(1) Seminar on “Receptor-mediated Modulation of Lymphocyte Growth and Activation”.
This meeting was held January 13-15, 1992 in San Diego, California. There were twelve U.S. participants and eight Japanese participants at this seminar which focussed upon receptor-mediated regulation of growth, differentiation, and activation of lymphocytes.
I. Signal Transduction in B Cells. Drs. John Cambier and Anthony DeFranco discussed the signal transduction mechanisms mediating activation of B cells through the membrane Ig (mIg) receptor complex. John Cambier described the structure of the molecular complex associated with mIg on the surface of B cells and the similarity between this complex and the previously characterized T cell receptor (TCR) complex. The possibility was discussed that there is tyrosine kinase activity in the MB1 and B29 gene products associated with the mIg receptor complex. In addition, the influence of CD45 on B cell signalling through mIg was described. Tony DeFranco analyzed the expression of phospholipase C!
!!PLC molecules by B cells. A role for early tyrosine phosphorylation of PLC!!!2 during B cell activation was described. In addition, evidence for the role of a G protein in mIgM-induced B cell activation was discussed. Kiyoshi Takatsu described the structure and function of the IL5 receptor on B cells. A role of IL5 was identified in early B cell development as well as in the activation of mature B cells. Both high and low affinity IL5 receptors have been characterized. It has been shown that receptors for IL5, GMCSF, and possibly IL3 all share a common p130 chain. Signal transduction through the IL5 receptor appears to involve protein tyrosine phosphorylation but no detectable calcium influx or PI hydrolysis Transgemc mice expressing IL5 have been produced and have been found to exhibit increased polyclonal B cell activation as well as massive eosinophilia, reflecting in vivo the activity of IL5 on these cells.
Richard Hodes described the expression of differential isoforms of CD44 and CD45 by activated B lymphocytes. The activation of B cells by IL5 was unique in inducing a subpopulation of B cells expressing a high surface density of CD44. This population had a dramatically increased affinity for the extracellular matrix component hyaluronic acid (HA). The CD44 molecules expressed on these cells exhibited a decrease apparent molecular weight on gel analysis as a result of differential N-glycosylation. Activation of B cells by several stimuli also resulted in distinct patterns of expression of CD45 isoforms. Complex changes in serologically detected CD45 determinants were correlated with changes detected by immunoprecipitation. In addition, unique patterns of differential splicing of variable CD45 exons were correlated with specifrc B cell stimuli.
II. Signal Transduction in T cells. The regulation of phosphorylation by kinases and phosphatases was discussed in several presentations. Bart Sefton described the src family of protein tyrosine knases and their function during lymphocyte activation. Several bands of phosphorylated proteins were described as mIg associated molecules following B cell activation. Multiple src kinases were identified in individual B cells. The substrates for CD45 phosphatase were analyzed, and it was found that CD45 had significant activity on the phosphorylation of p56lck with little or no effect on lyn or fyn, implying that CD45 may, directly or indirectly, differentially affect different kinases. Art Weiss described the ability of the T cell receptor complex zeta chain to mediate signalling through the T cell receptor. Within seconds after stimulation the!
!!chain is strongly associated with a 70kd tyrosine phosphoprotein (ZAP-70). The portion of the!!!chain critical to signal transduction involves a motif that is in common in the CD3!!!,!
!!, and!
!!chains as well as in the Y chain of the IgE FcR and in the MB-1 and B29 chains of the mIg complex. The defects in CD45 deficient Jurkat variants were analyzed. These mutants failed to activate tyrosine phosphorylation or PLC through the T cell receptor. This defect was partially reconstituted by transfection with the 180kDa CD45 isoform. Dr. Takashi Saito discussed the consequences of heterogeneity of signalling through the TCR complex. Variants of the 2B4 hybridoma were found which failed to show increases in intracellular [Ca2+] in response to TCR signalling despite having normal surface levels of TCR expression. When lines were compared which had variable degrees of expression of!
!!and!
!!, there was no correlation of responsiveness with!!!or!!!expression. Fc!!!R!!!transfectants were also made which supported TCR expression in negative cell lines. Stimulation of these lines with specific antigen plus APC or anti-TCR antibody produced significant PI hydrolysis responses. These transfectants also secreted IL2 in response to antigen or anti-TCR stimulation. In contrast, growth inhibition was not seen in response to antigen. Experiments using homologous recombination have successfully targeted!!!and!!!genes in a hybridoma resulting in significantly decreased levels of expression of the corresponding gene products and in overall expression of TCR. These cells are unresponsive to antigen. Shigeo Koyasu also described signal transduction through a variety of TCR isoforms. MA5.8 cells, which are defective in expression of!!!and!!!were transfected with cDNA for!!!,!!!, or both. There is no difference in calcium PI, IL2, cell cycle arrest, or PTK activation in these various transfectants. The!!!chain of the Fc receptor was also sufficient to reconstitute MA5.8 although it produced a lower level of responsiveness. These two presentations thus demonstrated that there is a family of at least 3 members!!!,!!!, and Fc!!!, each of which can function in association with the TCR. The role of phosphatase was further discussed by Drs. Ian Trowbridge and Matthew Thomas. Ian Trowbridge described that cross-linking CD4 increased tyrosine phosphorylation of p56lck. The co-crosslinking of cDNA4 with CD45 inhibits this phosphorylation. In vitro kinase assays of p56lck activity give parallel results. Dr. Trowbridge also reported the first example of regulated phosphatase activity in vivo. It was found that the activity of CD45 phosphatase activity in mouse or human cells was inhibited by in creased intracellular [Ca2+]. This appears to be an indirect effect, requiring approximately 40 minutes and reflected by a decreased serine phosphorylation of CD45. Matt Thomas described the growing list of protein tyrosine phosphatases expressed in the mouse. These include both transmembrane molecules and intracellular phosphatases. The role of CD45 was analyzed in CD4+ or CD8+ CD45 loss variants. In response to stimulation with antigen or anti-CD3, these lines showed dramatically decreased proliferation, lymphokhe production, or cytolytic activity. They also exhibited increased resting levels of several tyrosine phosphoproteins.
Carl June described the role of CD28 in signal transduction in T cells. Stimulation through CD28 has previously been found to be cyclosporin resistant and does not appear to act by influencing!
!!chain phosphorylation. Use of herbimycin, an inhibitor of protein tyrosine kinase activity, demonstrated that signalling through CD28 does involve some kinase activity. Anti-CD28 stimulation of T cells which have been pre-activated by anti-CD3 results in the appearance of a characteristic pattern of tyrosine phosphoproteins. This can be achieved as well by stimulation with B7 molecules expressed on B cells, which may represent the natural ligand for CD28. Bob Abraham described studies analyzing signal transduction through the IL2 receptor The early signals involved in activation of cells through the IL2 receptor are not known. Stimulation of T cells with IL2 clearly leads to rapid increases in protein tyrosine phosphorylation. However, the IL2 receptor appears to lack intrinsic PTK activity. Evidence was described for the role of p59fyn but not p56lck in the coupling of IL2 receptor ligation to induction of P13-kinase activity in T cells. Further evidence was provided from an lck deficient cell line which implicated a role for p56lck in the “priming” effect of IL2 for cytotoxic effector function. These studies thus provided evidence for two different PIX molecules in transducing divergent regulatory signal pathways in IL2 stimulated T cells.
