Development of ELISAs for quantification of HMFG1-specific human anti-mouse IgG and IgM antibodies.

The aim of this study was to develop and validate ELISAs for quantification of HAMA-IgM and HAMA-IgG in serum of patients with ovarian cancer who enrolled in a large international randomized phase III trial of intraperitoneal Yttrium-90-labeled HMFG1 murine monoclonal antibody therapy. The capture antibody of these 2 assays was the murine antibody HMFG1, while mouse anti-human IgM-HRP or mouse anti-human IgG(Fc)-HRP served as tracer antibodies. A pool of HAMA-positive serum samples was used to prepare a series of assay standards and another pool served as reference preparation. The analytical sensitivity of the HAMA-IgM assay was 2.5 arbitrary units per mL (AU/mL) and 4.7 AU/mL for the HAMA-IgG ELISA. Diluted serum samples showed good parallelism with the HAMA-IgM and HAMA-IgG standard dose-response curves. Within-assay coefficient of variation was 7.5% for HAMA-IgM and 6.5% for HAMA-IgG. Between-assay variation was 14.2% for HAMA-IgM and 15.3% for HAMA-IgG. The developed HAMA-IgM and HAMA-IgG ELISAs show satisfactory reliability criteria (sensitivity, parallelism and precision) and are suitable for monitoring of HAMA-IgM and HAMA-IgG responses in ovarian cancer patients. These ELISAs will be used to monitor the development of HAMAs in patients who received radioimmunotherapy with murine HMFG1.

Characterisation and internalisation of recombinant humanised HMFG-1 antibodies against MUC1.

The humanised HMFG-1 immunoglobulin has been extensively developed as a clinical immunotherapeutic agent for MUC1 expressing tumours. We have constructed a single-chain Fv (scFv) and Fab fragment from this antibody and shown that both these species retain their specificity for MUC1. The scFv was less stable and less soluble than the Fab. Detailed analyses of the binding kinetics of the whole IgG and Fab fragment show that the affinity for MUC1 synthetic peptides is low (approximately 100 nM for the IgG and 10 muM for the Fab), with particularly low but similar dissociation rate constants (0.031-0.095 s(-1)). Binding to native antigen on the cell surface is over two orders of magnitude better. Confocal immunofluorescence microscopy shows that both the IgG and Fab are internalised rapidly (the IgG is internalised within 15 min) and colocalise to early endosomes. This work provides an appreciation of the binding, internalising and trafficking kinetics, important for the development of future therapeutics based on this antibody.

Impact of antibody framework residue VH-71 on the stability of a humanised anti-MUC1 scFv and derived immunoenzyme.

Anti-MUC1 single-chain Fv (scFv) fragments generated from the humanised antibody huHMFG1 had adequate antigen-binding properties but very poor stability irrespective of the applied linker or domain orientation. Mutagenesis of heavy-chain framework residue V(H)-71, previously described as a key residue for maintaining the CDR-H2 main-chain conformation and thus important for antigen binding, markedly stabilised the scFv while having only a minor effect on the binding affinity of the molecule. Because of its improved stability, the engineered fragment exhibited immunoreactivity with tumour cells even after 7 days of incubation in human serum at 37 degrees C. It also showed, in contrast to the wild-type scFv, a concentration-dependent binding to the target antigen when displayed on phage. When fusing the scFv to the recombinant ribonuclease rapLRI, only the fusion protein generated with the stable mutant scFv was able to kill MUC1(+) tumour cells with an IC(50) of 80 nM. We expect this novel immunoenzyme to become a promising tool for the treatment of MUC1(+) malignancies.

Source of oncofetal ED-B-containing fibronectin: implications of production by both tumor and endothelial cells.

