Selective localisation of two radiolabelled anti-sarcoma monoclonal antibodies in human osteosarcoma xenografts

Two mouse monoclonal antibodies (MoAbs), TP-1 and TP-3, previously shown in immunohistochemical studies to react with osteosarcomas, were labelled with 125I or 131I and evaluated for their ability to localise to human osteogenic sarcoma xenografts after intravenous injection. The radiolabelled TP-1 and TP-3 MoAbs had immunoreactive fractions of 70% and 67%, respectively, and bound to target cells with binding constants of 8.5 X 10(8) M-1 and 4.0 X 10(9) M-1, respectively. After injection of labelled TP-3 IgG, approximately 16% of the dose X g-1 tissue was found in the tumour after 24 hours. Maximum tumour/blood radioactivity ratios of 6-7 were achieved 3-4 days after antibody injection, while the ratios for the normal tissues were less than 1. The tumours could be clearly visualised by whole-body gamma scintigraphy without the need for subtraction techniques. The TP-1 IgG accumulated to a large extent also in the spleen. Hence, with this antibody the tumour was less well delineated from the adjacent normal tissues. However, the F(ab')2 fragments, derived from the TP-1 IgG, gave tumour/blood ratios up to approximately 40 after 3-4 days and yielded sharp gamma scintigrams of the tumour. Specificity of the antibody localisation was indicated by the lack of accumulation in a contralateral melanoma xenograft and the failure of 2 isotype-matched irrelevant MoAbs to localise to the sarcomas. With the F(ab')2 fragments satisfactory images could be obtained already after 16 hours. The results suggest that this preparation may be useful in clinical radioimmunodetection of osteogenic sarcomas.

Mitozolomide (NSC 353451), a new active drug in the treatment of malignant melanoma. Phase II trial in patients with advanced disease

A phase II trial with mitozolomide was carried out in patients with malignant melanoma, since in preclinical studies this new imidazotetrazine had shown promising effects against human melanoma xenografts. Twenty-one evaluable patients with advanced malignant melanoma were treated with 115 mg m-2 of mitozolomide, given orally every 6 weeks. None of the patients had received prior chemotherapy. Two partial responses (10 and 7+ months) were observed. The responding patients had lung metastases, and one of them had, in addition, a huge (17 X 14 cm) lymph node metastasis in the groin. Also, one patient had a 48% tumour volume reduction of lung metastases. The dose limiting side effect of the treatment was bone marrow depression, with delayed leukopenia and thrombocytopenia. The median white blood cell counts and platelet nadirs were 2.5 X 10(9) 1(-1) (range 1.1-3.8) and 59 X 10(9) 1(-1) (range 14-95), respectively. Non-haematological adverse reactions were limited to mild or moderate nausea. It is concluded that orally administered mitozolomide is active against malignant melanoma and seems to have a response rate comparable to those of the most active established drugs.

Elimination of B-lymphoma cells from human bone marrow: model experiments using monodisperse magnetic particles coated with primary monoclonal antibodies

Conditions for removing B-lymphoma cells from human bone marrow using "immunobeads" (IBs) were investigated. The IBs were prepared by coupling monoclonal antibodies directly to a new type of monodisperse magnetic polymer particles (M 450). Two monoclonal immunoglobulin M antibodies, AB-1 (CD 19), a B-cell-specific antibody, and AB-4, an HLA-DR-specific antibody, were used. The IBs were incubated with Rael Burkitt lymphoma cells admixed to fresh, mononuclear human bone marrow cells. After incubation for 30 min at 4 degrees C, the IBs were removed using cobalt samarium magnets. The number of remaining clonogenic tumor cells was assayed by the Courtenay and Mills soft agar procedure, and the clonogenic capacity of the bone marrow progenitor cells was measured by granulocyte-monocyte and granulocyte-erythroid-monocyte-megakaryocyte assays. With a ratio of tumor cells to normal bone marrow cells of 0.1 or 0.01 and a ratio of immunobeads to tumor cells in excess of 75, a tumor cell depletion of more than 3 logs was achieved with the AB-4 IBs and slightly less with the AB-1 beads. After two consecutive cycles of purification with the AB-4 beads, no colonies were found, corresponding to more than 6 logs of purification. In the case of the AB-1 beads, 4 to 5 logs of purification were achieved. The concomitant reduction in clonogenic bone marrow progenitor cells was only 30 to 40%. Flow cytometric studies showed that the tumor cell population contained appreciable proportions of cells binding only small amounts of the antibodies used. The results indicate that the IB procedure is highly efficient and capable of removing tumor cells expressing low levels of antigen. Compared to other purging methods in use the procedure described seems to offer several advantages with respect to efficacy, speed, and simplicity. By the use of a panel of suitable antibodies the new immunobead procedure may be potentially useful in autologous bone marrow transplantation of B-lymphomas and non-T-leukemias with poor prognosis.

