Differential expression of mucin genes in mammary and extramammary Paget's disease

Paget's disease (PD) of the skin is characterized by intraepidermal adenocarcinoma cells, which contain clear cytoplasm and abundant mucin. Nearly all cases of mammary PD (MPD) are associated with underlying ductal carcinoma of the breast, whereas in the majority of cases of extramammary PD (EMPD) no underlying regional malignancy is identified. Mucins are high molecular weight glycoproteins produced by epithelial cells. Different mucin genes are expressed in various types of tissues such as mammary glands, intestinal mucosa, and adnexal structures of the skin. We studied the immunohistochemical expression of apomucin MUC1, MUC2, MUC5AC in MPD, and EMPD. MUC1 is commonly expressed in most cases of PD. MUC5AC is a unique mucin that is exhibited in the majority of cases of EMPD, but not in any MPD. Of the 13 patients with MPD who all had associated breast ductal carcinoma, both Paget cells and underlying ductal carcinoma exhibited the phenotype (MUC1+MUC2-MUC5AC-). This mucin phenotype is also expressed by Toker cells, which have been identified in the epidermis of five of 50 nipples in mastectomies without MPD. Of the three patients with perianal PD who all had associated rectal adenocarcinoma, Paget's cells expressed MUC2 constantly but expressed MUC1 and MUC5AC variably. Seven patients with intraepidermal vulvar PD and two patients with scrotal-penile PD had no identifiable underlying malignancy. Paget cells from all of these nine cases of EMPD expressed a uniform phenotype of mucin (MUC1+MUC2-MUC5AC+). One case of vulvar PD associated with underlying apocrine carcinoma had a phenotype (MUC1+MUC2-MUC5AC-) identical to that of normal apocrine glands. The skin appendage and Bartholin's glands from 20 normal-appearing vulvar skin samples and anal glands from 10 hemorrhoidectomies were also studied. Only Bartholin's gland expressed a mucin phenotype identical to that of intraepidermal EMPD. The results of the present study indicate that 1) MPD may arise from either mammary glands or epidermal Toker cells, 2) intraepidermal EMPD in the anogenital areas may arise from ectopic MUC5AC+ cells originating from Bartholin's or some other unidentified glands, and 3) unique expression of MUC2 in perianal PD indicates its origin from colorectal mucosa. We conclude that the study of mucin gene expression is useful in identifying the histogenesis of PD.

Hepatic effects of long-term methotrexate use in the treatment of inflammatory bowel disease

OBJECTIVE

Methotrexate is currently used as a treatment for refractory inflammatory bowel disease. This study sought to evaluate the hepatic effects of long-term methotrexate therapy in patients with inflammatory bowel disease and to determine whether the established guidelines for monitoring methotrexate-related hepatotoxicity with surveillance liver biopsy in patients with psoriasis or rheumatoid arthritis are applicable to these patients.

METHODS

Thirty-two patients with inflammatory bowel disease receiving cumulative methotrexate doses of > or = 1500 mg were studied. Liver chemistry tests were obtained before and during therapy. Twenty patients underwent liver biopsies as recommended for methotrexate-treated patients with psoriasis; the biopsies were reviewed and graded according to Roenigk's criteria for methotrexate-induced hepatotoxicity (a grading system for methotrexate hepatotoxicity in psoriasis patients) by a liver pathologist blinded to the methotrexate dose.

RESULTS

In patients who had liver biopsies, the mean cumulative methotrexate dose was 2633 mg (range, 1500-5410 mg), given for a mean of 131.7 wk (range, 66-281 wk). Nineteen of 20 patients (95%) had mild histological abnormalities (Roenigk's grade I and II), and one patient had hepatic fibrosis (Roenigk's grade IIIB). Abnormal liver chemistry tests, present in 6 of 20 (30%) patients, did not identify the patient with Roenigk's grade IIIB hepatotoxicity.

CONCLUSIONS

Cumulative methotrexate doses up to 5410 mg given up to 281 wk in patients with inflammatory bowel disease are associated with little hepatotoxicity. Surveillance liver biopsies based on cumulative methotrexate doses are not warranted in these patients.

