Inhibition of Phosphodiesterase Type IV Suppresses Human Immunodeficiency Virus Type 1 Replication and Cytokine Production in Primary T Cells: Involvement of NF-κB and NFAT

Abstract

Rolipram, a phosphosdiesterase type IV-specific inhibitor, prevented p24 antigen release from anti-CD3-activated human immunodeficiency virus (HIV)-infected T cells and CD4⁺-cell depletion associated with viral replication in a dose-responsive manner but minimally inhibited T-cell proliferation. Moreover, rolipram reduced the production of tumor necrosis factor alpha (TNF-α) and interleukin-10 (IL-10) by HIV-infected T cells. The transcriptional ability of a luciferase reporter gene under control of the HIV long terminal repeat, induced by phorbol myristic acetate plus ionomycin or by TNF-α, in primary T and Jurkat cells was also inhibited by rolipram. Rolipram inhibited NF-κB and NFAT activation induced by T-cell activation in Jurkat and primary T cells, as measured by transient transfection of reporter genes and electrophoretic mobility shift assays. Exogenous addition of TNF-α in the presence of rolipram restored NF-κB but not NFAT activation or p24 release. Addition of dibutyryl-cyclic AMP (dBcAMP) mimicked the effects of rolipram on p24 antigen release, NF-κB activation, and TNF-α secretion, but it did not affect NFAT activation or IL-10 production. The protein kinase A inhibitor KT5720 prevented the inhibition of TNF-α secretion but not that of HIV type 1 (HIV-1) replication caused by rolipram. Our data indicate that blockade of phosphodiesterase type IV could be of benefit against HIV-1 disease by modulating cytokine secretion and transcriptional regulation of HIV replication, and they suggest an important role of NFAT in HIV replication in primary T cells. Some of those activities cannot be ascribed solely to its ability to increase cAMP.

The pathogenic mechanisms underlying human immunodeficiency virus type 1 (HIV-1) infection and disease are extremely complex (5, 27, 28, 43). Virological as well as immunological factors contribute to pathogenesis. During infection, integrated provirus may stay in an inactive state until the appropriate activation signals stimulate viral transcription in the infected immune cells (13, 32, 74). Thus, the triggering of HIV expression by T-cell activation is dependent on the interplay of viral regulatory proteins and induced host factors that bind to their specific elements present both in the promoters of crucial genes involved in T-cell activation and in the long terminal repeat (LTR) of the virus (reviewed in references 17 and 33). In this way, at the same time, cellular activation and viral production are induced. The HIV-1 LTR contains binding sites for several mammalian transcription factors, including NF-κB (22, 53, 57), Ets (42, 67), NFAT (20), and Sp-1 (35, 48, 57). Among those, the main inducible element is the core enhancer element (position −104 to −81), which binds to nuclear factor kappa B (NF-κB) (53). This factor has been shown to play a very important role in LTR-driven transcription in primary T cells (2). NF-κB is composed of hetero- and homodimers of a family of proteins involved in gene regulation (7, 8, 44). T-cell activation by protein kinase C (PK-C) through the T-cell receptor or by tumor necrosis factor alpha (TNF-α) activates NF-κB by inducing this nuclear translocation (37, 46, 51). Very recently, nuclear factor of activated T cells (NFAT) has been also implicated in the control of HIV replication (10, 39). It seems to bind to the same NF-κB core element and not to the putative NFAT site (−255 to −217) in the LTR (39). NFAT activity has been defined as a complex family of transcriptional regulators distantly related to NF-κB through the Rel homology domain. Resting T cells express inactive NFAT molecules in their cytoplasm that upon T-cell activation translocate to the nucleus (16, 50, 55, 63).

On the other hand, HIV-1 infection is associated with increased production of a number of cytokines, which may be involved either in the induction of virus replication or in the pathogenesis of the immune dysregulation associated with disease progression (15, 27, 28). Several cytokines, including TNF-α, interleukin-1 (IL-1), IL-6, IL-2, and IL-12, have been shown to induce HIV replication when inhibited endogenously by neutralizing antibodies or added exogenously to chronically infected cells (reviewed in reference 61). More interestingly, studies with peripheral blood mononuclear cells (PBMC) from infected individuals or infected in vitro have indicated that HIV-1 replication is tightly regulated by autocrine secretion of some cytokines, such as TNF-α, IL-1β, and gamma interferon (40, 52, 72). Among all of the HIV-1-inducing cytokines, TNF-α is probably the most potent. TNF-α induces HIV-1 replication through activation of the transcription factor NF-κB, which binds to the LTR of HIV-1, increasing its transcription (23, 31, 56). On the other hand, opposite effects have been described for IL-10: IL-10 is able to inhibit HIV replication in macrophages by downregulating TNF-α secretion (73), whereas it stimulates HIV replication in other cell types (4, 9).

