Gastric mucosa abnormalities and tumorigenesis in mice lacking the pS2 trefoil protein

The pS2/TFF1 trefoil factor, from basic research to clinical applications

pS2/TFF1 trefoil factor is normally expressed in the stomach, and is found ectopically in gastrointestinal inflammatory disorders and in various carcinomas. It is involved in stomach ontogenesis and in the maintenance of the integrity of the mucosa, and may represent a pharmacological tool for prevention and healing of gastrointestinal ulcerations. In breast cancer, it can be used to select patients suitable for hormone therapy. pS2/TFF1 is a pleiotropic factor involved in mucin polymerization, cell motility, cell proliferation and/or differentiation, and possibly in the nervous system.

Short-chain fatty acids inhibit intestinal trefoil factor gene expression in colon cancer cells

Intestinal trefoil factor (ITF) gene expression was detected in five colon cancer cell lines. ITF was synthesized by mucous cells of LIM 1215 and LIM 1863 lines, from which it is secreted constitutively. The ITF mRNA transcript was estimated to be 0.6 kb. In LIM 1215 cells, the expression of ITF was potently and dose-dependently inhibited by short-chain fatty acids (butyrate > propionate > acetate) within 8 h of application. The inhibitory effect of butyrate was ablated by actinomycin D and preceded its effects on differentiation of LIM 1215 cells as indicated by induction of alkaline phosphatase activity and counting of periodic acid-Schiff-positive cells. The human ITF promoter contained an 11-residue consensus sequence with high homology to the butyrate response element of the cyclin D1 gene. Mobility shift assays show specific binding of this response element to nuclear protein extracts of LIM 1215 cells. We conclude that butyrate inhibits ITF expression in colon cancer cells and that this effect may be mediated transcriptionally and independently of its effects on differentiation.

Aspects of the biology of regeneration and repair in the human gastrointestinal tract

The main pathways of epithelial differentiation in the intestine, Paneth, mucous, endocrine and columnar cell lineages are well recognized. However, in abnormal circumstances, for example in mucosal ulceration, a cell lineage with features distinct from these emerges, which has often been dismissed in the past as 'pyloric' metaplasia, because of its morphological resemblance to the pyloric mucosa in the stomach. However, we can conclude that this cell lineage has a defined phenotype unique in gastrointestinal epithelia, has a histogenesis that resembles that of Brunner's glands, but acquires a proliferative organization similar to that of the gastric gland. It expresses several peptides of particular interest, including epidermal growth factor, the trefoil peptides TFF1, TFF2, TFF3, lysozyme and PSTI. The presence of this lineage also appears to cause altered gene expression in adjacent indigenous cell lineages. We propose that this cell lineage is induced in gastrointestinal stem cells as a result of chronic mucosal ulceration, and plays an important part in ulcer healing; it should therefore be added to the repertoire of gastrointestinal stem cells.

Dimerization of human pS2 (TFF1) plays a key role in its protective/healing effects

Human pS2 (trefoil factor family 1, TFF1), a 60-amino acid member of the trefoil peptide family, forms dimers via Cys58 and may stimulate gut repair. The effects of dimeric pS2-TFF1 and monomeric pS2-TFF1 (Cys58 replaced by Ser58) were compared in models of wound healing. Rats given dimeric pS2-TFF1 at 25 and 50 micrograms/kg per h had 50 per cent and 70 per cent reduction in gastric damage induced respectively by indomethacin (20 mg/kg subcutaneously) and restraint (P < 0.01). Monomeric pS2-TFF1, at the same doses, was significantly less effective at reducing injury (about half the amount of protection, P < 0.01 vs. same doses of dimeric). The rate of migration of cells at the leading edge of wounded monolayers of the human colonic cell line HT29 was increased by addition of dimeric or monomeric forms of pS2-TFF1 (0.65-325 micrograms/ml). Dimeric pS2-TFF1 had a greater effect than the monomeric form at all doses tested (P < 0.05). Cell migration induced by pS2-TFF1 was blocked by a pS2-TFF1 antibody, but not by a transforming growth factor beta neutralizing antibody. pS2-TFF1 did not influence cell proliferation as assessed by thymidine incorporation. The increased biological effects of dimeric pS2-TFF1 might be due to direct interaction of Cys58 with a putative trefoil receptor or, more likely, dimerization of pS2-TFF1 might stabilize the interaction with its receptor. This may involve a bivalent interaction of residues on the surfaces of the two trefoil domains.

