Modulation of Monocyte Type I Transforming Growth Factor+ Receptors by Inflammatory Stimuli*

The regulatory mechanisms which control the wide array of cellular responses to transforming growth factor B (TGFB) are not understood. This report pre-sents evidence that down-regulation of TGFj3 receptors on human monocytes may be one mechanism by which the effects of TGFB are regulated. Treatment of mono- cytes with interferon y (1FN-y) and lipopolysaccharide for 18 h reduced monocyte receptor number (approxi- mately 400/cell) in a dose-dependent fashion by 89 and 78%. respectively, as determined by '261-TGFB bind- ing. Incubation with other cytokines (granulocyte-macrophage colony-stimulating factor, macrophage colony-stimulating factor- 1, interleukin-1, tumor necrosis factor a) did not alter the amount of TGFB bound. The decrease in '261-TGF@ binding could not be attributed to competition for receptor sites by secreted TGFB. Instead, the decline in binding was due to a loss of type I TGFB receptors, the subtype primarily ex- pressed by monocytes, with no decrease in receptor affinity. Lipopolysaccharide-induced receptor loss was rapid (1-4 h), in contrast to the prolonged (12 h) decline induced by IFN-y. Loss of receptors was accom- panied by a diminished ability of the cells to respond to TGFj3 with an induction of TNFa mRNA. Thus, this monocyte system is the first example of a heterologous agent causing the down-regulation Monocyte Isolation-Human monocytes were isolated from hepa- rinized blood of healthy normal volunteers undergoing leukapheresis at the NIH Blood Bank. The leukocyte-rich preparation was sepa- rated by counterflow centrifugal elutriation as previously described (19). Monocytes obtained by this procedure have been shown to be >90% pure as judged by dual fluorescence microfluorometry using specific cell surface markers. The monocytes were maintained in suspension in Dulbecco's modified Eagle's medium (Mediatech, Hern-don, VA) containing 20 mM glutamine, 50 pg/ml of gentamicin (Whittaker, Walkersville, MD), and 0.1% fetal bovine serum (GIBCO, Grand Island, NY) at a concentration of 3 X IO6 cells/ml. The cells were incubated with or without the addition of various cytokines at 37 "C in a humidified atmosphere of 95% air, 5% COz. TGFB Assay-TGFp was assayed by inhibition of IL-1 dependent thymocyte proliferation as previously described (20). A portion of each sample was heated to 80 "C for 5 min to activate any latent TGFB (18). The concentration of TGFB present in heat-treated samples was equivalent to the total TGF@ concentration, active and latent; the concentration present in the non-heated samples was the amount of TGFB activated endogenously. TGFB concentrations were determined from a standard curve prepared with TGFp obtained from Collaborative Research (Bedford, MA). For some experiments, the amount of TGFB in the cell supernatant was also quantified by a modification of the A549 competitive binding assay previously described (21). 1261-TGF@ (100 PM) was incubated with A549 cultures at 4 "C for 3 h in the presence of either unlabeled TGFp (10 pM to 2 nM/well for standard curve generation, or 10 nM for nonspecific binding determination) or cell culture supernatant.

The multiple proinflammatory effects of TGFP on monocytes suggests the existence of regulatory mechanisms to antagonize or diminish the effects of TGFP, thus leading to the resolution of the inflammatory state. The release of TGFP in a latent form is an important regulatory mechanism (1); however, inflammatory macrophages have been shown to activate TGFP, probably through the release of a specific sialidase (13) or protease (14), resulting in active TGFP at the site of inflammation (15). Another possible regulatory mechanism is suggested by observations that sustained exposure to TGFP leads to suppression of some monocyte proinflammatory activities. For example, the generation of reactive oxygen intermediates is diminished (16), and the production of cytokines in response to a secondary stimulus is inhibited (17).
