The Ubiquitin Ligase c-Cbl Down-Regulates FcγRIIa Activation in Human Neutrophils¹

Polymorphonuclear neutrophils play an important role in the innate immune system where their major, although not unique, function is to internalize and degrade pathogens or immune complexes. This phagocytic process is greatly aided by opsonization of pathogens by complement fragments and IgGs. The latter are recognized, through their Fc tails, by FcγRs. Recognition of pathogens or immune complexes by FcγRs initiates intracellular signals that lead to multiple neutrophil physiological responses: degranulation, activation of the respiratory burst, and phagocytosis (1, 2, 3, 4, 5).

Human neutrophils constitutively express only two activating FcγRs, namely FcγRIIa (CD32a) and FcγRIIIb (CD16b). Several lines of evidence indicate that FcγRIIa is directly involved in the phagocytic process (6, 7, 8), whereas FcγRIIIb participates only indirectly and probably in a nonredundant manner in the phagocytic function of the neutrophil.

The structure of FcγRIIa is unique among FcγRs consisting of a single transmembrane polypeptide chain comprising two extracellular Ig domains, a single transmembrane domain and a short cytoplasmic segment containing an ITAM in which the signature tyrosines are 12 aa apart rather than 8 as in canonical ITAMs (9, 10). This unique molecular signature is likely to impact on the specific composition of the signaling complex through which FcγRIIa transmits the appropriate signals to neutrophils. It should be noted that this receptor is not expressed in mouse and, therefore, that mouse knockout models are not available to directly study its function and regulation. Mouse neutrophils express FcγRIIIa (a multimeric receptor with an ITAM in its accessory γ-chain) and FcγRIIb, the unique inhibitory member of the FcγRs family, which transmits inhibitory signals through an ITIM (11, 12, 13). FcγRIIb is known to negatively control the activation of ITAM-containing immunoreceptor. In the case of Ab-mediated phagocytosis, macrophages have the capacity to trigger a strong proinflammatory response, which must be tightly controled by FcγRIIb. However, the available data indicate that human neutrophils, in contrast to human macrophages, express very little (14) or no (our personal observations) FcγRIIb. So, the neutrophil appears to be the only human phagocyte that does not express ITIM-bearing FcγRs, and the mechanism involved in the down-regulation of FcγRIIa remains to be elucidated.

In systemic autoimmune diseases such as rheumatoid arthritis, a sustained recruitment of leukocytes in joints is observed, and high levels of autoantibodies form immune complexes with their cognate Ags and represent potent activators of effector cells (including neutrophils) via cross-linking of FcγRs. Since neutrophils are the major cells that infiltrate the synovial fluid in rheumatoid arthritis (15), the activation of FcγRs on neutrophils by IgG-containing immune complexes is thought to play an important role in the chronicity and the severity of the disease, leading to inflammation and joint destruction. Furthermore, in rheumatoid arthritis, as well as in systemic lupus erythematosus (16), the efficiency of anti-TNF-α therapy (17) or methotrexate treatment (18) is related to a loss of function of FcγRIIa. Thus, understanding the mechanism of activation of neutrophils via FcγRIIa at the molecular level is clearly important for the development of strategies aimed at modulating their involvement in host defense as well as in chronic inflammation. In the present study, we wanted to understand the regulation of FcγRIIa activation in human neutrophils in which ITIM-bearing FcγRs are absent.

The earliest events following cross-linking of FcγRIIa on human neutrophils include the translocation of the receptor to detergent-resistant membrane (DRM)³ domains (19, 20) often thought of as signaling platforms (21, 22, 23) and an enhancement of the tyrosine phosphorylation pattern (19, 20, 24). Among the tyrosine-phosphorylated substrates, the rapid and transient tyrosine phosphorylation of the FcγRIIa itself (6, 25, 26) and of the ubiquitin ligase/adaptor protein c-Cbl (27) has been described previously.

The Cbl protein family comprises c-Cbl, Cbl-b, and Cbl-3. These three homologs have been initially described as adaptors because they were shown to interact by via their tyrosine kinase-binding domains with numerous signaling proteins, (recently reviewed by Schmidt and Dikic (28)). More recently, these proteins were described to possess ubiquitin ligase (E3) activity that is mediated through their Ring finger domain. Hence, they can negatively regulate signaling by directing the ubiquitination and the degradation of activated receptors (29). EGFR is a well-studied model of the prominent role of c-Cbl in the down-regulation of receptor tyrosine kinases (30, 31, 32, 33, 34, 35). Recent findings demonstrated that Cbl-b and c-Cbl played different roles in EGFR regulation (36, 37, 38, 39). In human neutrophils, we have previously shown that c-Cbl is rapidly phosphorylated following the cross-linking of FcγRIIa and composed of the major band of 116 kDa that appears in the tyrosine phosphorylation pattern observed in response to ligation of this receptor (27). The tyrosine phosphorylation of c-Cbl was demonstrated to play a critical role in its ubiquitin ligase activity (40, 41, 42). The c-Cbl homologs (Cbl-b and Cbl-3) have not been described in human neutrophils.

Our previous studies have shown that cross-linking of FcγRIIa on human neutrophils led to a rapid Src-dependent loss of immunoreactivity of the fraction of the receptor that translocates to DRMs (19, 20). In this study, we investigated the biological and biochemical basis of the loss of signal of FcγRIIa in human neutrophils as well as its functional impact. We show that, in human neutrophils, FcγRIIa is ubiquitinated by c-Cbl, leading to the down-regulation of its surface expression and function and its degradation likely via the proteasomal pathway. Our hypothesis is that this stimulated degradation represents a non-ITIM-dependent termination step following the engagement of the FcγRIIa and, for this reason, could play an important role in the regulation of the activation of FcγRIIa in human neutrophils.

