Differential Effect of Cytokine Treatment on Fcα Receptor I- and Fcγ Receptor I-Mediated Tumor Cytotoxicity by Monocyte-Derived Macrophages

Macrophages represent an important effector cell for Ab-mediated tumor therapy. Previous studies have documented that cytokines can influence Fc receptor (FcR) expression and function. In this study we examined the tumoricidal activities of the type I receptors for IgG (FcγRI) and the IgA FcR (FcαRI) on monocyte-derived macrophages (MDM) cultured in the presence of IFN-γ, M-CSF, or GM-CSF. Bispecific Abs were used to target a Her2/neu breast carcinoma cell line, SKBR-3, to FcαRI or FcγRI on MDM. Although FcαRI and FcγRI share a common signaling pathway contingent on association with the γ-chain (FcRγ subunit), a marked difference in their efficiency in mediating tumoricidal functions was seen in response to specific cytokines. M-CSF- and GM-CSF-treated MDM mediated efficient phagocytosis of SKBR-3 cells by FcαRI and FcγRI; however, IFN-γ-treated MDM phagocytosed tumor cells only with the FcγRI-directed bispecific Abs. Similarly, IFN-γ-cultured MDM lysed tumor cells more efficiently via FcγRI then by FcαRI as measured in Ab-dependent cellular cytotoxicity assays. Conversely, GM-CSF-treated MDM mediated more efficient lysis of tumor cells via FcαRI than FcγRI, while M-CSF-cultured MDM were relatively less efficient in mediating Ab-dependent cellular cytotoxicity through either receptor. With the exception of IFN-γ-mediated enhancement of FcγRI expression and FcγRI γ-chain complexes, the regulation of FcγRI- or FcαRI-mediated activity occurred without significant change in either receptor expression or total complexes with γ subunit. These data suggest that the efficiency of Ab-mediated tumor therapy, which depends on FcR effector cell functions, may be modified by the use of specific cytokines.

M acrophages are a critical part of the immune system, participating in both natural and specific immunity. Macrophages can function as effector cells to eliminate pathogens and as accessory cells that recruit and activate other immune cells. Furthermore, macrophages have been implicated to be critical in mediating Ab-dependent tumor regression in some animal models (1,2). Many of these effector and accessory functions are mediated through Fc receptors (FcR) (2) that bind Agcomplexed Igs. Human macrophages constitutively express the Fc receptor for IgA (Fc␣RI), and the high and low affinity Fc receptors specific for IgG (Fc␥RI, Fc␥RIIA, and Fc␥RIIIA). These FcR mediate effector functions that are well documented, including Abdependent cellular cytotoxicity (ADCC) 2 and phagocytosis (3)(4)(5)(6)(7).
In contrast, polymorphonuclear cells (PMN), an other myeloid effector cell with phagocytic and cytolytic capacity, constitutively express Fc␣RI, but not Fc␥RI. However, Fc␥RI expression can be induced in vitro with IFN-␥ treatment or in vivo with either G-CSF or IFN-␥ (8 -10). The relative importance of these two FcR types in promoting cytotoxic effector functions of various myeloid pop-ulations (monocytes, macrophages, and PMN) has not been clearly delineated.
Fc␥RI (CD64) binds monomeric human IgG1 and IgG3 with high affinity (3,4), and Fc␥RIIIA (CD16) has intermediate affinity for monomeric IgG. Fc␥RIIA (CD32) efficiently binds to IgG immune complexes and IgG-opsonized particles, but not to monomeric IgG (3). On macrophages, a single class of IgA Fc receptor, Fc␣RI (CD89), has been characterized and binds both Ag-complexed and monomeric IgA1 and IgA2 (5,11). This suggests that in vivo Fc␣RI may be saturated with monomeric IgA in the same manner as Fc␥RI and Fc␥RIIIA are significantly occupied with IgG.
