HIV-1 envelope glycoprotein gp120 induces, independently of infection, the release of proinflammatory cytokines, including IL-1β from macrophages, that are implicated in the pathogenesis of HIV-associated dementia. However, the signal transduction pathways involved have not been fully defined. Previously, our laboratory reported that soluble gp120 activates multiple protein kinases in primary human monocyte-derived macrophages, including the Src family kinase Lyn, PI3K, and the focal adhesion-related proline-rich tyrosine kinase Pyk2. In this study we showed that gp120 induces IL-1β release from macrophages in a time- and concentration-dependent manner through binding to the chemokine receptor CCR5 and coupling to Giα protein. Using pharmacological inhibitors and small interfering RNA gene knockdown, we demonstrated that concomitant activation of Lyn, Pyk2, and class IA PI3K are required for gp120-induced IL-1β production. By coimmunoprecipitation and immunofluorescence confocal microscopy, we showed that CCR5 activation by gp120 triggered the assembly of a signaling complex involving endogenous Lyn, PI3K, and Pyk2 and is associated with PI3K and Pyk2 translocation from the cytoplasm to the membrane where they colocalized with Lyn. Finally, we demonstrated that virion-associated gp120 induced similar response, as structurally intact whole virions also triggered IL-1β release and re-localization of PI3K and Pyk2. This study identifies a novel signaling mechanism for HIV-1-induced IL-1β production by primary human macrophages that may be involved in the neuropathogenesis of HIV-associated dementia.
Human immunodeficiency virus-1-associated dementia (HAD)3 is a common neurological complication associated with HIV and is estimated to develop in ∼20% of infected patients (1). Unlike many viral encephalopathies, neurons are not productively infected by HIV. HAD is characterized by the infiltration of blood-derived macrophages at perivascular sites in the CNS, the formation of multinucleated giant cells, and neuronal death, as well as the widespread activation of brain macrophages (2). Although infection of macrophages/microglia in the CNS is a prerequisite for the development of HAD, macrophage activation appears to play a critical role as the number of activated macrophages in the CNS is the best correlate of HAD (3, 4, 5), and it is believed that the release of proinflammatory cytokines from activated macrophages/microglia is a principal mechanism of neuronal injury (6). Both infected and uninfected (bystander) macrophages are activated in HAD (7), but the mechanisms responsible are incompletely defined.
IL-1β and TNF-α are highly expressed in the CNS of patients with HAD, correlate with neuronal injury, and have been implicated as important proinflammatory cytokines involved in HAD pathogenesis (8, 9). IL-1β is up-regulated in vivo in both cerebral spinal fluid and brain of HAD patients (10, 11, 12), and macrophage/microglia is the primary cell type in CNS responsible for its production (7, 13). Exposure of macrophages to whole HIV-1 virions or envelope (Env) glycoprotein gp120 has been shown in many in vitro systems to induce IL-1β release independently of productive infection (14, 15). In animal models, intracerebral injection of gp120 can also induce neuronal apoptosis (16), which is mediated through IL-1β release from macrophages/microglia because it can be prevented by pretreatment with the blocker for the IL-1β receptor or the IL-1β-converting enzyme (17). Despite the important role of IL-1β in neuronal injury and HAD pathogenesis, how gp120 induces IL-1β release from macrophages is not well understood.