The pS2 protein in colorectal carcinomas and metastases

Expression of pS2 protein in 50 primary tumors, metastases and recurrent tumors of colorectal carcinomas has been analyzed by immunohistochemistry. Sixty percent of the primary tumors were at least focally positive for the antigen. There was no correlation between pS2 expression and histologic grade of the lesions. In contrast, pS2 expression in T4 and T3 tumors was significantly higher than in T2 carcinomas. Immunoreactions in carcinomas with distant metastases (MI) were stronger than in M0 cases. However, this difference did not reach statistical significance. The presence of lymph node metastases did not correlate with pS2 expression. High expression of pS2 in T4 and T3 carcinomas together with the finding of pronounced expression of the antigen at invasion fronts in single cases could be interpreted as a function in tumor cell invasion and motility. However, in metastases and recurrent tumors, pS2 expression did not differ from primary lesions (53% positive lesions). All in all, under consideration of the latter finding in particular and together with the randomly distributed immunopositive tumor cells and cell clusters in the majority of cases, it is more likely that the expression pattern of pS2 in colorectal carcinomas is a result of overall tumor cell heterogeneity.

The estrogen responsive element of the pS2 gene is recognized by a methylation sensitive DNA binding protein

The human pS2 gene is specifically expressed is a subclass of estrogen receptor containing human breast cancer cells. In the MCF7 cell line, its induction by estradiol is a primary transcriptional event. The exact location of its estrogen responsive element has been determined using a chimeric recombinant transfected into HeLa cells and a transient expression assay. In this study we found, using electrophoretic mobility shift experiments, that in HeLa cells the estrogen responsive element (ERE) of the pS2 gene is recognized by a methylation sensitive DNA binding protein (MSDBP) different from the estrogen receptor. Competition experiments have shown that the binding of this protein requires at least one CpG in the center of the palindromic sequence and that imperfect palindromic sequences are also recognized. Although the presence of CpG is necessary, CpG-rich oligonucleotides, containing consensus sequences for Sp1 or AP2, do not interfere with its binding to the pS2 oligonucleotide, indicating that the ERE sequence itself participates in the specificity of its binding. This protein binds the pS2 sequence with a relatively high affinity (apparent Kd = 10(-10) M) and its binding is strongly reduced by the methylation of the cytosines at CpG sites. UV cross-linking experiments and peptide mapping indicate that this protein has an apparent molecular weight of 46 kDa and is present in several cell lines, including non-human cell lines. Taken together, these data suggest that this protein might have a potential role in regulating gene activity or in chromatin structure of some genes possessing an ERE.

Mice lacking secretory phospholipase A2 show altered apoptosis and differentiation with Helicobacter felis infection

BACKGROUND & AIMS

Infection with Helicobacter pylori uniformly leads to a chronic superficial gastritis that may progress to atrophic gastritis, a premalignant process. A mouse model of Helicobacter felis infection was used to study possible genetic determinants of the response to infection.

METHODS

Three inbred mouse strains with known secretory phospholipase A2 (sPLA2) genotypes [BALB/c (+/+), C3H/HeJ (+/+), and C57BL/6 (-/-)] were orally infected with H. felis and examined longitudinally using routine histology, immunocytochemistry, electron microscopy, proliferating cell nuclear antigen, terminal deoxynucleotidyl transferase-mediated deoxyuridine triphosphate nick-end labeling, and Northern and Western blot studies.

RESULTS

Only the C57BL/6 strain showed increased gastric fundic proliferation and apoptosis in response to infection. In addition, the C57BL/6 mouse showed a marked loss of parietal and chief cells, along with a marked expansion of an aberrant gastric mucous cell lineage that stained positive for spasmolytic polypeptide. In contrast, no significant change in these cell types was observed in BALB/c and C3H/HeJ strains. Increased expression of sPLA2 was observed in BALB/c and C3H/HeJ after H. felis infection, whereas sPLA2 expression was absent in C57BL/6 mice.

