Antibody targeting of CD24 efficiently retards growth and influences cytokine milieu in experimental carcinomas.

BACKGROUND: The targeting of cancer stem cells by monoclonal antibodies offers new options for therapy. CD24 is a glycosylphosphatidylinositol-anchored membrane protein with a small protein core and a high level of glycosylation. It is overexpressed in many human carcinomas and is correlated with poor prognosis. CD24 is a marker for pancreatic and ovarian cancer stem cells, whereas breast cancer stem cells are negative for CD24. In cancer cell lines, changes of CD24 expression can alter cellular properties in vitro and tumour growth in vivo. We have shown before that monotherapy with monoclonal antibody (mAb) SWA11 to CD24 effectively retarded tumour growth in xenotransplanted mice. METHODS: Here, we have investigated in more detail the molecular mechanisms of mAb SWA11 therapeutic effects in A549 lung and SKOV3ip ovarian carcinoma models in scid/beige and CD1 mice, respectively. We focused on anti-proliferative, pro-apoptotic, anti-angiogenic and microenvironmental effects of SWA11 mAb treatment. RESULTS: We find that CD24 targeting is associated with changes in tumour cell proliferation and angiogenesis. The treatment lead to increased infiltration of tumour tissues with immune cells suggesting involvement of ADCC. We found that SWA11 mAb treatment strongly altered the intratumoural cytokine microenvironment. The addition of SWA11 mAb to gemcitabine treatment strongly potentiated its anti-cancer efficacy in A549 lung cancer model. CONCLUSION: Our data demonstrate that targeting of CD24 could be beneficial for the anti-cancer treatment combined with standard chemotherapy regimes.

Prognostic significance of L1CAM in ovarian cancer and its role in constitutive NF-κB activation.

BACKGROUND: Overexpression of L1-cell adhesion molecule (L1CAM) has been observed for various carcinomas and correlates with poor prognosis and late-stage disease. In vitro, L1CAM enhances proliferation, cell migration, adhesion and chemoresistance. We tested L1CAM and interleukin-1 beta (IL-1ß) expression in tumor samples and ascitic fluid from ovarian carcinoma patients to examine its role as a prognostic marker. PATIENTS AND METHODS: We investigated tumor samples and ascitic fluid from 232 serous ovarian carcinoma patients for L1CAM by enzyme-linked immunosorbent assay. L1CAM expression was correlated with pathoclinical parameters and patients' outcome. IL-1ß levels were measured in tumor cell lysates. Ovarian cancer cell lines were analyzed for the contribution of L1CAM to IL-1ß production and nuclear factor 'kappa-light-chain-enhancer' of activated B-cells (NF-κB) activation. RESULTS: We observed that L1CAM-expressing tumors show a highly invasive phenotype associated with restricted tumor resectability at primary debulking surgery and increased lymphogenic spread. Soluble L1CAM proved to be a marker for poor progression-free survival and chemoresistance. In ovarian carcinoma cell lines, the specific knock-down of L1CAM reduces IL-1ß expression and NF-κB activity. CONCLUSIONS: L1CAM expression contributes to the invasive and metastatic phenotype of serous ovarian carcinoma. L1CAM expression and shedding in the tumor microenvironment could contribute to enhanced invasion and tumor progression through increased IL-1ß production and NF-κB activation.

Metastatic potential severely altered by changes in tumor cell adhesiveness and cell-surface sialylation.

A plastic adherent variant line (ESb-M) of a highly invasive and metastatic murine T cell lymphoma (ESb) was found to have lost its metastatic potential while still being tumorigenic in normal syngeneic hosts. The variant retained most of its ESb-derived antigenic and biochemical characteristics but differed at binding sites for certain lectins with specificity for terminal N-acetylgalactosamine residues. Whereas such sites were masked by sialic acid on metastatic ESb cells, they became unmasked on the adherent variant line. Metastatic revertants of ESb-M cells did not express the respective lectin receptor sites because these were again masked by sialic acid. It is suggested that the masking of specific lectin receptors sites on the tumor cell surface is of crucial importance for metastatis. If freely exposed, these sites may change adherence characteristics of the cells possibly not only in vitro (to plastic) but also in vivo.

