siRNA-mediated silencing of PD-1 ligands enhances tumor-specific human T-cell effector functions.

Adoptive cell therapy using tumor-specific T cells is a promising strategy for treating patients with malignancy. However, accumulating evidences have demonstrated that optimal function of tumor-reactive T cells is often attenuated by negative regulatory signal(s) delivered through receptors, such as cytotoxic T-lymphocyte antigen 4 (CTLA-4), programmed death 1 (PD-1), and their cognate ligands. Although systemic blocking of these molecules needs careful attention on the risk of uncontrolled immune activation, selective inhibition of negative signals in tumor-specific T cells by their genetic modification is an attractive approach to overcome immunological suppression in cancer patients. Here, we demonstrate the improved effector functions of tumor-specific CD4(+) and CD8(+) human T cells by small interfering RNA (siRNA) -mediated silencing of PD-1 ligands, PD-L1 or PD-L2. Tumor antigen MAGE-A4-specific human T-cell clones upregulated the expression of PD-1 ligands upon activation. siRNA-mediated knockdown of PD-L1 or -L2 enhanced the interferon-γ production and antigen-specific cytotoxicity of these cells. Peripheral blood mononuclear cells transduced with a retroviral vector encoding MAGE-A4-specific T-cell receptor α/ß chains also increased their effector functions by this modification. These results suggest that siRNA-mediated knockdown of PD-1 ligands is an attractive strategy to inhibit a negative regulatory mechanism of tumor-specific T cells resulting in enhanced efficacy of adoptive T-cell therapy of cancer using genetically modified autologous lymphocytes.

Cell surface antigens of human malignant melanoma. II. Serological typing with immune adherence assays and definition of two new surface antigens.

Immune adherence assays revealed that 10 out of 18 melanoma patients had demonstrable antibody to surface antigens of autologous cultured melanoma cells, with serum titers ranging from 1/4 to 1/160. Autologous fibroblasts showed no reactions with these sera. Antibody from individual patients showed reproducible temperature preference for maximal reactivity. Two new melanoma antigenic systems were defined in this study. The first, BD, was restricted to autologous melanoma and could not be demonstrated in absorption tests on 12 allogeneic melanoma cell lines. The other, AH, was found on 5 of 12 melanomas and represents a class of shared melanoma surface antigens. Neither BD nor AH antigen was found on normal cells from autologous, allogeneic, or xenogeneic sources or on any nonmelanoma tumor cell line. Methods are now available to develop a comprehensive serological classification of the surface antigens of melanoma.

CD4-CD8- T cell receptor alpha beta T cells: generation of an in vitro major histocompatibility complex class I specific cytotoxic T lymphocyte response and allogeneic tumor rejection.

The generation of an in vitro major histocompatibility complex class I specific response of CD4-CD8- T cell receptor (TCR) alpha beta cytotoxic T lymphocytes (CTL) and their allogeneic tumor rejection were investigated. Inocula of BALBRL male 1 were rejected in C57BL/6 (B6) mice treated with minimum essential medium (MEM) (control), anti-L3T4 (CD4) monoclonal antibody (mAb) or anti-Lyt-2.2 (CD8) mAb and CTL against the tumor were generated in vitro. No rejection and no induction of CTL were observed in B6 mice treated with anti-L3T4 (CD4) plus anti-Lyt-2.2 (CD8) mAb. CTL with the classical Thy-1+ CD3+CD4-CD8+ TCR alpha beta phenotype were generated in mixed lymphocyte tumor cell culture (MLTC) spleen cells from B6 mice treated with MEM (control) or anti-L3T4 (CD4) mAb, whereas CTL with an unusual Thy-1+CD3+CD4-CD8- TCR alpha beta phenotype were generated in MLTC spleen cells from anti-Lyt-2.2 (CD8) mAb-treated B6 mice. Both types of CTL were reactive with both H-2Kd and Dd (Ld) class I antigen. These findings suggest that when CD4+ cells were blocked by anti-L3T4 (CD4) mAb, CD8+ CTL mediated rejection, and when CD8+ cells were blocked by anti-Lyt-2.2 (CD8) mAb, CD4+ cells were capable of mediating rejection, although less efficiently than CD8+ cells, by inducing CD4-CD8- TCR alpha beta CTL. The finding that adoptive transfer of CD4 and CD8-depleted MLTC spleen cells, obtained from anti-Lyt-2.2 (CD8) mAb-treated B6 mice that had rejected BALBRL male 1, resulted in rejection of BALBRL male 1 inoculated into B6 nu/nu mice confirmed the above notion. CTL clones with the CD4-CD8- TCR alpha beta phenotype specific for Ld were established.