Chromosome 13 instability and esterase D expression in an osteosarcoma cell line

In order to assess the involvement of the 13q14 region in the development of osteosarcoma, both osteosarcoma tumor cells and normal tissue from a retinoblastoma patient previously used in restriction fragment length polymorphism studies, and sarcoma cells and normal fibroblasts from other tumor patients, have been investigated with respect to esterase D (E.C. 3.1.1.1) expression and chromosome pattern. In spite of an increased number of apparently normal chromosomes #13, a 50% reduction in esterase D activity in osteosarcoma cells from the retinoblastoma patient was observed. This suggests that loss of the RB1 gene or an OSRC gene closely linked to the ESD and RB1 gene loci is involved in the development of the osteosarcoma tumor. No reduction in esterase D expression was seen in four other sarcoma cell lines.

Human melanoma cell lines showing striking inherent differences in sensitivity to immunotoxins containing holotoxins

The in vitro sensitivity of human melanoma cell lines to conjugates of whole abrin or ricin linked through disulfide bonds to the monoclonal antimelanoma antibody 9.2.27 was studied. After passage of the conjugates through a Sepharose 4B column to remove molecular species with exposed binding sites on their B-chains, toxicity of the conjugates to different melanoma cell lines and nonmelanoma tumor lines was assessed by measuring their ability to inhibit cellular protein synthesis. The abrin conjugate was far more toxic to the target cells than the corresponding ricin conjugate. The 8 melanoma cell lines studied differed widely in their sensitivities to the abrin conjugate. The differences were associated with concomitant large differences in the sensitivities of the cells to the native toxins, and the significance of the level of the antigen expression became apparent only when the sensitivities of the different cell lines were normalized with respect to their sensitivity to native abrin. The observed relationship could not be accounted for by unspecific binding via the B-chain binding site of the immunotoxin. The differential sensitivity of the melanoma cell lines to the immunotoxin seems to be related to inherent differences between the cells in their ability to internalize and to process immunotoxins and toxins. The findings may have considerable practical implications.

Characteristics of a cell line established from a patient with multiple osteosarcoma, appearing 13 years after treatment for bilateral retinoblastoma

An osteosarcoma cell line, OHS, was established from a patient with multiple skeletal manifestations of osteosarcoma, developing after bilateral retinoblastoma. The tumor cells expressed sarcoma-associated antigens and showed rapid growth in monolayers and as multicellular spheroids. They formed distinct colonies in soft agar, and subcutaneous tumors in nude mice. Morphological studies indicated that OHS cells had retained important characteristics of the cells of origin. No deletion of the retinoblastoma genes on chromosome 13q14 could be demonstrated with the banding techniques used. However, cytogenetic studies revealed double minute chromosomes, as evidence of gene amplification, as well as translocations involving chromosomes 1,6,11 and 13. The OHS line can be used to study the genetic basis of tumor initiation and growth, and to elucidate factors predisposing for second primary cancers in retinoblastoma patients.

