Electrostatic tuning of the proximity-induced exchange field in EuS/Al bilayers.

We demonstrate that the proximity-induced exchange field H(ex) in ferromagnetic-paramagnetic bilayers can be modulated with an electric field. An electrostatic gate arrangement is used to tune the magnitude of H(ex) in the Al component of EuS/Al bilayers. In samples with H(ex)~2 T, we were able to produce modulations of ±10 mT with the application of perpendicular electric fields of the order of ±10(6) V/cm. We discuss several possible mechanisms accounting for the electric field's influence on the interfacial coupling between the Al layer and the ferromagnetic insulator EuS, along with the prospects of producing a superconducting field-effect transistor.

Exchange field-mediated magnetoresistance in the correlated insulator phase of Be films.

We present a study of the proximity effect between a ferromagnet and a paramagnetic metal of varying disorder. Thin beryllium films are deposited onto a 5 nm thick layer of the ferromagnetic insulator EuS. This bilayer arrangement induces an exchange field, H(ex), of a few tesla in low-resistance Be films with sheet resistance R≪R(Q), where R(Q)=h/e2 is the quantum resistance. We show that H(ex) survives in very high-resistance films and, in fact, appears to be relatively insensitive to the Be disorder. We exploit this fact to produce a giant low-field magnetoresistance in the correlated-insulator phase of Be films with R≫R(Q).

Spin-resolved tunneling studies of the exchange field in EuS/Al bilayers.

We use spin-resolved electron tunneling to study the exchange field in the Al component of EuS/Al bilayers, in both the superconducting and normal-state phases of the Al. Contrary to expectation, we show that the exchange field H(ex) is a nonlinear function of applied field, even in applied fields that are well beyond the EuS coercive field. Furthermore, the magnitude H(ex) is unaffected by the superconducting phase. In addition, H(ex) decreases significantly with increasing temperature in the temperature range of 0.1-1 K. We discuss these results in the context of recent theories of generalized spin-dependent boundary conditions at a superconductor-ferromagnet interface.

Origin of excess low-energy states in a disordered superconductor in a Zeeman field.

Tunneling density of states measurements of disordered superconducting Al films in high Zeeman fields reveal a significant population of subgap states which cannot be explained by standard BCS theory. We provide a natural explanation of these excess states in terms of a novel disordered Larkin-Ovchinnikov phase that occurs near the spin-paramagnetic transition at the Chandrasekhar-Clogston critical field. The disordered Larkin-Ovchinnikov superconductor is characterized by a pairing amplitude that changes sign at domain walls. These domain walls carry magnetization and support Andreev bound states that lead to distinct spectral signatures at low energy.

Saturation of the anomalous Hall effect in critically disordered ultrathin CNi3 films.

We demonstrate that a distinct high-disorder anomalous Hall effect phase emerges at the correlated insulator threshold of ultrathin, amorphous, ferromagnetic CNi3 films. In the weak-localization regime, where the sheet conductance G>>e{2}/h, the anomalous Hall resistance of the films increases with increasing disorder and the Hall conductance scales as G{xy} proportional to G{phi} with phi=1.6. However, at sufficiently high disorder the system begins to enter the 2D correlated insulator regime, at which point the Hall resistance R{xy} abruptly saturates and the scaling exponent becomes phi=2. Tunneling measurements show that the saturation behavior is commensurate with the emergence of the 2D Coulomb gap, suggesting that e-e interactions mediate the high-disorder phase.

Emergence of intrinsic superconductivity below 1.178 K in the topologically non-trivial semimetal state of CaSn3.

Topological materials which are also superconducting are of great current interest, since they may exhibit a non-trivial topologically-mediated superconducting phase. Although there have been many reports of pressure-tuned or chemical-doping-induced superconductivity in a variety of topological materials, there have been few examples of intrinsic, ambient pressure superconductivity in a topological system having a stoichiometric composition. Here, we report that the pure intermetallic CaSn3 not only exhibits topological fermion properties, but also has a superconducting phase at ~1.178 K under ambient pressure. The topological fermion properties, including the nearly zero quasi-particle mass and the non-trivial Berry phase accumulated in cyclotron motions, were revealed from the de Haas-van Alphen (dHvA) quantum oscillation studies of this material. Although CaSn3 was previously reported to be superconducting with T c = 4.2 K, our studies show that the T c = 4.2 K superconductivity is extrinsic and caused by Sn on the degraded surface, whereas its intrinsic bulk superconducting transition occurs at 1.178 K. These findings make CaSn3 a promising candidate for exploring new exotic states arising from the interplay between non-trivial band topology and superconductivity, e.g. topological superconductivity (TSC).