ED-B fibronectin (FN) is a FN isoform derived from alternative splicing of the primary transcript of a single gene. Its expression on tumor stroma and neoformed tumor vasculature and its absence, with few exceptions, in normal adult tissues imply a prognostic and diagnostic value for ED-B FN. We investigated the location and source of ED-B FN because this will be of importance both in understanding its role in tumor development and in designing strategies to target this molecule. We have confirmed that ED-B FN is expressed in the majority of breast and colorectal carcinoma tissue samples, with strong immunohistochemical staining around the tumor cells and in the tumor stroma. No staining of tumor neovasculature was seen. ED-B FN is produced by a range of tumor and endothelial (both primary and transformed) cell lines, as detected by reverse transcription-PCR, but is not expressed at the plasma membrane. Strong expression of human ED-B FN is seen in tumor xenografts. These data indicate that neoplastic cells can act as the source of ED-B FN in tumors. The lack of cell surface expression on tumor cell lines has clear implications for the design of therapeutic strategies which target this molecule.

Imaging of metastatic melanoma utilising a technetium-99m labelled RGD-containing synthetic peptide.

Integrins are cell-surface glycoproteins found in different forms on all cells except erythrocytes. Integrins bind to cell adhesion molecules and to proteins found in the extracellular matrix. A tripeptidic sequence Arg-Gly-Asp (RGD) is often the primary site of recognition by integrins which are expressed on tumour cells and are responsible for tumour invasion and metastasis. A synthetic decapeptide designated alpha P2 containing two RGD sequences radiolabelled with technetium-99m was used to image malignant melanoma in vivo. Fourteen patients previously diagnosed with metastatic melanoma underwent gamma camera imaging 20-180 min following intravenous administration of the radiolabelled synthetic decapeptide alpha P2. Six out of eight (6/8) of the lymph node metastases (75%) and all other neoplastic sites (11 sites) were successfully imaged, with the exception of three sites in the mediastinal area which were not positively imaged. In two cases there was false positive uptake in the rounded pigmented areolar/nipple area. In three cases (seven sites) the peptide scan confirmed the absence of disease in suspected lesions (true-negative). The synthetic peptide was rapidly removed from the circulation by filtration through the kidneys and excretion in the urine. No toxicity or adverse events were recorded. Radiolabelled alpha P2 peptide, which binds specifically to adhesion molecules on tumours, can be used for the in vivo detection of neoplastic metastases.

Analysis of the kappa light chain variable region in multiple myeloma.

The study of immunoglobulin heavy chain gene rearrangements in multiple myeloma has revealed extensive divergence from the germline sequences, but no intraclonal diversity with disease evolution. Our study investigated the state of the rearranged kappa light chain variable region (V kappa) gene segments as well as abortive V kappa family gene usage in cases of multiple myeloma expressing lambda light chain. We studied 11 cases of kappa and five cases of lambda light chain-expressing multiple myeloma. Total cellular RNA was extracted from the bone marrow of patients with overt disease and subjected to reverse transcription-polymerase chain reaction (RT-PCR) analysis to amplify clonally rearranged variable region sequences. Direct nucleotide sequencing by the dideoxy-chain termination method was performed on the RT-PCR products. We did not observe preferential usage of certain V kappa gene families. Mutation frequencies of the V kappa segments varied in number. In the majority of cases, extensive somatic mutations occurred within the complementarity determining regions (CDRs) of V kappa, whereas only a limited degree of divergence from the germline was observed in others. In all cases studied. replacement mutations tended to cluster in the CDRs, a finding compatible with an antigen-driven somatic hypermutation process. In 3/5 cases of lambda light-chain expressing multiple myeloma, abortively rearranged V kappa gene segments were amplified from genomic DNA; in two cases a non-templated nucleotide insertion rendering the V kappa sequences out-of-frame was observed, and in the third a stop codon was identified in the open reading frame of the V kappa sequence. Somatic mutations were observed in all cases of abortive V kappa genes studied; however, their distribution does not suggest selection by antigen. We conclude that somatic mutations observed in the V kappa regions of myeloma cells are of variable extent and suggest operation of the antigen selection process. Lack of or minimal somatic hypermutation in a few cases may be in some way implicated in the biological heterogeneity of the disease.

Structure-function correlation and biostability of antibody CDR-derived peptides as tumour imaging agents.