Anti-CD2 antibodies induce T cell unresponsiveness in vivo.

The CD2 receptor functions as an adhesion and signal molecule in T cell recognition. Multimeric binding of CD2 on T cells to its physiologic ligand LFA-3 on cognate partner cells in vitro efficiently augments the antigen-specific T cell signal delivered by the T cell receptor/CD3 complex. The precise contribution of the antigen-nonspecific CD2-LFA-3 interactions to T cell immune responses in vivo, however, has been difficult to assess. Here we analyzed the role of CD2 in the murine immune response using a nondepleting anti-CD2 monoclonal antibody that induces a marked, reversible modulation of CD2 expression on murine T and B cells in situ. This modulation is dose and time dependent, specific for CD2, and does not require the Fc portion of the antibody. Anti-CD2 antibodies [rat IgG1 or F(ab')2] significantly inhibit the CD4+ T cell-mediated response to hen egg lysozyme and the cytotoxic CD8+ T cell response to a syngeneic tumor cell line. In both cases, anti-CD2 antibodies are only effective when administered before or within 24 h after antigen priming. The suppression of the antitumor response corresponds to a sixfold reduction of specific cytotoxic T lymphocyte precursor cells and results in the abrogation of protective antitumor immunity. Anti-CD2 antibodies also affect the humoral immune response to oxazolone: the isotype switch from specific IgM to IgG1 antibodies is delayed, whereas the IgM response is unaltered. In addition, a single antibody injection results in sustained polyclonal unresponsiveness of T cells irrespective of antigen priming and CD2 modulation. These results document that CD2-mediated signals induce a state of T cell unresponsiveness in vivo.

The alpha 4 integrin chain is a ligand for alpha 4 beta 7 and alpha 4 beta 1.

The heterodimeric alpha 4 integrins alpha 4 beta 7 lymphocyte Peyer's patch adhesion molecule ([LPAM]-1) and alpha 4 beta 1 (very late antigen-4) are cell surface adhesion molecules involved in lymphocyte trafficking and lymphocyte-cell and matrix interactions. Known cellular ligands include vascular cell adhesion molecule (VCAM)-1, which binds to alpha 4 beta 1 and alpha 4 beta 7, and the mucosal addressin cell adhesion molecule (MAdCAM)-1, which binds to alpha 4 beta 7. Here we show that the alpha 4 chain of these integrins can itself serve as a ligand. The alpha 4 chain, immunoaffinity purified and immobilized on glass slides, binds thymocytes and T lymphocytes. Binding exhibits divalent cation requirements and temperature sensitivity which are characteristic of integrin-mediated interactions, and is specifically inhibited by anti-alpha 4 integrin antibodies, which exert their effect at the cell surface. Cells expressing exclusively alpha 4 beta 7 (TK-1) or alpha 4 beta 1 (L1-2) both bound avidly, whereas alpha 4-negative cells did not. A soluble 34-kD alpha 4 chain fragment retained binding activity, and it inhibited lymphocyte adhesion to alpha 4 ligands. It has been shown that alpha 4 integrin binding to fibronectin involves an leucine-aspartic acid-valine (LDV) motif in the HepII/IIICS region of fibronectin (CS-1 peptide), and homologous sequences are important in binding to VCAM-1 and MAdCAM-1. Three conserved LDV motifs occur in the extracellular sequence of alpha 4. A synthetic LDV-containing alpha 4-derived oligopeptide supports alpha 4-integrin-dependent lymphocyte adhesion and blocks binding to the 34-kD alpha 4 chain fragment. Our results suggest that alpha 4 beta 7 and alpha 4 beta 1 integrins may be able to bind to the alpha 4 subunit on adjacent cells, providing a novel mechanism for alpha 4 integrin-mediated and activation-regulated lymphocyte interactions during immune responses.

Sulforaphane targets pancreatic tumour-initiating cells by NF-kappaB-induced antiapoptotic signalling.

