Lymphokine-activated killer cells targeted by monoclonal antibodies to the disialogangliosides GD2 and GD3 specifically lyse human tumor cells of neuroectodermal origin.

Monoclonal antibodies 14.18 (IgG3) and 11C64 (IgG3) directed against disialogangliosides GD2 and GD3, respectively, when used in conjunction with human peripheral blood mononuclear cells (PBMCs) stimulated with human recombinant interleukin (rIL-2) lyse both human melanoma and neuroblastoma cells by antibody-dependent cellular cytotoxicity. Such monoclonal antibody-"armed" effector cells are specifically directed to targets expressing the given disialoganglioside without detectable cross-reactivity. In addition, antibody-dependent cellular cytotoxicity as well as the natural killing ability of human PBMCs is augmented by a brief coincubation with rIL-2. PBMCs augmented by rIL-2 and armed with monoclonal antibodies significantly suppressed tumor growth in the xenotransplant nude mouse model. Our results suggest that once a threshold level of activation of PBMCs is achieved, additional rIL-2 (over three orders of magnitude of concentration) does not significantly enhance cytolytic augmentation. Furthermore, anti-GD3 monoclonal antibody 11C64 together with rIL-2-stimulated PBMCs from melanoma patients with widely differing tumor burdens effectively lyse melanoma tumor targets in antibody-dependent cellular cytotoxicity. Our results also suggest that GD2 and GD3 represent distinct and relevant immunotherapeutic target structures on melanoma whereas GD2 does the same for neuroblastoma tumors. Our data suggest that targeting of activated human effector cells may provide a new and effective cancer immunotherapy protocol.

Activation of human peripheral blood mononuclear cells by anti-T3: killing of tumor target cells coated with anti-target-anti-T3 conjugates.

Cytotoxic T-cell clones have been shown to be capable of killing unrelated target cells in the presence of heteroconjugates consisting of OKT3 (monoclonal antibody against cell surface antigen T3) and targeting antibody. In contrast, human peripheral blood T cells were not known to have this ability, although these cells have been shown to undergo proliferation in response to OKT3 and accessory signals. Here we report that OKT3-stimulated human peripheral blood mononuclear cells effectively killed human melanoma or Raji cells coated with anti-target-cell--OKT3 conjugates. The extent of target-cell killing by stimulated effector cells was much greater than that observed with unstimulated cells, in spite of down-regulation of the T3 structure on the OKT3-treated cells. Target-cell lysis occurred at physiological concentrations of effector cells and was inhibited by monomeric OKT3, indicating that killing was T-cell-mediated.

Generation of stable cellular phenotypes in a human malignant cell line conditioned by alterations in the cellular microenvironment.

Microenvironmental conditions may result in phenotypic changes of malignant cells and/or selection of preexisting variants with a growth advantage under the new growth conditions. The present study was initiated to evaluate the stability of changes in crucial cellular attributes as induced in vivo by nonimmunological mechanisms. Expression of major histocompatibility antigens, membrane immunoglobulin, in vitro growth rate, chromosome complement, and modal chromosome number were thus examined in a human B-lymphoma line (RH-L4) prior to and after short-term passage through the peritoneal cavity or spleen of newborn mice. Seven sublines (four from spleen and three from peritoneal cavity) of RH-L4 cells were established after the passage. These lines were found to differ phenotypically from the original line in respect to several or all of the attributes that were studied. These differences were stable for greater than 100 cell generations. Analyses of major histocompatibility complex Class II antigen expression indicated that the modulation of this antigen was independent of the immunological competence of the mice and unrelated to cell cycle-dependent variations. The emergence of the variant sublines seemed not to reflect a microenvironmentally determined selection of minority subpopulation in the original RH-L4 lymphoma line but, instead, induced by the growth conditions in the cellular microenvironment in the mouse. It is suggested that inductive processes may be of importance in the generation of phenotypic diversity within individual tumor cell populations, including the generation of phenotypic variants with high metastatic activity.

Disialoganglioside GD3 on human melanoma serves as a relevant target antigen for monoclonal antibody-mediated tumor cytolysis.

Monoclonal antibody MB3.6 (IgG3) recognizes disialoganglioside GD3, which represents a major surface marker on most human melanoma cells. We demonstrate that this antibody effectively lyses four human melanoma cell lines expressing significant levels of GD3 on their surface by either of two mechanisms: antibody-dependent cellular cytotoxicity (ADCC) or complement-mediated cytotoxicity. However, a melanoma cell line that expresses minimal levels of GD3 on 13% of the cells shows insignificant lysis by MB3.6 by either of these two mechanisms, suggesting that a threshold level of antigen expression may be required for effective in vitro cytolysis. In addition, monoclonal antibody (MAb) MB3.6 effectively inhibits establishment and growth of human melanoma tumors in the nude mouse when injected 24 hr after subcutaneous inoculation of tumor cells. Furthermore, MB3.6 produces specific regression of established melanoma tumors when injected 7 days after the subcutaneous inoculation of tumor cells. In contrast, tumor growth in animals injected with the melanoma cell line expressing minimal levels of GD3 was not affected by MAb MB3.6. These data indicate that once appropriate levels of the GD3 ganglioside are expressed on human melanoma cells, MAb MB3.6 can mediate tumor cell killing in vitro and in vivo and, thus, may prove useful for effective immunotherapy of human malignant melanoma.

Antibody producing human-human hybridomas. I. Technical aspects.

