Signal-Specific Activation and Regulation of Human Neutrophil Fc Receptors
http://www.100md.com
免疫学杂志 2005年第9期
Abstract
FcRs with the ITIM domain have been shown to regulate the inflammatory signal delivered by the ITAM-containing FcRs. In this study, we demonstrate that the function of human neutrophil FcR type IIA (CD32A) is regulated in a distinct manner by different cell activation signals at the ligand-binding stage. Activation of neutrophils with fMLP up-regulated the ligand-binding function of CD32A, whereas PMA-mediated activation completely abolished ligand binding without altering CD32A expression. Furthermore, PMA treatment also abolished CD16B-dependent ligand binding irrespective of the level of expression. The effect of PMA was cell type specific, because the ligand-binding function of CD32A expressed on cultured cells such as K562 and CHO-CD32A transfectants was not affected by PMA. Interestingly, phorbol 12,13-dibutyrate, another phorbol ester, and IL-8 up-regulated CD32A-dependent ligand-binding function. These results demonstrate that regulation of CD32A-dependent ligand binding in human neutrophils is not only cell type specific but also activation signal specific. Moreover, these results suggest the possibility that signals delivered to neutrophils by various inflammatory stimuli can exert opposing effects on the function of human FcRs, representing a novel inside-out regulatory mechanism of FcR ligand binding.
Introduction
The FcRs expressed on inflammatory cells mediate Ab-dependent cellular cytotoxicity and the immunophagocytosis of invading pathogens. FcR-mediated cellular toxicity has been implicated in many autoimmune and immune complex (IC)4-mediated diseases (1, 2, 3). Inflammatory cells in mice and humans express multiple FcRs with overlapping ligand-binding specificity (4, 5, 6, 7). However, the physiological significance and the mechanism of regulation of these coexpressed receptors are not clearly understood. Recently, gene knockout studies in mice have shown that the coexpression of an ITIM-containing FcR, CD32B, with an ITAM-containing CD16A-/ complex results in controlled immune responses and prevents IC-mediated diseases (1, 8, 9, 10). Thus, the inhibitory signal delivered by CD32B attenuates an excessive activation signal delivered by CD16A during normal immune responses. These observations suggest that opposing signals delivered by FcRs during responses to IC and invading pathogens provide an important mechanism to regulate inflammatory signals produced by ITAM-containing FcRs and that an imbalance in activation vs inhibitory signals may result in immunopathological reactions leading to tissue injury. However, it is not known how the functions of FcRs are regulated in human neutrophils.
The FcR system in mouse and human neutrophils are quite different, and therefore studies from mice cannot be directly applied to humans. Murine neutrophils express polypeptide-anchored CD16A and CD32B (11), whereas human neutrophils express GPI-anchored CD16B and polypeptide-anchored CD32A (12, 13). The CD16A expressed on mouse neutrophils can deliver the activation signal by associating with ITAM-containing or subunits (14), and this activation can be regulated by ITIM-containing mouse CD32B (15). In contrast, human CD32A has an ITAM motif in the cytoplasmic domain and is capable of transducing inflammatory signals (16, 17). Moreover, the signal delivered by CD16B is not inhibitory to CD32A; rather, studies have shown that CD16B augments the function of CD32A (16, 17). It is possible that CD32B may be regulating inflammatory signals in human neutrophils too, but studies so far have not shown the expression of CD32B molecule on human neutrophil surface, although CD32B mRNA has been documented (18). Previous studies in our laboratory (19) and others (20, 21) have shown that, in human neutrophils, IC binding can be completely blocked by a combination of mAbs specific to CD16B and CD32A, suggesting that functionally active CD32B protein may not be expressed on neutrophils. These observations suggest that CD32A might be regulated by alternative mechanisms in humans.
Recently, we have reported that neutrophils keep their strong signaling CD32A receptor in a low-avidity state (19), but once neutrophils are activated with agents such as fMLP, a chemotactic peptide derived from bacteria, CD32A binds Ab-coated target cells with high avidity. This up-regulation of ligand-binding function by cell activation is very similar to regulation of the LFA-1 integrin molecule (22, 23). Yet, unlike the integrin system, the data presented in this report demonstrate that two well-known neutrophil activators, fMLP and PMA, regulate CD32A ligand-binding function in an opposing manner. Such a signal-specific alteration in ligand binding represents a novel inside-out regulatory mechanism that may influence the function of FcRs in a distinct manner under various inflammatory conditions.
Materials and Methods
Reagents
Rabbit anti-DNP IgG, fMLP, PMA, phorbol 12,13-dibutyrate (PDBu), phorbol 12,13-diacetate (PDAc), 4--phorbol 12-myristate 13-acetate (4-PMA), C5a, 1,10-phenanthroline, PKH-26 (a lipophilic dye), and other chemicals were from Sigma-Aldrich. Human IL-8 (R&D Systems) was a kind gift from Dr. V. Udhayakumar at Centers for Disease Control (Atlanta, GA). SRBC were purchased from Colorado Serum Company. All cell culture reagents were purchased from Invitrogen Life Technologies. FITC-conjugated F(ab')2-goat anti-mouse IgG was from Jackson ImmunoResearch Laboratories.
Cell lines and Abs
K562, CHOK1, mouse hybridomas secreting anti-CD32 (IV.3), anti-CD64 (32.2), anti-CD11b (LM2.1.6.11), anti-CD3 (OKT3) mAbs, and mouse myeloma secreting a nonbinding IgG (X63) were purchased from the American Type Culture Collection. These cells were cultured in RPMI 1640 supplemented with 10% FBS (HyClone), 2 mM glutamax I, 1 mM sodium pyruvate, penicillin (100 U/ml), and streptomycin (100 μg/ml). CHOK1-CD32A transfectants and hybridoma secreting anti-CD16 (CLBFcgran-1) were described earlier (19, 24). Anti-CD16 (CLBFcgran-1) and anti-CD32 (IV.3) IgG were purified from corresponding hybridoma culture supernatant using protein G-Sepharose. Fab of purified IgG were prepared by pepsin digestion (custom service by Lampire Biological Laboratories) and further purification using protein G-Sepharose.
Preparation of neutrophils
Neutrophils were isolated from peripheral blood of normal volunteers by dextran sedimentation and Ficoll-Paque (1.077) density gradient centrifugation as described earlier (19). Contaminating erythrocytes in the neutrophil fraction were removed by hypotonic lysis in water for 20 s. Neutrophils were resuspended in RPMI 1640/0.3% BSA (activation buffer). Because change in temperature during preparation of neutrophils has been shown to activate neutrophils (25, 26), all procedures were conducted at room temperature. Neutrophils were prepared at room temperature and used as quickly as possible. All washing media and centrifuges were also kept at room temperature to avoid temperature fluctuations during neutrophil preparations. Blood collection from normal volunteers was done according to the approved protocols by Institutional Review Board of Emory University.
Activation of neutrophils
Neutrophil activation was performed as described earlier (19). Unless indicated, activation was done by incubating neutrophils in the activation buffer with the specified activators at 37°C for 45 min. After activation, neutrophils were washed once and resuspended in cold HBSS (Ca2+/Mg2+)/0.3% BSA/5 mM EDTA (binding buffer). In some experiments, immediately after activation, neutrophils were fixed with 0.04% paraformaldehyde for 10 min at room temperature. The fixed neutrophils were washed twice in binding buffer and resuspended in cold binding buffer, until used in the ligand-binding assays.
Rabbit Ab-coated sheep erythrocyte (EA) binding assay
EA binding assays were conducted using SRBC-coated with trinitrophenol followed by opsonization with rabbit anti-DNP IgG (hereafter referred to as EA), as described earlier (24). A minimum of 200 cells was examined under light microscopy for rosetting. Neutrophils with a minimum of five EA attached were scored as a rosette. In some of the assays, EA binding was analyzed by flow cytometry using PKH-labeled EA (27). PKH labeling of EA was performed as described earlier (27, 28). The PKH-labeled EA was stored in erythrocyte storage buffer EAS45 (29). EA stored in this buffer can be used for ligand-binding experiments for up to 10 days. Neutrophils (50 μl of 5 x 106/ml) in binding buffer were incubated with PKH-labeled EA (50 μl of 2 x 108) for 2–4 h at 4°C. To determine CD32A-dependent EA binding, neutrophils were preincubated with Fab (5 μg/ml) of anti-CD16 mAb (CLBFcgran-1) for 20 min on ice (19). PKH-labeled unopsonized erythrocytes were used as a control in all the experiments. Mean fluorescence intensity (MFI) at FL2 channel and the percentage of FcR+ cells bound to EA was determined in a FACScan or FACSCalibur flow cytometer (BD Biosciences). The EA binding was represented as percentage of rosette formation or the attachment index when rosettes were counted under a microscope or by flow cytometry, respectively. The attachment index was calculated by the following formula: % cells bound to EA x mean fluorescence/100. As reported by others (27), and from the results from the studies here, we observed that the quantitative analysis of rosette formation measured by both of the methods is comparable.
Flow cytometric analysis
The cell surface expression of FcRs and the neutrophil activation marker (CD11b) were analyzed by indirect immunofluorescence flow cytometry. Cells (5 x 105) were incubated with corresponding mAbs or control mIgG (X63) followed by FITC-conjugated F(ab')2-goat anti-mouse IgG. The samples were analyzed in a FACScan flow cytometer (BD Biosciences). For quantitation of cell surface markers, standard fluorescent-labeled beads (Bangs Laboratories) were used under the same settings.
