Identification, characterization, and homologous up-regulation of latent (cryptic) receptors for tumor necrosis factor-alpha in rat liver plasma membranes.

A population of latent (cryptic) receptors for tumor necrosis factor-alpha (TNF) has been characterized in the rat liver plasma membrane (PM). 125I-TNF bound to high (Kd = 1.51 +/- 0.35 nM) and low (Kd = 13.58 +/- 1.45 nM) affinity receptors in PM. Solubilization of PM with 1% Triton X-100 prior to incubation with 125I-TNF increased both high affinity (from 0.33 +/- 0.04 to 1.67 +/- 0.05 pmol/mg of protein) and low affinity (from 1.92 +/- 0.16 to 7.57 +/- 0.50 pmol/mg of protein) TNF binding without affecting the affinities for TNF. Digestion of intact PM with chymotrypsin abolished most of the TNF binding capacity of PM. However, substantial binding activity was recovered by solubilization of chymotrypsin-treated PM with 1% Triton X-100, suggesting the presence of a large latent pool of TNF receptors. The affinities of the high and low affinity sites recovered from chymotrypsin-treated membranes were similar to those of intact PM. Affinity labeling of receptors whether from PM, solubilized PM, or membranes digested with chymotrypsin and then solubilized resulted in cross-linking of 125I-TNF into Mr 130,000, 90,000, and 66,000 complexes. Thus, the properties of the latent TNF receptors were similar to those initially accessible to TNF. To determine if exposure of latent receptors is regulated by TNF, 125I-TNF binding to control and TNF-pretreated membranes was assayed. Specific binding was increased by pretreatment with TNF (p less than 0.05), demonstrating that hepatic PM contains latent TNF receptors whose exposure is promoted by TNF. Homologous up-regulation of TNF receptors may, in part, be responsible for sustained hepatic responsiveness during chronic exposure to TNF.

A population of latent (cryptic) receptors for tumor necrosis factor-a (TNF) has been characterized in the rat liver plasma membrane (PM). lz61-TNF bound to high (ZL = 1.51 f 0.35 nM) and low (& = 13.58  from adipose and other host tissues in response to infection (2, 3). Thus, TNF has been proposed to mediate the wasting (cachexia) that often accompanies chronic disease states. The demonstration that the immune system produces an oncolytic agent with therapeutic potential, which may also induce significant alterations of target tissue metabolism, has produced great interest in TNF.
Chronic exposure of target cells to hormones generally results in desensitization of responsiveness to further stimulation. Down-regulation of the number of hormone receptors expressed at the cell surface is a common mechanism through which cellular responsiveness may be attenuated (4,5). Therefore, chronic exposure to TNF would be expected to promote receptor down-regulation and desensitization of target organ responsiveness.
However, TNF has been proposed to induce metabolic changes in various host tissues during chronic diseases (3). In the liver, TNF induces the synthesis of acute phase proteins (6-9) and lipids (lo-12), and the liver remains responsive to TNF upon repeated administration over several days (13). To reconcile these observations, we have sought mechanisms that might sustain the responsiveness of the liver to TNF.
Maintenance of the level of the cell-surface receptor is one likely process through which tissue responsiveness to TNF could be sustained.
Target cells for a number of peptide hormones possess latent binding sites whose expression is regulated by homologous and heterologous factors (14). In this study, we have identified and characterized a latent pool of TNF receptors that reside within the rat hepatic plasma membrane (PM from the incubation mixtures were added to 150 ,.d of 0.1 M PO:-(pH 7.5). ""I-TNF-receptor complexes were precipitated during a 15.min incubation at 4 "C by addition of 500 ~1 of 25% polyethylene glycol and 500 ~1 of 0.1% rabbit globulin.
After autoradiography, lZ51-TNF was detected in M, 130,000, 90,000, and 66,000 complexes (Fig. 2, lane I). High concentrations of unlabeled TNF inhibited cross-linking of Y-TNF into each of these complexes (lane 2). Therefore, uptake of ""I-TNF into these complexes was specific. In some preparations of PM, an additional specific M, 117,000 complex was also observed (lane 1).

lz51-TNF Binding to Triton X-100~solubilized
Liver PM-To examine whether there are latent receptors for TNF in PM, membranes were solubilized into 1% Triton X-100 and incubated with 0.04-30 nM lz51-TNF (Fig. lA). Scatchard transformation (20) and computer analysis (21) of the equilibrium binding data resolved high and low affinity binding sites for TNF ( Fig. 1B and Table I). The Kd values with which these sites bound TNF were comparable to those of intact PM. However, TNF binding capacity was significantly increased by detergent solubilization of membrane proteins. The number of high affinity binding sites increased 5.1-fold (from 0.33 f 0.04 to 1.67 f 0.05 pmol/mg of protein); the number of low affinity binding sites increased 3.9-fold (from 1.92 f 0.16 to 7.57 f 0.50 pmol/mg of protein). Thus, a large population of latent TNF receptors reside within rat hepatocyte PM.
