicated chromatin from the Jurkat cells was immunoprecipitated with control IgG (IgG) or p65 (a kind gift from Al Baldwin) or anti-AcH3 by chromatin immunoprecipitation assay kit per the manufacturer’s instruction (Upstate, Millipore). Samples were purified and subject to real-time qPCR analysis. The relative p65 binding or acetylation of IL-2 or LTR promoter is expressed with ratios of ChIP-specific signal divided by signals from control 10% input chromatins. IL-2 promoter was analyzed with TaqMan reagents using the following primers: 5′-cac cta agt gtg tgg gct aat gta ac-3′and 5′-ctg atg act ctt tgg aat ttc ttt aaa cc-3′ and a FAM:TAM probe 5′-aga ggg att tca cct aca tcc att cag tca gtc-3′. This amplicon spans −226 to −133 in the human IL-2 promoter and spans the NFκB and AP-1 sites. HIV-1 LTR was analyzed with SYBR Green reagents (Abgene, UK) using the primers (5′-agc cct cag atc ctg cat ata agc a-3′ and 5′-gtt agc ca gaga gct ccc agg ctc a-3′) that yield an amplicon spanning −44 to −49.

EMSA

Standard electrophoretic mobility shift assays were performed (50). Briefly, probes containing the NFκB site (GGGATTCCCC) from the MHC gene or HIV-1 LTR were labeled with [α-³²P]dCTP using Klenow. 10 μg of nuclear extract was incubated with probe. For antibody supershift or blocking, 2 μg of α-p65, α-p50, or control IgG were added to nuclear extract and labeled probe.

Western Blot

1 × 10⁶ Jurkat T cells were lysed in 0.5% Nonidet P-40 lysis buffer and resolved by 10% SDS-PAGE. Approximately 25 μg of proteins was used, and FoxP3 was probed with an anti-FoxP3 antiserum (51), α-rabbit-IgG-horseradish peroxidase and visualized by ECL (Amersham Biosciences).

Statistical Analysis

For statistical analysis, a two-tailed Student t test is employed where p < 0.05 is considered significant.

FoxP3 Differentially Modulates IL-2 and HIV-1 Gene Expression by Inhibiting IL-2 but Enhancing LTR Promoter Activity

To investigate the mechanism of HIV-1 replication in FoxP3+ T cells, we expressed FoxP3 in the human CD4+ T leukemia (Jurkat) cells. When Jurkat cells were infected with VSV-G pseudotyped HIV-1 containing luciferase reporter gene, elevated levels of HIV-1 gene expression were detected in both unstimulated and stimulated FoxP3+ T cells (5- to 8-fold, Fig. 1A). Retrovirus-mediated expression of FoxP3 in Jurkat cells inhibited the expression of IL-2 upon activation (>6-fold, Fig. 1B). To determine if FoxP3 similarly regulates expression of IL-2 and HIV-1 in primary CD4+ T cells, we expressed FoxP3 in purified primary CD4+ T cells. HIV-1 gene expression was significantly enhanced in Th cells previously transduced with FoxP3-expressing retroviral vector (Fig. 1C) or by Amaxa nucleofection (data not shown). In contrast, ectopic expression of FoxP3 in CD4+ T cells efficiently suppressed expression from the endogenous IL-2 gene (Fig. 1D). Thus, FoxP3 inhibits IL-2 expression but enhances HIV-1 gene expression in human T cells.

To determine if FoxP3 regulates expression of IL-2 and HIV-1 at the promoter level, we transfected IL-2 or HIV-1 LTR promoter constructs driving the luciferase gene into Jurkat T cells. Similar to NL4-luciferase virus infection, FoxP3 also enhanced gene expression from HIV-LTR (Fig. 2A). FoxP3 efficiently suppressed IL-2 gene expression from both the endogenous IL-2 gene (Fig. 1, B and D) and from the transfected IL-2 promoter (Fig. 2B). The HIV-1 LTR co-transfection experiment was performed in the absence of Tat. Therefore, HIV-1 Tat is not required for the FoxP3 activity.

Forkhead Domain and N Terminus of FoxP3 Are Critical for FoxP3-mediated Enhancement of HIV-1 LTR

The Forkhead domain and distinct residues in FoxP3 are critical for FoxP3 function as indicated by mutations in the scurfy mouse and in human IPEX patients (8). To define the FoxP3 domains that contribute to this LTR-enhancing activity, we generated FoxP3 deletion mutants and transduced Jurkat T cells. Loss of the N terminus, including the proline-rich region, inhibited FoxP3 enhancement (Fig. 3A), whereas expression of the N terminus alone was not able to enhance LTR activity (data not shown). We also tested the known mis-sense mutations derived from IPEX patients by single site-directed mutagenesis introduction into full-length FoxP3. Interestingly, all four such mis-sense mutations tested also inactivated the FoxP3 activity in enhancing HIV-1 LTR (Fig. 3B). Similar levels of the mutant FoxP3 proteins were expressed (supplemental Fig. S1). Therefore, the FoxP3 activity in enhancing HIV-1 LTR depends on both the N-terminal domain and critical residues in the forkhead domain, which are also critical for the function of the FoxP3 protein in humans.

