Search

Monoclonal antibodies (MAbs) OC125, MOv2 and MOv8 recognize the CA125, CAMOv2 and CAMOv8 epitopes, respectively, which are associated with human ovarian carcinomas and are shed by these tumors. The biochemical analysis of the recognized antigens and epitopes has been performed to investigate the possibility of exploiting their expression, if any, on the same molecules. Cross-competition experiments clearly indicated that the 3 epitopes are distinct. Biochemical analysis was performed on a cyst fluid from an ovarian cystadenoma. Heat treatment and periodate oxidation indicated that CAMOv2 and CAMOv8 are saccharidic and suggested that CA125 is a conformational epitope to which both saccharides and the protein backbone could contribute. By gel-filtration chromatography CA125 activity was eluted into 2 peaks of high molecular weight, whereas CAMOv2 and CAMOv8 activities were associated with a single broad peak which included CA125-positive fractions. Isopycnic centrifugation showed that CA125, CAMOv2 and CAMOv8 carrying molecules had the same high density (1.46 g/ml) in the first peak, whereas CA125 molecules had a lower density (1.41 g/ml) in the second peak. Double-determinant immunoradiometric assays, carried out using different combinations of the MAbs, indicated that in the cyst fluid molecules expressing both CAMOv2 and CAMOv8 could be detected, whereas CA125 was carried by different molecules. CAMOv2 and CA125 could however be expressed on the same molecule in 2 out of 9 ascitic fluids from ovarian carcinoma patients. Taken together, these data indicate that CA125 and CAMOv2-CAMOv8 were only occasionally expressed on the same molecules. Therefore, the use of the CA125-CAMOv2 combination could not increase the sensitivity achievable by using each respective simultaneous immunoradiometric assay.

Fusion of spleen cells from a mouse immunized with a surgical specimen of a human renal carcinoma with murine P3 myeloma cells resulted in the establishment of a hybridoma cell line that secreted a monoclonal antibody (MKi-1), of IgG1 subclass, which preferentially reacted on kidney crude membrane (CM) preparations. This monoclonal antibody was tested by solid-phase radioimmunometric assay and immunofluorescence (IF) on a panel of tumor cell lines and on CM preparations and cell suspensions from surgical specimens of normal and neoplastic tissues. In addition, cryosections of normal and cancer tissues of various histologic types were tested by IF. The expression of the MKi-1 antigen was limited to normal kidney epithelium, renal cancers, some areas in the pancreas, the apical region of some breast ducts, and a proportion (5-50%) of activated lymphocytes. Electron microscopic study by the immunoperoxidase technique on fixed sections from normal kidney showed that MKi-1 stained the brush border of almost all proximal tubules. The molecule recognized by MKi-1 was a single polypeptide chain with a molecular weight of 140,000.

By a complement-dependent cytotoxicity test, 37 sera from 1- to 18-month-old BALB/c mice were tested for natural anti-tumor and anti-thymus immune reactivity. The level of anti-tumor and anti-thymus antibodies increased with age, and a high individual variability, particularly evident in the older animals, was found. The appearance of the spontaneous antibodies was correlated to the T-cell status of mice. Animals with the higher spleen T-cell content were those with the higher anti-thymus and the lower anti-tumor natural antibody level, and vice versa. In T-deprived mice the anti-tumor response increased, whereas the anti-thymus response was not influenced. An inoculum of T cells from young to older syngeneic mice yielded a decrease of the anti-tumor response and an increase of the anti-thymus response. Homeostatic immunologic mechanisms, in which the cellular and the humoral system alternate in the tumor cell control, are suggested.

A double-determinant radioimmunoassay for the detection of circulating antigens associated with human ovarian carcinoma was developed using two monoclonal antibodies: MOv2 and MOv8 employed respectively as catcher and tracer. The development of the method through three different procedures enabled us to detect the presence of CaMOv2-CaMOv8 carrying molecules in 14 out of 15 ascitic fluids from ovarian carcinoma patients whose tumors were found to be positive with MOv2 and MOv8 monoclonal antibodies by immunofluorescence. Moreover, 13 out of 15 ovarian carcinoma patients presented high levels of antigen in their serum (60-170 Ua/ml). Low levels of antigen were observed in the normal population, the values ranging from 30-40 Ua/ml. However, in 13 out of 100 apparently healthy women high levels of antigen were found in the serum

Ricin A chain was coupled to murine monoclonal antibodies MBr1 and MOv2 respectively raised against human breast and ovarian carcinomas. Inhibition of protein synthesis only occurred in those cultured human tumor cells bearing the appropriate target antigens, demonstrating that both components of the conjugate were unchanged in regards to specificity and toxicity. Conjugates were 125-200 times more efficient in inhibiting [3H]proline incorporation than the uncoupled ricin A chain. They were however unable to kill the entire population of the appropriate cells even after repeated treatment. Although the two monoclonal antibodies had similar binding kinetics, the conjugates differed in their cytotoxicity kinetics. The MBr1-ricin A chain conjugate had slow kinetics, and about 20 hours were needed to obtain a protein synthesis inhibition above 50% on the appropriate line (mammary carcinoma MCF-7). In contrast, the MOv2-ricin A chain conjugate showed very fast kinetics, reaching 50% inhibition after only 30 minutes of treatment on both appropriate cell lines SW626 and HT-29 from ovarian and colon carcinomas, respectively. Growth conditions of cell lines, i.e., adherent cells versus suspended cells, and plating time were found to greatly influence the conjugates' killing efficiencies. These studies confirm the possibility of preparing ricin A chain-antibody conjugates, which retain specific cytotoxicity against tumor cells; but they also underline the need for further in vitro studies of various parameters before one considers a therapeutic use of such conjugates.