Shin Yonehara reported on the structure and function of the cell surface Fas antigen. Antibodies to the Fas molecule induced apoptosis similar to that induced by TNF, but the Fas molecule is distinct from the TNF receptor. The cDNA encoding Fas has been characterized and shown to be similar to the extracellular domain of TNF receptors and NGF receptors.
Takushi Tadakuma characterized the conversion from IL2 production to IL4 production by T cells repeatedly stimulated with a variety of antigens. Different antigenic stimuli differ in the efficiency of conversion from IL2 to IL4 producton. Pre-treatment with IL1 allowed a conversion to IL4 production after stimulation with anti-CD3 antibody. Long term T cell lines which produced IL4 alone after antigen stimulation produced IL2 when they were stimulated in the presence of cycloheximide, suggesting that the switch from IL2 to IL4 secretion is mediated by a cycloheximide sensitive protein.
The mechanism of signal transduction in cell-cell adherence junctions was discussed by Shoichiro Tsukita It was shown that a number of unique proteins concentrate in the area of the adherence junction (AJ) Recently, it has been shown that regulatory enzymes such as protease and kinases are among the molecules which accumulate in this area and which may be important for signal transduction. Cell-cell AJ’s have recently been isolated and found to contain molecules including yes kinase and src knase. The possible roles of these molecules in adherence and signal transduction were discussed.
III. Signal transduction in T cell differentiation and repertoire selection.
Signal transduction events involved in thymocyte differentiation were discussed by three speakers. Craig Thompson analyzed expression of the RAG- I and RAG-2 genes which are involved in recombination of T cell and B cell receptors. It was shown that anti-D3 crosslinking terminates expression of RAG- I and RAG-2 in CD4+8+ human thymocytes. These findings suggest a role for ligand recognition by the TCR in terminating subsequent TCR gene recombination events. C0-crosslinking of CD3 with CD4 or CD8 enhances CD4+8+ thymocytes responses as measured by tyrosine phosphorylation, [Ca2+], or PL!
!!1 Phosphorylation. This is inhibited by crosslinking CD45 with CD3 and CD4. Thus, signalling through these surface molecules on immature thymocytes results in a highly regulated pattern of responses. Al Singer described the regulation of TCR expression on CD4+8+ thymocytes by CD4 mediated signals. Evidence was provided that CD4 engagement in vivo results in increased degradation of TCR and a consequent decreased cell surface expression. Culturing of thymocytes in vitro can release them from this influence, resulting in increased TCR expression by decreasing degradation in the endoplasmic reticulum. This increased TCR expression is also associated with an induced capacity for [Ca2+] signalling through the T cell receptor. Inhibition of RNA synthesis blocks the increase of TCR expression in vitro but does allow dephosphorylation of!
!!chain of the TCR, and under these circumstances one sees increased [Ca2+] signalling, suggesting that!!!phosphorylation may be an off signal for the TCR. It was also demonstrated that cultured thymocytes with increased TCR expression undergo apoptosis upon exposure to their ligand, providing a model for a negative selection of self-reactive cellsA role for Gi proteins in T cell signalling was discussed by Roger Perlmutter. Classical G proteins are heterotrimeric signal transduction molecules that mediate responses delivered by a variety of hormone receptors. Two forms of Gi heterotrimer are represented at high levels in thymocytes and mature T cells. Transgenic mice were constructed which active pertussis toxin was expressed under the proximal lck promoter in thymocytes. This toxin inactivates the Gi. The thymocytes from lck-PT mice matured normally within the thymus but failed to populate the peripheral lymphoid system. Thus a PT sensitive process, probably involving the Gi protein, appears to regulate thymocyte emigration. The thymi of these transgenic mice demonstrate a grossly disrupted thymic architecture, but nonetheless demonstrate normal thymic repertoire selection.
Hiromi Fujiwara and Yoshimoto Katsura presented work which analyzed the factors involved in inducing apoptosis in the thymus. Hiromi Fujiwara described the ability of a thymic stromal cell line to induce differentiation of thymocytes. In the presence of specific antigen, these stromal lines can induce death in antigen-specific T cell clones, providing a model for negative selection. A factor produced by these stromal cells allows the induction of apoptosis by splenic APC. This factor has been characterized and purified but not yet molecularly defined, and appears distinct from other known cytoknes including IL7. Yoshimoto Katsura employed a system of thymic organ culture to study the ability of various cell types to mediate apoptosis. Clonal deletion of V!!!6 T cells occurred only in thymi which had been repopulated by both B cells and dendritic cells Only the B cells were required to express the relevant Mlsa. Interaction between these two cell types is therefore required to induce negative selection in this in vitro model of T cell differentiation.
This meeting was perhaps the most exciting and productive of those held within this series to date. Participants included world experts in the cell biology of signal transduction in lymphoid cells. The recent expansion of information concerning signal transduction pathways, combined with studies of lymphocyte receptor structure and the ontogeny of lymphocyte differentiation provided an extremely fertile background for this meeting. Discussions were extremely intense and productive, and resulted in suggestions for f future avenues of research in this field
(2) Seminar on “Programmed Cell Death in Cancer & T Cell Repertoire Selection”
Hotel Nikko Fukuoka, Japan, February 18-19, 1992
Professor Takahiko Sasazuki welcomed all participants at the first session on Tuesday, February 18, 1992 covering Programmed Cell Death and its Genetic Control Dr. Taisei Nomura of Osaka University presented studies on the biological effects of low and high dose radiation. He had established a very effective thin section organ culture system for thymus in which radiation effects could be quantitated and followed longitudinally. A novel observation in this system was the presence of clustered thymocyte death.
Dr. Nomura also established the radiation sensitivity of various strains of mice noting that B6 and AKR are quite sensitive while C3H is resistant. A series of genetic backcrosses suggests that there are one dominant gene and two recessive genes involved in these strain differences. Of note, the cumulative incidence of leukemia in whole body irradiation is directly proportional to the radiation sensitivity in these strains. Dr. Takeshi Yamada of the National Institute of Radiologic Sciences in Chiba identified differences in low dose vs. high dose radiation. Exposure to up to 100 rads resulted in an interphase death in which no division was needed. This event is associated with a cell size reduction, the disappearance of microvilli, and internucleosomal degradation of chromatin, all of which will be blocked by pretreatment with cycloheximide. A high dose radiation results in a proliferation that resembles necrosis. Dr. Harumi Ohyama also of the National Institute of Radiologic Sciences at Chiba had established a system of in vivo radiation followed by in vivo culture assessment Ten gy resulted in an or none low dose effect whereas 100 gy proceeded by a necrosis pathway. This investigator presented an intriguing assay, “coma test” capable of detecting single cell apoptosis.
The last half of the morning session turned to the identification of genes within cell death systems. Dr. Robert H. Horvitz of the Department of Biology at MIT detailed his studies on the genetics of cell death in the nematode Caenorhavditis Elegans. In this nematode 131 starter cells are destined to die following cell divisions. Mutations in this pathway have identified two autosomal recessive genes, ced-3 and ced-4 representing effectors of programmed death. Their elimination results in the survival of all 131 cells. Ced-4 provocatively possesses an EF hand motif characteristic of calcium binding proteins whereas ced-3 has a serine arginine rich N-terminus similar to phosphorylation sites in other proteins. In addition an autosomal dominant gene entitled ced-9 has been identified in which a single allele upregulation saves all 131 cells.
