(1) Seminar on "Molecular Mechanisms of Cell:Cell Interactions Involved in Tumor Immunity"
I. The Immunology and Molecular Biology of Tumor Antigens
In the two sessions devoted to this topic area, a number of presentations were devoted to descriptions of cell surface antigens expressed on malignant cells. In particular, the molecular biology underlying the expression of these antigens and their relationship to the host immune response to tumors was emphasized. Dr. Tokunaga described the identification of "FT" antigens ex-pressed by embryonic mouse thymocytes which were also found on the surface of a number of mouse leukemias of diverse etiology. A biochemical characterization and study of ontogeny of expression of these antigens in normal thymocytes was reported. In addition, preliminary studies were described involving the use of these FT antigens as immunogens for tumor protection. Dr. Weissman outlined a hypothetical model for leukemogenesis in which viral receptors expressed on lymphocytes served as the receptors for infectious leukemogenic agents. Evidence in support of this hypothesis was described and included the existence of virus binding receptors on both T cell and B cell leukemias.
The relationship of IL-2 receptor expression to malignancies of lymphoid origin was discussed in presentations by Drs. Yodoi, Warner Greene, and Sugamura. Dr. Yodoi presented evidence that infection with human T cell leukemia virus (HTLV-I) leads to the activation of the endogenous IL-2 receptor TAC gene with resulting constitutive expression of that receptor on adult T cell leukemia cells. He further suggested that IL-2-independent growth of such cells resulted directly or indirectly from this event. Dr. Warner Greene presented detailed evidence concerning the genetic structure and sequence of the gene encoding the human IL-2 receptor in HTLV-infected cells. The signals regulating IL-2 receptor expression in normal T cells were characterized, and a model was discussed to explain the constitutive expression of the IL-2 receptor in HTLV-infected cells. Dr. Sugamura extended these findings by providing evidence for the ability of HTLV to induce IL-2 receptor expression in B cells as well as T cells. In addition, he demonstrated that, among a series of T cell lines, recombinant IL-2 was capable of stimulating growth in certain T cell lines while inhibiting growth in others. The mechanism underlying this difference is not yet understood.
Dr. Schreiber described his work characterizing unique tumor-specific antigens expressed on UV-induced tumors in the mouse. The use of highly efficient cytotoxic T cells as a probe for selecting antigenic variants has led to the finding that specific antigen loss variants can be selected in vitro. Moreover, stereotyped hierarchies of tumor antigen expression are demonstrated by this approach. Evidence was presented suggesting a relationship between these antigens and relatively public determinants expressed on class I MHC antigens. Dr. Mark Greene described experimental findings involving the isolation of an onc gene from rat neuroblastoma. This gene, termed the neu gene, was assayed by transfection of a 3T3 line. Evidence was presented that transfection with the neu gene resulted in both malignant transformation and the expression of a detectable 185-kDa cell surface protein. Antibodies directed at this transformation-related cell surface glycoprotein had a number of effects on in vitro tumor cell growth, anchorage-independent growth, and tumor cell differentiation. These studies suggested the possibility that this identified cell surface determinant was involved in the regulated growth and differentiation of the malignant line. Dr. Croce described his studies of the genetics of human B line neoplasms. His findings were interpreted in the context of a two-step model of B cell transformation in which the first step represented a chronic stimulus to B cell proliferation and the second step involved a translocation or other chromosomal change giving rise to a "high-grade malignancy. "
II. Biology of Host-Tumor Interactions
The nature and identity of T cells specific for tumor-related antigens were discussed by Drs. Kikuchi and Fujimoto. Dr. Kikuchi described the isolation of tumor-infiltrating lymphocytes from rats, their cell surface phenotypic characterization, and their successful long-term culture in vitro. In vitro cytotoxic T cell populations as well as long-term cultured T cells active in a Winn assay were isolated from these populations. In addition, the function of T cell-derived lymphokines in cytotoxic T cell differentiation was described. Dr. Fujimoto characterized the T cells active in both in vitro cytotoxic assays and in vivo Winn assays with specificity for the mouse S1509a tumor. Distinct populations mediating MHC-unrestricted versus -restricted cytotoxic activity were characterized, and a number of transformed CTL clones specific for different class I-related antigens on the S1509a tumor cells were described. In terms of their activation requirements, it was concluded that tumor-specific CTL may be different from alloantigen-specific CTL.
The characterization of lymphokines was discussed at the molecular biologic, biochemical, and functional levels by several speakers. Drs. Vitetta and Kishimoto discussed lymphokines active in the regulation of B cell activation. Dr. Vitetta reported characterization and purification of the lymphokine termed BCGFI, now termed BSFp1. Evidence was presented suggesting that this factor is not a growth factor but rather a differentiation factor rendering B cells responsive to subsequent growth-inducing stimuli. Dr. Kishimoto described studies of BCGFII effects on human B cell lines. Monoclonal antibodies against BCGFII appeared to inhibit BCGFII-induced human B cell line colony formation but had no effect upon stimulation of the same B cells by LPS, suggesting different activation signals in these two situations. Human BCDF purification was also described. The transfection of B cells with the c-myc gene was successfully accomplished and demonstrated to induce transient responsiveness to BCGF, suggesting that c-myc transfection can substitute for physiologic stimuli such as anti-immunoglobulin antibodies in the activation of B cells.
Dr. Taniguchi reported his most recent findings in characterizing the genes encoding beta interferon (IFN-!
!!) and IL-2. Both native and deletion mutant forms of the human IFN-!!!gene have been transfected and expressed in mouse cells. Similarly, the murine IL-2 gene has been employed to transfect human T cells. Okada described a "co-stimulator" distinct from IL-2 which functioned in the generation of cytotoxic T cells in response to UV-irradiated stimulating cells. The study of a number of human T cell hybridomas suggested the existence of a new killer helper factor which is endogenously produced by certain cytotoxic T cell hybridomas and may be required for their cytotoxic differentiation. Hashimoto described a lymphokine produced by activated rat lymphocytes which was capable of inhibiting proliferation in a variety of cell types including T and B lymphocytes. A human analog of this same lymphokine was also described. Kumagai described the activation of killer cells in human peripheral blood lymphocytes by, recombinant IL-2 alone and characterized the ability of these activated killers to lyse lymphoid and nonlymphoid tumor lines. IFN-!
