Modification of blood group A antigen expression in a pancreatic cancer cell line (PC-1) by inhibitors of N-glycan processing

A and B blood group antigens in human pancreatic ductal adenocarcinomas

OBJECTIVES

The aim of this study was to determine the clinicopathological significance of blood group expression patterns in patients with surgically resected pancreatic ductal adenocarcinomas.

METHODS

Pancreatic ductal adenocarcinomas from 459 patients were analyzed for immunohistochemical expression of A and B blood group antigens on tissue microarrays. Correlations were established by Fisher's test, χ(2) test, and logistic regression models, and relationships to postsurgical overall survival were analyzed by Kaplan-Meier method, log-rank test, and multivariate Cox models.

RESULTS

Tumors expressing blood group A antigen predicted fewer lymph node metastases (<3) and had longer postoperative survival by univariate and multivariate analysis. Loss of A antigen tumor expression in A blood group type patients correlated to vascular invasion.

CONCLUSION

In patients with resectable pancreatic ductal adenocarcinomas, blood group A antigen expression predicted fewer lymph node metastases and was associated with improved outcome after pancreaticoduodenectomy.

"Natural" antibodies and histo-blood groups in biological development with respect to histo-blood group A. A perspective review

The "inappropriate" A-specific ovarian glycosphingolipids discovered in unfertilized C57BL/10J female mice reflect growth processes, which suggest the activity of embryonic stem cells undergoing genetic polymorphism. And the responding anti-GalNAc antibody represents the first classical "natural" antibody, which was unmasked as a highly specific autoantibody. This murine anti-A is subspecifically distinct from the human antibody, discovering by a broader reactivity growth-dependent, xenoreactive A-specific structures also in non-reproductive murine tissues, where an equivalent of the human AB gene family as a cis AB-gene encodes A-and B glycotransferases. Expression of antigen is known to need always more than its encoded enzyme, and the special mechanism which in the C57BL/10J murine ovarian glycospingolipids blocks the expression of "B" still remains still unknown. A herewith arising postulation of a growth-predominating common biological activity may be supported by findings in rats. The number of A-genes here significantly exceeds those of B and in the Wistar rat the A-antigen is only expressed in the wild type, while B-expression requires the transfer of human B. Nevertheless in transgenic rats, the appearance of "A" still remains more pronounced. The observations lead to reports on animals, which do not show AB transferase production or a respective antigen expression in their normal tissues, but inconcistently display A activity in malignant tumors. And respective examples are delivered by phenotype independent neo expressions of "inappropriate" A-specific structures in human cancer. Although in comparison with epitope deletions they are rare, the ubiquitous "natural" (IgM and IgG) anti-A and anti-B levels, against self and not self, irrespective of the blood group in any normal human sera, may reflect invisible "inappropriate" A-specific growth. The role of the associated (auto) anti-B might be different, because B-neo expressions obviously not occur in cancer, and anti-gal-antibodies are supposed to originate primarily from environmental, cross-reactive stimulation, and beyond their functions in defense are otherwise engaged in physiology. In general natural antibody specificities undergo significant phylogenetical changes within the species. However, the in nature wide-spread "natural" anti-A agglutinin specificities survived or even predominated the long-term evolution from the brown trout up to man and still respond to the biological power, i.e. the products of a CAZY glycosyltransferase 6 (ABO) gene family. It is so hypothesized that both, the murine and human "natural" anti-A antibodies represent examples of a still to be analyzed polyclonal response to a provocative, species-independent evolutionary epitope, which arises or escapes by some enzymatic predominance from the genetical polymorphism in a consistent developmental process.

Localization of galectin-3 in normal and diseased pancreatic tissue

CONCLUSION

Galectin-3 is expressed in both human and hamster pancreatic tumors and tumor cell lines and this expression is increased over normal.

