Search

(51)Cr-prelabelled colon cancer cells (simulating 'circulating tumor cells', CTCs) were added to human peripheral blood and exposed to staurosporine (ST) to increase carcinoembryonic antigen (CEA) expression. CTCs were captured with immunomagnetic beads coated with Ber-EP4 monoclonal antibody, recognizing the common epithelial antigen present in the majority of cancer cells of epithelial origin, with capture efficiency of more than 80%. Moreover, ST treatment increased CEA expression without compromising Ber-EP4 capture efficiency. In a pilot clinical study on 37 patients, CTCs were captured using Ber-EP4 beads, and recognized by RT-PCR set for CEA or cytokeratin-19 (CK) mRNA detection. The results showed that: (a) the percentage of CEA-positive CTCs (CTC(CEA), 54.1%) was lower than that of CK-positive CTCs (CTC(CK), 70.3%); (b) in vitro ST treatment converted a significant number of CTC(CEA)-negative into CTC(CEA)-positive cases. Therefore, immunomagnetic capture combined with exposure to ST provides a feasible and sensitive technique for the detection of functionally-active CTCs responsive to ST-mediated CEA up-regulation.

Staurosporine (ST), a protein kinase C inhibitor, was found to produce antitumor effects against C22.20, a clonal subline derived from colon cancer HT-29 line, selected for low expression of carcinoembryonic antigen (CEA). However, as assessed by FACS analysis using propidium iodide, no apoptosis or cell cycle alteration was found on day 3 after treatment of C22.20 cells with ST (1-100nM). Exposure of cells to graded concentrations of the drug (i.e., from 1 to 25nM) resulted in a concentration-dependent increase in the percentage of CEA positive cells, as determined by flow cytometric analysis. However, when higher concentrations (i.e. 50nM - 100nM) of ST were used, the percentage of CEA positive cells declined compared to that detected in 25nM-treated tumor. Since these results were obtained in a clonal cell population, it is reasonable to hypothesize that induction rather than selection mechanism is involved in this phenomenon. The potential clinical interest of the present findings stems from the consideration that treatment with ST or its derivatives could improve sensitivity and efficacy of diagnostic and/or immunotherapeutic approaches based on CEA molecules.

Surgery, the most effective treatment for colon and rectal cancer, is based on empirical knowledge of the patterns of tumor spread, gross findings at laparotomy, and histologic confirmation of tumor-free margins. In spite of the many technical improvements in surgery, there has not been a significant change in cure rates for colon and rectal cancers. In fact, one-half of affected patients will not survive 5 years. It is in this arena of treatment for primary colon and rectal cancer patients that radioimmunoguided surgery (RIGS) technology may provide the most benefit. RIGS is an intraoperative procedure for detection of carcinoma lesions that are targeted with a radiolabeled monoclonal antibody (MAb) to provide the surgeon with immediate intraoperative definition of tumor margins and identification of occult disease. To optimize this technique, our studies were designed to increase tumor uptake by higher affinity CC-49 (a second-generation MAb) and to increase tumor antigen expression using biological response modifiers (BRMs). The ability of BRMs, such as interferons (IFNs), to enhance the expression of tumor-associated antigens, may play an important role in an adjuvant setting for MAb-based treatment. Preclinical and clinical data provided evidence for the use of IFN as an adjuvant to enhance MAb-targeting of human carcinoma lesions. A combination protocol with IFN and RIGS is ongoing at our institution.

5-Fluorouracil (5-FU) and human recombinant gamma-interferon (gamma-IFN) were found to increase the expression of carcinoembryonic antigen (CEA) in human cancer cells in vitro. In the present study, the antimetabolite was associated with gamma-IFN or folinic acid (FA), a biochemical modulator of cellular metabolism of 5-FU, able to increase its antineoplastic activity. Treatment of two human colon cancer cell lines (HT-29 and WiDr) with 5-FU + gamma-IFN resulted in an increase of CEA expression higher than that obtainable with both agents alone, although no synergistic effects were obtained. This was demonstrated in terms of: (a) mRNA transcripts (HT-29); (b) cytoplasm and membrane CEA protein levels detected by Western blot analysis (HT-29); and (c) plasma membrane reactivity determined by flow cytometry analysis (HT-29 and WiDr). Moreover, 5-FU + gamma-IFN increased HLA class I molecules in the HT-29 cell membrane over that obtainable with gamma-IFN alone. In contrast, both agents did not induce the expression of the costimulatory molecule B7-1. Treatment with FA enhanced the antitumor effect of 5-FU but not its ability to augment CEA expression. This suggests that the FA-sensitive biochemical mechanism of action of 5-FU is not involved in its effect on CEA expression. In vivo studies showed, for the first time, that 5-FU, alone or combined with gamma-IFN, increases the amount of CEA protein over controls, either in cancer cells or in peripheral blood of nude mice bearing HT-29 cells. These results could be of potential diagnostic and/or therapeutic value when CEA protein is the target of humoral or cell-mediated immunity.

