The Interaction of 1251-C~lony-stimulating Factor- 1 with Bone

THE JOURNAL OF BIOLOGICAL CHEMtSTRY 0 1986 by The American Society of Biological Chemists, Inc.

VOl. , 261, No. 9, Issue of March 25, pp. 40244032 1986 Printed in I~.s.A.

The Interaction of 1251-C~lony-stimulating Factor- 1 with Bone Marrow-derived Macrophages*

(Received for publication, August 19,1985)

Larry J. GuilbertS and E. Richard Stanley From the Albert Einstein College of Medicine, Bronx, New York 10461

The colony-stimulating factor, CSF-1, stimulates cultured quiescent murine bone marrow-derived macrophages (BMM) to enter DNA synthesis with a lag phase of 10-12 h. The binding, dissociation, internalization, and degradation of lZ5I-CSF- 1 by BMM during the lag phase were investigated. Quiescent BMM ex- press -5 x lo4 cell surface receptor sites/cell but contain additional cryptic sites (-lO"/cell) that can appear at the cell surface within 10 min at 37 "C. Studies of the binding reaction at both 2 OC (& I 2 x 1O-l' M) and 37 O C (& -4 X 10"' M) are consistent with the existence of a single class of cell surface sites. The disappearance of cell surface 12SI-CSF-1 following a 2-37 OC temperature shift results from two, competitive, first order processes, internalization and dissociation. Internalization (tllz = 1.6 min) is 6 times more frequent than dissociation ($Ilz = 9.6 min). Following internalization, 10-15% of the intracellular CSF-1 is rapidly degraded whereas the remaining 85-90% is slowly degraded by a chloroquin-sensitive first order process (tl,z > 3.5 h). These findings were confirmed and extended by studies of the uptake of 12SI-CSF-1 at 37 "C. Following addition of 12BI-CSF-1, cell surface receptors are rapidly down-regulated (tllz - 7 min) and their replacement does not commence until 20-60% of pre-existing surface receptor sites have disappeared. Despite receptor replacement, initially from the cryptic pool and later by de novo synthesis and/or receptor recycling (4 molecules/cell/s at steady state), the number of receptors at the cell surface remains low. The process results in the intracellular accumulation of large amounts of lZ6I-CSF-1 (>lo5 molecules/cell) by BMM. Thus, whereas the kinetics of association, dissociation, and internalization of CSF-1 with BMM and peritoneal exudate macrophages are similar, BMM, which exhibit a higher proliferative response, degrade growth factor 12 times more slowly.

-

CSF-1' is a growth factor which specifically stimulates the

* This work was supported by Grant CA26504 from the National Cancer Institute and the Albert Einstein College of Medicine Cancer Center Core grant. 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.

$Present address: Dept. of Immunology, University of Alberta, Edmonton, Alberta, Canada T6G 2H7.

CSF-1, the colony-stimulating factor subclass, defined by subclass-specific radioimmuno- and radioreceptor assays, that specifically stimulates mononuclear phagocyte proliferation (1); PEM, starch-induced peritoneal exudate macrophages; EGF, epidermal growth factor; BMM, bone marrow-derived macrophages; IMDM, Iscove's modified Dulbecco's medium; FCS, fetal calf serum; IB, IMDM without bicarbonate; pH 7.35; PBS, phosphate-buffered sa- line, 150 mM NaCl, 5 mM phosphate, pH 7.35; msr, molar specific radioactivity.

survival, proliferation, and differentiation of mononuclear phagocytes (colony-forming unit 4 monoblast * promono- cyte + monocyte +- macrophage) in tissue culture (reviewed in Ref. 1). The response to CSF-1 is pleiotropic but its nature is dependent on cell type. For example, CSF-1 regulates the survival of non-dividing macrophages but the survival, proliferation, and differentiation of the colony-forming unit (1, 2). Responses to CSF-1 are regulated by the CSF-1 receptor, a tyrosine kinase of M, - 165,000, which in membrane preparations of responsive cells mediates the CSF-1-induced phos- phorylation of several membrane proteins? The receptor is immunologically and functionally related to the product of the proto-oncogene c-fms (3).

Initial studies of the CSF-1-mononuclear phagocyte interaction were carried out with peritoneal exudate macrophages (PEM) which could be obtained in larger numbers and as more homogenous populations than developmentally earlier cells (4-6). However, fewer than 50% of quiescent PEM are capable of initiating DNA synthesis in response to CSF-1 and the time taken for all these cells to enter S phase is slow (-3 days, (7)). Despite this cellular heterogeneity and slow response, the binding of lZ5I-CSF-1 to PEM at 2 "C is consistent with existence of a single class of high affinity cell surface receptors. At 37 "C, 9/10 of the surface-bound 1251-CSF-1 molecules are internalized and the internalized 'z51-CSF-1 degraded intracellularly as a single reactant pool (4-6,8, 31). Thus the binding, uptake, and degradation of '251-CSF-1 by PEM is homogeneous.

Homogeneity in the handling of a growth factor by responsive cells is of interest in view of recent reports of functionally relevant heterogeneity in the target cell binding and intracellular compartmentalization of other growth factors. Unique classes of high affinity celI surface receptors for EGF and nerve growth factor appear to direct uptake of these factors to relatively stable intracellular compartments where their presence correlates with biological response (9-12). The CSF- 1-mononuclear phagocyte interaction may differ fundamen- tally from the interaction of these other growth factors with their target cells in its homogeneity. Alternatively, the observed homogeneity of the CSF-1-PEM interaction may be correlated with the relatively poor proliferative response of

This question has been addressed by studies of the interaction of CSF-1 with bone marrow-derived macrophages (BMM, (2)). BMM are a more homogeneous, developmentally earlier population with greater proliferative capacity than PEM. In contrast to PEM, 93-98% of BMM are cycling when cultured in the presence of CSF-1 and 280% of BMM rendered quiescent by removal of CSF-1 enter DNA synthesis synchronously within 24 h of CSF-1 addition (13). AS was

'Y. G. Yeung, P. T. Jubinsky, and E. R. Stanley, unpublished

PEM to CSF-1.

observations.

4024

Mononuclear Phagocyte Growth Factor Uptake and Destruction 4025

observed with PEM, results of the binding of 1251-CSF-1 to BMM at both 2 and 37 "C was consistent with the existence of a single class of binding sites. However, in contrast to the more mature and more slowly proliferating PEM population, BMM accumulate large amounts of internalized CSF-1 in a compartment in which it slowly decays.

MATERIALS AND METHODS

CSF-1 Preparations-The preparation of purified L-cell conditioned medium (8 x IO7 units/mg of protein) and of purified lZ5I-CSF- 1 (-400,000 cpm/ng of protein) (14) and the determination of the molar concentration and molar specific radioactivity (msr) of stock solutions were carried out as described (31). One unit of CSF-1 = 4.4 X mol (13).

Solutions and Media-IMDM was obtained from GIBCO and FCS from Flow Laboratories. IB was IMDM lacking bicarbonate, pH 7.35.