mAb have been developed that bind to Fc␥RI (mAb 32.2 and 22) and Fc␣RI (mAb A77) at sites distinct from their ligand binding domains (11)(12)(13). ADCC mediated by bispecific Abs (BsAb) prepared using these anti-FcR Abs is not blocked by human IgG or IgA (6, 7, 14 -16). In addition to triggering effector functions under physiologically relevant conditions, BsAb made from these mAbs bind exclusively to the targeted FcR and provide a suitable method to study the capacity of individual FcR in a variety of functional assays. Both Fc␥RI and Fc␣RI have been shown to be functionally associated with the ␥ subunit, which mediates signaling events following receptor clustering (17,18). Although these receptors share a common signaling component (the ␥ subunit), their expression on the surface of effector cells is differentially regulated. Fc␣RI expression on monocytes can be enhanced by TNF-␣ (19,20), IL-1␤, GM-CSF, and bacterial LPS (20), whereas TGF-␤1 has been shown to decrease Fc␣RI expression (21). Monocyte Fc␥RI expression is up-regulated with IFN-␥ and IL-10, and can be down-regulated with IL-4 (8,9,22,23).
Previous studies have demonstrated that treatment of monocytederived macrophages (MDM) with M-CSF or IFN-␥ differentially regulates FcR-mediated phagocytosis and lysis of tumor cells (22,24,25). These data showed that that M-CSF-cultured MDM were proficient in mediating phagocytosis of tumor cells via Fc␥RII and Fc␥RIII, whereas cells incubated with IFN-␥ were ineffective at mediating phagocytosis of tumor cells via these two low affinity IgG receptors. The reverse has been demonstrated with regard to ADCC. MDM propagated in IFN-␥ appear to be more efficient at mediating Fc␥RII-and Fc␥RIII-dependent tumor cell lysis than untreated or M-CSF-cultured MDM (22,24). However, the Fc␥RImediated effector functions of MDM (ADCC and phagocytosis) can be enhanced with IFN-␥ (15,26).
The cytotoxic capacity of the macrophage Fc␣RI has not been fully explored. Recently, we and Valerius et al. reported that Fc␣RI on monocytes, PMN, and macrophages is a potent trigger molecule for ADCC and phagocytosis of tumor cells (6,7). Because of the role of IgA and Fc␣RI in mucosal immunity, the potential for this receptor to mediate anti-tumor activity through systemic treatment has been largely overlooked. In this study we report for the first time the ability of MDM to mediate ADCC via the Fc␣R. In addition, we have investigated the impact of several immune-modulating cytokines on the ability of MDM to kill tumor cells by phagocytosis or ADCC through Fc␥RI and Fc␣RI.

Bispecific Abs
The preparation of the BsAb A77 ϫ 520C9 and H22 ϫ 520C9 has previously been described in detail (7,15). All the BsAb preparations were tested for endotoxin contamination by the chromogenic Limulus amebocyte lysate assay (BioWhittaker, Walkersville, MD) and were found to be free of endotoxin contamination (assay sensitivity limit, 0.1 EU/ml).

Phagocytosis assay
BsAb-mediated phagocytosis of SKBR-3 cells by (MDM) was examined by a modification of the method described by Munn and Cheung (27). Briefly, monocytes, purified from normal adult source Leukopacs (Advanced Biotechnologies, Columbia, MD), were differentiated in 24-well plates in macrophage serum-free medium (Life Technologies, Grand Island, NY) supplemented with 10% FBS and either M-CSF (2 ng/ml; R&D Systems, Minneapolis, MN), GM-CSF (10 ng/ml; R&D Systems), or IFN-␥ (1000 U/ml; Genzyme, Cambridge, MA) for 5-10 days. The cultures were given fresh medium containing the appropriate cytokines every 3-4 days. The cytokine-containing medium was removed before the phagocytosis assay. SKBR-3 cells were labeled with the lipophilic red fluorescent dye, PKH 26 (Sigma, St. Louis, MO). The labeled SKBR-3 cells were added to the wells containing MDM in the absence or the presence of BsAb and incubated at 37°C overnight (unless otherwise stated). MDM and nonphagocytized SKBR-3 cells were recovered with trypsin (0.025% for 30 -45 min at 37°C), transferred to FACS tubes, and stained with a cocktail of FITC-labeled anti-CD14 mAb (AML-2-23; 10 g/ml) and anti-CD33-FITC (10 g/ml) for 1 h on ice. Cells were washed and analyzed by two-color fluorescence using the FACScan. The percent phagocytosis was calculated as the number of dual-positive target cells (ingested by MDM) divided by the total number of target cells. Assays were performed in duplicate.