The structural properties of the latent receptors were char-  Membranes prepared as described under "Experimental Procedures" were incubated with increasing concentrations of "'I-TNF (0.04-30 nM) at 4 "C for 24 h, and specific binding was determined. Dissociation constants (Kd) and the number of binding sites (B,.,) were determined by the method of Scatchard (20) and computer analysis using LIGAND (21). Kd  Membranes, prepared as described under "Experimental Procedures," were incubated with 2.5 nM '""I-TNF in the absence (lanes I, 3, 5, and 7) or presence (lanes 2, 4, 6, and 8) of 1 pM unlabeled TNF at 4 "C for 24 h. After affinity labeling, membranes were fractionated on a 4-20% acrylamide gradient gel. Lanes I and 2, PM; lanes 3 and 4, solubilized PM; lanes 5 and 6, chymotrypsin-treated PM; lanes 7 and 8, membranes chymotrypsin-treated and then solubilized. Molecular weight standards were myosin (M, 200,000), &galactosidase (M, 116,250), phosphorylase b (M, 92,500), bovine serum albumin (M, 66,200), and ovalbumin (M, 45,000). acterized by affinity cross-linking (Fig. 2). As in PM, '""I-TNF was covalently incorporated into M, 130,000, 90,000, and 66,000 complexes in detergent extracts of PM (lane 3). However, the density of the ""I-TNF-receptor complexes was increased greatly by solubilization relative to that in intact PM (compare lanes 1 and 3). Thus, results from affinity crosslinking experiments as well as from receptor binding assays show that hepatic PM possesses latent high and low affinity TNF receptors that appear structurally indistinguishable from those initially available.
Chymotrypsin digestion of membranes before solubilization enabled us to further characterize the latent pool of receptors. As shown in Table I and Fig. 3, chymotrypsin treatment diminished the binding capacity of PM by -80% and increased the Kd of the lower affinity sites for TNF from 13.58 + 1.45 to 26.13 f 5.48 nM (p < 0.05). The ability of chymotrypsin to degrade exposed binding sites finds counterpart in affinity labeling experiments in which cross-linking of 1251-TNF into covalent complexes is all but abolished (Fig. 2, lane  5). Chymotrypsin-treated PM were solubilized and then incubated with increasing concentrations (0.04-30 nM) of lz51-TNF. As shown in Table I  affinities of both the high and low affinity sites in membranes solubilized from chymotrypsin-treated PM were not significantly different from those of intact PM (high affinity site: 1.51 + 0.35 nM in PM uersus 1.29 f 0.13 nM in membranes chymotrypsin-treated and then solubilized, low affinity site: 13.58 f 1.45 nM in PM uersus 11.37 f 1.34 nM in membranes chymotrypsin-treated and then solubilized). Additionally, affinity labeling shows that receptors solubilized from chymotrypsin-digested membranes were not structurally distinguishable from those detected on intact PM (Fig. 2, lane 7). These data further demonstrate that the properties of latent and exposed TNF receptors are very similar.
Up-regulation of TNF Receptors by Pretreatment with TNF-We next conducted experiments to determine if ex-posure of latent receptors could be regulated by TNF. Thus, PM were incubated in the absence or presence of TNF (50 nM) for 3 h at 23 "C (Phase I incubation).
Unbound TNF was binding capacity is substantially increased upon exposure of the latent receptor population.
High and low affinity receptors for TNF have been described in human monocytes (22), MCF-7 human breast carcinoma cells (33), murine adipocytes (34), and ML-l and HL-60 human myeloid leukemia cells (35), although a single class of receptors is present in a variety of normal and transformed cells (36). Ishikura et al. (35) have related high affinity binding of TNF by ML-l and HL-60 cells to differentiation and proposed that the lower affinity interactions inhibit this process. The functional significance of the heterogeneous TNF receptors in hepatic tissue has not been elucidated.
washed from the membranes, which were then incubated in TNF-free medium for 30 min at 23 "C to dissociate TNF bound to membranes during the Phase I incubation.
Control experiments (not shown) demonstrate that this condition permits dissociation of virtually all TNF bound to membranes during the Phase I incubation.