. ,

Starting medium was 15% FCS in IMDM containing 10% (v/v) medium conditioned by serum-free cultures of the myelomonocytic leukemic cell line WEHI-3 (15) and 1000 units/ml of stage 1 murine L-cell CSF-l(6). Growth medium was 15% FCS in IMDM containing 2000 units/ml stage 1 L-cell CSF-1. The pH 4.0 dissociation buffer was 0.1% (v/v) FCS in 100 mM acetic acid, 120 mM NaCl, pH 4.0.

BMM Preparation and Culture-BMM were prepared by a modification of a previously reported method (2). Briefly, a single cell suspension of femoral and tibial bone marrow cells (lo6 nucleated cells/ml) was cultured in startiig medium for 24 h. The nonadherent cells were collected, pronase-treated (15 min at 37 "C) in 9 p.u.k./ml Pronase (B grade, Calbiochem-Behring, in PBS) to break up clumps, washed by centrifugation through FCS, and recultured in fresh starting medium for another 2 days. The nonadherent cells were again collected, pronase-treated, and washed. The resulting single cell suspension was either frozen at -70 "C (in 20% (v/v) FCS, 10% dimethyl sulfoxide in IB) for later use or cultured in 2 ml of growth medium in 35-mm tissue culture dishes (Lux, 3 x 106/dish). After 2 days of culture, the nonadherent cells were discarded and the adherent cells (day 5 BMM) washed once with PBS prior to culture in fresh growth medium. Log-phase growth was maintained by periodically changing the medium, if necessary (see Ref. 2 for formula predicting the CSF- 1 consumption by BMM in log-phase growth). Log-phase day 7 BMM were rendered quiescent by washing once with PBS, replacing growth medium with FCS-IMDM, and incubating for 15 h. Quiescent day 7 BMM (2-6 X 10' cells/35-mm tissue culture dish) were used for all experiments.

Protease Treatments-Proteolysis was initiated by addition of 0.1 ml of fresh stock solution of filter-sterilized, ice-cold trypsin (Sigma, type 111) or Pronase (B-grade, 45,000 P.U.K./g, Calbiochem-Behring) in IB to rapidly cooled or prewarmed cultures containing protein-free medium. Rapid cooling was carried out as described under '"I-CSF- 1 binding at 2 "C (see below). The reaction was stopped by rapid washing with FCS followed by either ice-cold or pre-warmed PBS, depending on the temperature required for the next procedure. Com- plete CSF-1 receptor loss was observed within 10 min of incubation with 0.1% trypsin at 2 "C.

'251-CSF-l Binding at 2 "C-Dishes at 37 "C were rapidly washed with ice-cold PBS and transferred to damp tissue paper on ice, 1 ml of ice-cold FCS-IB was added, and the dishes were incubated for 20 f 1 min prior to commencement of binding. The binding reaction was started by addition of 0.1 ml of ice-cold '"I-CSF-1 in FCS-IB to cooled cells in 1.0 ml of FCS-IB (final 12'I-CSF-1 concentration, 1 nM). The reaction was stopped by rapidly washing the dishes 5 times with 2 ml of ice-cold PBS (total volume, 10 ml). Cells were removed from the dish for determination of radioactivity by addition of 0.1 M NaOH (see below). Net high affinity '"I-CSF-1 binding is the binding that is blocked by a 2-h preincubation with 2 nM unlabeled CSF-1 (6). The net high affinity binding following a 2-h incubation with 1 nM lZ5I-CSF-1 was used to determine the total number of cell surface binding sites.

'25Z-CSF-l Uptake at 37 "C-Dishes at 37 or 2 "C were rapidly washed with 37 "C PBS, given prewarmed and gassed (37 "C, 5% COz in air) FCS-IMDM (1.0 ml), and immediately transferred to prewarmed 150-mm tissue culture dishes containing damp tissue paper that were kept in a COP-regulated, fully humidified 37 "C incubator. The dishes were incubated for 20 f 1 min prior to CSF-1, '=I-CSF- 1, or 13'I-CSF-1 addition. Uptake was started by addition of less than 0.1 ml of '"I-CSF-1 in FCS-IMDM at 37 "C to prewarmed cells (final '"I-CSF-1 concentration, 300 p ~ ) . Uptake was stopped and cellular

radioactivity removed for counting as described for lBI-CSF-l binding at 2 "C. Net high affinity '"I-CSF-1 uptake is defined as the uptake that is blocked by a 1-h preincubation, and further coincubation, with IO-* M unlabeled CSF-1.

Analysis of Cell-associated '25Z-CSF-l-Cellrassociated '"I and DNA were removed from the washed dishes by overnight incubation at room temperature with 1 ml of 0.1 M NaOH. The contents of each dish were repeatedly pipetted up and down in a Pasteur pipette to reduce viscosity prior to transfer to a glass tube, together with two 0.5-ml washes of 0.1 M NaOH. The '"I content of the pooled extract was determined on a y counter and the DNA content of the same extract determined by a modification (2) of the method of Hinegard- ner (16). The number of cell-associated '"I-CSF-1 molecules was determined using the formula: molecules/cell = A X B/(msr X C ) where A = Avagadro's number (6.02 X loz3 molecules/mol), B = net high affinity '"I-CSF-1 bound (counts/min/Ng of DNA), C = cell number/pgof DNA = 1.7 X lo5 for quiescent and 1.3 X lo5 for growing BMM (13), and msr = molar specific radioactivity = F, X I/m (F, = fraction of stock '"I-CSF-1 preparation bound by anti-CSF-1 antibody, I = counts/min "'I/ml of the stock '"I-CSF-1 solution, and m = the number of moles of CSF-l/ml of stock '"I-CSF-1 solution determined by '311-CSF-1 radioimmunoassay (6). The msr of 13'1- CSF-1 stock solutions was similarly derived.

Determination of Cell Surface-bound '"I-CSF-I-Cells from 37 "C incubations were washed and rapidly cooled as described above for '"I-CSF-1 binding at 2 "C. Medium was removed from cells a t 2 "C and replaced with 1 ml of ice-cold pH 4 dissociation buffer. Following a 20-min incubation on ice, the supernatant was saved and pooled with two 0.5-ml washes. Cells were removed from the dish with 0.1 M NaOH for 7 counting and DNA analysis. The "'I in the pH 4 dissociated supernatant fraction was measured and normalized to the DNA content of the extract, and the fraction combining with anti- CSF-1 antibody determined as described (6).

Temperature Shift Experiment-BMM were rapidly cooled, incubated at 2 "C for 2 f 0.2 h with 360 PM '"I-CSF-1 and washed with 10 mi of ice-cold PBS as described above for lZ6I-CSF-1 binding at 2 "C. The washed cells were rapidly shifted from 2 to 37 "C by floating the cold dishes containing the cells into a shallow 37 "C water bath and immediately adding 1 ml of prewarmed (37 "C) FCS-IB. At the time points indicated (20 s to 22 h) supernatants from replicate dishes were removed and placed on ice for later analysis. They were replaced, in most dishes, with 1 ml of fresh 37 "C FCS-IB and the incubation was continued. The supernatants removed at each time point were analyzed for total '"I (Ti, counts/min) by y counting and for intact lBI-CSF-l (Pi, counts/min) by anti-CSF-1 antibody binding (6, 14). The amount of intact "'I-CSF-1 that had dissociated during a given period was calculated as the sum of Pi for all dishes processed during that period. Similarly, the amount of '"I-CSF-1 degraded and released by the cells up to a given time point was calculated as the sum of (Ti - Pi). At each time point, one or two plates were used to determine total cell surface '''1 (St, counts/min), cell surface-bound intact "'I- CSF-1 (Si, counts/min) and internalized '"I (&), as described above. The cumulative amounts of intact and degraded '"I-CSF-1 released by cells at the indicated time points were expressed as a percentage of the amount of '%I present on the dish at 2 "C prior to temperature Shift.

