Bone Marrow-derived Mononuclear Phagocytes Autoregulate Mannose Receptor Expression*

This study extends our previous observation that surface mannose receptor expression by pure populations of CSF-1-dependent bone marrow-derived mac- rophages increases with time We presently find, however, that the progressive enhancement of lZ6I-mannose-bovine serum albumin (‘”I-Man-BSA) bind- ing per cell reflects cell number rather than duration of culture. In fact, macrophages plated at high density bind &fold more ‘”I-Man-BSA than do their low density counterparts, with no difference in receptor-li- gand affinity. Furthermore, cells cultured at high density are ultimately subjected to lower levels of exoge- nously provided macrophage growth factor, and fewer are in interphase. By obtaining synchronous popula- tions of quiescent bone marrow macrophages, how-ever, we demonstrate that neither cell cycling nor at- tendant levels of colony stimulating factor- 1 influence mannose receptor expression. Our next series of experiments established that den-sity-related mannose receptor expression reflects re- moval, by marrow macrophages, of a “down-regulat-ing” factor contained in culture medium. To this end, we treated mononuclear phagocytes with either N precipitate The supernatant was aspirated and left at 4 "C for 1 week after which it was dialyzed against a-MEM overnight. Control material was treated identically except for acidification. A suboptimal down-regu- lating volume (10 pl) of material was tested for its effects on lZ5I-Man-BSA binding.

This study extends our previous observation that surface mannose receptor expression by pure populations of CSF-1-dependent bone marrow-derived macrophages increases with time (Clohisy, D. R., Bar-Shavit, Z., Chappel, J. C., and Teitelbaum, s. L. (1987) J. Biol. Chem. 262,16922-15929). We presently find, however, that the progressive enhancement of lZ6Imannose-bovine serum albumin ('"I-Man-BSA) binding per cell reflects cell number rather than duration of culture. In fact, macrophages plated at high density bind &fold more '"I-Man-BSA than do their low density counterparts, with no difference in receptor-ligand affinity. Furthermore, cells cultured at high density are ultimately subjected to lower levels of exogenously provided macrophage growth factor, and fewer are in interphase. By obtaining synchronous populations of quiescent bone marrow macrophages, however, we demonstrate that neither cell cycling nor attendant levels of colony stimulating factor-1 influence mannose receptor expression. Our next series of experiments established that density-related mannose receptor expression reflects removal, by marrow macrophages, of a "down-regulating" factor contained in culture medium. To this end, we treated mononuclear phagocytes with either macrophage-or control-conditioned medium and found that, via a fetal calf serum-residing protein(s), only control medium is capable of noncompetitively reducing lZ51-Man-BSA binding in a dose-dependent manner. Moreover, reconstituted 20-40% (NH4)zS04-precipitable fractions derived from either sham-conditioned medium or fetal calf serum are capable of downregulating mannose receptor expression. Alternatively, the same fraction obtained from macrophageconditioned medium contains no such activity. Finally, initial characterization of the down-regulating factor reveals it to be acid-activable and trypsin-sensitive, yet resistant to heating to at least 80 "C, ribonuclease A, or freezing and thawing. We conclude that bone marrow macrophages up-regulate expression of their own plasma membrane mannose receptor by inactivating a noncompetitive, serum-residing inhibitory protein(s).  The mannose receptor is a 175-kDa plasma membrane component, which in the marrow resides exclusively in cells of the monocyte/macrophage family (1). The protein is known to be involved in recognition and endocytosis of particles and other substances displaying terminal mannose residues. Thus, it is likely that the mannose receptor is pivotal to an activity which characterizes the macrophage phenotype, namely phagocytosis (2).
Recently, the mannose receptor has also been found to serve as a marker of macrophage differentiation. 1,25-Dihydroxyvitamin D, an agent known to promote monocytic differentiation of leukemic cells (3), also accelerates mannose receptor expression (4). Taken with evidence from others that the mannose receptor is differentiation-dependent (5), our findings indicate that this membrane-residing protein may be used as a hallmark of macrophage maturation. Furthermore, the apparent association of mannose-receptor expression and macrophage differentiation raises the possibility that both events may be functionally related. Hence, an understanding of the means by which developing mononuclear phagocytes express this protein may yield important clues into the fundamentals of macrophage differentiation.
