The macrophage capacity for phagocytosis

Macrophages can ingest great quantities of particles, yet not so much that they burst. When plated onto flat surfaces opsonized with IgG, they engage that surface as if to engulf it, a process termed frustrated phagocytosis (Henson, 1971; Takemura et al., 1986), and spread to some limit. It may be that the macrophage capacity for phagocytosis or its extent of spreading during frustrated phagocytosis is limited by the number of available phagocytic receptors, and when all of them have been internalized or engaged phagocytosis stops. Alternatively, some cellular structure which changes during the process may approach a limit, and that limit defines satiety.

The goals of the present experiments were to determine the phagocytic capacity of murine bone marrow-derived macrophages, to compare that capacity to frustrated phagocytosis, and to identify factors that limit those capacities. Capacity was measured using polystyrene beads that range in size from those that could easily be ingested by one cell to those that were significantly larger than the cell itself. The cell and bead concentrations were adjusted so that the usual condition consisted of one cell attempting to phagocytose one bead. Determining the largest bead that the macrophage could engulf under various conditions allowed for the examination of several interesting characteristics of phagocytosis. Cell surface area at the phagocytic limit was similar to that reached after frustrated phagocytosis. Fc-mediated phagocytosis stopped before all Fc-receptors were engaged or internalized, indicating that cell surface membrane became limiting. Moreover, the phagocytic capacity was lowered when microtubules were depolymerized. Comparisons with frustrated phagocytosis suggested that the radial symmetry of the phagocytic response enhanced the phagocytic capacity.

Murine bone marrow-derived macrophages were obtained by the method of Swanson (1989b). Bone marrow was removed from femurs of female ICR mice (Charles River, Cambridge, MA) and was washed 3× in cold Dulbecco’s modified essential medium plus 10% heat-inactivated fetal bovine serum (DME-10F). Cells were resuspended and plated at 2 × 10⁵ cells/ml on 100 mm Lab-Tek dishes in 25 ml or on 60 mm Lab-Tek dishes in 10 ml of complete bone marrow medium (DME plus 30% L-cell-conditioned medium plus 20% heat-inactivated fetal bovine serum). The cells were incubated at 37°C under 5% CO₂ for 6 days, and the macrophages adhered to the bottom of the dish during this period. Macrophages were harvested from dishes after brief washing with cold, sterile, divalent cation-free, phosphate-buffered saline (PD: 137 mM NaCl, 3 mM KC1, 7 mM phosphate buffer, pH 7.4). The resuspended cells were plated onto 12 mm coverslips in 24-well Costar dishes at 1 ×10⁵ cells/ml. After 30 minutes at 37°C, the PD was replaced with DME-10F. Cells were used within the next two days. Greater than 95% of the cells were macrophages, as determined by their ability to phagocytose opsonized sheep red cells.

Macrophages resuspended from Lab-Tek dishes into cold PD were examined microscopically (magnification, ×500). Diameters of 83 rounded, unspread cells were measured using a calibrated eyepiece graticle.

2 × 10⁶ polystyrene polybeads of diameter 21.1 μm (Polysciences Inc., Warrington, PA) were incubated in 10 mg/ml bovine serum albumin (BSA, Sigma Chem. Co., St. Louis) in PD for 1 hour at 4°C. The beads were then washed 5 times in PD, then resuspended in 1 ml PD. Rabbit anti-bovine albumin, IgG fraction (anti-BSA IgG, Cappel, West Chester, PA), was added to a final dilution of 1:500, then incubated for 30 minutes at 37°C and 10 minutes at 4°C. The beads were then washed 3 × in 1 ml PD and recounted. The 21.1 μm beads had a considerable variance in size (4.09 μm, according to the manufacturer) with a range of 13 μm to >30 μm in diameter, and this made the phagocytosis assay possible.

For other phagocytosis assays, 21.1 gm BSA-coated beads were incubated in 6 mg/ml 2,4-dinitrobenzene sulfonate (DNBS, Aldrich Chem. Co., Milwaukee), in 4% K₂CO₃ for 1 hour, then washed 3 × in 1 ml PD. lite beads were then incubated in anti-DNP IgG before or after they came in contact with the macrophages.

A quantitative measurement for the phagocytic capacity of macrophages was obtained by allowing the cells to engulf large latex beads, then using immunofluorescence to label beads that were not internalized, and measuring how many beads of a given size interval were positively labeled.

