Upregulation of DF3, in association with ICAM-1 and MHC class II by IFN-gamma in short-term human mammary carcinoma cell cultures

This study was designed to determine whether in vitro exposure of isolated short-term human primary and metastatic breast tumor cell cultures to interferon-gamma (IFN-gamma) could enhance expression of the breast tumor associated DF3 antigen in association with the intercellular adhesion molecule 1 (ICAM-1) and MHC class II molecules. Cell cultures were established from primary solid tumors and metastatic cells as previously described (Sgagias et al., 1995). Data show that recombinant human IFN-gamma treatment, in vitro, dramatically increased the breast tumor associated DF3 antigen, in association with ICAM-1, and MHC class II antigens in primary breast cancer cell cultures. All primary breast tumor cell cultures constitutively expressed high levels of HLA-class I antigen. Metastatic breast cancer cell cultures expressed high levels of DF3 and recombinant human IFN-gamma treatment, in vitro, upregulated ICAM-1 and MHC class II antigens before and after passage of the metastatic cells through the nude mouse. Metastatic breast cancer cells similar to primary breast cancer cells constitutively expressed high levels of MHC class I antigens. In addition, three LAK cell lines significantly lysed the primary and the metastatic breast tumor cell cultures to the same degree before and after passage of the metastatic cancer cells through the nude mouse. These data indicate the upregulation of the breast tumor associated DF3 antigen in vitro after IFN-gamma treatment and its persistence in vivo, after passage of the metastatic breast cancer cells through the nude mouse. The ability of IFN-gamma to upregulate the breast tumor associated DF3 antigen in association with the ICAM-1 and HLA class II antigens may play an important role in eliciting an immune response which may contribute to the immunodiagnosis, and immunotherapy of breast cancer.     

肿瘤细胞HLAⅠ类分子表达对NK抗性的影响及IFN-γ的调节作用

目的 探讨肿瘤细胞表面HLAⅠ类分子的表达与NK杀伤的关系及IFN－γ的调节作用。方法 用间接免疫荧光法检测7种肿瘤细胞表面HLA－ABC分子的表达;用鼠抗人HLA－ABC单抗封闭靶细胞后,观察NK杀伤的变化;用IFN-γ处理靶细胞后,观察瘤细胞表面HLA分子表达的变化及NK杀伤的变化。结果 不同肿瘤细胞株表达HLA－ABC分子的比例及荧光强度均不相同,除Karpas外,均表现为不同程度的下降。HLA－ABC表达阴性的K562细胞对NK杀伤最敏感,其他细胞的敏感性均有下降。肿瘤细胞表达HLA－ABC分子的表达多有不同程度的下降。用抗HLA单抗封闭相应位 点后,可使NK杀伤明显增强。用IFN－γ500U／ml处理48h后,白血病细胞K562、黑色素瘤细胞M21和高转移肺癌细胞PG表面HLAⅠ类分子表达明显增加,对NK杀伤的敏感性降低;而人T细胞淋巴瘤Karpas、髓样白血病细胞HL60和结直肠癌细胞HT29经处理后,HLAⅠ类分子表达无明显变化,对NK杀伤的敏感性反而明显增强。IFN－γ促进HL60和HT29细胞的凋亡。结论 NK细胞能识别HLA－ABC分子表达下降或缺陷的肿瘤细胞并发挥杀伤作用;INF－γ能恢复部分肿瘤细胞HLA－ABC分子的丢失,并能促进部分肿瘤细胞的凋亡,提高肿瘤对机体抗瘤机制的敏感性。     

Tumor aggressiveness and MHC class I and II antigens in laryngeal and breast cancer

HLA class I and II molecules play an important role in the immune system. Thus, alterations in their expression could affect the immunosurveillance against tumors. The expression of these molecules may therefore be one of the factors responsible for oncogenicity, due to their role as restriction elements in T cell recognition. A series of 78 primary laryngeal and 94 breast tumors were evaluated for the expression of MHC class I and II antigens, using monoclonal antibodies. We found a number of cases presenting total loss of HLA-ABC and selective losses of HLA-A and HLA-B antigens. These losses were statistically associated with clinical and pathological parameters such as T stage, degree of differentiation, scores according to the malignancy grading systems and degree of leukocytic infiltration. On the other hand, the de novo expression of DR antigens in laryngeal cancer was found in a group of neoplasms having a peculiar clinical behavior and excellent prognosis. Expression of class II molecules in breast tumors was also related to a better differentiation and prognosis. Thus, HLA class I and II antigen expression is closely related to aggressiveness and prognosis of laryngeal and breast cancer.     

