Immunoperoxidase localization of carcinoembryonic antigen in normal human intestinal mucosa.

The indirect, labeled antibody and peroxidase-antiperoxidase complex (PAP) methods were studied to determine their sensitivity in detecting carcinoembryonic antigen (CEA) in conventionally processed specimens of morphologically normal human colon mucosa. CEA-positive staining was demonstrated in 13 of 19 specimens reacted with the PAP method, whereas only 4 of these specimens stained positive with the labeled antibody procedure. Detection of CEA with either technique was unrelated to normal mucosa content of antigen as determined by radioimmunoassay. Tissue fixation in 95% ethanol 1% acetic acid (EA) resulted in an enhanced and defined cytoplasmic staining of the normal colon cell lining the mucosal surface and upper levels of the glandular crypts. Cytoplasmic localization in Formalin-fixed specimens was absent or markedly reduced. Colon goblet cells and the small intestinal epithelium were CEA-negative in both Formalin- and EA-fixed specimens. These results show that the PAP immunoperoxidase method is more sensitive than the indirect, labeled antibody procedure in detecting CEA in morphologically normal colon mucosa. Furthermore, staining of tissues fixed in EA demonstrated that CEA is a product of the columnar epithelial cell and is not associated with goblet cells.

Carcinoembryonic antigen in histopathology: immunoperoxidase staining of conventional tissue sections.

A triple-bridge, indirect peroxidase-antiperoxidase method for demonstrating carcinoembryonic antigen (CEA) in frozen, ethanol-fixed or formalin-fixed, paraffin-embedded specimens was evaluated. Examination of 359 tissue specimens--234 malignant tumors, 37 benign neoplasms, 41 nonneoplastic diseased tissues, and 47 normal specimens--showed that CEA could usually be demonstrated in a group of cancers. We could detect CEA in carcinomas of the stomach, colon, rectum, pancreas, lung, and cervix. However, malignant tumors of the breast, prostate, kidney, larynx, brain, lymphoreticular system, soft tissues, and skin proved negative for CEA by the immunoperoxidase test. CEA could be detected in ethanol- or formalin-fixed sections. The only nonmalignant specimens showing CEA staining were a few benign tumors, the mucosae of some cases of colitis, and the resection margins of 2 cases of colon cancer; however, these were commonly very weak reactions. Measurement of tumor CEA content by radioimmunoassay revealed two causes for this relative specificity of the immunoperoxidase test for CEA:1) a quantitative difference existed in tissue CEA among the various specimens, and 2) the threshold for CEA staining in malignant specimens was usually above that in nonmalignant specimens. An analysis of the formalin-paraffin-treated sections showed that immunoperoxidase-tested CEA positivity reflected CEA levels in tissue of at least 3.0-5.0 mug/g; this permitted retrospective estimates of minimal tissue CEA concentrations in older histopathologic specimens by the immunoperoxidase reaction method. Formalin-paraffin-treated sections as old as 10 years still had demonstrable CEA. Although tumor CEA concentration correlated well with immunoperoxidase staining for CEA, plasma CEA titer did not necessarily reflect tumor CEA content. CEA positivity in primary and secondary tumors was strongly correlated; it was less strongly correlated with level of tumor differentiation.

Representation of epitopes on colon-specific antigen-p defined by monoclonal antibodies.

Colon-specific antigen-p (CSAp) is a large molecular-sized protein restricted to normal and neoplastic gastrointestinal tissues and to some mucinous ovarian tumors. Murine monoclonal antibodies (MAbs) were raised against CSAp that was affinity purified with goat polyclonal antibodies from GW-39 human colonic carcinoma xenografts or against the CSAp-producing colon cancer cell line SW-948. Two of the MAbs, designated Mu-2 and Mu-4, recognized a CSAp determinant containing sialic acid, and this epitope was also expressed on bovine submaxillary mucin (BSM). Blocking experiments demonstrated that the Mu-2 and Mu-4 MAbs recognized different determinants. A third MAb, Mu-3, did not cross-react with BSM, but unlike Mu-2 and Mu-4, it did react with human saliva. Reactivity of Mu-3 with saliva did not correlate with major blood group and Lewis-related secretory blood group substances in saliva. This reactivity was not related to sialylated Lewis activity. The fourth antibody, Mu-1, appeared to react with a conformational determinant since its epitope was destroyed by heat treatment or thiol reduction. Enzyme immunoassays have demonstrated that all four epitopes may be expressed on one molecular species.

