An agent or agents produced by virus-transformed cells cause unregulated ruffling in untransformed cells.

KNRK cells (a normal rat kidney [NRK] cell line transformed by Kirsten murine sarcoma virus) in sparse culture exhibit a highly ruffled morphology, but the cause of this ruffling is unknown. In this study, we have demonstrated that the continuous, excess ruffling on KNRK cells is caused by one or more soluble agents secreted by the KNRK cells themselves. To do this study, an assay for ruffling responses in live cell cultures was defined, and its reproducibility was demonstrated. This assay permitted observation of the kinetics of ruffling responses (percentage of cells ruffled as a function of time after stimulation). This method was used to compare the kinetics of ruffling induced by insulin, epidermal growth factor, fibroblast growth factor, glucose, and KNRK cell conditioned medium (CM). Ruffling was elicited on NRK cells by each of the polypeptide mitogens and nutrients, but, in each case, this ruffling subsided spontaneously within an hour. CM from KNRK cells also caused ruffling movements on untransformed NRK cells, but this ruffling continued for at least 20 h. This response was largely blocked by premixing the KNRK cell CM with rabbit IgG against rat transforming growth factor, type alpha, (TGF-alpha). KNRK cells made quiescent (ruffle free) by a pH shift (from 7.4 to 8.4) responded to insulin, glucose, and KNRK cell CM with kinetics similar to those observed for each of these factors in NRK cells. The unusual feature for the ruffle-inducing agent(s) produced by KNRK cells was that this activity was not subject, in either NRK or KNRK cells, to the cellular off-regulation that limits the responses to insulin or glucose. Thus, the continuous ruffling of KNRK cells is caused by their own unregulated ruffle-inducing agent or agents, which appear to include TGF-alpha. This work also demonstrates that kinetic analysis of cellular responses to exogenous factors can provide new insights into the regulatory mechanisms involved in the normal limitation of these responses.

Cell-mediated co-action of transforming growth factors: incubation of type beta with normal rat kidney cells produces a soluble activity that prolongs the ruffling response to type alpha.

Intense, continuous ruffling is a characteristic of many transformed cells, but untransformed cells ruffle intensely only briefly after exposure to growth factors. We reported previously that cells of a normal rat kidney (NRK) cell line transformed by Kirsten murine sarcoma virus secrete their own ruffle-inducing agent(s) that cause sustained ruffling in either themselves or untransformed NRK cells. In the present study, we examined the roles of the transforming growth factors TGF-alpha and TGF-beta in the induction and maintenance of ruffling in untransformed NRK cells and observed the following: TGF-alpha caused a transient epidermal growth factor (EGF)-like response, which could be blocked by prior exposure of cells to EGF or by antiserum directed against the COOH-terminus of TGF-alpha. TGF-beta caused no ruffling and did not itself prolong TGF-alpha ruffling. A new, buffer-soluble (transferable) mediator activity produced by incubation of TGF-beta with NRK cells for 6-h extended the duration of maximal TGF-alpha-induced ruffling by several-fold. This study demonstrates that TGF-alpha alone causes an EGF-like, transient ruffling response, but neither TGF-alpha or TGF-beta alone, nor the two together, cause transformation-associated sustained ruffling. Rather, TGF-alpha acts in concert with a new, TGF-beta-dependent activity. This new activity appears to inhibit normal cellular off-regulation of TGF-alpha-induced ruffling. Inhibition of the cellular off-regulation of a growth factor response could play a key role in the unregulated growth associated with malignancy.

Modulation by normal serum factors of Kirsten murine sarcoma virus-induced transformation in adult rat cells infected in early passage.