Plasma low density lipoprotein transport kinetics in noninsulin-dependent diabetes mellitus.

Plasma low density lipoprotein (LDL) transport kinetics were determined from the disappearance of 125I-LDL injected into age- and weight-matched groups of 13 normal subjects, 20 mild diabetics, and 8 moderately severe diabetic patients (fasting plasma glucose less than 150 and greater than 150 mg/100 ml, respectively). In mild diabetics, LDL apo-lipoprotein-B (apo-B) synthetic rate (SR) was significantly greater than normal. The fractional catabolic rate (FCR), however, was also increased so that plasma LDL concentration remained normal. In moderately severe diabetics, LDL SR was normal but FCR was reduced resulting in increased plasma LDL cholesterol and apo-B concentrations. In normal subjects, moderate obesity was associated with increased LDL secretion. In diabetic subjects, however, changes in LDL turnover were of equal magnitude in obese and nonobese patients. In normolipemic and hyperlipemic mild diabetic subjects with equal degrees of glucose intolerance, both LDL apo-B SR and FCR were greater than normal. The magnitude of these increases, however, was lower in the hyperlipemic individuals. Stepwise regression analysis revealed that both LDL SR and FCR correlated positively and linearly with insulin response to glucose loading, but negatively and curvilinearly with fasting plasma glucose and glucose response. We propose that in noninsulin-dependent diabetes, mild hyperglycemia is accompanied by increased LDL turnover, despite normal plasma LDL levels, whereas moderately severe hyperglycemia is associated with decreased LDL catabolism, resulting in increased plasma LDL levels. These changes cannot be attributed to the presence of obesity or hypertriglyceridemia, and may relate to varying degrees of insulin resistance and decreased insulin secretion affecting plasma very low density lipoprotein (VLDL) secretion, VLDL conversion to LDL, and LDL catabolism. Both increased LDL turnover in mild diabetes and delayed removal of LDL in moderately severe diabetes could increase cholesterol ester availability to peripheral tissues, and may result in an increased risk of atherosclerosis.

Immunosuppressive properties of a virion polypeptide, a 15,000-dalton protein, from feline leukemia virus.

The 15,000-molecular-weight polypeptide (p15) of feline leukemia virus (FeLV) was shown to impair normal lymphocyte function in vitro and to abrogate immunity to feline oncornavirus disease in vivo. FeLVp15 suppressed concanavalin A-induced blast transformation of normal feline lymphocytes by 68%, while other virion proteins had no effect. p15 suppression was not due to toxicity, nor was p15 a competitive inhibitor of concanavalin A binding. Capping of receptors for concanavalin A on normal feline lymphocytes also was inhibited by either inactivated FeLV or FeLV p15. Groups of cats were immunized with either killed feline oncornavirus-associated cell membrane antigen bearing tumor cells or tumor cells plus FeLV p15. After challenge with feline sarcoma virus, three of four p15-treated cats developed progressive fatal fibrosarcoma as compared to one of five non-p15-treated cats. The cats receiving p15 also had lower cytotoxic antibody titers against feline oncornavirus-associated cell membrane antigen (mean peak titer, 1:6) than did the non-p15 group (1:74). These data support the hypothesis that the immunosuppression in cats infected with FeLV is mediated by FeLV p15.

Integrated regulation of very low density lipoprotein triglyceride and apolipoprotein-B kinetics in non-insulin-dependent diabetes mellitus.

