Cell surface antigens of human astrocytoma defined by mouse monoclonal antibodies: identification of astrocytoma subsets

Identification of glioblastoma-specific antigens expressed in patient-derived tumor cells as candidate targets for chimeric antigen receptor T cell therapy

Background

New therapies for glioblastoma (GBM) are urgently needed because the disease prognosis is poor. Chimeric antigen receptor (CAR)-T cell therapy that targets GBM-specific cell surface antigens is a promising therapeutic strategy. However, extensive transcriptome analyses have uncovered few GBM-specific target antigens.

Methods

We established a library of monoclonal antibodies (mAbs) against a tumor cell line derived from a patient with GBM. We identified mAbs that reacted with tumor cell lines from patients with GBM but not with nonmalignant human brain cells. We then detected the antigens they recognized using expression cloning. CAR-T cells derived from a candidate mAb were generated and tested in vitro and in vivo.

Results

We detected 507 mAbs that bound to tumor cell lines from patients with GBM. Among them, E61 and A13 reacted with tumor cell lines from most patients with GBM, but not with nonmalignant human brain cells. We found that B7-H3 was the antigen recognized but E61. CAR-T cells were established using the antigen-recognition domain of E61-secreted cytokines and exerted cytotoxicity in co-culture with tumor cells from patients with GBM.

Conclusions

Cancer-specific targets for CAR-T cells were identified using a mAb library raised against primary GBM tumor cells from a patient. We identified a GBM-specific mAb and its antigen. More mAbs against various GBM samples and novel target antigens are expected to be identified using this strategy.

Cancer immunoediting in malignant glioma

Significant work from many laboratories over the last decade in the study of cancer immunology has resulted in the development of the cancer immunoediting hypothesis. This contemporary framework of the naturally arising immune system-tumor interaction is thought to comprise 3 phases: elimination, wherein immunity subserves an extrinsic tumor suppressor function and destroys nascent tumor cells; equilibrium, wherein tumor cells are constrained in a period of latency under immune control; and escape, wherein tumor cells outpace immunity and progress clinically. In this review, we address in detail the relevance of the cancer immunoediting concept to neurosurgeons and neuro-oncologists treating and studying malignant glioma by exploring the de novo immune response to these tumors, how these tumors may persist in vivo, the mechanisms by which these cells may escape/attenuate immunity, and ultimately how this concept may influence our immunotherapeutic approaches.

Engineered cell surface expression of membrane immunoglobulin as a means to identify monoclonal antibody-secreting hybridomas

Monoclonal antibodies (mAbs) have proven to be effective biological reagents in the form of therapeutic drugs and diagnostics for many pathologies, as well as valuable research tools. Existing methods for isolating mAb-producing hybridomas are tedious and time consuming. Herein we describe a novel system in which mAb-secreting hybridoma cells were induced to co-express significant amounts of the membrane form of the secreted immunoglobulin (Ig) on their surfaces and are efficiently recovered by fluorescent activated cell sorting (FACS). Fusion of a novel myeloma parent, SP2ab, expressing transgenic Igalpha and Igbeta of the B-cell receptor complex (BCR) with spleen cells resulted in hybridomas demonstrating order of magnitude increases in BCR surface expression. Surface Ig levels correlated with transgenic Igalpha expression, and these cells also secreted normal levels of mAb. Hundreds of hybridoma lines producing mAbs specific for a variety of antigens were rapidly isolated as single cell-derived clones after FACS. Significant improvements using the Direct Selection of Hybridomas (DiSH) by FACS include reduced time and labor, improved capability of isolating positive hybridomas, and the ease of manipulating cloned cell lines relative to previously existing approaches that require Limiting Dilution Subcloning (LDS).

The significance of immunohistochemistry in the diagnosis and therapy of neoplasms

This review article details the diagnostical significance of immunohistochemistry, which has developed during the last quarter of the century. Certainly, the advancement of monoclonal antibody technology has been of great significance in assuring the place of immunohistochemistry in the modern accurate microscopic diagnosis of human neoplasms, as a method of choice in histopathology. The fact still remains that in order to properly assess any immunohistochemical reactivity used for differential diagnostic purposes, the target cells have to be identified as neoplastically transformed cells by routine histopathological techniques. Selected groups of target molecules of great significance in cancer biology are discussed. The discovery of neoplasm-associated antigens has not only made the more accurate diagnosis of human cancer feasible but has also shed light on the extensive immunophenotypical heterogeneity of even the most closely linked human malignancies. The identification of disseminated neoplastically transformed cells by immunohistochemistry has allowed for a clearer picture of cancer invasion and metastasis, as well as the evolution of the tumour cell associated immunophenotype towards increased malignancy. Some possibilities of neoplasm-associated antigen targeted, receptor-directed immunotherapy are discussed and reviewed in this manuscript. Future antineoplastic therapeutical approaches should see the inclusion of a variety of immunotherapies, in the form of an individualised 'cocktail' specific for the particular immunophenotypical pattern associated with each individual patient's neoplastic disease.

gp120-induced alterations of human astrocyte function: Na(+)/H(+) exchange, K(+) conductance, and glutamate flux

Many human immunodeficiency virus (HIV)-infected patients suffer from impaired neurological function and dementia. This facet of the disease has been termed acquired immunodeficiency syndrome (AIDS)-associated dementia complex (ADC). Several cell types, including astrocytes and neurons, are not productively infected by virus but are involved in ADC pathophysiology. Previous studies of rat astrocytes showed that an HIV coat protein (gp120) accelerated astrocyte Na(+)/H(+) exchange and that the resultant intracellular alkalinization activated a pH-sensitive K(+) conductance. The present experiments were conducted to determine whether gp120 affected human astrocytes in the same fashion. It was found that primary human astrocytes express a pH-sensitive K(+) conductance that was activated on intracellular alkalinization. Also, gp120 treatment of whole cell clamped human astrocytes activated this conductance specifically. Furthermore, gp120 inhibited glutamate uptake by primary human astrocytes. These altered physiological processes could contribute to pathophysiological changes in HIV-infected brains. Because the gp120-induced cell physiological changes were partially inhibited by dimethylamiloride (an inhibitor of Na(+)/H(+) exchange), our findings suggest that modification of human astrocyte Na(+)/H(+) exchange activity may provide a means of addressing some of the neurological complications of HIV infection.

Decreased expression of alpha 2 beta 1 integrin in scleroderma fibroblasts

Systemic scleroderma (SSc) is a complex connective tissue disorder of unknown etiology. In early stages of the disease, fibroblasts are activated to produce large amounts of collagen with subsequent fibrosis. Collagen metabolism of fibroblasts is modulated by their contact with the extracellular matrix (ECM), which involves distinct receptors on the cell surface, mainly belonging to the integrins. We investigated the expression of collagen receptor alpha 2 beta 1 in SSc and normal fibroblasts, since this receptor has been shown to be utilized by fibroblasts for adhesion to and reorganization of collagen I. 9 strains of scleroderma fibroblasts grown as monolayer cultures were first analyzed with respect to their collagen I expression. 6 of these strains were similar to controls "low" producers) and 3 strains showed up to 2-3 x higher levels of collage I mRNA expression ("high" producers). Northern hybridization using a cDNA probe specific for the alpha 2 integrin subunit revealed a decrease of the corresponding mRNA in SSc fibroblasts as compared to controls (75% versus 100%). "High" collagen producing cell strains displayed the lowest values for alpha 2 integrin mRNA. The decrease of alpha 2 integrin subunit expression at the mRNA level in selected fibroblasts was further substantiated by radioimmunoprecipitation using specific mAbs directed against alpha 2 integrin subunit. No significant changes in beta 1 integrin expression could be observed - neither at mRNA nor at the protein level. Our data indicate a correlation between excessive synthesis of collagen and low levels of alpha 2 integrin subunit expression in SSc fibroblasts. Further experiments should clarify whether this observation is a phenomenon specific for scleroderma or whether it reflects an "activated" state of fibroblasts.