Results
Signal-specific modulation of ligand-binding function of neutrophil CD32A
In our previous studies, we demonstrated that the CD32A-mediated neutrophil ligand-binding function is up-regulated by fMLP treatment (19). In this report, we investigated whether this up-regulation was a result of a general activation-associated phenomenon or was signal specific. Neutrophils were activated with PMA and then were tested for CD32A-mediated EA binding. fMLP was included as a control for neutrophil activation in these experiments. Total binding (contributed by both CD16B and CD32A) and CD32A-dependent binding were measured. Microscopic evaluation (Fig. 1A) and quantitative measurement of ligand binding showed that the total EA binding was not altered after fMLP activation (B). As expected, fMLP-activated neutrophils showed an increase in CD32A-dependent rosette formation (Fig. 1, A and B), whereas, surprisingly, PMA activation of neutrophils resulted in complete loss of both total and CD32A-dependent rosette formation (A and B). This was not due to the loss of CD32A expression, because flow cytometry analysis showed that neutrophils retained >70–80% of their CD32A expression after PMA activation (Fig. 1C). Most of the CD16B expression was lost after activation, possibly due to the shedding of CD16B during neutrophil activation (13, 30). Therefore, the loss of total binding could be due to loss of CD16B expression, because studies from our laboratory (19) and others (20, 21) have shown that CD16B is the major receptor for IC binding to resting neutrophils. CD11b expression, a neutrophil activation marker, was up-regulated about 3-fold, confirming that the neutrophils were activated with PMA (Fig. 1C). This observation of PMA-mediated complete down-modulation of FcRs ligand-binding function was confirmed with neutrophils from >30 individual donors. Collectively, these results have indicated that the fMLP-mediated CD32A up-regulation is not donor specific, although the level of CD32A-depenendent rosette formation may vary (19). Similarly, PMA-mediated inhibition of the rosette formation was observed in all of the donors we have tested so far.
FIGURE 1. Activation signal-specific modulation of CD32A-dependent ligand-binding function in neutrophils. Neutrophils from the control donor were treated with medium or fMLP (1 μM) or PMA (50 ng/ml) in serum-free RPMI 1640 medium for 45 min at 37°C. A, Microscopic evaluation of neutrophil (N) rosette formation with EA (E) was performed using light microscope, and the pictures were taken using Leica PhotoShop software. B, Quantitative rosette formation with resting and activated neutrophils with EA were conducted as described in Materials and Methods. Total binding (left panel) and CD32A-dependent EA binding (performed in the presence of Fab of anti-CD16 IgG) (right panel) were determined as described in Materials and Methods. Experiments were done in triplicate. Rosette formation data are representative of >20 independent experiments with different donors. C, Flow cytometry analysis of CD16B, CD32A, and CD11b expression. Unactivated (medium) or activated (fMLP or PMA) neutrophils were stained with appropriate mAbs (filled histogram) and a nonbinding mIgG (open histogram). Cells were analyzed using a FACScan flow cytometer. The number indicates the mean fluorescence value.
The inactivation of FcRs by PMA does not depend on the level of CD16B expression
Next, we determined whether shedding of CD16B was responsible for the loss of EA binding in PMA-treated neutrophils. It has been reported that the partial loss of CD16B expression due to shedding is caused by metalloproteinases and can be prevented by metalloproteinase inhibitor 1,10-phenanthroline (31, 32). To investigate this possibility, neutrophils were preincubated with 1,10-phenanthroline before activation with fMLP or PMA. Neutrophils activated with fMLP or PMA showed a reduced level of CD16B expression (only 30% that of unactivated neutrophils) (Fig 2A) without a change in CD32A expression (data not shown). As reported earlier (31, 32), pretreatment of neutrophils with 1,10-phenanthroline restored the level of CD16B expression (80% of the unactivated neutrophils). Yet, despite having near-normal levels of CD16B expression after 1,10-phenanthroline pretreatment, PMA-activated neutrophils did not bind to EA (Fig. 2B). Interestingly, even though the levels of CD16B expression were the same after PMA or fMLP activation, only fMLP-activated neutrophils retained their ability to bind to EA. 1,10-Phenonthroline pretreatment alone did not have any effect on the EA binding of FcRs in neutrophils (data not shown). These findings suggest that down-modulation of FcR-mediated ligand-binding functions of neutrophils by PMA is not due to the levels of CD16B or CD32A expression.
FIGURE 2. CD16B density does not influence the ligand binding on PMA-activated neutrophils. Neutrophils were incubated with or without 1,10-phenanthroline (5 mM) in serum-free RPMI 1640 (Medium) for 30 min at 37°C. Untreated (– phenon) and 1,10-phenanthroline-treated (+ phenon) neutrophils were further incubated with fMLP (1 μM) or PMA (50 ng/ml) in serum-free RPMI 1640 medium for 45 min at 37°C. A, Cell surface expression of CD16B was analyzed by flow cytometry using specific mAb or nonbinding mIgG (solid-line open histogram). Closed and dotted line histograms represent cells treated without and with 1,10-phenanthroline, respectively. B, Rosette formation of untreated and 1,10-phenanthroline-treated resting and activated neutrophils with EA were carried by flow cytometry-based rosetting assay, as described in Materials and Methods. A representative of two independent experiments from two donors is presented.
Kinetics of modulation of FcRs ligand binding by PMA and fMLP treatment
Next, the kinetics of the effect of PMA and fMLP treatment on total and CD32A-dependent ligand binding was determined. As shown in Fig. 3A, the total EA binding of PMA-activated neutrophils was only 28% at 20 min compared with 70% in resting neutrophils, and the EA binding was completely abolished at 40 min. However, the CD32A-dependent rosette formation in fMLP-activated neutrophils increased as early as 10 min and was at a maximum at 40 min (Fig. 3B). To determine the concentration of PMA that is necessary for PMA-mediated down-modulation of FcRs, the activation experiment was repeated using neutrophils from another donor. Neutrophils were treated with different concentration of PMA. The down-modulation of total CD32A-dependent EA binding to PMA-activated neutrophils was also dose dependent. Nearly 70–80% inhibition of rosette formation was observed at a concentration as low as 1–2 ng/ml PMA, and complete loss of rosette formation was seen at 10 ng/ml PMA (Fig. 3C). However, neutrophils activated with fMLP showed an increase in CD32A-dependent EA binding (Fig. 3D).
FIGURE 3. Kinetics and dose dependency of PMA-mediated down-modulation of EA rosette formation. A and B, Kinetics of down-modulation: Neutrophils (1 x 107) isolated from the control donor were treated without or with 1 μM fMLP () or 50 ng/ml PMA () in serum-free RPMI 1640 medium for various times at 37°C. After activation, neutrophils were fixed with RPMI 1640/10% FBS/0.04% paraformaldehyde for 10 min on ice. Total EA rosette formation (A) with neutrophils was done as previously described. To determine CD32A-dependent rosette formation (B), rosetting was performed in the presence of 2 μg/ml Fab of anti-CD16 IgG (CLB). C and D, Dose dependency of down-modulation: Neutrophils were activated with different concentrations of PMA (C) or fMLP (D) as described above. Total and CD32A-dependent EA rosette formation was analyzed as described in Materials and Methods. The figure is a representative of two independent experiments from two different donors.
Effect of C5a- and IL-8-mediated neutrophil activation on FcRs ligand binding
Because neutrophil activation with fMLP or PMA showed opposing effects on the ligand-binding functions of FcRs, we investigated whether neutrophil activation by inflammatory mediators such as C5a and IL-8 would also modulate the FcRs function. Neutrophils from different donor were treated with C5a or IL-8 at the indicated concentration, and expression of FcRs and CD11b was analyzed by flow cytometry. As a control, cells were also treated with PMA and fMLP. The level of CD16B was decreased upon activation, whereas the level of CD32A was not altered (Fig. 4A). Neutrophil activation marker CD11b was increased 2- to 3-fold, confirming the activated state of neutrophils (Fig. 4A). The EA binding to activated neutrophils was determined by flow cytometry. As shown in Fig. 4B, the total EA binding was not altered when neutrophils were activated with fMLP, IL-8, or C5a, whereas PMA activation completely abolished it. Neutrophils treated with fMLP and IL-8 showed 3- and 1.5-fold increases in CD32A-dependent EA binding, respectively, whereas C5a had no effect or moderate increase in CD32A-dependent binding (Fig. 4C). These results demonstrate that different neutrophil activators distinctly modulate the ligand-binding function of FcR.
FIGURE 4. Effect of neutrophil activators on FcR-mediated EA binding. Neutrophils were incubated with fMLP (1 μM) or PMA (2 ng/ml) or C5a (10 ng/ml) or IL-8 (10 ng/ml) for 45 min at 37°C. Cells treated with medium at 4°C were used as resting neutrophils. A, Flow-cytometric analysis of resting and activated neutrophils was performed as described in Materials and Methods using appropriate mAbs. B and C, Total (B) and CD32A-dependent EA (C) binding was conducted as described in Materials and Methods. The EA binding was analyzed by in a FACSCalibur flow cytometer. The attachment index was quantitated as described in Materials and Methods. Data from one of the two donors are presented.
