Colonic tumor CEA, CSAp and MUC-1 expression following radioimmunotherapy or chemotherapy.

Understanding the changes in tumor biology following cytotoxic therapy may lead to a better understanding of the properties of surviving tumor cell populations and to an improved ability to target and treat these cells. This report addressed the time-dependent dynamic alterations in the expression of three tumor-associated antigens: carcinoembryonic antigen (CEA), colon-specific antigen (CSAp) and mucin-1 (MUC-1) following chemotherapy with 5-fluorouracil (5-FU) or radioimmunotherapy (RAIT; (131)I-labeled anti-CEA IgG) in human colonic tumor xenografts. Immunoassay results show that CEA and MUC-1 expression all increase rapidly after either 5-FU or RAIT. GW-39 tumors show a 2.7-fold increase in CEA expression after a maximum tolerated dose of RAIT, being highest after 21 days, while LS174T and HT-29 tumors maximally increase expression 8.3- and 2.6-fold on day 7 after RAIT, respectively. The change in LS174T is short-term, whereas the change in HT-29 is maintained for at least 4 weeks. Serum CEA levels in these tumor- bearing mice also increase in parallel to the changes observed in tumor. MUC-1 increases 2.5-fold by day 5-7 following RAIT in LS174T tumors and 6-fold by day 14 following RAIT in GW-39 tumors, with a corresponding increase in serum MUC-1. Dramatic increases in CSAp after RAIT were also demonstrated in GW-39 tissue by immunohistochemistry. Thus, these data indicate that the response of tumor cells to low-dose-rate radiation from RAIT or to chemotherapy is associated with an increase of CEA, MUC-1 and CSAp.

Interferon-gamma upregulates MUC1 expression in haematopoietic and epithelial cancer cell lines, an effect associated with MUC1 mRNA induction.

Epithelial mucin-1 (MUC1) is an important target antigen that it is overexpressed in both epithelial and haematological cancers including multiple myeloma (MM) and some lymphomas and leukaemias. MUC1 has adhesive and immunosuppressive properties, which may promote cancer progression. These studies evaluated the effect of IFNs on MUC1 expression, since these agents are widely used in clinical cancer therapy. MUC1 and interferon (IFN) receptor expression were measured by radioligand binding. Changes in MUC1 mRNA levels in response to IFN-gamma were assessed by semi-quantitative reverse transcriptase-polymerase chain reaction (RT-PCR). IFN-gamma was found to be a more potent inducer of MUC1 expression than IFN-alpha. 125I-IFN binding studies indicated that both IFN receptors were expressed in most of the cell lines. With IFN-gamma treatment, there was upregulation of MUC1 mRNA. IFN-gamma has a more consistent and more potent effect upon MUC1 induction than IFN-alpha. The ability to upregulate MUC1 across a broad range of cancer types by a clinically available cytokine, IFN-gamma, has important implications for enhancing immunotherapeutic approaches targeting MUC1.

Therapeutic advantage of (90)yttrium- versus (131)iodine-labeled PAM4 antibody in experimental pancreatic cancer.

OBJECTIVES: Radioimmunotherapy studies using (131)I-PAM4 have demonstrated significant antitumor effects in mice bearing human pancreatic cancer xenografts. For several reasons (90)Y has been proposed as a more effective radionuclide for radioimmunotherapy of pancreatic cancer. The present study examined whether one radionuclide was more efficacious than the other in tumor-bearing mice. METHODS: Athymic nude mice bearing CaPan1 xenograft tumors ( approximately 1.0 cm(3)) were given increasing doses of either (90)Y-PAM4 or (131)I-PAM4 up to their respective maximal tolerated doses [MTDs (260 and 700 microCi, respectively)]. RESULTS: (90)Y-PAM4 provided significantly greater growth inhibition than the (131)I-PAM4 (P < 0.035). Median survival time for the untreated mice was 6 weeks, whereas median survival times for the (131)I-treated mice and (90)Y-treated mice at their respective MTDs were 17.5 weeks and >26 weeks (the end of the study period), respectively. Within the (131)I-PAM4-treated group, two of eight mice were responders (>50% decrease in tumor size) for a median of 14 weeks. At the end of the study (26 weeks), 1 mouse was alive with no sign of tumor. All of the (90)Y-PAM4-treated mice were responders with a median duration of response of 20 weeks. Six of the seven mice were alive at week 26, with four mice having no evidence of disease. CONCLUSIONS: These data demonstrate the advantage of (90)Y over (131)I as the radionuclide for PAM4-targeted radioimmunotherapy of xenografted pancreatic cancer. Furthermore, the duration and extent of the antitumor response suggests that multiple treatment cycles of (90)Y-PAM4 may provide an effective therapeutic for the control of pancreatic cancer.