At least seven phosphodiesterase (PDE) isoenzyme families have been described so far; these were identified on the basis of selectivity towards the substrate (cyclic GMP [cGMP] versus cyclic AMP [cAMP]) as well as sensitivity to pharmacologic inhibitors (11). PDE type IV (PDE IV) is the predominant isoenzyme expressed in myeloid and lymphoid cells, having 50-fold more affinity for cAMP than for cGMP (11), although lymphocytes also possess PDE III, a cGMP-inhibited cAMP PDE (64). PDE IV is highly selective for cAMP, whereas PDE III degrades cAMP or cGMP with similar kinetics (11). Rolipram (RP) [(±)-4-(3′-cyclopentyloxy-4′-methoxyphenyl)-2-pyrrolidone] is a selective inhibitor of PDE IV (18) that has been employed in clinical trials, as an antidepressant drug, with safety and efficacy (25). More recently, some anti-inflammatory properties of RP have been described (reviewed in reference 70), which have been linked to the ability of RP to downregulate TNF-α synthesis (66, 68, 70). At the molecular level, those actions have been attributed to the ability of RP to increase the intracellular concentration of cAMP as a consequence of its selective inhibition of PDE IV (18).

Recently, RP has been described as a potent inhibitor of HIV replication in chronically infected cells, although its mechanism of action was not elucidated (3). Here, we have investigated the effect of blocking of PDE IV by RP on HIV-1-infected primary T cells. RP is able to block LTR-dependent transcription and HIV replication as well as cytokine (IL-10 and TNF-α) synthesis. Moreover, it also inhibited activation of the transcription factors NF-κB and NFAT. Exogenous addition of TNF-α in the presence of RP restored NF-κB activation but neither NFAT activation nor HIV replication. Also, intracellular increases in cAMP mimicked RP inhibition of NF-κB but not of NFAT activation or LTR-dependent transcription, indirectly suggesting an important role for NFAT in HIV replication in primary T cells.

MATERIALS AND METHODS

Cell cultures.

PBMC from healthy HIV-1-seronegative donors were isolated from whole blood by Ficoll-Hypaque (Pharmacia Fine Chemicals, Uppsala, Sweden) centrifugation and resuspended in RPMI 1640 medium supplemented with 10% fetal calf serum (FCS), basically as described previously (54). PBMC were depleted from monocytes by incubation in plastic dishes at 37°C for 2 h, and T cells were further purified by passing the nonadherent population through a nylon fiber wool column as described previously (54). The purity of this population (detected by flow cytometry), was always greater than 95% CD3⁺ cells. Purified T cells (10⁶/ml in RPMI medium containing 10% FCS) were cultured in six-well dishes and stimulated with immobilized (1 μg/ml applied to wells) anti-CD3 antibody (SPV3Tb; kindly provided by J. E. de Vries, DNAX, Palo Alto, Calif.) purified from ascitic fluid. The cells were infected with 5,000 HIV-1 (strain SP61; kindly provided by C. Lopez Galindez, Instituto Carlos III, Madrid, Spain) infectious particles/ml (or mock infected), along with different concentrations of RP (a generous gift of Schering Spain) or cAMP (Sigma, St. Louis, Mo.) and/or human recombinant TNF-α (10⁷ U/mg) (a generous gift of Antibioticos-Pharma, Madrid, Spain) or the PK-A-specific inhibitor KT5720 (Kamiya Biomedical Co., Thousand Oaks, Calif.). The cultures were incubated at 37°C and maintained in a humidified atmosphere containing 5% CO₂. At the 3rd and 6th days after infection, 50% of the culture supernatants was harvested, and the wells were replenished with an equivalent volume of fresh medium containing 5 U of recombinant human IL-2 (a gift from Eurocetus, Madrid, Spain) per ml to maintain a viable culture, together with the same concentrations of the respective reagents. None of the reagents affected the viability of the cells at the concentrations used, as indicated by the trypan blue dye exclusion test.

Proliferation and cytokine production assays.

Purified T cells were cultured as described above for 3 days. Culture supernatants were harvested at 3 days postinfection, and their cytokine contents were quantified by commercially available specific enzyme-linked immunosorbent assays (ELISAs) (Innogenetics, Zwijnaarde, Belgium, for TNF-α; Bender Medsystems, Vienna, Austria, for IL-10). Cell proliferation was evaluated by incorporation of [³H]thymidine (New England Nuclear, Boston, Mass.) into DNA during the last 16 h of culture. The cells were pulsed with 1 μCi of [³H]thymidine and harvested in glass fiber filters by using an automatic cell harvester, and radioactivity incorporation was measured in a liquid scintillation spectrometer. The assay was carried out for triplicate cultures.

