Semiallogeneic cancer vaccines formulated with granulocyte-macrophage colony-stimulating factor for patients with metastatic gastrointestinal adenocarcinomas: a pilot phase I study.

The authors report the results of a phase I clinical study using semiallogeneic cancer vaccines formulated with granulocyte-macrophage colony-stimulating factor (GM-CSF) to treat patients with metastatic adenocarcinomas of the gastrointestinal tract. A specially engineered cell line, FO1-12, was used to generate semiallogeneic hybrids by fusion with patient-derived tumor cells; the hybrids express HLA class I and II haplotypes derived from both parental cells. For treatment, the vaccine was mixed with GM-CSF, irradiated, and injected intradermally into patients at weekly or biweekly intervals. Vaccinations were associated with minimal or no toxicity and showed that semiallogeneic hybrids formulated with GM-CSF can induce a specific antitumor immune response in some patients, as measured by a delayed-type hypersensitivity response to autologous tumor cells. Because of the simplicity, feasibility, and flexibility of this immunotherapeutic approach, semiallogeneic hybrid vaccines have the potential to be used in the treatment of virtually any type of cancer.

Semi-allogeneic cell hybrids stimulate HIV-1 envelope-specific cytotoxic T lymphocytes.

OBJECTIVE: The present study was designed to determine whether the HLA allogeneic T helper response stimulated by semi-allogeneic cell lines could be used as an in vitro model of immune-based therapy to stimulate HIV-specific cytotoxic T lymphocytes. DESIGN AND METHODS: Semi-allogeneic cell hybrids were obtained by the fusion of peripheral blood mononuclear cells from HIV-infected patients with the allogeneic beta2-microglobulin-deficient FO1-12 melanoma cell line. These hybrids were used as antigen presenting cells for HIV envelope peptide (env)-specific cytotoxic assays. RESULTS: The hybrid cell lines express HLA class I and II antigens from both parental cells, as well as the CD86 costimulatory molecule. HIV-specific cytotoxic T lymphocyte activity was obtained when patients' peripheral blood mononuclear cells were costimulated with env peptides plus semi-allogeneic hybrids, in contrast with stimulation with either env or hybrid cells alone. Thus, the semi-allogeneic hybrids enhanced HIV-specific killing of target cells. CONCLUSIONS: Irradiated, semi-allogeneic cell hybrids engineered for individual AIDS patients provide efficient and simultaneous co-recognition of HLA allogeneic determinants and viral antigenic determinants presented by self-HLA molecules on the same antigen presenting cells and results in the generation of enhanced HIV-specific cytotoxic T lymphocyte activity.

Semiallogeneic cell hybrids as therapeutic vaccines for cancer.

The authors have engineered a cell line that can be used in human studies as a universal donor cell for the formation of semiallogeneic cell hybrids after fusion with patient-derived tumor cells. These hybrids can be irradiated and injected as a patient-tailored therapeutic vaccine in patients affected by virtually any type of cancer. A crucial step in this research effort has been the derivation of an allogeneic cell line (FO1-12) that expresses both a dominant selectable marker (neomycin resistance) and a recessive selectable marker (sensitivity to hypoxanthine, aminopterin, and thymidine), which allows easy selection of semiallogeneic cell hybrids derived from the fusion of FO1-12 cells with patient-derived tumor cells. Tumor-infiltrating lymphocytes derived from select patients with melanoma and exposed to semiallogeneic cell hybrids from the same patient were better able to specifically lyse autologous tumor cells. Furthermore, FO1-12 cells express carcinoembryonic antigen, which is ubiquitous in adenocarcinomas, and fusion of FO1-12 cells with various patient-derived adenocarcinoma cells showed that the hybrid cells also express carcinoembryonic antigen. Because of the results of these preclinical studies, the authors were given permission to use semiallogeneic cell hybrids for immunotherapy of patients with metastatic melanoma or metastatic adenocarcinoma who had not responded to standard treatment regimens. Treatment with semiallogeneic vaccines is associated with minimal or no toxicity and can induce a specific anti-tumor immune response.

Characterization of a sustained-release delivery system for combined cytokine/peptide vaccination using a poly-N-acetyl glucosamine-based polymer matrix.

