EuroSCORE II and N-terminal pro-B-type natriuretic peptide for risk evaluation: an observational longitudinal study in patients undergoing coronary artery bypass graft surgery.

BACKGROUND: Postoperative heart failure remains the major cause of death after cardiac surgery. As N-terminal pro-B-type natriuretic peptide (NT-proBNP) is a predictor for postoperative heart failure, the aim was to evaluate if preoperative NT-proBNP could provide additional prognostic information to the recently launched EuroSCORE II. METHODS: A total of 365 patients with acute coronary syndrome (ACS) undergoing isolated coronary artery bypass graft (CABG) surgery were studied prospectively. Preoperative NT-proBNP and EuroSCORE II were evaluated with regard to severe circulatory failure after operation according to prespecified criteria. To assess what clinical outcomes are indicated by NT-proBNP levels in different risk categories, the patients were stratified according to EuroSCORE II. Based on receiver operating characteristics analysis, these cohorts were assessed with regard to preoperative NT-proBNP below or above 1028 ng litre(-1). The follow-up time averaged 4.4 (0.7) yr. RESULTS: Preoperative NT-proBNP≥1028 ng litre(-1) [odds ratio (OR) 9.9, 95% confidence interval (CI) 1.01-98.9; P=0.049] and EuroSCORE II (OR 1.24, 95% CI 1.06-1.46; P=0.008) independently predicted severe circulatory failure after operation. In intermediate-risk patients (EuroSCORE II 2.0-10.0), NT-proBNP≥1028 ng litre(-1) was associated with a higher incidence of severe circulatory failure (6.6% vs 0%; P=0.007), renal failure (14.8% vs 5.4%; P=0.03), stroke (6.6% vs 0.7%; P=0.03), longer intensive care unit stay [37 (35) vs 27 (38) h; P=0.002], and worse long-term survival. CONCLUSIONS: Combining EuroSCORE II and preoperative NT-proBNP appears to improve risk prediction with regard to severe circulatory failure after isolated CABG for ACS. NT-proBNP may be particularly useful in patients at intermediate risk according to EuroSCORE II. CLINICAL TRIAL REGISTRATION: NCT00489827.

Mixed venous oxygen saturation predicts short- and long-term outcome after coronary artery bypass grafting surgery: a retrospective cohort analysis.

BACKGROUND: Complications of an inadequate haemodynamic state are a leading cause of morbidity and mortality after cardiac surgery. Unfortunately, commonly used methods to assess haemodynamic status are not well documented with respect to outcome. The aim of this study was to investigate Sv(O2) as a prognostic marker for short- and long-term outcome in a large unselected coronary artery bypass grafting (CABG) cohort and in subgroups with or without treatment for intraoperative heart failure. METHODS: Two thousand seven hundred and fifty-five consecutive CABG patients and subgroups comprising 344 patients with and 2411 patients without intraoperative heart failure, respectively, were investigated. Sv(O2) was routinely measured on admission to the intensive care unit (ICU). The mean (sd) follow-up was 10.2 (1.5) yr. RESULTS: The best cut-off for 30 day mortality related to heart failure based on receiver-operating characteristic analysis was Sv(O2) 60.1%. Patients with Sv(O2) <60% had higher 30 day mortality (5.4% vs 1.0%; P<0.0001) and lower 5 yr survival (81.4% vs 90.5%; P<0.0001). The incidences of perioperative myocardial infarction, renal failure, and stroke were also significantly higher, leading to a longer ICU stay. Similar prognostic information was obtained in the subgroups that were admitted to ICU with or without treatment for intraoperative heart failure. In patients admitted to ICU without treatment for intraoperative heart failure and Sv(O2) ≥60%, 30 day mortality was 0.5% and 5 yr survival 92.1%. CONCLUSIONS: Sv(O2) <60% on admission to ICU was related to worse short- and long-term outcome after CABG, regardless of whether the patients were admitted to ICU with or without treatment for intraoperative heart failure.

Mixed venous oxygen saturation is a prognostic marker after surgery for aortic stenosis.

