Search

B7-2 is a recently discovered, second ligand for the CTLA-4/CD28, T cell signaling system. Using the GL-1 rat monoclonal antibody (mAb), we monitored expression of B7-2 on mouse leukocytes with an emphasis on dendritic cells. By cytofluorography, little or no B7-2 was detected on most cell types isolated from spleen, thymus, peritoneal cavity, skin, marrow, and blood. However, expression of B7-2 could be upregulated in culture. In the case of epidermal and spleen dendritic cells, which become highly immunostimulatory for T cells during a short period of culture, the upregulation of B7-2 was dramatic and did not require added stimuli. Lipopolysaccharide did not upregulate B7-2 levels on dendritic cells, in contrast to macrophages and B cells. By indirect immunolabeling, the level of staining with GL-1 mAb exceeded that seen with rat mAbs to several other surface molecules including intercellular adhesion molecule 1, B7-1, CD44, and CD45, as well as new hamster mAbs to CD40, CD48, and B7-1/CD80. Of these accessory molecules, B7-2 was a major species that increased in culture, implying a key role for B7-2 in the functional maturation of dendritic cells. B7-2 was the main (> 90%) CTLA-4 ligand on mouse dendritic cells. When we applied GL-1 to tissue sections of a dozen different organs, clear-cut staining with B7-2 antigen was found in many. B7-2 staining was noted on liver Kupffer cells, interstitial cells of heart and lung, and profiles in the submucosa of the esophagus. B7-2 staining was minimal in the kidney and in the nonlymphoid regions of the gut, and was not observed at all in the brain. In the tongue, only rare dendritic cells in the oral epithelium were B7-2+, but reactive cells were scattered about the interstitial spaces of the muscle. In all lymphoid tissues, Gl-1 strongly stained certain distinct regions that are occupied by dendritic cells and by macrophages. For dendritic cells, these include the thymic medulla, splenic periarterial sheaths, and lymph node deep cortex; for macrophages, the B7-2-rich regions included the splenic marginal zone and lymph node subcapsular cortex. Splenic B7-2+ cells were accessible to labeling with GL-1 mAb given intravenously. Dendritic cell stimulation of T cells (DNA synthesis) during the mixed leukocyte reaction was significantly (35-65%) blocked by GL-1.(ABSTRACT TRUNCATED AT 400 WORDS)

We tested whether the in vivo infusion of recombinant, soluble CTLA4 fused with Ig heavy chains, as a surrogate ligand used to block CD28/CTLA4 T-cell costimulation, could prevent efficient T-cell activation and thereby reduce graft-versus-host disease (GVHD). Lethally irradiated B10.BR recipients of major histocompatibility complex disparate C57BL/6 donor grafts received intraperitoneal injections of human CTLA4-Ig (hCTLA4-Ig) or murine CTLA4-Ig (mCTLA4-Ig) in various doses and schedules beginning on day -1 or day 0 of bone marrow transplantation (BMT). In all five experiments, recipients of CTLA4-Ig had a significantly higher actuarial survival rate compared to mice injected with an irrelevant antibody control (L6) or saline alone. Survival rates in recipients of hL6 or PBS were 0% at 29 to 45 days post-BMT. In recipients of CTLA4-Ig, survival rates were as high as 63% mice surviving 3 months post-BMT. However, protection was somewhat variable and recipients of CTLA4-Ig were not GVHD-free by body weight, clinical appearance, and histopathologic examination. There were no significant differences in the survival rates in comparing injection dose, injection duration, or species of CTLA4-Ig (hCTLA4-Ig v mCTLA4- Ig). Splenic and peripheral blood flow cytometry studies of long-term hCTLA4-Ig-injected survivors showed a significant peripheral B-cell and CD4+ T-cell lymphopenia, consistent with GVHD. A kinetic study of splenic reconstitution was performed in mice that received hCTLA4-Ig and showed that mature splenic localized CD8+ T-cell repopulation was not significantly different in recipients of hCTLA4-Ig compared with hL6, despite the significant increase in actuarial survival rate in that experiment. These data suggest that the beneficial effect of hCTLA4-Ig on survival is not mediated by interfering with mature donor- derived T-cell repopulation post-BMT. Neither hCTLA4-Ig nor mCTLA4-Ig interfered with hematopoietic recovery post-BMT. We conclude that CTLA4- Ig (most likely in combination with other agents) may represent an important new modality for GVHD prevention.

Haemophilus influenzae type b capsular polysaccharide (PRP) conjugate vaccines, which are thought to induce T cell-dependent antibody production, induce protective responses after a single dose in individuals under 15 months of age. However, multiple doses of these vaccines are required to induce protective antibody responses in infants, with the exception of PRP conjugated to meningococcal outer membrane proteins (OMPC), which does so after a single dose. The basis for this difference is not fully understood, although others have proposed that OMPC and porins, the major protein component of OMPC, act as adjuvants or mitogens. In this report OMPC is shown to enhance CD40 ligand-mediated, T cell-dependent antibody production in mice. This paralleled the induction by OMPC of CD86, CD80 and CD40 costimulatory molecules on human neonatal and murine B cells and of Th1 cytokines. Neither porins nor lipopolysaccharide fully reproduced the effects of OMPC. These studies indicate that OMPC acts both as carrier and adjuvant, and thereby enhances T cell-dependent antibody responses in human infants.

Resting B (rB) cells are known to be incompetent APCs in vitro, which alone can induce specific unresponsiveness to single minor histocompatibility (miH) Ags and, when combined with CD40 pathway blockade, can induce hyporesponsiveness to MHC molecules in vivo. Here we show that anti-CD40 ligand (CD40L) mAb does not prevent the expression of B7-2 on allogeneic rB cells in vivo but did prolong donor-specific cardiac allograft survival. Moreover, pretreatment with professional APCs combined with anti-CD40L mAb induced hyporesponsiveness to alloantigens in vivo. rB cells from CD40 knockout mice were unable to induce unresponsiveness, while graft prolongation was achieved in CD40L knockout recipients pretreated with wild-type rB cells. These data suggest that CD40-CD40L interactions in the recipient play a critical role in the induction of hyporesponsiveness to alloantigens in vivo and that the effect of the CD40 pathway may be independent of its effect on the B7 costimulatory pathway.

