The literature on the prognostic relevance of signet-ring cell (SRC) histology in gastric cancer (GC) is controversial which is most likely related to inconsistent SRC classification based on haematoxylin-eosin staining. We hypothesised that mucin stains can consistently identify SRC-GC and predict GC patient outcome.

We performed a comprehensive literature review on mucin stains in SRC-GC and characterised the mucin expression in 851 Caucasian GC and 410 Asian GC using Alcian Blue (AB)-Periodic Acid-Schiff (PAS), MUC2 (intestinal-type mucin), and MUC5AC (gastric-type mucin). The relationship between mucin expression and histological phenotype [poorly cohesive (PC) including proportion of SRCs, non-poorly cohesive (non-PC), or mucinous (MC)], clinicopathological variables, and patient outcome was analysed.

Depending on mucin expression and cut-offs, the positivity rates of SRC-GC reported in the literature varied from 6 to 100%. Patients with MUC2 positive SRC-GC or SRC-GC with (gastro)intestinal phenotype had poorest outcome. In our cohort study, PC with ≥ 10% SRCs expressed more frequently MUC2, MUC5AC, and ABPAS (p < 0.001, p = 0.004 and p < 0.001, respectively). Caucasians with AB positive GC or combined ABPAS-MUC2 positive and MUC5AC negative had poorest outcome (all p = 0.002). This association was not seen in Asian patients.

This is the first study to suggest that mucin stains do not help to differentiate between SRC-GC and non-SRC-GC. However, mucin stains appear to be able to identify GC patients with different outcome. To our surprise, the relationship between outcome and mucin expression seems to differ between Caucasian and Asian GC patients which warrants further investigations.

Leucine-rich repeat-containing G-protein-coupled receptor 5 (Lgr5) is a putative intestinal stem cell marker that is also expressed in various tumors. To analyze its pathological characteristics in mucosal gastric signet-ring cell carcinoma (SRCC), we investigated Lgr5 expression in 35 intramucosal gastric SRCC patients using RNAscope, a newly developed RNA in situ hybridization technique. Lgr5 expression in individual tumor cells was scored semi-quantitatively from 0 to 400. Ki67 was also examined by immunohistochemistry, with a linear arrangement of Ki67-expressing cells present in 20 of 35 cases. This area of Ki67-expressing cells was topographically divided into upper, middle, and lower regions. All cases with linear Ki67 expression patterns also had Lgr5-positive cells arranged in a linear fashion in the lower area-which was distinct from the area of high Ki67 expression. The rate of Ki67 positivity in Lgr5-positive cells was significantly lower than that of Lgr5-negative cells in areas of high Ki67 expression. In intramucosal SRCC, the low mitotic activity of Lgr5-positive cells suggests that they may represent cancer stem cells as seen in other types of stomach carcinomas. Intramucosal SRCC may therefore contain stem cells expressing Lgr5 in the lower area of the lamina propria, akin to normal gastric pyloric mucosa.

Adipophilin, a lipid droplet-associated protein that regulates lipid droplet structure and formation, is expressed in a wide variety of tumors. The aims of this study were to evaluate the frequency and distribution pattern of adipophilin expression in gastric epithelial neoplasia and to correlate these variables with clinicopathological features and the mucin phenotype. We retrospectively examined 159 cases of gastric epithelial neoplasia, which were classified according to the Vienna classification system as 52 noninvasive low-grade adenoma (category 3), 65 noninvasive high-grade neoplasia (category 4), and 42 invasive neoplasia (category 5). Immunohistochemistry for adipophilin was performed, and phenotypic marker expression was determined by immunohistostaining with MUC2, MUC5AC, CD10, MUC6, and villin. Adipophilin was expressed in 41 of the 52 (79 %) category 3 cases, in 42 of the 65 (65 %) category 4 cases, and in 23 of the 42 (55 %) category 5 cases. Expression of adipophilin was only present in lesions with complete or incomplete intestinal phenotypes. Adipophilin was expressed more frequently in the surface epithelium in category 3, whereas there was a stepwise increase in cryptal staining of adipophilin from category 3 to category 5. In conclusion, adipophilin expression is closely related to the intestinal differentiation of the tumor. The pattern of immunostaining for adipophilin might be a useful new marker for discriminating adenomas from adenocarcinomas.

This study was performed to determine the association between RUNX3 expression and Helicobacter pylori infection in premalignant gastric lesions.

We examined 107 patients with gastric epithelial dysplasia who had undergone endoscopic mucosal resection or submucosal dissection. All tissue samples were evaluated by RUNX3 staining and subclassified by immunophenotype. H. pylori infection in dysplastic lesions and the normal surrounding tissue was examined by silver staining, and cagA status was assessed by polymerase chain reaction.

The loss of RUNX3 expression was observed in 62 cases (57.9%), and an association with H. pylori infection was found in 54 cases (50.5%). The infection rate with the cagA-positive H. pylori strain was 63.0%. In RUNX3-negative lesions, the rate of H. pylori infection (p=0.03) and the frequency of category 4 lesions (according to the revised Vienna classification) were high (p=0.02). In addition, the gastric mucin phenotype was predominant. In RUNX3-negative category 4 lesions, the rate of cagA-positive H. pylori infection rate was high but not significantly increased (p=0.08).

Infection with H. pylori is associated with inactivation of RUNX3 in early gastric carcinogenesis. This mechanism was prominent in gastric cancer with a gastric mucin phenotype.

Undifferentiated-type early gastric adenocarcinomas are generally classified into two groups: pure undifferentiated-type adenocarcinomas, which naturally develop as undifferentiated-type without a glandular component; and mixed differentiated/undifferentiated-type adenocarcinomas, which are associated with some vestigial glandular component and presumably develop from differentiated-type adenocarcinoma. The differences in phenotypic expression between these two groups were examined using mucin core protein and CDX2.

A total of 210 lesions of undifferentiated-type early gastric adenocarcinoma less than 25 mm in diameter were classified into four categories (gastric type, gastrointestinal type, intestinal type, and null type) based on their MUC5AC, MUC6, MUC2, and CDX2 immunoprofiles.

Gastric type was significantly (p < 0.01) decreased and gastrointestinal type was significantly (p < 0.01) increased both in pure undifferentiated-type adenocarcinomas and in mixed differentiated/undifferentiated-type adenocarcinomas when CDX2 was applied to mucin core protein. In the pure undifferentiated-type adenocarcinomas, gastric type decreased and gastrointestinal type increased as tumor size increased (p < 0.05). In contrast, in the mixed differentiated/undifferentiated-type adenocarcinomas, gastrointestinal type was most common even in small-sized (≤10 mm) carcinomas and was generally stable regardless of tumor size. In submucosal carcinomas, gastrointestinal type decreased and gastric type and intestinal type increased during carcinoma invasion from the intramucosal to submucosal parts (p < 0.05). The positivity rates for all phenotypic markers, especially gastric markers, tended to decrease during submucosal invasion.