Three different Abs against FcγRIIa were used in this study. 1) mAb IV.3 was purified from ascites of mice inoculated with hybridoma HB-217 obtained from the American Type Culture Collection. This Ab recognizes a native extracellular epitope of FcγRIIa. It was used for all the cross-linking and flow cytometry experiments. Confocal microscopy experiments were performed with PE-conjugated IV.3 (catalog no. IM1935) from Beckman Coulter. 2) Polyclonal goat anti-human FcγRIIa/CD32a Ab (catalog no. AF1875) was purchased from R&D Systems. It recognizes an extracellular epitope on FcγRIIa and was used for immunoblotting (see Fig. 1 C only). 3) CT10 is an IgG fraction of a polyclonal rabbit serum against the cytoplasmic domain of FcγRIIa raised against the synthetic peptide whose sequence was published by Ibarrola et al. (43). It was used for immunoblotting.

The anti-phosphotyrosine (4G10; catalog no. 05-321) and anti-PI3K subunit p85 (catalog no. 06–195) Abs were purchased from Upstate Biotechnology. Goat anti-mouse F(ab′)₂ (anti-Fc, catalog no. 115-006-071; or anti-F(ab′)₂, catalog no. 115–006-072), HRP-labeled donkey anti-rabbit IgGs (catalog no. 711-035-152), and FITC-conjugated F(ab′)₂ fragment goat anti-mouse IgG (Fc fragment specific; catalog no. 115–096-071) were purchased from Jackson ImmunoResearch Laboratories. HRP-labeled sheep anti-mouse IgGs (catalog no. NXA931) were obtained from GE Healthcare. Polyclonal rabbit anti-ubiquitin Ab (catalog no. Z0458) was purchased from DakoCytomation. Anti-c-Cbl (catalog no. SC-170) and anti-SHIP-1 (catalog no.SC-8425) Abs were purchased from Santa Cruz Biotechnology. Monoclonal anti-flotillin-1 (catalog no. F65020-050) was purchased from BD Biosciences.

4-Amino-5-(4-chlorophenil)-7-(t-butyl)pyrazol(3,4-d) pyrimidine) (PP2) was obtained from Calbiochem. Sodium orthovanadate (Na₃VO₄), soybean trypsin inhibitor, Triton X-100, N,2′-O-dibutyryladenosine-3′,5′-cyclic monophosphate sodium salt (db AMPc), PMSF, chloroquine, ammonium chloride (NH₄Cl), DMSO, f-Met-Leu-Phe, and human IgGs were obtained from Sigma-Aldrich Canada. Dextran T-500, Percoll, and protein A-Sepharose were purchased from GE Healthcare. Aprotinin and leupeptin were purchased from Roche Applied Science. MG132 was obtained from BIOMOL and clasto-lactacystin β-lactone from Cayman Chemical. Western lightning chemiluminescence plus was obtained from PerkinElmer. Ficoll-Paque and HEPES were obtained from Wisent. Di-isopropyl fluorophosphate (DFP) was purchased from Serva Electrophoresis. Gelatin was obtained from Fisher Scientific. RPMI 1640 medium, Zymosan A Bioparticles, calcein-AM, fura 2-AM, and IgG-free FBS were obtained from Invitrogen Life Technologies.

The c-Cbl siRNAs were purchased from Dharmacon. The siGENOME SMARTpool Cbl (catalog no. M-003003-01) is a mixture of four designed siRNAs targeting c-Cbl. The negative control (catalog no. D-001210-01) is a nontargeting siRNA.

The myeloid cell line PLB-985 (from the German Collection of Microorganisms and Cell Cultures) was grown in RPMI 1640 medium containing 10% FBS at 37°C in a 5% CO₂ humidified atmosphere. The cells were maintained in culture for 12 passages before new batches were thawed. To induce differentiation to a neutrophil-like phenotype, PLB-985 cells were cultured in medium supplemented with 0.3 mM dibutyryl cAMP for 3 days before each experiment.

Neutrophils were collected from healthy adult volunteers and sterilely isolated as described previously (44). They were resuspended in Mg²⁺-free HBSS containing 1.6 mM CaCl₂.

One day following differentiation with dibutyryl cAMP, PLB-985 cells were transiently transfected using the Nucleofector system from Amaxa Biosystems. After centrifugation, 2 × 10⁶ cells were suspended in 100 μl of nucleofection buffer (25 mM HEPES (pH 7.4), 120 mM KCl, 2 mM MgCl₂, 10 mM K₂HPO₄, and 5 mM l-cysteine) containing 3 μg of siRNA. The samples were transferred into an electroporation cuvette, and transfections were performed with the program setting U-002. After nucleofection, the cells were immediately gently transferred into prewarmed complete RPMI 1640 medium containing 0.3 mM dibutyryl cAMP, 10 mM HEPES, and 1 mM sodium pyruvate and maintained at 37°C in a 5% CO₂ humidified atmosphere. Two days after nucleofection, cells were harvested and resuspended in Mg²⁺-free HBSS containing 1.6 mM CaCl₂ for analysis.

Freshly purified human neutrophils or PLB-985 were resuspended in HBSS at 20 × 10⁶ cells/ml (except when indicated). The cells were then incubated with Ab IV.3 (1 μg/ml for 20 × 10⁶ cells/ml) for 5 min and goat anti-mouse F(ab′)₂ anti-Fc (25 μg/ml) (or HBSS for negative controls) was used to cross-link FcγRIIa for the times and temperature indicated in the figure legends. For the experiments of Fig. 4, goat anti-mouse F(ab′)₂ anti-Fc was used at 1.25 μg/ml. For the flow cytometry and FcγRIIa immunoprecipitation experiments, a goat anti-mouse F(ab′)₂ anti-F(ab′)₂ was used for the FcγRIIa cross-linking to let the Fc portion of IV.3 free. The stimulations were stopped at the indicated times by transferring aliquots of the cell suspensions directly in the same volume of 2× boiling Laemmli’s sample buffer (composition of 1×: 62.5 mM Tris-HCl (pH 6.8), 4% (w/v) SDS, 5% (v/v) 2-ME, 8.5% (v/v) glycerol, 2.5 mM orthovanadate, 10 mM para-nitrophenylphosphate, 10 μg/ml leupeptin, 10 μg/ml aprotinin and, 0.025% bromophenol blue) and boiled for 7 min. Where indicated, cells were incubated with PP2 (10 μM) for 10 min before the addition of Ab IV.3.

