DOES THE DETERMINATION OF MUCINS M1 IN PANCREATIC CYSTIC LESIONS IMPROVE THE DIAGNOSIS ?

SOME QUESTIONS CONCERNING M1 MUCINS M1 MUCINS AS MARKERS OF PRECANCEROUS COLON LESIONS

History

• « Ex nihilo, nihil ». First data suggesting that mucins may be tumour markers were obtained by histochemical methods some twenty years ago. In the course of dimethylhydrazine-induced formation of tumours in the rat, Dr Filipe observed, by HID alcyan blue staining, an abnormal expression of sialomucins several weeks following the injection of the carcinogen.
• The expression of the sialomucins increased progressively until the apparition of adenocarcinomas. Similarly, in man, sialomucins were detected in precancerous lesions (adenomas) and in the colon mucosa adjacent to adenocarcinomas, while no sialomucins were visualised in normal colon mucosa.
• These observations clearly indicated qualitative alterations of mucin expression as an early event in colon cancerogenesis. Thus, the detection of mucins could be considered as a method to screen for colon cancer in an early stage. It was clear however that a better characterisation of the mucins was needed. Immunochemical methods appeared to be the best first approach, as the use of antibodies could allow to identify the mucins and to develop methods for their quantitation.

  Gastric M1 mucins

• At the Institute of Cancer Research, Villejuif, it was immunised rabbits against mucins of different origins (stomach, small intestine, colon,, mucinous ovarian cysts). In this manner, we have characterised three large families of immunochemically and biochemically different mucins termed M1, M2 and M3. They are produced respectively by mucus-secreting cells of the surface epithelium of gastric mucosa (M1), gastric glands and Brünner cells (M2) and by goblet cells of the intestinal mucosa (M3). Two different intestinal mucins were subsequently distinguished, M3C expressed in the colon and M3I found essentially in the small intestine. We have shown that out of these five groups of organ-specific mucinous antigens, M1 was expressed in the colon during fœtal development but was absent from the normal colon tissue after birth. The important observation was that M1 was re-expressed in precancerous states of the colon, both in the rat and man, and, although less frequently, in human colon adenocarcinomas. In the rat, M1-positive goblet cells were found 15 days after carcinogen injection, within histologically normal colon mucosa. In addition, the M1 pattern in histologically identifiable lesions observed in the rat in the course of chemically induced carcinogenesis, was similar to that of human colon mucosa in the vicinity of adenocarcinomas , suggesting that human colon carcinogenesis concerns a large portion of the colon tissue.
• M1 mucins were discovered in a rather unusual fashion, using solely immunohistochemical methods. The initial studies were followed by the production of monoclonal antibodies whose immunoreactivity displayed the same pattern in the normal gastro-intestinal tract and precancerous colon (adenomas) as that seen with polyclonal antibodies. In the eighties, this novel methodology allowed us to specify the biochemical nature of the M1-immunoreactivity. Seven epitopes were characterized (M1a-g), all associated with the peptide chain and sensitive to reduction with mercaptoethanol; thus, they can be considered as « conformational » epitopes.
• More recently, genes coding for mucins (MUC genes) were cloned, and organ specificities were characterised by hybridisation methods. It turned out that MUC2 codes for intestinal mucins with the distribution characteristic to that of the antigens M3, MUC5AC for the mucins produced by the surface epithelium of the gastric mucosa, thus corresponding to the M1 antigens, and MUC6 codes for the mucins secreted by gastric glands and Brünner glands, source of antigens M2.
• The expression of a portion of MUC5AC cDNA by infecting insect cells Sf9 with recombinant baculovirus has led to the confirmation of the conjecture that MUC5AC codes for (at least part of) M1-immunoreactivity. Indeed, in situ hybridisation studies have shown that MUC5AC is expressed in the same tissues as M1, i.e. fœtal colon mucosa and precancerous colon lesions, and is absent from normal colon mucosa. This approach confirmed the oncofœtal expression of gastric mucins in the colon. The promoter of MUC5AC has been recently characterised, and its further study should help to elucidate the reasons for its derepression in the precancerous colon tissue.

  ... AND THE PANCREAS

• Gastric mucins M1 are also associated with pancreatic carcinogenesis. Expressed in 93% (82/88) of ductal pancreatic adenocarcinomas, they are undetectable in the fœtal as well as adult normal pancreas, and are an example of the ectopic production of mucins in this type of the cancer of the pancreas.
• As in the colon, the M1 mucins are expressed in precancerous lesions such as mucinous pancreatic cystadenomas, but not in serous adenomas considered as benign. M1 mucins are thus early markers of pancreatic as well as colon cancerogenesis. This is the basis of their potential importance in the diagnosis of precancerous cysts. These observations have been verified by in situ hybridisation with a MUC5AC probe: absence of expression in normal pancreatic tissue, positive signal in 96% (24/25) of invasive ductal adenocarcinomas, as well as in the small number of mucinous cystadenomas studied so far.