Turnover rates of plasma very low density lipoprotein (VLDL) triglyceride (TG) and apolipoprotein-B (apo-B) were significantly increased in age- and weight-matched groups of normolipemic and hyperlipemic mild diabetics, and hyperlipemic moderately severe diabetics when compared with normolipemic controls. As in the normolipemic subjects, a significant correlation between VLDL TG and apo-B turnover rates was found in all diabetic groups, suggesting that integration of TG and apo-B production at synthetic and/or secretory sites is retained in diabetes, thus resulting in increased secretion of VLDL particles of normal composition. In normolipemic mild diabetic subjects, the fractional turnover rates of VLDL TG and apo-B were also significantly increased so that increased removal accompanied increased VLDL production. In the hyperlipemic diabetics, however, the fractional turnover rates were significantly reduced, hence the increased in VLDL removal was not sufficient to compensate for enhanced production. In normolipemic mild diabetic patients, low density lipoprotein (LDL) formation was increased, only a small fraction of VLDL apo-B being removed via a non-LDL pathway, presumably as remnant VLDL. In hyperlipemic mild diabetics, removal of VLDL apo-B via both the LDL and non-LDL pathways was increased. In hyperlipemic moderately severe diabetes, LDL formation was not increased; catabolism of VLDL apo-B through the non-LDL route was however, fivefold greater than normal. We conclude that increased VLDL secretion is a fundamental defect in non-insulin-dependent diabetes. In hyperlipemic individuals, VLDL removal is also impaired. The increase in LDL and/or VLDL remnant formation, regardless of prevailing plasma lipid levels or the severity of diabetes, provides a source of cholesterol which may account for the atherogeneity of this disorder.

Relation of body fat distribution to metabolic complications of obesity.

The importance of body fat distribution as a predictor of metabolic aberrations was evaluated in 9 nonobese and 25 obese, apparently healthy women. Plasma glucose and insulin levels during oral glucose loading were significantly higher in women with predominantly upper body segment obesity than in women with lower body segment obesity. Of the former group, 10 of 16 subjects had diabetic glucose tolerance results, while none of the latter group was diabetic. Fasting plasma triglyceride levels were also significantly higher in the upper body segment obese women. The site of adiposity in the upper body segment obese women was comprised of large fat cells, while in the lower body segment obese subjects, it was formed of normal size cells. In both types of obesity, abdominal fat cell size correlated significantly with postprandial plasma glucose and insulin levels. Thigh fat cell size gave no indication as to the presence of metabolic complications. Thigh adipocytes were also resistant to epinephrine-stimulated lipolysis, presumably due to an increase in alpha-adrenergic receptors. Thus, in women, the sites of fat predominance offer an important prognostic marker for glucose intolerance, hyperinsulinemia, and hypertriglyceridemia. This association may be related to the disparate morphology and metabolic behavior of fat cells associated with different body fat distributions.

The effect of cholestyramine on the faecal excretion of bile acids and neutral steroids in familial hypercholesterolaemia.

The faecal excretion of total bile acids was measured in two normal subjects and in seven patients with familial hypercholesterolaemia (four heterozygotes and three homozygotes) in the untreated state and during treatment with near-maximal doses of cholestyramine. There were no significant differences between the three groups. The increase in bile-acid excretion in response to cholestyramine was as great in the homozygotes as in the normal subjects. It is concluded that familial hypercholesterolaemia is not generally due to an inherited defect in the mechanisms for catabolizing cholesterol to bile acids.

Transport kinetics of plasma free fatty acid, very low density lipoprotein triglycerides and apoprotein in patients with endogenous hypertriglyceridaemia: effects of 2,2-dimethyl, 5(2, 5-xylyoxy) valeric acid therapy.

The effects of C1-719 on plasma lipid and lipoprotein concentrations have been examined in four patients with endogenous hypertriglyceridaemia maintained on an isocaloric diet for a period of 6 months. During therapy (400 mg/day) the mean plasma triglyceride and cholesterol concentrations were reduced by 35% and 15% respectively, while the administration of 800 mg/day reduced these by 49% and 31%. This hypolipidaemic effect was due to a reduction in the circulating level of very low density lipoproteins (VLDL) without a change in their composition. Before treatment the plasma VLDL triglyceride turnover, and FFA flux, were higher than that of normal subjects maintained on a similar diet. The plasma VLDL B-apoprotein turnover was similarly higher than in the controls. Administration of C1-719 decreased the plasma VLDL triglyceride turnover, FFA flux and VLDL B-approtein turnover. The drug reduced the insulin response following a glucose load with some decrease in glucose levels. The results suggest that the increase in plasma triglyceride concentration in patients with endogenous hypertriglyceridaemia is due to increased production of plasma VLDL triglyceride and its apoptein associated with an enhanced supply of FFA for hepatic triglyceride synthesis. C1-719 exerts a hypolipidaemic effect through a reduction of VLDL production, consequent upon inhibition of lipolysis as well as decreased synthesis of the apoprotein carrier. These effects could in part be explained by an improvement in peripheral tissue responsiveness to insulin and decreased exposure of the liver to high levels of insulin. However, a direct effect of the drug on adipose tissue and liver metabolism has to be considered.