Intensity of class I antigen expression on human tumour cell lines and its relevance to the efficiency of non-MHC-restricted killing.

A modified tetrazolium reduction assay (MTT) was used to assess the relation between HLA class I antigen expression on tumour cells and their susceptibility as a target for non-MHC restricted LAK/NK cytotoxicity using interleukin-2 activated peripheral blood mononuclear cells (MNC) from normal individuals. At 20/1 effector/target ratio this ranged from no killing to 77%. The efficiency of killing was dependent on duration of effector cell culture with IL-2, peaking at day 10 and declining thereafter. This killing could be enhanced by addition of other cytokines including interferons alpha, beta and gamma. Study of a panel of 15 tumour cell lines using a single effector showed that there was no statistically significant inverse correlation (using Spearman rank test) between the degree of tumour class I expression and LAK/NK killing at 20/1 (r = 0.23 P = 0.39) and 10/1 (r = 0.30, P = 0.27) and at 5/1 E/T ratio r = 0.47, P = 0.08) respectively. Lack of inverse correlation between these two parameters came from study of one bladder tumour line (FEN), whose absent class I antigens had been corrected by transfection with beta 2 microglobulin gene. At high E/T ratio (20/1) there was an increase in the susceptibility of target cells to lysis (36% parent cell, 45% transfected cell), whilst at lower E/T ratios (1/1) there was significantly more killing of the non-transfected cells (10% vs 31%). The addition of anti-class I antibody W6/32 increased killing by 18% but this was non-specific as the same increase occurred with a class II antibody. These data suggest that overall there was not an inverse correlation between class I expression and LAK/NK killing at high E/T ratios, whilst at low (5/1 or lower) E/T ratios this correlation nearly reached statistical significance suggesting that the conflicting literature reports may be due to a threshold levels of effector cells above which the masking effects of MHC antigens disappears.     

Multiple defects of the antigen-processing machinery components in human neuroblastoma: immunotherapeutic implications

Low expression of human leukocyte antigen (HLA) class I in human tumors may be related to defects of the antigen-processing machinery (APM) components. Neuroblastoma cells are virtually HLA class I negative, but (i) the underlying mechanisms are unknown, and (ii) expression of the APM components has never been investigated. Here we have used a panel of novel monoclonal antibodies to proteasomal and immunoproteasomal components, chaperons and transporter associated with antigen processing (TAP) to characterize 24 stroma-poor neuroblastoma tumors and six neuroblastoma cell lines. Primary tumors showed defects in the expression of zeta, tapasin, TAP1 or TAP2, HLA class I heavy chain and beta2 microglobulin, LMP2 and LMP7, as compared to normal adrenal medulla. Neuroblastoma cell lines displayed roughly similar patterns of APM expression in comparison to primary tumors. Incubation of neuroblastoma cell lines with interferon-gamma caused upregulation of HLA class I molecules and reduced lysis by killer inhibitory receptor HLA ligand-matched NK cells. Defects in APM components explain reduced peptide loading on HLA class I molecules, their instability and failure to be expressed on the cell surface. HLA class I upregulation by interferon-gamma, although enhancing neuroblastoma cell recognition by cytotoxic T cells, dampens their susceptibility to NK cells.     

Retaliation against tumor cells showing aberrant HLA expression using lymphokine activated killer-derived T cells.