Ly phenotype of cytotoxic T cells for syngeneic tumor.

Our present and previous findings may be summarized as follows: The phenotype of C57BL/6 (B6) cytotoxic cells for allogeneic target cells is Thy-1+, Ly-1- Ly-2/3+, MSLA+, and Ig-. the phenotype of B6 cytotoxic cells for syngeneic tumor cells is Thy-1+, Ly-1+, Ly-2/3+, MSLA+, and Ig-. The phenotype of B6 cytotoxic cells for syngeneic tumor cells is Thy-1+, Ly-1+, Ly-2/3+, MSLA+, AND Ig-. Thus, differences in Ly phenotype appear to be exhibited not only by cytotoxic T cells as opposed to helper T cells, but also within subcategories of cytotoxic T cells.

Cell surface antigens of human malignant melanoma. III. Recognition of autoantibodies with unusual characteristics.

The sera of three patients with malignant melanoma showing reactivity with surface antigens of cultured autologous melanoma cells were analyzed by mixed hemadsorption and immune adherence assays in conjunction with absorption tests. In contrast to the melanoma-specific antigens demonstrated previously, the surface antigens detected by these sera occurred on a broad range of nucleated cells, both normal and malignant, from human, monkey, mouse, and chicken sources. Each serum had a characteristic pattern of reactivity in absorption tests, indicating the detection of distinct antigenic systems. Two sera showed auto-, allo-, and xenoreactivity, as well as the capacity to distinguish different cell populations in the same individual. The other serum reacted with an antigen apparently universally present on nucleated cells from a variety of species, but absent on erythrocytes. As these patients had been treated with chemotherapy, this may have played a role in the emergence of these broadly reactive autoantibodies.

Cell surface antigens of chemically induced sarcomas of the mouse. I. Murine leukemia virus-related antigens and alloantigens on cultured fibroblasts and sarcoma cells: description of a unique antigen on BALB/c Meth A sarcoma.

As background for a serological definition of the unique antigens of chemically induced sarcomas, we have typed a series of fibroblast and sarcoma cell lines of BALB/c and C57BL/6 origin by cytoxicity and absorption tests for murine leukemia virus (MuLV)-related cell surface antigens and known alloantigens. 7 of the 17 cultured lines expressed the range of cell surface antigens associated with MuLV (GIX, GCSA, gp70, p30), and this was invariably associated with MuLV production. In nonproducer lines of C57BL/6 (but not BALB/c) origin, a MuLV-gp70-like molecule was found on the surface of fibroblasts and sarcoma cells. The alloantigenic phenotype of these MuLV+ and MuLV- cell lines was H-2D+, H-2K+, Thy-1.2+ or -, PC.1+ or -, Lyt-1.2-, Lyt-2.2-, Ia.7-, and TL.2-. A unique antigen was defined on the BALB/c ascites sarcoma Meth A with antisera prepared in BALB/c or (BALB/c X C57BL/6)F1 mice. Tissue culture lines derived from this tumor were MuLV-, which facilitated serological study of the antigen. Absorption analysis indicated that the antigen was restricted to Meth A; it could not be detected in normal or fetal BALB/c tissue MuLV+ or MuLV- fibroblast lines, 12 syngeneic or allogeneic sarcomas, or normal lymphoid cells from 13 different inbred mouse strains.