Localization of pancreatic cancer with radiolabeled monoclonal antibody PAM4.

Experimental animal studies were performed with (111)In-labeled PAM4 anti-MUC1 antibody along with (111)In-labeled control antibody. Tumor uptake of radiolabeled PAM4 was significantly higher than for the control antibody at all time points. When normalized to a blood dose of 1500 cGy as an estimate of myelotoxicity, (90)Y-labeled PAM4 would provide 5344 cGy to the tumor, whereas an equitoxic dose of (90)Y-labeled control antibody would provide only 862 cGy to the tumor. In addition to the animal studies, five patients with proven pancreatic cancer were administered either (131)I-PAM4 IgG (n=2) or 99mTc-PAM4 Fab' (n=3). Tumor targeting was observed in four out of five patients. By immunohistochemistry, PAM4 was non-reactive with tumor from the one patient not targeted. Dosimetry from the patients given (131)I-PAM4 predicted that tumors would receive 10-20 cGy/mCi with tumor/red marrow dose ratios ranging from 3 to 10. Based upon these results, we have established a phase-I (111)In-labeled PAM4 imaging and (90)Y-labeled PAM4 therapy trial.

Monoclonal antibody G47 engineered to be reactive with colorectal tumor mucin.

We have previously shown that the normal adult colon produces a sialomucin containing the core trisaccharide 1,3 N-acetylgalactosamine. This structure was shown to be the epitope for a polyclonal antiserum that demonstrated colon "specific" activity. Antiserum binding is dependent upon the presence of O-acetylated sialic acids present at high concentrations in normal adult colon tissue. However, O-acetylation of sialic acids is decreased in colorectal cancer. Indeed, approximately 50% of colorectal carcinomas are nonreactive with this antiserum. In the current work, we used a de-O-acetylated, normal colon mucin as immunogen to generate monoclonal antibody (MAb) G47. Untreated normal colon mucins having a high O-acetylated sialic acid content were essentially nonreactive with G47. Removal of O-acetyl groups by saponification generated a reactive mucin derivative while subsequent treatment with neuraminidase abolished reactivity. By immunoperoxidase procedures MAb-G47 was reactive with approximately 85% of colorectal tumors while exhibiting relatively low reactivity with normal colon tissue. Mucins isolated from normal colon had on average less than 10% of the specific epitope as compared with mucins derived from colorectal tumors (p < 0.01). Initial immunohistochemical studies on tumors of noncolonic origin revealed few positive cases. The potential of MAb-G47 to assist in the diagnosis and/or prognosis of colorectal cancer is now being studied.

Epithelial mucin-1 (MUC1) expression and MA5 anti-MUC1 monoclonal antibody targeting in multiple myeloma.