Differential expression of tumor-associated antigens in human colon carcinomas xenografted into nude mice.

The tumor-associated antigen profile of a number of different human colon carcinomas xenografted into inbred Swiss nude mice was examined to determine whether the tumors could serve as useful models for antigen purification, radioimmunodetection, and immunotherapy. Extreme heterogeneity was observed both by radioimmunoassay and immunohistochemical procedures for the expression of carcinoembryonic antigen (CEA), colon-specific antigens (CSAp), and colonic mucin antigen (CMA) within the tumors. Four of 10 tumors (DLD-2, DLD-3, DLD-5, and HCT-10) were high producers of CEA (greater than 75 micrograms/g wet tissue wt). Two of these (DLD-2 and HCT-10) correlated with a high production of CSAp and CMA, whereas the other 2 produced low quantities of these antigens. Another tumor, HCT-14-OM1, produced large amounts of CMA yet produced moderate amounts of CEA and low quantities of CSAp. Immunohistochemical analyses of CEA gave a mostly diffuse cellular and cell surface staining for all of the tumors. Staining for CSAp was very focal, as in DLD-5 where only a few of the tumor cells were stained. Staining of CMA was limited; however, DLD-2, HCT-10, and HCT-14-OM1 showed intense cystoplasmic and intraluminal staining. A determination of the tumor antigen profile may be useful in characterization and classification of the tumor as well as enabling the selection of the proper antibody or antibodies for immunodetection and immunotherapy.

Immunological considerations in the use of goat antibodies to carcinoembryonic antigen for the radioimmunodetection of cancer.

Preparation of radioantibodies from carcinoembryonic antigen (CEA) antisera raised in goats by conventional procedures necessitates extensive absorption procedures to remove antibody activity against CEA-related substances. Further purification by affinity chromatography increases the antibody content of the radiolabeled preparation and enhances antibody uptake by a xenogeneic tumor. In patients given injections of CEA radioantibody, immune complexes form with CEA or a human antibody reactive with goat gamma-globulin, but these complexes do not appear to prevent tumor radioimmunodetection. Monoclonal antitumor antibodies from appropriate animal species would be a more suitable source, eliminating many of the problems encountered with conventional antisera.

ABH and Lewis blood group expression in colorectal carcinoma.

The expression of ABO(H) and Lewis blood group antigens on 68 colorectal carcinomas from 63 patients was studied by immunohistochemical staining of tissue specimens. The pattern of antigen expression was as follows: ABH was expressed in normal tissue only in secretors and was expressed in the proximal but not distal colon. In tumors, there was net loss of ABH expression in the proximal and net gain in the distal colon. Some nonsecretor tumor tissue expressed ABH. Lewis expression was similar to but less strong than ABH. Its expression occurred only in secretors and in normal epithelium only in the proximal colon. In tumors there was net loss of antigen expression proximally and net gain in the distal colon. There was no expression of Lewisb in tumors of nonsecretors. Lewisa antigen was expressed throughout the normal colon in secretors and nonsecretors with no discernible difference between proximal and distal colon. In tumors, net loss of expression of Lewisa occurred throughout the colon. No inappropriate blood group expression was observed in this study. With few exceptions, H expression paralleled expression of A and B in non-0 patients. Metastatic tumor antigen expression was similar to that of the primary in most cases. Alterations in antigen expression were not clinically or histologically distinctive.

Immunological heterogeneity of carcinoembryonic antigen: purification from meconium of an antigen related to carcinoembryonic antigen.

