Search

A comparison was made between labeled antibody accumulations in nude mice having either single or multiple human xenografts. The LS174T tumors were implanted subcutaneously. All animals were given 2 micrograms of labeled murine anti-carcinoembryonic antigen (CEA) monoclonal antibody 111In-mT84.66. Some animals were also given specific antibody pretreatment (SAP) of 200 micrograms of unlabeled mT84.66 to reduce liver accumulation of activity. In order to represent these multiple tumor examples, a simple initial-phase pharmacokinetic model was first fitted to each of the two groups (SAP and PBS treated) of single-tumor animals. Using the resultant six non-adjustable parameters as constants, the n = 1 uptake model was then used to represent tumor, liver and blood accumulations (%injected dose/organ) in the multiple-tumor animals. The model was found to be a good representation; in particular, it had far better agreement than single tumor predictions in the PBS mice. Differences between the single-tumor accumulations and those seen in multiple tumor examples were generally between two- and three-fold. The model also demonstrated that the result of SAP was to essentially eliminate the effect of liver targeting of tumor-secreted CEA. We conclude that an initial-phase one-tumor model can describe the decrease of accumulation of activity in the case of multiple tumors in nude mice in both untreated (PBS) and pretreated conditions. Implications for clinical imaging and therapy with monoclonal agents are discussed.

A simple, water-soluble procedure for conjugation of monoclonal antibodies to 1,4,7,10-tetraazacyclododecane-N,N',N",N"'-tetraacetic acid (DOTA) has been improved by optimizing pH, buffer, and temperature conditions for the preparation of N-hydroxysulfosuccinimidyl DOTA and its conjugation to the human/murine chimeric anti-carcinoembryonic antigen antibody cT84.66. This improved method results in a 6-fold increase in conjugation efficiency, a 3-7-fold decrease in antibody cross-linking, a more homogeneous population of conjugate species, and a 5-fold decrease in the quantities of reagents needed for conjugation. The cT84.66-DOTA conjugate was labeled to high specific activity with 111In, 90Y, 88Y, 64Cu, and 67Cu, affording near-quantitative incorporation of the majority of these radiometals. This improved conjugation procedure facilitates large-scale production and radiometal labeling of cT84.66-DOTA for clinical radioimmunotherapy trials.

Three analytic indicators were used to compare five members of a monoclonal antibody (Mab) family. The cognates consisted of the genetically engineered intact chimeric IgGI (cT84.66) and related engineered fragments [scFv, diabody, minibody, F(ab')2] reactive against the same epitope of carcinoembryonic antigen (CEA). All analyses were based on radioiodinated Mabs targeting to colorectal xenografts of LS174T tumors in nude mice. Affinity constants were evaluated initially. A second indicator was the imaging figure of merit (IFOM) which determines how rapidly a statistically significant tumor image can be acquired. Finally, deconvolution was used to determine tumor temporal response to an arterial bolus. This last analysis gave the possible tumor accumulation in the absence of normal tissue sequestration. Affinities were all in excess of 10(8) M-1 and were highest for the divalent Mabs. Using the IFOM criterion, an 131I label was best suited as a radiolabel for the intact (IgG) T84.66, while an 123I label indicated optimal imaging with either minibody or F(ab')2. Deconvolution analyses showed that divalent members behaved similarly while the univalent member (scFv) had a tumor residence time smaller by an order of magnitude. The diabody had the largest impulse response function, but renal uptake may limit its present usefulness.

Chimeric T84.66 (cT84.66) is a genetically engineered human/murine chimeric IgG, with high affinity and specificity to carcinoembryonic antigen (CEA). The purpose of this Phase I dose escalation therapy trial was to evaluate the toxicities, biodistribution, pharmacokinetics, tumor targeting, immunogenicity, and organ and tumor absorbed dose estimates of cT84.66 labeled with 90Y. Patients with metastatic CEA-producing malignancies were first administered 5 mCi 111In-labeled DTPA-cT84.66 (5 mg), followed by administration of the therapy dose of 90Y-labeled DTPA-cT84.66 1 week later. The therapy infusion was immediately followed by a 72-h administration of DTPA at 250 mg/m2/24 h. Dose levels of administered activity ranged from 5 to 22 mCi/m2 with three to six patients per level. Serial nuclear scans, blood samples, and 24-h urine collections were performed out to 5 days after infusion. Human antichimeric antibody response was assayed out to 6 months. Patients were administered up to 3 cycles of therapy every 6 weeks. Radiation absorbed doses to organs were estimated using a five compartment model and MIRDOSE3. Twenty-two patients received at least one cycle of therapy, with one individual receiving two cycles and two receiving three cycles of therapy. All were heavily pretreated and had progressive disease prior to entry in this trial. Reversible leukopenia and thrombocytopenia were the primary dose-limiting toxicities observed. Maximum tolerated dose was reached at 22 mCi/ m2. In general, patients with liver metastases demonstrated more rapid blood clearance of the antibody. Thirteen patients developed an immune response to the antibody. Average radiation doses to marrow, liver, and whole body were 2.6, 29, and 1.9 cGy/mCi 90Y, respectively. Dose estimates to tumor ranged from 66 to 1670 cGy (8.7 to 52.2 cGy/mCi 90Y) for each cycle of therapy delivered. Although no major responses were observed, three patients demonstrated stable disease of 12-28 weeks duration and two demonstrated a mixed response. In addition, a 41-100% reduction in tumor size was observed with five tumor lesions. 90Y-labeled cT84.66 was well tolerated, with reversible thrombocytopenia and leukopenia being dose limiting. Patients with extensive hepatic involvement by tumor demonstrated unfavorable biodistribution for therapy with rapid blood clearance and poor tumor targeting. Average tumor doses when compared with red marrow doses indicated a favorable therapeutic ratio. Stable disease and mixed responses were observed in this heavily pretreated population with progressive disease. This trial represents an important step toward further improving the therapeutic potential of this agent through refinements in the characteristics of the antibody and the treatment strategies used. Future trials will focus on the use of peripheral stem cell support to allow for higher administered activities and the use of combined modality strategies with radiation-enhancing chemotherapy drugs. Further efforts to reduce immunogenicity through humanization of the antibody are also planned. Finally, novel engineered, lower molecular weight, faster clearing constructs derived from cT84.66 continue to be evaluated in preclinical models as potential agents for radioimmunotherapy.