Analysis of Temperature Shift Data-The simultaneous dissociation and internalization of cell surface-bound '"I-CSF-1 was treated as a system of first order reactions competing for a common reactant (17):

P-S-Z-H k~ kin kh

where S = cell surface-bound '251-CSF-l, P = intact lZ51-CSF-1 dissociated into the medium, Z = internalized lZ5I-CSF-1, H = lZ5I- CSF-1 degraded by cells and released into the medium, and kff, k, and kh are first order rate constants. The following relationships are derived from this treatment:

1) The rate of appearance of '"I-CSF-1 in the medium (dP/dt = k&) plus the rate of internalization (dZ/dt = kinS) equals the decay rate of cell surface-bound lz51-CSF-1, dS/dt = -k.S = -(& + kff)S.

2) The ratio of products a t any time during the reaction equals the ratio of rate constants (kin/koff = Z/P). The nature of one of the reactions (internalization followed by degradation and release of '"I- labeled product) allows the ratio of Z/P to be determined only when intracellular "'I-CSF-1 has been entirely degraded and released (Z =

Thus, lz. = lzi. + &E.

4026 Mononuclear Phagocyte Growth Factor Uptake and Destruction

0); i.e. a t the end of the reaction. The nature of the protocel (constant removal of dissociated '2'I-CSF-1) minimizes reutilization of dissociated lZ5I-CSF-1 during the reaction. Thus, equals the sum of degraded '"I-CSF-1 in individual samples @(Ti - Pi)) divided by the sum of intact 1251-CSF-1 in individual samples @(Pi)).

3) The individual components of a competing first order system can be displayed in linear form if the rate coefficient for reactant decay (k.) and the fractional concentration of the product(s) are known. A plot of the fractional concentration of product versus 1 - e-kd should be linear and intercept zero. The slope of the straight line equals the first order rate coefficient for the reaction plotted.

To analyze the temperature shift data, a value for k, was obtained from a first order plot of cell surface bound lZ5I-CSF-1 (Fig. 3) and &ff determined from a plot of Fp (fraction of "'I-CSF-1 bound by anti-CSF-1 antibody) versus 1 - e-ht (Fig. 5). A value for kin was then calculated as k. - &e. Appropriateness of the treatment was checked in several ways: (a) The rate of decay of cell surface-bound "'I-CSF- 1 was shown to be first order (Fig. 3), and (b) the plot of Fp uersus 1 - e-ht was linear and intercepted at the origin (Fig. 4), and (c) the ratio of calculated rate constants (k/kff) was very similar to the observed ratio of 2(Ti - Pi)/Z(Pi) (see text).

CSF-I Uptake Experiment-The uptake experiment (Figs. 5 and 6) was carried out using 37 "C BMM cultures, as described above for the "'I-CSF-1 uptake at 37 "C. The number of unoccupiedcell surface binding sites during the uptake was determined in two different ways: Cells were washed and rapidly cooled, and the net high affinity binding was determined by incubation with 1 nM 1311-CSF-1 as described above ("'I-CSF-1 binding at 2 "C). Alternatively, two uptake experiments were set up, one in 300 pM 1251-CSF-1, the other in 300 p~ CSF-1. Those dishes incubated with "'I-CSF-1 were used to determine the amounts of cell surface '9-CSF-1 and internalized '%I, and those incubated with CSF-1 for determination of the number of unoccupied cell surface sites by 1251-CSF-1 binding at 2 "C. While both methods gave similar results, data from the latter method, which reduced experimental scatter, were used.

Autoradiography-Methods for autoradiography of methanol- fixed cells in 35-mm tissue culture dishes have been described in detail elsewhere (4).

RESULTS

CSF-1 Binding Sites on BMM at 2 "C-At 2 "C, CSF-1 binds almost irreversibly to a specific and saturable cell surface site on PEM (4, 6). At this temperature, although the number of lZ5I-CSF-1 molecules bound at saturation remains stable for up to 72 h, the number of unoccupied sites decreases (tllz - 10.5 h) (6). Thus the number of sites can be accurately measured only if the rate of site occupancy is rapid relative to the rate of site loss (6). Although the number of binding sites can conveniently be determined by rapid saturation binding at high ligand concentrations, the heterogeneity of binding sites cannot be investigated by equilibrium methods because of the loss of unoccupied sites during the long incubation times necessary for attainment of equilibrium at low ligand concentrations (6). However, the number of site classes and their individual contribution to the total number of sites can be inferred from the nature of CSF-1 macrophage association kinetics. Irreversible binding under conditions such that ['251-CSF-1] >> [binding site] should display first order kinetics through several half-lives if one site class is predom- inant, and more complex kinetics if more than one class significantly contributes. The kinetic data for the binding of 1 nM 1251-CSF-1 to BMM at 2 "C is shown in Fig. 1, together with the theoretical curve based on the assumption of first order kinetics. The binding data fits the theoretical curve over more than three half-lives, consistent with the existence of a single class of binding sites.

It should be noted that artifactual results concerning binding site heterogeneity can be obtained using the kinetic method, if binding of lZ5I-CSF-1 to BMM at 2 "C occurs over an extended period of time at lower lWI-CSF-l concentrations. Under these conditions, both kinetic and protease treatment

4 I I I I I

Y I I I 0 0.5 1.0 1.5

TIME (h)

FIG. 1. Net high affinity binding of '261-CSF-1 to quiescent BMM at 2 "C. The binding reaction was carried out as described under "Materials and Methods" in the presence of 9.5 X 10-l" M '=I- CSF-1. The theoretical curve for first order binding described by the equation: counts/min bound = counts/min X e-kt, where k = 6.5 X lo-' min-l, is drawn through the points. The second order association constant, kn = k/[T-CSF-l] = 7 X lo7 M" min-'.

TABLE I The effect of trypsin pretreatment on the binding of '251_CSF-I to

BMM at 2 "C Quiescent day 7 BMM (approximately 6 X 105/35-mm dish) were

pretreated at 2 "C as indicated and the net 1251-CSF-1 binding was determined following incubation with 2.2 nM 1251-CSF-l for 2 h at 2 "C. The number of molecules of '"1-CSF-1 bound per cell was calculated from net high affinity binding at 2 "C normalized to the DNA content of the cell extract as described under "Materials and Methods."

2 "C pretreatment

0.1% 0.1% '251-CSF-1 bound Tmsin Recovery T m s i n

15 min 15 h 15 min molecules/cell - - - 49,100 f 8,670" + +

- 1,940 f 16 - 6,903 f 339

+ + + 468 f 182

- +

a Values are means of triplicate samples & S.D.