Multiple myeloma (MM) is the second most common hematological cancer in the United States. It is typically incurable, even with myeloablative chemotherapy and stem-cell transplantation. The epithelial mucin-1 (MUC1) glycoprotein is expressed by normal and malignant epithelial cells but has also been shown to be expressed by MM cells. MUC1 is a useful antigenic target in solid tumors for clinical diagnostic and therapeutic monoclonal antibody (mAb)-based approaches. The MA5 mAb, as well as other anti-MUC1 mAbs reactive with the MUC1 variable number tandem repeat domain, exhibited moderate to strong reactivity with both MM cell lines and clinical samples. To explore the biochemical nature and potential of MUC1 as an antigenic target in MM, studies were performed to: (a) compare the mRNA and the MUC1 glycoprotein species between epithelial cancer and MM cell lines; and (b) develop and use a human MM tumor xenograft model system to study the biodistribution of the MA5 mAb. MA5 mAb was strongly reactive with six of eight human MM cell lines by flow cytometry. In seven of eight MM patient samples (bone marrow and/or peripheral blood) reactivity was found in 10-90% of the cells, whereas normal control (n = 5) and leukemia and lymphoma (n = 5) cells showed only 0-6% reactivity. 125I-labeled MA5 whole-cell binding studies showed quantitatively similar amounts of binding between strongly positive MM lines and high-MUC1-expressing breast carcinoma lines. mRNA expression was assessed by Northern blotting and reverse transcription-PCR. MM cell lines were positive by both methods, with strong similarity in the sizes of the mRNAs and cDNAs that were obtained. Finally, biodistribution experiments were carried out with 131I-labeled MA5 versus a nonbinding control 125I-labeled mAb in a s.c. MM xenograft model. Selective MM tumor uptake of the MA5 mAb was demonstrated, with a potential for delivering a tumor radiation absorbed dose of 8540 cGy/mCi of injected dose compared with 3099 cGy/mCi of tumor-absorbed dose delivered by nonspecific antibody.

Co-secretion of two distinct kappa light chains by the mu-9 hybridoma.

Mu-9 is a monoclonal antibody (MAb) specific for the CSAp antigen (Ag) expressed by colorectal cancers. By using variable (V)-region-specific primers, the respective VH and VL sequences of Mu-9 were polymerase chain reaction (PCR)-amplified. However, chimeric Ab (cMu-9-1) constructed from these PCR-amplified V sequences failed to bind the CSAp Ag. Although the light chain of murine Mu-9 was not glycosylated, that of cMu-9-1 was found to be O-glycosylated, as confirmed by reducing SDS-PAGE analyses, glycoprotein blotting and O-linked specific deglycosylation studies. Removal of O-linked oligosaccharides either by enzymatic digestion or by blocking O-glycosylation with a specific inhibitor did not restore the immunoreactivity of cMu-9-1, indicating that light chain O-glycosylation was not the cause for lack of immunoreactivity. We reported earlier that screening of a Mu-9 cDNA library uncovered the presence of an additional light chain sequence that was later proven to be the authentic light chain of Mu-9. Analyses of the cDNA sequence encoding the nonimmunoreactive light chain, however, revealed no defects that would preclude the sequence from being translated and secreted by the murine hybridoma. By adapting the Mu-9 hybridoma culture to serum-free conditions, we confirmed the secretion of low levels of O-glycosylated light chain. The biological significance of the O-glycosylation as well as the cosecretion of both light chains with respect to allelic exclusion are discussed.

Radioimmunotherapy of experimental pancreatic cancer with 131I-labeled monoclonal antibody PAM4.

We examined the therapeutic efficacy of 131I-labeled murine monoclonal antibody (MAb) PAM4 against human pancreatic cancers carried as xenografts in athymic nude mice. Animals bearing the CaPan1 tumor (0.2 cm3) were either untreated or were given, 131I-labeled nonspecific Ag8 antibody. By week 7 mean tumor size had grown 16.5 +/- 8.4-fold and 4.2 +/- 2.5-fold for the untreated and 131I-Ag8-treated animals, respectively. In contrast, animals administered 131I-PAM4 exhibited marked regression of tumors to an average of 15% of initial tumor volume. Since most pancreatic cancer patients present with large tumor burdens, the limitation of 131I-PAM4 treatment with respect to initial tumor size was investigated in animals bearing tumors of approximately 0.5 cm3, 1.0 cm3 and 2.0 cm3. Significant extension of survival time (>3-fold increase) was noted for both the 0.5 cm3 and 1.0 cm3 131I-PAM4-treated groups, compared to their respective untreated controls. Even in the group bearing large 2.0-cm3 tumors, survival was increased 2-fold over the control group. To further improve anti-tumor effects in large tumors, 2 injections of 131I-PAM4 were administered at a 4-week interval to animals bearing tumors of approximately 1.0 cm3. Significant extended survival was noted for the group receiving 2 doses when compared to the group receiving only 1 dose.