Specificity of PMA-mediated loss in FcRs ligand-binding function
PMA-mediated inactivation of FcRs in this report contradicts a previous report showing that PDBu, another phorbol ester (33), activated the ligand-binding function of CD32A. This difference could be due to the use of different phorbol esters. To resolve this discrepancy, neutrophils (from another donor) were activated with PMA or PDBu, and EA binding was determined. Both PMA and PDBu showed 3-fold increase in CD11b expression, indicating that neutrophils were activated by both phorbol esters (Fig. 5A). Compared with the expression of unactivated neutrophils, CD16B expression of PMA- or PDBu-activated neutrophils were 34 and 68%, respectively (Fig. 5A). Similar results were observed with neutrophils obtained from other donors. The FcR-dependent EA binding was determined by flow cytometry as described in Materials and Methods. PMA-activated neutrophils did not bind to EA; however, PDBu-activated neutrophils bound to EA (Fig. 5B). In a separate experiment using neutrophils from another donor, we also determined the effect of other phorbol esters such as PDAc and 4-PMA (inactive analog of PMA) on CD32A ligand binding in neutrophils. As shown in Fig. 5C, PMA-activated neutrophils showed minimal EA binding, whereas PDAc and PDBu did not affect the EA binding. Furthermore, neutrophils treated with 4-PMA, an inactive analog of PMA, did not affect EA binding to neutrophils, suggesting that the PMA-mediated down-modulation of CD32A function is specific. Earlier studies have used higher concentrations of PDBu (33), thus raising the possibility that the difference observed in EA ligand binding in this study with PDBu or PMA was due to the difference in the concentration of phorbol esters used in both studies. To test this issue, neutrophils were activated with different doses of PMA or PDBu, and EA binding was determined. PMA-mediated inactivation of FcR-dependent EA binding was seen at a concentration as low as 1 ng/ml PMA. Yet, even at a higher concentration (100 ng/ml) of PDBu, EA binding was not altered (Fig. 5D). These findings indicate that the difference between the present observation and the earlier report on the effect of phorbol esters on FcR-dependent EA binding may be due to the difference in phorbol esters used for neutrophil activation.
FIGURE 5. Effect of different phorbol esters on neutrophil FcR-mediated EA binding. Neutrophils were incubated with fMLP (1 μM) or PMA (1 ng/ml) or other phorbol esters (2 ng/ml) as indicated in the figure. Cells incubated in medium at 4°C represent resting neutrophils. A, Expression of CD16B, CD32A, and CD11b on resting (medium) and activated neutrophils were analyzed by flow cytometry, and the mean fluorescent values are represented in the figure. B, Total EA rosette formation of resting and activated neutrophils was analyzed by flow cytometry. The attachment index was calculated as described in Materials and Methods. C, Effect of different phorbol esters on EA rosette formation. Cells incubated with PMA or other phorbol esters (PDAc, PDBu) or inactive analog 4-PMA were used in this experiment. EA rosette formation was determined by flow cytometry. M1 represents the histogram peak for rosetted cells. D, Effect of different concentrations of PMA and PDBu. Cells were treated with PMA or PDBu at indicated concentration. Flow-cytometric analysis of EA rosette formation of resting (medium) and activated neutrophils was performed as described in Materials and Methods. Representative of three independent experiments from three different donors is given. The numbers given in the figure represent MFI/percentage of neutrophils formed rosettes, respectively.
PMA-mediated FcRs inactivation is cell type specific
In our earlier studies, we observed that 53% of CHOK1 cell transfectants expressing CD32A (CHO-CD32A) formed rosettes with EA, and this binding was completely blocked by anti-CD32A mAb, IV.3 (19). The anti-CD32 mAb has shown to be specific for CD32A receptor (34). Similarly, CD32A expressed on K562, a human erythroleukemia cell line that naturally expresses CD32A, is able to bind EA efficiently, suggesting that CD32A expressed in this cell line is constitutively active (Fig. 6A). These results suggest that CD32A expressed on cultured cells such as K562 and CHO-CD32A cells, unlike neutrophils, is constitutively active. We tested whether the PMA-mediated inactivation of CD32A was also cell type specific. EA binding to both CHOK1-CD32A and K562 cells was determined after pretreating the cells with PMA. Incubation of CHO-CD32A and K562 cells with PMA before rosetting did not influence the EA rosette formation (Fig. 6A). These findings indicate that CD32A is fully functional in CHO cells, and the PMA-mediated inactivation of CD32A is cell type specific. Similarly, incubation of CHO-CD16B transfectants with PMA before rosetting did not abolish EA rosette formation (Fig. 6B), suggesting both FcRs expressed in CHO cells are not susceptible to inactivation by PMA.
FIGURE 6. PMA-mediated CD32A inactivation is cell type specific. A, CHO cell transfectants expressing CD32A, and K562 cells were incubated with medium () or PMA (5 ng/ml) () for 1 h at 37°C and allowed to form rosette with EA in ice for 2 h. Fab of IV.3 (5 μg/ml) completely blocked this binding (data not shown). B, Rosetting of CHO cell transfectants expressing CD16B was conducted as described above. The results are expressed as the mean of three experiments. mAbs specific to FcRs completely blocked this EA binding. Untransfected CHO cells did not form EA rosettes (data not shown).
It is possible that the difference in EA binding between cultured cells and unactivated neutrophils may be due to the difference in CD32A receptor density on the cell surface. To test this possibility, we measured the level of CD32A expression in CHO-CD32A transfectants and K562 cells. CD32A expression was analyzed by immunofluorescence flow-cytometric analysis, and the level of CD32A expression (MFI) was determined using standard beads run under identical instrument settings. The ligand-binding function was determined by EA rosette formation. As shown in Table I, CD32A expression was about the same in K562 but 10-fold higher in CHO-CD32A cells, compared with neutrophils. However, the surface area of neutrophils (300 μm2) was smaller than that of CHOK1 and K562 cells (591 and 584 μm2, respectively) (35, 36, 37). When the cell surface density of CD32A was normalized for its surface area, neutrophils expressed 1.6-fold higher levels of CD32A than K562 cells. Interestingly, CD32A-dependent EA binding to neutrophils was only 5% compared with 47.5% in K562. The EA binding was completely blocked by Fab of anti-CD32A mAb (data not shown). These results show that the higher level of EA binding mediated by CD32A expressed on K562 cell line is not due to higher receptor density, and suggest that these cells constitutively express a high-avidity form of CD32A. CHO-CD32A cell transfectants also formed 45% rosettes with EA. The surface density of CD32A expressed on CHO cells, however, is 6-fold higher than the neutrophils, and therefore, at present we cannot rule out the role of surface density in CHO-CD32A cell transfectants binding to EA.
Table I. Efficient CD32A-mediated EA binding by K562 cells is not due to higher receptor densitya
Discussion
Receptors involved in cell-cell interactions carry out two vital functions that are necessary for productive cell communication. First, they bind to counter receptors/ligands on the opposing cell, and second, after binding, they deliver activating or inhibitory signals inside the cell. Therefore, the regulation of these receptors at any one of their functions is critical for normal immune response. Abnormal regulation of these receptors has been found in many disease states, including fatal autoimmune disorders, such as lupus and arthritis (1, 38, 39, 40). FcRs are not only involved in binding of IgG-coated target cells, but also signal for subsequent effector functions of the cell. Many of the FcRs contain an ITAM in their cytoplasmic domain or associate with ITAM-containing subunits that transduce activation signals to cells. Interestingly, CD32B expressed on inflammatory cells has an ITIM, which delivers inhibitory signals to the cells. Studies using knockout mouse models suggest that the signals transduced by ITIM can regulate the activation signal delivered by ITAM-containing FcRs (41, 42, 43). Thus, coexpression of ITAM- and ITIM-containing receptors play an important role in regulating immune responses during humoral immunity. Recently, Ravetch and coworker (44) have shown that the signal delivered through ITAM and ITIM motifs may influence IC-induced maturation of dendritic cells, suggesting that the cross-regulation of ITAM and ITIM receptors may influence cellular immunity initiated by dendritic cells (44).
Although studies using gene knockout mice (8, 9, 10) have clearly demonstrated that the signal transduced by FcRs containing ITAMs can be regulated by CD32B-mediated ITIM signaling, it is not clear whether such a regulation or an alternative pathway of regulation exists in human neutrophils. As described earlier, the FcRs system in human and mice are quite different, especially on neutrophils, the major inflammatory cells in the blood. In humans, the cell-cell interaction mediated by neutrophil ITAM-containing CD32A with Ab-coated target cells is a critical step in Ab-dependent cellular cytotoxicity and immunophagocytosis (16, 45, 46). Our results show that the ITAM-containing CD32A can be regulated at the ligand-binding stage. Using neutrophils from a CD16B-deficient donor, we have observed that CD32A does not contribute to IC binding in resting neutrophils, because it is in a low-avidity state (19). However, once neutrophils are activated, CD32A becomes functionally active and binds IgG-coated particles efficiently. Interestingly, the results presented in this study show that this type of CD32A regulation is activation signal specific, because the neutrophil-activating compounds fMLP and PMA have opposing effects on CD32A ligand binding. Treatment of neutrophils with fMLP up-regulated the EA binding function of CD32A, whereas PMA-mediated activation of neutrophils completely abolished the ligand-binding function of CD32A without altering CD32A levels. Moreover, PMA treatment not only abrogated CD32A-dependent EA binding, but also abolished CD16B-dependent EA binding regardless of the level of receptor expression. Activation of neutrophils with other phorbol esters including PDAc and PDBu did not affect the FcR-dependent EA binding. These observations suggest that the PMA-mediated down-modulation of FcRs ligand binding is an activation signal-specific phenomenon. In an extensive literature survey, we found that, in 1986, Wright and Meyer (47) reported a similar effect of PMA on neutrophils IC binding, but they attributed the decreased IC binding to the shedding of CD16B molecule. No analysis of the function or expression of CD32A was conducted in their study.