Lymphokine activated killer (LAK) cells represent mixtures of natural killer (NK) and non-MHC-restricted CTLs that have the capacity to lyse a variety of tumor cells and MHC class I-negative target cells. Although it is clear that NK cells are negatively regulated by interactions with MHC class Ia or class Ib molecules, the regulation of LAK-derived T cells has not been clarified to date. In the studies presented here, we demonstrate that IFN-gamma treatment of tumor cells can induce their resistance to LAK-derived T cells in a manner similar to that seen for NK cells. The IFN-gamma-mediated suppression of LAK activity correlates with increased MHC class I expression by the tumor cells, and the inhibition of LAK-mediated cytotoxicity can be reversed in the presence of class I-specific antibody. Furthermore, the expression of MHC class Ia or class Ib molecules in class I-negative cell lines can reduce their susceptibility to LAK-mediated cytotoxicity. This principle of negative regulation by MHC class I molecules applies to LAK-derived T cells generated from peripheral blood lymphocytes of renal cell carcinoma patients and healthy, control donors. Although LAK-derived T cells can be inhibited in their lytic activity through interactions with MHC class Ia and class Ib molecules, they do not express the known inhibitory receptors specific for these ligands that are found on NK cells. Apparently, LAK-derived T cells are negatively regulated by as yet undefined inhibitory receptors.     

Comparative assessment of MHC antigen expression in bladder and testis tumour biopsies and established urological tumour cell lines: the relevance of cytokine and gene transfection for correction of defective MHC antigens

The aim of this study using radio-binding (RB), peroxidase-anti-peroxidase (PAP) and immunoprecipitation (IP) techniques was to investigate the pattern of major histocompatibility complex (MHC) antigen expression in urological malignancies and to compare the results with those seen in established urological human tumour cell lines. The results showed that using PAP technique, the percent bladder cases showing complete loss or cases with greater than 90% of tumour cells negative with W6/32 (detects all class I antigens), HC10 (detects free heavy chain) and BMM.1 (detects beta2-mirogobulin) monoclonal antibodies (Mab) were 16%, 44% and 2% respectively. In a subgroup of 37 cases, the intensity of MHC class II antigen expression for strong, weak and negative cases were 9 (24%), 8 (22%) and 20 (54%) respectively. The expression for class I antigens on testis tumours was mainly negative and when positive, it was present in a small percent of tumour cells. This was also observed for class II antigens where only 8% of cases showed some degree of positivity. Using RB technique, 10 of 12 (85%) of tumour lines expressed class I antigens spontaneously and following interferon gamma (IFNgamma) stimulation, the 2 negative lines one testis (Tera I) and one bladder (Fen) remained negative and 2 lines (both testis lines Tera II and Ep2102) showed a significant class I up-regulation. None of the lines expressed class II antigens spontaneously and following IFNgamma stimulation, 8 of 12 (66%) were induced. The absence of class I and II antigens in the negative lines was confirmed using IP technique. In the case of one class I negative bladder cell line i.e. Fen, the biochemical analysis showed the absence of beta2-m gene product which could not be restored by IFNgamma stimulation. However, transfection of the cells with beta2-m gene resulted in the expression of fully assembled class I antigens, indicating that the loss of antigens was due to the absence of functional beta2-m gene. These results indicated the similarity between the pattern of expression of MHC antigens on tumour biopsies and established tumour cell lines. They also demonstrated that both cytokine stimulation and gene transfection could be used to correct defective class I antigens in tumour cell lines. These approaches might have important implications for pre-selection of bladder cancer patients for cytokine or gene therapies.     

Decreased susceptibility of lined human gliosarcoma cells to lymphokine-activated killer cell cytolysis by gamma-interferon treatment

The susceptibility of the established cultured gliosarcoma line GI-1 to lymphokine-activated killer (LAK) cells was analyzed with and without interferon (IFN)-gamma treatment of target GI-1 cells. IFN-gamma treatment decreased the susceptibility of GI-1 cells to LAK cell cytolysis in a dose-dependent manner. Acid treatment of GI-1 cells increased their susceptibility to cytolysis compared with untreated cells. IFN-gamma treatment and acid treatment of GI-1 cells respectively increased and decreased the expression of class I HLA antigens on GI-1 cells. The susceptibility of GI-1 cells to LAK cell cytolysis and their expression of HLA class I molecules were inversely correlated. Subpopulation depletion experiments on the LAK cells with monoclonal antibodies and complement revealed that phenotypically natural killer type (CD16+) cells had a high cytotoxic activity against untreated GI-1 cells but a relatively low activity against IFN-gamma-treated GI-1 cells in both the precursor and effector phases. On the other hand, phenotypically T-type (CD3+) cells did not show these tendencies at all in both the precursor and the effector phases.     