Neutrophils (20 × 10⁶ cells/ml, 1 ml for each immunoprecipitation) were preincubated with 1 mM DFP for 10 min at room temperature, and the cells were incubated (or not) with PP2 (10 μM) for 10 min before FcγRIIa cross-linking at room temperature. The stimulations were stopped by transferring the cells to an ice bath, and polymorphonuclear neutrophils were quickly centrifuged (10 s at 15,000 × g). The cell pellets were lysed by adding 1 ml of cold lysis buffer (20 mM HEPES (pH 7.4), 137.2 mM NaCl, 1% Triton, 2 mM orthovanadate, 10 μg/ml leupeptin, 10 μg/ml aprotinin, and 3 mM DFP) for 2 min on ice. The cell lysates were sonicated on ice for 2 s at power level 1 in a Branson Sonifier 450 sonicator, and the insoluble material was discarded by two successive centrifugations at 400 × g for 2 min at 4°C. The supernatants (700 μl) were then collected, and FcγRIIa bound to IV.3 was immunoprecipitated by adding protein A-Sepharose beads for 2 h at 4°C. The beads were then collected and washed three times with cold lysis buffer, and 50 μl of Laemmli’s sample buffer was added before boiling for 7 min.

One milliliter of neutrophil suspension (40 × 10⁶ cells/ml, except for plasma membranes purification of dPLB-985 that were performed at 9 × 10⁶ cells/ml in 500 μl), was preincubated with 1 mM DFP for 10 min at room temperature, and FcγRIIa was cross-linked at room temperature. The cells were transferred to an ice bath to stop the stimulations. The cells were then quickly centrifuged (10 s at 15,000 × g) and resuspended in modified relaxation buffer (100 mM KCl, 3 mM NaCl, 10 mM HEPES (pH 7.4), 10 μg/ml aprotinin, 10 μg/ml leupeptin, 2 mM sodium orthovanadate, 250 μg/ml soybean trypsin inhibitor, 1 mM PMSF, and 3 mM DFP). The cells were sonicated on ice for 22 s at power level 1 in a Branson Sonifier 450 sonicator, and the lysates were then centrifuged at 400 × g for 2 min. The supernatants S400 (900 μl for neutrophils, 225 μl for dPLB-985) were applied on top of a two-step Percoll gradient composed of 1.4 ml (350 μl for dPLB-985) of a 1.12 g/ml Percoll solution layered beneath 1.4 ml (350 μl for dPLB-985) of a 1.05 g/ml Percoll solution, as previously described by Kjeldsen et al. (45). The Percoll gradients were centrifuged for 30 min at 37,000 × g at 4°C in a fixed-angle rotor (Beckman TLA100.4). The plasma membranes were found on top of the gradients under clear cytosol fractions. An aliquot of the cytosolic fractions was collected and boiled for 7 min in Laemmli’s buffer. The plasma membranes were collected and centrifuged at 100,000 × g for 45 min at 4°C to remove Percoll. Plasma membranes formed a visible disc above the Percoll pellet. They were collected, resuspended in relaxation buffer, and stored at −80°C. An aliquot was boiled for 3 min in Laemmli’s buffer.

Plasma membranes from control or FcγRIIa cross-linked neutrophils (40 × 10⁶ cell equivalent/ml) were solubilized in 1 or 2% Nonidet P-40 buffer (137 mM NaCl, 20 mM HEPES pH 7.4, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 2 mM sodium orthovanadate, 250 μg/ml soybean trypsin inhibitor, 1 mM PMSF, and 3 mM DPF) for 20 min on ice. Solubilized membranes were adjusted to 40% (v/v) OptiPrep with a stock solution (59.4% OptiPrep/10 mM HEPES (pH 7.4)). Then, 700 μl was transferred to 4-ml centrifuge tubes and overlaid with 700 μl of ice-cold solutions of 35, 30, 25, 20, and 0% OptiPrep. The gradients were centrifuged at 380,000 × g for 3 h at 4°C in a TLA 100.4 rotor, and subsequently, 13 fractions of 300 μl were collected from the top of the gradients. Proteins were chloroform/methanol-precipitated as described previously (46). The precipitates were resuspended in 60 μl of Laemmli’s buffer and boiled for 7 min.

Proteins were separated by SDS-PAGE on 10% acrylamide gels or on 7.5–20% gradient acrylamide gels. The proteins were then transferred to polyvinylidene difluoride membranes. For the anti-ubiquitin blots, the polyvinylidene difluoride membranes were boiled in water for 30 min following the transfer step (47). Blocking agents and Abs were diluted in a TBS-Tween solution (25 mM Tris-HCl (pH 7.8), 190 mM NaCl, and 0.15% v/v Tween 20). Blotto solution (5% w/v) was used to block nonspecific sites before anti-flotillin-1, anti-ubiquitin, and anti-FcγRIIa (CT10 Ab) immunoblotting. Gelatin solution (2% w/v) was used to block non specific sites before anti-c-Cbl, anti-human FcγRIIa/CD32a, anti-PI3K subunit p85, anti-SHIP-1, and anti-phosphotyrosine (4G10) immunoblotting. 4G10 and anti-ubiquitin Abs were diluted 1/4000; CT10 Ab, anti-c-Cbl, and anti-flotillin Abs were diluted 1/1000; anti-human FcγRIIa/CD32a Ab was diluted 1/2000; anti-PI3K subunit p85 was diluted 1/15,000; and anti-SHIP-1 was diluted 1/500. HRP-labeled donkey anti-rabbit IgGs and HRP-labeled sheep anti-mouse IgGs were diluted 1/20,000 in TBS-Tween solution. HRP-labeled donkey anti-goat IgGs were diluted 1/13,300 in TBS-Tween. Chemiluminescence reagents were used to detect Abs, with a maximal exposure time of 5 min. All the immunoblots presented were controlled for equal protein loading with immunoblots against the structural protein flotillin-1, the PI3K subunit p85 or the lipid phosphatase SHIP-1.