CONCLUSIONS

H. felis infection leads to increased apoptosis and altered cellular differentiation in the C57BL/6 mouse, a strain that lacks gastric sPLA2 expression. Because sPLA2 has been identified recently as the MOM1 (modifier of MIN) locus that influences polyp formation in the colon, these studies suggest that sPLA2 may also influence the gastric epithelial response to Helicobacter infection.

Estrogen receptor mediated inhibition of cancer cell invasion and motility: an overview

In this overview of results from our laboratory, we address the question of the role of estrogens during early steps of metastasis, involving cell invasion through the basement membrane and cell motility. The motility of several estrogen receptor (ER) positive breast (MCF7, T47D) and ovarian (BG-1, SKOV3, PEO4) cancer cell lines was studied using a modified Boyden chamber assay. We observed, in all cases, estradiol induced inhibition of cancer cell invasion and motility. A similar inhibitory effect of estradiol was found when the wild-type ER alpha was stably transfected in the ER-negative MDA-MB231 cells and 3Y1-Ad12 cancer cells. The mechanism of this inhibitory effect is unknown. In ovarian cancer, however, it may involve intermediary proteins such as fibulin-1, an extracellular matrix protein that strongly interacts with fibronectin and which is induced by estrogen and secreted by ovarian cancer cells. We conclude that estrogens in ER-positive breast and ovarian cancers have a dual effect, since they stimulate tumor growth but inhibit invasion and motility. This may be consistent with the good initial prognostic value of ER-positive breast cancers compared to ER negative breast cancers noted in several clinical studies.

Oral human spasmolytic polypeptide protects against aspirin-induced gastric injury in rats

Spasmolytic polypeptide (SP) is a member of the trefoil peptide family; gut peptides that participate in the protection and repair of the gastric mucosa. Previous studies have failed to agree on the mode of action of human SP (hSP). We investigated the effect of orally administered human SP on the protection and repair of rat gastric mucosa in an established in vivo model of damage induced by the non-steroidal anti-inflammatory drug aspirin (ASA). The integrity of the gastric mucosa was quantified in four ways: the temporal change in transmucosal potential difference (PD), area of macroscopic damage by planimetry, relative area of microscopic damage by histological morphometry, and the number of deep erosions per centimetre of mucosa sectioned. Human SP (200 micromol/L) administered orally before, or in combination with ASA significantly reduced the fall in PD, the area of microscopic damage, and the number of deep erosions (P < 0.05). The area of macroscopic damage was significantly reduced only in rats where hSP (200 micromol/L) was given in conjunction with ASA (P < 0.05). Human spasmolytic polypeptide (70 or 200 micromol/L) administered after ASA failed to hasten the re-establishment of PD or stimulate the repair of the gastric mucosa in the 90 min following injury (P > 0.05, compared with ASA alone). We conclude that hSP prevents gastric mucosal damage by its topical actions, probably by a rapid interaction with luminal mucins or epithelial cells, but fails to stimulate early restitution in the injured gastric mucosa.

Gene disruption in mice: models of development and disease

Gene targeting technology in mice by homologous recombination has become an important method to generate loss-of-function of genes in a predetermined locus. Although the inactivation is limited to irreversible alteration of chromosomal DNA and a surprising variety of genes have given unexpected and disappointing results, modification of the basic technology now provides additional choices for a more specific and variety of manipulations of the mouse genome. This includes conditional cell-type specific gene targeting, knockin technique and the induction of the specific balanced chromosomal translocations. In the past decade this technique not only generated a wealth of knowledge concerning the roles of growth factors, oncogenes, hormone receptors and Hox genes but also helped to produce animal models for several human genetic disorders. In the future it may provide more powerful and necessary tools to dissect the psychiatric disorders, understanding the complex central nervous system and to correct the inherited disorders.