Surface antigens of cell subpopulations in prenatal rat brain are expressed in a characteristic non-random pattern on their ethylnitrosourea-induced malignant counterparts

Selective induction of neural tumors in the rat by single-dose exposure of the immature nervous system to ethylnitrosourea (EtNU) is a model for the study of cell lineage-, differentiation stage-, and carcinogen-dependent mechanisms in neuro-oncogenesis. Overall yields and relative frequencies of different types of neural tumors vary with the developmental window chosen for the EtNU-pulse. Precursor cells belonging to different neural lineages and targeted by the carcinogen at distinct developmental stages may thus bear a differential risk of malignant conversion. To specify subpopulations of neural precursors in fetal (prenatal day 18) BDIX-rat brain, four monoclonal antibodies (mAbs) recognizing cell surface differentiation antigens were used: mAb RB13-2 directed against O-acetylated gangliosides and binding to approximately 36% of fetal brain cells (FBC); mAb RB13-6 recognizing a 130 kDa glycoprotein (expressed by approximately 8% of FBC); and mAbs RB21-7 and RB21-15 which bind, respectively, to embryonal neural cell adhesion molecules (N-CAM) and a 24 kDa protein (expressed by approximately 55% and 12% of FBC). Antigen expression profiles were compared with those of 14 primary brain tumors and 16 malignant neural cell lines, all of which had been induced by EtNU on prenatal day 18 in vivo. Monoclonal antibodies RB13-2 and RB21-7 did not bind to any of the tumors or cell lines. In contrast, mAbs RB13-6 and RB21-15 both reacted with 14/14 tumors, and with 16/16 and 10/16 cell lines, respectively. Expression of the latter antigens might thus specify lineage-specific stages of FBC development/differentiation particularly susceptible to EtNU-induced malignant transformation. Two-color fluorescence analyses revealed three subsets of FBC binding mAb RB13-6 (RB13-2+/RB13-6+/RB21-15-; RB13-2-/RB13-6+/RB21-15-; and RB13-2-/RB13-6+/RB21-15+), representing successive stages of differentiation.

Integrin expression and ability to adhere to extracellular matrix proteins and endothelial cells in human lung cancer lines

We examined the integrin expression in 19 human lung cancer cell lines with monoclonal antibodies to the integrin subunits alpha 1, alpha 2, alpha 3, alpha 4, alpha 5, alpha 6, beta 1, beta 2, and beta 4. We measured their ability to adhere to the extracellular matrix (ECM) and human umbilical vein endothelial cells (HUVECs). Almost all lines expressed the beta 1 subunit and approximately half of the lines expressed the beta 4 subunit; by contrast, none expressed the beta 2 subunit. Subunits alpha 2, alpha 3, alpha 5 and alpha 6 were frequently expressed, whereas very few lines expressed alpha 1 and alpha 4. Most lines adhered strongly to ECM (type I collagen, laminin and fibronectin) in correspondence to their expression of integrins. Binding by most lines to fibronectin was completely inhibited by arginine-glycine-aspartic acid (RGD) peptide. Three lines that expressed few or no integrins had very weak ability to adhere to ECM. Strong binding to HUVECs was found in most lines, but the three lines had very little ability to adhere to HUVECs. Binding to HUVECs was strongly inhibited at 4 degrees C, under divalent cation-free conditions and by antibodies to the beta 1 subunit. These results suggest that lung cancer cells adhere to ECM and endothelial cells through integrins, especially the beta 1 subfamily.

Processing of antibody-radioisotope conjugates after binding to the surface of tumor cells

BACKGROUND

Previous experiments indicated that most antibodies binding to cell surface antigens are internalized gradually and degraded within lysosomes, with a half-life of degradation of approximately 1 day, for most antibodies. The research discussed in this article extended our studies to eight additional antibodies reacting with six different antigens, including three antigens anchored in the membrane by glycosyl-phosphatidylinositol. The authors also tested antibodies labeled with 111indium, as well as 125iodine, to determine whether different radiolabels would be processed differently.

METHODS

Antibodies were radiolabeled with 125I or with 111In bound to benzyl-DTPA. After binding to the surface of tumor cells in vitro, excess antibody was washed away, and the fate of the radiolabel was investigated over periods of 3-7 days. Radiolabel released into the supernatant or retained by the cells was analyzed to determine whether it was still on intact antibody.

RESULTS

In 13 of the 15 antibodies that were tested, a similar pattern of irreversible binding and gradual catabolism was observed. Iodine conjugated to antibodies was released rapidly from the cell after antibody catabolism. In contrast, the 111In was retained within the cell much longer than 125I, with the rate of degradation and release into the medium being at least fivefold slower. More than 50% of the bound 111In was still present on the cells after 7 days. Biochemical analysis of the retained 111In extracted cells after 4-6 days demonstrated that it was no longer associated with antibodies and was in a low molecular weight form, probably still associated with the chelator benzyl-DTPA.

CONCLUSIONS

Different radiolabels are processed by tumor cells differently, after catabolism of the antibody to which they originally were conjugated. The data suggest that the prolonged retention of 111In, relative to that of 125I, is due not to deiodination of iodine conjugates, but rather to intracellular retention of catabolic products containing 111In, perhaps within lysosomes. The use of radioisotopes that are retained within cells after antibody internalization and degradation may improve both radioimmunodetection and radioimmunotherapy of cancer.

Cellular and molecular composition of human skin in long-term xenografts on SCID mice

We report on the immunophenotypical characterization of adult human skin transplanted onto severe combined immunodeficient (SCID) mice. Thirty animals were followed for up to 12 months after receiving split-thickness xenografts, of which 28 were tolerated for the whole test period. Antigen mapping revealed an almost complete preservation of human cellular and extracellular tissue components in long-term transplants including skin immune cells (Langerhans-cells, macrophages, lymphocytes) and also parts of the engrafted endothelium. Hence, xenografts on SCID mice offer a versatile experimental tool for the in vivo study of both human skin immune cell function and endothelial cell-mediated interactions in an environment completely devoid of interferences by adoptive host immune response.

VLA-4 molecules on tumor cells initiate an adhesive interaction with VCAM-1 molecules on endothelial cell surface

To elucidate the role of VLA-4 (alpha 4 beta 1 integrin) in tumor metastasis, we have transfected cDNA coding alpha 4 subunit into human fibrosarcoma (HT1080) cells. VLA-4-overexpressing HT-VC1 cells exhibited increased ability to interact with known ligands for VLA-4, such as CS1 peptide and VCAM-1 (vascular cell adhesion molecule-1). In addition, the in vitro invasive ability of HT-VC1 cells was augmented and the mRNA for type IV collagenase was increased in HT-VC1 cells. The induction of VCAM-1 molecules on lung endothelial cells of nude mice by tumor necrosis factor-alpha treatment resulted in augmentation of in vivo HT-VC1 cell adhesion to the lung endothelial cells. Thus, the VLA-4 molecules on tumor cells initiate an adhesive interaction with VCAM-1 molecules on endothelial cells, that is important for hematogenous metastasis.

