Tissue stem cells and cancer stem cells: potential implications for gastric cancer.

Gastric cancer remains the second leading cause of death in the world today, making the search for its molecular and cellular basis an important priority. Though recognition of the tight link between inflammation and tumorigenesis is centuries old, only recently are the pieces of the etiological puzzle beginning to fall together. Recent advances in gastric stem cell biology appear to be central to this slowly resolving puzzle. At least two types of stem cells may be important. Resident adult or tissue stem cells may, in a chronically inflamed environment, slowly acquire a series of genetic and epigenetic changes that lead to their emergence as ''cancer stem cells''. This scenario has not yet been proven experimentally, although the first step, prospective recognition of a gastric stem cell has recently been conquered. Alternatively, the setting of chronic inflammatory stress and injury may lead to loss of the indigenous gastric stem cells from their niches; bone marrow derived stem cells may then be recruited to and engraft into the gastric epithelium. Such recruited cells have the potential to contribute to the tumor mass. Indeed, evidence supporting this scenario has been published. Here, we review these recent findings and discuss implications for the future.

EGF receptor activation stimulates endogenous gastrin gene expression in canine G cells and human gastric cell cultures.

Gastrin release from the antral gastrin-expressing cell (G cell) is regulated by bombesin and luminal factors. Yet, these same extracellular regulators do not stimulate expression of the gene. Since the gastric mucosa expresses large quantities of EGF receptor ligands such as TGFalpha, we examined whether EGF receptor ligands stimulate gastrin gene expression in gastrin-expressing cell cultures. EGF receptor activation of primary cultures stimulated gastrin gene expression about twofold; whereas bombesin treatment of antral G cell cultures stimulated gastrin release but not gene expression. EGF and TGFalpha were weak stimulants of gastrin release. EGF receptor activation of AGS human gastric adenocarcinoma cell line stimulated gastrin gene expression nearly fourfold; and gastrin reporter constructs transfected into AGS cells were stimulated more than fourfold by EGF. EGF induction was conferred by the previously defined GC-rich gastrin EGF response element (gERE) element located at -68 to -53 bp upstream from the cap site since a mutation of the gERE element abolished both basal and EGF induction. Moreover, EGF treatment of AGS cells stimulated binding of the transcription factor Sp1 to this element. Collectively, these results demonstrate that gastrin gene expression and gastrin release are regulated by different signaling pathways: gene expression by EGF receptor activation and gastrin secretion by neuropeptides and luminal factors.

ZBP-89 promotes growth arrest through stabilization of p53.

Transcription factor p53 can induce growth arrest and/or apoptosis in cells through activation or repression of downstream target genes. Recently, we reported that ZBP-89 cooperates with histone acetyltransferase coactivator p300 in the regulation of p21(waf1), a cyclin-dependent kinase inhibitor whose associated gene is a target gene of p53. Therefore, we examined whether ZBP-89 might also inhibit cell growth by activating p53. In the present study, we demonstrate that elevated levels of ZBP-89 induce growth arrest and apoptosis in human gastrointestinal cell lines. The ZBP-89 protein accumulated within 4 h, and the p53 protein accumulated within 16 h, of serum starvation without changes in p14ARF levels, demonstrating a physiological increase in the cellular levels of these two proteins. Overexpression of ZBP-89 stabilized the p53 protein and enhanced its transcriptional activity through direct protein-protein interactions. The DNA binding and C-terminal domains of p53 and the zinc finger domain of ZBP-89 mediated the interaction. A point mutation in the p53 DNA binding domain, R273H, greatly reduced ZBP-89-mediated stabilization but not their physical interaction. Furthermore, ZBP-89 formed a complex with p53 and MDM2 and therefore did not prevent the MDM2-p53 interaction. However, heterokaryon assays demonstrated that ZBP-89 retained p53 in the nucleus. Collectively, these data indicate that ZBP-89 regulates cell proliferation in part through its ability to directly bind the p53 protein and retard its nuclear export. Our findings further our understanding of how ZBP-89 modulates cell proliferation and reveals a novel mechanism by which the p53 protein is stabilized.