John Cohen, M.D., Ph.D. from the Department of Microbiology of the University of Colorado, Denver followed describing a subtractive hybridization approach for the isolation of death pathway genes. He utilized cycloheximide treatment to prevent RNA degradation following the induction of dexamethasone death in thymocytes. 20 independent clones have been isolated that show transcriptional increases within 2-4 hours following dexamethasone treatment. Several of these genes have also been found to be active in growth factor deprivation death and one contains a zinc finger and another a membrane spanning segment His studies on the effects of cycloheximide in different programmed death systems argues for two separate categories of transduction and induction. Transduction would require no new RNA or protein synthesis following signal transduction and might include T cell K responses as well as hyperthermia whereas induction requires RNA and protein synthesis and would include dexamethasone and radiation induced death. Dr. Larry Schwartz of the University of Massachusetts, Amherst completed the session with a discussion of death pathways within the moth Mimducasexta and within T cells. At day 18 of development in Minducasexta the adult moth emerges after massive death of intersegmental muscles. A +/- cDNA screening approach revealed polyubiquidn genes that were massively induced at 3 and 5 hours on day 18. Multiple proteins in this system get ubiquinated and targeted for rapid massive degradation. In contrast he noted that dexamethasone treatment of thymocytes did not result in an induction of polyubiquitin. Consequently, a +/- hybridization screening was performed on a transgenic mouse model in which T cell receptors on thymocytes are specific for an ovalbumin peptide. One gene, apt-2 is particularly attractive as it is induced within 1-1/2 hours following treatment with peptide or anti-CD3 is blocked by cyclosporin A and actinomycin D and contains a zinc finger. Thus these two pathways of death seem distinctly different in which both require new gene expression but one is a polyubiquitin pathway and the other independent of that. T cell death is associated with apoptotic changes in genomic DNA degradation whereas the intersegmental death is not
Professor Sohei Kondo opened the afternoon session entitled “Programmed Cell Death in Cancer”. Dr. Shin Yonehara from the Tokyo Metropolitan Institute of Medical Science presented his fascinating work on the Fas antigen. He has noted that the P55 form of the TNF receptor type-1 has a conserved intracytoplasmic domain with Fas Transfection of Fas antigen into target cells results in their death with anti-Fas antibody. Dr. Yoneham believes a major role for Fas is in the elimination of activated peripheral T cells. Following PHA stimulation and treatment with IL2 some 50-60% of those cells are easily killed with anti-Fas. Of potential therapeutic importance Dr. Yoneham has noted that HIV infected cells are sensitive to anti-Fas monoclonal antibody. Dr. Stanley Korsmeyer of Washington University School of Medicine completed this session with a discussion of the Bcl-2 proto-oncogene and its role in blocking programmed cell death. Bcl-2 is novel as an oncogene in that its protein product is an integral mitochondrial membrane protein. Moreover, it functions to inhibit apoptosis rather than promoting proliferation. Two transgenic models were discussed which revealed normal physiologic roles of Bcl-2. In a B cell model the deregulated Bcl-2 extended immune responses including resting B cell memory. A T cell model redirected Bcl-2 to the immature thymocytes of the cortex. Those normally vulnerable thymocytes were now resistant to glucocorticoid, radiation and anti-CD3 induced death. Despite that three forms of negative selection against super-antigen, class I or class II antigens remained entirely intact. The mice selectively accumulated an intermediate population of T cells with medium intensity T cell receptors and CD4+8+ cells that were downregulating either CD4 or CD8. It is speculated that Bcl-2 may play a role in promoting positive selection resulting in this phenotype.
The Afternoon Session was opened by Dr. John Cohen and entitled “Gene Targeting and Chromosomal Transfers for Studying Programmed Cell Death” The first presentation was from Professor Takehiko Sasazuki from Kyushu University. He discussed knock-out and replacement of activated K-ras genes in human colon cancer cell lines. He observed that homologous recombination that eliminated the mutated allele of the K-ras gene resulted in decreased growth. Whereas, homologous knock-out of a normal allele of K-ras produced the reciprocal event of increased growth rate. Moreover, there was an in vivo effect in that the elimination of the mutated allele demonstrated a loss of tumorigenicity within a nude mouse model. These studies provided convincing proof of a direct etiologic role for Ras mutations in the tumorigenicity of colon carcinoma cells. In addition Dr. Sasazuki summarized the incidence of P53, N-ms and DCC abnormalities within the Japanese population which are very parallel to studies within the United States. Dr. Takeshi Yagi of Tsukuba Life Science Center presented fascinating data on targeted gene replacement of the C-fyn locus by homologous recombination in ES cells. He generated an embryonal stem cell line which produces a remarkable 80% chimeric rate of which 90% transmit to the germline. He utilized an exon III region construct of fyn and cleverly inserted a LacZ reporter with a polyA tail as well as a neomycin resistance gene. This had the additional benefit of producing blue cells at the site of fyn expression as well as knocking out the native protein. The investigator observed that fyn was expressed in the caudal neural tube as its first site of utilization. By day 13 it was in ganglion, thymus, and the olfactory bulb of the brain. On day 15 the dura major in the mesencephalon also expressed. Homozygous mice showed a decreased sized brain. These mice grew slowly as well and local examination revealed the possibility of neuronal cell death. Dr. Mitsuo Oshimura from Tottori University in Yonago completed this session with the discussion of the induction of cellular senescence of immortalized cells by chromosome transfer. He discussed the utilization of microcell transfers to define complimenting genes. The HT-1080 parental recipient tumor line showed markedly decreased cell growth, colony formation, elimination of tumorigenicity and an increased doubling time with the transfer of chromosome 1 but not with chromosome 11 or chromosome 2. Moreover in a bladder carcinoma cell line he indicated that chromosome 7, 9 and 1 2 prevented senescence. Overall Dr. Oshimura divides his types of genes into type l which induce cellular senescence: chromosome 1, 4 and 11; type 2 which reduce in vitro growth and produce morphologic alterations; and type 3, suppressors of tumorigenesis including chromosome I for neuroblastoma, chromosome 3 for renal cell carcinoma, chromosome 6 for uterine carcinoma, chromosome 9 for uterine carcinoma, and chromosome 11 for many carcinomas. Provocatively he feels that the area on chromosome segment 1p encodes a suppressor gene and may well be the same area in which neuroblastomas reveal loss of heterozygosity.