Confocal imaging
After fixation and flow cytometric analysis of the phagocytosis samples, dual-positive cells were sorted with a FACStar Plus flow cytometer (Becton Dickinson, Mountain View, CA) and examined with a Bio-Rad MRC1024 laser scanning confocal microscope (Hercules, CA). Cells were scanned for fluorescence using the 488-nm line from a 15-mW KR/AR laser and two photodetectors (522/32 nm dichroic for FITC fluorescence and 585 nm longpass for PKH-26 fluorescence). A ϫ63 Plan-Apo 1.4 NA objective (Carl Zeiss, Thornwood, NY) was used in conjunction with an iris setting of 2.5, which allowed for detection of optical sections of the fluorescence image that were ϳ1.5 m thick. A minimum of 100 cells/sample were examined for quantitative evaluation of phagocytosis.

Tumor cell survival assay
To determine the tumor cell survival following phagocytosis, samples were obtained from the trypsin-harvested cells and plated in 96-well tissue culture plates. Live tumor cells were allowed to adhere overnight in RPMI containing 10% FBS. The medium was removed from each well, and the cells were fixed with 0.25% glutaraldehyde. Plates were blocked with 5% BSA solution, then reacted with mAb 251.03, which binds HER2/neu at a different site than 520C9 mAb. The 251.03 mAb was detected with goat anti-murine IgG Fc-specific alkaline phosphatase probe (The Jackson Laboratory, Bar Harbor, ME). The plates were developed using p-nitrophenyl phosphate and read at a wavelength of 405-650 nm. The relative number of SKBR-3 cells was calculated from the formula: % cell survival ϭ [(sample OD Ϫ MDM only OD)/SKBR-3 only OD Ϫ MDM only OD)] ϫ 100%. Samples were analyzed in duplicate.

Ab-dependent cellular cytotoxicity
SKBR-3 cells were cultured as described above and used as targets for lysis by MDM. Targets were labeled with 100 Ci of 51 Cr for 1-2 h before combining with effector cells and BsAb in a U-bottom microtiter plate. After incubation for 16 -18 h at 37°C supernatants were collected and analyzed for radioactivity. Cytotoxicity was calculated by the formula: % lysis ϭ (experimental cpm Ϫ target leak cpm)/(detergent lysis cpm Ϫ target leak cpm) ϫ 100%. Specific lysis ϭ % lysis with BsAb Ϫ % lysis without BsAb. Assays were performed in triplicate.

FcR expression
Monocytes were differentiated in 24-well or 175-cm 2 plates in the presence or the absence of IFN-␥, M-CSF, or GM-CSF as described for the phagocytosis assays. The cells were harvested with trypsin and incubated at 4°C for 60 min with 10 g/ml of IV.3, 32.2, A77, or 251 mAbs to stain for CD32, CD64, CD89, and CD33, respectively. After incubation with antimurine IgG-FITC probe, cells were washed and fixed in 1% paraformaldehyde, and their fluorescence was analyzed by FACScan.