The prompt and vigorous immune response to xenogenic tissue remains a significant barrier to clinical xenotransplantation. Simultaneous blockade of the CD28 and CD40 costimulatory pathways has been shown to dramatically inhibit the immune response to alloantigen.. In this study, we investigated the ability of simultaneous blockade of the CD28 and CD40 pathways to inhibit the immune response to xenoantigen in the rat-to-mouse and pig-to-mouse models.Simultaneous blockade of the CD28 and CD40 pathways produced marked inhibition of the cellular response to xenoantigen in vivo and produced long-term acceptance of xenogeneic cardiac and skin grafts (rat-to-mouse), and markedly suppressed an evoked antibody response to xenoantigen. In addition, this strategy significantly prolonged the survival of pig skin on recipient mice.Long-term hyporesponsiveness to xenoantigen across both a concordant and discordant species barrier, measured by the stringent criterion of skin grafting, can be achieved using a noncytoablative treatment regimen.

In this study, using a murine model of aortic allotransplantation, the role of blockade of signaling through CD28/B7 and CD40/CD40 ligand costimulatory pathways in the evolvement of posttransplant vasculopathy was examined.Aortic allografts were transplanted across C57BL/1OJ (H2b)--C3H (H2k) strain combinations. Transient or more stable blockade of second signaling was achieved by either a single injection or multiple injections of CTLA4-Ig fusion protein (200 microg/dose i.p.) and/or anti-CD40 ligand (CD40L) monoclonal antibody (250 microg i.m.). At day 30 after transplantation, the grafts were harvested for histopathological and immunohistochemical examination.Similar to allografts of untreated animals, aortic allografts obtained from recipients treated with either CTLA4-Ig or anti-CD40L monoclonal antibody alone exhibited marked narrowing of the lumen primarily due to concentric intimal thickening caused by proliferation of alpha-smooth muscle actin-positive cells. Contemporaneous treatment, however, with either a single injection or multiple injections of CTLA4-Ig and anti-CD40L monoclonal antibody resulted in marked diminution of intimal thickening. Interestingly, concurrent prolonged inhibition of CD28/B7 and CD40/CD40L pathways resulted in complete abrogation of the development of posttransplant arteriopathy.These data suggest that a more stable disruption of signaling through costimulatory pathways may be required to obviate the development of posttransplant vasculopathy.

Immunotoxins have shown promise as antitumor agents in clinical trials. However, they have not become part of standard cancer therapy because of factors that include their inherent immunogenicity, which limits the duration of therapy. To address this issue, we evaluated in preclinical models the concomitant use of the immunosuppressive agent CTLA4Ig and BR96 sFv-PE40, a single-chain immunotoxin that binds to carcinoma cells expressing Le(y). Cotreatment with CTLA4Ig, an inhibitor of the CD28/CTLA4-CD80/CD86 costimulation pathway, blocked the production of Abs against BR96 sFv-PE40 in immunocompetent rodents and dogs. It also blocked hypersensitivity reactions in rats carrying colon carcinoma allografts during a second course of BR96 sFv-PE40 therapy, and the cotreatment with CTLA4Ig resulted in enhanced antitumor activity. Cotreatment with CTLA4Ig also prevented hypersensitivity reactions induced by repeat dosing of BR96 sFv-PE40 (q3dx5) in dogs. The production of anti-BR96-sFv-PE40 Abs was decreased in CTLA4Ig-cotreated rodents and dogs resulting in increased plasma levels of BR96 sFv-PE40 relative to non-CTLA4Ig-cotreated animals. These data show that cotreatment of immunotoxins with CTLA4Ig, by inhibiting the production of anti-immunotoxin Abs, can extend the duration of BR96 sFv-PE40 therapy to give greater exposure, reduced toxicities, and increased efficacy.

Blocking the CD28-B7 T cell costimulatory activation pathway protects animals from developing experimental autoimmune encephalomyelitis (EAE). In the mouse EAE model, selective blockade of B7-1 by specific mAbs has been shown to protect animals from EAE. In the Lewis rat model, we have shown that CD28-B7 blockade by systemic administration of CTLA4Ig prevents actively induced EAE. Since CTLA4Ig binds to both B7-1 and B7-2, we used a mutant form of CTLA4Ig (CTLA4IgY100F) that binds only B7-1, to study the role of B7-1 blockade in this model. Such a reagent avoids the potential of signaling by mAbs. Systemic administration of CTLA4IgY100F in several dosing regimens did not protect from EAE, and in some protocols worsened disease, while CTLA4Ig was always protective. In contrast, systemic injection of APCs preincubated ex vivo with the encephalitogenic peptide of myelin basic protein and either CTLA4Ig or CTLA4IgY100F protected recipients from disease. In vitro studies confirmed the in vivo observations and showed that primed lymph node cells from protected animals had decreased proliferative responses to myelin basic protein as compared with controls, while lymphocytes from animals treated with systemic CTLA4gY100F did not. More importantly, systemic administration of CTLA4IgY100F abrogated the protective effect of ex vivo treated APCs. These data suggest an important regulatory role for B7-1, perhaps through binding to CTLA4, in this model of EAE. Understanding the role and mechanisms of selective blockade of costimulatory molecules has implications for therapy of autoimmune disease.

Murine lupus in NZB/NZW F1 (B/W) mice can be retarded by sustained administration of CTLA4Ig and by brief treatment early in life with mAb that block CD40/gp39 interactions. We sought to determine whether brief therapy with CTLA4Ig could provide sustained benefit in B/W mice and whether a synergistic effect could be derived by blockade of both the B7/CD28 and the CD40/gp39 pathways. We found that a short course of CTLA4Ig at the onset of disease produced only short-term benefit. However, when CTLA4Ig was combined with anti-gp39, there was long-lasting inhibition of autoantibody production and renal disease. Ten months after the 2-wk course of therapy, 70% of these mice were alive, compared with only 18% and 0% of those that received only anti-gp39 or CTLA4Ig, respectively. These findings demonstrate that brief simultaneous blockade of the B7/CD28 and CD40/gp39 costimulation pathways can produce benefit that lasts long after treatment has been discontinued.