New monoclonal antibodies specific for human sarcomas

Monoclonal antibodies (MAbs) against sarcoma-associated cell membrane antigens were prepared by immunizing BALB/c mice with tumor cells from a human osteosarcoma, TPX, grown as a xenograft in athymic BALB/c nude mice. Spleen cells from immunized mice were hybridized with X-63 Ag. 8.653 mouse myeloma cells which yielded 260 growing hybridomas. Seven of these produced antibodies that bound to TPX cells and to cells from another osteosarcoma, but not to autologous skin fibroblasts. MAbs from 2 (TP-1 and TP-3) of these 7 clones did not cross-react with non-sarcomatous tumor cells or peripheral blood lymphocytes. Immunohistochemical studies on frozen tissue sections showed that the TP-1 (IgG-2a) and TP-3 (IgG-2b) antibodies had characteristic and identical specificity profiles. Binding of TP-1 (TP-3) was demonstrated to 15/15 (15/15) osteosarcomas, 3/3 (2/2) synovial sarcomas, 7/9 (6/8) malignant fibrous histiocytomas, 2/2 (1/1) malignant hemangiopericytomas, 1/2 (1/2) chondrosarcomas and 3/6 (1/3) unclassified sarcomas. The antibodies did not bind to any of 16 sarcomas belonging to other histological subtypes, including liposarcomas and leio- and rhabdomyosarcomas. Moreover, they failed to bind to sections of 66 different non-sarcomatous malignancies, or to any of a range of normal adult and fetal tissues, although some weak staining of proximal kidney tubules was seen. The restricted specificity of these antibodies to some major subtypes of human sarcomas makes them promising tools for identification and subclassification of sarcomas.

Unexpected expression of the 250 kD melanoma-associated antigen in human sarcoma cells

Images

Chemosensitivity profiles of human cancers assessed by the 6-day SRC assay on serially xenografted tumors

To assess the potential value of the 6-day subrenal capsule (SRC) assay in preclinical evaluation of new drugs using serially xenografted human tumors as source of tumor tissue, we studied the response of 31 human tumor lines (8 malignant melanomas, 12 sarcomas, 9 lung carcinomas and 2 colon carcinomas) to relevant standard drugs and to a new imidazotetrazine, Mitozolomide. Mitozolomide was found to be the most active drug tested in 50% of the lung carcinomas and as active as CCNU in melanomas. The activity of the standard anticancer drugs against subrenal grafts closely resembled the patterns seen with the same tumors in the clinic. In further attempts to validate the procedure, sensitivity profiles of some tumors were concurrently determined in the subcutaneous (s.c.) nude mouse model. In 11 out of the 12 tumors, the two assays selected the same drug as being the most active and in most of these tumors the two procedures gave the same ranking for the different drugs. Also, when the relative sensitivities of a series of melanoma xenografts to each of two drugs (DTIC and CCNU) were tested, the two assays gave the same ranking of the xenografts for each drug. The concordance between the two assays and the fact that the s.c. nude mouse assay reflects the chemosensitivity of the parent tumor in patients, suggest that the application of the 6-day SRC assay to xenografted tumors is a valid and useful procedure permitting rapid preclinical evaluation of new drugs to be carried out at relatively low cost.

Intra-arterial infusion of mitomycin C in treatment of breast cancer: occurrence of skin necrosis in irradiated patients

Eight previously irradiated breast cancer patients with local recurrences were treated with intra-arterial infusions of 8 mg/m2 mitomycin C given at 3-week intervals. The mean time interval between radiotherapy and intra-arterial chemotherapy was 38 months (range 2-60). In five cases a temporary reduction in tumour size was observed. However, in 3 of the 8 patients severe local pain, starting immediately after the third course of treatment, was followed 4 weeks later by the development of deep necrotic ulcers of the chest wall. These cases are reported here and discussed in relation to the results of previous studies.

Radiolocalization of xenografted human malignant melanoma by a monoclonal antibody (9.2.27) to a melanoma-associated antigen in nude mice

A murine monoclonal antibody (9.2.27), directed to a Mr 250,000 glycoprotein-chondroitan sulfate proteoglycan complex, was radiolabeled with 125I and assessed for radiolocalization in tumor and normal tissues of normal and tumor-bearing nude mice. The 125I-9.2.27 localized in vivo preferentially in Mr 250,000 antigen-expressing human melanomas (FMX-Met, SESX) but not in low antigen-expressing tumors (LOX-L) xenografted in nude mice. The imaging index of tumor cells was positively correlated with the antigen density of the various melanoma cell lines as measured by flow cytometry. The nonspecific immunoglobulin RPC-5 of the same IgG2a subclass as 9.2.27 did not specifically localize to xenografts of melanoma. The total amount of 125I-9.2.27 accumulated in the tumor was directly correlated with tumor size. However, the specific radioactivity (cpm/g) in smaller tumors was higher than that in larger tumors. Nonspecific uptake and circulating antibody levels differed between normals and tumor-bearers. The organs of the reticuloendothelial system of normal mice accumulated more labeled antibody than did those of tumor bearers, and conversely, tumor bearers had higher levels of circulating labeled antibody in the blood than normals. The circulating labeled antibody in tumor bearers was still monomeric but had no detectable antigen-binding capacity.