HER2/neu oncogene encodes a 185 kDa trans-membrane protein which is overexpressed in 20-30% of breast and ovarian cancers and portends a poor prognosis. We have studied the targeting and therapy of this oncoprotein with 4D5, a murine monoclonal antibody which recognizes a distinct epitope on the extracelluar domain of HER2/neu. We conjugated the antibody with an active ester of the macrocyclic chelating agent DOTA, radiolabeled the conjugate with either (111)In or (90)Y, and studied the antibody distribution and therapy, respectively, in athymic mice bearing xenografts of MCF7/HER2/neu, a human breast cancer cell line transfected with the HER2/neu oncogene. For the biodistribution of (111)In-labeled DOTA-4D5, a high specificity of tumor localization (30% ID/g) was seen with a tumor-to-blood ratio of greater than 2 at 48 h postinjection. Compared to a previously published study with (125)I-labeled 4D5 in beige nude mice bearing NIH3T3/HER2/neu xenografts [De Santes et al. (1992) Cancer Res. 52, 1916-1923], (111)In-labeled 4D5 antibody gave superior antibody uptake in tumor (30% ID/g vs 17% ID/g at 48h). In the therapy study, treatment of the nude mice bearing MCF7/HER2/neu xenografts with 100 microCi (3 microg) of (90)Y-labeled DOTA-4D5 caused a 3-fold reduction of tumor growth compared to untreated controls (injected with human serum albumin) in 40 days. Treatment of animals with 100 microCi of nonspecific antibody (90)Y-labeled DOTA-Leu16 (3 microg) had no tumor growth inhibition. Treatment with unlabeled DOTA-4D5 (3 microg) had a slight effect on tumor growth compared to untreated controls. When analyzed at the level of single animals, no effect was seen in seven of nine animals; however, in two of the animals, tumor growth inhibition was observed. Although a cold antibody therapeutic effect was unexpected at this dose level (3 microg), it may be possible that in some animals that 3 microg of antibody of (90)Y-labeled DOTA-4D5 augmented tumor growth reduction. To further explore the effects of cold antibody treatment alone, animals were treated with 100 or 400 microg of unlabeled 4D5 administered in two doses. These animals showed a 1.7-1.8-fold reduction in tumor growth over 28 days, a result less than that obtained with RIT only.

Background: Carcinoembryonic antigen (CEA) is a human tumor antigen with the domain structure N-A1-B1-A2-B2-A3-B3, in which each domain is predicted to have an Ig-like fold and is known to bind epitope specific anti-CEA antibodies., Objective: To determine the affinity constants of several domain specific anti-CEA antibodies using purified recombinant or synthetic domains., Results and Conclusion: We have determined the kinetic and affinity constants of several anti-CEA antibodies for CEA, CEA domains (A3-B3) expressed in HeLa cells, and a synthetic peptide corresponding to the A3 domain using a BIAcore biosensor. There was no difference in affinity for CEA among a murine (mT84.66), a mouse/human chimeric form (cT84.66) or a disulfide deleted version (delta SScT84.66) of this antibody. There was less than a five-fold drop in affinity of murine T84.66 for the A3-B3 domain expressed in HeLa cells compared to CEA. The synthetic A3 domain had an affinity constant for mT84.66 which was ten-fold less than for CEA. The affinity constants for CEA with several other anti-CEA monoclonal antibodies, including three antibodies which have almost identical CDR sequences (CEA.281, CEA.11 and CEM231) were also determined. CEM231 which had a two-fold higher affinity constant for CEA than either CEA.281 or CEA.11 had a two-fold faster on-rate which accounts for its higher affinity constant. This difference may be due to one or more of the amino acid differences present in H1 (N vs. S or D) and H3 (A vs. V).

Biodistributions of two radiometal chelate conjugates of the human/murine chimeric anticarcinoembryonic antigen monoclonal antibody cT84.66 were obtained in nude mice bearing LS174T human colorectal carcinoma xenografts. Derivatives of the macrocyclic chelating agent 1,4,7,10-tetraazacyclododecane-N,N',N",N"'-tetraacetic acid (DOTA) were covalently attached to the antibody by a stable amide linkage and by a maleimidocysteineamido side chain (MC-DOTA) that has been shown to be chemically labile at physiological temperature and pH. Biodistributions of both 111In and 90Y labels were obtained in these studies. At common biodistribution time points, it was found that the 111In label had greater uptake in the liver than 90Y for both conjugates. No significant differences were found with respect to bone uptake of 90Y using either chelate. Blood curves were generally lower at comparable time points for MC-DOTA, indicative of faster clearance as compared to DOTA. Tumor uptake was high for both conjugates (57-68% ID/g at 48 h), with a longer tumor residence time in the case of the DOTA conjugate, probably a result of its longer blood circulation times. We conclude that bone uptake of 90Y would be minimal if either DOTA or MC-DOTA were used as the bifunctional chelator. This would imply preference for these macrocyclic ligands if radiation doses to the bone marrow would be considered to be dominated by skeletal uptakes. Alternatively, if bone marrow radiation dose is dominated by circulating antibody, the chemically labile linker system employed by the MC-DOTA conjugate offers the advantage of enhanced blood clearance.