Freshly isolated neutrophils (20 × 10⁶ cells/ml) were resuspended in RPMI 1640 medium supplemented with 0.1% IgG-free FBS and prestained with 5 μg/ml calcein-AM for 30 min at 37°C. Cells were centrifuged for 2 min at 1500 × g at room temperature and plated on a glass slide coated with IgG-free 100% FBS. This purified serum was used to avoid neutrophil activation by the glass slide and FcγRs activation by serum IgGs. Neutrophils were stimulated as described in the corresponding legend and visualized live at 37°C in an environment chamber with 5% CO₂ through a spinning disc confocal microscope using ×63 objective (Quorum Spinning Disc Wave FX).

Neutrophils (10 × 10⁶ cells/ml) were preincubated with 1 μM fura 2-AM for 30 min at 37°C. Extracellular probe was removing by washing in HBSS, and cells were resuspended at 5 × 10⁶ cells/ml. Cells were stimulated as described in the corresponding legend. Fluorescence was monitored in a fluorescence spectrophotometer (Fluorolog-SPEX from Jobin Yvon) using two excitation wavelengths at 340 and 380 nm and an emission wavelength of 510 nm. The internal calcium measurement was calculated as the ratio of the fluorescence values obtained at 340 and 380 nm.

Alexa 488-conjugated zymosan A bioparticules were opsonized by incubation with commercial human IgGs at 37°C for 1 h and washed twice in HBSS. A ratio of 10 particules per cell was added at 25 μl of neutrophils (10 × 10⁶ cells/ml) to initiate phagocytosis. To achieve synchronisation of the interaction between particules and cells, the mixure was centrifuged at 400 × g for 30 s before incubation at 37°C. To quench the fluorescence of the noningested Alexa 488-conjugated zymosan, 0.4 mg/ml trypan blue in citrate buffer (pH 4.4) was added 2 min before flow cytometry analysis with an Epics XL flow cytometer from Beckman Coulter.

Because FcγRIIa and FcγRIIIb both recognize the Fc portion of IgGs (including when presented in immune complexes), cross-linking of FcγRIIa with a mAb is widely used to study the signaling pathways and the functions specifically downstream of this receptor. Ab IV.3 used in this study specifically recognizes the extracellular portion of FcγRIIa (48). Previous studies have shown that, following the cross-linking of FcγRIIa, a rapid and transient pattern of tyrosine phosphorylation was induced (19). The time course of this extensive increase in the level of tyrosine phosphorylation is illustrated in Fig. 1,A. The maximum level of tyrosine phosphorylation was observed 30 s following cross-linking of FcγRIIa. Within the next 10 min, a return to the baseline level of tyrosine phosphorylation was observed. Prominent tyrosine phosphorylation bands at 40, 55–70, and 120 kDa were consistently observed. When the same membrane was immunoblotted with CT10 (an Ab directed against the cytoplasmic tail of FcγRIIa), a decrease of the amount of FcγRIIa (a decreased immunoreactivity) was observed as soon as 1 min after cross-linking (Fig. 1,B and Ref. 19). This effect was not observed in unstimulated cells in which equal amounts of FcγRIIa were observed at 30 s and after 10 min of incubation. It should also be pointed out that aliquots of the whole-cell suspensions (cells and stimulation medium buffer) were directly transferred to the tubes containing boiling sample buffer. Therefore, the decrease in CT10 immunoreactivity cannot be explained by a release of intact FcγRIIa into the extracellular medium. When the same samples were immunoblotted with an anti-human FcγRIIa/CD32a Ab, which is directed against the extracellular portion of FcγRIIa, a similar decrease in FcγRIIa immunoreactivity was observed (Fig. 1,C). Therefore, the decrease in FcγRIIa immunoreactivity, which was observed with both anti-FcγRIIa Abs, cannot be explained by a decrease in FcγRIIa recognition by CT10, by a shedding of the FcγRIIa extracellular portion, or by a selective proteolysis of its cytoplasmic tail. Densitometric analysis (Fig. 1 D) indicated that the loss of immunoreactivity was maximal at 2 min following cross-linking, as suggested by previous data (19).

The decrease in FcγRIIa immunoreactivity observed with both Abs following cross-linking of FcγRIIa suggested an intracellular process of degradation for this receptor. To measure the internalization of the receptor following its cross-linking, we quantified the membrane expression of FcγRIIa by flow cytometry. Neutrophils were stimulated as described above by cross-linking FcγRIIa, and the stimulation was stopped on ice after 1, 10, and 15 min. After incubation with fluorescent anti-mouse Abs, neutrophils were analyzed by flow cytometry. As illustrated in Fig. 2,A, the fluorescence intensity detected with the anti-mouse Ab (a measure of surface expression of FcγRIIa) decreased in a time-dependent manner. This effect was obvious within the first minute of stimulation and increased for the next 15 min where up to 80% of FcγRIIa is internalized. To confirm that this decrease was not due to a dissociation of the IV.3 from the receptor, we performed another experiment whose result is presented in Fig. 2,B. The surface expression of FcγRIIa was measured as above (Fig. 2,B, ▪), and using PE-conjugated IV.3, which was cross-linked at 0, 2, or 5 min, we also measured the total amount of FcγRIIa (Fig. 2,B, □). The difference between these two labeling represents the presence of the ligand inside the cell. Finally, we wanted to visualize the presence of the receptor inside neutrophils using confocal microscopy. As shown in Fig. 2 C, before cross-linking (0 s), the receptor is uniformly distributed at the periphery of the cell. Forty-five seconds following cross-linking, the receptor is concentrated as small spots that appear clearly at the inner side of the plasma membrane and later on (90 s); these spots are detected inside the cell. At 300 s, the signal remains inside the cell albeit weaker.