HIV-1 p24 antigen assay.

Culture supernatants harvested at 9 days postinfection were assayed for viral p24 antigen content by using an antigen capture immunoassay (ELAVIA Ag; Pasteur Diagnostics, Paris, France).

Analysis of cell surface expression.

The percentage of the CD4⁺-T-lymphocyte subset present in T-cell cultures was evaluated by direct flow cytometry as previously described (54). Briefly, cells were incubated in RPMI supplemented with 5% FCS, in the absence or presence of different stimuli and reagents and under the same conditions described above. At 3, 6, and 9 days postinfection, the cells (1 × 10⁵ to 2 × 10⁵) were harvested and centrifuged three times in phosphate-buffered saline containing 2% bovine serum albumin (Sigma) and 0.1% sodium azide. They were subsequently incubated with fluorescein isothiocyanate-labeled anti-CD4 antibodies (Becton Dickinson), or with a fluorescein isothiocyanate-labeled irrelevant monoclonal antibody as negative control, for 30 min at 4°C. The cells were then washed in the above-described buffer, and surface fluorescence was determined in a FACScan spectrofluorimeter (Becton Dickinson). A minimum of 5,000 cells per point were analyzed.

Electrophoretic mobility shift assays (EMSA).

Nuclear extracts were obtained from T cells essentially by a previously described method (58, 59). The binding assays were performed as reported, using as labeled probes the double-stranded κB element of the HIV LTR (5′ TCCGCTGGGGACTTTCCGAGAG 3′) or the distal NFAT site from the IL-2 promoter (5′ GGAGGAAAAACTGTTTCATACAGAAGGCGT 3′). The binding complexes were separated in a 5% acrylamide gel, and their specificities were determined by competition with a 50× molar excess of the same unlabeled oligonucleotide (58, 59).

Transcription assays.

The reporter pLTRWT-luc expression plasmid was a generous gift of J. L. Virelizier and has been previously described (6). It carries the U3⁺R of the LTR of the LAI strain of HIV-1 from nucleotide −644 to +78. The cytomegalovirus (CMV)-Tat plasmid was a gift of J. Alcam and contains full-length HIV Tat under control of the CMV immediate-early promoter (2).

The reporter pNF-κB-luc expression vector contains three tandem copies of the NF-κB site of the conalbumin promoter driving the luciferase reporter gene and was also provided by J. Alcami (2). The NFAT-luc reporter plasmid contains three tandem copies of the NFAT site of the human IL-2 promoter; it was a generous gift of G. Crabtree and has been previously described (24). The AP1-luc reporter plasmid contains four tandem copies of the AP-1 site (−68 to −46) of the human CD11c promoter linked to the luciferase gene and was a generous gift of A. Corbi. The cAMP-responsive element (CRE)-luc expression plasmid contains four copies of the CRE site of the human choriogonadotropin α gene promoter (−147 to −129) fused to pT81 luc and has been previously described (65).

For transfection assays, resting purified T cells were resuspended in RPMI supplemented with 10% FCS and electroporated at 320 V and 1,500 μF by using a Bio-Rad Gene Pulser II with 1 μg of purified plasmid(s) per 10⁶ cells. After transfection, the cells were cultured at 37°C for 14 h before being activated with phorbol myristic acetate (PMA) (10 ng/ml) plus ionomycin (1 μM). Cells were incubated for an additional 12 h, harvested, and lysed. Luciferase activity was measured in a luminometer and expressed as relative luciferase units (RLU), calculated as (light emission from the experimental sample − light emission from untransfected cells)/10⁶ cells. In some experiments, data are represented as fold induction (observed experimental RLU/basal RLU in the absence of any stimulus).

For TNF-α-induced LTR transcription assays, transformed Jurkat T cells were used. For this the electroporation conditions were 280 V and 1,500 μF. Cells were stimulated with recombinant human TNF-α for 4 h, and luciferase activity was determined as described above.

RESULTS

Effect of PDE IV blockade on HIV-infected T cells.

Purified T cells, depleted of the great majority of monocytes, were activated through the T-cell receptor with immobilized anti-CD3 and infected or mock infected with HIV-1. We have shown previously that in this experimental system neither the absolute levels nor the kinetics of the proliferation of T cells in response to anti-CD3 were significantly altered by HIV infection during the first 6 days of culture (54). The addition of RP together with the virus to anti-CD3-stimulated T-cell cultures inhibited viral production measured as p24 antigen release (Fig. 1A). We also tested the effect of RP on the secretion of the cytokines TNF-α and IL-10, which were previously shown to influence HIV replication (4, 9, 23, 31, 40, 52, 56, 61, 72, 73), by the same cultures of purified human T cells infected with HIV-1 in response to anti-CD3 (Fig. 1B). Both cytokines were also strongly inhibited by RP. A hallmark of HIV infection is the depletion of CD4⁺ T cells. In agreement with this, we found that the number of viable CD4⁺ cells was lower in the cultures of HIV-infected T cells than in the controls and decreased over time. RP prevented this CD4⁺-cell depletion associated with HIV infection (Fig. 1C).

