Dosimetry of rhenium-186-labeled monoclonal antibodies: methods, prediction from technetium-99m-labeled antibodies and results of phase I trials.

Rhenium-186 is a beta-emitting radionuclide that has been studied for applications in radioimmunotherapy. Its 137 keV gamma photon is ideal for imaging the biodistribution of the immunoconjugates and for obtaining gamma camera data for estimation of dosimetry. Methods used for determining radiation absorbed dose are described. We have estimated absorbed dose to normal organs and tumors following administration of two different 186Re-labeled immunoconjugates, intact NR-LU-10 antibody and the F(ab')2 fragment of NR-CO-02. Tumor dose estimates in 46 patients varied over a wide range, 0.4-18.6 rads/mCi, but were similar in both studies. Accuracy of activity estimates in superficial tumors was confirmed by biopsy. Prediction of 186Re dosimetry from a prior 99mTc imaging study using a tracer dose of antibody was attempted in the NR-CO-02 (Fab')2 study. Although 99mTc was an accurate predictor of tumor localization and the mean predicted and observed radiation absorbed doses to normal organs compared favorably, 186Re dosimetry could not be reliably predicted in individual patients. The methods described nevertheless provide adequate estimates of 186Re dosimetry to tumor and normal organs.

Detection and staging of small cell lung carcinoma with a technetium-labeled monoclonal antibody. A comparison with standard staging methods.

Tumor-associated radiolabeled monoclonal antibodies (MoAb) can detect neoplasms in a variety of settings. The authors conducted a study comparing the ability to detect and stage small cell lung carcinoma by using a Tc-99m labeled monoclonal antibody (NR-LU-10 Fab) (NeoRx Corp, Seattle, WA) with standard staging methods. Standard staging included a physical examination, chest x-ray, a battery of radionuclide scans and/or computerized tomographic studies (head, abdomen, and bone), and bone marrow examination. A total of 22 comparisons were performed in 17 patients (five patients had reevaluations after therapy). Fifty-four (74%) of the 73 lesions defined by standard staging were detected by the radiolabeled MoAb. Seven of eight patients (88%) classified by standard staging as having "limited stage" disease on presentation were concordantly "limited stage" by radioimmunoimaging. One patient deemed "limited stage" by standard staging was correctly upstaged (bone marrow involvement) as a result of the radiolabeled MoAb. Two patients found to have extensive disease at diagnosis were characterized as "limited stage" by the MoAb, for an overall staging accuracy of 0.88. Thirteen of 19 missed lesions were smaller than 2 cm (10 were smaller than 1 cm; 3 measured 1 to 2 cm). This comparative study shows that radioimmunoimaging by Tc-99m labeled NR-LU-10 Fab antibody is capable of complementing standard staging methods used in the evaluation of small cell lung carcinoma.

Radionuclide imaging of bone marrow metastases with a Tc-99m labeled monoclonal antibody to small cell lung carcinoma.

The detection of metastatic disease confined to the bone marrow compartment has in the past been technically limited. We have identified excellent imaging of bone marrow metastases during the evaluation of a Tc-99m labeled monoclonal antibody (NR-LU-10 Fab) (NeoRx Corp., Seattle, WA). This occurred during a study to assess the monoclonal antibody's ability to detect sites of small cell cancer (primary and metastatic). The study by design compares areas seen by the monoclonal antibody scan with those found by standard staging methods in patients with small cell lung cancer. Standard staging included chest x-rays, bone scans, CT studies of the abdomen, and histologic examination of the bone marrow. Fifteen patients have been evaluated, four on two occasions, for a total of 19 monoclonal imaging studies. Metastasis to the marrow compartment was identified by the monoclonal imaging in all patients whose bone marrow biopsies were positive for small cell carcinoma, and it was primarily responsible for the eventual detection of extensive disease (marrow involvement) in one patient. Thus it appears that compartmental bone marrow imaging for metastatic disease is possible with immunoscintigraphy.