We have synthesized two bifunctional derivatives of the macrocyclic chelating agent 1,4,7,10-tetraazacyclododecane-N,N',N",N"-tetraacetic acid (DOTA) equipped with maleimide groups for conjugation to reduced disulfide bonds of monoclonal antibodies. Using water-soluble carbodiimide chemistry, DOTA was coupled to L-cysteine to incorporate both a "pendant-type" carboxyl group for metal coordination and an orthogonal thiol group for protein attachment. The homobifunctional reagent 1,6-bis(maleimido)hexane was then used to introduce the maleimide functionality via a sulfide linkage to the macrocycle, and alternatively, the sulfide group was converted to a sulfone side chain. Both maleimide derivatives were conjugated to the anticarcinoembryonic antigen chimeric monoclonal antibody cT84.66 after light reduction of the mAb with dithiothreitol. In this manner, antibody conjugates were prepared which afforded near-quantitative labeling with the radiometals 111In(III) and 90Y(III) as well as quantitative immunoreactivity. Radioimmunoconjugates prepared with the sulfide and sulfone compounds exhibited relatively rapid linker-dependent radiometal loss when incubated in human serum and aqueous solutions at physiological temperature and pH. The unconjugated maleimidocysteineamido-DOTA derivatives and their Y(III) complexes were incubated in aqueous solution at 37 degrees C, and the resulting decomposition products were analyzed by HPLC and mass spectrometry. These studies revealed that the two bifunctional chelating agents underwent linker-specific cleavage reactions which were considerably faster at pH 7.4 than at pH 5.4. The chemically labile linker systems are expected to release chelated radiometal from mAb conjugates in a pH-dependent manner. This property may impart favorable tumor uptake and normal tissue clearance on radioimmunoconjugates prepared with these reagents, on the basis of the observation that many solid tumors are significantly more acidic than normal tissues.

We have developed a method for attachment of the macrocyclic chelating agent 1,4,7,10-tetraazacyclododecane N,N',N",N"'-tetraacetic acid (DOTA) to proteins by activation of a single carboxyl group with N-hydroxysulfosuccinimide (sulfo-NHS). The sulfo-NHS active ester of DOTA was prepared in a single step using 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide (EDC), and DOTA conjugates of cytochrome c and the anti-carcinoembryonic antigen chimeric monoclonal antibody cT84.66 were prepared by adding the DOTA active ester reaction mixture to the proteins at pH 8.5-9.0. Mass spectrometry of the cytochrome c conjugates showed that as the molar ratio of DOTA active ester to protein in the reaction mixture was increased from 10:1 to 100:1, the average number of chelators attached to the protein molecule increased from 2.64 to 8.79. When DOTA active ester reacted with the antibody at a molar ratio of 100:1, the conjugate averaged 3.8 chelates per antibody. Immunoreactivity of the antibody conjugate radiolabeled with 111In(III) and 90Y(III) remained quantitative. Variation of the DOTA:sulfo-NHS:EDC activation stoichiometry from 2:2:1 to 10:10:1 revealed that the kinetic stability of the radioconjugates increased as the molar ratio of carbodiimide, relative to DOTA and sulfo-NHS, was decreased. Radiolabeling of the protein conjugates with 111In(III) and 90Y(III) proved to be sensitive to pH, buffer, and temperature effects. The optimum pH for the labeling reaction was different for each protein and may be related to the isoelectric point of the protein. Radiometal incorporation at high specific activity was accomplished in acetate and Tris buffers, but the presence of citrate inhibited the labeling reaction. Increasing the temperature of the radiolabeling reaction from 25 to 43 degrees C greatly increased both the efficiency of radiometal incorporation and the kinetic stability of the radioconjugates. Stability studies of the conjugates in human serum and in the presence of a 5000- to 250,000-fold excess of diethylenetriaminepentaacetic acid (DTPA) demonstrated that the radiolabeled proteins are kinetically inert under physiological conditions. In serum, the 111In(III)-labeled antibody showed a rate of radiometal loss of approximately 0.08% per day. In the presence of excess DTPA, both conjugates lost 111In(III) at a rate of about 0.3% per day. No loss of 90Y(III) from the conjugates was observed in serum, but in excess DTPA, both 90Y(III) labeled proteins showed a rate of radiometal loss of approximately 0.2% per day. Therefore, kinetic analysis of metal loss from a radiolabeled immunoconjugate in the presence of a vast excess of DTPA may provide a better indication of the in vivo stability of that immunoconjugate than serum stability studies.

The monoclonal anti-idiotypic antibody (maId) 6G6.C4, directed against the carcinoembryonic antigen (CEA)-specific T84.66 immunoglobulin, was recently shown to act as a surrogate antigen for CEA in experimental animals. In this report, we have extended our studies. 1) The kinetics of complex formation in this CEA-specific idiotypic cascade were investigated using Biosensor technology. 2) Bacterial expression studies show that the mimicked epitope can be delimited to the A3 domain of CEA. 3) We cloned and characterized the genes coding for maId 6G6.C4. 4) Comparison of this epitope-bearing domain with the hypervariable region sequences of 6G6.C4 yields substantial amino acid similarity. Sequence homology between maId 6G6.C4 and anti-CEA antibodies binding to the T84.66 epitope led us to investigate the interaction of maId 6G6.C4 and CEA. Surprisingly, 6G6.C4 specifically binds to CEA but not CEA-related antigens in Western blots. 6G6.C4 and T84.66 recognize different epitopes on CEA. Our results suggest that the T84.66 epitope functionally mimicked by maId 6G6.C4 may be involved in the homophilic binding between CEA molecules, and that heterophilic interactions in the CEA-family are mediated by different binding sites. A model for the intermolecular adhesion of CEA is presented.

Carcinoembryonic antigen (CEA) is one of the better-studied oncodevelopmental antigens to which numerous monoclonal antibodies (MoAbs) have been generated. Many of these MoAbs have been recently grouped (Gold classification) according to their epitope recognition. The present study was designed to immunocharacterize various MoAbs, each representative of the five Gold groups, on colorectal cancers and normal tissues using semiquantitative immunohistochemistry. Sensitivity, based on the number of colorectal cancer cases (n = 100) with positive reaction (> 5% cells), was greater with Gold groups 1 and 2 (93% each) than with groups 3, 4, and 5 (78%, 83%, and 87%, respectively). The intensity of the strain also correlated with the Gold groups, with 24%, 20%, and 18% of cancer cases displaying weak or negative staining (0 or 1+) when reacted with MoAbs from groups 3, 4, and 5, respectively, versus 6% and 12% with Gold 1 and 2 antibodies. Cross-reactivity of the anti-CEA antibodies with CEA-related molecules was found to be significant with Gold 5 antibody, which stained most of the normal lung, liver, stomach, and intestinal tissues tested. Strong staining also was seen in granulocytes when they were reacted with Gold 4 and 5 antibodies. The other antibodies showed much less and variable cross-reactivity with normal tissues, with a not statistically significant advantage for Gold 1 antibodies. In addition to lower sensitivity in CEA detection, Gold 3 to 5 MoAbs were less specific due to cross-reaction with one or more of the CEA-related macromolecules expressed by normal tissues. Based on these results and given the broad clinical applications of anti-CEA MoAbs, it is essential to characterize each MoAb to be used for clinical purposes in order to avoid interpretation errors of potential relevance resulting from poor sensitivity/specificity. The use of antibodies that recognize the epitope of group 1 or 2 is recommended to maximize sensitivity and specificity for CEA detection.