TABLE I1 Percentage of cell-associated 1251present as cell surface 1251-CSF-l

during incubation with '=I-CSF-l a t 2 and 37 "C Quiescent day 7 BMM (approximately 6 X 105/35-mm dish) were

incubated with 1.2 nM (2 "C) or 300 PM (37 "C) lZ5I-CSF-1. At the indicated times, cultures were washed 5 times with 10 ml of ice-cold PBS and cell surface '251-CSF-1 was determined by pH 4 dissociation as described under "Materials and Methods." All values are means of duplicates & range. Total cell-associated '''1 is the net high affinity counts/min/wg of DNA.

Incubation parameters Cell-associated '%I

TemDerature Time Total Cell surface "C countslmin X I O y 5 % 2 2 h 7.07 f 0.12 93.3 f 1.4

37 2.5 min 0.95 f 0.001 53.0 f 0.5 37 10.0 min 2.59 f 0.01 16.4 f 0.4 37 2 h 7.30 f 0.24 37 24 h 23.62 f 0.001 5.6 rt_ 0.3

8.6 & 0.3

data indicate that approximately 15% more sites appear by 15 h of incubation than are observed following a 2-h incubation. On appearance, these "cryptic" sites are also protease- sensitive (Table I) and kinetically indistinguishable from the sites measured by a 2-h incubation at 2 "C (data not shown). The data presented in Fig. 1 and Table I indicate that quies-


cent BMM contain approximately 50,000 cell surface receptor sites/cell.

Uptake of 1251-CSF-1 by BMM at 37 "C-In order to exam- ine the interaction of lZ51-CSF-1 with BMM at 37 "C, it was necessary to develop assays that distinguish cell surface "'1- CSF-1 from internalized lZ5I. The protease sensitivity of the unoccupied receptor sites on BMM demonstrates their cell surface location (Table I). However, protease treatment cannot be used to determine cell surface lZ5I-CSF-1 since the occupied cell surface sites are protease resistant (31). On the other hand, as in the case of PEM (31), the rapid dissociation of surface-bound "'I-CSF-1 at pH 4 could be used. Table I1 shows that >93% of lZ5I bound to BMM at 2 "C during a 2-h incubation with 12'I-CSF-1 dissociates at pH 4.

Following a 2-h incubation of BMM with lZ5I-CSF-1 at 37"C, only 9% of the cell-associated lZ5I was dissociated at pH 4 (Table 11). This suggested that lZ5I-CSF-1 is only transiently bound at the cell surface at 37 "C. Transient cell surface binding at 37 "C was confirmed in an experiment in which the percentage of the total cell-associated "'1-CSF-1 that could be dissociated at pH 4 was measured at different times after the addition of lZ5I-CSF-1 to BMM at 37 "C (Table 11). Greater than 50% of the lZ5I-CSF-1 bound by 2.5 min dissociated at pH 4. Between 2.5 min and 24 h, despite the accumulation of larger amounts of lZ5I-CSF-1 by the BMM, the dissociable fraction decreased. Thus at 37 "C, BMM appear to bind lZI-CSF-1 transiently at the cell surface and very rapidly internalize it.

The combination of rapid binding, rapid internalization, and relatively slow dissociation of the ligand renders quantitative treatment of the interaction at 37 "C complex. However, by reducing the number of simultaneous events, quantitative analysis is possible. Cells almost exclusively containing surface-bound "'1-CSF-1 can be prepared at 2 "C (Table 11) (31). By using these cells the process of 1251-CSF-1 binding to an unknown concentration of sites is removed from the analysis, and rapid increase in the temperature from 2 to 37 "C allows synchronous initiation of events determining the fate of the surface-bound lZ5I-CSF-1. Following such a temperature jump, surface-bound lWI-CSF-l can either remain bound to the surface, dissociate intact, or be internalized. All three of these alternative processes can be monitored The amount remaining at the surface by pH 4 dissociation, the dissociated '=I- CSF-1 as the fraction of supernatant '''1 combining with anti- CSF-1 antibody and the internalized lZ5I-CSF-1 as the difference between the cell-associated and the cell surface lZ5I- CSF-1. Changes in these three parameters with time are shown for a temperature shift experiment with BMM to which 1251-CSF-1 was bound by preincubation at 2 "C (Fig. 2). Fol- lowing the temperature shift, surface bound material disappeared very rapidly ( tl/z < 2 min) dropping to 5% of the initial amount bound by 1 h, then slowly decreasing to 3.5% over the next 20 h (Fig. 3). Only 13% of the total surface-bound 1251-CSF-1 dissociated rapidly as intact 12'I-CSF-1. With the exception of the 3.5% apparently bound at the surface at 22.5 h, all of the remaining lZ5I had appeared in the medium by 22.5 h. As most of the apparently surface-bound lZ5I at 22.5 h was shown autoradiographically to be bound to the tissue culture dish and not to cells (data not shown), it is clear that cell surface lZ5I-CSF-1 either dissociates as intact 1251-CSF-1 or is internalized.

The.dissociation and internalization of lZ5I-CSF-1 bound to the surface of PEM have been shown to be competitive first order processes (31). Because the disappearance of >go% of the 1251-CSF-1 bound to the surface of BMM exhibits first order kinetics (Fig: 3, rate constant for disappearance, k. =

h TIME (min)

FIG. 2. Fate of surface-bound 12'I-CSF-1 after rapid temperature shift from 2 to 37 OC. Conditions of binding at 2 "C: 2 h, 360 PM lZ5I-CSF-1, 5 x lo5 BMM/plate. All values are expressed as a percentage of the net high affinity lZ5I-CSF-1 bound at 2 "C. V, per cent at the cell surface; 0, per cent dissociated into the medium as intact 1251-CSF-1; 0, per cent released into the medium as degraded '"I-CSF-1. Values are means of 2-9 samples (9 at the earlier time points and 2 at the last point) with a mean relative standard deviation of 0.208 k 0.095.

TIME (min)

FIG. 3. First order disappearance of '2'I-CSF-1 bound to the surface of BMM. Data from Fig. 2. Ordinate: In So/& where So and S are the immunoreactive fraction of pH 4-dissociable lZ5I minus the fraction nonspecifically bound to the dish at t = 0 and the indicated times, respectively. The data fit the line drawn with a slope = & = 8.5 X s-l (correlation coefficient = 0.999).

8.5 X s-'), the entire process was also treated as a system of competing first order reactions. Such treatment allows rate constants for the competing reactions to be easily calculated (Ref. 17, see "Materials and Methods"). A plot of intact dissociated 1251-CSF-1 uersus 1 - e-&& was linear with a zero intercept indicating that the dissociation reaction is first order and that the treatment is valid (Fig. 4). The slope of this line yields the dissociation rate constant, koe = 1.2 x s-l. (For comparative purposes, the disappearance of cell surface "'1- CSF-1 is also plotted in Fig. 4.) Since the internalization rate constant, ki, plus kff equals k, it follows that kin = 7.3 X s-'. The above values can be independently verified since the ratio of rate constants should equal the ratio of products. The ratio of products is the amount of degraded to intact "'1-CSF- 1 at the end of the reaction and equals 6.1 (from Fig. 2). This very closely matches the ratio of rate constants, Izi,/koff = 7.3 X 10-3/l.2 x lop3 = 6.08. This analysis of the data presented in Fig. 2 indicates that internalization and dissociation of lZ5I- CSF-1 bound to the surface of BMM are parts of a single, internally competing first order process.