The data presented in Figs. 1 and 2 suggest that FcγRIIa is degraded subsequently to its internalization. A classical protein modification that precedes degradation is ubiquitination. However, this modification is poorly documented in primary neutrophils. To examine whether this possibility occurs in human neutrophils in the present context, we stimulated FcγRIIa on neutrophils as described above for 30 and 90 s, and we prepared plasma membranes as described in Materials and Methods. When these purified membrane fractions were immunoblotted with CT10, a smear of bands appeared in the 40–90-kDa region following stimulation of FcγRIIa (Fig. 3,A). These higher m.w. bands could represent polyubiquitinated and/or multiple monoubiquitinated forms of FcγRIIa. In this experiment, purified membranes were loaded on the gel, and the film was deliberately overexposed to be able to detect the modified forms of FcγRIIa. Because of the overexposure, the loss of immunoreactivity of FcγRIIa, which was observed in whole cells at 40 kDa (compare with Fig. 1), is not obvious in this figure. To determine whether this upward shift in immunoreactivity was caused by the ubiquitination of FcγRIIa, we stimulated neutrophils as described above. Following this stimulation, the receptor was immunoprecipitated, and its potential ubiquitination examined by immunoblotting with anti-ubiquitin Ab. A rapid ubiquitination pattern was induced, which persisted for up to 2 min following receptor cross-linking (Fig. 3,B, upper panel). When the same membrane was immunoblotted with CT10 (Fig. 3 B, lower panel) or with anti-human FcγRIIa/CD32 Ab recognizing the extracellular portion of the receptor (data not shown), a decrease in FcγRIIa immunoreactivity was observed. The time course of this decrease correlates with the ubiquitination pattern confirming that the residual 42-kDa band detected with CT10 Abs represents the nonubiquitinated form of FcγRIIa. No ubiquitinated bands were observed in immunoprecipitations of control cells either at 15 s or after 2 min of incubation. We observed a decrease in the ubiquitination pattern of the immunoprecipitated receptor after 5 min of cross-linking (data not shown).

The ability of 4-amino-5-(4-chlorophenil)-7-(t-butyl)pyrazol(3,4-d) pyrimidine) (PP2), a potent Src kinase family inhibitor, to inhibit the loss of immunoreactivity of FcγRIIa after its cross-linking (19) suggested that these tyrosine kinases are involved in the signaling pathways involved in this mechanism. To determine whether receptor ubiquitination is also dependent on the activity of Src kinases, we performed the same experiment in the presence of PP2. No ubiquitination of FcγRIIa was detected in the presence of PP2, correlating with an inhibition of the loss of immunoreactivity of the receptor (Fig. 3 B).

Ubiquitination generally precedes protein degradation via either the endosome/lysosome or the proteasome pathways. To discriminate between these two pathways, we tested the effect of different inhibitors on the loss of FcγRIIa signal. MG132 and clasto-lactacystin β-lactone were chosen for their specificity toward proteasome activity (49), whereas ammonium chloride (NH₄Cl) and chloroquine were used as inhibitors of the endosome/lysosome pathway (50, 51, 52). As shown in Fig. 4, MG132 (Fig. 4,A) and clasto-lactacystin β-lactone (Fig. 4,B) both inhibited the degradation of FcγRIIa following its cross-linking. The same result was observed with a third proteasome inhibitor, N-acetyl-Leu-Leu-Norleu-AL (data not shown). It is unlikely that, by inhibiting the proteasome pathway, the pool of free ubiquitin was depleted since the stimulation times were short (<10 min). By contrast, neither NH₄Cl (Fig. 4,C) nor chloroquine (Fig. 4 D) had any effect on the loss of FcγRIIa immunoreactivity. The efficacy of these two latter inhibitors was validated for each experiment by measuring the superoxide production in response to f-Met-Leu-Phe as described elsewhere (53). We observed a significant decrease in superoxide production when neutrophils were preincubated with 20 μM NH₄Cl or 200 μM chloroquine (data not shown).

We next wanted to identify the ubiquitin ligase involved in the ubiquitination of FcγRIIa. We first examined the translocation of c-Cbl to the plasma membranes in response to FcγRIIa cross-linking in human neutrophils. c-Cbl is a ubiquitin ligase involved in the down-regulation of the EGFR (29, 40, 54, 55). It is regulated by tyrosine phosphorylation, and in human neutrophils, we have previously shown that c-Cbl is a prominent tyrosine-phosphorylated substrate in response to FcγRIIa cross-linking (27). Neutrophil plasma membranes derived from resting cells or from FcγRIIa-stimulated cells were thus isolated as described in Materials and Methods. They were then immunoblotted with anti-c-Cbl and anti-flotillin1 (a plasma membrane marker) Abs. The addition of IV.3 alone did not induce any translocation of c-Cbl to the plasma membranes (data not shown). Following FcγRIIa cross-linking, we observed a rapid increase in the levels of c-Cbl in the plasma membranes (Fig. 5). This response was observed as early as 15 s following stimulation and was transient. After 5 min of cross-linking, the amounts of c-Cbl in the plasma membranes returned close to their baseline level. Flotillin-1 is a plasma membrane marker (56, 57) whose cell localization is not modified upon FcγRIIa signaling in human neutrophils (20). The anti-flotillin-1 immunoblots indicate that equal amounts of membranes were loaded in each lane. The translocation of c-Cbl to the plasma membranes of stimulated cells was inhibited by PP2 (Fig. 5). These results indicate that c-Cbl is recruited to plasma membranes in a Src kinase-dependent manner in response to cross-linking of FcγRIIa in human neutrophils. They suggest that c-Cbl could, potentially, be the enzyme responsible for the ubiquitination of this receptor.

Our previous studies provided evidence that FcγRIIa was recruited to DRM-H of the plasma membrane after its cross-linking (20). In the present study, we used an OptiPrep gradient, which was recently used for the isolation of neutrophil DRMs (20, 58). After FcγRIIa cross-linking, plasma membranes were purified and solubilized in Nonidet P-40. An OptiPrep step gradient was overlaid on 700 μl of the solubilized membranes and ultracentrifuged as described in Materials and Methods. An analysis of the gradients derived from resting and cross-linked membranes is presented in Fig. 6. In the absence of cross-linking, FcγRIIa was detected in the last fractions (fractions 10–12) of the gradient representing the soluble proteins. Following cross-linking, a shift of the receptor to lighter-density fractions (fractions 8–10) was observed as described previously (20). This receptor shift following FcγRIIa stimulation correlated with the appearance and detection of c-Cbl in fractions 8–10. These results indicate that FcγRIIa and c-Cbl cofractionated in the same DRM-H fractions. The distribution of flotillin-1, a structural raft component (56, 57), remained the same before and after FcγRIIa cross-linking (data not shown and Ref. 20). c-Cbl is absent in the control samples (Fig. 6, upper panel) as it is localized in the cytosol of unstimulated cells.

