CXCR4 Is Down-Regulated in Cells Infected with the CD4-Independent X4 Human Immunodeficiency Virus Type 1 Isolate m7NDK

Abstract

Macrophages and T cells infected in vitro with CD4-dependent human immunodeficiency virus type 1 (HIV-1) isolates have reduced levels of CD4 protein, a phenomenon involved in retroviral interference. We have previously characterized the first CD4-independent HIV-1 X4 isolate m7NDK, which directly interacts with CXCR4 through its mutated gp120. We thus investigate CXCR4 expression in cells infected with this m7NDK CXCR4-dependent HIV-1 mutant. We present evidence of the down-regulation of CXCR4 membrane expression in CD4-positive or -negative cells chronically infected with the HIV-1 m7NDK, a phenomenon which is not observed in the CD4-dependent HIV-1 NDK parental strain. This down-regulation of CXCR4 was demonstrated by fluorescence-activated cell sorter analysis and was confirmed by the absence of CXCR4 functionality in m7NDK-infected cells, independently of the presence of CD4 protein. Furthermore, a drastic reduction of the intracellular level of CXCR4 protein was also observed. Reduced levels of CXCR4 mRNA transcripts were found in m7NDK-infected HeLa and CEM cells, reduced levels that could not be attributed to a reduced stability of CXCR4 mRNA. Down-regulation of CXCR4 on m7NDK-infected cells may thus be explained by transcriptional regulation.

Expression of CD4 (20, 22) and the chemokine receptors CCR5 and CXCR4 at the target cell surface is essential for human immunodeficiency virus (HIV) entry (2, 24). HIV type 1 (HIV-1) cell entry is mediated by a first interaction between envelope (Env) glycoprotein gp120 and CD4, which induces a conformational change in gp120, exposing the coreceptor binding site or creating the conformational coreceptor binding site, leading to membrane fusion (6, 21, 23, 32).

Macrophages and T cells infected with HIV in vitro have reduced surface CD4 expression (8, 13, 16). The reduction of CD4 surface expression is due to the combined action of three viral proteins: Env, Vpu, and Nef. The HIV envelope protein precursor gp160 forms a complex with CD4 in the endoplasmic reticulum (ER) of infected cells (7, 18, 36), and Vpu triggers the degradation of ER-retained CD4 molecules (37, 38). The auxiliary Nef protein triggers the accelerated internalization of CD4 molecules that have already reached the cell surface (1, 30, 33).

We have previously reported the characterization of the first HIV-1 strain that no longer requires the presence of CD4 to enter its target cells (10). This CD4-independent isolate was derived spontaneously from the X4 HIV-1 isolate NDK after a long-term culture (average of 200 days) in the CD4⁺ T-cell line CEM and has been named m7NDK. This new tropism has been shown to correlate with seven specific amino acid changes in critical regions of gp120, C2, V3, and C3. We have postulated that this mutant envelope subunit has either a predisposed conformation or a greater binding affinity for CXCR4 and overcomes the need for CD4-induced conformational modifications (10).

Our interest focused on CXCR4 receptor expression, in cells infected with the CD4-independent CXCR4-dependent m7NDK HIV-1. Down-regulation of CXCR4 has been described in CD4⁺ T cells following infection with the human herpesvirus 6 (HHV-6) and HHV-7 (34, 39); however, it is worth noting that these viruses do not use CXCR4 as a receptor (40). It has been established that CXCR4 is down-regulated by a few HIV-2 isolates which use CXCR4 as their primary receptor (12), although down-regulation of the coreceptor CXCR4 by X4 CD4-dependent HIV-1 viruses has never been characterized. A variant of HIV-1/IIIB termed HIV-1/IIIBx has been characterized (17) that is both replication competent and fusogenic for a CD4-negative subclone of SupT1 cell line. However, it failed to down-regulate CXCR4 in chronically infected cells (17).

Recently, several studies have shown that regulation of CXCR4 mRNA expression depends on cell activation and oxidative stress, as well as cell type (5, 26, 31). Furthermore, signaling and internalization of CXCR4 protein can be regulated by receptor phosphorylation-dependent and -independent mechanisms (15), and alternative trafficking of CXCR4 can be induced by several chemical or pathogenic agents (3, 35). Nevertheless, regulation of CXCR4 membrane expression has not yet been described after infection with a CXCR4-dependent virus.

