Abstract
Context.—The biologic effects of Notch1 and Notch2 vary with cancer types and their potential role(s) in gastric cancers (GCs) remains largely unknown.
Objectives.—This study aimed to address the previously mentioned issue by checking the expression of Notch1, Notch2, and Notch target gene Hes1 in GCs, premalignant gastric lesions, and noncancerous endoscopic gastric mucosa and by inhibiting Notch signal transduction in GC cells.
Design.—The status of Notch1, Notch2, and Hes1 expression in 74 GC surgical specimens, 10 endoscopic samples, and 4 human GC cell lines was evaluated by tissue microarray–based immunohistochemical staining, Western blotting, and reverse transcription-polymerase chain reaction, and the importance of Notch signaling was elucidated by treating 2 GC cell lines with 2 γ-secretase inhibitors.
Results.—Notch1 was undetectable in noncancerous gastric mucosa but was expressed with nuclear translocation in 16.7% (4 of 24) of chronic gastritis, 50.0% (9 of 18) of intestinal metaplasia, 54.2% (26 of 48) of intestinal GC, and 23.1% (6 of 26) of diffuse GC, showing distinct differences of Notch1 detection rates between either intestinal metaplasia and chronic gastritis or intestinal GCs and diffuse GCs (P = .03; P = .005, respectively). Notch2 nuclear translocation frequencies were 10.0% (1 of 10) in noncancerous endoscopic mucosa, 71.4% (30 of 42) in premalignant lesions, and 97.3% (72 of 74) in GC tissues, demonstrating a correlation of Notch2 expression with both intestinal GC and diffuse GC formation (P < .001). The rates of nuclear-Hes1 labeling were 1 of 10 among noncancerous, 42.9% premalignant, and 81.1% cancer tissues, which were closely correlated with Notch2 (P < .001) rather than Notch1 (P = .42) nuclear translocation. Only Notch2 was expressed accompanied with Hes1 nuclear labeling in the 4 GC cell lines established from diffuse GC cases. Inhibition of Notch signaling with γ-secretase inhibitors, L-685,458 and DAPT, prevented Hes1 nuclear translocation but neither suppressed growth nor induced cell death.
Conclusions.—This study demonstrated a close correlation of Notch2 expression with GC formation and the potential link of Notch1 upregulation with intestinal-like phenotypes of gastric lesions. Although inhibition of Notch activity failed to achieve anti-GC effects, the activated Notch signaling may reflect a potential GC risk.
Notch signaling plays pivotal roles in determining the manner of cell differentiation, tissue/organ formation, apoptosis in the early stages of development, and the processes of wound and healing of adult tissues.1–4 So far, 4 members have been found in the Notch family, of which Notch1 and Notch2 are well characterized genetically and functionally.5,6 Recent data demonstrated that the biologic effects of Notch signaling varied in cell type–dependent fashions.7,8 Moreover, Notch1 and Notch2 proteins may function differently and even in reversed manners in the same cell types. These 2 proteins have different roles in the commitment and differentiation of neuronal and glial lineages of the olfactory epithelium during development.9 In addition, Notch1 rather than Notch2 is required for neuronal and glial differentiation in the cerebellum.10 This evidence suggests that the signaling and functioning of Notch family proteins are not as simple as what has been supposed.
Cancer is the problem of mal-differentiation due to genetic and epigenetic alterations.11–14 A body of evidence showed that Notch1 and Notch2 were involved in tumorigenic processes in the forms of altered upregulation or downregulation.15,16 Because the biologic effects of individual Notch members may not be equivalent on different cancer types, contradictory results have been reported. For instance, Notch1 was regarded as a tumor suppressor, whereas Notch2 was regarded as a tumor promoter for medulloblastoma cells and B-cell chronic lymphocytic leukemia,16,17 whereas a reverse situation was found in myeloma and breast cancers.18,19 Furthermore, these 2 proteins work in a coordinated manner to promote cutaneous malignant melanoma20 but to inhibit varicocele formation.21 Apparently, a cautious evaluation should be performed on a cancer system in which the roles of the Notch pathway remain largely unknown.