HIV-1 entry is initiated by gp120 binding first to cellular CD4 followed by structural changes that enable interaction with the chemokine receptors CCR5 (R5 strains), CXCR4 (X4 strains), or both (R5X4 strains). Most HIV-1 isolates from the CNS or those associated with HAD use CCR5 (R5 strains). Our laboratory has previously reported that HIV-1 R5 gp120 can activate CCR5-mediated signaling pathways in monocyte-derived macrophages (MDMs) including the Src family kinase (SFK) Lyn, PI3K, the focal adhesion-related proline-rich tyrosine kinase 2 (Pyk2), and several ionic currents and that Lyn and PI3K are involved in gp120-induced TNF-α production by macrophages (18, 19, 20, 21). However, the signaling pathways regulating IL-1β release by macrophages in response to gp120 and the mechanisms of interaction among signaling molecules activated by gp120 have not been fully defined. In this study, we set out to determine the role of Lyn, PI3K, and Pyk2 in gp120-induced IL-1β release by macrophages and find out whether these kinases physically associate to form a signaling complex in response to activation of CCR5 by HIV-1 gp120. We also investigated the effect of virion-associated gp120, because macrophages in vivo may be exposed to both monomeric (shed) and virion-associated Env glycoproteins. We found that HIV-1 gp120 induces IL-1β release by primary macrophages through binding to CCR5 and the coupling of Giα protein and requires concomitant activation of Lyn, PI3K, and Pyk2 and their subcellular redistribution and physical association to form a multimeric signaling complex. We further demonstrated that, like monomeric gp120, virion-associated gp120 also triggers IL-1β release as well as the translocation of PI3K and Pyk2 from the cytoplasm to the plasma membrane where they colocalize with Lyn in a CCR5-dependent manner.
Materials and Methods
Reagents and cells
Recombinant gp120 (CM235) was obtained from the National Institutes of Health AIDS Reagent Repository (Bethesda, MD) and tested negative for endotoxin contamination (limit of detection = 1.5 pg/ml). CCL4 (MIP-1β) was from PeproTech. The CCR5 antagonist M657 was kindly provided by M. Miller of Merck (22). Pertussis toxin (PTX) was purchased from Sigma-Aldrich. Isoform-specific PI3K inhibitors were from Echelon. All other kinase inhibitors, their inactive analogues, and Pansorbin beads were from EMD Chemicals. The Lyn-specific peptide inhibitor KRX-123.302 and control peptide KRX-107.110 were synthesized by Bachem according to previously published peptide sequences (23).
Affinity-purified anti-gp120 polyclonal IgG derived from pooled AIDS patient sera and anti-gp120 rabbit polyclonal Ab were provided by the late K. Steimer of Chiron. Rabbit polyclonal Abs to Lyn, Hck, PI3Kp85, PI3Kp101, and anti-phosphotyrosine Abs were from Millipore. Mouse monoclonal Pyk2 Ab was from BD Biosciences. Rabbit polyclonal β-actin Ab and HRP-conjugated anti-rabbit and anti-mouse IgG were from Cell Signaling Technology. Goat polyclonal PI3Kp85 Ab and control mouse, rabbit, and goat IgG were from Santa Cruz Biotechnologies. Alexa Fluor 555-, 488-, and 633-conjugated IgG were purchased from Molecular Probes.
Monocytes were isolated by elutriation (24) or by selective adherence (25) from healthy donors who were screened for the CCR5Δ32 mutation by PCR (26). Only donors homozygous for the wild-type allele were used unless otherwise specified. Monocytes were cultured in 100-mm petri dishes for 7 days to differentiate into macrophages under conditions described previously (18), harvested by gentle scraping, resuspended, and replated for IL-1β ELISA, kinase activation, coimmunoprecipitation, or confocal immunofluorescence microscopy studies.
Quantitation of IL-β release
MDMs plated in 24-well plates (5 × 105 cells/well) were incubated in serum-free medium overnight, exposed to 20 nM gp120 for 16 h (unless stated otherwise), and IL-1β levels in cell culture medium were determined by ELISA (R&D Systems) according to the manufacturer’s instructions. The limit of detection for IL-1β ELISA is <1 pg/ml. For blocking studies, antagonists, inhibitors, or control vehicle (0.1% DMSO) were added 15 min to 2 h before and maintained throughout the period of stimulation. To ensure that the inhibitory effects of the pharmacological agents on IL-1β production were not due to cytotoxicity, whenever inhibitors were used a cell viability assay was performed using the CellTiter 96 AQueous One Solution cell proliferation assay (Promega) in parallel. MDMs (105) in 100 μl of serum-free medium were added to the 96-well plates and incubated with the inhibitors for the same duration as that used for IL-1β stimulation. Cell viability was measured spectrophotometrically according to the manufacturer’s instructions.