␥ Subunit association with Fc␥RI and Fc␣RI
Monocytes from three donors were prepared and differentiated into macrophages exactly as described for the phagocytosis experiments, except they were cultured in 175-cm 2 tissue culture flasks. Cells were harvested by scraping using cold 0.02% EDTA solution, then were washed with cold PBS. A portion of these cells was analyzed by flow cytometry. Macrophages were adjusted to 10 7 /ml of lysis buffer (1% digitonin, 20 nM triethanolamine, 150 nM NaCl, 0.12% Triton X-100, 2 mM PMSF, and 0.5 trypsin-inhibitory unit/ml aprotinin) and incubated for 1 h on ice. The lysates were centrifuged at 15,000 ϫ g for 30 min at 4°C, and the supernatants were adjusted to 1.0 mg/ml protein, then frozen at Ϫ80°C. The lysates were added directly to microtiter wells coated with A77 F(abЈ) 2 , M22 F(abЈ) 2 , or 520C9 F(abЈ) 2 and incubated overnight at 4°C. After washing, the wells were incubated for 2 h at ambient temperature with a 1/2,000 dilution of ␥ subunit-specific rabbit serum (donated by Dr. Don Durden, Indiana University School of Medicine, Indianapolis, IN). The assay was developed with an alkaline phosphatase-conjugated goat anti-rabbit IgG probe with p-nitrophenyl phosphate and read at a wavelength of 405-650 nm.

Statistics
Statistical analysis was performed using Student's t test and calculated by SIGMAPLOT software (Jandel, San Francisco, CA). p Ͻ 0.05 was considered significant. All experiments were performed a minimum of three times.

Results
The role of tumor cell phagocytosis by macrophages in vivo remains speculative, yet in vitro studies clearly demonstrate therapeutic potential for this form of anti-tumor activity (24 -28). The development of stable dyes and new methods has been essential to study this phenomenon in vitro. We have adapted a flow cytometry method, initially described by Munn and Cheung (27), to study tumor cell phagocytosis by MDM Fc␣RI and Fc␥RI. This twocolor method employs a red lypophilic dye to stain SKBR-3 carcinoma cells, and FITC-conjugated mAbs to label MDM. Phagocytosis can be evaluated by quantitative analysis of the number of single-colored cells (red or green) and the number of dual-colored cells, which represent tumor cells engulfed by MDM. Initial experiments using confocal microscopy were performed to confirm that dual-positive cells represented phagocytosis of target cells and not merely binding of target with effector cells. The BsAb A77 ϫ 520C9 and H22 ϫ 520C9, which specifically targeted the HER2/ neu-expressing SKBR-3 cells to Fc␣RI and Fc␥RI, respectively, were used to mediate phagocytosis. These BsAb contained only the FabЈ of the individual mAbs; therefore, binding to FcR was restricted to the specificity of the Ab.
As shown in Table I, ϳ20 -25% of the dual-positive events were E:T cell conjugates at 30 min. However, by 300 min this value was 3% or less, and Ͼ90% of dual-positive macrophages contained partial or whole tumor cells. The images in Fig. 1 show conjugates from the 30 min and phagocytosis at the 300 min points.
To further demonstrate the difference in Fc␣RI-and Fc␥RImediated phagocytosis using IFN-␥-treated MDM, we studied tumor cell survival following phagocytosis. Samples of cells from the phagocytosis assay were taken (just before fixation), washed, and allowed to adhere overnight to microtiter plates in growth medium. The ability of the tumor cells to adhere was the primary requirement for viability, and relative cell survival was assessed by a HER2/neu-specific ELISA. The SKBR-3 cells cultured without MDM served as the 100% survival control, and MDM cultured without SKBR-3 cells was used to determine 0% survival. Fig. 3 shows that the percent reduction in SKBR-3 cell survival was similar for Fc␣RIand Fc␥RI-targeted BsAb when MDM were cultured with M-CSF. Both BsAb reduced cell survival ϳ50 -60% from control values. In contrast, with IFN-␥-treated MDM, the Fc␥RI-targeted BsAb, and not the Fc␣RI-targeted BsAb, resulted in a significant loss of tumor cells (ϳ50 -60% reduction). These results make clear the divergence in cytotoxic function of Fc␥RI and Fc␣RI under specific conditions.