Identification of tumor-associated antigens (TAAs) and their class I MHC-restricted epitopes now allows for the rational design of peptide-based cancer vaccines. A biocompatible system capable of sustained release of biologically relevant levels of cytokine and TAA peptide could provide a more effective microenvironment for antigen presentation. Our goal was to test a sustained-release cytokine/TAA peptide-based formulation using a highly purified polysaccharide [poly-N-acetyl glucosamine (p-GlcNAc)] polymer. Granulocyte-macrophage colony-stimulating factor (GM-CSF; 100 microgram) and MART-1(27-35) peptide (128 microgram in DMSO) were formulated into p-GlcNAc. Peptide release was assayed in vitro using interleukin 2 production from previously characterized MART-1(27-35)-specific Jurkat T cells (JRT22). GM-CSF release was assayed via ELISA and proliferation of M-07e (GM-CSF-dependent) cells. Local bioavailability of MART-1(27-35) peptide for uptake and presentation by antigen-presenting cells was demonstrated for up to 6 days (>0.5 microgram/ml). More than 1.0 microgram/ml GM-CSF was concomitantly released over the same period. Biocompatibility and local tissue response to p-GlcNAc releasing murine GM-CSF was determined in C57BL/6 mice via s.c. injection using murine GM-CSF (0. 2 microgram/ml) in 200 microliter of a 2.5% polymer gel. Significant lymphocytic and eosinophilic infiltration was observed 2-7 days after injection with polymer containing murine GM-CSF. The results of our studies show that this biocompatible system is capable of a sustained concomitant release of biologically active peptide and cytokine into the local microenvironment. These findings support further studies to validate a p-GlcNAc delivery system vehicle for a cytokine/TAA peptide-based cancer vaccine.

Melanoma-associated tumor antigens and their clinical relevance to immunotherapy.

The last few years have witnessed the publication of a large body of evidence demonstrating conclusively the existence of tumor-associated antigens. A large majority of these studies focused on melanoma-associated tumor antigens because of the collective evidence that the immune system can influence the pathogenesis of melanoma, and because of the well-documented, although limited, success of immunotherapeutic modalities in melanoma patients. This review summarizes what is known about melanoma-associated antigenic peptides: their identity, presentation by human leukocyte antigen class I molecules to cognate T cell receptors, and their potential to induce an effective immune response. The inability of melanoma patients to mount an efficacious antitumor response and the distinction between antigenicity (i.e., the ability to express a tumor antigen) and immunogenicity (i.e., the ability to elicit an effective immune response) are discussed. Recruitment of antigen-presenting cells at the tumor site is suggested as a way to overcome tumor-induced immunotolerance. The importance of developing or perfecting laboratory and/or clinical correlates of response to immunotherapeutic modalities is emphasized because of the pressing need for reliable tests that are predictive of clinical outcome.

delta 12-Prostaglandin-J2 is cytotoxic in human malignancies and synergizes with both cisplatin and radiation.

We have been investigating the synergistic cytotoxic interactions between tamoxifen (TAM) and cisplatin (DDP) in human malignant cell lines. Recent data have demonstrated that TAM activates phospholipase D, which can increase the production of prostaglandin D2. Prostaglandin D2 has been shown to have growth inhibitory properties in several malignant cell lines. delta 12-Prostaglandin-J2 (delta 12-PG J2) is a derivative of prostaglandin D2 that has been shown to have similar inhibitory properties. We hypothesized that TAM may increase the production of delta 12-PG J2, which in turn may synergize with DDP. To begin our investigation of this interaction, we sought to determine if delta 12-PG J2 was cytotoxic and synergistic in our melanoma system and then expanded our observations to include a wide range of malignant cells. We have demonstrated that delta 12-PG J2 is cytotoxic to multiple malignant cell lines including melanoma, ovarian, prostate, colon, pancreas, small cell lung cancer, and breast cancer lines. The IC50s ranged from 0.70 microM (small cell lung cancer) to 3.30 microM (DDP-resistant melanoma). Additionally, delta 12-PG J2 exhibited synergistic cytotoxicity with both DDP and ionizing radiation. These data suggest that delta 12-PG J2 should be further evaluated in an in vivo model to confirm activity.

