Effects of ketoconazole on the proliferation and cell cycle of human cancer cell lines.

The growth-inhibitory effects of ketoconazole, an antifungal agent which inhibits arachidonic acid lipoxygenases and cytochrome P-450 enzymes, were tested in human colon and breast cancer cell lines. In the serum independent HT29-S-B6 colon cell clone, ketoconazole reduced cell proliferation and [3H]thymidine incorporation in a dose-dependent fashion, with a 50% inhibitory concentration of approximately 2.5 microM. Flow cytometry showed an accumulation of cells in the G0-G1 phase of the cell cycle and a concomitant decrease of the percentage of cells in S phase. Ketoconazole also inhibited [3H]thymidine incorporation in the hormone-independent breast cancer cells MDA-MB-231 and Evsa-T, with respective 50% inhibitory concentration of approximately 13 and 2 microM. The mechanism of action of ketoconazole is unknown. However, another lipoxygenase inhibitor, BW755C, inhibited only weakly [3H]-thymidine incorporation and accumulated the cells in S and G2. Conversely, clotrimazole and SKF525A, inhibitors of cytochrome P-450 enzymes, had effects similar to those of ketoconazole on HT29-S-B6 cells whereas metronidazole and secnidazole, other azole derivatives which do not inhibit cytochrome P-450 enzymes, had no effect. The results suggest that cytochrome P-450 enzyme(s) activity(ies) could be implicated in the antiproliferative effects of ketoconazole.

Proliferation of the human colon carcinoma cell line HT29: autocrine growth and deregulated expression of the c-myc oncogene.

Human colon adenocarcinoma cell (line HT29) are able to proliferate in a defined (serum-free) medium containing no added growth factors; in such conditions, their doubling time is 3 to 4 days (on serum-coated dishes) or 2 to 3 days (on an autologous extracellular matrix) compared with 1 day in the presence of fetal calf serum. In the presence of suramin, a polyanion disrupting the binding of growth factors to their receptors, the incorporation of [3H]thymidine in serum-free cultures is reduced (27.0 +/- 2.9% of control after 3 days of culture), suggesting involvement of autocrine growth factors in the autonomous proliferation of the cells. The expression of the proliferation-related oncogene c-myc was examined during various stages of growth and differentiation of the HT29 cells. The cellular contents of c-myc mRNA were similar in all experimental conditions studied: exponential phase; stationary phase; nondifferentiated as well as differentiated cells (by glucose deprivation); and also in serum-free medium containing or not suramin. An approximately 2-fold increase in the level of c-myc mRNA was observed in cells cultured for 3 days in suramin-containing medium and then incubated during 3 h in the absence of suramin (with or without 10% fetal calf serum). Southern blot analysis of the genomic DNA of HT29 cells did not reveal any rearrangement within the region containing the c-myc gene and the flanking sequences (approximately five kilobases upstream and approximately three kilobases downstream). The c-myc locus was weakly amplified (four to six copies per cell). These results indicate that the c-myc gene expression in HT29 cells is deregulated and does not require growth factor stimulation. The deregulation of the c-myc gene may be related to the reduced growth factor requirement of the HT29 cell line.

Growth-promoting effect, biological activity, and binding of insulin in human intestinal cancer cells in culture.