Adult rat adrenal cells, infected with Kirsten murine sarcoma virus in early passage, transform consistently in 10% fetal bovine serum (FBS)-supplemented medium. Substitution of 3% horse serum (HS) for FBS reverses early foci and delays transformation. The influence of the serum on DNA synthesis, anchorage dependence, tumorigenicity, and subcellular Mr 21,000 transforming protein (p21) distribution was followed from infection in passage 1 to complete transformation. In FBS, increased expression of p21 preceded other evidence of transformation. Subsequently, p21-positive cells transformed morphologically, but initially their growth parallelled that of coexisting untransformed cells. Foci formed at passages 5 to 10, and the cells became anchorage independent and tumorigenic at passages 10 to 20. As transformation in FBS progressed, p21 relocated from a diffuse distribution to sites of retraction from substrata and then to ruffles and lamellae on cellular processes. Early in transformation, HS-medium reduced proliferation of morphologically normal and morphologically transformed p21-positive cells. This effect was counteracted by the addition of FBS or the Mr 50,000 to 100,000 fraction of FBS. Fully transformed, tumorigenic cells grew rapidly in both sera but, if transferred from FBS to HS, became more anchorage and density dependent, and p21 relocated from cell processes to the cell bodies. In immortal lines, the substitution of HS for FBS accelerated rather than delayed the progression of Kirsten murine sarcoma virus-induced transformation. These results show that Kirsten murine sarcoma virus-induced transformation of adult presenescent cells is controlled by physiological factors to which immortal cells appear refractory. The changes in subcellular distribution of p21 during transformation parallel the expression of some, but not other, transformation parameters and suggest a possible association of p21 with increased membrane activity.

Time-resolved confocal analysis of antibody penetration into living, solid tumor spheroids.

The in vivo function of a biologically active molecule is governed in part by the dynamics of its distribution within its target tissue. To enhance our ability to probe living cells, we have endeavored to improve live confocal microscopy methods and to develop analytical methods that simplify the handling of the resulting complex data sets. To do this we attached a recently developed micro-incubation system to the stage of a Leica confocal laser scanning microscope and were able to maintain physiologic culture conditions over several hours. Axial stability was achieved by modifying the room air conditioning. Laser illumination was low enough to retain cell viability through several hours of continuous scanning. With this setup, planar, time-resolved data sets (xyt) were produced by continuously rescanning a single xy plane at the rate of one scan/min. As an alternative, volumetric data sets (xyz) were acquired by stepping the scanned plane through the z axis. In both types of data sets, a semi-quantitative determination of the concentration of a fluorescent reporter molecule (e.g., FITC) over a gray level range of 0.255 was recorded along with the positional information. Thus, concentration (as intensity of fluorescence, or i) gave a fourth variable by either scan method, resulting in high-density xyti or xyzi data sets. The biological model we used to examine these methods was the penetration of a FITC-labeled, anti-carcinoma monoclonal antibody into cultured spheroids of tumor cells bearing the antibody-binding epitope. In one case, the distribution of antibody-FITC conjugate was compared with that of a long wavelength membrane dye, DiIC18(5). Several different software analyses were compared, including examining xyt data sets as "volumes". We observed that by increasing the displayed resolution of one variable, the demonstrable resolution of the other variables was reduced. For example, with high temporal resolution, either quantitative or positional resolution had to be sacrificed. Thus, we needed to perform several different analyses of a single data set to compare all of the variables properly. In these experiments, the dynamic aspects of the changes in antibody-FITC distribution were examined. Along with comparison of antibody-FITC penetration with that of DiI, these data suggest an as yet unexplained biological transport of antibody into a tumor spheroid, which is not consistent with mere passive diffusion through the fluid of extracellular clefts. Using this model system, we have performed and analyzed highly time-resolved confocal microscopy on living specimens maintained under physiologic conditions.

Regional variations in tuberculosis policy in the Cape and Ciskei.

Regional variations in tuberculosis policy in Cape Town, Paarl and Ciskei, as well as problems experienced by health workers, are examined. South African policy is compared with modern trends and recommendations drawn from the international literature and is found to conform in many respects, although tuberculosis services are not always integrated with curative services. The most marked variation occurs in Ciskei, where policy requires hospitalization of as many patients as possible. The majority of problems in implementing policy relate to lack of funds and medical infrastructure.

The implementation of tuberculosis policy in three areas in South Africa.

The implementation of tuberculosis policy at hospital and clinic level was examined in three areas (Cape Town, Paarl and Ciskei). Investigation showed that bacteriological diagnosis, standardized treatment regimens, supervision of therapy and contact tracing were not being correctly implemented. Compliance was also poor.

Tuberculosis--the patients' perspective.

This, the last of three studies examining aspects of tuberculosis (TB) control in South Africa, ascertained the views of the consumer--the TB patient. Aspects such as the impact of contracting TB on employment opportunities and social life, difficulties in getting treatment, and knowledge of TB were studied in a sample population of urban black working-class TB patients. Patients had difficulties in getting treatment at centralized points, and health education was not effective. Poor socio-economic conditions were also important in the experience of the TB sufferer.

p21ras. Heterogeneous localization in transformed cells.