!!secretion was induced by IL-2 under these circumstances, and antibody to IFN-!!!suppressed cytotoxic T cell generation in response to IL-2, suggesting a role for IFN in activation of killer cells. Recent advances in the use of molecular biologic approaches to the study of T cell activation and receptor specificity were described in presentations by Drs. Davis, Burakoff, Hodes, Germain, and Fathman. Dr. Davis described the isolation and composition of the murine T cell receptor!
!!and!!!genes. The homology of these genes with immunoglobulin structures was discussed, and the unique features of these T cell recognition structures were outlined, including a description of V region amino acid sequences and gene expression in thymocyte populations as well as in a series of T cell lines and hybridomas. Dr. Burakoff reported the use of cloned and "shuffled" MHC class I genes in transfecting target cells for T cell recognition. The nature of the domain structure involved in T cell recognition of these cell surface class I determinants was discussed. Dr. Hodes described the successful microinjection of a mouse embryo with a cloned swine class I MHC gene and the establishment of a stable transgeneic mouse line. The expression of the transfected gene product in these mice was characterized, and the ability of these transfected gene products to serve as transplantation antigens both in vivo and in vitro was described. Dr. Germain has conducted extensive studies involving the transfection of L cells with native, mutant, and shuffled murine MHC genes. Transfection with class II (Ia) genes resulted in the expression by L cells of la determinants capable of functioning in antigen presentation. Through the use of mutant and exon shuffled class II genes, the nature of critical elements of domain structure in the cell surface molecules was analyzed. Fathman reported the result of the studies in normal and mutant strain mice which allowed a comparison of la sequences and a determination of critical molecular structures involved in antigen presentation. Studies of human HLA-D region products were also reported and led to speculation concerning the relationship of CTL-identified determinants in relation to human autoimmune disease processes. Dr. Singer described the repertoires of functionally distinct class I- and II-specific T cells. In particular, the relationship between class I-specific helper and killer T cell populations was described and contrasted, leading to the proposal of alternative hypotheses to explain the apparently distinct repertoires of these two functionally different T cell subpopulations.
III. Preclinical Approaches to Tumor Immunotherapy
Dr. Hamaoka described a regimen employing tumor-specific helper T cells in in vivo tumor resistance. After pretreatment to eliminate suppressor cells and priming of the host to hapten, unmodified tumor challenges could be treated successfully through intratumor injection of the specific hapten. In adoptive transfer experiments, the cells involved in tumor protection were found to be Lyt1+2- T cells which work independently of other T cell populations via the activation of macrophages with antigen-nonspecific anti-tumor effector function. Muramatsu described both direct and indirect effects of interferon on a variety of tumor cells. It was found that different tumor lines were either sensitive or resistant to the direct effects of interferon. Drs. Vitetta and Fathman described the use of immunotoxins. Vitetta outlined a strategy for the use of ricin-coupled tumor-specific antibodies in vitro and in vivo. Fathman described the use of an analogous approach in the treatment of autoimmune processes by the elimination of autoantigen-reactive T cells by ricin-coupled antigen. Application of these principles to transplantation problems was also outlined.
(2) Seminar on "Diagnosis"
The US-Japan Cooperative Cancer Research Program Conference on "Tumor Markers and Their Genes" was held on January 14-16, 1985, in Building 31 at the NIH in Bethesda, Maryland, USA. The initial meeting on Monday, January 14, 1985, was utilized to introduce the participants and exchange presentation abstracts and reprints on topics of interest. Following review of these materials and the initial informal, productive exchanges between participants, the program began the following morning.
Dr. Tom Vogt of Fox Chase Cancer Center in Philadelphia led off the first session with a summary of his and Dr. Shirley Tilghman's work on alphafeto-protein (AFP). Dr. Vogt presented the tandemly linked genomic organization of the albumin and AFP genes on mouse chromosome number 5. The amino acid homology of these products is 32%, but the nucleotide homology is fully 50%. They constructed an AFP minigene from genomic sequences by retaining 5' and 3' potential regulatory regions but deleting an internal region so that the RNA product was distinctive (YZE). This construct was introduced into F9Ec terato-carcinoma cells. The tissue specificity of expression was examined by inducing visceral versus parietal differentiation of this cell line. Five of 11 transformants showed selective induction in only the visceral endoderm, demonstrating the same qualitative pattern noted in normal differentiation. Further deletion mutant constructs have been generated, and losses 7 kb 5' to AFP are functionally regulated, whereas those only 1 kb 5' to the gene are not functional. He then presented further new data on the expression of introduced AFP in transgenic mice. They noted the proper tissue-specific expression in mouse fetus and neonatal liver, and it was appropriately shut off at days 3, 7, and 14. Overall, they noted an interesting correlation between multiple integration and the inducibility of the YZE RNA. This and the pattern of regulation suggest the presence of 5' flanking DNA sequences required for activation. This work represents a pioneering step in dissecting the expression of oncofetal proteins in early development.
Dr. Shinzo Nishi presented his group's detailed analysis of the rat AFP cDNA. Their clone, RA1, covered almost the entire rat mRNA. The cDNA was 1.9 kb in size and contained 5' flanking, ATG, and AATAA 15 bp upstream of the poly A tail. The first 18 amino acids are felt to the be signal peptide. Considerable homology with the mouse was noted, and the data are consistent with up to 15 exons again. Their cross species comparison suggests that the common functions of fatty acid binding may be encoded in domain III, while disparate functions such as estrogen binding may related to domain I or II. To analyze functions, an expression vector construct was made in pMC 1513, which generates a galactosidase fusion product. About 20% of colonies produced AFP by RIA and Western blotting. They also created a full-size AFP by recloning and once again putting it into a lac Z fusion vector. Dr. Nishi even noted independent production of AFP when put into pKK 223.3 vector. These different constructs are of great importance in allowing this group to cleanly dissect the location of functional epitopes in this molecule that are responsible for estrogen binding, fatty acid binding, and other functions.
Dr. Kazya Higashino then turned our attention to the human alkaline phosphatases (ALP). There are three classes of ALP which are encoded by three different genes. The biochemical differences are well characterized as are their developmental expression; however, in carcinomas the products of all three genes reappear. Dr. Hagashino and his colleagues found an ALP isoenzyme, kasahara, in hepatomas that was an isoenzyme of the intestinal class. Two variants were noted which varied only in their neuraminic acid content. The kasahara isoenzyme proved to be much more restricted to only the liver disease patients with cancer. However, while specific, its prevalence in hepatomas is only up to 30%. This particular tumor marker does show promise because of its remarkable specificity to hepatic cancer. Dr. Higashino will be seeking to improve its sensitivity by generating type-specific monoclonal antibodies for use in an RIA.