Tumor localization was obtained in a murine system by use of 131I-labeled monoclonal antibodies, both of IgG and IgM class. The A6 IgG2 monoclonal antibody (which recognizes the gp70 of MuLV) and the B3 IgM monoclonal antibody (which recognizes a proteic structure widely exposed on chemically induced tumors), which both manifest an in vitro cytotoxic activity for various types of murine lymphomas, were injected in tumor-bearing (B6 × BALB/c)F1 mice and B6 mice, respectively, at the dose of 1 μg per mouse. The radioactivity count demonstrated an optimal tumor accumulation of the radiolabeled monoclonal antibodies 96 h after iodine injection, although an initial accumulation was already present 48 h after injection. The scanning detection was less sensitive, since at 48 h no tumor localization was possible. In the experiments with the A6 antibody, the presence of the gp70 in the circulating form, demonstrated by the detection of immunocomplexes in kidney and spleen of tumor-bearing mice injected with the A6, did not prevent radioactivity accumulation in the tumor. This accumulation was found to increase with tumor size only up to 1 g of tumor weight, then a decreased binding index was observed, whereas in the kidney the accumulation was progressive and paralleled the increase in tumor weight.

The monoclonal antibody MBr1 defines the blood group H determinant with beta 1----3N-acetylgalactosamine linkage (Fuc alpha 1----2Gal beta 1----3GalNAc----R) carried by type 3 or 4 backbone. The distribution of the antigen detected by this antibody was studied immunohistochemically in liver tissues. Although bile ducts with a diameter of more than about 100 microns normally expressed the MBr1-reactive antigen supranuclearly, smaller bile ducts and bile ductules did not express the antigen. In cirrhotic liver, proliferated bile ductules extensively expressed the MBr1-reactive antigen. In spite of the absence in normal liver cells, the antigen was expressed membranously in some cirrhotic liver cells. Under subcellular fractionation, MBr1 reactivity was almost exclusively recovered in the microsomal fraction. By HPTLC immunostaining, the major MBr1-reactive antigen was shown to be carried by type 4 chain H glycolipid (globo-H, Fuc alpha 1----2Gal beta 1----3GalNAc beta 1----3Gal alpha 1----4Gal beta 1----4Glc beta 1----1Cer). MBr1 reactive glycoprotein was not found. In conclusion, although type 4 chain H glycolipid is not expressed by normal bile ductules and liver cells, it is actively synthesized and expressed by proliferated bile ductules and some of the liver cells in cirrhosis in the absence of any neoplastic change.

Monoclonal Antibodies (Mabs) represent a promising tool for cancer diagnosis and theraphy. Administration of MAbs alone or conjugated to cytotoxic agents has been attempted but has significant limitations. Another potentially effective approach is the use of bispecific or bifunctional antibodies where the capacity to recognize the tumor cell and the toxic agent or lymphocyte activation molecule are united in one MAb. The hybrid molecule can be produced by chemical linkage between the two parentalantibodies, or alternatively by a biological approach that consists in the fusion of the two selected hybridomas. In the resulting quadroma cell the hybridoma immunoglobulin chains recombine randomly to form the bifunctional MAb. In different in vitro and in vivo models, bifunctional MAbs against tumor and CDS at nanomolar concentration has been shown to promote tumor cell killing by cytotoxic T cells. Specific localization of chemotherapeutic drugs in xenografted tumors has been demonstrated in mice pretreated with hybrid MAbs. The advantages of the hybrid MAb approach are that it should reduce the MAb biodistribution problem and that it involves no chemical manipulation between the functional agent and the MAb molecules.

The murine monoclonal antibody (MAb) MOv2 was found to be directed against the carbohydrate moieties of different kinds of molecules expressed on a human ovarian cystoadenocarcinoma. To define further the glycoconjugates carrying the MOv2-defined epitope, different procedures were used to analyze materials from surgical specimens and carcinoma cell lines. SDS-PAGE and immunoblotting showed glycoprotein molecules migrating in the gel as high and intermediate molecular weight components. A low-molecular-weight band, migrating approximately with the dye front, was also immunostained by MOv2. On the other hand, the immunostaining of high-performance thin-layer chromatography (HPTLC) of the total glycolipid extract and its neutral and acid fractions, after DEAE chromatography, showed selective reactivity with a neutral glycolipid. Reanalysis by immunoblotting of this glycolipid band scraped off the HPTLC plate indicated that it corresponds to the low-molecular-weight component. Periodate oxidation and Pronase digestion further demonstrate the saccharide nature of the determinant on both types of glycoconjugates. In conclusion, we report evidence that with a single analytical procedure, i.e., immunoblotting, it is possible to recognize the same carbohydrate determinant carried on both protein and lipid molecules.

We studied the influence of different in vitro culture conditions on the growth characteristics of certain human and murine tumor cell lines maintained in either conventional medium supplemented with FCS or in serum-free synthetic medium. A lower growth rate was observed for all of the cell lines maintained in serum-free conditions. Three human melanoma cell lines which were Ia-positive and showed an absence of pigmentation in the presence of FCS, without it produced melanin and lost their Ia antigens. In serum-free conditions, a murine hybridoma lost its membrane IgM and two human carcinoma cell lines expressed and secreted an increased amount of several tumor markers identified by monoclonal antibodies. The results of the study are in keeping with the hypothesis that tumor cell lines cultured in serum-free medium acquire differentiation characteristics.