BACKGROUND AND AIMS: Emerging evidence suggests that highly treatment-resistant tumour-initiating cells (TICs) play a central role in the pathogenesis of pancreatic cancer. Tumour necrosis factor-related apoptosis-inducing ligand (TRAIL) is considered to be a novel anticancer agent; however, recent studies have shown that many pancreatic cancer cells are resistant to apoptosis induction by TRAIL due to TRAIL-activated nuclear factor-kappaB (NF-kappaB) signalling. Several chemopreventive agents are able to inhibit NF-kappaB, and favourable results have been obtained--for example, for the broccoli compound sulforaphane-in preventing metastasis in clinical studies. The aim of the study was to identify TICs in pancreatic carcinoma for analysis of resistance mechanisms and for definition of sensitising agents. METHODS: TICs were defined by expression patterns of a CD44(+)/CD24(-), CD44(+)/CD24(+) or CD44(+)/CD133(+) phenotype and correlation to growth in immunodeficient mice, differentiation grade, clonogenic growth, sphere formation, aldehyde dehydrogenase (ALDH) activity and therapy resistance. RESULTS: Mechanistically, specific binding of transcriptionally active cRel-containing NF-kappaB complexes in TICs was observed. Sulforaphane prevented NF-kappaB binding, downregulated apoptosis inhibitors and induced apoptosis, together with prevention of clonogenicity. Gemcitabine, the chemopreventive agents resveratrol and wogonin, and the death ligand TRAIL were less effective. In a xenograft model, sulforaphane strongly blocked tumour growth and angiogenesis, while combination with TRAIL had an additive effect without obvious cytotoxicity in normal cells. Freshly isolated patient tumour cells expressing markers for TICs could be sensitised by sulforaphane for TRAIL-induced cytotoxicity. CONCLUSION: The data provide new insights into resistance mechanisms of TICs and suggest the combination of sulforaphane with TRAIL as a promising strategy for targeting of pancreatic TICs.

The neural cell adhesion molecule N-CAM enhances L1-dependent cell-cell interactions.

On neural cells, the cell adhesion molecule L1 is generally found coexpressed with N-CAM. The two molecules have been suggested, but not directly shown, to affect each other's function. To investigate the possible functional relationship between the two molecules, we have characterized the adhesive interactions between the purified molecules and between cultured cells expressing them. Latex beads were coated with purified L1 and found to aggregate slowly. N-CAM-coated beads did not aggregate, but did so after addition of heparin. Beads coated with both L1 and N-CAM aggregated better than L1-coated beads. Strongest aggregation was achieved when L1-coated beads were incubated together with beads carrying both L1 and N-CAM. In a binding assay, the complex of L1 and N-CAM bound strongly to immobilized L1, but not to the cell adhesion molecules J1 or myelin-associated glycoprotein. N-CAM alone did not bind to these glycoproteins. Cerebellar neurones adhered to and sent out processes on L1 immobilized on nitrocellulose. N-CAM was less effective as substrate. Neurones interacted most efficiently with the immobilized complex of L1 and N-CAM. They adhered to this complex even when its concentration was at least 10 times lower than the lowest concentration of L1 found to promote adhesion. The complex became adhesive for cells only when the two glycoproteins were preincubated together for approximately 30 min before their immobilization on nitrocellulose. The adhesive properties between cells that express L1 only or both L1 and N-CAM were also studied. ESb-MP cells, which are L1-positive, but N-CAM negative, aggregated slowly under low Ca2+. Their aggregation could be completely inhibited by antibodies to L1 and enhanced by addition of soluble N-CAM to the cells before aggregation. N2A cells, which are L1 and N-CAM positive aggregated well under low Ca2+. Their aggregation was partially inhibited by either L1 or N-CAM antibodies and almost completely by the combination of both antibodies. N2A and ESb-MP cells coaggregated rapidly and their interaction was similarly inhibited by L1 and N-CAM antibodies. These results indicate that L1 is involved in two types of binding mechanisms. In one type, L1 serves as its own receptor with slow binding kinetics. In the other, L1 is modulated in the presence of N-CAM on one cell (cis-binding) to form a more potent receptor complex for L1 on another cell (trans-binding).