Studies have shown that PMA-mediated activation of neutrophils enhances ligand-binding functions of many cell surface proteins. For example, PMA activation enhanced LFA-1/ICAM-1-mediated cell adhesion (48), and the binding of complement-opsonized target cells to CD11b was up-regulated by PMA activation of neutrophils (48). These findings suggest that the PMA-mediated down-modulation of FcRs ligand binding on neutrophils is not due to global changes in the cellular membrane organization that affects the function of cell surface molecules.
The opposing effect of different neutrophil activating phorbol esters, particularly PDBu and PMA on FcR-dependent ligand-binding functions, is intriguing. An earlier study (33) and findings from this study show that PDBu, in contrast to PMA, up-regulated CD32A-dependent ligand-binding function. Although these two phorbol esters have been shown to transduce similar signals in many cell types, differences in their action on neutrophils have been reported (49, 50). It has been shown that PMA and PDBu differ in their use of cytosolic protein kinase C in generating superoxide anion in neutrophils (49). PDBu and PMA also differ in their ability to prime neutrophils for fMLP-induced superoxide anion generation (49). The difference in priming has been attributed to the specific effect of PDBu on fMLP binding to its receptor (50). Although, at present, the mechanisms underlying the opposing effects of these two phorbol esters on FcRs are not clear, these previous reports on the difference in signaling to neutrophils by PMA and PDBu suggest that the differential effect of phorbol esters have on FcRs ligand binding could be due to the difference in signaling mechanisms.
The results presented in this report show that, in humans, the strong signaling CD32A is regulated at the ligand-binding stage. Moreover, this activation-dependent functional regulation of FcRs, in particular CD32A, appears to be neutrophil specific, because CD32A expressed on K562 cells binds ligand efficiently without activation. This type of regulation of FcRs at the ligand-binding stage is distinct from CD32B-mediated regulation, which regulates the function of activating FcRs at the signaling stage. However, our studies here do not rule out the possibility that CD32A signaling can also be regulated by negative signaling receptors in human neutrophils.
Because the EA binding function of FcRs expressed on cultured cells such as K562 and CHO cell transfectants is not affected by fMLP or PMA, the signal-specific modulation of neutrophil FcRs is cell specific. Precedents for cell-specific avidity modulation of cell surface receptors by cell activation have been reported for receptors, such as integrins (51, 52). This type of cell-specific regulation of receptor affinity has been observed with a 2 integrin, the LFA-1 molecule (22, 23). The inactive LFA-1 expressed on PBL can be activated by PMA to bind ICAM-1, whereas LFA-1 expressed on COS cell transfectants can bind ICAM-1 without activation. This observation suggests that avidity modulation by cell activation is a cell-specific phenomenon. Studies on the integrins expressed on neutrophils have shown that both fMLP- and PMA-induced cell activation up-regulated the ligand-binding function of LFA-1 and CD11b molecules (48). Interestingly, as shown in our results, unlike integrins CD32A activation is signal specific, i.e., up-regulated by fMLP but down-modulated by PMA.
Although neutrophil CD32A function can be modulated by cell activation, the molecular mechanisms that convert CD32A from a low- to high-avidity state are not known. The avidity modulation of CD32A may be a physiologically important phenomenon. Many studies have shown that neutrophil CD32A plays an important role both in clearing IC from the circulation and in immunophagocytosis (53, 54, 55, 56, 57, 58). Therefore, it is possible that the aberrant regulation of CD32A, the major phagocytic FcR, may result in IC-mediated diseases (59) and poor defense against bacterial infections. For example, excessive binding of IC by strong signaling CD32A due to a high-avidity state may result in tissue injury, whereas lack of binding to IC will result in deficient phagocytosis of IgG-coated bacteria and defective clearance of IC. Thus, IC-mediated tissue injury may occur if neutrophil CD32A is constitutively active due to either an abnormality in its regulation or the persistent presence of inflammatory cytokines or other neutrophil-activating factors. Therefore, it can be hypothesized that the flare-up of IC-mediated vasculitis and the exacerbation of certain autoimmune diseases following bacterial infections (60) may be due to the high-avidity state of CD32A. Analysis of CD32A function expressed on neutrophils obtained from patients with infection-induced vasculitis will be necessary to test this hypothesis. It is also possible that the aberrant regulation of CD32A ligand-binding function in neutrophils can result in exacerbated autoimmune or IC-mediated diseases. In a preliminary study using neutrophils from five patients, we observed that CD32A expressed on neutrophils from some infectious and autoimmune disease patients showed higher CD32A-dependent EA binding than neutrophils from controls (S. Nagarajan and P. Selvaraj, unpublished observation). This observation needs to be substantiated using a large number of patients to demonstrate whether the dynamic regulation of CD32A ligand binding occurs in vivo during infections and inflammatory conditions.
Acknowledgments
We thank Dr. Peter Jensen, Dr. Aron Lukacher, and Ms. Saranya Selvaraj for their comments on this manuscript.
Disclosures
The authors have no financial conflict of interest.
Footnotes
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.
1 This work was supported by National Institutes of Health Grant R01AI49400.
2 Current address: Arkansas Children’s Nutrition Center, Department of Microbiology and Immunology, University of Arkansas for Medical Sciences, Little Rock, AR 72202.
3 Address correspondence and reprint requests to Dr. Periasamy Selvaraj, Department of Pathology and Laboratory Medicine, Room 7309, Woodruff Memorial Research Building, 101 Woodruff Circle, Emory University School of Medicine, Atlanta, GA 30322. E-mail address: pselvar@emory.edu
4 Abbreviations used in this paper: IC, immune complex; PDBu, phorbol 12,13-dibutyrate; PDAc, phorbol 12,13-diacetate; EA, rabbit Ab-coated sheep erythrocyte; MFI, mean fluorescence intensity.
Received for publication June 11, 2004. Accepted for publication February 28, 2005.
References
Ravetch, J. V., S. Bolland. 2001. IgG Fc receptors. Annu. Rev. Immunol. 19:275.
Hogarth, P. M.. 2002. Fc receptors are major mediators of antibody based inflammation in autoimmunity. Curr. Opin. Immunol. 14:798.
Takai, T.. 2002. Roles of Fc receptors in autoimmunity. Nat. Rev. Immunol. 2:590.
Unkeless, J. C., E. Scigliano, V. H. Freedman. 1988. Structure and function of human and murine receptors for IgG. Annu. Rev. Immunol. 6:251.
Kimberly, R. P., J. Wu, A. W. Gibson, K. Su, H. Qin, X. Li, J. C. Edberg. 2002. Diversity and duplicity: human Fc receptors in host defense and autoimmunity. Immunol. Res. 26:177.
Raghavan, M., P. J. Bjorkman. 1996. Fc receptors and their interactions with immunoglobulins. Annu. Rev. Cell Biol. 12:181.
Indik, Z. K., J. G. Park, S. Hunter, A. D. Schreiber. 1995. Structure/function relationships of Fc receptors in phagocytosis. Semin. Immunol. 7:45
Ravetch, J. V., R. A. Clynes. 1998. Divergent roles for Fc receptors and complement in vivo. Annu. Rev. Immunol. 16:421.
Sylvestre, D. L., J. V. Ravetch. 1994. Fc receptors initiate the Arthus reaction: redefining the inflammatory cascade. Science 265:1095.
Clynes, R., C. Dumitru, J. V. Ravetch. 1998. Uncoupling of immune complex formation and kidney damage in autoimmune glomerulonephritis. Science 279:1052.
Ravetch, J. V., A. D. Luster, R. Weinshank, J. Kochan, A. Pavlovec, D. A. Portnoy, J. Hulmes, Y. C. Pan, J. C. Unkeless. 1986. Structural heterogeneity and functional domains of murine immunoglobulin G Fc receptors. Science 234:718
Selvaraj, P., W. F. Rosse, R. Silber, T. A. Springer. 1988. The major Fc receptor in blood has a phosphatidylinositol anchor and is deficient in paroxysmal nocturnal hemoglobinuria. Nature 333:565.
Huizinga, T. W. J., C. E. Van Der Schoot, C. Jost, R. Klaassen, M. Kleijer, A. E. G. K. Von dem Borne, D. Roos, P. A. T. Tetteroo. 1988. The PI-linked receptor FcRIII is released on stimulation of neutrophils. Nature 333:667.
Wirthmueller, U., T. Kurosaki, M. S. Murakami, J. V. Ravetch. 1992. Signal transduction by FcRIII (CD16) is mediated through the chain. J. Exp. Med. 175:1381
Daeron, M.. 1997. Fc receptor biology. Annu. Rev. Immunol. 15:203.
Edberg, J. C., R. P. Kimberly. 1994. Modulation of Fc and complement receptor function by the glycosyl-phosphatidylinositol-anchored form of FcRIII. J. Immunol. 152:5826.
Zhou, M.-J., E. J. Brown. 1994. CR3 (Mac-1, m2, CD11b/CD18) and FcRIII cooperate in generation of a neutrophil respiratory burst: requirement for FcRII and tyrosine phosphorylation. J. Cell Biol. 125:1407.