A recombinant vaccinia virus encoding the gene for granulocyte-macrophage colony-stimulating factor (rV-GM-CSF) was used to infect the poorly immunogenic murine colon adenocarcinoma cell line, MC-38. Infection of MC-38 tumor cells with rV-GM-CSF completely suppressed the growth of the MC-38 primary tumors, whereas progressively growing tumors were formed in mice injected with MC-38 cells infected with wild type V-Wyeth. Irradiation of the recipient B6 mice before implantation of rV-GM-CSF-infected tumor cells resulted in the development of progressively growing tumors. Moreover, in vivo T-cell depletion studies revealed that growth suppression of the rV-GM-CSF-infected tumor cells was dependent on the presence of both CD4+ and CD8+ T-cell subsets. Subsequent studies established that this immunity was long-lasting and antigen specific, as demonstrated by the protection of rV-GM-CSF-immunized mice from MC-38 tumor challenge but not from challenge with another syngeneic tumor cell type. No such effects were observed when MC-38 tumor cells were infected with recombinant vaccinia viruses expressing interleukin (IL)-2 or IL-6. The results demonstrate that paracrine release of biologically active murine GM-CSF by tumor cells infected with rV-GM-CSF enhances the intrinsic immunogenicity of a poorly immunogenic murine tumor. Presumably the augmentation of tumor immunogenicity induces an antigen-specific T-cell-dependent antitumor response that prevents the formation of primary tumors and protects mice from tumor challenge. Thus in this experimental model, GM-CSF functions as a highly effective vaccine adjuvant.

Interleukin-2 (IL-2) has been an effective immune modulator in several active-specific immunotherapy experimental protocols using either viral or oncolysate-based vaccines. In this report, data indicate that IL-2 administration can appreciably augment the therapeutic effect of a single immunization of a recombinant vaccinia virus-carcinoembryonic antigen (rV-CEA) vaccine using a CEA-expressing syngeneic experimental murine model system. A single rV-CEA immunization of C57BL/6 mice bearing palpable CEA-positive colon adenocarcinoma tumors results in complete tumor regression in approximately 20% of the mice. The addition of a course of low-dose IL-2 results in complete tumor regression in 60-70% of the mice. Moreover, the combination of rV-CEA and IL-2 induces systemic immunity, which protects those tumor-free mice from subsequent rechallenge with the CEA-expressing tumor cells. No such tumor regression or protection was observed in those mice immunized with the wild-type vaccinia vaccine (V-Wyeth) alone or with IL-2 administration alone. Cellular immune assays revealed that the addition of IL-2 to rV-CEA immunization significantly increased the CEA-specific T-cell proliferative responses as well as the cytolytic T-cell responses when compared with rV-CEA immunization alone. The enhanced CEA-specific immune response, coupled with the improved experimental therapeutic outcome following IL-2 administration, suggests that treatment with that cytokine may effectively substitute for multiple rV-CEA immunizations in active-specific immunotherapy clinical protocols directed at CEA-expressing tumors.