Here we present evidence of CXCR4 down-regulation in the CD4⁺ T-cell line CEM, as well as in the nonlymphocytic CD4⁻ HeLa cell line, infected with the m7NDK mutant. We demonstrate the absence of CXCR4 surface expression and functionality in these cells. Analysis of CXCR4 mRNA transcripts revealed a decreased CXCR4 mRNA steady-state level which was not observed in the CD4-dependent parental strain or uninfected cells. However, these results did not correlate with a reduction of CXCR4 mRNA transcript stability. In addition, we demonstrate that it is an active phenomenon since we observe CXCR4 down-regulation upon acute infection of CEM cells.

Taking all of these findings together, our results suggest that down-regulation of CXCR4 upon m7NDK infection might be explained by transcriptional regulations and may provide a mean for the m7NDK isolate to monitor viral interference.

MATERIALS AND METHODS

Cell lines and viruses.

The CD4-positive human lymphoid cell line CEM was a gift from J. L. Virelizier (Institut Pasteur, Paris, France) and was grown in RPMI 1640 (Life Technologies) medium supplemented with 5% fetal calf serum, antibiotics, and glutamine (Life Technologies). The NDK isolate was a gift from F. Barré-Sinousi (Institut Pasteur) (11) and was propagated in CEM cells. The previously described NDK mutant m7NDK was obtained after a long-term culture in CEM cells (10).

HeLa, P42 cells (HeLa CD4⁺ long terminal repeat [LTR]-lacZ) and Z24 (HeLa LTR-lacZ) have been previously described (9) and were kindly provided by M. Alizon, ICGM, Paris, France. Adherent cell lines were grown in Dulbecco modified Eagle medium supplemented with 5% fetal calf serum, antibiotics, and glutamine (Life Technologies).

Three clones (14, 48, and 108) of HeLa cells chronically infected with m7NDK isolate were kept for further analysis, after endpoint dilution cloning of chronically infected HeLa cells.

Infection of CEM cells.

Virus were added to CEM cells at 16 ng per 10⁶ cells and incubated at 37°C for 4 h in a minimum volume. After this period of time, supplemented RPMI medium was added to attain a final concentration of 10⁶ cells/ml. Infection was then followed by fluorescence-activated cell sorter (FACS) analysis, as well as by cell fusion assays.

Flow cytometry analysis.

Aliquots of 10⁶ cells were subjected to direct or indirect label staining to analyze the surface and/or intracellular expression of antigens. Nonadherent cells were washed with ice-cold Cell Wash (Becton Dickinson) and stained with the primary antibodies for 1 h at 4°C. Adherent cells were first harvested using 1× phosphate-buffered saline (PBS; Life Technologies)–citrate (0.01 M) and then treated as the nonadherent cell lines. They were washed with Cell Wash to remove unbound antibody and stained with the secondary antibody for 1 h at 4°C. The cells were then washed and resuspended in Cell Wash containing 1% formaldehyde (Merck), kept at 4°C, and analyzed. The chemokine receptor CXCR4 was detected by indirect staining with anti-CXCR4 MAb12G5, MAb171, MAb172, or MAb173 (R & D Systems) monoclonal antibodies (MAbs), followed by treatment with phycoerythrin (PE)-conjugated rabbit anti-mouse immunoglobulin G (IgG) (Dako) or by direct staining with PE-conjugated anti-CXCR4 MAb173 or with fluorescein isothiocyanate (FITC)-conjugated anti-CXCR4 MAb12G5.

For the intracellular detection of CXCR4, cells were first saturated with unconjugated anti-CXCR4 (MAb173), followed by staining with the secondary antibody FITC-conjugated rabbit anti-mouse IgG (Amersham). After staining, cells were fixed with 3% formaldehyde, washed with Cell Wash (Pharmingen) and quenched to saturate free radicals with glycine (20 mM). Cells were permeabilized in Cell Wash containing 0.05% saponin and then stained with PE-conjugated anti-CXCR4 MAb173. Nonspecific fluorescence was determined by using irrelevant isotype-matched MAbs (Dako).

Transferrin receptor was detected using a MAb R-Trf (Roche), CD4 was detected using MAb MT310 (Dako) or MAb OKT4, and the HIV-1 Env protein was detected using a seropositive serum.

