Establishment, molecular rescue, and expression of 123AV16-1, a tumor-reactive human monoclonal antibody.

The human monoclonal antibody (mAb) 123AV16-1 was generated by Epstein-Barr virus transformation of peripheral blood lymphocytes from a colorectal patient undergoing active specific immunotherapy with an autologous tumor cell-Bacille Calmette-Guérin vaccine. Direct immunohistochemical staining of tumor and normal pairs of tissues indicated that this human IgA1, lambda 2 mAb preferentially reacted with colon tumor epithelium. To generate a recombinant derivative of this Epstein-Barr virus-transformed cell line, we isolated the expressed complete heavy and light chain genes by a novel strategy and cloned them into modified pSV2-neo and pSV2-gpt expression vectors. The recombinant 123AV16-1 human mAb was expressed in both a murine myeloma and a human-murine heteromyeloma and was secreted as both monomers and dimers. The recombinant 123AV16-1 mAb expressed by both cell lines reacted with human colon tumor xenografts in a manner similar to the mAb derived from the Epstein-Barr virus-transformed cell line, indicating that the antibody specificity was not appreciably altered during the molecular rescue, cloning, or expression.

New chelating agent for attaching indium-111 to monoclonal antibodies: in vitro and in vivo evaluation.

111In possesses excellent radiophysical properties suitable for use in immunoscintigraphy of cancerous tissues when attached to an antitumor antibody. However, 111In has a tendency to accumulate in normal tissues such as liver. Instability of the linkage between 111In and antibody may contribute to this problem. To avoid this, we developed a new bifunctional chelating agent, 1,3-bis[N-[N-(2-aminoethyl)-2-aminoethyl]-2-aminoacetamido]-2-(4- isothiocyanatobenzyl)propane-N,N,N',N'',N''',N'''',N''''',N'''''- octaacetic acid (LiLo), that forms a kinetically stable chelate with metal ions such as indium. Using LiLo, indium-111 was conjugated to a human monoclonal antibody, 16.88. Competitive binding analysis revealed that the 16.88-LiLo conjugate is as immunoreactive as the unconjugated native antibody. This conjugate was compared with 111In-16.88, where diethylenetriaminepentaacetic acid dianhydride (DTPAa) was used as the chelating agent. In vitro stability studies showed that 111In was more stably bound to 16.88-LiLo than to 16.88-DTPA. Biodistribution studies in athymic mice bearing colorectal tumor xenografts indicated less liver retention with 16.88-LiLo than with 16.88-DTPA. These results demonstrate that LiLo is superior to DTPAa for attachment of 111In to the monoclonal antibodies.

Quantitation of human tumor-reactive monoclonal antibody 16.88 in the circulation and localization of 16.88 in colorectal metastatic tumor tissue using murine antiidiotypic antibodies.

Detection of administered human monoclonal antibodies in the tissues and circulation of patients requires special reagents to overcome interference by normal endogenous immunoglobulin. A practical approach is the development of antiidiotypic antibodies to the human monoclonal antibody and their application in immunoassays specific for the human monoclonal antibody. Accordingly, antiidiotypic antibodies were made to the monoclonal antibody 16.88, a human IgM class anti-colon carcinoma antibody being developed for applications in antibody-targeted immunotherapy of cancer. Three stable clones were obtained that produced antiidiotypic antibodies reactive with 16.88 but nonreactive with human polyclonal IgM or 16.52, a patient-matched IgM monoclonal antibody with different specificity than 16.88. One antiidiotypic antibody, MID 65, was used in a capture format radioimmunoassay to detect 16.88 in the sera of patients who had received 108-mg doses of unlabeled 16.88 coadministered with trace doses of 131I-16.88. Using this assay it was demonstrated that unlabeled 16.88 antibody and 131I-labeled 16.88 antibody did not differ significantly in blood retention for up to 24 h after administration, the period during which the immunoreactivity of the administered antibody remained over 90%. Indirect microautoradiography using exogenously applied 125I-MID 65 to localize 16.88 in frozen metastatic tumor tissue from patients given 16.88 8 days prior to surgery demonstrated the accumulation of 16.88 in areas of apparently healthy tumor cells. Much less 16.88 was detected in stroma or areas of tumor cell necrosis. The accumulation of antibody in nonnecrotic tumor sites encourages the further development of 16.88 for radioimmunotherapy of colon cancer and provides support for further development of human anticytokeratin monoclonal antibodies for cancer therapy.