Antibody-dependent cellular cytotoxicity (ADCC) has been suggested to be an important defense mechanism against tumors. The effects of recombinant human macrophage colony-stimulating factor (rhM-CSF) on ADCC activity of human monocytes were investigated. Human peripheral monocytes were pre-incubated for 72 h with rhM-CSF at various concentrations (50, 100, 200, 400 U/ml) and then used as effector cells in a 24-h 111-Indium release assay. Human carcinoma cell lines LS-174T, CBS, and KLE were used as targets to react with anti-carcinoma monoclonal antibodies (mAbs: murine D612, murine CC49, and chimeric CC49). A significant increase in ADCC activity was observed after monocytes were incubated in 100-400 U/ml of human rhM-CSF. Variation in ADCC activity of monocytes among donors was observed. The enhancement of ADCC activity was blocked by the addition of a neutralizing antibody to rhM-CSF. Less D612 mAb was required for the rhM-CSF-treated monocytes to mediate an equivalent level of ADCC activity as compared to the untreated monocytes. Because of the low levels of rhM-CSF required in these studies to enhance ADCC, treatment of monocytes alone with comparable levels of rhM-CSF did not enhance antibody-independent cytotoxicity. Moreover, it is demonstrated here that recombinant human interleukin-4 (rhIL-4) and rhM-CSF can have a synergistic effect of monocyte-mediated ADCC on human tumor cells. These results thus indicate that rhM-CSF augments ADCC of human peripheral blood monocytes using mAbs to human carcinomas, suggesting a potential role for rhM-CSF in cancer immunotherapy.

Comparative studies on the suppressive effects of recombinant interferon-alpha (IFN-alpha), 5-fluorouracil (5-FU), or IFN-alpha + 5-FU have been performed in vitro on colon carcinoma cells (HT-29 cell line) and PHA-stimulated mononuclear cells (MNC) of peripheral blood obtained from healthy donors. IFN-alpha was used at 500 U/ml against HT-29 cells and at 1000 U/ml against MNC on day 1 of culture; 5-FU was used at 250 microM against HT-29 and at 1400 microM against MNC on day 2 of culture. The results show that: (a) IFN-alpha inhibited MNC and HT-29 cells by 13.4% and 32.9%, respectively; (b) 5-FU inhibited MNC and HT-29 cells by 54.7% and 87.0%, respectively; (c) IFN-alpha + 5-FU resulted in a stronger inhibition of HT-29 cells (i.e., 96.1%). In contrast, that combination was significantly less suppressive than 5-FU alone when MNC were used as targets (i.e., 35.9% inhibition). Natural cell-mediated cytotoxic activity relative to 10(6) MNC was not markedly altered by all agents alone or in combination. Moreover, treatment with IFN-alpha, 5-FU or IFN-alpha + 5-FU resulted in a marked increase in the number of HT-29 cells positive for the CEA surface antigen. These data seem to provide further rational support of the clinical use of IFN-alpha + 5-FU in colorectal cancer, based on the differential toxicity of this drug combination on tumor versus normal immunocompetent cells.

Purpose: This study was designed to determine whether the ability of interferon-gamma to upregulate the expression of a human tumor antigen improved the therapeutic efficacy of a radionuclide-conjugated monoclonal antibody., Methods: Tumor xenografts of the moderately differentiated human colon tumor cell line HT-29 were grown in athymic mice. Constitutive levels of the human tumor antigen, tumor-associated glycoprotein-72, were measured before and after treatment with interferon-gamma. Antitumor effects of an 131I-labeled antitumor-associated glycoprotein-72 monoclonal antibody, CC49, were determined by measuring changes in tumor volumes in the respective groups of athymic mice., Results: Interferon-gamma induced a time-dependent and dose-dependent increase in tumor-associated glycoprotein-72 expression in the HT-29 tumors. Immunohistochemical staining revealed a more homogeneous tumor-associated glycoprotein-72-positive tumor cell population in tumors isolated from mice treated for eight days with interferon-gamma, which accounted for the enhanced tumor localization of 131I-CC49 in mice. That experimental model was used to examine the antitumor effects of combining interferon-gamma with 131I-CC49. Administration of 300 microCi of 131I-CC49 to mice bearing HT-29 tumors induced a transient suppression of tumor growth. Conversely, a long-term, sustained HT-29 tumor growth suppression was achieved in mice given 300 microCi of 131I-CC49 and interferon-gamma. In fact, the cytokine/radioimmunoconjugate combination eradicated any evidence of tumor in approximately 30 percent of the mice., Conclusion: The ability of interferon-gamma to enhance tumor-associated glycoprotein-72 expression substantially augmented the antitumor effects of the radioimmunoconjugate. Those observations provide additional argument for use of a radioimmunoconjugate in combination with a cytokine to improve tumor diagnosis and therapy.