186Re radioimmunotherapy of small cell lung carcinoma xenografts in nude mice.

A 186Re-labeled monoclonal antibody (MAb), NR-LU-10, was used for the radioimmunotherapy of a subcutaneous human small cell lung carcinoma xenograft, SHT-1, in nude mice. Biodistribution with specific and irrelevant labeled MAb demonstrated peak tumor uptake of 8% and 3% of the injected dose/g at 2 days, respectively. Dosimetry analysis predicted tumor:whole-body radiation-absorbed dose ratios of 2.43:1 for NR-LU-10 and 0.62:1 for irrelevant MAb. Single-dose toxicity screening estimated a 50% lethal dose within 30 days of 600 microCi (880 cGy of whole-body radiation). As anticipated, a multiple-dose regimen of 490 microCi in four doses over 10 days (720 cGy of whole-body radiation, eight of eight surviving greater than 30 days) was less toxic than a single bolus dose of 430 microCi (644 cGy of whole-body radiation), six of eight surviving greater than 30 days). A multidose radioimmunotherapy regimen was initiated in nude mice bearing 66-mm3 tumors (total dose, 500 to 600 microCi). Complete remissions (greater than 140 days) were achieved in three of 16 mice, and the remainder showed a mean tumor growth delay of 53 days. Matched doses with irrelevant MAb produced one remission, one treatment-related death, and a mean growth delay of only 20 days in six of eight mice. Thus, in this nonoptimal radioimmunotherapy model, significant antitumor responses were observed using a mildly toxic multiple dosing regimen.

Staging non-small cell carcinoma of the lung using technetium-99m-labeled monoclonal antibodies.

A technetium-labeled monoclonal antibody was administered to 52 patients with non-small cell lung carcinoma, either to stage the mediastinum preoperatively or to detect distant metastases. Results from planar and tomographic imaging are compared to CT and histologic confirmation. Differences in detection rates and predictive values between imaging modalities are discussed. The authors conclude that imaging with a technetium-labeled monoclonal antibody is safe and accurate and may be useful for staging patients with either operable or inoperable non-small cell lung cancer.

Immunotoxins of Pseudomonas exotoxin A (PE): effect of linkage on conjugate yield, potency, selectivity and toxicity.

Conjugates of monoclonal antibodies and Pseudomonas exotoxin A (PE) were formed with disulfide or thioether bonds. Thioether conjugates which formed with succinimidyl 4-(N-maleimidomethyl)-cyclohexane-1-carboxylate (SMCC) modified PE and reduced antibody formed with an 80% yield of equimolar conjugate within 30 min with an offering of one to one (toxin:antibody). The efficiency and kinetics of thioether formation were much higher with SMCC than with other maleimide reagents as well as more efficient than disulfide linkers. Thioether linkage resulted in immunotoxin consistently more potent and more selective in vitro than disulfide bonded conjugate. Thioether bonded conjugates also proved to have other favorable in vivo properties compared to disulfide conjugates: (1) a longer half-life in serum; (2) increased tumor localization; and (3) reduced toxicity. Toxicity of thioether linked holotoxin conjugates was directed at the liver hepatocyte but was easily monitored by serum liver enzymes. The conjugates are currently undergoing clinical evaluation for treatment of ovarian cancer with intraperitoneal administration. Research is ongoing to further decrease residual toxicity without reducing the potency of the conjugate.

Development of a stable radioiodinating reagent to label monoclonal antibodies for radiotherapy of cancer.