FIG. 1.

Effect of RP on activated HIV-1-infected T cells. Human T cells were stimulated with immobilized anti-CD3 antibody and infected with HIV-1 or mock infected. RP (100 μM) was added to the cultures as indicated. At 3, 6, and 9 days after infection, cultures were assayed for the p24 viral antigen (Ag) content in the supernatants by using an antigen capture immunoassay (A), TNF-α and IL-10 contents in the supernatants of infected cells by using specific ELISAs (B), or the number of CD4⁺ cells in infected or mock-infected cultures by direct flow cytometry (C). Data shown are the means ± standard deviations for triplicate cultures.

A summary of the dose-response effects of RP, using T cells from three different donors, is shown in Fig. 2. Doses of RP as low as 1 μM exerted a significant inhibitory effect (70%) on p24 release, and 30 μM was completely inhibitory in all experiments performed (Fig. 2). There was a dose-responsive inhibition of the production of IL-10 and TNF-α, and their sensitivities to RP inhibition were similar to that observed in p24 antigen release (Fig. 2). By contrast, RP poorly inhibited the proliferation of HIV-1-infected T cells. The effect of the drug (up to 1 mM) was not due to a toxic effect, since no decrease in viable cell number, tested by trypan blue exclusion, was observed (62).

FIG. 2.

Dose-response effect of RP on activated HIV-1-infected T cells. Human cells were stimulated with immobilized anti-CD3 antibody and infected with HIV-1. RP, at the indicated concentrations, was added to the cultures. Culture supernatants were assayed for p24 viral antigen (Ag) content 9 days after infection, using an antigen capture immunoassay. Proliferation was evaluated by [³H]thymidine incorporation during the last 16 h of culture, 3 days after infection. Cytokines in the supernatants were evaluated by ELISA 3 days postinfection. Results shown are the means ± standard deviations from three experiments with different donors, each one carried out in triplicate, standardized as percentages of control values. Unstimulated HIV-infected T cells did not produce detectable amounts of either p24 antigen or cytokines and did not proliferate (not shown).

Effect of cAMP on HIV-infected T cells.

So far, all of the actions of RP have been attributed to increases in intracellular cAMP due to its ability to block PDE IV (18). To test whether cAMP elevations were responsible for the above-described effects, we studied the effect of a permeable analog, dibutyryl-cAMP (dBcAMP), in our system. As seen in Fig. 3, dBcAMP exerted a strong dose-dependent inhibitory effect on TNF-α production and p24 antigen release (70% inhibition at around 10 μM). By contrast, it had a weak inhibitory effect on IL-10 secretion by anti-CD3-activated T-cell cultures, which was observed only at doses higher than 100 μM (Fig. 3). Interestingly, although also weak, dBcAMP had a more pronounced inhibitory effect than RP on cell proliferation.

FIG. 3.

Effect of dBcAMP on activated HIV-1-infected human T cells. Human T cells were stimulated with immobilized anti-CD3 antibody and were infected with HIV-1 as described in the legend to Fig. 1. dBcAMP at the indicated concentrations was added to the cultures. The results shown (means ± standard deviations for triplicate cultures) correspond to the same experiments as in Fig. 1 standardized as percentages of control values. Ag, antigen.

To further study the involvement of the cAMP–PK-A pathway in RP activities, we tested the ability of KT5720, a specific PK-A inhibitor, to overcome the inhibitory effect of RP on HIV-1 replication and cytokine secretion. As expected, KT5720 prevented dBcAMP inhibition of HIV-1 replication and TNF-α production. It also prevented the inhibition by RP of TNF-α secretion, but surprisingly, it minimally affected the inhibition of HIV replication (Fig. 4) or IL-10 (62) caused by RP. Taken together, these results indicate that activation of PK-A by intracellular cAMP cannot solely explain the inhibitory effect of RP on HIV replication.

FIG. 4.

Effect of PK-A inhibitors on RP and dBcAMP activities. Human T cells were stimulated with immobilized anti-CD3 antibody and infected with HIV-1. TNF-α secretion and p24 antigen (Ag) release were evaluated 3 and 9 days after infection as described in Materials and Methods. dBcAMP (300 μM), RP (100 μM), and KT5720 (200 nM), alone or in combination, were added to the cultures. Results are means ± standard deviations.