A GC-rich element confers epidermal growth factor responsiveness to transcription from the gastrin promoter.

Epidermal growth factor (EGF) and transforming growth factor alpha are important determinants of mucosal integrity in the gastrointestinal tract, and they act both directly and indirectly to prevent ulceration in the stomach. Consistent with this physiological role, EGF stimulates transcription of gastrin, a peptide hormone which regulates gastric acid secretion and mucosal growth. EGF stimulation of gastrin transcription is mediated by a GC-rich gastrin EGF response element (gERE) (GGGGCGGGGTGGGGGG) which lies between -54 and -68 in the human gastrin promoter. The gERE sequence also confers weaker responsiveness to phorbol ester stimulation. The gERE sequence differs from previously described EGF response elements. The gERE DNA sequence specifically interacts with a GH4 DNA-binding protein distinct from previously described transcription factors (Egr-1 and AP2) which bind GC-rich sequences and mediate transcriptional activation by growth factors. Furthermore, the gERE element does not bind the Sp1 transcription factor even though the gERE sequence contains a high-affinity Sp1-binding site (GGCGGG).

ZBP-89, a Krüppel-like zinc finger protein, inhibits epidermal growth factor induction of the gastrin promoter.

We have shown previously that a GC-rich element (GGGGCGGGGTGGGGGG) conferring epidermal growth factor (EGF) responsiveness to the human gastrin promoter binds Sp1 and additional undefined complexes. A rat GH4 cell line expression library was screened by using a multimer of the gastrin EGF response element, and three overlapping cDNA clones were identified. The full-length rat cDNA encoded an 89-kDa zinc finger protein (ZBP-89) that was 89% identical to a 49-kDa human factor, ht(beta), that binds a GTGGG/CACCC element in T-cell receptor promoters. The conservation of amino acids between the zinc fingers indicates that ZBP-89 is a member of the C2H2 zinc finger family subclass typified by the Drosophila Krüppel protein. ZBP-89 is ubiquitously expressed in normal adult tissues. It binds specifically to the gastrin EGF response element and inhibits EGF induction of the gastrin promoter. Collectively, these results demonstrate that ZBP-89 functions as a repressor of basal and inducible expression of the gastrin gene.

ZBP-89-induced apoptosis is p53-independent and requires JNK.

ZBP-89 induces apoptosis in human gastrointestinal cancer cells through a p53-independent mechanism. To understand the apoptotic pathway regulated by ZBP-89, we identified downstream signal transduction targets. Ectopic expression of ZBP-89 induced apoptosis through the mitochondrial pathway and was accompanied by activation of all three MAP kinase subfamilies: JNK1/2, ERK1/2 and p38 MAP kinase. ZBP-89-induced apoptosis was markedly enhanced by ERK inhibition with U0126. In contrast, inhibiting JNK with a JNK1-specific peptide inhibitor or dominant-negative JNK2 expression abrogated ZBP-89-mediated apoptosis. The p38 inhibitor SB202190 had no effect on ZBP-89-induced cell death. Protein dephosphorylation assays revealed that ZBP-89 activates JNK via repression of JNK dephosphorylation. Oligonucleotide microarray analyses revealed that ectopic expression of ZBP-89 downregulated expression of the dual-specificity phosphatase MKP6. Overexpression of MKP6 blocked ZBP-89-induced JNK phosphorylation and PARP cleavage. In addition, ectopic expression of ZBP-89 repressed Bcl-xL and Mcl-1 expression, but had no effect on Bcl-2. Silencing ZBP-89 with small interfering RNA enhanced both Bcl-xL and Mcl-1 expression. Taken together, ZBP-89-mediated apoptosis occurs via a p53-independent mechanism that requires JNK activation.

Pilot study of Biomarkers for predicting effectiveness of ramosetron in diarrhea-predominant irritable bowel syndrome: expression of S100A10 and polymorphisms of TPH1.