A second day of the conference on February 19 was chaired by Dr. David Ucker and entitled “Program Cell Death in T Cell Repertoire Selection”. Dr. Ucker initiated the program with a discussion of physiologic cell death in the cell cycle. This work was performed at the Molecular Biology Institute at La Jolla He performed a detailed analysis of nuclear lamins (the intermediate filaments of the nuclear envelope). He examined the ratio of soluble to insoluble fractions during proliferation as well as with the induction of death. From this work he noted a marked relationship between genome digestion and nuclear dissolution. In another system he looked at the attack of cytotoxic T cells on NH3T3 cells which initially shift DNA to 100-200 kb mean size in pulsed field gels. He found one variant of a 3T3 line that did not digest DNA even though it died an apparent apoptotic death. The nuclease that Dr. Ucker is examining has a mobility of approximately 40 kD and activities much more like DNAse I. Of interest the DNA degradation resistant target cell lacked this endonuclease activity. Dr. Ucker concludes that genome digestion may be a dispensable consequence in apoptosis and that no fall in ATP occurs in that setting. His attention is now focused upon whether apoptosis represents an abortive attempt at mitosis. The next talk was provided by Dr. Takushi Tadakuma from Keiogijuku University of Tokyo. This talk dealt with TCR mediated apoptosis in T cell repertoire selection and used analysis of CD4/CD8 thymic lymphomas reactive to SEB. He noted that CD4+8+ thymic lymphomas were more sensitive to anti-CD3 induced death and could find examples of cells that were CD4+8+ and immature by HSA that were also killed by SEB, dexamethasone, cyclic AMP, calcium ionophore. He documented the need for antigen presenting cells within the thymus and found that epithelial cells induced with interferon which expressed either IA or IE would work in this system. He noted culture systems in which treatment with anti-CD3 would result in death of some cells at 48 hours and then the selective emergence of either CD4+ or CD8+ cells at day 3 to day 5. This was an intriguing model of. in vitro thymic selection. The next talk was by Professor Takeshi Watanabe from Kyushu University in Fukuoka He presented work with a fascinating thymic stromal line from neonatal thymus. This line within 24 hours would show marked endocytosis of BALB/C thymocytes very reminiscent of thymic nurse cells. Moreover the immature cells in the presence of the TEL-2 line progressed to CD4 and CD8 cells which demonstrated high T cell receptor intensity Scanning EM showed a potential process of migration of some thymocytes into the cytoplasm of these cells. With the correct use of MHC selecting haplotype this appeared to be a working model of in vitro positive selection. Moreover the addition of IL-7 to these cultures appears to augment this selection. The final presentation of the conference was related to the liomzanon affects in T cell repertoire selection. This very succinct talk was given by Dr. Fukui, a worker in Professot Takehiko Sasazuki’s laboratory at Kyushu University. These investigators took advantage of a very novel integration of a DR trans gene into the X chromosome of C57BL/6 mice. These workers followed the X inactivation of the transgene from 120 cell stage embryo on. Female mice showed a large variation in transgene expression in peripheral blood lymphocytes presumably representing random X inactivation. Of interest homozygous mice showed 1.4% DR high cells whereas hemizygous females showed 4. 1%. Curiously female hemizygous mice showed no response in a primary MLR. This raises the possibility that they are self-tolerant in the presence of a self-reactive repertoire. These cells leak into the peripheral compartment but are non-responsive as a tolerant peripheral state.
Overall this was a highly successful meeting that combined investigators interested in regulation of cell death from perspectives of T cell development radiation induced cell death, genes like Fas and Bcl-2 which are differential effectors of apoptosis, pathways of death in other organisms such as C. elegans, Manduca Sexta, and also novel transgenic and ES approaches to this topic. Without question perspectives were broadened by the type of meeting which initiated multiple discussions and collaborations.
(3) Seminar on “Mucin and Other Cancer-associated Carbohydrate Antigens Related to Cell Biology”
This seminar was held on February 24, 25 and 26 at the Sheraton Makaha Resort, Oahu. Hawaii. The organizers were Dr. Ikuo Yamashina (Kyoto Sangyo University, Kyoto, Japan) and Dr. Young S. Kim (VA Medical Center and UC San Francisco, USA). There were 12 speakers from Japan and 11 from the United States. The purpose of the seminar was to exchange recent findings and views on recent advances in search on structure. biosynthesis, cell biology and immunology of cancer-associated crearbaroehaydrate antigens and to discuss uncertainties and future direction of research in this area.
In the session on characterization of mucin antigens, Dr. Ikuo Yamashina (Kyoto Sangyo University) talked about the epitopic structures recognized by monoclonal antibodies (MAb) established in his laboratory. He observed that Sialyl-Tn and Tn antigens are not the simple structures, such as Sia-GalNAc- Ser/Thr and GalNAc-Ser/Thr, respectively, but are cluster structures of these monomeric structures. Although the cluster structure of Sialyl-Tn antigen has not been demonstrated, his group succeeded in isolating glycopeptides involving (GalNAc)Ser-(GalNAc)-Thr-(GalNAc)-Thr (Tn) by immunoaffinity chromatography. These glycopeptides strongly bound to all the anti-Tn MAbs so far established. Interestingly, this consecutive sequence structure with nonsialylated GalNAc is expressed on mucins of cancer cell surfaces and on ovine submaxillary mucins (OSM) where most of the carbohydrate chains are sialylated. Since Tn antigen is cancer-associated, the formation of the cluster structure with nonsialylated GalNAc is of extreme biological importance.
He then reported on the structures of a large number of novel oligosaccharides isolated by immunoaffinity chromatography using MSW 113 MoAb from human milk and from mucin glycoproteins and glycolipids by NMR analysis. MSW 113 binds not only to Sialyl-Lea oligosaccharides, but also to other Type 1 oligosaccharides such as Sialyl-Lea hexasaccharide (CA19-9) minus internal glucose or lactose residues to the same extent as it binds to the Sialyl-Lea hexasaccharide, while NS 19-9 binds very poorly to the latter type of oligosaccharides. The broader binding property of MSW 113 makes this antibody more useful than NS 19-9 for in vitro cancer diagnosis.
The analysis of oligosaccharides of mucin glycoproteins and glycolipids on cancer cell surfaces showed that mucin oligosaccharides reactive with MSW 113 are extremely large in size and very heterogeneous with a variety of structures which cannot be explained by extensive sialylation nor by the involvement of repeating lactosamine units. In addition, glycolipids with unique oligosaccharides such as Lea-X antigen was isolated and a MAb directed towards its asialo derivative has been shown to be useful for cancer diagnosis.
Dr. Shigeyuki Fukui (Kyoto Sangyo University) discussed the characterization of mucin antigens using MAbs recognizing cancer antigen (MLS 102) and normal antigen (MLS 103). The mucins were isolated from cell lysates of a colon cancer cell line LS 180 which was used as immunogens. The MLS 102 antigen has been shown to be a cluster of sialyl alpha 2-GalNAc-Ser/Thr and the MLS 103 antigen has recently been shown to be Lea antigen. The MLS 102 antigen carries GalNAc and Sia 2-6 GalNAc as major carbohydrate chains attached to polypeptide chain, whereas the MLS 103 antigen carries mainly long carbohydrate chains, most of which possess Lea structure at the nonreducing terminus. These findings are consistent with the idea that cancer-associated glycoproteins or glycolipids are characterized by aberrant glycosylation.
Dr. Kenneth Lloyd (Memorial Sloan-Kettering Cancer Center, New York) discussed three aspects of mucin antigen expression in ovarian epithelial cancer. First he discussed his recent studies on MAb MT334 which reacts with a subset of benign and malignant ovarian tumors. The MT334 antigen is a typical high molecular weight mucin with a characteristic amino acid and sugar composition. the precise epitope recognized by the antibody, however, is not easily elucidated as it seems to consist of both protein and carbohydrate components. The epitope is also unusual in that it is sensitive to reduction by 2-ME and DTT, indicating the involvement of disulfide-linked regions of the protein. He then compared the antigentc profiles of two classes of mucins obtained from benign and malignant ovarian cells as determined by a panel of MAbs. Lastly, he descrived a new HPLC method for the identification of reduced oligosaccharides from mucins. This method is based on anion exchange chromatography in strong base and a very sensitive pulsed amperiometric detector (Dionex Co.). The procedure was applied to three model mucins (OSM, BSM, and PSM). It is particularly suitable for the analysis of sialylated oligosaccharides. Also, NeuAc- and NeuGc-containing oligosaccharides are readily distinguished using this method.