We tested the effects of blocking CD28-B7 T cell costimulation by using CTLA4Ig in an established transplantation model in which LBNF1 cardiac allografts are rejected in an accelerated manner (36 h) by LEW rats presensitized with Brown-Norway skin grafts. Treatment with CTLA4Ig with or without donor alloantigen in the sensitization phase (between skin and cardiac engraftment) minimally delayed accelerated rejection. However, adjunctive infusion of CTLA4Ig and donor alloantigen in the effector phase (after cardiac engraftment) resulted in long term graft survival and donor-specific tolerance in 30 to 50% of the recipients. The mutant form of CTLA4Ig, which blocks B7-1 but not B7-2, was ineffective. The tolerant state was accompanied by reduction of cell-mediated (MLR/CTL) responses and depression of humoral (circulating IgM/IgG allo-Abs) alloreactivity in vivo. Hence, the binding of CD28 on T cells to both CD80 and CD86 ligands represents a crucial initial costimulatory step leading to accelerated graft rejection. CTLA4Ig-mediated early blockade of the CD28 signaling pathway combined with transfusion of donor cells in the perioperative period interrupts sensitization and may produce transplantation tolerance. This regimen inhibits T cell costimulation and activation to provide help to CD8+ cytotoxic T and B cells, perhaps, via CTLA4Ig-induced clonal anergy or deletion.

Blockade of CD28-mediated T cell costimulatory signals produces effective immunosuppression of a variety of T cell-dependent in vivo immune responses, including autoimmune disorders and transplant rejection. The soluble fusion protein CTLA4Ig, which competitively blocks CD28 ligands B7-1 and B7-2, can prevent allograft and xenograft rejection and in some circumstances induce transplantation tolerance. To determine the relative roles of B7-1 and B7-2 in graft rejection, we have performed islet and cardiac allografts with normal and B7-1(-/-) mice in conjunction with selective blocking reagents. We found that the absence of B7-1 on donor or recipient tissues leads to a slight prolongation of islet allograft survival, but has minimal or no effect on cardiac allograft survival. Allograft function is further prolonged in the islet model when both donor and recipient lack B7-1, although cardiac allograft survival is not prolonged. In the cardiac model, treatment with CTLA4Ig induces long term survival in B7-1(-/-) recipients regardless of donor status. In contrast, anti-B7-2 mAb leads to indefinite allograft survival only when the recipient and donor both lack B7-1, indicating that even in the absence of available B7-2, B7-1 molecules on the donor or recipient cells alone are sufficient to induce graft rejection. These data also indicate that B7-1 and B7-2 are the only CD28 ligands relevant to cardiac allograft rejection in mice.

Blockade of the B7/CD28 costimulation pathway with the fusion protein, CTLA4-Ig, has been shown to prolong allograft survival in numerous rodent models, suggesting that this pathway is functionally important in the allograft rejection response. This pathway is complex and consists of at least the B7-1, B7-1a, B7-1cyt II, and B7-2 molecules on the antigen-presenting cell and CD28 and CTLA4 molecules on the T cell.The intragraft transcript expression of the B7 molecules and their counterreceptors was defined using reverse transcriptase-polymerase chain reaction in the vascularized mouse cardiac allograft model. In addition, the functional significance of these molecules was investigated both in vitro in the mixed leukocyte response (MLR) and in vivo in the vascularized mouse cardiac allograft model.Intragraft expression of B7-1, B7-1a, B7-1cyt II, B7-2, CD28, and CTLA4 transcripts is up-regulated in allografts when compared with both normal untransplanted hearts and syngeneic transplants at between 5 and 12 days after transplant. Both anti-B7-1 and anti-B7-2 monoclonal antibodies alone inhibited T-cell proliferation in the MLR, however, equivalent maximal inhibition was obtained by a combination of these agents or by CTLA4-Ig. Likewise, in the mouse cardiac allograft model, both anti-B7-1 and anti-B7-2 modestly prolonged graft survival. However, an increased survival was obtained with either a combination of anti-B7-1 and anti-B7-2 or CTLA4-Ig. Blockade of the B7/CD28 pathway in the MLR using T cells from CD28 knockout mice had no effect on the proliferative response. Likewise, blockade of the B7/CD28 pathway did not effect the rate of rejection of cardiac allografts by CD28 knockout recipients.These data suggest that both B7-1 and B7-2 have an important role in allograft rejection in the mouse vascularized cardiac allograft model.

We examined the role of the CD28/B7 interaction in regulation of NK cell activity. Cells transfected with B7 enhanced IL-12-induced production of IFN-gamma by IL-2-activated, CD28+ NK cells, but not by resting CD28- NK cells. The ability of B7 transfectants to enhance NK cell production of IFN-gamma was dependent on the intracellular adhesion molecule-1/LFA-1 interaction and could be inhibited by TGF-beta, but not IL-10. Since IL-12-induced production of IFN-gamma by NK cells is associated with resistance to certain infections, we examined whether the CD28/B7 interaction regulated NK cell responses during infection. Infection of SCID mice with Toxoplasma gondii resulted in the appearance of a population of CD28+ NK cells, NK cell production of IFN-gamma, and increased NK cell cytolytic activity. Administration of CTLA4-Ig to SCID mice infected with T. gondii inhibited these latter two effects and resulted in a significant increase in parasite burden. The stimulus for CD28 expression by NK cells in SCID mice infected with T. gondii appeared to be independent of IL-2. However, mRNA for IL-15, a cytokine with properties similar to those of IL-2, was detected in tissues of SCID mice infected with T. gondii. In vitro experiments demonstrated that IL-15 could stimulate resting NK cells to express functionally active CD28 as well as enhance the production of IFN-gamma by SCID splenocytes stimulated with T. gondii. Together our data demonstrate that the interaction of CD28+ NK cells with B7 regulates NK cell production of IFN-gamma associated with resistance to infection and that IL-15 may be involved in these events.

We have previously demonstrated that the development of allergen-induced airway hyperresponsiveness in a murine model is CD4+ T cell dependent. In the present study, we examined the role of the B7/CD28-CTLA4 costimulatory T cell activation pathway in the pathogenesis of allergen-induced airway hyperresponsiveness in this murine model. Sensitized A/J mice develop significant increases in airway responsiveness, bronchoalveolar lavage eosinophils, serum IgE levels, and Th2-associated cytokine production following aspiration challenge with OVA. Administration of CTLA4-Ig either before Ag sensitization or before pulmonary Ag challenge abolished Ag-induced airway hyperresponsiveness and pulmonary eosinophilia. Examination of cytokine protein levels in the bronchoalveolar lavage showed a significant decrease in the level of the Th2 cytokine, IL-4, after CTLA4-Ig treatment either before sensitization or before challenge, with no significant change in the concentration of the Th1 cytokine, IFN-gamma. Further, the Ag-specific Ab isotypes IgG1 and IgE were significantly decreased in animals treated with CTLA4-Ig before challenge, while there was no significant change in the IgG2a Ab isotype. These data demonstrate that administration of CTLA4-Ig is effective in ablating allergen-induced airway dysfunction concomitant with a significant reduction in the Th2 response. We conclude that B7/CD28-CTLA-4 costimulation is required for the development of many of the immunologic and physiologic features of asthma, possibly by promoting a pathologic type 2-associated response.