The biological action and binding of insulin were tested in two human intestinal cancer cell lines originating from the duodenum (HUTU 80) and the colon (HT 29). After serum deprivation for 24 hr, insulin stimulated cell division and the incorporation of labeled precursors into RNA, protein, and DNA for both cell lines. The action on the RNA and protein was rapid and significantly different (1.5 to 2 times that of control) 1 hr after adding insulin. These effects were dose dependent, present at physiological concentration in vivo (10(-10) M), and independent of the transport of precursors. For thymidine incorporation, the stimulation was delayed up to 8 hr and culminated with cell division 20 hr later. As previously shown for HT 20, HUTU 80 cells exhibited insulin-specific binding sites. Binding of 125I-insulin was saturable; reversible; and time, temperature, and pH dependent. Scatchard analysis of the binding data of the two cell lines gave curvilinear plots. Assuming the presence of two independent binding sites, the high-affinity constants were 6 to 8 X 10(8) M-1, and the number of high-affinity receptors was similar and accounted for 2000 to 3000 receptors/cell. For both cell lines, the effect of insulin on protein and RNA synthesis was significantly different from control at 1 hr when binding reached a maximum at 37 degrees. The biological action of insulin on growth and macromolecular synthesis was dose dependent and maximum at about 10(-8) M insulin, which corresponds to 70% displacement of 125I-insulin binding. Furthermore, the binding and the biological action of proinsulin were about 2% that of native insulin in the two cell lines studied. These results show that insulin acts as a growth factor for these two cell lines and that these effects are probably mediated by the interaction of insulin with specific receptors.

Development of insulin and epidermal growth factor receptors during the differentiation of rat preadipocytes in primary culture.

An homogeneous cell population isolated from the inguinal tissue of 3-day-old rats is able to proliferate in primary culture. In the presence of a physiological concentration of insulin (1.5 nM) it converts into cells exhibiting the morphology and the biochemical characteristics of adipocytes. Insulin and epidermal growth factor (EGF) receptors were studied during both the exponential growth and the adipose conversion phases of these cells. Binding experiments with 125I-labelled peptides were performed directly in the culture dishes. The number of high affinity insulin binding sites increased, during the entire culture period studied, reaching 18 days after plating the value of 10,600 x 2360. Control cells (cultured in the presence of anti-insulin antibody) exhibited an increase of the concentration of insulin binding sites from no more than 500 sites/cell to 6880 +/- 1710 sites/cell between dat 0 and 9 (corresponding to the exponential growth phase); this increase was followed by a rapid reduction in insulin receptors during the stationary phase. The density of EGF binding sites increased between day 0 and 4 (one cell cycle), whether the cells were maintained or not with insulin, and plateaued thereafter. Mature adipocytes freshly isolated from the inguinal tissue of 3-day-old rats had no detectable EGF binding sites, but their content in high affinity binding sites for insulin was similar to that of cells after complete adipocyte conversion in primary culture.

Characterization and repartition of epidermal growth factor-urogastrone receptors in gastric glands isolated from young and adult guinea pigs.

Physiological studies indicate that epidermal growth factor-urogastrone (EGF) acts on stomach epithelium as mitogen and modulator of acid secretion. Here, we studied the binding of 125I-EGF to gastric glands isolated from the guinea-pig fundus (acid-secreting part) and antrum. At 20 degrees C, the association of 125I-EGF to gastric glands was time-dependent (plateau at 90 min) and reversible (75-85% dissociation in 1 h). No degradation of the peptide was detected, but a time-dependent loss of binding capacity was observed. At apparent equilibrium (90 min, 20 degrees C) unlabelled EGF (80 pM to 80 nM) competed with 125I-EGF-binding in the same manner in antrum and fundus (50% inhibition, with 0.6 nM EGF). Whereas kinetics properties were similar in antrum and fundus, the binding capacity was 40-55% lower in fundus than in antrum in young animals (6-8 weeks). By contrast, in adult animals (20-30 weeks), binding was the same in both parts of stomach. Scatchard analysis showed that two orders of binding sites were present in all cases (Ki 0.34-0.47 nM, Ki 2.2-3.4 nM), and that the differences observed were only accounted for by number of binding sites. These results show that EGF possess high affinity binding sites on gastric epithelium. These sites, dependent upon the age of the animals, may be related to the modulations by EGF of gastric trophism and secretions.

Enterocyte differentiation is compatible with SV40 large T expression and loss of p53 function in human colonic Caco-2 cells. Status of the pRb1 and pRb2 tumor suppressor gene products.

