HIV-1 proteins, including Tat, gp120, and Nef, activate macrophages (MΦ), which is consistent with the fact that HIV-1 infection is characterized by sustained immune activation. Meanwhile, MΦ are functionally classified into two types: proinflammatory M1-MΦ and anti-inflammatory M2-MΦ. We show that HIV-1 proteins, particularly Nef, preferentially activate M2-MΦ. Extracellular Tat, gp120, and Nef activated MAPK and NF-κB pathways in human peripheral blood monocyte-derived MΦ. However, the activation was marked in M-CSF–derived M2-MΦ but not GM-CSF–derived M1-MΦ. Nef was the most potent activator, and its signaling activation was comparable to that by TNF-α. Indeed, Nef was internalized more rapidly by M2-MΦ than by M1-MΦ. The myristoylation and proline-rich motif of Nef were responsible for the observed signaling activation. Consistent with the activation of MAPK/NF-κB pathways, Nef stimulated the production of a number of proinflammatory cytokines/chemokines by M2-MΦ. However, Nef reduced the expression of CD163 and phagocytosis, the characteristic markers of M2-MΦ, indicating that Nef drives an M2-like to M1-like phenotypic shift. Because the differentiation of most tissue MΦ depends on M-CSF and its receptor, which is the essential axis for the anti-inflammatory M2-MΦ phenotype, the current study reveals an efficient mechanism by which HIV-1 proteins, such as Nef, induce the proinflammatory MΦ.
Human immunodeficiency virus-1 infection is characterized by sustained activation of the immune system, which is a significant predictor of HIV-1 disease progression (1, 2). However, the underlying triggers of this immune activation remain unclear. Macrophages (MΦ), one of the major targets of HIV-1, are thought to contribute to the process, because HIV-1–infected MΦ display an activated phenotype and altered cytokine/chemokine production profiles (3–5). Although the activation of MΦ is critical for the induction of an effective immune response, their inappropriate or sustained polarization can lead to immune dysfunction (6). MΦ are functionally classified into two types: M1 (classic) MΦ, which produce proinflammatory cytokines and contribute to tissue destruction and resistance to pathogens, and M2 (alternative) MΦ, which produce anti-inflammatory cytokines and promote tissue repair and remodeling, as well as tumor progression (7, 8). Indeed, in vitro studies showed that HIV-1 infection drives MΦ toward an M1-like phenotype (4, 5).
Recently, we demonstrated that, when expressed in human myeloid leukemia cells, Nef, a multifunctional pathogenetic protein of HIV-1, inhibits the growth response of the cells to a MΦ-specific cytokine, M-CSF, but not to another cytokine, GM-CSF (9, 10). This was due to the inhibition of the cell surface expression of M-CSFR, but not GM-CSFR, by Nef (11, 12). These results imply that Nef accelerates M1-like MΦ polarization, because it is well established that GM-CSF and M-CSF induce the production of M1-MΦ and M2-MΦ, respectively (13–16).
Of interest, it was shown that even uninfected MΦ can be activated by soluble HIV-1 proteins (17). For instance, the envelope glycoprotein gp120 was found to activate multiple signaling molecules, including MAPK, in MΦ (18–20). Likewise, the trans-activating protein Tat was found to activate p38 and JNK MAPK (21–23). In addition, the pathogenetic protein Nef was found to activate the MAPK, NF-κB, and Stat pathways (24–26). It was also reported that the treatment of human monocyte-derived MΦ with soluble Nef proteins induced the expression of a number of proinflammatory cytokines and chemokines, including IL-1β, IL-6, TNF-α, MIP-1α, and MIP-1β (27). These viral proteins are detected in the sera of HIV-1–infected patients (17, 28, 29) and may be released by infected/apoptotic cells. Soluble exogenous Tat and Nef proteins have been shown to enter MΦ (17, 27, 28), although the mechanisms by which they enter MΦ are not fully understood. More importantly, it remains to be determined whether these soluble HIV-1 proteins differentially activate M1-MΦ and M2-MΦ. This study provides evidence that these soluble HIV-1 proteins, particularly Nef, preferentially activate and drive M-CSF–derived anti-inflammatory M2-MΦ toward MΦ with an M1-like phenotype. It was shown that the differentiation of most tissue MΦ depends on M-CSF and its receptor Fms (30, 31), which is the essential axis for the anti-inflammatory M2-MΦ phenotype (13–16). Therefore, the current study appears to identify a mechanism by which HIV-1 proteins, such as Nef, efficiently induce the production of proinflammatory MΦ.