Organ-specific metastases in melanoma: experimental animal models.

Metastases from certain primary tumors frequently exhibit specific organ preference. Animal models have been developed to induce in a reproducible fashion the formation of organ-specific metastases by malignant melanoma cells. Some of these models rely on the use of immunodeficient mice, which can support the growth of murine as well as human malignant melanomas. Moreover, immunodeficient mice, because of their diminished ability to mount an effective immune response, allow the expression of malignant properties (e.g., preferential colonization of certain organs), which are intrinsic to transplanted melanoma cells. This review discusses the relevant factors (and limitations) of some of the animal models used to study the in vivo properties of melanoma cells.

Organ-specific metastases in immunodeficient mice injected with human melanoma cells: a quantitative pathological analysis.

Pathological and morphometric techniques were used to investigate the potential of two human melanoma cell lines for organ colonization in three different immunodeficient mouse strains; nude (nu/nu), NIH triple immunodeficient (TID: nu/nu, bg/bg, xid/xid) and severe combined immunodeficient (SCID) mice. The MM-RU cell line gave rise exclusively to lung metastases, whereas the MM-AN cell line gave rise to lung and extrapulmonary metastases. Although the TID mice showed more pancreatic and brown fat lesions than nude or SCID mice, the overall pattern of distribution of organ metastases among the strains was similar, suggesting that cellular properties intrinsic to the melanoma cells are important for the colonization of specific organs. The metastatic nodules were well circumscribed in all organs and exhibited peripherally located macrophages, except for brain metastases, where a more invasive pattern along vasculature was observed. The differences in cellular infiltrate and infiltrative patterns of the tumors implicate features of the host microenvironment (organ-specific factors) which are, at least in part, independent of the host's genetic background or degree of immunodeficiency. Our findings suggest that intrinsic malignant cellular properties play an important role in organ-specific colonization by haematogenously metastasizing cells.

Modulation by MHC class I antigens of the biology of melanoma cells. Non-immunological mechanisms.

Many human as well as experimental tumours, including melanoma, express reduced levels of major histocompatibility complex (MHC) Class I antigens. Decreased MHC Class I antigen expression may be selected during neoplastic progression because it allows tumour cells to escape killing by cytotoxic T lymphocytes. Furthermore, the regulatory role of MHC Class I antigens in the proliferation of T cells suggests that abnormalities in MHC Class I antigen expression may play a role in the disordered proliferation of malignant cells and in their metastatic potential by non-immunological mechanisms. This paper reviews some of the available evidence supporting the concept of non-immune functions of MHC Class I antigens in the biology of malignant cells, with emphasis on experimental models for metastatic melanoma.

Selective loss of human leukocyte class I allospecificities and staining of melanoma cells by monoclonal antibodies recognizing monomorphic determinants of class I human leukocyte antigens.

Immunoperoxidase staining of frozen sections from surgically removed melanoma lesions showed that anti-human leukocyte antigen (HLA)-A2, A28 monoclonal antibody (mAb) stained keratinocytes, but did not stain melanoma cells in 21% of the 14 primary and 44% of the 9 metastatic lesions tested. The loss of HLA-A2 and/or A28 allospecificities did not affect the staining patterns with mAb recognizing monomorphic determinants of HLA Class I antigens, in terms of percentage of stained melanoma cells and intensity of staining. This finding is not likely to reflect the sensitivity of the immunoperoxidase technique, since cytofluorographic analysis detected no significant difference in the staining pattern by mAb to monomorphic determinants of HLA Class I antigens between a melanoma cell line and an autologous transfectant that had acquired HLA-A2 antigens following gene transfer. The results of the present study imply that the frequency of abnormalities in HLA Class I antigen expression by melanoma cells is higher than that described in the literature, since selective losses of HLA Class I allospecificities are not detected by staining of melanoma cells with mAb to monomorphic determinants of HLA Class I antigens. The latter reagents have been used in most of the published studies to characterize the expression of HLA Class I antigens in melanoma lesions. Furthermore, the present results provide a mechanism for the unexpected resistance to cytotoxic T-cell-mediated lysis and the unexpected poor clinical course of the disease in some patients despite a high expression of HLA Class I antigens as measured by staining of melanoma cells with mAb to monomorphic determinants.