Cell surface antigens of murine leukemias induced by radiation leukemia virus. Recognition of individually distinct cell surface antigens by cytotoxic T cells on leukemias expressing crossreactive transplantation antigens.

The specificity of transplantation immunity and T cell cytotoxicity against leukemias induced by RadLV was examined. Subcutaneous inoculation of two RadLV leukemias induced in BALB/c mice, BALBRVB and BALBRVD, resulted in initial tumor growth in CB6F1 mice, followed by complete tumor regression. Mice that had rejected leukemias BALBRVB or BALBRVD were subsequently challenged with various tumors of BALB/c origin. The growth of all five RadLV leukemias tested, and of one radiation-induced leukemia, was significantly inhibited. Another radiation-induced leukemia, a methylcholanthrene-induced sarcoma, and a leukemia induced by the Moloney leukemia virus, were not inhibited. The results indicate that RadLV leukemias share cell surface antigens that induce transplantation immunity in vivo. Cytotoxic lymphocytes were generated by coculturing spleen cells from mice that had rejected leukemia BALBRVB or BALBRVD with the corresponding leukemia cells. Direct tests and inhibition tests showed that such cytotoxic cells recognized individually specific antigens on leukemias BALBRVB and BALBRVD, distinct from the shared antigens detected in transplantation experiments. The effector cells in cytotoxicity assays were Thy-1+, Lyt-1+,-, Lyt-2+, and Lyt-3+ T cells.

Cell surface antigens of human renal cancer defined by autologous typing.

Sera from 28 patients with renal cancer were tested for reactivity with surface antigens of cultured autologous renal cancer cells. Four serological assays were used to survey sera for autologous antibody. Immune adherence, protein A, and C3-mixed hemadsorption assays detected reactivity in a high percentage of patients (80-100%), whereas mixed hemadsorption assays were negative with sera from all but one patient. Reactive sera from six patients were analyzed by absorption tests with autologous, allogeneic, and restricted to autologous renal cancer cells; class 2 antigens, present on certain allogeneic renal and nonrenal cancer cells; and class 3 antigens, found on a wide variety of normal and malignant cell types. The sera of one patient detected class 1, 2, and 3 antigens, the sera of three patients detected class 2 antigens, and the sera of two patients detected class 3 antigens. This analysis of renal cancer, with the recognition of three classes of surface antigens recognized by autologous sera, resembles the results of autologous typing of three other human malignancies: malignant melanoma, acute leukemia, and astrocytoma. Evidence provided by autologous typing of these cancers indicates that class 1 and class 2 antigens are tumor-restricted and that under certain circumstances these antigens are immunogenic for the autologous host.

Expression of T-cell differentiation antigens on effector cells in cell-mediated cytotoxicity in vitro. Evidence for functional heterogeneity related to the surface phenotype of T cells.

