Monoclonal antibodies that inhibit IgE binding

Four monoclonal antibodies were produced that inhibit IgE binding to the high affinity IgE receptor (Fc epsilon R) on rat basophilic leukemia cells. The four monoclonal antibodies (mAb) fall into two groups. The first group was comprised of 3 antibodies (mAb BC4, mAb CD3, and mAb CA5) that reacted with the Fc epsilon R at epitopes close or identical to the IgE-binding site. With 125I-labeled antibodies there was reciprocal cross-inhibition between the antibodies and IgE. The antibodies activated both RBL-2H3 cells and normal rat mast cells for histamine release. The 3 antibodies immunoprecipitated the previously described alpha, beta, and gamma components of the receptor. The number of radiolabeled Fab fragments of 2 of these antibodies bound per cell was similar or equal to the number of IgE receptors. In contrast, the mAb BC4 Fab bound to 2.1 +/- 0.4 times the number of IgE receptor sites. Therefore, the portion of the Fc epsilon R exposed on the cell surface must have two identical epitopes and an axis of symmetry. These 3 monoclonal antibodies recognize different but closely related epitopes in the IgE-binding region of the Fc epsilon R. The fourth monoclonal antibody (mAb AA4) had different characteristics. In cross-inhibition studies, IgE and the other 3 monoclonals did not inhibit the binding of this 125I-labeled monoclonal antibody. The number of molecules of this antibody bound per cell was approximately 14-fold greater than the Fc epsilon R number. This monoclonal antibody caused the inhibition of histamine release and it appears to bind to several cell components.     

The fate of IgE bound to rat basophilic leukemia cells.

The present study investigates the fate of the cell-bound IgE by using a well-characterized rat basophilic leukemia cell line and a purifed IgE myeloma protein. Both histamine-releasing and nonreleasing cell lines were examined. In both cases, no evidence for cell-mediated IgE catabolism could be elicited. Both the dissociated IgE and the receptors remained intact for prolonged periods of time, as demonstrated by binding assays. Internalization and/or recycling of membrane-bound IgE could not be demonstrated by E. M. autoradiography. We found only limited time-dependent changes in accessibility to anti-IgE antibody, trypsin, or elution at low pH (2.9 to 3.1). A biphasic dissociation of cell-bound 125I-IgE during incubation in the presence of excess unlabeled IgE was reproducibly observed; the more slowly dissociated IgE was also less readily dissociated at pH 3.4. These studies lead us to conclude that, in vitro, IgE resides in a functional orientation on the surface of RBL-1 cells, for prolonged periods of time.     

Characterization of the target cell receptor for IgE. II. Polyacrylamide gel analysis of the surface IgE receptor from normal rat mast cells and from rat basophilic leukemia cells.

Purified rat peritoneal mast cells (RMC) and cultured rat basophilic leukemia (RBL) cells were surface labeled with 125I by using lactoperoxidase, incubated with unlabeled rat monoclonal IgE and subjected to solubilization by treatment with Nonidet P-40 (NP-40). With both cell types significant amounts of radioiodinated material could be specifically precipitated by a "sandwich" system consisting of rabbit anti-rat epsilon-chain and goat anti-rabbit Ig. The precipitates were dissociated with sodium dodecyl sulfate (SDS) and urea and subsequently analyzed by SDS-polyacrylamide gel electrophoresis. With RMC three radioactive bands were seen. One corresponded to IgE present on the RMC at the time of isolation. A small band migrating in the region of light chain was seen with both sepcific (anti-IgE) and control precipitates. It showed no demonstrable relationship to IgE. The major radioactive band corresponded to a m.w. of 62,000. This band was dependent upon the presence of IgE and was not found when non-IgE binding control cells were used. With RBL cells, only the IgE-dependent 62,000 dalton peak was present. Saturation of the IgE receptor sites of the RMC or RBL cells before lactoperoxidase labeling almost totally eliminated this radioactive band, indicating that cell-bound IgE rendered this membrane component inaccessible to the radiolabel. These results strongly suggest that this cellular component is identical, at least in part, with the target cell surface receptor for reaginic antibody. The data also further support the hypothesis that the neoplastic RBL cells have a normal surface receptor for IgE.     

Characterization of the IgE receptor isolated from human basophils.