A FACSCalibur (Becton Dickinson Immunocytometry Systems, San Jose, Calif.) or a FACScan (Epics Elite; Coulter, Miami, Fla.) was used for cytometry analysis. The excitement radius was 488 nm, and the emission radius band-pass was at 575 nm for 10,000 cell events.

Immunofluorescence microscopy.

Adherent cells were plated in Lab-Tek chamber slides (Polylabo). Characterization of intracellular chemokine receptor expression was achieved by fixation of cells in 3% formaldehyde (15 min) at room temperature (RT), quenching in 0.1 M glycine-PBS, and saturation with PBS containing 0.2% bovine serum albumin and 0.05% saponin (Sigma). In the same buffer, an overnight incubation at 4°C was performed with MAb against CXCR4 (MAb173). The cells were washed and subsequently incubated with anti-mouse cyanin 3 (Cy3)-conjugated secondary antibody (Caltag) in the same buffer for 1 h at RT, followed by extensive washing prior to mounting. The staining of cell surface proteins was as described above except for the use of saponin, which was excluded from the buffer. Cells were double stained for CXCR4 and Env protein, with MAb173 and a seropositive serum, followed by treatment with anti-mouse antibody–Cy3 and anti-human antibody–FITC.

Omission of the primary antibody and substitution with an isotype-matched MAb served as a control. After mounting, the cells were observed with a laser confocal microscope (MRC 1000; Bio-Rad, Hercules, Calif.).

Cell fusion assays.

Cell fusion assays were performed between adherent or nonadherent cells chronically infected by HIV-1, as previously described (9, 10). Fusion efficiency was analyzed 11 h later. Measurements of β-galactosidase enzyme activities were done as previously described by staining in situ with 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal; Life Technologies) as a substrate or by using a quantitative assay using chlorophenol red–β-d-galactopyranoside (CPRG; Roche) as a substrate (10).

Measurement of calcium mobilization.

Changes in the cytosolic free-calcium concentration were measured in cells loaded with 1 μM Fura 2-acetoxymethylester (Fura 2-AM) (Sigma) at 37°C for 1 h (14, 27, 29). Cells were washed and resuspended in 20 mM HEPES-Hanks balanced salt solution (HBSS; Life Technologies). Fura-2 fluorescence assays were performed with aliquots of 4 × 10⁷ cells in 2 ml of HBSS, using a fluorimeter (Jobin Yvon 3D; Jobin Yvon, Lonjumeau, France) equipped with a thermally controlled cuve holder and a magnetic stirrer. After we recorded the baseline [Ca²⁺]_i levels, stromal-cell-derived factor 1α (SDF-1α) at 10 nM (R & D Systems) was added. The excitation and emission wavelengths for Fura-2 fluorescence assays were 340 and 510 nm, respectively. Cytosolic calcium concentrations were calculated as described previously (14). Tracings were reproduced and scanned using an Agfa Snap CAM, with version F-3.0 Color It software (Apple).

Chemotaxis assays.

Cell migration in response to SDF-1α (R & D Systems) was measured in 3.0-μm-pore-size Transwell cell culture chambers (Costar). In the upper chamber, 10⁶ cells were suspended in 100 μl of complete RPMI 1640 and placed on top of the lower chamber containing 500 μl of complete RPMI 1640 with different concentrations of SDF-1α. Plates were incubated at 37°C in CO₂ for 5 h. The upper chamber was then carefully removed, and the numbers of viable cells present on the lower chamber were counted using trypan blue exclusion. The percentages of the transmigrations were determined for each concentration of SDF-1α.

Northern blot analysis.

Total cellular RNA extraction and purification was performed using an RNA B isolation system (Bioprobe) according to the manufacturer's protocol. RNA was extracted from uninfected or chronically infected cells, and 10 μg of each preparation was denatured with formaldehyde and size fractionated by electrophoresis on a 1% agarose gel. The RNAs were then transferred to a hybridization transfer membrane and hybridized with a ³²P-labeled CXCR4 cDNA or glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cDNA probes. To determine the posttranscriptional stability of CXCR4 mRNA, actinomycin D (Sigma) was added at 5 μg/ml to uninfected or infected cells to block transcription. Cells were collected after various incubation periods and were used for RNA extraction. The half-life of CXCR4 mRNA was estimated by plotting the densitometric ratios of CXCR4 mRNA versus GAPDH mRNA, determined using Image Quant Tools, version 1.0, for Power Macintosh.