Effect of PDE blockade on the transcriptional activity of the HIV LTR.

To further characterize the mechanism of action of RP on HIV replication, we studied the ability of the drug to modulate LTR transcription. For this, we used a method that allows transfection of reporter genes in normal resting T cells and stimulated them with PMA plus ionomycin, a treatment that mimics T-cell activation through the T-cell receptor. Due to the nature of the electroporation-transfection assays, this pharmacological stimulus works better in transfected T cells than immobilized CD3. This treatment enhanced LTR activity by an average of sixfold. RP at 100 μM strongly inhibited this induction, whereas dBcAMP (300 μM) showed a minimal inhibitory effect (Fig. 5A).

FIG. 5.

Effect of PDE IV inhibition on nuclear factor and HIV-1 LTR-driven transcription. Human resting T cells were transfected with the reporter plasmid HIV-LTR-luc alone or cotransfected with pCMV-Tat (A) or with reporter plasmid NF-κB-luc, NFAT-luc, AP-1-luc, or CRE-luc (B). After 14 h, the cells were stimulated or not with PMA (10 ng/ml) plus ionomycin (IONO) (1 μM) in the presence or absence of RP (100 μM), dBcAMP (300 μM), or CsA (100 ng/ml) as indicated, and 12 h later the amount of luciferase in the cells was estimated. Shown are the means ± standard deviations from three experiments using T cells from three different donors.

Some reports indicate that LTR-controlled transcription in primary T cells requires the presence of the Tat protein of HIV (2, 45). Thus, in an attempt to reproduce experimental conditions closer to those occurring in HIV-infected T cells, we cotransfected T cells with a CMV-Tat plasmid expressing HIV-1 Tat under the control of the CMV promoter, which is able to drive Tat synthesis in the absence of T-cell activation. Under those conditions, a strong enhancing effect (an average of 22-fold) of ectopic Tat expression alone on the LTR promoter was seen, as previously reported (2). Interestingly, none of the drugs significantly inhibited Tat-mediated LTR transcription in primary T cells (Fig. 5A). Activation by PMA plus ionomycin together with Tat expression synergistically augmented LTR transcription (100-fold), and this enhancing effect of T-cell activation was inhibited by RP to a similar extent (Fig. 5A), compared to in the absence of Tat (Fig. 5A). Again, dBcAMP had no significant effect. Notably, the addition of the immunosuppressant cyclosporin A (CsA), an inhibitor of NFAT activation (29), strongly inhibited T-cell activation-induced LTR-dependent transcription in either the absence or presence of Tat. Similar effects of RP, CsA, and dBcAMP were observed in LTR-transfected Jurkat T cells in either the presence or absence of Tat expression (62).

During T-cell activation, several transcription factors required for HIV LTR-dependent transcription are induced (12, 33). Among those, NF-κB is considered one of the most important (2). Recently, NFAT has been also implicated (10, 29). Moreover, cytokine transcription in T cells is also dependent on the activity of several transcription factors, including NF-κB, NFAT, and AP-1 (16). Therefore, we also tested the effect of RP on primary T cells transiently transfected with reporter genes under control of NF-κB, NFAT, and AP-1 sites and stimulated with PMA plus ionomycin. Again, a good activation of those reporter genes can be detected in transfected normal resting T cells upon activation with PMA plus ionomycin (Fig. 5B). RP (100 μM) was able to inhibit by 60 to 80%, depending on the donor cells, the induction of NFAT activity. It also inhibited the activation of NF-κB, although always slightly less than that of NFAT. By contrast, AP-1 induction was enhanced by RP over the low levels already induced by PMA plus ionomycin. As expected, the activity of a reporter gene under the control of the CRE was also enhanced (Fig. 5B). Interestingly, dBcAMP had effects similar to those of RP on the activation of NF-κB, AP-1, and CRE reporter genes. In contrast, it minimally affected the induction of NFAT activity. As specificity controls, we used CsA and pyrrolidine dithiocarbamate, which completely inhibited NFAT and NF-κB, respectively (62).

To corroborate the observed inhibition of NF-κB and NFAT activation in the presence of RP, we tested the effect of RP in EMSA with normal T cells. Two NF-κB–DNA complexes (specifically competed by an oligonucleotide corresponding to the κB sequence of the HIV-LTR), were detected in the nuclei of HIV-infected and anti-CD3-activated T cells by EMSA (Fig. 6). The lower band has been previously shown to correspond to inactive p50 NF-κB homodimers, whereas the upper one contains the p50–c-Rel and p50-p65 heterodimers of the NF-κB family, which have strong transactivating activity (58, 59). In contrast, resting T cells expressed detectable levels of only p50 NF-κB homodimers in the nuclei. We also found that the continuous presence of RP in the culture inhibits the appearance of NF-κB binding activity in the nuclei of T cells at 14 h (Fig. 6) or any time (62) after activation. Similar inhibition by RP was found when the cells were stimulated with PMA plus ionomycin (62). Addition of dBcAMP to the cultures also produced inhibition of NF-κB binding.