Dr. Eiichi Sato (Kagoshima University Medical School) described the immunohistochemistry of the early stages of cancer in the digestive system. Antibodies were used to localize various carbohydrate antigens including sialyl Tn, Tn, sialyl Lewisx, and H antigens. In cancerous lesions of the esophagus, UEA-1 lectin staining in situ revealed H antigen expression in the basal layer. In the stomach, sialyl Lewisx antigen was expressed in 43% of the gastric adenomas examined, 29% of the minute carcinomas, and 9% of the benign cases. The level of sialyl Lewisx expression appeared to correlate with the survival rate of the patients. Sialyl Tn antigen was highly expressed in the minute carcinomas and in the cancerous foci of the adenomas. High levels of the above two sialyl antigens were detected in goblet cells of intestinal metaplasia, which suggests involvement of the antigens in the metaplasia MUC1 type mucin core protein antigen was highly expressed in pancreatic invasive ductal carcinoma but not in mucin producing pancreatic tumors. Conversely, MUC2 type mucin core protein antigen was expressed in all the mucin producing tumors but not in the ductal carcinoma of the pancreas. These data suggest that the mucin core protein antigens may be associated with the biological properties of pancreatic cancer cells. In the large bowel, H-like blood group antigen stained positive in invasive carcinoma, focal carcinoma in adenoma, an adenoma with severe atypia In contrast, normal mucosa of the distal large bowel gave negative staining. The origin of the mucin-producing tumor and its immunological characteristics were discussed. Analysis of the expression of the above antigens may provide further information on the properties of pancreatic cancer cells and may be useful in determining the extent of malignancy and the prognosis of the patients with digestive tract cancers.
Dr. Yasunori Kozutsumi (Faculty of Pharmaceutical Sciences, Kyoto University) described the biosynthesis of glycoconjugates containing N- glycolylneuraminic acid (NeuGc) residue on the glycoconjugates which is an epitope of the Hanganutziu-Deicher (H-D) antigens. A key step in the production of the NeuGc residue is hydroxylation of CMP-acetylneuraminic acid (CMP-NeuAc) but not direct hydroxylation of NeuAc. The hydroxylase activity is found in the cytosolic fraction of the mouse liver. The enzyme requires at least two components. One of these was purified from horse erythrocytes and identified as cytochrome b on the basis of immunological reactivity and sequencing of the protein. The other factor, designated Z component, has not been purified as yet, but may regulate the expression of the enzyme. Dr. Kozutsumi proposed that an electron transport system from CMP-NeuAc via NADH and NADPH forms CMP-NeuGc. Purification and characterization of azide-sensitive factor Z and salt sensitive factors X and Y are needed to elucidate the mechanism by which NeuGc-containing glycoconjugates including H-D antigens are expressed. A greater understanding of this process may open a new field of glycoconjugate research.
Dr. Nobutaka Wakamiya (Research Institute for Microbial Disease, Osaka University) described the expression of H-D antigen in normal and cancerous human tissues. An H-D antigenic, 29 kD glycoprotein was expressed in T lymphocytes stimulated with IL-2 and phorbol ester (PMA) asjudged by SDS-PAGE and immunoblotting. Similarly, two glycoproteins of 50 kD and 70 kD were induced by treatment with PMA plus A 23187. M-CSF also induced a 100 kD glycoprotein in activated monocytes/macrophages. These glycoproteins await further characterization. In sperm and testicular tissues, H-D glycoproteins were found. An agglutination assay using anti-H-D antigen showed the expression to be limited to the sperm head. Highly elevated titers of antibody to the H-D antigen were observed in serum of patients with autoimmune diseases such as systemic lupus erythematosus and rheumatoid arthritis. A possible role of the antigen to infertility was discussed.
In the session on glycosyitransferases, Dr. Naoyuki Taniguchi (Osaka University School of Medicine) discussed his recent data on cDNA cloning of UDP-N-acetylglucosamine: beta mannoside beta 1-4N-acetylglucosaminyltransferase III (GnT-III). The GnT-III activities in several lines of rat ascites hepatoma were elevated and AH-66 was found to have very high activity. The enzyme activity was induced by feeding the rats with 3’-methyl DAB and the time course experiment exhibited two peaks, early and late, which corresponded to hyperplastic nodule formation and hepatoma development, respectively. The elevation of the GnT-III activity in human hepatoma cultured cells were also demonstrated.
In order to determine the induction mechanism of the enzyme and its relationship to carcinogenesis, his group decided to clone cDNA for this enzyme and started to purify the enzyme from rat kidney and determine amino acid sequences of peptides derived from the purified enzyme. Oligonucleotide primers designed on the basis of the amino acid sequences of tryptic peptides of purified enzyme was used for the polymerase chain reaction. A rat kidney cDNA library was screened with a PCR product and a cDNA clone was obtained. The sequence indicates that the GnT-III is a type II transmembrane protein and the deduced amino acid sequence is 536 amino acids long and contains three putative N-glycosylation sites. No sequence homology is found to other previously cloned glycosyltransfemses. The transfection of the cDNA to COS-1 and HeLa cells were successfully achieved, demonstrating that the cloned gene encodes the GnT-III.
He used the pyridylaminated oligosaccharide (GlcNAc beta 1-2Man alpha 1-2 (GlcNAC beta 1-2Man alpha 1-6) Man heta 1-4 GlcNAc beta 1-4GlcNAc-pyridylamine) as the substrate for the enzyme and the products of the enzyme reaction and remaining substrates were analyzed in one run of HPLC with a fluorescence detector. This method is quite sensitive and can also be applied to the enzyme assay for GnT-IV and V using the same substrate oligosaccharide derivative. The time course of the enzyme activity in the liver of rats fed with the carcinogen demonstrated two peaks and two questions were raised about the difference in molecular mechanisms and the possibility of elevation due to phosphorylation of the enzyme protein. Studies on the correlation between the levels of mRNA and the enzyme activity are in progress. GnT-V activity was reported to correlate with metastatic potential and a possibility that the substrate level for GnT-V may be regulated by GnT-III was raised.
Dr. Akemi Suzuki (Tokyo Metropolitan Institute of Medical Science) reported the genetic regulation of Gal beta 1-3Gb4Cer: beta 1-6GlcNAc transferase. His group found that DBN2 mice lack the expression of Gal beta 1-4 (Fuc alpha 1-3) GlcNAc beta 1-6 (Gal beta 1-3) Gb4Cer (GL-Y) in the kidney. However, they express Gal beta 13Gb4Cer (GL-X), which is the precursor for GL-Y. A mating experiment indicated the involvement of an autosomal single gene (Gsl-5), suggesting that Gsl-5 controls the GlcNAc transferase activity. This hypothesis was supported by the measurement of the GlcNAc transferase activity and the analysis of GL-Y expression in backcross mice obtained on the mating including DBN2 mice. Gsl-5 gene was mapped 7.5 CM telomeric to Ly-1 on mouse chromosome 19. In the course of mapping experiment, another gene (Gs1-6), which controls the expression of NeuGc alpha 2-3Gal beta 1-3Gb4Cer and its disialoderivative, was found and mapped 1.9 cM centromeric to the Gsl-5 gene.
During the discussion, a question was raised on the possibility that the genetic regulation may not be fully responsible for the regulation of carbohydrate chain expression and that the modification of transferase activities through biosignals produced by growth factors and cell to cell and cell to matrix recognitions may be more important. Dr. Suzuki answered that the comment is quite reasonable but the genetic approach is quite useful to find genetic background, which builds up a genetically defined machinery for the production of carbohydrate chains, and fine regulations are achieved by many factors through the modulation of genetically defined machinery. Another question was raised on whether these polymorphic glycolipids are functionally involved or whether mice lacking the glycolipid have any impaired functions of the kidney. He answered that gross impairment in kidney functions have not been found so far, but it is still possible that DBN/2 mice have some weakness when they are exposed to unfavorable conditions.
