Radiolabeled antibody therapy for squamous cell carcinoma of the head and neck.

Immunohistochemical staining for ferritin was performed on 11 human head and neck squamous cell carcinomas transplanted in nude mouse xenografts. Seven tumors were found to be positive. Using ferritin as a tumor antigen target, escalating doses of yttrium-90-labeled antiferritin antibodies were injected intravascularly into nude mice that were transplanted with a ferritin-positive human squamous cell carcinoma. Forty-five days after injection, the mean treated tumor size was 25.6% of control in the 100-microCi group, and 20.6% of control in the 200-microCi group. Ninety days after injection the mean tumor size was 27.5% that of control in the 100-microCi group and 31.7% in the 200-microCi group. Higher doses of radiation (300 and 400 microCi per mouse) caused death of most of the animals due to radiation toxicity. Presensitization of the animal, before antibody injection, with a bolus intraperitoneal injection of 7.5 mg of cisplatin per kilogram of body weight, resulted in further reduction in tumor size when compared with antibody alone or cisplatin alone. This study demonstrates that radioimmunotherapy with selected doses of yttrium-90-labeled antiferritin antibodies is effective against human head and neck squamous cell carcinoma xenografts in nude mice.

Targeting and therapy of human glioma xenografts in vivo utilizing radiolabeled antibodies.

Radiolabeled antibodies provide a potential basis for selective radiotherapy of human gliomas. We have measured tumor targeting by radiolabeled monoclonal antibodies directed against neuroectodermal and tumor-associated antigens in nude mice bearing human glioma xenografts. Monoclonal P96.5, a mouse IgG2a immunoglobulin, defines an epitope of a human melanoma cell surface protein and specifically binds the U-251 human glioma as measured by immunoperoxidase histochemistry. IIIIn-radiolabeled P96.5 specifically targets the U-251 human glioma xenograft and yields 87.0 microCi of tumor activity/g/100 microCi injected activity compared to 4.5 microCi following administration of 100 microCi radiolabeled irrelevant monoclonal antibody. Calculations of targeting ratios demonstrate the deposited dose to be 11.6 times greater with radiolabeled P96.5 administration compared to irrelevant monoclonal antibody. The dose found in normal organs is less than 20% of that in the tumor, further supporting specific targeting of the human glioma xenograft by this antibody. Monoclonal antibody ZME018, which defines a second melanoma-associated antigen, demonstrates positive immunoperoxidase staining of the tumor, but comparatively decreased targeting. To test the therapeutic potential of 90Y-radiolabeled P96.5 and ZME018, tumors and normal sites were implanted with miniature thermoluminescent dosimeters. Average absorbed doses of 3770 +/- 445 (SEM) and 645 +/- 48 cGy in tumor, 353 +/- 41 and 222 +/- 13 cGy in a contralateral control i.m. site, 980 +/- 127 and 651 +/- 63 cGy in liver, and 275 +/- 14 and 256 +/- 18 cGy in total body were observed 7 days following administration of 100 microCi 90Y-radiolabeled P96.5 and ZME018, respectively. Calculations of absorbed dose by the medical internal radiation dose method confirmed thermoluminescent dosimeter absorbed dose measurements. To test the therapeutic potential, tumor-bearing nude mice were given intracardiac injections of either buffer or 90Y-radiolabeled P96.5 or ZME018. Tumor regression was measured in 1 of 12, 9 of 10, and 12 of 12 compared to 0 of 10, 1 of 10, and 2 of 10 animals following administration of 50, 100, or 200 microCi 90Y-labeled P96.5 and ZME018, respectively. Average maximal decreases in tumor volume were 42.7 +/- 11.9 and 94.2 +/- 3.3% 28 and 58 days following 100 and 200 microCi 90Y-radiolabeled P96.5 administration, respectively. In contrast, no average decrease in tumor volume was noted following 50, 100, or 200 microCi 90Y-labeled ZME018.(ABSTRACT TRUNCATED AT 400 WORDS)

Yttrium-90 and iodine-131 radioimmunoglobulin therapy of an experimental human hepatoma.