We found, during our prior studies, that appearance of the mannose receptor on bone marrow macrophage precursors increases with time in culture (4). While this finding may reflect the differentiation-associated properties of the receptor, they also raise the possibility that its expression is regulated by extracellular (medium-contained) components which are progressively modified by the developing cell. In fact, we demonstrate herein that a serum-residing protein factor "down-regulates" plasma membrane expression of the mannose receptor and that bone marrow macrophages progressively inactivate this inhibitory agent, thereby leading to enhanced binding of mannosylated radioligand. MATERIALS  The abbreviations used are: BSA, bovine serum albumin; lz5I-Man-BSA, '251-mannosylated bovine serum albumin; CSF-1, colony stimulating factor-1; MEM, minimum Eagle's medium; PBS, phosphate-buffered saline: Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid. media and stirred overnight. The supernatant was recovered and the gel rinsed twice with 3 mM N a p 0 4 (pH 6.5) and once with 94 mM Na3P04. Supernatants were collected from the high molarity phosphate rinses and dialyzed against deionized water. The Lowry assay was used for protein determination (7). Final specific activities were typically lo6 units/mg protein.
Marrow Cells-Nonadherent cells were obtained from bone marrow cultures of 9-12-week-old male A/J mice (Jackson Labs, Benton Harbor, MI) and prepared as described previously (4). The ends of freshly harvested femurs were excised, and the cells collected by flushing the medullary cavity with ice-cold a-MEM through a 25gauge needle. The marrow plug was dispersed by several passages through an 18-gauge needle, and the cells were pelleted (800 x g for 7 min at 4 "C). The pellet was resuspended in ice-cold a-MEM and the number of nucleated cells determined by counting an aliquot of the resuspended cells in 2% acetic acid. Cells (1 X lo6 cells/ml) were then seeded into tissue culture dishes (Falcon Plastics) at a density of 3.4 X lo5 cells/cm2 in the presence of complete medium containing 900 units/ml of Stage I CSF-1. After a 24-h incubation, nonadherent cells were collected and the adherent cells discarded. The nonadherent population was then pelleted (800 X g for 7 min at 4 "C), resuspended in 1 ml of Pronase solution (0.02% w/v Pronase, B grade, Calbiochem, 1.5 mM EDTA in PBS/107 nucleated cells) and incubated for 15 min at 37 "C. Pronase treatment was stopped by the addition of horse serum (0.2 m1/10 ml Pronase solution), and the suspension layered onto 15 ml of ice-cold horse serum and incubated on ice for 15 min. The cell suspension was then pelleted (1200 X g for 7 min at 4 "C), resuspended in complete medium and nucleated cell number determined. This group of cells, which is 24 h post-bone marrow cell isolation and freshly recovered from Pronase treatment, is designated "marrow cells." When these cells are cultured in the presence of 500 units/ml CSF-1 for 7 days, all colony forming units are positive for the monocyte-specific enzyme, a-napthyl acetate esterase.
Adherent Bone Marrow-derived Mononuclear Phagocytes-Marrow cells were cultured at the concentrations indicated in 24-well plates (NUNC) in 2 ml of complete medium/well. After designated periods, adherent cells were used for cell cycle analysis and binding studies.
Adherent Cell Counts (6)-Adherent mononuclear phagocytes were washed with cold PBS, detached from the tissue culture plastic after a 5-min incubation at 20 "C in 0.005% Zwittergent (Calbiochem), and counted by hemocytometer.
Analysis of DNA Contentper Nuclei-Adherent mononuclear phagocytes in 24-well plates were rinsed three times with PBS and 1.0 ml of Krishan's reagent (9). After the cells detached, they were harvested and rinsed twice with fresh Krishan's reagent. The resulting nuclei were placed on ice and DNA content per nucleus determined in a Coulter EPICS V fluorescence-activated cell sorter.