Small interfering RNA (siRNA) gene knockdown
Predesigned siRNA specific for Lyn, PI3Kp85, Pyk2, and nontargeting control siRNA were purchased from Dharmacon. Transfection of siRNAs in primary MDMs was performed using the Human Macrophage Nucleofector Kit (Amaxa) according to the manufacturer’s instructions. Seven-day-old MDMs were resuspended (5 × 106 cells/ml) in 100 μl of nucleofection solution with 1–3 μg of siRNA and electroporated using program Y-10. After transfection, MDMs were cultured for 2 days (chosen for maximal protein knockdown) before harvesting to assay in parallel both functional response by ELISA and protein expression by immunoblot analysis.
Immunoprecipitation and immunoblot analysis
MDMs plated in 12-well plates (106 cells/well) were incubated in serum-free medium overnight, exposed to stimuli for indicated durations, and lysed in immunoprecipitation buffer supplemented with protease inhibitor mixture as previously described (18). Protein concentrations of whole cell lysates were determined by BCA protein assay (Pierce). For immunoprecipitation, equal amounts (400 μg) of cell lysates were incubated at 4°C with 1 μg of Lyn or Pyk2 Ab for 2 h followed by 50 μl of Pansorbin beads for 2 h. Immune complexes were washed twice with lysis buffer, eluted by boiling in sample buffer for 5 min, and subjected to immunoblot analysis.
Immunoblot analysis was conducted on immunoprecipitates generated as described above or directly on cell lysates containing 20 μg of protein. Samples were denatured by boiling in Laemmli buffer for 5 min, resolved by electrophoresis on 8% SDS-polyacrylamide gel, and transferred to nitrocellulose membranes. The conditions for blocking, washing, and Ab dilution were based on instructions provided by the manufacturers of the Abs. Immunoreactive proteins were visualized by ECL detection reagents (GE Healthcare) on autoradiographic films. Multiple exposures of the films were used to ensure that quantitation was conducted within the linear range of the images. The films were digitally scanned and the relative intensities of the protein bands were quantitated by NIH Image J software. In some experiments, membranes were stripped with Restore Western Blot stripping buffer and reprobed with a second Ab.
Immunofluorescence laser-scanning confocal microscopy
MDMs were plated onto glass coverslips and cultured overnight in serum-free medium, exposed to stimuli, then fixed in 4% paraformaldehyde for 20 min and permeabilized with 0.2% Triton X-100 for 10 min. Fixed cells were blocked with donkey serum for 20 min before incubation for 1 h with the following primary Abs (all 1/100 dilution): rabbit polyclonal anti-Lyn, goat polyclonal anti-PI3Kp85, and mouse monoclonal anti-Pyk2. Controls consisted of replacing the primary Abs with nonspecific IgG from the corresponding species or omitting primary Abs. MDMs were then incubated for 30 min with the corresponding secondary Abs (all 1/100 dilution) conjugated to the fluorophores. Between all incubation steps, cells were washed three times for 5 min with PBS. All incubations were conducted at room temperature. Coverslips were mounted onto slides with VectorShield mounting medium (Vector Laboratories,). Confocal images were captured with a Zeiss LSM 510 Meta laser-scanning confocal device attached to an Axioplan 2 microscope using a ×63 Plan-Apochromat oil objective. To avoid the bleed-through effects, each fluorophore was scanned independently using the multitracking function of the LSM 510 device. Images were electronically merged using the LSM Image Browser software and saved as TIFF files.
Virions were generated using an Env trans-complementation method. HEK293T cells were cotransfected with the Env-deleted pNL-Luc backbone (27) and a plasmid expressing the Env gene of the HIV-1 R5 strain BaL. To make Env-negative (bald) virions as a control, the pNL-Luc backbone was cotransfected with an empty plasmid vector. Cell culture supernatants were harvested 72 h posttransfection and concentrated by ultracentrifugation through 20% sucrose (150,000 × g for 2 h). The pelleted virions were resuspended and quantitated using a p24gag ELISA (PerkinElmer). Incorporation of Env into pseudotype but not control virions was confirmed by Western blotting. Equal amounts of each virion preparation based on p24gag Ag content were subjected to SDS-PAGE alongside known amounts of recombinant gp120 as standards and subjected to immunoblot analysis with rabbit anti-gp120 and mouse anti-p24 Abs.