Degradation of Internalized 125Z-CSF-1-In the course of its interaction with macrophages, lZ5I-CSF-1 is intracellularly degraded to 1251-tyrosine (18). The medium above the cells during the interaction therefore contains either intact, biolog- ically active lWI-CSF-l or 1251-tyrosine (31). Since iodotyro- sine is not reutilized for de nouo protein synthesis and rapidly


I I I I

$ 1 g 20

0 0.2 0.4 0.6 0.8 1.0 1-e-"

FIG. 4. Dissociation of '"1-CSF-1 as a competitive first order process. Data from Fig. 2. Ordinate: per cent of immunoreactive cell surface-bound IZSI-CSF-l at t = 0 which either dissociates intact or disappears. 0, dissociation of immunoreactive '"I-CSF-I; abscissa, 1 - e-+ t where &s = 1.2 X lo9 s-l (correlation coefficient = 0.931). 0, disappearance of surface-bound '251-CSF-1; abscissa, 1 - e-kt where ka = b,, + kos = 8.5 X s-' (from Fig. 3).

diffuses into the medium (19), the course of intracellular degradation can be conveniently followed as the amount of supernatant lZ5I that is not bound by anti-CSF-1 antibody, The time course of 1251-CSF-1 degradation by PEM after internalization in a temperature shift experiment indicated that the process was kinetically first order with a tlIz of 16 min (Table 111) (31). Under identical conditions (following uptake of 15-25% of the maximum amount of lZ5I-CSF-1 bound by the cells at 2 "C), BMM also degrade 1251-CSF-1 with first order kinetics. However, the degradation rate is more than 12-fold slower than for PEM (tl12 = 3.5 h, Table 111). Furthermore, when approximately 4 times more lZ5I-CSF- 1 is taken up by BMM following incubation with 1251-CSF-1 at 37 "C, the rate of degradation of this larger internalized pool of lZ5I-CSF-1 was again first order with an even longer half-life. When the analysis of the degradation was carried out in the presence of CSF-1, which would have resulted in a still larger internalized pool (see Fig. 6), the degradation rate was further decreased. Thus although BMM, like PEM, degrade lSI-CSF-l intracellularly with first order kinetics, they do so much more slowly. Furthermore, the rate of degradation appears to be inversely related to the amount of lZ5I-CSF-1 internalized.

In order to further investigate the degradation of internal-

ized lZ5I-CSF-1, the analysis was carried out in the presence of the lysosomotropic amine, chloroquin, which is known to slow lysosomal proteolysis (20). Consistent with the intracellular destruction of lZ5I-CSF-l in lysosomes, chloroquin slowed the rate of degradation, but the rate remained first order (Table 111).

Accumulation of 1251-CSF-l ut 37OC"The rapid uptake (Table 11) and slow intracellular degradation (Table 111) of IZ5I-CSF-1 by BMM is reflected in the large intracellular accumulation of 1251-CSF-1 observed during long-term incubation (Table 11, Figs. 5 and 6). Within 2 h, BMM accumulate more 1251-CSF-1 than the maximum amount that can be bound by cell surface receptors on quiescent cells (Fig. 5). By 16 h, the cells accumulate -4 times more 1z251-CSF-1 and degrade -6 times more 1251-CSF-l than the molar equivalent of the original cell surface receptor (extent of degradation calculated from the level of 1251-CSF-1 in the supernatant in Fig. 6). Thus, by 16 h, BMM have internalized more than 10 times the amount of 1251-CSF-1 than can be bound by the original cell surface receptor complement.

These results indicate the existence of either binding site recycling, rapid de novo receptor synthesis, a large number of cryptic binding sites, alternate uptake mechanisms, or com- binations of these. The presence of a relatively small number of cryptic binding sites that appear after trypsin pretreatment during incubation at 2 "C has already been demonstrated (Table I). This same experimental approach was used to estimate the number of cryptic sites available at 37 "C (Table IV). The relative number of cell surface sites before and after trypsinizations or 37 "C incubations (rests) was estimated by lSI-CSF-1 binding at 2 "C. The relative 15-min, 37 "C uptake of '251-CSF-1 by the treated BMM was also determined. As previously observed (Table I), trypsin pretreatment at 2 "C removed 95% of 2 "C binding. However, the same treatment removed only half the 37 "C uptake. Since trypsin pretreatment precludes receptor recycling and the short-term 37 "C uptake minimizes a contribution of de IWUO synthesis, the difference in uptake between 2 "C and 37 "C has to be due to either cryptic sites appearing during the 37 "C uptake period or alternate (high affinity) uptake mechanisms. The reap- pearance of all of the 2 "C binding during incubation of the trypsin-pretreated cells at 37 "C for 15 min indicates that the difference is due to the appearance of a large number of cryptic sites. Trypsin treatment a t 37 "C (10 min) removed 94% of 37 "C uptake and 99% of 2 "C binding, and each successive 2 "C trypsinization reduced 37 "C uptake by 50 to

TABLE I11 Intracellular degradation of 1251-CSF-l

either for 2 h with 360 pM '251-CSF-1 at 2 "C (Experiments a and b) or for 1 h with 300 p~ '251-CSF-1 at 37 "C Data are from three separate experiments each using 5-7 X lo5 adherent macrophages. Cells were preincubated

(Experiment c) then washed with PBS, and the measurement of the degradation rate at 37 "C carried out as described under "Materials and Methods."

First-order parameters from linear region of plot of

Experiment Cells Internalized In[Zo / ( lo - &)] against timeb 1zsp Degradation medium

Linear range Correlation coefficient Half-lie

a PEM 15-25 FCS-IMDM 15 min-2 h 0.998 16.1 min b BMM 15-25 FCS-IMDM 30 min-4 h C BMM FCS-IMDM

0.997 3.5 h

C BMM 80-90

FCS-IMDM + CSF-1 1-22 h 0.999 1.2 h

80-90 1-22 h C BMM

0.999 10.0 h 80-90 FCS-IMDM + chloroquin 1-22 h 0.999 13.0 h

"Intracellular lZ5I at the beginning of the analysis of degradation rate expressed as the percentage of the maximum amount of 1z51-CSF-1 that can be bound by the cell surface binding sites on quiescent cells (see Figs. 2 and 6).

4 is the amount of '%I internalized at the commencement of the degradation rate measurements and Ht is the amount of degraded lZ5I-CSF-1 in the supernatant at various times thereafter.


15 30 45 60 120

TIME (min)

FIG. 5. Uptake of 1261-CSF-1 by quiescent BMM at 37 "C. 0, Net high affinity uptake; 0, cell surface 1251-CSF-1; A, unoccupied cell surface binding sites during the first 120 min of incubation with 300 PM 1251-CSF-1 (-6 X lo5 cells/dish, 1-ml cultures). Data points represent the means of duplicate samples and error bars denote the range.