Two NFκB Sites in the LTR Enhancer Core Are Both Required for FoxP3 Enhancement

We have mapped the cis-acting elements in HIV-1 LTR that respond to FoxP3 by employing the 26 linker scanning mutants (LS1–LS26) of HIV-1 LTR (48). The mutations in LS20 (−111 to −94) and LS21 (−93 to −76), which inactivate the two NFκB binding sites in the LTR enhancer core, significantly reduced response to FoxP3 (Fig. 4A). Deletion of the NFκB/NFAT enhancer core completely abolished LTR response to FoxP3 (Fig. 4B). The two NFκB sites also overlap with the sites for NFAT binding (41). Point mutations in either NFκB site (LTR mκB1 and mκB2) significantly reduced FoxP3-mediated enhancement, whereas mutations in both sites (mnκB) led to complete loss of enhancement by FoxP3 (Fig. 4B). Therefore the two NFκB sites in the LTR enhancer core are both required for FoxP3 enhancement.

FoxP3 Enhances HIV-1 LTR Activity in NFκB Sequence- and Cell Type-specific Fashions

To further elucidate the role of NFκB, we blocked NFκB activation using the IκB super-repressor in T cells. Consistent with results that the NFκB sites in LTR were critical for FoxP3-mediated activation, FoxP3 was unable to enhance HIV-1 LTR activity when NFκB function was inhibited (Fig. 5A). We also determined if FoxP3 enhancement was specific to the LTR-derived NFκB sequences. A luciferase reporter gene driven by a minimal promoter with three NFκB sites derived from HIV-1 LTR (LTR-3κB-luc) or from the MHC gene (MHC-3κB-luc) was studied in Jurkat cells. Under the same experimental conditions, FoxP3 enhanced LTR-3κB-luc expression but not the MHC-3κB-luc expression in T cells (Fig. 5B). Interestingly, both LTR-3κB-luc and MHC-3kB-luc, as well as the LTR-luc were inhibited by FoxP3 in 293T cells (Fig. 5, C and D), corroborating results previously described for FoxP3 in 293T cells (39, 40). Thus, FoxP3 modulates HIV-1 LTR activity via its specific NFκB sites in T cells.

FoxP3 Differentially Alters the Acetylation Level of the IL-2 Promoter and HIV LTR

We tested if FoxP3 affects the chromatin structure at the IL-2 promoter and the HIV-1 LTR by ChIP assays with anti-acetylated histone H3-K9 (AcH3) antibodies. Relative levels of AcH3 were determined by qPCR of the NFκB and NFAT/AP-1 region (−226 to −133) in the human IL-2 promoter after ChIP (Fig. 6A). FoxP3 reduced the AcH3 level or the “openness” of the IL-2 promoter in Jurkat cells either unstimulated or after stimulation. When the stably integrated HIV-1 LTR in the same Jurkat A82 (52) cell was analyzed, we discovered that FoxP3 either had no effect (unstimulated) or increased (stimulated) AcH3 acetylation level at the NFκB/Sp1 enhancer core sequences of the LTR (1.8-fold, p < 0.05, Fig. 6B). Thus, FoxP3 reduced the histone acetylation level of the IL-2 promoter but had no effect or increased that of the HIV-1 LTR in T cells.

FoxP3 Does Not Alter NFκB Nuclear Translocation but Modulates NFκB Occupancy at the HIV-1 LTR Promoter

Given that FoxP3 enhancement is both NFκB sequence- and cell type-specific, it is unlikely that a simple increase in the nuclear levels of NFκB is involved. Indeed, FoxP3 did not enhance NFκB nuclear localization in Jurkat cells by Western blot analysis (Fig. 7A). Using a consensus NFκB probe, similar levels of p65/p50 and p50/p50 NFκB DNA binding activity were detected in FoxP3+ and vector control cells (Fig. 7B). We next wanted to determine if FoxP3 was able to modulate NFκB occupancy at the HIV-1 LTR promoter in vivo by ChIP using specific anti-p65 antibodies. In a Jurkat T cell line stably integrated with the HIV-1 LTR (53), FoxP3 enhanced the association of p65 at the LTR promoter (~2-fold, Fig. 7C) but not at the IL-2 promoter (Fig. 7D). Therefore, although NFκB nuclear translocation was unchanged, FoxP3 enhanced p65 occupancy specifically at the HIV-1 LTR in T cells.