Differential Regulation of HIV-1 Fusion Cofactor Expression by CD28 Costimulation of CD4+ T Cells

See allHide authors and affiliations

Science  11 Apr 1997:
Vol. 276, Issue 5310, pp. 273-276
DOI: 10.1126/science.276.5310.273


Activation of CD4+ T lymphocytes from human immunodeficiency virus–type 1 (HIV-1)–infected donors with immobilized antibodies to CD3 and CD28 induces a virus-resistant state. This effect is specific for macrophage-tropic HIV-1. Transcripts encoding CXCR4/Fusin, the fusion cofactor used by T cell line–tropic isolates, were abundant in CD3/CD28-stimulated cells, but transcripts encoding CCR5, the fusion cofactor used by macrophage-tropic viruses, were not detectable. Thus, CD3/CD28 costimulation induces an HIV-1–resistant phenotype similar to that seen in some highly exposed and HIV-uninfected individuals.

Infection of CD4+ T lymphocytes with HIV-1 is initiated by binding of the viral envelope glycoprotein gp120 to the CD4 receptor on the lymphocyte surface, followed by fusion of the virus with the cell membrane (1). Virus-cell fusion is mediated by members of the chemokine receptor family, with different members serving as fusion cofactors for macrophage-tropic (M-tropic) and T cell line–tropic (TCL-tropic) isolates of HIV-1 (2).

HIV-1 infection is also influenced by the host T cell activation state. Commitment to T cell activation requires T cell receptor (CD3) engagement as well as a costimulatory signal. Interaction of the costimulatory molecule CD28 with its ligands CD80 or CD86 provides a necessary costimulus to induce an immune response (3). In addition to promoting the long-term polyclonal proliferation of CD4+ T cells in the absence of exogenous cytokines or feeder cells, activation with immobilized antibodies to CD3 (anti-CD3) and CD28 (anti-CD28) specifically induces a potent anti-HIV effect (4). This resistance is due in part to an enhanced production of the β-chemokines RANTES, MIP-1α, and MIP-1β (5), which block infection by M-tropic isolates (6). However, CD3/CD28-stimulated cells express an additional, cis-acting component of resistance specific to costimulation with immobilized anti-CD28 (5). This intrinsic CD3/CD28-specific antiviral effect was examined by comparing the HIV-1 infection process in cells stimulated with either immobilized anti-CD3 and anti-CD28 or with the mitogenic lectin phytohemagglutinin (PHA) and interleukin-2 (IL-2).

Purified CD4+ lymphocytes obtained from uninfected donors were cultured either with PHA and IL-2 or with beads coated with the CD3 monoclonal antibody OKT3 and the CD28 monoclonal antibody 9.3. The cells were infected with either the M-tropic isolate HIVUS1or the TCL-tropic isolate HIVNL4-3 (7) 3 days after stimulation, and the kinetics of virus replication were assessed by measuring p24Gag antigen production (Fig.1). Incubation of PHA/IL-2–activated CD4+cells with either HIVUS1 or HIVNL4-3 resulted in a productive infection. When CD3/CD28-activated CD4+cells were infected with HIVUS1, p24Gagantigen production was virtually undetectable throughout the experiment, in agreement with our previous observation that CD3/CD28-activated cells are resistant to infection with the M-tropic isolate HIVBa-L (4). However, when CD3/CD28-stimulated cells were infected with the TCL-tropic isolate HIVNL4-3, a productive infection ensued (Fig.1).

Figure 1

The CD28 antiviral effect is restricted to M-tropic isolates of HIV-1. Peripheral blood lymphocytes were obtained by apheresis of healthy donors, and CD4+cells were purified by negative selection as described (16). The purified cells were stimulated with DYNAL M-450 beads coated by tosyl conjugation with equal quantities of anti-CD3 [OKT3, mouse immunoglobulin 2a (IgG2a); American Type Tissue Collection], and anti-CD28 (9.3, mouse IgG2a) in medium containing IL-2 (100 U/ml; Boehringer Mannheim) (4). Alternatively, purified CD4+ cells were stimulated with PHA [5 μg/ml; Sigma] and IL-2 (100 U/ml; Boehringer Mannheim). Three days after stimulation, 7 × 106 CD4+ cells stimulated with anti-CD3/anti-CD28 (open symbols) or PHA/IL-2 (filled symbols) were infected with 1 × 104TCID50 (median tissue culture infectious dose) of HIVUS1 (squares) or with 1 × 104 MAGI (17) infectious doses of HIVNL4-3 (circles). After 2 hours at 37°C, the cells were washed three times and refed with 50% conditioned medium to a final concentration of 1 × 106 cells per milliliter. At the designated time points, cleared supernatant was analyzed for the presence of p24Gagantigen by enzyme-linked immunosorbent assay (Coulter). Each experiment was done at least five times, and representative values are shown.