CDX2 is a sensitive marker for assessing intestinal phenotypic expression, and it is likely that there are two different pathways of tumor progression in undifferentiated-type adenocarcinoma of the stomach, according to phenotypic expression.

A close association between patterns identified by magnifying narrow-band imaging (M-NBI) and histological type has been described. M-NBI patterns were also recently reported to be related to the mucin phenotype; however, detials remain unclear.

We investigated the cellular differentiation of gastric cancer lesions, along with their mucosal distribution observed by M-NBI. Ninety-seven depressed-type early gastric cancer lesions (74 differentiated and 23 undifferentiated adenocarcinomas) were visualized by M-NBI. Findings were divided into 4 patterns based on abnormal microvascular architecture: a chain loop pattern (CLP), a fine network pattern (FNP), a corkscrew pattern (CSP), and an unclassified pattern. Mucin phenotypes were judged as gastric (G-type), intestinal (I-type), mixed gastric and intestinal (M-type), and null (N-type) based on 4 markers (MAC5AC, MUC6, MUC2, and CD10). The relationship of each pattern of microvascular architecture with organoid differentiation indicated by cancer cell differentiation and its distribution in each histological type of early gastric cancer was investigated.

All CLP and FNP lesions were differentiated. The cancer cell distribution showed organoid differentiation in 84.2% (16/19) and 61.1% (22/36) of the two types of lesions, respectively, and there was a significant difference from the unclassified pattern with organoid differentiation (p<0.001). Almost all (94.7%; 18/19) CSP lesions were undifferentiated, and organoid differentiation was observed in 72.2% (13/18). There was a significant difference from the unclassified pattern with organoid differentiation (p<0.05).

Cellular differentiation and distribution are associated with microvascular architecture observed by M-NBI.

Helicobacter pylori infection is an important factor for gastric carcinogenesis in human. In carcinogen-treated Mongolian gerbils, H. pylori infection enhances stomach carcinogenesis, while infection alone induced severe hyperplasia called heterotopic proliferative glands. A high-salt diet or early acquisition of the bacteria exacerbates inflammation and carcinogenesis. Oxygen radical scavengers or anti-inflammatory chemicals as well as eradication of H. pylori are effective to prevent carcinogenesis. H. pylori-associated inflammation induces intestinal metaplasia and intestinalization of stomach cancers independently. It is necessary to control cancer development not only in H. pylori-positive cases but also in H. pylori-negative metaplastic gastritis.

The results of a randomized controlled study and meta-analysis study have recently proved that Helicobacter pylori eradication has a preventive effect against the development of metachronous and primary gastric cancer. However, gastric cancer is sometimes detected after successful eradication. There is a lack of study about gastric cancers in eradicated patients. To clarify the characteristics of gastric cancers detected after H. pylori eradication, we analyzed the clinicopathological features of these cancers.

The subjects were 18 early-stage gastric cancer specimens resected from 17 patients who had received successful eradication of H. pylori from February 1995 to March 2009. The control group consisted of 36 specimens from noneradicated patients with persistent H. pylori infection who were matched with the subjects in age, sex, and depth of invasion. Clinicopathological features and mucin phenotypes of gastric cancer were clinically and immunohistologically evaluated.

The average diameter of gastric cancer was smaller and Ki-67 index was lower in the eradication group. The morphological distribution of depression types was significantly lower in the control group. Immunohistochemical phenotyping revealed that 72.2% of the lesions in the eradicated group were complete gastric type or gastric predominant mixed type, whereas the percentages of gastric type and intestinal type in the control group were similar.

Our findings indicate that the clinicopathological characteristics of gastric cancers detected after H. pylori eradication are different from those of gastric cancers in patients with persistent H. pylori infection. H. pylori eradication may suppress intestinalization during the development of gastric cancer.

To investigate a relationship between the clinicopathological features and mucin phenotypes in advanced gastric adenocarcinoma (AGA).

Immunohistochemical staining was performed to determine the mucin phenotypes in 38 patients with differentiated adenocarcinomas (DACs), 9 with signet-ring cell carcinomas (SIGs), and 48 with other diffuse-type adenocarcinomas (non-SIGs) of AGA. The mucin phenotypes were classified into 4 types: gastric (G), gastrointestinal (GI), intestinal, and unclassified.

The G-related mucin phenotypes were highly expressed in all the histological subtypes of AGA. The expression of the GI phenotype in SIG patients was lower than that in DAC patients (P = 0.02), and this phenotype was observed in 56% of the non-SIG patients in the intramucosal layer. Among non-SIG cases, the expression of the GI phenotype was significantly higher in patients with extended adenocarcinomas and those with positive rates of lymph node metastasis. There was no difference between the expressions of the G and other GI phenotypes factors. Among DAC and non-SIG patients, there were no differences between the survival rates of the corresponding patient groups.

The GI phenotype might possess more invasive characteristics than the G phenotype in non-SIG. Neither of the phenotypes indicated a poor prognosis of DAC and non-SIG.

Resistin-like molecule beta (RELMbeta), an intestinal goblet cell-specific protein, is a biomarker of intestinal metaplasia in Barrett's esophagus and over-expressed in colon cancer. Since gastric adenocarcinomas can arise through a process of intestinalization, we hypothesized that RELMbeta might be aberrantly expressed in gastric cancer. This study was undertaken to examine the RELMbeta expression in gastric cancer and correlate it with clinical outcome. One hundred and thirty-six gastric cancer patients were evaluated for the RELMbeta expression by immunohistochemistry. The RELMbeta transcripts were measured by real-time quantitative PCR. In normal gastric mucosa, RELMbeta expression was absent, whereas areas of intestinal metaplasia revealed RELMbeta reactivity. Eighty-nine patients of gastric cancer (65.4%) were positive for RELMbeta expression. In a subtotal of 20 patients, RELMbeta transcripts were positively correlated with protein levels in gastric cancer tissues, but absent in normal gastric mucosa. The expression rate of RELMbeta was higher in intestinal-type carcinomas than in diffuse-type carcinomas (P < 0.001). RELMbeta positivity in gastric cancer was positively correlated with tumor differentiation (P = 0.001) and inversely correlated with tumor infiltration (P = 0.007), lymph node metastasis (P = 0.035), and heparanase expression (P < 0.001), without correlation with age, gender, tumor location and size, tumor-node metastasis stages, and Ki-67 expression. Patients showing positive RELMbeta expression had a significantly longer overall survival than those with negative expression (P = 0.001). These results provide evidences that the RELMbeta expression in gastric cancer is correlated with clinicopathological features and may be a useful prognostic factor for predicting the outcome of gastric cancer patients.