Flow cytometry was used to determine whether changes in FcR expression may account for the differences observed in tumor cytotoxicity by MDM. Previous studies have shown that IFN-␥ significantly up-regulated Fc␥RI on myeloid cells (8 -10), while Fc␣RI levels were relatively unaffected by IFN-␥ (20). In contrast, GM-CSF has been shown to increase Fc␣RI on monocytes (20). Fig. 5 illustrates the relative expression of FcR compared with that by MDM cultured without cytokines (averaged from six independent donors). The only statistically significant change was the increase in Fc␥RI expression with IFN-␥-treated MDM (4.8-fold increase over control, p Ͻ 0.05). Treatment of MDM with various cytokines did not result in significant differences in Fc␣RI expression. No effect was demonstrated on Fc␥RII or CD33 expression, a nontriggering macrophage surface receptor used as a control.
Because FcR surface expression could not account for most of the differences observed in tumoricidal activities of MDM cultured with different cytokines, we hypothesized that the effect may be at the level of association between the signal-transducing ␥ subunit and Fc␣RI and Fc␥RI. To study the level of association between the ␥ subunit and Fc␣RI and Fc␥RI, lysates were prepared from cytokine-differentiated MDM and were analyzed by ELISA. These assays were performed by capturing FcR ␥-chain complexes with F(abЈ) 2 of FcR-specific mAbs or an isotype control, followed by detection of ␥ subunit with ␥-specific rabbit serum (Fig. 6). As expected, the results from three independent MDM preparations  showed a significant enhancement of total Fc␥RI ␥-chain complexes in IFN-␥-treated MDM. Interestingly, GM-CSF-cultured MDM also had elevated Fc␥RI ␥-chain complexes compared with MDM cultured without exogenous cytokine or with M-CSF ( p Ͻ 0.05). No significant differences were observed for Fc␣RI ␥-chain complexes, although MDM cultured in GM-CSF and IFN-␥ had somewhat higher values than control MDM or MCSF-treated MDM. There was no reactivity when lysates were added to plates coated with irrelevant F(abЈ) 2 molecules (maximum OD 405 was Ͻ0.06 in all cases). These data imply that most of the differential activities of cytokine cultured MDM are not manifested by changes in total FcR ␥-subunit complex formation. However, the effect of IFN-␥ on enhanced Fc␥RI expression and Fc␥RI␥ subunit complexes may play a significant role in the Fc␥RI-mediated tumoricidal activity of IFN-␥-treated MDM.

Discussion
In this study we investigated two cytotoxic mechanisms, phagocytosis and ADCC, mediated by Fc␣RI and Fc␥RI on macrophages derived from monocytes in the presence of M-CSF, IFN-␥, or GM-CSF. Although the cytotoxic effector functions of Fc␥RI have been well studied, relatively little is known regarding the ability of Fc␣RI to mediate tumor cell killing. To study the activities of individual FcRs, we used chemically linked FabЈ ϫ FabЈ BsAb, which excluded the possibility of interactions with other FcR through Ab Fc domains. Specifically, we used BsAb A77 ϫ 520C9 and H22 ϫ 520C9, which targeted the HER2/neu-expressing SKBR-3 tumor cells to Fc␣RI and Fc␥RI, respectively. These BsAb have been previously described (7,15), and H22 ϫ 520C9 is currently being investigated in clinical trials for treatment of HER2/neu-expressing malignancies (31,32). We found significant differences in tumor cell cytotoxicity that were dependent on which FcR was targeted as well as the presence of specific cytokines during the differentiation of monocytes into MDM.