反义IGF-I基因对人7402肝癌细胞肿瘤免疫原性影响的研究

本研究利用基因治疗方法中的反义技术,通过脂质体包裹反义胰岛素样生长因子-I(IGF-I)基因导入人7402肝癌细胞株,研究肿瘤细胞表面MHC分子的表达和其对LAK细胞杀伤的敏感性。结果表明,反义IGF-I基因转染的人7402肝癌细胞,经潮霉素筛选后,Northern Blot分析显示反义IGF-I RNA表达阳性,MHC Ⅰ类和Ⅱ类抗原的表达水平高于亲本7402细胞。LAK细胞对反义IGF-I基因转染的人7402肝癌细胞的杀伤活性明显高于亲本7402细胞组(P<0.05)。说明反义IGF-I基因转移至该肝癌细胞,可提高肿瘤的MHC Ⅰ类和Ⅱ类分子的表达,显示肿瘤细胞的免疫原性增强,并可提高LAK细胞对其杀伤的敏感性。     

Renal cell carcinoma-infiltrating natural killer cells express differential repertoires of activating and inhibitory receptors and are inhibited by specific HLA class I allotypes

Among tumor-infiltrating lymphocytes (TILs) directly isolated from renal cell carcinomas (RCCs), we found substantial numbers of natural killer (NK) cells in most tumor tissues. They could be identified reliably in situ with an antibody directed against the activating receptor (AR) NKp46 that is exclusively expressed by all NK cells. NK-enriched TILs (NK-TILs) showed cytotoxicity against major histocompatibility complex (MHC) class I-negative cell lines. The ability to detect lysis of target cells was dependent on the percentage of NK cells within the TILs, and cytotoxicity was only observed after overnight activation with low-dose interleukin-2 (IL-2). Infiltrating NK cells were found to express various inhibitory receptors (IRs); among these the CD94/NKG2A receptor complex was overrepresented compared to the autologous peripheral blood mononuclear cell (PBMC) population. Other IRs were underrepresented, indicating that NK subpopulations vary in their tumor-infiltrating capacity. IRs expressed by NK-TILs are functional since receptor engagement with MHC class I ligands presented by human leukocyte antigen (HLA)-transfected target cell lines was able to inhibit NK-mediated cytotoxicity. NK-TILs were also able to lyse autologous or allogeneic tumor cell lines in vitro. This activity correlated with low HLA class I surface expression since lysis could be inhibited by interferon (IFN)-gamma-expressing RCC transductants that displayed a higher surface density of HLA class I molecules. Therefore, NK cells infiltrating tumor tissues have an inherent ability to recognize transformed cells, but they require cytokine activation and are sensitive to inhibition by IR ligands.     

Flow cytometric investigation of immune-response-related surface molecules on human colorectal cancers

Our purpose was to clarify whether human colorectal cancer cells are equipped to present tumour-associated-antigens to the immune system, and whether this ability correlates with lymphoid infiltration, the Dukes' stage and Jass classification. Enzymatically dissociated tumour cells from 70 different colorectal cancers were monitored by multiparameter flow cytometry. Gating on EP4⁺ cells, the expression of the surface molecules HLA class I, HLA class II, CD80 (B7-1), CD54 (ICAM-1) and CD58 (LFA-3) was evaluated. In 60 of 70 tumours, all tumour cells expressed HLA class I, in 10 tumours 15-96% of the tumour cells expressed HLA class I. In 1 tumour, all tumour cells expressed HLA class II, in 67 tumours some expressed HLA class II, in 2 tumours none expressed HLA class II. Expression of CD58 was heterogeneous, and there was no or only sparse expression of CD80 and CD54. Expression of the HLA class I molecules, but not the class II, was correlated with lymphoid infiltration and the Jass classification. Expression of these surface molecules was not correlated with the Dukes' stage. The tumour cells were generally equipped to present antigens to the effector arm of the immune system since HLA class I is expressed, but the tumour cells were not optimal in stimulating an immune response, since HLA class II and CD58 were only marginally expressed and CD80 and CD54 were absent. Int. J. Cancer (Pred. Oncol.) 79:283–287, 1998.© 1998 Wiley-Liss, Inc.     