Quantitative data are presented as mean ± S.E. of three independent experiments using cells from different blood donors. Multiple group comparisons were analyzed by one-way ANOVA followed by the Bonferroni test for comparison of means. For all tests, p < 0.05 was considered statistically significant. The immunoblot and immunofluorescent staining images shown are representative of experiments done using cells from at least three different donors.
HIV-1 Env glycoprotein gp120 induces IL-1β release by primary human macrophages
To characterize the effect of HIV-1 gp120 stimulation of IL-1β in primary human macrophages, we first performed time course and concentration-response experiments. As shown in Fig. 1,A, MDM stimulation with gp120 from an R5 HIV-1 strain triggered an increase in IL-1β release in a time-dependent manner with the maximal response at 16 h, a 3.5-fold increase over the basal level. IL-1β production by MDMs in response to gp120 was also concentration dependent (Fig. 1,B) with maximal stimulation observed at 20 nM gp120 and no further increase with higher concentrations. As shown in Fig. 1 C, IL-1β release was completely abrogated if gp120 was heat inactivated (ΔHI; 100°C for 30 min), confirming that it was not due to contamination with a heat-stable component such as endotoxin. IL-1β release was also eliminated if gp120 was preincubated with affinity-purified anti-gp120 polyclonal IgG derived from pooled AIDS patient sera, whereas incubation with control IgG exhibited no blocking effect. These results confirmed that IL-1β release by primary human macrophages is elicited by HIV-1 gp120, which is consistent with long-standing data from previous studies (14, 15).
HIV-1 gp120 induces IL-1β release by MDMs through binding to CCR5 and coupling of Giα protein
We used both pharmacological and genetic approaches to define the role of the chemokine receptor CCR5 in gp120-induced IL-1β release in MDMs. As shown in Fig. 1,D, pretreatment of macrophages with the CCR5 antagonist M657 completely abolished the IL-1β response to gp120. In addition, CCR5-deficient MDMs from donors homozygous for the Δ32 deletion allele failed to release IL-1β in response to gp120, although they responded normally to LPS stimulation (Fig. 1 E). Macrophages lacking CCR5 also failed to produce IL-1β in response to CCL4 (MIP-1β), the most specific chemokine agonist of CCR5, whereas MDMs from wild-type donors responded to CCL4 with IL-1β production at levels similar to that seen with R5 gp120.
Although Giα is the primary G protein coupled to CCR5 (28), other G proteins have been implicated in CCR5 signaling (29, 30). Therefore, we determined whether CCR5-mediated IL-1β release by MDMs in response to gp120 requires coupling of Giα. Pretreatment with the Gi/o inhibitor PTX markedly impeded the gp120-induced IL-1β production, whereas the Gs-specific inhibitor NF449 (31) showed no effect (Fig. 1 D). Together, these results indicate that gp120 induces IL-1β release by MDMs through binding to CCR5 and the coupling of Giα protein.
gp120-induced IL-1β release by MDMs requires activation of Src family kinase Lyn
We previously reported that gp120-triggered, CCR5-mediated TNF-α production in MDMs was mediated through SFK member Lyn, based on pharmacological inhibition (18). To determine whether Lyn is involved in gp120-induced IL-1β release, MDMs were pretreated with the broad spectrum SFK inhibitor PP2. As shown in Fig. 2 A, PP2 completely abrogated gp120-induced IL-1β production whereas its inactive analog PP3 showed no blocking effect. For greater specificity, we also tested a pseudosubstrate peptide inhibitor specific to Lyn, KRX-123.302 (23), as well as a negative control peptide, KRX-107.110. IL-1β release was also markedly attenuated by KRX-123.302 while no inhibition was observed with the negative control peptide KRX-107.110. Parallel experiments to evaluate the viability of MDMs in the presence of pharmacological inhibitors ruled out inhibitor-induced cytotoxicity as a cause of the decreased IL-1β response (data not shown).