The most striking finding in this study was the lack of significant ADCC or phagocytosis of SKBR-3 cells by the BsAb targeted to the Fc␣RI when MDM cultured in the presence of IFN-␥ were used as effector cells (maximum Fc␣RI activity, 10% phagocytosis and 14% ADCC). In contrast, the IFN-␥-treated MDM mediated both phagocytosis and ADCC through Fc␥RI (maximum Fc␥RI activity, 23% phagocytosis and 34% ADCC). Decreased phagocytic function with IFN-␥-treated MDM have previously been reported with anti-tumor Abs that mediate their activity via Fc␥RII and Fc␥RIII (22,24,25). Interestingly, our results and those reported by Ely et al., which demonstrate enhanced phagocytosis with IFN-␥-treated MDM (26), suggest that the Fc␥RI is unique from other FcRs. This observation is intriguing, because both Fc␣RI and Fc␥RI appear to require association with the FcR␥ subunit to mediate intracellular signaling (17,18).
The difference between Fc␣RI and Fc␥RI activity using IFN-␥-derived MDM was most pronounced when we examined tumor cell survival following coculture of MDM with tumor cells. Addition of the Fc␥RI-targeted BsAb to these cultures resulted in 50 -60% reduction in SKBR-3 cells, whereas no significant loss of tumor cells was noted (Ͻ10%) with the Fc␣RI BsAb. These results correlated well with the ability of Fc␥RI on IFN-␥-treated MDM to mediate both phagocytosis and ADCC and the lack of such activity for Fc␣RI.
We speculated that the variation in Fc␣RI-and Fc␥RI-mediated activity of cytokine-cultured MDM may have been the result of changes in their surface expression or the ability of ␥-chain to functionally associate with the FcR. Recent studies by Launay et al. (33), which demonstrated the presence of Fc␣RI with and without ␥-chain on the surface of monocytes and neutrophils, implicated that receptor activity may be regulated by the magnitude of ␥-chain association. However, when we examined the effects of cytokines on MDM FcR surface expression or total levels of FcR␥ subunit complexes, we did not find a complete correlation with these parameters and the observed effects of cytokines on tumoricidal function via Fc␣RI and Fc␥RI. Consistent with previous reports, IFN-␥ treatment of MDM resulted in increased levels of surface Fc␥RI and Fc␥RI␥ subunit complexes (8 -10). These effects probably contributed to Fc␥RI-mediated phagocytosis and ADCC with IFN-␥-treated MDM. However, Fc␥RI-mediated phagocytosis was greater when MDM were generated in the presence of M-CSF and GM-CSF despite lower Fc␥RI expression and lower total Fc␥RI␥ subunit complexes. The fact that Fc␥RI-mediated phagocytosis with GM-CSF-or M-CSF-cultured MDM was greater than phagocytosis with IFN-␥-treated MDM, supports the MDM were generated exactly as described for the phagocytosis and ADCC experiments. MDM were washed and incubated with Abs specific for Fc␥RI (m32.2), Fc␣RI (A77), Fc␥RII (AT10), and CD33 (251). The primary Abs were detected with goat anti-murine IgG-PE, and samples were analyzed for mean fluorescence intensity (MFI) using FACScan. The relative expression was calculated by setting the MFI value without cytokine treatment at 100%. The data demonstrate the mean percent change in expression Ϯ SE from six experiments with different donors. ‫,ء‬ p Ͻ 0.05 comparing expression of a receptor under different cytokine conditions. FIGURE 6. The association of ␥-chain with Fc␣RI and Fc␥RI in cytokine-treated MDM. Lysates prepared from MDM cultured with cytokines for 5-7 days were incubated at specified concentrations in microtiter wells coated with F(abЈ) 2 of anti-Fc␥RI (M22) or anti-Fc␣RI (A77). The FcRassociated ␥ subunit was detected with a ␥ subunit-specific rabbit serum and alkaline phosphatase-conjugated anti-rabbit IgG probe. The data represent the mean Ϯ SE of lysates processed from three independent MDM preparations for each cytokine condition. ‫,ء‬ p Ͻ 0.05 comparing ␥ subunit-FcR association under different cytokine conditions to that in control cultures.