Functional cooperation between the neural adhesion molecules L1 and N-CAM is carbohydrate dependent.

The neural cell adhesion molecules L1 and N-CAM have been suggested to interact functionally by formation of a complex between the two molecules (Kadmon, G., A. Kowitz, P. Altevogt, and M. Schachner. 1990. J. Cell Biol. 110:193-208). To determine the molecular mechanisms underlying this functional cooperation, we have studied the contribution of carbohydrates to the association of the two molecules at the cell surface. Aggregation or adhesion between L1- and N-CAM-positive neuroblastoma N2A cells was reduced when the synthesis of complex and/or hybrid glycans was modified by castanospermine. Fab fragments of polyclonal antibodies to L1 inhibited aggregation and adhesion of castanospermine-treated cells almost completely, whereas untreated cells were inhibited by approximately 50%. Fab fragments of polyclonal antibodies to N-CAM did not interfere with the interaction between castanospermine-treated cells, whereas they inhibited aggregation or adhesion of untreated cells by approximately 50%. These findings indicate that cell interactions depending both on L1 and N-CAM ("assisted homophilic" binding) can be reduced to an L1-dominated interaction ("homophilic binding"). Treatment of cells with the carbohydrate synthesis inhibitor swainsonine did not modify cell aggregation in the absence or presence of antibodies compared with untreated cells, indicating that castanospermine-sensitive, but swainsonine-insensitive glycans are involved. To investigate whether the appropriate carbohydrate composition is required for an association of L1 and N-CAM in the surface membrane (cis-interaction) or between L1 on one side and L1 and N-CAM on the other side of interacting partner cells (trans-interaction), an L1-positive lymphoid tumor cell line was coaggregated with and adhered to neuroblastoma cells in the various combinations of castanospermine-treated and untreated cells. The results show that it is the cis-interaction between L1 and N-CAM that depends on the appropriate carbohydrate structures.

The L1 adhesion molecule is a cellular ligand for VLA-5.

The L1 adhesion molecule is a member of the immunoglobulin superfamily shared by neural and immune cells. In the nervous system L1 can mediate cell binding by a homophilic mechanism. To analyze its function on leukocytes we studied whether L1 could interact with integrins. Here we demonstrate that VLA-5, an RGD-specific fibronectin receptor on a wide variety of cell types, can bind to murine L1. Mouse ESb-MP cells expressing VLA-5 and L1 could be induced to aggregate in the presence of specific mAbs to CD24 (heat-stable antigen), a highly and heterogeneously glycosylated glycophosphatidylinositol-linked differentiation antigen of hematopoietic and neural cells. The aggregation was blocked by both mAbs to L1 and VLA-5, respectively. Aggregation was blocked also by a synthetic RGD-containing peptide derived from the Ig-domain VI of the L1 protein. ESb-MP subclones with low L1 expression could not aggregate. In heterotypic binding assays mouse bone marrow cells could adhere in an L1-dependent fashion to platelets that expressed VLA-5. Also purified L1 coated to polystyrene beads could bind to platelets. The binding of L1-beads was again inhibited by mAbs to L1 and VLA-5, by soluble L1 and the L1-RGD peptide in a dose-dependent manner. Thymocytes or human Nalm-6 tumor cells expressing VLA-5 could adhere to affinity-purified L1 and to the L1-derived RGD-containing peptide coated to glass slides. The adhesion was strongly enhanced in the presence of Mn(2+)-ions and blocked by mAbs to VLA-5. We also demonstrate a direct L1-VLA-5 protein interaction. Our results suggest a novel binding pathway, in which the VLA-5 integrin binds to L1 on adjacent cells. Given its rapid downregulation on lymphocytes after induction of cell proliferation, L1 may be important in integrin-mediated and activation-regulated cell-cell interactions.

Nectadrin, the heat-stable antigen, is a cell adhesion molecule.