Transfer of the SV40 large-T (LT) oncogene into isolated human and murine intestinal epithelial cells induced alterations of the ultrastructural organization and polarization of the resulting immortalized cell lines. We now demonstrate that the functional expression of the SV40 LT antigen in Caco-2 cells did not alter phenotypic markers of differentiation, including expression of villin, sucrase-isomaltase, brush border and dome formation. As compared to parental cells, the transfected Caco-2 LT9 cells exhibited similar growth curves and no invasive properties in vitro. The major oncogenic function of the SV40 LT antigen in transfected Caco-2 cells is associated with reduced latency times necessary for the manifestation of tumors in athymic nude mice. The Caco-2 cell line contained deleted and mutant p53 alleles (stop codon in position 204) and has no detectable truncated p53 protein by Western blot. Molecular complexes between the SV40 LT antigen and the retinoblastoma-related proteins pRb1 and Rb2 were clearly identified at the different phases of the growth curve. When compared to normal human colonic crypts, Caco-2 cell differentiation is related to partial redistribution of pRb1 into hypophosphorylated, antiproliferative forms. The pRb2 protein is found elevated in a subset of human colorectal tumors and their corresponding liver metastases. We conclude that: (1) Caco-2 cells exert a dominant control against the oncogenic functions of the LT antigen; (2) loss of p53 function is not restrictive for the establishment of polarity and differentiation of the enterocyte lineage; (3) the levels and phosphorylation status of the Rb1 and Rb2 proteins may play important roles in the proliferation and differentiation of normal and neoplastic human colonic mucosa.

Insulin binding by a cell line (HT 29) derived from human colonic cancer.

Insulin binding was demonstrated in cultured HT 29 cells originating from a human colon carcinoma. At 37 degrees and in complete medium, the binding of [125I]insulin (1-4x10-10M) reaches a maximum in 40 min and the cell associated radioactivity remains constant for at least 4 h. No degradation of the hormone is observed under these conditions. The binding is proportional to the number of cells and its pH optimum is 7.8. In the presence of excess insulin 50% of the [125I]insulin is dissociated from the complex after 10 min. At equilibrium, insulin binding is specific: proinsulin is 25 times less potent than native insulin in competing with [125I]insulin and related polypeptide hormones are inactive. Scatchard analysis indicates two classes of binding sites (1400 sites/cell of "high affinity" e.g. 4.7 x 108 M-1, and 20 000 sites of "low affinity" e.g. 4 x 107 M-1). The binding of insulin to this non-target cell shows the same kinetic characteristics and specificity as found for insulin in its target cells, except that HT 29 cells do not degrade the hormone. The problem of the correlation between insulin binding and a biological effect in these cells remains to be elucidated.

Histamine and VIP interactions with receptor-cyclic AMP systems in the human gastric cancer cell line HGT-1.

In HGT-1 cells incubated at 20 degrees C for 15 min with 1 mM 3-isobutyl-1-methylxanthine (IBMX), histamine (10(-4)M) increased basal cAMP levels from 2.12 +/- 0.14 to 22.9 +/- 2 pmol per 10(6) cells, with a potency of 6.4 X 10(-6)M. IBMX was added in order to inhibit cAMP degradation by low and high Km cAMP-phosphodiesterases (cAMP-PDE). The use of specific H1, H2 agonists or antagonists indicated that the histamine effect was due to an interaction with typical H2 -receptors that are involved in gastric acid secretion. Cyclic AMP levels were also increased (10-fold) by vasoactive intestinal peptide VIP (3 X 10(-11) - 10(-8)M). Porcine peptide having N-terminal histidine and C-terminal isoleucine amide (PHI) and secretin were respectively 80 and 3600 times less potent than VIP and did not produce additive effect when tested in combinations with VIP. This observation indicates that these two peptides, structurally related to VIP, are acting through the recognition sites for VIP. Combination of VIP and histamine results in additive stimulation on intact cells as well as on membrane-bound adenylate cyclase, suggesting the existence of two cell populations bearing respectively the two sets of receptors. Two other human cancer cell lines originating from nongastric tumors (HT-29 and HL-60) possess only VIP or histamine receptors, respectively, indicating the gastric cellular originality of the HGT-1 cells. It is concluded that HGT-1 cells possess both VIP and histamine H2 receptors with similar pharmacological properties to those characterized in normal human fundic glands (1,2). Therefore, this cell line can be a good model to study drugs used therapeutically during the treatment of patients for gastric ulcer or cancer.