Because of the difficulties associated with the transfection of human neutrophils, we opted for the use of a human neutrophil-like cellular model, the dibutyryl cAMP-dPLB-985 (59), to further test the involvement of c-Cbl in the stimulated ubiquitination and degradation of FcγRIIa. The responses of these cells to FcγRIIa cross-linking were characterized and found to closely resemble those of primary neutrophils (see Discussion). To confirm the role of c-Cbl in the down-regulation of FcγRIIa expression, silencing of its expression by siRNA transfections of dPLB-985 cells was attempted. This ubiquitin ligase is present endogenously in dPLB-985 cells as shown in Fig. 7,A, upper panel. Following the nucleofection with specific siRNAs, the expression level of c-Cbl was significantly decreased as shown by c-Cbl immunoblotting (Fig. 7,A, upper panel). The decreased levels of expression of c-Cbl correlated with an inhibition of the loss of immunoreactivity of FcγRIIa (Fig. 7,A, lower panel). The same samples were immunoblotted with anti-phosphotyrosine Abs. A stronger and more sustained tyrosine phosphorylation pattern was observed in response to FcγRIIa cross-linking in cells transfected with the siRNA against c-Cbl than in those in which the negative control siRNA was introduced (Fig. 7,B). These results also indicate that the inhibition of FcγRIIa degradation is not due to a lack of stimulation. The data in Fig. 7 C show that silencing c-Cbl had no effect on the internalisation of FcγRIIa. The fluorescence intensity of the cells detected with an anti-mouse Ab (a measure of FcγRIIa surface expression) decreased to a similar manner in negative control cells and in c-Cbl-silenced cells.

Additional evidence that c-Cbl is the ubiquitin ligase involved in the ubiquitination (and down-regulation) of FcγRIIa was sought after by monitoring directly the ubiquitination of FcγRIIa in dPLB-985 cells in which the expression of c-Cbl was reduced by siRNA silencing. FcγRIIa was cross-linked and immunoprecipitated as described in Materials and Methods. The resulting immunoprecipitates (Fig. 8) were probed for ubiquitin in the upper panel, which represents the ubiquitinated forms of cross-linked FcγRIIa. The immunoprecipitates were also probed with CT10 Abs in the lower panel, which represents the 42-kDa nonubiquitinated form of cross-linked FcγRIIa. The ratio of the two panels demonstrates that cross-linked FcγRIIa is mainly ubiquitinated in the stimulated siRNA control-transfected cells. In contrast, the proportion of the ubiquitinated form of the receptor is strongly decreased in stimulated cells transfected with c-Cbl siRNA. Whole-cell samples were probed for c-Cbl expression and confirmed the efficiency of the transfection and of the silencing of c-Cbl expression (data not shown). These results provide additional support for the involvement of c-Cbl in FcγRIIa ubiquitination.

Two experiments were performed to show that FcγRIIa degradation following its cross-linking is relevant to dampening neutrophil activation (Fig. 9). First, we observed that following calcium mobilization in response to FcγRIIa cross-linking, a subsequent stimulation of the receptor by addition of the secondary cross-linking Ab (Fig. 9,A, upper panel) or by addition of IV.3 and the secondary cross-linking Ab (Fig. 9,A, middle panel) did not trigger a second wave of calcium mobilization whereas another agonist f-Met-Leu-Phe did (Fig. 9,A, lower panel). This result demonstrates a desensitization of FcγRIIa following its engagement. Finally, we analyzed the consequence of this desensitization on one of the main functions of neutrophil, i.e., phagocytosis (Fig. 9 B). Using fluorescent zymosan opsonized with human IgGs, we measured the number of neutrophils that internalized zymosan. We observed a 75% decrease of the percentage of cells having internalized zymosan particles when FcγRIIa was previously cross-linked. A mean of two fluorescent particles were internalized with or without FcγRIIa prestimulation in those cells in which internalized zymosan was detected. This result illustrates the role of FcγRIIa down-expression in dampening a functional response downstream of this receptor.

We previously showed that the amount of FcγRIIa that can be detected by immunoblotting of whole human neutrophils rapidly decreases following its ligation (19, 20). In this study, we further characterized this event and showed that it is associated with a rapid internalization and ubiquitination of the receptor followed by a proteasome-dependent degradation. We also observed that the ubiquitin ligase c-Cbl is recruited to the plasma membrane where it cofractionates in the same subdomains as the receptor. Silencing of c-Cbl in dPLB-985 cells provided direct data indicating that c-Cbl plays a prominent role in the ubiquitination and the degradation of the receptor. Finally, we showed a functional significance of FcγRIIa down-expression as intracellular mobilization and phagocytosis are inhibited following a previous round of receptor cross-linking.

In the present study, we investigated the molecular mechanism(s) involved in the previously reported loss of immunoreactivity of FcγRIIa following its cross-linking in human neutrophils (19). Several explanations for this observation were conceivable. It should be noted that extracellular shedding of the receptor could not account for the decreased immunoreactivity as we analyzed by SDS-PAGE whole-cell suspensions containing the extracellular medium. The stimulated decreased immunoreactivity could have resulted from a loss of reactivity of CT10 with the tyrosine-phosphorylated receptor. This possibility was eliminated by the observation that a similar decrease in detection of FcγRIIa was observed using a separate Ab directed against its extracellular domain. Additionally, we never detected low m.w. bands, whatever the Ab used for immunoblotting. These last observations indicate that a selective proteolysis of the cytoplasmic tail of the receptor can also be ruled out. A stimulated degradation (loss of immunoreactivity) of FcγRIIa has previously been noted in HL60 cells (60) and in FcγRIIa-transfected CHO-ts20 cells (61). It should be noted that, in the latter model, the degradation was observed only after more than 60 min of stimulation and that, in sharp contrast to the situation in human neutrophils, it was insensitive to inhibition by PP2 (62), thereby suggesting that its molecular basis may be different. We also observed a lack of calcium response to the cross-linking of FcγRIIa in cells in which FcγRIIa had been cross-linked 5 min previously. FcγRIIa was also found to be degraded subsequently to the stimulation of human neutrophils with heat-aggregated human IgGs (data not shown). Taken together, these observations strongly suggest that cross-linking of FcγRIIa in human neutrophils leads to the rapid degradation of the receptor and that the down-regulation mechanism herein described is of physiological relevance.