BACKGROUND

Galectin-3 is overexpressed in many gastrointestinal tumors. This study examined the expression of galectin-3 in human and hamster pancreatic tumors to determine if galectin-3 could be used as a marker for pancreatic cancer.

METHODS

Membranes were prepared from human and hamster pancreatic tumor cell lines. Galectin-3 was visualized by immunoblot analysis of separated membrane proteins using the monoclonal antibody (MAb) M3/38. Paraffin-embedded sections from normal, pancreatitis, and cancerous human pancreatic tissue and normal, N-nitrosobis(2-oxopropyl)amine (BOP)-treated hyperplastic, and cancerous hamster pancreatic tissues were processed immunohistochemically for galectin-3 using the MAb M3/38.

RESULTS

Galectin-3 was heavily expressed in cytoplasmic and nuclear regions of 50% of normal human pancreatic tissue. Expression of galectin-3 in ductal cells in chronic pancreatitis and cancerous pancreatic tissue was increased over normal and was more uniform (>95% cells/duct stained). Normal hamster pancreatic ducts showed weak or no expression of galectin-3. Hyperplastic pancreatic ductal cells from BOP-treated hamsters heavily expressed galectin-3 (60-95% cells/duct stained). Galectin-3 expression in ductal cells in cancerous pancreatic lesions was increased to >95%. Galectin-3 was also detected in the pancreatic nerves in all human tissue specimens tested.

Modification of blood group A expression in human pancreatic tumor cell lines by inhibitors of N-glycan processing

Pancreatic adenocarcinomas induced in Syrian hamsters by N-nitrosobis (2-oxopropyl)amine (BOP) treatment express blood group A (BGA) antigen, which was previously shown by this lab to be expressed on multiantennary asparagine (Asn)-linked glycans attached to membrane glycoproteins. To determine if a similar expression pattern was found in humans, three human pancreatic ductal adenocarcinoma cell lines (CD18, CD11, and Capan 1) from individuals of blood type A were analyzed and shown to express BGA antigen on membrane glycoproteins similar in molecular mass to those found in hamster tumor cells. The BGA antigen was located on Asn-linked oligosaccharides in all three human cell lines, as indicated by loss of activity after peptide:N-glycosidase F (PNGase F) treatment. Also, as shown previously in hamster pancreatic tumor cells, BGA expression at the surface of the human cell lines was blocked by growth of the cells in media containing deoxymannojirimycin (dMM), an inhibitor of mannosidase I. These results demonstrate that the BGA antigen is on Asn-linked glycans in human pancreatic adenocarcinoma cells and that these glycoproteins are processed similarly to the BGA glycoproteins in hamster pancreatic adenocarcinoma.

Relationship between blood group-A antigen expression and malignant potential in hamster pancreatic cancers

The loss of expression of the ABH blood group antigens is suggested to be associated with more aggressive behavior of cancers. We have compared the growth behaviors of two hamster pancreatic cancer cell lines with different blood group-A expressions. PC-1.0 cells, which expressed blood group-A antigen poorly, showed a faster growth in vitro and in vivo when implanted into the pancreas of homologous animals, whereas PC-1.2 cells, all of which express the antigen, had a slower growth rate both in vitro and in vivo. PC-1.0 also tended to metastasize, whereas PC-1.2 cells grew primarily locally. The allografts of both PC-1.2 cells (PC-1.2AG) and PC-1.0 cells (PC-1.0AG) and the metastases of PC-1.0 cells expressed blood group A antigen in a similar rate. There was no significant difference in the number of A-antigen positive cells (A+) between the PC-1.2AG and PC-1.0AG, although the expression of A antigen in PC-1.0AG showed a greater heterogeneity. The combined immunohistochemistry and autoradiography did not show any significant differences in the labeling index of A+ or A- cells between the two allografts. Thus, the results indicate that blood group A antigen expression is unrelated to malignancy in this model. The faster growth rate of PC-1.0 cells may be due to their shorter cell cycle.