The session on apomucin had two speakers. Dr. Young S. Kim (Veterans Affairs Medical Center and Univ. of Calif., San Francisco) stated that considerable progress has recently been made on cloning and sequencing of mucin genes in epithelial tissues from various species. Two predominant forms of mucin or mucin-like molecules have been described; 1) membrane-associated mucin- like molecules such as MUC1 mucin, which has a membrane spanning domain and, 2) secreted mucins which lack membrane spanning domain. Two distinct types of intestinal mucin cDNAs have been isolated and partially sequenced in his laboratory. These intestinal mucin genes have been designated as MUC2 and MUC3. The structural feature of this apomucin is the presence of central core of repetitive amino acid sequence (23 amino acids for MUC2 and 17 amino acids for MUC3). These tandem repeats in the control core region are rich in hydroxy amino acids which are potential O-glycosylation sites and are flanked on either side by unique sequences which also contain a very high density of potential O- glycosylation sites and also high cysteine residues. Both MUC2 and MUC3 genes are polymorphic and are located on different chromosomes CMUC2, chromosome 11 and MUC3, chromosome 7). Their recent studies on the expression of MUC1, MUC2 and MUC3 apomucin peptides and the levels of their message in various normal and cancerous epithelial tissues, indicate that mucin gene expression varies with epithelial cell types. These studies also demonstrate that malignant transformation is associated with increased expression and heterogeneity of immunoreactive mucin peptides. Thus, their data suggest that increased detection of core peptide epitopes in colon, stomach, and ed with normal tissue may be due to increased exposure of lung adenocarcinoma compar the core peptides due to impaired glycosylation rather than to an increase in mucin RNA synthesis.
Dr. Kohzoh Imai (Sapporo Medical College) prepared a new MAb, MUSE 11 using ascites from cancer patients as immunogen. MUSE 11 antigen is preferentially expressed in virtually all pancreatic adenocarcinoma tissues and the molecular weight of this antigen was estimated to be about 300 kDa. When the sandwich enzyme immunoassay using MAb MUSE 11 in the homologous assay system was used, the higher sensitivity was obtained in the sera of pancreatic cancer (60.5%) compared to normal (5.3%) or the patient with benign diseases (11.6%). Their data indicate that MAb MUSE 11 bound to the synthetic peptide corresponding to the tandem repeat unit of a human epithelial mucin core protein (MUC1).
Thus, it is clear that impaired glycosylation exposes the peptide core region of mucin molecules in cancerous tissues which sometimes enter the circulation. However, the mechanisms of the impaired glycosylation and how the membrane-associated vs. secreted mucins are processed in the cancer cells remain an area of future research.
In the first half of the session on cell adhesion, Dr. Reiji Kannagi (Aichi Cancer Center, Nagoya) presented data that sialyl Laa can also serve as a ligand for ELAM- 1. Findings in support of this include: l) A cultured human colon cancer cell line, Colo201, exhibits a clearly ELAM-1-dependent adhesion to IL 1 beta-activated human umbilical vein endothelial cells (HUVECs). 2) This adhesion is completely inhibited by several anti-sialyl Lea antibodies, but not by anti- sialyl Lex antibodies, although the cells express both antigens. 3) Pretreatment of HUVECs with either sialyl Lex or sialyl Lea glycolipid results in the complete inhibition of adhesion. They observed that as many as 14 out of the 15 cancer cell lines exhibited a clearly ELAM-1-dependent adhesion to HUVECs.
Dr. Hiroshi Nakada (Faculty of Engineering, Kyoto Sangyo University) presented his recent data on beta 1, 4 Galactosyltransferase (GT) which has been used as a marker enzyme for Golgi apparatus and located in the trans-Golgi cisternae of eukaryotic cells based on electron microscopic immunohistochemistry. Under certain conditions, however, GT has been found also on the cell surface and suggested to be involved in a number of important biological processes.
When LS180 cells, human colon cancer cell line, were cultured on matrigel coated plates, most of the cells formed densely packed colonies. With this colony formation, GT was induced remarkably. In the presence of the antibody or alpha-lactoalbumin in the culture medium, however, the cells failed to form the densely packed cell colonies, suggesting that GT on the cell surface is involved in the colon cancer cell adhesion. Electron microscopical study shows that GT was located on tightly associated surfaces of the adjacent cells but not on the surface both at the apical and basal parts of the cells. He concluded that GT induced by components of basement membrane played a significant role in colon cancer cell adhesion.
Dr. John L. Magnani (Gaithersburg, Maryland) examined the specificity of the E selectin (endothelial cell leukocyte adhesion molecule-1) for binding to a library of immobilized carbohydrate sequences as determined by a sensitive cell binding assay. Human E-selectin cDNA transfectants of mouse L cells were found to bind Sialyl Lea (NeuAc alpha 2-3Gal beta 1-3(Fuc alpha 1-4) GlcNAc) as well as or slightly better than the structural isomer Sialyl Lex (NeuAc alpha 2-3Gal beta 1-4(Fuc alpha 1-3)GlcNAc). A monoclonal antibody, HECA 452, which recognizes E-selectin ligands also binds both carbohydrate sequences. Hard sphere exoanomeric calculations were performed on these two hexasaccharides and indicate a three-dimensional domain common to both carbohydrate sequences which consists of the positioning of the NeuAc alpha 2-3Gal to the Fucose at the nonreducing terminal. Preliminary experiments indicate that some mucins containing this carbohydrate domain bind E-selectin transfected cells. By interacting with E-selectin expressed on endothelial cells lining the human vasculature, these mucins may interfere with the normal recruitment of leukocytes and contribute to the immunodepressed state observed in many cancer patients. Furthermore, tumor cells expressing high levels of Sialyl Lea/Sialyl Lex may have a distinct advantage initiating successful metastasis by specifically adhering to E-selectin on the endothelium and mimicking the normal trafficking of leukocytes through the vasuclar wall.
Dr. Masayuki Miyasaka (Tokyo Metropolitan Institute of Medical Science) reported their data on the role of LECAM-1 in lymphocyte homing. First, they cloned a cDNA encoding the rat LECAM-1 (L-selectin), and found that rat LECAM-1 contains an amino-terminal lectin domain, an EGF-related domain, and 3 short concensus repeats of C’ regulatory proteins instead of the two found in mouse and human LECAM-1. He then produced a soluble fusion protein of rat LECAM-1 and human IgG (LEC-IgG), and immunized a hamster with the chimeric protein to produce MAb against rat LECAM-1. MAb HRLI inhibited binding of PPME, a pseudoligand for LECAM-1, to lymphocytes, whereas MAb HRL2 did not, although both MAbs recognized LECAM-1transfected cells specifically. The use of LEC-IgG revealed that ligands for LECAM-1 are selectively accumulated in high endothelial cells (HEC) in lymph nodes and also the white matter of the CNS. A HEC-derived cell line, Ax, bound to LEC-IgG fixed on plastic plate, which was inhibitable with anti-rat LECAM-1 MAb, indicating that the ligand for rat LECAM- 1 is expressed on the Ax. The rat LEC-IgG immunoprecipitated from lymph node lysates 150 kDa as well as 250 kDa glycoproteins, suggesting that an as yet unidentified glycan structure(s) recognized by LECAM-1 is borne on a particular type of cell-surface glycoprotein(s).