Hepatocyte nuclear factor 4alpha (HNF4alpha) isoforms in the human stomach have not been fully investigated. The purpose of the present study was to evaluate the expression of P1 and P2 promoter-driven HNF4alpha (P1 and P2-HNF4alpha) in differentiated-type early gastric carcinomas (DEGC). P1- and P2-HNF4alpha expression was examined immunohistochemically both in non-neoplastic mucosa and carcinoma from surgical specimens. In all samples of non-neoplastic mucosa, foveolar, cardiac, fundic and pyloric gland epithelium was negative for P1-HNF4alpha, but was positive for P2-HNF4alpha. Intestinal metaplasia was positive for P1 and P2-HNF4alpha in all cases. Gastric carcinomas were classified into four mucin phenotypes based on the pattern of mucin expression: gastric, intestinal, mixed and null type. DEGC showed striking differences in the staining pattern for P1-HNF4alpha according to the mucin phenotype. Gastric carcinomas of intestinal, mixed and null type showed high positivity for P1-HNF4alpha, but the gastric type was negative for P1-HNF4alpha in all but one tumor. In contrast, P2-HNF4alpha was expressed in all tumors regardless of the mucin phenotype. Negative expression of P1-HNF4alpha was indicated as one of the useful immunohistochemical markers in the classification of mucin phenotype of both non-neoplastic mucosa and cancers of gastric phenotype.

We histopathologically re-evaluated the histogenesis of gastric carcinomas from comparative studies between Helicobacter pylori-positive and H. pylori-negative cases using the gastritis score from the Updated Sydney System. The incidence of H. pylori-negative gastric carcinomas was 3.11% (12/386); they are likely to develop in the fundic gland mucosa, and show a gastric phenotype by mucin immunohistochemistry. Even in cases of completely gastric and predominantly gastric phenotypes, CDX2 protein was expressed in most cases (90.9% of pT1 and 100% of pT2-3), indicating a possibility that intestinalization of carcinoma cells occurs independently of the background mucosa. Regarding the degree of gastritis of background mucosa surrounding 143 H. pylori-positive differentiated-type adenocarcinomas, the mean score ranged from 1.497 to 1.713. Our data support the hypothesis that intestinal metaplasia is not a precancerous but a paracancerous lesion, and most gastric adenocarcinomas develop in mildly to moderately atrophic mucosa with H. pylori-infection, i.e., ongoing atrophy.

Gene expression in tumours is regulated by environmental as well as genetic/epigenetic factors. This study assessed the environmental factors in intestinal expression of gastric cancers.

We immunohistochemically examined intratumoural heterogeneity in the expression of Cdx2, MUC2, MUC5AC and MUC6 in 39 intramucosal and 49 extramucosally invasive undifferentiated-type gastric carcinomas (UGCs), consisting of signet ring cell carcinomas showing a layered structure (LS) in the mucosa and dedifferentiated tubular adenocarcinomas without LS and with minor tubular components (TC). The LS retains mucosal vertical polarity with superficial MUC5AC expression. Loss of this polarity was independent of intestinal expression and associated with extramucosal invasion. In LS(+) UGCs, intestinal expression was enhanced as the size of mucosal spread increased and was significantly reduced with deeper extramucosal invasion, whereas, in LS(-)/TC(+) UGCs, intestinal expression was frequent and predominant in the mucosa and was insignificantly reduced with deeper extramucosal invasion.

In LS(+) UGCs, intestinal expression showed dynamic alteration probably by environmental induction and progression-related loss of induction, whereas it was relatively stable in LS(-)/TC(+) UGCs. Thus, intestinal expression in UGCs is not useful as a marker of tumour progression because it is also affected by environmental factors and genetic lineage.

Gastric and intestinal mucin phenotype cell markers are widely expressed in gastric carcinoma cells, irrespective of their tumor histological type. In the present study, we tried to reveal the clinicopathological significance of mucin phenotype in human gastric carcinomas. Moreover, we investigated the clinical significance of RUNX3 in association with mucin phenotype.

The mucin expression of MUC5AC, MUC6, MUC2, and CD10 was evaluated in 97 gastric carcinomas by immunohistochemistry. Tumors were classified into gastric (G), gastric and intestinal mixed (GI), intestinal (I), and null (N) phenotype according to combination of mucin expression.

The rate of G, GI, I, and N phenotype was 40.0%, 38.1%, 10.3%, and 19.6%, respectively. Mucin phenotype was also significantly correlated with several clinicopathological findings. Patients with I phenotype had a significantly poorer prognosis than those with any other phenotypes. They also had a higher rate of postoperative liver metastasis. Multivariate analysis revealed that mucin phenotype was a significant independent prognostic factor. We suggested that Loss of RUNX3 expression might correlate with intestinal phenotype and postoperative outcome.

Mucin phenotype has a significant prognostic value and may be a useful marker for the treatment of human gastric carcinoma.

Although incidences of stomach cancer have decreased over the past several decades, the disease remains an important public health problem. To identify pathological and molecular biochemical mechanisms, various experimental animal models have been established in rats and mice with chemical carcinogens including N-methyl-N'-nitro-N-nitrosoguanidine (MNNG) and N-methyl-N-nitrosourea (MNU). Helicobacter pylori(H. pylori) is one of the most important factors for human stomach disorders, including neoplasia, and the H. pylori-infected and carcinogen-treated Mongolian gerbil (MG) has proven very useful for analyses of underlying processes. The findings with this model support the hypothesis that intestinal metaplasia is important not as a precancerous lesion but rather as a paracancerous condition and that intestinalization of stomach cancer progresses with chronic inflammation. Furthermore, dose-dependent enhancing effects of salt on stomach carcinogenesis could be demonstrated in MGs treated with MNU and H. pylori modifying surface mucous gel layer. H. pylori itself only causes chronic inflammation and acts as a promoter of stomach carcinogenesis in experimental models. Based on the precise pathological diagnosis of stomach lesions such as noncancerous heterotopic proliferative glands (HPG) and adenocarcinomas, a basis for understanding mechanisms of carcinogenesis has been established on which chemoprevention can be modeled.

Abnormal localization of beta-catenin is frequently observed in human gastric cancers. The aim of the present study was to evaluate relationships among gastrointestinal differentiation phenotypes, beta-catenin localization and mutations of Wnt signalling genes.

Seventy-seven regions in 39 gastric adenocarcinomas were classified according to beta-catenin localization and gastric and intestinal phenotypes. Cases with membranous beta-catenin localization showed a gradual decrease from gastric (G) (55% = 6/11) and gastric-and-intestinal-mixed (GI) (17% = 5/29) to intestinal (I) (0% = 0/21) phenotypes, while those with nuclear localization showed a concomitant increase: 18% (2/11), 41% (12/29), 95% (20/21) and 63% (10/16) for G, GI, I and null type (N), respectively (P < 0.001, membranous versus nuclear localization in G, GI through I). Mutations in exon 3 of the beta-catenin gene were found in G (50% = 1/2), GI (67% = 8/12), I (45% = 9/20) and N (0% = 0/10) regions with nuclear beta-catenin localization (GI versus N, P < 0.01; I versus N, P < 0.05). Adenomatous polyposis coli (APC) gene mutations were demonstrated only in GI, I and N types, irrespective of beta-catenin localization. Molecular analysis of these genes revealed 10 tumours to be heterogeneous out of 16 informative cases (62.5%).

