HIV-1 gp120 Induces IL-4 and IL-13 Release from Human FcεRI⁺ Cells Through Interaction with the V_H3 Region of IgE¹

The existence of functionally polarized human CD4⁺ T helper responses based on their profile of cytokine secretion has been established (1). T_H1 cells produce IFN-γ and IL-2, whereas T_H2 cells produce IL-4, IL-5, and IL-13. A type 2 cytokine profile secreted by CD4⁺ T cells stimulates IgE levels (2, 3). Serum IgE levels are increased in adults and in children with HIV-1 infection (4, 5, 6, 7, 8). Elevated IgE levels in HIV-1-infected children and adults have been associated with the progression of HIV-1 disease (9, 10, 11). Thus, IgE levels could be a marker of poor prognosis in some patients in the early or late stages of HIV-1 infection (9, 10, 11).

These observations generated great interest and some controversy. Clerici et al. (12, 13) suggested that, during the early stages of HIV-1 infection, there is a switch from “T_H1-like” toward a “T_H2-like” pattern of cytokine production. This hypothesis was indirectly supported by the observation that IL-4 gene-targeted mice lacking T_H2 responses do not develop murine AIDS (14). However, the association with another gene product is a prerequisite for nondevelopment of murine AIDS (15). Graziosi et al. (16) did not detect an overall shift in the cytokines pattern toward the T_H2 subset in lymph nodes of HIV-1-infected individuals . In another study, Maggi et al. (17) did not find a bias toward T_H2-like cytokine patterns in T cell clones from HIV-1-infected individuals during the progression to AIDS . They demonstrated a preferential depletion of CD4⁺ T_H2-type cells in the advanced phases of HIV-1 infection and found that HIV-1 replicates preferentially in T_H2 rather than in T_H1 clones (17). Subsequent reports have added to the controversy (18, 19, 20, 21). The apparently conflicting results could be due to 1) technical reasons, 2) the production of T_H2-like cytokines by cell types other than lymphocytes, 3) stimulation by specific superantigens, or 4) cytokines other than IL-4. In fact, most studies have focused on IL-4 and IL-10, whereas recent data show that other cytokines such as IL-13 are critical for T_H2 cell polarization (22, 23, 24).

Basophils and mast cells are the only cells that synthesize histamine and express high-affinity receptors for IgE (FcεRI)³ (25, 26). Immunologic activation of human basophils generate and secrete a restricted profile of cytokines (IL-4 and IL-13) (27, 28, 29, 30) that are critical for T_H2 cell polarization (1, 2, 3, 22, 23, 24). Also, immunologically activated human mast cells synthesize IL-4 and IL-13 (31, 32, 33). Moreover, HIV Ags induce histamine release from basophils (34).

HIV-1 and HIV-2 destroy CD4⁺ lymphocytes, which leads to AIDS (35, 36). The entry of HIV into host cells is mediated by sequential interaction of the viral envelope glycoprotein (gp) gp120, with the CD4 gp (36, 37) and chemokine receptors on the cell surface (38, 39, 40). HIV-1 gp120 is a new member of the Ig superantigen family (41, 42, 43). Ig V_H3 gene products are the ligand for gp120 (44), and this interaction might explain the superantigen activation of human B lymphocytes in patients with AIDS (45). Protein Fv, an endogenous superantigen stimulated by viral infections in humans (46), interacts with the V_H3 domain of IgE to induce the release of IL-4 and IL-13 from human FcεRI⁺ cells (47, 48). In this study, we demonstrate that HIV-1 gp120 interacts with the V_H3 domain of IgE to induce the release of IL-4 and IL-13 from human FcεRI⁺ cells, thus acting as a viral superantigen.