Two antigens cross-reactive with carcinoembryonic antigen (CEA) and distinct from the nonspecific cross-reacting antigen were identified in meconium by double immunodiffusion with a conventional goat anti-CEA antiserum. These two antigens together competitively inhibited cross-reacting antibodies against them in CEA radioimmunoassay and contributed to the measurement of meconium CEA levels which averaged 6 times higher than that determined with anti-CEA specific antibody. A purification method for one of these antigens, tentatively designated meconium antigen, is described and uses a combination of ethanol fractionation, ion-exchange and molecular sieve chromatography, and adsorption to an immunoadsorbent containing a cross-reactive murine monoclonal antibody to CEA. Preliminary characterization of the purified meconium antigen showed it to be a glycoprotein, migrating as an alpha-globulin and having a molecular size similar to that of CEA (Mr 185,000 versus 200,000). Antigenically, it lacks at least one determinant present on CEA and differs further from CEA by being weakly reactive with concanavalin A and resistant to proteolytic digestion with Pronase E. Although these properties of meconium antigen suggest that it may be nonspecific cross-reacting antigen 2, additional chemical and antigenic studies are required to establish its relationship to CEA and other CEA-related antigens in meconium.

Immunological heterogeneity of carcinoembryonic antigen: immunohistochemical detection of carcinoembryonic antigen determinants in colonic tumors with monoclonal antibodies.

The immunoperoxidase localization of carcinoembryonic antigen (CEA) determinants was studied in colonic adenocarcinomas using four murine monoclonal antibodies to CEA in a bridged avidin:biotin technique. One of the monoclonal antibodies, NP-1, recognizes a CEA epitope shared with the nonspecific cross-reacting antigen and meconium antigen. Two others, NP-2 and NP-3, discriminate two separate CEA epitopes shared with meconium antigen only, whereas NP-4 reacts with a unique determinant expressed on a subpopulation of CEA molecules. The monoclonal antibodies and polyclonal goat antisera against CEA and nonspecific cross-reacting antigen stained columnar epithelial cells in morphologically normal mucosa. Neutrophils were stained by only the NP-1 monoclonal antibody and goat anti-nonspecific cross-reacting antigen antiserum. All moderately differentiated colorectal adenocarcinomas and most of their nodal and liver metastases reacted with the goat antisera and cross-reactive monoclonal antibodies. Thirty % of these primary tumors and most of the nodal and/or liver metastases from six patients with NP-4-positive primary tumors failed to stain with NP-4. These results suggest heterogeneity in the expression of a CEA variant and/or determinant recognized by the NP-4 monoclonal antibody that perhaps identifies a subgroup of colonic cancers which differ in their functional differentiation.

Comparison of tumor targeting of mouse monoclonal and goat polyclonal antibodies to carcinoembryonic antigen in the GW-39 human tumor-hamster host model.

We have evaluated 4 radioiodinated mouse monoclonal anticarcinoembryonic antigen antibodies (MAbs) by using the GW-39 human colorectal tumor xenograft transplanted i.m. in immunocompetent hamsters to determine whether there were any differences in their tumor localization properties. Additional comparisons were made to affinity-purified goat anticarcinoembryonic antigen antibody. Statistically significant differences were found in the percentage/g of tumor uptake and tumor/nontumor ratios among the antibodies, so that the antibodies could be ranked according to their tumor localization properties (NP-2 greater than NP-4 = goat antibody greater than NP-1 greater than NP-3). Although statistical differences were found, tumor/nontumor values generally were not distinguished by a factor of more than 1.5, suggesting that these differences may not be biologically significant. F(ab')2 fragments of NP-2 were found to be superior to NP-4 F(ab')2 fragments, giving tumor/liver and tumor/blood ratios of 16 and 11.5, respectively, within 3 days, in comparison to 5.4 and 3.8 for NP-4 F(ab')2 fragments. Mixtures of all of the MAbs or a mixture of NP-2 and NP-4 did not improve tumor localization, in comparison to NP-2 alone. These studies suggest that mixtures of these anticarcinoembryonic antigen MAbs may not afford better tumor imaging than the use of a certain single antitumor MAb.