Compelling experimental evidence suggest that the use of differentiation-inducing agents which enhance tumor antigen expression may play a pivatal role in MAb-based approaches to tumor diagnosis and/or therapy. In particular, interferons can up-regulate CEA and TAG-72 expression on the surface of human carcinoma cells which leads to an enhanced MAb tumor localization. The ability to target additional MAb to the tumor site has subsequently been shown to augment radioimmunodetection and therapy in experimental models. Initial clinical studies reported that interferon administration to patients diagnosed with malignant melanoma improved the tumor distribution of antimelanoma MAb without an accompanying increase of antibody in normal tissues. The findings from our phase 1A trial clearly showed that intracavitary IFN-gamma administration augmented TAG-72 and CEA expression on carcinoma cells isolated from malignant ascites from patients diagnosed with ovarian or gastrointestinal cancer. Future efforts need to further investigate the molecular mechanisms involved in the regulation of these and other human tumor antigens by the interferons as well as other differentiation-inducing agents. Additional clinical studies should evaluate the overall effectiveness of combining an antigen up-regulation protocol with a conjugated MAb in the hope of improving tumor diagnosis and therapy.

The moderately differentiated human colon tumor cell line, HT-29, constitutively expresses low levels of the high molecular weight mucin, tumor-associated glycoprotein 72 (TAG-72), and the M(r) 180,000 carcinoembryonic antigen (CEA) when grown as s.c. tumors in athymic mice. We report that the in vivo administration of gamma-interferon (IFN-gamma) resulted in a time- and dose-dependent increase in both TAG-72 and CEA expression in the HT-29 tumors. Immunohistochemical staining revealed a more homogeneous TAG-72-positive tumor cell population after IFN-gamma. Furthermore, both anti-TAG-72 and anti-CEA monoclonal antibodies (MAbs) showed enhanced localization to the HT-29 tumors in mice treated with IFN-gamma. Using that experimental model, subsequent studies presented evidence showing that the combination of IFN-gamma with 131I-CC49, an anti-TAG-72 MAb, resulted in a statistically significant improvement in therapeutic efficacy when compared with 131I-CC49 alone. For example, treatment with 300 microCi of 131I-CC49 initially suppressed HT-29 tumor growth; however, that reduction in tumor growth was transient as evidenced by the emergence of additional tumor growth at later time points. In contrast, an 8-day treatment with IFN-gamma in combination with 300 microCi 131I-CC49 resulted in sustained suppression of HT-29 tumor growth. Thus, IFN-gamma in vivo can substantially increase the TAG-72 expression in human colon tumor xenografts which leads to an increased tumor localization of anti-TAG-72 MAbs and seems to be responsible for the enhanced antitumor effects when IFN-gamma was combined with 131I-CC49. The results provide further evidence for including a biological response modifier, such as IFN-gamma, which can increase the expression of specific tumor antigens (i.e., TAG-72 and CEA) subsequently leading to a dramatic improvement in the antitumor efficacy of a radionuclide-conjugated MAb.

Treatment of human colorectal tumor cells (LS174T, HT-29, and WiDr) with analogues of cyclic AMP (cAMP) (dibutyryl-cAMP and 8-Cl-cAMP) selectively enhances the expression of carcinoembryonic antigen (CEA). Dose and temporal kinetics results revealed that 8-Cl-cAMP was approximately 100-fold more potent than dibutyryl-cAMP for increasing CEA expression. Results demonstrated that 8-Cl-cAMP treatment of LS174T quantitatively increased CEA levels in cell extracts 2-fold, increased anti-CEA monoclonal antibody (MAb) binding to the tumor cell surface, and induced the appearance of CEA-related mRNA transcripts. The findings suggest that 8-Cl-cAMP is capable of regulating CEA expression at transcriptional and/or post-transcriptional levels. Other human tumor cells, as well as normal cell types which do not constitutively express CEA, remained CEA-negative following 8-Cl-cAMP treatment. Moreover, the level of expression of other human tumor antigens as well as antigens of the major histocompatibility complex were not changed by 8-Cl-cAMP treatment, suggesting some selectivity for CEA regulation by this cAMP analogue. In vivo administration of 8-Cl-cAMP to athymic mice bearing LS174T tumor xenografts increased the amount of anti-CEA MAb bound to tumor extracts as well as the tumor localization of a radionuclide-conjugated anti-CEA MAb. The results indicate that 8-Cl-cAMP can selectively upregulate CEA expression on human colorectal tumor cells in vitro and in vivo. Interestingly, IFN-gamma treatment of the LS174T cells fails to enhance or induce expression of CEA or any of the histocompatibility leukocyte antigens. Thus, 8-Cl-cAMP treatment regulates CEA expression through another cellular pathway which may involve cAMP-dependent protein kinase.