RESULTS

Surface CXCR4 protein expression.

The level of cell surface expression of CXCR4 in uninfected, Wild-type NDK (wtNDK)-infected, or m7NDK chronically infected CEM cells was analyzed by flow cytometry (Fig. 1A). Two different MAbs were used, directed against different CXCR4 epitopes, MAb12G5 and MAb173. Binding of MAb12G5 and MAb173 to m7NDK-infected cells was negative compared to that for uninfected or wtNDK-infected cells. The same result was obtained with two other antibodies directed to different CXCR4 epitopes (data not shown). The expression of CD4 by flow cytometry was also monitored to insure that down-modulation of CD4 was observed on wtNDK-and m7NDK-infected CEM cells (Fig. 1B). Expression levels of the transferrin receptor and the HIV envelope protein were also measured by flow cytometry in order to verify the integrity and the infected state of these cells, respectively (Fig. 1B).

FIG. 1.

Analysis of cell surface expression of CXCR4. (A) Flow cytometric analysis of CXCR4 at the surface of uninfected, wtNDK-infected, or m7NDK chronically infected CEM cells. Cells were incubated with two different MAbs directed against CXCR4, MAb12G5 and MAb173, or with an isotype-matched control (CTRL; shaded area). (B) Uninfected, wtNDK-infected, and m7NDK-infected CEM cells were stained for flow cytometric analysis of CD4 (MAb MT310), transferrin receptor (MAb R-Trf), or the HIV-1 Env protein (serum from seropositive patient). (C) Flow cytometric analysis of surface CXCR4 of uninfected or three different HIV-1 m7NDK-infected HeLa cell clones using MAb12G5 and MAb173. The control represents each one of the cell lines stained with a respective isotype-matched control.

Surface expression of CXCR4 was also determined on a CD4-negative cell line to determine whether down-regulation of CXCR4 upon m7NDK infection was a CD4-dependent phenomenon. CXCR4 expression was evaluated on uninfected HeLa cells and three different clones of m7NDK chronically infected HeLa cells, by flow cytometry using MAb12G5 and MAb173 (Fig. 1C). Surface expression of CXCR4 was detected on uninfected HeLa cells but was not detected on any of the three different HeLa m7NDK clones, thus indicating that the CXCR4 modulation is independent of CD4 cellular expression and is not specific to the CEM CD4-positive lymphocytic cell line.

A coculture test using LTR-lacZ indicator cells either positive or negative for CD4 expression (9, 10) was also performed in parallel to the FACS assays to confirm Env protein expression and fusion ability (data not shown).

Functionality of CXCR4.

The natural ligand for CXCR4 is SDF-1α (4, 28). To test CXCR4 functionality at the cell surface, we analyzed both the intracellular Ca²⁺ flux and chemotactic response following SDF-1α stimulation on m7NDK-infected CEM cells compared to uninfected or wtNDK-infected CEM cells.

As shown in Fig. 2A, m7NDK-infected CEM cells presented an elevation of [Ca²⁺]_i in response to SDF-1α that was >60% reduced in comparison to uninfected or wtNDK-infected CEM cells. These last two cell lines presented a similar response to SDF-1α, excluding the possibility that the chronically infected condition might alter cell responsiveness to SDF-1α. Each of these three cell lines responded strongly and in a similar fashion to the nonspecific Ca²⁺ ionophore, ionomycin (14; data not shown).

FIG. 2.

Analysis of CXCR4 functionality in uninfected, wtNDK-infected, and m7NDK-infected CEM cells. (A) Measure of intracellular calcium concentration in response to 10 nM SDF-1α. Cells were loaded with Fura 2-AM and incubated in HEPES-buffered Na⁺ solution, and the [Ca²⁺]_i was determined fluorimetrically. After the baseline [Ca²⁺]_i levels were recorded, SDF-1α (10 nM) was added as indicated. The cytosolic calcium concentrations were calculated as described previously (14). Tracings are shown for one population of cells and are representative of at least three independent experiments. (B) Chemotaxis response to SDF-1α. Chemotaxis assays were performed in transwell chemotaxis chambers in the absence or presence of SDF-1α (10 and 100 nM). The results are shown as the number of migrated cells related to the SDF-1α concentration. These results are representative of at least three independent experiments.