Preclinical studies on the pharmacokinetic properties of human monoclonal antibodies to colorectal cancer and their use for detection of tumors.

We studied the pharmacokinetic properties of two human monoclonal antibodies to colon carcinoma cells and their ability to detect tumors in nude mice bearing primary human colon carcinoma xenografts. The 16-88 and 28A32 monoclonal antibodies are immunoglobulin M class human antibodies produced by cell lines derived from peripheral blood lymphocytes from patients with colon carcinoma. The patients received an autologous tumor cell vaccine as part of an active specific immunotherapy protocol. The 125I-labeled antibodies were cleared from the circulation of non-tumor-bearing and tumor-bearing nude mice with a 6-8-h half-life. The half-life of the antibodies in tumor tissue was 48 to 72 h compared to 8 to 12 h for normal tissues. Tumor:normal tissue ratios were highest 4 to 7 days postinjection with tumor:blood ratios of 12:1 for 16-88 and 10:1 for 28A32 antibody. Experiments with a control human immunoglobulin M myeloma protein confirmed the specificity of the human monoclonal antibodies. Radioimmunoscintigraphic studies using nude mice bearing contralateral antibody-reactive and nonreactive colon tumor xenografts further confirmed that the antibodies specifically localized in tumor tissues. The antibody-reactive tumors were clearly visible by radioimmunoscintigraphy within 4 days of injection. These experiments, undertaken as a preliminary step to clinical trials, demonstrated for the first time that i.v. administered human immunoglobulin M monoclonal antibodies could be taken up by human colon tumor tissue and retained to a sufficient extent to easily permit tumor detection by external radioimmunoscintigraphy. These studies also demonstrated that the nude mouse human colon tumor xenograft model is a useful in vivo system for comparison studies of human monoclonal antibodies as part of a selection process for clinical trials and for evaluating immunoconjugates containing these antibodies for relative pharmacokinetic properties and potential diagnostic or therapeutic efficacy.

Natural cytotoxicity against mouse hepatitis virus-infected target cells. I. Correlation of cytotoxicity with virus binding to leukocytes.

Spleen cells from uninfected control mice selectively lysed BALB/c 3T3 fibroblasts infected with mouse hepatitis virus (MHV), a murine coronavirus. Lysis of infected cells occurred within 3 hr, and histocompatibility between effector and target cells was not required. This natural, cell-mediated, virus-associated cytotoxicity differed from NK cell- and T cell-mediated lysis. Spleen cells from animals infected with MHV were enriched in NK activity and were more cytotoxic to YAC-1 target cells, but did not show enhanced cytotoxicity for MHV-infected target cells. Spleen cells from beige mice, which are deficient in NK cell activity, were able to lyse MHV-infected target cells, as were spleen cells from nude mice, which are deficient in T cell activity. Lysis of MHV-infected target cells could be mediated by cells from the spleen and, to a lesser extent, by cells from the bone marrow, but not by resident peritoneal cells or thymocytes. We suggest the term "virus killer (VK) activity" for this phenomenon. VK activity of splenocytes from different mouse strains correlated with the ability of the splenocytes to bind purified radiolabeled MHV virions. MHV virions caused agglutination of spleen leukocytes from susceptible mouse strains, indicating that leukocyte agglutination or adsorption may provide a useful assay for coronaviruses such as MHV which lack hemagglutinating activity. SJL mouse splenocytes did not bind MHV and did not lyse infected targets. MHV bound relatively well to splenocytes of other mouse strains, but poorly to thymocytes and erythrocytes. Binding of MHV to leukocytes was not influenced by 6 mM EDTA or EGTA, indicating a lack of requirement for Mg++ or Ca++. VK activity was also resistant to EDTA and EGTA, in contrast to NK activity, which was sensitive to those chelating agents. VK activity was also unaffected by actinomycin D, cycloheximide, or puromycin, indicating that new protein synthesis was not required for lysis. Antibody to interferon-alpha/beta did not block lysis, nor was there substantially enhanced lysis mediated by leukocytes from mice infected with virus and thus exposed to high levels of interferon. VK activity was blocked by antibody directed against the peplomeric glycoprotein E2 of MHV. VK activity required infected target cells, because cells with adsorbed MHV virions were not lysed by splenocytes.(ABSTRACT TRUNCATED AT 400 WORDS)

Natural cytotoxicity against mouse hepatitis virus-infected cells. II. A cytotoxic effector cell with a B lymphocyte phenotype.