We have previously shown that the colon carcinoma (LS174T) xenografts that emerged shortly after radioimmunotherapy with 90Y-labeled anti-CEA monoclonal antibody (MAb) ZCE025 lacked significant expression of CEA in comparison with the untreated tumors. The present study was designed to establish if the immunophenotype of the treated tumors was the result of CEA specific therapy and if the effect was permanent. Athymic mice bearing LS174T tumors were treated either with 120 mu Ci of 90Y-ZCE025, an equal dose of 90Y-96.5 (nonspecific MAb), or received no treatment. When the treated tumors grew to approximately 1.5 cm in diameter (6 weeks after therapy), they were resected and aliquoted to be transplanted to other mice, plated in tissue culture, fixed in formalin, and homogenized for CEA quantitation. The procedure was repeated 3 times (a total of 4 months after treatment). The CEA content was evaluated 2 and 6 weeks after therapy and when the tumors were transplanted. We confirmed a 4-fold decrease of CEA in the resurgent tumors 6 weeks after specific 90Y-ZCE025 therapy, which was twice the decrease experienced by the tumors treated with nonspecific 90Y-96.5, indicating substantial and specific killing of CEA-expressing cells. The CEA content slowly but progressively increased with each new pass of the tumor in the mice, reaching approximately one-half the value of the controls at the end of the study. The resurgent tumors were also studied by immunohistochemistry with MAbs detecting different epitopes of CEA, keratin, TAG-72, and epithelial membrane antigen to evaluate possible additional immunophenotypic changes induced by radioimmunotherapy. Only the expression of TAG-72 (recognized by MAb B72.3) increased immediately after therapy, but it returned to the original levels by the end of the study. These results suggest that: (a) specific radioimmunotherapy with 90Y-ZCE025 selectively kills cells that express higher levels of CEA; (b) the immunophenotype of the surviving fraction of the tumor appears to slowly revert to its original form; and (c) other tumor markers unrelated to CEA can also be affected. These observations have important implications for the design of radioimmunotherapy trials.

A novel transition metal complex, Venus flytrap cluster (VFC), is described for the preparation of radio-labeled antibodies. VFC contained 57Co, which was held tightly between the faces of two covalently bridged carborane ligands by cluster bonding of the metal with appropriate ligand orbitals. Anti-carcinoembryonic antigen monoclonal antibody T84.66 was conjugated to 57Co-VFC with full retention of immunological activity. Biodistribution studies in nude mice bearing carcinoembryonic antigen-producing tumors showed excellent tumor localization of 57Co-VFC-T84.66. The accumulation of radionuclide in normal liver was low and independent of dose, which may reflect the stability of the radionuclide complex. These results presage the use of VFC systems for binding transition metals that are clinically useful for radioimmunodiagnosis and radioimmunotherapy.

Genes encoding the four principal polypeptide domains (N, A1-B1, A2-B2, and A3-B3) of carcinoembryonic antigen (CEA) were synthesized and expressed in Escherichia coli as fusion products with bacterial CMP-KDO synthetase (CKS). The four synthetic fusion proteins were purified in high yield and used as targets in Western blots for 11 anti-CEA MAbs and to compete with immobilized CEA for binding to four of these MAbs. Each of the MAbs showed strong binding to one or more of the fusion proteins. In Western blots, MAbs H19C91 and 4230 bound only to CKS-N. MAbs H8C2 and H11C35 bound only CKS-A1-B1, and MAbs T84.66, H46C136, and H21C83 appeared to be specific for CKS-A3-B3. None of the MAbs tested bound only to CKS-A2-B2. However, two MAbs bound both CKS-A1-B1 and CKS-A3-B3 and one MAb (3519) bound to all three of the repeated domains. Since these three domains exhibit over 90% amino acid sequence homology, the latter results were not surprising. The competition studies largely confirmed the results of Western blots but did show some MAb-fusion protein interactions not observed in Western blots. These competition studies also allowed estimation of the relative affinities of the MAbs for the synthetic domains and for native CEA. These studies demonstrated that epitopes in CEA recognized by the MAbs in this study are peptide in nature and that the fusion proteins are of utility in the localization of the epitopes on the polypeptide chain of CEA.

Carcinoembryonic antigen (CEA) is one of the best characterized tumor-associated antigens and is extensively used in the in vitro immunodiagnosis of human colon adenocarcinomas. Among a number of anti-CEA monoclonal antibodies, the murine monoclonal antibody T84.66 shows the highest specificity and affinity for CEA and has been used successfully for in vivo tumor imaging in mice and humans. We report here the cloning and sequencing of the genes coding for monoclonal antibody T84.66 and the amino acid sequence of the variable regions for the heavy and light chains. We also report the construction of mouse/human chimeric IgG1 antibody genes using T84.66 variable region genes and human constant region genes. The resulting chimeric gene constructs were transfected into murine myeloma cells (Sp2/0) by electroporation and into Chinese hamster ovary cells by lipofection. The chimeric antibodies obtained exhibited the same specificity and affinity for CEA as that of the T84.66 immunoglobulin produced by the murine hybridoma cell line. Antibody concentrations in culture medium supernatants were clonally variable but similar (15-480 ng/ml) for both Sp2/0 and Chinese hamster ovary transfectants; the average production by Chinese hamster ovary transfectants was only 3-5-fold less than Sp2/0 transfectants. Ascites production of Sp2/0 transfectants is sufficiently high (900 micrograms/ml) for initial in vivo studies with humanized T84.66.