TFF1 gene expression in human medullary thyroid carcinoma

Medullary thyroid carcinoma (MTC) is an uncommon tumour of calcitonin-secreting C-cells of the thyroid gland. This cancer represents an important potential model for the study of mechanisms of human epithelial cell transformation. Although recent studies have identified the gene involved in familial forms of MTC, little is known about the molecular pathogenesis of the sporadic variants of this tumour. The biological and prognostic significance of TFF1 expression, particularly in diverse human malignancies, suggests that the TFF1 protein could have a role in human neoplasia. Furthermore, in prostate cancer it has been demonstrated that TFF1 expression is closely associated with premalignant changes and neuroendocrine differentiation. In the present study, the expression of TFF1 was analysed in 18 human MTCs, comprising sporadic and familial tumours, C-cell hyperplasia, and one case of lymph gland metastasis. TFF1 expression was also examined in the cultures of a human MTC-derived tumour cell line (TT cell line). The results showed that ten sporadic tumours, three hereditary tumours (including C-cell hyperplasia), and one lymph gland metastasis displayed TFF1 immunoreactivity. Indirect fluorescence immunocytochemistry and Western blotting revealed that the TFF1 protein was strongly expressed in the TT cells. Northern analysis revealed that tumours and TT cells expressed the TFF1 transcript. Although the function of TFF1 protein in the carcinogenesis of MTC remains to be elucidated, its expression in the majority of cases of both sporadic and hereditary tumours, metastatic tumours, and in C-cell hyperplasia suggests that it may contribute to the pathogenesis of MTC.

Loss of pS2 protein expression is an early event of intestinal-type gastric cancer

To investigate the prevalence of pS2 expression in gastric cancer with respect to tumor histopathology, intestinal metaplasia and Helicobacter pylori (H. pylori) infection, pathologic specimens of 91 patients with gastric cancer were immunostained for pS2. Such immunoreactivity was correlated with the status of H. pylori infection, tumor staging, histology, subtyping, and associated intestinal metaplasia. Positive pS2 staining was seen throughout all non-neoplastic epithelia, and in all 9 patients with the complete type of intestinal metaplasia. In contrast, 21 of 45 incomplete type of intestinal metaplasia had negative pS2 staining (P < 0.001), and 54 out of 91 tumors (59.3%) showed loss of pS2 expression in the cancer tissues proper. There was no correlation of pS2 expression with age, gender, depth of invasion, duodenal involvement, lymph node metastasis, venous invasion or H. pylori infection. Negative pS2 staining was significantly higher in the intestinal (74.5%) and Borrmann type I, II, III (64.2%) tumors than the diffuse (43.2%, P < 0.005) and Borrmann type IV (20%, P < 0.05) tumors. Our results indicate that loss of pS2 expression may occur as an early event in the malignant transformation process of intestinal-type tumors.

Trefoil peptides: mitogens, motogens, or mirages?

Healing of mucosal damage occurs in two phases: restitution of mucosal integrity followed by remodeling with recreation of mucosal architecture. Models of these phenomena include cryoprobe-induced ulcers, NSAID lesions, and surgical anastomosis. Three trefoil peptides are expressed constitutively by epithelial cells in specific regions of the GI tract: pS2 (gastric), spasmolytic polypeptide (SP, gastric and Brunner's glands), and intestinal trefoil factor (ITF, goblet cells). Altered expression occurs in reparative epithelium and adjacent mucosa. In cryoprobe ulceration, rSP mRNA abundance doubles within 2 h, with rITF mRNA becoming detectable after 2-3 days. TGF-alpha and EGF mRNAs do not increase as rapidly as rSP or to the same extent as rITF. Indomethacin lesions of gastric mucosa show increased SP immunoreactivity deep in damaged glands within hours. Surgical anastomotic damage increases rITF mRNA levels at the ulcer edge and sometimes rSP mRNA and peptide in para-anastomotic crypts. Initially, trefoil peptides were viewed as mitogens. However, they are in fact motogens, able to promote cell migration, and may possibly be morphogens. Interactions occur between trefoils and other wound healing peptides (FGFs and EGF). Trefoil peptides appear to be of considerable importance to mucosal healing and might constitute a biologic target of therapeutic relevance.