Materials and Methods
Recombinant cytokines and HIV-1 proteins
Recombinant human (rh)M-CSF with a molecular mass of 85 kDa (32) was a gift from Morinaga Milk Industry (Kanagawa, Japan). rhGM-CSF and rhTNF-α were purchased from PeproTech. The 86-amino acid HIV-1 Tat protein (clade B) produced in Escherichia coli was obtained from Dr. John Brady through the National Institutes of Health AIDS Research and Reference Reagent Program (Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD). The gp120 protein (LAV strain) produced in insect cells was purchased from Protein Sciences (Meriden, CT). The C-terminal His-tagged 210-amino acid HIV-1 Nef proteins (SF2 strain) produced in E. coli were purchased from Jena Bioscience (Jena, Germany). The preparation was shown to be free of endotoxin (<0.03 U/ml), as determined by a Limulus assay (33, 34). In selected experiments (Figs. 4B, 6B), Nef mutants (G2A, AxxA, and LL/AA; Jena Bioscience) were also used (see Fig. 4A for details). Polymyxin B sulfate was purchased from Wako (Osaka, Japan).
Preparation of primary human monocyte-derived MΦ
Heparinized venous blood was collected from healthy donors after informed consent had been obtained in accordance with the Declaration of Helsinki. The approval for this study was obtained from the Kumamoto University Medical Ethical Committee. MΦ were prepared essentially as described previously (10, 35). Briefly, mononuclear cells obtained using Pancoll reagent (PAN Biotech, Aidenbach, Germany) were suspended in RPMI 1640 medium-1% FCS at 1 × 106 cells/ml and seeded in 24-well plates. Monocytes were enriched by adherence to plates for 1 h at 37°C, and nonadherent cells were removed by extensive washing with PBS. Then, the adherent monocytes were differentiated into MΦ by culturing them with RPMI 1640-10% FCS containing 100 ng/ml rhM-CSF or 10 ng/ml rhGM-CSF (36). After 3 d, the cultures were replaced with fresh complete media after extensive washing with PBS to remove nonadherent cells and incubated for another 2 d. The purity of the day-5 MΦ prepared by this method was routinely >95%, according to the expression of CD14 (data not shown). The differentiated MΦ were stimulated and subjected to subsequent analyses. Recombinant Nef was used at 100 ng/ml, unless otherwise stated.
In a selected experiment (see Fig. 4E), the MΦ were incubated for 1 h at 37°C with pharmacological inhibitors prior to their treatment with Nef. PP2 (Src kinase inhibitor; Wako), SU6656 (Src kinase inhibitor; Calbiochem), Src Kinase Inhibitor I (Calbiochem), LY294002 (PI3K inhibitor; Calbiochem), and IPA-3 (p21-activated kinase inhibitor; Calbiochem) were used in this study. All of the inhibitors were dissolved in DMSO (Wako) and added to the cultures at a final concentration of 10 μM (0.1% v/v). The same volume of DMSO was used as a vehicle control.
Western blotting was performed essentially as described previously (37). Briefly, the MΦ were stimulated and lysed on ice with Nonidet P-40 lysis buffer (1% Nonidet P-40, 50 mM Tris, and 150 mM NaCl) containing protease inhibitors (1 mM EDTA, 1 μM PMSF, 1 μg/ml aprotinin, 1 μg/ml leupeptin, and 1 μg/ml pepstatin) and phosphatase inhibitors (1 mM Na3VO4 and 1 mM NaF). Total-cell lysates were then subjected to Western blotting. The following Abs were used: anti–p-specific p38 (Thr180/Tyr182; Cell Signaling), anti-p38 (#A-12; Santa Cruz Biotechnology), anti–p-specific JNK (Thr183/Tyr185; Cell Signaling), anti-JNK (#FL; Santa Cruz Biotechnology), anti–p-specific ERK (Tyr204; Santa Cruz Biotechnology), anti-ERK (#K-23; Santa Cruz Biotechnology), anti–p-specific IκB kinase (IKK)α/β (Ser176/Ser180; Cell Signaling), anti-IKKα (#2682; Cell Signaling), anti–IκB-α (#C-21; Santa Cruz Biotechnology), anti–p-specific Hck (Tyr411; Santa Cruz Biotechnology), anti-Hck (#18; BD Transduction), anti–p-specific Akt (Ser473; Cell Signaling), anti–p-specific Stat1 (pY701; BD Biosciences), anti-Stat1 (#610185; BD Biosciences), anti–p-specific Stat3 (pY705; BD Biosciences), anti-Stat3 (#610189; BD Biosciences), anti-phosphotyrosine (PY99; Santa Cruz Biotechnology), anti-actin (#C-2; Santa Cruz Biotechnology), and anti-His tag (#H-15; Santa Cruz Biotechnology). Detection was performed with HRP-labeled secondary Abs (GE Healthcare), Immunostar LD Western blotting detection reagent (Wako), and an image analyzer (ImageQuant LAS 4000; GE Healthcare).