In these measurements, we quantitated, via surgical samples, human primary tumor uptake of the anti-carcinoembryonic antigen monoclonal antibody T84.66. Uptake was measured in units of percentage of injected dose/kg with 111In as the radiolabel. All 11 colorectal lesions were nonnecrotic and were visualized upon scanning. Tumor volume was calculated using the three orthogonal dimensions as described by pathology. Uptake mean +/- SD was 6.55 +/- 3.55% injected dose/kg with a range of 1.2 to 10.4% injected dose/kg. Lesion mean volume was 36 cm3 with a range of 1.5 to 304 cm3. Using mean values, assuming no biological clearance and that the biodistribution of the monoclonal is independent of its radiolabel, the predicted human tumor 90Y beta dose was 0.12 Gy/mCi. Therefore a 10-Gy tumor dose would require 83 mCi of i.v. activity. Using least and highest uptake results, requisite activity values were 3-fold larger and smaller respectively. Thus, there was approximately an order of magnitude variation in the amount of 90Y predicted to achieve a given tumor dose in colorectal cancer patients. Murine and human uptake values were consistent if lesion mass and carcinoembryonic antigen content were taken into account.

Over a 4-year period, 108 patients with known or suspected colorectal cancer were studied by radioimmunoconjugate scintigraphy prior to operative procedures. Study subjects received 0.2 to 40 mg i.v. of murine anti-carcinoembryonic antigen monoclonal antibody labeled with 2-5 mCi of 111In (Indacea). Resected tissues were analyzed for 111In and carcinoembryonic antigen content. Tumor, liver, and draining lymph nodes had over 10% injected dose/kg compared to less than 2.5% injected dose/kg for other normal tissues. Primary tumors that were successfully imaged were significantly larger and had higher 111In and carcinoembryonic antigen content. In 54 patients, primary tumors were visualized with a sensitivity of 78%. Hepatic metastases (58 patients) were visualized as negative filling defects (sensitivity, 45%). Extrahepatic (intraabdominal) metastases (25 patients) were visualized (sensitivity, 48%) as areas of increased uptake. Extraabdominal metastases were uncommon (10 patients; sensitivity, 80%). Of 56 patients with known or suspected hepatic metastases who presented with no evidence of extrahepatic disease by conventional tests (X-ray, computerized tomographic scan), 20 (36%) were documented to have extrahepatic metastases at exploratory surgery and 10 of these (50%) had the extrahepatic disease localized by the Indacea scan. The management of these 10 patients was, or could have been, modified by the scan findings and unnecessary surgery eliminated.

The effects of external beam radiation on tumor uptake of radiolabeled monoclonal antibody were investigated. Nude mice bearing carcinoembryonic antigen (CEA)-producing subcutaneous human tumor xenografts (LS174T) were irradiated (60Co) with a single fraction of 0, 2 or 20 Gy, 6 or 11 days after tumor inoculation. An Indium-111 labeled anti-CEA monoclonal antibody (T84.66) was injected 1-2 hrs after irradiation. Biodistribution studies performed at 48 hrs showed a statistically significant (p less than 0.01) inverse correlation between tumor mass and tumor antibody uptake. Tumor age was also a significant factor with 11 day old tumors having significantly less uptake (p less than 0.0001) compared to 6 day old tumors for a given mass. Radiation increased tumor antibody uptake only in those tumors where growth inhibition also occurred. Multiple regression analysis showed that this inverse correlation between tumor mass and antibody uptake was the same for irradiated and nonirradiated tumors. We conclude that, in this model system, radiation does not act independently to enhance tumor antibody deposition. Radiation's primary effect is to reduce tumor mass, with mass reduction then resulting in an increase in antibody uptake. The clinical implications of this study are discussed.

The epitope specificities of monoclonal antibodies against carcinoembryonic antigen (CEA) were determined in a solid-phase, biotin-avidin enzyme immunoassay. Constant amounts of biotin-labeled antibodies were incubated with the immobilized antigen in the presence of decreasing amounts of unlabeled antibodies. The biotinylated antibodies bound to the antigen were then reacted with avidin-peroxidase conjugates. The activity of the bound peroxidase was determined by the use of o-phenylenediamine and hydrogen peroxide.

Patients with primary and/or metastatic colorectal cancer who had been scheduled for operative intervention were injected intravenously with 200 micrograms of a high-affinity anti-carcinoembryonic antigen (CEA) monoclonal antibody labeled with 2 mCi of 111-indium (Indacea). Patients were imaged by gamma camera at 24 and 48 hours. Primary tumors were identified in 3/10 cases and were not visualized in 3/10 cases. Four scans were considered equivocal. Hepatic metastases were identified as image defects in 5/13 cases and were not visualized in 8/13 cases. All tumors contained CEA by immunoperoxidase staining. In all cases, the primary tumor uptake (5.44 +/- 1.07% ID/kg) was much higher than the uptake of the adjacent fat (0.18 +/- 0.04% ID/kg). There was a direct correlation between tumor CEA content, tumor radioactivity, and the imaging of primary tumor by Indacea. High liver uptake (30.3 +/- 3.0% ID/kg), seen when scanning all patients, was the main limitation of imaging and led to photopenic visualization of hepatic metastases. These results suggest that selection of patients with colorectal carcinoma on the basis of tumor CEA content will lead to improved rates of tumor imaging by Indacea in post-surgical scanning.

A systematic approach for the determination of epitope specificities of monoclonal antibodies to a complex antigen system is described. After initial screening to identify antigen-binding monoclonal antibodies, one or more of the clones are isolated by limiting dilution cloning, grown in ascites, and the resulting antibodies secreted into the ascitic fluid are affinity purified on Sepharose-bound protein A, radiolabeled, and cross-compared with antibodies from other clones by a solid-phase competitive immunoassay. In this work, BALB/c mice were immunized with either purified carcinoembryonic antigen (CEA) or the CEA-producing cell line HC 84S. Spleen cells were fused with the mouse myeloma cell line Sp2/0-Ag14. The supernatants from 25 hybrids showed a significant binding of 125I-CEA (greater than or equal to 15%). Nine hybrids were cloned, resulting in 33 different clones. The antibodies produced by the different cloned hybrids and the remaining uncloned hybrids recognized a total of five different epitopes on CEA. All of the epitopes reside on the protein moiety of the molecule as determined by antibody binding to deglycosylated CEA. The monoclonal antibodies with five different epitope specificities were reacted with tissue sections of normal and cancerous tissues and with peripheral blood smears. Each of the five monoclonal antibodies reacted with tissue sections from colonic, gastric, lung, and mammary carcinomas, as well as from a benign colonic polyp and a resection margin from a colonic carcinoma. Four monoclonals reacted with normal liver tissue. Granulocytes in peripheral blood smears bound three antibodies strongly and one antibody weakly, and one antibody was not bound. One monoclonal antibody that reacted with normal liver tissue was not bound by granulocytes. The ability of these five monoclonal antibodies to differentially detect three different CEA-related antigens in normal and malignant tissues may have clinical utility.