The cellular targets for the Kirsten murine sarcoma virus (KiMSV)-transforming protein, p21ras, are unknown. Other studies have indicated that the mature form of p21 is distributed diffusely on the cytoplasmic face of the plasma membrane. However, after fixation without buffer washes, indirect immunofluorescent staining of sparse cultures revealed a particularly well preserved cellular architecture and a strikingly heterogeneous subcellular distribution of p21 in transformed normal rat kidney (NRK) cells but not in their untransformed counterparts. The transformed cells included A KiMSV-transformed NRK line. NRK cells newly transformed with KiMSV. A temperature-sensitive (ts) KiMSV-transformed NRK line. An uninfected, spontaneously transformed NRK line in which p21 was neither phosphorylated nor overproduced. In the tsKNRK line p21 was abundant at both permissive and non-permissive temperatures; however, its distribution was heterogeneous at the permissive temperature only. Observation of this array of cells indicates that the transformation-associated p21 distribution does not require overexpression of the gene, nor phosphorylation of the protein, nor the viral oncogene. Furthermore, it is reversible in the tsKNRK cells, and so appears to be highly correlated with acquisition of a transformed morphology. Accumulations of p21 occurred preferentially in subcellular locations similar to those where ruffles were observed by phase contrast microscopy and lamellar and villous extensions were observed by scanning electron microscopy (SEM). Since enhanced ruffling is a morphological correlate of transformation in a variety of cells, the distribution of p21 observed here may relate to its function as a transforming molecule.

Concanavalin A increases spontaneous beat rate of embryonic chick heart cell aggregates.

The plant lectin concanavalin A (Con A), at concentrations of 5-200 micrograms/ml, induced a twofold to fivefold increase in spontaneous beat rate of cultured aggregates of ventricular cells from seven-day chick embryos. This response was time, dose, and temperature dependent and was accompanied by a decrease in transmembrane potential. It could be blocked or reversed by alpha-methyl-D-mannoside but was not reversed by dilution alone. Binding of the lectin occurred in the cold, but a temperature-dependent process was also necessary to produce the response. Divalent (succinyl) Con A did not cause a beat rate increase. Whole heart aggregates responded similarly but less intensely than ventricular aggregates. Atrial aggregates, and whole heart aggregates treated with 5 microgram/ml of Con A, produced a biphasic chronotropic response, first decreasing then increasing their beat rates. These results suggest that saccharide-bearing macromolecules on the heart cell surface play a role in regulating spontaneous beat rate.

Attitudes to the provision of primary health care at the day hospitals in Cape Town.

Day hospitals in Cape Town are examined against the criteria in the definition of primary health care of the World Health Organization's Declaration of Alma Ata. A survey of patients attending a day hospital was undertaken to ascertain the consumer perspective regarding access to and quality of the service offered. The provider's perspective was obtained from secondary sources and in-depth interviews. It was found that the services are generally acceptable to the users, but that factors such as waiting time and transport create problems for patients. Community participation is also very limited. The separation of preventive and curative services also places limitations on the day hospitals--they provide only curative health care. The most important reason given by patients for using the day hospitals is the low cost.

Recloning of SV40 early gene transfected human endothelial cells repeatedly recovers subpopulations with low passage characteristics and morphologies.

Cultured human umbilical vein endothelial cells (HUVECs) are a valuable model for investigation of endothelial functions, but they enter senescence at low passage. Transfection of early passage HUVECs with the early genes of SV40 greatly extends the replicative potential of these cells, but eventually results in marked changes in growth, morphology, and biochemistry. Here we report a modified approach that appears to have overcome the problem of late passage decline after transfection. Plasmid pX-8 containing the SV40 early genes was transfected into passage four HUVECs. At passage five, these transfectants were cloned by limiting dilution and selected on the basis of both morphological and biochemical resemblance to their untransfected counterparts. Two clones that expressed factor VIII and in which the basal and the tumor necrosis factor-alpha inducible levels of interleukin 6 and endothelial adhesion molecules were normal were chosen. Vimentin and fibronectin distribution in these clones resembled untransfected cells. At passage 25, growth pattern changes were becoming evident, but recloning these late passage clones recovered numerous subclones of normal, cobblestone appearance. Two of these were further characterized and found to resemble their original parental clone by all of the biochemical criteria listed above. These subclones appeared to transform more rapidly than the parental clone, but repeated subcloning again rescued clones with normal morphologies and normal biochemical characteristics. We conclude that periodic recloning may indefinitely perpetuate lines that are useful equivalents of their original counterparts.