Modulation of monocyte expression of the TGFp receptor is an additional mechanism by which the proinflammatory effects of TGFp could be regulated. Preliminary evidence from this laboratory has suggested that activated monocytes exhibit a reduction in the specific binding of '"1-TGFP to monocytes (9). In this study, that observation is expanded to show that interferon y (IFN-,) and lipopolysaccharide (LPS), two substances potentially present within the inflammatory site, reduce the number of cell surface receptors to <25% of the number of receptors expressed on control cells. In addition, receptor loss is shown to correspond to a loss in capacity to be stimulated by TGFP. Thus, the proinflammatory recruitment and stimulatory influence of TGFP on blood monocytes may be moderated once the cells have become activated, thereby favoring resolution of the inflammatory response.
Monocyte Isolation-Human monocytes were isolated from heparinized blood of healthy normal volunteers undergoing leukapheresis at the NIH Blood Bank. The leukocyte-rich preparation was separated by counterflow centrifugal elutriation as previously described (19). Monocytes obtained by this procedure have been shown to be >90% pure as judged by dual fluorescence microfluorometry using specific cell surface markers. The monocytes were maintained in suspension in Dulbecco's modified Eagle's medium (Mediatech, Herndon, VA) containing 20 mM glutamine, 50 pg/ml of gentamicin (Whittaker, Walkersville, MD), and 0.1% fetal bovine serum (GIBCO, Grand Island, NY) at a concentration of 3 X IO6 cells/ml. The cells were incubated with or without the addition of various cytokines at 37 "C in a humidified atmosphere of 95% air, 5% COz.
TGFB Assay-TGFp was assayed by inhibition of IL-1 dependent thymocyte proliferation as previously described (20). A portion of each sample was heated to 80 "C for 5 min to activate any latent TGFB (18). The concentration of TGFB present in heat-treated samples was equivalent to the total TGF@ concentration, active and latent; the concentration present in the non-heated samples was the amount of TGFB activated endogenously. TGFB concentrations were determined from a standard curve prepared with TGFp obtained from Collaborative Research (Bedford, MA).
For some experiments, the amount of TGFB in the cell supernatant was also quantified by a modification of the A549 competitive binding assay previously described (21). 1261-TGF@ (100 PM) was incubated with A549 cultures at 4 "C for 3 h in the presence of either unlabeled TGFp (10 pM to 2 nM/well for standard curve generation, or 10 nM for nonspecific binding determination) or cell culture supernatant. Active versus latent TGFB was distinguished as above. Following the incubation, the cultures were solubilized with the Triton X-100 buffer (211, and counted in a 1277 Gamma-master counter (Pharmacia, LKB Biotechnology Inc., Piscataway, NJ). TGFP ReceptorAssay"TGFj3 receptors were quantified by binding of '261-TGF@ to monocytes using a modification of the method of Wakefield (22). TGFp was iodinated to a specific activity of 1.7-2.2 pCi/pmol usinga modified chloramine-T method (23). TGFp receptor number was quantified for freshly isolated monocytes and monocytes incubated with various agents at 4 or 37 "C for 1-18 h, followed by 2 washes with ice-cold binding buffer (above) to remove the stimuli. For the assay, cells were suspended in binding buffer (Dulbecco's modified Eagle's medium with 0.1% bovine serum albumin (Sigma) and 25 mM HEPES, pH 7.4) at a concentration of 20 X 106/ml, and aliquots of the cell suspension (100 pl) were incubated with 100 p~ lz6I-TGFp in a final volume of 200 pl for 3 h at 4 "C with constant agitation. Previous experiments had shown that this concentration of TGFB was sufficient to saturate monocyte receptors (9), and that the incubation conditions did allow binding to reach equilibrium. Nonspecific binding was determined in the presence of a 400-fold excess of unlabeled TGFB. Following the binding incubation, the cells were pelleted and an aliquot of the supernatant was removed to determine the amount of unbound lZ5I-TGFj3. Following removal of the residual supernatant, the cell pellet was resuspended in 200 pl of ice-cold binding buffer and spun through a silicon oil/paraffin oil (84:16) cushion. The resultant pellet was counted in a 1277 Gammamaster counter to determine the amount of bound ligand.