Gastric cancer (GC) is one of the commonest malignancies in the world. The gastric epithelial cells undergo stepwise morphologic changes before malignant transformation.22 Multiple signaling pathways have been known to be involved in gastrocarcinogenic processes. Concurrent activation of wingless-type MMTV integration site family member 2, signal transducer and activator of transcription 3, and nuclear factor of κB pathways were frequently found in the 2 major subtypes of GCs, intestinal GC (i-GC) and diffuse GC (d-GC),23–25 which may have certain cross-talks with Notch signaling in the form of coactivation,26 the molecular components shared by them,27 and similar biologic consequence of their alterations to the cells.28 However, the corresponding study of the Notch pathway in GCs has not yet been available. Our current study thus aimed to address this issue by checking the status of Notch signaling in GCs and their noncancerous and premalignant counterparts and by evaluating the influence of Notch inhibition in GC cells.
MATERIALS AND METHODS
Sample Collection
The detailed description including the human samples used and experimental protocols of the current study were reviewed by the institutional review board of Dalian Medical University, Dalian, China. After getting institutional review board approval and patients' consent, 74 GC surgical specimens and 10 endoscopic specimens were collected from the operating rooms at Dalian Medical University First and Second Affiliated Hospitals. The tissues were chosen from the tumor mass and, where possible, tissue surrounding the tumor, and then trimmed to suitable sizes on ice. After getting patients' consent, fresh endoscopic gastric biopsies from 10 GC-free patients were collected from the Gastroendoscopic Department of Dalian Medical University Second Affiliated Hospital. The samples were initially used for Helicobacter pylori testing and then for research purposes. Part of the tissue samples was snap-frozen immediately in liquid nitrogen and stored at −85°C until use; the remaining tissues were fixed with 10% formaldehyde for preparing paraffin blocks. Staining with hematoxylin-eosin was performed on each of the tissue blocks and a pathologic diagnosis determined before further experiments.
Tissue Microarray Construction and Immunohistochemical Staining
The representative donor tissues in tissue blocks were sampled with individual coring needles and orderly embedded into a recipient paraffin block according to the methods described elsewhere.29 The tissue microarrays with densities of 56 spots/cm2 were sectioned to 5-µm thickness for immunohistochemical profiling of Notch1, Notch2, and Hes1 expression and intracellular distribution, respectively. The antibodies used were a mouse and a goat anti-human Notch1 antibody (1∶100; MS-1339-P0, NeoMarkers, Inc, Fremont, California and 1∶120; sc-6014, Santa Cruz Biotechnology, Inc, Santa Cruz, California), 2 rabbit anti-human Notch2 polyclonal antibodies (1∶100; SC-5545, Santa Cruz Biotechnology and 1∶200; ab8926, Abcam plc., Cambridge, United Kingdom), and a rabbit anti-human Hes1 polyclonal antibody (1∶1000; a generous gift of Tetsuo Sudo, PhD, Toray Industries, Tokyo, Japan). The tissue microarray sections lacking incubation with individual primary antibodies were used as background controls. The staining results were evaluated by 2 researchers (H.L. and Y.S.), with the intensity of immunolabeling scored as negative (−), weakly positive (+), moderately positive (++), or strongly positive (+++).
Cell Culture and Analyses
Human GC cell lines AGS30 and HGC-2731 were purchased from the American Type Culture Collection (ATCC, Manassas, Virginia), and MGC80332 and BGC823 cell lines33 were kindly provided by Beijing Cancer Institute, Peking University, Beijing, China. According to the literature, all of the 4 cell lines were established from poorly differentiated GCs.30–33 The HGC-27, MGC803, and BGC823 cells were cultured in Dulbecco modified Eagle medium (Invitrogen Co, Grand Island, New York) containing 10% fetal bovine serum (Invitrogen) and AGS cells in Ham F-12 medium (Invitrogen) supplemented with 10% fetal bovine serum at 37°C in 5% carbon dioxide. For immunocytochemical (ICC) and immunofluorescence (IF) studies, gelatin-coated coverslips were put onto the dishes before initial cell seeding. When they grew to a semiconfluent state, the cell-bearing coverslips were collected for ICC and IF staining; the remaining cells were harvested for RNA and protein preparations, followed by Notch1-, Notch2-, and Hes1-oriented reverse transcription-polymerase chain reaction (RT-PCR) and Western blot analyses.