Control and TNF-treated membranes were finally incubated (Phase II incubation) with 5 nM ""I-TNF for 24 h at 4 "C. The low temperature was used to "freeze" the membranes, thereby preventing possible return to latency of receptors "exposed" during the Phase I incubation. As shown in Fig. 4, significantly more '*"I-TNF bound to TNF-pretreated PM than to control PM (p < 0.05, twotailed t test). These results show that TNF had recruited receptors into the exposed pool from the latent pool.

DISCUSSION
Peptide hormone receptors are continuously removed from and added to PM of target cells. The balance between upand down-regulation determines the number of receptors expressed at the cell surface and, consequently, cellular responsiveness to hormonal stimulation (4, 5). Homologous downregulation induced by TNF (22,23) or heterologous downregulation induced by interleukin-1 (24), endotoxin (25), or activators of protein kinase C (24, 26-28) results in redistribution of TNF receptors from PM into the cell. In MCF-7 human breast adenocarcinoma cells, after internalization, cell-surface receptors are replenished by recycling and recruitment from an intracellular pool (29). In various transformed cell lines, interferon-y (30, 31) and lectins (32) increase the complement of receptors at the cell surface, the former by promoting synthesis and the latter by diminishing the rate at which TNF-receptor complexes are internalized. In this study, we have identified a novel population of latent TNF receptors that reside in rat liver PM. Isolation of latent TNF receptors within PM distinguishes these sites from newly synthesized receptors in transit to the cell surface and from receptors in the process of recycling. The exposed and latent populations of hepatic TNF receptors were characterized as high affinity, low capacity and low affinity, high capacity TNF receptors. TNF receptors solubilized from liver PM and from chymotrypsin-treated membranes faithfully reflect the receptor heterogeneity of intact PM, although Affinity labeling experiments present further insight into the nature of the latent and exposed populations of TNF receptors. ""I-TNF was detected in M, 130,000, 90,000, and 66,000 complexes after incubation with liver PM, solubilized PM, and PM that had been digested with chymotrypsin and then solubilized. Consistent with the presence of latent receptors severalfold more abundant than those initially accessible to TNF, solubilization of control or chymotrypsin-digested membranes before affinity labeling resulted in increased cross-linking of lZ51-TNF into all three complexes relative to uptake by intact PM. In control experiments, ' we have found that a variety of protease inhibitors increase the specific binding of '*"I-TNF to PM and its cross-linking into each of the covalent complexes. Thus, it is unlikely that the M, 66,000 complex is produced by proteolysis of the higher M, species. Reduction with dithiothreitol fails to release lower M, species from any of the ligand-receptor complexes. This result shows that hepatic TNF-binding proteins do not contain subunits retained in high molecular weight aggregates by interchain disulfide bonds. Thus, affinity labeling and equilibrium binding assays demonstrate the presence of a heterogeneous pool of latent TNF-binding proteins in hepatic PM. Affinity labeling has been employed to characterize TNF receptors in many types of cells (33,(37)(38)(39). In a recent study (38), the selective immunoprecipitation with antireceptor antibodies of M, 75,000 and 95,000 cross-linked complexes from epithelial cells, but not an M, 100,000 cross-linked complex from myeloid cells, provided evidence that different types of cells contain different receptors for TNF. The cDNAs for two types of TNF receptors have recently been cloned (40-42). These receptors differ in transcript size, amino acid sequence, potential glycosylation sites, and cellular distribution. The relationship between these cloned receptors and the receptors we have characterized in rat liver PM remains to be determined.
Another novel observation described in this study is the demonstration that TNF promotes exposure of latent receptors from a pre-existing pool present in the hepatic cell membrane. Once solubilized from the membrane, latent and exposed receptor populations were indistinguishable.
Receptors in their latent and exposed states differed only in their susceptibility to proteolysis by chymotrypsin. We have previously shown (43,44) that hormone-induced conformational changes alter the affinity with which insulin is bound by its receptors and also the lability of such sites to proteolysis by exogenous trypsin. The present results lead us to speculate that latent receptors, with little or no affinity for TNF, exist in aggregates or as conformers resistant to proteolysis by exogenous proteinases.
TNF binding to accessible receptors may induce allosteric conformational changes leading to exposure of latent sites.
TNF is elaborated by macrophages in response to bacteria, viruses, parasites, and also cancer (l-3). In response to such 23. stimuli, breakdown of skeletal muscle and fat occurs coordinately with a shift to anabolic metabolism in the liver and other tissues (l-3). The latent receptors described by this 24 ' study may be important for the sustained response of the liver 25. to TNF during infection. 26.