The IgE receptor of human basophils was purified by using simple and repetitive affinity chromatography on human IgE-Sepharose. Basophils were partially purified from peripheral blood of patients with chronic myelogenous or basophilic leukemia. Cells were labeled with 125I by using the lactoperoxidase method and were solubilized with nonionic detergent. Elution of IgE-Sepharose with 0.5 N acetic acid, 1% NP-40 allowed recovery of active IgE receptor. Analysis of human IgE receptor by SDS polyacrylamide gel electrophoresis with 10% gels demonstrated one major radioactive peak with an apparent m.w. of 58,000 to 68,000, somewhat larger than rat IgE receptor. The purified human IgE receptor was active since approximately 10 to 42% of labeled receptor could specifically rebind to insolubilized human IgE. Rebinding was blocked by nanomolar concentrations of soluble human IgE or rat IgE but not by human or rat IgG, heat-inactivated human IgE, or heat-aggregated human IgG; thus it appears that rat IgE receptor. The relative abilities of active rat IgE and active human IgE to inhibit human IgE receptor rebinding could not be precisely determined because of the limitations in assessing the proportion of human IgE that retains receptor-binding activity.     

An intestinal galactose-specific lectin mediates the binding of murine IgE to mouse intestinal epithelial cells.

A number of lactose-binding lectins have recently been identified in the rat and mouse intestine, one of which corresponds to the C-terminal domain of IgE-binding proteins, originally identified in rat basophilic leukemia (RBL) cells and mouse 3T3 fibroblasts. In the present report, we describe the affinity purification of a rat intestinal lactose-specific lectin which binds murine IgE antibodies. This binding most likely occurs via the immunoglobulin carbohydrate chains, as it is inhibited by lactose. This intestinal lectin molecule is also immunologically related to the previously described IgE-binding protein (epsilon BP) isolated from RBL cells, since it is recognized by antibodies raised against recombinant epsilon BP. This intestinal form of epsilon BP has a molecular mass of 17.5 kDa, which is much lower than that of its RBL cell analogue (31 kDa). The attachment of IgE to the mouse intestinal epithelium was demonstrated by immunohistochemistry, along with the presence of a corresponding mouse intestinal epsilon BP. The carbohydrate-dependent nature of this attachment was established by demonstrating that IgE binding to mouse epithelium was specifically abolished by lactose (4 mM) and by a blood-group-A-active tetrasaccharide (0.2 mM), but not by mannose (10 mM). Finally, the association of IgE with the mouse intestinal epithelium was prevented by competition with the purified IgE-binding lectin isolated from rat intestine. Although the physiological function of this intestinal protein is still unknown, the finding that IgE binds to a lectin in the intestinal epithelium pinpoints a possible novel mechanism for the regulation of IgE-mediated disorders, such as food allergy.     

Inhibition of IgE binding to RBL-2H3 cells by a monoclonal antibody (BD6) to a surface protein other than the high affinity IgE receptor.

A mAb was isolated (mAb BD6) that recognized a surface glycoprotein on rat basophilic leukemia cells (RBL-2H3). The antibody bound to 2 x 10(6) molecules/cell and specifically blocked IgE binding (50% inhibition with 3.48 +/- 0.51 micrograms/ml; mean +/- SEM), although neither IgE nor anti-high affinity IgE receptor (anti-Fc epsilon RI) mAb blocked mAb BD6 binding to the cells. mAb BD6 did not affect the rate of dissociation of cell-bound IgE, nor did it induce or inhibit the internalization of IgE. mAb BD6 did not release histamine. However, it did cause rapid spreading of the cells. By 1 h the cells had retracted to a spherical shape with their surface covered with membranous spikes, and they could easily be detached from the tissue culture plate. These changes differed from those observed after Fc epsilon RI activation. mAb BD6 immunoprecipitated a complex of two proteins, 38 to 50 kDa and 135 kDa from 125I-surface labeled rat basophilic leukemia cells that are not subunits of Fc epsilon RI. Chemical cross-linking studies showed that these molecules are associated on the cell surface. By immunoblotting, mAb BD6 reacted with a 40-kDa protein. Therefore, mAb BD6 binds to a surface protein that is close to the Fc epsilon RI and sterically inhibits 125I-IgE binding.     

Murine and rat IgE: relationships in terms of binding to cell receptors and to antibodies against rat epsilon chain.