Limiting dilution analysis of human, tetanus-reactive helper T lymphocytes. A rapid method for the enumeration of helper T lymphocytes with specificity for soluble antigens.

The number of helper T lymphocytes (HTL) in human peripheral blood with specificity for the soluble protein, tetanus toxoid, was estimated by limiting dilution analysis (LDA). HTL were detected via antigen-induced interleukin-2 (IL-2) production, as measured by incorporation of tritiated thymidine by an IL-2-dependent indicator cell line, CTLL-20. Culture conditions optimizing assay sensitivity are described, and the ability to detect antigen-specific HTL using this LDA technique are demonstrated. Observed HTL frequencies in healthy human donors tested for tetanus-reactive helper T cells ranged from less than 1 HTL/268,749 peripheral blood mononuclear cells (PBMC) (undetectable) to 1 HTL/1486 PBMC. The LDA technique was also used to detect frequency shifts in human peripheral blood HTL following challenge with antigen. This assay offers distinct advantages over proliferative LDA techniques in that it is rapid (requiring only 2 days), and defines an antigen-reactive T cell subset with defined function (IL-2 secretion). Furthermore, the LDA technique can be adapted for the detection of other soluble protein antigens, such as PPD and Candida albicans. In general, this LDA technique provides a reliable, quantitative index of human HTL reactivity to any of a variety of soluble protein antigens, and has clinical as well as experimental applicability.

Transperineal ultrasound-guided radioactive seed implantation for organ-confined carcinoma of the prostate.

PURPOSE: This retrospective study was undertaken to: (a) determine the prognostic significance of pretreatment and 1-year nadir serum prostate specific antigen (PSA) levels in organ-confined carcinoma of the prostate treated with ultrasound-guided radioactive 125I seed implantation; (b) determine the factors associated with postimplant morbidity and whether modification of the technique would reduce morbidity; (c) evaluate the local control rate and disease-free survival of patients undergoing seed implantation. METHODS AND MATERIALS: From October 1988 through December 1992, 142 patients with organ-confined adenocarcinoma of the prostate and a Gleason score < or = 7 underwent ultrasound-guided radioactive 125I seed implantation as an alternative to radical prostatectomy. Patients were considered to have persistent or progressive disease if there was evidence of local progression on digital exam, or if there were two consecutive increases in the PSA level. Patients suspected of persistent or progressive disease underwent restaging to include CT scan of the pelvis, bone scan, and ultrasound-guided prostate biopsy. Patients with increasing PSA levels in which active disease could not be confirmed were considered biochemical failures with occult systemic disease and were offered hormone ablation. RESULTS: With 1-6-year follow-up, median 30 months, the relapse patterns were prostate 4 (2.8%), bone 4 (2.8%), rising PSA 16 (11%). Pretreatment PSA level correlated with subsequent recurrence; pretreatment PSA < or = 4 (0), 4.1 to 10 (14%), 10.1 to 20 (21%), 20.1 to 50 (58%). Disease free survival at 2 years was 90% and at 5 years 76%. Nadir PSA (nPSA) at 1 year also correlated with recurrence: nPSA < or = 1 (3%), nPSA 1 < or = 4 (50%), and nPSA > or = 4 (100%). Seed implantation was well tolerated with 31% of patients experiencing RTOG morbidity > or = Grade 2, which typically consisted of transient radiation urethritis, which resolved with conservative measures. Eleven (8%) experienced RTOG morbidity > or = Grade 3. There was no correlation between number of seeds or total millicuries implanted and subsequent morbidity. However, reduction in the periurethral seed intensity reduced > or = Grade 3 morbidity from 11 to 4%. CONCLUSION: Ultrasound-guided radioactive seed implantation provides excellent local control of 97%, with a median 30 month follow-up. Morbidity is comparable to other curative modalities and by modifying Blasko's technique to reduce radioactive seed strength in the periurethral area, significant morbidity is rare. Pretreatment PSA and the nadir PSA at 1 year are important predictors of subsequent disease outcome. With a liberal definition of systemic recurrence as two consecutive increases in PSA levels, the 5-year disease-free survival is 76%.

Detection of donor-reactive alloantibody in the early posttransplant period. Elimination of therapeutic ALG as a complicating factor.