BACKGROUND: Adequate monitoring of the hemodynamic state is essential after cardiac surgery and is vital for medical decision making, particularly concerning hemodynamic management. Unfortunately, commonly used methods to assess the hemodynamic state are not well documented with regard to outcome. Mixed venous oxygen saturation (SvO(2)) was therefore investigated after cardiac surgery. METHODS: Detailed data regarding mortality were available on all patients undergoing aortic valve replacement for isolated aortic stenosis during a 5-year period in the southeast region of Sweden (n=396). SvO(2) was routinely measured on admission to the intensive care unit (ICU) and registered in a database. A receiver operating characteristics (ROC) analysis of SvO(2) in relation to post-operative mortality related to cardiac failure and all-cause mortality within 30 days was performed. RESULTS: The area under the curve (AUC) was 0.97 (95% CI 0.96-1.00) for mortality related to cardiac failure (P=0.001) and 0.76 (95% CI 0.53-0.99) for all-cause mortality (P=0.011). The best cutoff for mortality related to cardiac failure was SvO(2) 53.7%, with a sensitivity of 1.00 and a specificity of 0.94. The negative predictive value was 100%. The best cutoff for all-cause mortality was SvO(2) 58.1%, with a sensitivity of 0.75 and a specificity of 0.84. The negative predictive value was 99.4%. Post-operative morbidity was also markedly increased in patients with a low SvO(2). CONCLUSION: SvO(2), on admission to the ICU after surgery for aortic stenosis, demonstrated excellent sensitivity and specificity for post-operative mortality related to cardiac failure and a fairly good AUC for all-cause mortality, with an excellent negative predictive value.

Myocardial metabolism before and after valve replacement for aortic stenosis.

AIM: Post ischemic disturbances of myocardial metabolism that may contribute to postoperative heart failure and are accessible to metabolic treatment have been identified early after coronary surgery. Knowledge derived from these studies may not be applicable to other patient groups. Therefore we studied myocardial energy metabolism in patients operated for isolated aortic stenosis. METHODS: Twenty patients undergoing isolated aortic valve replacement (AVR) because of aortic stenosis without significant regurgitation were studied before and immediately after surgery. Myocardial uptake of oxygen and energy substrates was assessed with coronary sinus catheter technique. RESULTS: Free fatty acids (FFA) were the main source of myocardial energy before and after AVR. A significant uptake of lactate was only recorded preoperatively. A significant uptake of glutamate of the same magnitude as previously described in coronary patients was found pre- and postoperatively. Postoperatively a relative decrease of myocardial oxygen extraction ratio (P<0.001) and oxygen consumption (P=0.14) by approximately 20% was observed. CONCLUSION: Preoperative and postoperative metabolic adaptation with substantial uptake of glutamate, previously claimed to be due to chronic or repetitive ischemia, was demonstrated. The results indicate that oxidative metabolism had not fully recovered when the procedure was completed. However, the potentially unfavorable postoperative metabolic state with predominant reliance on FFA as energy source was out-balanced by the unloading effect of AVR with a reduction in myocardial oxygen extraction.

A multicenter initial clinical experience with right heart support and beating heart coronary surgery.

BACKGROUND: During coronary surgery without CPB, exposure of posterior vessel via sternotomy can cause deterioration of cardiac hemodynamics requiring inotrope drugs support. Recent animal experiments demonstrate hemodynamic benefit of right heart support (RHS) with the AMED system. The purpose of this study was to evaluate the hemodynamic effects during cardiac manipulation to expose the posterior coronary arteries, and determine the effect of RHS in restoring hemodynamics, increasing anastomotic exposure and reducing inotropic requirements. MATERIAL AND METHODS: From July 28 to December 29, 32 patients (25 men/7 women), mean age of 63.4 (+/- 6.2 years, ages: 49-78) received coronary revascularization with the A-Med RHS device. They were divided into two groups of 16 patients, A and B. Group A patients had at least one circumflex branch bypassed. The anterior wall was systematically bypassed off-pump without RHS. The right coronary artery (RCA) and the obtuse coronary artery (OM) were completed utilizing RHS. In group B patients, all vessels including anterior vessels were bypassed with the RHS. Mean arterial pressure (MAP), mean pulmonary arterial pressure (PAP), cardiac output (CO) and the average pump flow (APF) were recorded during the OM and RCA bypass for group A, and for group B LAD data was also recorded. RESULTS: Elective beating heart coronary artery bypass graft (CABG) was successfully accomplished in 32 patients with RHS. Data measurements recorded in Group A showed the improved hemodynamic recovery for OM and RCA bypass with RHS. The MAP increased from 44 to 68 mmHg (OM) and from 63 to 81 mmHg (RCA); the CO from 2.1 to 4.4 L/min (OM) and from 3.3 to 4.7 L/min (RCA). In group B, the data recorded showed the stability of the MAP in all vessels bypassed (LAD, OM and RCA). No device-related patient incidents occurred. All 32 patients were discharged to their homes. CONCLUSIONS: The AMED system, as RHS support, facilitated coronary bypass without CPB to posterior vessels, restoring hemodynamics, providing better exposure to anastomotic sites and apparently reducing inotropes need. Prospective randomize trials are necessary to confirm this initial experience.