The cell-mediated cytotoxicity (CMC) of nonadherent cells from the peritoneal cavity (NAPC) of alloimmunized mice can be measured by the [3H]proline microassay. The exhibition of thymus-derived (T) cell antigens on these killer cells was studied by incubating them with the relevant T-cell antisera and complement (C), under optimal conditions for lysis, before performance of the CMC assay. Under these conditions, the following T-cell antigens were demonstrable on the killer population in terms of percent reduction in CMC by the respective antisera: (a) Thy-1.1 (83%) and Thy-1.2 (100%), (b) MSLA (86%), (c) NTA-RA (a T-cell antigen recognized by naturally occurring autoantibody of NZB mice) (62%), (d) Ly1.1 )58%, (e) Ly-2.1 (11%; considered a marginal result) and Ly-2.2 (63%), and (f) Ly-3.2 (77%). The following were not demonstrable: (g) TL, and (h) Ly-1.2. (i) The antigen Ly-3.1 was not studied. Omission of C deprived all T-cell antisera tested of their capacity to suppress CMC, indicating that the cell components recognized by such antisera may perform no direct function in CMC. On the assumption that all Ly+ cells are Thy-1+, it is clear that the T-cell members of the immune NAPC population must be heterogenous. This follows from the fact that the proportions of T cells lysed by different Ly antisera did not correspond with ensuing degree of loss of CMC capacity. The extremes were represented by anti-Ly-1.2 (74% Thy-1+ cells lysed, but no reduction in CMC) and Ly-3.2 (54% Thy-1+ cells lysed, with 77% reduction in CMC). From this initial survey it appears that the C57BL/6 mice killer T-cell population active in CMC in vitro is relatively rich in surface antigens of the Ly-2/Ly-3 category and relatively poor in representation of the Ly-1 surface antigens. It remains to be seen whether this killer cell phenotype, poor in Ly-1 and rich in Ly-2/Ly-3, is characteristic of the mouse generally. From these results it appears that subsets of T cells with different immunological functions may exhibit qualitative or quantitative differences in surface antigens specified by different Ly loci; this will be easier to assess in the future when the results of experiments with the same Ly antisera but dealing with T-cell functions other than CMC become available.

Differential involvement of CD4+ cells in mediating skin graft rejection against different amounts of transgenic H-2K(b) antigen.

Differential involvement of CD4+ cells in mediating class I-disparate skin graft rejection was investigated using quantitatively different Kb transgenic mice as donors under conditions in which CD8+ cells were blocked in vivo by administration of anti-CD8 monoclonal antibody (mAb). Tg.H-2Kb-1 and -2 are C3H transgenic mice with 14 and 4 copies, respectively, of the H-2Kb gene. Cell surface expression of Kb antigen and the Kb antigenicity of skin for eliciting graft rejection with homozygous and heterozygous transgenic mice were correlated with the copy number. In vivo administration of anti-Lyt-2.1 (CD8) mAb markedly prolonged survival of heterozygous and homozygous C3H Tg.H-2Kb-2 skin grafted onto C3H mice, but prolonged survival of heterozygous Tg.H-2Kb-1 skin grafts much less and did not prolong survival of homozygous Tg.H-2Kb-1 grafts. Administration of anti-L3T4 (CD4) mAb alone did not have any effect on skin graft rejection. Administration of anti-L3T4 (CD4) mAb with anti-Lyt-2.1 (CD8) mAb blocked rejection in all combinations. These findings indicate that a quantitative difference of class I antigen caused differential activation of CD4+ cells under conditions in which CD8+ cells were blocked.

Roles of CD4+ and CD8+ cells, and the effect of administration of recombinant murine interferon gamma in listerial infection.

Studies were made on the effects of in vivo administration of anti-CD4 mAb, anti-CD8 mAb, or a combination of both mAbs on multiplication of bacteria, the levels of serum transaminases, and mortality in mice infected with Listeria monocytogenes. Results showed that in sublethal infection, CD8+ cells enhanced the peak of bacterial multiplication and liver cell necrosis, and CD4+ cells suppressed CD8+ cell-mediated enhancement. Results also showed that either CD4+ or CD8+ cells were necessary for, and capable of, mediating clearance of the bacteria. CD8+ cells were more efficient than CD4+ cells, but for optimal clearance both were necessary. In lethal listeriosis, treatment of mice with anti-CD8 mAb or a combination of both anti-CD4 and anti-CD8 mAbs, but not anti-CD4 mAb only, protected mice from death by decreasing multiplication of bacteria in the liver and spleen after a peak of approximately 10(8) CFU, and lowering the elevated serum levels of transaminases. These findings indicated that CD8+ cells were responsible for causing irreversible systemic Listeria infection and severe liver necrosis. In lethal listeriosis, administration of rMuIFN-gamma markedly prolonged survival by decreasing multiplication of bacteria and promoting recovery from liver necrosis.