A method of radioiodinating monoclonal antibodies such that the labeled antibodies do not undergo in vivo deiodination has been studied. The method utilizes conjugation of succinimidyl para-iodobenzoate to the antibody. The iodobenzoate was radiolabeled by using an organometallic intermediate to facilitate the reaction. Thus, succinimidyl para-tri-n-butylstannylbenzoate was radiolabeled in 60-90% radiochemical yield and subsequently conjugated to the antibody in 80-90% yield. Animal biodistribution studies were carried out with two separate anti-melanoma antibodies (9.2.27 and NR-M1-05) labeled by this method, and examined in nude mice bearing human melanoma tumor xenografts. Very large differences in the localization of radioactivity were observed in the thyroids and stomachs of mice when the iodobenzoyl-labeled antibodies were compared with the same antibodies labeled using the chloramine-T method of radioiodination. Few other significant differences in the tissue distribution of the radioiodinated antibodies were seen.

Successful imaging of malignant melanoma with technetium-99m-labeled monoclonal antibodies.

F(ab')2 and Fab fragments of murine monoclonal antibody 9.2.27, that recognizes the 250 kD melanoma-associated antigen, were labeled with 99mTc using the bifunctional chelate method of Fritzberg et al. Twenty-seven (27) patients received, intravenously, 10 mg of either F(ab')2 (8), or the Fab (27), labeled with up to 30 mCi of 99mTc. These doses were preceded by an infusion of cold irrelevant antibody. The average serum T1/2 of the F(ab')2 and the Fab were 11 hr and 2 hr, respectively. Twenty-two percent (22%) of the total injected F(ab')2 dose was excreted in the urine in 20 hr, compared to 55% for the Fab group. Imaging was optimal 6-9 hr postinjection for the Fab patients. No nonspecific uptake in liver, spleen, bone marrow, or lung was observed for either antibody form. Overall, (43/53) 81% of known metastases were seen with visualization of tumors as small as 250 mg and tumor localization as high as 0.03% injected dose/g. Immunoperoxidase staining of freshly-frozen tumor nodules removed 24 hr postinjection confirmed antibody deposition in the tumor. Thirty-six previously unknown ("occult") metastatic sites were detected. To date, 12/36 of these sites have been confirmed. We conclude that 99mTc-labeled antibody to melanoma produces high resolution images with a high sensitivity of detecting metastatic melanoma. The detection of previously unknown sites of disease has proven helpful in directing additional diagnostic studies (i.e., CT) as well as planning of therapeutic options.

Specific and stable labeling of antibodies with technetium-99m with a diamide dithiolate chelating agent.

Technetium-99m labeling of antibodies has been suboptimal because of low affinity adventitious binding, nonspecific labeling, and loss of immunoreactivity. The diamide dithiolate ligand system (N2S2) forms highly stable, well-defined tetradentate complexes with Tc(V). Antibodies and their fragments have been labeled by conjugation of preformed 99mTc-4,5-bis(thioacetamido)pentanoate active ester to protein amine groups to give a chemically known 99mTc-N2S2 complex covalently linked to antibody. Evaluations of the 99mTc-N2S2-bound antibodies and their fragments have shown high stability and retained immunoreactivity.

Effect of antibody dose on the imaging and biodistribution of indium-111 9.2.27 anti-melanoma monoclonal antibody.

Eleven patients with metastatic melanoma underwent serial gamma camera imaging and biodistribution measurements after i.v. injection of escalating doses of [111In]9.2.27, an antimelanoma murine monoclonal antibody. Patients received a fixed dose of 1 mg of [111In]9.2.27, with no additional 9.2.27 (five patients), or co-infused with 49 mg (five patients) or 99 mg (one patient) of unlabeled, unconjugated 9.2.27. Higher doses resulted in prolonged blood-pool retention, less uptake in spleen and bone marrow, and appeared to have a positive effect in improving tumor imaging. A dose of 1 mg of 9.2.27 permitted detection of tumors in two of five patients and two of ten lesions, while with greater than or equal to 50 mg, tumors were detected in all patients and in 24 of 32 lesions. Human gamma globulin injected prior to administration of [111In]9.2.27 failed to block the prominent liver, spleen, and bone marrow uptake. No toxicity was observed. These results indicate the feasibility of imaging metastatic melanoma with [111In]9.2.27 and suggest that antibody dose may be a critical determinant of biodistribution and tumor uptake.