TIME (h)

FIG. 6. Uptake of CSF-1 by quiescent BMM at 37 OC. Con- tinuation of the uptake experiment depicted in Fig. 5. 0, net high affinity uptake; A, unoccupied cell surface binding sites; m, supernatant CSF-1 levels, during initiation of DNA synthesis in the presence of 300 p~ 151-CSF-1 or CSF-1. In order to buffer the decrease in '9- CSF-1 concentration due to receptor-mediated supernatant CSF-1 degradation, dishes harvested from 0 to 4 h contained 1 ml of medium, from 4 to 12 h, 2 ml of medium, and from 12 to 24 h, 3 ml of medium. The 'Z51-CSF-1 concentration at harvest was determined by 1311-CSF- 1 radioimmunoassay in duplicate on pooled duplicate samples. Data points represent the mean from duplicate samples and error bars denote the range. In this experiment the maximum down-regulation of CSF-1 receptors was 80%. Complete down-regulation under the same conditions occurs a t a concentration of 2 nM CSF-1.

70%, indicating that the 37 "C uptake is mediated by trypsin- sensitive sites. The 37 "C rests did not greatly increase 37 "C uptake, indicating that the apparently trypsin-insensitive uptake at 37 "C was mediated only by cryptic sites rapidly appearing during the 15-min uptake .assay. The results in Table IV indicate that there are approximately twice as many cryptic as surface sites in BMM at 37 "C and that the cryptic sites can be rapidly (in 10 min) recruited to the cell surface.

DISCUSSION

The interaction of 1251-CSF-1 with BMM may be divided into three stages: the ligand binding reaction, ligand internalization, and the lysosomal degradation of internalized ligand. The formation of the CSF-1 receptor complex is the initial step in the action of CSF-1. The results of rapid saturation binding assays (Fig. 1) and sequential trypsinization studies (Table I) -show that CSF-1 binds to a single class of cell

TABLE IV '251-CSF-I binding sites on bone marrow-deriued.macrophages during

cycles of trypsinization at 2 "C and recovery at 37 "C Quiescent day 7 BMM (-6 X lo5 cells/35-mm dish) were treated

as indicated and assayed for net high affinity '251-CSF-1 binding at 2 "C and lZ51-CSF-1 uptake at 37 "C (15 min). Details are described under "Materials and Methods."

Sequence Of "' trYPsini- Cell-associated '%I after incubation zations (T) and 37 "C

incubations (R) with '%I-CSF-l at:

T R T R T 2 'C 31 "C countslminlpg DNA

" " - 51,901 f 3,775 117,724 f 6,837 + " " 2,742 f 1,223 56,586 f 2,574 + + - " 66,342 f 5,042 66,623 f 889 + + + - - 1,107 +. 289 19,025 k 2,136 + + + + - 28,830 f 294 28,650 k 1,216 + + + + + ND" 11,245 f 496 1 X 37 "C trypsiniza- 320 f 7 7,153 +. 886

ND, not done. tion (10 min)

surface receptors a t 2 "C. Furthermore, the observation that internalization and dissociation of cell surface-bound lZ5I- CSF-1 at 37 "C comprises a single internally competitive first order event (Figs. 3 and 4), is consistent with the existence of a single class of cell surface CSF-1-receptor complexes at the physiological temperature.

The very rapid internalization of cell surface-bound lZ5I- CSF-1 (Table 11) and the large number of cryptic receptors available (Table IV), precluded either equilibrium or accurate kinetic analyses of the binding reaction at 37 "C. Despite these limitations, two approximate values for kOn can be calculated from independent sets of data. One estimate of k,, at 37 "C is the product of the ratio of initial binding rates at 37 "C and 2 "C (2.3, Table IV) and the k,, at 2 "C (1.2 X lo6 M-' s-', Fig. l), or 22.7 X lo6 M-' s-'. A second estimate of k,, at 37 "C can be obtained from the uptake during the first 2.5 min (Fig. 5) which is equal to or less than the initial uptake velocity, un = k,, [1251-CSF-1], [receptor],. Since the concentrations of both ligand and receptor are known (Fig. 6), knn 2 1.24 X lo3 molecules cell-' ml-' min-'/(2.95 X lo-'' M X 2.32 X lo4 molecules cell-' ml-') = 3 X lo6 M-' s-'. From these estimates of KO,, and kff (=1.2 x lo-' s-I, Fig. 4) an estimated value for Kd = knff/kon 5 4 X IO-" M can be calculated. This value for Kd at 37 "C is within the concentration range for CSF-1 stimulation of BMM proliferation (half- maximal concentration -3 X lo-'' M (2)). Thus the kinetic data are consistent with a physiologically relevant, homogeneous equilibrium between CSF-1 and its cell surface receptor at 37 "C.

Although CSF-1 is reversibly bound to the cell surface at 37 "C (Figs. 2 and 4), its rate of dissociation is 6 times less than its rate of internalization (Figs. 2-4). Once internalized, CSF-1 is no longer available for surface dissociation (Fig. 2) and is degraded intralysosomally (tl,z = 3.5 h, Table 111). Thus CSF-1 binds to a varying number (Fig. 6) of cell surface receptors and spends an average of l/k. = 2 min at the cell surface. Six of every seven surface-bound CSF-1 molecules become irreversibly committed to degradation by rapid ( tll2 = 90 s) internalization. Ten to fifteen minutes are required for degradation to begin. After an initial burst of degradation (approximately 15 min), the remaining internalized CSF-1 (80-90%) is slowly degraded as a single pool (average time for degradation, l/kh = 14 h). The above scheme, largely con- structed from temperature shift data, is corroborated by observations that during mitogenic stimulation, CSF-1 contin-


ues to accumulate inside the cell with very little ever being present at the cell surface (Table 11, Fig. 5).

During mitogenic stimulation, the number of molecules of CSF-1 accumulated and degraded by BMM is many times the number of receptors at the surface of quiescent cells (Fig. 6 and "Results"). Furthermore, although a large cryptic pool exists (approximately twice the size of the cell surface receptor complement of quiescent cells, Table 111), receptor-mediated CSF-1 uptake and degradation exceed the estimated total number of receptors (cell surface + cryptic) by 1.6-fold at 8 h, 2.6-fold at 12 h, and 3.3-fold at 16 h (Fig. 6). Thus, after 8 h of uptake, receptors originally present on the quiescent cells have either been degraded and replaced by de novo synthesis or recycled.