Factors influencing the pharmacokinetics, dosimetry, and diagnostic accuracy of radioimmunodetection and radioimmunotherapy of carcinoembryonic antigen-expressing tumors.

The aim of this study was to examine factors that may influence the pharmacokinetics, diagnostic accuracy, and dosimetry in radioimmunodetection and radioimmunotherapy with anti-carcinoembryonic antigen (CEA) monoclonal antibodies (mAbs). Data from 275 patients with CEA expressing tumors were analyzed retrospectively. Of these, 69 patients devoid of human antimouse antibody (i.e., 31 colorectal, 9 lung, 7 breast, 4 ovarian, 6 pancreatic, 9 medullary thyroid, 1 gallbladder, and 1 salivary gland cancer, and 1 primary tumor of unknown origin) underwent a low-protein-dose diagnostic study (0.3-2.6 mg of protein; 6.8-28.8 mCi 131I-labeled IgG or fragments), followed within 4 weeks by a high-protein-dose therapy injection (4.0-27.5 mg of protein; 29.8-238.9 mCi). The anti-CEA antibodies NP-4 (Ka=10(8)M-1) and MN-14 (ka=10(9)M-1) were used. Plasma clearance, the molecular composition of radioactivity in the plasma, and the cumulated activity in organs and tumors were determined. Radiation doses were derived from the Medical Internal Radiation Dose scheme. At a low-protein dose and over a similar range of plasma CEA, a significantly higher percentage of MN-14 than of NP-4 was complexed with circulating CEA, consistent with its higher affinity. Complexation was reduced with increasing protein doses. However, the targeting sensitivity was not affected. Profound differences were found in the clearance of the antibody between different types of cancer. Colorectal cancer patients cleared the antibody significantly faster from blood (T1/2=17.6+/-12.6 versus 44.2 +/- 23.7 h) and whole body (t1/2= 53.2 +/- 30.1 versus 114.6+/-59.7 h) than all other tumor types (P <0.001). Consequently, significantly lower red marrow (2.1 +/- 1.0 cGy/mCi versus 4.3 +/- 1.6 cGy/mCi) and whole-body doses (0.5 +/- 0.3 cGy/mCi versus 1.0 +/- 0.4 cGy/mCi) were seen in colorectal cancer patients as compared with other tumor types (P < 0.001). This clearance is probably due to hepatic metabolism of the immune complexes. Clearance rates were especially high in patients with colorectal cancer having large liver metastases and elevated liver enzymes (rapid hepatic clearance with liberation of free I-). In contrast, a disease-stage and plasma CEA-matched cohort of colorectal cancer patients, examined with the 131 I-labeled anti-colon-specific antigen p mAb Mu-9, showed normal murine IgG pharmacokinetics (n=22;3 of them compared intraindividually to MN-14). Only in colorectal cancer patients did complexes between mAb and CEA tend to clear rapidly, whereas Mu-9 had normal kinetics in these patients. This suggests that different CEA-expressing cancer types may produce heterogeneous CEA molecules and that the variability in mAb clearance is due to varying clearance rates of these different circulating CEA subspecies. Disease-related alterations in antibody metabolism are unlikely, given that only anti-CEA antibodies exhibit this phenomenon.

Initial studies of monoclonal antibody PAM4 targeting to xenografted orthotopic pancreatic cancer.