Colon signet ring cell adenocarcinoma: immunohistochemical characterization and comparison with gastric and typical colon adenocarcinomas

Colon signet ring cell adenocarcinomas are uncommon, high-grade neoplasms. Given their rarity, the question of primary colon or metastatic gastric adenocarcinoma frequently arises when signet ring cell carcinoma is seen in a colonoscopic biopsy or in biopsies procured from other regions of the body. A second related question regarding colon and gastric signet ring cell carcinomas is their immunophenotypic similarities with the glandular form of adenocarcinoma in each site. We studied the immunohistochemical phenotype of 14 colonic signet ring cell adenocarcinomas and compared them with immunophenotype of 27 gastric signet ring cell adenocarcinomas. We also compared the immunophenotype of the 27 gastric signet ring cell with the immunophenotype of 19 gastric gland-forming adenocarcinomas, and the immunophenotype of the 14 colonic signet ring cell adenocarcinomas to the immunophenotype of 20 colonic gland-forming adenocarcinomas to identify staining differences in the neoplastic cells of the two architectures. Antibodies studied were cytokeratins 7, 17, 19, and 20, CA 19-9, CA-125. estrogen receptor, and gross cystic disease fluid protein 15. Sixty-four percent of colon signet ring cell adenocarcinomas had either no staining or focal staining with cytokeratin 7 compared with diffuse staining in 63% of gastric signet ring cell adenocarcinomas (P = 0.016). Seventy-two percent of colon signet ring cell adenocarcinomas had diffuse staining with cytokeratin 20 compared with no or focal staining in 50% of gastric signet ring cell adenocarcinomas (P = 0.019). Fifty-seven percent of the colon signet ring cell adenocarcinomas had a cytokeratin 7 (-)/cytokeratin 20 (+) staining pattern compared with 11% of gastric signet ring cell adenocarcinomas (P = 0.004). Forty-four percent of gastric signet ring cell adenocarcinomas had a cytokeratin 7 (+)/cytokeratin 20 (-) pattern, compared with none of the colon signet ring cell adenocarcinomas (P = 0.004). The staining distribution of the antibody battery was similar in colon signet ring cell and colon glandular adenocarcinoma and gastric signet ring cell and gastric glandular adenocarcinomas. When signet ring cell adenocarcinoma is encountered in a colon biopsy, a colon primary is supported if the neoplastic cells have a cytokeratin 7 (-)/cytokeratin 20 (+) staining pattern, and a gastric primary is supported if they have a cytokeratin 7 (+)/cytokeratin 20 (-) staining pattern. The signet ring morphology at each site had an identical immunophenotype as the cells forming their glandular counterpart.

Differential expression of MUC2 and MUC5AC mucin genes in primary ovarian and metastatic colonic carcinoma

Colonic adenocarcinoma, the most common tumor metastatic to the ovary, may closely mimic primary ovarian adenocarcinoma, especially that of mucinous or endometrioid histology. The differential diagnosis is important for therapeutic considerations. Mucin gene expression is relatively organ-specific and may therefore have use in distinguishing between colonic carcinomas metastatic to the ovary and primary ovarian tumors. In this study, we compared the expression of MUC2 and MUC5AC apomucins in 10 colonic adenocarcinomas metastatic with the ovary, 10 ovarian endometrioid carcinomas (4 primary, 6 metastatic), and 32 primary mucinous ovarian tumors (12 cystadenomas, 10 borderline tumors, and 10 cystadenocarcinomas). Monoclonal antibodies CCP58 and 45M1 were used for immunostains of MUC2 and MUC5AC apomucin, respectively. All but 1 of the 10 metastatic colon adenocarcinomas expressed MUC2, whereas none expressed MUC5AC. None of the 10 endometrioid carcinomas expressed MUC2, and only 2 showed weak immunoreactivity with MUC5AC. All 32 primary mucinous ovarian tumors expressed MUC5AC. The percentages of MUC2-positive immunostaining for cystadenomas, borderline tumors, and cystadenocarcinomas were 0% (0/12), 50% (5/10), and 70% (7/10) respectively. These studies show that MUC2 and MUC5AC are useful markers in the distinction between colonic carcinoma metastatic to the ovary and primary ovarian carcinoma.

Donor liver uridine diphosphate (UDP)-glucuronosyltransferase-1A1 deficiency causing Gilbert's syndrome in liver transplant recipients

BACKGROUND

Uridine diphosphate-glucuronosyltransferase-1A1 deficiency, causing Gilbert's syndrome, has been attributed to two extra (TA) bases in the TATAA-box of the promoter region of its gene, where the A(TA)6TAA allele corresponds to the normal gene and A(TA)7TAA corresponds to a gene with reduced expression. Our aim was to determine whether isolated hyperbilirubinemia in liver transplant recipients was due to Gilbert's syndrome acquired through the liver allograft.