We have developed an immunoabsorbent reagent that can differentially remove ALG from human serum samples in vitro--i.e., goat antihorse IgG covalently linked to Sepharose beads. When used according to protocol, this immunoabsorbent can effectively remove up to 0.78 mg/ml of ALG from human serum mixtures. To demonstrate that immunoabsorption is selective, an HLA-B7-specific alloserum was mixed with a known amount of ALG and absorbed with antibody-conjugated Sepharose beads. The addition of ALG to the serum sample caused high degrees of nonspecific lympholysis in standard microcytotoxicity assays. When this serum/ALG mixture was immunoabsorbed detectable ALG activity was lost but the original HLA alloantibody titer (1:2) against specific target lymphocytes was retained.

Cytolytic activity against allogeneic human endothelia: resistance of cytomegalovirus-infected cells and virally activated lysis of uninfected cells.

BACKGROUND: Cytomegalovirus (CMV) has been implicated as an exacerbating agent in the development of transplant vascular sclerosis; however, specific etiologic mechanisms remain unresolved. Based upon our previous observations that CMV-infected endothelial cells (ECs) stimulate proliferation and cytokine production by allogeneic T cells, we now test the hypothesis that CMV-driven cytolytic activity may contribute to graft endothelial injury. METHODS: Limiting dilutions of CMV-seropositive or -seronegative donor-derived T cells were stimulated with CMV-infected or uninfected allogeneic ECs in the presence of interleukin-2. T-cell proliferation was monitored by assay of [3H]thymidine incorporation and stimulated T cells were tested for lytic activity against CMV-infected or uninfected radiolabeled EC targets by 51Cr release assay. Natural killer (NK) cell activity was examined by incubating freshly isolated peripheral blood mononuclear cells with 51Cr-labeled targets, followed by assay of radiolabel release. RESULTS: CMV-infected ECs were resistant to T cell- and NK-mediated cytolysis regardless of donor serostatus, nature of stimulation, or level of T-cell proliferation. In contrast, although uninfected ECs were unharmed by NK cells, these targets experienced significant lysis by T cells stimulated with either uninfected or CMV-infected ECs. CONCLUSIONS: These results implicate CMV-infected graft endothelium as a persistent source of infectious virus, a chronic stimulus for potentially destructive host inflammatory activity, and a potential trigger for the generation of lytic injury to uninfected bystander endothelia, suggesting multiple mechanisms by which this virus might perturb equilibrium at the graft/host interface.

Frequency of human alloantigen-reactive T lymphocytes. II. Method for limiting dilution analysis of alloantigen-reactive helper T cells in human peripheral blood.

Quantitative immunologic techniques for analysis of human alloreactivity are currently lacking in transplantation immunology. We report a rapid, sensitive, and quantitative limiting dilution analysis technique that provides a minimal estimate of the number of peripheral blood mononuclear cells (PBMC) capable of secreting interleukin-2 (operationally defined as helpter T lymphocytes) when cultured in vitro with allogeneic PBMC bearing serologically identified MHC disparities. Using this LDA technique, we have estimated that approximately 1/500 to 1/2000 (0.2% to 0.05%) of the PBMC from various individuals can secrete IL-2 after in vitro contact with completely major-histocompatibility-complex-disparate PBMC. Under normal conditions the HTL frequency in human peripheral blood appears quite stable, based on serial analysis of HTL frequency in a healthy human donor. This LDA technique is more rapid and informative than the MLR, and may be useful for pretransplant evaluation and posttransplant monitoring of donor reactivity in transplant recipients.

Frequency of human alloantigen-reactive T lymphocytes. III. Evidence that cyclosporine has an inhibitory effect on human CTL and CTL precursors, independent of CsA-mediated helper T cell dysfunction.

We have used limiting dilution analysis to study the behavior of alloantigen-reactive cytolytic T lymphocytes derived from human peripheral blood. During these studies, we found that the presence of cyclosporine in limiting dilution microcultures significantly impairs the subsequent development of alloantigen-reactive cytolytic T cell activity. As a result, CsA reduces the estimate of CTL precursor frequency by limiting dilution analysis. CTL frequency estimates are reduced by CsA in a dose-dependent manner, and concentrations of CsA that are readily achieved in human peripheral blood (100-1000 ng/ml) are capable of reducing estimates of CTL frequency by 90% to 100%. Further studies revealed that (1) human CTL derived either from fresh peripheral blood or from primary mixed lymphocyte cultures are sensitive to the suppressive effects of cyclosporine in limiting dilution microcultures, indicating that CsA influences both alloantigen-primed CTL and CTL precursors; (2) CsA impairs an immunologic event or events, that occurs for at least the first four days of limiting dilution microculture incubation; (3) CsA-mediated suppression is eliminated by separation of CTL from cyclosporine; (4) CsA blocks development of CTL generation, but not cell proliferation in limiting dilution microcultures; and (5) the CsA-mediated suppression is not reversed by supraoptimal concentrations of IL-2, high concentrations of gamma-IFN, or supplementation with the multiple lymphokines present in MLC supernatants. These data suggest that CsA may have a direct inhibitory influence on the differentiation of human CTL precursors that is independent of helper T cell dysfunction.