Therapeutic trials were performed on the HepG2 human hepatoblastoma implanted s.c. in the athymic nude mouse. Animals were treated with polyclonal and monoclonal antiferritin and control antibodies labeled with either iodine-131 (131I) or yttrium-90 (90Y). Administration of 400 muCi of 131I-labeled polyclonal antiferritin or 300 muCi of 90Y-labeled polyclonal antiferritin significantly increased survival (P less than 0.001). There were no tumor cures with radiolabeled polyclonal antibody therapy. Animals treated with 200 or 300 muCi of 131I-labeled monoclonal antiferritin (QCI054) did not show increased survival compared to controls. Although 400 muCi of 131I-labeled QCI significantly prolonged survival, treatment resulted in no long-term survivors. Monoclonal antiferritin labeled with 90Y significantly prolonged survival of animals (P less than 0.001) at doses of 100, 200, or 300 muCi compared with untreated controls. Fifty % of the animals treated with 200 muCi and 75% of the animals treated with 300 muCi showed no evidence of disease at 140 days following treatment. Four hundred muCi of 90Y-labeled QCI proved toxic to the animals. Increased survival was accompanied by a decrease in tumor mitotic rate and increase in cellular polymorphism as determined by pathological examination. The radiation dose absorbed in the tumor correlated directly with tumor response following treatment. The absorbed dose in tumors for complete decay of the isotope ranged from 165 and 330 cGy at the periphery and center of small tumors for an administered activity of 200 muCi of 131I-labeled polyclonal antiferritin, to 7,573 and 12,400 cGy for 300 muCi of 90Y-labeled monoclonal antiferritin QCI.

Detection of specific anti-antibodies in patients treated with radiolabeled antibody.

Over 100 patients have received cyclic treatment with polyclonal 131I labeled anti-ferritin and anti-carcinoembryonic antigen (CEA) antibodies from different animal species (rabbit, pig, cynomolgous monkey, bovine, and baboon). Because survival was prolonged from original cyclic treatment, retreatment with original antibodies (recycling) became a necessary consideration. An assay using autoradiography of Ouchterlony gels, with diffusion of patients' sera against the varied radiolabeled antibodies, was developed to detect anti-antibody precipitin bands. Anti-antibody could be detected with a sensitivity to the 60 ng level. Sera from 35 patients given from 1 to 7 separate cycles (2 injections/week, total antibody 6 mg/cycle) of radiolabeled foreign antibody were studied for the production of anti-antibodies. Anti-antibodies were detected in 11 of 22 primary hepatoma patients studied, 3 of 4 intrahepatic biliary cancer patients, and 0 of 9 Hodgkin's disease patients. In all but two of the patients, the anti-antibodies produced were specific for the species used in the treatment of the patient. Eight patients were reinjected (recycled) with previously used antibodies and the presence or absence of precipitin bands correlated with the ability of these antibodies to deposit in the tumor or to be rapidly degraded. The importance of this assay is its simplicity, sensitivity, and the rapid detection of anti-antibody activity for patients requiring treatment with radiolabeled antibodies.

Ferritin concentration and 131I-antiferritin tumor localization in an experimental hepatoma.

Polyclonal 131I-labeled rabbit anti-rat ferritin was shown to specifically localize in the H4IIe rat hepatoma model. Tumor targeting was shown to be maximal in primary tumors or metastatic lesions less than 1 gram. Radiolabeled antiferritin tumor targeting decreased with increasing tumor size. In this study, ferritin levels were measured in H4IIe tumors grown both in vitro and in vivo. In vitro tumor cells synthesized and secreted ferritin into the medium as measured by radioimmunoassay, and confirmed by the incorporation of 14C-leucine into ferritin synthesis. The concentration of ferritin in the tumor cells as measured by radioimmunoassay remained relatively constant over this same time period. In vivo tumor ferritin levels in whole tumor extracts were highest in small tumors (less than 1 gram) and decreased as the tumors became larger. Serum ferritin levels of tumor-bearing animals paralleled the level in the tumors themselves. The elevated serum ferritin levels in animals with small tumors did not inhibit tumor targeting with radiolabeled antiferritin antibody. These findings are a foundation for understanding the selective tumor targeting of tumor associated proteins by radiolabeled antibodies, which includes factors such as tumor size, vascularity, antigen content, and circulating antigen.

A model system that predicts effective half-life for radiolabeled antibody therapy.

Radiolabeled antibodies to tumor associated proteins localize in both experimental and clinical cancers. In the therapeutic applications of radiolabeled antibody, tumor effective half-life (composite of biological and physical half-lives), along with the concentration of isotope deposited and energies of the isotope used, determine the tumor dose. Antibodies directed against the same antigenic specificity but derived from different species have varied tumor and whole body effective half-lives and as a result, achieve different tumor doses. In vitro testing does not evaluate the in vivo differences in effective half-life that affect tumor dose. We have developed an animal model to evaluate the effective half-life and biodistribution of radiolabeled immunoglobulin (IgG) from diverse species. To determine the relevance of such a model, the effective half-lives and tissue distributions of the different immunoglobulins in the model were compared to those obtained from the clinical program using the same radiolabeled antibody preparations. In both the experimental model and in the clinical trials, radiolabeled immunospecific and normal IgG derived from monkey, rabbit, and porcine sources had the longest effective half-lives, goat and sheep had intermediate effective half-lives, and chicken and turkey had the shortest effective half-lives. Prescreening of bovine and baboon normal IgG predict long half-lives and similar organ distributions. These species have been immunized for clinical use. Bovine IgG has a long clinical half-life and has been added to our other successful antibodies. Baboon IgG is now ready for clinical testing. The value of this model system is that it appears to be an effective in vivo preclinical screen for tumor effective half-life of antibodies and IgG from diverse species, thus guiding potential clinical use.