BACKGROUND: Serotonin type 3 receptor (5-HT3 R) antagonists are potentially useful therapeutic agents for diarrhea-predominant irritable bowel syndrome (IBS-D). To identify biomarkers predicting effectiveness of the 5-HT3 R antagonist (ramosetron) in IBS-D. METHODS: Irritable bowel syndrome-D Japanese subjects received 2.5 or 5 µg of ramosetron once daily for 4 weeks. Colonic mucosal S100A and tryptophan hydroxylase (TPH) mRNA expression levels were measured before treatment. Genomic DNA was extracted from blood and polymorphisms of TPH1 and TPH2 were analyzed. KEY RESULTS: Forty-two patients (27 men and 15 women, mean age 42 years) with IBS-D were included for analysis. Improvement of IBS symptoms was seen in 26 (61.9%). Baseline S100A10 (p = 0.02) and TPH1 (p = 0.02) expression were significantly higher in the ramosetron responders than in the non-responders. The frequencies of the TPH1 rs4537731G allele in linkage disequilibrium with the TPH1 rs7130929 T allele (11.5% vs 50%, p = 0.003; OR: 12; 95% CI: 2.1-69) along with TPH1 rs211105 C allele (3.8% vs 43.8%, p = 0.0003; OR: 19; 95% CI: 2.1-181) were significantly lower in the responders than in the non-responders. The mean scores of diarrhea at baseline were significantly higher (5.2 vs 3.7, p = 0.005) in patients with TPH1 rs211105 T/T than those with the G allele. CONCLUSIONS & INFERENCES: TPH1 gene polymorphisms and S100A10 expression, which correlate with 5-HT signaling were associated with ramosetron effectiveness in IBS-D, and may possibly lead to prospective identification of the resistance to treatment.

RAP1-like binding activity in islet cells corresponds to members of the Sp1 family of transcription factors.

Deletion and mutational analyses of the gastrin promoter have identified a binding site for the yeast transcription factor RAP1 relevant for transcriptional activation in islet cells. We here report that the mammalian transcription factors binding to this site in islet cells are the Sp transcription factor members Sp1 and Sp3. Furthermore, functional analyses revealed Sp1- and Sp3-mediated transcriptional activation of gastrin. These data reveal that the zinc finger proteins Sp1 and Sp3 do have similar binding specificities as the multifunctional yeast RAP1 protein.

EGF receptor activation of the human gastrin gene: a tale of two zinc finger transcription factor families.

Gastrin is one of the oldest and most actively studied of the gastrointestinal peptides. Adult expression of the mature peptide is restricted to the stomach and duodenum. However, it is antral gastrin that is subject to feedback regulation from the counter-regulatory peptide somatostatin as a function of gastric pH. Yet little is known about the molecular steps required to modulate the synthesis of this peptide. Using cell culture models, a GC-rich regulatory element called gERE has been identified that binds two families of zinc finger transcription factors--Sp and ZBP. Competitive binding of these two transcription factor families in association with changes in their phosphorylation state appears to mediate the positive and negative activation of this gene in response to EGF receptor ligands. These findings contribute to our understanding of gastrin gene expression in the presence of the activated ras oncogene, e.g., during colonic transformation.

Transient transcriptional activation of gastrin during sodium butyrate-induced differentiation of islet cells.

Transient expression of pancreatic gastrin corresponds to a period of rapid islet cell development. After birth gastrin expression silencing is coincidental with islet cell terminal differentiation, while persistent expression is accompanied with nesidioblastosis and reexpression observed in islet cell tumors. Experiments with transgenic animals suggested that gastrin might act synergistically with growth factors to stimulate islet cell development. The present study intended to establish an in vitro cell culture model to analyse the molecular events controlling gastrin gene activation and repression dependent on islet cell differentiation. Sodium butyrate, a proliferation-arresting compound has previously been shown to differentiate insulinoma cells while increasing insulin production. The present paper demonstrates concomitant transient increase in gastrin mRNA, intracellular and secreted gastrin during sodium butyrate treatment. Increased gastrin expression was due to activation or derepression of gastrin promoter activity as revealed by promoter analyses. This in vitro model mimics the expression pattern of gastrin and insulin observed during fetal islet cell development and provides an excellent tool to analyse the molecular mechanisms controlling gastrin gene activation and selective repression during islet cell differentiation.