PHlThymidine Incorporation Assay-Adherent mononuclear phagocytes were pulsed with 50 pCi of [3H]thymidine (ICN) and incubated at 37 "C. After a 60-min incubation, the cells were rinsed with PBS and incubated (37 "C) for 30 min in 10% trichloroacetic acid. They were then re-rinsed with an ethano1:ether (3:l) solution and extracted for counting in 0.1 N NaOH.
Preparation of Control-and Macrophage-conditioned Medium-Marrow cells were plated into 24-well plates at 1.5 X lo6 cells/ml and 2 ml/well in complete medium. After 48 h in culture, the macrophageconditioned medium from each well was collected, concentrated 1:lO (Amicon-YM5), and stored at 4 "C. Control-conditioned medium was generated in a similar but cell-free manner. Mannose-Bouine Serum Albumin Iodination (10)-100 pg of mannose-bovine serum albumin (Man-BSA) were mixed with 1 mCi of NalZ5I (Amersham Corp.) and 300 pg of chloramine T in 80 pl of 0.1 M Na3P04 buffer (pH 7.6). The reaction was terminated after 10 min on ice by addition of 190 pl of sodium metabisulfate (2.4 mg/ml) and 190 p1 of potassium iodine (10 mg/ml). The sample was then run on a Sephadex G-50 column (1 X 20 cm) buffered in 10 mM Tris-HCL (pH 7.5). 0.4-ml samples were collected and active fractions identified by y counting. Protein determination was performed by the Miller method (11) and specific activity was typically 5-8 X IO6 cpm/pg Man-BSA with >95% of total counts trichloroacetic acid-precipitable. Ligand was used within 2 weeks of iodination.
lZ51-Man-BSA Binding Assay-Binding determinations at 4 "C of lZ51-Man-BSA to adherent bone marrow mononuclear phagocytes involved slight modifications of techniques previously described (10).
Such procedures result in nonspecific binding representing 10-20% of total cell-associated counts.
The cells were washed three times (0.4 ml/well/wash) with HHBG (Hank's Balanced Salt Solution, 10 mM Hepes, 10 mM Tris, 0.1% glucose, and 10 mg/ml BSA, pH 7.1) and incubated with 0.2 ml of various concentrations of lZ5I-Man-BSA in HHBG plus 0.2 ml of HHBG & 4 mg/ml mannan (final volume 0.4 ml/well). Equilibrium binding was achieved after a 48-h incubation, and the level of cellassociated ligand determined. After 48 h, the incubation medium was aspirated, and cell layers quickly rinsed six times with Hank's Balanced Salt Solution. Cells were dissolved in 1.0 N NaOH (0.5 ml/ well) and cell-bound radioactivity of NaOH-solubilized material was measured by y counting. Duplicate values were determined for all binding points. Transformation of binding data to determine dissociation constants and estimate the number of available binding sites was performed by methods of Scatchard (12).
Evaluation of the Influence of Conditioned Medium and Fetal Calf Serum on lZ5I-Man-BSA Binding-Marrow cells were plated at 0.3 X lo6 cells/ml in 24-well plates (1 ml/well). Twenty-four hours later, various aliquots of either macrophage-or sham (control)-conditioned medium or fetal calf serum were added to appropriate wells. After the indicated additional culture period (12-36 h), cells were placed at 4 "C and the specific binding of lZ5I-Man-BSA per mg of cell protein determined by incubation with 2 pg/ml "'1-Man-BSA & 2 mg/ml mannan. Cell-associated binding was assessed by the method of Lowry (7) and standardized per unit of protein.
CSF-1 Levels-Radioimmunoassay determinations of CSF-1 levels were kindly performed by E. R. Stanley

(Albert Einstein Medical
Center, Bronx, NY). Ammonium Sulfate Precipitation-All ammonium sulfate samples were precipitated in a percent-to-percent methodology described by Green and Hughes (13). The precipitated samples were dissolved in PBS, which had been diluted 1:lO at a concentration 10 times that of the original sample. The solutions were dialyzed against a-MEM prior to use.