One of the most provocative presentations of the meeting was that of the pancreatic secretory trypsin inhibitor (PSTI) by the Osaka University group of Drs. Mori, Ogawa, and Nakamura. PSTI is an immunoreactively distinct inhibitor which is significantly elevated in patients with pancreatic and gastric carcinoma and some with esophageal, hepatoma, and thyroid cancer. By immunoperoxidase staining PSTI was found in normal acinar cells, foveolar epithelium of stomach and kidney tubule. Staining of malignant tissues and liver tissue surrounding hepatocarcinomas was dramatic. This group then successfully screened a pancreatic cDNA library with a 14-mer mixed probe (16-fold degeneracy) generated from a partial amino acid sequence. The structure looked like a secretory peptide with signal peptide sequence and a further 168 bp. Strikingly, this gene had a 46% homology with epidermal growth factor. This remarkable finding raises the distinct possibility that this peptide may be serving functions other than preventing the trypsin-catalyzed premature activation of zymogens. Perhaps this represents a tumor growth factor that might even be functioning in an autocrine-like mode. This group now has the necessary tools to dissect this important issue.
Dr. Irvine Boime presented an analysis of human placental lactogen (hPL) and chronic gonadotropin (hCG) expression during placental development and in tumorigenesis. The subunits are encoded by a single gene with a 5' enhancer region and 3' polymorphic sites for HindIII and EcoRI. Dr. Boime has exploited this restriction fragment length polymorphism to examine inheritance in tissues. He demonstrated that hydatidiform moles have a homozygous male pattern. Curiously, four of six choriocarcinomas had a rare polymorphism which is rather striking. Moles and choriocarcinomas will be followed prospectively for this polymorphism, as hCG expression may be linked to tumorigenicity. Developmentally, the hCG and subunit decrease at term pregnancy, while hPL remains constant. Furthermore, Dr. Boime demonstrated differences in placental hormone expression at different stages of trophoblast differentiation. His results suggested that hCG!
!!and!!!expression may be linked to differentiating cyto- trophoblast elements, while hPL expression requires a certain level of trophoblast differentiation. Dr. Boime's work nicely illustrated the analogous portions of choriocarcinomas and hydatidiform moles with their normal counterpart areas of placenta. Dr. Takashi Muramatsu presented a detailed description of cell surface glycoprotein markers in teratocarcinomas. He has concentrated on the carbohydrate markers carried on embryoglycan. He has a series of monoclonal anti-bodies: SSEA 1 (Fuc!
!!1-3GleNAc), ECMA-2(!!!-Gal) ECMA-2, and various lectin receptors. Patients with germ cell tumors were found to frequently have anti-body reactive with embryoglycan, and Dr. Muramatsu has used these to follow their clinical courses. He has also produced monoclonals against the receptors for Ficinus communis agglutinin-1 of teratocarcinoma cells. Three oncofetal antigens were detected: brushin, TC antigen, and OR17. These were scarcely detected in adult murine tissues except for the kidney. These studies were an important demonstration of the tumor relevance of very basic carbohydrate biochemistry. Dr. Takehiko Tanaka of Osaka University detailed the expression of the four types of pyruvate kinase isoenzyme in oncofetal tissue. The four isoenzymes fell within two groups: L (L and R) and M (M1 and M2). Type M2 proved to be most widely distributed in adult tissues, fetal tissues, and in hepatic regenerating nodules. Dr. Tanaka also noted different patterns of isoenzymes in poor, well, or highly differentiated rat hepatomas. His group has recently cloned the cDNAs for type L, M1, and M2. They had determined 92% of the sequence of M2 and 60% of M1. Curiously, the M2 mRNA was 3-fold increased following partial hepatectomy, but about 350-fold elevated in Yoshida ascites hepatoma. No gene amplification had occurred, and gene run-off experiments indicated an increased transcription rate. Excitingly, the initial data on the M1 and M2 sequence look as if there may be an alternative form of splicing which places different internal exons into the final M1 and M2 product. This would prove to be a very important and novel mechanism of molecular switching of gene expression. Dr. Tanaka thus has developed an extremely important system in understanding the regulation of such a developmental program.
Dr. Adi Gazdar of the Naval Medical Oncology Branch of the NCI concluded the first day's scientific session with an indepth presentation of their work characterizing small cell lung cancers (SCLC). He detailed their ability to start cell lines from the various histologic types of lung cancer. The SCLC lines were of APUD or neuroendocrine origin. These were then divided into two classes: classic SCLC which express high levels of L-dopa decarboxylase (DDC), bombesin (BCI) neuron-specific enolase (NSE) and BB isoenzyne of creatine kinase (CK-BB), and the variant SCLC which have no DDC or BLI. However, six of eight variants showed increased levels of myc RNA. This was usually due to an amplification which can be detected with classic c-myc or n-myc or this group's recently described L-myc. These lines also contain double minutes and associated homogenous staining regions. The correlation of myc expression and the variant morphology suggest a role for myc in the conversion that is seen in SCLC.
Dr. Richard Metzgar of Duke University initiated the presentations on the second full day of the conference. He presented a very lucid review of their dissection of the antigens of normal and malignant human exocrine pancreatic cells. He has produced a series of monoclonal antibodies (DU-PAN 1,2,3,4,5) to the human pancreatic tumor cell line (HPAF). One of the antigens (DU-PAN-2) was on normal pancreatic ductal epithelial cells and showed a particularly restricted distribution on tumor cells. This proved to be a heavily glycosylated antigen that was rather polydisperse and sensitive to neuraminidase and alkaline reduction. Taro broad bands were seen on immunoprecipitation and immunoblotting. Therefore, it is felt that the DU-PAN-2 epitope is likely on a mucin-like molecule. A competitive RIA has revealed the presence of detectable antigen in serum and ascites of patients with pancreatic gallbladder and stomach cancers. This antigen is also proving useful in following the clinical course and response to therapy of pancreatic cancer. Once again, these antigens appear to be oncofetal or differentiation associated antigen. Differences in expression and tissue distribution were demonstratable in fetal versus adult tissues. Moreover, the competitive RIA could even demonstrate a fall in antigen levels from cord blood to the post natal period. This set of antigens that Dr. Metzgar's group has defined is proving of interest not only for basic glandular development but also for its clinical utility.