An immunohistological study of HLA antigen expression by gestational choriocarcinoma

Two cases of gestational choriocarcinoma have been examined for the expression of HLA and trophoblast antigens using the indirect immunoperoxidase technique on frozen tissue sections. Approximately 40-70% of tumour cells express MHC Class I antigen as detected by a panel of antibodies to monomorphic, or framework, MHC Class I antigenic determinants. Evidence in one case suggests that most of these cells may not express the paternal, polymorphic HLA antigenic determinants but that a small subpopulation do carry the fully antigenically active molecule. These latter may give rise to patient anti-paternal HLA antibodies. Class II (HLA DR or DC) antigens are expressed by none, or very few, tumour cells.     

HLA antigens in colorectal tumours--low expression of HLA class I antigens in mucinous colorectal carcinomas

Expression of HLA antigens and beta 2-microglobulin was studied by immunoperoxidase staining of frozen sections of 9 mucinous and 10 nonmucinous colorectal adenocarcinomas, 1 cloacogenic carcinoma, 12 colorectal adenomas and 4 samples of normal colorectal mucosa using monoclonal antibodies (MAbs). Staining results were related to histopathological features. HLA Class I antigens were strongly expressed in morphologically normal colorectal epithelium, in all adenomas tested and in all non-mucinous carcinomas. In contrast, expression of HLA class I antigens by the majority of tumour cells was present in only 2 of the 9 mucinous carcinomas, whereas 2 of these mucinous carcinomas were completely negative. In the mucinous carcinomas a striking scarcity of mononuclear inflammatory infiltrate, especially around the mucus accumulations, was observed. HLA class II antigen expression was not detected in normal epithelium and was only focally present in 1 of the 12 adenomas. In 6 out of the 20 carcinomas tested between 20% and 90% of the tumour cells were stained by MAbs against HLA class II antigens. Apart from the low expression of HLA class I antigens in mucinous carcinomas no relationship was found between expression of HLA antigens and histological features of the tumours. The relative poor prognosis of mucinous colorectal carcinoma as reported in the literature may be associated with low expression of HLA class I antigens and scant mononuclear inflammatory infiltrate, which may be a reflection of a weak immune response to the tumour cells.     

Increased susceptibility of IFN-gamma-treated neuroblastoma cells to lysis by lymphokine-activated killer cells: participation of ICAM-1 induction on target cells.

We have investigated the effect of interferon-gamma (IFN-gamma) treatment of neuroblastoma cells on the susceptibility to lysis by lymphokine-activated killer (LAK) cells and examined the participation of cell-adhesion molecules on the target cells in LAK cell lysis. Untreated neuroblastoma cells expressed lymphocyte-function-associated antigen 3 (LFA-3) and neural-cell-adhesion molecule (NCAM), but did not express MHC-class-I, MHC-class-II, or intercellular-adhesion molecule I (ICAM-I). IFN-gamma treatment of neuroblastoma cells induced the expression of MHC-class-I and ICAM-I antigens, but did not affect the expression of MHC-class-II, LFA-3, and NCAM. This was accompanied by an increased susceptibility to lysis by LAK cells. Anti-ICAM-I antibody inhibited partially the increased sensitivity of IFN-gamma-treated neuroblastoma cells to LAK cell lysis, and blocked completely the increase in binding of LAK cells observed after IFN-gamma treatment of the target cells. These results suggest that the increased LAK sensitivity of IFN-gamma-treated neuroblastoma cells is partially attributable to the induction of ICAM-I on neuroblastoma cells and indicate that post-binding events also play a role in the increased sensitivity to LAK cell lysis observed after IFN-gamma treatment.     