Human glioma-specific antigens detected by monoclonal antibodies

Three murine monoclonal antibodies, designated GA-17, GB-4, and GC-3, were prepared by the hybridization of murine myeloma cells (NS-1) and spleen cells of BALB/c mice immunized with the crude membrane fraction of cultured human gliosarcoma cells (GI-1). Two of them (GA-17 and GB-4) reacted exclusively with the membrane of glioma cells, and the other (GC-3) reacted with the membrane of glioma cells and a T cell line (MOLT-4). Although these antibodies reacted with almost all of the gliomas, the reactions differed. GA-17 reacted equally well with all glioblastoma (17 cases) and low-grade astrocytoma (10 cases), whereas GB-4 reacted poorly with 7 cases of glioblastoma and GC-3 did not react with 7 cases of low-grade astrocytoma. The antigens, exclusively expressed on the cell surface, were analyzed by surface labeling with 125I followed by a cell lysis and immunoprecipitation with these antibodies. The findings obtained by sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed that GA-17, GB-4, and GC-3 reacted with Mr 140,000-145,000, Mr 160,000, and Mr 145,000-150,000 proteins, respectively. Some evidence has been obtained indicating that these antigens are composed of the same polypeptide chain (Mr 120,000) with the carbohydrate chains being different.

Monoclonal antibody against ependymoma-derived cell line

Mouse myeloma cells were fused with spleen cells from mice that had been immunized with a human ependymoma derived cell line, KMS II. Hybridomas producing monoclonal antibodies (MAbs) were screened and cloned. Specificity of the antibody was determined by enzyme-linked immunosorbent assay (ELISA) and/or indirect immunofluorescence assay. One of the MAbs, designated Ep-C4 (subclass = IgG1), reacted with two cell lines derived from ependymoma but did not react with 17 cell lines derived from other types of brain tumor nor with 4 neuroblastoma cell lines or 19 cell lines derived from carcinoma, hematopoietic tumors and amnion. Indirect immunofluorescence and immuno-electron microscopy studies revealed that the antigen recognized by MAb Ep-C4 was located on cell surface membrane. The membrane antigen of KMS II cells, immunoprecipitated by MAb Ep-C4, was a protein of 81,000 dalton. The reactivity of MAb Ep-C4 was further examined using immunofluorescence and/or immunoperoxidase methods and frozen sections and short-term cultures of various types of brain tumors. No cross-reactivity with normal adult or fetal brain tissues was detected by absorption assay and immunoperoxidase staining. Our results suggest that the antigen defined by MAb Ep-C4 is specific for ependymoma cells, and different from the antigens of glioma cells or other neuroectodermal-derived cells previously described.

TGF beta 1 and TGF beta 2 are potential growth regulators for low-grade and malignant gliomas in vitro: evidence in support of an autocrine hypothesis

Low-grade astrocytomas, anaplastic astrocytomas and glioblastomas in vitro were found to ubiquitously produce the mRNA of transforming growth factor-beta (TGF beta). TGF beta 1 and TGF beta 2 mRNA were expressed to a lesser degree among the hyperdiploid malignant gliomas. By radioreceptor assay of conditioned medium, TGF beta was secreted predominantly in latent form, in both latent and active form, or only in active form within a panel of low-grade and malignant gliomas. The TGF beta receptor (types I, II, and III) was evident among the glioma lines. Many near-diploid gliomas were growth-inhibited by TGF beta 1 and TGF beta 2 in vitro. Most hyperdiploid glioblastomas showed a positive mitogenic response to exogenous TGF beta 1 and TGF beta 2. A synergistic or additive mitogenic interaction with epidermal growth factor and insulin was observed among some. Under serum-free conditions, anti-TGF beta antibody neutralized the expected growth-regulatory effect of endogenous TGF beta, thus establishing the specificity of the response in vitro. TGF beta 1 also enhanced the clonogenicity of certain gliomas which had been growth-stimulated in monolayer. Thus, basic elements in support of an autocrine hypothesis have been demonstrated: TGF beta mRNA was expressed among low-grade and malignant gliomas, TGF beta was secreted in latent and/or active form into conditioned media and appeared to serve as an endogenous regulator of glioma proliferation in vitro. The mitogenic response, either positive or negative, correlated with the degree of anaplasia and karyotypic divergence.

Preparation and characterization of antisera and of murine monoclonal antibodies to human glioma-associated antigen(s)

Human glioma-associated markers can be exploited for the development of new diagnostic strategies and treatment modalities for these malignancies. A goat antiserum was first raised against human anaplastic astrocytoma (AC or AA) and glioblastoma multiforme (GB or GBM) extracts. Extensive sequential absorptions with normal brain tissue, normal serum, and human serum albumin (HSA) gave an antibody fraction specific for glioma. Balb/c mice were subsequently immunized with these glioma extracts. B-cell hybridomas from these mice were then cloned and subcloned by limiting dilution, yielding six monoclonal antibodies (MAbs) that were entirely specific for tumor tissues, and did not react with normal human serum or with normal human brain, liver, kidney, spleen, or muscle. Moreover, the murine MAbs did not cross-react with certain other human tumors, including melanoma. The fully absorbed antiserum and the murine MAbs both identify a polypeptide pattern possibly related to human glial fibrillary acidic protein (GFAP) or other intermediate filament proteins on immunoblots. These immunological reagents could serve as powerful tools for the diagnosis and possibly therapy of these uniformly fatal tumors.

Factor analysis of the immunophenotypes of astrocytomas and malignant gliomas: correlations with tumor grade and patient survival

Our previous investigations correlated the degree of cytogenetic and immunophenotypic heterogeneity of cultured normal glia, astrocytomas and malignant gliomas. The possible significance was suggested by the statistical correlation of individual antigens with diagnosis and patient survival. The present study has established the patterns of covariation of titers of monoclonal antibody reactivity with a panel of cell surface antigens among normal glia (8), astrocytomas (4), anaplastic astrocytomas (12), mixed malignant gliomas (8) and glioblastomas (21). A mean aggregate titer across 43 antigens was computed for each culture and then subtracted from the observed individual titers. Factor analysis was performed to determine a small number of Factors, derived as the weighted average of the 43 mean-adjusted antigens, which accounted for a significant proportion of the covariation of immunophenotypic expression in the sample of 53 cultures. Clusters of antigens were found to independently segregate in their deviation from the aggregate phenotype. Adjusting for age and diagnosis, Factors 1 and 4 correlated with patient survival among recurrent and primary neoplasms, respectively. Factor 2 additionally discriminated between primary and recurrent gliomas. Factor 3 was associated with age at diagnosis. Factors 1 and 2 correlated with the histopathologic grade of glial tumor. Scatter plots of Factor 1 vs. 2 revealed the minimal immunophenotypic diversity of the normal glia. Astrocytomas were similar but not identical. Progressive divergence was evident between the immunophenotypes of anaplastic astrocytomas, mixed gliomas and glioblastomas. These data suggest that qualitative and quantitative differences in antigenic heterogeneity may identify stages in glial tumor progression.(ABSTRACT TRUNCATED AT 250 WORDS)

Expression of 38-kD cell-surface glycoprotein in transformed human keratinocyte cell lines, basal cell carcinomas, and epithelial germs

In this study, we attempted to identify and characterize transformation-induced cell-surface glycoproteins of human keratinocytes. Therefore, we first searched for glycoproteins which are significantly elevated in human keratinocytes after transformation and immortalization by SV40 virus and which are also found at high levels in keratinocytic cell lines derived from squamous cell carcinomas of the skin. Out of at least 80 different cell-surface antigenic systems of human tumor cells, only three glycoproteins showed elevated expression in transformed keratinocytes. Among these, a 38-kD glycoprotein (gp 38) was highly increased in all transformed keratinocyte cell lines tested, but was not elevated in transformed fibroblasts. The expression of gp 38 was further characterized in normal epidermis and in its benign and malignant hyperproliferative disorders: gp 38 was generally not expressed in normal epidermis and in benign hyperproliferative disorders. In contrast, strong and homogeneous reactivity was found in solid and fibrosing basal cell carcinomas whereas no or low reactivity was detected in squamous cell carcinomas and in those parts of BCC revealing keratotic differentiation. Interestingly, high expression of gp 38 was also found in primary epithelial germs of fetal skin, secondary germ cells of the telogenic hair follicle and secretory tubules of sweat glands. The immunohistologic data suggest that gp 38 is preferentially expressed by epidermal cells which lack squamous and pilosebaceous differentiation.