The antigenic variations on the sugar moieties of CEA molecules were described by Dr. Yuji Matsuoka of Fukuoka University. There are a number of related antigenic glycoproteins which crossreact with most anti-CEA antibodies. These consist of the normal fecal antigens (NFA 1,2), normal fecal crossreacting antigens (NFCA), and NCA-2 of meconium. Recently, Dr. Matsuoka's group produced over 300 monoclonal antibodies in order to better distinguish their specificities. He concentrated on five monoclonals: F3-30 was anti-NFA-1; F4-82 was anti-common determinant; F4-11 was against a determinant on some groups of CEA molecules; F8-52 was anti-NFA-2; and F48-60 was anti-NCA-2. F4-11, F8-52, and F48-60 recognized sugar moieties on the CEA molecules. Curiously, the F4-11 antigen determinant was expressed only on some CEA molecules (3 of 20 antigens and approximately 40% of patient sera), raising the possibility that it might be allotypic. When over 60 patients sera were examined, the F3-30 and F4-82 showed quite universal reactivity comparable to the conventional polyclonal antibody testing. The other three antibodies showed variable patterns. This group is further pursuing the genetic implications of the heterogeneity in carbohydrate epitope expression as well as refining the specificity and utility of CEA as a clinical tumor marker.
Dr. Ann Thor of the Laboratory of Tumor Immunology and Biology of the NIH discussed their studies of human carcinoma cell surface antigens and oncogene products that were detected by monoclonal antibodies. She presented work with their monoclonal B723 which recognizes a 220- to 400-kDa ductular antigen glycoprotein complex with selectivity for human colon and breast carcinomas. This has been particularly effective as an adjunct to cytology preparations on effusions and has shown good specificity for those with carcinoma. She then presented some striking staining of human tissues with monoclonal antibodies against synthetic peptides (aa 10-17) of Hu-ras T24. Enhanced ras p21 protein was evident in the majority of mammary and colon carcinomas by immuno-peroxidase techniques. A dramatic example of a crypt was shown in which the cells at the surface were negative while p21 became increasingly intense as one sectioned further into dysplastic and frankly neoplastic areas of the same biopsy. Most ductal carcinomas were positive (19 of 30), while zero of 11 fibrocystic disease but 2 of 10 recurrent fibroadenomas expressed p21. Antibodies to such oncogene products promise to improve our understanding of neoplastic progression on an in situ basis.
Dr. Alonzo Ross of Dr. Hilary Koprowski's lab at the Wistar Institute in Philadelphia presented their exciting work on the nerve growth factor (NGF) receptor. This group had prepared a large number of monoclonal antibodies to melanoma cells. Two of these antibodies, ME 82-11 and ME 20.4, blocked the binding of iodinated NGF to cells. These antibodies immunoprecipitated a 75-kDa glycoprotein with serine residue phosphorylation and both N- and 0-linked glycosylation. Purified receptor was prepared by lectin affinity chromatography and immunoaffinity columns, and 30 N-terminal amino acids were determined. Meanwhile, the antibodies had been successfully used to screen L cells transfected with high molecular weight DNA. Secondary transformants expressing the NGF receptor are now available and hybridize with an Alu sequence probe and are being checked with an oligonucleotide probe derived from the amino acid sequence information. Thus, the analysis of melanoma-associated antigens has provided a tool for examining the molecular biology of a growth factor receptor. Such findings will have importance that transcends its initial application to melanoma.
The final section of the meeting covered leukemia and lymphoma biology. Dr. Ryuzo Uedo of the Aichi Cancer Center focused his presentation on three groups of antigens of the many monoclonal antibodies they had prepared to hematopoietic tumors. He discussed three Null acute lymphoblastic leukemia-associated antigens, NL-22, NL-1, and HL-47. NL-1 appeared to be another antibody recognizing CALLA (g p100). NL-22 recognized 58% of cases; HC-47 recognized 48% and may be a new antigen. Neither showed cytotoxicity against CFU-C or BFU-E and are being employed for autologous bone marrow transplantation of Null-ALL patients. Two antibodies Ta60a and Ta60b recognized the human IL-2 receptor. Ta60a sees apparently the same epitope as anti-Tac and thus blocked the binding and function of IL-2, but Ta60b did not. Ta60b recognized all adult T cell malignancies, but not other T cell neoplasms. He also generated an antibody to a pan-T antigen on pre-T ALL, Tp40. Large fetal thymocytes were strongly positive as were all T-ALL and lymphoblastic lymphomas. They have successfully transfected HPB-ALL DNA into LtK- cells and have stable secondary transformants that express the Tp40 antigen. Thus, the molecular cloning of a number of these leukemia-associated antigens is under way.
Dr. Lee Nadler discussed his utilization of monoclonal antibodies prepared against the cell surface determinants of B cell malignancies to determine the road map of B cell differentiation. He has two remarkably B cell lineage-restricted monoclonals, B1 (pp35) and B4 (p40/80). He finds the primary follicle bears HLA-DR, B4, B2, B1, and lgM/IgD, but upon activation the germinal center cells change to IgG, lose B2, and begin to acquire T10, PCA-1, and PC-1. Correspondingly, nodular poorly differentiated lymphomas are 100% B4, 69% CALIA, HIA-DR+, and B2+, while diffuse PDL is CALLA¯. The B1+, B2+ (C3d receptor ) cells respond to mitogens. This work is aimed at determining the normal developmental sequence of cell surface antigen expression and its function and identifying the cells’ neoplastic counterparts.
Dr. Michael McGrath presented his work investigating whether cell surface antigen receptor on T cells and B cells also served as a receptor for transforming viruses. He reviewed the data for the murine BCL1 tumor, where an anti-idiotype blocks the binding of the env gene product to that tumor.
Similarly, Dr. Allison's anti-clonotypic antibody for T cells recognizes the T cell receptor and blocks MuLV binding. Anti-chicken immunoglobulin will block the binding of ALV to bursal lymphomas. Most recently, they have pre-pared au N-terminal synthetic peptide for the V!