Pricop, L., P. Redecha, J. L Teillaud, J. Frey, W. H. Fridman, C. Sautes-Fridman, J. E. Salmon. 2001. Differential modulation of stimulatory and inhibitory Fc receptors on human monocytes by Th1 and Th2 cytokines. J. Immunol. 166:531.(Shanmugam Nagarajan, Nimi)
FcRs with the ITIM domain have been shown to regulate the inflammatory signal delivered by the ITAM-containing FcRs. In this study, we demonstrate that the function of human neutrophil FcR type IIA (CD32A) is regulated in a distinct manner by different cell activation signals at the ligand-binding stage. Activation of neutrophils with fMLP up-regulated the ligand-binding function of CD32A, whereas PMA-mediated activation completely abolished ligand binding without altering CD32A expression. Furthermore, PMA treatment also abolished CD16B-dependent ligand binding irrespective of the level of expression. The effect of PMA was cell type specific, because the ligand-binding function of CD32A expressed on cultured cells such as K562 and CHO-CD32A transfectants was not affected by PMA. Interestingly, phorbol 12,13-dibutyrate, another phorbol ester, and IL-8 up-regulated CD32A-dependent ligand-binding function. These results demonstrate that regulation of CD32A-dependent ligand binding in human neutrophils is not only cell type specific but also activation signal specific. Moreover, these results suggest the possibility that signals delivered to neutrophils by various inflammatory stimuli can exert opposing effects on the function of human FcRs, representing a novel inside-out regulatory mechanism of FcR ligand binding.
Introduction
The FcRs expressed on inflammatory cells mediate Ab-dependent cellular cytotoxicity and the immunophagocytosis of invading pathogens. FcR-mediated cellular toxicity has been implicated in many autoimmune and immune complex (IC)4-mediated diseases (1, 2, 3). Inflammatory cells in mice and humans express multiple FcRs with overlapping ligand-binding specificity (4, 5, 6, 7). However, the physiological significance and the mechanism of regulation of these coexpressed receptors are not clearly understood. Recently, gene knockout studies in mice have shown that the coexpression of an ITIM-containing FcR, CD32B, with an ITAM-containing CD16A-/ complex results in controlled immune responses and prevents IC-mediated diseases (1, 8, 9, 10). Thus, the inhibitory signal delivered by CD32B attenuates an excessive activation signal delivered by CD16A during normal immune responses. These observations suggest that opposing signals delivered by FcRs during responses to IC and invading pathogens provide an important mechanism to regulate inflammatory signals produced by ITAM-containing FcRs and that an imbalance in activation vs inhibitory signals may result in immunopathological reactions leading to tissue injury. However, it is not known how the functions of FcRs are regulated in human neutrophils.
The FcR system in mouse and human neutrophils are quite different, and therefore studies from mice cannot be directly applied to humans. Murine neutrophils express polypeptide-anchored CD16A and CD32B (11), whereas human neutrophils express GPI-anchored CD16B and polypeptide-anchored CD32A (12, 13). The CD16A expressed on mouse neutrophils can deliver the activation signal by associating with ITAM-containing or subunits (14), and this activation can be regulated by ITIM-containing mouse CD32B (15). In contrast, human CD32A has an ITAM motif in the cytoplasmic domain and is capable of transducing inflammatory signals (16, 17). Moreover, the signal delivered by CD16B is not inhibitory to CD32A; rather, studies have shown that CD16B augments the function of CD32A (16, 17). It is possible that CD32B may be regulating inflammatory signals in human neutrophils too, but studies so far have not shown the expression of CD32B molecule on human neutrophil surface, although CD32B mRNA has been documented (18). Previous studies in our laboratory (19) and others (20, 21) have shown that, in human neutrophils, IC binding can be completely blocked by a combination of mAbs specific to CD16B and CD32A, suggesting that functionally active CD32B protein may not be expressed on neutrophils. These observations suggest that CD32A might be regulated by alternative mechanisms in humans.
Recently, we have reported that neutrophils keep their strong signaling CD32A receptor in a low-avidity state (19), but once neutrophils are activated with agents such as fMLP, a chemotactic peptide derived from bacteria, CD32A binds Ab-coated target cells with high avidity. This up-regulation of ligand-binding function by cell activation is very similar to regulation of the LFA-1 integrin molecule (22, 23). Yet, unlike the integrin system, the data presented in this report demonstrate that two well-known neutrophil activators, fMLP and PMA, regulate CD32A ligand-binding function in an opposing manner. Such a signal-specific alteration in ligand binding represents a novel inside-out regulatory mechanism that may influence the function of FcRs in a distinct manner under various inflammatory conditions.
Materials and Methods
Reagents
Rabbit anti-DNP IgG, fMLP, PMA, phorbol 12,13-dibutyrate (PDBu), phorbol 12,13-diacetate (PDAc), 4--phorbol 12-myristate 13-acetate (4-PMA), C5a, 1,10-phenanthroline, PKH-26 (a lipophilic dye), and other chemicals were from Sigma-Aldrich. Human IL-8 (R&D Systems) was a kind gift from Dr. V. Udhayakumar at Centers for Disease Control (Atlanta, GA). SRBC were purchased from Colorado Serum Company. All cell culture reagents were purchased from Invitrogen Life Technologies. FITC-conjugated F(ab')2-goat anti-mouse IgG was from Jackson ImmunoResearch Laboratories.
Cell lines and Abs
K562, CHOK1, mouse hybridomas secreting anti-CD32 (IV.3), anti-CD64 (32.2), anti-CD11b (LM2.1.6.11), anti-CD3 (OKT3) mAbs, and mouse myeloma secreting a nonbinding IgG (X63) were purchased from the American Type Culture Collection. These cells were cultured in RPMI 1640 supplemented with 10% FBS (HyClone), 2 mM glutamax I, 1 mM sodium pyruvate, penicillin (100 U/ml), and streptomycin (100 μg/ml). CHOK1-CD32A transfectants and hybridoma secreting anti-CD16 (CLBFcgran-1) were described earlier (19, 24). Anti-CD16 (CLBFcgran-1) and anti-CD32 (IV.3) IgG were purified from corresponding hybridoma culture supernatant using protein G-Sepharose. Fab of purified IgG were prepared by pepsin digestion (custom service by Lampire Biological Laboratories) and further purification using protein G-Sepharose.
Preparation of neutrophils
Neutrophils were isolated from peripheral blood of normal volunteers by dextran sedimentation and Ficoll-Paque (1.077) density gradient centrifugation as described earlier (19). Contaminating erythrocytes in the neutrophil fraction were removed by hypotonic lysis in water for 20 s. Neutrophils were resuspended in RPMI 1640/0.3% BSA (activation buffer). Because change in temperature during preparation of neutrophils has been shown to activate neutrophils (25, 26), all procedures were conducted at room temperature. Neutrophils were prepared at room temperature and used as quickly as possible. All washing media and centrifuges were also kept at room temperature to avoid temperature fluctuations during neutrophil preparations. Blood collection from normal volunteers was done according to the approved protocols by Institutional Review Board of Emory University.
Activation of neutrophils
Neutrophil activation was performed as described earlier (19). Unless indicated, activation was done by incubating neutrophils in the activation buffer with the specified activators at 37°C for 45 min. After activation, neutrophils were washed once and resuspended in cold HBSS (Ca2+/Mg2+)/0.3% BSA/5 mM EDTA (binding buffer). In some experiments, immediately after activation, neutrophils were fixed with 0.04% paraformaldehyde for 10 min at room temperature. The fixed neutrophils were washed twice in binding buffer and resuspended in cold binding buffer, until used in the ligand-binding assays.
Rabbit Ab-coated sheep erythrocyte (EA) binding assay
EA binding assays were conducted using SRBC-coated with trinitrophenol followed by opsonization with rabbit anti-DNP IgG (hereafter referred to as EA), as described earlier (24). A minimum of 200 cells was examined under light microscopy for rosetting. Neutrophils with a minimum of five EA attached were scored as a rosette. In some of the assays, EA binding was analyzed by flow cytometry using PKH-labeled EA (27). PKH labeling of EA was performed as described earlier (27, 28). The PKH-labeled EA was stored in erythrocyte storage buffer EAS45 (29). EA stored in this buffer can be used for ligand-binding experiments for up to 10 days. Neutrophils (50 μl of 5 x 106/ml) in binding buffer were incubated with PKH-labeled EA (50 μl of 2 x 108) for 2–4 h at 4°C. To determine CD32A-dependent EA binding, neutrophils were preincubated with Fab (5 μg/ml) of anti-CD16 mAb (CLBFcgran-1) for 20 min on ice (19). PKH-labeled unopsonized erythrocytes were used as a control in all the experiments. Mean fluorescence intensity (MFI) at FL2 channel and the percentage of FcR+ cells bound to EA was determined in a FACScan or FACSCalibur flow cytometer (BD Biosciences). The EA binding was represented as percentage of rosette formation or the attachment index when rosettes were counted under a microscope or by flow cytometry, respectively. The attachment index was calculated by the following formula: % cells bound to EA x mean fluorescence/100. As reported by others (27), and from the results from the studies here, we observed that the quantitative analysis of rosette formation measured by both of the methods is comparable.
Flow cytometric analysis
The cell surface expression of FcRs and the neutrophil activation marker (CD11b) were analyzed by indirect immunofluorescence flow cytometry. Cells (5 x 105) were incubated with corresponding mAbs or control mIgG (X63) followed by FITC-conjugated F(ab')2-goat anti-mouse IgG. The samples were analyzed in a FACScan flow cytometer (BD Biosciences). For quantitation of cell surface markers, standard fluorescent-labeled beads (Bangs Laboratories) were used under the same settings.