Based on the CDR3 V(H) sequence of a monoclonal antibody (ASM2) raised against epithelial cancer cells, the synthetic peptide YCAREPPTRTFAYWG (EPPT1) has been found to have an appreciable affinity (Kd = 20 microM) for the deglycosylated mucin-derived peptide antigen YVTSAPDTRPAPGST (PDTRP). The technetium-radiolabelled form of this peptide has been found to be a good tumour-imaging candidate for diagnosis of breast carcinoma. Several EPPT1 peptide analogues were synthesised. A differential biostability was obtained blocking the end groups of EPPT1. The susceptibility to proteolytic degradation was significantly decreased for the C-amidated form of EPPT1 than the N-acetylated form. Using resonant mirror biosensor technique, the EPPT1 analogues were classified as active and non-active peptides according to their PDTRP-binding properties. The binding of EPPT1 to PDTRP in free solution was also determined unambiguously by CD spectroscopy. CD spectra of both active and non-active peptides showed the presence of irregular conformations in H2) and SDS above cmc. In TFE, significant degree of ordered conformations of alpha-helix or beta-turn type were induced but did not correlate well with their binding properties. In SDS below cmc a conformational difference was observed between the active and non-active peptides. The active peptides exhibited CD spectra of aggregation of beta-strand type whilst the non-active showed CD spectra similar to those in H2O and SDS above cmc, critical micelle concentration. A good correlation between the extended conformation of beta-strand type and the binding affinity of the active peptides suggests this conformation as the binding feature of the EPPT tumour-imaging peptides. These information are vital for the design of novel EPPT analogues. Any modification to improve binding affinity must retain the ability of the peptides to adopt the extended conformation of beta-strand type.

Breast cancer imaging with radiolabelled peptide from complementarity-determining region of antitumour antibody.

Specific tumour imaging with radiolabelled monoclonal antibodies has been extensively investigated. Although some success has been reported, there are many limitations due to the slow kinetics, poor extravasation, catabolism by the reticuloendothelial system, and non-specific uptake of macromolecules such as antibodies. We have tried to overcome some of the problems associated with monoclonal antibodies while retaining their specificity by using an antibody-derived synthetic peptide. A synthetic pentadecapeptide (alpha M2) derived from the third heavy-chain complementarity-determining region (CDR-3H) of a tumour-associated monoclonal antibody was produced and shown to retain its specificity against the pan-carcinoma cell-surface antigen, polymorphic epithelial mucin, detected by the parent antibody. The peptide was radiolabelled with technetium-99m and injected intravenously to image malignant lesions in 26 women with primary, recurrent, or metastatic breast cancer. Visualisation of breast tumours and their metastases was obtained shortly after administration of alpha M2, and was optimum by 3 h. Overall, 57 (77%) of 74 sites were visualised. Successful imaging was achieved in 14 of 15 primary tumour sites and all of eight local recurrences. Five of six metastases in the opposite breast, eight of 15 metastatic axillary lymph nodes, and all of six metastatic supraclavicular lymph nodes were imaged. Metastatic sites in the lungs, mediastinum, chest wall, and liver were poorly visualised because of background cardiac blood pool. alpha M2 detected small lesions ( < 2 cm) as efficiently as larger ones. The peptide was rapidly (3 h) cleared from the circulation. No acute or chronic adverse reactions due to the alpha M2 were observed. Specific tumour targeting with the radiolabelled anticancer peptide alpha M2 offers new opportunities for breast cancer imaging and possibly therapy.

Generation of monoclonal antibodies utilizing the host anti-idiotypic network.

In accordance with Jerne's idiotypic network theory, it was attempted to generate antibodies utilizing the host's anti-idiotypic reactions. A murine anti-tumour monoclonal antibody was administered to rats; we examined the possibility of whether the animals would develop anti-idiotypic antibodies, the paratope of which would be the "internal image" of the tumour associated antigen that the original murine monoclonal antibody recognizes. These anti-idiotypic antibodies would generate a second generation of anti-idiotypic antibodies, which would have the same specificity as the immunogen murine antibody. It was found that, while all rats had no anti-tumour response prior to immunization with the anti-tumour murine monoclonal antibody, they all developed antitumour antibodies following administration. The latter observation came as a consequence of the development of anti-mouse immunoglobulin and anti-idiotypic antibodies. The animals were sacrificed and their spleen cells were fused with myeloma cells for the production of monoclonal rat anti-tumour antibodies. Finally, a monoclonal antibody was selected which appeared to recognize the same tumour-associated antigen as the originally administered murine antibody, as shown by cell-binding and competition assays, as well as immunohistochemistry. This new approach allows for "replication" of polyclonal and monoclonal antibodies without the need of the antigen and it may also serve as an immunotherapy strategy for the augmentation of the antitumour immune response of patients receiving monoclonal antibody treatment.