Human melanoma antigen AH is an autoantigenic ganglioside related to GD2.

AH antigen, initially defined by an antibody present in a melanoma patient, is a cell surface antigen found on approximately 65% of melanoma cell lines. Absorption analysis indicates that AH is a differentiation antigen marking normal and malignant cells of neuroectodermal origin. The AH determinant has been found to be related to GD2 ganglioside.

Rapid alphabeta TCR-mediated responses in gammadelta T cells transduced with cancer-specific TCR genes.

Adoptive T-cell transfer of in vitro cultured T cells derived from cancer patients with naturally developed immune responses has met with some success as an immunotherapeutic approach, although only a limited number of patients showed spontaneous immune responses. To find alternative ways, such as cancer-specific T-cell receptor (TCR) gene transfer, in preparation for sufficient numbers of antigen-specific T cells is an important issue in the field of adoptive T-cell therapy. Given the inherent disadvantage of alphabeta TCR transfer to other alphabeta T cells, namely the possible formation of mixed TCR heterodimers with endogenous alpha or beta TCR, we employed gammadelta T cells as a target for retroviral transfer of cancer-specific TCR and examined whether gammadelta T cells were useful as an alternative population for TCR transfer. Although retroviral transduction to gammadelta T cells with TCR alphabeta genes alone, isolated from a MAGE-A4(143-151)-specific alphabeta CD8(+) cytotoxic T lymphocyte (CTL) clone, did not provide sufficient affinity to recognize major histocompatibility (MHC)-peptide complexes due to the lack of CD8 co-receptor, gammadelta T cells co-transduced with TCR alphabeta and CD8 alphabeta genes acquired cytotoxicity against tumor cells and produced cytokines in both alphabeta- and gammadelta-TCR-dependent manners. Furthermore, alphabeta TCR and CD8-transduced gammadelta T cells, stimulated either through alphabeta TCR or gammadelta TCR, rapidly responded to target cells compared with conventional alphabeta T cells, reminiscent of gammadelta T cells. We propose alphabeta TCR-transduced gammadelta T cells as an alternative strategy for adoptive T-cell transfer.

Long-term phenotypic, functional and genetic stability of cancer-specific T-cell receptor (TCR) alphabeta genes transduced to CD8+ T cells.

In adoptive T-cell transfer as an intervention for malignant diseases, retroviral transfer of T-cell receptor (TCR) genes derived from CD8(+) cytotoxic T-lymphocyte (CTL) clones provides an opportunity to generate a large number of T cells with the same antigen specificity. We cloned the TCR-alphabeta genes from a human leukocyte antigen (HLA)-A(*)2402-restricted CTL clone specific for MAGE-A4(143-151). The TCR-alphabeta genes were transduced to 99.2% of non-TCR expressing SupT1, a human T-cell line, and to 12.7-32.6% of polyclonally activated CD8(+) T cells by retroviral transduction. As expected, TCR-alphabeta gene-modified CD8(+) T cells showed cytotoxic activity and interferon-gamma production in response to peptide-loaded T2-A(*)2402 and tumor cell lines expressing both MAGE-A4 and HLA-A(*)2402. A total of 24 clones were established from TCR-alphabeta gene-transduced peripheral blood mononuclear cells and all clones were functional on a transduced TCR-dependent manner. Four clones were kept in culture over 6 months for analyses in detail. The transduced TCR-alphabeta genes were stably maintained phenotypically, functionally and genetically. Our results indicate that TCR-transduced alphabeta T cells by retroviral transduction represent an efficient and promising strategy for adoptive T-cell transfer for long term.

Clinical and prognostic significance of cytokine receptor expression in adult acute lymphoblastic leukemia: interleukin-2 receptor alpha-chain predicts a poor prognosis.