Neurological injury after surgery for ischemic heart disease: risk factors, outcome and role of metabolic interventions.

OBJECTIVES: Neurological complication remains a feared and increasing problem in association with cardiac surgery. The aim of this study was to analyze risk factors for neurological complications in a cohort of patients in whom inotropes for weaning from cardiopulmonary bypass was gradually replaced by metabolic treatment. METHODS: The records of 775 consecutive patients undergoing coronary artery bypass grafting (CABG) or combined CABG+valve procedures were examined. Forward stepwise multiple logistic regression analysis was used for statistical evaluation of independent risk factors. RESULTS: The incidence of neurological injury was 1.8% in patients undergoing isolated CABG and 5.4% after combined CABG+valve procedures. After cross-validation multivariate analysis identified history of cerebrovascular disease, advanced age and aortic cross-clamp time as independent risk factors for postoperative cerebral complications. Chronic obstructive pulmonary disease and number of bypasses also emerged as risk factors in the primary analysis. CONCLUSIONS: In general, markers for advanced atherosclerosis, with history of cerebrovascular disease as the most important, emerged as predictors for neurological injury. Although it did not enter the final risk model, the results also suggest that postoperative heart failure deserves further surveillance as a potential risk factor for neurological complications. However, no evidence for untoward neurological effects associated with glutamate or glucose-insulin-potassium treatment was found.

Tumor surveillance: expression of the transporter associated with antigen processing (TAP-1) in ex-vivo human tumor samples and its elevation by in vitro treatment with IFN-gamma and TNF-alpha.

We have analyzed 30 human tumor specimens (two breast, six lung, and 21 ovarian carcinomas and one malignant melanoma) for the expression of the transporter associated with antigen processing, TAP-1. Cell suspensions judged to contain negligible contamination with non-tumor cells were tested for reactivity with the antibody directed to TAP-1 in Western blot. According to these results all lysates prepared from the tumor samples contained the protein, but with considerable quantitative variations. Tumors exposed in vitro for a short time to IFN-gamma and TNF-alpha had elevated TAP-1 levels in 12 out of 30 experiments. In accordance with previous results, the tumor cell populations were heterogeneous with regard to MHC class I expression, as analyzed by indirect immunofluorescence on viable cell suspensions. Short term in vitro treatment with IFN-gamma and TNF-alpha elevated MHC class I expression in several tumors. In several mixed cultures, cytokine treated tumor cells induced cytotoxic activity in autologous or allogeneic lymphocyte populations. In vitro treatment with IFN-gamma and TNF-alpha upregulated thus the expression of MHC class I and TAP-1 in a fraction of the tumor samples. However the potentiation of the capacity to interact with T cells induced by the cytokines was not always parallel with the changes in these two parameters. It is therefore likely that the cytokine treatment induced additional changes in the tumors which contributed to their capacity to elicit the T cell response.

Assessment of myocardial glutamate requirements early after coronary artery bypass surgery.

Glutamate is an important substrate for the intermediary metabolism of the heart, particularly in association with ischemia. Early after coronary artery bypass surgery (CABG) myocardial uptake of glutamate seems to be limited by substrate availability (arterial levels). However, glutamate is not an innocuous substrate. As arterial levels of glutamate are important both for myocardial uptake and adverse effects, an attempt was made to determine a minimum dose of glutamate sufficient to supply the needs of the heart after CABG. Ten patients received and infusion of 220-240 ml of 0.1 M L-glutamic acid solution at varying rates during two 30-min periods, starting 2 h after uncomplicated elective CABG. Intravenous glutamate infusion caused a dose-dependent linear increase in arterial glutamate and an increased myocardial uptake of glutamate. However, myocardial uptake of glutamate correlated with arterial levels only at lower infusion rates. Although maximal peak uptake in individual patients (6.6 +/- 1.1 mumol/min) occurred at an average increase of arterial whole blood glutamate of 172 +/- 34 mumol/L, the greatest impact on myocardial glutamate uptake was achieved by increasing arterial whole blood glutamate by less than 100 mumol/L. This implies that an infusion rate of 30-40 mg glutamate/kg BW/h could suffice to achieve a maximal or near maximal myocardial glutamate uptake in most patients after CABG. The adequacy of this dosage remains to be confirmed in high-risk patients.