Transgenic mouse models for tumor suppressor genes

The identification and cloning of tumor suppressor genes has mostly relied on familial human cancer predisposition syndromes and reverse genetics. Recent advances in manipulating the mouse genome by gene targeting techniques in embryonic stem (ES) cells has led to the generation of mutant mouse models mimicking many human syndromes. Mice lacking one or both alleles of known tumor suppressor genes have been generated to evaluate the normal function of these genes in vivo. These mice have proven to be highly susceptible to tumor development, indicating that the mouse is a potent in vivo assay system for tumor suppressor genes. The initiation of gonadal tumor development in mice lacking both copies of the alpha-inhibin gene demonstrates that this assay is also useful for identifying new tumor suppressor genes. In the future, murine ES cell/gene targeting strategies will continue to be used to identify novel tumor suppressors and analyze their in vivo roles in growth control.

Estrogen receptor alpha and beta expression in upper gastrointestinal tract with regulation of trefoil factor family 2 mRNA levels in ovariectomized rats

This study compared expression of two estrogen receptor (ER alpha and ER beta) genes in the rat upper gastrointestinal tract and the effects of 17 beta-estradiol administration on gastric trefoil factor family (TFF) mRNA steady-state levels in ovariectomized rats. Estrogen receptor alpha and beta cDNA fragments from fundic mucosa were cloned by reverse transcriptase polymerase chain reaction (RT-PCR) and sequenced. Both ER subtypes were detected in fundus, antrum and duodenum by RT-PCR. Northern analysis of poly(A)+ mRNA from fundic mucosa showed that ER alpha mRNA is expressed as a single transcript at 6.5 kb and ER beta is expressed as multiple transcripts with major transcripts ranging from 1.1-4.7 kb. ER beta mRNA was expressed in greater abundance than ER alpha mRNA. Fundic TFF2 mRNA steady-state levels were increased by 17 beta-estradiol administration in ovariectomized rats with no significant change in TFF1 mRNA levels. These studies show expression of both ER subtypes in the rat upper gastrointestinal tract with regulation of TFF2 mRNA by 17 beta-estradiol. These results suggest that estrogens, probably acting via ER beta, have a direct role in regulating gastric physiology.

Characterization of dopuin, a polypeptide with special residue distributions

A 62-residue polypeptide, dopuin, has been isolated from pig small intestine. It is distinguished by an N-terminal part with a high content of proline (7 in a 26-residue segment), a C-terminal part with a high proportion of histidine (3 in a 9-residue segment), and six half-cystine residues in three intrachain disulphide bridges (connecting positions 22-25, 23-54 and 35-44). The Cys and Pro distributions suggest a tight and special conformation. In contrast to PEC-60 and somatostatin, it has no established inhibitory effect on insulin secretion. At 10 nM concentration, a weak inhibitory tendency is less than half of that of the other two peptides. Like gastrointestinal trefoil peptides, dopuin has three disulphide bridges, Ala-Pro segments, and many charged residues, but they are differently distributed and dopuin belongs to a separate, apparently novel family. However, dopuin is similar to a peptide corresponding to an expressed-sequence-tag cDNA of human fetal liver and spleen, establishing the nature of the mature form of the product of this cDNA, and showing a general tissue, age, and species distribution of this peptide. A truncated form of vimentin, composed of its C-terminal 37 residues, vimentin-C37, was also purified and structurally characterized. These two peptides increase the complexity of known intestinal polypeptides and at least dopuin has properties compatible with specific biofunctions.

Homodimerization and hetero-oligomerization of the single-domain trefoil protein pNR-2/pS2 through cysteine 58

The single-domain human trefoil proteins [pNR-2/pS2 and human intestinal trefoil factor (hITF)] have seven cysteine residues, of which six are involved in maintaining the structure of the trefoil domain. The seventh does not form part of the trefoil domain and is located three residues from the C-terminus. The ability of the pNR-2/pS2 single trefoil domain protein to dimerize was examined by using recombinant protein with either a cysteine or a serine residue at this position by equilibrium ultracentrifugation, laser-assisted desorption MS, gel filtration and PAGE. pNR-2/pS2 Cys58 formed dimers, whereas pNR-2/pS2 Ser58 did not. Experiments in which the dimer was treated with thiol agents demonstrated that the dimer was linked via a disulphide bond and that the intermolecular disulphide bond was more susceptible to reduction than the intramolecular disulphide bonds. To examine whether dimeric pNR-2/pS2 was secreted by oestrogen-responsive breast cancer cells, which are known to express pNR-2/pS2 mRNA, conditioned medium was separated on non-denaturing polyacrylamide gels, transferred to PVDF membrane and reacted with antiserum against pNR-2/pS2. Monomeric and dimeric pNR-2/pS2 were detected but the majority of the protein reactivity was associated with a larger protein. Treatment of this protein with thiol agents suggested that it is an oligomer containing pNR-2/pS2 linked to another protein by a disulphide bond. These studies suggest that the biological action of pNR-2/pS2 single-domain trefoil protein might involve the formation of homodimers or oligomers with other proteins.