Acid Actiuation-The 20-40% (NH4)2SO4-precipitable fraction of fetal calf serum was redissolved and concentrated 10 times in PBS and dialyzed against a-MEM. The solution was then titrated to pH 2.0 with 1 N HC1 for 5 min at 4 "C, during which time a precipitate formed which was removed by microfugation (12,000 X g for 5 min). The supernatant was aspirated and left at 4 "C for 1 week after which it was dialyzed against a-MEM overnight. Control material was treated identically except for acidification. A suboptimal down-regulating volume (10 pl) of material was tested for its effects on lZ5I-Man-BSA binding.

RESULTS
Effect of Cell Number on the Binding of lZ5I-Man-BSA by Bone Marrow Mononuclear Phagocytes-We recently reported that lZ5I-Man-BSA binding by bone marrow mononuclear phagocytes increases with time in culture (4). Since these cells are actively dividing, we explored the possibility that such binding actually relates to cell number and not culture duration. To this end, we plated various concentrations of marrow cells (1.5, 7.5, and 15 X lo6 cells/ml) and after 2 days assessed lZ5I-Man-BSA binding at 4 "C.
As seen in Fig. L4 and 20 x lo4 receptors/cell, respectively. These data demonstrate that mannose receptor expression is influenced by cell number and not duration of culture. Effect of Cell Density on CSF-1 Levels and Cell Cycling-Having determined that lZ51-Man-BSA binding per cell increases with cell number, we next sought to identify distinct characteristics of high and low density populations which may alter radioligand binding. Two parameters which we consid-  Table I demonstrates that at high cell density 1) ambient levels of CSF-1 are lower (106 versus 344) and 2) fewer cells are in interphase (10% high density versus 30% low density).
As '251-Man-BSA binding increases dramatically with increasing cell density (Fig. l ) , these data raised the possibility that CSF-1 levels and/or cell cycling may dictate lZ5I-Man-BSA binding.
Characterizing a Model to Isolate Cycling and Noncycling Populations of Bone Marrow-derived Mononuclear Phagocytes-With the intent of ultimately defining the role of cell cycling in regulating '251-Man-BSA binding, we adopted previously described techniques designed to yield populations of bone marrow-derived macrophages enriched with quiescent (GoGI phase) or cycling (S-phase) cells (14). Such methods are predicated on the finding that cell survival and proliferation are dependent on the macrophage growth factor, CSF-1 (8). To obtain quiescent cells, we determined that 50 units of CSF-l/ml is sufficient to maintain marrow macrophage survival without stimulating mitogenesis (data not shown). More specifically, 24 h after treating asynchronously dividing cells with 50 units of CSF-l/ml, we found a greater than 90% decrease in [3H]thymidine incorporation and more than 90% of the cells in GoGl (i.e. pre-DNA synthetic) phase by flow cytometry analysis (Fig. 2, insert A ) . These data characterize such cells as a population of noncycling, bone marrow-derived macrophages.
With the capacity to produce quiescent cells in hand, we turned to the isolation of their cycling counterparts. To this end, we treated quiescent cells with increasing quantities of CSF-1 and determined that 1000 units/ml induces maximal mitogenesis (data not shown). Further, as shown in Fig. 2, when quiescent cells (point A ) are exposed to 1000 units of CSF-1, they exhibit a time-dependent increase in [3H]thymidine incorporation. Most notably, after addition of the mitogen (Fig. 2, point A), [3H]TdR incorporation, and thus entry into S-phase, commences after 12-16 h with the peak in DNA synthesis occurring at 36 h. Accompanying inserts of analysis of DNA content per nucleus estimate that after a 36-h exposure to 1000 units/ml CSF-1, 45% of bone marrow macrophages are in S-phase (Fig. 2, insert D) as compared to less than 5% exposed for the same time period to 50 units/ml CSF-1 (Fig. 2, insert C).  Fig. 3. The data demonstrate that regardless of cell cycling dissociation constants as well as estimated number of available binding sites are similar.