Transforming growth factor beta, (TGFbeta) secreted by immunogenic ex vivo human carcinoma cells, counteracts the activation and inhibits the function of autologous cytotoxic lymphocytes. Pretreatment with interferon gamma and tumor necrosis factor alpha reduces the production of active TGFbeta.

We present the results obtained with an in vitro model system that resembles the in vivo tumour micro-environment, where malignant cells are in close contact with the infiltrating lymphocytes. Unmanipulated blood lymphocytes were cytotoxic against the autologous ex vivo tumour cells in 3/19 patients and this function was generated in 6-day mixed cultures in five additional cases. Production of transforming growth factor beta (TGFbeta) by the freshly separated tumour cells was determined in parallel. Cytotoxicity was generated by a small number of tumour cells (2-5/100 lymphocytes), while a large number (10-20/100 lymphocytes) inhibited not only the generation but also the existing lytic activity. The presence of a neutralising TGFbeta-specific mAb (2G7) potentiated the activation of lymphocytes and suspended the suppression inflicted by the tumour cells. In those tumours, which expressed relatively high levels of MHC class I and ICAM-1 molecules, the quantity of secreted TGFbeta interfered with the ability of tumour cells to generate cytotoxic lymphocytes. In the tumours with low expression of class I, such a correlation was not detected, indicating the primordial role of MHC class I expression in the regulation of autologous tumour recognition. Our results demonstrate the involvement of TGFbeta in the impaired lymphocyte-mediated reactivity against immunogenic tumours and support a mechanism that contrasts the tolerance or anergy. Since presence of TGFbeta in the microenvironment of tumours counteracts the function of cytotoxic T lymphocytes, production of this cytokine can contribute to the failure of immunotherapy.

Addition of a thiazide: an effective remedy for furosemide resistance after cardiac operations.

BACKGROUND: The frequent use of diuretic drugs in cardiac surgical practice contrasts with the lack of documentation regarding diuretic treatment in this setting. The aims of this study were to delineate the need for diuretic drugs in adult cardiac surgical practice and to evaluate the impact of adding a combination of 50 mg hydrochlorothiazide and 5 mg amiloride orally to patients responding poorly to furosemide. METHODS: Two hundred ten consecutive patients, 159 undergoing coronary artery bypass grafting procedures and 51 having valve operations, were studied. RESULTS: Seventy-seven patients received large doses of furosemide (> or = 80 mg/24 h) at some time during the postoperative course, and of these 20 responded poorly to furosemide (weight loss 0.3 +/- 0.2 kg) despite considerable fluid retention. The addition of hydrochlorothiazide and amiloride provided a prompt and effective remedy to relative furosemide resistance. Average weight loss was 2.3 +/- 0.2 kg (p < 0.01 compared with response to furosemide) and average diuresis was 2,949 +/- 156 mL in the following 24 hours. CONCLUSIONS: Relative furosemide resistance is common after cardiac operations. Thiazides, although they are mild diuretic agents, may serve as useful adjuncts in this setting.

Mechanisms of escape from CD8+ T-cell clones specific for the HER-2/neu proto-oncogene expressed in ovarian carcinomas: related and unrelated to decreased MHC class 1 expression.

We have developed an in vitro model to study mechanisms by which ovarian tumor cells that over-express the HER-2/neu proto-oncogene escape recognition by TCD8+. Nine tumor-specific, HLA A2-restricted TCD8+ clones were isolated from 2 ovarian tumor-specific TCD8+ lines derived from tumor-infiltrating or -associated lymphocytes. Of these, 2 clones recognized the previously defined HER-2/neu epitope E75 (a.a. 369-377) and one recognized the C85 epitope (a.a. 971-979), whereas the specificity of the remaining 6 clones was unknown. Three different tumor escape variants (EVC8, EVC22 and EVC36) were produced by co-culturing an ovarian tumor line over-expressing HER-2/neu with these autologous TCD8+ clones. Cell surface expression of HLA A2 was markedly decreased on all 3 escape variants, relative to the parental tumor line, while no significant decrease in their expression of the HER-2/neu, ICAM-1 or LFA-3 molecules was found. There was a correlation between the level of tumor-specific recognition and HLA A2 expression among the tumor clones isolated from 2 of the escape variants (EVC8 and EVC36). In contrast, high HLA A2-expressing tumor clones isolated from the EVC22 variant, or EVC22 which had regained high HLA A2 expression through IFN-gamma treatment, were not recognized by the HER-2/neu-specific TCD8+ clone C-22. No mutations were found in the cDNA or the genomic DNA derived from the PCR product corresponding to a 496 bp fragment including the region coding for the E75 epitope of the HER2/neu gene in the EVC22 variant. Collectively, this in vitro model underlines the importance of decreased expression of the HLA restriction element for escape from tumor-specific TCD8+ but also demonstrates that additional mechanisms exist.