The protein toxin restrictocin, isolated from the mould Aspergillus restrictus, inactivates protein synthesis in eukaryotic cells by blocking the ribosome elongation cycle. This protein acts as a specific nuclease that cuts off a small fragment from the 28-S rRNA. Biochemical and biological characterization of this toxin indicated that it is a non-glycosylated polypeptide of Mr 16836, exhibiting in cell-free systems a protein synthesis inhibition capacity similar to that of the ricin A chain. This polypeptide seemed unable to penetrate most of the cancer cell lines tested, as measured by its low in vitro cytotoxicity. In addition in vivo studies in BALB/c mice demonstrated that restrictocin toxicity was very low and that in rabbits, after intravenous injection 15% of the toxin was still present in the blood stream 24 h later. After derivatization with N-succinimidyl 3-(2-pyridyldithio)propionate and reduction by dithiothreitol, the restrictocin maintained its protein synthesis inhibitory activity, as assayed in a cell-free system. This derivatized toxin was then coupled to monoclonal antibodies (MBr1, MLuC1, MLuC2, MOv17, MOv18, MOv19) which exhibited a restricted spectrum of reactivity against human carcinomas. The biochemical and biological characterization of the immunoconjugates indicated that (a) when restrictocin was coupled to monoclonal antibodies with an average molar ratio of about 2, the immunoconjugates maintained the binding activity of the antibody and protein synthesis inhibition activity of the toxin; (b) four immunoconjugates were tested for cytotoxicity and three of them obtained with the MBr1, MLuC1 and MOv17 monoclonal antibodies exhibited a good level of cytotoxicity for relevant target cells and low or no toxicity for the irrelevant cell lines. The MLuC2 monoclonal antibody which gave rise to a completely ineffective immunoconjugate, induced internalization of less than one tenth of the antigenic sites whereas the MBr1, MLuC1 and MOv17 monoclonal antibodies exhibited about one third of the antigenic sites internalized. From these data it is concluded that, providing an appropriate target antigen and coupling procedure are selected, restrictocin can be considered a suitable toxin for immunoconjugate generation.

BALB/c mice repeatedly injected with syngeneic or allogeneic lymphoma cells blocked by mitomycin-C had a statistically significant decrease of lymphoma and increase of lung adenoma spontaneous incidence in comparison to untreated controls. When tested for the number of plaque-forming spleen cells, animals inoculated with mitomycin-C-treated cells revealed a severe immunodepression suggesting a role of the immunosurveillance in adenomagenesis. Also, one group of mice injected with untreated allogeneic normal immunocompetent cells revealed an increased lung adenoma incidence most likely due to a graft versus host phenomenon; this result seems to exclude the fact that a direct carcinogenic effect of mitomycin-C on lung tissue could be responsible for the increased adenomagenesis.

Thirty-six lung adenomata induced in mice by urethane followed or not by cortisone, all had an adenomatous morphology at first s.c. transplant in syngeneic hosts. Seventeen of them acquired a sarcomatous structure within a few s.c. transplant generations whilst the other 19 remained adenomatous for as long as tested, i.e. at least 10 transplant generations. The change of morphology was not dependent on s.c. growth, since tumours also transformed when allowed to grow in the lung, and was not correlated to the capacity of a tumour for growth or metastasis. The 2 types of tumours were antigenically different, since only tumours that after few transplants changed their morphology were found at the first s.c. transplant to possess tumour-associated membrane antigens as revealed by an in vitro test. In addition, only the tumours which acquired a sarcomatous morphology were found gs-positive. The majority of antigenic primary tumours arose in mice belonging to the groups of treatment which induced the strongest immunodepression. It is suggested that a predisposition to sarcoma progression is related to an immunological control, at the time of adenoma induction, of an oncornavirus, responsible for the superimposed sarcomatous change. Images Fig. 2 Fig. 3 Fig. 4 Fig. 5

A leukemogenic treatment with urethan in infant (SWR x C57BL)F1 mice (1 mg/g body weight, 5 times, once every second day) induced a marked reduction of the thymus and spleen weight, of the number of RBC, WBC, and bone marrow cells, with an inverted lymphoid/myeloid cell ratio in the bone marrow. An analysis of the bone marrow population showed that the carcinogen acted selectively on the lymphoid cells and, to a lesser extent, on the erythroblastic ones, while the myeloid cells did not appear to be involved. The damage was still present 10 days after the end of treatment. A single intracardiac inoculum of 15–20 × 106 syngeneic normal bone marrow cells, 3 days after the last urethan injection, determined a recovery of the spleen weight and of the cell ratio in the bone marrow, whereas the thymus weight and the blood and bone marrow cell counts were uneffected. Moreover the bone marrow inoculum failed to inhibit the induction of thymic lymphosarcomas.

Five injections of urethane, 1 mg./g. body weight to suckling mice markedly reduced the primary immune response against sheep red blood cells assessed by splenic plaque forming cells (PFC) determination and haemagglutinin (HA) titration. The immunological impairment lasted for about 50 days after the end of the treatment. The secondary response tested by HA titration was not affected. A lower dose of urethane (0.5 mg./g.) produced only a delay of the primary HA response. A single neonatal dose of N-nitrosomethylurea (NMU) caused a profound immunodepression evaluated as HA titre and number of PFC. Both primary and secondary responses were still depressed when tested at 50 and 90 days of age respectively. No clear correlation between the degree of immunodepression and lymphoma development was found.