TGFB Receptor Affinity Labeling and Molecular Mass
Determination-Monocytes were incubated with 100 pM '251-TGFp in binding buffer as above. The labeled TGFj3 was cross-linked to its receptor using a modification of the procedure described in Ref. 24. Cells were incubated for 15 min at 4 "C with a freshly prepared solution of 0.25 mM disuccinimidyl suberate (Pierce Chemical Co.) in phosphatebuffered saline (pH 7.4). The reaction was stopped by the addition of Tris-HC1 (pH 6.8) and EDTA to final concentrations of 0.3 and 0.6 M, respectively. The cells were washed three times with phosphatebuffered saline and then solubilized with a solution of 0.1% SDS and 1 mM phenylmethylsulfonyl fluoride and stored at -20 "C. Monocyte proteins were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on an 8% polyacrylamide gel using a ["Clmethylated protein mixture (Amersham) as standards. The gel was dried and exposed to X-Omat AR film (Kodak, Rochester, NY) and the resultant autoradiograph was scanned with an Ultroscan XL Laser Densitometer (Pharmacia LKB, Piscataway, NJ).
RNA Blot Analysis-Total RNA was isolated from 1 X lo7 elutriated monocytes cultured for 18 h in supplemented Dulbecco's modified Eagle's medium with or without GM-CSF, IFNy, or LPS, then for an additional 3 h in the presence or absence of TGFj3. Total RNA was extracted by the acid guanidinium thiocyanate/phenol/ chloroform extraction procedure (25). A portion of the total RNA from each sample (5 pg) was fractionated by electrophoresis on a 1% agarose gel containing formaldehyde and transferred to a nitrocellulose filter. Blots were prehybridized for 4 h at 42 "C, and then hybridized overnight at 42 "C with a 32P-labeled cDNA probe for TNFa (26). Filters hybridized with the TNFa probe were washed with 2 X SSC, 0.1% SDS twice at room temperature, then once at 42 "C with 0.1 X SSC, 0.1% SDS.

RESULTS
Radiolabeled TGFP Binding to Monocytes-Human monocytes were assayed to quantify TGFB receptors immediately after isolation or after overnight culture in the presence or absence of specific cytokines or LPS. As detailed in Table I, freshly isolated human monocytes bind 0.59 * 0.06 fmol of TGFP/106 cells (average * S.E., n = 4); this corresponds to about 360 receptors/cell, assuming a 1:l binding stoichiometry. Because monocytes cultured overnight without serum or growth factors lose viability (27) and exhibit increased nonspecific '"I-TGF@ binding, GM-CSF or M-CSF, cytokines which promote cell survival, were added to the cultures. These greatly reduced the loss of monocyte viability and thus the increase in nonspecific binding. Cells cultured for 18 h with GM-CSF (300 units/ml) bound approximately the same amount of TGFp as freshly isolated monocytes, 0.57 -I-0.05 fmol/106 cells (average -+ S.E., n = 9) corresponding to 340 receptors/cell. Equivalent amounts of TGFB were bound at other GM-CSF concentrations (200, 500 units/ml); however, concentrations 5100 units/ml were insufficient to maintain monocyte viability and the nonspecific binding increased to approximately 70% of the total. The amount of TGFP bound by monocytes cultured with 100 units/ml of M-CSF (0.55 f 0.07 fmol/106 cells, 330 receptors) was similar to that bound by the GM-CSF-treated monocytes. Since the addition of colony-stimulating factors aided in maintaining cell viability without altering the number of TGFP receptors, monocytes incubated with GM-CSF (300 units/ml) served as the 18-h control.