The following were purchased: 60% HClO₄ (Baker, Deventer, The Netherlands); BSA, PIPES, hyaluronidase, collagenase, chymopapain, elastase type I, human serum albumin (Sigma, St. Louis, MO); HBSS, IMDM, and FCS (Life Technologies, Grand Island, NY); deoxyribonuclease I (Calbiochem, La Jolla, CA); RPMI 1640 with 25 mM HEPES buffer, Eagle’s MEM (Flow Laboratories, Irvine, Scotland); and Dextran 70, Percoll, and protein A-Sepharose (Pharmacia Biotech, Uppsala, Sweden). Rabbit anti-human-Fcε Ab was a generous gift from Drs. Teruko and Kimishige Ishizaka (La Jolla Institute for Allergy and Immunology, La Jolla, CA). The mAb against the α-chain of FcεRI was a generous gift from Dr. John Hakimi (Roche Research Center, Hoffman-LaRoche, Nutley, NJ).

The PIPES buffer used in these experiments was made up of 25 mM PIPES (pH 7.37), 110 mM NaCl, and 5 mM KCl. The mixture is referred to as P. PCG contained, in addition to P, 5 mM CaCl₂ and 1 g/L dextrose (47); pH was titrated to 7.4 with sodium bicarbonate. PBS contained 8.0 g/L NaCl, 2.89 g/L Na₂HPO₄ · 7H₂O, 0.2 g/L KH₂PO₄, 0.2 g/L KCl, (pH 7.3); TCF contained 8.0 g/L NaCl, 0.2 g/L KCl, 0.05 g/L NaH₂PO₄, 0.28 g/L NaHCO₃, and 1.0 d-glucose (pH 7.3); TGMD contained 0.25 g/L MgCl₂ · 6H₂O, 10 mg/l DNase, and 1 g/L gelatin in addition to TCF (pH 7.3) (46).

Monoclonal IgM were purified from the sera of patients with Waldenstrom’s macroglobulinemia by gel permeation, as described (46). Variable regions of these monoclonal IgM were determined using a well-characterized panel of primary sequence-dependent V_H and V_K family-specific reagents that identify framework regions previously described (46).

Basophils were purified from peripheral blood cells of normal subjects, aged 19–47 years (mean age, 36.1 ± 5.2 years), undergoing hemapheresis. Buffy coat cell packs from healthy volunteers provided by the Immunohematology Service at the University of Naples Federico II were reconstituted in PBS containing 0.5 g/l human serum albumin and 3.42 g/L sodium citrate and loaded onto a countercurrent elutriator (model J2-21; Beckman Instruments, Fullerton, CA). Several fractions were collected, and fractions containing basophils in large numbers (>20 × 10⁶ basophils) and of improved purity (>15%) were further enriched by discontinuous Percoll gradients (49). Yields by this technique ranged from 3 to 10 × 10⁶ basophils, with a purity from 74 to 98%, as assessed by basophil staining with Alcian blue and counting in a Spiers-Levy eosinophil counter (49).

Macroscopically normal lung tissue obtained from patients undergoing thoracotomy and lung resection, mostly for lung cancer, was dissected free from pleura, bronchi, and blood vessels, minced into 3- to 8-mm fragments, and dispersed into single-cell suspensions as described (46). Yields with this technique ranged between 3 × 10⁶ and 20 × 10⁶ mast cell, with a purity between 1 and 8%. The cells were resuspended and incubated overnight in RPMI 1640 containing 25 mM HEPES, 2 mM l-glutamine, 1% gentamicin, and 10% FCS as described previously (46). Mast cells isolated from lung parenchyma were fractionated in a Beckman elutriator (Beckman Instruments). Elutriation fractions with the greatest percentage of mast cells were pooled, and further purified by flotation in Percoll density gradients as described (46). The fractions rich in mast cells were then counted by Alcian blue stain.