The effect of epidermal growth factor on chronotropic response in cardiac cells in culture.

Cardiac chronotropic response to epidermal growth factor (EGF) was assessed in chick embryonic ventricular cell aggregates. EGF at a concentration of 10 ug/mL but not at 5 ug/mL produced a significant (p less than 0.05) increase in cardiac beating rate. This was evident within 10 min, reached a peak at about 15 min and remained at that level for 1.5 hr or the rest of the observation period. The effect of EGF on cardiac automaticity was reduced but not abolished at a lower temperature (22oC) that is known to decrease the affinity of the EGF receptor and reduce the internalization of EGF. Hypothermia did not change the maximum increase in heart rate response from isoproterenol although it altered the pattern of the response. Beta adrenoreceptor blockade with metoprolol only slightly altered the response to EGF. These data indicate that EGF produces functional effects on the heart that may be mediated through EGF receptor linked mechanisms.

Human ovarian surface epithelium in primary culture.

The ovarian surface epithelium (OSE) represents a minute fraction of the cell mass of the ovary but gives rise to over 80% of human ovarian carcinomas. No experimental models for the study of human OSE exist. To characterize OSE cells in culture, explants of ovarian surface from normal ovary of premenopausal women were grown on plastic, glass, and collagen gel in 25% fetal bovine serum/Waymouth's medium 752/1. About 25% of explants produced epithelial outgrowths. Morphologically, these outgrowths resembled OSE in vivo and endothelial and mesothelial cells in culture, but they differed from cultured ovarian stromal, granulosa, and luteal cells. Only OSE among ovarian cell types were intensely keratin positive by immunofluorescence. Keratin also distinguished OSE cells from the keratin-negative endothelial cells. Most but not all OSE colonies tested showed 17 beta-hydroxysteroid dehydrogenase (HSD) activity, which was absent in peritoneal mesothelial cells. Colonies from most patients were limited to a few millimetres and became stationary within a few weeks. Changes that accompanied cessation of growth included senescence, increased keratin content, or the formation of multicellular papillary aggregates. With time, OSE cells tended to assume a fibroblast-like morphology but remained keratin positive and continued to resemble OSE by scanning electron microscopy (SEM). Subcultured OSE cells persisted in a stationary keratin-positive form for many weeks. Throughout this study, all pavementlike epithelial outgrowths that were contiguous with an explant stained for keratin; thus, such colonies can be assumed to be OSE. Conversely, fibroblast-shaped cells may represent OSE as indicated by keratin content and SEM appearance. The methods presented here permit culture of normal human OSE under conditions in which the cells exhibit morphologic plasticity, variable 17 beta-HSD activity, and presence of keratin.

Lineage-specific induction of B cell apoptosis and altered signal transduction by the phosphotyrosine phosphatase inhibitor bis(maltolato)oxovanadium(IV).

Protein tyrosine phosphorylation is known to play key roles in lymphocyte signal transduction, and phosphotyrosine phosphatases (PTP) can act as both positive and negative regulators of these lymphocyte signals. We sought to examine the role of PTP further in these processes by characterizing the effects of bis(maltolato)-oxovanadium(IV) (BMLOV), previously known to be a nontoxic insulin mimetic agent in vivo. BMLOV was found to be a potent phosphotyrosine phosphatase inhibitor. BMLOV induced cellular tyrosine phosphorylation in B cells in a pattern similar to that observed following antigen receptor stimulation, whereas little tyrosine phosphorylation was induced in T cells. In B cells, BMLOV treatment resulted in tyrosine phosphorylation of Syk and phospholipase C gamma 2, while sIgM-induced signals were inhibited. By contrast, T cell receptor signals were moderately increased by BMLOV, and the cells displayed greater induction of IL-2 receptor without toxicity. The compound selectively induced apoptosis in B cell lymphoma and myeloid leukemia cell lines, but not in T cell leukemia or colon carcinoma cells. Interleukin-4 plus anti-CD40 antibody treatment of normal human peripheral B cells rescued the cells from BMLOV-induced death. These results suggest that phosphotyrosine phosphatase inhibitors can activate B cell signal pathways in a lineage-specific manner, resulting in desensitization of receptor-mediated signaling and induction of apoptosis.