Between 8 and 16 h of uptake, unoccupied cell surface receptor levels remain relatively constant. Since cell surface levels should reflect total unoccupied receptor levels at steady state (21), the total cellular pool of unoccupied receptor would appear to be constant between 8 and 16 h. The estimated receptor internalization rate between 8 and 16 h (assuming that receptor internalization is equimolar with respect to the disappearance of 1251-CSF-1 from the medium, Fig. 6 (5)) averages 4 molecules/cell/s. As the number of unoccupied receptors remains constant over this period, the replacement rate equals the internalization rate. The data do not allow us to determine the relative contributions of de nouo synthesis and receptor recycling to receptor replacement. However, if one assumes that replacement is entirely mediated by receptor recycling, then the recycling time is very slow (>4 h, see "Appendix") compared with the recycling times observed for receptors which are primarily replaced by recycling, e.g. the low density lipoprotein, mannosyl-glucosyl or asialoglycopro- tein receptors (22-24). If one assumes that replacement is entirely mediated by de nouo receptor synthesis, then the maximum rate of de novo receptor synthesis is 4 molecules cell-' s-l. De nouo receptor synthesis is a normally occurring event during log-phase BMM growth, resulting in a doubling of the receptor pool every 24 h. Assuming the rate of CSF-1 receptor synthesis is constant throughout the cell cycle, it can be calculated to be 0.6-1.7 molecules cell-' s-'. As the rate of total protein synthesis is maximally stimulated within 2 h of addition of CSF-1 to quiescent BMM (25), it appears that de nouo receptor synthesis significantly contributes to receptor replacement between 8 and 16 h of uptake. The results of a related study, in which the cell surface receptor replacement on PEM after down-regulation by CSF-1 ( t l / g = 3 h) or after modulation downwards by TPA (t1l2 = 1.2 h) was blocked by the protein synthesis inhibitor cycloheximide (18, 26), are consistent with this conclusion. However, it is not entirely clear from this work whether cycloheximide inhibits receptor synthesis or the synthesis of a protein necessary for receptor recycling.

Examination of the data presented in Fig. 5 reveals the timing of the replacement of internalized receptors. Prior to 10 min, all the lZ5I taken up by BMM remains inside the cells (Fig. 2). During the first 10 min of uptake (Fig. 5 ) , the average ratio of lZ5I-CSF-1 taken up to cell surface receptors lost of approximately 1 (0.91), indicates that surface receptors are not being replaced. After 10 min of uptake, the cells begin to rapidly release '''1 (Fig. 2). Thus between 15 and 60 min (Fig. 5 ) , the actual rate of 1251-CSF-1 internalization (or receptor loss) is greater than the observed average rate of lZ5I accumulation which equals 3.6 molecules cell-' s-l. However, after 15 min, the rate at which unoccupied receptors were lost markedly decreased (Fig. 5) to an average of 1.2 molecules cell-' s-l. Thus between 15 and 60 min, receptors are being

replaced at a rate of >2.4 molecules cell-' s-'. Receptor replacement between 15 and 60 min by recycling is unlikely in view of the hypothetical recycling time of >4 h, estimated above. It must occur either from the large cryptic receptor pool (maximal rate of appearance, 226 molecules cell-' s-', calculated from Table 111) or by de nouo synthesis (maximal rate -4 molecules cell-l s-').

Both macrophages and fibroblasts are strongly adherent, highly pinocytic cells. As a similar set of parameters to those generated for CSF-1 and BMM and CSF-1 and PEM have been obtained for the interaction of EGF with primary fibroblasts (21,27-29), comparison of the EGF data with the CSF- 1 data is of interest (Table V). There are slightly more EGF receptors on the fibroblast surface than CSF-1 receptors on the macrophage surface. At 37 "C, the values for k,,, and lz, (= kin + kod in both systems are the same. After correcting for differential cell-mediated degradation during the stimulation period, both growth factors are active in approximately the same concentration range (10-11-10-9 M (2, 21, 29)). Despite these similarities, there are striking differences. Cell surface CSF-1 is internalized more rapidly and dissociates more slowly than EGF bound to fibroblasts (compare kn/lzoff, Tables V and VI), the CSF-1 macrophage Kd is 15-fold lower than the EGF-fibroblast Kd (entirely due to a lower koe), and the degradation rate constant (k,,) is much lower for CSF-1 degradation by BMM (but not PEM) than for EGF degradation by fibroblasts. Important consequences of these differences are: (1) CSF-1 uptake by macrophages is saturated at a lower concentration than EGF uptake by fibroblasts (2,21) and (2) even though the relative ratios of first order constants for association to cell-surface loss (ken X [growth factor]/k,) at the same factor concentrations closely agree, and are much less than unity at physiological concentrations (lo-'' M), BMM maintain relatively far less bound factor at the cell surface and accumulate much more inside the cell (21, 27).

The present study of the interaction of CSF-1 with BMM, together with the accompanying study of its interaction with PEM (31), are helpful in deciding which states of the cell- associated ligand are necessary for the entry of the cells into S-phase. For example, it is possible that a high affinity cell surface component of cell-associated CSF-1 represents a re- liable and constant source of mitogenic signal. However, the constant amount of cell surface-bound CSF-1 (1600 molecules/cell) during the lag phase of incubation with 300 p~ CSF-1 (Fig. 5) drops to 51% of the BMM receptor pool (<500 molecules/cell, data not shown) in the presence of concentrations of CSF-1 which result in the maximum rate of entry into S-phase. This inverse relationship between amount of cell surface-bound CSF-1 and rate of entry into S-phase suggests that the cell surface-bound CSF-1 is not sufficient for mitogenic stimulation.

Another potential source of mitogenic stimulation is a large internalized CSF-1 pool. Twenty-four hours after addition of CSF-1 to quiescent BMM, -80% of the cells have relatively synchronously entered S-phase (lag phase - 12 h, 13). In contrast, PEM not only commence entry into S-phase after a 24-h lag phase but also enter more slowly (20% by 48 h, (8)). Cells of both populations have the same number of high affinity CSF-1 receptors/cell and similar parameters for the binding and internalization reactions (Table VI). However, they differ in their metabolism of CSF-1. Measurements made under identical conditions indicate that, compared with PEM, BMM accumulate intracellular CSF-1 and degrade it >13 times more slowly. Thus, accumulation of CSF-1 into a slowly degraded interior compartment is correlated with mitogenic response. This same correlation has been reported for EGF-


TABLE V Parameters of the EGF-primary target cell interaction

Parameter (units) 4 "C 37 "C

k,,(M-' S-l) 1.5 X lo5 (27)o 2.9 - 3.1 X lo6 (21) kffK1) 1.9 X 10-4* 1.2 - 2.5 X lo-' (21) kin(8-l) 2.5 - 5.3 X 10-~ (21)

ks = kin + &S(S-') 21.9 x lo4 1.5 - 3 X lo-' (21) kill/kff(s-l) 0.2 - 0.21 (21)

K d = kff/kn(M) 1-2 X 10-~ (27) 4.3 - 8 X lo4 (21)

kds-l) 2 - 8.3 X lo4 (21, 27, 28) ki& 6.4 - 9.7 (21) Surface binding sites (molecules/cell) 7 - IO X lo4 ( 2 7 , ~ ) Cryptic binding sites (molecules/cell) 4 . 5 X lo5 (27) Active concentration range (M) IO-" - 10-9 ( 2 1 , ~ )

a Source references are in parentheses. Absence of brackets indicates that values have been calculated from primary constants listed.

Calculated from the tlIz for dissociation given in Ref. 27.