To further substantiate the role of Lyn in gp120-induced IL-1β release, we then used siRNA-mediated gene silencing. As shown in Fig. 2,B, siRNA knockdown suppressed Lyn protein expression in primary MDMs by an average of 91% based on the quantitation of immunoblots, whereas other proteins such as Pyk2, PI3K, the SFK Hck, and β-actin were unaffected. Down-regulation of Lyn significantly impaired the ability of MDMs to produce IL-1β in response to gp120 (Fig. 2 B). Suppression of cytokine production was not the result of decreased cell viability, as the viability of both control and Lyn siRNA-transfected MDMs were equivalent (data not shown). We also confirmed that Lyn is activated by gp120 based on tyrosine phosphorylation (data not shown), which is consistent with results previously observed by kinase assay (18).
Activation of PI3K is necessary for gp120-triggered IL-1β induction in MDMs
PI3K has been implicated as a mediator of CCR5-elicited signals, leading to both survival and TNF-α production in macrophages (19, 32). Therefore, we tested whether PI3K is important for gp120-induced IL-1β production. As shown in Fig. 3 A, LY294002 and wortmannin, two potent PI3K inhibitors with different mechanisms of action, significantly attenuated the macrophage IL-1β response to gp120. We verified that gp120 activated PI3K in MDMs based on phosphorylation of Akt, a major downstream target of PI3K (data not shown), which is consistent with our prior published results (19).
Classically, class IA PI3K isoforms (p85-p110α,β,δ) are activated by receptor tyrosine kinases, whereas class IB PI3Ks (p101-p110γ) are activated by G protein-coupled receptors (GPCRs), including chemokine receptors (33, 34). However, some reports suggested that class IA PI3K might also be activated by GPCRs and may play a role in chemokine receptor-mediated IL-1β induction in primary human monocytes (35, 36, 37, 38). To define the role of specific PI3K isoforms on gp120-stimulated IL-1β release, we used the class IA PI3K-specific inhibitor PI-103 (39) and the class IB PI3K specific inhibitor AS605240 (40, 41). As shown in Fig. 3 B, inhibition of class IA PI3K markedly attenuated the IL-1β response to gp120 while the IB inhibitor had no effect.
We then used siRNA directed against the p85 subunit of PI3K, which achieved ∼85% knockdown of PI3K protein expression as shown in Fig. 3,C, while the expression of class IB PI3Kp101 and other proteins including Pyk2, Lyn, and Hck were unaffected. IL-1β induction by gp120 was markedly attenuated in PI3Kp85-depleted MDMs relative to MDMs transfected with control siRNA (Fig. 3 C), confirming a role for PI3K in gp120-stimulated IL-1β release. Together, these results suggest that class IA PI3K, rather than the more typically GPCR-associated class IB isoforms, is the primary isoform responsible for IL-1β production in response to gp120.
IL-1β induction by gp120 involves focal adhesion-related kinase Pyk2
Previous data from our laboratory and others have reported that CCR5 activates the tyrosine kinase Pyk2 (20, 42, 43), but the role of Pyk2 in gp120-triggered IL-1β release in primary macrophages has not been examined. Because specific inhibitors for Pyk2 are currently not available, we used two pharmacological agents that target upstream activators of Pyk2 through different mechanisms. Dantrolene inhibits Pyk2 by blocking ryanodine receptor-mediated intracellular calcium release, which is required for Pyk2 activation (44, 45, 46). Another agent often used as a Pyk2 inhibitor is AG17, which blocks calcium release-activated calcium (CRAC) channel-mediated intracellular calcium (47, 48, 49). As shown in Fig. 4 A, both dantrolene and AG17 significantly inhibited IL-1β response whereas the inactive analog AG43 exhibited no effect.