RNA Isolation and RT-PCR
The cell pellets collected from each of the GC cell lines was placed immediately in a 1.5-mL Eppendorff tube containing 400-µL ice-cold TriZol reagent (Invitrogen). RNA isolation was conducted according to the method described elsewhere.34 RT-PCR was performed using a Takara RNA PCR kit (Avian Myeloblastosis Virus version 2.1, Takara Inc Ltd, Dalian Branch, Dalian, China) under the following conditions: 0.8 µg of total RNA was reverse-transcribed at 55°C for 30 minutes in a 20-µL solution containing random primer, and the reaction was terminated by incubating at 99°C for 5 minutes. β-Actin, a quantitative and qualitative control, was polymerase chain reaction–amplified according to conditions described elsewhere34 to generate a 326-bp fragment, followed by amplification of Notch1, Notch2, and Hes1 transcripts with 3 pairs of specific polymerase chain reaction primers (Takara), respectively. The sequences of polymerase chain reaction primers for each of the gene transcripts are as follows: Notch1 (109 bp) upstream 5′-CCG CAG TTG TGC TCC TGA A-3′, downstream 5′-ACC TTG GCG GTC TCG TAG CT-3′16; Notch2 (533 bp) upstream 5′-TCG TGC AAG AGC CAG TTA CCC-3′, downstream 5′-AAT GTC ATG GCC GCT TCA GAG-3′35; Hes1 (254 bp) upstream 5′-CCA GTT TGC TTT CCT CAT TCC-3′, downstream 5′-TCT TCT CTC CCA GTA TTC AAG TTC C-3′. Polymerase chain reaction products were separated on 1.5% agarose gels stained with 0.5 µg/mL ethidium bromide and photographed with the automatic ultraviolet illuminator (UVP, Inc, Upland, California).
Protein Preparation and Western Blotting
The cell pellets were collected from the 4 GC cell lines, respectively. Each of them were mixed with 150-µL prechilled 2 X Laemmli sodium dodecyl sulfate sample buffer containing 0.125M Tris-HCI, pH 6.8; 4% sodium dodecyl sulfate; 20% v/v glycerol; and 0.2M dithiothreitol. After being homogenized by ultrasonic vibration (UP200s; Hielscher Ultrasonics GmbH, Teltow, Germany), the samples were heated in boiling water for 5 minutes. The protein-containing supernatant liquids were removed to fresh tubes for experimental purposes.
For Western blot analysis, the protein samples (100 µg/sample) were separated by 10% sodium dodecyl sulfate–polyacrylamide gel electrophoresis and transferred onto a polyvinylidene difluoride membrane (Amersham Biosciences Inc, Bucks, United Kingdom). The prepared membrane was blocked with 5% skim milk (Becton Dickinson France SAS, le Pont de Claix, France) in Tris-buffered saline Tween-20 at 4°C overnight. The membrane was then incubated at room temperature for 1 hour with anti-Notch1 antibody (C-20; sc-6014), followed by 3 washes in blocking solution for 10 minutes each. Secondary antibody incubation was conducted for 1 hour with horseradish peroxidase-conjugated rabbit anti-mouse IgG (1∶3000). Bound antibody was detected using the enhanced chemiluminescence solution (Roche Diagnostic GmbH, Mannheim, Germany) and exposing the membrane to x-ray film (Kodak Co, Rochester, New York). After removing the labeling signal by incubation with stripping buffer (62.5mM Tris-HCI, pH 6.7; 100mM 2-Mercaptoethonal; 2% sodium dodecyl sulfate) at 55°C for 30 minutes, the membrane was reprobed with anti-Notch2 polyclonal antibody (ab8926) and thereafter with a rabbit anti-human Hes1 polyclonal antibody.
Treatments With γ-Secretase Inhibitors
To elucidate the importance of Notch1 and Notch2 in the growth of GC cells, Notch2-expressing AGS and BGC823 cell lines were treated by 2 γ-secretase inhibitors, respectively. The experiment was set in 4 groups as group 1, normal culture; group 2, treatment with 8µM L-685,458 (γ-secretase inhibitor X; EMD Biosciences, Inc, La Jolla, California); group 3, treatment with 10µM N-[N-(3,5-difluorophenacetyl)-l-ananyl]-S-phenyglycine t-butyl ester (γ-secretase inhibitor IX/DAPT; EMD Biosciences)36,37; and group 4, treatment with 0.2% dimethyl sulfoxide, the organic solvent for L-685,458 and DAPT. Total cell numbers and cell viability were determined in 12-hour intervals by staining the suspended single cells with 0.25% trypan blue.38 The stained and unstained cells were counted by 3 individuals with hemocytometers. Meanwhile, the cell-bearing coverslips were collected at 4-hour intervals and fixed in cold acetone for Hes1-oriented ICC staining and IF. The treatments lasted for 48 hours and were repeated at least 3 times.