The uptake of Nef by M-CSF–derived or GM-CSF–derived MΦ was determined by flow cytometry on a FACSCalibur using Cell Quest software (Becton Dickinson). The day-5 MΦ were incubated with FITC-labeled Nef (Fitzgerald, Acton, MA) at a final concentration of 100 ng/ml for 15, 30, or 60 min, detached from the wells using 0.25% trypsin (Wako) or enzyme-free Cell Dissociation Buffer (Life Technologies), and immediately subjected to flow cytometric analysis. FITC-labeled Tat (ImmunoDiagnostics, Woburn, MA) was also used at a final concentration of 100 ng/ml as a reference. In a selected experiment (see Fig. 4C), recombinant Nef proteins labeled with FITC using a Fluorescein Labeling Kit-NH2 (Dojindo, Kumamoto, Japan) were used. The fluorescent signals were also visualized with a BZ-8000 fluorescence microscope (Keyence, Osaka, Japan) equipped with Plan-Fluor ELWD 20×/0.45 objective lenses (Nikon). Image processing was performed with a BZ-Analyzer (Keyence) and Adobe Photoshop software (Adobe Systems).
The expression of surface molecules on M-CSF–derived MΦ cultured in the presence or absence of Nef for 2 d was also determined by flow cytometry. The following labeled Abs were used: FITC-labeled anti–HLA-DR (#HLDR01; Caltag), FITC-labeled anti-CD204, FITC-labeled anti-CD163 (38), and PE-labeled anti–M-CSFR (#3-4A4; Santa Cruz Biotechnology). The phagocytic activity of MΦ was determined by measuring the uptake of fluorescent microspheres (Fluoresbrite Carboxylate Microspheres, 0.7 μm in diameter; Polysciences, Warrington, PA). MΦ cultured on 24-well tissue culture plates were incubated with the fluorescent microspheres (1/2000 dilution) for 1 h at 37°C, washed extensively with PBS, detached from the wells using enzyme-free Cell Dissociation Buffer, and immediately subjected to flow cytometric analysis (38).
The relative levels of multiple cytokines and chemokines in the supernatants of MΦ were analyzed using a human cytokine array (R&D Systems), according to the manufacturer’s instructions. Briefly, M-CSF–derived day-5 MΦ were cultured in the presence or absence of Nef for 2 d, and the culture supernatants collected after centrifugation (100 μl) were added to dot blots onto which the capture Abs had been spotted in duplicate. After incubation with the secondary Ab mixture, the signals were detected using Immunostar LD Western blotting detection reagent and an ImageQuant LAS 4000 image analyzer. The intensity of the spots was quantified using ImageQuant TL software (GE Healthcare).