Radiation-enhanced expression of major histocompatibility complex class I antigen H-2Db in B16 melanoma cells.

Exposure of eukaryotic cells to ionizing radiation induces several cellular responses including DNA repair, arrest of DNA synthesis, and increased synthesis of specific cellular proteins. We derived from the murine melanoma cell line B16-F10 a clonal isolate (M1) that was exposed to a total dose of 5000 cGy in 25 fractions, according to a protocol that reflects the standard for current radiotherapeutic regimens. We measured, by flow cytometry of fluorescence-stained cells, the surface expression of the two major histocompatibility complex class I antigens H-2Db and H-2Kb in irradiated M1 cells and untreated M1 controls. We found that after 2000 cGy, expression of H-2Db antigen was enhanced in irradiated cells versus controls. Radiation-induced expression of H-2Db antigen appeared to be selective, since no up-regulation of the H-2Kb antigen was detectable, and persisted for at least 5 weeks following the last irradiation. Enhanced H-2Db antigen expression correlated with increased steady-state levels of H-2Db mRNA in irradiated cells. These results are consistent with the notion that enhanced expression of major histocompatibility complex class I antigens is part of a long-lasting stress response elicited in cells by radiation.

Role of beta-1 integrins in organ specific adhesion of melanoma cells in vitro.

BACKGROUND: Recent data suggest that the extracellular matrix of organs and heterogeneous integrin expression of tumor cells may influence metastasis distribution. EXPERIMENTAL DESIGN: Three human melanoma cell lines were characterized for integrin expression, in vitro binding to cryostat sections of different organs, and ability to generate experimental metastases in triple immunodeficient mice. RESULTS: The three cell lines exhibited heterogeneous expression of integrins, binding to cryostat sections, and organ colonization. A primary melanoma cell line (PM-WK) did not give rise to experimental metastases, showed scant or mild attachment to only a few organ tissue sections, and showed absent or minimal expression of alpha-integrin subunits tested (VLA 1-6) and alpha v beta 3. In contrast, two lymph node derived lines exhibited distinct patterns of organ colonization: MM-RU colonized only the lungs and expressed predominantly alpha 2 beta 1 and alpha v beta 3 integrin, whereas MM-AN colonized lung and extrapulmonary sites including pancreas and subcutaneous brown fat and expressed predominantly alpha 2 beta 1 and alpha 6 beta 1 integrin. In vitro, MM-RU exhibited marked attachment to lung, brown fat, kidney, and adrenal with no binding to liver, pancreas, brain, or muscle tissue sections, whereas MM-AN had a similar binding profile but with additional attachment to liver and pancreas. Function blocking anti-beta 1 monoclonal antibody inhibited the attachment of MM-RU and MM-AN cells to these tissues (p < 0.001), whereas function blocking anti-alpha 5 and an unrelated monoclonal antibody (HLA class I) did not. Function blocking anti-alpha 2 monoclonal antibody inhibited MM-RU cell adhesion (p < 0.001) but not MM-AN adhesion. However, the function blocking monoclonal antibody alpha 6 beta 1 significantly inhibited the binding of MM-AN to these tissues. CONCLUSIONS: These data suggest that alpha 2 beta 1 and alpha 6 beta 1 mediate differential melanoma cell attachment to organ tissue sections in vitro and that differences in integrin expression of these melanoma cells may be involved in differential organ colonization in vivo.

Actin organization and cell migration of melanoma cells relate to differential expression of integrins and actin-associated proteins.

We have recently described marked differences in cell migration rates and organization of actin in human melanoma cell lines isolated from various stages of tumor progression. Metastatic lines derived from lymph node metastases organized actin into stress fiber arrays and had high mean migration rates in vitro when compared to lines from other stages. Melanoma cells also reveal marked differences in localization of alpha-actinin and beta 1 integrins at stress fiber termination sites (focal contacts). Disruption of this organization is induced by antibodies against beta 1 integrins, alpha-actinin, recently postulated as having a role in linkage of actin to beta 1 integrins, is differentially expressed in melanoma cells by Northern blot analysis and a relatively high alpha-actinin to actin ratio is associated with stress fiber formation and increased cell migration. Furthermore, actin-binding protein, which cross-links actin filaments, is also significantly increased in lines exhibiting high migration rates. Control of migration and actin organization may be mediated by extracellular matrices and/or modulation of actin-associated proteins including alpha-actinin and actin binding protein. These findings provide evidence that an interaction of transmembrane adhesion molecules and elements of the cytoskeleton in melanoma cells may be responsible for differences in migration rates and capacity for metastasis.