Owing to the calcium dependence of gp120-induced Pyk2 activation suggested by these results and to the ability of gp120 to open ion channels to increase intracellular calcium levels in MDMs demonstrated in previous studies (20, 21), we then asked whether gp120-induced IL-1β release involves calcium/calmodulin-dependent protein kinase II (CaMKII), which has been implicated in ATP-induced IL-1β release by MDMs (50) and Pyk2 activation in some cell models (51, 52). We also examined the role of glycogen synthase kinase 3 (GSK3), a downstream target of PI3K that modulates cytokine production in several model systems (53, 54, 55). Neither the CaMKII inhibitor KN62 nor the GSK3 inhibitor BIO affected gp120-induced IL-1β induction in MDMs (Fig. 4 B). In addition to excluding a role for CaMKII and GSK3 in gp120-induced IL-1β production, these results also serve to illustrate, in part, the specific requirement for Lyn, PI3K, and Pyk2 activation in IL-1β induction by gp120.
Because the available pharmacological inhibitors of Pyk2 are not completely specific, we next confirmed the role of Pyk2 in gp120-induced IL-1β production by transfecting primary MDMs with Pyk2-specific siRNA. This knockdown suppressed Pyk2 protein expression by ∼72% and was gene specific as the expression of other proteins were unaffected (Fig. 4,C). Pyk2 knockdown markedly attenuated gp120-induced IL-1β induction as shown in Fig. 4 C. Immunoprecipitation/immunoblot analysis confirmed tyrosine phosphorylation of Pyk2 in a time-dependent manner concordant with activation in response to gp120 exposure to MDMs (data not shown) as reported previously (20).
Lyn, PI3K, and Pyk2 physically interact in response to HIV-1 gp120 stimulation in primary MDMs
Our data demonstrated that Lyn, PI3K, and Pyk2 are each important for gp120-induced IL-1β production in MDMs. All three kinases contain structural motifs that facilitate protein-protein interactions including Src homology (SH)2 and SH3 domains of Lyn and PI3K, as well as the proline-rich motifs of PI3K and Pyk2. Therefore, we determined whether gp120 triggered a physical association among the endogenous Lyn, PI3K, and Pyk2 in macrophages (Fig. 5).
When endogenous Lyn protein was immunoprecipitated with Lyn Ab in unstimulated MDMs (Fig. 5,A), low levels of PI3K and Pyk2 could be detected in the immune complex (left lane), suggesting some constitutive association among these proteins. Exposure of MDMs to gp120 resulted in a marked increase in coprecipitation (Fig. 5 A, second lane from left), indicating that gp120 stimulation up-regulated the formation of this complex.
We next used both PI3K and Pyk2 Abs for immunoprecipitation to perform the reciprocal experiments (Fig. 5, B and C). Immunoprecipitation with an anti-PI3K Ab demonstrated gp120-enhanced Pyk2 and Lyn coprecipitation (Fig. 5,B), and similar results were observed with Pyk2 coimmunoprecipitation studies (Fig. 5,C). In contrast, immunoprecipitation with Ab specific to Hck, another SFK member that is abundantly expressed in MDMs, showed no constitutive or stimulation-induced association with PI3K or Pyk2 protein (Fig. 5,D), nor did probing the Lyn, PI3K or Pyk2 immunoprecipitation blots with anti-Hck Ab yield any signal (bottom panels of Figs. 5, A–C). Of note, the rapidity of complex formation (10 min) indicates an increase in the complex-associated fraction for each kinase following gp120 exposure rather than up-regulated total protein levels, a mechanism that is concordant with our prior results showing an increase in activated but not total Lyn and Pyk2 following gp120 exposure (18, 20). Thus, Lyn, PI3K, and Pyk2 specifically associate to form a signaling complex in response to gp120 stimulation.
HIV-1 gp120 induces colocalization of Lyn, PI3K, and Pyk2 in primary macrophages
Because a physical association among Lyn, PI3K, and Pyk2 was suggested by our coimmunoprecipitation results, we next performed triple-labeling immunocytochemistry to determine the subcellular distribution of these kinases and the putative signaling complex (Fig. 6). In unstimulated MDMs (Fig. 6, A–D, top panels), Lyn was primarily localized at the plasma membrane (Fig. 6,A) and PI3Kp85 was distributed both at the plasma membrane and in the cytoplasm (Fig. 6,B), whereas Pyk2 was mainly found in the cytoplasm (Fig. 6,C). Colocalization of the three kinases, which should appear as white, was not observed in the merged image in unstimulated MDMs (Fig. 6 D). Control experiments in which the primary Abs were omitted or substituted with the isotype control IgG of the appropriate species showed no fluorescent staining (data not shown).