Divergence in patterns of invasion among subpopulations derived from a human carcinoma clone: roles of intercellular contacts and of cell-substratum adhesion.

In tumor progression, populations of cancer cells with different patterns of growth and invasion arise within the same tissue and within individual neoplasms. We tested the hypothesis that, even in histologically undifferentiated carcinomas, such diversity may be influenced by differentiation-dependent adhesive mechanisms. We used as prototypes two cell lines that originated in the same clone of a poorly differentiated cervical carcinoma, but express strikingly different phenotypes. Cells of line C-4I express select characteristics of the spinous stage of stratified squamous epithelial differentiation while cells of line C-4II resemble basal cells. C-4I cells form rapidly expanding compact tumors in vivo and multilayered cohesive colonies in culture, while C-4II cells form slow-growing infiltrating tumors in vivo and dispersed, monolayered colonies in culture. In suspension culture which prevented any cell-substratum interactions, C-4I cells formed aggregates that were significantly larger and more compact than those formed by C-4II. Thus, greater intercellular adhesion between the 'spinous' C-4I cells contributed significantly to the phenotypic divergence of the lines. Upon disruption of intercellular adhesion with the glutamine analogue 6-diazo-4-oxo-norleucine (DON), C-4I cultures on plastic and in suspension assumed forms resembling C-4II. On plastic, single 'basal' C-4II cells adhered more rapidly and migrated more slowly than C-4I cells, in keeping with the capacity of C-4II, but not C-4I, to secrete fibronectin (FN) substrata. However, on exogenous FN matrices, migration and cell dispersion were accelerated in both lines. Both lines expressed similar integrin profiles. Thus, the lines had diverged in extracellular matrix production, but not in the receptors for extracellular matrix components. The properties of the C-4 lines mimic those of specific cell types in normal stratified squamous epithelia, where intercellular adhesion increases but FN secretion diminishes with progression from the basal to the spinous stage of differentiation. Our results demonstrate a direct influence of differentiation-associated adhesive mechanisms on growth patterns and suggest that similar mechanisms may be responsible for variations in invasiveness among neoplastic clonal subpopulations. An awareness of these correlations may help to interpret the modes of local invasion by poorly differentiated carcinomas in terms of specific, well-defined cell properties.

Localized elaboration of IgA in normal and neoplastic murine mammary cell cultures: relationship to fibronectin matrix.

IgA secretion by the lactating mammary gland culminates a complex sequence of biologic events both within the gland and at distant sites. Because of the many extraglandular influences, it is difficult to investigate IgA secretion at the tissue and cellular levels in the intact animal. In this study, with the use of immunohistofluorescence, we have observed elaboration of IgA by primary monolayer cultures of mammary cells from the glands of mid-pregnant mice and from mammary tumors. In cultures of normal cells, IgA appeared first in vesicular structures on the upper surfaces of the monolayers. Vesicular IgA was maximal at day 5 in culture, and at that time, was observed only over mounds (domelike structures) that were covered with fibrillar fibronectin (FN), and eventually developed a subpopulation of fusiform cells. It appears that the IgA observed was secreted in vitro, that normal mammary epithelial cells must form multicellular FN-bearing structures to secrete IgA in culture, and that by mid-pregnancy the murine mammary gland contains all of the lymphoid cells required for IgA secretion. In contrast, primary cultures of mammary tumors displayed minimal amounts of IgA and FN. The small amount of cell-associated IgA that was detected on tumor cultures was not localized to any multicellular structures nor was it associated with FN, but instead appeared as minute, punctate accumulations on individual cells scattered across the epithelioid areas. This finding is consistent with the loss of specialized functions and the increased autonomy typical of malignant cells. The study in cultured cells of a function as specialized as IgA secretion should permit greater understanding of both the process itself and the despecializations that accompany malignancies of secretory epithelia.

Costimulation of T lymphocytes with integrin ligands intercellular adhesion molecule-1 or vascular cell adhesion molecule-1 induces functional expression of CTLA-4, a second receptor for B7.

