Immunohistochemical Markers of the Epithelial-to-Mesenchymal Transition (EMT) Are Related to Extensive Lymph Nodal Spread, Peritoneal Dissemination, and Poor Prognosis in the Microsatellite-Stable Diffuse Histotype of Gastric Cancer

Abstract

Background: Although the prognostic value of the epithelial-to-mesenchymal transition (EMT) in gastric cancer has been reported in several studies, the strong association with the diffuse type may represent a confounding factor. Our aim is to investigate potential correlations among EMT status, tumor advancement, and prognosis in diffuse gastric cancer. Methods: Between 1997 and 2012, 84 patients with microsatellite-stable (MSS) diffuse-type tumors underwent surgery. The EMT phenotype was assessed with the E-cadherin, CD44, and zinc finger E-box binding homeobox 1 (ZEB-1) immunohistochemical markers. Results: Forty-five out of 84 cases (54%) were EMT-positive; more advanced nodal status (p = 0.010), pTNM stage (p = 0.032), and vascular invasion (p = 0.037) were observed in this group. The median numbers of positive nodes (13 vs. 5) and involved nodal stations (4 vs. 2) were higher in the EMT-positive group. The cancer-related survival time was 26 months in EMT-positive cases vs. 51 in negative cases, with five-year survival rates of 17% vs. 51%, respectively (p = 0.001). The EMT status had an impact on the prognosis of patients with <70 years, R0 resections, or treatment with adjuvant chemotherapy. Tumor relapses after surgery and peritoneal spread were significantly higher in the EMT-positive tumors. Conclusions: EMT status, when assessed through immunohistochemistry, identified an aggressive phenotype of MSS diffuse-type tumors with extensive lymph nodal spread, peritoneal dissemination, and worse long-term outcomes.

Keywords: diffuse histotype; epithelial-to-mesenchymal transition; gastric cancer; immunohistochemistry; microsatellite-stable tumors; outcome; prognosis.

Figure 1

Representative samples of positivity for epithelial-to-mesenchymal transition markers: (a) A poorly cohesive signet ring cell phenotype (acc. to WHO) or diffuse type (acc. to Lauren) infiltrating the muscular layer is presented; (b) a 8/18 cytokeratin stain confirming the epithelial origin of the neoplastic proliferation is shown; (c) E-cadherin negativity of neoplastic cells is depicted; the positive control is represented by the normal gastric epithelium on the right side expressing E-cadherin; (d) neoplastic cells show membranous positivity for CD44 and nuclear expression of Zeb-1 (inset, bottom right) in more than 10% of cells. (a) Hematoxylin and eosin; (b) CK8/18 stain; (c) E-cadherin stain; (d) CD44 stain; inset on the bottom right, Zeb-1 stain. Original magnification (OM): (a–d), 4×; (d), inset, 10×.

Figure 2

Representative sample of negativity for epithelial-to-mesenchymal transition markers: (a) A poorly cohesive other cell type (WHO) or diffuse type (Lauren) is presented; (b) a 8/18 cytokeratin stain confirming the epithelial origin of the neoplastic proliferation is shown; (c) E-cadherin positivity of neoplastic cells is depicted; (d) neoplastic cells do not show positivity for CD44 and Zeb-1 (inset, bottom right); the positive control is represented by lymphocytes (arrows). (a) Hematoxylin and eosin; (b) CK8/18 stain; (c) E-cadherin stain; (d) CD44 stain; inset on the bottom right, Zeb-1 stain. Original magnification (OM): (a–d), 4×; (d), inset, 10×.

Figure 3

(a) Box-plot representation of the correlation between EMT status and the number of positive lymph nodes (Mann–Whitney U-test). The box represents the interquartile range (IQR), which contains 50% of the values; the lines extending from the box indicate the highest and lowest non-outlier values, while the line across the box indicates the median value. Circles represent “mild” outliers, i.e., values between 1.5 and 3.0 times the IQR; asterisks represent “extreme” outliers, i.e., values more than 3.0 times the IQR. (b) Box-plot representation of the correlation between EMT status and the number of metastatic lymph nodal stations classified according to the JGCA guidelines (Mann–Whitney U-test). A median number of 10 lymph node stations for patients were removed (range 6 to 15); a total number of 924 lymph node stations were analyzed. (c) Linear correlation between the number of removed and positive lymph nodes in EMT-positive and EMT-negative cases.

Figure 4

Cancer-specific survival according to EMT status in MSS diffuse-type tumors; the difference is statistically significant (Log-rank test).

Figure 5

Cancer-specific survival according to EMT status: (a) MSS diffuse-type tumors that did not involve the serosa and (b) MSS diffuse-type tumors that involved the serosa; stratified for age groups: (c) <70 years old and (d) ≥70 years old (Log-rank test).