Methodological aspects of the 6-day subrenal capsule assay for measuring the response of human tumors to anticancer agents

Methodological aspects of testing chemosensitivity by the 6-day subrenal capsule (SRC) assay in immunocompetent mice were investigated. Human tumor xenografts, serially transplanted in athymic nude mice, were used as source of material. All drugs were given by the intravenous route. Administration of the drugs on days 1 and 2 gave the same results as when they were given daily for 5 days in equitoxic total doses, and a clear dose-response relationship was demonstrated. High reproducibility was found with different anti-cancer agents when 15 different tumors (4 melanomas, 7 soft tissue sarcomas, 2 colon carcinomas, and 2 lung carcinomas) were tested repeatedly over a period of several years. The tumors examined showed individual chemosensitivity profiles. The same ranking of drugs was found when the results in the SRC assay were compared with those obtained in the sc nude mouse model, using the same tumors (a colon carcinoma and a leiomyosarcoma), supporting the validity of the SRC assay. Altogether, the results strongly support the view that the 6-day SRC assay in immunocompetent mice is a useful method for assessing the response of human tumors to anticancer agents.

The use of multicellular spheroids in establishing human sarcoma cell lines in vitro

By a new procedure stable monolayer cultures were derived from spheroids in 8 out of 17 different human sarcomas (16 soft-tissue and I osteogenic sarcoma). Eleven of the sarcomas were obtained from patients undergoing surgery, and 6 from BALB/c nude mice carrying s.c. growing xenografts. The new procedure involves aggregation of single-cell suspensions into spheroids and cultivation of these in agar-coated flasks until the growth rate levels off, at which time the spheroids are transferred to uncoated flasks. Cells proliferating from the rim of adhering spheroids are trypsinized and aggregated to form new spheroids. By 3 to 5 such alternations, monolayer cultures were obtained that have now been subcultured for about 6 months. The cell lines all gave rise to colonies in a clonogenic soft-agar system, and upon s.c. injection into athymic nude mice 3 lines tested formed growing tumors. The histology of spheroids formed from late monolayer passages closely resembled that of the original tumors. That the new procedure is superior to other methods of establishing sarcoma cell lines is indicated by the fact that a stable monolayer culture could be obtained directly from the tumors in only 1/8 cases where the above procedure was successful, and in only 2 instances from soft-agar colonies derived from the tumors.

Activity of mitozolomide (NSC 353451), a new imidazotetrazine, against xenografts from human melanomas, sarcomas, and lung and colon carcinomas

The chemosensitivity of human tumor xenografts to mitozolomide, 8-carbamoyl-3-(2-chloroethyl)imidazo[5-1-d]-1,2,3,5-tetrazin-4(3H) -one, was studied in 3 different assay systems. In concentrations of 1 to 500 micrograms/ml, mitozolomide completely inhibited the colony-forming ability in soft agar of cell suspensions from sarcomas, melanomas, lung and colon cancers, and a mammary carcinoma. When a panel of tumors of the different histological types was tested for its sensitivity to mitozolomide in vitro, in the 6-day subrenal capsule assay in conventional mice, and, in some cases, as s.c. growing tumors in nude mice, good agreement between the different assay systems was seen. In most cases, a very pronounced antitumor effect was observed. The efficacy of mitozolomide was as good or better than that of the drugs clinically used against the tumor types tested. Tumor size measurements and histological examinations indicated that nude mice carrying a melanoma, a small cell lung cancer, and an osteosarcoma were cured of their tumors. The approach here used for evaluating the effect of a new drug on human cancers may be useful for selecting the tumor types which primarily should be studied in clinical trials. The results indicate that clinical responses to mitozolomide may be anticipated in sarcoma, melanoma, small cell lung cancer, and possibly in colon cancer.