CD28-mediated resistance to M-tropic viruses requires costimulation with bead-immobilized anti-CD28 because stimulation of CD4+ cells with bead-immobilized anti-CD3 and soluble anti-CD28, or with bead-immobilized anti-CD3 and IL-2, renders the cells sensitive to infection with M-tropic viruses (4). To further investigate the specificity of the CD28-mediated antiviral effect, we prepared beads containing anti-CD3 in combination with antibodies to the cell surface coreceptors CD2, CD4, CD5, CD7, and major histocompatibility complex (MHC) class I. Binding of antibodies to these coreceptors in conjunction with anti-CD3 treatment increases cellular proliferation (8). Purified CD4 cells were stimulated with beads coated with anti-CD3 and antibodies to the various surface receptors. Interleukin-2 was added to all cultures to ensure that all combinations of immobilized antibodies resulted in equivalent cell proliferation (9). Only cells stimulated with immobilized anti-CD3/anti-CD28 were resistant to infection with the M-tropic isolate HIV-1Ba-L (7) (Fig.2). Quantitative analysis indicated that the resistance was robust (9).

Figure 2

Resistance of CD4+ cells to infection by M-tropic HIV isolates requires costimulation with immobilized antibodies to CD28. Purified CD4+ cells were stimulated with either PHA and IL-2 (lanes 1 to 4) or with beads coated with equal quantities of anti-CD3/anti-CD28 (lanes 5 to 8), anti-CD3/anti-MHC class I (lanes 9 to 12), anti-CD3/anti-CD2 (lanes 13 to 16), anti-CD3/anti-CD4 (lanes 17 to 20), anti-CD3/anti-CD5 (lanes 21 to 24), and anti-CD3/anti-CD7 (lanes 25 to 28) in medium containing IL-2 (5). Three days after stimulation, 5 × 106 CD4+ cells stimulated by each method were infected with 1 × 104 TCID50 of HIVBa-L. Cells (1 × 106) were harvested immediately after virus addition (hour 0), after virus washout (hour 2), and at designated time points thereafter. HIV-1 gag DNA sequences present in crude cell lysates were quantitated by means of a previously described PCR-based assay (18). Quantitative results are shown (9). All exposures were for 1 hour.

To determine the nature of the M-tropic–specific block, we examined early events in the viral replication cycle by monitoring reverse transcription in HIV-1–infected CD3/CD28- and PHA/IL-2–stimulated CD4+ cells (Fig. 3). DNA polymerase chain reaction (PCR) was performed with oligonucleotide primers designed to detect reverse transcription products synthesized before the first-strand switch [“strong stop” DNA (10)]. Cells stimulated with anti-CD3/anti-CD28 or PHA/IL-2 were infected with HIVBa-L, HIVUS1, or HIVNL4-3. In PHA/IL-2–treated cells, strong stop DNA was detectable shortly after infection with all three viruses, and the level increased for the duration of the experiment. Furthermore, in CD3/CD28-stimulated cells infected with HIVNL4-3, strong stop DNA products accumulated rapidly, confirming that the CD28 antiviral effect was restricted to M-tropic viruses. In contrast, little or no strong stop DNA was detected in anti-CD3/anti-CD28–treated cells infected with either HIVBa-L or HIVUS1.