Flow cytometric analysis of tumour infiltrating lymphocyte activation and tumour cell MHC class I and II expression in breast cancer patients.

The primary tumour cells and tumour infiltrating lymphocytes (TILs) of 31 breast cancer patients have been analysed by dual colour flow cytometry to determine whether the phenotype and/or activation status of the TILs bears any relationship to the expression of MHC antigens on the tumour cells. The phenotype and activation status of 5000 TILs were studied using Mabs to CD4, CD8, HLA DR, CD25 (the low affinity inducible IL-2 receptor) and the transferrin receptor and related to Class I and Class II MHC expression on 5000 primary tumour cells. On the tumour cells, Class I MHC expression ranged from 1-74%, averaging 12.9%. HLA DR expression ranged from 1-69% averaging 14.3%. When the phenotypic proportions of the lymphocytic infiltrate were analysed there was found to be a correlation between tumour expression of Class I MHC and the proportion of both CD4+ (P less than 0.05) and CD8+ (P less than 0.02) T cells within the tumour. No such relationship was found with the MHC Class II antigen. When TIL activation markers were analysed, the percentage of CD8+ TILs positive for HLA DR expression correlated strongly with the expression of Class I (P less than 0.001) and Class II (P less than 0.001) antigens on the tumour cells. The percentage of CD4+ TILs positive for HLA DR expression also correlated significantly, but less strongly with the expression of Class I (P less than 0.01) and Class II (P less than 0.02) antigen expression on the tumour cells. The percentage of CD4+ TILs positive for CD25 expression correlated with both Class I (P less than 0.05) and Class II (P less than 0.03) expression on the tumour cells while the percentage of CD8+ TILs positive for CD25 did not. The percentage of TILs bearing the transferrin receptor showed no measurable correlation with the expression of either class of MHC antigen on the tumour. The data suggest that MHC expression on the tumour cells has a selective effect on the response capacity of different parts of the immune system.     

Low-dose IFN-gamma induces tumor MHC expression in metastatic malignant melanoma.

Specific antitumor immune responses require expression of MHC class I or II molecules on tumor cells, and MHC antigen down-regulation is a presumed tumor growth promoting mechanism. Because IFN-gamma up-regulates tumor MHC antigen expression in vitro, in this Phase II trial of an immunologically active dose and schedule we evaluated whether this was the case in vivo. Twenty-three patients with metastatic melanoma were treated with IFN-gamma 100 microg/m(2) s.c. once weekly for a maximum of 6 months. There were three complete responses, now maintained for 53, 36, and 25 months. The remainder had progressive disease. The treatment was well tolerated, with no toxicity exceeding National Cancer Institute Common Toxicity Criteria grade II. Immunohistochemical analysis of tumor biopsies during treatment was performed using monoclonal antibodies to HLA class I (W/632) and class II (CR3/43) monomorphic determinants. HLA class I was down-regulated in 2 of 19 patients pretreatment and up-regulated by IFN-gamma in both. HLA class II was down-regulated pretreatment in 14 of 18 patients and up-regulated by IFN-gamma in 6 (43%). The HLA up-regulation persisted throughout the study. IFN-gamma induced significant but short-lived up-regulation of surrogate markers of monocyte activation (serum neopterin) and class I up-regulation (serum beta-2-microglobulin) in most patients. There was no consistent relationship between surrogate marker up-regulation, tumor antigen up-regulation, and responses. The study shows that the significant immune modulation induced by IFN-gamma does not correlate with tumor responses and that the serum surrogate marker changes do not reflect tumor events. The durable and long-lived responses, clear demonstration of tumor MHC up-regulation, and low toxicity suggest that weekly IFN-gamma 100 microg/m(2) would be a useful addition to chemoimmunotherapeutic regimens.