FIG. 6.

Inhibition of NF-κB and NFAT activation by RP. Human HIV-infected T cells were stimulated with immobilized anti-CD3 antibody (αCD3) or with PMA (10 ng/ml) plus ionomycin (1 μM) (P+I) in the presence of RP (100 μM), dBcAMP (300 μM), CsA (100 ng/ml), or TNF-α (30 ng/ml) as indicated. The binding activity of NFAT (A and B) or NF-κB (C) in the nuclei of T cells was assayed 14 h later, using a κB-HIV or distal IL-2 NFAT site labeled probe as described in the text. Control specific binding was detected by using as a competitor a 50-fold excess of unlabeled oligonucleotide. uns, unstimulated.

Furthermore, activation by immobilized anti-CD3 or PMA plus ionomycin also induced the appearance of an NFAT complex in the nucleus, which can be competed out by the specific oligonucleotide. Its induced expression was blocked by CsA, an inhibitor of NFAT translocation (29). The induction of this complex was also completely inhibited by RP at 100 μM (Fig. 6). By contrast, dBcAMP did not inhibit NFAT activation, in agreement with the reporter data.

Effect of exogenous cytokines on RP activity.

Some of the actions of RP have been linked to its ability to downregulate TNF-α synthesis through cAMP elevation (66, 70). Moreover, autocrine TNF-α production is required for HIV-1 replication in our cultures (52). Therefore, TNF-α inhibition could be the mechanism by which RP inhibits HIV-1 replication. If this were the only mechanism by which RP affects viral replication, exogenous TNF-α should prevent RP inhibition in PMA plus ionomycin. To test this, we added exogenous TNF-α to the cultures and assayed its ability to inhibit RP activity. Although addition of TNF-α to the cultures had a significant enhancing effect on HIV-1 replication, it could not prevent the inhibitory effect of RP on HIV replication (Fig. 7). In contrast, TNF-α prevented NF-κB inhibition in EMSA (Fig. 6C). Moreover, it did not prevent the RP blockade of NFAT activation (Fig. 6B).

FIG. 7.

Effect of exogenous TNF-α on the inhibitory effect of RP. Human T cells were stimulated with immobilized anti-CD3 antibody and infected with HIV-1. RP (500 μM) and/or TNF-α (100 ng/ml) was added to the cultures as indicated. Shown is the mean p24 antigen (Ag) released (± standard deviation) in triplicate culture supernatants at day 9 after infection from two independent experiments.

To further study the relationship between PDE blockade and TNF-α-induced LTR transcription, transformed Jurkat T cells were used, since primary T cells do not express cell surface TNF receptors unless they are activated through the T-cell receptor (59). An inhibitory effect of RP on LTR-driven transcription was observed in Jurkat cells stimulated with TNF-α (Fig. 8). It is well established that activation of LTR by TNF-α (23, 56) depends on NF-κB activation. Therefore, this indicates that TNF-induced LTR-driven transcription mediated by NF-κB activation is also sensitive to RP inhibition. However, the inhibition depended on both the RP and TNF-α concentrations and could be reversed by increasing the TNF-α concentration (Fig. 8).

FIG. 8.

Effect of RP on TNF-α-induced LTR transcription. Jurkat cells were transfected with LTR-luc. After 14 h, the cells were stimulated with TNF-α (10 or 30 ng/ml) in the presence or absence of RP (50 or 100 μM), and 12 h later the amount of luciferase in the cells was estimated. Results are means ± standard deviations.

DISCUSSION

Establishing an experimental system that mimics or serves as surrogate for a physiological system, similar to those in which HIV replicates “in vivo,” is of great importance for understanding AIDS pathogenesis and the mechanisms of action of potentially antiretroviral drugs. However, most of the studies on HIV replication were carried out in already-activated T cells such as transformed T-cell lines or T-cell clones (27, 28, 61), and very little was known about the mechanism of HIV activation in normal resting T cells. In contrast to the case for resting T cells, HIV infection of lymphoblastoid cell lines results in intense replication even in absence of additional stimuli (1). Moreover, recent results indicate that requirements for HIV replication in normal and transformed cell lines are quite different (2), emphasizing the need to analyze the mechanism of viral activation in models more physiologically relevant that those represented by chronically infected cells. On the other hand, the use of inhibitors of the various steps of T-cell activation may also help in the elucidation of the basic mechanism underlying HIV replication. Here, we have employed primary T-cell cultures and efficient systems of transfections of normal resting peripheral blood T cells (2). These transfected normal resting human T cells provide a sensitive and physiologically relevant model for study of HIV transcription, and we used RP as an inhibitor of PDE IV (18) to clarify its role in HIV replication and T-cell activation.