Activation of cellular immunity after intracavitary monoclonal antibody therapy of ovarian cancer.

BACKGROUND: Murine monoclonal antibodies (MoAbs) are now used for targeted tumor therapy. A major obstacle in their successful application is the development of a humoral antiglobulin response, which limits the use of repeated cycles of therapy. The cellular aspects of that response are not well understood. METHODS: Fifteen patients who had one (12 patients) or two (3 patients) courses of MoAb treatment, 13 age-matched patients with the same histologic types of tumors who had not received MoAbs, and 4 healthy control subjects were studied. Peripheral blood mononuclear cells (PBMCs) were obtained and tested for the ability of T-cells to proliferate in vitro in the presence of the MoAb administered for therapy (HMFG1), a control antibody (11.4.1), and, in some cases, their F(ab')2 fragments. In addition, PBMCs from these patients were phenotyped after in vitro MoAb stimulation with antibodies against CD2, CD3, CD4, CD8, CD16, CD20, CD25 (interleukin-2 receptor [IL-2R]), CD45RA, and UCHL1, and the production of interleukin-2 (IL-2) was evaluated by the CTLL-2 bioassay. RESULTS: A dose-dependent in vitro T-cell proliferation was observed in 13 of the 15 patients after MoAb therapy. This was not observed in the pretherapy group of patients or healthy control subjects. The mean stimulation index (SI) in the posttherapy group was significantly higher than that of the pretherapy patients and that of healthy control subjects (P = 0.007). When the in vitro T-cell proliferative responses of these patients were measured in the presence of HMFG1 MoAb (IgG1) and 11.4.1 MoAb, there was no significant difference in the mean SI for HMFG1 versus 11.4.1 for the whole group of treated patients (P = 0.67). A significant increase in the mean SI was observed in the presence of HMFG1 over 11.4.1 and their F(ab')2 fragments (P = 0.02) in patients treated twice. A significant increase in the percentage of cells expressing IL-2R was observed after in vitro MoAb stimulation. CD4+ lymphocytes, particularly the CD4+/UCHL1+ memory, the CD4+/IL-2R+ subpopulation, and the CD4/CD8 ratio, increased in all the cases studied after MoAb stimulation, where B-cell and natural killer-cell numbers remained relatively constant (< 2-3%). A sixfold increase was found in the production of IL-2 in PBMC supernatants after MoAb stimulation. CONCLUSIONS: Mouse MoAbs administered to patients with cancer can lead to the generation of T-cells, which can recognize these MoAbs as antigens and therefore refocus the host's cellular immune response against the targeted tumor. The main proliferating population appears to be CD4+ T-lymphocytes, which after stimulation can release IL-2. Multiple treatments may lead to the generation of T-cells with specificity for the idiotypic component of the administered MoAb.

Effects of phorbol esters and cytokines (interleukin-2,-4, and -6) on the proliferation and surface phenotype of Epstein-Barr virus immortalised human B lymphocytes.