The effector cell in mouse spleen which mediates natural cytotoxicity against mouse hepatitis virus (MHV)-infected target cells was characterized. The target cells were MHV-infected BALB/c 3T3, and the assay time was 3 hr. The effector cell, designated virus killer (VK) cell for the purpose of discussion, had the following phenotype: lymphocyte morphology, plastic-nonadherent, nylon wool-adherent, nonphagocytic, cyclophosphamide-sensitive; by antibody plus complement (C) depletion studies, it was asialo GM1-, NK 1.2 alloantigen-negative, Thy-1.2-, Lyt-5-, and macrophage antigen-negative; by rosetting techniques, it was Fc receptor-positive and surface Fab+; by flow cytometry (FACS) analysis, it was Lyt-2-, MAC-1-, Ia+, IgG (gamma)+, IgM (mu)+, IgD (delta)+, and B cell lineage antibody B-220+. NK cells, measured for cytotoxicity on YAC-1 cells, were similarly tested and were found to differ from the VK cell in the following properties: nylon wool-nonadherent, asialo GM1+, NK alloantigen-positive, Lyt-5+, surface Fab-, MAC-1+, Ia-, IgG-, IgM-, IgD-, and B-220-. The VK effector cell had a phenotype highly distinguishable from NK cells, effectors most commonly associated with antiviral natural cytotoxicity. The VK cell had a phenotype identical to that of a B lymphocyte and was identified as such. Although the effector cells displayed cell surface antibody, the antibody did not appear to be involved in lysis, because lysis could not be blocked by F(ab)'2 directed against Fab, mu, or delta. Cytotoxicity was more likely associated with recognition of the B lymphocyte surface by the MHV glycoprotein E2, as shown in the accompanying companion paper. This is the first demonstration that natural cytotoxicity can be mediated by B lymphocytes.

Generation of tumor cell-reactive human monoclonal antibodies using peripheral blood lymphocytes from actively immunized colorectal carcinoma patients.

The use of human monoclonal antibodies (MCA) in the detection and treatment of human cancer has been limited by the apparent scarcity of MCA to tumor cell surface antigens. Using peripheral blood lymphocytes from autologous tumor-immunized patients, we isolated 36 MCA that react to sections of colorectal carcinoma. Twenty of these human MCA appear to be directed against cell surface antigens. Two-thirds of the human MCA-producing cell lines were diploid human B-cells rather than human-mouse heterohybridomas. Direct antibody-binding assays performed with the MCA indicated that they recognized antigenic determinants preferentially expressed on tumor cells. Experiments with paired specimens of air-dried, dissociated colon tumor cells and normal colonic mucosa cells suggested that the MCA bound significantly more to the cell surfaces of tumor cells than to the surfaces of normal colonic mucosa cells. Similarly, tests with a panel of cryostat sections of paired colon tumor and normal colonic mucosa showed that MCA bound to the tumor cells and not to the normal colonic mucosa. None of the MCA bound to cells from frozen sections of normal breast, stomach, liver, skeletal muscle, or skin. Furthermore, the human MCA did not react with carcinoembryonic antigen and human erythrocyte antigens as measured by various techniques. Our data also demonstrated that these transformed B-cells and hybridomas were stable producers of human MCA. Thus, our studies show that these tumor-specific human MCA may have the specificity and stability necessary for in vivo evaluation of their use in the detection and treatment of cancer.

A diagnostic-prognostic test for bladder cancer using a monoclonal antibody-based enzyme-linked immunoassay for detection of urinary fibrin(ogen) degradation products.