Effects of VIP on the regulation of mucin secretion in cultured human pancreatic cancer cells (Capan-1).

The effects of Vasoactive intestinal peptide (VIP) on mucin secretion in the pancreatic cancer Capan-1 cell line were studied by Enzyme-linked-immunosorbent-assay (ELISA), and by light and electron microscopy using immunocytological methods. During the exponential growth phase, mucins were accumulated in the cytoplasm of cells and slowly exocytosed. In contrast, there was enhanced exocytosis of mucins during the stationary phase when the cells were well-polarized. Moreover, during this phase, VIP induced a dose-dependent rise in mucin content in the extracellular medium. The reaction with anti-M1 monoclonal antibodies, which recognize specifically the peptide core of gastric mucins, showed an accumulation of secretion granules near the apex of well-polarized cells together with fusion of the granule and plasma membranes after VIP stimulation. Moreover, mucin exocytosis was stimulated by Pituitary adenylate cyclase activating polypeptide (PACAP) and secretin. It was also increased after forskolin treatment suggesting that this mechanism was cAMP-dependent. Our results suggested that exocytosis of mucins could be under the control of VIP in pancreatic duct cells of the Capan-1 cell line.

Early lesions induced in rat colon epithelium by N-methyl-N'-nitro-N-nitrosoguanidine.

The development of ACF (aberrant crypt foci), adenoma and cancer following intrarectal administration of the alkylating agent N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) has been described. However, microscopic lesions not previously reported were observed as soon as two weeks following carcinogen treatment. These lesions protrude slightly over the epithelial lining of the colon, with a micropolyp-like appearance. Oriented sections show that the centre of these lesions present pseudo-"cystic" appearance, with disorganized crypts made of normal cells. The chorion of the lesion is invaded by numerous inflammatory cells and some ACF may be present nearby. The epithelium lining the cysts and the distorted crypts shows expression of gastric mucin M1/MUC5AC, an early marker of colonic carcinogenesis which is not present in normal colon. This mucin is retained within the "cysts" together with some inflammatory cells. The micropolyps observed contain in a minute form some histological elements described in ulcerative colitis or short-term radiotherapy (distortion of crypts, crypt abscesses, increase of chorion cellularity, infiltration by immune cells). In addition, the presence of bifid crypts nearby suggests mucosal regeneration. Our hypothesis is that these modifications are steps in a normal healing pathway that may in some cases degenerate into precancerous lesions and cancer.

Insulin binding in epithelial cells isolated from human gallbladder.

The binding of 125I-insulin was studied in epithelial cells isolated from human gallbladder. Kinetic studies at 22 degrees C showed that binding of 125I-insulin to gallbladder cells was rapid (maximum 1 h.) and reversible. In the presence of added excess of unlabelled insulin, 50% of initially bound insulin is dissociated in 10-15 min. At apparent equilibrium (1 h at 22 degrees C) unlabelled insulin (10(-10) to 10(-7) M) inhibited competitively tracer (5 X 10(-11) M) binding, with 50% inhibition at about 2 X 10(-9) M. Scatchard analysis gave curvilinear plots, that may be attributed either to negative cooperativity, or to two orders of binding sites: In the first case, the extreme average affinities are Ke = 1.8 X 10(8) M-1, and Kf = 3.2 X 10(7) M-1 for 49,000 sites/cell. In the latter case, gallbladder cells present 8,000 sites/cell of high affinity (Kd = 1.16 X 10(-9) M) and 42,000 sites/cell of low affinity (Kd = 3.3 X 10(-8) M). In addition, in the same cellular preparation at 37 degrees C, insulin at a concentration of 10(-8) M significantly stimulated (p less than 0.01) 3H-leucine incorporation into acid-insoluble fraction by 1.4 +/- 0.1 fold at 30 min. A maximal effect is obtained at 10(-7) M insulin (2.11 +/- 0.06 fold). Our results suggest a stimulatory effect of insulin or related "insulin-like" peptides on growth in human gut.