Intestinal phenotypic expression is accompanied by a shift from membranous to cytoplasmic/nuclear accumulation of beta-catenin. In contrast, N-type regions may progress along a different pathway.

Gastric and intestinal phenotypic cell markers are expressed in gastric carcinomas, irrespective of their histologic type. In the present study, we determined the clinicopathologic significance of phenotypic marker expression in early-stage gastric differentiated-type tumors and the association between marker expression and genetic alterations.

Phenotypic marker expression was determined by examining the expressions of human gastric mucin (HGM), MUC6, MUC2, and CD10 in 63 gastric adenomas, 133 early differentiated-type carcinomas, and 24 follow-up cases with gastric adenoma. Tumors were classified into gastric, gastric and intestinal mixed, or intestinal phenotypes according to the immunopositivity of the above markers. The presence of mutations in APC, K-ras, and p53 and the microsatellite instability status were also determined in all tumors.

The expressions of HGM and MUC6, representing gastric or gastric and intestinal mixed phenotypes, were significantly associated with high-grade atypia in the 63 gastric adenomas. Among the 133 early differentiated-type carcinomas, HGM expression was significantly associated with mixed-type (with an undifferentiated-type component) tumors and lymph node metastasis. MUC2 expression was inversely associated with submucosal invasion. A multivariate analysis revealed that gastric adenomas were significantly associated with the intestinal phenotype and were inversely associated with p53 mutation compared with early differentiated-type carcinomas. Among all 196 tumors, APC mutation was significantly associated with CD10 expression and the intestinal phenotype and was inversely associated with the expressions of HGM and MUC6. The microsatellite instability status was significantly associated with MUC6 expression. Malignant transformation from gastric adenoma to carcinoma was shown in 5 of the 24 follow-up cases of gastric adenoma. The malignant transformation was significantly associated with the gastric and intestinal mixed phenotype and was inversely associated with APC mutation. No malignant transformation was found in intestinal phenotype gastric adenomas with APC mutation.

Our present findings show that phenotypic marker expression is associated with tumor aggressiveness during the early stage of gastric differentiated-type tumors. Differences in the biological behavior of tumors with different phenotypes may result from differences in the genetic backgrounds during the incipient phase of gastric tumorigenesis.

CDX2 is an intestinal transcription factor that might be involved in the regulation of proliferation and differentiation of intestinal epithelial cells. It is well known that both gastric and intestinal phenotypic cell markers are expressed in gastric cancers. The aims of this study were to analyze the CDX2 expression and its relationship with the patients' clinicopathological characteristics and the mucin phenotypes by performing immunohistochemistry.

A total of 259 gastric cancer cases (122 early and 137 advanced cancers) were evaluated histologically and phenotypically. CDX2 expression was assessed by immunohistochemistry.

Increased CDX2 expression correlated with a higher proportion of intestinal-type cancers of Lauren and early gastric cancers (p<0.001 and p<0.001, respectively) and a lower proportion of perineural invasion and lymph node metastasis (p<0.001 and p=0.003, respectively). Increased expressions of intestinal mucin (MUC-2, CD10) and decreased gastric mucin (MUC5AC) were associated with an increased CDX2 expression (p<0.001, p=0.045 and p=0.004, respectively). MUC6 expression was not associated with CDX2 expression. There was a significantly increased CDX2 expression in the intestinal phenotype compared with the other phenotypes (p<0.001).

These results suggest that CDX2 might be a useful marker in predicting the clinical outcome for patients with gastric cancers.

Gastric marker mucins (MUC5AC and MUC6) and intestinal marker molecules (MUC2 and CD10) have been used to determine the cell lineage of epithelial cell of gastric carcinoma (GC).

To clarify the characteristics of Epstein-Barr virus (EBV)-associated GC, 18 cases were immunohistochemically evaluated along with 56 cases of EBV-negative GC.

MUC2 expression was lower in EBV-associated GC: immunostaining grades 0, 1, 2, 3, and 4 were observed in 10, 6, 1, 1, and 0 cases of EBV-associated GC, respectively, and in 18, 11, 15, 6, and 6 cases of EBV-negative GC, respectively (P = 0.013). CD10 positivity (grades 2-4) in EBV-associated GC was 6%, significantly lower than in EBV-negative GC (34%) (P = 0.030). When phenotypes of GC were categorized by the combined positivities of gastric markers (either MUC5AC or MUC6) and intestinal markers (either MUC2 or CD10), EBV-associated GC included primarily null (44%) and gastric (39%) types, but EBV-negative GC comprised null (7%), gastric (30%), intestinal (27%), and mixed (36%) types. The age of patients with gastric types was significantly younger for both EBV-associated GC and EBV-negative GC cases.

Neoplastic epithelial cells of EBV-associated GC did not express MUC2 or CD10, and most of them were categorized as null or gastric types. EBV infection may occur in the epithelial cells of null or gastric phenotypes, which may be devoid of transdifferentiation potential toward intestinal phenotypes.

It has been proved that some differentiated-type gastric carcinomas have a gastric phenotype. Similarly, it can be conjectured that some undifferentiated-type gastric carcinomas have an intestinal phenotype and that there are biological differences between undifferentiated-type gastric carcinomas with a gastric phenotype and those with an intestinal phenotype. We classified the phenotypes of early undifferentiated-type gastric carcinomas and investigated the relationship between their biological behavior and the phenotypes.

Sixty lesions of intramucosal undifferentiated-type gastric carcinoma were classified into four phenotypes; gastric type, incomplete-intestinal type, complete-intestinal type, and unclassified type, according to the expression of CD10, MUC2, small-intestinal mucinous antigen (SIMA), human gastric mucin (HGM), or concanavalin A (ConA).

The incidence of gastric-type carcinoma, incomplete-intestinal-type carcinoma, and complete-intestinal-type carcinoma was 33% (20 cases), 65% (39 cases), and 2% (1 case), respectively. There was no significant difference in any of the clinicopathological factors examined between the 20 gastric-type carcinomas and the 40 intestinal-type carcinomas, but there were significant differences in the morphological findings. Intestinal-type carcinomas had a glandular structure more frequently than the gastric-type carcinomas. The spreading pattern of gastric-type carcinomas showed a middle-layer type more frequently than the intestinal-type carcinomas.

Undifferentiated-type gastric carcinomas frequently expressed an intestinal phenotype. There were differences in the growth patterns between undifferentiated-type gastric carcinomas with a gastric phenotype and those with the intestinal phenotype.