Blocking the CD28-B7 T lymphocyte costimulatory pathway with the recombinant protein CTLA4Ig induces long term allograft survival in rodents. It has been suggested that this results from selective activation of the Th2 immune pathway. To test this hypothesis, we compared vascularized cardiac allograft survival in wild-type (IL-4 +/+) and homozygous IL-4 gene-knockout (IL-4 -/-) mice. We report in this study that long term survival (100 days) of fully allogeneic grafts can be induced readily in IL-4 -/- recipients treated with a short course of CTLA4Ig. We also demonstrate that IL-4 -/- mice are deficient in Th2-type cytokine expression following in vitro or in vivo allostimulation. These results suggest that IL-4 production and subsequent generation of a Th2-type immune response are not obligatory for CTLA4Ig-induced long term acceptance of vascularized allografts.

During the differentiation of thymocytes to mature T cells the processes of positive and negative selection result in signals that either protect thymocytes from cell death, or delete, through apoptosis, thymocytes with self-reactive T cell receptors (TCR). Glucocorticoids have been shown to induce thymocyte apoptosis and are produced within the thymic microenvironment. Furthermore, steroid-induced apoptosis of thymocytes has been suggested as a potential mechanism for removal of nonselected thymocytes. In this report, we demonstrate that thymocytes can be rescued from glucocorticoid-induced apoptosis by incubation with cells that express high levels of B7-1 or B7-2. In addition, the ability to be rescued by B7-1 and/or B7-2 can precede expression of the TCR. We demonstrate that CD3(+)-depleted or CD3+/ TCR-beta(+)-doubly depleted thymocytes can be rescued from glucocorticoid-induced apoptosis through the interaction of CD28 or CTLA-4 on thymocytes with cells bearing high levels of B7-1 or B7-2. Furthermore, these transfected cells are major histocompatibility complex (MHC) class II negative and, while they may express MHC class I, there is no preferential rescue of CD8+ thymocytes in the presence of glucocorticoids. Together, these data suggest that the rescue of thymocytes from glucocorticoids can be independent of the TCR. We also demonstrate that, in addition to CD28, CTLA-4 is expressed on thymocytes, suggesting that rescue from glucocorticoid-induced cell death can be mediated by both CD28 and CTLA-4. A CTLA-4Ig fusion protein which binds to both B7-1 and B7-2 was shown to completely block the rescue of thymocytes from glucocorticoid-induced cell death. Therefore, we conclude that interactions between B7-1/B7-2 and CD28/CTLA-4 are sufficient and necessary for rescue of thymocytes from glucocorticoid-induced cell death.

B7 molecules (CD80 (B7-1) and CD86 (B7-2/B70)) on APCs provide costimulatory signals for T cell proliferation. We previously described the presence of an alternatively spliced form of murine CD80 (previously termed MB7-2 and renamed as B7-1a) that completely lacks the second Ig-like domain coded by exon 3 in activated murine B cells. in this study, we first examined whether B7-1a mRNA can be detected in vivo by RNase protection assay. The expression of B7-1a mRNA was only detected in lymphoid organs although the level of expression was lesser than that of CD80 mRNA. However, we demonstrated that the expression of B7-1a mRNA like CD80 mRNA was considerably augmented in spleen cells treated with either LPS in vitro or OVA/CFA conjugate in vivo. We next determined the functional activity of B7-1a using Chinese hamster ovary (CHO) cells transfected by B7 genes. When resting T cells were cocultured with CHO cells expressing B7-1a molecules in the presence of PMA/ionomycin, T cell proliferation was not detected, while CHO cells either expressing CD80 or CD86 could promote the proliferation of resting T cells. in contrast to resting T cells, CHO cells expressing B7-1a could support the proliferation of activated T cells. Thus, costimulatory activity of B7-1a molecules was dependent upon the activation stage of T cells. Therefore the IgV-like region of CD80 contains a critical region for functional interaction with its ligands and can transduce a costimulatory signal for T cell proliferation.

Allograft rejection is dependent on T cell activation, which requires both the engagement of the T cell receptor by antigen in the context of the MHC molecules and costimulatory signals delivered by cell surface molecules such as B7-CD28/CTLA4 pathway. CTLA4-Ig is a fusion protein that blocks this pathway and has previously been shown to prolong both allograft and xenograft survival. The current study demonstrates markedly prolonged murine cardiac allograft survival and specific prolongation of secondary skin grafts using a combination of CTLA4-Ig plus donor bone marrow. A role for hematopoietic chimerism in the establishment of CTLA4-Ig-induced transplantation tolerance was investigated using reverse transcriptase polymerase chain reaction analysis of recipient tissues. Expression of donor-specific MHC class II transcripts in both peripheral and lymphoid tissues was demonstrated at greater than 200 days after transplant. To investigate the functional significance of this observation, heart donors, and donor bone marrow were irradiated before transplantation in CTLA4-Ig-treated recipients. A reduction in allograft survival was associated with irradiation of both the donor heart and the bone marrow. These results suggest that there may be a donor-derived radiosensitive element that enhances allograft survival in this model. Reverse transcriptase polymerase chain reaction analysis of allografts of tolerant and control animals at days 5, 8, and 12 after transplantation failed to demonstrate a dramatic difference in the expression of interleukin (IL)-2, IL-4, IL-10, and interferon-gamma message. Cytotoxicity effector transcripts were largely intact in CTLA4-Ig + bone marrow-treated recipients as they showed no decrease in intragraft granzyme, perforin, Fas, or Fas ligand transcripts during thr first 8 days after transplant. These results imply that complex mechanisms may be important for the induction and maintenance of transplantation tolerance in the CTLA4-Ig plus bone marrow murine cardiac allograft model.