Results
Signal-specific modulation of ligand-binding function of neutrophil CD32A
In our previous studies, we demonstrated that the CD32A-mediated neutrophil ligand-binding function is up-regulated by fMLP treatment (19). In this report, we investigated whether this up-regulation was a result of a general activation-associated phenomenon or was signal specific. Neutrophils were activated with PMA and then were tested for CD32A-mediated EA binding. fMLP was included as a control for neutrophil activation in these experiments. Total binding (contributed by both CD16B and CD32A) and CD32A-dependent binding were measured. Microscopic evaluation (Fig. 1A) and quantitative measurement of ligand binding showed that the total EA binding was not altered after fMLP activation (B). As expected, fMLP-activated neutrophils showed an increase in CD32A-dependent rosette formation (Fig. 1, A and B), whereas, surprisingly, PMA activation of neutrophils resulted in complete loss of both total and CD32A-dependent rosette formation (A and B). This was not due to the loss of CD32A expression, because flow cytometry analysis showed that neutrophils retained >70–80% of their CD32A expression after PMA activation (Fig. 1C). Most of the CD16B expression was lost after activation, possibly due to the shedding of CD16B during neutrophil activation (13, 30). Therefore, the loss of total binding could be due to loss of CD16B expression, because studies from our laboratory (19) and others (20, 21) have shown that CD16B is the major receptor for IC binding to resting neutrophils. CD11b expression, a neutrophil activation marker, was up-regulated about 3-fold, confirming that the neutrophils were activated with PMA (Fig. 1C). This observation of PMA-mediated complete down-modulation of FcRs ligand-binding function was confirmed with neutrophils from >30 individual donors. Collectively, these results have indicated that the fMLP-mediated CD32A up-regulation is not donor specific, although the level of CD32A-depenendent rosette formation may vary (19). Similarly, PMA-mediated inhibition of the rosette formation was observed in all of the donors we have tested so far.
FIGURE 1. Activation signal-specific modulation of CD32A-dependent ligand-binding function in neutrophils. Neutrophils from the control donor were treated with medium or fMLP (1 μM) or PMA (50 ng/ml) in serum-free RPMI 1640 medium for 45 min at 37°C. A, Microscopic evaluation of neutrophil (N) rosette formation with EA (E) was performed using light microscope, and the pictures were taken using Leica PhotoShop software. B, Quantitative rosette formation with resting and activated neutrophils with EA were conducted as described in Materials and Methods. Total binding (left panel) and CD32A-dependent EA binding (performed in the presence of Fab of anti-CD16 IgG) (right panel) were determined as described in Materials and Methods. Experiments were done in triplicate. Rosette formation data are representative of >20 independent experiments with different donors. C, Flow cytometry analysis of CD16B, CD32A, and CD11b expression. Unactivated (medium) or activated (fMLP or PMA) neutrophils were stained with appropriate mAbs (filled histogram) and a nonbinding mIgG (open histogram). Cells were analyzed using a FACScan flow cytometer. The number indicates the mean fluorescence value.
The inactivation of FcRs by PMA does not depend on the level of CD16B expression
Next, we determined whether shedding of CD16B was responsible for the loss of EA binding in PMA-treated neutrophils. It has been reported that the partial loss of CD16B expression due to shedding is caused by metalloproteinases and can be prevented by metalloproteinase inhibitor 1,10-phenanthroline (31, 32). To investigate this possibility, neutrophils were preincubated with 1,10-phenanthroline before activation with fMLP or PMA. Neutrophils activated with fMLP or PMA showed a reduced level of CD16B expression (only 30% that of unactivated neutrophils) (Fig 2A) without a change in CD32A expression (data not shown). As reported earlier (31, 32), pretreatment of neutrophils with 1,10-phenanthroline restored the level of CD16B expression (80% of the unactivated neutrophils). Yet, despite having near-normal levels of CD16B expression after 1,10-phenanthroline pretreatment, PMA-activated neutrophils did not bind to EA (Fig. 2B). Interestingly, even though the levels of CD16B expression were the same after PMA or fMLP activation, only fMLP-activated neutrophils retained their ability to bind to EA. 1,10-Phenonthroline pretreatment alone did not have any effect on the EA binding of FcRs in neutrophils (data not shown). These findings suggest that down-modulation of FcR-mediated ligand-binding functions of neutrophils by PMA is not due to the levels of CD16B or CD32A expression.
FIGURE 2. CD16B density does not influence the ligand binding on PMA-activated neutrophils. Neutrophils were incubated with or without 1,10-phenanthroline (5 mM) in serum-free RPMI 1640 (Medium) for 30 min at 37°C. Untreated (– phenon) and 1,10-phenanthroline-treated (+ phenon) neutrophils were further incubated with fMLP (1 μM) or PMA (50 ng/ml) in serum-free RPMI 1640 medium for 45 min at 37°C. A, Cell surface expression of CD16B was analyzed by flow cytometry using specific mAb or nonbinding mIgG (solid-line open histogram). Closed and dotted line histograms represent cells treated without and with 1,10-phenanthroline, respectively. B, Rosette formation of untreated and 1,10-phenanthroline-treated resting and activated neutrophils with EA were carried by flow cytometry-based rosetting assay, as described in Materials and Methods. A representative of two independent experiments from two donors is presented.
Kinetics of modulation of FcRs ligand binding by PMA and fMLP treatment
Next, the kinetics of the effect of PMA and fMLP treatment on total and CD32A-dependent ligand binding was determined. As shown in Fig. 3A, the total EA binding of PMA-activated neutrophils was only 28% at 20 min compared with 70% in resting neutrophils, and the EA binding was completely abolished at 40 min. However, the CD32A-dependent rosette formation in fMLP-activated neutrophils increased as early as 10 min and was at a maximum at 40 min (Fig. 3B). To determine the concentration of PMA that is necessary for PMA-mediated down-modulation of FcRs, the activation experiment was repeated using neutrophils from another donor. Neutrophils were treated with different concentration of PMA. The down-modulation of total CD32A-dependent EA binding to PMA-activated neutrophils was also dose dependent. Nearly 70–80% inhibition of rosette formation was observed at a concentration as low as 1–2 ng/ml PMA, and complete loss of rosette formation was seen at 10 ng/ml PMA (Fig. 3C). However, neutrophils activated with fMLP showed an increase in CD32A-dependent EA binding (Fig. 3D).
FIGURE 3. Kinetics and dose dependency of PMA-mediated down-modulation of EA rosette formation. A and B, Kinetics of down-modulation: Neutrophils (1 x 107) isolated from the control donor were treated without or with 1 μM fMLP () or 50 ng/ml PMA () in serum-free RPMI 1640 medium for various times at 37°C. After activation, neutrophils were fixed with RPMI 1640/10% FBS/0.04% paraformaldehyde for 10 min on ice. Total EA rosette formation (A) with neutrophils was done as previously described. To determine CD32A-dependent rosette formation (B), rosetting was performed in the presence of 2 μg/ml Fab of anti-CD16 IgG (CLB). C and D, Dose dependency of down-modulation: Neutrophils were activated with different concentrations of PMA (C) or fMLP (D) as described above. Total and CD32A-dependent EA rosette formation was analyzed as described in Materials and Methods. The figure is a representative of two independent experiments from two different donors.
Effect of C5a- and IL-8-mediated neutrophil activation on FcRs ligand binding
Because neutrophil activation with fMLP or PMA showed opposing effects on the ligand-binding functions of FcRs, we investigated whether neutrophil activation by inflammatory mediators such as C5a and IL-8 would also modulate the FcRs function. Neutrophils from different donor were treated with C5a or IL-8 at the indicated concentration, and expression of FcRs and CD11b was analyzed by flow cytometry. As a control, cells were also treated with PMA and fMLP. The level of CD16B was decreased upon activation, whereas the level of CD32A was not altered (Fig. 4A). Neutrophil activation marker CD11b was increased 2- to 3-fold, confirming the activated state of neutrophils (Fig. 4A). The EA binding to activated neutrophils was determined by flow cytometry. As shown in Fig. 4B, the total EA binding was not altered when neutrophils were activated with fMLP, IL-8, or C5a, whereas PMA activation completely abolished it. Neutrophils treated with fMLP and IL-8 showed 3- and 1.5-fold increases in CD32A-dependent EA binding, respectively, whereas C5a had no effect or moderate increase in CD32A-dependent binding (Fig. 4C). These results demonstrate that different neutrophil activators distinctly modulate the ligand-binding function of FcR.
FIGURE 4. Effect of neutrophil activators on FcR-mediated EA binding. Neutrophils were incubated with fMLP (1 μM) or PMA (2 ng/ml) or C5a (10 ng/ml) or IL-8 (10 ng/ml) for 45 min at 37°C. Cells treated with medium at 4°C were used as resting neutrophils. A, Flow-cytometric analysis of resting and activated neutrophils was performed as described in Materials and Methods using appropriate mAbs. B and C, Total (B) and CD32A-dependent EA (C) binding was conducted as described in Materials and Methods. The EA binding was analyzed by in a FACSCalibur flow cytometer. The attachment index was quantitated as described in Materials and Methods. Data from one of the two donors are presented.