The Helicobacter pylori-infected Mongolian gerbil (MG) has been established as an appropriate animal model for studies of stomach cancer development. However, there have hitherto been no data on the phenotypic classification of glandular stomach cancers in H. pylori-infected and non-infected MG. We therefore examined the phenotypes of 50 and six advanced glandular stomach cancers in H. pylori-infected and non-infected MG, respectively, as well as adjacent non-neoplastic mucosa, using several gastrointestinal epithelial phenotypic markers. The lesions were divided phenotypically into 21 gastric, 24 gastric-and-intestinal mixed, four intestinal and one null types, with 90.0% of the lesions harboring gastric elements and 56.0% demonstrating intestinal phenotypic expression in H. pylori-infected MG. All six lesions were classified as gastric type in non-infected MG. There was no clear correlation with the presence of intestinal metaplasia in surrounding mucosa. In conclusion, our data suggest that most advanced adenocarcinomas retain a gastric cellular phenotype in the glandular MG stomach. Thus, it might be proposed that intestinal metaplasia is a paracancerous phenomenon rather than a premalignant condition. H. pylori infection may trigger intestinalization of both stomach cancers and non-neoplastic mucosa.

It has been suggested that the prevalence of microsatellite instability (MSI) is high in intramucosal differentiated gastric cancers with gastric foveolar phenotypic expression, and that these tumors are prone to lose their glandular structures and progress to undifferentiated-type lesions. To test this hypothesis, we examined the relationships among human MutL homologue 1 (hMLH1) expression (which is linked to MSI), the phenotype, and the histological type in patients with advanced and intramucosal gastric cancer.

We analyzed hMLH1 expression by immunohistochemistry in 70 advanced and 30 intramucosal gastric cancers with histological evaluation and assessment of the phenotype, and Cdx2 expression determined by immunohistochemistry. The MSI status was also examined in 20 cases.

Thirteen (18.6%) advanced and 5 (16.7%) intramucosal gastric cancers were judged to be hMLH1-negative. In the advanced cases, no association was observed between the histological type and the phenotype and loss of hMLH1. In the intramucosal cases, MUC5AC expression was observed in all 5 hMLH1-negative differentiated-type cancers. However, no hMLH1-negative lesions were detected in the intramucosal undifferentiated cancers (0/14; P < 0.05 vs differentiated types). In the advanced cases, MSI-positivity (MSI +) and loss of hMLH1 expression did correlate (P < 0.0001), while no association was observed between MSI +, histological type, and phenotype.

Our data support the hypothesis that, phenotypically, some MSI-positive differentiated gastric cancers of gastric foveolar phenotypic expression may easily change, from gastric to intestinal phenotypic expression, also changing, histologically, from differentiated to undifferentiated type with progression.

Adenocarcinomas are malignant tumors with glandular growth and/or supposed intracellular mucin as identified by periodic acid-Schiff (PAS) positivity. Gastric signet ring cell carcinomas are classified as diffuse type. A proportion of diffuse-type adenocarcinomas have previously been suggested to be of neuroendocrine origin. In the present study we examined gastric signet ring cell carcinomas for neuroendocrine differentiation. Of 11 gastric signet ring cell carcinomas, 8 contained areas with PAS-positive signet ring cells that also were immunoreactive for one or several neuroendocrine markers: synaptophysin, chromogranin A, and histidine decarboxylase, the latter an enterochromaffin-like (ECL) cell marker. Whereas PAS positivity was located in the central cytoplasm, neuroendocrine immunoreactivity was often located as a rim surrounding an otherwise non-immunoreactive cytoplasm, presumed to represent the area with PAS-positive material. These findings indicate that signet ring cell carcinomas could be of neuroendocrine origin. We propose that signet ring cell carcinomas develop by gradual dedifferentiation from ECL cells via signet ring cells with neuroendocrine immunoreactivity toward signet ring cells where the cytoplasm mainly consists of PAS-positive material. This finding could have implications for the classification and understanding of gastric carcinogenesis.

To evaluate the relationship between phenotypic expression and tumour progression as represented by macroscopic features, submucosal invasion and lymph node metastasis in early differentiated gastric adenocarcinoma.

One hundred and fifty-five cases of early gastric differentiated adenocarcinoma without any poorly differentiated components were studied. The mucosal and submucosal components of carcinomas and lymph node metastatic lesions were classified into four categories, gastric type (G-type), incomplete intestinal type (I-type), complete intestinal type (C-type) and unclassified type (U-type), based on the combination of the phenotypic expression of HGM (gastric foveolar epithelium), MUC 6 (gastric pyloric glands), MUC 2 (intestinal goblet cells) and CD 10 (small intestinal brush border). In addition, a new classification representing a phenotypic shift from mucosa to submucosa or from primary lesion to lymph node metastasis was established with the categories of preserved group (P-group), loss group (L-group) and acquired group (A-group).

(1) In submucosal carcinoma, U-type was more common in the submucosa (39%) than in the mucosa (9%). (2) U-type was more common in the metastatic lesions (42%) than in the primary lesions (5%). (3) The submucosal component and lymph node metastatic lesions were classified as P-group in 48% and 43% of cases, respectively, and as L-group in 50% and 52% of cases, respectively. (4) Lymph node metastatic lesions comprising undifferentiated carcinoma were classified as L-group in 100% of cases.

During the course of tumour progression, early differentiated adenocarcinoma at first tends to lose its phenotypic expression despite preserving its morphology, but subsequently morphological dedifferentiation occurs.

The clinicopathological significance of colonic and small-intestinal phenotypes has hitherto remained unclear in gastric cancers. The purpose of the present study was therefore to examine 86 gastric carcinomas histologically and phenotypically using several phenotypic markers, including colon-specific carbonic anhydrase 1 (CA1) and sucrase as small-intestine specific marker. Of 86 gastric cancers, sucrase and CA1 expression was observed in 12 (14.0%) and only in two cases (2.3%), respectively, associated with other intestinal markers such as villin and mucin core protein (MUC)2. In the sucrase cases, expression appeared independent of the stage. However, CA1 expression was observed only in two advanced cases. No association was observed between colonic and small-intestinal phenotypes, and lymph node metastasis and postoperative survival in the advanced gastric cancer cases with intestinal phenotypic expression. Cdx2 appeared to be linked to upregulation of both CA1 and sucrase. In conclusion, the data suggest that colonic phenotype occurs rarely in gastric carcinogenesis. Colonic and small-intestinal phenotypes appear with expression of several intestinal phenotypic markers under the control of Cdx2 and presumably other related transcription factors.

It is well known that both gastric and intestinal phenotypic cell markers are expressed in gastric cancers. This study was aimed at investigating the correlation between gastric and intestinal phenotypic marker expression patterns of tumors and the clinicopathologic characteristics of gastric carcinomas.

We evaluated phenotypic marker expression by immunohistochemical staining with monoclonal antibodies. All tumors were classified as gastric (G), gastric and intestinal mixed (GI), intestinal (I), or null (N) phenotype.