Modulation of lz5I-TGFP Binding-To investigate if inflammatory mediators could influence expression of monocyte TGFB receptors, monocytes were incubated for 18 h with IFN--y (200 units/ml), a known monocyte activator, before assaying for TGFD receptors. This resulted in a substantial decrease in the amount of specifically bound '251-TGFP, from 0.57 & 0.05 to 0.14 -+ 0.04 fmol/106 cells (average k S.E., n = 9) or 24% of the amount bound in fresh cells or 18-h GM-CSF control cells (Table I). Other inflammatory cytokines known to modulate monocyte function, TNFa and IL-1, were   Table I, neither of these cytokines changed the amount of TGFP bound during an 18-h incubation relative to the control cells incubated with GM-CSF or the freshly isolated monocytes. At the indicated concentrations, these cytokines also maintained cell viability. Simultaneous addition of an equivalent amount of GM-CSF, also did not change TGFB binding. LPS, an inflammatory mediator and macrophage activator present at sites of infection, was next evaluated for its effects on TGFP binding. Overnight incubation of monocytes with 1 pg/ml of LPS dramatically reduced the amount of TGFP specifically bound to 0.065 k 0.010 fmol/106 cells (average k S.E., n = 9) or 11% of the amount bound by fresh cells and 18-h control cells. Both of these activators, LPS and IFNy, effectively maintained monocyte viability, and the simultaneous addition of GM-CSF (300 units/ml) did not prevent receptor loss, nor change its magnitude. Although the TGFB assay is preceded by a washing step to remove the activating agents, residual IFN-y or LPS may remain which may compete with or bind the lZ5I-TGFP, thus preventing its interaction with the receptor. To address this possibility, LPS (1 pg/ml) and IFNy (200 units/ ml) were directly added to freshly isolated monocytes suspended in binding buffer with 100 PM lZ5I-TGF/3 at the onset of the lZ5I-TGFP binding incubation. Even at these high concentrations, neither LPS nor IFNy reduced the binding of lz5I-TGFP to its receptor (data not shown).
Additional studies examined the concentration range over which IFNy and LPS were effective in decreasing the amount of labeled TGFp specifically bound. As shown in Fig. 1 drop in TGFP binding. Binding declined by 60% after only a 1-h incubation. By 4 h, the loss of binding was maximal with only 10% of the total binding remaining. The kinetics of IFNy-induced decline in TGFB binding were more protracted, with maximal reduction in lZ5I-TGFp binding occurring after a 12-h incubation with the IFNy.
Affinity Labeling of Monocyte TGFP Receptors with lZ5I-TGFP-Monocyte cell surface TGFP-binding proteins were affinity labeled by first equilibrating equal numbers of fresh or 18-h cultured cells with lZ5I-TGF/3 at 4 "C, then crosslinking the lZ5I-TGFp to its receptor with the bifunctional cross-linker, disuccinimidyl suberate. Following separation of the solubilized cell proteins by SDS-PAGE, the labeled ligand/receptor protein bands were visualized by autoradiography. Freshly isolated human monocytes were found to express primarily the type I TGFP receptor (for review, see Ref. 1). This appears as the 65-kDa protein band representing the receptor-ligand complex (Fig. 3, lane 1 ) . The specificity of the interaction between '251-TGFP and the receptor protein is shown by the ability of a 400-fold excess of unlabeled TGFB to completely eliminate labeled TGFP binding (lane 2). Control monocytes incubated with GM-CSF for 18 h at 4 "C (lane 3 ) and at 37 "C (lane 5 ) also specifically bound a similar amount of "'I-TGF@ in a 65-kDa protein band corresponding to the type I receptor-ligand complex. As before, for both the 4 and 37 "C incubations, the addition of a 400-fold excess of unlabeled TGFB eliminated the 65-kDa band, indicating the specificity of the '251-TGF@/protein interaction (lanes 4 and 6, respectively). Fresh and GM-CSF-treated cells from some experiments also occasionally yielded a faint band at approximately 100 kDa which was not readily discernable without overexposure of the autoradiograph, and could not be readily distinguished by densitometric scan. When identified, this band was only present in the absence of an excess of unlabeled TGFP. This protein band may represent the type I1 TGFp receptor-ligand complex (1). When monocytes were incubated with LPS for 18 h at 37 "C, specific binding of radiolabeled TGFP to the 65-kDa complex was virtually eliminated (lane 9). Nonspecific binding (background) was also greatly reduced ( l a n e 10). When the 18-h LPS incubation was carried out at 4 "C (lane 7) rather than 37 "C, monocytes were found to specifically bind approximately the same amount of '251-TGFP in the 65-kDa protein band as did the freshly isolated (lane I ) and GM-CSF-treated cells (lanes 3 and 5). This suggests that LPS does not directly interfere or compete with TGFp  C (lanes 3 and 4 ) or 37 "C (lunes 5 and 6 ) , with LPS at 4 (lanes 7 and 8 ) or 37 "C (lanes 9 and IO), or with IFNy at 4 (lanes 11 and 12) or 37 "C (lanes 13 and 14) were washed, then equilibrated with 100 PM '2sII-TGFfi in the presence (+) or absence (-) of a 400-fold excess of unlabeled TGFp for nonspecific binding determination. After chemically cross-linking the '""ITGFB to its receptors, cell proteins from equal numbers of monocytes were separated by SDS-PAGE, except in lanes 5 and 6 in which there is about 40% less cell protein. The dried gels were exposed to x-ray film for 7 days. The autoradiograph of the gels are shown with molecular mass markers indicated to the left and the receptor bands indicated to the right. This analysis was repeated 3 times with similar results.