To resemble the clinical presentation of pancreatic cancer in an animal model more closely, we developed an orthotopic xenograft of CaPan-1 human pancreatic cancer in athymic nude mice. Within 3 weeks after implantation into the body and head of the pancreas, animals had palpable tumors. By 8 weeks, metastases to the liver and spleen were observed, and at 10-14 weeks, ascites formation, with and without seeding of the diaphragm, and jaundice were evident. Thus, this tumor model exhibited many of the most common features of human pancreatic cancer. Radiolabeled monoclonal antibody PAM4 showed specific localization of the primary orthotopic and metastatic tumors. On day 3, PAM4 accumulation within the primary tumor (0.5 g) was 11.3 +/- 5.1% injected dose/g with a localization index of 11.3 +/- 4.0. The estimated tumor:blood radiation dose ratio for PAM4 was 4:1, whereas a nonspecific antibody (Ag8) would provide only 40% of the blood dose to the tumor. Based on these observations, animals bearing 4-week-old orthotopic tumors (estimated volume, 0.25 cm3) were administered either 131I-labeled PAM4, 350 microCi, or nonspecific Ag8, 350 microCi, and compared with an untreated control group. Radiolabeled PAM4 provided a significant (P < 0.001) increase in survival time with less morbidity compared with the untreated control group, whereas nonspecific Ag8 was not significantly different from the control group. These studies provide a rationale for initiating a Phase I clinical study for detection and therapy of pancreatic cancer with PAM4.

Initial tumor targeting, biodistribution, and pharmacokinetic evaluation of the monoclonal antibody PAM4 in patients with pancreatic cancer.

This pharmacokinetic study was performed to assess the potential usefulness of the murine monoclonal antibody (MoAb) PAM4-IgG1 as an immunotargeting agent for pancreatic cancer imaging or therapy. This MoAb reacts specifically with mucin purified from human pancreatic cancer. 131I-labeled PAM4-IgG1 was injected i.v. into five patients with suspected pancreatic cancer. Whole-body scans and spot views of the abdominal area were recorded with a computerized gamma camera, and specific regions of interest were drawn over the liver and spleen to define the kinetics of activity in these organs. Blood samples taken from 0.1-144 h after injection served to define the kinetics of plasma distribution and removal of activity from the body. Surgery confirmed pancreatic cancer in four of the five patients, whereas chronic pancreatitis was present in the fifth patient; in all four pancreatic cancer patients, immunostaining with the MoAb PAM4 demonstrated the presence of the specific antigen, with a cytoplasmic and endoluminal/secretory pattern of distribution. Nonspecific radioactivity accumulation in the liver, spleen, and bone marrow was low, linked essentially to the blood pool effect of circulating activity in these organs. The overall quality of scintigraphic maps recorded over the abdomen was quite satisfactory due to the low liver and spleen activity, with good scintigraphic demonstration of the pancreatic cancers (either primary or metastatic); the patient subsequently found to have pancreatitis failed to show PAM4 targeting. Except in one patient with widespread peritoneal metastases (in whom these tumor implants were detected scintigraphically already 24-48 hours after tracer injection), scintigraphic evidence of the tumor lesions was usually late, starting at about 72-96 h after tracer injection. The results obtained in this preliminary study indicate the potential usefulness of MoAb PAM4 for immunoscintigraphy in patients with either primary and/or recurrent pancreatic cancer while also suggesting that the use of the faster-clearing Fab fragments of this MoAb probably would result in improved immunoscintigraphic properties.

Targeting of xenografted pancreatic cancer with a new monoclonal antibody, PAM4.

We have examined the ability of murine monoclonal antibody PAM4, directed against a pancreatic cancer-derived mucin, to target human pancreatic cancers carried as xenografts in athymic nude mice. Four tumor lines were used representing the range of expected differentiation; CaPan1, AsPc1, Hs766T, and BxPc3. In each case tumor uptake of PAM4 (range, 21-48% injected dose/g on day 3) was significantly higher than concomitantly administered, nonspecific, isotype-matched Ag8 antibody (range, 3.6-9.3% injected dose/g on day 3). Based upon the biodistribution data the estimated potential radiation dose delivered to the tumors when normalized to the blood dose as an estimate of dose-limiting myelotoxicity would be 13.1-, 2.2-, 3.4-, and 3.3-fold higher than to blood, respectively. PAM4 showed no evidence of targeting to normal tissues, except within the CaPan1 tumor model, where a small but consistent splenic uptake was observed. Splenic targeting was abolished by use of an increased PAM4 protein dose. Targeting of PAM4 to other normal tissues was not affected by the increased protein dose; however, tumor uptake of PAM4 (percentage of injected dose/g) was significantly increased by as much as 3-fold. The ability of PAM4 to target the CaPan1 tumor compared favorably to that of MN14, an anti-carcinoembryonic antigen murine monoclonal antibody. Tumor uptake of PAM4 was much greater than that for MN14 at days 1 and 3, whereas at later time points equivalent accumulations of activity were noted. Estimates of potential radiation doses to the tumor when normalized to the blood dose were 3.0 for MN14 and 9.6 for PAM4. These studies have shown that PAM4 is able to target pancreatic cancer with high specificity, achieving high concentrations at the tumor site. A rationale exists, then, for the performance of a clinical trial of radiolabeled PAM4 in the detection and localization of pancreatic cancer.