The tumors were phenotypically divided into G-phenotype tumors (33.2%), GI-phenotype tumors (25.7%), I-phenotype tumors (26.8%), and N-phenotype tumors (14.3%). N-phenotype tumors were associated with more corporeal location than GI- and I-phenotype tumors (o=0.009 and p=0.007, respectively), a larger size than I-phenotype tumors (0=0.007), a higher proportion of advanced gastric cancers than G-, GI-, and I-phenotype tumors (0=0.003, p<0.001, and p<0.001, respectively), more perineural invasion than G-, GI-, and I-phenotype tumors (p=0.076, p=0.003, and p=0.003, respectively), and more lymph node metastasis than GI-phenotype tumors (p=0.017). I-phenotype tumors were associated with a higher proportion of intestinal-type tumors than G-, GI-, and N-phenotype tumors (p<0.001, p=0.011, and p<0.001, respectively).

Our results indicate that the gastric and intestinal phenotypic marker expression pattern of tumors is prognostically useful for patients with gastric carcinoma.

Admixture of signet-ring cell carcinoma component (SIG) and glandular component (GC) is sometimes found in a gastric cancer. The aim of this study was to elucidate the biological behavior of the gastric cancer containing either SIG only (SIG-type tumor) or SIG admixed with GC (SIG-GC-type tumor). Forty-three cases of SIG-type tumor and 26 cases of SIG-GC-type tumor, less than 2 cm in largest diameter, were examined. Twenty-six cases (60%) of SIGs were less than 10 mm in size, whereas only 3 cases (12%) of SIG-GCs were less than 10 mm in size (p<0.0001). Submucosal invasion by cancer cells was found in 12 cases of SIG-GCs but was not found in any of the SIGs. Those SIG-GCs showed gastric phenotype. It is important to note SIG-GC-type tumor for histopathological diagnosis.

Other than ectopic expression of intestinal transcription factors, Cdx1 and Cdx2, the molecular mechanisms underlying gastric and intestinal phenotypes of human stomach adenocarcinomas have yet to be clarified in detail. We have reported that Sox2, an HMG-box gastric transcription factor, is expressed in normal gastric mucosa and down-regulated in intestinal metaplasia.

We analysed mRNA levels of Sox2 and other differentiation markers in 50 surgically resected stomach adenocarcinomas, immunohistochemically classified into gastric (G), gastric-and-intestinal (GI)-mixed, solely intestinal (I), and null (N) types. Sox2 was found to be transcribed in G and GI-mixed type adenocarcinomas in accordance with MUC5AC and MUC6 expression, while Cdx1 and Cdx2 were up-regulated in GI-mixed and I types along with the expression of MUC2 and villin. In the N type, both gastric and intestinal transcription factors were suppressed. Immunohistochemistry confirmed expression of Sox2 in MUC5AC+ lesions and Cdx2 localization together with MUC2. A stomach adenocarcinoma cell line, KATOIII, demonstrated both MUC5AC and Sox2, although MUC5AC mRNA was not detected in the Sox2+ AGS cell line.

Sox2 may play an important role in maintaining a gastric phenotype in stomach cancers as well as in normal tissue, in cooperation with other cofactor(s).

The aim of this study was to examine the clinical significance of mucin phenotypes of early undifferentiated gastric carcinoma, and to identify variables that might be used to select patients suitable for minimally invasive surgery.

A total of 129 patients with early undifferentiated gastric carcinoma were studied. The mucin phenotype was determined immunohistochemically using markers for M1, apomucin (MUC) 6 and MUC2. Tumours were classified into gastric (G), intestinal, gastrointestinal (GI) or unclassified type. Undifferentiated carcinomas were classified into signet-ring cell carcinoma (SIG) and non-SIG. The immunoreactivity of matrix metalloproteinase (MMP) 7 and beta-catenin was also investigated.

GI-type tumours more commonly expressed non-SIG than SIG histology. The GI phenotype was associated with a higher incidence of submucosal invasion, lymphatic invasion, MMP-7 expression and nuclear accumulation of beta-catenin than the G type. Non-SIG histology, and the combination of GI type and nuclear accumulation of beta-catenin were independent predictors of submucosal invasion. The combination of GI type and MMP-7 expression independently predicted lymphatic invasion. MMP-7 expression correlated with lymph node metastasis.

GI-type early undifferentiated carcinomas and those with non-SIG histology had increased potential for invasion and metastasis. GI type, MMP-7 expression and nuclear accumulation of beta-catenin might prove useful markers in the selection of patients for less invasive surgery.

Helicobacter pylori infection is well accepted to be a very important factor for the development of gastric carcinogenesis in the human stomach. In Mongolian gerbils treated with chemical carcinogens, H. pylori infection enhances glandular stomach carcinogenesis, and eradication of infection and results in curtailment of enhancing effects, particularly at early stages of associated inflammation. A high-salt diet exacerbates the effects of H. pylori infection on gastric carcinogenesis, and these two factors act synergistically to promote the development of gastric cancers in this animal model. However, the bacterium exerts the greater effects. Early acquisition significantly increases gastric chemical carcinogenesis in Mongolian gerbils, as compared to later infection. While heterotopic proliferative glands, hyperplastic and dilated glands localized beneath the muscularis mucosae, frequently develop with H. pylori infection alone in this animal model, they obviously regress on eradication, suggesting a relation to severe gastritis, rather than a malignant character. Furthermore, endocrine cells, positive for chromogranin A, are observed in the heterotopic proliferative glands, in contrast to cancerous lesions which lack endocrine elements. In conclusion, H. pylori is not an initiator, but rather a strong promoter of gastric carcinogenesis, whose eradication, together with reduction in salt intake, might effectively prevent gastric cancer development.

It has previously been reported that Cdx2 is the useful prognostic and intestinal phenotypic marker in advanced gastric cancers (GC). In this study, Cdx2 expression and phenotype in early GC and non-neoplastic surrounding mucosa were examined. A total of 130 early GC (70 intramucosal and 60 submucosally invasive cancers) histologically and phenotypically were evaluated. The expression of Cdx2 was assessed by immunohistochemistry. The lesions were phenotypically divided into 44 gastric (G), 42 gastric and intestinal mixed (GI), 30 intestinal (I), and 14 null (N) types, independent of the histological classification. Most of the early GC were Cdx2-positive, nuclear staining being strongly associated with intestinal phenotypic expression. Early differentiated cancers tended to feature both Cdx2 and intestinal phenotypic expression, while their undifferentiated counterparts were more likely to demonstrate only gastric phenotypic expression (P < 0.05). The phenotypes of six intramucosal microcarcinomas did not correlate with those of adjacent normal glands. These data suggest that Cdx2 is expressed in the very early stage of gastric carcinogenesis in association with the shift from gastric to intestinal phenotypic expression. This appears to occur in differentiated cancers at an earlier stage than in undifferentiated ones, and may be linked to suppression of expansion of malignant cells.

The objectives of this study were to evaluate the clinical utility of MUC2-specific reverse transcriptase-polymerase chain reaction (RT-PCR) in gastric cancer patients and to compare MUC2-specific RT-PCR to carcinoembryonic antigen (CEA)-specific RT-PCR.