An enzyme-linked immunosorbent assay (ELISA) using a monoclonal antibody was developed to determine the clinical value of urinary fibrinogen/fibrin degradation product levels for the identification and management of patients with bladder cancer. Assays were performed on 286 serial urine specimens from 56 bladder carcinoma patients. Specimens were grouped according to whether the patient had an evident tumor at the time of specimen collection (134 specimens, 41 patients) or was clinically disease-free following treatment (152 specimens, 38 patients). Many patients contributed specimens to both groups as determined by their clinical status at the time of collection. In addition, 45 specimens from 33 patients with inflammation of the urogenital tract and 81 specimens from 19 patients with renal or prostatic cancer were assayed for urinary fibrin degradation products. The ELISA, using a high-sensitivity procedure, identified 83% of the specimens from bladder cancer-positive patients with an overall accuracy with all specimens of 78% and a false-negative rate of 5% for all specimens tested. The high-sensitivity ELISA appeared most appropriate for monitoring bladder cancer patients for recurrence of tumor after surgery. The ELISA using a high-specificity procedure appeared most appropriate for screening. The high-specificity ELISA accurately identified 96% of urine specimens from non-bladder cancer patients with a false-positive rate of only 5%. These results demonstrate that the ELISA is an efficient, reliable, quantitative, and noninvasive immunoassay that can be useful both for the identification of bladder cancer patients and for monitoring the course of the disease.

Multiple organ-reactive monoclonal autoantibodies.

Autoantibodies directed against a wide range of normal tissue antigens have been found in the sera of patients with autoimmune diseases. It is generally thought that different and specific autoantibodies react with different tissues but the possibility exists that some autoantibodies may react with common antigens found in different tissues and organs. Recently, we showed that mice infected with reovirus developed a polyendocrine disease with autoantibodies to the pancreas, anterior pituitary, thymus and gastric mucosa. Using hybridoma technology, we obtained a number of monoclonal autoantibodies which reacted with antigens in single organs. We now report the production and pattern of reactivity of seven multiple organ-reactive monoclonal autoantibodies. By using antibody-affinity columns, autoantigens also have been isolated and their molecular weights determined. The results suggest that monoclonal multiple organ-reactive autoantibodies react either with the same molecule present in several organs or with common antigenic determinants on different molecules in multiple organs. In either case, the existence of multiple organ-reactive antibodies may be a partial explanation for multiple organ autoimmunity.

Virus-induced autoimmunity: monoclonal antibodies that react with endocrine tissues.

Mice infected with reovirus type 1 develop an autoimmune polyendocrine disease. Spleen cells from these mice were fused with myeloma cells and the culture fluids were screened by indirect immunofluorescence for autoantibodies reactive with normal mouse tissues. A large panel of cloned, stable antibody-producing hybridomas has been obtained. Fourteen of the hybridomas make autoantibodies that react with cells in the islets of Langerhans, 24 with cells in the anterior pituitary, 11 with cells in gastric mucosa, and 5 with nuclei. Except for the antibodies to nuclei, the monoclonal autoantibodies are organ-specific. Some, however, show broad cross-species reactivity, recognizing similar antigenic determinants in mouse, rat, pig, and human organs, whereas other recognize determinants only in rodent tissues. Several of the antigens recognized by these monoclonal autoantibodies have been identified as hormones (for example, glucagon, growth hormone, and insulin).

Acute and persistent viral infections of differentiated nerve cells.

Within the nervous system the highly specialized structure and function of nerve cells renders the pathogenesis of viral infections amazingly complex. In vivo and in vitro studies reveal that viruses may display tropism for distinct types of cells such as neurons, myelin-forming cells, or astrocytes. In neurons, RNA viruses mature in the cell body and in dendrites close to synapses, from which they can spread to synaptic endings. Undefined host factors and stage of differentiation may favor defective viral assembly, which, in turn, results in persistent infections of neurons. In myelin-forming cells, lytic infection results is persistent infections of neurons. In myelin-forming cells, lytic infection results in degeneration of myelin and, consequently, in altered conduction in those axons that are ensheathed by a myelin-forming cell. In addition, breakdown of myelin may induce an autoimmune response, which then leads to further demyelination. Autoimmune demyelination may also occur when glial cells other than myelin-forming cells are infected. Astrocytes are prone to persistent infection or viral transformation.

Genetic analysis of murine hepatitis virus strain JHM.