Characterization of monoclonal antibody PAM4 reactive with a pancreatic cancer mucin.

A monoclonal antibody (MAb), PAM4, having reactivity with pancreatic carcinoma has been developed. PAM4 is an IgG1 immunoglobulin produced by immunization of mice with mucin purified from the xenografted RIP1 human pancreatic carcinoma. An immunohistochemical study of normal adult tissues showed the PAM4 reactive epitope to be restricted to the gastrointestinal tract and absent from normal pancreas. In neoplastic tissue, PAM4 was reactive with 85% of the pancreatic carcinomas, approximately half of the colon cancers and none of the breast, ovarian, prostate, renal and liver cancers. PAM4 was, in general, non-reactive with pancreatitis specimens whereas CA19.9 and DUPAN2 were strongly reactive with each one. Treatment of the mucin antigen by heating, reduction of disulfide bonds, or protease digestion abolished immunoreactivity with PAM4. Treatment of the mucin by neuraminidase or periodate oxidation reduced immunoreactivity but did not completely abolish it. Our data are consistent with the proposal that the PAM4 epitope is a conformationally dependent peptide epitope and that certain carbohydrate structures are necessary in order to maintain the correct peptide conformation. The high specificity and intense reactivity of PAM4 with pancreatic carcinoma tissue suggests that the antibody may prove useful for in vitro diagnostic assays as well as in vivo targeting of diagnostic and therapeutic agents.

Microheterogeneity of a purified IgG1 due to asymmetric Fab glycosylation.

A murine monoclonal anti-granulocyte IgG1, IMMU-MN3, was seen to exhibit heterogeneity. On reduced SDS-PAGE, the purified antibody appeared as two heavy-chain bands of unequal intensity, and only one light-chain band. Hydrophobic interaction chromatography (HIC) also resolved two populations of the IMMU-MN3 antibody. Based on Concanavalin A affinity chromatography, enzymatic digestion with Endoglycosidase F and carbohydrate analysis, it was found that the heterogeneity detected by SDS-PAGE and HIC was due to differences in glycosylation. Furthermore, sequential gel analysis (non-reduced/reduced) demonstrated that the upper heavy-chain band was asymmetrically glycosylated.

Murine monoclonal antibodies to colon-specific antigen p.

A previous report described the production of monoclonal antibodies Mu-1, -2, -3, and -4 against colon-specific antigen p. We now report the tissue distribution of the reactive epitopes. Each of the monoclonal antibodies was reactive with epitopes shared by the entire length of the intestine, with Mu-2 and Mu-3 determinants being present in the stomach as well. Mu-1 and Mu-4 appeared to have a more restricted distribution than did Mu-2 and Mu-3, which were present in several nongastrointestinal tissues. Unfortunately, these MAbs were not ideal for in vivo use against colorectal carcinoma. Mu-1 was not present in neoplastic colon tissue, Mu-2 and Mu-4 were reactive with granulocytes, and Mu-3 was reactive with kidney tubules and connective tissue. An additional fusion was performed from which Mu-9 was selected for further study. Among normal adult tissues this monoclonal antibody gave reactivity restricted to the intestines. The stomach and all nonintestinal tissues were negative by immunoperoxidase. Among colon tumors 60% were positive for the Mu-9 epitope. The data suggest Mu-9 may be a good candidate for in vivo targeting of diagnostic and therapeutic agents to colon cancer.

Comparison of tumor targeting in nude mice by murine monoclonal antibodies directed against different human colorectal cancer antigens.