binding to its receptor since such an interaction would take place at 4 "C, as well as 37 "C. Instead, LPS appears to reduce "'I-TGFP binding following energy-requiring cellular events. In a similar manner, incubation of monocytes for 18 h at 37 "C with IFNy also resulted in a decrease in the amount of specifically bound ligand (Fig. 3, lane 13). By scanning densitometry it was determined that the IFNy-treated cells exhibited a 75-85% decrease in labeled 65-kDa TGFP receptor relative to that expressed by freshly isolated or 18-h GM-CSF-treated cells. As with the LPS treatment, IFNy did not decrease l2'1-TGFP binding if the 18-h incubation proceeded at 4 "C (lane 11 ). As before, this suggests that IFNy does not directly inihbit or compete with TGFP binding to its receptor.

Analysis of the Inducible Decline in Specific lZ5I-TGFP Bind-
ing-To determine if the decline in specific binding of 12'1-TGFP was due to a loss of TGFB receptors from the cell surface or a decrease in the affinity of the receptors, Scatchard analysis was performed. Freshly isolated monocytes or monocytes treated for 18 h with GM-CSF, IFNy, or LPS were equilibrated in binding buffer containing concentrations of '2sI-TGFP ranging from 0.3 to 100 PM. The amount of TGFP bound per lo6 cells was plotted against the ratio of bound over free TGFB (Fig. 4). Fresh monocytes and cells treated for 18 h with GM-CSF both bound approximately 0.6 fmol of TGF@/106 cells (360 receptors/cell) and with similar affinity, KD = 7 pM. Following treatment with LPS, the amount of TGFP the monocytes bound dramatically decreased to 0.06 fmol/106 cells (40 receptors/cell). The residual TGFP binding was of somewhat greater affinity, KD = 2 PM. Incubation with IFNy also reduced the amount of TGFP bound to 0.17 fmol (100 receptors/cell); however, the binding affinity of the remaining receptors was comparable to that of fresh or cultured control cells, approximately 5 PM. Thus the reduction of specific binding observed following LPS or IFNy treatment . . 0 0.1 0.2 0.3 0.4 0.5 0.6 0 is an apparent loss of cell surface receptors rather than a decline in affinity.

0.
An apparent decrease in the number of cell surface receptors as determined by radiolabeled ligand-binding studies cannot distinguish between a true decrease in receptor number and an increase in receptor sites occupied by cold ligand. Monocyte activators have been shown to induce the production of TGFp (3, 5, 6), although the protein is often released in a latent form (1) which does not bind to the receptor (28). Supernatants from the same 18-h suspension cultures of cytokine-treated monocytes used for receptor measurements were assayed for active and latent TGFP. This was done to determine if the decrease in 'z'I-TGF/3 binding observed with LPS and IFNy-treated cells was due to the release of active TGFP. As shown in Fig. 5, LPS and IFNy do stimulate the release of TGFP; some of the TGFP is found in the latent form, measurable only after heat activation. The concentration of TGFP already in the active form in the culture medium (0.15 and 0.06 ng/ml or 6.0 and 2.4 PM, respectively) is fairly low relative to the concentration of the '"I"TGF/3 present during the receptor-binding assay (2.5 ng/ml or 100 p~) . This concentration of TGFP is much lower than is required to saturate receptor binding and therefore, the potential error introduced by the presence of the unlabeled TGFP is not more than 5%. Incubation with GM-CSF also resulted in the release of some active TGFP which could potentially compete with the '251-TGF@ for receptor binding. However, as before, the concentration of the active TGFB was low (0.11 ng/ml or 4.4 PM) and therefore would introduce only small error into the determination of TGFB receptor number. In fact, the receptor number determined for GM-CSF-treated cells was not different from that of freshly isolated monocytes. The concentration of TGFP in monocyte supernatants derived from cultures stimulated with M-CSF was also determined and found to be similar to that of GM-CSF.