Statistical Analyses
Mann-Whitney tests were used to analyze the status of Notch1, Notch2, and Hes1 expression in different histologic groups. The Spearman correlation test was used to determine the correlation of Hes1 expression with Notch1 expression and Notch2 expression and Notch1, Notch2, and Hes1 expression with malignant transformation. The level of statistical significance was always set at P < .05. All P values are 2-sided. Statistical analysis was performed using the statistical software SPSS version 11.5 (SPSS, Inc, Chicago, Illinois).
RESULTS
Histologic Examination and Classification
Gastric cancer tissue and, when possible, the premalignant mucosa surrounding the tumor, were taken from the same surgical specimen during sample collection. Histologic reevaluation was performed on them prior to further experiments. A total of 42 GC surrounding premalignant tissues (24 chronic gastritis and 18 intestinal metaplasia [IM]) and 74 cancer tissues (48 i-GC and 26 d-GC) were examined in this study. Ten gastroendoscopic biopsies with mild inflammatory changes were obtained from GC-free patients and used as relatively normal controls.
Preferable Notch1 Expression and Nuclear Translocation in IM and i-GCs
The results of tissue microarray–based immunohistochemical staining are summarized in Table 1 and shown in Figure 1, A through E and Figure 2. It was found that Notch1 expression and nuclear translocation was absent in all of the 10 noncancerous endoscopic gastric mucosa specimens and became detectable in 20.8% (5 of 24) and 16.7% (4 of 24) of chronic gastritis and in 61.1% (11 of 18) and 50.0% (9 of 18) of IMs, respectively. The rates of Notch1 detection were 58.3% (28 of 48) and 54.2% (26 of 48) in i-GCs and the same 23.1% (6 of 26) in d-GCs of cytosolic and nuclear stainings, respectively. Statistical analyses revealed that the frequencies of Notch1 expression and nuclear translocation were significantly different between (1) the noncancerous group and premalignant (cytoplasm, P = .02; nuclear, P = .046) or GC group (cytoplasm, P = .007; nuclear, P = .01), (2) the chronic gastritis and IM tissues (cytoplasm, P = .009; nuclear, P = .03), and (3) the i-GC and d-GC group (cytoplasm, P = .002; nuclear, P = .005). In comparison with the results of normal gastric tissues, the incidence of Notch1 detection was distinctly increased in IM and i-GC (cytoplasm, r = 0.236, P = .04; nuclear, r = 0.290, P = .01) but not in chronic gastritis and d-GC (cytoplasm, r = 0.160, P = .22; nuclear, r = 0.191, P = .14).
Tissue microarray–based immunohistochemical staining for Notch1, Notch2, and Hes1. A, F, and K, Noncancerous gastric mucosa. B, G, and L, Chronic gastritis. C, H, and M, Intestinal metaplasia. D, I, and N, Intestinal-type gastric cancer. E, J, and O, Diffuse-type gastric cancer (original magnifications ×400).
Tissue microarray–based immunohistochemical staining for Notch1, Notch2, and Hes1. A, F, and K, Noncancerous gastric mucosa. B, G, and L, Chronic gastritis. C, H, and M, Intestinal metaplasia. D, I, and N, Intestinal-type gastric cancer. E, J, and O, Diffuse-type gastric cancer (original magnifications ×400).
The positive rates of Notch1, Notch2, and Hes1 expression in different histologic groups of the stomach. Abbreviations: ChG, chronic gastritis; d-GC, diffuse gastric cancer; i-GC, intestinal gastric cancer; IM, intestinal metaplasia; N, noncancerous gastric mucosa.
The positive rates of Notch1, Notch2, and Hes1 expression in different histologic groups of the stomach. Abbreviations: ChG, chronic gastritis; d-GC, diffuse gastric cancer; i-GC, intestinal gastric cancer; IM, intestinal metaplasia; N, noncancerous gastric mucosa.