Epstein-Barr virus (EBV)-induced in vitro infection of peripheral blood mononuclear cells (PBMCs) leads to a polyclonal proliferation and immortalisation of B lymphocytes. In the present study we determined the effects of three different cytokines, interleukin-2 (IL-2), interleukin-4 (IL-4) and interleukin-6 (IL-6), and the tumour promoting phorbol ester 12-O-tetradecanoyl-phorbol-13-acetate (TPA) on EBV-immortalised B lymphocytes. These factors have known activities on normal B cells. IL-4 and IL-6 increased significantly EBV-B cell proliferation after 3 and 5 days of culture, where IL-2 had no effect. The effect of IL-4 and IL-6 on EBV-B cells was abolished after pre-incubation with anti-IL-4 and anti-IL-6 neutralising antisera, respectively. TPA induced a dose dependent inhibition of proliferation both in serum free and 10% fetal calf serum (FCS) supplemented culture medium. Combinations of TPA and interleukins did not restore lymphoblastoid cell proliferation to background levels. All possible combinations of the three cytokines showed no synergistic or antagonistic effect on proliferation. TPA induced significant phenotypic changes of EBV immortalised B lymphocytes, by increasing IL-2 receptor (IL-2R) expression and decreasing CD20 and CD23 antigen expression. Other B cell differentiation antigens; HLA-DR, CD19, and transferrin receptor (CD71), did not demonstrate significant changes. A dose dependent inhibition of CD21 and increase in CD22 expression was observed in 2 out of 3 lymphoblastoid cell lines tested.

Development of humoral immune responses against a macrocyclic chelating agent (DOTA) in cancer patients receiving radioimmunoconjugates for imaging and therapy.

The development of stable immunoconjugates by the advent of macrocyclic metal chelating agents (DOTA) has enabled us to study the ability of 111In-DOTA-labeled monoclonal antibodies to detect tumor lesions in a pilot radioimmunolocalization study, as well as to evaluate the kinetics, toxicity, and efficacy of i.p. administered 90Y-DOTA-labeled murine monoclonal antibody in a Phase I/II clinical trial of advanced ovarian cancer. The development of serum sickness-like reactions in three of six treated patients, in the absence of previous monoclonal antibody administration, led us to study the potential immunogenicity of the new chelate. Six patients with ovarian cancer received 25 mg of HMFG1 monoclonal antibody coupled with 90Y-DOTA (doses of radioactivity, 15 to 25 mCi), administered i.p. Eight patients with various malignant tumors received low doses (220 micrograms to 1 mg) of monoclonal antibodies, labeled with 111In-DOTA, i.v. for imaging studies. Using a solid-phase enzyme-linked immunosorbent assay method, the immunogenicity of DOTA was evaluated. Serial dilutions of patients' sera, before and after imaging or therapy with DOTA-coupled monoclonal antibodies, as well as sera from patients who did not receive DOTA-coupled antibody, were screened on enzyme-linked immunosorbent assay plates coated with human serum albumin (HSA), HSA-2-iminothiolane, and HSA-2-iminothiolane-benzyl-DOTA. All patients treated with i.p. monoclonal antibody developed anti-DOTA antibodies. Four of eight patients who received i.v. "imaging" doses of DOTA-coupled monoclonal antibody developed antibodies against DOTA. The levels of anti-DOTA response correlated with the amount of injected radioimmunoconjugate (r = 0.889, P less than 0.001). None of the patients who received DOTA-coupled antibody had detectable antibodies against the macrocycle before immunoconjugate administration. We then addressed further the restriction of the immune response against the macrocycle. We found that there was no or very low response against the aromatic ring attached to DOTA. Most, if not all, of the immune response is directed against the DOTA ring structure. Affinity purification of anti-DOTA antibody from serum enabled quantitation of these antibodies in the serum of patients. An inverse, statistically significant correlation was observed between the percentage of binding inhibition of a patient's serum to DOTA, by HSA-2-iminothiolane-DOTA (100 micrograms/ml) and the level of anti-DOTA immunoglobulin in the serum.(ABSTRACT TRUNCATED AT 400 WORDS)

Patients receiving murine monoclonal antibody therapy for malignancy develop T cells that proliferate in vitro in response to these antibodies as antigens.

