Activation of the Human Immunodeficiency Virus-1 Long Terminal Repeat by Respiratory Burst Oxidants of Neutrophils

PHAGOCYTES (neutrophils, eosinophils, monocytes/macrophages) respond to stimulation with a burst of oxygen consumption and much, if not all, of the extra oxygen consumed is converted initially to the superoxide anion and then to H₂O₂. Among the potent inducers of a respiratory burst in phagocytes is phorbol myristate acetate (PMA), which is a direct stimulus of protein kinase C (PKC), and opsonized zymosan, which acts as a phagocytic stimulus. Myeloperoxidase (MPO) (released from the cytoplasmic granules of neutrophils and monocytes) and eosinophil peroxidase (released from eosinophil granules) can react with the H₂O₂formed to oxidize a halide to form potent oxidants, which are toxic to ingested microorganisms and adjacent extracellular targets.

We have previously reported that neutrophils,1monocytes,2 and eosinophils,3 when appropriately stimulated, are viricidal to cell-free human immunodeficiency virus type 1 (HIV-1), and we have provided evidence for the involvement of the peroxidase-H₂O₂–halide antimicrobial system in this toxicity. Thus, the viricidal effect was inhibited by catalase, but not by heated catalase or superoxide dismutase (SOD) implicating H₂O₂ formed by the stimulus-induced respiratory burst. This conclusion was supported by the absence of a viricidal effect when neutrophils which lack a respiratory burst, ie, from patients with chronic granulomatous disease (CGD), were used, unless a source of H₂O₂ was added. Azide, a potent inhibitor of peroxidase, also was inhibitory, which is compatible with the involvement of the phagocyte peroxidase. This was supported by the finding that stimulated neutrophils and monocytes which lack MPO, ie, from patients with hereditary MPO deficiency, were not viricidal to HIV-1 unless MPO was added.

Biologic oxidants formed by phagocytes may also influence HIV-1 in other ways. The long terminal repeat (LTR) of HIV-1 contains promoter and enhancer sequences required for the initiation of gene transcription. H₂O₂ can activate the HIV-1 LTR introduced by transfection into HeLa,4Jurkat,5,6 and THP-1⁶ cells and can increase HIV-1 replication in the chronically infected human monomyelocytic cell line U-14,6 and the chronically infected T-lymphocyte cell line Jurkat.5 This effect of H₂O₂may be due, at least in part, to activation of Nuclear Factor κB (NF-κB), which can recognize NF-kB binding sites in the HIV-1 LTR and thus facilitate gene transcription.5 Other mechanisms, however, cannot be excluded. The activation by H₂O₂ of the HIV-1 LTR in THP-1 or Jurkat cells and of viral replication in U-1 cells is greatly potentiated by vanadate.6 Vanadate reacts with H₂O₂ to form peroxides of vanadate with heightened biologic activity.

We report here on the activation of the HIV-1 LTR in a T-lymphocyte–derived cell line by intact neutrophils stimulated either by PMA or opsonized zymosan.

PMA, sodium orthovanadate, and H₂O₂ (30% wt/wt) were obtained from Sigma Chemical Co (St Louis, MO). The H₂O₂ was assayed by its absorption at 230 nm, using an extinction coefficient of 81 mol/L⁻¹cm⁻¹. Human recombinant tumor necrosis factor-α (TNF-α) was generously provided by Genentech (South San Francisco, CA), and recombinant human interleukin-2 (IL-2) by Hoffman-La Roche (Nutley, NJ). Bovine liver catalase (CTS, 57,622 U/mg) was obtained from Worthington Biochem Corp (Freehold, NJ), and human erythrocyte catalase (50,000 U/mg) was obtained from Calbiochem (La Jolla, CA). The bovine liver catalase was dialyzed against water overnight and the catalase preparations were heated at 100°C for 20 minutes where indicated. Endotoxin contamination of the catalase preparations was determined using a Limulus Amebocyte Lysate kit (BioWhittaker, Inc, Walkersville, MD). Endotoxin was removed from the bovine liver catalase preparation by passage through a column containing polymyxin B immobilized on agarose (Detoxi-gel; Pierce, Rockford, IL). Superoxide dismutase (SOD, bovine erythrocytes, 5,000 U/mg) was obtained from Boehringer Mannheim Biochemicals (Indianapolis, IN). Bisindolylmaleimide 1 (BIM) and Gö 6976 were obtained from Calbiochem.