We quantitatively assessed the expression of cytokine receptors (interleukin-2 receptor (IL-2R), IL-3R, IL-4R, IL-5R, IL-6R, IL-7R, granulocyte-macrophage colony-stimulating factor R (GM-CSFR), G-CSFR, c-fms, c-mpl, c-kit and FLT3) in cells from 211 adults with acute lymphoblastic leukemia (ALL) by flow cytometry and determined their prevalence and clinical significance. Although all cytokine receptors were expressed to various degrees, the levels of IL-3R alpha-chain (IL-3Ralpha), IL-2Ralpha, IL-2Rbeta, IL-7Ralpha, common-Rgamma(gammac), c-mpl, c-kit and FLT3 exhibited a wide spectrum > or =2000 sites/cell. Among them, IL-3Ralpha, IL-2Ralpha and FLT3 were highly expressed in B-lineage ALL, whereas IL-7Ralpha, gammac and c-kit predominated in T-lineage ALL. Higher levels of IL-3Ralpha, IL-2Ralpha, c-kit and FLT3 correlated with the expression of CD13/33. Increased IL-2Ralpha levels related to the presence of Philadelphia chromosome (Ph), leukocytosis and shorter event-free survival (EFS). C-kit preferred in male. Elevated FLT3 levels correlated with age > or =60 years. Multivariate analysis in B-lineage ALL revealed only IL-2Ralpha (P=0.028) and Ph (P=0.020) as independent factors for EFS. These findings suggest that several cytokine receptors associated with certain cellular and clinical features, but IL-2Ralpha solely had a prognostic value and should be considered as a major prognostic factor for adult ALL that is comparable with Ph.

Clonal expansion of CD8+ cytotoxic T lymphocytes against human T cell lymphotropic virus type I (HTLV-I) genome products in HTLV-I-associated myelopathy/tropical spastic paraparesis patients.

Short-term culture of peripheral blood mononuclear cells (PBMC) derived from patients with human T cell lymphotropic virus type I-associated myelopathy (HAM)/tropical spastic paraparesis resulted in dominance by DR+ activated CD8+ T cells. Variations in the T cell receptor (TCR) V alpha and V beta chains in these cells were analyzed, and in all 10 patients examined, 2-3 V gene families were dominant in both TCR V alpha and V beta. In five patients we examined, cultured lymphocytes contained cytotoxic lymphocytes for p40tax (patients HAM2, 3, 7, and 8) or env protein (patient HAM4) of human T lymphotropic virus type I. In patients HAM2 and HAM8, cultured lymphocytes contained a large proportion of V beta 8+ CD8+ and/or V beta 12+ CD8+ cells. The sequence of V beta 8+ and V beta 12+ cDNA revealed that they were oligoclonal with identical or similar sequences in each patient. Elimination experiments with monoclonal antibodies for TCR V beta 8 and V beta 12 showed that they were CD8+ cytotoxic T lymphocytes (CTL) for p40tax. In addition, flow cytometry and sequencing analysis of uncultured PBMC revealed that in HAM2, V beta 8+ CTL and their precursors account for 7% and V beta 12+ CTL and their precursors account for 18% of total CD8+ cells. This indicates the presence of two markedly expanded clones in vivo. No common dominant TCR V alpha or V beta were observed among 10 HAM patients analyzed.

Cytotoxic reactions of murine lymphoid cells studied with a tritiated proline microcytotoxicity test.

A lymphocyte cytotoxicity (CTX) test with 3-H-proline-prelabeled target cells was used to detect the immune response of murine lymphoid cells to H-2 and tumor antigens. The specificity of the reaction was determined by simultaneous tests on unrelated target cells and, for reactions directed against H-2 antigens, by blocking experiments with alloantiserum directed against the H-2 antigens of the target cells. After a single intraperitoneal (ip) injection of allogeneic spleen cells, CTX of unfractionated peritoneal cellswas strong, with a sharp peak on day 5. Repeated ip immunization markedly increased the CTX of unfractionated peritoneal cells. The reaction was strongest when the test was done at 37 degrees C. Sometimes CTX should be detected after as little as 6 hours' incubation. CTX depended primarily on the absolute number of effector or target cells per area rather than on the ratio of effector to target cells. Both nonadherent and adherent peritoneal cells destroyed target cells specifically. The CTX of nonadherent peritoneal cells was increased by 2-mercaptoethanol. The CTX reaction depended on effector cells bearing Tyl-1. Destruction of "innocent bystander" target cells was seen with one of four combinations of unfractionated and nonadherent peritoneal cells from hyperimmune animals.