!!chain of the Molt-3 cell line. They have an antibody (M3) that immunoprecipitates the!!!chain of the receptor and is expressed on only 0.5% of thymus cells. However, eight of nine human T cell malignancies are positive, and M3 blocks the binding of HTLV. The receptor-mediated leukemogenesis theory is a very intriguing hypothesis for which some supportive evidence does exist. Importantly, the molecular tools now exist to allow it to be directly confirmed or its role lessened in a number of lymphoid neoplasms. Dr. Stanley Korsmeyer of the NCI closed the meeting with a discussion of how immunoglobulin and T cell receptor gene rearrangements serve as molecular markers of clonality, lineage, differentiation, and translocation. This group detailed how the immunoglobulin genes were sequentially rearranged, heavy before light and!
!!before!
!!, during early B cell precursors. Because of the stochastic process of immunoglobulin gene joining, the pattern of gene rearrangement in each malignancy is unique and serves as a tumor-specific marker. This has revealed evidence of clonal evolution with ALLs, CMLs, and lymphomas. In addition, these rearrangements are sensitive markers, capable of identifying even minority populations (1-5%) of clonal cells within tissues of mixed cellularity. This approach has detected the persistence of large clonal populations of cells in bone marrows of patients felt to be in remission from ALL. Dr. Korsmeyer's lab has recently exploited an unanticipated rearrangement of an immunoglobulin heavy chain gene to clone the chromosome breakpoint of the t(14;18) (q32, q21) bearing human lymphomas. The isolated region of chromosome segment 18q21 mediates the translocation in other follicular lymphomas (60%) but does not rearrange in normal B cells or other B cell malignancies. The breakpoints are clustered on chromosome 18 and around the JH region on chromosome 14. A new transcriptional unit is introduced onto the derivative 14 chromosome near the immunoglobulin enhancer element in all cases. The cloning of this 18q21 element provides the opportunity to characterize a potentially new transforming gene that alters the growth and differentiative capacity of these lymphomas. This conference on tumor antigens and their genes was quite successful both in its formal scientific session and the informal exchange it created among the various participants. The topics covered a broad area, but all were central to a theme of normal differentiation and its malignant counterparts. This is an area in which the techniques of monoclonal antibodies, glycoprotein biochemistry, and molecular biology are making rapid inroads. Dr. Higashino and myself appreciate the opportunity to have organized this international exchange, and I am sure that we can testify for the group that it was productive.
(3) Seminar on "Mechanism of Growth Regulation in Normal and Neoplastic Cells"
The seminar was organized by Drs. Yoji Ikawa and Thomas F. Deuel and held January 14-16, 1985, in the East-West Center, University of Hawaii, Honolulu, Hawaii. The principal advisors (Drs. I. Pastan and H. Sugano) and the co-organizers (Drs. T. Deuel and Y. Ikawa) had the specific goal of a very exciting scientific exchange of information both through formal and informal communications. It was planned that intensive presentations would be followed by free indepth discussions of the scientific matters presented. Discussions were to continue through luncheons and through an informal dinner together as well as informally outside of the meeting center. It was thought also that plans for further exchanges of scientific information and additional scientific interactions would result.
The success of the program depended largely on the very high quality of the presentations but equally importantly on many formal and informal discussions which arose among the participants as a direct result of the presentations. Topics covered during the program included the role of growth factors in malignant transformation, with emphasis upon the platelet-derived growth factor (PDGF) sis oncogene and on the epidermal growth factor (EGF) receptor gene/v-erb B oncogene; recent investigations relating to mechanisms and/or sites of activity of the ras oncogene; the role of protein kinase C in signal transduction and its potential for mediating oncogenic responses; the regulation of the c-myc gene expression, of heat-shock gene, Go-specific, temperative-sensitive mutants; the role of the pX gene product in leukemogenesis ; and the role of specific sugar chains in glycoproteins in cellular transformation and as modifiers of growth factor receptor activity.
Dr. Deuel opened the session discussing growth factor activity and viral transformation. The PDGF is the major mitogenic protein in human serum for cells derived from mesenchyme. Recently, it has also been shown that PDGF is a potent chemotactic agent for cells involved in vascular inflammation and repair. Both the mitogenic and chemotactic activities of PDGF are optimally expressed at concentrations of PDGF well below those in human serum. Thus, PDGF is ideally suited to mediate important events in inflammation and repair at sites of blood vessel injury. Recently, PDGF also was shown to have striking homology with the predicted amino acid sequence of p28v-sis, the transforming protein of the simian sarcoma virus (SSV), and the product of the viral sis oncogene. The homology between the partial amino acid sequence of human PDGF and that of p28v-sis (<90%) suggests that expression of a PDGF-like growth factor activity is required to initiate and maintain transformation in SSV-transformed cells.
A growth factor activity identical to PDGF in mitogenic and immunological properties has been identified in SSV-NIH-3T3 cell lysates; this activity is immunoprecipitated by anti-PDGF antiserum and was shown to have a molecular weight essentially identical to the major intracellular processed product of the v-sis gene. SSV-transformed NIH-3T3 cells and SSV NRK cells secreted a PDGF-like growth factor activity which competed with 125I-PDGF for receptor binding. Transformed cells contain a limited number of specific high affinity 125I-PDGF receptors, suggesting that secreted p28v-sis may interact with receptors on the surface of SSV-transformed cells to stimulate autocrine cell growth. Culture with anti-PDGF antisera and p28v-sis blocks the incorporation of [3H]thymidine into SSV-NIH-3T3 cell and SSV-NRK cell DNA, suggesting that the transforming protein of the SSV stimulates the autocrine growth of these SSV-transformed cells through PDGF cell surface receptors.
The rate of tumor cell growth in athymic nude mice injected with SSV-transformed cells was compared with levels of secreted growth factor activity. The rate of tumor cell growth in nude mice corresponded directly with the levels of p28v-sis secreted by SSV-transformed cells.
SSV NP-1 cells do not secrete identifiable levels of growth-promoting activity, nor is the incorporation of thymidine into growing SSV NP-1 cell DNA inhibited by anti-PDGF antiserum. Thus, transformation per se may be mediated by an intracellular receptor and may not have an obligate requirement for secretion of transforming protein. Such intracellular receptors potentially may be found in the endoplasmic reticulum where the newly processed PDGF receptor protein and newly processed p28v-sis may interact in a manner very similar to the interactions occurring at the cell surface.