IG5 (Jurkat LTR_luc), a Jurkat T-cell–derived cell line containing a stably integrated HIV LTR-luciferase construct was obtained through the Research and Reference Reagent Program, Division of Acquired Immune Deficiency Syndrome (AIDS), National Institute of Allergy and Infectious Diseases (NIAID), National Institutes of Health (NIH) and maintained in RPMI-1640 containing 10 mmol/L Hepes buffer, 2 mmol/L L-glutamine, 50 U/mL penicillin, 50 μg/mL streptomycin sulfate, and 10% heat inactivated fetal bovine serum (RPMI-FBS) (GIBCO, Grand Island, NY). On the day of the experiment, the cells were suspended in RPMI without antibiotics or FBS (RPMI).

Neutrophils were isolated from venous blood collected from normal human volunteers and a CGD patient as previously described.7 The preparation contained greater than 97% neutrophils of which greater than 96% were viable as measured by trypan blue exclusion. The neutrophils were suspended in RPMI and used immediately.

Into 12 × 75-mm polypropylene tubes (Falcon 2063; Becton Dickinson Labware, Lincoln Park, NJ) were added 2 × 10⁶ Jurkat LTR_luc, the components indicated in the legends to the figures and table, and RPMI to a final volume of 2.0 mL. The tubes were incubated at 37°C in a CO₂ incubator (5% CO₂, 95% air) for 6 hours and the luciferase activity was determined using a luciferase assay kit (Promega Biotec, Madison, WI) and a Monolight 1500 luminometer (Analytical Luminescence Laboratories, San Diego, CA). Photons were counted over a 30-second period and designated as relative light units (RLU).

The components indicated in the legend to the figure were placed in 12 × 75 mm polypropylene tubes (Falcon 2063), which were incubated at 37°C in a CO₂ incubator (5% CO₂, 95% air). At intervals, chemiluminescence was measured over a 10-second period using the Monolight 1500 luminometer and the results expressed as RLU.

The results are presented as mean ± standard error of mean (SEM) and the Mann Whitney U rank-sum test (unpaired, two-tailed) was used to analyze differences for significance unless otherwise indicated (NS, not significant P > .05).

Figure 1 demonstrates the activation of the HIV-1 LTR in Jurkat T cells by normal neutrophils + PMA. Neutrophils alone had no effect on the HIV-1 LTR at concentrations ranging from 3 × 10⁴ to 10⁷/mL. The Jurkat cell concentration was 10⁶/mL so that the ratio of neutrophils:Jurkat cells ranged from 0.03:1 to 10:1. The LTR in Jurkat cells was activated by PMA alone (background RLU 20,218 ± 2,725; PMA 373,473 ± 65,262, n = 9, P < .001). However, when neutrophils and PMA were combined, activation of the LTR was significantly increased at a neutrophil concentration of 3 × 10⁵/mL, reached a maximum at 10⁶/mL (neutrophil:Jurkat ratio 1:1) and then decreased to a level significantly below that of PMA alone when the neutrophil concentration was increased to 10⁷/mL. Activation of the LTR in the presence of PMA was significantly greater than that in the absence of PMA (P < .05) at all the neutrophil concentrations used except 10⁷/mL. When normal neutrophils were replaced by neutrophils from a patient with CGD, the stimulatory effect in the presence of PMA was lost. The decrease in activity at high neutrophil concentration, however, was still evident.

A PMA dose-response curve is shown in Fig2. A significant effect of PMA alone was observed at concentrations down to 1 ng/mL, whereas when PMA and 10⁶ neutrophils/mL were combined, activation of the LTR was significantly greater than that produced by PMA alone at PMA concentrations down to 3 ng/mL. The decrease in activity at high neutrophil concentrations (10⁷/mL) was observed at all of the PMA concentrations used.

Catalase had no effect on the LTR when added alone or with either neutrophils (10⁶ or 10⁷/mL; data not shown) or PMA (Fig 3). However, when catalase was added to cells exposed to a combination of neutrophils (10⁶/mL) and PMA, a significant inhibition was observed, with the activation of the LTR decreasing to the level observed with PMA alone. This inhibitory effect of catalase was lost on its heat inactivation. At the high neutrophil concentration (10⁷/mL) where activation of the LTR by neutrophils + PMA was lower than that of PMA alone (Fig 1), the addition of catalase had no effect (Fig 3). SOD at 1 μg/mL had no effect on the activation of the LTR under any of our experimental conditions.