Urethan was given to four inbred strains of mice and to one hybrid, either in the drinking water at 0.4 per cent for 15–20 days to 5-week old animals, or subcutaneously 3 mg once to newborns, or intraperitoneally 1 mg/g body weight, 5 times, once every second day, to 10-day old animals. The 5 treatments starting at 10 days of age induced a high percentage of rapidly developing thymic lymphosarcomas, while in the groups treated at birth or at 5 weeks, few tumors of this type were observed. The incidence of non-thymic malignant lymphomas or leukemias was not influenced by urethan, apart from a reduction of latent period. A small number of mammary adenocarcinomas occurred among the treated females of the C57BL and SWR strains but none in the control groups. In C3Hf and BC3F1 females there was a high incidence of mammary tumors when treatment was started at 5 weeks; treatment at birth or at 10 days did not modify the control incidence. On the contrary, hepatomas developed more rapidly and in higher percentage in the animals treated at birth or at 10 days. The incidence of pulmonary adenomas was increased by treatment in all strains parallel with the spontaneous occurrence. Tumors of harderian glands, skin, kidney, thyroid, adrenals and ovaries were also more frequent among treated animals.

Groups of outbred albino CTM mice were either thymectomized or splenectomized at 4 weeks of age and administered urethan 0.4 percent in the drinking water for 10 days at 5 weeks. In comparison with intact, urethan-treated animals, the thymectomized, urethan-treated mice had a slightly lower incidence of malignant lymphomas (19 against 24 %), a marked decrease of mammary tumors (31 against 62%), and increased incidence of lung adenomas (84 against 48%) and of skin papillomas (10 against 1%). Thymectomized, but otherwise untreated mice developed less lymphomas and mammary tumors than intact, untreated controls. Splenectomy did not modify significantly the tumor incidence. In a second experiment, groups of CTM mice were administered either 1 mg of cortisone s. c. daily for 10 days or urethan in the drinking water for 5 days followed by 5 days of cortisone or viceversa, or urethan alone. No major differences in the tumor incidence in the various groups were observed, but for a decreased incidence of thymic lymphosarcomas in the group with cortisone after urethan. In a third experiment, CTM mice were given 2 mg urethan within the first day after birth and thymectomized or sham operated during the anesthesia caused by the urethan. Both groups had a very high incidence of hepatomas and lung adenomas, and developed only few lymphomas and mammary tumors. Finally, a group of C3H mice thymectomized at birth under Nembutal anesthesia, developed a considerably lower incidence of hepatomas and mammary tumors than sham operated animals.

The human ovarian carcinoma cell line, IGROV1, produces two forms of folate binding protein (FBP), the membrane form that is anchored to the cell surface by a glycosylphosphatidylinositol tail and the soluble form that is shed into the tissue culture medium. Both forms are recognized by the monoclonal antibodies MOv18 and MOv19. Here we describe their purification and biochemical characterization. The purified soluble protein appeared as a single band with an apparent Mr of 36 kDa after SDS-PAGE, whereas the membrane form appeared as a single band with an apparent Mr of 38 kDa. The size difference between the two forms of FBP was confirmed by gel filtration of both the native and the N-glycanase-treated proteins. Both purified proteins had equal capacity to bind folic acid. The immunological cross-reactivity and the folic acid binding capability of the FBPs extracted from IGROV1 gave more evidence of the possible existence of a precursor-product relationship between them.

We have recently selected MAbs specific for different molecular forms of the TCR gamma/delta (expressed by distinct cell subsets), able to activate TCR gamma/delta+ cells. Two of these reagents (G1 and A13) were used for the construction of bispecific MAbs in conjunction with a MAb (Mov19) directed to ovarian carcinoma cells, using the hybrid hybridoma technique. The G1-derived bispecific MAb GM33.9 efficiently induced lysis of Mov19+ ovarian carcinoma cell lines (IGROVI and SW626) by G1+ clones in a 4-hr 51Cr-release assay. On the other hand, it was ineffective when Mov19- target cells were used. Comparable results were obtained with the A13-derived AM18.4 bispecific MAb when A13+ clones were used as effector cells. Bispecific MAbs were also able to induce secretion of IL-2 and TNF-alpha by TCR gamma/delta+ clones when Mov19+ target cells were present.

Recently we reported that cytotoxic T-cell clones can be retargeted to unrelated tumor cells by bispecific monoclonal antibodies (MAbs), anti-CD3 and anti-ovarian carcinoma (alpha OC/TR) (Mezzanzanica et al., 1988). In the perspective of in vivo tumor immunotherapy, as an alternative to cytotoxic T lymphocytes (CTL) from T-cell clones, since human peripheral blood mononuclear cells (PBMCs) without stimulation were quite ineffective, a suitable in vitro activation method was developed to render PBMCs lytic for relevant targets in the presence of the bispecific hybrid MAb alpha OC/TR. This activation protocol was applied to PBMCs from 9 healthy donors (HD) and 6 ovarian carcinoma patients (P) and to tumor-associated lymphocytes (TAL) from 4 ovarian carcinoma P. The method consisted of in vitro stimulation with phytohemagglutinin (PHA) for 2 days, followed by culture with a low dose of recombinant human interleukin-2 (rIL-2) for 6 to 10 days. The antibody-mediated lysis of CTL from HD PBMCs was found to be specifically directed against cells expressing the relevant ovarian tumor antigen when different tumor cell lines and short-term cultures of tumor and normal cells were tested. The antibody-mediated lysis of CTL from P PBMCs or TAL was efficient both on autologous and allogeneic ovarian tumor cells, whereas no reactivity with autologous normal cells was observed and LAK activity was only evident in 1 out of 4 cases. The hybrid antibody induced cytotoxic activity of CTL from P was, however, lower than that of CTL from HD.