Immunological heterogeneity of carcinoembryonic antigen: antigenic determinants on carcinoembryonic antigen distinguished by monoclonal antibodies.

Murine monoclonal antibodies against carcinoembryonic antigen (CEA) derived from a colonic tumor were analyzed by radioimmunoassay for reactivity with CEA and the CEA-related antigens, meconium antigen (MA) and nonspecific cross-reacting antigen. Antibody-antigen binding profiles revealed three classes of hybridomas. The Class I antibody, NP-1, recognized an epitope shared among all three antigens, and its affinity for CEA and MA was high but low for nonspecific cross-reacting antigen. The Class II antibodies reacted with sites shared between CEA and MA, while those of the Class III type only bound CEA. The Class III antibody, NP-4, bound less than 50% of the CEA molecules recognized by goat specific anti-CEA antibody and the other classes of monoclonal antibodies. Two Class II antibodies, NP-2 and NP-3, bound similar amounts of CEA and MA with moderate but different affinities for CEA. Studies using labeled monoclonal antibodies for CEA epitope blocking revealed that NP-2 and NP-3 recognize two separate epitopes on CEA within the Class II category. Thus, monoclonal antibodies to CEA can differentiate at least four antigenic sites on colonic cancer CEA. One is shared by CEA, MA, and nonspecific cross-reacting antigen; two others are shared by CEA with MA; and a fourth appears specific for a subpopulation of CEA molecules.

Localization of GW-39 human tumors in hamsters by affinity-purified antibody to carcinoembryonic antigen.

With the paired-labeled antibody technique, the in vivo localization of radioiodinate, affinity-purified antibody to carcinoembryonic antigen (CEA) was studied in GW-39, a xenografted, CEA-producing tumor model. When compared to the whole immunoglobulin G fraction, a 4-fold greater tumor accumulation of radioantibody was obtained with affinity-purified specific CEA antibody. The degree of increased tumor localization of affinity-purified antibody was similar to its improved immunoreactivity as observed in radioimmunoassay and binding to CEA immunoadsorbent. Affinity-purified antibody cross-reactive with CEA and colon carcinoma antigen III was as equally effective in tumor localization as was specific CEA antibody prepared similarly. It thus appears that affinity-purified CEA radioantibody will provide a superior tumor-imaging agent for clinical use.

Characterization of the colorectal carcinoma-associated antigen defined by monoclonal antibody D612.

Monoclonal antibody (MAb) D612 recognizes an antigen expressed on the cell surface of normal and malignant gastrointestinal epithelium. It is a murine IgG2a/kappa which has been previously shown to mediate killing of human colon carcinoma cells using human effector cells (which could be enhanced in the presence of interleukin-2). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analyses of MAb D612 immunoprecipitates of extracts of L-[35S]methionine-, L-[3H]leucine-, and D-[3H]glucosamine-labeled human colon carcinoma cells showed that the D612 antigen is a Mr 48,000 glycoprotein. Similar estimates of molecular mass were obtained from SDS-PAGE analyses of MAb D612 immuno-precipitates of radioiodinated extracts of surgically resected colon carcinoma and adjacent normal colonic mucosa. D612 antigen was not detectable in immunoprecipitates of supernatant media from radiolabeled cell cultures, suggesting that the antigen is not readily shed from the surface of cultured cells. The D612 antigen was shown to be clearly distinct from previously described gastrointestinal carcinoma-associated glycoproteins: the D612 antigen shows a migration pattern of SDS-PAGE distinct from those of the antigens recognized by MAbs KS1/4 and GA733, and reciprocal immunodepletion analyses of D-[3H]glucosamine-labeled colon carcinoma cells utilizing MAbs D612 and GA733 revealed no cross-reactivity between these antibodies. Similarly, competitive binding studies between MAbs 17-1A and KS1/4 and MAb D612 revealed no similarity between the epitopes recognized by MAb D612 and MAbs 17-1A and KS1/4. MAbs D612 and 17-1A were also titered in immunoperoxidase staining assays on serial frozen sections of normal and malignant colon. MAb D612 showed a higher titer of immunostaining reactivity with both normal and malignant colon than did MAb 17-1A. MAb D612 showed roughly equivalent immunostaining titers against normal and malignant colon; whereas MAb 17-1A showed higher titer of immunostaining reactivity against the normal colon tissue than against the malignant colon. Flow cytometric analysis of phosphatidylinositol-specific phospholipase C-treated colon carcinoma cells revealed no loss of D612 antigen from the cell surface, suggesting that the mechanisms of attachment of the D612 antigen to the cell surface does not involve linkage to a phosphatidylinositol glycan.(ABSTRACT TRUNCATED AT 400 WORDS)