Figure 3

Infection of CD3/CD28-costimulated cells by M-tropic HIV-1 isolates is blocked at or before the initiation of reverse transcription. Purified CD4+ cells were stimulated with PHA/IL-2 (left) or anti-CD3/anti-CD28 (right) as described in Fig. 1. Three days after stimulation, 5 × 106 cells were infected with 1 × 104TCID50 of the M-tropic isolates HIVUS1 and HIVBa-L or 1 × 104 MAGI infectious doses of the TCL-tropic isolate HIVNL4-3, as described in Fig. 1. Virus stocks were treated with deoxyribonuclease (Boehringer Mannheim) before harvesting to degrade contaminating viral DNA. Cells were harvested immediately after infection (lanes marked 0), as well as 2, 6, 12, 24, and 72 hours after infection, as indicated. HIV DNA was detected as described in Fig. 2. Early reverse transcription (strong stop) products were amplified by use of the following primers: 5′-GGCTAACTAGGGAACCCACTG-3′ (sense) and 5′-CTGCTAGAGATTTTCCACACTGAC-3′ (antisense) (10). Products were detected by liquid hybridization with an end-labeled oligonucleotide probe (5′-CCGTCTGTTGTGTGACTCTGGTAACTAGAG-3′). The small amount of strong stop DNA present in time zero samples most likely represents reverse transcription products initiated within the virion (19). Input cell equivalents were standardized by amplification of human β-globin DNA sequences. Amplified β-globin DNA sequences are shown immediately underneath the HIV DNA panels.

The failure of M-tropic HIV-1 isolates to initiate reverse transcription in CD3/CD28-stimulated CD4+ cells suggested that a prior event in the replication cycle, such as viral binding or entry, was impaired. Because CD3/CD28-stimulated cells and PHA/IL-2–stimulated cells express equivalent levels of surface CD4 (4), the ability of activated CD4+ T cells to support membrane fusion by envelope glycoproteins from different viral isolates was analyzed with the use of a β-galactosidase reporter gene–based cell fusion assay (11). PHA/IL-2–stimulated CD4+ cells fused with cells expressing either TCL-tropic or M-tropic HIV-1 envelopes. In contrast, whereas CD3/CD28-activated cells fused efficiently with cells expressing TCL-tropic envelopes, they failed to fuse with cells expressing M-tropic envelopes. This experiment demonstrated that the block in the ability of M-tropic viruses to enter CD3/CD28-stimulated cells was at the level of envelope-mediated membrane fusion. The high level to which CD3/CD28-stimulated cells fused with cells expressing the LAV envelope is consistent with the susceptibility of CD3/CD28-activated cells to infection by TCL-tropic isolates.

The recent identification of CXCR4/Fusin and CCR5 as fusion cofactors for TCL-tropic and M-tropic HIV-1 isolates, respectively (2), prompted us to examine CD4+ cells stimulated with PHA/IL-2 or anti-CD3/anti-CD28 for transcripts encoding these chemokine receptors (Fig. 4). RNA was harvested from cells 24 to 72 hours after stimulation with either anti-CD3/anti-CD28 or PHA/IL-2, as well as from unstimulated cells. The RNA was analyzed by Northern blotting with probes specific for CXCR4/Fusin and CCR5. CXCR4/Fusin transcripts (1.8 kb) were detected at low abundance in unstimulated CD4+ cells, and stimulation with either anti-CD3/anti-CD28 or PHA/IL-2 induced a rapid increase in CXCR4/Fusin transcript levels (Fig. 4). Transcripts encoding CCR5 were not detected in unstimulated cells. Although 4.0-kb CCR5-specific transcripts were detected shortly after PHA/IL-2 stimulation of CD4+ cells, no CCR5 transcripts were detected at any point in CD3/CD28-stimulated cells.

Figure 4

Expression of chemokine receptor transcripts is differentially regulated in CD3/CD28-stimulated CD4+ cells. RNA was isolated from 5 × 107 unstimulated CD4+ cells (lane 1) or from CD4 cells stimulated with PHA/IL-2 (lanes 2 and 3) or anti-CD3/anti-CD28/IL-2 (lanes 4 and 5) at the indicated times after stimulation with RNAstat (Teltest). Total RNA (20 μg) was separated on agarose-formaldehyde gels and transferred to Zeta-probe membranes (Bio-Rad). The membranes were hybridized initially with an end-labeled oligonucleotide probe specific for CCR5 (5′-CTTGATAATCCATCTTGTTCCACCCTGTGC-3′) (20). The sequence of the CCR5-specific oligonucleotide probe was chosen to distinguish between CCR5 and the closely related transcripts encoding CCR2A and CCR2B (2). The blots were stripped and then rehybridized with a random-primed 1.3-kb Eco RI CXCR4/Fusin gene fragment (21). The membranes were then stripped and hybridized with an end-labeled oligonucleotide probe specific for 28S ribosomal RNA (rRNA) (Clonetech) to ensure that equivalent amounts of RNA were loaded in each lane. Transcripts were visualized by means of a Molecular Dynamics PhosphorImager. The positions of 18S and 28S rRNA are indicated by open arrows, whereas probe-specific bands are indicated by closed arrows. The image in (A) was obtained with a 2-hour exposure, and the image in (C) with a 10-min exposure, whereas the image in (B) was obtained with a 48-hour exposure.