Intracellular Distribution of Notch1, Notch2, and Hes1 in Different Gastric Histologic Groups

Correlation of Notch2 Nuclear Translocation With GC Formation
As shown in Table 1, Figure 1, F through J, and Figure 2, Notch2 was expressed in 20.0% (2 of 10) of noncancerous endoscopic gastric mucosa, and only 1 (10.0%) was found with nuclear translocation; in contrast, 73.8% (31 of 42) of premalignant lesions were positive for Notch2 expression and 71.4% (30 of 42) with nuclear staining, showing a significant difference between these 2 groups in expression and nuclear translocation (cytoplasm, P = .002; nuclear, P = .001). The tendency of Notch2 upregulation became more distinct among GC tissues in increased rates and levels of expression (98.6%; 73 of 74) and nuclear translocation (97.3%; 72 of 74; Table 1). Statistical analysis revealed that the frequency of Notch2 expression and nuclear translocation in the GC group was significantly different from that of noncancerous (cytoplasm, P < .001; nuclear, P < .001) and premalignant samples (cytoplasm, P < .001; nuclear, P < .001), whereas no statistical significance of Notch2 expression was found between i-GC and d-GC (cytoplasm, P = .39; nuclear, P = .38). Correlative study showed that Notch2 expression (r = 0.474, P < .001) and nuclear translocation (r = 0.585, P < .001) were closely linked with both i-GC and d-GC formation (P < .001).
Close Correlation of Hes1 and Notch2 Nuclear Translocation
The rates of nuclear Hes1 detection were 10.0% (1 of 10) in noncancerous endoscopic epithelial tissues, 42.9% (18 of 42) in premalignant lesions, and 81.1% (60 of 74) in GC cases (Table 1; Figure 1, K through O), which were significantly different between GCs and the premalignant lesions (P < .001) or the noncancerous mucosa (P < .001). Mann-Whitney tests revealed distinct differences between noncancerous mucosa and premalignant lesions (P = .04) but not between i-GCs and d-GCs (P = .59). Correlative study showed a close relation of Hes1 nuclear translocation with malignant transformation of gastric epithelia (r = 0.474, P < .001). Activation of Notch signaling was decided when Hes1 appeared in the nuclei accompanied with Notch1 and/or Notch2 nuclear translocation (Figure 2). As showed in Table 2, nuclear labeling of Hes1 was correlated with Notch2 (r = 0.442, P < .001) rather than Notch1 (r = 0.072, P = .42) expression.
Constitutive Hes1 and Notch2 Expression in GC Cell Lines
Four established human GC cell lines were used to validate the previously discussed in vivo findings, which had been derived from poorly differentiated adenocarcinomas.30–33 They were checked with the same parameters for tissue samples by ICC and IF staining combined with RT-PCR and Western blot analyses. Both ICC and IF demonstrated that Notch2 and Hes1 were detected in BGC823, MGC803, AGS, and HGC-27 cells, whereas none of those cell lines was found with Notch1 expression (Figure 3, A through C). Consistent with the previously mentioned findings, RT-PCR generated sufficient amounts of Notch2 and Hes1, whereas Notch1 was almost undetectable in the 4 cell lines (Figure 3, D). Accordingly, Western blot analyses clearly demonstrated a 100-kDa Notch2 band and a 45-kDa Hes1 band but no sign of Notch1 production in the 4 cell lines (Figure 3, D).
Immunocytochemical illustration of Notch1 (A), Notch2 (B), and Hes1 (C) expression and intracellular distribution in AGS human gastric cancer cell line derived from a poorly differentiated gastric cancer (diffuse gastric cancer; original magnifications ×400). The results of reverse transcription–polymerase chain reaction and Western blot analyses of Notch1, Notch2, and Hes1 expression in BGC823, MGC803, AGS, and HGC-27 cell lines. β-actin was used as qualitative and quantitative control (D). The antibodies that recognize the intracellular domains of Notch1 (sc-6014) and Notch2 (ab8926) were used in Western blotting.
Immunocytochemical illustration of Notch1 (A), Notch2 (B), and Hes1 (C) expression and intracellular distribution in AGS human gastric cancer cell line derived from a poorly differentiated gastric cancer (diffuse gastric cancer; original magnifications ×400). The results of reverse transcription–polymerase chain reaction and Western blot analyses of Notch1, Notch2, and Hes1 expression in BGC823, MGC803, AGS, and HGC-27 cell lines. β-actin was used as qualitative and quantitative control (D). The antibodies that recognize the intracellular domains of Notch1 (sc-6014) and Notch2 (ab8926) were used in Western blotting.