The effect of iron dextran on ferritin content and 131I-antiferritin localization in experimental hepatomas.

Polyclonal 131I rabbit antirat ferritin localizes in certain hepatoma models. The effect of intraperitoneal iron dextran on tumor and sera ferritin content and tumor and normal tissue localization with 131I antiferritin was studied. Separate groups of 10-12 animals were injected with escalating doses of 131I-antiferritin IgG, or nonspecific IgG, one week after injection with iron dextran or normal saline. The results demonstrate that tumor, serum, and normal tissue ferritin content was increased after iron dextran administration but tumor localization increased after administration of 131I-antiferritin in the H4II-E and 7800 models. The 3924A and 7777 models showed no tumor localization with or without iron dextran but did show an increase in normal tissue localization after iron dextran. Immunoperoxidase staining of tissues with antiferritin revealed increased staining in the liver and spleen and only a slight increase in the tumors after iron dextran was administered. The results demonstrate that tumor localization is a complex phenomenon that depends on normal tissue, sera, and tumor-antigen distribution.

Factors that affect antiferritin localization in four rat hepatoma models.

The effect of tumor size, vascularity, ferritin content and the amount of injected 131I-antiferritin on tumor localization was studied in four hepatoma models with varying growth rates, histology, vascularity, and ferritin content. Separate groups of 12 animals with the H-4-II-E, 7800, 7777, and 3924A rat hepatomas with less than 2 g or greater than 2 g tumors were injected with escalating doses of 131I-antiferritin or 131I nonspecific IgG (control). Tumor vascularity was measured by 51Cr-labeled erythrocyte injection, ferritin content of tumors by radioimmunoassay and immunoperoxidase staining, and the histological location of 131I-antiferritin by autoradiography. 131I-antiferritin specifically localized in the H-4-II-E and 7800 models and correlated with the tumor size, vascular content, and amount of injected antiferritin. No localization took place in the 7777 or 3924A tumors despite the presence of ferritin in these models. The only factor that correlated with localization in the models was vascularity. The vascularity of 3924A and 7777 tumors was significantly reduced in comparison to the H-4-II-E and 7800 tumors. The dependence of targeting on vascularity was demonstrated with autoradiography as well. These findings indicate the correlation of vascularity and tumor localization with 131I-antiferritin.

Variables affecting the tumor localization of 131I-antiferritin in experimental hepatoma.

Ferritin is both a normal tissue- and tumor-associated protein. The in vivo localization of 131I-radiolabeled antitumor ferritin and normal IgG antibodies in the H-4-II-E rat hepatoma model was investigated in both tumor and normal tissues over a dose range of 0.67 micrograms to 5 mg of normal and antiferritin IgG and at labeling ratios (microCi 131I per micrograms IgG) of 15:1, 5:1, and 1:10. The total dose from nonpenetrating radiation in rads was calculated and demonstrated a maximum of 2.9 times greater dose deposition (rads) of antiferritin than normal IgG in hepatoma without specific increase in binding in normal tissues. The maximum tumor targeting achieved was dependent on the amount of injected IgG and not on the labeling ratio or procedure. The binding in tumor could be inhibited by unlabeled antiferritin but not by unlabeled normal rabbit IgG and demonstrated the requirement of specificity for tumor binding. Normal tissues did not target with antiferritin. Most normal tissues have a capacity to bind normal and antiferritin IgG nonspecifically that is linear in relationship to the amount of injected IgG. The results demonstrate that 131I-antiferritin selectively targets ferritin-secreting hepatoma over normal tissues and that the amount of targeting is dependent on the amount of antiferritin injected. The physiologic reasons for such selective localization is not known, but the term "biologic window" has been used to describe the differential availability of tumor ferritin for binding.

Selective tumor localization in experimental hepatoma by radiolabeled antiferritin antibody.

The in vivo localization of 131I-radiolabeled antiferritin and normal IgG in the H-4-II-E rat hepatoma model was investigated by serial necropsy. Groups of 14 to 18 animals were injected with 500 microCi (200 micrograms) of normal and antiferritin IgG. The total dose from the nonpenetrating beta radiation was calculated for tumor and normal tissue, and expressed as a targeting ratio of antiferritin to normal IgG for each organ studied. The results demonstrate 2.9 times greater dose deposition in tumors of animals treated with 131I-antiferritin than with 131I-normal IgG. 131I-antiferritin deposited equivalently in primary tumors and metastatic lesions of similar size. The specific binding in tumors could be competitively inhibited by the addition of unlabeled antiferritin but not unlabeled normal IgG. Specific targeting with 131I-antiferritin comparison to 131I-normal IgG did not occur in any normal tissue. There was considerable variation in the dose deposition in different normal tissues.