TABLE VI Parameters of the CSF-1-macrophage interaction

Value

37 "C

PEM BMM

1.2 x 10-3 1.2 X 10-3 1.2 x lo-* 7.3 X 10-3

1.2 x 5 4 x 10-l0

2 3 x lo6

8.5 x 10 6.1

7.3 X 1 0 - ~ 55.8 x 10-~ 16

Surface binding sites (molecules/cell) Cryptic binding sites (molecules/cell) >5 X 103 105 Active concentration range (M)b 10-11 - 10-9 10-11 - 10-9

2126 7 X 104 4.9 X lo4

"Values for BMM are from this paper. Values for PEM are from other sources (31). Because of the high rate of degradation of CSF-1 by macrophages, lower limits of these ranges can only be

determined at very low cell concentration. Lower limits are estimated from data contained in Refs. 2 and 30.

stimulated initiation of DNA synthesis in quiescent fibroblasts (11, 12). Cells unable to respond bind and internalize, but fail to accumulate EGF, due to degradation. In contrast, responding cells appear to route a significant fraction of internalized EGF to a unique organelle where it very slowly decays. Although these correlations could reflect a role of internalized growth factor or receptor in the mitogenic process, it appears that significant accumulation of CSF-1 into a slowly decaying interior pool is not sufficient for response. By 8 h after CSF-1 addition, quiescent BMM have accumulated 60,000 CSF-1 molecules/cell (Fig. 6). This accumulation decays with a half-life of 10 h (Table 111). Yet, the cells will not initiate DNA synthesis at 12 h if extracellular CSF-1 is removed at 8 h (13).

Recent studies raise the possibility that the higher CSF-1 degradation rate exhibited by PEM compared with BMM is simply a specific function that is characteristic of their more differentiated state. Circulating levels of CSF-1 have been shown to be specifically regulated by mononuclear phagocytes in U ~ U O . ~ The mechanism of CSF-1 clearance is by the receptor mediated internalization and intralysosomal destruction doc- umented in this paper and elsewhere (31). The bulk of this clearance is handled by the Kupffer cells of the liver, which as a population resemble PEM, both in terms of the frequency of cells capable of proliferation in response to CSF-1 and their kinetics of entry into S-phase (30). Thus a slow rate of intracellular CSF-1 degradation need not be causally related

A. Bartocci, D. Mastrogiannis, and E. R. Stanley, manuscript in preparation.

to a strong proliferative response but rather due to poor expression of a differentiated function by immature cells.

APPENDIX

If all turnover between 8 and 16 h of uptake in Figure 6 is due to recycling of pre-existing receptor, the average length of time for an internalized receptor to reappear (t,) would equal the number of internalized receptors (Ri) divided by the replacement rate (4 molecules cell" s-l).

t, = RJ4

The total number of internalized receptors is unknown but minimal and maximal estimates can be made. If the only receptors internalized were those available at t = 0, the number inside the cell between 8 and 16 h in Figure 6 would be 3 times the number initially at the cell surface minus the number of unoccupied cell surface receptors measured between 8 and 16 h = 6 x lo4 molecules/cell. Thus a minimum estimate for t, = 6 X 104/4 = 1.5 X lo4 s = 4.2 h. A maximum estimate would assume that all the cell-associated CSF-1 between 8 and 16 h is receptor-associated. Thus t, = 8.6 X 104/4 = 2.2 X lo4 s = 6 h.

Acknowledgments-We thank Michelle Hankins and Susanne Saltzman for excellent technical assistance and Lenore Grollman for help in the preparation of the manuscript.

REFERENCES 1. Stanley, E. R., Guilbert, L. J., Tushinski, R. J., and Bartelmez,

S. H. (1983) J. Cell. Biochem. 21, 151-159

4032 Mononuclear Phagocyte Growth Factor Uptake and Destruction 2. Tushinski, R. J., Oliver, I. T., Guilbert, L. J., Tynan, P. W., 17. Frost, A. A., and Pearson, R. G. (1962) Kinetics and Mechanism,

3. Sherr, C. J., Rettenmier, C. W., Sacca, R., Roussel, M. F., Look, 18. Chen, B. D-M., Hsu, S., and Lin, H-S. (1982) Biochim. Biophys.

4. Guilbert, L. J., and Stanley, E. R. (1980) J. Cell Biol. 85, 153- 19. Alexander, N. M. (1964) Endocrinology 74, 273-278

5. Guilbert, L. J., and Stanley, E. R. (1984) J. Immunol. Methods Biochemistry 23,2495-2534

6. Stanley, E. R., and Guilbert, L. J. (1981) J. Immunol. Methods Bwl. Chem. 259,5623-5631

7. Mauel, J., and Defendi, V. (1971) Exp. Cell Res. 65,377-385 L. (1982) Cold Spring Harbor Symp. Quant. Biol. 46,713-721 8. Chen, B. D-M., Kuhn, C . 111, and Lin, H-S. (1984) J. Cell Sci. 23. Stahl, P., Schlesinger, P. H., Sigardson, E., Rodman, J. S., and

9. King, A. C., andcuatrecasas, P. (1982) J. Bwl. Chem. 257,3053- 24. Ciechanover, A., Schwartz, A. L., and Lodish, H. F. (1983) Cell

10. Shooter, E. M., Yankner, B. A., Landreth, G. E., and Sutter, A. 25. Tushinski, R. J., and Stanley, E. R. (1983) J. Cell. Physiol. 116,

11. Miskimins, W. K., and Shimizu, N. (1982) J. Cell. Physwl. 112, 26. Chen, B. D-M., Lin, H-S., and Hsu, S. (1983) J. Cell. Physiol.

12. Miskimins, W. K., and Shimizu, N. (1984) J. Cell. Physwl. 118, 27. Dunn, W. A., and Hubbard, A. L. (1984) J. Cell Biol. 98, 2148-

13. Tushinski, R. J., and Stanley, E. R. (1985) J. Cell. Physiol. 122, 28. Stoscheck, C. M., and Carpenter, G. (1984) J. Cell Bwl. 98,1048-

14. -Stanley, E. R., and Heard, P. M. (1977) J. Bwl. Chem. 252, 29. Carpenter, G., and Cohen, S. (1976) J. cell. Physwl. 88,227-237

15. Guilbert, L. J., Nelson, D. J., Hamilton, J. A., and Williams, N. 168-170

16. Hinegardner, R. T. (1971) Anal. Biochem. 39, 197-201 in press

Warner, J. R., and Stanley, E. R. (1982) Cell 28, 71-81 pp. 160-162, John Wiley & Sons, New York

A. T., and Stanley, E. R. (1985) Cell 41,665-676 Acta 72,366-373

159 20. DeDuve, C., DeBarsy, T., Trouet, T., and van Hoff, E. (1974)

73,17-28 21. Knauer, D. J., Wiley, H. S., and Cunningham, D. D. (1984) J.

42,253-284 22. Brown, M. S., Anderson, R. G. W., Basu, S. K., and Goldstein, J.

70,147-166 Lee, Y. C. (1980) Cell 19,207-215

3060 32,267-275

(1981) Prog. Horm. Res. 37,417-446 67-75

327-338 116,207-212

305-316 2159

221-228 1053

4305-4312 30. Stanley, E. R., Chen, D-M., and Lin, H-S. (1978) Nature 274,

(1983) J. Cell. Physiol. 115, 276-282 31. Guilbert, L. J., Tynan, P. W., and Stanley, E. R. J. Cell. Biochem.,

Documents

The Interaction of 1251-C~lony-stimulating Factor- 1 with Bone