Fresh DME-10F (1 ml) was added to day 7 or day 8 macrophages. 2 × 10⁴ beads (21.1 /mt diameter) were added to each well. Cells were allowed to phagocytose the beads for one hour. Dishes were then placed on ice, washed gently 3 × in cold PD, the wells aspirated dry, and rhodamine-labeled goat anti-rabbit IgG (rhodamine-GAR: heavy and light chain specific, Cappel), 300 μl at 1:50 in PD, was added. The cells were then incubated on ice for 30 minutes and washed 3 x in PD. Control experiments showed that only beads which had both BSA and anti-BSA on them could be labeled with the rhodamine-labeled secondary antibody (data not shown). Even though 30 minutes at 4°C was allowed for the rhodamine-labeling step of the assay, beads were completely labeled in less than 10 minutes at that temperature.

For microscopic analysis, the coverslips were inverted, mounted on a glass slide on top of glass coverslip fragments, and the cavity filled with PD and sealed with valap to prevent drying (Swanson, 1989a). The cells were studied with a Zeiss Photomicroscope III equipped for epi-illumination of fluorescent specimens. To measure bead diameters, a Wert ×10 ocular micrometer was calibrated with a stage micrometer. The divisions on the ocular micrometer were 1.8 micrometers apart at ×500. For each bead, the diameter was measured and the bead scored as plus or minus for rhodamine fluorescence with rhodamine filters. Only in cases where there was one bead per macrophage, and one macrophage per bead, were measurements taken. Although a phase 3 lens was used to view the cells, a condenser lens without phase rings was used during bead diameter measurement to avoid optical distortion of bead dimensions. The diameter of the beads could be measured to the nearest micrometer with considerable accuracy. For each 1.8 μm gradation, 10 beads were scored for fluorescence.

All of the experiments described in this section were executed in their entirety at least twice.

The beads were assayed as above, except that phagocytosis was allowed to proceed for shorter periods of time: 5,10, 20, 30 and 60 minutes. Coverslips were removed from the dish and placed in a new dish on ice with 1 ml PD in each well. Rhodamine-GAR was added after the cells were chilled and washed.

BSA/anti-BSA-coated beads, BSA/DNP-coated beads (no anti-DNP), and BSA/DNP beads that had been opsonized with rabbit anti-DNP IgG were added to two wells. After one hour of phagocytosis, anti-DNP was added to the unopsonized bead wells (300 μl of 50 μg/ml in PD) for 30 minutes on ice, while the others were kept in PD. After 3 washes, all were incubated in the rhodamine-GAR.

Macrophages were incubated in 10 μM nocodazole (Sigma Chemical, St. Louis, MO) in DME-10F for one hour before 21.1 μm beads were added. The phagocytosis and labeling procedures remained the same.

Macrophages were incubated overnight (15 hours) in 10 mg/ml or 20 mg/ml sucrose in DME-10F according to Swanson et al. (1986). Phagocytosis and labeling protocols were unchanged. For invertase recovery, cells that had been incubating in 20 mg/ml sucrose for 15 hours were washed twice in medium, then 1 ml of 0.5 mg/ml invertase (Sigma Chem. Co.) in medium was added for 2, 4 or 6 hours before the phagocytosis assay was performed. In control experiments, vacuolated cells were maintained in sucrose for 4 hours before allowing phagocytosis.

The extent of cellular spreading of macrophages that engage in frustrated phagocytosis was determined. 12 mm coverslips were treated with DNP according to the method of Michl et al. (1979). Coverslips were opsonized with 50 μg/ml anti-DNP IgG in PD for 30 minutes and then washed 3 x in PD. Macrophages that had been chilled and resuspended from Lab-Tek dishes were plated onto the coverslips at 1–2 × 10⁴/well. After 1 hour of spreading, the cells were chilled on ice and then incubated with 300 μl rhodamine-GAR (1:100) for 30 minutes to label areas of the coverslip which were not masked by the macrophages. The cell diameters were measured with the ocular micrometer.

For other studies, coverslips coated with BSA were opsonized with anti-BSA IgG for 30 minutes, then washed thoroughly before plating macrophages. Macrophages resuspended from Lab-Tek dishes were plated onto coverslips in Ringer’s saline ± 10 μM nocodazole.

Sheep red blood cells (SRBC: Cappel) were opsonized with goat anti-SRBC IgG (Cappel) at a final dilution of 1:1000 (30 minutes 37°C, 10 minutes 4°C, washed 3 × in PD). Macrophages that had phagocytosed test beads were incubated for 1 hour with 3 × 10⁶ opsonized SRBC. Cells were subsequently chilled and incubated in rhodamine-GAR. In separate experiments, opsonized SRBC were added to macrophages 1 hour after plating onto opsonized coverslips (DNP-anti DNP). SRBC were incubated with the macrophages for 15 minutes at 4°C, then unbound SRBC were washed away. The number of red cells on the surface of 100 randomly selected macrophages was measured.