γ-Secretase Inhibitors Suppressed Hes1 Expression But Not Cell Proliferation
To evaluate potential impact(s) of Notch activation in GC cells, 2 specific inhibitors of γ-secretase, L-685,458 (γ-secretase inhibitor X) and DAPT, were used to inhibit proteolytic activation of Notch signaling in AGS and BGC823 cells, respectively. The results revealed that after 8µM L-685,458 treatment for 36 hours or 10µM DAPT treatment for 48 hours, Notch activation was blocked efficiently in reduced cytoplasmic distribution and almost diminished nuclear translocation of Hes1 proteins in both AGS (Figure 4, A through C) and BGC823 cells (data not shown). However, these treatments neither brought about growth arrest nor cell death (Figure 4, D).
Immunocytochemical staining (original magnifications ×400) and immunofluorescence staining (insets for A, B, and C, original magnifications ×400) of Hes1 demonstrated that Notch activation in AGS gastric cancer cells (A) was blocked efficiently after treatment with 8µM L-685,458 for 36 hours (B) or with 10µM N-[N-(3,5-difluorophenacetyl)-L-ananyl]-S-phenyglycine t-butyl ester (γ-secretase inhibitor IX/DAPT) for 48 hours (C) in reduced cytoplasmic distribution and almost diminished nuclear translocation of Hes1. D, Sequential examination of survival rates of AGS cells incubated in normal culture medium (1) and the medium containing 0.2% dimethyl sulfoxide (2), 8µM L-685,458 (3), and 10µM DAPT (4).
Immunocytochemical staining (original magnifications ×400) and immunofluorescence staining (insets for A, B, and C, original magnifications ×400) of Hes1 demonstrated that Notch activation in AGS gastric cancer cells (A) was blocked efficiently after treatment with 8µM L-685,458 for 36 hours (B) or with 10µM N-[N-(3,5-difluorophenacetyl)-L-ananyl]-S-phenyglycine t-butyl ester (γ-secretase inhibitor IX/DAPT) for 48 hours (C) in reduced cytoplasmic distribution and almost diminished nuclear translocation of Hes1. D, Sequential examination of survival rates of AGS cells incubated in normal culture medium (1) and the medium containing 0.2% dimethyl sulfoxide (2), 8µM L-685,458 (3), and 10µM DAPT (4).
COMMENT
Notch signaling is activated dynamically during evolution and plays active roles in cell fate determination and differentiation during embryonic development. Therefore, alteration of this signaling may lead to some abnormalities such as cancers39–41 and developmental disorders.42 Multiple Notch family members have been identified, of which Notch1 and Notch2 are well-categorized. Extensive investigations revealed that these 2 Notch proteins functioned in tumor type–related fashion and their roles may not be necessarily identical in different kinds of cancers.16–21 The multifaceted features of Notch family members suggest the necessity to check the activation pattern(s) and potential roles of this signaling in different tumor types without any initial impression.
The biologic functions of Notch1 vary greatly with cancer types. For example, Notch1 was regarded as a target gene of p53 and involved in human keratinocyte tumor suppression,43 whereas it was proved to directly regulate c-Myc gene transcription and promote the growth of leukemia and mammary cancer cells.44–46 More recent study revealed that activated Notch signaling has little influence in the growth and survival of medulloblastoma cells.47 According to the results of the current study, Notch1 was rarely expressed in relatively normal gastric mucosa and appeared in premalignant and cancer tissues. Interestingly, Notch1 was found to be more preferably expressed with distinct nuclear translocation in IM and well-differentiated i-GC, whereas the 4 GC cell lines derived from poorly or undifferentiated GCs were negative in its expression. Because Notch signaling is important in regulating the regeneration, differentiation, and proliferation of the intestinal epithelial cells,48 we therefore speculate that Notch1 may play certain role(s) in promoting metaplastic transition of gastric epithelial cells and/or in maintaining constant proliferation of intestinalized epithelial cells.