A step that classically precedes the degradation of a plasma membrane protein is its internalization. To verify this hypothesis, we analyzed by flow cytometry the amount of FcγRIIa on the plasma membrane of the cells following cross-linking. The results indicate that cross-linked FcγRIIa is rapidly internalized together with its ligand (within 1 min) (Fig. 2). The kinetics of this observation suggest that receptor internalization precedes its intracellular degradation. Confocal microscopy pictures confirm this rapid internalization. An internalization of FcγRIIa was previously suggested on the basis of microscopical techniques (63) and radioactivity measurements (64) but was never quantified as we did in the current study. Furthermore in these two previous studies, FcγRIIa was described inside the cell after 10 min (62) or 2 min (63) cross-linking. Our flow cytometry results indicate that 50% of surface FcγRIIa is internalized as soon as 1 min after cross-linking.

Cross-linking of FcγRIIa results in a time-dependent tyrosine phosphorylation (Fig. 1 and Ref. 19). Among the tyrosine-phosphorylated substrates, the receptor itself, the tyrosine kinase Syk, the lipid phosphatase SHIP-1 (65), and the ubiquitin ligase c-Cbl have been characterized (27). The tyrosine phosphorylation of these substrates is strongly inhibited following a preincubation of neutrophils with 4-amino-5-(methylphenyl)-7-(t-butyl)pyrazolo[3,4-d]-pyrimidine (PP1) or PP2, two Src kinase inhibitors (19). The loss of signal of the receptor is inhibited in the presence of these inhibitors (19) and as shown in Fig. 3,B, ubiquitination of the receptor is also Src kinase dependent. Our results differ from those observed in FcγRIIa-transfected CHO-ts20 cells where the delayed ubiquitination of FcγRIIa was reported to be independent of tyrosine phosphorylation (62). These differences outline the difficulties of extrapolating from model cell systems to primary cells. Our present data, together with those we have previously published (19), confirm the major role Src kinases play in the early steps of the signaling cascades activated following the cross-linking of FcγRIIa. Furthermore, the dephosphorylation of the receptor is not required for its degradation since preincubation of neutrophils with pervanadate, a tyrosine phosphatase inhibitor, had no effect on the degradation of FcγRIIa (data not shown). MG132 and clasto-lactacystin β-lactone both clearly inhibit the stimulated degradation of FcγRIIa, whereas chloroquine and NH₄Cl had no effect (Fig. 4). Although these latter results do not rule out a potential, late involvement of the lysosomal compartment, they do provide solid evidence that this mechanism is not predominant during the early responses monitored in this study and that the proteasome pathway appears responsible for the down-regulation of FcγRIIa. Specific signals are required to target proteins to proteasome pathway. Ubiquitination is a classical posttranslational protein modification that precedes protein degradation via the endosome/lysosome or the proteasome pathway (29). In the current study, we detected the ubiquitination of FcγRIIa, which explains the loss of immunoreactivity observed at early cross-linking times (Figs. 1, B and C, and 3) and targets the receptor to the proteasome pathway for its degradation. At later cross-linking times (Fig. 1, B and C), the loss of immunoreactivity of FcγRIIa may be explained by both its ubiquitination and proteasome-dependent degradation.

Very few membrane receptors have been described to be regulated via the proteasome. In the case of the EGFR, inhibitors of endosomal/lysosomal and of proteasomal degradation both inhibit EGFR degradation, indicating that the two pathways may be interrelated (31). By contrast, the lysosomal pathway does not seem to be involved in our system, at least at the earlier times. In neutrophils, which are widely described as a professional phagocyte, the endosome/lysosome pathway (leading to the phagosome maturation) and the proteasome pathway may represent two distinct and highly regulated pathways. In the absence of large opsonized pathogens or particles, as is the case in systemic autoimmune diseases and in our model of stimulation, the use of the proteasomal pathway for FcγRIIa down-regulation may serve to limit an excessive activation of neutrophils and the resulting potentially deleterious consequences.

A common point between the two pathways mentioned above is the ubiquitination of the receptor. Three classes of enzymes are required for the ubiquitination process: ubiquitin-activating enzyme (E1), ubiquitin-conjugating enzyme (E2), and ubiquitin ligase (E3) (reviewed in Ref. 66). A likely candidate for attachment of ubiquitin on FcγRIIa following cross-linking is the ubiquitin ligase c-Cbl. This enzyme, constitutively expressed in human neutrophils, is tyrosine phosphorylated following FcγRIIa stimulation (27). This tyrosine phosphorylation is required for its ubiquitin ligase activity (29). To affix ubiquitin molecules, ubiquitin ligases have to recognize their target by protein-protein interactions. Because c-Cbl is localized in the cytosol in resting neutrophils whereas FcγRIIa is a membrane protein, we first examined whether c-Cbl translocates from the cytosol to the plasma membrane following FcγRIIa stimulation. The results of our experiments provide evidence for a rapid recruitment and a sustained localization of c-Cbl in membrane fractions following FcγRIIa cross-linking (Fig. 5). The kinetics of translocation (detectable within 15 s of stimulation) of this E3 ligase correlate with the tyrosine phosphorylation pattern that was observed following FcγRIIa engagement (Fig. 1). The translocation of c-Cbl (Fig. 5) precedes the degradation of the receptor and is correlated with FcγRIIa ubiquitination (Figs. 1, B and C, and 3). Furthermore, the stimulated recruitment of c-Cbl to the plasma membrane was, as previously shown for the loss of immunoreactivity of FcγRIIa, sensitive to the Src kinase inhibitor PP2 (Fig. 5). The lack of degradation of FcγRIIa previously observed in plasma membranes isolated from resting neutrophils (20) (which do not contain much if any detectable levels of c-Cbl) is consistent with the requirement of cytosolic elements, c-Cbl among others, in the down-regulation of this receptor expression. Taken together, these data confirm the important function of c-Cbl, which is relocalized to the plasma membrane to affix ubiquitin molecules on FcγRIIa before its degradation.