Specificity of PMA-mediated loss in FcRs ligand-binding function
PMA-mediated inactivation of FcRs in this report contradicts a previous report showing that PDBu, another phorbol ester (33), activated the ligand-binding function of CD32A. This difference could be due to the use of different phorbol esters. To resolve this discrepancy, neutrophils (from another donor) were activated with PMA or PDBu, and EA binding was determined. Both PMA and PDBu showed 3-fold increase in CD11b expression, indicating that neutrophils were activated by both phorbol esters (Fig. 5A). Compared with the expression of unactivated neutrophils, CD16B expression of PMA- or PDBu-activated neutrophils were 34 and 68%, respectively (Fig. 5A). Similar results were observed with neutrophils obtained from other donors. The FcR-dependent EA binding was determined by flow cytometry as described in Materials and Methods. PMA-activated neutrophils did not bind to EA; however, PDBu-activated neutrophils bound to EA (Fig. 5B). In a separate experiment using neutrophils from another donor, we also determined the effect of other phorbol esters such as PDAc and 4-PMA (inactive analog of PMA) on CD32A ligand binding in neutrophils. As shown in Fig. 5C, PMA-activated neutrophils showed minimal EA binding, whereas PDAc and PDBu did not affect the EA binding. Furthermore, neutrophils treated with 4-PMA, an inactive analog of PMA, did not affect EA binding to neutrophils, suggesting that the PMA-mediated down-modulation of CD32A function is specific. Earlier studies have used higher concentrations of PDBu (33), thus raising the possibility that the difference observed in EA ligand binding in this study with PDBu or PMA was due to the difference in the concentration of phorbol esters used in both studies. To test this issue, neutrophils were activated with different doses of PMA or PDBu, and EA binding was determined. PMA-mediated inactivation of FcR-dependent EA binding was seen at a concentration as low as 1 ng/ml PMA. Yet, even at a higher concentration (100 ng/ml) of PDBu, EA binding was not altered (Fig. 5D). These findings indicate that the difference between the present observation and the earlier report on the effect of phorbol esters on FcR-dependent EA binding may be due to the difference in phorbol esters used for neutrophil activation.
FIGURE 5. Effect of different phorbol esters on neutrophil FcR-mediated EA binding. Neutrophils were incubated with fMLP (1 μM) or PMA (1 ng/ml) or other phorbol esters (2 ng/ml) as indicated in the figure. Cells incubated in medium at 4°C represent resting neutrophils. A, Expression of CD16B, CD32A, and CD11b on resting (medium) and activated neutrophils were analyzed by flow cytometry, and the mean fluorescent values are represented in the figure. B, Total EA rosette formation of resting and activated neutrophils was analyzed by flow cytometry. The attachment index was calculated as described in Materials and Methods. C, Effect of different phorbol esters on EA rosette formation. Cells incubated with PMA or other phorbol esters (PDAc, PDBu) or inactive analog 4-PMA were used in this experiment. EA rosette formation was determined by flow cytometry. M1 represents the histogram peak for rosetted cells. D, Effect of different concentrations of PMA and PDBu. Cells were treated with PMA or PDBu at indicated concentration. Flow-cytometric analysis of EA rosette formation of resting (medium) and activated neutrophils was performed as described in Materials and Methods. Representative of three independent experiments from three different donors is given. The numbers given in the figure represent MFI/percentage of neutrophils formed rosettes, respectively.
PMA-mediated FcRs inactivation is cell type specific
In our earlier studies, we observed that 53% of CHOK1 cell transfectants expressing CD32A (CHO-CD32A) formed rosettes with EA, and this binding was completely blocked by anti-CD32A mAb, IV.3 (19). The anti-CD32 mAb has shown to be specific for CD32A receptor (34). Similarly, CD32A expressed on K562, a human erythroleukemia cell line that naturally expresses CD32A, is able to bind EA efficiently, suggesting that CD32A expressed in this cell line is constitutively active (Fig. 6A). These results suggest that CD32A expressed on cultured cells such as K562 and CHO-CD32A cells, unlike neutrophils, is constitutively active. We tested whether the PMA-mediated inactivation of CD32A was also cell type specific. EA binding to both CHOK1-CD32A and K562 cells was determined after pretreating the cells with PMA. Incubation of CHO-CD32A and K562 cells with PMA before rosetting did not influence the EA rosette formation (Fig. 6A). These findings indicate that CD32A is fully functional in CHO cells, and the PMA-mediated inactivation of CD32A is cell type specific. Similarly, incubation of CHO-CD16B transfectants with PMA before rosetting did not abolish EA rosette formation (Fig. 6B), suggesting both FcRs expressed in CHO cells are not susceptible to inactivation by PMA.
FIGURE 6. PMA-mediated CD32A inactivation is cell type specific. A, CHO cell transfectants expressing CD32A, and K562 cells were incubated with medium () or PMA (5 ng/ml) () for 1 h at 37°C and allowed to form rosette with EA in ice for 2 h. Fab of IV.3 (5 μg/ml) completely blocked this binding (data not shown). B, Rosetting of CHO cell transfectants expressing CD16B was conducted as described above. The results are expressed as the mean of three experiments. mAbs specific to FcRs completely blocked this EA binding. Untransfected CHO cells did not form EA rosettes (data not shown).
It is possible that the difference in EA binding between cultured cells and unactivated neutrophils may be due to the difference in CD32A receptor density on the cell surface. To test this possibility, we measured the level of CD32A expression in CHO-CD32A transfectants and K562 cells. CD32A expression was analyzed by immunofluorescence flow-cytometric analysis, and the level of CD32A expression (MFI) was determined using standard beads run under identical instrument settings. The ligand-binding function was determined by EA rosette formation. As shown in Table I, CD32A expression was about the same in K562 but 10-fold higher in CHO-CD32A cells, compared with neutrophils. However, the surface area of neutrophils (300 μm2) was smaller than that of CHOK1 and K562 cells (591 and 584 μm2, respectively) (35, 36, 37). When the cell surface density of CD32A was normalized for its surface area, neutrophils expressed 1.6-fold higher levels of CD32A than K562 cells. Interestingly, CD32A-dependent EA binding to neutrophils was only 5% compared with 47.5% in K562. The EA binding was completely blocked by Fab of anti-CD32A mAb (data not shown). These results show that the higher level of EA binding mediated by CD32A expressed on K562 cell line is not due to higher receptor density, and suggest that these cells constitutively express a high-avidity form of CD32A. CHO-CD32A cell transfectants also formed 45% rosettes with EA. The surface density of CD32A expressed on CHO cells, however, is 6-fold higher than the neutrophils, and therefore, at present we cannot rule out the role of surface density in CHO-CD32A cell transfectants binding to EA.
Table I. Efficient CD32A-mediated EA binding by K562 cells is not due to higher receptor densitya
Discussion
Receptors involved in cell-cell interactions carry out two vital functions that are necessary for productive cell communication. First, they bind to counter receptors/ligands on the opposing cell, and second, after binding, they deliver activating or inhibitory signals inside the cell. Therefore, the regulation of these receptors at any one of their functions is critical for normal immune response. Abnormal regulation of these receptors has been found in many disease states, including fatal autoimmune disorders, such as lupus and arthritis (1, 38, 39, 40). FcRs are not only involved in binding of IgG-coated target cells, but also signal for subsequent effector functions of the cell. Many of the FcRs contain an ITAM in their cytoplasmic domain or associate with ITAM-containing subunits that transduce activation signals to cells. Interestingly, CD32B expressed on inflammatory cells has an ITIM, which delivers inhibitory signals to the cells. Studies using knockout mouse models suggest that the signals transduced by ITIM can regulate the activation signal delivered by ITAM-containing FcRs (41, 42, 43). Thus, coexpression of ITAM- and ITIM-containing receptors play an important role in regulating immune responses during humoral immunity. Recently, Ravetch and coworker (44) have shown that the signal delivered through ITAM and ITIM motifs may influence IC-induced maturation of dendritic cells, suggesting that the cross-regulation of ITAM and ITIM receptors may influence cellular immunity initiated by dendritic cells (44).
Although studies using gene knockout mice (8, 9, 10) have clearly demonstrated that the signal transduced by FcRs containing ITAMs can be regulated by CD32B-mediated ITIM signaling, it is not clear whether such a regulation or an alternative pathway of regulation exists in human neutrophils. As described earlier, the FcRs system in human and mice are quite different, especially on neutrophils, the major inflammatory cells in the blood. In humans, the cell-cell interaction mediated by neutrophil ITAM-containing CD32A with Ab-coated target cells is a critical step in Ab-dependent cellular cytotoxicity and immunophagocytosis (16, 45, 46). Our results show that the ITAM-containing CD32A can be regulated at the ligand-binding stage. Using neutrophils from a CD16B-deficient donor, we have observed that CD32A does not contribute to IC binding in resting neutrophils, because it is in a low-avidity state (19). However, once neutrophils are activated, CD32A becomes functionally active and binds IgG-coated particles efficiently. Interestingly, the results presented in this study show that this type of CD32A regulation is activation signal specific, because the neutrophil-activating compounds fMLP and PMA have opposing effects on CD32A ligand binding. Treatment of neutrophils with fMLP up-regulated the EA binding function of CD32A, whereas PMA-mediated activation of neutrophils completely abolished the ligand-binding function of CD32A without altering CD32A levels. Moreover, PMA treatment not only abrogated CD32A-dependent EA binding, but also abolished CD16B-dependent EA binding regardless of the level of receptor expression. Activation of neutrophils with other phorbol esters including PDAc and PDBu did not affect the FcR-dependent EA binding. These observations suggest that the PMA-mediated down-modulation of FcRs ligand binding is an activation signal-specific phenomenon. In an extensive literature survey, we found that, in 1986, Wright and Meyer (47) reported a similar effect of PMA on neutrophils IC binding, but they attributed the decreased IC binding to the shedding of CD16B molecule. No analysis of the function or expression of CD32A was conducted in their study.