In the second half of the session on cell adhesion, Dr. Roger Laine (Louisiana University and Glycomed Inc.) introduced the work on the quest of ELAM-1 ligands done by Dr. Brian Brandley in Glycomed Inc. Their group developed a modified method of TLC-immunostaining for detection of the ELAM-1 ligand carbohydrates. First they chromatographed the glycolipid mixture prepared from human neutrophils on TLC plates, and stained the plates with radiolabeled COS cells which were transfected with ELAM-1. The glycolipids in the resulting positive bands were purified and analyzed for their carbohydrate structures, and by doing this, they could identify the so-called VIM-2 glycolipid as an ELAM-1 ligand in human neutrophils. Dr. Laine showed that, by computed modeling analysis, the GlcNAc residue in the ligand can be Glc residue without affecting the binding activity to ELAM-1, and the last three carbons in NeuAc seem to be not necessary for the binding. He also took notice of the findings that complement-binding domains in ELAM-1 interact with phosphatidylserine and with sulfated glycoconjugates, and this interaction is not Ca2+-dependenL This interaction interferes the binding of the lectin domain to ligand carbohydrates.
Dr. John B. Lowe (University of Michigan School of Medicine) reported on cDNA cloning and characteristics of the human 3-fucosyltransferases. FucT III, which was the first cloned 3-fucosyltransferase, can synthesize Lex, sialyl Lex, Lea and sialyl Lea, and seems to correspond to the Lewis type enzyme. On the other hand, the cloned FucT IV (which is the same enzyme as ELFT) can synthesize only Lex (and VIM-2 when transfected to CHO cells), and this looks like the myeloid type enzyme. The newly cloned third 3-fucosyltransferase, FucT V, can synthesize Lex, sialyl Lex, and a very small amount of Lea, and this may correspond to the plasma type enzyme. The sequence homology between FucT III and FucT V was more than 95%, while that between FucT IV and FucT IV was less than 60%. Dr. Lowe also mentioned that his group recently cloned another 3-fucosyltransfemse, FucT VI, which can synthesize Lex and sialyl Lax. With regard to the ELAM-1 ligand, he showed that the VIM-2 glycolipid does not necessarily serve as a ligand for ELAM-1, while the FucT IV is the only 3- fucosyltransferase in human myeloid cells. Therefore, it was discussed whether some accessory molecules besides FucT IV are needed for myeloid cells to synthesize the effective ELAM-1 ligands.
In the session on metastasis, there were 3 speakers. Numerous factors have been proposed to affect the metastatic potential of tumor cells. In this session, expression of mucin with regard to mucin gene expression and with regard to carbohydrate antigens was discussed. A full-length cDNA of human mucin (MUC1) with human beta actin promoter was transfected in hamster pancreatic tumor cell line HP-1 by Dr. Richard Metzgar (Duke University Medical Center, Durham, NC). Two cell lines, HP4 and 5, both transfected with vector containing MUC1 gene clearly showed high tumorigenicity as compared with those transfected with vector alone or transfected with MUC cDNA encoding the sequence in the opposite orientation. HP-5, which has a 42-tandem repeat sequence, showed higher tumorigenicity than PH-4 which was transfected with a lower tandem repeat sequence. Growth rate of non-transfected and the transfected cell lines was nearly identical.
Dr. Tatsuro lrimura (Faculty of Medicine, The University of Tokyo and The University of Texas M.D. Anderson Cancer Center) applied blotting analysis with antibodies directed to MUC1gene product. Colorectal carcinoma specimens from patients at advanced stages contained higher levels of MUC1 than non-metastatic tumors. This is due to an increased mRNA level encoding MUCl. A low molecular weight factor producing enhanced mucin gene expression has been detected and termed “mucomodulin”. He also tried to determine if there was an increased expression of sialyl-Lex defmed by FH6 on mucins. This carbohydrate epitope is expressed more in metastatic deposits than in the original colon tumors. There is a clear difference in mortality between colonic cancers with high expressers of sialyl Lex versus high and low expressers of sialyl Lex This difference appears to be ascribable to the interaction of tumor cells with activated endothelial cells expressing ELAM-1 as was discussed extensively in the previous section, particularly by Dr. Reiji Kannagi.
Dr. Sen-ichiroh Hakomori (The Biomembrane Institute, The University of Washington, Seattle, WA) discussed two major topics related to metastatic potential. The first was the role of glycolipid-glycolipid interaction in tumor cell adhesion on non-activated endothelial cells in a dynamic flow system. The second topic was about various factors affecting selectin-dependent adhesion. Melanoma adhesion on endothelial cells based on GM3 (expressed on melanoma) and LacCer (expressed on endothelial cells) interaction predominates over lectin- or integrin-mediated adhesion in a dynamic flow system. This is probably because integrin-dependent adhesion is a slower process than the carbohydrate-mediated process.
Cell adhesion through E- or P-selectin involves both sialyl Lex (S-Lex) and sialyl Lea (S-Lea) P-selectin binding to S-Lex and S-Lea was inhibited by sulphated glycan and both E- and P-selectin mediated adhesion were greatly diminished by preincubation of target cells with henzyl-alpha-GalNAc which inhibited chain elongation of O linked sugars but not N-linked sugar chains. N- glycosylation modifier did not affect selectin-dependent adhesion. In addition, he discussed the involvement of N,N-dimethyl-sphingosine or other sphingolipid catabolites in inhibiting selectin expression at the cell surface. Thus, a possibility was discussed that if selectins were indeed involved in tumor cell metastasis, metastatic potential might be inhibited by not only selectin epitope or antibodies but also by O-glycosylation modifier and selectin expression inhibitor.
The session on Immunology began with a presentation of data by Dr Ralph A. Reisfeld (The Scripps Research Institute, La Jolla, CA), which emphasized the potential of genetically engineered anti-ganglioside GD2 antibodies for treatment of neuroblastoma and melanoma GD2 was shown to be excellent for antibody and complement mediated killing of these tumor cells in part, because of the large number of GD2 epitopes (14X106/cell) and also because the application of anti-GD2 antibodies tends to potentiate immune responses to tumor cells, Both a murine anti-GD2 antibody and a gemnetically engineered human/mouse chimeric form of this antibody were highly active in killing tumor cells by ADCC, particularly when granulocytes, stimulated by GM-CSF, were used as effector cells. In clinical phase I trials of neuroblastoma patients, both mouse and human/mouse chimeric antibodies were effective, as six complete, long-term remissions of bone marrow metastases were achieved among 16 patients treated thus far. Further genetic engineering resulting in deletion of the CH2 domain of the chimeric anti-GD2 antibody (ch 14.18) produced a construct, which is potentially useful for tumor imaging. Specifically, this mutant maintained its full antibody activity, has reduced immunogenicity, is cleared rapidly from the circulation, and has increased targeting specificity. Finally, a genetically engineered antibody fusion protein between chl4.18 and IL-2 was shown to target this cytokne and to stimulate killing of autologous GD2+ tumors by tumor infiltration lymphocytes (TIL). An optimal tumor therapy might combine IL-2 activation of immune effector cells (TIL) and presentation of large amounts of tumor antigen together with a tumor specific antibody like chl4.18 that also mediates both complement-dependent cytotoxicity and antibody-dependent cellular cytotoxicity.
In summary, genetically engineered anti-tumor antibodies already facilitate improvements in the immunotherapy of neuroblastoma and will likely lead to further advancements in the therapy of a variety of solid human tumors.
