Expression and interconversion of neuron-specific enolase in patient sera and extracts from small-cell lung cancer cells.

The isoforms of gamma-enolase were characterized in serum from patients with small-cell lung cancer (SCLC) and in extracts from SCLC cell lines and malignant melanoma tumor tissue. Large variations in the expression of the 3 gamma-isoforms of enolase were observed. These forms probably represent the homodimeric gamma gamma-enolase, the heterodimeric alpha gamma-enolase and the monomeric forms of gamma-enolase. Only the dimeric forms are enzymatically active. The predominant gamma-enolase in the cell lines is the heterodimeric alpha gamma-enolase. The SCLC cell lines can be divided into two groups: one with negligible gamma gamma-enolase expression and considerable amounts of the nonneuronal alpha alpha-enolase and a second group with a large fraction of gamma gamma-enolase concomitant with a low expression of alpha-enolase. Similar patterns are observed in tissue extracts from malignant melanoma. When changing buffer conditions by increasing the ionic strength and decreasing the Mg2+ concentration, interconversions between the isozymes occur. In contrast to the predominant alpha gamma-enolase in extracts from cell lines, the multiple forms of gamma-enolase in serum might be caused by a subunit exchange facilitated by the low Mg2+ concentration in plasma. However, there seems to be a stable equilibrium between the isoforms in undiluted patient serum. The induction of subunit exchange by perturbation in ionic strength and/or Mg2+ concentration indicates a need for caution when choosing diluents for use in assays for neuron-specific enolase.

Targeted therapy with immunotoxins in a nude rat model for leptomeningeal growth of human small cell lung cancer.

Metastasis to the central nervous system in patients with small cell lung cancer is not uncommon, and a fraction of the cases have leptomeningeal disease for which no effective therapy is available. To establish an experimental model for evaluation of new therapeutic approaches for such tumor lesions, 1 x 10(6) human H-146 cells were injected directly into the cerebrospinal fluid in the cisterna magna of nude rats. Small, superficial leptomeningeal tumors developed, consistently resulting in symptoms of central nervous system involvement after a mean latency of 20 days. The model was used to study the efficacy of intrathecal targeted therapy with immunotoxins. The monoclonal anti-carcinoma antibodies MOC-31 and NrLu10 and the growth factor transferrin were conjugated to Pseudomonas exotoxin A (PE), and 1 day after tumor cell inoculation instilled in the cisterna magna as a single bolus dose of 1.5 micrograms. The antibody conjugates, which were highly cytotoxic to target cells in a protein synthesis inhibition assay in vitro, increased the symptom-free latency by 35-46%. PE had no effect, reflecting a lower in vitro cytotoxicity and possibly also a down-regulation of transferrin-receptor expression in the meningeal H-146 tumors. Delayed or repeated treatment with MOC-31-PE was less effective than day 1 administration, whereas the addition of 10% glycerol to the injection solution increased the symptom-free period to 72%. The efficacy of MOC-31-PE is superior to reported effects obtained in similar models with other therapies, and the results support the development of this immunotoxin towards clinical evaluation in small cell lung cancer patients with leptomeningeal carcinomatosis.

Nude rat models for human tumor metastasis to CNS. Procedures for intracarotid delivery of cancer cells and drugs.

Models for hematogenous spread of human cancer to the central nervous system (CNS) were established by injecting human tumor cells into the internal carotid artery of nude rats. With 4 out of 10 cell lines, belonging to four different tumor types, metastases developed in all injected animals. Tumor growth manifested clinically as neurological symptoms which appeared after a median latency ranging from 19-87 days for the different tumors. The H-146 and DMS-273 small cell lung cancers and the LOX melanoma almost exclusively gave meningeal tumors, whereas with FEMX-I melanoma cells bone metastases in the skull dominated. For these tumor types a correlation was found between the capacity for experimental metastasis formation and the s.c. tumorigenicity. In agreement with clinical experience, none of the 2 sarcoma and 2 glioblastoma lines gave CNS metastases. With a modified microsurgical technique, allowing for repeated ipsilateral intracarotid injections, we analyzed the drug concentrations obtained in tumor and surrounding brain tissue after i.v. treatment with doxorubicin. The concentration in the LOX tumor reached therapeutic levels and was approximately 100 x higher than in normal brain tissue, both with and without intraarterial pretreatment with arabinose. In the same model, the tissue concentrations of 9.2.27-abrin immunotoxin 10 min after intracarotid injection were examined. Although the levels were low, a tumor to brain concentration ratio of up to 9 was achieved. The data demonstrate that clinically relevant tumor models can be established with the techniques described, and these models may successfully be used to evaluate the pharmacokinetics and effect of intravenous or intraarterial therapy.