Helicobacter pylori-induced atrophic gastritis progressing to gastric cancer exhibits sonic hedgehog loss and aberrant CDX2 expression.

BACKGROUND: The loss of sonic hedgehog is an early change that occurs in the mucosa prior to neoplastic transformation and correlates with the type of intestinal metaplasia. Aberrant expression of CDX has also been shown to correlate with the development of intestinal metaplasia. AIM: To examine CDX2 expression in the non-cancerous mucosa of patients with gastric cancer and compared it to CDX2 expression in controls with intestinal metaplasia. METHODS: Sixty patients who had undergone endoscopic mucosal resection for early gastric cancer and 60 gender- and age-matched controls were studied. Two specimens each were obtained from the greater and lesser curves of the corpus and from the greater curve of the antrum. Expression of CDX2 and sonic hedgehog were evaluated by immunostaining. RESULTS: Gastric cancer was associated with a higher frequency of incomplete intestinal metaplasia (OR = 8.3; 95%CI, 3.7-18.9, P < 0.001). CDX2 negatively correlated with sonic hedgehog expression, however, multivariate analysis revealed that CDX2 correlated with the intestinal metaplasia scores. Sonic hedgehog indices were lower and CDX2 staining in the corpus lesser curve was higher in the cancer group than in the controls. Sonic hedgehog indices in the corpus decreased and CDX2 indices in both areas increased in patients in the ascending order of those without intestinal metaplasia, those with complete intestinal metaplasia and those with incomplete intestinal metaplasia (P < 0.001). CONCLUSIONS: Loss of sonic hedgehog expression and aberrant expression of CDX2 correlates with the type of intestinal metaplasia and may play a role in carcinogenesis.

cAMP regulates gastrin gene expression.

Gastrin is one of the most potent regulators of acid secretion and gastrointestinal cell growth. A variety of signals regulate gastrin release from the antral G cell. However, whether these secretagogues also stimulate gastrin gene expression has not been established. Dramatic increases in gastrin gene expression occur in the stomach after birth and in response to chronic achlorhydria. Moreover, gastrin gene expression in malignant islet cell tumors (gastrinomas) appears to represent reactivation of the fetal pattern of expression in the pancreas. Thus differential expression of the gastrin gene is a reflection of differences in transcriptional control. Since various luminal and humoral factors stimulate the production of cAMP and gastrin secretion in both gastrinoma and antral G cells, we studied the effect of cAMP on gastrin gene expression. Using stable and transient transformants of a pituitary cell line containing the human gastrin gene, we found that cAMP stimulates a three-fold increase in gastrin mRNA levels and that the response maps to elements located between -148 and -40 base pairs upstream from the cap site. Collectively, these studies link an important regulator of gastrin secretion to regulation of gastrin gene expression.

3-Hydroxy-3-methylglutaryl coenzyme A reductase in isolated villous and crypt cells of the rat ileum.

The localization of 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMG-CoA reductase, E.C. 1.1.1.34) in the villous and crypt cells of the small intestine was accomplished after separating these cells from the mucosal layer by sequential dissociation in a "dual-buffer" system. Consistent separation was demonstrated by using the marker enzymes alkaline phosphatase, specific to the villous cell, and thymidine kinase, specific to the crypt cell. Cells obtained were 95-100% viable, and no relative difference in lability was observed, as evidenced by the equal distribution of acid phosphatase. This method of cell separation was an improvement over the "scraping" technique which damaged cells severely and produced villous preparations that contained little or no reductase activity. The HMG-CoA reductase specific activity in whole cell homogenates of the ileal villi was 0.47 and of the crypts was 0.27 nmol/min per mg of protein, considerably higher values than have been reported earlier. Also in comparison to the crypts, the villi incorporated 1.5-fold more [(14)C]-acetate into sterols, a ratio similar to that describing the distribution of HMG-CoA reductase in the two cell populations. These results unequivocally establish that the villi have higher HMG-CoA reductase activity than the crypts and confirm an earlier report from this laboratory that the villi are a major site of sterol synthesis. The sterol bio-synthetic capacity of the small intestine was highest in the ileum and decreased towards the jejunum. The HMG-CoA reductase specific activity of the ileum averaged 0.30 and that of the jejunum 0.10 nmol/min per mg of protein; however, the cholesterol content of the ileum was slightly lower than the jejunum. These results are discussed to suggest the possibility that the sterol content of the ileum may largely be due to in situ synthesis.