To further demonstrate that monocytes incubated with LPS, IFNy, or GM-CSF do not significantly activate latent TGFO, exogenous latent TGFp (400 PM) was added to the monocyte cultures at the onset of the 18-h incubation. The concentration of active and latent TGFP in the resultant culture supernatant was then determined by two TGFP assays: thymocyte growth inhibition and displacement of lZ5I-TGFO binding to A549 cultures as described under "Materials and Methods.'' The concentration of active TGFP in the supernatant of the LPS, IFNy, and GM-CSF cultures was only 2-8 PM (data not shown). As before, this suggests that these monocyte cultures do not appreciably convert latent TGFB to active TGFP. Separate controls showed that the presence of LPS, IFNy, or GM-CSF in a cell-free system also does not activate latent TGFP. Heat treatment of the various supernatants confirmed that activation of the exogenous latent TGFP does result in a biologically active protein which displays normal TGFP receptor binding kinetics.
Since an assay of the concentration of active TGFP in the supernatant cannot measure TGFP already bound to the monocytes, a neutralizing antibody to TGFp was added to the monocyte cultures immediately prior to GM-CSF, LPS, and IFNy addition. After an 18-h incubation, receptor number was determined for the various cultures. The presence of the TGFP neutralizing antibody was found not to prevent the loss of monocyte TGFP receptors following LPS and IFNy treatments (data not shown). Separate control experiments confirmed that the antibody could neutralize the biological effects of active exogenous TGFP. Thus, these data suggest that the release of active TGFp by monocytes and its binding to the monocyte TGFP receptors is not the mechanism by which TGFO receptors are reduced.
Effect of TGFP Receptor Down-regulation on Monocyte Responsiveness to TGFP-Decreasing the number of TGFB receptors on monocytes may provide a mechanism to modulate the response of the cells to TGFP. Addition of TGFP to freshly isolated monocytes in vitro has been shown to cause the induction of mRNA for various cytokines including TNFa (5,6, 9, 10). Monocytes were incubated with GM-CSF, LPS, or IFNy for 18 h, and then challenged with exogenous TGFp (10 ng/ml) for an additional 3 h. After the second incubation, total cellular RNA was isolated and analyzed to determine if the cells could respond to the TGFB with an increase in the expression of TNFa mRNA. As shown by Northern blot analysis, TGFB induced TNFa mRNA expression by cells incubated with GM-CSF for 18 h (Fig. 6), as it does for freshly isolated monocytes (5, 6,9,10). In contrast, when LPS was added to the monocyte cultures, it resulted in the complete elimination of TGFP-induced TNFa mRNA expression detectable by Northern blot. Examination of the 28 S and 18 S rRNA bands shows that this change in TNFa mRNA expression is not due to unequal quantities of RNA in the lanes (Fig. 6, lower panel). Receptor 2,4, and  6) or with no additions (lanes 1,3, and 5). The total RNA from each sample was analyzed for TNFu mRNA expression by Northern blot analysis as outlined under "Materials and Methods." Ribosomal RNA (28 S and 18 S) is shown to demonstrate equivalent quantities of RNA among the lanes. indicated that the TGFP receptor level on the LPS-treated cells was only 5% of that of the GM-CSF control cells. In a similar manner, 18 h incubation with IFNy caused a 47% decrease in the TGFP-stimulated expression of TNFa mRNA, as quantified by densitometric scan (Fig. 6). Concurrent measurements of TGFP receptors on IFNy-treated cells revealed a 71% loss of TGFP receptors. This experiment was repeated three times with identical results for the LPS treatment, and somewhat more variability seen with the 1FN-y treatment: 30-50% decrease in mRNA expression. Thus, the data demonstrate reduced functional responses to TGFB in cells exhibiting decreased receptor expression.