These observations suggest that the susceptibility of CD3/CD28-stimulated cells to TCL-tropic viruses results from up-regulation of CXCR-4/Fusin mRNA expression, consistent with the high level of fusion between CD3/CD28-stimulated cells and cells expressing TCL-tropic envelope glycoproteins. Furthermore, the resistance of CD3/CD28-activated cells to infection by M-tropic viruses and primary isolates of HIV-1 correlates with the absence of detectable CCR5 mRNA expression. This result is consistent with the inability of CD3/CD28-stimulated cells to fuse with cells expressing M-tropic envelope glycoproteins.

Although the mechanism by which CCR5 expression is inhibited in CD3/CD28-stimulated CD4+ cells is unknown, CD28 costimulation exerts many effects on gene expression in general and cytokine expression in particular (3, 12). CD28-induced down-regulation of β-chemokine receptors may be a general feature in T cells, as Loetscher and colleagues recently reported that costimulation of CD4+ cells with anti-CD3/CD28 induced the down-regulation of CCR1 and CCR2 (13). In recent studies we have found that CD3/CD28-stimulated CD4+ cells produced high levels of β-chemokines in comparison to lectin-stimulated cells, and that the levels are similar to those in cells stimulated with a variety of costimuli such as CD3/CD5 (5). Thus, high levels of β-chemokines are not sufficient to mediate down-regulation of CCR5. Together, these results indicate that chemokine receptor expression is regulated by distinct forms of T cell activation and that chemokine receptor expression is not a consequence of T cell activation in general.

The progression to acquired immunodeficiency syndrome (AIDS) is associated with a shift from an M-tropic to a TCL-tropic viral phenotype (14). Although the selective forces driving this phenotypic transition are not well defined, T cell activation itself could be a selective force. CD3/CD28-stimulated CD4+ cells may exert selective pressure in favor of TCL-tropic isolate production through the combination of high levels of β-chemokine production (5) and lack of CCR5 expression. Presumably antigen/B7-stimulated CD4 T cells have a similar M-tropic resistance phenotype; however, further studies will be required to establish this. At least early in infection, selective forces other than the susceptibility phenotype of CD4+ target cells may be dominant, as individuals who are homozygous for a defective allele of CCR5 remain resistant to infection and heterozygotes are suggested to have a delayed progression of infection (15). These findings may have important consequences for immune reconstitution or gene therapy initiatives in HIV-infected individuals.

Our findings reveal a paradox. CD4 cells stimulated in vitro with anti-CD3 and anti-CD28 on a bead as an “artificial antigen-presenting cell” become resistant to HIV infection, whereas antigen presentation in vivo, as for example after vaccination, leads to enhanced viral replication. Further investigation will be required to establish the relevance, if any, of our findings to the in vivo pathogenesis of HIV infection.

Stimulation of cells with immobilized antibodies to CD3 and CD28 permits the large scale ex vivo expansion of primary CD4+cells, thus removing one of the largest obstacles to gene therapy or immune replacement therapy for HIV-1–infected individuals (4). Indeed, persistent increased CD4 counts and a lack of spikes in viral load have been noted in a clinical trial that is currently in progress to test the CD28 antiviral effect in patients with intermediate stage HIV infection. Our results demonstrate that lymphocyte proliferation and HIV-1 fusion cofactor expression can be unlinked. Such cellular activation regimes could permit the development of new therapeutic approaches for HIV-1–infected individuals.

  • * To whom correspondence should be addressed. E-mail: rin0cxj{at}


Stay Connected to Science

Navigate This Article