The objective of our experimental protocols was to develop a metastatic model for a human colon carcinoma xenograft in congenitally athymic nude mice. This model would be useful in evaluating the efficacy of radiolabeled monoclonal antibodies for detection and treatment of regional and distant metastases. The LS-174T human colon carcinoma line was used to establish primary subcutaneous tumors in nude mice. Mice were killed at varying time intervals to establish the incidence of spontaneous metastases. Only lung micrometastases were observed during the 2-month observation period. To increase the metastatic rate, the site of primary implantation was varied and/or surgical manipulations were performed. Excision of small primary tumors resulted in a low incidence of local recurrence and no distant metastases. However, with excision of large primary tumors, a high local recurrence rate was noted and over 30% of mice had gross metastases. Mice bearing hind footpad tumors underwent excision when tumors were at least 1 cm in size. There were no local recurrences, but by 8 weeks over 40% had large pulmonary metastases. The LS-174T tumor was also established as a primary implant in the spleen from which 10 to 15% of the mice developed liver or lung metastases. The LS-174T tumor can metastasize in the nude mice and the latter two models may prove very useful in imaging and therapy studies.

Heterogeneity in the expression of tumor-associated antigens, as defined by the binding of monoclonal antibodies, is a characteristic common to most, if not all, human carcinoma cell populations. Antigen-negative cells within the population can escape detection and therapy by their failure to bind the appropriate antibody. Therefore, the extent of antigenic heterogeneity is an important consideration when designing protocols for the management of cancer by administration of monoclonal antibodies. One approach to counteracting the effect of antigenic heterogeneity is the use of clone A of recombinant human leukocyte interferon (Hu-IFN-alpha A). Administration of Hu-IFN-alpha A in vivo effectively increased the amount of tumor antigen expressed by a human colon xenograft in situ and augmented the localization of a radiolabeled monoclonal antibody to the tumor site. Concomitant administration of Hu-IFN-alpha A and monoclonal antibody may thus be effective in overcoming the antigenic heterogeneity of carcinoma cell populations and in enhancing the efficacy of monoclonal antibodies in the detection and treatment of carcinoma lesions.

The LS-174T human colon carcinoma line and A375 human melanoma line were used to establish primary tumor xenografts at three sites (subcutaneous, spleen, and kidney) in congenitally athymic mice. A monoclonal antibody (MAb) reactive with the LS-174T line, B72.3 IgG, was labeled with iodine 125, and an isotype-identical control antibody MOPC-21, was labeled with iodine 131. Labeled antibodies were injected intravenously in tumor-bearing mice, and animals were killed at varying intervals. Tumor-to-blood and tumor-to-organ ratios of MAb 72.3 indicated no significant difference at any of the three primary tumor sites in LS-174T tumor-bearing mice. The percent injected dose per gram was higher at visceral sites on day 3, but was similar on days 5 and 7 at all sites. Localization indices on all days ranged from 4 to 1 to greater than 16 to 1, confirming the specificity of the B72.3 reactivity at all sites. Athymic mice bearing tumor xenografts were scanned on day 7, and the LS-174T spleen and kidney tumors were imaged, with efficacy similar to that of the subcutaneous site. The visceral tumor model is more representative of the human disease, and may therefore be a better model for evaluation of monoclonal antibodies for radioimmunodetection and therapy for cancer in intra-abdominal organs.