DISCUSSION
The cellular response to cytokines and hormones is often regulated at the level of receptor expression; exposure to homologous and heterologous agents causes a rapid alteration in receptor affinity or number. Evidence for the existence of this regulatory mechanism has been noticeably limited with respect to TGFP. One cell type which does modulate TGFp receptor expression is the T lymphocyte, a cell which is growth inhibited by TGFB. T cells have been shown to increase the expression of type I, 11, and I11 TGFp receptors following stimulation with mitogens for 48 h (29). Similarly, B lymphocytes exhibit a 5-6-fold increase in TGFP receptor expression following in vitro activation (7,30). Adrenal cortical cells have also been reported to increase the expression of the type I11 TGFP receptor subtype following stimulation with adrenocorticotrophic hormone (31). Whether the increased expression of TGFP receptors by these cells results in a greater sensitivity to TGFB has not yet been established. Some cell types which are known to alter their responsiveness to TGFP in conjunction with developmental changes do not stimultaneously exhibit a significant change in TGFp receptor expression (32). The monocyte system described here represents a model in which, for the first time, TGFP receptors are shown to be down-regulated by exposure to heterologous agents. In addition, these studies demonstrate that receptor loss may affect cell function since it results in a diminished ability of the monocytes to produce cytokines in response to TGFP.
Both LPS and 1FN-y were found to decrease monocyte TGFp receptors and this receptor modulation could not be attributed to occupancy of the binding sites with TGFP released by the cells. The rate of receptor loss initiated by these two agents varied considerably, suggesting the process was occurring by different mechanisms. Incubation with IFNy resulted in a steady decline in receptor number over 12 h; this may be due to an inhibition of steady state TGFP receptor synthesis. In contrast, the rapid reduction in receptor number induced by LPS suggests an alternative mechanism. Although the decrease in nonspecific, as well as specific, binding following LPS treatment (Fig. 3, lanes 9 and 10) might indicate an increase in fluid phase pinocytosis or membrane turnover, numerous studies have shown that this interpretation is unlikely since LPS (and IFNy) treatment increase, as well as decrease, many plasma membrane proteins, yet have no effect on other proteins. For example, stimulation of monocytes with LPS and IFNy has been shown to down-regulate C5a receptors, but receptors for another chemotaxis factor, formyl-methionyl-leucyl-phenylalanine, are not altered (33). In addition, treatment with LPS, IFNy, or GM-CSF has been shown not to alter the monocyte marker protein CD14, nor do they change the expression of CD16 (FcyRIII) (11). Ding et al. (34) directly addressed the question of the selectivity of the LPS effect in conjunction with their observation that TNFa receptors were rapidly down-regulated by LPS. These researchers showed that the LPS effect was not a generalized internalization of the plasma membrane receptors because the surface expression of complement receptor type 3 did not change during the length of the incubation. In contrast to LPS, other researchers have found that IFNy increases the number of TNFa receptors (35,36). LPS and IFNy have also been shown to increase the expression of the IL-2 receptor and major histocompatability complex class I1 antigen (HLA-DR) (37). Thus, these studies indicate that the down-regulation of TGFP receptors by LPS and IFNy is a selective effect on a specific population of membrane proteins and not the result of generalized membrane internalization.
The decline in number of cell surface receptors was coupled with a diminished ability of the cells to respond to TGFP as evaluated by induction of cytokine mRNA. Pre-treatment with LPS yielded cells which did not express TNFa mRNA in response to TGFB stimulation; this corresponded to an almost complete loss of TGFP receptors on these cells. IFNytreated cells also down-regulated TGFP receptors, although to a lesser extent. These cells displayed a partial reduction of TNFa mRNA expression following TGFP stimulation. This would suggest that certain functional responses of the monocyte to TGFP are limited by receptor number. Possibly, other cell types which express a greater number of TGFP receptors (21) would not be restricted in this manner.