Immunophenotyping of childhood astrocytomas with a library of monoclonal antibodies

Immunophenotype (IP) analysis of 14 childhood glial tumors was performed with a library of 16 monoclonal antibodies (MAbs) using biotin-streptavidin-alkaline phosphatase immunohistochemical detection technique. Presence of glial or neuronal differentiated cells within the tumors was evaluated with MAbs against cell-lineage-specific markers: high-, medium- and low-molecular-weight neurofilament protein (NFP) and glial fibrillary acidic protein (GFAP). Intense expression of GFAP was demonstrated in 14/14 astrocytomas. The three NFs were detected in 10-50% of the cells in 6/14 cases. The pan-neuro-ectodermal antigen defined by MAb UJ 13/A was present in 7/14 astrocytomas on more than 10% of the cells. Thy-1 was expressed in 14/14 tumors on more than 50% of their cells. The GQ ganglioside antigen detected by MAB A2B5, was found in 12/14 tumors. Shared antigens exist among morphologically benign and malignant glial tumor cells and leukocytes detectable with the following four MAbs: Thy-1, PI 153/3, UJ 308 and anti-HLe, common leukocyte antigen (CLA). CLA-expressing cells were demonstrated in 8/12 astrocytomas, and in 4/12 cases more than 90% of the cells were positive. We have shown that cells within childhood astrocytomas can express neuronal IP. The most common expressed phenotype for glial tumors was: GFAP+, Thy-1+, A2B5+, UJ 167.11+, UJ 223.8+, NF (H,M)+, UJ 13/A+, UJ 127.11-, and NF (L)-.

Immunohistochemical analysis of the A4 and AO10 (gp110) cell-surface antigens of human astrocytoma

The A4 and AO10 (110 kd glycoprotein) cell-surface antigens are biochemically distinct markers of cultured human astrocytomas that are expressed by only a limited number of other cultured cell types. To further characterize these two antigens, the authors used immunohistochemical methods to determine their expression in normal human tissues, astrocytomas, and over 100 tumors of other histologic types. They found that A4 is expressed 1) throughout the central (CNS), but not peripheral nervous system (PNS); 2) in smooth muscle and a small number of epithelial tissues; and 3) in reactive glia and in astrocytomas, but not in most tumors of other histologic types. In contrast, the AO10 antigen is expressed 1) in a small subset of CNS neurons, but not in astrocytes, PNS neurons, or other normal tissues; 2) in astrocytomas and reactive glia; and 3) in some additional neuroectodermal tumors, but not melanomas, carcinomas, or sarcomas. These findings show that A4 and AO10 are restricted markers for human astrocytomas in vivo. Furthermore, the antigens show distinct patterns of expression in normal human CNS but appear to be coordinately expressed in astrocytomas and astrocytoma-derived cell lines.

Radioimmunoassay of glioma-associated antigen in cerebrospinal fluid and its usefulness for the diagnosis and monitoring of human glioma

Quantitative determination of human glioma-associated antigen in cerebrospinal fluids (CSFs) obtained from 66 patients with a variety of neurological diseases was performed by solid-phase radioimmunoassay with a monoclonal antibody (G-22). In this system, the minimum detectable amount of the antigen in the CSF was 8 ng/ml. It was demonstrated that CSF diagnosis of glioblastoma might be possible in the case of small tumors with a diameter of less than 2 cm. CSFs obtained from all 18 patients with glioma were positive and the level varied from 11.2 to 186.1 ng/ml. The antigen level in the cystic fluid of the tumor was higher than that in CSF. There was a tendency for the antigen level in CSF to be correlated with the tumor size and the type of histology. The malignant types of glioblastoma or medulloblastoma showed higher levels than the benign type of ependymoma and astrocytoma. Most types of non-gliomatous brain tumor were negative except immature teratoma, meningioma with central neurofibromatosis, and metastatic brain tumor from lung cancer. We also noted that tumor progression or regression of malignant glioma could be predicted by the monitoring of the antigen in the CSF.

Immunobiology of brain tumors

In summary, many actual interactions between tumors in the CNS and the immune system have been demonstrated. The normal brain does not possess a lymphatic system and is partially hidden from the systemic immune system by the BBB, furthermore brain cells do not express MHC antigens which are necessary for the initiation of an immune response. In pathological conditions however, immunocompetent cells may find their way through transformed endothelial cells. Microglia and astrocytes may function as antigen presenting cells. Glioma cells when stimulated by cytokines such as IFN gamma can be induced to express MHC class I and class II antigens, thus making them more susceptible to an immune attack. In addition glioma cells are capable of secreting several cytokines including IL 1, IL 3 and IL 6 also involved in the generation of an immune response. Indeed, a functional analysis of lymphocytes infiltrating gliomas has revealed the accumulation at the tumor site of cytotoxic T lymphocytes as well as NK cells. However host-immune responses against gliomas seem to be weak in comparison to other cancers. Glioma cells are known to secrete TGF beta 2 and PGE 2 which may in part be responsible for this lack of immune response, thus shielding themselves from immune attack. In order to be recognized by the immune system the tumor cells must express TAA in addition to MHC antigens, and such TAA have been identified by MAbs. These MAbs can be used for "targeted" therapy when coupled to toxic agents or radionuclides. Preclinical studies have shown that, after intravenous or intracarotid injection, there is specific accumulation of the MAb in the tumor but in insufficient amounts for therapeutic use. The relatively small amount of MAb binding to the tumor in vivo can be due to several factors: not all the cells in a single tumor express a given tumor-associated antigens, the MAb may have a low affinity for the antigen, the BBB may hinder the passage of the MAb. Attempts have been made to overcome these drawbacks by opening the BBB for example. In addition MAbs can readily be used for the treatment of carcinomatous meningitis. There has been little success in the development of immunotherapy with IFN beta 1 and even less with adoptive immunotherapy using LAK cells plus IL 2. TIL as well as LAK cells can be expanded in vitro with IL2 and it is feasible to reinject these cells into the tumor site.(ABSTRACT TRUNCATED AT 400 WORDS)

Integrin VLA-3: ultrastructural localization at cell-cell contact sites of human cell cultures

The integrin VLA-3 is a cell surface receptor, which binds to fibronectin, laminin, collagen type I and VI (Takada, Y., E. A. Wayner, W. G. Carter, and M. E. Hemler. 1988. J. Cell. Biochem. 37:385-393) and is highly expressed in substrate adherent cultures of almost all human cell types. The ligand specificity of VLA-3 and the inhibition of cell adhesion by anti-VLA-3 monoclonal antibodies suggest its involvement in cell-substrate interaction. In normal tissues, VLA-3 is restricted to few cell types, notably the kidney glomeruli and basal cells of the epidermis. In the epidermis, VLA-3 is generally strongly expressed on the entire plasma membrane of basal cells and is not polarized towards the basement membrane (Klein, C. E., C. Cardon-Cardo, R. Soehnchen, R. J. Cote, H. F. Oettgen, M. Eisinger, and L. J. Old. 1987. J. Invest. Dermatol. 89:500-507). Based on this finding we speculated that, in addition to a role of VLA-3 for adhesion of cells to substrate, it could also be relevant for cell-cell interaction. To investigate this, we ultrastructurally localized VLA-3 on the surface of cultured cells by immunoelectron microscopy. In accordance with our concept, we found VLA-3 strongly associated with intercellular contact sites. Interestingly, very little immunoreactivity was detected at the under-surface of cells which had been cultured for 18-32 h. This observation was unexpected but is consistent with previous findings (Kantor, R. R. S., M. J. Mattes, K. D. Lloyd, L. J. Old, and A. P. Albino. 1987. J. Biol. Chem. 262:15158-15165) which suggest that the association of VLA-3 with the basal surface of substrate adherent tumor cells is a late event occurring after days of culture under confluent conditions. However, we cannot formally rule out VLA-3 expression at the undersurface of cells under our experimental conditions, since VLA-3 molecules at this location could be inaccessible for in situ labeling of unfixed cells because of spatial interferences. In conclusion, our results demonstrate the expression of VLA-3 at intercellular contact sites of cultured cells supporting the concept that it may be relevant for intercellular interactions also.