We performed a genetic analysis of 37 temperature-sensitive mutants of murine hepatitis virus strain JHM. Of our mutants, 32 did not induce murine hepatitis virus-specific RNA synthesis in infected cells at the restrictive temperature, 39 degrees C. By complementation testing we have identified at least seven nonoverlapping complementation groups. Six of the genes identified in this way are required for murine hepatitis virus-specific RNA synthesis. The seventh complementation group is made up of five mutants which induced virus-specific RNA synthesis at 39 degrees C.

Mouse hepatitis virus type 4 (JHM strains). induced fatal central nervous system disease. I. genetic control and murine neuron as the susceptible site of disease.

Mouse hepatitis virus (JHM strain) type 4 induces acute encephalitis followed by death in many strains of laboratory mice. Immunohistochemical study in vivo and analysis of mouse neuronal cells in vitro both indicate that the target cells in this infection is the neuron. Further, examination of several inbred mouse strains and neuronal cells from them shows that disease expression is controlled by a single autosomal gene action at the level of the neuronal cell. Susceptibility is dominant but not H-2 linked. However, cultured neuronal cells and macrophages from SJL/J mice, which are resistant to this infection, fail to make significant amounts of infectious virus after an appropriate viral inoculation. Apparently the defect is not at the level of the virus-cell receptor, because these cells, in part, express viral antigens.

Selective localization of wild type and mutant mouse hepatitis virus (JHM strain) antigens in CNS tissue by fluorescence, light and electron microscopy.

Demyelination may be induced by several different pathogenetic mechanisms. We have been utilizing mouse hepatitis virus (MHV) to study virus-induced demyelination in the central nervous system (CNS). To learn whether the different disease phenotypes in 4-week-old mice, caused by wild type (a model for fatal encephalomyelitis) or mutant ts8 (a model for primary demyelination), is due to an altered cellular tropism, we have developed an immunolabeling technique to evaluate critically the localization of MHV antigens in the unique cells of the CNS. Using mouse derived L-cells and primary neuronal cells in vitro, we determined an appropriate fixative (4% paraformaldehyde and 0.5% glutaraldehyde) that both preserved MHV antigenicity and cell structure. These studies in vitro showed the presence of MHV antigens on the surface of cells. Utilizing immunoperoxidase labeling as developed, we studied the localization of MHV antigens in vivo. MHV antigens associated with wild type (wt) virus were localized in neuronal cells as well as oligodendrocytes, which might account for the encephalomyelitis and primary demyelination, respectively. In contrast, MHV antigens associated with ts8 were localized rarely in neurons but commonly in oligodendrocytes. This might account for the uncommon occurrence of fatal encephalomyelitis, but the frequent presence of primary demyelination. Of interest was the finding of viral antigens during MHV infection in the cytoplasmic processes of oligodendrocytes surrounding intact myelin sheaths. We conclude that the different disease phenotypes caused by wt and mutant ts8 reflect differences in the cellular tropism of the two viruses for cells in the CNS.

Temperature-sensitive mutants of mouse hepatitis virus produce a high incidence of demyelination.

Mutagenesis of mouse hepatitis virus with 5-azacytidine or 5-fluorouracil yielded several temperature-sensitive mutants. Mutants have been isolated that dramatically enhance the production of demyelinating disease over that previously noted with the wild-type virus. This reproducible model should now make possible the precise elucidation of the pathogenic mechanism and molecular basis of this virus-induced demyelination.

Human histocompatibility determinants and virus antigens: effect of measles virus infection on HLA expression.

Histocompatibility antigens on the surface of human lymphoblastoid cells were quantified by a microadsorption technique. During the course of measles virus infection, no quantitative or qualitations in surface HLA antigens were observed. In contrast, infection with poliovirus type 1 or vesicular stomatitis virus, or treatment with puromycin (50 microgram/ml) resulted in a significant decrease in surface HLA. These experiments suggest that an inhibition of host protein synthesis rather than the insertion of virus-specificied antigens into the membrane results in a net decrease in amounts of this cell surface antigen. The HLA antigens also appear to be both functionally and structurally distinct from measles virus surface antigens. Pretreatment of cells with HLA-directed antibody did not prevent the infection of these cells by measles virus, thus HLA antigens appear unrelated to the measles virus receptor site on the plasma membrane. Electron microscopic studies revealed that measles virus maturation occurs at membrane sites devoid of demonstrable HLA. Furthermore, HLA antigens could not be detected on the surfaces of mature infectious virions.