Tumor targeting of five radioiodinated murine monoclonal antibodies (MAbs) directed against human colorectal cancer were studied in nude mice bearing the GW-39 human colonic tumor xenograft. All of the MAbs are of the IgG1 isotype. Two of the MAbs (NP-4 and MN-14) are directed against a Class III carcinoembryonic antigen-specific epitope, but they differ 10-fold in their affinity. The other three MAbs recognize mucins found in colonic cancer. Mu-9 recognizes a peptide determinant similar to that described previously for a polyclonal goat anti-colon-specific antigen p. G9 identifies an organ-specific, tumor-associated carbohydrate epitope. The tumor targeting of these MAbs was compared to that of B72.3, another anti-mucin MAb. The tumor uptake of all the MAbs were similar on days 1, 3, and 7, with an average maximum accretion of between 30 and 40%/g tumor occurring by day 3. This tumor uptake was maintained for 14 days with the anti-mucin MAbs, whereas the percentage of injected dose/g in the tumor for 2 anti-carcinoembryonic antigen MAbs decreased 2-fold by day 14. Although no statistical difference could be found between the percentage of injected dose/g in the tumor for NP-4 and MN-14 (carcinoembryonic antigen MAbs), in a paired-labeled study using 131I-MN-14 and 125I-NP-4, MN-14 uptake in the tumor was consistently 1.3 times higher than that of NP-4 on all days tested. F(ab')2 fragments showed lower tumor uptake (maximum uptake for NP-4 and Mu-9 was 11% on day 1), but the faster clearance resulted in a 4- to 40-fold increase in tumor/blood ratios on day 3 in comparison to the whole IgGs. Fab' fragments had the lowest tumor uptake of the 3 forms of immunoglobulin, with a maximum of only 5%/g 6 h after injection. However, tumor/blood ratios on day 1 for the Fab' fragments were improved 3-fold over that of F(ab')2. All of these MAbs, except Mu-9, identify epitopes that can be detected in plasma, but none of the MAbs complexed appreciably when mixed in vitro with plasma containing antigen at antigen/MAb ratios anticipated to be encountered most frequently in imaging or therapy applications in humans. However, complexation of the MAbs will occur if antigen in the plasma is markedly elevated. These studies suggest that if used clinically, tumor targeting of colorectal cancer with any of these MAb IgG may be similar if these antigens are present at relatively similar concentrations in tumor.(ABSTRACT TRUNCATED AT 400 WORDS)

Generation of a monoclonal antibody (G9) reactive with an organ-specific, tumor-associated epitope of human colon carcinoma.

A monoclonal antibody (G9), having an organ-specific and tumor-associated reactivity with colon carcinoma, has been generated. Monoclonal antibody G9 is an IgG1 immunoglobulin produced by immunization of mice with mucin which had been purified from a liver metastasis of a moderately differentiated human colon carcinoma. Examination of normal adult tissues, by enzyme immunoassay and immunohistochemical procedures, showed the G9-reactive epitope to be restricted to the gastrointestinal tract. Within the gastrointestinal tract the colon produced the highest amount of the epitope. A sharp, decreasing gradient of reactivity was observed, ending in the small intestine. Although the normal colonic epithelium did produce the G9-reactive determinant, there was a significant quantitative increase of the epitope in neoplastic colonic tissue; mucins derived from normal colon contained less than 10% of the specific epitope as compared with mucins derived from colon cancer tissues (P less than 0.01). In addition, a tumor xenograft contained 100 times the amount of epitope as did normal colonic tissue. By immunohistochemical procedures 70% of all colon carcinomas were positive. A relationship with differentiation was noted, with 80% of well differentiated and 83% of moderately differentiated tumors being positive, whereas only 1 of 6 poorly differentiated tumors were stained. The organ specificity was noted in neoplastic as well as normal tissues. Monoclonal antibody G9 was nonreactive with breast, lung, and ovarian tumors. The data suggest that at the level of sensitivity obtained by the immunoassays used, and within the range of tissues examined, monoclonal antibody G9 is organ specific and highly tumor associated in its reactivity.

Studies on the structure of the organ-specific determinant of human colonic mucin.