Plasmids have been constructed that direct the synthesis in Escherichia coli of heavy chains and/or light chains of an anti-carcinoembryonic antigen (CEA) antibody. Another plasmid was constructed for expression of a truncated form of heavy chain (Fd' fragment) in E. coli. Functional CEA-binding activity was obtained by in vitro reconstitution in E. coli extracts of heavy chain or Fd' fragment mixed with extracts containing light chain.

Splenocytes from mice immunized with purified, papain-solubilized HLA B27 antigen and/or human lymphocytes bearing the B27 specificity were fused with myeloma cell lines NSI or Sp2. The screening strategy employed a protein A binding assay in which various target cells were used. First, the hybrid cell supernatants were screened against B lymphocyte cell lines of known HLA specificities and the Daudi cell line, which does not express HLA-A, B, or C antigens. Second, a panel of PBLs were used as target cells. It was necessary to refine the protein A binding assay by preabsorbing the radiolabeled protein A with PBLs and by precoating the test wells with ovalbumin. Clones selected by these criteria were further tested by indirect immunoprecipitation and by inhibition of binding or microcytotoxicty to target call lines with purified HLA antigens or beta 2m. Forty-four clones were selected which showed varying degrees of specificity for allo- and nonallo-specific determinants and one clone was selected which was specific for beta 2m. Clone 27M1 which was previously shown to be specific or HLA-B27 as judged by conventional microcytotoxicity testing (Grumet et al., Lancet No. 8239, II:174, 1981) was compared with other clones using the above parameters for evaluation. Antibody from clone 27M1 showed preferential binding to B27 positive cell lines and PBLs, lesser binding to B7 positive target cells, and no binding to B40 positive target cells. Purified B27 antigen (papain) from two sources including the B27 target cell line, was able to inhibit the binding of antibodies from 27M1 to target cells. The extension of the protein A binding assay to PBLs has made it possible to more accurately quantitate the binding or inhibition of binding of antibodies to panels of PBLs.

Inverse correlations of tumor uptake (u), measured in percent injected dose per gram, with tumor mass (m) are demonstrated for phospholipid vesicle, nonspecific and specific monoclonal antibody tracers. Correlation coefficients implied u = B mA in 11 different animal experiments. Experimental exponent (A) values lay in the range -0.28-0.64 with a mean of -0.43 while intercept (B) values varied from 3 to 18. Spherical and cylindrical tumor models implied exponents of -0.33 and -0.5, respectively. Comparison of three implantation sites of the human LS174T xenograft revealed a narrow range of exponents (-0.38- -0.46) indicating a consistent geometry for this tumor. Blood flow to the lesion site and inside its volume (as dictated by tumor size) are factors in tumor uptake. Our results indicate that biodistribution data should include the variation of tumor uptake with mass. For less than 10 g lesions, we predict that radiation absorbed dose will be highly dependent upon tumor size.

A general method is described for the determination of affinity constants and antigen cross-reactivities of monoclonal antibodies. The method employs biotin-labeled antibody, radiolabeled antigen, and avidin as a precipitating agent in a homogeneous phase, competitive radioimmunoassay. This method eliminates incomplete or variable precipitation of antigen-antibody complexes often encountered in immunoassays in which monoclonal antibodies are employed. Using this assay system, we were able to rapidly determine the affinity constants for a number of monoclonal antibodies elicited to carcinoembryonic antigen (CEA). In the preceding paper it was shown that five of the monoclonal antibodies recognized distinct epitopes on CEA. In antigen-binding experiments with these five monoclonal antibodies, the percent of radiolabeled CEA bound in antibody excess ranged from 30 to 92%. The CEA cross-reacting antigens, normal cross-reacting antigen (NCA), and tumor-extracted, CEA-related antigen (TEX) were significantly bound by one, and to a lesser degree, by two of the five antibodies. Two antibodies did not bind significant amounts of NCA or TEX. In inhibition studies, the amount of unlabeled CEA leading to 50% inhibition of 125I-labeled CEA-binding was in the range of 3.7 to 760 ng per tube. The amount of TEX showing the same degree of inhibition was 23-fold greater than the amount of CEA for two antibodies and 351-fold greater than the amount of CEA for a third antibody. The affinity constants for CEA were in the range of 1.0 x 10(8) to 5.1 x 10(10) M-1. The affinity constants for NCA and TEX, determined for one of the antibodies, were three orders of magnitude lower in comparison to CEA. The heterogeneity of radiolabeled CEA as indicated by the low fraction bound by one of the monoclonal antibodies is shown to be most probably an artifact resulting from radioiodination damage. The application of the approach described in this report should eliminate the problems most commonly encountered in the determination of affinity constants for monoclonal antibodies or the use of monoclonal antibodies in competitive, homogeneous-phase immunoassays.

Patients with primary, recurrent, or metastatic colorectal adenocarcinoma were given injections of 200 micrograms of anticarcinoembryonic antigen (CEA) monoclonal antibody labeled with 2 mCi of 111In (Indacea). Patients were imaged at 24 and 48 h. Celiotomy was performed on 40 patients between 3 and 17 days post-Indacea injection. Of 16 primary tumors, 11 (69%) were imaged. Of six extrahepatic recurrences, none was imaged. Intrahepatic metastases were visualized as negative images in 10 of 24 (42%) patients. On the basis of the activity in tissue expressed as a percentage of the total radioactive dose per kg injected into the patient (% ID/kg), extrahepatic tumors that were imaged using Indacea had a significant uptake of radiolabel in the tumor [5.99 +/- 0.91% ID/kg (SE)] and in the associated normal mesenteric lymph nodes (12.0 +/- 2.4% ID/kg). The CEA content of these tumors was high (13.3 +/- 4.7 micrograms/g), and, histologically, the CEA was located primarily apically or intraluminally. Intrahepatic tumor imaging correlated only with tumor size. The greatest Indacea uptake was seen in normal liver (22.1 +/- 3.2% ID/kg). Low Indacea uptake was seen in fat (0.21 +/- 0.05% ID/kg) and bowel wall (1.11 +/- 0.17% ID/kg). In conclusion, Indacea imaging of colorectal carcinoma is specific for high concentrations of accessible CEA in CEA-bearing tumors or in lymph nodes draining these tumors. The successful clinical use of monoclonal antibodies for tumor imaging and therapy will require careful selection of patients for a number of antigen-related parameters including antigen content and distribution in tumors. This information will only come from careful correlation between image results and tissue analysis. High uptake by normal liver tissue is the major unresolved problem with labeled antibody imaging.