We have shown in this report that specific blockade of PDE IV by RP inhibits viral replication in acutely infected T cells and prevents the depletion of CD4⁺ cells associated with HIV-1 infection. Furthermore, RP exerts a direct effect on LTR-dependent transcription, likely due to its inhibition of NF-κB as well as NFAT activation. Moreover, RP inhibits the production of several cytokines involved in controlling HIV replication, such as IL-10 and TNF-α, by HIV-infected human T cells (4, 9, 23, 31, 40, 52, 56, 61, 72) in response to activation by anti-CD3 but minimally affects T-cell proliferation. Similar poor sensitivity of T-cell proliferation to RP in uninfected cells has been reported (26).

At the molecular level, the most obvious mechanism leading to those effects caused by RP may involve cAMP-dependent pathways resulting from PDE inhibition (18). Moreover, the fact that the 50% inhibitory concentrations for IL-10, TNF-α, and HIV replication were very similar further suggested that all of these activities could be mediated by the same intracellular effect. However, augmentation of intracellular cAMP by dBcAMP cannot mimic some RP activities. Thus, dBcAMP had a very weak effect on the induction of the transcriptional activity of the LTR and on NFAT activation by PMA plus ionomycin in primary T cells as well as in Jurkat cells. In addition, IL-10 secretion by anti-CD3-activated T cells was also poorly inhibited by dBcAMP, in agreement with previous reports (60). Furthermore, inhibition of TNF-α but not of IL-10 secretion or of HIV replication by PDE IV blockade can be prevented by the PK-A inhibitor KT5720, indicating that only the RP effect on TNF-α can be exclusively ascribed to its ability to increase cAMP.

Elevated cAMP has been shown to inhibit NF-κB activation in transformed T cells, as measured by EMSA or by transient transfections of reporter genes (14, 34, 71). In contrast, elevation of intracellular levels of cAMP did not generally inhibit NFAT activation (14, 34), except when dBcAMP was used at a very high concentration (500 μM) (73), although it stimulated AP-1 (14, 34, 36, 69) or CRE binding factors (34). Our results with dBcAMP in primary T cells are in agreement with those. In contrast, RP inhibits both NF-κB and NFAT, whereas it stimulates AP-1 and CRE binding factors. The NF-κB site is missing in the IL-10 promoter (60), whereas it is present in the TNF-α gene (7). In contrast, NFAT is clearly required for TNF-α transcription, and IL-10 synthesis is sensitive to CsA, which is generally accepted as proof of NFAT involvement (63). Therefore, this unique NFAT inhibition by RP may explain why this drug and not dBcAMP inhibits IL-10 production. Recently, the synthesis of IL-5, which is dependent on NFAT (41, 63), was also shown to be inhibited by RP (30). Taken together, these results may suggest that some actions of RP (i.e., TNF and NF-κB inhibition) are mainly due to stimulation of the cAMP–PK-A pathway, whereas others are due to its ability to block other cellular functions, such as NFAT activation, or a combination of both. Experiments are in progress to elucidate the molecular link between PDE IV inhibition and decreased NFAT activity.

A previous report by Angel et al. (3) showed that RP inhibited HIV replication in chronically infected U1 cells. Those authors hypothesized that RP inhibited p24 antigen release by blocking TNF-α production. We have previously shown that blocking of autocrine TNF-α secretion decreases NF-κB activation (58, 59) and concomitantly HIV-1 replication (52) in our cultures of T cells. Also, many potential clinical uses of RP have been related to its ability to inhibit TNF-α synthesis (66, 70). Thus, the hypothesis that RP may decrease HIV replication by preventing TNF-α production was attractive. However, our results do not support such a simple hypothesis. This is based in the following findings: (i) supplementation of the cultures with TNF-α did not restore RP inhibition of p24 production, (ii) TNF-α but not inhibition of p24 antigen release in acutely infected cells was reversed by PK-A inhibitors, and (iii) TNF-α-induced LTR transactivation was also inhibited by RP. Taken together, our results pointed to a more complex mechanism of action by RP, suggesting that additional mechanisms for inhibition of HIV replication are operating.