Influence of CSF-1 on lZ5I-Man-BSA Binding by Bone Marrow Macrophages-We next sought to determine if CSF-1 influences expression of the lZ5I-Man-BSA binding site. Hence, we cultured quiescent cells (Fig. 2, point A ) with either 1000 units of CSF-l/ml, or 50 units of CSF-l/ml, and after 8 h (Fig. 2, point B) assessed their binding of lZ5I-Man-BSA. Fig. 4 demonstrates that both binding affinity and capacity are nearly indistinguishable regardless of CSF-1 concentration. It should be noted that since the cells in this particular experiment bound relatively high levels of lZ5I-Man-BSA (1 ng/105 cells), we also examined the influence of CSF-1 on those expressing fewer receptors (i.e. low density cultures) and found that in this circumstance CSF-1 also fails to impact on lZ5I-Man-BSA binding (data not shown).
The Influence of Medium-contained Factors on lZ5I-Man-BSA Binding-To evaluate the possibility that bone marrow macrophages may alter marrow binding by modifying their environment, we incubated complete culture medium for 48 h in either the presence of bone marrow macrophages (macrophage-conditioned medium) or in their absence (controlconditioned medium) concentrated 10-fold. One hundred p1 of either macrophage-or control-conditioned medium was then added to bone marrow macrophage cultures, and after 12, 24, or 36 h, '251-Man-BSA bound per mg cell protein was determined at a saturating dose of the ligand (2 pg/ml). Results displayed in Table I1 document that the specific binding of lZ5I-Man-BSA by cells treated with either control-

Effect of control versus macrophage conditioned medium on
Iz5I-Man-BSA binding After 24 h in culture, cells were treated with 1OO-gl aliquots of either control-conditioned or macrophage-conditioned medium. At the times indicated, Iz5I-Man-BSA (2 pg/ml) was added at 4 "C and nanograms bound/mg of cell protein determined. All values shown are the means of auadruulicate determinations. or macrophage-conditioned medium increases in a stepwise fashion with time in culture. More importantly, however, those cells exposed to macrophage-conditioned medium bind more radioligand at each time point than do their sham medium-treated counterparts. Fig. 5 documents that the enhanced binding capacity of the macrophage-conditioned medium-exposed cells reflects a greater than 2-fold increase in the number of sites per cell with unaltered affinity. These differences in lZ5I-Man-BSA binding capacity could, however, represent either "up-regulation" of available binding sites by macrophage-conditioned medium or "down-regulation" by control-conditioned medium. To resolve this issue, cells were exposed to increasing aliquots of either form of medium for 36 h. As before, lZ5I-Man-BSA specific binding was determined at a saturating concentration of the ligand (2 pg/ml). The results of this experiment demonstrate that macrophages incubated with control-conditioned medium bind progressively less lZ5I-Man-BSA (Fig. 6). Specifically, untreated cells bind 103 f 11 ng of '251-Man-BSA/mg protein, those exposed to 150 pl of control-conditioned medium, only 23 f 8 ng/mg protein, and macrophages incubated with up to 150-p1 aliquots of macrophage-conditioned medium bind as much radioligand (96 f 15 ng/mg protein) as do virgin cells.
Influence of Fetal Calf Serum on lZ5I-Man-BSA Binding-We next turned to identifying the inhibitor of lZ5I-Man-BSA binding present in control-conditioned culture medium, which conceivably could include components of modified Eagle's Medium, Stage I CSF-1, or fetal calf serum. Data not shown demonstrate that the inhibitory factor is maintained within M, 14,000 exclusion dialysis tubing, precluding the possibility that the factor of interest is modified Eagle's Medium.
In addition, data presented under "Influence of CSF-1 on lZ5I-Man-BSA Binding by Bone Marrow Macrophages," excludes the possibility that CSF-1 or any other element of our Stage I CSF-1 preparation is responsible for regulation of lZ5I-Man-BSA binding.
Consequently, we explored the effect of fetal calf serum on lZ5I-Man-BSA binding. Thus, marrow cells were cultured for 24 h in complete medium followed by addition of various volumes of 10-fold concentrated fetal calf serum (0-100 pl). After an additional 36 h, lZ5I-Man-BSA binding was determined at a saturating quantity of the ligand (2 pg/ml).