The similarity between murine and rat IgE was examined in terms of their fixation to target cells and interaction with monospecific antibodies to rat epsilon-chain (anti-epsilon). Purified rat monoclonal IgE (IgEr) was found to block the fixation of murine reagin (IgEm) to mouse and rat skin and to rat basophilic leukemia (RBL) cells. The capacities of mouse reaginic serum (MRS), rat reaginic serum, and IgEr to inhibit the binding of radiolabeled 125I-IgEr to RBL cells were shown to be similar. These results suggest that the binding of IgE of either species occurs on the same or on adjacent receptor sites of mast cells and RBL cells. The antigenic cross-reactivity between IgEm and IgEr was established by depletion of the reaginic activity from MRS by treatment of MRS with anti-epsilon. The reaginic activity of MRS could be recovered by the addition of IgEr to anti-epsilon:IgEm complexes. From these findings it may be inferred that i) IgEm and IgEr share some antigenic determinants and ii) the regions of the immunoglobulins responsible for fixation to receptors on mast cells and RBL cells are identical or similar.     

Characterization of the radioiodinated analogue of SCH 23390: in vitro and in vivo D-1 dopamine receptor binding studies

A new radioiodinated molecule, 125I-SCH 38840 (previously referred to as 125I-SCH 23982), has been recently reported to be a D-1 dopamine receptor ligand. The current study confirms and expands the characterization of both the radiolabeled and unlabeled forms of this compound, as well as describing the development of an in vivo D-1 receptor binding assay utilizing the 125I-SCH 38840. The binding of 125I-SCH 38840 to rat striatal membranes, in vitro, was saturable and exhibited a KD of 1.47 nM. Competition studies using 125I-SCH 38840 exhibited a pharmacological profile consistent with the proposal that 125I-SCH 38840 was binding to the D-1 receptor. Further studies with the unlabeled SCH 38840 demonstrated that it inhibited dopamine-stimulated adenylate cyclase with a KI of 66.1 nM, indicating that SCH 38840 was acting as a D-1 antagonist. Behavioral studies demonstrated that SCH 38840 (MED = 1.0 mg/kg, s.c.) blocked conditioned avoidance responding in rats, a measurement considered predictive of anti-psychotic activity in man. In vivo binding of 125I-SCH 38840 to rat striatum following s.c. administration was specific. Peak striatal levels were observed 1 h after injection, with measurable binding observed out to 8 h post-treatment. The displacement of the in vivo binding by unlabeled standards again suggested a D-1 selective interaction. The half-life of the in vivo binding of 125I-SCH 38840 was approximately 1.25 h, and was nearly equivalent to the half-life of the anti-CAR activity of unlabeled SCH 38840. These results clearly demonstrate the D-1 nature of SCH 38840's behavioral activity and strengthen the anti-psychotic potential of a D-1 antagonist.     

Surface expression of functional IgE binding protein, an endogenous lectin, on mast cells and macrophages.

IgE-binding protein (epsilon BP) is a galactoside-specific lectin containing an S-type carbohydrate-recognition domain. It was originally identified in rat basophilic leukemia cells and is now known to be identical to a macrophage surface Ag, Mac-2, and lectins designated as CBP 35/L-34/RL-29. It has also been related to a nonintegrin laminin-binding protein isolated from mouse macrophages. In this report we have shown the following: epsilon BP is present in variable amounts in several mast cell lines, and the surface expression of epsilon BP in these cell lines is quite variable and does not correlate with the total amount of epsilon BP in the cell. epsilon BP is displayed on the cell surface in a manner that is reversible by lactose, most likely through attachment to yet unidentified glycoconjugates. The putative epsilon BP binding sites on the cell surface can be readily demonstrated by using radiolabeled epsilon BP, and the sites are present in comparable amounts in various cell lines. Expression of epsilon BP on the cell surface can be regulated; the most notable example is the upregulation of surface epsilon BP on RBL cells activated through the high-affinity IgE receptor by IgE immune complexes. Cell-surface epsilon BP is functional as measured by its ability to promote adhesion of trypsinized rabbit erythrocytes to mast cells and macrophages. On the basis of these results and reported properties of related lectins, we propose that the lectin represented by epsilon BP is a new class of cell-adhesion protein.     