In the BXSB autoimmune disease-prone mouse strain, male mice develop severe lupus-like symptoms and die early in life (4-6 mo), whereas females do not. We have previously demonstrated that profound phenotypic and functional changes occur with age in CD4+ cells from BXSB males. CD4+ cells from males (4 mo old) were predominantly CD44high, CD45RBlow, and MEL-14low (activated/memory phenotypes), while the reciprocal phenotypes characteristic of naive cells were prevalent in age-matched females and young adult males (2 mo old). CD4+ cells from older males proliferated less and produced less IL-2 and IFN-gamma than cells from either females or young males in response to immobilized anti-CD3 mAb. We tested the effect of CTLA4Ig treatment on the progression of disease in BXSB males. CD4+ cells from CTLA4Ig-treated mice at 4 mo of age were predominantly CD44low, CD45RBhigh, and MEL-14high phenotypes that were identical with those observed in CD4+ cells from young (3-mo-old) females. In contrast, control male mice treated with IgG2a accumulated the CD4+ memory phenotype. CD4+ cells from 4-mo-old male CTLA4Ig-treated mice proliferated and produced IL-2 at levels similar to those of cells from females in response to immobilized anti-CD3 mAb. Furthermore, in contrast to IgG2a-treated mice, female and CTLA4Ig-treated male mice at 4 mo of age produced no anti-chromatin Abs. Three of four male mice injected with CTLA4Ig until 6 mo of age appeared healthy at 8 mo of age, whereas all five of IgG2a-treated control males died by 6 mo of age. These 8-mo-old CTLA4Ig-treated males showed variable resistance to autoimmunity as well as function and phenotype marker expression, and a less striking glomerulonephritis than 4-mo-old untreated males. The results of this study demonstrate that the rampant T cell activation and T cell dysfunction that occur in male BXSB mice by 4 mo of age are abrogated by blocking the CTLA4-dependent costimulatory signal(s). They also show that treatment with CTLA4Ig can suppress the pathogenesis of disease and increase longevity.

Engagement of the TCR/CD3 complex together with ligation of CD28 by its counterstructures B7-1 (CD80) and B7-2 (CD86) on APC are required for mitogenic T cell activation. After activation, T cells not only express B7-1 and B7-2 molecules, but a second receptor for the B7 ligands, CTLA-4, can be found on their surfaces. We here show that B cells can be induced to express CTLA-4 on the plasma membrane. Similar to what has been reported for T cells, CTLA-4 expression on B cells was transient. Purified B cells did not express CTLA-4 when mitogenically activated with alpha IgM and CD40 Ab, but did express the molecule when cultured in the presence of membranes from activated T cells, which suggests that induction of CTLA-4 expression on B cells was dependent on direct cell-cell contact of B lymphocytes and activated T cells. CTLA-4 molecules isolated from either T or B cells were biochemically indistinguishable. Moreover, because the ability of chimeric B7-1/Ig proteins to bind to activated B cells was correlated with CTLA-4 expression levels on these cells, we conclude that B cell-expressed CTLA-4 has ligand binding capacity. These data suggest that costimulatory receptors and their specific ligands not only play a role in T cell stimulation, but contribute in a direct fashion to the regulation of B cell responses.

We have examined the ability of CTLA4Ig to induce long-lasting Ag-specific tolerance to a T-dependent Ag. Treatment of mice with murine CTLA4Ig following immunization with SRBC induced immune unresponsiveness to SRBC that was only reversible upon repeated Ag exposure. This unresponsiveness was Ag specific, lasted90 days, was observed in thymectomized mice, and was maintained even during challenge with another strong T-dependent Ag. A third challenge with SRBC, however, led to an Ab response. Splenic T cells and B cells from unresponsive mice were functional upon transfer to irradiated hosts, indicating that they had been neither depleted nor rendered permanently "tolerant" by CTLA4Ig. The mechanism of immunosuppression by CTLA4Ig was investigated by measuring cytokine transcripts in spleens of immunized mice. CTLA4Ig treatment following primary immunization blocked the induction of IL-2 transcripts in splenic T cells and IL-4 transcripts in both T cells and non-B, non-T cells. Splenocytes from CTLA4Ig-treated, SRBC-unresponsive mice showed altered induction of IL-2 and IL-4 transcripts, but T cells nonetheless became primed for IL-4 mRNA synthesis despite the lack of a measurable Ab response. Anti-IL-4 mAb and human CTLA4Ig were synergistic in their ability to induce unresponsiveness, indicating that incomplete suppression of IL-4 production by CTLA4Ig limited its effectiveness.

Signals initiated through both the TCR complex and CD28 are required for optimal activation of T lymphocytes. Recently, it has been demonstrated that CD28 interacts with two different ligands, designated CD80 (B7/B7-1) and CD86 (B70/B7-2). We have produced stable transfectants that express CD80, CD86, or both ligands and have examined their ability to costimulate T cell proliferation, cytokine production, and the generation of CTL. When we used small, resting human peripheral blood T cells as responders, both CD80 and CD86 transfectants efficiently costimulated anti-CD3 mAb-induced proliferation and the secretion of IL-2 and IFN-gamma. Additionally, both CD80 and CD86 transfectants were able to generate functional CTL. The magnitude and kinetics of these responses were similar, which indicates that both ligands provide efficient costimulatory signals. Because many APCs coexpress both CD80 and CD86, we compared the ability of anti-CD80 and anti-CD86 mAbs to inhibit allogeneic MLR stimulated with B lymphoblastoid cell lines and showed that it is necessary to inhibit interactions with both ligands to optimally block CD28-dependent proliferation. Given the limited homology of CD80 and CD86, it was surprising that the binding of CD28-Ig fusion protein to CD80 and that to CD86 transfectants were essentially indistinguishable. Binding of CTLA-4-Ig fusion protein to both transfectants also was quite similar, but was of higher affinity than CD28-Ig binding. Results from these studies indicate that both CD80 and CD86 are potent and similar costimulators of T lymphocytes. Therefore, the role of CD80 and CD86 in an immune response may be determined primarily by their differential expression on APC.