Basophils (∼6 × 10⁴ basophils/tube) or mast cells (∼3× 10⁴ cells/tube) resuspended in PCG (histamine release) or IMDM (cytokine release) and 0.1 ml of the cell suspension were placed in 12- × 75-mm polyethylene tubes and warmed to 37°C; 0.1 ml of each preformed stimulus for release was added, and incubation was continued at 37°C for 45 min (histamine release), 4 h (IL-4 secretion), or 18 h (IL-13 secretion) (47, 48). At the end of this step, the reaction was stopped by centrifugation (1000 × g, 22°C, 2 min), and the cell-free supernatants were stored at −20°C for subsequent assay of histamine and cytokine content. The cell-free supernatants were assayed for histamine with an automated fluorometric technique (50). Total histamine content was assessed by lysis induced by incubation of cells with 2% perchloric acid before centrifugation. To calculate histamine release as a percentage of total cellular histamine, the spontaneous release of histamine from basophils (0–8% of the total cellular histamine) and mast cells (4–18% of the total cellular histamine) was subtracted from both the numerator and denominator (49). All values are based on means of duplicate or triplicate determinations. Replicates differed from each other in histamine content by <10%.

The harvested supernatants were assayed for IL-4 or IL-13 by using the IL-4 or IL-13 Quantikine high sensitivity kit (R&D Systems, Minneapolis, MN). The standard curve for these kits was run in the same medium used for the release experiments (47, 48).

RNA was isolated by harvesting the basophils from culture wells and centrifuging for 30 s at 10,000 × g. After removal of the supernatants, the cell pellet was extracted with RNAzol B (Tel-Test, Friendswood, TX), which is a modified guanidinium thiocyanate single-step procedure, as described (47).

An aliquot of total cellular mRNA was reversed transcribed to cDNA and PCR expanded using the GeneAmp RNA PCR Core kit (Perkin-Elmer International, Nieuwerkerk, The Netherlands), as described (47). In this protocol, RT was performed on 2 μl of the RNA extract (10–20% of the total RNA extracted). The RT mix (5 mM MgCl₂) was incubated for 20 min in a Perkin-Elmer/Cetus thermocycler (Perkin-Elmer/Cetus, Norwalk, CT) followed by 2 min at 95°C to inactivate the RT. Buffers, dNTPs (final concentration, 0.4 mM each), Amplitaq polymerase (1 U/50 μl reaction), and paired primers (0.5 μM each) were added to RT tubes (bringing MgCl₂ to 2 mM), and the PCR reaction was cycled according to the following protocol: denaturation at 95°C for 15 s, annealing at 60°C for 15 s and at 72°C for 30 s. IL-4 was cycled 56 times before a 15-min incubation at 72°C for a final extension. The primers for IL-4 and IFN-γ was performed and synthesized with commercial source (Life Technologies) based on the known cDNA sequences for the cytokines.

An aliquot (10 μl) of the reaction product was visualized on 2% agarose in buffer containing 0.5 μg/ml of ethidium bromide. As a negative control, an aliquot of each RNA sample was subjected to PCR amplification without the RT step.

In some experiments, an aliquot of total cellular mRNA was reversed-transcribed to cDNA and PCR expanded for quantitative PCR using a Cytopress Detection kit (BioSource International, Camarillo, CA). This technique is a competitive PCR in which a known copy number of an exogenous synthesized DNA, known as the internal calibration standard (ICS). The ICS was constructed to contain PCR primer binding sites identical to the IL-4 cDNA and a unique capture binding site that allows the resulting ICS amplicon to be distinguished from the IL-4 amplicon. The Cytopress kit contains IL-4 primers, one of, which is biotinylated, to be included in the PCR mix. During amplification, biotin-labeled primer is incorporated into both ICS and IL-4 amplicons. After PCR, the amplicons are denatured and hybridized to either ICS or IL-4 sequence-specific capture oligonucleotides. Capture oligonucleotides are prebound to microtiter wells. The captured biotinylated sequences are detected and quantified by the addition of an enzyme-streptavidin conjugate HRP, followed by the addition of the substrate. The signal generated in the reaction is proportional to the amount of amplicon present. Because the ICS is amplified at an efficiency identical to the IL-4 cDNA, it can serve as a standard for IL-4 cDNA quantitation. The number of copies of IL-4 in each PCR reaction is calculated from the ratio of the total OD for the IL-4-specific well to the total OD for the ICS well and the input copy number of the ICS. The following formula is used to calculate the starting copies of IL-4 cDNA in the PCR reaction: (total IL-4/total ICS OD) × 2 × input copy number of ICS = starting copy number of IL-4 cDNA. Factor 2 is used to correct for double-stranded DNA ICS. The copy number is adjusted for any dilution done on cDNA before amplification according the manufacturer’s protocol.