Fate of gamma-interferon-activated killer blood monocytes adoptively transferred into the abdominal cavity of patients with peritoneal carcinomatosis.

Five patients with colorectal cancer widely metastatic to peritoneal surfaces have been treated i.p. with infusions of autologous blood monocytes made cytotoxic by in vitro incubation with human gamma-interferon. The monocytes were purified by a combination of cytapheresis and counter-current centrifugal elutriation procedures; each week approximately 350 million activated monocytes were given to patients as adoptive immunotherapy by a single i.p. instillation. On the eighth cycle of treatment the trafficking of i.p. infused blood monocytes was studied in two patients by prelabeling the cells with 111In. These activated cells became distributed widely within the peritoneal cavity. Two and 5 days after infusion their position within the peritoneum had not changed. When peritoneal specimens were obtained 36 h after 111In-labeled monocyte infusion, labeled monocytes were demonstrated to be associated with the serosal surfaces by autoradiographic analysis. Scintiscanning structures outside the abdominal cavity revealed that 111In-labeled monocytes infused i.p. did not traffic to other organs during the 5 days of the study. We conclude that i.p. adoptive transfer of autologous killer blood monocytes is an effective way of delivering these cytotoxic cells to sites of tumor burden on peritoneal surfaces in these cancer patients.

Monoclonal antibody therapy in malignant melanoma: factors effecting in vivo localization.

Thirteen patients with metastatic malignant melanoma received intravenous therapy with the murine antimelanoma monoclonal antibody 9.2.27. Five patients were entered on a dose escalation protocol with twice weekly escalating doses of 10-500 mg, in an extension of a previously reported trial. These patients demonstrated near saturation of available antibody binding sites in vivo following the 500 mg dose, with minimal toxicity. The remaining patients were entered onto a dose schedule comparison study, with a 500 mg dose administered either in a single 2 h infusion or as five daily 2 h infusions of 100 mg to examine the effects of different dose schedules and of an interrupted schedule on subsequent therapy with the same antibody. Intratumor localization of the monoclonal antibody did not appear to vary with respect to the dose schedule; however, interruption in therapy for 4 weeks was accompanied by somewhat poorer localization of antibody. This effect appeared to be primarily attributable to development of human antimurine antiglobulin in 25-30% of patients with resultant decrease in intratumor localization of antibody and more rapid clearance of the 9.2.27 antibody from the circulation. Earlier reports with other antibodies notwithstanding, initial infusions of 500 mg of 9.2.27 did not induce tolerance to the murine immunoglobulin. This study confirms and extends the findings of our initial trial of the 9.2.27 antibody by demonstrating that, although clinical responses were not observed, the antibody can be safely administered at doses up to 500 mg, with good intratumor localization of antibody. The diminished localization of antibody associated with antiglobulin responses indicates the importance of monitoring antiglobulin levels during therapy, and the necessity of controlling or preventing this phenomenon when monoclonal antibodies are administered in multiple doses as drug, toxin, or radionuclide immunoconjugates.

Kinetic model for the biodistribution of an 111In-labeled monoclonal antibody in humans.