Although RP at doses higher than 100 μM partially inhibited proliferation, arresting a proportion of the cells in the G₀/G₁ phase of the cell cycle, supplementation of the RP treated-cultures with IL-2 restored cell proliferation, but not NF-κB or NFAT (62), suggesting that the observed effects on transcription were not indirectly due to the inhibition of cell proliferation. Moreover, RP inhibition of NF-κB or NFAT activity was observed at any time tested after activation (as early as 4 h), which also does not support cell cycle effects being responsible for the observed activity.

As shown here, inhibition of NF-κB activation (induced by T-cell receptor, PK-C, or TNF-α) could be another mechanism by which PDE IV blockade may inhibit HIV-1 replication, since the activation of this nuclear factor is required for HIV-1 LTR transactivation (2). However, exogenous TNF-α prevented RP inhibition of NF-κB activation but not inhibition of antigen p24 release from the same T-cell cultures. Moreover, the inhibition by RP of TNF-α-induced LTR transcription was reversed by increasing TNF-α concentrations, and it is well established that activation of the LTR by TNF-α depends exclusively on NF-κB activation (23, 56). Together, those results suggest that other transcription factors important for HIV replication should be affected by PDE IV blockade, in addition to NF-κB. The disparate effects of dBcAMP and RP on LTR and NFAT activation, although they produced a similar inhibition of NF-κB activation, observed in the same cultures of primary T cells both in transient transfections of reporter genes and EMSA, as well as the LTR sensitivity to CsA, point to NFAT as a likely candidate. The fact that TNF could not restore either NFAT activation or p24 antigen release is also in agreement with this hypothesis.

There has been speculation that NFAT plays a role in HIV replication for some time, based on CsA sensitivity of HIV replication (21, 38). However, it has been reported that transactivation of the HIV enhancer is not dependent on NFAT (49), and deletion of the putative NFAT binding site (position −255 to −217) in the HIV LTR had no measurable effect on T-cell-dependent HIV gene expression in transformed Jurkat T cells (47). In contrast, very recently, Kinoshita et al. (39) have clearly shown that NFAT is a positive activator of HIV replication. This factor synergizes with NF-κB and binds to the core enhancer element (−104 to −81) instead of the NFAT site of the LTR (−255 to −217). In agreement with that, an interaction of NF-κB, NFAT, and Ets has been recently shown to be required for activation of HIV enhancers (10). According to those results, transactivation of the HIV LTR requires the formation of a trimolecular complex between NF-κB, NFAT, and Ets. This complex probably binds via NFAT to one and via NF-κB to the other of the two adjacent κB sites on the HIV-1 LTR. Both factors are inhibited by PDE IV blockade. Therefore, if NFAT remains inhibited despite recovery of NF-κB activation after exogenous addition of TNF-α, the active complex cannot form and transcription is blocked, as was observed in the RP cultures. Taken together, our results with RP, dBcAMP, and CsA, although indirect, strongly support an important role for NFAT, besides NF-κB, in HIV replication. Moreover, in contrast to the above-described reports, they do so in a more physiological model of primary T cells.

However, it is noteworthy that the concentrations of RP required to inhibit HIV replication in T-cell cultures were lower than those required to inhibit NF-κB, NFAT, or LTR transcription. The reason for those discrepancies may simply lie in the different sensitivities of the two types of assays. Alternatively, it may suggest that in normal T cells additional inhibitory mechanisms are operating, such as the blocking of cytokine (TNF-α and IL-10) secretion. As mentioned above, blocking of TNF clearly blocks HIV replication in many systems (40, 72), including our cultures (52). Since, IL-10 has been also shown to induce HIV replication by a TNF-α-dependent (9) as well as by a TNF-independent (4) mechanism, RP may also contribute to inhibit HIV replication by inhibiting IL-10 production. Thus, it seems likely that the inhibitory effect of PDE blockade on HIV replication in primary T cells is due to a combination of effects on TNF-α, NF-κB, and NFAT activation.

T-cell activation is critical for the immune response against infection. Paradoxically, this event is also critical for HIV replication and the progressive immune dysfunction associated with AIDS progression. Since HIV replication is dependent on NF-κB and NFAT, they could constitute alternative targets for therapy. Thus, therapeutic agents that inhibit them, such as RP or other PDE IV inhibitors, may be considered. In addition, the use of drugs for neutralizing inappropriately elevated production of certain cytokines, such as TNF-α and IL-10, in HIV-infected individuals is an objective of immune-based therapy for AIDS (12). Pentoxifylline, which is also a PDE inhibitor, although nonspecific, has similar properties (54) and has been considered for the treatment of AIDS (19). As RP is more potent than pentoxifylline, it may be a more effective inhibitor of HIV replication in patients when given at equivalent doses. Our results indicate that RP, by affecting HIV replication and cytokine production, may therefore be helpful as adjunctive therapy in AIDS.