Peripheral blood mononuclear cells (PBMCs) were obtained from patients receiving radioactive murine monoclonal antibody (MAb) therapy for malignant epithelial tumours, as well as normal controls, and were tested for the ability of T cells to proliferate in vitro in the presence of the MAb administered for therapy (HMFG1), and another isotypically matched antibody of irrelevant specificity (11.4.1). We studied 13 patients who had one (ten patients) or two (three patients) courses of MAB treatment, 11 age matched patients with the same histologic types of tumours, that had not received MAbs, and four normal controls. There was a consistent dose dependent in vitro T cell proliferation in 11 of the 13 patients after MAb therapy. This was not observed in the pre-therapy group of patients or normal controls, where the T cell proliferative responses remained baseline. The mean stimulation index (S.I.) in the post-therapy group was significantly higher than that of the pre-therapy patients and that of normal controls. When the in vitro T cell proliferative responses of these patients were measured in the presence of HMGF1 MAb (IgG1) and an isotypically identical, but idiotypically unrelated 11.4.1 MAb (IgG1), there was no statistically significant difference in the mean S.I. For HMFG1 vs 11.4.1 for the whole group of treated patients. When patients were separated into those who received one and those who received two MAb treatments, a significant increase in the mean S.I. was observed in the presence of HMFG1, in the group of patients receiving two treatment courses, suggesting the generation of T cells with specificity for the idiotypic component of the administered murine immunoglobulin. In order to further characterise these in vitro cellular responses we incubated PBMCs with and without an optimal concentration of the MAb (100-300 micrograms ml-1), as defined by the proliferation assay, and compared the differences in cell subpopulations. A significant increase in the percentage of cells expressing interleukin-2 receptors (IL-2R) was observed after MAb stimulation. The percentage of CD4+ lymphocytes and the CD4/CD8 ratio increased in all the cases studied, after MAb stimulation, where the percentages of B cells and NK cells remained relatively constant at less than 2-3% of the total population. We therefore conclude that murine MAbs administered to patients with cancer can lead to the generation of T cells which can recognise these MAbs as antigens when presented appropriately in vitro. The main proliferating population appears to be T helper CD4+ lymphocytes which following stimulation can release interleukin-2 leading to the expression of high levels of IL-2R.

Immunolocalisation of testicular tumours using radiolabelled monoclonal antibody to placental alkaline phosphatase.

Tumour associated monoclonal antibody against placental alkaline phosphatase (H17E2) was radiolabelled with Indium-111 and Iodine-123 and administered intravenously in 33 patients with primary and/or metastatic testicular tumour, as well as in 8 patients who were in complete remission after surgical excision of the tumour. The presence of a tumour was confirmed and correlated well with conventional diagnostic techniques and, in addition, the antibody scan revealed the presence of active disease in 2 patients with negative conventional imaging and with elevated serum markers. In addition, in one patient the CT produced a false positive result where the antibody scan was negative. Finally, the absence of tumour was confirmed in all 8 cases of patients in complete remission. All patients studied with Indium-labelled antibody had observable concentrations of the radiolabel in the liver (estimated to be approximately 30% of the administered dose), as well as in the kidneys and spleen. The patients studied with the Iodine-123 labelled antibody had observable concentrations in the thyroid gland and the stomach. The best images were seen at 48 and 24 hrs after the Indium and Iodine radiolabelled antibody respectively. No human anti-mouse antibody was detected in any of our patients, even in those who received 2 and 3 administrations, with the highest amount of administered protein being 800 micrograms. No toxicity was encountered in any of our patients in 4 months of follow-up. This method may be of clinical value in patients with testicular neoplasma and represents a new addition to current imaging techniques. A positive scan indicates the definite presence of a tumor.(ABSTRACT TRUNCATED AT 250 WORDS)

The role of humoral and cellular immunity in patients developing human anti-murine immunoglobulin antibody responses after radioimmunotherapy.

Epstein-Barr virus (EBV)-immortalised B cell lines were established from patients receiving multiple administrations (two or more) of radiolabelled murine monoclonal antibodies for the treatment of ovarian cancer. Cells secreting anti-id2 Ig, an immunoglobulin with binding specificities comparable to the administered murine monoclonal antibody, were isolated by using magnetic beads coated with tumour-associated antigen, incubated with the cells and concentrated with a magnetic particle concentrator. Cross-linking of the immunoglobulin receptors by the antigen-coated beads appears to stimulate proliferation, resulting in increased secretion of the human anti-tumour-associated antigen antibodies. The T cell responses were studied and it was found that monoclonal antibody therapy appears to lead to an increase in the population of T cells committed to proliferate in response to both specific antigen and non-specific mitogens. Multiple administrations of monoclonal antibody induce the generation of T cells which proliferate in vitro following stimulation with murine antibodies. The relevant (administered) monoclonal antibody induces higher proliferation rates than an idiotypically unrelated antibody of the same isotype, indicating the generation of idiotypically restricted T cell responses in these patients.