Abnormal expression of the erb B proto-oncogene in human cancer cells was then discussed by Dr. Yamamoto. Avian erythroblastosis virus strain H (AEV-H) contains the v-erb B oncogene which is responsible for the induction of both erythroblastosis and sarcomas. Analysis of the nucleotide sequence revealed that the primary product of the v-erb B gene of AEV-H is a protein with a molecular weight of 67,000. This protein, p67erb B, is glycosylated in vivo to form a glycoprotein gp72erb B. Inspection of the amino acid sequence deduced from the nucleotide sequence of the v-erb B genes of both AEV-H and its mutant, td-130, pointed out the following features: (1) the erb B protein has a domain characteristic of the gene products of the src gene family, which is the kinase domain; and (2) a domain that follows the kinase domain at the carboxy-terminal side has a crucial role for the induction of erythroblastosis.
To understand functions of the erb B proto-oncogene, the group has been searching the human gene library for the v-erb B-related sequence and could identify two genes, c-erb B-1 and C-erb B-2. As was reported, one of the erb B proto-oncogenes, c-erb B-1, encodes EGF receptor. Preliminary analysis of the c-erb B-2 gene showed that it could encode a protein similar to EGF receptor. Data showed that amplifications of the c-erb B-1 gene and the c-erb B-2 gene are associated with some human cancers. These results indicated that both the two v-erb B-related genes have roles in some stage of tumorigenesis.
Dr. Feramisco's topic was microinjection of ras oncogene proteins and inhibitory antibodies specific for the ras oncogene proteins into living normal and transformed cells. Through the use of microneedle injection procedures, the group introduced both purified ras oncogene proteins and purified anti-bodies directed against the ras oncogene protein into living normal and transformed cells. The human H-ras oncogene protein and the H-ras proto-oncogene protein were made from E. coli after expression of the genes in the pAS1 vector. The oncogene protein causes the induction of DNA synthesis and proliferation after injection into quiescent normal cells, while the normal form of the protein has little or no effect on the growth state of the cells; the oncogene protein obviates the need for certain growth factors. For complementary experiments, antibodies were made to synthetic peptides covering the region of the v-K-ras oncogene protein (near amino acid 12) that distinguishes it from the normal form of the ras proteins. The antibodies bound only to the oncogene form of the protein and inhibited the guanine nucleotide binding activity. Injection of these antibodies into v-K-ras-transformed cells made them revert to a normal phenotype for 2 days.
Isao Uno concentrated on functions of ras gene products in yeast. The regulatory roles of cAMP in yeast have been studied by using mutants defective in adenylate cyclase (AC) or cAMP-dependent protein kinase (A-kinase). cAMP-requiring mutants arrested at G1 phase of the cell cycle in the absence of cAMP and phosphorylation of cellular proteins by A-kinase has essential functions in passing throughG1 to S phase. cyrl gene is a structural gene of catalytic subunit of AC, and G protein binds GTP leads AC stimulation in yeast as mammalian cell. The bcyl mutant isolated as a suppressor of cyrl mutant indicated the continuous activation of A-kinase and suppressedG1 arrest by nutrient limitation or by cyrl mutation.
On the other hand, yeast contains two ras genes (ras1, ras2) homologous to H-ras, and replacement of two genes with nonfunctional versions lost the ability of spores to germinate. Human ras gene can substitute for the inactivated endogenous yeast ras genes. The bcyl mutation could suppress ras1 ras2 mutation, and triple mutant showed nondetectable level of AC activity and cAMP. Activation of transforming potential of a mammalian ras gene is brought about by a single amino acid change in the critical locations. When yeast gene with an alteration analogous to one known to activate the mammalian gene was introduced into yeast cells like ras2val, they indicated high levels of AC activity and cAMP and could not arrest atG1 phase. Further, the ras gene product had GTPase activity as human p21 and G protein. Finally, GTP-dependent AC could be reconstituted with plasma membrane fraction containing ras gene product and catalytic subunit of AC. These results indicate that ras gene product is G protein and that it regulates AC activity. Then cAMP controls cell proliferation through protein phosphorylation.
The last paper of the day was Akiro Kishimoto’s on the role of protein kinase C in signal transduction. Information of a variety of extracellular signals appears to flow from the cell surface into the cell interior through two routes, protein kinase C activation and Ca2+ mobilization. It is becoming clearer that the signal-dependent breakdown of inositol phospholipid, particularly phosphatidylinositol (PIP2), is a key event for initiating these processes. Under physiological conditions, one of the earliest products of PIP2 breakdown, diacylglycerol, activates protein kinase C selectively whereas the other product, inositol trisphosphate, appears to mobilize Ca2+ from its intracellular reservoir (M. J. Berridge). Under appropriate conditions, these two routes may also be opened separately and independently by the exogenous addition of synthetic diacylglycerol and Ca2+-ionophore, respectively, to intact cells without interaction with any cell surface receptors. It is worth noting that tumor-promoting phorbol esters act as a substitute for diacylglycerol and that protein kinase C is a prime target of tumor promoters. By using these procedures, it is found that the activation of cellular responses by the route of protein kinase C is separate from and synergistic to those elicited by Ca2+ mobilization. Recent evidence will be summarized that these two synergistic routes are of crucial importance in cellular functions and proliferation observed in various cell types exposed to extracellular signals.
Dr. Ichiro Yahara opened the second day's session by discussing the involvement of heat-shock proteins in regulation of cell growth. A distinction was noted betweenG1 arrests that lead cells to enter G0 and those that do not.
Durable synthesis of heat-shock proteins (hsps) in G0 cells of the yeast and other eucaryotes led them to hypothesize that these hsps might function in the cellular transition from the proliferating state to G0 and/or in the maintenance of the G0 state. To examine this hypothesis, his team attempted to isolate heat-shock-resistant mutants of Saccharomyces cerevisiae, which constitutively express hsps, and examine them for their properties concerning growth control.
A heat-shock-resistant mutant of S. cerevisiae was isolated showing constitutive synthesis of two hsps and altered growth. Unusual phenotypes revealed by this mutant are attributed to a single nuclear, recessive mutation, designated hsrl. Exponentially growing hsrl cells constitutively synthesize six proteins including two hsps, hsp48A and hsp48B, and are approximately 1000-fold more resistant to lethal heat shock than hsrl cells. The hsrl mutation also elongated theG1 period and affected both growth arrest by sulfur starvation and growth recovery from it.