METHODS

From 305 patients followed in our Liver Transplant Clinic, five patients with isolated unconjugated hyperbilirubinemia in the absence of hemolysis, recurrent viral hepatitis, and biliary tract pathology were identified; 10 other post-orthotopic liver transplantion patients with normal liver chemistry tests were randomly selected as a control group. DNA was extracted from paraffin-embedded liver allograft tissue and peripheral lymphocytes and was genotyped for the TA repeat at the uridine diphosphate glucononosyltransferase-lA1 promoter region by polymerase chain reaction and acrylamide gel electrophoresis. Homozygosity for the (TA)7 allele was considered diagnostic of Gilbert's syndrome.

RESULTS

The mean serum total bilirubin level of the study patients was 2.28 mg/dl (range 1.8-3.0), consisting predominantly of the unconjugated form; that of the control patients was 0.76 mg/dl (range 0.4-1.1). The liver tissue from all five patients in the study group possessed the homozygous A(TA)7TAA genotype that was not observed in their lymphocytes. None of the liver tissue from the control patients demonstrated homozygosity for the A(TA)7TAA allele.

CONCLUSION

Uridine diphosphate-glucuronosyltransferase-1A1 deficiency, causing Gilbert's syndrome, may be carried by the donor liver and present with isolated unconjugated hyperbilirubinemia in liver transplant recipients.

Viral RNA in duodenal bile of cirrhotic patients with chronic hepatitis C

BACKGROUND

Hepatitis C virus (HCV) has been detected in blood, saliva, urine, semen, breast milk, and tears. To our knowledge, bile has not yet been investigated. We observed histologic immunoreactivity in bile with an antibody to c100 protein in four of five HCV-positive cirrhotic livers, but also in two HCV-negative controls owing to a focally present cross-reacting antigen.

METHODS

We collected duodenal bile from 13 cirrhotic patients during endoscopic evaluation of varices (10 HCV, three controls) and assayed for HCV by reverse transcriptase polymerase chain reaction.

RESULTS

Viral RNA was detected in the bile of 8 of 10 seropositive patients and in 0 of 3 seronegative controls.

CONCLUSION

Hepatitis C virus RNA and an antigen immunoreactive with anti-c100 protein are present in bile in a proportion of cirrhotic patients with chronic HCV. It remains to be determined whether the virus is intact or degenerate, and whether it is shed into bile from hepatocytes or is a contaminant from blood or other secretions.

Cathepsin D in intestinal ganglion cells. A potential aid to diagnosis in suspected Hirschsprung's disease

There is still a need for a better method of detecting immature ganglion cells in paraffin sections of colorectal luminal biopsies in cases suspected of Hirschsprung's disease. The lysosomal aspartic proteinase cathepsin D has been immunolocalized to various cell types, including ganglion cells. We investigated its expression in intestinal ganglion cells to determine whether it could be used as an aid in the detection of immature ganglion cells in rectal biopsies from children suspected of having Hirschsprung's disease. Routinely processed tissues of eight adult intestines resected for gunshot wounds and six ganglioneuromas (for mature ganglion cells), of six colons resected for neonatal necrotizing enterocolitis (for immature ganglion cells), and of 11 cases of suspected and three cases of known Hirschsprung's disease were immunostained with a polyclonal antibody to cathepsin D using the avidin-biotin-peroxidase method. In all cases, all ganglion cell bodies present showed intense granular cytoplasmic reactivity for cathepsin D. The granules crowded the cytoplasm and formed a collarette around the nucleus. In the submucosa, the only other immunoreactive cells were histiocytes, but they could be distinguished from ganglion cells by their characteristic nuclear features and their occurrence singly and unassociated with nerves. The three resection specimens with Hirschsprung's disease showed a clear transition between the ganglionic and the aganglionic segments. We conclude that cathepsin D is a promising marker of immature ganglion cells in cases suspected of Hirschsprung's disease.

Glutaraldehyde colitis following endoscopy: clinical and pathological features and investigation of an outbreak

Although potentially noxious compounds are used routinely to disinfect endoscopes, reports of their inadvertent introduction to the gastrointestinal tract, usually attributed to the retention of disinfectant within endoscope channels, are rare. This case report describes the clinical features of glutaraldehyde-induced colitis and the pathology of the mucosal injury in four patients, in at least one of whom the disinfectant was not retained in the endoscope itself. Within 3 months, three patients experienced severe acute proctocolitis < 6 hours after a sigmoidoscopy showing no abnormalities, performed in a small endoscopy unit. Investigation of the unit's protocols suggested that the most likely cause was retention of 2% glutaraldehyde disinfectant in the endoscope channels, and changes were made to prevent this. When a fourth case occurred 5 months later, the source of the glutaraldehyde was found to be the tubing connecting water bottles to the endoscopes, which was disinfected rigorously but flushed inconsistently between cases. Glutaraldehyde-induced colitis seems similar to ischemic colitis in biopsy specimens and cannot be diagnosed by histological analysis alone. Acute colitis occurring within 24 hours of a colonoscopy showing no abnormalities should be considered iatrogenic and should lead to an investigation of procedures in use for cleaning and disinfecting endoscopic equipment.