We had previously shown that the human colon produces at least two immunochemically distinct mucins, one neutral and the other a sialomucin [Gold et al. J. biol. Chem. 256, 6354-6358 (1981)]. In addition, the sialomucin was shown to contain an immunodeterminant restricted to colonic epithelium and may thus prove useful as a tissue-specific marker. In the current study we have shown that a specific linkage of sialic acid to the oligosaccharide backbone has a major role in the organ-specific immunodeterminant structure. Treatment of intact colonic mucin with sialidase (Cl. perfringens) cleaved 20-80% of the sialic acid as measured colorimetrically. Immunoreactivity was decreased by 0-42% with respect to the untreated material. Saponification (0.1 N KOH, 20 min at room temp) caused an approximate 90% decrease in immunoreactivity for each mucin. Subsequent to saponification, neuraminidase cleaved most of the sialic acid from the mucins. The majority of sialic acid was observed to be O-acetylated, thus making it sialidase-insensitive. Gas chromatography-mass spectrometric analyses of the trimethylsilyl sialic acid derivatives indicated the presence of NeuNAc; NeuNAc, 9-OAc; and NeuNAc, 7,9 diOAc as the major sialyl derivatives. The radioimmunoassay data appeared to indicate that O-acetylated sialic acid was necessary for immunoreactivity. It should be noted that jejunal mucin and bovine submaxillary mucin also contain O-acetylated sialic acid, but did not inhibit in our radioimmunoassay. This may have been due to differences in the O-acetylation pattern or the linkage of sialic acid to the core carbohydrate. Analyses of the partially methylated alditol acetate derivatives by gas chromatography-mass spectrometry of the untreated, as well as the saponified and neuraminidase treated, mucins revealed that sialic acid was attached to the carbohydrate core either to galactose, N-acetylglucosamine, and/or N-acetylgalactosamine. Linear regression analyses comparing immunoreactivity with specific epitope concns, in conjunction with RIA analyses of known structures, suggested that the organ-specific immunodeterminant was (or was dependent upon the presence of) the structure GlcNAc (1,3)[O-acetylated Neu5Ac(2,6)] GalNAc.

Monoclonal antibody B72.3 reacts with a core region structure of O-linked carbohydrates.

Monoclonal antibody (moAb) B72.3 appears to be exceptional among the many mouse moAbs reactive with human tumor cells in its degree of tumor specificity. The antigen recognized was characterized as a mucin-like molecule. We report here that B72.3 reacts strongly with ovine submaxillary mucin and that reactivity is abolished by desialylation, suggesting that the determinant recognized is the disaccharide N-acetylneuraminic acid alpha (2----6)-N-acetylgalactosamine, linked to serine or threonine, which might be designated sialylated Tn. A variety of human mucin preparations, including human salivary mucins, and other glycoproteins containing O-linked or N-linked carbohydrates were nonreactive. The data suggest that the determinant recognized is a core structure which is normally not exposed due to chain elongation at C-3 of the N-acetylgalactosamine, which appears to be less frequent in carcinoma cells than in normal cells.

Organ specific antigens of the human gastrointestinal tract.

We previously reported the characterization of a normal adult colonic mucin antigen which contained an organ specific immunodeterminant [Tissue Antigen 11, 362 (1978)]. In the present study we have investigated mucins produced at other levels of the gastrointestinal tract in order to determine if regional specificities exist. Mucins were isolated from normal adult stomach, jejunum, ileum and colon and used to prepare antisera in rabbits. By radioimmunoassay at least four distinct specificities were observed. Gastric, ileal and colonic mucins were shown to contain immunodeterminants which were organ specific. Antiserum directed toward jejunal mucin determinants was reactive with the entire gastrointestinal tract. However, by heterologous inhibition analyses employing purified mucins as inhibiting antigens, the anti-jejunum antiserum was shown to be capable of discriminating a determinant present in much higher epitope density within small intestinal mucins as compared to mucins of the stomach and colon. Thus, it appeared that immunologic determinants present within mucin type glycoproteins of the gastrointestinal tissues exhibit anatomic specificity. In each case the structure of the immunodeterminant was, or was dependent upon the presence of a sialic acid derivative.