Studies have shown that PMA-mediated activation of neutrophils enhances ligand-binding functions of many cell surface proteins. For example, PMA activation enhanced LFA-1/ICAM-1-mediated cell adhesion (48), and the binding of complement-opsonized target cells to CD11b was up-regulated by PMA activation of neutrophils (48). These findings suggest that the PMA-mediated down-modulation of FcRs ligand binding on neutrophils is not due to global changes in the cellular membrane organization that affects the function of cell surface molecules.
The opposing effect of different neutrophil activating phorbol esters, particularly PDBu and PMA on FcR-dependent ligand-binding functions, is intriguing. An earlier study (33) and findings from this study show that PDBu, in contrast to PMA, up-regulated CD32A-dependent ligand-binding function. Although these two phorbol esters have been shown to transduce similar signals in many cell types, differences in their action on neutrophils have been reported (49, 50). It has been shown that PMA and PDBu differ in their use of cytosolic protein kinase C in generating superoxide anion in neutrophils (49). PDBu and PMA also differ in their ability to prime neutrophils for fMLP-induced superoxide anion generation (49). The difference in priming has been attributed to the specific effect of PDBu on fMLP binding to its receptor (50). Although, at present, the mechanisms underlying the opposing effects of these two phorbol esters on FcRs are not clear, these previous reports on the difference in signaling to neutrophils by PMA and PDBu suggest that the differential effect of phorbol esters have on FcRs ligand binding could be due to the difference in signaling mechanisms.
The results presented in this report show that, in humans, the strong signaling CD32A is regulated at the ligand-binding stage. Moreover, this activation-dependent functional regulation of FcRs, in particular CD32A, appears to be neutrophil specific, because CD32A expressed on K562 cells binds ligand efficiently without activation. This type of regulation of FcRs at the ligand-binding stage is distinct from CD32B-mediated regulation, which regulates the function of activating FcRs at the signaling stage. However, our studies here do not rule out the possibility that CD32A signaling can also be regulated by negative signaling receptors in human neutrophils.
Because the EA binding function of FcRs expressed on cultured cells such as K562 and CHO cell transfectants is not affected by fMLP or PMA, the signal-specific modulation of neutrophil FcRs is cell specific. Precedents for cell-specific avidity modulation of cell surface receptors by cell activation have been reported for receptors, such as integrins (51, 52). This type of cell-specific regulation of receptor affinity has been observed with a 2 integrin, the LFA-1 molecule (22, 23). The inactive LFA-1 expressed on PBL can be activated by PMA to bind ICAM-1, whereas LFA-1 expressed on COS cell transfectants can bind ICAM-1 without activation. This observation suggests that avidity modulation by cell activation is a cell-specific phenomenon. Studies on the integrins expressed on neutrophils have shown that both fMLP- and PMA-induced cell activation up-regulated the ligand-binding function of LFA-1 and CD11b molecules (48). Interestingly, as shown in our results, unlike integrins CD32A activation is signal specific, i.e., up-regulated by fMLP but down-modulated by PMA.
Although neutrophil CD32A function can be modulated by cell activation, the molecular mechanisms that convert CD32A from a low- to high-avidity state are not known. The avidity modulation of CD32A may be a physiologically important phenomenon. Many studies have shown that neutrophil CD32A plays an important role both in clearing IC from the circulation and in immunophagocytosis (53, 54, 55, 56, 57, 58). Therefore, it is possible that the aberrant regulation of CD32A, the major phagocytic FcR, may result in IC-mediated diseases (59) and poor defense against bacterial infections. For example, excessive binding of IC by strong signaling CD32A due to a high-avidity state may result in tissue injury, whereas lack of binding to IC will result in deficient phagocytosis of IgG-coated bacteria and defective clearance of IC. Thus, IC-mediated tissue injury may occur if neutrophil CD32A is constitutively active due to either an abnormality in its regulation or the persistent presence of inflammatory cytokines or other neutrophil-activating factors. Therefore, it can be hypothesized that the flare-up of IC-mediated vasculitis and the exacerbation of certain autoimmune diseases following bacterial infections (60) may be due to the high-avidity state of CD32A. Analysis of CD32A function expressed on neutrophils obtained from patients with infection-induced vasculitis will be necessary to test this hypothesis. It is also possible that the aberrant regulation of CD32A ligand-binding function in neutrophils can result in exacerbated autoimmune or IC-mediated diseases. In a preliminary study using neutrophils from five patients, we observed that CD32A expressed on neutrophils from some infectious and autoimmune disease patients showed higher CD32A-dependent EA binding than neutrophils from controls (S. Nagarajan and P. Selvaraj, unpublished observation). This observation needs to be substantiated using a large number of patients to demonstrate whether the dynamic regulation of CD32A ligand binding occurs in vivo during infections and inflammatory conditions.
Acknowledgments
We thank Dr. Peter Jensen, Dr. Aron Lukacher, and Ms. Saranya Selvaraj for their comments on this manuscript.
Disclosures
The authors have no financial conflict of interest.
Footnotes
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.
1 This work was supported by National Institutes of Health Grant R01AI49400.
2 Current address: Arkansas Children’s Nutrition Center, Department of Microbiology and Immunology, University of Arkansas for Medical Sciences, Little Rock, AR 72202.
3 Address correspondence and reprint requests to Dr. Periasamy Selvaraj, Department of Pathology and Laboratory Medicine, Room 7309, Woodruff Memorial Research Building, 101 Woodruff Circle, Emory University School of Medicine, Atlanta, GA 30322. E-mail address: pselvar@emory.edu
4 Abbreviations used in this paper: IC, immune complex; PDBu, phorbol 12,13-dibutyrate; PDAc, phorbol 12,13-diacetate; EA, rabbit Ab-coated sheep erythrocyte; MFI, mean fluorescence intensity.
Received for publication June 11, 2004. Accepted for publication February 28, 2005.
References
Ravetch, J. V., S. Bolland. 2001. IgG Fc receptors. Annu. Rev. Immunol. 19:275.
Hogarth, P. M.. 2002. Fc receptors are major mediators of antibody based inflammation in autoimmunity. Curr. Opin. Immunol. 14:798.
Takai, T.. 2002. Roles of Fc receptors in autoimmunity. Nat. Rev. Immunol. 2:590.
Unkeless, J. C., E. Scigliano, V. H. Freedman. 1988. Structure and function of human and murine receptors for IgG. Annu. Rev. Immunol. 6:251.
Kimberly, R. P., J. Wu, A. W. Gibson, K. Su, H. Qin, X. Li, J. C. Edberg. 2002. Diversity and duplicity: human Fc receptors in host defense and autoimmunity. Immunol. Res. 26:177.
Raghavan, M., P. J. Bjorkman. 1996. Fc receptors and their interactions with immunoglobulins. Annu. Rev. Cell Biol. 12:181.
Indik, Z. K., J. G. Park, S. Hunter, A. D. Schreiber. 1995. Structure/function relationships of Fc receptors in phagocytosis. Semin. Immunol. 7:45
Ravetch, J. V., R. A. Clynes. 1998. Divergent roles for Fc receptors and complement in vivo. Annu. Rev. Immunol. 16:421.
Sylvestre, D. L., J. V. Ravetch. 1994. Fc receptors initiate the Arthus reaction: redefining the inflammatory cascade. Science 265:1095.
Clynes, R., C. Dumitru, J. V. Ravetch. 1998. Uncoupling of immune complex formation and kidney damage in autoimmune glomerulonephritis. Science 279:1052.
Ravetch, J. V., A. D. Luster, R. Weinshank, J. Kochan, A. Pavlovec, D. A. Portnoy, J. Hulmes, Y. C. Pan, J. C. Unkeless. 1986. Structural heterogeneity and functional domains of murine immunoglobulin G Fc receptors. Science 234:718
Selvaraj, P., W. F. Rosse, R. Silber, T. A. Springer. 1988. The major Fc receptor in blood has a phosphatidylinositol anchor and is deficient in paroxysmal nocturnal hemoglobinuria. Nature 333:565.
Huizinga, T. W. J., C. E. Van Der Schoot, C. Jost, R. Klaassen, M. Kleijer, A. E. G. K. Von dem Borne, D. Roos, P. A. T. Tetteroo. 1988. The PI-linked receptor FcRIII is released on stimulation of neutrophils. Nature 333:667.
Wirthmueller, U., T. Kurosaki, M. S. Murakami, J. V. Ravetch. 1992. Signal transduction by FcRIII (CD16) is mediated through the chain. J. Exp. Med. 175:1381
Daeron, M.. 1997. Fc receptor biology. Annu. Rev. Immunol. 15:203.
Edberg, J. C., R. P. Kimberly. 1994. Modulation of Fc and complement receptor function by the glycosyl-phosphatidylinositol-anchored form of FcRIII. J. Immunol. 152:5826.
Zhou, M.-J., E. J. Brown. 1994. CR3 (Mac-1, m2, CD11b/CD18) and FcRIII cooperate in generation of a neutrophil respiratory burst: requirement for FcRII and tyrosine phosphorylation. J. Cell Biol. 125:1407.
Pricop, L., P. Redecha, J. L Teillaud, J. Frey, W. H. Fridman, C. Sautes-Fridman, J. E. Salmon. 2001. Differential modulation of stimulatory and inhibitory Fc receptors on human monocytes by Th1 and Th2 cytokines. J. Immunol. 166:531.(Shanmugam Nagarajan, Nimi)