Using an assay based on the binding of a carcinoembryonic antigen (CEA)-specific monoclonal antibody, we have examined the expression of carcinoembryonic antigen genes in human colon tumor and normal fibroblast cell lines. CEA expression was not detectable in the normal fibroblast cell lines, whereas varying levels of high CEA expression were found in the colon tumor cell lines LS-174T, GEO, and WIDR. We have used a 550-base pair CEA probe derived from cloned complementary DNA to carry out Southern analysis of the DNA isolated from the normal and colon tumor cell lines. At high stringency, the CEA probe detected seven BamHI fragments in all DNAs analyzed. At low stringency, however, 14 BamHI fragments ranging from 1.5 to 23 kilobases were detected. Results of the Southern analysis demonstrate no amplification or rearrangement of the CEA genes in tumor cells. We used methylation-sensitive restriction endonucleases, HpaII and HhaI, to compare the degree of methylation of CEA family of genes in normal and colon tumor cell lines. Our results demonstrate that the CEA family of genes exists in a state of hypermethylation in the normal cell lines. In contrast, the CEA gene(s) are relatively hypomethylated in the tumor cell lines, suggesting a correlation between the state of methylation and degree of expression of the CEA gene(s). A comparison of the state of methylation of the CEA gene(s) in cells before and after treatment with the gamma-interferon (which up-regulates CEA steady-state mRNA levels) showed no detectable difference in the degree of DNA methylation. The segments of CEA genes that are hypermethylated in normal cells, but are hypomethylated in tumor cells, were also identified. Thus, these studies may help identify the sites of methylation that are crucial for the control of CEA gene regulation.

Treatment of human breast or colon carcinoma cells with recombinantly derived human leukocyte (clone A) interferon (IFN-alpha A) increases the surface expression of specific tumor-associated antigens (TAAs) recognized by monoclonal antibodies (MAbs). The MAbs used, B1.1, B6.2, and B72.3, recognize three distinct TAAs, i.e., the Mr 180,000 carcinoembryonic antigen, a Mr 90,000, and a Mr 220,000 to 400,000 glycoprotein, respectively. The binding of the MAbs to the surface of tumor cells increased in a dose-dependent manner, with optimal levels of TAA enhancement at 100 to 1,000 units IFN-alpha A/ml. Higher concentrations of IFN-alpha A that were cytostatic or cytotoxic were also less effective in enhancing TAA expression. Human melanoma (A375) cells and normal fibroblasts (WI-38 and Flow 4000) do not express any of the three TAAs, either before or after interferon treatment. The ability of IFN-alpha A to increase the expression of TAAs on human carcinoma cells was also temporally dependent, with optimal enhancement occurring after 16 to 24 hr. The enhancement of specific TAAs at the surface of the carcinoma cells by IFN-alpha A was confirmed, using fluorescence-activated cell sorter analysis. These data demonstrate that the IFN-alpha A-mediated increase of surface antigen is a result of both an accumulation of more antigen per cell, and an increase in the percentage of cells expressing the antigen. The ability of recombinant interferon to enhance specific TAAs on human carcinoma cells may be exploited in designing protocols for the in situ detection and therapy of human carcinoma lesions by MAbs, as well as in further defining the role of specific TAAs in the expression of the transformed phenotype.

We describe the generation and characterization of a monoclonal antibody (MAb), designated D612, with selective reactivity for malignant and normal gastro-intestinal epithelium. MAb D612, a murine IgG2a, was generated using a membrane-enriched fraction of a human colon carcinoma biopsy as immunogen. Employing radioimmunoassays (RIAs) of biopsy extracts to a range of normal and neoplastic tissues, and both immunofluorescence and immunoperoxidase assays on frozen sections of a range of normal and neoplastic tissues, we have shown that MAb D612 binds to 82% of colorectal carcinomas tested (n = 67) and to normal gastro-intestinal epithelium, but does not bind similarly to either neoplastic or normal tissues from a wide range of other sites. Western blotting has shown MAb D612 to react with a high-molecular-weight antigen. Live cell RIAs and FACS analyses demonstrate the reactive epitope to be present on the surface of colon carcinoma cells. Immunohistochemical studies have shown intense membrane staining of colon adenocarcinomas with MAb D612; the vast majority of both primary and metastatic colon adenocarcinomas from a variety of sites were positive with many lesions showing homogeneous staining of virtually all cells present. Using human effector cells, we also showed that MAb D612 mediates antibody-dependent cell-mediated cytotoxicity (ADCC) of human colon carcinoma cells; this activity was enhanced in the presence of interleukin (IL-2). Radiolabelled D612-purified IgG selectively binds a human colon carcinoma xenograft in situ. The pattern of membrane-associated staining, the molecular weight of the reactive antigen, the IgG2a isotype, the ability to mediate ADCC in the presence of IL-2, and the immunohistochemical and RIA studies demonstrating highly restricted reactivity to malignant and normal gastro-intestinal tissue, all distinguish MAb D612 from other MAbs thus far described.