Macrophages were allowed to spread for 40 minutes onto coverslips opsonized with anti-BSA IgG ± 10 μM nocodazole. Cells were fixed with 3.7% formaldehyde plus 0.25 M sucrose, 0.5 mM EDTA, 1 mM EGTA, 20 mM HEPES, pH 7.4, then were washed and observed by phase-contrast microscopy. Video images of 35 cells for each condition were recorded and digitized for image processing. Outlines of cells were traced using interactive software; these tracings were used to generate a binary mask for each cell, then these binaries were analyzed to measure area (A), perimeter (P) and the shape parameter (P²/4πA) for each cell using the image processing capabilities of a Tracor Northern TN 8500 system (Noran Inst., Middletown, WI).

The phagocytosis assay measured the percentage of latex beads within a given size range that could be engulfed by macrophages. Only instances in which one macrophage engaged one bead were acceptable for measurements. Internalized beads were identified as such by their inaccessibility to fluorescently labeled antibodies against their surfaces. Beads bound to cells were almost always completely dark or brightly fluorescent. Exceptions occurred when beads were partially engulfed and thus only partially labeled; such beads were scored positive for fluorescence (i.e. they had failed to engulf the bead). The basic assay to determine the phagocytic capacity of bone marrow-derived macrophages was performed as a separate experiment and as a control for each of the subsequent experimental variations. The data were averaged and expressed as the percentage of positively labeled beads in each bead size category (Fig. 1). An operative definition for the phagocytic limit was given as the bead diameter at which 50 per cent of the beads scored positive for rhodamine labeling. In order to determine this point, a logit transformation was performed on the data to make the sigmoidal curve linear, and then first-order regression analysis was performed. In the control condition, the 50% mark occurred at 19.8 μm, with a 95% confidence interval of 17.0 μm to 22.5 gm. Macrophages resuspended from a culture dish assume a spherical morphology. The diameters of resuspended macrophages were found to be 13.8 ± 2.3 μm (s.d.). The calculated phagocytic capacity of 19.8 μm indicates that those cells are capable of ingesting particles 1.44 times their diameter, or 3 times their volume.

The time course of phagocytosis was determined by varying the time that the macrophages were exposed to the beads at 37°C. As Fig. 2 indicates, macrophages reached their phagocytic capacity by the 30 minute time (50% =20.8, P>0.5). The phagocytic size limit at 20 minutes was already not significantly different from the control values (50%=18.4 μm, P<0.40).

To examine the role of opsonization in the phagocytosis of larger beads, BSA-coated beads were labeled with dinitrobenzene sulfonate (DNBS) before incubation with the cells, and phagocytosis was measured by chilling and labelling the beads with anti-DNP and then rhodamine-GAR. Macrophages phagocytosed DNP/anti-DNP coated beads (50% = 19.1 μm, P>0.50) to the same extent as they did BSA/anti-BSA coated beads, but unopsonized beads had a 50% point of 14.0 μm (PcO.001, data not shown). This implicates receptorligand interactions as necessary for the phagocytosis of larger beads.

We compared the phagocytic capacity for opsonized beads with the extent of macrophage spreading on opsonized substrata. This frustrated phagocytosis (Henson, 1971; Michl et al., 1979; Takemura et al., 1986) is analogous to the phagocytosis of an infinitely large particle. A significant difference in values obtained for phagocytosis of beads versus those for spreading would indicate that the cells can distinguish shapes of particles during phagocytosis. Due to the tightness of the seal that the spread cells formed with the coverslip, and the fact that cells that were not touching any other cells would spread into circular profiles (Fig. 3A), a simple measurement of the diameter of the circular profile permitted estimation of cell spreading and surface area. Since cells either tended to clump together in suspension, or to contact each other as they spread, only the 10 largest, isolated and circular macrophage profiles on each coverslip were measured. The average diameter of the 10 largest cells on coverslips from several different cultures of macrophages was 44.3 μm, with a standard deviation of 3.8 μm. A circle with this diameter has an area of 1,541 μm². To account for both the upper and lower faces of the cell, the area should be doubled to yield a cell surface area of 3,080 μm². This estimate is a minimum value for the surface area, since it considers the cell as an infinitely thin disk. It is a fair estimate, however, since cell ruffling is considerably reduced in this condition, and the cells are very flat (Fig. 3A). Moreover, our estimate is in good agreement with other determinations of surface area obtained in thioglycolate-elicited peritoneal macrophages (Phaire-Washington, 1980). This limit is similar to the surface area required to ingest the largest beads. To engulf a bead of 19.8 pm, the macrophage plasma membrane must reach a surface area of at least 2,463 μm² (2 × 1232 μm²). This similarity between bead phagocytosis and frustrated phagocytosis indicates that at their limit these two processes attain the same total cell surface area, and that macrophages do not discriminate particle shapes in these processes. The difference of 600 μm² may be explained by the difference in the way the limits are defined. For frustrated phagocytosis we measured the 10 largest circular cell profiles, essentially identifying champions, whereas for bead phagocytosis we attained the half maximal bead size for phagocytosis, which measures the capacities of the whole population. The latter method would be expected to yield lower numbers.