The catalase used in Fig 3 was from bovine liver. It contained 58 ng endotoxin per mg protein as measured by the Limulus Amebocyte Lysate assay. Passage of the catalase preparation through a polymyxin B column resulted in the removal of greater than 99% of the endotoxin. The resultant endotoxin-poor catalase preparation inhibited the activation of the HIV-1 LTR by neutrophils + PMA to a degree comparable to that of the original catalase preparation, as did erythrocyte catalase, which contained 4.0 ng endotoxin per mg protein.

Vanadate greatly potentiates the activation of the HIV-1 LTR in Jurkat cells by H₂O₂.6 Under the conditions used in Fig 4, vanadate at 3 × 10⁻⁵ mol/L had no effect on the activation of the LTR by a combination of neutrophils (10⁶/mL) and PMA. However, when the neutrophil concentration was decreased to 10⁵/mL, activation by neutrophils + PMA was decreased and now a significant stimulation by vanadate was observed.

Azide is an inhibitor of hemeproteins, such as catalase and MPO, which degrade H₂O₂. Under the conditions used in Fig 5A,activation of the LTR by neutrophils + PMA was significantly increased by azide at 3 × 10⁻⁶ mol/L, whereas, when the azide concentration was increased to 10⁻⁴ to 10⁻³ mol/L, activation of the LTR was inhibited. Azide at concentrations ranging from 3 × 10⁻⁶ mol/L to 10⁻³mol/L had no effect on the activation of the LTR in Jurkat cells induced by PMA alone (Fig 5A). The stimulation by azide at 3 × 10⁻⁶ mol/L was not seen when normal neutrophils were replaced by CGD neutrophils.

These findings suggested that the H₂O₂ formed by the respiratory burst of PMA-stimulated neutrophils can activate the HIV-1 LTR in Jurkat T cells. It was, therefore, surprising that another potent stimulus of the respiratory burst of neutrophils, opsonized zymosan, when combined with neutrophils, had only a small effect on the LTR (Table 1). The combination of neutrophils and opsonized zymosan was significantly different from background by paired (P < .01), but not by unpaired, analysis. Azide at concentrations ranging from 3 × 10⁻⁶ to 3 × 10⁻⁴ mol/L significantly increased the activation of the LTR by normal neutrophils + opsonized zymosan, with an optimum effect observed at 3 × 10⁻⁵ mol/L azide (Fig 5B). Azide had no effect on the activation of the LTR either alone (data not shown), with normal neutrophils alone (data not shown), with opsonized zymosan alone (Fig5B), or with CGD neutrophils + opsonized zymosan (Fig 5B). The activation of the LTR by opsonized zymosan-stimulated normal neutrophils in the presence of azide was strongly inhibited (P< .001), but not abolished (P < .001) by catalase, whereas heated catalase was without effect (Table 1), implicating H₂O₂.

The considerably greater responsiveness of the LTR to neutrophils stimulated by PMA as compared with neutrophils stimulated by opsonized zymosan, despite both acting as strong stimuli of the neutrophil respiratory burst appears to be due, at least in part, to a synergism between H₂O₂ and PMA in the activation of the LTR. Under the conditions used in Fig 6, H₂O₂ alone at the optimum concentration of 10⁻⁴ mol/L increased luciferase activity from a background of 23,379 RLU to 110,956 RLU and PMA alone at 100 ng/mL increased luciferase activity to 403,823 RLU. When H₂O₂ and PMA were combined, luciferase activity was 3,096,165 RLU, which was significantly greater than the effect of H₂O₂ (P < .005) or PMA (P < .005) alone or than the additive effect of H₂O₂and PMA (P < .005). Thus, under these conditions, H₂O₂ and PMA act together to produce an effect, which is considerably greater than the sum of the effects of each alone. Synergism with H₂O₂ was not observed when PMA was replaced by TNF-α or IL-2 at 100 U/mL (Fig 6). IL-2 alone had no effect on the LTR in Jurkat cells under our experimental conditions.