Heterogeneity and modulation of tumor-associated antigens in human glioblastoma cell lines

Seven human glioblastoma cell lines established in vitro from primary tumor explants were studied. A marked heterogeneity of glial fibrillary acidic protein was observed whereas vimentin was uniformly expressed by all cell lines. Indirect immunofluorescence and flow cytofluorometry revealed a heterogeneous distribution of surface GE 2 and CG 12 tumor-associated antigens (TAA's): three cell lines were positive (greater than 69% TAA-positive cells) and three cell lines were negative (less than 9% TAA-positive cells). One cell line (Hu 228) was moderately positive at early culture passages and subsequently acquired a TAA-negative phenotype. The difference in the relative amounts of surface TAA's of the three positive cell lines was less than twofold. In spite of the heterogeneous distribution of surface TAA's, all cell lines exhibited considerable amounts of intracellular TAA. Treatment with phorbol esters and density-dependent growth arrest decreased the percentage of the TAA-positive cells and the amount of cell-surface TAA's in one cell line (Hu 195). Interferon-gamma treatment in vitro increased the percentage of CG 12-positive cells by 12% and the amount of cell-surface CG 12 antigens by 38% as compared to untreated cells. The percentage of TAA-positive cells among phorbol ester-treated cells of the Hu 195 cell line was lowest 48 hours after treatment, but returned to normal values within the next 48 hours. Reduction of 3H-thymidine incorporation preceded the decrease in number of TAA-positive cells by about 18 hours. Two-color fluorescence analysis performed in positive cell lines for simultaneous determination of surface TAA's and deoxyribonucleic acid content or reactivity with the proliferation-associated Ki67 intracellular marker indicated that GE 2 and CG 12 antigens are expressed preferentially by actively proliferating glioma cells. The results of this study indicate the existence of two different phenotypes in cultured human glioblastoma cells: surface TAA-positive/cytosol TAA-positive and surface TAA-negative/cytosol TAA-positive cell populations. In addition, modulation of TAA expression was dependent on the cell-cycle differentiation stage, culture conditions, and proliferative state of the cells.

Induction of pigmentation in mouse fibroblasts by expression of human tyrosinase cDNA

A distinguishing characteristic of cells of the melanocyte lineage is the expression of the melanosomal enzyme tyrosinase that catalyzes the synthesis of the pigment melanin. A tyrosinase cDNA clone, designated BBTY-1, was isolated from a library constructed from the pigmented TA99+/CF21+ melanoma cell line SK-MEL-19. Expression of BBTY-1 in mouse L929 fibroblasts led to synthesis and expression of active tyrosinase, and, unexpectedly, to stable production of melanin. Melanin was synthesized and stored within membrane-bound vesicles in the cytoplasm of transfected fibroblasts. BBTY-1 detected a 2.4-kb mRNA transcript in nine of nine pigmented, tyrosinase-positive melanoma cell lines. Tyrosinase transcripts of the same size and abundance were detected in a subset (three of eight) of nonpigmented, tyrosinase-negative melanoma cell lines, suggesting that post-transcriptional events are important in regulating tyrosinase activity. Two melanocyte antigens, recognized by mAbs TA99 and CF21, that are specifically located within melanosomes and are coexpressed with tyrosinase activity, did not react with transfected mouse fibroblasts expressing human tyrosinase, supporting the conclusion that these antigenic determinants are distinct from the tyrosinase molecule coded for by BBTY-1.

Immunology of gliomas

The treatment of cerebral gliomas continues to challenge neurosurgeons and research scientists. The lack of major success with chemotherapy and radiation therapy has given rise to further investigation into the biology of these tumors and host reactions to them. Much of this research has centered upon the evaluation of tumor cell antigenicity and on both the humoral and cellular immune responses to gliomas. Contrary to previous considerations, evidence suggests that astrocytes, glioma cells, and tumor endothelial cells may all have pivotal roles in the initiation and prolongation of both local and systemic immune responses to the tumor. In this review we will discuss the immunobiology of the glioma with specific reference to the interactions between the tumor and the host immune system. In addition, ideas for potential therapeutic manipulation of the host-glioma immune interactions will be reviewed, stressing potential pitfalls and risks.

Cell surface molecules of human melanoma. Immunohistochemical analysis of the gp57, GD3, and mel-CSPG antigenic systems

The rapidly expanding list of monoclonal antibodies (MAbs) to human cell surface antigens provides reagents to probe the biology of malignant melanoma and to develop new diagnostic and therapeutic approaches to this disease. The criteria used to select MAb-defined antigens as targets for passive immunotherapy or immunolocalization of melanoma include: 1) consistent antigen expression in melanomas, 2) restricted antigen distribution in normal tissues and nonmelanocytic tumors, and 3) cytotoxic activity of the MAb or MAb conjugates. The present study examined the tissue distribution of three prototype melanoma cell surface antigens, the Mr 57,000 glycoprotein (gp57) recognized by MAb A42, the GD3 ganglioside, and the mel-CSPG chondroitin sulfate proteoglycan. The avidin-biotin immunoperoxidase method was used to examine a large panel of normal tissues and over 150 malignant tumors. It was found that A42 has a highly restricted distribution in normal tissues and is expressed in subsets of melanomas and nonmelanocytic tumors. It was also found that GD3 and mel-CSPG are more widely distributed in normal tissues and among tumors than was thought previously. These immunohistochemical patterns provide an essential data base to evaluate the ongoing clinical trials employing MAbs to GD3 and mel-CSPG for the therapy and immunolocalization of melanomas, and they identify gp57 as a potential marker for subsets of normal and transformed melanocytic cells.

A transformation-associated 130-kD cell surface glycoprotein is growth controlled in normal human cells

Two characteristics of cell surface molecules involved in the regulation of cell proliferation are altered expression in relation to growth phase in normal cells and overexpression in transformed cells. Here, we describe a similar pattern of expression for a 130-kD cell surface glycoprotein (gp 130) in human cells. Synthesis and cell surface expression of gp130 were greatly increased in both virally and chemically transformed fibroblasts, fibrosarcomas, a squamous cell carcinoma of the skin, and T cell leukemia lines. Furthermore, gp130 expression was induced in serum-starved fetal fibroblasts by serum stimulation, and in fresh T cells by various activating agents. Expression in response to serum stimulation was associated primarily with the transition from a quiescent state (G0) into the cell cycle (G1).