The mechanism of effectiveness of allogeneic versus syngeneic immunization in mounting a humoral specific immune response against tumor-associated antigens of murine urethan-induced thymic lymphosarcomas was studied. C57BL mice immunized with untreated or blocked C3Hf lymphoma cells yielded sera with complement-dependent cytotoxicity for C57BL lymphosarcomas, whereas they did not respond to repeated injections of blocked syngeniec lymphoma cells, indicating a helper activity of alien alloantigens in the reflex of tumor antigens. Thymectomized, lethally-irradiated and T, but not B-reconstituted mice produced active sera, suggesting a T-dependent response. The anti-tumor activity of the serum ran parallel with the anti-histocompatibility one. The tumor-associated and the alien histocompatibility determinants had to be present on the same cell membrane. In addition, active sera were obtained when immunizing syngeneic lymphoma cell inocula were followed by allogeneic normal immunocompetent cells. It is suggested that either a helper mechanism with B and T cells cooperating in recognizing haptenic tumor-associated and carrier histocompatibility antigens or a graft-versus-host reaction determining an abnormal induction of the immune system is needed in order to detect the weak tumor-associated antigens of urethan-induced lymphomas.

In the perspective of therapeutic approaches the monoclonal antibody, MBrl, with a quite restricted spectrum of reactivity for human breast carcinoma, was coupled to restrictocin (Res), a ribosome inactivating protein produced by Aspergillus restrictus. In a cell-free system this toxin was found to have an activity comparable to that of other plant toxins, but its in vitro toxicity was shown to be low on different cell lines. Three batches of MBr1-Res conjugate were prepared and their specificity, efficiency, and maximum level of cytotoxicity were analyzed on the cell line MCF-7 expressing the relevant antigen, on several irrelevant tumor cell lines, and on normal cells. Conjugates were from 600 to 1500 times more efficient than the uncoupled derivatized Res towards MCF-7 cells and were completely ineffective on the other target cells. The antigen-driven cytotoxicity was confirmed by the nontoxicity of an irrelevant conjugate on MCF-7 cells. The cytotoxic efficiency of MBr1-Res was low when compared to the binding level of MBr1 at the same concentration and a portion of treated cells (from 10% to 30%) survived the treatment. The heterogeneity of expression of the relevant antigen, together with its only partial internalization, could account for these limitations. The lysosomotropic agent ammonium chloride and the carboxylic ionophore monensin were tested as potentiating agents but in both cases the cytotoxicity remained unmodified. A neutralization assay performed on a xenogenic model indicated that the MBr1-Res conjugate was capable of reducing the tumor take. These data indicate the possibility of using the Res to prepare a reproducible and highly selective breast cancer conjugate. However, there are still a number of problems which must first be solved before we can consider its clinical application.

In order to increase the availability of SCLC cells derived from biopsies, in vivo and in vitro growth methods were investigated. The cells grown in both conditions were periodically monitored for reactivity with 2 monoclonal antibodies (MAbs): MLuC1 directed against SCLC cells and IM1 which recognizes the class II antigen on activated lymphocytes and macrophages. About 50 % of the 28 analyzed SCLC specimens were found to proliferate in one or both systems. The in vitro-grown cells exhibited the same heterogeneity found in the original cell suspensions and moreover, in some cases only normal cells were recovered after several in vitro passages. From the subcutaneous transplanted tumors a large number of MLuC1-positive tumor cells could easily be recovered, thus indicating the validity of the in vivo methodology. The MBr1 MAb, directed against an epithelial antigen, was found to react with about 50 % of the 26 tested tumors, mainly those which demonstrated in vivo and/or in vitro growth capacity. These data suggest that only some tumors, presumably with peculiar biological characteristics, can efficiently grow in these artificial systems.

In order to investigate in vivo clinical applications of murine monoclonal antibodies directed against human ovarian carcinoma a preclinical in vivo model was developed using BALB/c athymic mice. Three human carcinoma cell lines (MCF7, HT29, and SW626) were injected into the peritoneal cavity of pristane-primed animals and the biological and antigenic characteristics of the i.p. grown tumors were studied. The animals were killed when moribund or 6-8 weeks after tumor injection. At autopsy tumor take was observed in 85% of the injected animals, whereas palpable nodules were evident in only 83%. Examination of the peritoneal cavity revealed intraabdominal carcinomatosis with tumor masses varying in size between 0.2 and 0.5 cm in diameter and tumor sheets. The most frequently affected organs were the diaphragm, the liver, and the reproductive system. Ascitic fluid formation was rare and no animal developed tumors outside the peritoneal cavity. To determine whether the in vivo tumors retained the same antigenic characteristics as the in vitro cell lines, four monoclonal antibodies (MBr1, MOv2, MOv8, and MOv15) directed against ovarian carcinoma-associated antigens and two different experimental approaches (immunofluorescence and immunoblotting) were used. Variations at either a quantitative or a qualitative level were observed for some antigens, whereas no evident changes were apparent for others. In particular, the antigens detected by MBr1 and MOv15 on the MCF7 line both maintained high levels of expression and immunoblotting staining pattern, whereas the antigens detected by MOv2 on the HT29 and SW626 lines, although present at a high level, clearly changed their staining pattern. As regards the antigens recognized by MOv8 and MOv15 on the HT29 and SW626 lines, we observed a drastic decrease in the level of their expression and in many cases a drop below the threshold of detectability of the test. The intraabdominal carcinomatosis described partially mimics the growth characteristics of human ovarian cancer and maintains the expression of some antigenic markers associated with epithelial tumors of the ovary and may therefore be useful in devising immunodiagnostic and/or immunotherapeutic strategies for ovarian carcinoma.