Relationship of radioantibody localization and cell viability in a xenografted human cancer model as measured by whole-body autoradiography.

The simultaneous distribution of monoclonal 131I-labeled anti-carcinoembryonic antigen (CEA) immunoglobulin (IgG) (NP-2) or 131I-labeled irrelevant myeloma IgG (Ag8) and [3H]thymidine was studied in hamsters bearing transplants of the GW-39 human colon carcinoma by qualitative double-tracer whole-body autoradiography. Autoradiography showed that large solid GW-39 tumors are characterized by heterogeneity of radioantibody retention and uneven [3H]thymidine accumulation, reflecting zonal variations in antibody reactivity and tumor cell proliferation, respectively. The autoradiographic images showed that both 131I-labeled-monoclonal antibody and control 131I-labeled IgG targeted nonproliferating tumor zones, suggesting a mechanism of nonspecific tumor uptake of radioantibodies in these areas. Absence of tumor center labeling with [3H]thymidine, associated with cellular necrosis, was confirmed by histology and microautoradiography in separate animal studies. In confirmation of earlier reports, 131I-labeled anti-CEA monoclonal antibody gave higher tumor-to-non-tumor labeling patterns than did control 131I-labeled IgG, at both 3 and 7 days following treatment. Immunohistochemical localization of CEA in GW-39 tumors with necrotic centers showed the presence of CEA in nonviable cells, but CEA antigen concentrations were diminished as compared to cells located in the tumor's periphery. The results indicate that double-tracer whole-body autoradiography is well suited for studying the kinetics of radioantibody localization in relation to regional tumor cell viability.

Antiidiotype antibodies in cancer patients receiving monoclonal antibody to carcinoembryonic antigen.

The initial 10 patients of a Phase I clinical trial involving multiple injections of murine monoclonal anti-carcinoembryonic antigen (CEA) antibody, NP-2, were studied for the presence in their sera of antiidiotypic antibody. Most patients had advanced gastrointestinal adenocarcinoma and received 1 mg/m2 monoclonal antibody three times weekly, or once a week, resulting in five to 13 injections over 12 to 240 days. Antiidiotype antibody was detected with a blocking radioimmunoassay using [125I]NP-2-F(ab')2 binding to CEA-coated microwells and [125I]NP-4-F(ab')2 as a control antibody. Five out of 10 patients demonstrated 65-96% inhibition of NP-2 binding at 1/20 dilution of serum compared to NP-2 binding in the presence of pretreatment sera. The inhibitory activity was preserved after adsorption over a polyclonal mouse IgG immunoadsorbent whereas exposure to a NP-2 affinity column completely depleted the activity. Specificity testing, including the blocking effect of patient sera on the control antibody NP-4, and interference by the possible presence of circulating NP-2, circulating CEA, and human anti-CEA activity, confirmed that the inhibition observed was specific for NP-2 and was caused by an agent with CEA-like characteristics. Longitudinal studies demonstrated that elevated titers of antiidiotypic antibody appeared later in the course of immunization than did antibody against mouse immunoglobulin. These studies indicate that patients can be sensitized to the idiotype (anti-combining site and/or combining site-related) of monoclonal antibodies to CEA following multiple infusions.