Technical aspects of generation of antibody-secreting human-human hybridomas are evaluated as based on 100 human-human fusions with a human B-lymphoma cell line (RH-L4) or the SKO-007 myeloma cell line as malignant fusion partners, and compared with similar fusion conditions in the mouse hybridoma system. The yield of hybrids was significantly lower when normal peripheral blood lymphocytes were used as fusion partners as compared with spleen lymphocytes, but could be substantially improved by increasing the amount of mitotic active B-lymphocytes by mitogen stimulation of the lymphocytes, preferably in HAT medium, prior to fusion. Furthermore, human hybrids grew slower and had a higher degree of chromosomal instability than usually observed in the mouse hybridoma system. Thus, out of 72 fusions, only 3 stable hybrids with antibody production against a predefined antigen were established. The importance of improved sources of human B-lymphocytes for human-human hybridoma production is discussed and methods of obtaining such improvement suggested.

Phenotypic characterization of mice of thymus target cells susceptible to productive infection by the radiation leukemia virus.

The spread of virus replication was studied by electron microscopy in the thymuses of inbred C57BL/Ka mice after intrathymic inoculation of the radiation leukemia virus (RadLV). The first type C-budding virus particles appeared in scarce blast cells of the subcapsular zone. Most of these blast cells were "X-cells," i.e., the thymus lymphoid cells most actively engaged in DNA synthesis. Virus replication spread to the entire cortical blast cell population and, from day 7 on, to the small cortical lymphocytes. The first virus-producing cells were derived from a very few target cells (approximately 0.001-0.003% of thymocytes) susceptible to RadLV infection. For determination of the phenotypes of these target cells, various thymocyte subpopulations obtained through a battery of cell separation methods were tested for their ability to support the replication of RadLV/VL3 virus in short-term culture. Most of these target cells were sensitive to the lytic effect of hydrocortisone and migrated in the fastest fraction of a 1Xg sedimentation gradient, together with the majority of [3H] thymidine-incorporating blast cells. They exhibited an intermediate density and expressed H-2 and Thy 1,2 cell surface antigens, although they were not found preferentially among the high Thy 1,2 population to which most of the cortical blast cells belonged. The spread of RadLV within the thymus and the surface phenotype characteristics of target cells indicate that these cells correspond to the thymocyte subset at the earliest stage of thymic lymphopoiesis and may be transitional between the prothymocytes and the subcapsular blast cell population.

Marrow-thymus interactions during radiation leukemogenesis in C57BL/Ka mice.

Transplantation of thymus and bone marrow cells from irradiated C57BL/Ka mice demonstrated the presence of potentially neoplastic cells in the thymus at 30 to 60 days postirradiation. During the same interval, no such cells could be detected in the bone marrow; moreover, the capacity of bone marrow cells to repopulate the thymus was impaired severely. These observations suggest that the primary site of neoplastic transformation in irradiated C57BL/Ka mice is the thymus rather than the bone marrow and that impaired thymic regeneration is a critical step in radiation leukemogenesis in mice.

Macrophage and lymphocyte-depleted thymus reticuloepithelial cell cultures: establishment and functional influence on T-lymphocyte maturation, C-type virus expression and lymphomatous transformation in vitro.

A method of cultivation has been developed that yields thymus reticuloepithelial (TRE) cell monolayers composed almost entirely of epithelial-like cells, with no detectable macrophages or lymphocytes. These cultures show the capacity to induce responsiveness to lectin mitogens in immature thymocytes. Such monolayers induce Thy-1 expression in nude mouse spleen and bone marrow cells. The functional macrophage-lymphocyte depleted TRE monolayers can be cryptically infected by radiation leukemia virus (RadLV). However, RadLV-infected TRE cells lack the capacity to induce neoplastic transformation in normal weanling and newborn thymocytes. Cultures prepared from C57BL/Ka mice treated with a leukemogenic course of irradiation produce a B-fibrotropic nonleukemogenic virus, and normal thymocytes cocultivated with these monolayers do not undergo in vitro transformation. It is concluded that TRE monolayers depleted of macrophages and lymphocytes: (1) exhibit the functional capacity to induce some of the properties of mature T cells; (2) are infectable by RadLV; (3) are one of the sites for expression of ecotropic virus after leukemogenic fractionated irradiation; and (4) are not capable of transforming normal thymocytes in vitro.

Detection of infectious centers in C57BL/Ka lymphoid cell populations infected in vitro by the radiation leukemia virus.

The cocultivation of nonproducer lymphoma cells derived from a radiation-induced lymphoma of the C57BL/Ka mouse with cultures of lymphoid cell populations from the thymus, spleen, and marrow of the same strain 48 hr after their infection by the C57BL/Ka leukemia viruses permits the detection of infectious centers in these cultures. A quantitative assay is described which allows the estimation in lymphoid cell subpopulations of the numbers of target cells susceptible to productive infection by the thymotropic and leukemogenic viruses of C57BL/Ka mice in vitro. This assay should greatly facilitate the identification and characterization of such target cells.

In vivo activity of lymphocytes sensitized in vitro by antigen-fed macrophages: inhibition of lymphoma growth.

In previous studies, we have shown that macrophages fed with radiation leukemia virus could induce primary in vitro sensitization of lymphocytes which could be measured by their cytotoxic activity against target cells. In the present study, we tested the in vivo influence on tumor growth of such lymphocytes. We found that macrophage-mediated sensitized lymphocytes could protect mice against tumor growth if injected into normal recipients four days prior to challenge with lymphoma cells. The protective function of such lymphocytes was not affected by their irradiation, but no protection occurred in sublethally irradiated recipients. This indicated that the sensitized lymphocytes did not inhibit tumor cell growth directly but recruited an effector protective response in the recipient mice. This protective activity was different from the one elicited by lymphocytes which had been sensitized directly against cells carrying antigens cross-reacting with RadLV. The protective activity of the directly sensitized lymphocytes was radiosensitive and was probably mediated by their direct cytotoxic activity against tumor cells.