It has been reported that Notch2 acts as an oncogene in some types of cancers but its status in GCs remains obscure. It was found in this study that Notch2 expression and nuclear translocation was uncommon in endoscopic noncancerous mucosa with mild inflammatory changes and became frequent in both i-GC (98.6%) and d-GC (97.3%) cases. Western blot analysis demonstrated that the original and proteolytic forms of Notch2 were detected in all of the 4 GC cell lines, suggesting the close correlation of Notch2 expression and proteolytic activation with stepwise gastrocarcinogenesis. Hes1 is a famous target gene of Notch signaling49 because its expression can be triggered when the fragments of cleaved Notch intracellular domain translocalize to nucleus, associate with the DNA-binding-protein, and convert the latter to Hes1 activator.6 Therefore, Hes1 expression and nuclear ending have been regarded as the indicator of Notch activation.50 Unlike the situation of Notch1, coexpression and, more importantly, the nuclear cotranslocation of Notch2 and Hes1 were more frequent in vivo and in vitro. Because all the d-GC cases so far checked are positive in Notch2 and only 23.1% of them are positive in Notch1, it is reasonable to consider that Notch2-mediated signaling would be more important for GC formation and progression.
Recent findings have revealed that Notch signaling can variously promote or suppress tumor growth through regulating differentiation, cell cycle progression, angiogenesis, and possibly cancer stem cell self-renewal and immune function.7,51,52 Nevertheless, the knowledge about the functional aspect of Notch in gastrocarcinogenesis is still limited. Given the evidence of prominent Notch-associated Hes1 expression in GCs and their related lesions, the importance of Notch signaling in the growth and survival of Notch2-activating GC cells was evaluated by inhibiting proteolytic cleavage of Notch transmembrane protein with 2 γ-secretase antagonists. Although Hes1 expression and nuclear translocation were inhibited efficiently by the 2 reagents, both cell lines showed neither growth arrest nor cell death. These phenomena suggest that Notch2 signaling may not be the critical molecular machinery for the maintenance and survival of GC cells. Multiple signaling pathways are activated during gastrocarcinogenesis, of which some such as wingless-type MMTV integration site family member and nuclear factor of κB signaling have certain functional cross-links with Notch signaling.53–56 Therefore, it would be possible that other signaling pathways may compensate Notch signaling when its biologic activity is suppressed. In this context, simply blocking Notch signaling may not achieve promising inhibitory outcome in a GC experimental system.
Taken together, the Notch1 and Notch2 expression patterns and the status of Notch signaling during stepwise gastrocarcinogenesis were profiled in this study, which revealed a close correlation of Notch activation with GC formation. Because blockage of Notch proteolytic activation failed to inhibit growth and to induce cell death of GC cells, it would be possible that beyond Notch signaling, other more powerful cell maintaining systems may be present in GCs. However, the favorable Notch1 expression in gastric lesions with intestinal phenotypes and the frequent Notch2 and Hes1 expression in different premalignant and cancer tissues can still be regarded as adverse biologic indicators of GC risk. As the next research step, it would be worthwhile to check the expression and functional aspect(s) of Notch1 in i-GC cell lines and to elucidate whether inhibition of all Notch signaling could offset the effect of either Notch1 or Notch2 alone on proliferation and survival of GC cells.
This work is supported in part by grants from the National Natural Science Foundation of China (30527002, 30670946, 30700365, and 81071971). We thank the surgeons and nurses in the Department of General Surgery, Dalian Medical University First and Second Affiliated Hospitals for providing gastric specimens; the doctors in the Department of Pathology, Dalian Medical University First Affiliated Hospital for pathologic consultation; and Qi-Gui Liu, PhD, in the Department of Medical Statistics, Dalian Medical University for his help with statistical analyses. We thank Tetsuo Sudo, PhD, from Toray Industries, Tokyo, Japan for providing us with anti-Hes1 antibody.
References
Author notes
From the Liaoning Laboratory of Cancer Genomics and Department of Cell Biology (Drs Sun, Gao, Liu, Q. Wang, and Li, Mr Kong, and Mss Xiao-W. Wang and Chen) and the Faculty of Clinical Medicine, College of Basic Medical Sciences (Mss Cheng and Qu), Dalian Medical University, Dalian, China; and the Department of Pathology, the First Affiliated Hospital of Dalian Medical University, Dalian, China (Dr Gao).
The authors have no relevant financial interest in the products or companies described in this article.