T cell activation is the result of antigen-specific interactions with the TCR/CD3 complex and costimulation via other T cell surface receptors. Prevention of costimulation can result in clonal anergy. CTLA4Ig is a fusion protein that binds with high-affinity to the B7/BB1 ligand and blocks the interaction of this ligand with CD28 and CTLA4. We explored the immunosuppressive effects of CTLA4Ig in a murine nonvascularized heterotopic cardiac transplant model and in a model of cell mediated immunity. CTLA4Ig administered in vivo for two days at the time of transplantation resulted in significant prolongation of allograft survival (55 +/- 2.0 vs. 12.2 +/- 0.5 days for control, P0.03). Administration at later times or to previously primed animals produced no prolongation of graft survival. CTLA4Ig administered during in vivo immunization to the hapten TNP suppressed the contact sensitivity response and inhibited the subsequent in vitro generation of secondary TNP-specific CTL. CTLA4Ig administered in vivo had no effect on subsequent primary alloantigen-specific CTL or MLR responses--however, when added to culture the fusion protein inhibited the MLR response by 80%, but not the alloantigen-specific CTL response. CTLA4Ig inhibited CD4+ and CD8+ proliferative and cytokine responses to alloantigen. Flow cytometry showed no changes in distribution of subpopulations of T cells. These results confirm the immunosuppressive activity of CTLA4Ig in vivo in an allograft model and show that both CD4+ and CD8+ T cells are suppressed by CTLA4Ig. The most efficacious time of administration is during priming of the immune response at the time of antigen presentation.

Infection of inbred strains of mice with Leishmania major is a well-characterized model for analysis of the development of effector CD4+ subsets of the Th1 and Th2 types in vivo. We co-administered a fusion protein, CTLA4Ig, that blocks the CD28-B7 costimulatory pathway important for optimal T cell activation, to assess the relative role for this pathway during maturation of Th1 and Th2 cells in vivo. Surprisingly, CTLA4Ig administered within the first week of infection completely abrogated progressive disease in susceptible BALB/c mice while having no effect on the protective immune response developed by resistant C57BL/6 mice. The protective effect in BALB/c mice was increasingly lost if administration of CTLA4Ig was delayed longer than 1 wk after infection. As in other protective interventions used in this model, control of infection was associated with down-regulation of IL-4 mRNA transcripts in lymph node cells recovered 5 wk after infection together with abrogation of IgE production and enhanced parasite-specific IgG2a relative to IgG1. Although a single dose of CTLA4Ig was protective, sustained delivery abolished the capacity of BALB/c mice to contain infection, suggesting that costimulation through this pathway is required at later stages of the immune response. Taken together, the data demonstrate that the priming of Th2 cells is more dependent upon the CD28-B7 pathway than the priming of Th1 cells, and suggest that the development of Th subsets in vivo may be influenced by limiting CD28-B7 costimulation.

Activation of Ag-specific T cells often requires costimulatory signals in addition to the primary signal mediated through the TCR. The CD28-B7 interaction provides one important costimulatory signal. Previous studies have shown that a soluble CD28 homologue, CTLA4lg, binds B7 with high affinity and can inhibit CD28-B7-mediated costimulation in vitro and in vivo. In this study we examined the ability of soluble human CTLA4lg to inhibit autoantibody production in vivo. For this purpose we used a novel transgenic (Tg) model of autoantibody production. Hepatitis B eAg-expressing Tg mice can be induced to produce autoantibody to the circulating autoantigen (HBeAg) by the injection of a T cell recognition site that fails to elicit T cell tolerance in these mice. Autoimmunity in this model is quantitative because serum autoantibody and autoantigen concentration are inversely correlated and easily measurable by ELISA. In this system a single regimen of CTLA4lg treatment significantly suppressed primary autoantibody production and variably led to long term unresponsiveness. Furthermore, in vivo treatment with CTLA4lg inhibited both T cell activation and T cell-B cell interactions.

The rejection of the transplanted allograft is dependent on T cell activation, which requires T cell receptor engagement by antigen and costimulatory signals delivered by T cell surface molecules such as CD28. CTLA4-Ig is a fusion protein that has previously been shown to block the CD28-mediated costimulatory signal and inhibit immune responses in vitro and in vivo. In this report we show that treatment of the C3H/He recipient of a BALB/c vascularized cardiac allograft with a 12-day course of CTLA4-Ig produced indefinite graft survival (100 days) in the majority of recipients. In addition, these recipients demonstrated donor-specific transplantation tolerance when tested with donor-specific (BALB/c) and third-party (C57BL/10) skin grafts. These results demonstrate that CTLA4-Ig can induce transplantation tolerance in the adult murine cardiac allograft model.

Dendritic cells (DC) play a critical role in the initiation of T cell-mediated immune responses, and express costimulatory molecules that are required for optimal activation of unprimed T cells. Studies on the regulation of the costimulatory molecules on DC have produced evidence from several systems that GM-CSF can up-regulate expression of CTLA4 counter receptor (CTLA4-CR) (but not intercellular adhesion molecule 1 (ICAM-1) and heat stable Ag (HsAg)) on DC. This is demonstrated on splenic DC, Langerhans cells, kidney DC in culture, and in a skin-explant culture system, in which the increased expression of CTLA4-CR on Langerhans cells (LC) occurs concomitantly with their migration out of skin. Interestingly, despite the ability of both GM-CSF and IFN-gamma to increase CTLA4-CR and maintain similar levels of ICAM-1, HsAg, and MHC molecule expression, the functional consequences of these cytokines on splenic DC are distinctly different. GM-CSF enhances the ability of DC to stimulate both T cell proliferation and cytokine release, whereas IFN-gamma causes no increase in immunostimulatory function. Further analysis of the CTLA4-CR on these cell populations by using the GL-1 and IG10 mAbs has shown that GM-CSF-cultured DC express high levels of both B7-1 and B7-2, whereas IFN-gamma-cultured DC express increased levels of only B7-2. These results suggest that optimal stimulation of unprimed T cells to proliferate and release cytokines may require participation of both of these CTLA4 counter receptors, and confirm the importance of GM-CSF for the maturation of DC into potent stimulators of T cell activation.

The role of costimulation in the activation of TCR-gamma delta cells in normal mice and mice transgenic (tg) for a TCR-gamma delta receptor was investigated. Activation of TCR-gamma delta cells required two signals. One signal was mediated by TCR occupancy, whereas a second signal was provided by accessory cells. The importance of the CD28/B7 interaction in the delivery of the second signal was demonstrated in multiple ways. First, addition of a soluble fusion protein homolog of CD28, CTLA4Ig, significantly inhibited the activation of G8 tg splenic TCR-gamma delta lymphocytes and intestinal epithelial TCR-gamma delta lymphocytes by Ag-bearing lymphocytes during primary stimulation. Similarly, both proliferation and IFN-gamma production were inhibited by addition of CTLA4Ig to secondary antigenic stimulation of G8 tg TCR-gamma delta cells. Second, an Ag-bearing thymoma, EL-4, was only able to stimulate expanded G8 tg TCR-gamma delta cells when the thymoma expressed B7. This stimulation was blocked by both CTLA4Ig and anti-B7 antibody. Third, antibodies to CD28 were able to mimic the costimulatory affect of APC. TCR-gamma delta cells cultured with either Ag-bearing fixed stimulator cells or submitogenic concentrations of immobilized anti-pan TCR-gamma delta mAb proliferated only in the presence of anti-CD28 mAb. Finally, G8 tg cells produced IL-2 only in the presence of APC costimulation or anti-CD28 antibodies, and the addition of exogenous rIL-2 overcame the need for costimulation. Thus, autocrine IL-2 production is one of the major consequences of TCR-gamma delta cell costimulation. Together these data demonstrate that costimulation is necessary for the activation of TCR-gamma delta cells and can occur through CD28 interaction.