A human colon cancer metastasis model for radioimmunodetection.

One important issue in radioimmunodetection is how well the current methods can locate and disclose small metastatic foci in visceral sites. We have developed a human colonic tumor metastasis model by surgically implanting GW-39 tumor cells in the liver of unconditioned hamsters. Tumors were produced in 71 of 73 animals and were macroscopically apparent within 1 wk. In addition, multiple nodular lung metastases of GW-39 were found in about 80% of the animals given implants of tumor in the liver, but implantation of tumor in the spleen failed to show lung metastases even after 4 wk. Hamsters bearing GW-39 liver and cheek pouch grafts or normal hamsters were given injections of a mixture of 131I-labeled anti-carcinoembryonic antigen antibody and 125I-labeled irrelevant immunoglobulin G. After 7 days, tumor was localized by external scintigraphy without subtraction techniques in both the liver and cheek pouch, but even in animals with extensive lung metastases we failed to unequivocally detect tumor in the lungs by external imaging or by comparing tissue counting data from uninvolved and tumor-bearing lungs. However, whole-body autoradiography confirmed specific localization of anti-carcinoembryonic antigen antibody in the tumors at all sites indicating that tissue counting and external imaging were not sensitive enough to reveal micrometastatic tumors. Thus, the current methods used for this model appear to be useful for further investigation of the radioimaging of tumors growing in visceral organs.

Self-reactive antibody expression by human carcinoma cells engineered with monoclonal antibody genes.

The purpose of this study was to determine if human colon cancer cells transduced with monoclonal antibody (MAb) genes become sensitive to immune destruction through coexpression of both the MAb and its reactive antigen. Murine retroviral expression vectors were constructed with the heavy or light chain genes of an anti-human colon carcinoma MAb, D612, that mediates antibody-dependent cell-mediated cytotoxicity (ADCC). Transduction of D612 MAb genes into the D612 antigen-positive (> 95%) human colon carcinoma cell line, LS-174T, was carried out by sequential cocultivation with PA317 packaging cells producing infectious virions containing the light or heavy chain expression vectors. Six cultures survived drug selection, two of which were found to have elevated levels of both light and heavy immunoglobulin chain activity in their supernatants. IgG secretion levels (24 h) were 1-2 ng/1 x 10(6) cells. Low but definite antigen reactivity was also present in supernatants obtained from these LS-174T transductants. Immunocytochemical staining of transduced tumor cells revealed that > 95% of the cells were positive for IgG expression. Thus, LS-174T transductants were capable of producing both the D612 MAb and D612-reactive antigen. Analysis of transductants by flow cytometry further revealed that > 95% of the cells had murine immunoglobulin on their surfaces. ADCC mediated by human natural killer cells against nontransduced tumor cells was observed when the latter cells were co-cultivated in the presence of transductants producing both D612 heavy and light chains but not in the presence of tumor cells transduced with light chain only. LS-174T cells transduced with both D612 heavy and light chain genes were more sensitive to cytotoxicity mediated by natural killer cells than were light chain gene only transductants. ADCC contributed to the greater sensitivity of the former transductants to cytotoxicity based on its inhibition by anti-FcR gamma III antibody. Thus, these studies demonstrate that tumor cells transduced with genes encoding for MAbs that can participate in ADCC reactions are able to sensitize nontransduced tumor cells to immune destruction as well as to direct killer cells against themselves. These studies may lead to a new immunotherapeutic approach for the treatment of cancer based on MAb gene therapy.

Antibody to carcinoembryonic antigen in hamsters bearing GW-39 human tumors.