Characteristics of human tumour xenografts transplanted under the renal capsule of immunocompetent mice

Human tumour lines established in athymic nude mice were grafted under the renal capsule of immunocompetent mice. Grafts from 27 human tumour lines comprising 9 malignant melanomas, 10 sarcomas, 2 colon carcinomas, 4 lung carcinomas and 2 mammary carcinomas, grew well under the renal capsule of the immunocompetent mice and retained morphological and functional characteristics of the parent tumours, as judged by light and electron microscopy and immunohistochemical examinations. Numerous mitoses were detected. Granulation tissue and necrosis were not predominant features. After Day 4, the grafts became infiltrated from the periphery by mouse inflammatory cells. The infiltration could be prevented by pretreatment of the animals with cyclophosphamide. Anti-human antibodies were detected after Day 3. Single cell suspensions from the subrenal grafts were able to form colonies in soft agar. and upon reimplantation in nude mice, subcutaneous tumours were formed showing that the grafted tumour tissue had also retained its malignant character. Altogether the results support the view that human tumour xenografts grow well under the renal capsule of immunocompetent mice and that the grafts retain important characteristics of the original tumour.

Human tumor xenografts transplanted under the renal capsule of conventional mice. Growth rates and host immune response

The growth of 29 different human tumor lines under the renal capsule of immunocompetent mice was investigated. The tumors, previously established in athymic mice, included malignant melanomas, colon carcinomas, soft-tissue sarcomas, lung cancers and a mammary carcinoma. The growth rates of 17 tumors, measured repeatedly over a period of several years, were highly reproducible. The different grafts exhibited distinctly different and individual growth rates for up to 6 days. In animals pretreated with an immunosuppressive dose of cyclophosphamide, the growth rate was the same as in non-pretreated animals, but the growth continued for several more days. In the case of 9 different grafts, the subrenal growth rates were compared with those observed when the same tumors were growing subcutaneously in athymic, nude mice. The relative growth rates of the different tumors were practically the same in the two systems. The results indicate that the growth conditions under the renal capsule permit the grafts to express their inherent growth potentials and that the subrenal grafts do not represent a selected sub-population of the tumor cells. The extent of infiltration of the grafts by mouse inflammatory cells was measured by flow cytometry on single-cell suspension as well as by quantitative analysis of serial histological sections. In most cases the mouse cells occupied 15-25% of the total graft volume on day 6. The results indicate that the effect of mouse cell infiltration on the growth of established tumor lines is slight and that it is unnecessary to use athymic mice as host animals when testing new investigational drugs by the SRC assay. The use of established tumor lines in the SRC assay in immunocompetent mice may be useful also in the study of factors influencing the anti-cancer activity of current drugs.

Stable quantitative differences of antigen expression in human melanoma cells isolated by flow cytometric cell sorting

The stability of expression of a membrane antigen on human FME melanoma cells was investigated by means of flow cytometric cell sorting and analysis. The melanoma-associated 250 kd antigen was strongly expressed on all cells, as recognized by binding of the monoclonal antibody 9.2.27. By flow cytometric cell sorting, cells of high and low antigen expression were isolated, and the difference in antigen expression between the two populations was examined as a function of time in culture. Immediately after sorting, the median fluorescence intensities of the two populations differed by a factor of 2.7. After the first few days in culture, much of the range in antigen expression of the parent population was regenerated. However, a lasting difference in antigen expression was established, corresponding to 50% higher density of antigen on the cells sorted for high fluorescence intensity, compared to those sorted for low intensity. After trypsin treatment, which removed the antigen from the cell surface, normal antigen expression was regained after 2-3 days in culture, with the same difference between the two populations as before the trypsin treatment. The stability of the established difference in antigen expression between the two sorted subpopulations indicates that expression of this antigen is a precisely controlled, heritable characteristic of the FME melanoma cells.