Nectadrin, the cell surface glycoprotein recognized by the novel mAb 79, was found to be immunologically identical to the heat-stable antigen (HSA). It is a glycoprotein with a polypeptide core of only 30 amino acids and a very high carbohydrate content (Wenger, R. H., M. Ayane, R. Bose, G. Köhler, and P. J. Nielsen. 1991. Eur. J. Immunol. 21:1039-1046). Immunocytological studies using cultured splenic B-lymphocytes, neuroblastoma cells, and cerebellar cells indicated that nectadrin is preferentially expressed at sites of cell-cell contact. Purified nectadrin and monoclonal nectadrin antibody 79, but not other monoclonal nectadrin antibodies, inhibited the aggregation of B-lymphocytes by 70%, suggesting that nectadrin may act as a cell adhesion molecule. Nectadrin was purified from a mouse lymphoma cell line in two forms of 40-60 and 23-30 kD. The lower molecular weight form appears to be generated from the higher molecular weight form by degradative removal of saccharide residues characteristic of complex type oligosaccharide side chains. Latex beads coated with purified nectadrin aggregated and the rate of their aggregation depended on the molecular form of nectadrin, with the larger form being more potent than the smaller one in mediating bead aggregation. Nectadrin thus appears to be a self-binding cell adhesion molecule of a structurally novel type in that its extensive glycan structures may be implicated in mediating cell adhesion.

Ectodomain shedding of L1 adhesion molecule promotes cell migration by autocrine binding to integrins.

The L1 adhesion molecule plays an important role in axon guidance and cell migration in the nervous system. L1 is also expressed by many human carcinomas. In addition to cell surface expression, the L1 ectodomain can be released by a metalloproteinase, but the biological function of this process is unknown. Here we demonstrate that membrane-proximal cleavage of L1 can be detected in tumors and in the developing mouse brain. The shedding of L1 involved a disintegrin and metalloproteinase (ADAM)10, as transfection with dominant-negative ADAM10 completely abolishes L1 release. L1-transfected CHO cells (L1-CHO) showed enhanced haptotactic migration on fibronectin and laminin, which was blocked by antibodies to alpha v beta 5 and L1. Migration of L1-CHO cells, but not the basal migration of CHO cells, was blocked by a metalloproteinase inhibitor, indicating a role for L1 shedding in the migration process. CHO and metalloproteinase-inhibited L1-CHO cells were stimulated to migrate by soluble L1-Fc protein. The induction of migration was blocked by alpha v beta 5-specific antibodies and required Arg-Gly-Asp sites in L1. A 150-kD L1 fragment released by plasmin could also stimulate CHO cell migration. We propose that ectodomain-released L1 promotes migration by autocrine/paracrine stimulation via alpha v beta 5. This regulatory loop could be relevant for migratory processes under physiological and pathophysiological conditions.

Drug-induced expression of the cellular adhesion molecule L1CAM confers anti-apoptotic protection and chemoresistance in pancreatic ductal adenocarcinoma cells.

Pancreatic ductal adenocarcinoma (PDAC) is characterized by rapid tumor progression, high metastatic potential and profound chemoresistance. We recently reported that induction of a chemoresistant phenotype in the PDAC cell line PT45-P1 by long-term chemotherapy involves an increased interleukin 1 beta (IL1beta)-dependent secretion of nitric oxide (NO) accounting for efficient caspase inhibition. In the present study, we elucidated the involvement of L1CAM, an adhesion molecule previously found in other malignancies, in this NO-dependent chemoresistance. Chemoresistant PT45-P1res cells, but not chemosensitive parental PT45-P1 cells, express high levels of L1CAM in an ILbeta-dependent fashion. PT45-P1res cells subjected to short interfering RNA (siRNA)-mediated L1CAM knock-down exhibited reduced inducible nitric oxide synthase expression and NO secretion, as well as a significant increase of anti-cancer drug-induced caspase activation, an effect reversed by the NO donor S-nitroso-N-acetyl-D,L-penicillamine. Conversely, overexpression of L1CAM in PT45-P1 cells conferred anti-apoptotic protection to anti-cancer drug treatment. Interestingly, L1CAM ectodomain shedding, in example, by ADAM10, as reported for other L1CAM-related activities, seemed to be dispensable for anti-apoptotic protection by L1CAM. Neither the shedded L1CAM ectodomain was detected in chemoresistant L1CAM-expressing PT45-P1 cells nor did the administration of various metalloproteinase inhibitors affect L1CAM-dependent chemoresistance. Immunohistochemical analysis revealed L1CAM expression in 80% of pancreatic cancer specimens, supporting a potential role of L1CAM in the malignancy of this tumor. These findings substantiate our understanding of the molecular mechanisms leading to chemoresistance in PDAC cells and indicate the importance of L1CAM in this scenario.