A total of 305 lymph nodes from 28 patients with gastric cancer were histologically examined and analyzed by MUC2 RT-PCR and CEA RT-PCR.

MUC2 and CEA were expressed in 17.1 and 7.0% of the 286 histologically negative lymph nodes, respectively. The detection rate of MUC2 was significantly higher than that of CEA (P < 0.01). MUC2 RT-PCR revealed no false positive results in control specimens. Lymph node micrometastases in pT1 gastric cancer were expressed only in perigastric lymph nodes near the tumor and were not detected in tumor less than 30 mm in patients with mucosal cancer.

MUC2-specific RT-PCR is a sensitive and specific method to detect lymph node micrometastases in gastric cancer patients. The distribution of lymph node micrometastases detected by this method may be useful in minimally invasive procedures for gastric cancer, especially pT1 cases.

It is well known that gastric cancers (GCs) at early stages, independent of the histological type, mainly consist of gastric phenotype malignant cells, while those at advanced stage tend to have a more intestinal phenotype with progression. However, the connection between this shift and expression of homeobox genes, which are important factors to maintain tissue character, has remained unclear. We therefore evaluated the expression of Cdx1/2 in relation to the phenotype of GCs.

We analyzed the expression of Cdx1/2 mRNAs by Northern blotting and Cdx2 protein by immunohistochemistry in seventy advanced GCs, and evaluated phenotypically using mucin- and immunohistochemistry.

Seventy GCs were divided phenotypically into 16 gastric (G type), 18 gastric and intestinal mixed (GI type), 18 intestinal (I type), and 18 null (N type) phenotypes, independent of the histological classification. Cdx1 and Cdx2 mRNAs statistically demonstrated an increase with shift from G to I ( P=0.042 and P=0.0082, respectively). Cdx2 nuclear staining was observed immunohistochemically in the intestinal phenotypic cancer cells, but could not be detected in those with only the gastric phenotype.

These results show that Cdx1 and Cdx2 might be indispensable for intestinal phenotypic expression even in gastric cancer cells.

It is well known that both gastric and intestinal phenotypic markers are expressed in gastric carcinomas, irrespective of their histological type. In the present study, the associations among phenotypic marker expression of gastric carcinomas, tumor thymidylate synthase (TS) expression, and the chemotherapeutic response to 5-fluorouracil (5-FU) were examined.

The gastric and intestinal phenotypic marker expression of the tumor was determined by the combination of the expression of human gastric mucin (HGM), MUC6, MUC2, and CD10, and was evaluated in comparison with tumor TS expression in 137 advanced gastric carcinomas in 137 patients (75 with postoperative chemotherapy with 5-FU and 62 without postoperative chemotherapy). Tumors were classified into the gastric- (G-), gastric and intestinal mixed- (GI-), intestinal- (I-), or unclassified- (UC-) phenotype according to the immunopositivity of HGM, MUC6, MUC2, and CD10 stainings. The associations among the gastric and intestinal phenotypic marker expression of the tumor, tumor TS expression, effect of postoperative chemotherapy with 5-FU, and the patient's prognosis were examined.

Of the 137 gastric carcinomas, 48 (35.0%), 58 (42.3%), 23 (16.8%), and 8 (5.8%)were classified as the G-, GI-, I- and UC-phenotype, respectively. The high TS expression of more than 25% tumor cell positivity was found in 25 (52.1%) of the 48 G-phenotype tumors, 39 (67.2%) of the 58 GI-phenotype tumors, 18 (78.3%) of the 23 I-phenotype tumors, and 4 (50.0%) of the 8 UC-phenotype tumors. The I-phenotype tumors were significantly correlated with the higher rate of the high TS expression as compared with the G-phenotype tumors (P<0.05). Among 48 patients with the G-phenotype tumor, the 5-year survival rate in patients with and without postoperative chemotherapy was 39.7 and 27.8%, respectively. The patients with postoperative chemotherapy had a significantly better prognosis than those without postoperative chemotherapy (P<0.05). Conversely, there were no significant correlations between the presence of postoperative chemotherapy and the patient's prognosis among patients with GI-, I-, and UC-phenotype tumors.

These results indicate that postoperative chemotherapy with 5-FU could be effective for patients with the G-phenotype tumor, since the incidence of intratumoral expression of TS, the target enzyme of 5-FU, is significantly low in G-phenotype tumors.

The clinicopathologic significance of the phenotype and Cdx2 expression has hitherto remained unclear in gastric cancers. In the present study, we therefore examined the correlation between prognosis, phenotype, and Cdx2 expression in advanced cases.

We evaluated 177 advanced gastric carcinomas histologically and phenotypically. The expression of Cdx2 was also assessed by immunohistochemistry.

The lesions were phenotypically divided into 32 gastric (G type), 36 gastric and intestinal mixed (GI type), 53 intestinal (I type), and 56 null (N type) types, independent of the histological classification. Cdx2 nuclear staining demonstrated a close relation to intestinal phenotypic expression, not with the histological classification. Kaplan-Meier analysis of Cdx2 expression and the phenotype showed that the Cdx2 positive groups had a significantly better outcome than the negative ones (P=0.0013), and the patients with GI type cancers had significant better survival than those with N type (P=0.0052).

Our results suggest that Cdx2 is a useful prognostic marker. In addition, advanced gastric cancers with both intestinal and gastric phenotypic expression have a relatively good prognosis. Combined evaluation of gastric and intestinal epithelial cell markers, including Cdx2, is clinically useful to predict outcome in patients with advanced neoplasm of the stomach.

Hepatoid adenocarcinoma of the stomach is a highly malignant neoplasm. Most gastric hepatoid adenocarcinomas coexist with tubular adenocarcinoma. However, the relationship between hepatoid adenocarcinoma and tubular adenocarcinoma is still unclear. In the present study, the characteristics of the coexistent tubular adenocarcinomas were determined by examining their profiles of mucin production. Subsequently, molecular pathological techniques were applied to determine the clonality of 15 mixed hepatoid and tubular adenocarcinomas of the stomach. Mucin analysis suggested that the coexistent tubular adenocarcinomas with hepatoid adenocarcinoma were of the intestinal type. The patterns of chromosome X inactivation were identical between the hepatoid adenocarcinoma component and tubular adenocarcinoma component in all of 3 informative female cases. Mutations in the p53 gene were found in 5 cases. The sequences were identical within both tumor components in all 5 cases. Microsatellite analysis indicated more than 3 common patterns of loss of heterozygosity in 8 cases. These observations strongly suggested that hepatoid adenocarcinomas were of identical origin to coexistent tubular adenocarcinomas.