Immunological and biochemical strategies for the identification of brain tumor-associated antigens

Various strategies have been used to identify and characterize the antigens associated with human brain tumors. These approaches have included the raising of polyclonal and monoclonal antibodies against tumor antigens and, more recently, efforts toward the direct biochemical identification of such proteins. This review summarizes the progress made in this area, suggests reasons for the broad antigenic cross-reactivity and heterogeneity revealed by these studies, and proposes additional methods for deciphering the complex antigenic composition of human brain tumors.

Human glioma cell lines: tumour associated antigens distribution and sensitivity to antibody-toxin or ligand-toxin conjugates. A preliminary report

We have investigated the phenotype of seven human glioma cell lines established in vitro from primary tumour explants. Indirect immunofluorescence and flow cytofluorimetry revealed a heterogeneous distribution of surface GE 2 and CG 12 Tumour Associated Antigens (TAA). In one group of cell lines TAA were detected both at the cell surface and in the cytosol, whereas in a second group of glioma cell lines TAA were found only in the cytosol. We have also investigated the sensitivity of glioma-derived cell lines to antibody-toxin and ligand-toxin conjugates (Immunotoxins). Monoclonal antibodies anti GE 2 antigen linked to ricin toxin A subunit (RTA) showed poor cytotoxicity, which increased about 50 fold when the whole toxin was linked to anti GE 2 monoclonals. Treatment with human recombinant interferon gamma (IFN-gamma) greatly augmented the percentage of HLA-DR+ cells and the amount of HLA-DR antigens per cell. IFN-gamma treatment resulted in a net increase of sensitivity to anti HLA-DR Immunotoxins (IT). Human diferric transferrin linked to RTA exhibited a potent cytotoxic effect against human glioma-derived cells when used in the presence of the lysosomotropic carboxylic ionophore monensin.

Changes in cell surface glycoprotein expression during differentiation of human keratinocytes

Six cell surface glycoproteins defined by monoclonal antibodies were selected for study on human epidermal cells. In tests on tissue sections, three of the glycoproteins [J143 (gp140/30); T43 (gp85/36); H99 (gp38)] were expressed in the basal cell layer of the epidermis, whereas the other three glycoproteins [T179 (gp140/95); T16 (gp40/50); BT15 (gp80)] were preferentially expressed in maturing keratinocytes above the basal layer. We compared synthesis of these glycoproteins in fresh epidermis and in primary epidermal short term cultures using [35S]methionine for metabolic labeling. Synthesis of J143 was 8- to 20-fold higher and synthesis of T43 was 4- to 10-fold lower in cultured cells compared with fresh epidermis. BT15, an antigen strongly expressed on terminally differentiating keratinocytes, was synthesized at 5- to 15-fold higher levels in fresh epidermis than in cultured cells. Biosynthesis levels of H99, T179, and T16 did not change in cultured epidermal cells. Based on our findings, we propose a model of surface antigenic changes that occur during keratinocyte differentiation in vivo.

Monoclonal antibodies in urologic oncology

The sensitivity and specificity of immune reactions have long been recognized. However, since the description of the monoclonal antibody technique certain technical hurdles have been overcome. Monoclonal antibody (mAb) technology allows far more precise understanding of the humoral immune response by allowing dissection of this response into its individual B-lymphocyte populations. Furthermore, the ability to select and expand a particular B-cell clone allows for production of unlimited amounts of a pure antibody "reagent." Each of these reagents may be readily used as a "probe" for its respective antigenic determinant. Panels of these reagents may be used to probe complex biologic structures (e.g., neoplastic cells) and thereby "dissect" them at a molecular level. Murine mAbs are produced by hyperimmunizing a mouse with the antigen of interest. The spleen provides a rich source of B-lymphocytes. These normal B-lymphocytes are incapable of surviving in culture. However, mouse B-lymphocyte tumor cell lines (myelomas) have been previously immortalized in tissue culture. A hybridoma is formed by fusing the normal immunized B-cells with myeloma cells. The hybridoma combines the best features of its parent cells. The immunoglobulin product of each resulting clone is screened against a panel of antigens. This allows selection of those clones producing antibody to the desired antigens. The mAb may be used to purify and/or characterize its respective antigen including definition of the chromosomal site of its gene and factors involved in the regulation of its expression. A large number of mAbs have been produced against renal, bladder, and prostate cancer antigens. These mAbs are allowing a more precise, molecular subclassification of these cancers and providing improved predictability of the natural history of each patient's cancer. Evidence in experimental animals demonstrates that mAbs accumulate at the site of tumor and can lead to the destruction of tumor cells. These encouraging results have prompted clinical trials in patients with metastatic cancer. Such a clinical trial in patients with metastatic renal cancer is currently underway at New York Hospital and Memorial Sloan-Kettering Cancer Center.

Phenotypic heterogeneity of melanoma. Relation to the differentiation program of melanoma cells

Phenotypic heterogeneity is a characteristic feature of tumor lesions in patients with melanoma. Variability can be observed in cell morphology, pigmentation, and antigen expression. To test whether phenotypic heterogeneity could be the result of events regulated during cell differentiation, we evaluated the expression of a panel of differentiation traits on melanoma cells. Metastatic melanoma lesions from two patients, designated FD and AP, were examined histologically and found to contain mixed populations of cells. Established melanoma cell lines derived from each of these lesions were subcloned at early passage in culture (passages 7 and 8) to create a panel of clones derived from each tumor. There was heterogeneity in the expression of differentiation-related traits in clones, corresponding to distinct phenotypes observed within the original tumors. Clones from patient FD corresponded to early to intermediate stages of melanocyte differentiation, and clones from patient AP ranged from intermediate to late stages. The influence of cholera toxin and PMA on differentiation of parental cultures and subclone was studied. Results of induction studies demonstrated a number of features of differentiation of melanoma cells: regulation of differentiation traits is coordinated as a program of traits expressed sequentially at specific stages; early traits, such as the epidermal growth factor receptor and the melanoma chondroitin sulfate proteoglycan antigen, are downregulated as melanoma cells differentiate, whereas late markers, including melanin, tyrosinase activity, and antigens expressed in mature melanosomes, are upregulated; Ia (class II major histocompatibility) antigens are characteristically expressed on melanomas corresponding to early or intermediate stages of differentiation and are regulated as part of the differentiation program; minimal changes in stage of differentiation were observed during induction of parental cultures with either cholera toxin or PMA, whereas definite shifts in differentiation could be induced in selected cloned subpopulations. We conclude that melanoma cells are not frozen at a specific stage of differentiation, but rather are capable of differentiating when exposed to appropriate signals. Diversity in the differentiation state of melanoma cells can account for much of the phenotypic heterogeneity observed in melanoma lesions.

Characterization of a kidney antigen defined by a mouse monoclonal antibody K2.7

Analysis by indirect immunoperoxidase staining of tissues demonstrated that monoclonal antibody (mAb) K2.7 derived from a mouse immunized with a renal cancer cell line OS-RC-2 was reactive with the tissues of all 36 normal kidneys examined, including two fetal kidneys, 22 of 25 renal cell carcinomas (RCCs) and one of 15 gastric cancers, but not with 23 normal tissues from 11 different organs or 41 malignant tissues from 10 different organs. mAb K2.7 was reactive with the kidneys of several animal species. Positive staining was seen on tubular epithelial cells of the kidney. Thus, the antigen recognized by mAb K2.7 appeared to be an interspecies kidney specific antigen. Results obtained by protein A (PA) assays with a cultured cell panel were consistent with those of staining and confirmed the restricted expression of the antigen. Immunochemical analysis revealed that mAb K2.7 reacted with molecules of 63,000, 60,000 and 43,000 daltons.