A second single nuclear mutation, designated suhl, was isolated which suppresses all the phenotypes derived from the hsrl mutation.
Dr. Merino discussed cloning and characterization of EGF receptor gene sequences in A431 human carcinoma cells. A 2.4-kbp cDNA (pE7) isolated from an A431 human carcinoma cDNA library was sequenced which is homologous to AEV v-erb B and identical to known EGF receptor peptides. Using this cloned cDNA as probe for RNA blot analysis of EGF receptor RNA, an excellent correlation was found in a variety of normal and malignant cell lines between major RNA species (10 and 5.6 kb) and receptor protein (Mr 170 kDa) immunoprecipitated from [35S]methionine-labeled cell extracts. A431 cells contained the highest levels of both RNA and protein; this was due in part to an N 30-fold amplification of EGF receptor gene sequences in A431 cells relative to other cell lines tested by DNA blot analysis. A squamous carcinoma cell line also contained amplified EGF receptor genes. Other cell types appeared to contain a low copy number as determined by comparison to the hybridization signal of single copy plasmid DNA (probably one to two copies per haploid genome). A431 cells also possess an overexpressed, aberrant 2.9-kb EGF receptor RNA containing divergent sequences at the 3' end; a cDNA encoding this aberrant RNA (pE15) was isolated. Major differences in the banding pattern of A431 EGF receptor gene sequences suggested that a rearrangement is responsible for the production of this aberrant RNA. Recently, it was discovered that the pE7 cDNA probe hybridizes in situ to the p14/12 region of chromosome 7 in normal cells, near a heritable fragile site. In A431 cells the pE7 probe hybridizes to the chromosome 7-related translocation marker chromosomes as well. A probe from the divergent region of pE15 also hybridizes to the p14/12 region of chromosome 7, suggesting that an intrachromosomal rearrangement has occurred.
ln vivo regulation of myc gene expression was discussed by Kenshi Hayashi. Observations of the structural changes of myc and the resultant deregulations of expression of the gene in diverse types of tumor-derived cell lines strongly suggest that myc gene function is involved in one of the common characteristics of tumors, that is the cell growth. In vitro studies showed that myc is transiently expressed when cells such as B lymphocytes, T lymphocytes, or fibroblasts progress from resting state of G0 phase to growth cycle. One could ask how myc would respond in vivo to the proliferative stimulus. One suitable in vivo system for investigating this question is liver regeneration. In a rat, surgical removal of approximately two-thirds of the liver prompts the remaining liver to proliferate, and synchronized DNA replication starts approximately 20 hours after the operation. A remarkable but transient increase in myc mRNA at an extremely early phase of liver regeneration was found. The sharp rise followed by the quick decline of the myc mRNA level suggests that myc is regulated by a feedback mechanism as in the case of some bacterial genes, e.g., heat-shock genes of E. coli. Supposing such a mechanism exists, it follows that inhibition of protein synthesis would block the suppression of myc. Indeed, i.p. injection of cycloheximide induced a dramatic and steady increase of myc mRNA in rat liver, reaching a level about 600 times that of the uninduced level at 6 hours after injection. Similar induction myc mRNA was also observed in spleen or kidney but not in brain, muscle, or lenkocytes. The interpretation of these observations is that, in vivo, at least two different mechanisms can operate to suppress expression of myc.
Toshinori Ide reported on a G0-specific ts mutant. Under conditions restrictive for growth, normal cells in culture leave cell cycle and become arrested in the G0 phase. Under the same restrictive conditions, transformed cells do not arrest but continue to proliferate, or die under more severe conditions. Proto-oncogenes in normal cells and oncogenes in transformed cells have been assumed to participate in the regulation of the cell cycle, especially from G0 to S phase. The purpose of this study was to find the specific function(s) which is required solely for the stimulation of cells from G0 to S phase but is not required for the progression from M to S phase. A ts mutant clone, tSJT60, was isolated from Fischer rat cell line 3Y1. During the exponential growth at both 34 and 39.5°C, tsJT60 did not appear as ts mutant cells. However, once entered resting state (G0) under serum deprivation at the confluent state, they could reenter S phase at 34°C but could not at 39.5°C following the stimulation of cells either by the addition of fetal bovine serum or by trypsinization and replating. These and other results suggested that tsJT60 is a G0-specific ts mutant, i.e., the cells have ts defect(s) in the function which is required for the stimulation from the resting state to S phase but not for the progression of cell cycle in an exponential growth phase. The execution point was estimated at about 9 hours after the stimulation, and the mutated function appeared to be required for the progression from G0 through 9 hours post stimulation. Though total RNA and protein synthesis did not increase at 39.5°C after the stimulation of G0 cells, expression of c-myc did around 2 hours after the stimulation at both temperatures.
The last topic covered by Dr. Y. Ikawa was on the complete nucleotide sequence of bovine leukemia virus (BLV) proviral DNA and its evolutionary relationship with human T cell leukemia virus (HTLV). BLV is a causative agent of enzootic leukosis of cattle and shares several biological features common to HTLV.
The BLV proviral genome was detected in an oligoclonal manner in the enzootic leukemia tissue among Japanese cattle. A proviral DNA of BLV was then cloned in a Charon 4A g-phage vector. Comparison in restriction cleavage sites of the genome of the isolate with other BLV isolates revealed highly conserved regions near the 3’ end.
The complete nucleotide sequence of the BLV genome was obtained and was compared by computer analysis with that of the HTLV-I genome, revealing the similar genomic organization, unique LTR-gag-independent protease gene pol-env-pX frames-unique LTR, which distinguishes the two viruses from other retrovirus groups (tentatively proposed as "Type E"). The computer-assisted two-dimensional matrix analysis between BLV and HTLV-I genomes on amino acid and nucleotide sequence level revealed highly homologous regions, extensively in the gag and pol and partially in the env regions.
Spanning the gag-pol junction of the BLV genome is an independent open reading frame coding for a protease which might cleave gag polyprotein. This protein actually expressed BLV-infected cells as an independent protein using the above frame. The HTLV-I genome sequenced in pgATK clone also showed the corresponding open reading frame but with stop codons.
As to the open reading frames between env and 3' LTR termed pXs, first described in the HTLV-I pgATK clone, the BLV clone also had the similar open reading frames (pXBL), the longer one of which showed only about 25% homology with pX4 of the above HTLV-I, clone. The pX frames of BLV were hypomethylated, and there were preferential third letter changes in the longest pX frames’ codons between this isolate and Belgian isolate.