We next verified the capacity of surface CXCR4 to initiate cell migration in response to an SDF-1α gradient (Fig. 2B) (19). Uninfected, wtNDK-infected, or m7NDK chronically infected CEM cells were layered on a transwell upper chamber, and their migration, induced by two different doses of SDF-1α (10 and 100 nM), was evaluated 5 h later. No migration was observed with m7NDK CEM cells even at the highest SDF-1α concentration (100 nM). In contrast, wtNDK-infected or uninfected CEM cells presented a similar pattern of migration whatever SDF-1α concentration was used. This finding indicates that the chronic status of infection is not responsible for m7NDK-infected cell unresponsiveness.

In summary, CXCR4 expressed at the cell surface of uninfected or wtNDK-infected CEM cells displayed equally the characteristics of functional receptors. In contrast, on m7NDK-infected cells, CXCR4 functionality for chemotaxis, as well as for intracellular calcium flux responses, was severely reduced. This clearly correlates with previous FACS analyses and is indicative of the absence of CXCR4 membrane expression.

Intracellular CXCR4 protein expression.

The intracellular CXCR4 level was determined by flow cytometry in m7NDK-infected cells, as well as in uninfected or wtNDK-infected cells. Figure 3 shows the intracellular versus surface presence of CXCR4 detected by MAb12G5. This experiment allowed the detection of intracellular CXCR4 in m7NDK-infected CEM cells, although its expression was drastically reduced compared to those in uninfected or wtNDK-infected cells (Fig. 3A). Similar results were observed for CD4-negative HeLa cells, since the cellular clones of m7NDK chronically infected HeLa cells showed a drastic intracellular CXCR4 reduction compared to uninfected HeLa cells (Fig. 3B).

FIG. 3.

Analysis of intracellular versus surface expression of CXCR4 protein. Flow cytometric analysis of surface expression of CXCR4 and intracellular expression of CXCR4 in uninfected, wtNDK-infected, and m7NDK chronically infected CEM cells (A) and in uninfected and m7NDK-infected HeLa cells (B). Surface CXCR4 expression was determined by measuring the FITC fluorescence intensity, and intracellular CXCR4 expression was determined by measuring the PE fluorescence intensity. In both cases, anti-CXCR4 MAb12G5 (black tracings) and an isotype-matched control (shaded area) were used.

Immunofluorescent microscopy experiments were carried out in both uninfected (Fig. 4A) and m7NDK-infected (Fig. 4B) HeLa cells to analyze intracellular CXCR4 protein. Cells were permeabilized and stained with anti-CXCR4 MAb173, followed by treatment with anti-mouse antibody–Cy3 (red). In infected cells the intracellular expression of CXCR4 presented a pattern of distribution similar to that for the uninfected HeLa cells, although only scarce amounts could be detected.

FIG. 4.

Detection of intracellular CXCR4 by immunofluorescence microscopy. Uninfected (A) and m7NDK chronically infected (B) HeLa cells were grown on LabTek coverslips, fixed, permeabilized, and stained for CXCR4. Staining for anti-CXCR4 antibody (MAb173), followed by staining for Cy3-conjugated secondary antibody (red), was photographed at ×80 magnification. Double staining was performed as a control in uninfected (C) and m7NDK-infected (D) HeLa cells. Cells were left intact and double stained for CXCR4 and Env protein using MAb173 and a seropositive serum, respectively. Staining with anti-CXCR4 antibody was followed by staining with Cy3-conjugated secondary antibody (red) and Env, followed by staining with FITC-conjugated secondary antibody (green). Photographs were then obtained at ×32 magnification. Cells were photographed after analysis using a confocal microscope (MRC 1240; Bio-Rad).

As a control to verify surface CXCR4 expression and the infected condition of these cells, cells were left intact and double stained for CXCR4 and Env, using MAb173 followed by treatment with anti-mouse antibody–Cy3 (red) and a seropositive serum followed by treatment with anti-human antibody–fluorescein isothiocyanate (FITC) (green), respectively. In uninfected HeLa cells (Fig. 4C), CXCR4 is clearly observed and no viral proteins were detected; as expected, CXCR4 was not detected on the membrane of m7NDK-infected HeLa cells (Fig. 4D), and thus only the staining for viral Env proteins was determined.