Establishment and characterization of five cell lines derived from human malignant gliomas

We established and characterized five cell lines derived from human malignant gliomas (four glioblastomas multiforme and one highly anaplastic astrocytoma). All cell lines exhibited tumor cell morphology and growth kinetics, and anchorage-independent growth in soft agar. Cytogenetic analysis revealed significant aneuploidy in all five cases as well as clonal chromosomal alterations unique to each cell line. No cell line was tumorigenic in athymic mice. Two of the cell lines were sensitive to carmustine (BCNU) in monolayer and soft-agar cultures. Electron microscopy showed marked variability between cell lines in the number and structure of intracytoplasmic organelles; SF-126 formed collagen fibers in vitro. Immunohistochemical analysis of the surgical specimens showed variable expression of glial fibrillary acidic protein (GFAP) in malignant astrocytes; positive immunostaining for glycoproteins of the extracellular matrix was found predominantly in perivascular regions. In early-passage cultures, only cell line SF-295 expressed GFAP; at establishment, none of the cell lines expressed GFAF or glutamine synthetase. Fibronectin and laminin were expressed by all cell lines in early-passage culture, but expression of these glycoproteins at establishment was variable. Only SF-126 was positively identified by immunostains for procollagen III; this was also the only cell line in which DEAE-cellulose chromatography and SDS-PAGE demonstrated interstitial collagen synthesis. These well-characterized glioma-derived cell lines may now serve as useful tools with which to study the cell biology of gliomas. The synthesis of interstitial collagen by a glioma-derived cell line may suggest a derivation from vascular mesenchymal elements, either reactive or transformed, in the original heterogeneous malignant glioma, rather than from a glial precursor cell.

Monoclonal antibodies against antigens expressed on human hepatocellular carcinoma cells

Monoclonal antibodies with selectivity for human hepatoma cell lines were produced by immunizing BALB/c mice with human hepatoma cell lines, HA22T/VGH or Hep 3B, and fusing sensitized mouse spleen cells with mouse myeloma cells. Two monoclonal antibodies recognizing antigens present only on human hepatoma cell lines were investigated. The monoclonal antibody IB1 was found to react with 3 of 9 hepatoma cell lines. Monoclonal antibody 9B2 reacted with all nine hepatoma cell lines. None of the other 20 cell lines tested was bound by IB1 and 9B2. The immunoperoxidase staining of monoclonal antibodies on frozen sections of paired hepatoma and normal liver tissues from the same individuals were studied. Antibody IB1 reacted with 3 of 13 hepatoma tissues, but with none of the normal liver and other tissues, and antibody 9B2 was reactive with antigens appearing on the bile canalicular domain of hepatoma and normal liver tissues. The antibody 9B2 stained no normal tissues with the exception of proximal tubules of kidney. Radioimmunoprecipitation tests identified two antigens reacting with 9B2. The major antigen had an apparent molecular weight of 140,000 and a minor one of 130,000. Therefore, antibody IB1 seems to be specific for antigens present on a group of human hepatoma cells and may be useful for classification and diagnosis of human hepatomas. Antibody 9B2 is quite specific to human liver cells and may be used to provide clues for the characterization of tumor cell lines, identification of metastatic tumors with hepatocytic origin, and study of the structure and function of bile canaliculi.

Extracellular matrix-modulated expression of human cell surface glycoproteins A42 and J143. Intrinsic and extrinsic signals determine antigenic phenotype

We have used serologic, biochemical, and genetic methods to characterize two stage-specific human differentiation antigens of neural and melanocytic cells: A42 (57,000 Mr glycoprotein) and J143 (140,000/30,000 Mr glycoprotein). The genes determining A42 and J143 cell surface expression in rodent-human hybrids were chromosomally mapped, and the respective human chromosomes were introduced into rodent cells derived from distinct differentiation lineages. Serologic analysis of the resulting hybrid clones has permitted the identification of two types of regulatory signals determining A42 and J143 expression. First, both antigens are expressed in hybrids constructed with antigen-positive human cells and also in certain hybrids constructed with antigen-negative human cells, indicating that intrinsic signals provided by the differentiation program of the rodent fusion partner induce antigen expression. Second, a series of human-mouse neuroblastoma hybrids, which are A42- or J143- when cultured on plastic surfaces, can be induced to express the antigens when cultured on substrates coated with extracellular matrix (ECM) produced by bovine corneal endothelial cells or fibronectin. This induction of antigen expression by extrinsic, ECM-derived signals is accompanied in the neuroblastoma hybrids by increased substrate adhesiveness and cell spreading and by characteristic changes in cell morphology. A similar program of phenotypic changes is also seen in spontaneous variants of human neuroblastoma and Ewing's sarcoma cells and in ECM-induced Ewing's sarcoma cells. These findings suggest that ECM-derived signals have a role analogous to mitogens and soluble differentiation factors in modulating differentiation phenotypes and tissue-specific patterns of cell surface antigen expression.

Detection of human glioma-associated antigen by rat monoclonal antibody raised against syngeneic rat glioma cells

A monoclonal antibody termed "FR77" was obtained from a hybridoma clone established by fusion between P3x63Ag8.653 mouse myeloma cells and spleen cells of a Fischer F344 rat hyperimmune to syngeneic 9L/R3 glioma cells. Immunoperoxidase staining of various cultured cells showed that FR77 was reactive to both rat and human glioma cells, but was not reactive with other nonglioma cells. Immunohistochemical examination of paraffin-embedded or cryostat-frozen sections of various human tissues revealed that FR77 was strongly reactive with glioblastoma, grade III astrocytoma, and craniopharyngioma; partially reactive with intracerebral primitive neuroectodermal tumor, pineoblastoma, and desmoplastic medulloblastoma; and weakly reactive with low-grade astrocytoma. It was not reactive with other types of brain tumors and normal human tissues tested. The FR77-defined antigen was observed to be predominantly localized in the cytoplasm of antigen-bearing cells as suggested by the immunostaining pattern, but part of it was also expressed on the cell surface of glioma cells as demonstrated by a complement-mediated cytotoxic test. Fractionation of the antigenic component and periodic acid treatment of tumor tissue bearing the FR77-defined antigen indicated that the antigen is of a neutral glycolipid nature and that the antigenic determinant to FR77 is present on its sugar portion.

Human melanoma proteoglycan: expression in hybrids controlled by intrinsic and extrinsic signals

Human malignant melanoma cells express specific chondroitin sulfate proteoglycans (mel-CSPG) on the surface, both in vivo and in vitro. Melanocytes in normal skin show no detectable mel-CSPG but can be induced to express the antigen when cultured in the presence of cholera toxin and the tumor promoter 12-O-tetradecanoylphorbol-13-acetate. Most other cell types do not express mel-CSPG either in vivo or in vitro. A study was designed to examine regulatory signals controlling mel-CSPG expression. The gene encoding mel-CSPG was mapped to human chromosome 15, and this chromosome was introduced into rodent cells derived from distinct differentiation lineages. Three types of mel-CSPG--expressing hybrids were found: (i) hybrids derived from human melanomas; (ii) hybrids derived from human cells that do not express mel-CSPG; and (iii) hybrids derived from human cells expressing mel-CSPG that are antigen-negative but that are induced to express mel-CSPG when cultured on extracellular matrix instead of plastic surfaces. Thus, mel-CSPG expression can be controlled both through intrinsic signals, provided by the differentiation program of the rodent fusion partner, and through extrinsic signals, provided by specific cell-matrix interactions.