Comparison of two antibody-based methods for elimination of breast cancer cells from human bone marrow.

Three monoclonal antibodies reactive with antigens abundantly expressed on human carcinoma cells were used to develop and compare the efficacy of immunotoxins (ITs) and immunobeads for purging breast cancer cells from bone marrow. ITs constructed as conjugates of the monoclonal antibodies and Pseudomonas exotoxin A showed high specific cytotoxicity against three breast cancer cell lines, inhibiting protein synthesis by 50% at concentrations of 4 x 10(-13) M to 1 x 10(-10) M. Tested in a reproducible clonogenic assay, two of the ITs used at a concentration of 0.1 microgram/ml killed > 5 log units of MCF7 cells, the maximal sensitivity for assessing cytotoxic effects, and 1.5 log of T-47D tumor cells. At 1 microgram/ml, each of the three ITs eliminated > 5 log of both cell lines. The immunobead procedure removed 2.0-4.1 log of tumor cells with one purging cycle and up to 6.0 log with two cycles. The mixture of the three ITs or immunobeads was not clearly superior in efficacy, compared to the use of individual molecules, probably reflecting an overlap in expression of the respective antigens in these cell lines. For both methods, the purging efficacy was not reduced when the tumor cells were admixed with normal bone marrow cells at a ratio of 1:10. The survival of colony-forming units, granulocyte/macrophage, was 49-86% with the immunobeads and 44-75% even at high concentrations (up to 2.5 micrograms/ml x 3) of the ITs. The results indicate that each of the two immunological methods can be safely used for effective elimination of tumor cells from the graft of breast cancer patients undergoing autologous bone marrow transplantation.

Eradication of small cell lung cancer cells from human bone marrow with immunotoxins.

The potential of autologous bone marrow transplantation to improve the treatment results for patients with small cell lung cancer (SCLC) may be limited by the presence of tumor cells in the graft. We constructed immunotoxins (ITs) involving 4 monoclonal antibodies and Pseudomonas exotoxin A and investigated the cytotoxicity of the ITs to H-146 SCLC cells in the presence and absence of normal human bone marrow (BM) cells. The Pseudomonas exotoxin A conjugate with the MOC-1 antibody, which recognizes an NCAM antigen, was inactive, as tested in a reproducible soft agar assay. Conjugates involving the monoclonal antibodies MOC-31, NrLu10, and MLuC1 killed about 3.5 log tumor cells at 0.1 microgram/ml and > 5.0 log at 1 microgram/ml. In the absence of BM cells, the combination of the 3 ITs was not superior to each IT used individually. However, when H-146 cells were admixed to nucleated BM cells at the ratio of 1:10, > 5 log tumor cell kill was obtained at a concentration as low as 0.1 microgram/ml of each IT. Survival of normal BM progenitor cells was only moderately reduced by the IT treatment, even in experiments in which the 3 IT3s were used at 2.5 micrograms/ml each. Freezing and thawing of the BM, as required in a clinical setting, reduced the colony-forming unit, granulocyte-macrophage, and colony-forming unit, granulocyte-erythroid-macrophage-megakaryocyte, by 30-60% in both treated and untreated cultures. We conclude that the use of a mixture of the 3 ITs provides a safe, rapid, and effective method for eradicating SCLC cells from BM used for autologous bone marrow transplantation following high-dose chemotherapy.

Effective removal of SCLC cells from human bone marrow. Use of four monoclonal antibodies and immunomagnetic beads.