Covalently linking radiolabeled TGFP to its high affinity cell surface-binding proteins followed by electrophoretic analysis, revealed that monocytes primarily express the 65-kDa type I receptor. Thus, the analysis and characterization of '251-TGFP binding can be attributed to this single receptor subtype. Freshly isolated human monocytes and monocytes maintained in GM-CSF-supplemented suspension culture were found to express approximately 360 receptors with an affinity of 7 p~. These numbers fall within the lower end of the range reported for type I TGFP receptors on other cells: 300-4000 binding sites/cell with an affinity in the range of 5-50 PM (32). Incubation with LPS or IFNy reduced receptor number to approximately 40 and 80 receptors/cell, respectively. Treatment with LPS also resulted in a measurable increase in TGFB receptor affinity (2 pM); however, the limited magnitude of this change suggests that it may not be functionally significant. Cross-linking '"I-TGF@ to monocyte receptors also revealed that the cells may express a small population of 100-kDa type I1 receptors; however, the sporadic appearance of the band suggests it may instead be due to changes in the percent of contaminating T lymphocytes. T cells do express the type I1 TGFP receptor (29) and are the main contaminant of monocyte preparations (1-8%). Others have identified both type I and type I1 TGFP receptors on human monocytes (38,39); however, no details were given as to the relative proportion of the receptor subtypes. Analysis of receptor subtypes expressed by the U937 human monocytic leukemia cell line showed no evidence of type I1 receptors (32), however, these cells express the 250-350-kDa type I11 receptor along with the type I receptor, possibly indicating that expression of the type I1 receptor also varies from that of normal cells.
The absence of type I11 receptor expression on monocytes may have facilitated our ability to measure changes in receptor number. The number of type I11 receptors on other cell populations is typically much greater than type I or type I1 receptors, up to lo5 sites/cell (32) and is generally not considered to fluctuate. A current hypothesis holds that type I, and possibly type I1 receptors are involved in signal transduction (32). If this is the case, then the type I11 receptor may not be subjected to up-and down-regulation in response to environmental stimuli. Thus any changes in number of type I or I1 receptors may not be distinguishable against a high background of a constant number of type I11 receptors. It is also possible that cells which express the type I11 receptors do not undergo down-regulation of TGFP receptors. The type I11 receptor is believed to function as a reservoir or clearance system for bioactive TGFP (32); thus, binding and storage of TGFP by type I11 receptors may take the place of receptor down-regulation.
Recent advances in our understanding of the central role of TGFP in mediating the inflammatory response supplies insight into the physiological significance of monocyte TGFP receptor modulation by inflammatory mediators. Femtomolar concentrations of TGFP initiate a motility response by monocytes. Local release of TGFP by platelets and other inflammatory cells at a site of inflammation is important in recruiting monocytes to the area. Once the cells have arrived at the site, they are thought to be exposed to higher concentrations of TGFB and other regulatory mediators (2). In this regard, TGFP is required at picomolar concentrations to stimulate additional monocyte functions which promote and augment the inflammatory process. Furthermore, TGFB up-regulates its own synthesis and secretion by monocytes (5, 6), contributing to the build-up of TGFB within an inflammatory site. These observations suggest that TGFP is an extremely potent proinflammatory cytokine, a conclusion consistent with recent studies investigating the effects of in vivo administration of TGFP within the synovial space (40). Thus, the continued secretion of TGFP and the resultant monocyte activation could result in prolonged inflammatory sequelae. One mechanism which could possibly control these effects may be the loss of sensitivity of monocytes to TGFP stimulation by the down-regulation of their receptors.
As mentioned above, modulation of chemotactic factor receptors is not limited to the TGFP receptor. Stimulation of monocytes with IFNy or LPS selectively down-regulates the cell surface receptor for another monocyte chemotactic peptide, C5a, but not for formyl-methionyl-leucyl-phenylalanine (33). The purpose of down-regulation of receptors for monocyte chemotactic factors after activation is of interest; under physiologic conditions, C5a or TGFP-directed chemotaxis would no longer be required once the monocytes have been recruited and activated at an inflammatory site. In support of this hypothesis, macrophages isolated from inflammatory lesions (33) or IFNy-treated animals (41, 42) exhibit signifi-