Abnormal expression of gastric mucin in human and rat aberrant crypt foci during colon carcinogenesis.

Colorectal cancer is a major cause of death in Europe and the USA, and much effort is therefore devoted to improve its early detection. In this article, we report the abnormal expression of gastric mucin in aberrant crypt foci (ACF) that appear in the colon mucosae removed from colorectal cancer patients and rats treated with methyl-N'-nitro-N-nitroso-guanidine (MNNG). We performed the immunoperoxidase test using monoclonal antibodies raised against gastric M1 mucin encoded by the MUC5AC gene and against rat gastric mucins (MAb 660), respectively. In both human and rat colon, these anti-gastric mucin MAbs stained specifically goblet cells within ACF. In humans, the M1/MUC5AC mucin was expressed in the upper part of the glands in hyperplastic ACF and in the typical ACF. In addition, the anti-gastric mucin MAbs stained some rare, scattered, histologically normal glands in the human and rat colon mucosae. These glands may be regarded as precursors of ACF. The abnormal expression of the MUC5AC gene constitutes a novel change in addition to genetic modifications already observed in ACF, and supports our previous findings demonstrating the potential of this gastric mucin as an early marker of human and rat colon carcinogenesis.

Antiproliferative effects of the arotinoid Ro 40-8757 in human gastrointestinal and pancreatic cancer cell lines: combinations with 5-fluorouracil and interferon-alpha.

The arotinoid Ro 40-8757 was previously shown to inhibit the growth of a variety of human cancer cell lines derived from breast, lung and uterus. In view of the high incidence of human digestive cancers, and the slow progress in the development of new therapy, we examined in this paper several combinations between the new arotinoid Ro 40-8757, 5-fluorouracil (5FU) and interferon alpha-2a on the growth of nine human cancer cell lines derived from the gastrointestinal and pancreatic system. Half-maximal inhibition of cell proliferation by Ro 40-8757 was observed at concentrations ranging between 0.18 and 0.57 microM, and increased up to 4.7 microM in retinoid-resistant CAPAN 620 pancreatic cells. All-trans-retinoic acid was 70 times less potent. The sensitivity of HT29-5FU-resistant colonic cells was similar to that observed in the parental cells, suggesting an action independent of pyrimidine metabolism. Ro 40-8757 did not induce any differentiation on HT29 cells, as suggested by ultrastructural analysis. The arotinoid did not interact with receptor signal transduction pathways under the control of serum components, such as growth factors as half-maximal inhibiton of growth was similar in HT29-S-B6 cells cultured in the absence or presence of serum. Cell cycle analysis showed that Ro 40-8757 was not acting at a phase-specific transition in HT29 cells and, accordingly, did not induce overexpression of the protein kinase C (PKC)alpha isoform, or conversion of hyperphosphorylated p105 Rb into hypophosphorylated forms. However, the arotinoid induced significant accumulation of the dephosphorylated, active form of the tumour-suppressor protein. Combinations of Ro 40-8757 with 5FU and interferon alpha 2a resulted in an additive but not synergistic antiproliferative action in HT29 cells. Our data support the interest in Ro 40-8757 as a potent anti-cancer drug, especially in combination therapy with 5FU and interferon, in gastrointestinal and pancreatic cancers, where new active therapeutic modalities are urgently needed.

Devazepide (L-364718) inhibits growth and increases expression of tumor markers in HT29-S-B6 cells.