As shown in Fig. 7, lZ5I-Man-BSA binding by bone marrow macrophages falls with increasing volumes of serum. In fact, whereas nonsupplemented cells bind 43 f 3 ng of '"I-Man-BSA/mg of protein, those treated with 100 pl of concentrated calf serum bind only 27 k 6 ng/mg protein.
These observations raised the possibility that fetal calf serum simply competes for available 'T-Man-BSA binding sites and is, in effect, acting as cold ligand. This concern was investigated by simply preincubating bone marrow macro- phages at 4 "C in 1 ml of binding buffer (HHBG) plus 0-100 pl of fetal calf serum for 6 h, rinsing the wells, and then measuring lZ5I-Man-BSA binding per mg cell protein by our standard assay.
Data shown in Fig. 8 document that fetal calf serum does not compete with lZ5I-Man-BSA for available binding sites. Taken together, these findings demonstrate that a component of fetal calf serum truly "down-regulates" lZ5I-Man-BSA binding sites.

Comparison of 20-40% Ammonium Sulfate-precipitable Fractions in Control and Macrophage-conditioned Medium-
Because the 20-40% (NH)2S04-precipitable fraction of fetal calf serum appeared to contain the component of culture medium responsible for down-regulating the mannose receptor, we examined the effect of a similar fraction from controland macrophage-conditioned medium on expression of the mannose binding site. We found that cells cultured with the 20-40% fraction from control-conditioned medium bind 16 Ifr 4 ng of lZ5I-Man-BSA/mg of protein which those exposed to  serum fraction on '25Z-Man-BSA binding 100 pl of reconstituted and dialyzed 20-40% (NH&S04-precipitable serum fractions, either untreated or exposed to 10 pg/ml of ribonuclease A, a quantity capable of degrading at least 15 pg of total RNA (data not shown), for 1 h at 37 "C, were added to cells cultured for 48 h. Forty-eight hours later, the cells were incubated with a saturating quantity of "%Man-BSA (2 pg/ml) at 4 "C and specific binding per well determined. Negative control represents cultures to which ribonuclease A but no (NH4)2S04-precipitable protein was added. similarly treated macrophage-conditioned medium bind 28 f 6 ng (p < 0.05) (Fig. 10). These data demonstrate that bone marrow-derived mononuclear phagocytes inactivate a medium-residing protein factor contained in the 20-40% (NH)2S04-precipitable fraction of fetal calf serum responsible

TABLE V Effect of freezing and thawing of 20-40% fNH4)2S04-precipitable serum fraction on "'I-Man-BSA binding
100 pl of reconstituted and dialyzed 20-40% (NH4)2S04-precipitable serum fractions, either untreated or frozen and thawed, were added to cells precultured for 48 h. Forty-eight hours later, the cells were incubated with a saturating quantity of I2'I-Man-BSA (2 pg/ml) at 4 "C and specific binding per well determined. Negative control represents cultures to which no (NH4)2S04-precipitable protein was added.    10 pl of reconstituted and dialyzed 20-40% (NH4)2S04-precipitable serum fractions, either untreated or maintained at pH 2.0 for 1 week, were added to cells precultured for 48 h. Forty-eight hours later the cells were incubated with a saturating quantity of "'I-Man-BSA (2 pg/ml) at 4 "C and specific binding per well determined. Negative control represents cultures to which no (NH&S04-precipitable protein was added. Pre-acidified

5,372
for down-regulation of '251-Man-BSA binding sites. Partial Characterization of Mannose Receptor Down-reguluting Factor-Initial characterization of the serum factor responsible for down-regulating the mannose receptor was performed on the 20-40% (NH4),S04-precipitable fraction of fetal calf serum. The inhibitory activity is trypsin-sensitive in a dose-dependent manner (Table 111) but ribonuclease Aresistant (Table IV). It also resists freezing and thawing ( Table V) and temperatures of at least 80 "C, but is inactivated by boiling (Table VI).