Lectin-mediated interactions of surfactant protein D with alveolar macrophages

Surfactant protein D (SP-D) is a calcium-dependent carbohydrate-binding protein that is secreted into the pulmonary airspaces by type II epithelial and Clara cells. Previous studies have shown that SP-D can bind to specific surfactant phospholipids and to glycoconjugates associated with the surface of various microorganisms, consistent with possible roles in surfactant metabolism and pulmonary host defense. We now describe specific saccharide-mediated interactions of SP-D with alveolar macrophages in lung tissue and in vitro. Biotinylated rat SP-D showed specific binding to alveolar macrophages in sections of rat lung; this labeling was inhibited by competing saccharides or EDTA. In addition, the binding of 125I-SP-D to isolated alveolar macrophages in the presence of calcium was time-dependent, saturable, and reversible and was preferentially inhibited by known monosaccharide and disaccharide ligands for SP-D. Scatchard analysis gave an apparent single class of binding sites with a Kd = 1.4 x 10(-6) M. We speculate that the multivalent structure of SP-D mediates bridging interactions between microbial glycoconjugates or surfactant phospholipids and specific glycosylated ligands expressed on the surface of phagocytic cells.

Surfactant protein D: subcellular localization in nonciliated bronchiolar epithelial cells

Surfactant protein D (SP-D, CP4) is a collagenous surfactant-associated carbohydrate binding protein that was initially characterized as a biosynthetic product of type II pneumocytes. Immunoperoxidase studies of formaldehyde solution-fixed and paraffin-embedded rat lung demonstrated staining for SP-D in the cytoplasm of a subpopulation of bronchiolar epithelial cells as well as type II cells. Accordingly, immunogold-labeling techniques were used to further examine the cellular distribution and subcellular localization of SP-D in the small airways. Lung tissues were fixed with 0.5% glutaraldehyde-3% paraformaldehyde and embedded in LR White resin. Sections were reacted with affinity purified polyclonal antibodies to SP-D, and sites of antibody binding were demonstrated using a biotinylated secondary antibody-streptavidin-gold detection system. Anti-SP-D selectively decorated secretory compartments of nonciliated bronchiolar cells (Clara cells) with strong and specific labeling of apical electron-dense secretory granules. Almost all of the granules in nonciliated columnar cells were labeled; however, labeling was typically nonuniform, with preferential decoration of the periphery of the granule. The largest numbers of immunoreactive epithelial cells were observed in the distal membranous bronchioles, with progressively smaller numbers of cells in more proximal bronchioles. There was no detectable labeling of cells lining the large cartilagenous airways or trachea. These studies provide evidence that SP-D is a secretory product of nonciliated bronchiolar cells. We suggest that Clara cell-derived SP-D is a component of bronchiolar lining material, consistent with our hypothesis that SP-D contributes to surfactant metabolism and/or host defense within small airways.

Interactions of surfactant protein D with bacterial lipopolysaccharides. Surfactant protein D is an Escherichia coli-binding protein in bronchoalveolar lavage

Surfactant protein D (SP-D) is a collagenous glycoprotein that is secreted into the pulmonary airspaces by alveolar type II and nonciliated bronchiolar cells. SP-D exhibits Ca(++)-dependent carbohydrate binding in vitro and is structurally related to the collagenous C-type lectins, including serum conglutinin, serum mannose-binding proteins, and surfactant protein A. Preliminary studies showed calcium- and saccharide-dependent binding of fluorescein-conjugated or radioiodinated SP-D to a variety of microorganisms, including Gram-negative bacteria and fungi. A laboratory strain of Escherichia coli (Y1088) was chosen to further examine the mechanism(s) of binding. Binding of SP-D to Y1088 was time dependent, saturable, and inhibited by cold SP-D or competing saccharides; Scatchard analysis gave a Kd of 2 x 10(-11) M. At higher concentrations, SP-D also caused Ca(++)-dependent agglutination of Y1088 that was inhibited by alpha-glucosyl-containing saccharides, antisera to the carbohydrate-binding domain of SP-D, or Y1088 LPS. Lectin blots showed specific binding of 125I-SP-D to Y1088 LPS, as well as LPS from other several strains of enteric Gram-negative bacteria. Immunogold studies demonstrated strong and uniform surface labeling of the bacteria. Rat and human bronchoalveolar lavage (BAL) caused Ca(++)-dependent agglutination of E. coli that was dose dependent and inhibited by competing saccharides or anti-SP-D. SP-D was selectively and efficiently adsorbed from rat BAL by incubation with E. coli, and incubation of E. coli with radiolabeled rat type II cell medium revealed that SP-D is the major E. coli-binding protein secreted by freshly isolated cells in culture. We suggest that SP-D plays important roles in the lung's defense against Gram-negative bacteria.