When MDMs were stimulated with gp120, most of the PI3K and Pyk2 was relocalized at the plasma membrane (Fig. 6, F and G), while Lyn remained in its previous membrane distribution (Fig. 6,E). This gp120-induced PI3K and Pyk2 translocation resulted in colocalization of the three kinases at the membrane as highlighted by the white arrowheads seen in the merged image (Fig. 6,H). Relocalization was blocked if MDMs were incubated with the CCR5 antagonist M657 before stimulation with gp120 (Fig. 6, I–L), confirming that gp120-induced translocation of these kinases and complex formation were mediated through CCR5.
Virion-associated gp120 also triggers redistribution of PI3K and Pyk2, colocalization with Lyn, and IL-1β release in primary macrophages
These results demonstrated that soluble monomeric gp120 triggers colocalization of Lyn, PI3K, and Pyk2 in MDMs. In addition, to shed monomeric gp120, macrophages in vivo may also be exposed to gp120 on the surface of HIV-1 virions where it exists as a trimer (56). Therefore, to extend our findings to this physiologically relevant stimulus, we then asked whether virion-associated gp120 would also alter PI3K and Pyk2 subcellular distribution and colocalization with Lyn. Pseudotype virions were generated either carrying Env from the HIV-1 R5 strain BaL or lacking Env (bald virus) to serve as a control. MDMs were exposed to equal amounts of Env-containing and bald virus based on p24gag Ag content and subjected to triple-labeling immunofluorescent staining and confocal microscopy (Fig. 7,A). Compared with unstimulated macrophages (Fig. 7,A, micrographs a–d), MDMs exposed to gp120-containing virions showed a marked increase in PI3K and Pyk2 distribution at the membrane, resulting in colocalization of Lyn, PI3K, and Pyk2 as shown in the merged image (Fig. 7,A, micrographs e–h). This pattern was identical with that seen following MDM exposure to monomeric gp120 (Fig. 6). In contrast, virion particles lacking Env did not induce colocalization (Fig. 7,A, micrographs i–l), confirming gp120 dependence of the stimulus. Of note, bald virions also serve as important controls for potential virion-incorporated host membrane proteins (57). Treatment of macrophages with the CCR5 antagonist M657 completely blocked virion-induced colocalization (Fig. 7 A, micrographs m–p), confirming further that it is mediated through CCR5.
We also demonstrated that virion-associated gp120 induces IL-1β release in macrophages (Fig. 7 B). MDM exposure to gp120-containing virions resulted in IL-1β production at levels comparable to that observed with soluble gp120. This IL-1β induction was both gp120- and CCR5-dependent, as bald virions failed to trigger cytokine release and IL-1β release in response to Env-containing virions was abolished by the pretreatment of macrophages with the CCR5 antagonist M657.
IL-1β is one of the important inflammatory and neurotoxic products secreted by immune-activated macrophages/microglia that has been implicated in the pathogenesis of HAD (6, 58). In this study we demonstrated that HIV-1 gp120-induced IL-1β release by macrophages is mediated through CCR5 and the coupling of Giα protein; that it requires concomitant activation of Lyn, PI3K, and Pyk2; that these kinases act coordinately through the formation of a multi-kinase signaling complex; and that both monomeric and virion-associated gp120 activate these pathways (Fig. 8). IL-1β release by HIV-1 gp120- or virion-stimulated macrophages is a long-standing observation believed to be important in pathogenesis (9, 14, 15), and this is the first study to define the signal transduction mechanism responsible.