Although it appears that membrane availability is the limiting factor with respect to phagocytic capacity, another reason why the cells might not be able to ingest more could be that they deplete from their surfaces all receptors for binding with the opsonized particles. To test this, opsonized sheep erythrocytes were added to either normally spread macrophages that had phagocytosed a large bead or macrophages that were engaging a coverslip in frustrated phagocytosis. As shown in Fig. 4, macrophages spread onto opsonized surfaces could bind red cells on their upper surface. A count of erythrocytes per 100 cells yielded 15 (±6) bound erythrocytes per macrophage. Similarly, macrophages engaging large, opsonized beads could also bind opsonized erythrocytes (data not shown). These results indicate that the phagocytic capacity is not limited by the availability of receptors.

The cell surface areas reached by phagocytosis were considerably greater than the reported macrophage membrane surface area (825 μm²\Steinman et al., 1976), indicating that internal membranes are recruited to the plasma membrane during phagocytosis. We asked how expanding internal membranous compartments affected the engulfment of larger beads. To do this, the endocytic and lysosomal compartments were vacuolated during a 15 hour incubation in sucrose. Macrophages lack invertase, so pinocytosed sucrose accumulates in lysosomes, where it expands that compartment osmotically (Cohn and Ehrenreich, 1969). Cells were incubated overnight in various concentrations of sucrose and then the phagocytosis assay was performed (Table 1). The ability of macrophages to ingest beads was drastically reduced after overnight vacuolation in 30 mg/ml and 20 mg/ml sucrose. 10 mg/ml sucrose also had an appreciable effect, reducing the 50% labeled value to 16.4 μm (P<0.02).

To determine if this decrease in the size limit was not due to general impairment of cells, the vacuolation was reversed with invertase. Invertase degrades sucrose to its component monosaccharides, which exit lysosomes and permit shrinkage of the lysosomal compartment (Cohn and Ehrenreich, 1969; Knapp and Swanson, 1990; Swanson et al., 1986). Cells were vacuolated in sucrose for 15 hours and then washed in medium. They were then incubated in 0.5 mg/ml invertase for up to 6 hours (Table 1). Within 2 hours the phagocytic capacity had nearly returned to pre-vacuolation levels. A small increase was seen at the 4 hour and 6 hour time points. The invertase-treated cells approached but did not reach pre-vacuolation levels.

To determine if microtubules were necessary for phagocytosis of the beads, cells were incubated in 10 μM nocodazole, which causes reversible depolymerization of macrophage microtubules (Swanson et al.,1987), for 1 hour before and during incubation with the beads. Nocodazole produced a small but significant decrease in the amount the cells could ingest (50% = 16.9 μm, P<0.05).

Because depolymerization of cytoplasmic microtubules reduced the phagocytic capacity, we compared this response to frustrated phagocytosis and found that nocodazole did not limit the extent of spreading, but instead made that spreading response irregular. Whereas the usual frustrated phagocytic response created circular spreading profiles, with the nucleus centrally placed, macrophages spreading in nocodazole formed more irregular shapes (Fig. 3). We quantified these observations by tracing the profiles of cells spread onto opsonized coverslips for forty minutes in the presence or absence of 10 μm nocodazole. Using digital image processing, we determined spreading area and the deviation of the profile from a circle (Table 2). Cells were not significantly different in spread areas but were significantly different in shape. This indicates that phagocytosis of the largest particles is enhanced by a microtubule-dependent organization of cytoplasm.