The results are means ± SEM. The data subjected to linear regression were calculated by the least-squares method (y = a + bx) in which a was the y-axis intercept and b the slope of the line (51).

For our experiments we purified (>98%) human peripheral blood basophils from healthy individuals who were seronegative for Abs to HIV-1 and HIV-2. We then cultured basophils with recombinant human IL-3 (10 ng/ml for 16 h), washed the cells, and challenged them with recombinant gp120 from HIV-1 (10 nM) or anti-IgE (1 μg/ml). Fig. 1 shows a representative result of four experiments in which we examined the effects of gp120 and anti-IgE on the levels of extracellular IL-4 protein, specific IL-4 mRNA, and secretion of histamine. These experiments demonstrated that gp120 and anti-IgE increased specific IL-4 mRNA copies. gp120 and anti-IgE stimulated the release of IL-4 parallel to the secretion of histamine from basophils. In contrast, IFN-γ mRNA was not detected in any of the basophil preparations stimulated with gp120, suggesting that gp120-mediated stimulation of FcεRI⁺ cells induced only cytokines of the T_H2 profile (data not shown).

To verify that HIV-1 envelope glycoprotein gp120 has the capacity to induce basophil activation, we compared four recombinant gp120 (gp120_MN, gp120_SF2, gp120_LAV, and gp120_CM) derived from divergent HIV-1 isolates from different viral clades (B and E) of various geographical origins (United States, France, and Thailand). These highly divergent samples of gp120 induced IL-4 release from basophils. This implies that the capacity to induce cytokine release from basophils is a general feature of gp120, which has been maintained throughout the evolution of the virus (Fig. 2). Two unrelated gps, BSA and human serum albumin (1–100 nM), did not activate any of the basophil preparations tested (data not shown).

Fig. 3,A compares the kinetics of histamine, IL-4, and IL-13 release from basophils challenged with gp120. Histamine release was complete within 5 min, IL-4 release was complete in 3–4 h, whereas IL-13 release plateaued at 18 h. Lower concentrations of gp120 consistently exerted a more potent effect on the secretion of IL-13 than on IL-4 and histamine release (Fig. 3 B). Similar results were obtained with the four gp120 derived from different clades (data not shown). These data demonstrate that nanomolar concentrations of gp120 induce the secretion of IL-13, an important cytokine for the polarization of T_H2 cells (22, 23, 24), from basophils.

Two immunophilin-binding drugs, CsA and tacrolimus (FK-506), are potent inhibitors of the IgE-dependent release of proinflammatory mediators from human basophils and mast cells (52, 53, 54). We compared the effects of preincubation of low concentrations of CsA (24–800 nM) and tacrolimus (1–30 nM) on the release of histamine and the secretion of IL-4 from purified basophils activated by gp120. CsA concentration-dependently inhibited the gp120-induced release of histamine and IL-4 from basophils at concentrations as low as 24 nM (Fig. 4,A). The inhibition of IL-4 release ranged from ≈30% at 24 nM to 90–95% at 240 nM, with an IC₅₀ of 31.6 ± 6.4 nM, similar to that calculated from the inhibition of histamine release (46.0 ± 9.3 nM). The effects of tacrolimus on the release of histamine and the secretion of IL-4 from human basophils were similar to those observed with CsA (Fig. 4 B). However, tacrolimus was more potent than CsA with an IC₅₀ for IL-4 release of 2.3 ± 0.2 nM, similar to that calculated from the inhibition of histamine release (2.0 ± 0.4 nM).