In the brain, IL-1β is expressed at a low level under physiological conditions and is induced in response to both acute and chronic inflammation (59). This is in agreement with immunohistochemical studies of autopsy brain tissues from AIDS patients showing intense IL-1β staining in macrophages/microglia only from those with pathological evidence of HIV encephalitis, but not in those without neurological complications or in healthy individuals (10, 11). Similarly, cerebral spinal fluid IL-1β levels are also increased in HIV-infected individuals with clinical HAD but not in those without HAD (12). Results from our primary human macrophage cell model are consistent with these observations showing the release of low level IL-1β in unstimulated MDMs, which is increased by stimulation by gp120 and gp120-coating virions. Hence, primary MDMs provide a physiologically relevant primary human cell model to examine the signaling pathways mediating gp120-induced IL-1β release.
Src family kinase members and PI3K have each individually been implicated in modulating IL-1β production in both primary monocyte/macrophage cells and monocytoid cell lines in response to LPS, which acts through TLR4 (60, 61, 62, 63). Our results extended these data to chemokine receptor-elicited IL-1β response and further identify Pyk2 as a component of the pathway.
Multiple kinases often organize to form a signaling complex to modulate both efficiency and specificity of signaling response. In this study we identify a gp120/CCR5 interaction that promotes physical association between Lyn, PI3K, and Pyk2 to assemble a new signaling complex. Lyn, PI3K, and Pyk2 all possess structural motifs known to facilitate protein-protein interactions. Lyn and PI3K each contain both SH2 domains that may interact with phosphorylated tyrosine residues and SH3 domains that can bind to proline-rich motifs (64, 65), whereas PI3K and Pyk2 possess proline-rich motifs that could interact with SH3 domains (66, 67). Furthermore, all three kinases contain tyrosine residues that undergo phosphorylation upon activation (68, 69, 70). Hence, it is possible that these three kinases could directly interact with each other to form a complex in response to CCR5 activation in macrophages. Alternatively, Lyn, PI3K, and Pyk2 can be associated indirectly through binding to a common docking protein. We recently found that arrestin scaffolding proteins are involved in CCL4-induced Lyn/PI3K/Pyk2 complex formation (R. Cheung and R. G. Collman, unpublished observations) and thus it is likely that gp120-triggered complex assembly is also arrestin dependent, although a role for the direct interaction between these three kinases cannot be excluded.
How activation of PI3K, Lyn and Pyk2 regulate IL-1β release in MDMs is not clear. One likely mechanism is that activation of these kinases may lead to phosphorylation of transcription factors that interact with the IL-1β promoter region to induce transcriptional activation. NF-κB and AP-1 are two main cis-acting elements identified in the IL-1β promoter region (71, 72), and both Lyn and PI3K have been shown to activate NF-κB whereas Pyk2 is involved in AP-1 activation (73, 74, 75). Defining the downstream pathways by which each of these kinases regulate IL-1β in macrophages requires further investigation.
In summary, we have identified specific signaling pathways and evidence for interactions of multiple kinases that are activated by the HIV-1 Env glycoprotein both in soluble and native virion-associated forms and lead to the release of inflammatory and neurotoxic IL-1β by primary human macrophages. These findings provide a mechanism that may be involved in the development of the HIV-elicited bystander macrophage activation and neuronal injury that contribute to the pathogenesis of HAD.
We thank Dr. M. Malik for advice and discussions, M. Kiezle and J. Issacman-Beck for technical assistance, and the Virus/Molecular and Immunology Cores of the Penn Center for AIDS Research for technical support.
The authors have no financial conflict of interest.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by National Institutes of Health Grants MH061139, NS027405, AI 35502 (to R.G.C.), and CA108552 (to A.P.). R.C. is the recipient of a Canadian Institutes of Health Research postdoctoral fellowship award.
Abbreviations used in this paper: HAD, HIV-associated dementia; CaMKII, calcium/calmodulin-dependent protein kinase II; Env, envelope; GPCR, G protein-coupled receptor; GSK3, glycogen synthase kinase 3; MDM, monocyte-derived macrophage; PTX, pertussis toxin; Pyk2, proline-rich tyrosine kinase 2; SFK, Src family kinase; SH, Src homology; siRNA, small interfering RNA.