Human ex vivo carcinoma cells produce transforming growth factor beta and thereby can inhibit lymphocyte functions in vitro.

We tested 20 human carcinoma samples for the production of transforming growth factor beta (TGFbeta) in vitro. Tumour cell suspensions without obvious contamination with non-malignant cells were kept in culture conditions for 16 h and their supernatants were added to CCL-64 cells. The proliferation of these cells is inhibited by TGFbeta. According to this assay, the supernatants contained both active and latent TGFbeta. In addition, the supernatants were found to suppress the spontaneous cytotoxic function and activation of T-cell-enriched lymphocyte populations. A specific monoclonal antibody (mAb) counteracted these effects and therefore we concluded that they were mediated to a large extent by TGFbeta. In line with the results obtained with the supernatants, activation of lymphocytes could also be inhibited by tumour cells and their inhibitory effect was weaker in the presence of the TGFbeta-specific mAb. It is important to note that, when TGFbeta-specific mAb was added to autologous mixed lymphocyte/tumour cell cultures, lymphocyte activation occurred more often. These results thus substantiate the assumption that production of TGFbeta may help the survival of potentially immunogenic tumour cells in immunocompetent patients.

Interferon-gamma and tumor necrosis factor-alpha treatment of ex vivo human carcinoma cells potentiates their interaction with allogeneic lymphocytes.

Short-term exposure of ex vivo carcinoma and sarcoma cells to IFN-gamma and TNF-alpha induced or elevated to detectable levels the surface expression of MHC class I, class II, and ICAM-1 (CD54), but only rarely the B7 (CD80) molecules. The cytokine-treated tumor cells interacted more efficiently with allogeneic blood lymphocytes collected from healthy donors compared with untreated cells. This was demonstrated (1) by the induction of DNA synthesis and generation of cytotoxic activity in mixed cultures and (2) by the elevated susceptibility to the cytotoxic effectors. Although the cytokine-induced increase in MHC and ICAM-1 on the low-expressor tumors were probably important to the interaction with lymphocytes, it is likely that other properties were also induced that contributed to the phenomenon. This was indicated by the results obtained with several tumors that expressed indigenously high levels of these molecules but reacted with the allogeneic lymphocytes only or more efficiently after treatment with IFN-gamma and TNF-alpha. In these experiments B7 expression did not influence the efficiency of interactions between lymphocyte and tumor cells. The results also showed that, under the conditions used, the untreated tumor cells that did not activate allogeneic lymphocytes were sensitive to appropriately activated effectors. Thus the afferent and efferent arms of lymphocyte-tumor cell interactions appeared to have different requirements.

Detection of polypeptides associated with the histopathological differentiation of primary lung carcinoma.

Two-dimensional polyacrylamide gel electrophoresis combined with a non-enzymatic sample preparation technique is useful for analysing clinical tumour material. Using these techniques, we analysed the relationship between the histopathological findings in primary lung malignancies and the expression of a number of unidentified polypeptides that were detected in the molecular weight region 20-35 kDa. In this study 45 cases of primary lung cancer (PLC) (21 cases of adenocarcinoma, ten cases of squamous cell carcinoma, five cases of large-cell carcinoma, one case of adenosquamous cell carcinoma, five cases of small-cell carcinoma and three cases of carcinoid tumour) were examined. For reference, a human diploid fibroblast cell line (W138) and normal peripheral lymphocytes were used. Sixteen polypeptides were judged to be associated with histopathological features. These polypeptides seem to be valuable as differentiation markers. The simultaneous evaluation of these polypeptides and some other proliferation markers (e.g. PCNA, PCNA 'satellite', Numatin/protein B23 and lamin B) seems to clarify the characteristics of each case of PLC. Furthermore, it is possible to classify PLC based on the two-dimensional electrophoresis findings, and this classification of PLC is suggested to reflect the biological features of the tumour more precisely than that based only on morphology.