Intestinal trefoil factor (TFF 3) and pS2 (TFF 1), but not spasmolytic polypeptide (TFF 2) mRNAs are co-expressed in normal, hyperplastic, and neoplastic human breast epithelium

pS2-TFF 1 is expressed in breast cancers and has been investigated as a potential prognostic factor reflecting oestrogen dependence. The relationship to the expression of other trefoil peptides, human spasmolytic polypeptide (hSP-TFF 2) and intestinal trefoil factor (hITF/hPI.B-TFF 3) is documented here. Fifty-seven breast specimens were selected from surgical pathology archives and included five normal breasts (two lactating), seven benign proliferative lesions, 11 ductal carcinomas in situ (DCIS), three lobular carcinomas in situ (LCIS), 24 invasive ductal carcinomas (IDC), and seven invasive lobular carcinomas (ILC). The comparative distribution of trefoil mRNAs was assessed by in situ hybridization using 35S-labelled riboprobes and immunohistochemical staining for pS2-TFF 1 and hSP-TFF 2. pS2-TFF 1 and hITF/hPI.B-TFF 3 mRNA were focally present at low signal intensity in normal and benign breast. Both pS2-TFF 1 and hITF/hPI.B-TFF 3 were expressed in all DCIS, LCIS and ILC, and 21/24 IDC. Overall, expression patterns of pS2-TFF 1 and hITF/hPI.B-TFF 3 coincided, but hITF/hPI.B-TFF 3 mRNA was usually found in a greater proportion of cells. Expression of hSP-TFF 2 peptide or mRNA was not detected in any of these cases. MCF 7 breast carcinoma cells also expressed hITF/hPI.B-TFF 3 and pS2-TFF 1 mRNAs but not hSP-TFF 2. hITF/hPI.B-TFF 3 co-expression with pS2-TFF 1 may act as a prognostic factor, but also raises questions about the regulatory pathway for pS2-TFF 1 hITF/hPI.B-TFF 3. Trefoil factors have effects on cell motility and spreading in vitro, and co-expression of hITF/hPI.B-TFF 3 with pS2-TFF 1 could be functionally significant if they form a heterodimer or compete for receptor binding. Absence of hSP-TFF 2 expression may be of equal relevance to tumour cell biology.

Trefoil proteins: their role in normal and malignant cells

The three human trefoil proteins pS2, human intestinal trefoil factor (hITF), and human spasmolytic polypeptide (hSP) are expressed principally in the mucosa of the gastrointestinal tract. They are also expressed in a variety of other normal tissues and tumours. This review discusses the pattern of expression of trefoil proteins in cancer, current views on the biological functions of trefoil proteins, and the way in which the expression of trefoil proteins may influence the behaviour of cancer cells.

Expression of a breast-cancer-associated protein (pS2) in human neuro-endocrine tumours

pS2 protein expression has been demonstrated in a range of malignant tissues in an oestrogen-independent pathway. Recently, it has been demonstrated that pS2, in prostate cancer, is closely associated with neuro-endocrine differentiation. In the present study, we have analyzed, by immunohistochemistry along with microwave antigen retrieval, the expression of pS2 protein in a retrospective series of 236 human primary neuro-endocrine tumours and attempted to correlate this with the clinicopathologic features of patients and the presence of oestrogen receptor (ER). pS2 immunoreactivity was detected in 42% of small-cell lung carcinomas, 36% of lung carcinoids, 33% of phaeochromocytomas, 38% of carotid-body tumours, 31% of pancreatic neuro-endocrine tumours, 60% of stomach carcinoids, 55% of ileal carcinoids, 23% of appendiceal carcinoids and 86% of rectal carcinoids respectively in more than 10% tumour cells. No pituitary tumours displayed pS2 immunoreactivity. pS2 transcript was also detected in lung carcinoid and carotid-body tumours by Northern-blot analysis. There was a statistically higher incidence of pS2 expression in carcinoid tumours of the ileum and rectum than in those of the appendix. No association was observed between pS2 expression and the occurrence of the carcinoid syndrome; nor was any correlation observed between the occurrence of pS2 immunoreactivity and that of ER. Our results suggest that the expression of the pS2 protein in a wide spectrum of neuro-endocrine tumours may be implicated in the pathogenesis and progression of some neuro-endocrine tumours in an oestrogen-independent pathway.