We next turned to the effects of acidification on inhibitory activity. Thus, we exposed macrophages to 10% optimal inhibitory volumes (10 p1) of the 20-40% (NH4)2S04-precipitable fraction of fetal calf serum resulting in an approximately 50% decrease in lZ5I-Man-BSA binding. When the same vol-ume of material was subjected to a pH of 2.0 for 1 week and then dialyzed against a-MEM, however, down-regulating activity increased more than 6-fold (Table VII).

DISCUSSION
The mannose receptor is among the most studied of the macrophage surface proteins and, based upon its enhanced expression with time in culture, and its induction by 1,25dihydroxyvitamin D, it appears to be a marker of monocytic differentiation. Furthermore, its appearance parri passu with monocytic maturation suggests that the phagocytic capacity of macrophages is indeed a developmentally related event. Thus, with the potential importance of progressive expression of the mannose receptor in mind, we explored the means by which bone marrow-derived mononuclear phagocytes bind increasing quantities of 1251-Man-BSA with time.
Our first step was to examine the possibility that the timerelated increase in radioligand binding reflects an event other than duration of culture. As the cells are actively dividing, we questioned whether the enhanced appearance of the mannose receptor is a manifestation of increased cell number. To this end, we determined the relative numbers of available '"I-Man-BSA binding sites at three different cell densities on the same day of culture. If mannose receptor expression is related only to duration of culture, then one would expect each macrophage to display similar lZ5I-Man-BSA binding, regardless of cell density. We found, however, that after 48 h, mannose receptor expression per cell mirrors cell number (Fig. 1). Hence, the previously reported phenomenon of increased mannose receptor expression with duration (4) appears to reflect the influence of events sensitive to cell density and not necessarily time.
Two circumstances which are usually cell density-dependent and can impact on ligand binding by macrophages are entry into interphase and exposure to the macrophage-specific growth factor, CSF-1 (8,13). We found, in fact, that similar to mannose receptor expression, both the distribution of cells in interphase and the attendant levels of CSF-1 vary with cell number (Table I). With this information in hand, we developed a method for simultaneously altering the cell cycling patterns of bone marrow macrophages (Fig. 2). Thus, asynchronously dividing cells were placed in GoGl by exposure to low levels of CSF-1 for 24 h. These quiescent and synchronous cells were then recultured in either low or mitogenic concentrations of the macrophage growth factor for an additional 24 h yielding noncycling or cycling populations, respectively. Utilizing this system, we found mannose receptor expression to be uninfluenced by entry into interphase (Fig. 3) or ambient levels of CSF-1 (Fig. 4).
Having determined that neither entry into the cell cycle nor exposure to CSF-1 are responsible for density-dependent expression of the mannose receptor, we examined the possibility that the phenomenon reflects macrophage-mediated modification of the extracellular environment. These experiments involved exposure of bone marrow mononuclear phagocytes to aliquots of macrophage-conditioned medium which led to enhanced lZ5I-Man-BSA binding when compared to control-conditioned medium-treated cells (Table 11, Fig. 5).
We also noted that this differential expression of radioligand binding reflects down-regulation of the mannose receptor by cells exposed to control medium (Fig. 6). Furthermore, the down-regulating activity is localized to the protein fraction of both fetal calf serum and control medium precipitable with 20-40% (NH4)&04 (Fig. 9). The fact that the same fraction recovered from macrophage-conditioned medium does not inhibit mannose receptor expression indicates that the cells progressively inactivate the down-regulating factor. Failure of pretreatment with fetal calf serum at 4 "C to reduce radioligand binding indicates that our observations do not merely reflect competition for the mannose receptor by a nonradioactive moiety.
Finally, we have begun to characterize the down-regulating factor. It is a protease-sensitive, acid-activable moiety with what appears to be exceptionally stable characteristics. For example, the activity resists freezing, thawing, and heating to at least 80 "C.
Thus, our experiments document the presence of a serumresiding protein factor capable of down-regulating membrane expression of the mannose receptor and that macrophages are capable of inactivating this regulatory protein(s). These findings raise important issues regarding the phagocytic capacity of macrophages and underscore the importance of carefully defining culture conditions in studies involving expression of this surface binding site.