High dose chemotherapy with autologous bone marrow transplantation (ABMT) has shown promise in several types of cancer. There is, however, a risk of transfusing contaminating tumour cells with the bone marrow cells, e.g. in patients with small cell lung carcinoma (SCLC). To eliminate SCLC cells from normal human bone marrow, four monoclonal antibodies reactive with SCLC cells were used with immunomagnetic beads in model experiments. With two cycles of immunomagnetic elimination the individual antibodies removed 2.5-4.4 log of H-146 tumour cells from a single cell suspension, as assessed in a highly reproducible soft agar assay. Different combinations of two antibodies were only marginally more effective than the individual MAbs, whereas 5-6 log removal was obtained with a combination of all four antibodies. The method was equally effective when the tumour cells were mixed with bone marrow cells at a ratio of 1:10. The immunomagnetic procedure did not significantly affect the survival of normal progenitor cells, assessed in CFU-GM and CFU-GEMM assays. The results indicate that the procedure safely and effectively can be used to eliminate tumour cells from the bone marrow in conjunction with ABMT in patients with SCLC.

Detection of bone marrow metastases in small cell lung cancer patients. Comparison of immunologic and morphologic methods.

An immunocytochemical method, involving four monoclonal antibodies (MAbs) previously selected for their specific binding to small cell lung cancer (SCLC) cells in human bone marrow, was used for detection of bone marrow metastases in 81 patients with diagnosed SCLC. This procedure was compared with two routine morphologic methods with regard to diagnostic efficiency and sensitivity. Bone marrow involvement was found in 26 patients (32%), one of which had limited disease according to conventional clinical criteria. Eight of the positive cases were exclusively diagnosed by immunocytochemistry, whereas the histologic and cytologic methods separately identified two patients each. Immunocytochemistry had a detection level of tumor cells in the mononuclear cell fraction of approximately 1-2%, whereas no patients with less than 10% immunocytologically detectable tumor cells were diagnosed by cytomorphologic examination of bone marrow aspirates. Evidence was obtained that the diagnostic efficiency of any method increased with the number of samples examined. Of the four MAbs used, the anti-NCAM antibody, MOC-1, labeled tumor cells in all immunologically positive patients, and in all but one of these patients all cytologically confirmed tumor cells were stained. The antibodies MOC-31, which recognize a cluster-2 antigen, and NrLu10 bound nearly all tumor cells in most cases, whereas MLuC1 only diagnosed tumor cells in a fraction of the patients. The results show that the immunocytochemical application of these antibodies is superior to morphologic techniques in detecting SCLC bone marrow metastases. Further use of the method might provide prognostically and therapeutically useful information.

Selection of anti-SCLC antibodies for diagnosis of bone marrow metastasis.

The binding profile of 17 monoclonal antibodies (MAbs) to small lung cancer cell lines and normal bone marrow and peripheral blood cells was studied by immunocytochemistry and flow cytometry. At antibody concentrations that stained PJD tumour cells, only four MAbs (MOC1, MOC31, NrLu10, 81A6) were devoid of cross-reactivity with normal cells, whereas significant binding to subtypes of bone marrow and blood cells was seen for 13 antibodies. For the eight most promising MAbs the binding to ten SCLC cell lines was moderate to strong in 47 MAb/cell line combinations, and low or insignificantly in 33 combinations. Three cell lines lacked antigen for all MAbs studied. Flow cytometry was significantly less sensitive than immunocytochemistry in assessing MAb binding to both normal and tumour cells. The antigen expression was for several MAbs higher in exponentially growing tumour cells than in cells in stationary growth phase. Seven of the MAbs, which originally showed low to moderate cross-reactivity with normal cells, were titrated down to the lowest concentrations at which they stained H-146 tumour cells with high levels of antigen expression. At these concentrations five (MLuC1, Oat-1, SM-1, NCC-Lu-243, LAM2) of the seven Mabs showed acceptably low binding to bone marrow cells. At optimal concentrations altogether four to nine of the 17 antibodies studied may be used to detect tumour cell involvement in bone marrow of SCLC patients.