Antibody guided diagnosis and therapy of brain gliomas using radiolabeled monoclonal antibodies against epidermal growth factor receptor and placental alkaline phosphatase.

Twenty-seven patients with brain glioma were scanned using 123I-labeled monoclonal antibodies against epidermal growth factor receptor (EGFR1) or placental alkaline phosphatase (H17E2). Successful localization was achieved in 18 out of 27 patients. Eleven out of 27 patients were also studied using a nonspecific control antibody (11.4.1) of the same immunoglobulin subclass and observable tumor localization was also achieved in five patients. The specificity of targeting was assessed by comparing images obtained with specific and nonspecific antibodies and by examining tumor and normal tissue biopsies after dual antibody administration. Ten patients with recurrent grade III or IV glioma who showed good localization of radiolabeled antibody were treated with 40-140 mCi of 131I-labeled antibody delivered to the tumor area intravenously (n = 5) or by infusion into the internal carotid artery (n = 5). Six patients showed clinical improvement lasting from 6 mo to 3 yr. One patient continues in remission (3 yr after therapy), but the other five who responded initially relapsed 6-9 mo after therapy and died. No major toxicity was attributable to antibody-guided irradiation. Targeted irradiation by monoclonal antibody may be clinically useful and should be explored further in the treatment of brain gliomas resistant to conventional forms of treatment.

Intraperitoneal radioimmunotherapy for ovarian cancer: pharmacokinetics, toxicity, and efficacy of I-131 labeled monoclonal antibodies.

Thirty-six patients with ovarian cancer were treated with intraperitoneal I-131 labeled monoclonal antibodies to tumor associated antigens. The activity of I-131 administered was increased from 20 mCi to 158 mCi and the pharmacokinetics and toxicity evaluated. Five patients who had developed HAMA (Human Antimouse Antibodies) were retreated, and the pharmacokinetics and toxicity of the first and second treatment compared. Patients receiving their first therapy (HAMA negative), had a maximum of 25% (range 19.8-39.8%) of the injected activity in their circulation. This was accompanied by severe marrow suppression at I-131 activities over 120 mCi. The 5 HAMA positive patients had only 5% injected activity in the systemic circulation (range 3.8-6%), with rapid urinary excretion and neglible marrow suppression. In 31 patients with assessable disease there were no responses in 8 patients with gross disease (nodules greater than 2 cms), partial responses in 2 out of 15 patients with nodules less than 2 cms, and complete responses in 3 out of 6 patients with microscopic disease. The non specific radiation dose to the peritoneal cavity was estimated to be less than 500 cGy by lithium fluoride TLD, and could not be expected to account for the responses seen.

Development of anti-idiotypic antibodies against tumour antigens and autoantigens in ovarian cancer patients treated intraperitoneally with mouse monoclonal antibodies.

Repeated intraperitoneal administration of therapeutic amounts of radiolabelled (iodine-131) murine monoclonal antibodies leads to the development of an immune response in the recipient, part of which is directed against the variable region (idiotype) of the administered antibody (anti-Id1 response). Human immunoglobulin purified from these patients inhibits binding of the original murine monoclonal antibody to its target tumour antigen and therefore represents an "internal image" of the tumour antigen. Furthermore, this study recorded the development of human antibodies that themselves bind to the tumour antigen, with a specificity identical or similar to that of the injected monoclonal antibody. These human antitumour antibodies are probably generated by way of the idiotypic network and confirm the existence of the idiotypic network. Accompanying this antitumour response the transient development of autoantibodies that react with connective tissue components of liver, kidney, spleen, and diaphragm was also observed.