Costimulation by the CD28 ligand B7/BB1 plays an important role during T cell proliferation primarily by augmenting synthesis of IL-2 and other cytokines. Resting CD4+ T cells express CD28 but not CTLA-4 on their surface. Costimulation of T cells with ICAM-1 or VCAM-1 induced CTLA-4 expression and up-regulated CD28 expression. CD28 and CTLA-4 were independently distributed on the surface of activated T lymphoblasts. When co-immobilized with anti-TCR mAb both anti-CD28 and anti-CTLA-4 mAb augmented T cell proliferation. Although anti-CD28-mediated augmentation of T cell proliferation was stronger than that seen with anti-CTLA-4 mAb, together these two mAb caused supraadditive augmentation of T cell proliferation. The augmentation of the effects of anti-CD28 mAb by anti-CTLA-4 mAb was greater at low occupancy of CD28 by anti-CD28 mAb. Costimulation of CD28+ CTLA-4+ T cells with anti-CTLA-4 caused three- to fivefold increase in IL-2 production, whereas similar treatment with anti-CD28 caused > 40-fold increase. The costimulatory effect of B7 on primed T cells was partially inhibited by Fab anti-CD28 mAb. Anti-CTLA-4 mAb alone did not inhibit B7-induced response but caused modest increase in the inhibitory effect of anti-CD28 Fab. On integrin-mediated costimulation, Ag-specific CD4+ T cell lines also up-regulated their CTLA-4 expression, and proliferation of these cells was augmented by anti-CTLA-4 mAb. Unlike that of CD28, ligation of CTLA-4 alone failed to mobilize intracellular [Ca2+]. However, coligation of CTLA-4 and TCR induced stronger [Ca2+] response in Ag-specific T cell lines than that seen with TCR alone. These results suggest that integrin-costimulated T cells express CTLA-4 and can be costimulated via CTLA-4. Optimal development of various immune functions may involve combined costimulation via both CD28 and CTLA-4.

Intracellular CD22 rapidly moves to the cell surface in a tyrosine kinase-dependent manner following antigen receptor stimulation.

CD22 is a key accessory molecule for Ag receptor signaling in B cells that becomes tyrosine phosphorylated in the signaling process. CD22 associates with sIg and strongly amplifies sIg-induced signals. During B cell development, CD22 is initially expressed intracellularly, with surface expression appearing with IgD expression. We used confocal laser-scanning microscopy and flow cytometry to analyze CD22 translocation responses to signaling events. Cross-linking surface IgM (sIgM) induced rapid movement of CD22 to the cell surface in both Ramos and Daudi B cells, with a 50 to 100% increase in surface expression observed within 5 min of stimulation. The increase in CD22 surface expression was specific in that CD19 expression was not affected. Both cell surface and intracellular CD22 were directed toward the site of sIgM stimulation. Treatment with the phosphotyrosine phosphatase inhibitor bis(maltolato)oxovanadium(IV) also increased CD22 surface expression. The tyrosine kinase inhibitor tyrphostin A10 inhibited CD22 movement at concentrations that inhibited tyrosine phosphorylation of CD22 and other cellular proteins. In contrast to the B cell lines, mature peripheral blood B cells contained very little intracellular CD22 and showed no significant increase in surface expression following sIgM stimulation. The rapid directed movement of intracellular CD22 provides a new mechanism to dynamically regulate Ag receptor signaling, as well as CD22-mediated adhesion.

T-cell activation by the CD28 ligand B7 is required for cardiac allograft rejection in vivo.

Organ graft rejection is a T-cell-dependent process. The activation of alloreactive T cells requires stimulation of the T-cell receptor/CD3 complex by foreign major histocompatibility complex (MHC)-encoded gene products. However, accumulating evidence suggests that, in addition to T-cell receptor occupancy, other costimulatory signals are required to induce T-cell activation. Previously, the CD28 receptor expressed on T cells has been shown to serve as a surface component of a signal transduction pathway that can provide costimulation. In vitro, interaction of CD28 with its natural ligand B7 expressed on the surface of activated B cells or macrophages can act as a costimulus to induce proliferation and lymphokine production in antigen receptor-activated T cells. We now report evidence that stimulation of T cells by the CD28 ligand B7 is a required costimulatory event for the rejection of a MHC-incompatible cardiac allograft in vivo. These results demonstrate that the B7/CD28 activation pathway plays an important role in regulating in vivo T-cell responses.