DRM microdomains are described as platforms serving to concentrate signaling proteins (67). We have previously observed that FcγRIIa, which is mostly localized in detergent-soluble fractions of resting neutrophil plasma membranes, rapidly moves into a subset of DRMs following its engagement (20). This result is also illustrated in Fig. 6. We have now observed that cross-linking of FcγRIIa results in the translocation of c-Cbl to the same DRM-H as the receptor itself (Fig. 6). This is in agreement with our previous results (68), which showed the presence of the tyrosine-phosphorylated fraction of c-Cbl in the insoluble fraction when whole neutrophils were lysed in Nonidet P-40 solutions. Taken together, these observations further support an implication of the ubiquitin ligase c-Cbl in the signaling pathways initiated upon engagement of FcγRIIa.

Confirmation of the implication of c-Cbl in the stimulated ubiquitination of FcγRIIa was obtained in the human myeloid cell line PLB-985. We first obtained detailed evidence that the basic molecular mechanisms involved in the responses of db-cAMP-dPLB-985 cells closely approximated those we previously characterized in human neutrophils. Of particular relevance were the observations that cross-linking of FcγRIIa in dPLB-985 cells induced its internalization (Fig. 7,C), its ubiquitination (Fig. 8), its loss of immunoreactivity (Fig. 7,A) as well as a translocation of c-Cbl to plasma membranes (data not shown), and an increase in the global tyrosine phosphorylation pattern (Fig. 7,B). These data provide significant weight to the argument that these cells represent an adequate model to use in our investigations. Nucleofection of siRNA oligomers directed at c-Cbl very effectively silenced the expression of c-Cbl in db-cAMP-dPLB-985 cells. This resulted in a decrease in the amount of FcγRIIa ubiquitinated forms following its cross-linking leading to a significant inhibition of the degradation of FcγRIIa. These results are consistent with the interpretation that c-Cbl is critically involved in the ubiquitination of stimulated FcγRIIa and that this event plays a major role in directing the latter toward the degradation pathway. We also observed that the tyrosine phosphorylation response to the cross-linking of FcγRIIa is prolonged in time in the absence of c-Cbl, thereby indicating that the ubiquitination of FcγRIIa (and its degradation) are involved in the turning off of the signals initiated upon engagement of this receptor. However, FcγRIIa internalization was not modified in dPLB-985 cells in which c-Cbl was silenced (Fig. 7 C). The mechanisms underlying the internalization of this receptor differ from those of the EGFR, where the binding of c-Cbl to EGFR is indispensable for its internalization (69, 70). However, as for FcγRIIa, ubiquitination of the EGFR is not necessary for its internalization (71). These results may be explained by the double role of c-Cbl. First, this protein possesses a ubiquitin ligase activity, and second, it contains several well-characterized protein-protein interaction domains (Src homology 2-like and proline-rich regions), which confer scaffolding properties (29). Taken together, these results confirm the functional ability of dPLB-985 following nucleofection and are in agreement with a specific role of c-Cbl in the ubiquitination of the receptor and the cellular mechanisms downstream of this event rather than in the internalization process.

The functional role of ubiquitination in endocytosis and down-regulation of receptors remains a challenging subject and is not documented in human neutrophils. To our knowledge, this study is the first one demonstrating the ubiquitination of a receptor and its degradation by the proteasome in human neutrophils. Other immunoreceptors have been shown to be ubiquitinated and down-regulated by the ubiquitin ligase c-Cbl. In the case of FcεRI on the rat basophilic leukemia cell line RBL-2H3, c-Cbl is responsible for the ubiquitination of both FcεRI and Syk (whose kinase activity is induced after receptor engagement), leading to the down-regulation of the engaged receptor complexes and targeting the receptor to degradation (72). Furthermore, another study demonstrated that FcεRI engagement leads to the recruitment of c-Cbl in lipid raft domains (73). We also observed a translocation of c-Cbl to a specific subset of lipid rafts following the engagement of FcγRIIa. The TCR is another receptor whose function is modulated by c-Cbl through the ubiquitination of the TCRζ-chain (74). Finally, c-Cbl was shown to down-modulate the phagocytic function mediated by FcγRs (75, 76). These examples illustrate the involvement of the ubiquitin ligase activity of c-Cbl in the down-regulation of various immunoreceptors.

Our present results supplement previous data concerning the fate of FcγRIIa following its engagement on neutrophils. First, FcγRIIa is recruited from detergent-soluble regions of the plasma membrane to DRMs and more specifically to DRM-H (19, 20). Afterwards, it is internalized, ubiquitinated by c-Cbl and likely moves to the proteasome complex where it is degraded. The inhibition of calcium mobilization and phagocytosis that we observed following a previous round of FcγRIIa cross-linking indicates that this mechanism leading to FcγRIIa degradation is likely to allow the neutrophil, which does not possess a classical ITIM-dependent (FcγRIIb) mechanism for FcγRIIa down-regulation, to avoid overactivation during IgG complexes clearance. Neutrophil activation through FcγRs is exacerbated in systemic autoimmune diseases such as systemic lupus erythematosus, rheumatoid arthritis, or antineutrophil cytoplasmic Ab-associated vasculitis (Wegener’s granulomatosis) (77). This justifies a concerted investigation of the different mechanisms involved in the regulation of the signaling events that follow FcγRs stimulation. In the present study, we characterized novel mechanisms involved in the regulation of the expression of the opsonic receptor FcγRIIa in human neutrophils and their role in the functional responses elicited upon the occupation and activation of the latter. This mechanism may suggest novel strategies for the treatment of autoimmune pathologies.

We thank Dr. Maurice Dufour for expert technical assistance with the flow cytometric analysis and Dr. Maria J. Fernandes and Valérie Gagné for expert technical assistance with confocal microscopy.

The authors have no financial conflict of interest.