Soluble HIV-1 proteins activate MAPK and NF-κB pathways in M-CSF–derived M2-MΦ more strongly than in GM-CSF–derived M1-MΦ
M-CSF–derived and GM-CSF–derived MΦ prepared in this study showed spindle-like and fried egg-like morphologies, respectively (Supplemental Fig. 1A). Phagocytic activity was higher in M-CSF MΦ, whereas the expression of transferrin receptor (CD71) was higher in GM-CSF MΦ (data not shown), which were consistent with previous reports (36, 39). Using the widely used method to prepare M2-MΦ (M-CSF derived) and M1-MΦ (GM-CSF derived) (13–16), we examined whether soluble Tat, gp120, or Nef differentially activated these MΦ by comparing the activation of MAPK, such as p38, JNK, and ERK. As reported (18–27), all soluble HIV-1 proteins clearly activated p38 (Supplemental Fig. 1B). However, we found that p38 activation was more marked in M2-MΦ for all HIV-1 proteins (Supplemental Fig. 1B). Neither Tat- nor Nef-induced p38 activation was inhibited by polymyxin B, the concentration of which (10 μg/ml) completely inhibited 10 ng/ml LPS-induced p38 activation (data not shown), indicating that the observed effect was not due to endotoxin contamination. The recombinant Nef preparation did not induce p38 activation after immunodepletion with anti-Nef Abs (Supplemental Fig. 1C), confirming that the observed effect was specific to Nef. Because gp120 required a higher concentration (2.5 μg/ml) to activate p38 compared with Tat (50 ng/ml) or Nef (100 ng/ml), and Nef was more potent in signaling activation than Tat in our system, we focused on Nef in the subsequent experiments.
To confirm the marked signaling activation in M2-MΦ by Nef, we performed more detailed time-course analyses. M1-MΦ showed higher basal ERK activation, which was not enhanced by Nef, whereas M2-MΦ showed lower ERK activation, which was enhanced by Nef (Fig. 1, p-ERK blot). The activation of p38 was marked and rapid in M2-MΦ (Fig. 1, p-p38 blot). Moreover, the activation of JNK was detectable only in M2-MΦ (Fig. 1, p-JNK blot). We also analyzed the activation of the NF-κB pathway. In NF-κB pathway activation, IKKα/β are activated through serine phosphorylation, which phosphorylates IκB, leading to its degradation and the nuclear translocation of NF-κB (26). Thus, we assessed NF-κB pathway activation by IKKα/β phosphorylation and IκB degradation. As a result, we found that IKKα/β phosphorylation by Nef was detectable only in M2-MΦ (Fig. 1, p-IKKα/β blot), and IκB degradation was obvious in M2-MΦ but not in M1-MΦ (Fig. 1, total IκB blot). Therefore, these results indicated that Nef activated MAPK and NF-κB pathways in M2-MΦ more strongly than in M1-MΦ.
Activation of MAPK and NF-κB pathways in M2-MΦ by soluble Nef is comparable to that by TNF-α
We next analyzed the signaling activation in M2-MΦ by Nef by varying its concentration. The activation of p38, JNK, and ERK was detectable at 3 ng/ml (Fig. 2A). In contrast, higher concentrations were required to detect IKKα/β phosphorylation (100 ng/ml) or IκB degradation (30 ng/ml) (Fig. 2B). Importantly, the activation of these signaling pathways by Nef was comparable to that by TNF-α, the well-defined proinflammatory cytokine (40). Compared at the peak (30 and 10 min for Nef and TNF-α, respectively), the degrees of p38 activation, ERK activation, and IκB degradation induced by Nef were similar to those induced by TNF-α (Fig. 2C).
Rapid uptake of Nef by M2-MΦ
We next sought to clarify why Nef preferentially activated M2-MΦ. It was shown that soluble Nef enters MΦ or dendritic cells and accumulates in perinuclear regions (41, 42) through endocytosis, pinocytosis, or as-yet-unknown mechanisms (26). Of importance, we found that FITC-labeled Nef entered M2-MΦ more rapidly than it entered M1-MΦ (Fig. 3A, left and middle panels, respectively), which correlated with greater phagocytosis (Supplemental Fig. 2A) and pinocytosis activities (Supplemental Fig. 2B) of M2-MΦ. Such a strong FITC-Nef signal was not observed when M2-MΦ were incubated at 4°C (Fig. 3A, right panels), as previously reported (41). MΦ collected using trypsin showed weaker FITC-Nef signals than did those collected using enzyme-free buffer; however, the difference was modest (Fig. 3A), suggesting that most signal was not due to an association of FITC-Nef proteins to the surface of MΦ. Indeed, a strong fluorescent signal was detected in the perinuclear regions of M2-MΦ (Fig. 3B). Furthermore, we found that FITC-labeled Tat also entered M2-MΦ more rapidly than it entered M1-MΦ in the flow cytometric analysis, but the signals in both MΦ populations were significantly lower than those observed with FITC-Nef (data not shown). Therefore, it was likely that Nef rapidly and efficiently entered M2-MΦ and, thereby, induced the strong activation of MAPK and NF-κB pathways in M2-MΦ.