Alloantigenicity of human endothelial cells. III. Quantitated indirect presentation of endothelial alloantigens to human helper T lymphocytes.

We have investigated the contribution of monocytes (Mo) to the activation of purified human T cells by allogeneic human vascular endothelial cells (HUVEC). We have previously demonstrated that allogeneic HUVEC stimulate IL-2 production by CD8+ helper T lymphocytes (HTL), but not CD4+ HTL, in the absence of accessory Mo. We now show that addition of responder-autologous Mo to such cultures stimulates a high frequency of CD4+ HTL (1/6500), but no additional CD8+ HTL (1/35,000), as detected by limiting dilution analysis (LDA). The CD4+ HTL production of IL-2 increased with increasing numbers of Mo. Monoclonal antibodies to MHC class II interfered with HTL responses to allogeneic HUVEC in the presence, but not in the absence of autologous Mo. In contrast, IFN-gamma-treated HUVEC stimulated a high frequency of CD4+ HTL in the absence of autologous Mo. However, deletion experiments revealed that the population of HTL responsive to IFN-gamma-treated HUVEC is distinct from the population that responds to HUVEC in the presence of autologous Mo. These data suggest that Mo promote IL-2 production by presenting HUVEC-derived alloantigens via MHC class II molecules to CD4+ HTL, rather than by providing cytokines that promote more efficient IL-2 production by CD8+ HTL, or by inducing MHC class II expression on the HUVEC. In general, these data demonstrate that autologous Mo can play a significant role in the response of T cells to allogeneic HUVEC. Further, they demonstrate that the activation of human HTL by allogeneic HUVEC is complex, and can occur by at least three pathways: (1) direct stimulation of a small number of CD8+ HTL, but no CD4+ HTL, by quiescent HUVEC, (2) direct stimulation of a large number of CD4+ HTL by IFN-gamma-treated HUVEC, and (3) indirect stimulation of a different subset of CD4+ HTL by HUVEC in the presence of autologous monocytes. These three pathways of alloactivation are not unique to allogeneic HUVEC, but this experimental system provides a convenient and relevant model with which each pathway can be easily and independently investigated.

Alloantigenicity of human endothelial cells. II. Analysis of interleukin 2 production and proliferation by T cells after contact with allogeneic endothelia.

In organ allograft recipients, the first point of contact between the host immune system and the graft alloantigens is the graft vascular endothelia. Our previous experiments have established that there are multiple pathways by which T cells can be activated in vitro by endothelial alloantigens. In this report, we focus on the first of these pathways: the direct activation of T cells by endothelial MHC class I molecules. Using conventional mixed cell cultures and limiting dilution analyses, we demonstrate that "resting" allogeneic human umbilical vein endothelial cells (HUVEC) stimulate IL-2 production, but not proliferation of purified CD3+ PBMC. Transwell experiments demonstrate that soluble suppressive factors are not responsible for the lack of proliferation. Instead, they suggest that the production of growth factors, such as IL-2, is suboptimal in this system. Indeed, submitogenic concentrations of IL-2 synergize with HUVEC to induce strong T cell proliferative responses. This proliferation is associated with a detectable increase in T cell IL-2R expression, which is not apparent after stimulation with HUVEC or IL-2 alone. In conjunction with previous data, these observations characterize the first direct pathway of endothelia-induced T cell activation. Via this pathway, a small number of CD8+ T cells can be activated by the allogeneic MHC class I molecules displayed by resting allogeneic endothelial cells. This activation results in the elaboration of IL-2, among other things, in concentrations that are too small to promote IL-2R up-regulation, and thus T cell proliferation. Proliferation readily occurs if sufficient IL-2 is available in the environment to overcome this cytokine deficit. These studies suggest that endothelial MHC class I alloantigens are mildly antigenic to a small subset of T cells. However, in an inflammatory environment that is rich in cytokines and growth factors, these endothelial alloantigens may become potent direct stimulators of T cell activation and clonal expansion.