Next Dr. Michael Longenecker (University of Alberta and Biomira, Inc., Edmonton, Canada) developed a rationale for active specific immunotherapy (ASI) of carcinoma patients using, as immunogens, synthetic carbohydteradte (TF and S-Tn) and peptide (MUC l) antigens expressed on carcinoma-associated mucus. The “proof of principle” that synthetic carbohydrate and peptide antigens may be useful for ASI of carcinoma patients was first established in an animal model, i.e. mice bearing a highly lethal and transplantable mammary carcinoma cell line (Ta3-Ha). Such cells secrete and express large amounts of Epiglycanin. a cancer-associated mucin carrying multiple Thomsen Friedenreich (TF) epitopes. Furthermore, Epiglycanin was shown to be immunosuppressive when injected i.v. into mice and even generated specific T-suppressor cells in these animals which could, however, be inhibited by low dose cyclophosphamide. More important studies In this mouse model revealed an effective “vaccine” formulation consisting of synthetic TG hapten conjugated to keyhole lympet hymocyanm (KLH). When such a “vaccine” was injected together with the RIBI adjuvant DETOX, followed by low dose cyclophosphamide, it provided long-term (120 days) survival of mammary carcinoma bearing mice. In contrast, control mice died within 21 days. Furthermore, MUCI based synthetic vaccines were found to be effective for ASI of mice bearing mammary carcinoma cells that had been transfected with cDNA encoding MUC1.
Based on these encouraging results in animal models, phase I clinical trials were initiated in ovarian cancer patients to test the efficacy of ASI with synthetic carbohydrate “vaccine” formations. Ovarian cancer patients with extensive metastatic disease were immunized with TF-KLH, plus DETOX adjuvant following low dose cyclophosphamide. In another arm of this clinical phase I trial, the same protocol was followed with synthetic sialyl Tn-KLH. Both “vaccines” proved to be non-toxic and highly immunogenic, as they stimulated formation of cytotoxic anti-cancer antibodies and cellular immune responses. The initial results of the clinical phase I trial were considered sufficiently encouraging to test the efficacy of a similar "vaccine" formulations in a phase II trial.
Dr. Olivera Finn (University of Pittsburgh School of Medicine) discussed the expression of tumor-associated mucin epitopes on Burkitt’s lymphoma and EBV- immortalized B cell lines, transfected with MUC1 and their recognition of mucin- specific cytotoxic T-lymphocytes (TCL) Initial work by Finn et al. showed that the epitope recognized by these CTL is present on mucin produced by malignant breast and pancreatic cells but not on mucin produced by normal cells. Also, this CTL epitope is closely related to a tumor-specific mucin epitope defined by MAb SM-3. A major focus of the studies de scribed was to assess the relative importance of genetic changes versus incomplete glycosylation in the generation of tumor-specific mucin epitopes. An expression vector (pDKOF.muc1) was constructed to facilitate mucin production in EBV-immortalized B cell lines and in Burkitt’s lymphoma Maintenance of the mucin gene was shown after seven and eight weeks in culture and this highly multivalent form of the MUC1 gene was indeed stably maintained. Of seven cell lines transfected in this manner each expressed a number of tumor-specific mucin epitopes recognized by specific MAbs These findings suggest that defects in mucin synthesis in breast and pancreatic tumors are also seen in other transformed cells. Also, specificity of the mucin epitopes for tumors is not due to alterations in the mucin gene, since each cell line, transfected with the same expression construc~ expressed a different subset of tumor-specific epitopes. The role of incomplete glycosylation in the tumor specificity of these epitopes was confirmed, since an inhibitor of O- linked glycosylation (phenyl-N-acetyl-alpha-galactosaminide) increased the expression of tumor-specific epitopes. Convincing evidence was presented that glycosylation-inhibited transfectants maintained proliferation of mucin specific T cells and served as targets for CTL killing. Taken together, the data presented by Dr. Finn suggested that expression of novel epitopes may be a consequence of altered translational modification of normal gene products. Findings such as these may provide for a better understanding of tumor immunology and of autoimmune disease processes.
The last presentation of the session on Immunology was by Dr. Anil Singhal (Biomembrane Institute and University of Washington, Seattle, WA). He discussed a model for active immunotherapy of mammary carcinoma, i.e. the induction of the Tn antigen-mediated T cell responses in mice. Specifically, a block in carbohydrate chain elongation of O-glycosylated mucins causes the accumulation of alpha-GalNAc O linked to serine or threonine, i.e. Tn antigen, in many human adenocarcinomas. Immunization of mice with desialylated ovine submaxillary mucin (A-OSM), which contains much Tn antigen, was shown to protect the animals against a challenge of highly invasive, Tn+ syngeneic mouse mammary tumor, Ta3- Ha. Furthermore, immunization with A-OSM, but not with deglycosylated mucin produced a high anti-Tn antibody response in mice, indicating that protection was imported by the carbohydrate Tn antigen. The tumor immunity evoked by the Tn antigen is thought to be due to the generation of humoral and/or cellular immune response, since high serum anti-Tn titers of both IgG and IgM isotype were detected in A-OSM immunized mice. This study also showed that Tn-ontaining A- OSM caused in vitro T cell proliferation in a dose-dependent fashion; however, this proliferation was lower when deglycosylated OSM core protein was used, indicating that the T cells responded to the carbohydrate epitope. Yet, the possibility cannot be ruled out that carbohydrates confer a preferred conformation for the protein core to interact with T cells. Specifically, since many mucins contain Ser-Thr-Thr sequences with Tn and Th epitopes attached, it is possible that peptides containing these sequences interact with T cells when Ser-Thr-Thr sequences are glycosylated with Nacetyl-galactosamine (Tn antigen). Indeed, on deglycosylation of OSM, optimal conformation may be lost, leading to decreased interactions of peptide sequences with T cells. Additional data were presented which indicated that synthetic Tn antigen coupled to different proteins as backbone was also effective in preventing tumor growth in mice. Taken together, these data indicate that carbohydrate Tn antigen can bind to T cell receptor and stimulate T cells, providing a basis for an effective model for active immunotherapy of cancer.
In conclusion, data presented in the session in Immunology indicated that monoclonal antibodies directed against carbohydrate antigens, such as disialoganglioside GD2, can be modulated by genetic engineering to provide useful reagents for the treatment of neuroectodermal derived malignancies. Furthermore, it is evident from preclinical studies, as well as from initial results of a phase I clinical trial, that active specific immunotherapy of adenocarcinoma can be achieved with synthetic carbohydrate, i.e. Thomsen- Friedenreich (TF) and sialylated Tn antigens, as well as with selected peptide epitopes of MUC1. The importance of mucins as tumor-associated antigens was demonstrated by the finding of cytotoxic T lymphocytes (CTL) that specifically recognized epitopes on mucin produced by malignant breast and pancreatic cells, but not on mucin produced by normal cells. Transfection experiments with an expression vector containing the MUCI gene indicated that the specificity of different subsets of tumor-specific epitopes expressed on several EBV transformed cell lines and on Burkitt’s lymphoma is not due to genetic alterations in the mucin gene. This finding was attributed to incomplete glycosylation, since an inhibitor of 0-linked glycosylation increased the expression of tumor-specific epitopes on these cells that were transfected with the same MUC1 expression construct.
SEMINAR AGENDA AND PARTICIPANTS
(1) Seminar on “Receptor-mediated Modulation of Lymphocyte Growth and Activation”
San Diego, CA January 13-15, 1992
AGENDA