These results, along with the intracellular FACS analysis results, confirm the intracellular down-modulation of the CXCR4 protein in m7NDK-infected HeLa and CEM cells.

CXCR4 mRNA analysis.

Since a clear reduction of intracellular CXCR4 protein was observed in m7NDK-infected CEM and HeLa cells, we performed CXCR4 mRNA transcript analysis (Fig. 5). Northern blot experiments, using a cDNA CXCR4-specific 1-kb probe, revealed an almost undetectable level of CXCR4 mRNA transcripts on both CEM m7NDK-infected (Fig. 5) and HeLa m7NDK-infected (data not shown) cells, while CXCR4 transcripts were present in similar amounts either in uninfected HeLa cells or in uninfected and wtNDK-infected CEM cells. Nevertheless, a more sensitive assay, reverse transcription-PCR, confirmed that a specific CXCR4 cDNA could be identified (data not shown).

FIG. 5.

Northern blot analysis of CXCR4 mRNA in uninfected, wtNDK-infected, and m7NDK-infected CEM cells. Samples of total cellular RNA were hybridized with a ³²P-labeled CXCR4 cDNA probe or a GAPDH cDNA probe.

We next assessed the posttranscriptional stability of CXCR4 mRNA in m7NDK-infected CEM cells (Fig. 6). Actinomycin D was added to uninfected and wtNDK-or m7NDK-infected CEM cells to block transcription, and mRNAs were extracted from each sample at different times (Fig. 6A). Northern blotting experiments were performed using a CXCR4-specific cDNA probe. The half-life of CXCR4 mRNA was estimated by plotting the densitometric ratio of the CXCR4 mRNA levels to the GAPDH mRNA levels. The estimated mean half-lives of the CXCR4 mRNAs of three independent experiments were 3.7, 3.8, and 3.9 h in uninfected, wtNDK-infected, and m7NDK-infected CEM cells, respectively (Fig. 6). Similarly, a half-life of 3 h was found for both uninfected and m7NDK-infected HeLa cells (data not shown). These data demonstrate that infection by m7NDK virus in either CEM or HeLa cells does not reduce the posttranscriptional stability of CXCR4 mRNA.

FIG. 6.

Stability of CXCR4 mRNA transcripts in uninfected, wtNDK-infected, and m7NDK-infected CEM cells. (A) Actinomycin D was added at a concentration of 5 μg/ml for the indicated times, and CXCR4 mRNA expression was determined by Northern blot analysis. Since the CXCR4 mRNA transcriptional level in m7NDK-infected CEM cells was low, longer exposure times were used to estimate the half-life (data not shown). (B) The half-life of CXCR4 mRNA was determined by plotting the densitometric CXCR4 mRNA/GAPDH mRNA ratio versus time (in hours). These results are representative of three independent experiments.

Kinetics of CXCR4 down-regulation.

To preclude the possibility of positive selection of CXCR4 low-expressing cells and in order to verify that CXCR4 down-regulation is an active process induced by the m7NDK isolate, CEM cells were infected either with m7NDK or its wild-type counterpart isolate. The expression kinetics of CD4, CXCR4, and HIV Env protein were monitored by FACS analysis for 29 days postinfection (Fig. 7). Cells were infected with 16 ng/10⁶ cells of m7NDK (CEM+m7NDK) and wtNDK (CEM+wtNDK) isolates. As a control, uninfected cells (CEM) or m7NDK chronically infected cells (CEM+m7NDK) were subjected to the same FACS analysis. Fusion test assays were also performed to ensure HIV Env protein expression and fusion ability (data not shown).

FIG. 7.

Kinetics of CXCR4 down-modulation upon m7NDK infection. Cells were infected with 16 ng of each of the indicated viral supernatants per 10⁶ cells. The expression levels of CD4, CXCR4, and HIV-1 Env protein were measured postinfection by FACS analysis as indicated. CD4 was detected using MAb OKT4, CXCR4 detection was based on MAb12G5, and HIV-1 Env protein was detected using a seropositive serum. The results are presented as a percentage of the positive cells. The results of one of two representative experiments are shown.