Effect of mutation in Nef and chemical inhibitors on Nef-induced signaling activation in M2-MΦ
We next sought to clarify the molecular mechanism by which Nef induced the strong activation of MAPK and NF-κB pathways in M2-MΦ. To this end, we tried several approaches. First, we found that pretreatment of M2-MΦ with TNF-α completely prevented the activation of p38 and JNK by subsequent TNF-α stimulation, but it had no effect on subsequent Nef stimulation, including the activation of p38 and JNK, or IKKα/β phosphorylation (Supplemental Fig. 3), suggesting that Nef does not use TNF-αR to activate these signaling pathways. We next examined whether Nef induced the activation of receptor tyrosine kinases (RTK), using a p-RTK Ab array and a p-MAPK Ab array as a reference (both from R&D Systems), because the activation of RTK often leads to MAPK activation. However, in contrast to the strong MAPK activation (Supplemental Fig. 4A), we found no evidence of RTK activation by soluble Nef (Supplemental Fig. 4B).
We then compared the ability of three Nef mutants (Fig. 4A) to activate MAPK and NF-κB pathways. When expressed endogenously, Nef downregulates the cell surface expression of MHC class I and the HIV-1R CD4 (43–46). The former and latter functions are believed to diminish the recognition of the infected cells by CTL and allow efficient viral release from host cells, respectively (43–46). The G2A mutant, which lacks a myristoylation site, is defective in both functions; the AxxA mutant, in which the proline-rich PxxP motif is mutated, is defective in MHC class I downregulation; and the LL/AA mutant, in which the di-leucine motif is mutated, is defective in CD4 downregulation (45). The LL/AA mutant retained stimulatory activity (Fig. 4B). In contrast, the G2A and AxxA mutants failed to induce the activation of p38 and JNK, IKKα/β phosphorylation, or IκB degradation (Fig. 4B), even at a high concentration (300 ng/ml). It was likely that the signaling-activation inability of these mutants was a postinternalization event, because there was no obvious difference in the internalization efficiency between the wild-type (WT) and these mutants (Fig. 4C). When expressed endogenously, Nef binds and activates the Src family tyrosine kinase Hck (47–49), which is highly expressed in MΦ. Hck activation is mediated through the PxxP motif of Nef (47–49). Indeed, we found that soluble Nef also induced the activation of Hck (Fig. 4D, p-Hck blot). Of importance, such Hck activation was obvious in M2-MΦ but not in M1-MΦ (Fig. 4D). Nevertheless, its kinetics was slow, and it peaked later did than the activation of p38 and JNK (Fig. 4D), implying that Hck activation was not the cause of Nef-induced MAPK and NF-κB pathways. Furthermore, Src family kinase inhibitors (PP2, SU6656, and Src Inhibitor I) failed to block Nef-induced p38 activation, IKKα/β phosphorylation, and IκB degradation (Fig. 4E), although at least SU6656 markedly reduced the basal protein tyrosine-phosphorylation level (total pTyr blot). The PI3K inhibitor (LY294002) and the p21-activated kinase inhibitor (IPA-3) also failed to block the signaling activation induced by Nef (Fig. 4E). The concentration of LY294002 used in this study (10 μM) was sufficient to inhibit the activation of Akt, a downstream molecule of PI3K (Fig. 4F), and the concentration of IPA-3 (10 μM) was shown to inhibit the activation of p21-activated kinase (50). These results indicated that the strong activation of MAPK and NF-κB pathways in M2-MΦ by Nef was dependent on its proline-rich PxxP motif but independent of the activation of Hck, PI3K, or p21-activated kinase.