Devazepide (L-364718, a non-peptide antagonist of CCK-A receptors), inhibits the proliferation and induces morphologic changes in the mucous-secreting, autonomously proliferating human cancer colon cell line (HT29-S-B6. Addition of devazepide (10 microM) for at least 3 days in the exponential phase of growth enhanced the baseline production of gastric M1 mucins 2-3-fold and that of carcinoembryonic antigens 5-fold. Moreover, devazepide induced an increase in the amount of the MUC-5AC mRNA expressed by HT29-S-B6 cells. The increased in mucins secretion induced by devazepide was persistent after removal and independent of the presence of serum. In conclusion, devazepide-L-364718 behaves as a maturation agent in the cell clone HT29-S-B6.

The presence of epidermal growth factor receptors in cultured human adipocyte precursors.

The characteristics of 125I-epidermal growth factor urogastrone (125I-EGF) binding to human omental adipocyte precursors over a period of early differentiation in culture, is reported. The results show the presence of cell surface EGF receptor sites (38,000 per cell) that bind 125I-EGF with a high affinity (Ka = 2.7 X 10(9) l/mol). Their presence is not appreciably altered during a 28-day growth period in culture. The existence of such receptors suggests the possibility that growth factors might intervene in the regulatory processes involved in adipose tissue development.

Evidence for the presence of insulin binding sites in isolated rat intestinal epithelial cells.

Insulin receptors have been demonstrated in isolated rat intestinal epithelial cells. The specific binding of 125I-insulin was time--and temperature--dependent, the optimal temperature of study being 15 degrees. Dissociation of bound 125I-insulin by an excess of unlabelled hormone was rapid and attained 66 +/- 2% in 2 h. When initiated by dilution, the dissociation attained 35 +/- 4% in 2 h, and 72 +/- 1% in 2 h when 10(-7) mol/l unlabelled insulin was added. The pH optimum for the binding process was between 7.5 and 8, and the binding increased proportionally to cell protein concentration up to 1.5 mg/ml. Under standard conditions (2 h at 15 degrees) the degradation of the labelled hormone in the medium accounted for 20--50% of total tracer, depending on the concentration of cells. At apparent equilibrium (2 h at 15 degrees), unlabelled insulin in the range of 10(-10) to 10(-7) mol/l inhibited competitively the binding of 4.3--7 X 10(-11) mol/l 125I-insulin; fifty per cent inhibition was obtained with 3 X 10(-9) mol/l native insulin. Scatchard analysis, after correction for degradation, gave curvilinear plots, that may be explained by two orders of binding sites, with 2,000 +/- 200 sites/cell of high affinity (Ka = 2.2 +/- 0.2 X 10(9) l/mol) and 39,000 +/- 3,000 sites/cell of low affinity (Ka = 5.6 +/- 1.6 X 10(7) l/mol). The potency of proinsulin to compete with 125I-insulin for the binding site was 3% that of insulin, unrelated peptides were inactive. Such results give a molecular basis to different reports suggesting that the intestine could be a target-tissue for insulin.

Gastric M1 mucin, an early oncofetal marker of colon carcinogenesis, is encoded by the MUC5AC gene.

Gastric M1 mucin and the MUC5AC gene show a similar oncofetal expression in the colon. Our aim was to determine whether M1 mucin is the product of the MUC5AC gene. A recombinant baculovirus encoding the C-terminal portion of the MUC5AC gene as a fusion protein was isolated and the immunoreactivity of the recombinant mucin (rM) toward M1 antibodies studied. Chicken antibodies also were raised against purified rM. Besides its reactivity with L56/C, a serum recognizing the bacterially expressed MUC5AC gene product, rM was endowed with M1 immunoreactivity: (i) rM-expressing cells were stained specifically with anti-M1 serum and with the monoclonal antibody (MAb) 21M1, defining the M1-f epitope; (ii) both L56/C and anti-M1 antibodies recognized the same bands in immunoblots of rM-containing cell extracts; (iii) the 21M1 antibody reacted with rM in an immunoradiometric assay. Among the 7 M1 epitopes, M1-f was the only one encoded by the 3' portion of the MUC5AC gene. It was the only epitope detected in a native mucin M1-derived 170 kDa bromelain proteolytic fragment. Furthermore, the staining patterns of human tissues obtained with either anti-rM chicken antibodies or anti-M1 antibodies were identical. We conclude that M1 immunoreactivity is encoded at least in part by the MUC5AC gene.