L1 cell adhesion molecule (L1CAM) as a pathogenetic factor in endometriosis.

BACKGROUND: Endometriosis is a benign and progressive disease with a high prevalence. Women with endometriosis, especially with atypical endometriosis, have a higher probability for developing ovarian cancer compared with women without endometriosis. The L1 cell adhesion molecule (L1CAM) is over expressed in ovarian and endometrial carcinomas and is associated with a bad prognosis. Here, we have analysed L1CAM expression in endometriosis. METHODS AND RESULTS: In our study with the samples from 79 patients with, and 37 patients without, endometriosis, we found that endometriosis cell lines and short-term cultures of endometrium from women with endometriosis expressed L1CAM at the mRNA and protein level. Quantitative RT-PCR analysis showed that L1CAM was expressed at significantly higher level in the epithelial compartment from patients with endometriosis compared with healthy controls (P = 0.0126). By immunohistochemical staining, 15 of 31 ovarian endometriotic lesions (48%) were shown to have L1CAM-positive staining. Of these 15 L1CAM-positive samples, 13 were atypical endometriotic lesions. Soluble L1 present in the conditioned medium of epithelial endometrium cultures from women with endometriosis was able to stimulate neurite outgrowth as measured in a chicken ganglion assay. CONCLUSIONS: We propose that L1CAM could promote endometriosis development by increasing enervation and aggravation. L1CAM expression is higher in atypical endometriosis compared with normal endometriosis.

Extracellular vesicles from human saliva promote hemostasis by delivering coagulant tissue factor to activated platelets.

Essentials Human salivary extracellular vesicles (EVs) expose coagulant tissue factor (TF). Salivary EVs expose CD24, a ligand of P-selectin. CD24 and coagulant TF co-localize on salivary EVs. TF+ /CD24+ salivary EVs bind to activated platelets and trigger coagulation. SUMMARY: Background Extracellular vesicles (EVs) from human saliva expose coagulant tissue factor (TF). Whether such TF-exposing EVs contribute to hemostasis, however, is unknown. Recently, in a mice model, tumor cell-derived EVs were shown to deliver coagulant TF to activated platelets at a site of vascular injury via interaction between P-selectin glycoprotein ligand-1 (PSGL-1) and P-selectin. Objectives We hypothesized that salivary EVs may deliver coagulant TF to activated platelets via interaction with P-selectin. Methods We investigated the presence of two ligands of P-selectin on salivary EVs, PSGL-1 and CD24. Results Salivary EVs expose CD24 but PSGL-1 was not detected. Immune depletion of CD24-exposing EVs completely abolished the TF-dependent coagulant activity of cell-free saliva, showing that coagulant TF and CD24 co-localize on salivary EVs. In a whole blood perfusion model, salivary EVs accumulated at the surface of activated platelets and promoted fibrin generation, which was abolished by an inhibitory antibody against human CD24. Conclusions A subset of EVs in human saliva expose coagulant TF and CD24, a ligand of P-selectin, suggesting that such EVs may facilitate hemostasis at a site of skin injury where the wound is licked in a reflex action.

Different expression levels of alpha3/4 fucosyltransferases and Lewis determinants in ovarian carcinoma tissues and cell lines.