The antibodies used for this analysis were MAb OKT4 for CD4, which does not compete with gp120 for CD4 binding, and MAb12G5 for CXCR4 detection. HIV Env protein expression was determined using serum from a seropositive patient. Upon infection with the m7NDK isolate, Env protein expression could be detected on the surface of infected cells at day 1 postinfection (Fig. 7A). This drastic increase was accompanied by a marked decrease in CD4 expression (Fig. 7B). As expected, CXCR4 expression presented a drastic reduction, reaching an undetectable level by approximately 7 days postinfection (Fig. 7C), which was maintained until chronic infection was established.

Cells infected with wtNDK isolate did not present the same type of infection pattern, since the envelope expression increased slowly upon infection to reach a maximum at approximately 5 days postinfection. This progressive increase in the envelope expression was accompanied by a progressive decrease in CD4 expression. Furthermore, CXCR4 expression was not affected by infection with this isolate, as was expected. Fusion tests were performed along throughout the kinetic analysis to ensure correct HIV Env protein expression and fusion ability (data not shown).

During this kinetics analysis and independent of the viral isolate, cells behaved as a whole, and no subpopulations were observed by FACS analysis for CXCR4 expression (data not shown). This result clearly precludes the coexistence of heterogeneous high- and low-CXCR4-expressing cells in the CEM population.

DISCUSSION

We present here evidence of CXCR4 down-regulation in m7NDK-infected cells, a phenomenon that is independent of cellular CD4 expression status. Surface expression of CXCR4 was not detected on these cells using two antibodies directed against different epitopes of CXCR4, MAb12G5 and MAb173 (Fig. 1). This was confirmed using two other antibodies directed to two other different CXCR4 epitopes (data not shown).

The functionality of CXCR4 was analyzed by its ability to respond to SDF-1α-induced signalization. Chemotactic assays demonstrated a complete insensitivity of m7NDK CEM cells to SDF-1α-induced migration (Fig. 2B), and the intracellular calcium elevation was 60% reduced compared to that for uninfected or wtNDK-infected CEM cells (Fig. 2A). This reduction, given the cascade nature of this type of signalization, might be correlated with a loss of more than 90% of the cell surface receptor level. These results, together with the FACS analysis of surface CXCR4 expression, strongly indicate the absence of CXCR4 at the surface of m7NDK-infected cells.

Surface expression of CXCR4 was also determined on a CD4-negative cell line to verify if down-regulation of CXCR4 was a phenomenon dependent on surface CD4 expression. CXCR4 expression was evaluated by flow cytometry on uninfected HeLa cells and on three HeLa cell clones chronically infected with m7NDK (Fig. 1B). Surface expression of CXCR4 was not detected in any of these three different HeLa m7NDK clones (Fig. 2B), thus indicating that the phenomenon is independent of CD4 cellular expression and is not a particularity of a CD4-positive lymphocyte cell line.

The intracellular content of CXCR4 in m7NDK-infected HeLa and CEM cells was determined both by intracellular FACS analyses and immunofluorescent microscopy (Fig. 3 and 4). In both cases, a very reduced level of intracellular CXCR4 was found in m7NDK-infected HeLa and CEM cells compared to that in uninfected or wild-type-infected cells.

Analyses of CXCR4 mRNA transcripts in m7NDK-infected HeLa and CEM cells were performed. A Northern blot assay revealed a drastic reduction of the steady-state level of CXCR4 mRNA compared with that of uninfected or wtNDK-infected cells (Fig. 5). This drastic reduction was clearly not a consequence of decreased CXCR4 mRNA transcript stability (Fig. 6), since their half-life was not altered in m7NDK-infected cells. The transcriptional activity of the CXCR4 gene may probably be affected in these cells, even though no regulatory sequences in the CXCR4 gene promoter has yet been described as a target site for viral protein inhibition of transcription initiation (5, 25).