Somatostatin receptors on rat cerebrocortical membranes. Structural characterization of somatostatin-14 and somatostatin-28 receptors and comparison with pancreatic type receptors

Somatostatin binding and cross-linking to its receptors on rat cerebrocortical membranes were characterized with [125I-Tyr1]somatostatin-14 and [125I-Leu8, D-Trp22, Tyr25]somatostatin-28. When [125I-Tyr1]somatostatin-14 was cross-linked to its receptors with the photoreactive cross-linker, N-(5-azido-2-nitrobenzoyloxy)succinimide, the hormone was specifically associated with a Mr = 72,000 protein band in the presence or absence of reducing agents. Affinity labeling of the Mr = 72,000 protein band was decreased with increasing concentrations of unlabeled somatostatin-14 and nonhydrolyzable guanine nucleotide analog, guanyl-5'-yl imidodiphosphate (Gpp(NH)p). Pretreatment of cerebrocortical membranes with islet-activating protein resulted in a decrease in subsequent labeled somatostatin-14 binding and affinity-labeling of the protein and abolished an inhibitory effect of somatostatin-14 on vasoactive intestinal peptide-stimulated increase in adenylate cyclase activity. When the affinity-labeled protein was solubilized with Zwittergent 3-12 and adsorbed to wheat germ agglutinin-agarose, it was eluted by N-acetylglucosamine. [125I-Leu8, D-Trp22, Tyr25]somatostatin-28 cross-linking to cerebrocortical and pancreatic membranes with the same photoreactive agent revealed specifically labeled protein bands of a Mr = 74,000 in cerebrocortical membranes and a Mr = 94,000 in pancreatic membranes, respectively. These results suggest that: 1) somatostatin receptor on cerebrocortical membranes is a monomeric glycoprotein with a Mr = 70,000 binding subunit, coupled to guanine nucleotide regulatory protein, and 2) the Mr = 70,000 protein may be a common receptor for somatostatin-28 and somatostatin-14 and is distinct from a common pancreatic type receptor.     

Purification of a receptor for formaldehyde-treated serum albumin from rat liver

A membrane-associated receptor involved in a specific uptake of formaldehyde-treated serum albumin (f-Alb) was purified from rat livers by Triton X-100 solubilization of a 105,000 X g membrane preparation and affinity chromatography on an f-Alb-Sepharose column. The purified receptor exhibited Mr = 125,000, consisting of two noncovalently linked glycoprotein components with Mr = 53,000 and Mr = 30,000, respectively. Incubation of 125I-receptor with f-Alb, but not with native albumin, resulted in a marked shift of pI value from 5.9 to 5.1, reflecting the presence of a specific ligand-receptor interaction. The receptor incorporated into liposomes displayed a saturable binding to 125I-f-Alb and the binding was effectively replaced by the presence of unlabeled f-Alb, with binding parameters being similar to those obtained from 125I-f-Alb binding to the sinusoidal liver cell membrane (Horiuchi, S., Takata, K., and Morino, Y. (1985) J. Biol. Chem. 260, 475-481). Reaction of anti-f-Alb receptor antibody with extracts of sinusoidal cells resulted in a specific precipitation of two proteins whose molecular weights were identical to those for the purified receptor. The anti-receptor IgG fraction effectively blocked 125I-f-Alb binding to the sinusoidal cell membranes. These results indicate that the purified protein represents the membrane-associated receptor which is presumably involved in a specific uptake of this ligand from the circulation.     

Triggering of cultured neoplastic mast cells by antibodies to the receptor for IgE.

Cell surface receptors for IgE were isolated from detergent lysates of iodinated, IgE-saturated, rat basophilic leukemia cells by precipitation with anti-IgE antibodies followed by chromatography at acid pH. The isolated material showed a single 125I-band (m.w. approximately 58,000) on gel electrophoresis in sodium dodecyl sulfate and was used to immunize a rabbit. The resulting anti-serum was reacted with lysates of surface iodinated mouse or rat tumor mast cells. Analysis of the precipitates on (10%) gel electrophoresis revealed one major peak comprising greater than 80% of the detectable counts and having an estimated m.w. of approximately 58,000. The antiserum reacted with detergent-solubilized and cell-bound receptors in the presence or absence of excess IgE; it also inhibited the binding of 125I-IgE. Cultured mouse mastocytoma cells never exposed to IgE released 3H-serotonin when incubated with F(ab')2, but not Fab' fragments of the antiserum, which had been rigorously freed of IgE and anti-IgE. The release was inhibited in the presence of excess IgE, was Ca++ dependent, and equaled 80% of the maximum obtained with IgE and anti-IgE. We conclude that aggregation of the receptors for IgE provides the critical signals for cell activation.     