Brief exposure to low pH removes most of the IgE bound on FcεRI⁺ cells, thus greatly reducing the activating properties of IgE-mediated stimuli (46). Fig. 5 shows that brief exposure to lactic acid completely blocks the effect exerted by gp120 and by anti-IgE on IL-4 secretion from basophils. In contrast, the response to the mAb cross-linking the α-chain of FcεRI (47) was not affected by this treatment. These data are compatible with the hypothesis that gp120 activates FcεRI⁺ cells through the interaction with IgE bound on basophils.

To evaluate the mechanism whereby gp120 activates basophils from healthy individuals, we incubated gp120 with monoclonal IgM of different V_H families according to a recently described procedure (47). In these experiments, gp120 (10 nM) was preincubated (15 min, 37°C) with increasing concentrations (0.1–10 μg/ml) of three different preparations of monoclonal IgM V_H3⁺ (IgM M3, IgM M11, and IgM LAN) or monoclonal IgM V_H6⁺ (M14). Basophils isolated from HIV-1- and HIV-2-negative subjects were then added, and the incubation was continued for an additional 4 h at 37°C. At the end of this incubation, IL-4 in the supernatants was assayed. Fig. 6 shows that preincubation of three preparations of monoclonal IgM (M3, M11, and LAN), which possess the V_H3 domain, concentration-dependently inhibited the effect of gp120 on IL-4 secretion. In contrast, a monoclonal IgM (M14), which possesses a V_H6 domain, had no effect.

gp120 from different clades also induced mediator release from mast cells isolated and purified from human lung parenchyma, and the releasing activity was inhibited by preincubation with the three V_H3⁺ monoclonal IgM, but not by IgM V_H6⁺ (Fig. 7). Thus, binding to the V_H3 domain inhibits the interaction of gp120 with IgE bound to FcεRI on basophils and lung mast cells.

To map the IgE binding sites on gp120, we tested the inhibitory capacity of a panel of synthetic peptides encompassing the gp120_MN sequence (55). The synthetic peptides of the gp120_MN core motif (peptides 1959, 1960, 1985, 1988, and 1989), concentration-dependently inhibited the effect of gp120_MN on basophils. In contrast, peptides 1922 and 2015, which span the gp120_MN amino- and COOH-terminal region, respectively, failed to induce inhibition. Each of these peptides alone neither induced nor inhibited anti-IgE-induced cytokine or histamine release from human FcεRI⁺ cells (data not shown). Our results are consistent with the hypothesis that the superantigen-binding site(s) on gp120 is formed by protein sequences of at least two regions, which span in a discontinuous fashion the constant and variable domains of the molecule (41).

This study shows that a variety of gp120 from diverse clades induced the synthesis and release of IL-4 and IL-13 from human FcεRI⁺ cells. The activity of gp120 is mediated by interaction with the V_H3 region of IgE present on human basophils and mast cells. This is the first demonstration that gp120 triggers the release of two cytokines critical for T_H2 polarization from human FcεRI⁺ cells.

The relevance of this finding is 3-fold. It suggests that during the early phase of HIV infection, which is associated with high levels of viremia and spreading of the virus (35, 36), basophils exposed to virus-bound or shed gp120 (56) might represent an initial source of IL-4 and IL-13, thereby favoring a shift from a T_H0 toward a T_H2 phenotype. In advanced HIV-1 infection when CD4⁺ T cells are decreased, FcεRI⁺ cells might also represent a significant source of T_H2-like cytokines. Therefore, during both early and advanced HIV-1 infection, basophils and mast cells might be a source of cytokines that contribute to the polarization of CD4⁺ cells toward T_H2 cells. Finally, our findings might be significant also from a quantitative viewpoint. T_H2 cells represent 0.2–2% of CD4⁺ cells (57), whereas basophils represent 1% of peripheral blood leukocytes (25). Viral Ags interact with individual T_H clones; viral superantigen gp120 can produce a rapid and massive activation of basophils via V_H3⁺ IgE. Because the V_H3 family is the largest in the human repertoire (∼50%; see Refs. 42, 43, 44), it is likely that shed or virus-bound gp120 interacts with a high frequency with V_H3⁺ IgE bound to basophils of normal or early-infected individuals. Finally, the levels of IL-4 produced by human lymphocytes are about 10–20% of those generated by immunologically challenged basophils (58).