The appearance of anti-carcinoembryonic antigen (CEA) activity in the sera of hamsters bearing i.m. heterografts of GW-39, a CEA-producing colonic tumor of human origin, was demonstrated in radioimmunoassay using radioiodinated CEA purified from GW-39. Column chromatography, sucrose gradient ultracentrifugation, and radioimmunoelectrophoresis of pooled sera from tumor-bearing hamsters showed that the CEA antibody was of the immunoglobulin M type. The CEA antibody in tumor sera contained two specificities that reacted with different determinants on the CEA. On specificity reacted with a blood group-related determinant, whereas the remaining specificity was against a unique site on the CEA. Sequential analysis of individual tumor-bearing animals revealed that an increase in tumor growth paralleled the concomitant appearance of circulating CEA and highest antibody titer, suggesting that immunoglobulin M antibody, antigen, or complexes thereof influence tumor growth in this model.

Mice transgenic for human carcinoembryonic antigen as a model for immunotherapy.

Mice transgenic for the human carcinoembryonic antigen (CEA) gene were prepared for use as a preclinical model for immunotherapy. A 32.6-kb fragment containing the complete human CEA gene and flanking sequences was isolated from a genomic cosmid clone and used to produce transgenic C57BL/6 mice. A homozygous line was established that was designated C57BL/6J-TgN(CEAGe)18FJP. Southern blot analysis showed that this line contained intact copies of the cosmid clone, with approximately 19 integrated copies at one chromosomal location. A mouse-human chimeric anti-CEA monoclonal antibody was used to examine CEA expression by immunohistochemical staining of frozen tissue sections. In the cecum and colon, approximately 20% of the luminal epithelial cells had strong cytoplasmic staining, whereas occasional glands showed intense staining. CEA was also expressed in gastric foveolar cells, whereas small intestine villi had only a few (<1%) positive cells. CEA was not found by immunohistochemistry in other tissues of the digestive tract, nor was it found in a wide range of other tissues or organs. Concordance in results was obtained between immunohistochemistry and analysis of tissue extracts by enzyme immunoassay. The lone exception was the testis, which was positive only by enzyme immunoassay. Expression of human CEA was not observed in tissues derived from nontransgenic mice. The fecal content of CEA in transgenic mice was approximately 100-fold less than that observed for humans. Circulating CEA was not detected. A CEA-transfected syngeneic murine colon carcinoma cell line, MC-38, was prepared that had stable expression of CEA in vitro and in vivo. The molecular size of CEA produced by CEA-transfected MC-38 cells and by the colon of transgenic mice was similar to that obtained with CEA purified from human colon tumors. Anti-CEA antibody appeared in nontransgenic but not transgenic mice bearing transfected MC-38 tumors. These findings demonstrate that CEA distribution and its properties in tissues of mice transgenic for the human CEA gene are similar to that observed in human tissues. As in humans, immune responsiveness to CEA, as reflected by antibody formation, was not detectable in transgenic mice bearing CEA-positive tumors. Thus, CEA transgenic mice may serve as a useful model for studying the efficacy and safety of various immunotherapy strategies directed at this tumor self-antigen.

Targeting and therapy of carcinoembryonic antigen-expressing tumors in transgenic mice with an antibody-interleukin 2 fusion protein.