The fate of IgE bound to rat basophilic leukemia cells. II. Endocytosis of IgE oligomers and effect on receptor turnover

We have assessed the internalization of variously sized oligomers of IgE bound to rat basophilic leukemia (RBL) cells by measuring their accessibility to the extracellular environment, and by direct visualization of the radiolabeled ligands. We also followed the fate of the internalized ligands and their receptors, as well as the fate of the free receptor on cells internalizing oligomers. In contrast to monomeric IgE, surface-bound oligomeric IgE was internalized. Notably, dimers provided an effective signal for internalization, although larger oligomers seem to be internalized more efficiently. In our experiments, 48% of the cell-bound dimers and 67% of the trimers were eliminated from the cell surface in 180 min. One-half of the maximal internalization observed with dimers and trimers occurred in 25 and 11 min, respectively. Release of radioactivity into the supernatant followed internalization; the released radioactivity did not bind to fresh cells and was only partially TCA-precipitable. Radioactive ligands remaining associated with the cells were unchanged as judged by m.w; they also were shown to remain receptor-bound. During either internalization or release of substantial amounts of the originally cell-bound oligomers, there was no increase in IgE-binding activity. In contrast, there was a transient drop (25%) in the number of free surface receptors suggesting internalization of the free receptors together with the oligomer-occupied receptor. Cells that failed to release histamine (RBL-I) processed dimeric and trimeric IgE similarly to histamine-releasing (RBL-2H3) cells. We conclude that dimeric and trimeric IgE are internalized by RBL cells and later are released to the medium in a partially degraded form. The ligand-bound receptor seems to be internalized with the ligand, along with some free receptor, and does not appear to be reusable or to recycle rapidly to the cell surface.     

Lactoperoxidase-catalyzed iodination of the receptor for immunoglobulin E at the cytoplasmic side of the plasma membrane

The structural organization of the high affinity receptor for immunoglobulin E has been investigated in plasma membrane vesicles from rat basophilic leukemia cells using lactoperoxidase-catalyzed 125I-iodination to label exposed polypeptide regions. Intact vesicles are predominantly right-side-out in orientation, and lactoperoxidase iodination of these vesicles results in labeling of the alpha subunit of receptor but not the beta and gamma subunits. Lysis of these vesicles to expose the cytoplasmic face of the membrane by two different methods permits labeling of the beta and gamma subunits with no increase in labeling of alpha. The results indicate that both the beta and gamma subunits of the receptor have segments exposed at the cytoplasmic side of the plasma membrane. These studies have also revealed a previously unidentified IgE binding component in the membrane vesicles; its 125I-labeling characteristics and some other properties are described.     

Evidence for vasoactive intestinal peptide receptors in apical membranes from tracheal epithelium

[125I]VIP (vasoactive intestinal peptide) bound to apical membranes isolated from the bovine tracheal epithelium with a half maximal inhibition by unlabeled VIP (IC50) of 0.6 x 10(-9)M and binding was reversible. Glucagon did not affect [125I]VIP binding to the membranes. [125I]VIP was covalently cross-linked to tracheal membrane proteins using disuccinimidyl suberate. SDS-polyacrylamide gel electrophoresis of labeled tracheal membranes revealed one major [125I]-receptor complex of Mr = 71,000 to which binding of [125I]VIP was inhibited by 10 microM unlabeled VIP. These results are consistent with the presence of a specific, high-affinity receptor for VIP, with a Mr = 71,000, in apical membrane vesicles isolated from the bovine tracheal epithelium.     