Immunobiology of brain tumors

The treatment of the cerebral glioma continues to challenge neurosurgeons and basic scientists. The lack of major success with chemotherapy and radiation therapy has spurred further investigation into the biology of this tumor and host reactions to it. Much of this research has centred upon evaluation of tumor cell antigenicity and the cellular immune responses to the glioma. Contrary to previous considerations evidence suggests that astrocytes, glioma cells and tumor endothelial cells may all have pivotal roles in the initiation and prolongation of both local and systemic immune responses to the tumor. In this review we will discuss the immunobiology of the glioma with specific reference to the interactions between the tumor and the host immune system. In addition, ideas for potential therapeutic manipulation of the host-glioma immune interactions will be reviewed stressing potential pitfalls and risks.

Cell-surface antigens determined by human chromosomes 1 and 12: comparative serological and somatic cell genetic analysis of eight antigenic systems

We have previously assigned several genes controlling expression of cell-surface antigens to human chromosomes 1 and 12. In the present study, we characterize three additional cell-surface antigens determined by these chromosomes. Two monoclonal antibodies, AbSR75 and AbMG6, define antigens expressed on a wide range of cultured human cells. In contrast, AbK66 defines an antigen with a more restricted distribution which is expressed on normal and malignant human epithelial cells but not on neuroectoderm-derived cells or hematopoietic cells. Normal adult fibroblasts are K66-, whereas fetal fibroblasts are K66+. Serological analysis of rodent-human hybrid cells permitted the assignment of MSK31, controlling SR75, and MSK32, controlling K66, to chromosome 1, whereas MSK27, controlling MG6, maps to chromosome 12. Analysis of hybrids containing only deleted copies of chromosome 1 or chromosome 12 established regional assignments for MSK31, MSK32, and MSK27 and also for the previously defined genes MSK1, MSK4, and MSK7. The MSK1 gene, coding for a 140,000 mol wt cell surface glycoprotein, was found to map to the same band (1p22) as the NRAS protooncogene and the gene encoding the beta-subunit of nerve growth factor (NGFB).

Normal and malignant neural cells: a comprehensive survey of human and murine nervous system markers

Tumor-associated neural markers are finding increased application in diagnostic histopathology and in the development of brain tumor therapy. The major cell-type-specific markers and monoclonal antibodies that identify murine and human neural cells are reviewed in this study. Monoclonal antibodies, raised against fetal and adult neural tissue, neuroectodermal tumor tissue, or cell line immunogens which recognize epitopes on brain tumors are comprehensively described including antigens common to the nervous, hematopoietic, and immune systems. The clinical application of neural cell markers and monoclonal antibodies for the diagnosis, localization, and treatment of neuroectodermal tumors is reviewed.

Comparison of tissue culture cells and histological sections for use in screening monoclonal antibodies

Monoclonal antibodies have been screened using either tissue culture cells or histologic sections as a source of antigen. We evaluated and compared these two methods using six murine monoclonal antibodies on human kidney cultures and histologic sections of human normal kidney and renal carcinoma. Antibody binding demonstrated shared antigens on kidney culture cells, renal tubular cells and renal carcinoma cells. The patterns of antibody binding were generally similar with either cultured cells or histologic sections. However, certain antibodies demonstrated significant differences in antibody binding depending on the method used for screening.

Growth regulation of human melanocytes: mitogenic factors in extracts of melanoma, astrocytoma, and fibroblast cell lines

Melanocytes derived from fetal or adult skin do not propagate in vitro unless cultured in the presence of factors such as 12-O-tetradecanoylphorbol 13-acetate (TPA). In a search for physiological factors regulating the growth of melanocytes, extracts of various cultured cell types were tested. Factors produced by melanoma and astrocytoma cell lines support continued proliferation of melanocytes in the absence of TPA. WI-38, a fibroblast cell line derived from human embryonic lung, was the most active source of melanocyte growth factors. No melanocyte growth-promoting activity was found in extracts of cultured neuroblastoma, renal cancer, normal keratinocytes, or renal epithelium. Nerve growth factor, epidermal growth factor, melanocyte-stimulating hormone, transforming growth factor-beta, and platelet-derived growth factor did not have growth-promoting activity for melanocytes. The presence of melanocyte growth factors and TPA together resulted in the strongest mitogenic activity for melanocytes, permitting the recovery (at 20 days) of 4 to 20 times as many cells as in growth factor or TPA alone.

Pigmentation-associated glycoprotein of human melanomas and melanocytes: definition with a mouse monoclonal antibody

Pigmented melanoma cells and cultured melanocytes express a differentiation-related glycoprotein designated as pigmentation-associated antigen (PAA) of Mr 70,000-80,000. As described previously, PAA was initially defined by reactivity with antibodies in the serum of a patient with melanoma. Here we describe the production and characterization of a mouse monoclonal antibody to PAA. This antibody (TA99, an IgG2a) was shown by sequential immunoprecipitation experiments to react with the same component as the human antibody. Ab TA99 immunoprecipitated PAA from lysates of cells radiolabeled with [35S]methionine, [3H]glucosamine, [3H]fucose, and [3H]mannose as well as 125I. Using Ab TA99, the distribution of PAA was examined in frozen sections of 19 normal tissues and quantitatively in 68 tissue culture cell lines. In frozen sections, only melanin-containing cells were positive, including epithelial cells in the basal layer of the epidermis, in which pigment originates from melanocytes by transfer of melanosomes, and pigmented cells of the eye. In tissue culture cell lines, only pigmented melanoma cells were positive. PAA is an intracellular antigen, with a distribution very similar to that of melanosomes. This evidence confirms the close association of PAA with melanin production, and suggests that PAA may be a melanosome component. PAA was shown to be different from tyrosinase, the enzyme involved in melanin synthesis, but it was found to be identical to the previously recognized glycoprotein, gp75, characteristic of pigmented melanomas and melanocytes.

Cell surface antigens of human trophoblast and choriocarcinoma defined by monoclonal antibodies

Six distinct cell surface antigens of human trophoblast and choriocarcinoma were defined with MAbs. The distribution of the antigens was determined by MHA assays on 150 tumor cell lines and normal cell cultures and by immunofluorescence tests with a wide range of normal adult and fetal tissues and a tumor panel. Antigen LK26 is expressed on all cultured choriocarcinoma, teratocarcinoma and renal cancer lines but is absent from most cell lines derived from other tumor types and from cultures of normal kidney epithelium and fibroblasts. LK26 expression in normal tissues is restricted to the trophoblast. No other adult or fetal tissue was found to express the antigen, but choriocarcinoma and teratocarcinoma tissues were LK26+. SV19 is expressed on cultured choriocarcinomas and teratocarcinomas and on subsets of breast and colon cancer lines, but not on 120 additional cultures tested. In tissues, SV19 is detected in normal placenta, mammary gland and colon epithelium as well as in tumors of breast, colon and lung. Two antibodies, AbSV63 and AbK8, react with PLAP and AbSV63 also reacts with the intestinal form of the enzyme. AbLK24 defines a heat-stable determinant present on choriocarcinoma and breast cancer cell lines but absent from most other cultured cells. It is expressed on a small range of normal and malignant epithelial tissues, including normal trophoblast, normal breast epithelium and urothelium and tumors derived from these tissues. One antigen, K66, showed a wide distribution on cultured epithelial cells but was not found in any normal or malignant tissue. Finally, S4, a previously described marker of normal and malignant kidney epithelial cells, was also expressed on the choriocarcinoma cell lines. Four of the antigens are glycoproteins that could be immunoprecipitated from radiolabelled extracts of choriocarcinoma cells: LK26 (Mr 35,000), SV19 (Mr 40,000), PLAP (Mr 68,000) and S4 (Mr 160,000). The highly restricted distribution of LK26, SV19, S4, and PLAP in normal tissues and their expression in tumors make these antigens potential diagnostic markers of gestational choriocarcinoma and germ-cell tumors and, possibly, targets for immunotherapy.