IL-4 and IL-13 released from human FcεRI⁺ cells might also indirectly (i.e., through chemokine receptors) play a role in the entry of HIV-1 into CD4⁺ cells. The chemokine receptors CXCR4 and CCR5 are major coreceptors for HIV-1 entry into CD4⁺ T cells (38, 39, 40). Up-regulation of CXCR4 by IL-4 (59) facilitates HIV-1 infection of T cells and might even be sufficient to trigger CD4⁺ T cell depletion (60). IL-4 and IL-13 trigger monocytes to produce the T cell chemoattractant protein-1 STCP-1 (or macrophage-derived chemokine), a chemokine selectively active on CCR4 receptor expressed on T_H2 cells (61). These results evoke an amplification loop of polarized T_H2 responses based on induction of chemokine receptors by gp120-stimulated IL-4 and IL-13 from basophils. The latter observation is important because HIV-1 replicates preferentially in T_H2 rather than in T_H1 cells (17). In this context, basophils may play a role since they can rapidly produce IL-4 and IL-13 in a restricted manner without synthesizing T_H1-type cytokines (e.g., IFN-γ).

We found that the FcεRI⁺ cell-activating property is well conserved in gp120 from divergent HIV isolates from clades of different geographic locations. Consequently, conservation of these V_H3 binding sites of gp120 between viral clades could be an important mechanism by which the virus can elude the specific immune surveillance of the host.

Our previous in vitro (47, 52, 53) and in vivo (54) studies demonstrated that immunophilin-binding drugs (i.e., CsA and tacrolimus) exert anti-inflammatory effects by inhibiting the IgE-dependent release of proinflammatory mediators from human FcεRI⁺ cells. In this study, we demonstrate that CsA and tacrolimus concentration-dependently inhibit also the gp120-induced release of cytokines from human basophils. These findings imply that some of the effects exerted by this class of anti-inflammatory/immunosuppressive drugs on HIV-1 replication (62, 63) reflect their actions on IL-4 secretion from FcεRI⁺ cells. T cells with a T_H2-oriented cytokine profile are more susceptible to HIV-1 infection than are T_H1-oriented cells (17). This is due to up-regulation of CXCR4 receptors induced by IL-4 on T cells (59). The inhibitory effect of CsA and tacrolimus on IL-4 secretion from FcεRI⁺ cells might be an additional mechanism whereby this class of compounds exerts antiviral activity (62, 63).

In conclusion, we provide the first evidence that gp120 from different viral clades induces the release of IL-4 and IL-13 from human FcεRI⁺ cells. Because HIV-1 enters the body predominantly through mucosal surfaces and because the early phases of infection are associated with high levels of viremia, mast cells and basophils can be exposed to shed or virus-bound gp120. This suggests that FcεRI⁺ cells might be a novel source of T_H2 cytokines, thus contributing to the dysregulation of the immune system in HIV-1 infection. The latter observation might reconcile the apparently conflicting results of several investigators (12, 13, 16, 17, 18, 19, 20, 21). In fact, it highlights the importance of a specific viral superantigen, gp120, acting on cell types other than lymphocytes in the production of T_H2-like cytokines. This novel observation might be relevant in the design of drugs selectively acting on FcεRI⁺ cells in the treatment of HIV-1-infected subjects.

We thank Dr. J.-P. Bouvet for providing human monoclonal IgM V_H3⁺ and IgM V_H6⁺. We thank the AIDS National Institute of Allergy and Infectious Disease, Research and Reference Reagent Program, Division of AIDS, National Institute of Health for the gift of recombinant gp120s, and Jean Gilder for editing the text. This paper is dedicated to Rita Levi-Montalcini who first suggested an involvement of FcεRI⁺ cells in HIV-1 infection.