Pharmacological studies of ricin in mice and humans

A highly sensitive enzyme-linked immunosorbent assay (ELISA) for determination of ricin in serum is presented. Using this method it was found that IV-injected ricin disappeared from the plasma of mice and cancer patients according to first-order kinetics. DBA mice were found to be more sensitive to ricin than C3H and B6D2 mice. When mice of the different strains were given the same dose of ricin, the concentrations found in liver, spleen, and kidneys were highest in the most sensitive mice. Ricin disappeared most rapidly from serum of the mice with the highest sensitivity. The inverse correlation between the rate of disappearance of ricin from serum and the tissue concentrations reached may be due to the fact that ricin is rapidly and firmly bound to cell surface receptors. Whole-body autoradiography after IV injection of 125I-labeled ricin showed the highest amount of radioactivity in liver, spleen, and adrenal cortex. Considerable amounts of radioactivity were also present in bone marrow, showing that the lack of myelosuppressive activity of ricin previously found in mice and dogs cannot be accounted for by the failure of ricin to reach the bone marrow. Part of the ricin in the tissues was present in the form of free chains, the highest fraction being present in the liver. In this tissue both the free A-chains and those present in whole ricin were found to be modified. However, the modified A-chains had retained their full capacity to inhibit protein synthesis in vitro. In cancer patients, toxicity appeared at about the same initial serum levels as in the mice, supporting the view that mouse data have a good predictive value for man. At each dose level the individual variations were modest, a finding that is important for eventual clinical use of this potent drug.

Lack of correlation between natural killer activity and tumor growth control in nude mice with different immune defects

To elucidate the in vivo role of natural killer (NK) cells, the growth of several murine and human tumors was studied in four variants of athymic, nude mice with different levels of NK activity. Beige-nude mice, homozygous for both the beige and the nude genes, had very low levels of NK activity, and their response to the B-cell mitogen, bacterial lipopolysaccharide, was lower than that of high-NK, adult NIH nude mice. Young and adult NIH nudes had different NK levels and showed different response in assays for K-cell, T-cell, and B-cell activity. The B-cell-defective NIH-II mice had slightly lower NK levels than adult NIH animals, but much lower response in the antibody-dependent cell-mediated cytotoxicity assay. No correlation was found between host NK activity and the s.c. growth of various human (LOX, CEM, K562) and murine (YAC-1) tumor cells. Low NK activity was not associated with increased lung colony formation in a metastasis model using i.v.-injected human (LOX) and murine (B16F10) melanoma cells. No relationship was found between host NK activity and the rate of elimination of i.v.-injected 5-iodo-2'-deoxyuridine-labeled LOX, B16F10, and YAC-1 cells from lungs, liver, or spleen. The results fail to support the view that NK cells exert significant direct effects on tumor cells in vivo.

Reduced antineoplastic activity in mice of cisplatin administered with high salt concentration in the vehicle

The effect of high NaCl concentration in the vehicle on the toxic and antineoplastic activities of cisplatin [cis-dichloro-diammineplatinum(II)] (CDDP) was reinvestigated in mice. The toxicity, as measured by the survival of mice given CDDP iv, was reduced by 50-60% when the NaCl concentration in the vehicle was raised from 0.9 to 4%. In ascitic P388 leukemia the antineoplastic activity of CDDP given ip was not reduced significantly. However, in all other systems studied the antitumor activity was reduced when the CDDP was dissolved in high NaCl solution. The tumor models studied included systemic P388 and L1210 leukemias, Lewis lung carcinoma, and 5 human tumor xenografts. The human tumors were studied by the subrenal capsule assay. In the case of a malignant melanoma and a malignant fibrous histiocytoma, the effect also was demonstrated in the subcutaneous nude mouse model. In one of the malignant melanomas a 50% increase in the CDDP dose did not compensate for the reduced antitumor activity caused by the high NaCl concentration in the vehicle. These results, which stand in contrast to current views, question the experimental basis for the use of high-NaCl vehicles in the "high-dose" CDDP regimens.

Immune characteristics of the beige-nude mouse. A model for studying immune surveillance

Double immunodeficient mice, homozygous for both the beige and the nude genes, were developed on C57BL/6N-beige and N:NIH(S)-nude background through five steps of mating. The animals were healthy and had a life span comparable to that of regular nude mice. Beige and regular nude mice showed a significant difference in natural killer (NK) activity, with mean +/- SE values of 6 +/- 1.0% and 25 +/- 2.8%, respectively. The response to the T-cell mitogens phytohaemagglutinin and concanavalin A was similar in beige and NIH mice of the same genotypes for the nude genes. Both groups of nu/nu mice also showed some response in these assays, as well as in the plaque-forming cell assay, indicating that nude mice are not completely devoid of functional T or T-like cells. The nude genes were found to increase the response to the B-cell mitogen lipopolysaccharide in both the beige and NIH animals, whereas the response was reduced in mice of the beige genotypes. The viable, low NK beige-nude mice here reported on may be a valuable tool for studying factors involved in host defence against tumours. Interestingly, no spontaneous tumours have so far been observed in such mice.