Both BLV and HTLV-I LTRs possessed long R regions, a part of which showed about 50% nucleotide sequence homology. Poly(A) signal and poly(A) site were as far apart as 260 bp in both LTRs, and a large hairpin structure may be formed in the RNA transcript so that these two signal sequences could come closer to each other. The presence of enhancer element in BLV-LTR resembling that of an immunoglobulin gene may explain why BLV can target B lymphocytes.
Structural changes in glycoprotein sugar chains in cellular transformation were discussed by Akira Kobata. As clearly shown in the case of the monoclonal antibodies CA 19-9 and CA 125, the abnormal sugar chains of glycoconjugates produced by tumors can be useful markers for the diagnosis of cancer. Since no template is included in the biosynthetic machinery of sugar chains in the cell, they are formed by the concerted action of glycosyltransferases. Therefore, there are considerable possibilities that abnormal gene regulation in tumor cells produces sugar chains which are not found in normal cells. Accordingly, the sugar chains of various tumor glycoconjugates may provide useful markers for the diagnosis and therapeutic management of patients with malignant disease. q-Glutamyl transpeptidase (q-GTP) and human chorionic gonadotropin (hCG) are two such examples which underline the usefulness of this line of study.
q-GTP is a membrane-bound glycoenzyme widely distributed in the plasma membrane of epithelial cells of various mammalian organs. In malignant tissue, the enzyme is known to behave differently from its physiological counterpart. Comparative studies of the sugar chains of q-GTPs purified from rat liver and from rat ascites AH-66 hepatoma revealed that about half of the sugar chains of the hepatoma enzyme contain bisecting N-acetylglucosamine residues which are not found in those of the liver enzyme. The structural change associated with this malignant transformation is enabling the development of a new method for the diagnosis of hepatoma.
Another interesting example of sugar chain transformation is found in hCG. This glycoprotein hormone is produced by trophoblast and secreted into the blood and urine. Patients with trophoblastic diseases such as hydatidiform mole and choriocarcinoma excrete large amounts of hCG in their urine. Comparative investigations of the sugar chains of hCG samples purified from this source in normal pregnant women and patients with trophoblastic diseases revealed the following results. Mole hCGs have the same complement of sugar chains as normal hCG, while all choriocarcinoma hCGs examined contain two triantennary, two unusual biantennary, and one monoantennary complex-type asparagine-linked sugar chains which are not found in normal hCGs. The structural characteristics of these oligosaccharides indicate that the appearance of the N-acetylglucosaminyl transferase responsible for the formation of GlcNAc!
!!I 4Manal 3 group is the enzymatic basis of this transformational change. Because the unusual biantennary sugar chains have never been detected in other glycoproteins so far reported, the change could provide the basis for immunotherapy besides serving as a diagnostic marker. The final topic, the role of glycolipids in oncogenesis and regulation of cell proliferation, was addressed by Sen-itiroh Hakomori. Dramatic changes of glycolipid composition and metabolism have been observed in many tumor cells transformed by oncogenic viruses and in spontaneous tumors, including human cancer. A reduction of GM3 and GM1 gangliosides as well as other higher glycolipids has been observed in many transformed cells due to incomplete synthesis, whereby precursors are accumulated. On the other hand, some tumor cells are characterized by activation of synthesis of a new glycolipid that is absent in progenitor cells (neosynthesis). Either precursor accumulation or neosynthesis can lead to the formation of tumor-associated glycolipid markers. Such markers have been characterized imuunochemically in various experimental tumors, and recent studies with monoclonal antibodies have defined a large variety of glycolipid markers characteristic of specific types of tumors. Gangliotriaosylceramide (Ga1Nac
1$4Ga11$4G1c1$1Cer) in murine sarcoma and lymphomas, globotriaosylceramide (Ga1a1$4Ga11$4G1c1$1Cer) in Burkitt lymphoma, and GD3 ganglioside (NeuAca2$8NeuAca2$3Ga11$4G1c1$1Cer) in human melanoma have been identified as characteristic markers for these tumor cells and are absent or minimally expressed in various normal tissues. Other common antigens expressed in a large variety of human adenocarcinomas (lung, pancreatic, liver, gastric, colonic, and breast cancers) have been identified recently as modified blood group antigens, i.e., sialylated and fucosylated type 1 or type 2 chain. Of particular interest is the accumulation of difucosyl and trifucosyl type 2 chain and their sialyl derivatives, which are absent or minimally expressed in normal tissue, highly expressed in various adenocarcinomas, but not expressed in undifferentiated tumors. Systematic immunohistological studies indicate that these antigens are maximally expressed in fetal gastrointestinal mucosa during organogenesis (50 to 110 days of gestation). These antigens are, therefore, typical oncofetal antigens. The oligo-saccharide sequence with X hapten structure multivalently linked to lysyllysine inhibits embryogenesis. Thus, these structures are important recognition sites for cell-cell interaction during development and differentiation. A deletion or reduction of GM3 or GM1 ganglioside associated with oncogenic transformation can be correlated with a loss of growth control through dysfunction of growth factor receptors. Exogenous addition of GM1 or GM3 can arrest cell growth at theG1 phase, and GM3- or GM1-enriched cells become refractory to growth factor stimulation. In 3T3 cells GM1- or GM3-fed cells were refractory to PDGF stimulation. PDGF-dependent tyrosine phosphorylation of the PDGF receptor was inhibited by GM1 or GM3 but not by other gangliosides or neutral glycolipids. A similar inhibition of tyrosine phosphorylation of the EGF receptor by GM3 (but not by GM1) was observed in A431 cells. Both EGF receptor phosphorylation and PDGF receptor phosphorylation were promoted by phosphatidyl-ethanolamine but not by other phospholipids so far tested. Thus, the glycolipid/phospholipid balance may determine the lipid bilayer environment and affect the growth factor receptor. A reduction or enrichment of GM3 or GM1 causes an unbalanced lipid environment, which may lead to dysfunction of growth factor receptors.
SEMlNAR AGENDA AND PARTICIPANTS
(1) SEMINAR ON MOLECULAR MECHANISMS OF CELL-CELL INTERACTIONS INVOLVED IN TUMOR IMMUNITY
Honolulu, Hawaii, November 26-28, 1984
AGENDA