The hypothesis that, during acute infection, cells with abnormally low CXCR4 gene expression are positively selected for m7NDK isolate infection could be raised. In order to verify that the phenomenon of CXCR4 down-regulation is actually an active process and not a selection of CXCR4 low-expressing cells, we performed infections of CEM cells with either m7NDK or wtNDK viral isolates (Fig. 7). We observed a rapid appearance of HIV envelope protein expression on the surface of infected cells (Fig. 7A). This increase was followed by a reduction of CD4 expression also from the first day postinfection (Fig. 7B). A total loss of CXCR4 surface expression was observed approximately 7 days after infection, and this loss was conserved thereafter (Fig. 7C). The parental isolate wtNDK presented a slower process of expression; the envelope protein expression appeared gradually on infected cells, and the loss of CD4 was also gradual, with a total loss by 4 days postinfection. The expression of CXCR4 was unaltered in wtNDK-infected cells compared to uninfected cells. The loss of CXCR4 in m7NDK-infected cells occurred rapidly, and the whole population behaved similarly, which means that no subpopulations appeared with lower amounts of CXCR4 surface expression. This clearly precludes the hypothesis of positive selection of a low-CXCR4-expressing CEM cell clone after m7NDK infection. Furthermore, lower-CXCR4-expressing cells, in order to be positively selected for m7NDK virus infection, should present growth advantages to overcome high-CXCR4-expressing cell growth.

We performed cocultures between wtNDK- or m7NDK-infected CEM cells and uninfected HeLa CD4⁺ CXCR4⁺ cells. We then measured specific cell fusion inhibition in the presence of three different MAbs to CXCR4. Fusions of CEM m7NDK-infected cells were less inhibited by the three antibodies than was wtNDK-induced fusion (data not shown). The inhibitory effect was dose dependent, which is consistent with a reversible, competitive inhibition. This suggests a greater affinity of the m7NDK isolate Env protein for CXCR4, allowing it to strongly compete with the antibodies for CXCR4 binding, and correlates with kinetic infection data in which m7NDK Env expression is detectable as early as 1 day postinfection. Moreover, we can then suppose that its affinity for CXCR4 or its favorable conformation enables it to bypass the CD4-induced conformational change necessary for target cell entry (23, 32).

Down-regulation of receptor expression is a classical mechanism used to allow viral interference. An example of this is the ability of HIV and simian immunodeficiency virus to down-regulate the cell surface expression of CD4, their primary receptor. Down-regulation of the coreceptor CXCR4 by X4 CD4-dependent HIV-1 has never been characterized, although it has been established that CXCR4 is down-regulated by a few HIV-2 isolates which use CXCR4 as their primary receptor (12). A CD4-independent HIV-1 isolate, HIV-1/IIIBx, has recently been derived from the parental isolate. However, it failed to down-regulate CXCR4 on chronically infected cells (17). Other viral families induce down-regulation of CXCR4, as is the case with HHV-6 and HHV-7. These viruses induce a markedly decreased level of CXCR4 gene transcription, without any significant alteration of the posttranscriptional stability of CXCR4 mRNA (34, 39). Nevertheless, unlike the m7NDK HIV-1 isolate, these viruses do not use CXCR4 as a receptor for viral entry (40).

A down-regulation of CXCR4 in cells chronically infected with m7NDK isolate was expected to occur to allow cell survival. If this coreceptor, or main receptor in this case, was present on the surface of infected cells, syncytium formation would result, leading to cell death.

The results here presented support the concept of retroviral interference. They show that a virus, which derived spontaneously and which uses CXCR4 as a primary receptor, must down-regulate this receptor to maintain chronic expression in the infected cell line. However, the down-regulation of CXCR4 here described brings new insights into the mechanisms used by the viruses to achieve this. While CD4 is down-regulated in CD4-dependent HIV-1 by a number of different proteins that interfere with its stability and subcellular localization, CXCR4 is down-regulated by the m7NDK isolate primarily at the transcriptional level. We do not exclude the hypothesis of a retention of gp120-CXCR4 complex in the ER followed by degradation of CXCR4. However, since the steady states of CXCR4 mRNA and proteins levels are very much diminished, this mechanism, although possible, would thus occur with relatively low or undetectable efficiency.

Besides the relevance of these findings in relation to new aspects of CXCR4 expression regulation, the future identification of the mechanism used by the m7NDK HIV-1 isolate to achieve this modulation may provide new insights into HIV-1-cell interactions and could be a useful tool for the development of new prophylaxis concepts against HIV-1 infection.