Effect of soluble Nef and its mutants on phenotypes of M2-MΦ
Finally, we examined how Nef modulated the phenotypes of M2-MΦ. M2-MΦ cultured with soluble Nef for 2 d showed obvious morphological changes (i.e., the appearance of a number of MΦ with protrusions) (Fig. 5A, arrows). Based on the finding that soluble Nef strongly activates MAPK and NF-κB pathways in M2-MΦ (Figs. 1, 2), we next analyzed whether the treatment of M2-MΦ with Nef led to the production of proinflammatory cytokines or chemokines using the semiquantitative array. Nef-treated M2-MΦ produced higher amounts of proinflammatory cytokines, such as TNF-α and IL-6, and chemokines, such as MIP-1α and MIP-1β, in two donors tested (Fig. 5B). It was shown that Nef activates Stat family proteins in MΦ through the release of proinflammatory cytokines and chemokines (24, 27). Indeed, we found the activation of Stat1 and Stat3 in Nef-treated M2-MΦ (Fig. 5C), the kinetics of which were delayed compared with those related to the activation of MAPK and NF-κB pathways (Fig. 1).
Of importance, we found that Nef significantly reduced the surface expression of CD163 scavenger receptor in M2-MΦ (Fig. 6A), the high expression of which is one of the characteristics of M2-MΦ (51, 52). Although the surface expression levels of another scavenger receptor CD204, as well as HLA-DR, were also reduced, the degree was moderate compared with that of CD163 (Fig. 6A). The downregulation of CD163 was also observed with the Nef LL/AA mutant (Fig. 6B). However, the G2A and AxxA mutants failed to induce CD163 downregulation (Fig. 6B). Moreover, we found that Nef also reduced phagocytic activity (Fig. 6B), which is another characteristic of M2-MΦ (13, 36). As was the case for CD163 downregulation, the reduced phagocytic activity was observed with the LL/AA mutant but not with the G2A or AxxA mutants (Fig. 6B). In this study, we found that CD163 downregulation and reduced phagocytic activity by soluble Nef were not blocked by the pretreatment of M2-MΦ with p38 inhibitor (SB 239063), JNK inhibitor (JNK inhibitor II), or NF-κB inhibitor (DHMEQ) (data not shown). Instead, soluble Nef strongly downregulated the surface expression of M-CSF receptor (Fig. 6C), as observed with Nef-expressing MΦ (10). Therefore, it was possible that CD163 downregulation and reduced phagocytic activity by soluble Nef were due to the downregulation of M-CSFR, followed by an impaired response of MΦ to M-CSF, which is important for the M2-phenotype of MΦ (7, 8, 51, 52).
In summary, the current study strongly suggests that HIV-1 proteins, particularly Nef, efficiently enter M2-MΦ, activate their MAPK and NF-κB pathways through the PxxP motif, and drive them toward MΦ with an M1-like phenotype.
Studies with different mice (M-CSF–deficient op/op mice, M-CSFR knockout mice, and GM-CSF knockout mice) clearly demonstrated that the development or survival of most tissue MΦ is dependent on the M-CSFR system (30, 31, 53) and its ligands, including M-CSF and possibly the newly identified alternative ligand IL-34 (54). Moreover, it was suggested that, under normal conditions, peripheral blood monocytes are predisposed toward an M2 phenotype and are mostly devoted to tissue repair as a result of their stimulation by the relatively high levels of M-CSF present in sera (6, 55). Indeed, a transcriptome analysis showed that M2 polarization involved a minimal alteration in MΦ steady-state mRNA expression compared with M1 polarization (14). It was also shown that, unlike T cells, MΦ polarization is transient and highly reversible (56). Therefore, more marked activation of MAPK and NF-κB pathways in M2-MΦ by soluble HIV-1 proteins, such as gp120 (Supplemental Fig. 1B), Tat (Supplemental Fig. 1B), and Nef (Fig. 1), appears to be an efficient mechanism by which HIV-1 induces the production of proinflammatory MΦ. Such a response of M2-MΦ to soluble HIV-1 proteins might be a rapid process, because it occurs independently of viral replication within the MΦ.