previous finding that IFN-␥ treatment reduces the overall phagocytic capacity of macrophages (22,24,25). In contrast to Fc␥RI, no significant change in the total level of surface Fc␣RI or Fc␣RI␥ subunit complexes were noted among MDM cultured under the influence of different cytokines. Therefore, these parameters were unable to explain the enhanced Fc␣RI-mediated ADCC with GM-CSF-treated MDM or the low tumoricidal activity of Fc␣RI using IFN-␥-treated MDM. These data suggest that other molecules and interactions probably contribute significantly to the differential effect of cytokines on Fc␣RI-and Fc␥RI-mediated tumoricidal activity of MDM.
Although IFN-␥ treatment has been shown to promote ADCC activity, we found that this did not apply for Fc␣RI on MDM. We observed Ͻ15% Fc␣RI-dependent ADCC under these conditions. This result was surprising considering our recent findings that demonstrated Fc␣RI to be a potent mediator of ADCC when fresh (untreated) monocytes or PMN were used as effector cells (7). On the other hand, MDM generated in the presence of GM-CSF were efficient at Fc␣RI-mediated ADCC. In fact, MDM cultured with GM-CSF mediated greater ADCC activity via Fc␣RI than via Fc␥RI ( p Ͻ 0.05), although there was no significant up-regulation of Fc␣RI on MDM cultured with GM-CSF. Therefore, the positive or negative effects on Fc␣RI activity induced by cytokine treatment of MDM appear to work downstream from receptor binding. These results further imply that Fc␣RI and Fc␥RI (and probably other FcR) interact with different molecules (in addition to the ␥ subunit), which uniquely regulate their function.
Fc␣RI and Fc␥RI both mediated efficient phagocytosis of SKBR-3 cells when MDM were cultured with MCSF or GM-CSF. Ab-dependent phagocytosis was also demonstrated when MDM were cultured without exogenous cytokines. On the other hand, neither BsAb mediated efficient lysis of the tumor cells with untreated MDM or M-CSF-cultured MDM. The fact that MDM cultured without cytokines were unable to mediate significant ADCC is intriguing, because monocytes (before differentiation into MDM) have been shown to mediate efficient ADCC via both Fc␣RI and Fc␥RI (7,15). Previous studies have demonstrated that monocytes rapidly lose ADCC function in culture (30,34). Therefore, the lack of MDM ADCC activity without addition of cytokines does not imply that macrophages in vivo require cytokines for this function. The data we have presented demonstrate that MDM cultured under the influence of different cytokines result in effector cells with varying capacity to mediate cytotoxic function via Fc␣RI and Fc␥RI. These results suggest that Ab-dependent cytotoxic activity of human macrophages is dependent on the FcR that can be engaged by the anti-tumor Ab and the physical state of the macrophages. Further, the data imply that the use of specific cytokines to regulate macrophage function may enhance or reduce specific cytotoxic mechanisms in a clinical setting.
This study has focused on macrophages; however, in the use of cytokines in combination with Ab treatment for tumor therapy, other effector cell populations that may be recruited for ADCC and phagocytosis need to be considered. For example, the effects of G-CSF and IFN-␥ (in vitro and in vivo) on up-regulation of Fc␥RI levels and cytotoxic activity on granulocytes are well documented (8 -10). On the other hand, granulocytes constitutively express Fc␣RI, which mediates potent ADCC without addition of exogenous cytokines (6,7). Further, because PMN constitute an abundant effector cell population, they may play an important role in FcR-mediated immunotherapy. Having a collective understanding of the important effector cell populations and their different cytotoxic mechanisms will allow for greater exploitation of immunotherapy. Further studies that help optimize the effector functions of FcR in the course of immunotherapy will probably lead to more successful protocols for treatment of human cancers.