Ovarian carcinoma is the leading cause of death from gynecological cancers in many countries. Fucosylated glycoconjugates have been associated with various carcinomas. In the present study, we have characterized the expression of alpha3/4 fucosyltransferases transcripts and their products, the Lewis carbohydrate determinants, and their in vitro specificity towards synthetic acceptors using ovarian carcinoma cell lines OVM, m130, GG and SKOV3. We found different expression patterns: GG cells expressed mostly Lewisx (Lex), Lewisy (Ley), sLea and Leb, and m130 cells expressed mostly Lex and Ley. The detection was on the plasma membrane and in intracellular vesicles. OVM and SKOV3 cells had very low amounts of staining. From RT-PCR studies, enzyme specificity of cellular extracts towards a panel of synthetic carbohydrate acceptors and Western blot analysis we concluded that Lea, sLea and Leb were synthesised by FUT3, whereas Lex and Ley were synthesized by FUT4 and FUT9 in both cell lines. The GG and m130 cell lines are adequate models to investigate the role of Lex, Ley, sLea and Leb in ovarian carcinoma development.

Effects of mutagens on the immunogenicity of tumor cells: comparison of cell surface changes induced by 5-azacytidine and N-methyl-N'-nitro-N-nitrosoguanidine.

The reduced tumorigenicity of mutagenized cells appears to be due to the induction or alteration of cellular determinants leading to a more effective immune recognition. We previously reported that clones of the murine T-lymphoma Eb obtained after N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) treatment (Eb-MNNG) show a moderate to strong decrease in tumorigenicity and in addition altered expression of several cell surface proteins. We report here that treatment of Eb cells with 5-azacytidine (5-Aza) induces cell surface changes that are very similar to those seen in MNNG clones; likewise, Eb-5-Aza clones were decreased in tumorigenicity and increased in immunogenicity. Regressor mice that had been inoculated with Eb-5-Aza or Eb-MNNG cells could be shown to be protected against a subsequent challenge of Eb cells but not against the syngeneic mastocytoma P 815. Analysis of cytosine methylation of Eb-MNNG and Eb-5-Aza clones by high pressure liquid chromatography revealed decreased levels of methylcytosine content in several clones. Because of these similarities and because both drugs are DNA hypomethylating and gene inducing agents we assume that hypomethylation is involved in causing amplified protein expression. The possible role of amplified proteins in increased immunogenicity is discussed.

High-frequency generation of altered Mr 70,000 env glycoproteins in N-methyl-N'-nitro-N-nitrosoguanidine-treated murine tumor cells.

In previous studies we have characterized variant clones established following treatment of mouse Eb lymphoma cells with N-methyl-N'-nitro-N-nitrosoguanidine (MNNG). Some of these clones were impaired in tumorigenicity due to an increased immunogenicity (tum- phenotype). In this paper we investigated the mutagenic effect of MNNG on cell surface molecules. The results show that retroviral Mr 70,000 glycoprotein (gp70) antigens undergo extensive alterations following MNNG treatment. In five of five mutant clones analyzed, the two-dimensional gel electrophoretic patterns of gp70 antigens were altered in comparison to the parental Eb cells. Peptide mapping analysis of immunoprecipitated gp70 molecules using three different enzymes revealed detectable changes in amino acid sequence in three of five mutant clones. In contrast, no alterations were detected in H-2Kd and H-2Dd antigens of the same clones. The gp70 antigens expressed by mutant clones could be resolved in three distinct clusters. Only one cluster induced antibodies in the syngeneic host. When genomic DNAs of MNNG clones were investigated by Southern blot analysis using a gp70-specific probe, an additional 4.5-kilobase hybridizing band could be detected that was not present in parental Eb cells and 5'-azacytidine-treated Eb clones. Collectively, our results show that gp70 antigens and genes are affected by MNNG treatment with high frequency. The possible role of structurally altered gp70 molecules in the immunogenicity of mutagenized tumor cells is discussed.

Endogenous retroviral env genes after N-methyl-N'-nitro-N-nitrosoguanidine treatment of mouse tumor cells: stable DNA amplification and rearrangement.