Transcription factor ZBP-89 cooperates with histone acetyltransferase p300 during butyrate activation of p21waf1 transcription in human cells.

Inducible p53-independent regulation of the cyclin-dependent kinase inhibitor p21(waf1) transcription is mediated through proximal GC-rich sites. Prior studies have shown that Sp1, Sp3, and the histone acetylase co-activator p300 are components of the complexes binding to these sites. Although Sp1 and Sp3 collaborate with p300, a direct interaction between Sp1 and p300 does not occur. This study sought to determine whether ZBP-89 rather than Sp1 is the direct target of p300 during butyrate induction of p21(waf1). ZBP-89 (BFCOL1, BERF-1, ZNF 148) is a Krüppel-type zinc finger transcription factor that binds to GC-rich elements and represses or activates known target genes. Adenoviral-mediated expression of ZBP-89 in HT-29 cells revealed that ZBP-89 potentiates butyrate induction of endogenous p21(waf1) gene expression. Further, cotransfection of a ZBP-89 expression vector with a 2.3-kilobase p21(waf1) reporter recapitulated the potentiation by butyrate. DNase I footprinting analysis of the human p21(waf1) promoter with recombinant ZBP-89 identified a binding site at -245 to -215. Electrophoretic mobility shift assays confirmed that both recombinant and endogenous ZBP-89 and Sp1 bind to this element. The potentiation was abolished in the presence of adenoviral protein E1A. Deletion of the N-terminal domain of ZBP-89 abolished the potentiation mediated by butyrate treatment. This same deletion mutant abolished the ZBP-89 interaction with p300. Cotransfection of p300 with ZBP-89 stimulated the p21(waf1) promoter in the absence of butyrate. p300 co-precipitated with ZBP-89 but not with Sp1, whereas ZBP-89 co-precipitated with Sp1. Together, these findings demonstrate that ZBP-89 also plays a critical role in butyrate activation of the p21(waf1) promoter and reveals preferential cooperation of this four-zinc finger transcription factor with p300.

Correlation of Na+,K+-ATPase content and plasma membrane surface area in adapted and de-adapted salt glands of ducklings.

During salt-water adaptation, an increase occurs in Na+,K+-ATPase content and surface area of the basolateral plasma membrane of the principal cell of the duck salt gland. To determine the degree to which these changes are correlated, accepted morphometric methods were used to determine numerical cell densities and plasma membrane surface densities of peripheral and principal cells. After adaptation, the plasma membrane surface area per principal cell was five times greater than in controls. Following de-adaptation, the plasma membrane content in principal cells returned to 1.9 times control levels. Two other cell constituents, mitochondria and lipid droplets, displayed similar quantitative changes. Na+,K+-ATPase content increased about fourfold with adaptation and decreased to near control levels with de-adaptation. Thus, changes in Na+,K+-ATPase content and basolateral plasma membrane surface area in adapting and de-adapting secretory epithelia of the salt gland occur nearly in parallel. These quantitative data enable Na+,K+-ATPase synthesis and degradation to be investigated in relation to membrane biogenesis.

Fos is required for EGF stimulation of the gastrin promoter.

Gastrin gene expression is regulated by developmental cues, pH, and inflammation. These processes are mediated by various extracellular ligands, e.g., growth factors, cytokines, and neuropeptides that also stimulate c-fos gene expression. Therefore, to determine whether Fos is required for stimulation of the gastrin promoter, a c-fos sense expression vector was coexpressed with a gastrin reporter construct in a GH4 rat pituitary cell line. We found that epidermal growth factor (EGF) and tumor necrosis factor-alpha (TNF-alpha) transiently stimulate an increase in Fos protein that precedes stimulation of the gastrin promoter. However, the induction mediated by TNF-alpha was weaker than that mediated by EGF, indicating minimal overlap of the signaling pathways activated by EGF and TNF-alpha. Accordingly, overexpression of c-fos mRNA facilitated primarily EGF rather than TNF-alpha induction of the gastrin promoter. Expression of the c-fos gene in the absence of ligand did not stimulate the gastrin promoter. Thus c-fos gene expression is required but is not sufficient for induction of the gastrin promoter by EGF.