Using data from 12 patients, we have analyzed the pharmacokinetics of 111In-9.2.27, an antimelanoma monoclonal antibody, following i.v. infusion. Plasma data and scintillation camera images obtained from patients receiving either 1, 50, or 100 mg of monoclonal antibody indicated dose-dependent (i.e., saturable) kinetics. Based on these observations and known immunoglobulin kinetics, we developed a nonlinear compartmental model to describe the biodistribution of 111In-9.2.27 and the other coinjected 111In-associated compounds. The model included (a) three compartments representing intact 111In-9.2.27 ("plasma," "nonsaturable," and "saturable binding" compartments), (b) four compartments representing 111In-diethylenetriaminepentaacetic acid, and (c) one compartment representing 111In in an undetermined chemical form ("extravascular delay" compartment). Analysis of the rate of urinary excretion relative to plasma concentration indicated that the saturable binding compartment was a site for catabolism of monoclonal antibody. Further examination of the urinary data, together with previous studies of the site(s) of immunoglobulin catabolism, suggested that additional elimination took place from either the plasma or the nonsaturable compartment. The model indicated that to fill the saturable sites would require a dose of approximately 0.5 mg and suggested that greater than 3.5 mg would maintain saturation for 200 h. Computer integration of gamma camera counts over the spleen revealed a clear saturable component of uptake, whereas integration over the liver showed no such pattern. The proposed model was fitted to the liver and spleen imaging data by summing fractions of model simulations of each compartment. That analysis confirmed the suspected saturable uptake by the spleen (21% of the saturable binding compartment) and revealed a quantitatively important component of saturation in the liver (35% of the saturable binding compartment) that was not obvious from initial examination of the images. When the results were expressed on a concentration basis, the spleen accounted for 247% of the saturable compartment per kg, whereas the liver accounted for 25%/kg. The bone marrow also showed saturable uptake; hence, the saturable uptake may relate to the sinusoidal blood supply characteristic of liver, spleen, and marrow. The model predicts the dose levels required to overcome saturable background, suggests appropriate doses and schedules for cold loading strategies, and provides a format for explicit inclusion of tumor antigen.

Relationship of the clinical response and binding of recombinant interferon alpha in patients with lymphoproliferative diseases.

Patients with hairy cell leukemia (HCL) and chronic lymphocytic leukemia (CLL) were treated with recombinant interferon alpha A (rIFN-alpha A). The binding of iodinated recombinant interferon-alpha to baseline samples of peripheral blood mononuclear cells (PBMCs) from the leukemia patients was compared with clinical responsiveness to rIFN-alpha A. HCL patients (8/10) responded to rIFN-alpha A therapy, whereas none (0/10) of the CLL patients studied responded. The PBMCs from the eight responsive HCL patients bound approximately twice as much iodinated interferon as the PBMCs from nonresponsive CLL patients. This difference was due to more high-affinity receptors per cell with no difference in the affinity of the interferon-receptor interaction. However, because PBMCs from HCL patients were larger than PBMCs from CLL patients, the cell surface receptor density was similar. The leukemic cells from one of the two nonresponsive HCL patients bound iodinated interferon similarly to the cells from the responsive HCL patients, whereas the leukemic cells from the other nonresponsive HCL patient bound considerably less. The rapidity of response of the HCL patients did not correlate with the level of binding of iodinated interferon. Our results suggest that the absolute number of interferon receptors per cell may be only one of several important parameters in the response to rIFN-alpha A therapy, and that the responsiveness of a particular lymphoproliferative disease or a particular patient to rIFN-alpha A therapy cannot be predicted or explained solely by the degree of interaction between IFN and its cell surface receptor.

Recombinant leukocyte A interferon therapy for advanced hairy cell leukemia. Therapeutic and immunologic results.

Hairy cell leukemia is a lymphoproliferative disorder characterized clinically by cytopenias. Standard therapy following variable periods of disease stability consists of splenectomy that often restores normal hematologic parameters for periods ranging from weeks to years. Fifteen patients (five without prior splenectomy or chemotherapy) were treated with 3 X 10(6) units per day of recombinant leukocyte A interferon and 14 of 15 patients completed eight weeks of therapy and were evaluated for response. There was one complete and 12 partial responses for an overall response rate of 93 percent. All of these patients' conditions have remained in complete or partial remissions and they continue to receive interferon with a median follow-up of six months. Coincident with the normalization of peripheral blood counts was a return of natural killer activity and normalization of immunologic surface markers as determined by monoclonal antibodies. This study confirms and extends earlier observations with natural alpha-interferon and indicates that recombinant leukocyte A interferon in low daily doses is also very effective treatment for hairy cell leukemia. In fact, it may be the best single modality of therapy for inducing both hematologic and immunologic recovery of these patients and deserves consideration as initial therapy.