Immunogenic tumor variant clones derived by N-methyl-N'-nitro-N-nitrosoguanidine treatment of Eb lymphoma cells showed structurally altered gp70 env proteins at the cell surface. To further investigate this observation we screened for complementary DNA clones encoding gp70 antigens from a lambda gt11 expression library constructed from mRNA of a mutant cell clone. Using gp70-specific antibodies, a total of 10 complementary DNA clones were identified and analyzed. DNA-sequence analysis revealed the presence of both xeno/or mink cell focus-forming-type (clones 1, 10) and endogenous ecotropic gp70 (clones 2, 3, 4). Southern blot experiments using clone 2 as a probe detected new restriction fragments and DNA amplification with high copy number in the DNAs of N-methyl-N'-nitro-N-nitrosoguanidine-treated but not in parental or 5'-azacytidine-treated control cells even after greater than 120 days in tissue culture. No alterations in the restriction fragment pattern of the genomic DNAs could be detected using H-2K-, beta-actin-, or dehydrofolate-reductase-specific gene probes. The DNA amplification of specific gp70 sequences may be related to the enhanced expression of a subset of gp70 in mutant cells that was previously found to induce syngeneic antibody production.

Effects of mutagens on the immunogenicity of murine tumor cells: immunological and biochemical evidence for altered cell surface antigens.

Eb lymphoma cells were subjected to treatment in vitro with the alkylating mutagen N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) and then cloned by limiting dilution. When tested in vivo for tumorigenicity in groups of syngeneic DBA/2 mice, 6 from 18 clones were found to be strongly reduced (tum- phenotype). The other clones showed only moderate or no change in tumorigenicity compared to the untreated control. All clones were able to grow in 400-rad-irradiated mice. Mice in which MNNG clones had regressed were able to generate tumor-specific cytolytic T-lymphocytes in vitro. Limiting dilution analysis indicated that 3 of 4 MNNG clones analyzed in detail displayed additional antigenic determinants that were detected by cytolytic T-lymphocytes. These data thus provided evidence for increased immunogenicity of some of the MNNG clones. Membrane proteins of MNNG clones and original Eb cells were compared biochemically after metabolic labeling with [35S]methionine, TX114 solubilization, and electrophoretic separation. Two-dimensional gel maps revealed a general quantitative decrease in the expression of membrane proteins in MNNG clones. In addition, several proteins were only found in MNNG clones but not in untreated cells. Two membrane proteins of molecular weight 22,000 and 38,000 were greatly increased in expression in all MNNG clones but could be detected at a low level in the original Eb cells. MNNG is known to be a strong mutagenic agent, but it can also interfere with DNA methylation and cause transcriptional activation of genes. We suggest that amplified cell surface structures may be the consequence of such transcriptional activation and could be involved in altered immunogenicity.

Resistance of metastatic tumor variants to tumor-specific cytotoxic T-lymphocytes not due to defects in expression of restricting major histocompatibility complex molecules in murine cells.

Metastases of ESb lymphoma cells in syngeneic DBA/2 animals frequently are selectively immunoresistant to lysis by syngeneic cytotoxic T-lymphocytes (CTL). This immunoresistance of tumor cells to CTL lysis could be due to a defect in either of two structures: (a) tumor-associated transplantation antigens; or (b) mouse major histocompatibility complex (H-2) antigens serving as restricting elements. In this study, we have analyzed the possible involvement of major histocompatibility complex Class I antigens in the immunoresistance of ESb tumor variant cells. Syngeneic anti-ESb CTL appeared to be H-2Kd restricted since only antibodies to Kd but not D,Ld molecules could inhibit CTL lysis. Comparison of H-2 antigens expressed on immunosensitive and resistant ESb sublines by immunofluorescence and flow cytofluorography, alloreactive CTL, two-dimensional gel analysis did not reveal any differences either qualitatively or quantitatively. Southern blotting of tumor-derived DNA with H-2-specific probes did not reveal differences either. Serologically detectable cell surface differentiation antigens were expressed very similarly on immunosensitive and resistant tumor lines, and only minor differences were noted by biochemical analysis of plasma membrane proteins. C-Type viral Mr 70,000 glycoprotein antigens were also similar in both types of cells. We conclude that cell surface changes on immunoresistant ESb variant cells are very selective and involve only CTL-defined tumor-associated transplantation antigen determinants.