In the perspective of therapeutic in vivo targeting for T-cell attack, the monoclonal antibody (MAb) MOv18, selected for its restricted reactivity with human ovarian carcinoma, and an anti-T3 MAb were used to produce heteroconjugate or hybrid antibodies derived by fusion of relevant hybridomas. Specificity and activity of bispecific MAbs were analyzed by solid-phase RIA, immunofluorescence and a 51Cr-release assay on the ovarian carcinoma cell line OVCA 432, which expresses the relevant tumor-associated antigen, and on several irrelevant tumor cell lines. Both reagents efficiently promoted, at picomolar concentration, target cell lysis by cytotoxic T-cell (CTL) clones. Although the pattern of tumor cell lines which were lysed was wider than that predicted by binding studies, further studies using a double-determinant immunoradiometric assay confirmed the specificity of MAb targeting. Analysis of reagents indicated that the hybrid MAb was superior to the heteroaggregate as far as purification recovery and storage stability were concerned. Besides CTL clones, peripheral blood lymphocytes could also be used as cytolytic effectors, provided that a suitable in vitro activation scheme was used.

Antitumor antibodies were produced in vitro by spleen cells harvested from 12- to 18-month-old C57BL/6J and BALB/c mice with a naturally occurring complement-dependent serum cytotoxicity for tumor cells. The antibodies were of the IgM class and had a complement-dependent cytotoxic reactivity on EL4 cells, which was not absorbed by normal thymus cells. No natural antibodies were produced by untreated spleen cells from 1- or 3-month-old mice of the same two strains, which had no natural serum reactivity. However, after a treatment with an anti-Thy-1 serum and complement before culturing, spleen cells from 3-month-old mice, but not 1-month-old mice, produced the natural antitumor cytotoxic antibodies in vitro. When antibody-producing spleen cells from 12-month-old mice were cultured in vitro together with spleen cells from the unreactive 3-month-old syngeneic mice, inhibition of the antibody production occurred. This inhibiting capacity increased between 1 and 6 months of age and was abrogated by a treatment with an anti-Thy-1 serum and complement or with an anti-Ly-2 serum, whereas the passage of the inhibiting spleen cells on an anti-IgG column did not modify the inhibiting capacity. The in vitro data confirm previous in vivo findings on the nature, specificity and systems of regulation of the natural antitumor cytotoxic antibodies.

Absorption experiments on C3Hf normal mouse sera followed by cytotoxic tests on EL4 lymphoma cells were done to investigate a possible identity between natural antilymphoma antibodies (NAA) and various types of autoantibodies known to be present in normal mouse sera. Single C3Hf normal sera were also tested both by cytotoxicity on EL4 cells and by radioimmune precipitation assay (RIP) on 125I-labelled AKR ecotropic virus to ascertain whether or not viral antigen are the target structures of the NAA activity. The study provides evidence that NAA coexist with autoanticorpal and antiviral activities although they are distinct entities.

The possibility of using immunologic methods for detecting metastatic cells in bone marrow samples from breast cancer patients was investigated. The MBr1 monoclonal antibody, which recognizes a membrane antigen on breast carcinoma cells and is unreactive on normal bone marrow cells, seemed to be an adequate reagent for this kind of approach. When human leukocyte suspensions artificially contaminated with mammary tumor cells were tested by MBr1 immunofluorescence, it was demonstrated that the added tumor cells could be specifically discriminated from normal cells and that as little as one tumor cell in 200,000 could be detected. With the same methodology we screened bone marrow biopsies from breast cancer patients with apparently uninvolved lymph nodes at the moment of surgery. Immunoreactive tumor cells were detected by the MBr1 antibody in 17% of N- patients. None of the bone marrow samples showed any evidence of tumor involvement by conventional histologic analysis.

First cycle transformants of NIH 3T3 cells transfected with metastatic human thyroid carcinoma DNA were used as immunogen to obtain monoclonal antibodies (MAbs) against normal and transformation-related antigens. The transformed cell line (M33) was shown to contain Alu sequences. Two MAbs were selected on the basis of their differential reactivity toward untreated NIH 3T3 cells or the transformed M33 cell line. By immunofluorescence, immunoelectronmicroscopy and biochemical analysis, the first MAb (MTr1) was demonstrated to recognize an epitope on cytoskeletal filaments of proliferating murine fibroblasts. Similar MTr1-labelled filaments were also found to accumulate into cytoplast-like structures spontaneously produced by M33 cells. The characterization by immunofluorescence of MTr2, the second MAb, indicates that it recognizes a specific human antigen associated with normal thyroid epithelial cells and differentiated thyroid tumors.

In the perspective of in vivo therapeutic applications, the monoclonal antibody (MAb) MOv18 was selected for its restricted reactivity with human ovarian carcinoma. Using the pH 2.8 desorption assay, we found that the antigen recognized by MOv18 had a high stability on the cell membrane and poor internalization. Therefore, a therapeutic approach which does not require internalization, i.e., the re-targeting of cytotoxic T lymphocytes (CTL) by bispecific MAbs, was investigated. MOv18 and anti-CD3 MAbs were used to produce bispecific reagents, obtained either by chemical cross-linkage (heteroconjugates) or by somatic hybridization techniques (hybrid MAbs). The maintenance of the binding reactivity and specificity of the bispecific MAbs was analyzed by solid-phase radioimmunoassay, immunofluorescence and cross-competition tests on the relevant target cells (ovarian carcinoma cell line OVCA 432 for MOv18 and PHA-stimulated peripheral blood mononuclear cells for anti-CD3 MAbs), and on 2 irrelevant tumor cell lines. Bv a 51Cr-release assay the bispecific MAbs were found to efficiently promote, at picomolar concentration, cell lysis by CTL clones, but the specificity pattern was wider than that predicated by the binding studies. The F(ab')2 fragment of one hybrid MAb mediated a lysis which was just as efficient as the entire MAb on the relevant target cells and allowed specific lysis to be distinguished from Fc-receptor-mediated lysis. Human immunoglobulins were unable to compete with the Fc receptor binding of the hybrid MAbs and therefore, in the perspective of in vivo applications, Fc fragment removal seems to be an essential step. Analysis of the bispecific reagents indicated that hybrid MAbs are superior to the heteroconjugate as far as storage stability is concerned.