We have previously demonstrated that the inhibitory effects of IL-10 on ConA-induced T cell proliferation or IL-2 production by resting murine T cells were only observed when macrophages, but not when activated B cells, dendritic cells, or L cells, were used as accessory cells. To further elucidate the mechanism of action of IL-10 on the inhibition of macrophage costimulatory activity, we have used a system in which macrophages can develop into effective costimulator cells and the effect of IL-10 on this process can be studied in the absence of T cells. After fixation, resting macrophages have no costimulatory activity for soluble anti-CD3-induced T cell proliferation nor do they express the activation Ag B7/BB1. In contrast, macrophages activated by culture alone, or by culture with IFN gamma or LPS for 24 h, and then fixed, were effective accessory cells, expressed B7, and their costimulatory activity correlated with their level of cell surface B7 expression. Addition of IL-10 during the process of macrophage activation resulted in both a marked reduction in costimulatory activity and in B7 expression. IL-4 and transforming growth factor-beta that suppress many macrophage functions did not inhibit the induction of B7 expression. The inhibitory effect of IL-10 on the up-regulation of B7 was selective because the up-regulation of intercellular adhesion molecule-1 and MHC class II Ag was not affected. Direct evidence that the lack of B7 is the relevant limiting defect for IL-10-treated macrophage accessory cell function was obtained from studies in which the costimulatory capacity of IL-10-treated macrophages could be completely restored by the addition of B7 transfected, but not nontransfected, L cells to the assays.

Costimulatory molecules on the APC regulate T cell growth by providing signals that regulate responses to TCR occupancy. One such molecule is B7/BB-1, which triggers a T cell activation pathway by binding the CD28 and/or CTLA-4 cell-surface molecules. Expression and signaling activity of CD28 have been shown to increase after T cell activation by various polyclonal activators. Here we show that CD28 expression and signaling activity in activated T cells decrease after ligand binding to CD28. Stimulation of CD28 on PHA- or PMA-activated T cells by cross-linked mAb 9.3 or by co-culture with B7+ Chinese hamster ovary (CHO) cells caused a marked reduction of CD28 mRNA levels within 4 h. The decrease in CD28 mRNA was transient, and by 24 h of CD28 stimulation, CD28 mRNA was found at approximately initial levels. In contrast, CTLA-4 mRNA levels were usually up-regulated by CD28 triggering. Cell-surface expression of CD28, but not CD2 or CD3, decreased by 12 to 24 h after addition of B7+ CHO cells, but returned to initial levels or higher by 48 h. The ability of CD28 cross-linking on PMA-activated CD4+ cells to trigger calcium mobilization was also reduced by treatment with B7+ CHO cells, and remained reduced even after cell-surface expression of CD28 returned to normal levels. Thus, engagement of the CD28 receptor by its natural ligand B7/BB-1 leads to a transient down-regulation of CD28 synthesis and a prolonged unresponsiveness to CD28 signaling. This represents a novel mechanism for regulation of costimulatory signals delivered by interactions of CD28 with the B7/BB-1 counter receptor.

Staphylococcal enterotoxins, also known as superantigens (SAg), bind class II MHC molecules on APC and upon direct cell-to-cell contact stimulate proliferation of T cells expressing appropriate V beta gene products. The T cell surface molecule CD28 binds its costimulatory counter-receptor, B7 expressed on APC, and augments IL-2 production and T cell growth. Although the role of B7 costimulation during Ag-specific responses of T cells is established, its involvement during the activation of T cells with SAg has not been examined. Using a soluble Ig C gamma 1 chimera of CTLA-4, a second receptor for B7 and a homologue of CD28, this study examines the role of B7 expressed on APC during the induction of proliferation of CD4+ T cells upon stimulation with SAg (SAg/staphylococcal enterotoxins). CTLA-4lg, which has a higher avidity for B7 than CD28, had no effect on the synthesis of IL-2 as well as proliferative responses of CD4+ T cells induced by SAg presented on allogeneic EBV-transformed B cells, and IFN-gamma-activated endothelial cells. In contrast, T cell proliferation induced by alloAg presentation by the same APC was significantly inhibited by CTLA-4lg. mAb directed at the CD11a/CD18 molecule inhibited both SAg-induced and alloAg-induced proliferation of T cells. AlloAg-primed CD4+ T cells, which expressed both class II MHC and intercellular adhesion molecule-1 but not B7, presented SAg to and induced proliferation of both resting and SAg-primed T cells. These responses were inhibited by mAb directed at CD11a/CD18 but not by CTLA-4 Rg. These results suggest that SAg-induced responses differ from those induced by alloAg in that they are not obligatorily dependent on the costimulation by B7. In contrast, adhesive interaction between CD11a/CD18 on T cells and its counter-receptor on SAg-presenting cells is necessary and probably sufficient to support SAg-induced proliferation of T cells.