Rolling in the clover: trefoil factor family (TFF)-domain peptides, cell migration and cancer

Trefoil factor family (TFF)-domain peptides 1-3 are mucin-associated molecules, largely found in epithelia of gastrointestinal tissues. Structurally similar, resistant to enzymatic degradation, they are up-regulated around areas of epithelial damage such as ulcers. Transgenic expression or exogenous peptide ameliorates or prevents gastric mucosal damage due to indomethacin and some are rapidly up-regulated after cryogenic burns. A role in promoting cell migration is strongly suggested. Knockout mice lacking TFF1 or TFF3 show significant pathology, with the former developing gastric tumours. A recent Conference Philippe Laudat agreed upon a new nomenclature for these peptides.

Involvement of DNA methylation in the control of the expression of an estrogen-induced breast-cancer-associated protein (pS2) in human breast cancers

pS2 gene has been used to investigate the relationship between alterations of DNA methylation patterns in human tumors and gene expression. The expression of pS2, which is transcriptionally controlled by estrogens in breast cancer cell lines, is restricted to estrogen-receptor-rich human breast tumors. We found that the CCGG site within the promoter/enhancer sequence of pS2 was hypomethylated in estrogen-receptor-rich breast tumors expressing this gene. The amount of DNA molecules unmethylated at this site was related to the amount of pS2 mRNA detected in the samples. The demethylation of this region, which contains the estrogen responsive element, was confirmed by genomic sequencing. Transient expression of functional human estrogen receptors stimulated the expression of the endogenous pS2 in HeLa cells, but failed, in BT-20 cells, to stimulate expression of this gene. Since the promoter/enhancer region of pS2 is unmethylated in HeLa cells and methylated in BT-20 cells, these data also support the hypothesis that DNA methylation might be involved in the control of pS2 expression.

High-resolution solution structure of human pNR-2/pS2: a single trefoil motif protein

pNR-2/pS2 is a 60 residue extracellular protein, which was originally discovered in human breast cancer cells, and subsequently found in other tumours and normal gastric epithelial cells. We have determined the three-dimensional solution structure of a C58S mutant of human pNR-2/pS2 using 639 distance and 137 torsion angle constraints obtained from analysis of multidimensional NMR spectra. A series of simulated annealing calculations resulted in the unambiguous determination of the protein's disulphide bonding pattern and produced a family of 19 structures consistent with the constraints. The peptide contains a single "trefoil" sequence motif, a region of about 40 residues with a characteristic sequence pattern, which has been found, either singly or as a repeat, in about a dozen extracellular proteins. The trefoil domain contains three disulphide bonds, whose 1-5, 2-4 and 3-6 cysteine pairings form the structure into three closely packed loops with only a small amount of secondary structure, which consists of a short alpha-helix packed against a two-stranded antiparallel beta-sheet. The structure of the domain is very similar to those of the two trefoil domains that occur in porcine spasmolytic polypeptide (PSP), the only member of the trefoil family whose three-dimensional structure has been previously determined. Outside the trefoil domain, which forms the compact "head" of the molecule, the N and C-terminal strands are closely associated, forming an extended "tail", which has some beta-sheet character for part of its length and which becomes more disordered towards the termini as indicated by (15)N{(1)H} NOEs. We have considered the structural implications of the possible formation of a native C58-C58 disulphide-bonded homodimer. Comparison of the surface features of pNR-2/pS2 and PSP, and consideration of the sequences of the other human trefoil domains in the light of these structures, illuminates the possible role of specific residues in ligand/receptor binding.