Antigen expression in human melanoma cells in relation to growth conditions and cell-cycle distribution

The expression of a melanoma-associated antigen, recognized by the monoclonal antibody 9.2.27, has been studied in the human FME melanoma cell line, grown as a monolayer under various conditions in vitro and as tumours in athymic mice. Two-parameter flow cytometric measurements of DNA and immunofluorescence showed that the antigen expression was uniform throughout the cell cycle. Highest expression of the melanoma-associated antigen was found in cultures in which the medium was frequently renewed (4 times in 6 days) and no contact inhibition was present. In comparison, the antigen expression of cells subjected to medium starvation (6 days without medium renewal) was reduced to 44%, and the cell size, as measured by forward angle light scatter, to 82% of the values found for optimally growing cells. Intermediate conditions, such as scanty medium supply or contact inhibition of growth, gave smaller effects. Two-parameter measurements of cell size and immunofluorescence demonstrated a positive correlation between cell size and antigen expression. The reduction in cell size, therefore, could explain part, but not all, of the reduction in antigen expression on medium-starved and contact-inhibited cells. The antigen expression on FME cells grown as xenografts in athymic mice varied between individual tumours and was lower than on cells from optimal in vitro growth. Due to smaller cell size, the antigen density was, however, comparable to that on cells from optimal in vitro growth.

Phase I study of the plant protein ricin

A Phase I study was carried out with ricin, a plant toxin acting by inhibiting protein synthesis, on 54 cancer patients with advanced disease. Ricin was given as i.v. bolus injections every two weeks at dose levels ranging from 4.5 to 23 micrograms/sq m of estimated body surface area. Ricin was well tolerated at doses up to 18 to 20 micrograms/sq m. At these levels and at higher levels, flu-like symptoms with fatigue and muscular pain appeared and, in some patients, nausea and vomiting occurred also. No myelo-suppression was seen. Antibodies to ricin were detected in serum after two to three ricin injections. Ricin was eliminated from blood according to first order kinetics. At each dose level, the plasma concentrations, as well as the side effects, showed only minor differences between patients. The highest dose given, 23 micrograms/sq m, gave plasma concentrations twice those found previously to be therapeutically effective in tumor-bearing mice. Of 38 evaluable patients, one patient with lymphoma had a partial response. Stable disease was observed in four patients with renal cancers, in two with soft tissue sarcomas, and in one patient each with mesothelioma, thyroid, and rectal cancer. A dose of 23 micrograms/sq m is recommended for Phase II trials of ricin.

Antibody formation against the cytotoxic proteins abrin and ricin in humans and mice

Antibody formation may limit the therapeutic use of cancerostatic proteins. To study the significance of antibody formation against abrin and ricin, highly sensitive ELISA procedures for determination of anti-abrin and anti-ricin were developed. In mice treated weekly with therapeutic doses of ricin, antibodies appeared after 2-3 weeks and then rose rapidly, whereas after abrin treatment the antibody formation was slower. Ricin A-chain was found to be more immunogenic than either intact ricin or human serum albumin (HSA). Cyclophosphamide inhibited the antibody response to both abrin and ricin and a combination of cyclophosphamide and prednisolone totally inhibited both anti-abrin and anti-ricin formation during the 6-week observation period. In mice treated weekly with HSA, abrin treatment strongly reduced the anti-HSA formation, showing that abrin has an immunosuppressive effect which appeared to be stronger than that of cyclophosphamide. The existence of circulating antigen-antibody complexes could be demonstrated in the sera of toxin-treated mice by precipitation with polyethyleneglycol, whenever antibodies were detectable with ELISA. The life-span of animals given lethal ricin doses was appreciably enhanced in animals having antibody levels in excess of 10-20 ng/ml. In cancer patients treated i.v. every second week with therapeutic toxin doses, the 10-20 ng/ml levels of anti-ricin and anti-abrin were reached 6-8 weeks and 7-10 weeks after the first injection of ricin and abrin, respectively. The data indicate that the effective therapeutic use of abrin and ricin as single agents may be limited to these time frames, but that the period of effective use may be substantially prolonged if the toxins are given together with conventional cytostatic agents having immuno-suppressive activity.