Mapping of SOMU1 and M1 epitopes on the apomucin encoded by the 5' end of the MUC5AC gene.

We have developed 11 monoclonal antibodies (MAbs) against human gastric mucin, (1-13M1, 2-11M1, 2-12M1, 9-13M1, 58M1, 19M1, 21M1, 45M1, 463M, 589M, 62M1), which specifically stained by immunohistochemisty both the human gastric surface mucosa and colon adenoma. Among them, five (19M1, 21M1, 463M, 589M, 62M1) immunoreacted with the peptide encoded by the 3' region of the MUC5AC gene (Nollet et al: Int J Cancer 2002;99:336-343). In this study, we identified in the 5' region of this gene the nucleotide fragments encoding peptides immunoreacting with three other anti-M1 MAbs (1-13M1, 2-11M1 and 9-13M1), as well as the SOMU1 MAb (Sotozono et al: J Immunol Methods 1996;192:187-196). 1-13M1 MAb immunoreacts with peptides, including the Cys 2 and Cys 4 domains. The SOMU1 MAb recognized the Cys 5 domain, and the MAbs 2-11M1 and 9-13M1 the globular D1/D2 and D3 domains, respectively. Using serial sections of the mucosae adjacent to colon adenocarcinomas and colon adenomas, we observed that the anti-M1 and anti-SOMU1 MAbs displayed the same immunostaining patterns. The three anti-M1 MAbs (2-12M1, 58M1, and 45M1) did not react with the products of the MUC5AC gene tested until now. The MUC5AC apomucin is now well characterized by MAbs immunoreacting against seven different epitopes belonging to the different main cystein globular domains of this macromolecule. Such antibodies are useful tools for studying the biosynthesis, polymerization, and degradation of mucin.

Detection of gastric mucins (M1 antigens) in cyst fluid for the diagnosis of cystic lesions of the pancreas.

Mucinous cystic tumors of the pancreas must be distinguished from other cystic lesions because of their potential malignancy. Our purpose was to assess the reliability of gastric M1 mucin analysis in the fluid of cystic lesions of the pancreas in comparison or association with carcinoembryonic antigen. M1 mucin and carcinoembryonic antigen were measured in cyst fluid obtained preoperatively by fine-needle aspiration. The lesions consisted of 12 serous cystadenomas, 9 mucinous cystadenomas, 8 cystadenocarcinomas and 6 intraductal mucinous hypersecreting neoplasms. Thirty pancreatic pseudocysts complicating well-documented chronic pancreatitis were also examined. In addition, M1 mucins were localized by immunoperoxidase staining in fetal and normal adult pancreas and in mucinous and serous tumors. Carcinoembryonic values of > 20 ng/ml and M1 mucin values of > 50 U M1/ml represented 82 and 78% sensitivity, respectively, as well as 100% specificity for distinguishing mucinous lesions from serous cystadenomas; the sensitivity for this purpose was 100% using these criteria in combination. Carcinoembryonic antigen values of > 300 ng/ml and M1 mucin values of > 1,200 U M1/ml represented 56 and 30% sensitivity, respectively, as well as 100% specificity for distinguishing mucinous lesions from pseudocysts; the sensitivity for this purpose was 60% using these criteria in combination. By immunohistology, M1 mucins were detected in the wall of mucinous lesions but not in fetal and normal adult pancreas and in serous cystadenomas. Measurement of M1 mucin antigen in cyst fluid could thus improve the diagnosis of mucinous cystic lesions of the pancreas.