Inhibition of mucin glycosylation by aryl-N-acetyl-alpha-galactosaminides in human colon cancer cells

Specific inhibitors of the glycosylation of O-glycosidically linked glycoproteins have not previously been described. When tested for their effects on mucin glycosylation in a mucin-producing colon cancer cell line, LS174T, benzyl-, phenyl-, and p-nitrophenyl-N-acetyl-alpha-galactosaminide inhibited the formation of fully glycosylated mucin in a dose-dependent manner. Free aryl-oligosaccharides were found in the medium of treated cells labeled with [3H]glucosamine, [3H]galactose, [3H]fucose, [3H]mannosamine, or phenyl-alpha-[6-3H] N-acetylgalactosamine. UDP-Gal:GalNAc-beta 1,3-galactosyltransferase was inhibited by aryl-N-acetyl-alpha-galactosaminides but not by a number of other aryl-glycosides. Treatment with these inhibitors also causes reversible morphologic changes including formation of intercellular cysts. Aryl-N-acetyl-alpha-galactosaminides can be useful for the structural and functional studies of mucin macromolecules and other O-linked glycoproteins.

Deglycosylation of mucin from LS174T colon cancer cells by hydrogen fluoride treatment

Mucin from xenografts of LS174T human colon cancer cells was treated with anhydrous HF for 1 h at 0 degree C to give a product (HFA) with over 80% of the glucosamine and hexose removed, but retaining some galactosamine, and for 3 h at room temperature to give a product (HFB) devoid of carbohydrate. Rabbit antibodies against HFA bound to HFA much more than to HFB, and bound to native mucin to an intermediate extent. Antibodies to HFB bound to HFB more than to HFA, and did not bind to native mucin. Both HFA and native mucin bound a number of lectins, but HFB did not. By SDS/polyacrylamide-gel electrophoresis and size-exclusion h.p.l.c., native mucin and HFA are of apparent molecular mass greater than 400 kDa, whereas HFB is heterogeneous and of low molecular mass. On Western blots, antibody to HFA detected both high-molecular-mass mucin and a 90 kDa protein in homogenates of LS174T cells. Antibody to HFB detected a major 70 kDa band as well as higher-molecular-mass species. In tissue sections of normal colon and colon cancers, antibody to HFA showed both cytoplasmic and extracellular staining, whereas antibody to HFB generally stained only cytoplasmic antigens. These results indicate that anti-HFB antibody is specific for apo-mucin, whereas anti-HFA antibody is specific for GalNAc-apo-mucin.

Characterization of quantitative mucin variants from a human colon cancer cell line

Colonic mucins are high molecular weight glycoproteins produced by goblet cells of colonic epithelium. Some studies have indicated that patients with colonic cancers that produce high amounts of mucin have a poorer prognosis than patients whose tumors produce low amounts of mucin. At present, however, the role of mucin in affecting the behavior of colon cancer cells is not well understood. To further elucidate the relationship between cellular mucin content and the growth characteristics and morphology of tumor cells, we utilized a replica plating technique and immunoscreening method to identify and purify variant clones of the human colon cancer cell line LS174T that produce high and low levels of mucin. This procedure enabled us to isolate two high mucin-containing variants (HM3 and HM7) and one low mucin-containing variant (LM12). These variants exhibited different morphology. Both high mucin variants tended to form cell aggregates and suspended cells with adjoining mucoid threads. The low mucin variant formed spread monolayers on the substratum with the formation of cell processes. Metabolic labeling using [3H]glucosamine demonstrated that high mucin variants synthesized 2-fold more mucin in the cell layer and secreted 3-fold more mucin into the culture medium than the low mucin variant. The colony-forming efficiency in semisolid agar for these variants positively correlated with their mucin content. High mucin variant cells when injected into athymic nude mice formed tumors 2-fold larger than those of the parental cells while the low mucin variant formed tumors only one-half as large as those of the parental cell line. These mucin variants should provide a useful model for understanding the biological behavior of mucinous colon cancer cells in vivo and in vitro.