Two isotypes of murine nm23/nucleoside diphosphate kinase, nm23-M1 and nm23-M2, are involved in metastatic suppression of a murine melanoma line.

A series of sublines of a murine melanoma B16 of C57BL/6 origin were established and examined regarding their metastatic capacity and expression of nm23. The number of pulmonary metastases developed by these sublines was inversely correlated with the expression of two isotypes of nm23, nm23-M1 and nm23-M2. The cDNAs of nm23-M1, nm23-M2, and a combination of both were transfected into the highly metastatic melanoma subline FE7, with low nm23 expression. FE7 transfectants of any of these cDNAs expressed transfected genes, and their metastatic capacity was suppressed when compared with parental FE7 or FE7 transfected with a control neo gene. These cell lines, however, did not change in terms of in vitro growth in the presence of 3 or 10% fetal bovine serum and in vivo growth when injected s.c. into C57BL/6-nu/nu mice. Similar experiments were also performed using FE7 transfectants of human nm23 genes. Transfectants of nm23-H1, nm23-H2, and their combination did not present altered metastatic potential. These findings indicated that two murine isotypes of nm23 but not those of humans are intimately related with the suppression of metastasis in the murine body.

Genetic and enzymatic basis for the differential expression of GM2 and GD2 gangliosides in human cancer cell lines.

Using beta 1,4-N-acetylgalactosaminyltransferase (EC 2.4.1.92) complementary DNA, the correlation of gene expression, enzyme activity, and expression of ganglioside antigens was analyzed in 20 human tumor cell lines. In many lines, GM2 and/or GD2 were the most complex structures examined. Northern blot analysis revealed 5.2- and 3.0-kilobase mRNAs in almost all cell lines expressing GD2 and/or GM2. Some melanoma lines, however, showed no bands although they expressed fairly high levels of GD2. These cell lines expressed very high levels of alpha 2,8-sialyltransferase and the resulting product, GD3. Semiquantitative RT-PCR demonstrated that even cell lines with no bands in Northern blot contained 0.4-2.5% of mRNA level in the highest expressing cell line. These results indicate that GD2 expression on individual cell lines is regulated not only by the expression level of the N-acetylgalactosaminyl transferase but also by the amount of its precursor structure (GD3) and alpha 2,8-sialyltransferase present in the cells. beta 1,4-N-acetylgalactosaminyltransferase activities and mRNA levels generally correlated quite closely. A few lines, however, showed lower enzyme activities than expected from their mRNA levels, indicating the possibility that the enzyme is being regulated by translational or posttranslational modification such as phosphorylation and glycosylation as well as by transcriptional regulation. Depending on their patterns of ganglioside synthesis and expression, the lines examined were classified into 6 groups which were characteristic of different tumor cell types.

Isolation of human nm23 genomes and analysis of loss of heterozygosity in primary colorectal carcinomas using a specific genomic probe.

Genomic clones of nm23-H1 and -H2 were isolated. The nm23-H1 and -H2 genes were located in a tandem array 4 kilobases apart. Each genome contained 5 exons and most of the splicing sites in the exon-intron junctions of two isotypes were essentially identical. A probe derived from intron 4 of nm23-H1 was used to examine the loss of heterozygosity (LOH) in primary colorectal carcinomas. Twenty-nine of 42 samples were informative and LOH was found in 3 of these 29. In all three samples with LOH, cancer tissues expressed lower levels of nm23-H1 mRNA when compared with those without LOH.