Optimal strategies for developing human-human monoclonal antibodies.

Human monoclonal antibodies are desirable, especially as therapeutic agents, but the best means of producing them is still a matter of investigation. It is clear that human antibodies of predicted specificity from patients with autoimmune disease can be derived, and this may help unlock some of the mysteries of these illnesses. Human monoclonal antibodies against tumor-specific antigens for use in in vivo diagnosis and therapy remain desirable goals. Problems involved in their routine development include the lack of available, adequately immunized, and differentiated lymphocytes and the nature and paucity of the available human "myeloma" cell lines. These lines have been compared now by a number of authors who have reached similar conclusions. Our study directly compared the greatest number of cell lines and found UC729-6 and HF2 to be the best; on the other hand, our success in developing IgG-secreting hybridomas from U-266, using hyperimmunized lymphocytes, suggests that this line may only be capable of secretion with the more differentiated cell, the human equivalent of those hyperimmunized murine spleens. Hence both sides of the fusion equation must be made optimal. Two new approaches to circumvent this problem involve the use of either a human-murine myeloma chimera as the parental myeloma line or, more recently, genetic engineering techniques to substitute human constant regions for the murine while retaining the murine hypervariable region, preserving the binding specificity of the murine antibody.

A phase I evaluation of poly(I,C)-LC in cancer patients.

Twenty-five patients with metastatic carcinoma were entered into a Phase I clinical trial using poly(I,C)-LC at either 1 mg/m2 or 4 mg/m2 intravenous, twice weekly, for 4 weeks. None of the 15 patients entered at the 1 mg/m2 dose had an objective response; three had progressive disease. Similarly, no objective responses were observed among the 10 patients treated at the 4 mg/m2 dose of poly(I,C)-LC; one patient was removed from the study due to progressive disease. Toxicities observed at the 1 mg/m2 dose were mild hypotension, fever, nausea, vomiting, fatigue, and headache. The first patient treated at the 4 mg/m2 dose was taken off of the study for severe hypotension. In the subsequent nine patients treated at this dose, a pretreatment with one dose at 1 mg/m2 was given, and no further problems with hypotension were encountered. The other toxicities at 4 mg/m2 were similar to those seen at 1 mg/m2.

The generation of monoclonal anti-idiotype antibodies to human B cell-derived leukemias and lymphomas.

We developed murine anti-idiotype monoclonal antibodies for each of four patients with B cell-derived leukemias and lymphomas. Idiotypic immunoglobulin was isolated from mouse X human tumor-cell hybridomas or from patients' serum and was used to immunize mice for the development of murine anti-idiotype monoclonal antibodies. Each patient's anti-idiotype antibodies demonstrated reactivity restricted to the immunizing immunoglobulin, thereby limiting their therapeutic utility to a single individual. In addition, we isolated isotype switch variants of hybridomas producing monoclonal anti-idiotypic antibody. The restricted specificity of these antibodies was found to be of value for the analysis of the extent of malignant B cell infiltration in a variety of tissues from several patients. Large populations of idiotype-bearing cells were detectable in biopsy specimens from patients K.T. and L.H. In contrast, although bone marrow specimens from patient G.D. were apparently devoid of morphologically abnormal cells, a small, highly fluorescent population of cells was demonstrable underscoring the potential utility of these antibodies for posttreatment evaluation as well as for therapy. In a fourth patient, H.M., anti-idiotype antibodies developed against the circulating macroglobulin isolated from his plasma failed to react with either his circulating or bone marrow hairy cell leukemia cells. However, examination of an enlarged inguinal lymph node revealed the presence of a large number of idiotype-bearing cells. Thus, the presence of two distinct malignant B cell clones were discovered in this individual through the use of anti-idiotype monoclonal antibodies. Anti-idiotype antibodies, therefore, represent a highly specific tool for the evaluation and potential therapy of B cell malignancies in individual patients.