One hundred patients with known or suspected colorectal cancer were studied by radioimmunoconjugate scintigraphy prior to operation. Study subjects received murine monoclonal anticarcinoembryonic antigen labeled with indium 111 (Indacea). Sensitivity of imaging was 76 percent for primary tumors, 44 percent for hepatic metastases, 38 percent for extrahepatic abdominal metastases, and 78 percent for extraabdominal metastases. Seventeen of 46 patients (37 percent) with known or suspected hepatic metastases and no evidence of extrahepatic disease by conventional imaging methods had extrahepatic metastases at exploratory surgery. Nine of the 17 patients had disease accurately predicted by the Indacea scanning. The management of each of these nine patients was, or could have been, modified by the scan findings and unnecessary surgery eliminated. A number of patients without post-operative disease had an unexplained increase in plasma carcinoembryonic antigen level due to production of human antimouse antibody. The addition of excess mouse immunoglobulin to the plasma prior to assay blocked this artifactual increase.

This study was undertaken to determine the effect of tumor size and tumor carcinoembryonic antigen (CEA) content on the uptake of indium 111 (111In)-labeled anti-CEA monoclonal antibody in nude mice bearing xenografts. The tumor cell lines were WiDr, SW403, and LS174T, human colon cancer derivatives. The murine breast carcinoma cell line EMT-6 was used as a control. Tumor CEA levels (ng/g of tumor +/- standard error of the mean [SEM], measured by enzyme immunoassay (EIA) were: EMT-6, 0; WiDr, 105 +/- 5.7; LS174T, 2052 +/- 198; SW403, 17,575 +/- 1,785. The 111In-labeled monoclonal antibody was injected intravenously into mice bearing a single tumor. At 48 hours postinjection, scintiscan was performed, and the mice were killed so that biodistribution studies could be performed. The uptake of the monoclonal antibody was expressed as percent injected counts per minute per gram of tissue +/- SEM. The non-CEA-producing tumor, EMT-6, showed the lowest tumor uptake (1.4 +/- 0.3). WiDr, an intermediate CEA-producing tumor, showed some tumor uptake (16.4 +/- 1.5). The high CEA-producing tumors, SW403 and LS174T, had high tumor uptake (29.5 +/- 5.0 and 51.1 +/- 6.1, respectively). Biodistribution and scintiscan quality were closely related. Although LS174T had the best tumor uptake, SW403 had the highest CEA tumor content, indicating tumor CEA content cannot entirely predict scintiscan and biodistribution results. Tumor-to-blood (T/B), tumor-to-liver (T/L), and liver-to-blood (L/B) ratios were calculated for each animal and compared with tumor size. It was found that T/L had a negative correlation with tumor size (r = -0.72) and L/B had a positive correlation with tumor size (r = 0.94). These ratios may be useful clinically to follow response to therapy.

A monoclonal antibody to carcinoembryonic antigen (CEA) labeled with 111In (Indacea) was used to image tumors in patients with colorectal cancer. The anti-CEA antibody has a high affinity (2.6 X 10(10) M-1) for CEA and does not cross-react with normal cross-reacting antigen, biliary glycoprotein-1, or tumor-extracted, CEA-related antigen. During the course of these studies, it was noted that a significant number of male patients (20 of 27, 74%) showed uptake of Indacea in the testes. In order to determine if the Indacea uptake was specific, 20 testicular specimens were analyzed by immunohistological methods using five different anti-CEA CEA monoclonal antibodies recognizing five different epitopes on CEA. In 18 cases (90%) germ cells were uniformly stained by all five antibodies. Fresh frozen testis tissue was homogenized in water and precipitated with 1.0 M perchloric acid. The supernatant contained a CEA-like material as measured by an enzyme immunoassay specific for CEA. The same supernatant was radiolabeled with 125I and immunoprecipitated with anti-CEA monoclonal antibodies. Sodium dodecyl sulfate:polyacrylamide gel electrophoresis of the immunoprecipitates revealed a single species (Mr 180,000) which was indistinguishable from CEA. This study documents the first description of CEA in the germ cells of normal testis. The CEA in the testis was accessible to circulating monoclonal antibodies in the majority of male patients tested.

We produced a monoclonal antibody to the human mu-chain. Its specificity was proved by amino acid-sequence analysis of immunoabsorbent purified cell surface antigen. In binding studies an affinity constant (K-aff) of 5.7 X 10(9) M-1 was determined. When compared with a polyclonal anti-IgM antiserum it was found that this antibody may react with a determinant of the mu-chain differently exposed in clonal B-cell disorders rendering it to a reagent of particular diagnostic interest.