  HOW MUCH IS KNOWN ABOUT THE GENE CODING FOR M1 MUCINS ?

• As mentioned above, MUC5AC gene codes for at least one component of M1 immunoreactivity. The MUC5AC mRNA is 16.6 kb long, coding for 5525 aminoacids, so that the molecular weight of the peptide backbone of this mucin molecule is approximately 550000 Da. A large portion of the corresponding cDNA has been sequenced. The central portion contains repeated sequences of 24 bp characteristic of this gene. These repeated sequences code for peptides (8 aminoacids) rich in serines and threonines, corresponding to the hyperglycosylated region of the molecule (T-domain). The structure of the MUC5AC gene is not yet completely known. The 3' portion contains 18 exons and displays a strong resemblance with the MUC5B gene. The MUC5AC gene is situated on the chromosome 11p15.5, together with the genes MUC2, MUC5B and MUC6. All these genes are likely derived from a common ancestor.
• In the promoter of the MUC5AC gene, several consensus transcription-regulatory sequences have been identified: TATA-box, NF1, NFkB. This last sequence may play a role in the regulation of the expression of MUC5AC in inflammation.

  WHERE ARE SITUATED THE EPITOPES M1?

  Structure of gastric mucins

• Gastric mucins are very high molecular weight glycoproteins (>106 Da) containing 80% carbohydrates (branched saccharide chains) attached to a long peptide backbone (5525 aminoacids for the MUC5AC gene product; see above). These molecules contain two types of structures; i. hyperglycosylated, relatively rigid linear structures consisting of tandem repeats of 8 aminoacids rich in serines and threonines (T-domain); and ii. globular structures, weakly glycosylated « naked regions »), rich in cysteins implicated in the polymerisation and gel formation. The « naked regions » in the C- and N-terminal portions of the molecule are particularly important, but sparse, small clusters of « naked regions » can be found also in the central, hyperglycosylated portion where T-domains are predominant.

  Localisation of epitopes on M1 mucins

  • The M1 epitopes are situated in the « naked regions » . This can be deduced from the fact that they are destroyed by reduction with ß-mercaptoethanol (depending on S-S bonds) and by proteolytic enzymes (incapable of digesting T-domains). They are localised essentially in the N- and C-terminal regions. The position of one epitope, M1-f, has been more precisely established: it is situated in the C-terminal portion of the molecule. However, the « naked regions » situated in the central portion of the mucin molecule could also carry M1 epitope.
  • The full length cDNA not being available, their number and exact positions have not been established. It may also be difficult to identify the peptide sequences carrying the individual epitopes because of their conformational character: a given epitope may be constituted by two or more elements of the primary structure, distant on the backbone but proximal in space.
    

    HOW TO ASSAY M1 MUCINS ?

    The method used is based on the common principle of ImmunoRadioMetric Assay (IRMA), routinely employed in clinical biochemistry laboratories to determine tumor markers. It has been described in a recent issue of Mucus Dialogue.

    Antibodies

  • Anti-gastric mucin antibodies are essential for this assay. Organ specificity is very important since, as mentioned above, different organs secrete different mucins encoded by specific genes. To ensure the specificity of the assay, we have chosen to work with monoclonal anti-M1 antibodies raised in our laboratory.
  • These antibodies recognize a gastric mucin known by now to be encoded by the MUC-5AC gene, expressed in certain precancerous states (colon, pancreas).

    Reference M1 antigen

  • Certain biological fluids, for instance those present in ovarian mucinous cysts (particularly those of the « pure endocervical » type, according to the histological classification defined by Fenoglio ) contain exclusively M1 mucins, in large amounts. This material is used as standard solution since more than 10 years, because of its high concentration of M1.

  • We have arbitrarily defined the solution of 1 ng/ml (weight of lyophilised powder) as 1M1 unit.

    Conditions of sampling of biological fluids

  • Because of the sensitivity of M1 mucins to proteases, it is recommended to take samples as much as possible in a sterile fashion, store them at 4°C or frozen at -18°C, and deliver them as soon as possible to the laboratory where the analysis is to be carried out. One ml is sufficient for the assay.
  • However, it is always useful to dispose of larger quantity of the material, in order to repeat the assay if necessary, and to determine the contents of other markers (CEA, CA19.9, CA72-4).
    
    

    CONCLUSION

    The data available up to now show that a very high level of CEA (> 400 ng/ml) or of M1 mucins (> 5000 units/ml) in the pancreatic cyst fluid strongly suggest the presence of a mucinous cystadenoma of the pancreas. A very low level of these markers (CEA < 20 ng/ml, M1 < 50 units/ml) is a strong indication of a serous cystadenoma. In this latter situation, other biological parameters (activities of pancreatic enzymes) have to be considered, and the differential diagnosis between the mucinous and serous cystadenoma is a more delicate problem. Further studies are needed to determine the place of the biological markers in the diagnosis of pancreatic cystic lesions.

    
    Index de la page .
    coproweb@free.fr