Immunogenicity and immunosensitivity of ex vivo human carcinomas: interferon gamma and tumour necrosis factor alpha treatment of tumour cells potentiates their interaction with autologous blood lymphocytes.

Human carcinoma cells vary appreciably in the expression of MHC class I, class II, ICAM-1 (CD54) and B7 (CD80) molecules. Short-term in vitro exposure of ex vivo carcinoma cells to interferon gamma and tumour necrosis factor alpha elevated/induced the surface expression of MHC class I, class II and ICAM-1, but only rarely of B7. We found that cytokine treatment elevated the cytotoxic susceptibility and the stimulatory potential of ex vivo tumour cells. This was demonstrated (a) by the increased frequency and elevated level of auto-tumour lysis and (b) by induction of DNA synthesis and generation of cytotoxic lymphocytes in autologous mixed lymphocyte/tumour cell culture (MLTC). The MHC class I and ICAM-1 molecules on the tumour cells were required for interaction with the lymphocytes as indicated by the inhibitory effect of specific mAb both in the stimulation and in the cytotoxic tests. While the cytokine-induced increases in MHC and ICAM-1 on the low-expression tumours were probably important for the modification of functional interaction with the autologous lymphocytes, it is likely that alterations in other properties of tumour cells were also induced which contributed to the phenomenon. This was indicated by the results obtained with several tumours, which expressed indigenously high levels of these molecules but activated the autologous lymphocytes only after cytokine treatment. In several experiments the untreated targets that did not activate the lymphocytes were sensitive to the cytotoxicity of the effectors activated in MLTC. The results show that the afferent and efferent arms of the immune response have different requirements for functional interactions between lymphocytes and tumour cells.

Cytotoxic susceptibility and defective MHC class I expression do not correlate with mutation of p53 in human carcinomas.

Twenty-nine samples of ex vivo ovarian and lung carcinomas were investigated for the relationship between the presence of mutated protein 53 (mp53) and cytotoxic susceptibility. Unaltered expression of MHC class I alleles was required for the cytotoxic susceptibility of tumour cells to the autologous ex vivo blood lymphocytes, i.e. all 4 sensitive tumours belonged to the group of 11 tumours without defect in MHC class I expression. In contrast, the susceptibility did not correlate with the presence of mp53, i.e. cases with mp53 were randomly distributed between the sensitive and resistant tumours (2/4 and 10/17 respectively). There was no correlation either between the p53 mutation and down-regulation of MHC class I alleles. The results suggest that in these tumours the mutated p53 is not the source of immunogenic peptides and that the lack of recognition of the tumours with mp53 is not caused by a defect in the expression of MHC class I molecules.

Increased expression of MHC class I molecules on human cells after short time IFN-gamma treatment.

Human cell lines and blood lymphocytes were treated for short time periods with IFN-gamma. This treatment increased the amount of the assembled MHC class I molecules on the plasma membrane after 30 min. This early increase of the membrane expression subsided in the next few hours. A second wave of elevation occurred after 8-24 hr. Analysis of cytoplasmic and membrane molecules in pulse chase experiments showed that the cytokine enhanced both the assembly of available heavy and light chains and the transport of the complex to the plasma membrane. The membrane level of the HLA-A2 molecules showed similar kinetics. Addition of an A2 specific binding peptide stabilized the IFN-gamma induced molecules on the cell surface. It seems that IFN-gamma alone or together with a binding peptide can influence MHC class I expression solely through post-transcriptional events utilizing an available pool of free heavy and light chains already after a short time, before the enhancement of the synthesis starts.

Selectively down-regulated expression of major histocompatibility complex class I alleles in human solid tumors.