Expression of matrix metalloproteinase genes in transformed rat cell lines of high and low metastatic potential.

The enzymes that comprise the family of matrix metalloproteinases (MMPs) share the capacity to degrade extracellular matrix components. A large body of evidence indicates that certain members of this metalloproteinase gene family play critical roles in determining the malignant phenotype of solid tumors. We previously have derived transformed cell lines with vastly different metastatic potentials by transfecting different combinations of oncogenes into primary rat embryo cells. Conditioned medium from those cell lines was assayed by Western blot analysis for the production of four separate matrix metalloproteinases to see whether a correlation could be found between protease expression and the metastatic phenotype. The transformed rat embryo cell lines with high metastatic potential were found to produce high levels of the stromelysin 1 (MMP-3) and stromelysin 2 (MMP-10) proteases, while the nonmetastatic lines produced low or undetectable levels of these two enzymes. No correlation was seen between the metastatic phenotype of the cell lines and the level of expression of two other matrix metalloproteinases, the M(r) 72,000 type IV collagenase (MMP-2) and the M(r) 92,000 gelatinase (MMP-9). These data suggest that the differential regulation of the stromelysin proteases may contribute to the difference seen in the metastatic potential of these cell lines.

Expression of a transfected H-2Kb gene in B16 cells correlates with suppression of liver metastases in triple immunodeficient mice.

In vivo experiments performed with NIH (nu/nu, bg/bg, xid/xid) triple immunodeficient (TD) mice revealed the striking ability of i.v. injected B16-F1 and B16-F10 murine melanoma cells to colonize not only the lungs but also the liver of TD mice. Subsequently, B16 melanoma cell cultures, which express very low levels of H-2Kb antigen, were cotransfected with plasmids pRSVneo, containing the neomycin resistance gene, and 6-2B1pMT, expressing the H-2Kb complentary DNA under the control of the metallothionein enhancer-promoter. Several neomycin-resistant clones were analyzed for H-2Kb and H-2Db expression by RNase protection and flow cytometry assays. All parental lines and transfected clones expressed normal levels of H-2Db mRNA, while only some of the transfected clones expressed easily detectable levels of H-2Kb mRNA. Moreover, in these clones H-2Kb expression could be enhanced in the presence of Zn2+, indicating that the metallothionein enhancer was functioning properly. Parental cells and transfected clones were injected i.v. in TD mice to assess the possible involvement of H-2Kb antigen in regulating the metastatic potential of B16 melanoma cells. We observed a remarkable correlation between expression of H-2Kb antigen and suppression of liver-specific metastases in TD mice. Identical results were obtained when we gave TD mice injections of mixed populations of transfectants expressing H-2Kb antigen, obtained by fluorescence-activated cell sorting. These experiments allowed us to rule out the possibility that the observed changes in metastatic potential were due to clonal variability among individual transfected clones. Taken together, the results of our in vivo studies with immunodeficient mice support the notion that specific major histocompatibility complex Class I molecules modulate the metastatic potential of malignant cells also by mechanisms which are independent of their well-established role in antigen presentation.

Beta 2-microglobulin gene is mutated in a human colon cancer cell line (HCT) deficient in the expression of HLA class I antigens on the cell surface.