Because PMA activates PKC, the synergism between PMA and H₂O₂ may reflect a priming effect of PKC activation on the activation of the LTR by H₂O₂. This prompted a study of the effect of the PKC selective inhibitors BIM and Gö 6976 on the activation of the HIV-1 LTR in Jurkat T cells. BIM strongly inhibited the activation of the LTR by PMA, H₂O₂, H₂O₂ + PMA, and neutrophils + PMA, with an inhibitory effect observed at a concentration of 10⁻⁷mol/L, and in some instances, lower (Fig7). Gö 6976 was less effective with an inhibition observed at 10⁻⁶ mol/L in all instances. BIM and Gö 6976 at concentrations ranging from 3 × 10⁻⁹ to 10⁻⁶ mol/L were without effect on the activation of the LTR by TNF-α.

The effect of BIM on the respiratory burst of neutrophils stimulated with either PMA or opsonized zymosan as measured by either luminol or lucigenin-enhanced chemiluminescence is shown in Fig 8. Luminol-enhanced chemiluminescence by neutrophils is predominantly dependent on peroxidase-catalyzed reactions, whereas lucigenin-enhanced chemiluminescence is largely peroxidase-independent and related to superoxide anion release.8 Opsonized zymosan was more effective than PMA as a stimulus of luminol-enhanced chemiluminescence by neutrophils, whereas, when lucigenin-enhanced chemiluminescence was measured, PMA was more effective than opsonized zymosan as the stimulus. This may reflect a greater MPO release by the phagocytic stimulus opsonized zymosan. With both luminol and lucigenin, BIM almost completely inhibited PMA-induced chemiluminescence, whereas opsonized zymosan-induced chemiluminescence was decreased approximately 50%.

Activation of the HIV-1 LTR occurs by a variety of mechanisms, of which reaction of the product of the TAT gene with the TAR sequence in the LTR appears to be the most important. Activation of the LTR by H₂O₂ is an additional mechanism, raising the possibility that biologic sources of H₂O₂ such as stimulated phagocytes, may influence the replication of HIV-1. We report here that neutrophils, when combined with the stimulus PMA, strongly activate the HIV-1 LTR introduced by transfection into the Jurkat T-lymphocyte cell line. This activation by neutrophils and PMA is due to, at least, three processes: (1) the direct activation of Jurkat LTR_luc by PMA; (2) the activation of the LTR by H₂O₂ formed by the PMA-induced respiratory burst of neutrophils; and (3) synergism between H₂O₂ and PMA in the activation of the LTR.

The direct activation of the LTR in Jurkat cells by PMA could account for only 12% to 13% of the total luciferase synthesis induced by neutrophils (10⁶/mL) + PMA. Activation by PMA alone was unaffected by catalase or by vanadate (see also Kazazi et al6), which would argue against the involvement of H₂O₂, at least H₂O₂accessible to these agents. PMA is a potent activator of PKC. The involvement of PKC in the activation of the LTR by PMA is suggested by the inhibitory effect of the selective PKC inhibitors BIM and Gö6976. BIM had an IC₅₀ for PMA-induced LTR activation of 0.02 μmol/L and Gö 6976 was approximately 10-fold less effective.

That H₂O₂ formed by the PMA-induced respiratory burst of neutrophils is also responsible, in part, for the activation of the LTR is supported by the inhibitory effect of catalase, the stimulatory effect of vanadate at suboptimal neutrophil concentrations and the loss of activity when normal neutrophils were replaced by neutrophils from a CGD patient. The effect of catalase is lost on its heat-inactivation and is unrelated to endotoxin contamination. Catalase reduces the activation of the LTR by neutrophils + PMA to the level observed with PMA alone, as did the substitution of CGD for normal neutrophils, suggesting that the total additional effect on the addition of neutrophils to PMA is due to the H₂O₂ formed.