The purpose of this study was to engineer a bivalent single-chain anticarcinoembryonic antigen (CEA) antibody and an interleukin 2 (IL-2) fusion protein derivative for selective tumor targeting of cytokines. The variable domains of a high affinity anti-CEA antibody, T84.66, were used to form a single-gene-encoded antibody [single-chain variable fragment joined to the crystallizable fragment, Fc (scFvFc)]. The fusion protein (scFvFc.IL-2) consisted of mouse IL-2-fused to the COOH-terminal end of the scFvFc. The engineered proteins were assembled as complete molecules and were similar to the intact anti-CEA monoclonal antibody (Mab) in antigen-binding properties. Based on IL-2 content of the fusion protein, its ability to support proliferation of CTLL-2 cells was identical with that of IL-2. Despite a molecular size similar to that of the intact Mab, the blood clearance of the fusion protein was markedly faster than that of the intact Mab or scFvFc. Incubation of radiolabeled scFvFc.IL-2 but not the intact or scFvFc antibodies in mouse serum was accompanied by the appearance of complexes, suggesting that the latter may contribute to the accelerated clearance of the fusion protein. Biodistribution and tumor targeting studies were carried out in CEA-transgenic mice bearing CEA-positive murine tumors as well as the antigen-negative parental tumor. The bivalent anti-CEA scFvFc had tumor localization properties similar to those of the intact Mab. Although fusion of IL-2 to the COOH-terminal end of the bivalent scFvFc altered its pharmacokinetic properties, the fusion antibody was able to target tumors specifically. Maximum uptake of the intact Mab, scFvFc, and scFvFc.IL-2 in CEA-positive tumors was 29.3 +/- 5.0, 19.5 +/- 2.1, and 6.6 +/- 0.9% injected dose/g, respectively. Maximum tumor localization ratios (CEA-positive/CEA-negative tumor) were similar for all three antibody types (4.6-6.0), demonstrating the antigen specificity of the tumor targeting. Significant antigen-specific targeting to CEA-positive normal tissues of transgenic mice was not observed. Although the tumor-targeting properties of the fusion protein were low, the growth of CEA-expressing (P = 0.01) but not antigen-irrelevant (P = 0.22) syngeneic tumor cells was inhibited after treatment of transgenic mice with the anti-CEA-IL-2 antibody. Therapy of CEA-expressing tumors was improved after i.v. administration of the fusion protein (P = 0.0001). These studies indicate that anti-CEA antibody-directed cytokine targeting may offer an effective treatment for CEA-expressing carcinomas. The availability of an immunocompetent CEA transgenic mouse model will also help to determine the immunotherapeutic properties of these fusion proteins.

Generation and characterization of a recombinant/chimeric B72.3 (human gamma 1).

We report here the generation and characterization of a recombinant/chimeric construct of murine gamma 1 monoclonal antibody (MAb) B72.3, containing the murine variable region and a human gamma 1 constant region [designated cB72.3(gamma i)]. cB72.3(gamma 1) was generated by first isolating functionally rearranged VH and VL genes of B72.3 from partial genomic libraries in phage vectors. Construction of mouse-human chimeric heavy and light chain genes was performed by inserting restriction fragments carrying VL and VH regions of B72.3 into unique sites of expression vectors which contains sequences encoding constant regions of human kappa and gamma 1, respectively. The expression constructs were subsequently electroporated into SP2/0 cells. The transfected SP2/0 murine cell line has been shown to synthesize cB72.3(gamma 1) at a level of 10-20 micrograms/ml. Reciprocal competition radioimmunoassays demonstrated that cB72.3(gamma 1), a previously described cB72.3(gamma 4), and native B72.3 (designated nB72.3) competed similarly. A rat anti-idiotype MAb made against nB72.3 was shown to bind equally well to cB72.3(gamma 1) and to the nB72.3. Immunochemical studies of the nB72.3, cB72.3(gamma 4), and cB72.3(gamma 1) revealed slight differences in size among the three MAb forms on sodium dodecyl sulfate gels and revealed a higher isoelectric point for the cB72.3(gamma 1). Antibody-dependent cell-mediated cytotoxicity experiments using human lymphokine-activated killer effector cells indicated better tumor cell killing by the cB72.3(gamma 1) than the nB72.3 or cB72.3(gamma 4). Dual label studies of coinjected cB72.3(gamma 1) and nB72.3 revealed that both MAbs could efficiently localize human tumor xenografts in athymic mice. Pharmacokinetic studies, analyzing the blood clearance of cB72.3(gamma 1), cB72.3(gamma 4), and nB72.3 in mice, showed that the nB72.3 beta phase of clearance was slower than that of other MAb forms. However, when the pharmacokinetic patterns of these three MAbs forms were analyzed in monkeys, the cB72.3(gamma 1) and the nB72.3 showed similar clearance curves, while the cB72.3(gamma 4) showed a much slower plasma clearance. In view of the binding properties of nB72.3 and its ability to localize a range of carcinomas in clinical trials, the studies reported here demonstrate that the cB72.3(gamma 1) may serve as a potentially useful diagnostic and/or therapeutic reagent.