EGF stimulates gastrin promoter through activation of Sp1 kinase activity.

Epidermal growth factor (EGF) receptor activation stimulates gastrin gene expression through a GC-rich element called gastrin EGF response element (gERE). This element is bound by Sp1 family members and is a target of the ras-extracellular signal-regulated kinase (Erk) signal transduction cascade. This raised the possibility that Sp1 may be phosphorylated by kinases of this signaling pathway. Erk is capable of phosphorylating other mitogen-inducible transcription factors, e.g., Elk and Sap, suggesting that Erk may also mediate EGF-dependent phosphorylation of Sp1. This possibility was tested by studying Sp1-dependent kinase activity in extracts prepared from EGF-activated AGS cells by use of solid-phase kinase assays and immunoprecipitation of metabolically labeled Sp1. The results revealed that Sp1 kinase activity (like gastrin promoter activation) is inhibited by PD-98059 and, therefore, is dependent on mitogen-activated protein kinase kinase 1 (Mek 1). However, EGF-dependent activation of endogenous Erk did not account for most of the Sp1 kinase activity, since Erk and additional Sp1 kinase activity analyzed in a solid-phase kinase assay eluted from an ion-exchange column in different fractions. Phosphoamino acid analysis of in vivo radiolabeled Sp1 demonstrated that the kinase phosphorylates Sp1 on Ser and Thr in response to EGF. Therefore, most EGF-stimulated Sp1 kinase activity is Mek 1 dependent and distinct from Erk.

Gastritis and hypergastrinemia due to Acinetobacter lwoffii in mice.

In mouse models and humans, Helicobacter pylori is associated with an increase in serum gastrin and gastrin-expressing (G) cells with a concomitant decrease in somatostatin-expressing D cells. Inflammation of the gastric mucosa can progress to metaplastic changes in the stomach and to decreased colonization by H. pylori and increased colonization by non-H. pylori organisms. In addition, about 20% of individuals with chronic gastritis are H. pylori negative, suggesting that other organisms may induce gastritis. Consistent with this hypothesis, we report here that Acinetobacter lwoffii causes the same histologic changes as does H. pylori. Gastric epithelial cells were isolated from the entire stomach by an enzymatic method for quantitation by both flow cytometry and morphometric analysis. Two months after mice were inoculated with H. pylori or A. lwoffii, the mucosal T- and B-cell numbers significantly increased. After 4 months of infection, there was a threefold increase in the number of G cells and a doubling in the number of parietal cells. A threefold decrease in the number of D cells occurred in H. pylori- and A. lwoffii-infected mice. Plasma gastrin levels increased after both H. pylori and A. lwoffii infection. Histology revealed the presence of inflammation in the gastric mucosa with both A. lwoffii and H. pylori infection. A periodic acid-Schiff stain-alcian blue stain revealed mucous gland metaplasia of the corpus. Collectively, the results demonstrate that gastritis and hypergastrinemia are not specific for H. pylori but can be induced by other gram-negative bacteria capable of infecting the mouse stomach.

Sp1 affinity for GC-rich elements correlates with ornithine decarboxylase promoter activity.

The highly conserved ornithine decarboxylase (E.C.4.1.1.17) promoter contains multiple binding sites for Sp1 within the first 400 bp of the cap site. Therefore the ability of individual Sp1 elements to confer transactivation alone or in combination was tested. We show that different Sp1 sites vary in their affinity for Sp1 and that increasing affinity correlates with enhancer activity. In addition, several adjacent Sp1 sites synergistically enhanced promoter activity. Thus, the strength of promoter transactivation correlated with both the number of GC-rich elements and their affinity for Sp1 protein.