Clustering, mobility, and triggering activity of small oligomers of immunoglobulin E on rat basophilic leukemia cells

We have recently shown that small oligomers of IgE bound to univalent receptors for IgE on the surface of rat basophilic leukemia cells induce extensive aggregation of the receptors at 4 degrees C into patches resolvable by fluorescence microscopy and that this does not occur with monomeric IgE (Menon, A. K., D. Holowka, and B. Baird, 1984, J. Cell Biol. 98:577-583). Here we use fluorescence photobleaching recovery measurements to show that receptor oligomerization by this means is accompanied by a dramatic reduction of receptor lateral mobility, and that this immobilization occurs even when the clustering is not microscopically detectable. Furthermore, the degree of immobility induced by a particular oligomer fraction from a gel filtration column correlates positively with its ability to trigger cellular degranulation, whereas receptors labeled with monomeric IgE have no triggering activity and exhibit typical membrane protein mobility. The slow, large-scale oligomer-induced clustering appears to be a long term consequence of earlier selective interactions that result in receptor immobilization, and this highly clustered state provides a competent, noninhibitory triggering signal resulting in cellular degranulation upon warming to 37 degrees C. We conclude that even limited clustering of IgE receptors on rat basophilic leukemia cells induces interactions with other cellular components that constrain receptor mobility and eventually cause massive coalescence of the clusters. These primary selective interactions occurring at the level of receptor oligomers or small clusters of oligomers that result in immobilization may play a role in triggering cellular degranulation.     

A cell surface glycoprotein of rat basophilic leukemia cells close to the high affinity IgE receptor (Fc epsilon RI). Similarity to human melanoma differentiation antigen ME491.

A monoclonal antibody (mAb), AD1, was isolated that recognized a cell surface protein on rat basophilic leukemia cells (RBL-2H3). At high concentration, this antibody inhibited IgE-mediated but not calcium ionophore-induced histamine release (49% inhibition at 100 micrograms/ml). The mAb AD1 did not inhibit the binding of IgE or of several antibodies directed to the high affinity IgE receptor (Fc epsilon RI). Likewise, IgE did not inhibit mAb AD1 binding. However, several anti-Fc epsilon RI antibodies did inhibit mAb AD1 binding as intact molecules but not as Fab fragments. Therefore, the sites on the cell surface to which mAb AD1 binds are close to Fc epsilon RI. The mAb AD1 immunoprecipitated a broad, 50-60-kDa band from 125I-surface-labeled RBL-2H3 cells that upon peptide N-glycosidase F treatment was transformed into a sharp 27-kDa band. A similar 27-kDa protein was immunoprecipitated from surface-radiolabeled cells after culture with tunicamycin. Thus, the protein recognized by mAb AD1 is highly glycosylated with predominantly N-linked oligosaccharides. The N-terminal sequence of 43 amino acids was found to be different from any subunit of Fc epsilon RI but nearly identical to that of the human melanoma-associated antigen ME491. Therefore, mAb AD1 binds to a surface glycoprotein on RBL-2H3 cells sterically close to the Fc epsilon RI but distinct from the recognized subunits of the receptor.     

Multimeric structure of the tumor necrosis factor receptor of HeLa cells

The tumor necrosis factor (TNF) receptor of HeLa cells was solubilized in Triton X-100 and characterized by gel filtration, affinity labeling, and ligand blotting studies. Receptors solubilized with Triton X-100 eluted in gel filtration as a major peak of Mr = 330,000 and retained high affinity binding (KD = 0.25 nM). Affinity labeling of soluble receptor/125I-TNF complexes using the reversible, bifunctional bis[2-(succinimidooxycarbonyl-oxy)ethyl] sulfone resulted in the formation of cross-linked species of Mr = 310,000, 150,000-175,000, 95,000, and 75,000. The formation of these complexes was competitively inhibited by unlabeled TNF. Partial reversal of cross-linking in these complexes and their analysis by two-dimensional sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) resolved 125I-TNF dimers cleaved from the 95,000 band and 125I-TNF monomer cleaved from the 75,000 band, providing evidence for a Mr approximately 60,000 subunit. In addition, the 95,000 and 75,000 bands were resolved as components of larger complexes (Mr = 150,000-175,000), which presumably contain two receptor subunits. The Mr 95,000 and 75,000 bands were also released from the Mr 310,000 complex by reduction with dithiothreitol, suggesting a role for disulfide bond stabilization. To investigate the association of the putative receptor subunits, Triton X-100 extracts from HeLa membranes were fractionated by SDS-PAGE without reduction and transferred electrophoretically to nylon membranes for TNF binding assays. Only two bands of Mr = 60,000 and 70,000 specifically bound TNF, and higher Mr binding activity was not observed. These results indicate that TNF receptors in HeLa cells are high molecular weight complexes containing Mr = 60,000 and 70,000 subunits each capable of binding TNF and that the complexes are primarily stabilized by non-covalent, hydrophobic interactions.