Measurements using the phagocytic assay described here indicate that frustrated phagocytosis and particle phagocytosis are limited in similar ways, but are different from cell spreading onto unopsonized surfaces. Grinnell and Geiger (1986) described the phagocytosis by fibroblasts of various sizes of fibronectin-coated beads and found that fibroblasts could phagocytose 6 μm diameter beads easily, but could not engulf beads larger than 12–14 μm diameter. In the light of earlier work in which prior spreading on fibronectin-coated surfaces inhibited phagocytosis of fibronectin-coated beads (Grinnell, 1980), they concluded that phagocytosis and cell spreading on flat surfaces were fundamentally the same process, which was limited either by the number of fibronectin receptors or by some other physical constraint.

We found that particle capacity was increased by opsonization, but was not limited by Fc-receptor availability. Macrophages at their phagocytic capacity, either with a bead inside or spread to their limit, remained capable of binding opsonized erythrocytes on their upper surfaces. The limit instead appeared to be in the amount of cell surface membrane available for spreading. Macrophages in suspension, or plated onto unopsonized surfaces, contain prominent ruffles of membrane. When engaging an opsonized surface, ruffling membrane moves along that surface and flattens against it. As the limit is approached, surplus membrane is recruited into the phagocytic response, and the cell surface is drawn smooth (Fig. 5). The limit seems therefore set by the amount of available membrane. An alternative explanation has been offered by Rabinovitch et al. (1975), that frustrated phagocytosis is not limited by spreading, but rather by a paralysis of Fc-receptors for phagocytosis. In their studies, murine peritoneal macrophages plated onto opsonized coverslips could bind but not ingest opsonized erythrocytes. They argued that this inhibition was not a consequence of cell spreading because other conditions which increased cell spreading did not inhibit phagocytosis. It therefore remains possible that macrophage frustrated phagocytosis is limited by a specific inactivation of Fc-receptor function, an inhibition mediated by receptor ligation.

Our estimates of the cell surface area at capacity are similar for both bead phagocytosis and frustrated phagocytosis, and are also similar to values reported in earlier studies of phorbol ester-stimulated spreading (Phaire-Washington et al., 1980). Those surface areas are greater than those measured in unstimulated macrophages (825 μm²; Steinman et al., 1976), indicating that intracellular membranes are brought to the surface during phagocytosis. Occupation of internal membranes, such as occurred during sucrose vacuolation, decreased the phagocytic capacity reversibly, indicating that these membranes contribute to the spreading. It remains possible, however, that macrophage surface area does not change during phagocytosis; that the highly folded surface simply smooths out to enclose particles, and that filling endocytic compartments with sucrose or smaller particles depletes the cell surface of membrane which would otherwise be available for the spreading response.

The similar limits of particle phagocytosis and frustrated phagocytosis indicate that they are not affected by particle shape or by the size of the phagosomal lamellipod. The advancing edge of the pseudopod decreases beyond the equator of a spherical particle, but continually increases on a flat surface. Pseudopod advance therefore appears to be regulated locally, by the segmental interactions between Fc-receptors and the opsonized surface, and with the limit of advance set by the amount of available surface membrane.

That both particle phagocytosis and frustrated phagocytosis proceed in the absence of cytoplasmic microtubules suggests that the process is independent of this cytoskeletal element. Macrophages plated onto opsonized surfaces in the presence of nocodazole engage those surfaces in a phagocytic response, and reach the same spreading surface area as control cells. However, the phagocytic capacity for particles is diminished by nocodazole treatment, indicating that microtubules contribute to particle phagocytosis in some way. We suggest that microtubules provide a coordinating function. Such coordination is indicated by the striking symmetry of cell spreading during frustrated phagocytosis. Macrophages spread into nearly perfect circles, with the nucleus centrally placed and with microtubules radiating to the cell periphery from a central point. In the absence of microtubules, such spreading may be initially circular, but soon rearranges to an irregular profile. The radial symmetry that is lost during nocodazole treatment may reflect a process that coordinates pseudopod closure around large particles. Without microtubules, uncoordinated pseudopod advance around particles would produce irregularly shaped gaps that close less readily than the small, circular gaps left by symmetrical spreading.

In contrast to Fc-mediated phagocytosis (Bhisey and Freed, 1971), complement-mediated phagocytosis in macrophages is sensitive to microtubule-destabilizing drugs (Wright and Silverstein, 1982). Complement-mediated phagocytosis therefore may be more similar to simple spreading processes, which are sensitive to nocodazole, than to Fc-mediated phagocytosis. Measurements of the phagocytic capacity of complement-opsonized particles might be instructive in this regard.

The authors gratefully acknowledge the assistance and advice of Philip Knapp and Esther Racoosin. This work was supported by the NIH (CA44328).