The aim of this study was to investigate the relationship between MUC1 and MUC2 mucin expressions and clinicopathologic variables in gastric carcinomas with regard to survival times. MUC1 and MUC2 expressions were revealed immunohistochemically in 143 gastric carcinomas. Of these 143 patients, follow-up data were available for 45 (median survival time of 30 months, ranging from 2 to 80 months). MUC1 was detected in 82 (58%), and MUC2 in 60 (42%) out of 143 cases. Papillary adenocarcinomas showed significantly higher MUC1 and MUC2 immunoreactivity than did signet-ring cell and mucinous tumors (p = 0.045 and p = 0.01, respectively). MUC1 was highly positive in intestinal-type carcinomas (p = 0.006), whereas intestinal and diffuse carcinomas did not differ in MUC2 expression. There was a positive correlation between tumor differentiation and MUC1 expression. However, no correlation was found between MUC1 and MUC2 expressions and angiolymphatic invasion. According to the TNM classification, stage 1A tumors have significantly lower rates of MUC1 reactivity compared to higher stages (p = 0.04). The patients with gastric carcinomas expressing MUC1 showed significantly poorer survival than those without MUC1 expression (p = 0.04). The present study suggests that MUC1 expression be a useful prognostic factor for predicting the outcome of gastric carcinoma patients, whereas the role of MUC2 expression is still unclear.

All of the different types of stomach epithelial cells are known to be derived from a single progenitor cell in each gland. Similarly, cancers develop from single cells, based on data from clonality analysis in C3H/HeN<-->BALB/c chimeric mice. Using gastric and intestinal epithelial cell markers, intestinal metaplasia (IM) can be divided into two major types: a gastric and intestinal (GI) mixed type, and a solely intestinal (I) type. Ectopic expression of Cdx genes and down-regulation of Sox2 in isolated single GI mixed IM glands suggests abnormal differentiation of stem cells that can produce both gastric (G) and I type cells. Similarly, phenotypic expression of gastric cancer cells of each histological type can be clearly classified into G and I type epithelial cells. The heterogeneity of phenotypic expression of gastric cancer cells in individual cancers is assumed to reflect this intrinsic potential for differentiation in two directions. Gastric cancers at early stages, independent of the histological type, mainly consist of G type cells, and phenotypic shift from G to I type expression is clearly observed with progression. The data thus suggest IM may not be a preneoplastic change in gastric carcinoma, but rather that cells of the I type may appear independently in the gastric mucosa in IM and in gastric cancers. Intestinalization of gastric mucosa and cancer cells may represent a kind of homeotic transformation. Whether disturbance of the regulation of Sox2 and Cdx genes may be of importance to the biological behavior of gastric cancers should therefore be clarified in future studies.

Strain differences in susceptibility regarding stomach carcinogenesis due to N-methyl-N-nitrosourea were examined in males of six strains of mice: BALB/cA (BALB), C57BL/6N (C57BL6), CBA/JN (CBA), C3H/HeN (C3H), DBA/2N (DBA/2), and CD-1 (ICR). The frequency of pepsinogen-altered pyloric glands (PAPGs), putative precancerous lesions, was highest (19.6+/-9.9%) in the BALB and lowest in the ICR (12.3+/-5.7%) mice (P<0.05). Incidences of adenocarcinomas at week 52 were 59.3% (16 of 27) and 18.5% (5 of 27), respectively (P<0.005). Invasion also tended to be deepest in BALB compared with the other strains. Intestinal alkaline phosphatase-positive intestinal type cells were observed heterogeneously in some hyperplasias, adenomas and adenocarcinomas consisting of gastric type cells. Thus, intestinalization appeared to occur at random in both non-neoplastic and monoclonal neoplastic lesions, making it unlikely that IAP-positive cells could be precursors of gastric tumors. In contrast, the data suggest a direct histogenetic role for the PAPG, a useful preneoplastic marker lesion in mouse strains.

A well-differentiated (intestinal-type) early gastric cancer characterized by multifocal areas containing adenocarcinoma cells with eosinophilic cytoplasm is presented. The electron microscopy revealed that cytoplasmic eosinophilia was mainly due to the presence of numerous mitochondria. In addition, individual cells showed features of preparietal cell differentiation such as numerous mitochondria, tubulovesicles, intracellular canaliculi and rare mucin granules. There were also adenocarcinoma cells with a pyloric gland (mucopeptic) differentiation as revealed by the presence of granules with a dense protein core eccentrically located in the mucus matrix. These ultrastructural data suggested a gastric epithelial cell differentiation. The implications with regard to the classification and histogenesis of this tumor are discussed.

The expression of trefoil peptides (TFF1 and TFF2) and mucins (MUC1, MUC2, MUC5AC, and MUC6) has previously been described in gastric polyps. In the present study, the expression profile of these trefoil peptides and mucins was characterized in 96 gastric carcinomas, in an attempt to further the understanding of the histogenesis and cell differentiation of gastric carcinoma. Taking together the co-expression of trefoil peptides and mucins, three phenotypes were defined: complete gastric, incomplete gastric, and non-gastric phenotype. Gastric differentiation (complete and incomplete) was observed in 30 out of 33 (90.9%) diffuse carcinomas and in 38 out of 53 (71.7%) intestinal carcinomas. Non-gastric differentiation was observed in only three (9.1%) diffuse carcinomas and in 15 (28.3%) intestinal carcinomas. The phenotypes observed in intestinal carcinomas were similar to those previously observed in adenomatous polyps, whereas most diffuse carcinomas mimicked the phenotype of hyperplastic polyps. The percentage of cases displaying a non-gastric phenotype was higher, though not significantly, in tumours that had invaded the gastric wall than in T1 tumours, regardless of histotype. It is concluded that gastric-type differentiation is retained in the majority of gastric carcinomas, being more prominent in diffuse than in intestinal carcinomas, and in early than in advanced carcinomas.

The expression of trefoil peptides (TFF1 and TFF2) and mucins (MUC1, MUC2, MUC5AC, and MUC6) has previously been described in gastric polyps. In the present study, the expression profile of these trefoil peptides and mucins was characterized in 96 gastric carcinomas, in an attempt to further the understanding of the histogenesis and cell differentiation of gastric carcinoma. Taking together the co-expression of trefoil peptides and mucins, three phenotypes were defined: complete gastric, incomplete gastric, and non-gastric phenotype. Gastric differentiation (complete and incomplete) was observed in 30 out of 33 (90.9%) diffuse carcinomas and in 38 out of 53 (71.7%) intestinal carcinomas. Non-gastric differentiation was observed in only three (9.1%) diffuse carcinomas and in 15 (28.3%) intestinal carcinomas. The phenotypes observed in intestinal carcinomas were similar to those previously observed in adenomatous polyps, whereas most diffuse carcinomas mimicked the phenotype of hyperplastic polyps. The percentage of cases displaying a non-gastric phenotype was higher, though not significantly, in tumours that had invaded the gastric wall than in T1 tumours, regardless of histotype. It is concluded that gastric-type differentiation is retained in the majority of gastric carcinomas, being more prominent in diffuse than in intestinal carcinomas, and in early than in advanced carcinomas.