The human colon cancer cell line HCT does not express any detectable HLA class I antigens on the cell surface. RNA blot analyses showed that HCT cells synthesize easily detectable levels of heavy chains as well as beta 2-microglobulin (beta 2m) transcripts. Experiments of immunoprecipitation revealed the presence of intracellular HLA heavy chains and the absence of beta 2m molecules. Sequencing studies, performed on polymerase chain reaction-mediated amplification of beta 2m-specific complementary DNAs, indicated that in HCT cells both beta 2m genes are mutated. The first mutation consists of an 11-base deletion, corresponding to the first 11 base pairs of the second exon of the beta 2m gene. This mutation alters the reading frame, starting from the third amino acid residue of the mature beta 2m protein, resulting in the synthesis of a 31-amino acid peptide with no remarkable homology to any of the sequences stored in the protein database. The second mutation is a point mutation (C----A), resulting in a UAA stop codon corresponding to the 10th amino acid residue of the mature beta 2m. Therefore, it would appear that in HCT cells the beta 2m genes have undergone two different mutational changes. This is the first molecular demonstration of beta 2m mutations in a human epithelial cell line.

Expression of HLA-A2 antigen in human melanoma cell lines and its role in T-cell recognition.

Previous studies have suggested that, in human melanoma, expression of HLA-A2 antigen is important for tumor cell recognition by autologous T-lymphocytes. Because of the recent demonstration that expression of HLA Class I antigens may be selectively lost in several human tumors, including melanoma, we derived pairs of tumor infiltrating lymphocytes (TIL) and melanoma cell lines from 4 human lymphocytic antigen (HLA)-A2+ patients with metastatic melanoma. We observed that, although all 4 TIL cultures expressed HLA-A2 antigen, only 2 melanoma cell lines did so. Melanoma cells derived from the other 2 patients showed neither surface expression of the HLA-A2 antigen nor presence of the corresponding mRNA. We also observed some correlation between loss of HLA-A2 expression and level of c-myc transcription. TIL derived from patients whose melanoma cell lines had normal expression of HLA-A2 had a CD8 phenotype and were capable of lysing autologous melanoma cells. Melanoma cell killing was CD3 and major histocompatibility complex Class I restricted in both cases, but HLA-A2 restricted in only one case. On the other hand, TIL derived from the 2 patients whose melanoma cell lines had lost expression of HLA-A2 had a predominant CD4 phenotype and virtually no cytotoxic activity. Preincubation of the HLA-A2 negative melanoma cell lines with alpha- or gamma-interferon did not induce the re-expression of the HLA-A2 antigen. In an attempt to restore HLA-A2 antigen expression in one of the melanoma cell lines that were HLA-A2 negative, we transfected these cells with the HLA-A2 gene subcloned in the pSV2-neo vector. Four transfected clones, with high levels of HLA-A2 antigen expression, were expanded and characterized. Proliferative and cytotoxic activities of TIL against the autologous transfected clones as well as the untransfected parental melanoma cell line were measured and compared. CD4+ TIL showed no difference in the proliferative response to autologous parental and HLA-A2 transfected clones. However, we observed selective recognition of the HLA-A2 expressing clones by autologous cultured peripheral blood lymphocytes (which contained CD8 cells) as well as allogeneic CD8+ TIL with a HLA-A2 restricted pattern of recognition. In contrast, virtually no cytotoxic activity was detected against either parental or HLA-A2 transfected clones. Overall, our data suggest that selective down-regulation of HLA-A2 antigen expression in melanoma cells may represent one of the mechanisms by which tumor cells escape immunological recognition.

Epithelial glycoprotein is a member of a family of epithelial cell surface antigens homologous to nidogen, a matrix adhesion protein.

The cell surface antigen, epithelial glycoprotein, defined by the monoclonal antibody HEA 125, is expressed on virtually all epithelial cell membranes but not on mesodermal or neural cell membranes. The cDNA encoding epithelial glycoprotein was isolated by HEA 125 antibody enrichment of colon tumor cDNA expressed transiently in COS cells. The sequence of the epithelial glycoprotein antigen is identical to the cell membrane protein recognized by the monoclonal antibody KS 1/4 and is homologous to the tumor-associated antigen GA733. These proteins share sequence homology to nidogen, an extracellular matrix component that appears to participate in cell-matrix adhesion. These proteins also share a homologous domain found in the B1 chain of laminin, a matrix adhesion protein, and placental protein 12, an insulin-like growth factor I binding protein secreted during pregnancy that has been implicated in regulation of fetal growth. This common domain is also repeated multiple times within the thyroglobulin precursor. These findings suggest epithelial glycoprotein is a cell surface molecule involved in cell-cell or cell-matrix interaction.