Although the relative 5-year survival rate of breast cancer patients has improved over the past few years, the age-adjusted mortality rate has remained stationary because the age-adjusted incidence rate has increased. A local recurrence or the development of distant metastases after therapy is part of the natural course of the disease. Reliable factors capable of predicting which patients will develop metastatic cancer, and when, are not available yet. Clinical, pathological and biochemical studies contribute to the identification of a number of parameters which can help us predict early recurrences of the disease after radical surgery in women with operable disease. In fact, prognosis seems to be correlated, to a certain extent, with some of them, evaluated separately or in combination. The prognostic factors most commonly taken into consideration include the clinical stage, nodal status and hormone receptor levels (1,2). However, even well-established parameters are not entirely dependable.

Normal adult C57BL/6J fibroblasts, cultured to saturation density and therefore in resting phase, were harvested by EDTA, treated with trypsin at various concentrations for different lengths of time and then tested for absorbing capacity of a C57BL/He anti-embryo serum, the residual cytotoxic activity being measured by a 51Cr release assay on a C57BL/He lymphosarcoma known to carry embryonic antigens. A weak proteolytic treatment (0.05 – 0.25 % trypsin for 5–10 min) uncovered structures which absorbed over 40 % activity of the anti-embryo serum. The treated fibroblasts partially retained the absorbing capacity for an antihistocompatibility serum. Higher doses or longer exposure to trypsin progressively inactivated the absorbing capacity of fibroblasts for both antisera. Trypsin treatment of fibrosarcoma cells, which when untreated completely absorbed the anti-embryo serum, decreased their absorbing capacity. Additionally, normal untreated fibroblasts in growing phase were found to absorb the anti-embryo serum.

131I-labelled anti-CEA monoclonal antibody was tested in an animal model to evaluate: influence of antibody type (whole versus F(ab’)2 fragments), administration route (i.v. versus i.p.), dose of tracer (100 μCi versus 1000 μCi), growth site (s.c. versus i.p.) and size of tumor. Athymic mice bearing CEA-producing human colon carcinoma (HT-29) or human melanoma as an irrelevant tumor (MeWo) received tracer and immunoscintigraphy and the localization ratios (LR) were compared. In HT-29 bearing animals F(ab’)2 fragments localized better than the whole antibody. The LR were higher after i.p. administration of the tracer, independently of the tumor characteristics or the injected dose. The highest values were achieved when the radioactivity remaining in the whole body was below 2% of the injected dose. The images were negative when the i.p. injected dose was low or tumor growth was i.p. but positive in the other conditions (i.v. administration, high tracer dose, s.c. tumor growth). In the animals bearing melanoma, images scored positive or negative when the tumor weight was respectively above or below 400 mg, but the LR were always low.

The reactivity profile of an IgM monoclonal antibody, MBR1 , raised against the human breast cancer cell line MCF7 , was studied in a variety of human tumours and non-neoplastic tissues by light microscopic immunohistochemistry. The range of reactivity included specific types of non-neoplastic epithelial cells and a number of epithelial tumours. Most mammary carcinomas reacted with MBR1 , but adenocarcinomas and squamous carcinomas from different sites were also strongly positive. Different patterns of immunoreactivity were apparent in microscopically normal tissues, in tissues with inflammatory changes and in carcinomas. Heterogeneous staining, despite morphological similarities, was documented in neoplastic and non-neoplastic epithelial cells. The reactivity of MBR1 was different from that reported for other monoclonal antibodies, but revealed similarities to that of monoclonal antibodies and polyclonal sera against human milk fat globule membrane.

An antiserum obtained by the immunization of C57BL/HeDp mice with a pool of C3HF/Dp 7,12-dimethylbenz[alpha]anthracene (DMBA)-induced fibrosarcomas exerted a specific cytotoxic activity in vitro on C57BL/HeDp chemically induced lymphosarcomas. Conversely, C57BL/HeDp spleen cells sensitized against C3Hf/Dp chemically induced lymphosarcomas or embryo cells were cytotoxic for plated cells of syngeneic DMBA-induced fibrosarcomas. Absorption studies with antiembryo and antilymphoma antisera showed that embryonic antigens were shared between lymphosarcomas and fibrosarcomas and that all serologically defined antigens present on lymphoma cells, including virus-related antigens, were also on fibrosarcoma cells.

A natural cytotoxic reactivity directed against syngeneic or allogeneic tumor cells was demonstrated in serum of BALB/c mice by an in vitro cytotoxicity test using rabbit serum as the source of complement. The reactivity, studied on syngeneic fibrosarcoma cells, was found to be minimal in mice less than 10 weeks old and to increase progressively with age. T-deprivation determined an increase of reactivity in young mice to levels reached spontaneously only by the serum of 40-week-old mice. The BALB/c serum also revealed natural anti-thymus antibodies. Non-identity between anti-tumor and anti-thymus antibodies was demonstrated by direct cytotoxicity and absorption tests. An inoculum of syngeneic fibrosarcoma cells increased the level of anti-tumor serum reactivity in both normal and T-deprived young mice. The natural anti-tumor cytotoxicity revealed in vitro seemed to exert a specific in vivo protection as suggested by the indirect correlation found between the level of the natural anti-tumor reactivity and the growth of a transplanted fibrosarcoma.