In addition to TCR-derived signals, costimulatory signals derived from stimulation of the CD28 molecule by its natural ligand, B7, have been shown to be required for CD4+8- T cell activation. We investigate the ability of B7 to provide costimulatory signals necessary to drive proliferation and differentiation of virgin CD4-8+ T-cells that express a transgenic TCR specific for the male (H-Y) Ag presented by H-2Db class I MHC molecules. Virgin male-specific CD4-8+ T cells can be activated either with B7 transfected chinese hamster ovary (CHO) cells and T3.70, a mAb specific for the transgenic TCR-alpha chain that is associated with male-reactivity, or by male dendritic cells (DC). Activated CD4-8+ T cells proliferated in the absence of exogenously added IL-2. IL-2 activity was detected in supernatants of CD4-8+T3.70+ cells that were stimulated with T3.70 and B7+CHO cells. The response of CD4-8+T3.70+ cells to T3.70/B7+CHO or to male DC stimulation were inhibited by CTLA4Ig, a fusion protein comprising the extracellular portion of CTLA4 and human IgG C gamma 1. It has been previously shown that CTLA4Ig binds B7 with high affinity. Staining with CTLA4Ig revealed that DC express about 50 times more B7 than CD4-8+ T cells. CTLA4Ig also specifically blocked the proliferation of male-reactive cells in vivo. We have also used an in vitro deletion assay whereby immature CD4+8+ thymocytes expressing the transgenic male-specific TCR are deleted by overnight incubation with either immobilized T3.70 or male DC to investigate the participation of the CD28/B7 pathway in the negative selection of immature thymocytes. Staining with B7Ig established that both immature murine CD4+8+ and mature CD4-8+ thymocytes express a high level of CD28. However, despite the high expression of CD28 on CD4+8+ thymocytes, it was found that deletion of CD4+8+ thymocytes expressing the male-specific TCR by the T3.70 mAb was not inhibited by B7+CHO cells. Furthermore, the deletion of these thymocytes by DC also was not inhibited by CTLA4Ig. These findings provide evidence that although signaling through CD28 can costimulate a primary anti-male response in mature CD4-8+ T cells, the CD28/B7 pathway does not appear to participate in the negative selection of immature CD4+8+ thymocytes.

The B cell activation antigen B7/BB1 has been shown to co-stimulate growth of human T cells by binding the T cell molecule CD28. In mice, the heat-stable antigen (HSA) has also been shown to act as a co-stimulator for T cell growth. In this study, we have evaluated the contributions of B7 and HSA to the co-stimulatory activity of antigen-presenting cells (APC). Mouse B7 provides co-stimulatory activity for murine CD4 T cells in anti-CD3-induced proliferation. Human CTLA4Ig, a chimeric molecule comprising the extracellular region of CTLA-4 fused to an immunoglobulin C gamma fragment, binds to murine B7. We, therefore, use human CTLA4Ig and the hamster anti-HSA monoclonal antibody 20C9 to analyze the relative contributions of B7 and HSA to the co-stimulatory activity of murine spleen APC. Our data reveal that both murine B7 and HSA are expressed by dendritic cells and by low-density spleen B cells. Either CTLA4Ig alone or anti-HSA alone inhibited CD4 T cell proliferation to anti-CD3 by > 90%, while CTLA4Ig and anti-HSA together were far more efficient in inhibiting clonal expansion of CD4 T cells. These results demonstrate that functionally defined co-stimulation involves at least B7 and HSA and suggest that signals delivered by B7 and HSA synergize in promoting T cell growth.

Whereas dendritic cells (DC) are known to be potent activators of T cells both in vitro and in vivo, the critical costimulatory molecules expressed on DC are not well characterized. Using immunocytochemical and molecular techniques we find that splenic DC express B7/BB1, the counter-receptor for CD28. Moreover, expression of B7/BB1 is upregulated on epidermal Langerhans cells (LC) during their functional maturation into potent T cell stimulators. In blocking experiments, we find that participation of B7/BB1 is required for optimal proliferation of unprimed, allogeneic T cells in DC-driven, primary mixed leukocyte reactions. These data demonstrate that the regulated expression of B7/BB1 on DC may be important in the initiation of a primary T cell response.

The frequency of phenotypic expression of the herpes simplex virus type 1 tk and Escherichia coli gpt genes was compared with the frequency of genotypic transformation after calcium phosphate-mediated DNA transfection of a number of tk- and hprt- cell lines. In three of the five lines tested, the frequency of phenotypic expression was at most 10-fold higher than that of genotypic transformation as indicated by frequency of HAT resistance. The remaining two lines showed phenotypic responses which were 50- to 100-fold greater than the genotypic responses. The data indicate that the efficiency of DNA-mediated transformation with some cell lines can be limited by events after the uptake and expression of transfected DNA.

Oncostatin M is a polypeptide growth regulator produced by activated T cells and phorbol ester-treated U937 cells. To identify specific cellular receptors for this factor, we have characterized the binding of 125I-labeled oncostatin M to a variety of normal and malignant mammalian cells. Recombinant oncostatin M was labeled with 125I with full retention of growth inhibitory activity on A375 melanoma cells. 125I-Oncostatin M bound to sensitive cells in a time- and temperature-dependent fashion. Binding was specifically inhibited by unlabeled native or recombinant oncostatin M, but not by other polypeptide growth factors tested. Binding to human leukemic and normal blood cells was generally less than to nonhematopoietic cells. With four different cell lines, maximal growth inhibition by oncostatin M was achieved at less than maximal binding site occupancy. Scatchard graphs of direct binding data were curvilinear and indicated that 125I-oncostatin M bound with higher apparent affinity at lower 125I-oncostatin M concentrations. Using a two binding site model, affinity constants of Kd1 = 11 +/- 11 pM and Kd2 = 1000 +/- 380 pM were extrapolated from binding data with A375 cells, and values of Kd1 = 3 +/- 2 pM and Kd2 = 400 +/- 44 pM from A549 cells. The major 125I-oncostatin M binding species in a number of mammalian cell lines was identified by chemical cross-linking as a specific protein(s) of Mr = 150,000-160,000. 125I-Oncostatin M was internalized (t1/2 = 30 min) and degraded subsequent to binding to a responsive cell line.(ABSTRACT TRUNCATED AT 250 WORDS)

Epidermal growth factor (EGF) initiates a wide variety of events when added to responsive cultured cells. These range from early events requiring only brief exposure to EGF, e.g., stimulation of transport of amino acids or ions, to later events such as commitment of cells to a round of DNA synthesis, a process requiring 6 h or more of continuous exposure to hormone. EGF binding is followed first by phosphorylation of EGF receptors, which can be detected in purified membranes and permeabilized cells, and then by internalization and proteolytic processing of receptors in lysosomes. Native 160,000-dalton EGF receptors contain a site that is not exposed on the cell surface and is highly sensitive to cleavage by an endogenous protease, which yields a 145,000-dalton receptor fragment that retains phosphate acceptor activity. Cleavage of receptor at a trypsin-sensitive site, also not exposed to the cell surface, yields a 115,000-dalton fragment that binds EGF, but contains no phosphorylated species. The data indicate that the phosphate acceptor sites on EGF receptors are localized on a 45,000-dalton cytosolic region.