Although both PMA and H₂O₂ can individually activate the HIV-1 LTR in Jurkat cells,6 the effect of their combination is considerably greater than the sum of each alone. The combination of PMA and reagent H₂O₂increases the activation of the LTR to the level observed with neutrophils + PMA. This suggests that PMA acts not only directly and by inducing H₂O₂ formation by the neutrophils, but also by increasing the sensitivity of the LTR in Jurkat cells to activation by H₂O₂. Some selectivity for PMA was observed, as H₂O₂ did not synergize with either TNF-α or IL-2 in the activation of the LTR. A synergism between PMA and H₂O₂ in the activation of NF-κB has been reported,9 and this may form the basis for the synergistic activation of the LTR by these agents. Activation of the LTR by PMA in the presence of neutrophils or H₂O₂ is inhibited by BIM and Gö 6976 suggesting the involvement of PKC in the synergism. PKC consists of a number of isoforms, which have in common catalysis of the phosphorylation of protein serine/threonine residues. PKC has been implicated in the replication of HIV-1 at a number of sites including the activation of the HIV-1 LTR10,11 where PKC may act in part through its activation of NF-κB.12-14 Activation of the LTR by H₂O₂ alone also is inhibited by the PKC inhibitors raising the possibility of an autostimulatory mechanism by which activation of PKC by H₂O₂ amplifies the subsequent response to H₂O₂. The PKC inhibitors at concentrations up to 10⁻⁶ mol/L did not inhibit the activation of the LTR by TNF-α, indicating some selectivity for PMA- and/or H₂O₂-induced activation.

The strong activation of the LTR by neutrophils + PMA reached a maximum at 10⁶ neutrophils/mL and then decreased to levels below that of PMA alone when the neutrophil concentration was increased to 10⁷/mL. The mechanism of this decrease in activity at high neutrophil concentration is unclear. The generation of reactive oxygen species as measured by chemiluminescence was increased at high neutrophil concentration raising the possibility of a toxic effect of products of the respiratory burst on the Jurkat LTR_luccells; however, the decrease in LTR activation was not prevented by catalase suggesting that H₂O₂ was not a toxic species under these conditions. Neutrophils heated at 65°C for 15 minutes lose their ability to respond to PMA with a respiratory burst, but remain inhibitory of the LTR at high concentration (0.9 × 10⁷ heated neutrophils + 10⁶ normal neutrophils/mL) (data not shown). It is possible that a product of neutrophils unrelated to the respiratory burst is toxic to Jurkat cells at high neutrophil concentrations.

When opsonized zymosan was used instead of PMA as the stimulus of the neutrophil respiratory burst, activation of the LTR was considerably decreased. A small, but significant, activation was observed, which was less than 1% of that observed with neutrophils + PMA. Unlike PMA, opsonized zymosan did not have a direct effect on the Jurkat LTR_luc nor did opsonized zymosan act synergistically with H₂O₂ to activate the LTR (data not shown). Azide is a potent inhibitor of the heme enzymes catalase and myeloperoxidase, which degrade H₂O₂ and, thus, the amount of H₂O₂ detected in the extracellular fluid on stimulation of neutrophils by PMA or opsonized zymosan is greatly increased by azide.15 The addition of azide thus would be expected to increase the availability of H₂O₂ for LTR activation. When Jurkat LTR_luc were exposed to neutrophils + opsonized zymosan in the presence of azide, increased activation of the LTR was observed, which was strongly inhibited by catalase, but not by heated catalase, and was not seen when normal neutrophils were replaced by CGD neutrophils. In contrast, addition of azide to neutrophils stimulated by PMA had only a small stimulatory effect at low (3 × 10⁻⁶ mol/L) azide concentration, possibly due to the strong activation in the absence of azide.

Our findings suggest that H₂O₂ generated by the respiratory burst of neutrophils may influence the survival and replication of HIV-1 in two opposing ways: (1) by reaction with peroxidase and a halide to produce a viricidal effect1; and (2) by activation of the HIV-1 LTR with consequent stimulation of gene transcription and viral replication. The latter effect is favored by the addition of azide (which promotes the accumulation of H₂O₂ and its release into the extracellular fluid) or vanadate (which forms a bioactive complex with H₂O₂). Further, activation by H₂O₂ is amplified by procedures, such as the addition of PMA, which activates PKC. Activation of PKC occurs when receptor-mediated hydrolysis of inositol phospholipids induced by a variety of stimuli, including mitogens, growth factors, and cytokines results in the release of diacylglycerol.16H₂O₂, as well as other forms of oxidant stress, can also activate PKC.17-21 Activation of PKC by these and other mechanisms would be expected to increase the sensitivity of the HIV-1 LTR to further activation by H₂O₂generated by adjacent phagocytes, certain microorganisms such as vaginal lactobacilli, as well as from other sources.

We thank Peggy Sue O’Brien for the preparation of the manuscript.