The application of classical hybridoma technology to the study of human mammary carcinoma resulted in the generation of a variety of monoclonal antibodies that identify several antigens associated with this type of tumor (1–3). Monoclonal reagents with selected specificities are therefore available in increasing number, and their potential for applications in the diagnosis and characterization of breast cancer is under investigation.

In clinical oncology monoclonal antibodies have been proven useful for the diagnosis of breast carcinoma. Cells from 111 pleural and peritoneal effusions from patients with neoplastic and non-neoplastic diseases were studied using indirect immunofluorescence with MBrl and MOv2 monoclonal antibodies. The results, compared with those obtained by conventional cytology, demonstrated that the immunologic test allowed to retrieve 37% false negative cases among 49 breast carcinoma patients. MBrl was also used to examine bone marrow biopsies from T1–3NO breast cancer patients at the time of surgery. Reactive cells were found in about 20% of the 54 patients pathologically N- examined. A prospective prognostic study to evaluate the clinical significance of such a finding is in progress. The use of a pool of monoclonal reagents in comparison to a single monoclonal antibody was found to increase the efficiency of tumor detection from 70% to 100% when primary tumors or metastatic effusions were examined by immunofluorescence.

The nature and some biological properties of the natural antitumor cytotoxins previously found in normal C3Hf serum, have been investigated. By immunoelectronmicroscopy study, employing rabbit hybrid antibodies with specificity for mouse Ig and ferritin, the C3Hf cytotoxins were shown to belong to immunoglobulins. By gel filtration, by inhibition experiments of the cytotoxic serum activity with monospecific anti-IgG and anti-IgM sera, and by 2-mercaptoethanol treatment of the C3Hf serum, the cytotoxic immunoglobulins were demonstrated to belong to the IgM class. They were not inactivated by heating until 60 °C and were able to activate guinea pig, rabbit, and human complement. The highest cytotoxic activity of the normal C3Hf serum was found when cells and serum were incubated at the low temperature, suggesting a low binding affinity of the cytotoxic IgM.

Using an avidin-gold conjugate, the binding and internalization of the biotinylated anti-breast cancer monoclonal antibody MBr1 in MCF-7 cells were examined. After labelling MCF-7 cells at 0 degrees C, MBr1 was found to specifically bind to the entire cell surface. MBr1 distribution was diffused in cells fixed before labelling, but appeared in patches of different sizes in cells immediately fixed after labelling. On warming prelabelled cells at 37 degrees C for 2 min, MBr1 was internalized through non-coated small invaginations and vesicles an also through smooth invaginations with coated regions at the bottom. After warming for 15 and 30 min to 37 degrees C, the MBr1 internalization sites were observed as large membrane invaginations, large vesicular structures and lysosomes. The early steps of MBr1 internalization which consists of non-coated micro-invaginations seem to be correlated to the glycolipidic nature of the MBr1 recognized molecule. The MBr1 ability to enter MCF-7 cells suggests that this anti-tumour monoclonal antibody may be used as a toxin ability to enter MCF-7 cells suggests that this anti-tumour monoclonal antibody may be used as a toxin carrier agent. Images Figure 1

SummaryA leukemogenic treatment with urethan in infant mice induced a prolonged depression of primary hemagglutinin response against sheep red blood cells. The development of lymphosarcomas was found to be correlated with the degree of immunological impairment.

A microcytotoxicity test for cell-mediated immunity against target cells cultured in plastic microplates is described. The cytotoxic effect is measured by 51Cr release from target cells after 24-hr incubation with effector cells, and can be controlled by visual evaluation of the surviving target cells attached to the floor of the microwells. The assay requires a very low number of target and effector cells and allows a higher number of replicates. The method was found suitable for the detection of tumor-specific antigens

Normal, preleukemic and leukemic thymuses of CTM, C3Hf, C57BL and SWR mice were examined for their chromosome constitution. In untreated infant and young adult animals deviation from diploidy was rare. During and shortly after a brief course of a highly leukemogenic urethan treatment, little variation in chromosome number was found but diploid metaphases had many different chromosome aberrations which became rare a week after completion of treatment. No deviation from normal was observed in thymuses examined 5 and 11 weeks after urethan. The total analysis of 77 thymic lymphosarcomas showed a 40–41 bimodal distribution, although 7 cases had 80% or more diploid metaphases and 3 had 80% or more metaphases with 41 chromosomes. The additional chromosome was always identical or similar to those of the smallest pair.

Absorbed serum from a rabbit immunized with an extract of a pool of 5 breast carcinomas yielded a single precipitation line when tested on the pooled tumors used to immunize and on single breast carcinomas, whether participating in the immunization or not. No precipitation was seen with control normal mammary tissues from the same patients. Antisera from 5 out of 9 rabbit immunized with different single breast carcinomas gave one precipitin line against the tumor used to immunize and also in cross-reaction against the other 4 mammary carcinomas positive in the direct test. No bands were obtained with normal mammary tissue extracts or with a fibroadenoma extract. A rabbit antiserum against the fibroadenoma was always negative on all the breast carcinoma extracts tested.

Hybridomas which secrete monoclonal antibodies (MAbs) reacting with doxorubicin (DXR), a widely used anthracycline cytotoxic antibiotic, have been obtained by fusing NSO/P3 mouse myeloma cells with spleen cells from BALB/c mice immunized with DXR-BSA conjugate. The best producer among the several clones obtained was expanded and the secreted MAb (MAD2) purified and characterized. MAD2 cross-reacts to varying degrees with anthracycline compounds such as some DXR analogues and derivatives, but does not recognize anthracene and anthraquinone structures, with the exception of weakly reacting Mitoxantrone. MAD2 and the panel of MAbs which are at present being purified may become a tool for studying the relevance of different domains of the anthracyclin molecule in terms of biologic activity.