We studied the expression of major histocompatibility complex (MHC) class I molecule in 52 ex-vivo tumor samples comprising 29 ovarian, 15 lung, 1 breast, and 4 colon carcinomas; 1 midgut carcinoid; and 2 malignant mesenchymal tumors obtained from surgical specimens or from malignant effusions. The allelic products were visualized in untreated and interferon gamma + tumor necrosis factor alpha treated aliquots of tumor cells and in the patient's blood lymphocytes by the one-dimensional isoelectric focusing method. Generally, the tumor cells contained lower amounts of MHC class I molecules compared to the lymphocytes. In vitro exposure to interferon gamma and tumor necrosis factor alpha elevated the level of MHC class I expression in 24 of 52 tumors and corrected the assembly defect seen in 2 cases. In 20 tumors one or several human leukocyte antigen alleles were undetectable even after cytokine treatment. Correlation was seen with the grade of differentiation; the proportion of tumors with selective losses in poorly, moderately, or well differentiated tumors were 16 of 30, 3 of 13, and 1 of 9, respectively. Selective losses occurred in ovarian carcinoma cells collected from malignant effusions (12 of 22, 54%) but not in 7 primary tumors. In primary lung carcinomas the frequency was 36% (5 of 14 cases). Thirty-nine patients were serologically typed; thus the MHC alleles on the tumor cells could be identified. In this panel of tumors 30 human leukocyte antigen alleles were represented. Among these the expression of 15 was found to be down-regulated in some but not all tumors of the same histology.

HLA-B5-restricted auto-tumor-specific cytotoxic T cells generated in mixed lymphocyte-tumor-cell culture.

T-cell-enriched lymphocyte populations derived from the malignant exudate of a patient with ovarian carcinoma were exposed to autologous tumor cells in the mixed lymphocyte-tumor-cell culture (MLTC) and propagated for 42 days. Proliferation of lymphocytes depended on exposures to autologous tumor cells and on the presence of IL-2. After 7 days, the MLTC-lymphocytes lysed K562 and the autologous tumor cells. The latter effect was not inhibited by monoclonal antibodies (MAbs) reactive with MHC class-I antigens or with CD3. After 7 restimulations, the culture was enriched in CD8+ cells (92%) and showed selective lytic activity against the autologous tumor. This function was inhibited by the alpha-class I or alpha-CD3 MAbs, and also by antibodies reactive with the HLA B locus or B5 allele products. The antibodies reactive with HLA A molecules had no such effect. It seems therefore that the function of the CTLs was restricted by HLA B5. Analysis of the TCR beta genes indicated clonal T-cell expansion in this culture. This MLTC was 1 of 21 initiated with 11 blood- and 10 tumor-derived lymphocyte (TIL) populations prepared from the malignant effusions of ovarian carcinoma patients. None of these ex-vivo lymphocytes lysed autologous tumor cells. In 17 MLTCs the lymphocytes did not proliferate, and in 3 cultures the proliferation was maintained only for 2-3 weeks. In 3 of 4 cultures auto-tumor cytotoxicity was induced.

MHC class-I-restricted auto-tumor-specific CD4+CD8- T-cell clones established from autologous mixed lymphocyte-tumor-cell culture (MLTC).

Autologous mixed lymphocyte-tumor cell cultures (MLTC) were initiated with cytokine (IFN gamma and TNF alpha)-treated ex-vivo tumor cells of lung, ovarian, breast and stomach carcinomas. The cytokine-treated tumors expressed class-I but not class-II molecules. Although the proportion of CD8+ lymphocytes increased in the bulk culture of MLTCs, in 5/7 experiments the majority of the established T-cell clones were CD4+. Among the CD8+ clones a high proportion (77%) was cytotoxic, while the proliferative response was more frequent among the CD4+ clones (70%). In 4/26 cytotoxic T-lymphocyte (CTL) clones (3/17 CD4+ and 1/9 CD8+), derived from a patient with class I+ class II- stomach carcinoma, lysis was restricted to the autologous tumor cells. These auto-tumor-specific clones did not lyse the autologous ConA blasts, the 5 allogeneic ex-vivo tumors, the NK-sensitive K562 or the relatively sensitive Daudi cells. The cytotoxicity of these clones was inhibited by pre-incubation of the tumor cells with W6/32 (alpha-class I) MAb, or by preincubation of the lymphocytes with OKT3 (alpha-CD3) MAb. The alpha-CD4 (OKT4) MAb had only a marginal effect on the CD4+ clones, while the lytic function of the CD8+ clone was inhibited by the alpha-CD8 (OKT8) MAb. The 3 CD4+ CTL clones also responded with proliferation to the autologous tumor cells. This proliferative response was inhibited by the presence of W6/32 MAb. Our results indicate that the auto-tumor lysis exerted by CD4+ CTL clones was restricted by the class-I antigens, and that the CD4 molecules of the clones were not essential for the lytic interaction.