Tumor imaging and biodistribution of an indium-labeled monoclonal antibody (MAB) to carcinoembryonic antigen (CEA) [anti-CEA MAB-diethylenetriaminepentaacetic acid (DTPA)-111In] have been investigated using LS174T human colon cancer xenografts in nude mice. Antibody specificity, dose, and specific activity were examined with respect to tumor uptake and quality of scintiscans at different times following injection. The CEA-bearing LS174T tumors were imaged specifically with anti-CEA MAB-DTPA-111In. Using 62.5 ng of indium-labeled MAB (50 microCi/micrograms) the ratio of activity in tissue expressed as a percentage of the total radioactive dose injected into the animal per gram tissue for tumor:blood increased from 0.66 +/- 0.02 (SE) at 1 h to 14.8 +/- 1.1 at 72 h. Scintiscan quality improved with the rise in tumor:blood ratio until 48 h. At longer intervals insufficient counts remained for imaging. The tumor:blood ratio and the scintiscan quality were not improved by increasing the MAB dose to 625 or 6250 ng but good images were obtained at longer times postinjection. By decreasing the 111In from 50 to 10 microCi/micrograms of MAB, the unbound 111In was decreased from 7 microCi/micrograms (14%) to 0.2 microCi/micrograms (2%). Even with the lower specific activity (9.8 microCi/micrograms) of the 10-microCi/micrograms preparation, scintiscan quality at the 62.5-ng dose was maintained. This anti-CEA MAB-DTPA-111In preparation was stable, retained immunological activity, did not require column chromatography to remove unbound 111In, was specific for a CEA-bearing tumor, and was effective for tumor imaging over a wide range of antibody doses (3 to 300 micrograms MAB/kg body weight). This anti-CEA MAB-DTPA-111In preparation is feasible and practical for imaging CEA-bearing tumors in humans.

Twenty patients with 21 primary colorectal adenocarcinomas were studied with 2 mCl (7.6 X 10(7) becquerels) of indium-labeled monoclonal antibody (200 micrograms) specific for carcinoembryonic antigen (CEA). Fifteen lesions (71%) were visualized by gamma camera scintigraphy at 48 hours postinjection. Tumors that were identified by immunoscintigraphy were large (38.10 +/- 17.76 cm3 vs 6.00 +/- 1.65 cm3), had a grossly fungating component, had a high content of CEA by enzyme immunoassay (12.9 +/- 3.6 micrograms/g vs 3.3 +/- 1.7 micrograms/g), and had an apical and/or intraluminal staining pattern on immunohistologic section. Patients whose tumors were visualized had a low serum CEA level (1.9 +/- 0.4 ng/mL vs 14.6 +/- 8.0 ng/mL). Prospective selection of patients for follow-up imaging or therapy with radiolabeled monoclonal antibodies may be feasible using these measurements.

Retrospective analysis of 108 patients who received indium 111-labeled murine monoclonal antibodies for imaging of cancer was performed. Most patients had operative procedures for colorectal carcinoma following completion of scintiscanning. Eleven patients had markedly elevated carcinoembryonic antigen (CEA) levels postoperatively without evidence of residual or recurrent disease. The laboratory method of measuring CEA levels was a commercially available double mouse monoclonal antibody enzyme immunoassay. It was postulated that the unexplained elevation of CEA was a reflection of the presence of human anti-mouse antibody (HAMA) induced by the administration of radiolabeled mouse antibody. A competitive assay for HAMA was undertaken by incubation of these patients' sera with a high dose of nonspecific mouse immunoglobulin prior to CEA determinations, and subsequent CEA levels were normal. The presence of HAMA was confirmed by a noncompetitive solid-phase enzyme immunoassay in 73% of tested patients who received murine monoclonal antibodies for imaging. Identification of artifactual CEA elevations is important in the treatment of cancer patients.

A new method has been developed for conjugating diethylenetriaminepentaacetic acid (DTPA) to proteins using the N-hydroxysuccinimide active ester of DTPA. The DTPA-active ester was prepared using diisopropylcarbodiimide in a simple single step synthesis. DTPA-conjugated proteins were prepared by adding the DTPA-active ester reaction mixture to protein solutions (5 mg/ml) buffered at pH 7.0 and purified by Sephadex G-50 chromatography. A monoclonal antibody directed against carcinoembryonic antigen was reacted with four different amounts of the DTPA-active ester. Solid-phase enzyme immunoassay showed that the immunological activity of the antibody conjugate was not altered when the active ester: antibody molar ratio was 36:1 or 72:1; however, it decreased when the ratio was 180:1 or 360:1. The antibody heavy and light chains had slightly decreased electrophoretic mobilities when analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, a result consistent with the covalent attachment of DTPA to the protein. Sephadex G-200 chromatography showed that the native and conjugated antibodies were the same size. When the DTPA-conjugated antibody was incubated with 10, 50, and 100 microCi of 111In/micrograms of protein, specific activities of 9.8, 43.1, and 56.3 microCi/micrograms were obtained. Enzyme immunoassay and radioimmunoassay of the 111In-labeled antibody showed that it retained its full immunological activity. The high specific activity of the 111In-labeled antibody makes it suitable for imaging carcinoembryonic antigen-bearing tumors using low doses of antibody.

Administration of a large dose (0.2 mg) of unlabeled specific anticarcinoembryonic antigen (anti-CEA) monoclonal antibody (MAB) to nude mice bearing LS174T human colon cancer xenografts significantly decreased normal liver uptake of 111In-labeled anti-CEA MAB (Indacea). Mice bearing tumors of approximately 1 g showed liver accumulation of indium-111 at 48 h following injection of 2 micrograms/10 microCi Indacea of 33.8 +/- 1.5% injected dose per gram (%ID/g) (N = 25). Treatment with 0.2 mg unlabeled anti-CEA MAB reduced this to 8.9 +/- 0.5% ID/g (N = 22; P less than 0.001). The dose of pretreatment was found to be critical. Increasing the amount of unlabeled MAB to 2.0 mg did not significantly improve the liver level of indium-111, but did compromise the tumor uptake of Indacea (15.9 +/- 1.3 versus 12.4 +/- 0.4% ID/g; P less than 0.05). Lowering the dose of pretreatment 10-fold resulted in increased (P less than 0.001) liver uptake of the label (26.5 +/- 2.8% ID/g). The unlabeled anti-CEA MAB treatment given as a single dose or fractionated over several days gave the same results. The decrease in liver uptake was the same for i.v. administration of the unlabeled MAB given 1 week prior to Indacea injection or mixed together with Indacea. With i.p. administration, simultaneous injection of the unlabeled MAB with Indacea was not as effective as pretreatment (20 min to 7 days) in decreasing the liver uptake of 111In (P less than 0.05). Epitope specificity and affinity were shown to be important considerations in the choice of MAB combinations used for pretreatment and imaging. Pretreatment with nonspecific MAB was ineffective in decreasing liver uptake of Indacea.