Among those tested in our system (gp120, Tat, and Nef), Nef was the most potent activator of MAPK and NF-κB pathways of M2-MΦ, the degree of which was comparable to that of TNF-α (Fig. 2C). Although the concentration of Nef required for the optimal activity (100–300 ng/ml) was higher than that detected in patients’ sera (1–10 ng/ml) (29), the activation of p38 was detectable at a minimal concentration of 3 ng/ml (Fig. 2A). The mechanism and significance of the sustained activation of ERK in M2-MΦ by Nef remain unclear (Fig. 1A), but such sustained ERK activation has been found in several stimuli such as TPA-induced megakaryocytic differentiation of K562 cells (57–59). Soluble Nef was shown to activate MAPK and NF-κB pathways in MΦ (25–27). However, the in vitro preparations of differentiated MΦ used in these studies varied; for instance, monocytes were cultured with GM-CSF and then FCS alone (26) or were cultured with a high concentration of FCS alone (27). To our knowledge, the current study is the first report in which the response to exogenous Nef was compared between two major MΦ populations. In this study, we showed that Nef markedly stimulated the production of proinflammatory cytokines/chemokines, such as TNF-α, IL-6, MIP-1α, and MIP-1β, in M2-MΦ (Fig. 5B). Of interest, Nef did not stimulate the production of IL-12 or IL-23 in M2-MΦ (Fig. 5B), the higher expression level of which is one of the characteristics of M1-MΦ (13). More functional and transcriptional analyses, including the expression of transcription factors, such as IRF4 (60) and IRF5 (61), are needed to understand the phenotypic changes induced in M2-MΦ by Nef. Despite the unresolved issue, our novel finding that Nef reduced the expression of CD163 and phagocytic activity of M2-MΦ (Fig. 6), both of which are characteristics of M2-MΦ (13, 36, 51, 52), strongly suggests that Nef drives anti-inflammatory M2-MΦ toward MΦ with an M1-like phenotype.
The reason why soluble Nef activates MAPK and NF-κB pathways in M2-MΦ more strongly than in M1-MΦ might be because Nef rapidly and efficiently enters M2- MΦ (Fig. 3). Indeed, it was shown that the N-terminal myristoylation of Nef at the glycine residue is required for Nef to target the cellular membrane (62), and we found that the nonmyristoylated Nef G2A mutant neither activated the MAPK/NF-κB pathways (Fig. 4B) nor reduced the surface expression of CD163 or phagocytic activity (Fig. 6B). Of interest, the proline-rich PxxP motif-disrupted AxxA mutant also lost these abilities (Figs. 4B, 6B). The proline-rich motif of Nef was shown to bind the Src homology 3 (SH3) domains of a subset of cellular Src family tyrosine kinases, such as Hck, Lyn, and possibly c-Src (47–49, 63). Among them, the interaction with Hck is important because it causes the activation of Hck kinase activity (47–49). However, the activation kinetics (Fig. 4D) and the results of our pharmacological inhibition analyses (Fig. 4E) did not support the idea that the activation of Hck or other Src kinases was involved in the soluble Nef-mediated activation of MAPK and NF-κB pathways. Because the proline-rich sequence acts as a canonical SH3 domain-binding motif, these results suggest that an unidentified SH3-containing cellular protein mediates the signaling activation of MΦ by soluble Nef. When expressed endogenously, Nef activates PI3K (64) and p21-activated kinase (65). However, our pharmacological inhibition analysis did not support the idea that these kinases were involved in the soluble Nef-mediated activation of MAPK and NF-κB pathways (Fig. 4E). Mangino et al. (66) recently showed that the pathway involving TNFR-associated factors was required for soluble Nef-induced activation of Stat family proteins, which may provide another explanation for why Nef preferentially activates M2-MΦ and clarify why the PxxP motif-disrupted Nef mutant fails to activate M2-MΦ.
The activation of MΦ by soluble HIV-1 proteins is thought to contribute to the sustained activation of the immune system observed in HIV-1 infection (6, 17). Our finding that HIV-1 proteins, particularly Nef, preferentially target M2-MΦ, which predominate under normal conditions, and drive them toward MΦ with an M1-like phenotype provides a novel mechanism by which HIV-1 efficiently and rapidly induces the sustained immune activation. More detailed phenotypic characterization of MΦ treated with gp120, Tat, and Nef will clarify the pathological significance of the activation of MΦ by HIV-1 proteins and the molecular mechanisms by which MΦ differentiation is physiologically regulated.
We thank F. Koutaki and H. Motoyama for secretarial assistance.
This work was supported by a grant from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (to S.S.). This work was also supported by the Global Center of Excellence program “Global Education and Research Center Aiming at the Control of AIDS,” which was commissioned by the Ministry of Education, Culture, Sports, Science, and Technology of Japan (to S.O. and S.S.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
receptor tyrosine kinase
Src homology 3
The authors have no financial conflicts of interest.