signal-regulating 1 and cyclin D1 compose a positive feedback loop contributing to tumor growth in gastric cancer

Yoku Hayakawaa, Yoshihiro Hirataa, Hayato Nakagawaa, Kei Sakamotob, Yohko Hikibab, Hiroto Kinoshitaa, Wachiko Nakataa, Ryota Takahashia, Keisuke Tateishia, Motohisa Tadac, Masao Akanumab, Haruhiko Yoshidaa, Kohsuke Takedad, Hidenori Ichijod, Masao Omataa, Shin Maedaa,e,1, and Kazuhiko Koikea

aDepartment of Gastroenterology, Graduate school of Medicine, University of Tokyo, Tokyo 113-8655, Japan; bDivision of Gastroenterology, Institute for Adult Diseases, Asahi Life Foundation, Tokyo 100-0005, Japan; cDepartment of Medicine and Clinical Oncology, Graduate School of Medicine, Chiba University, Chiba 260-0856, Japan; dLaboratory of Cell Signaling, Graduate School of Pharmaceutical Sciences, University of Tokyo, Tokyo 113-0033, Japan; and eDepartment of Gastroenterology, Yokohama City University, Yokohama 236-0004, Japan

Edited by Michael Karin, University of California, San Diego School of Medicine, La Jolla, CA, and approved December 2, 2010 (received for review August 4, 2010) -activated protein kinase (MAPK) pathways regulate mul- In this study, we examined the role of ASK1 in gastric tu- tiple cellular functions and are highly active in many types of human morigenesis using both human GC samples and ASK1-deficient − − cancers. Apoptosis signal-regulating kinase 1 (ASK1) is an upstream (ASK1 / ) mice. We demonstrated that ASK1 is important for MAPK involved in apoptosis, inflammation, and carcinogenesis. gastric tumorigenesis through the regulation of cyclin D1 ex- This study investigated the role of ASK1 in the development of pression. Furthermore, we validated an ASK1-dependent posi- gastric cancer. In human gastric cancer specimens, we observed tive feedback loop controlling cyclin D1 expression in GC cells as increased ASK1 expression, compared to nontumor epithelium. a potential anticancer target. Using a chemically induced murine gastric tumorigenesis model, we observed increased tumor ASK1 expression, and ASK1 knockout Results mice had both fewer and smaller tumors than wild-type (WT) mice. Increased Expression of ASK1 in Human Gastric Cancer. To analyze ASK1 siRNA inhibited cell proliferation through the accumulation of the involvement of ASK1 in GC, we first examined expression cells in G1 phase of the cell cycle, and reduced cyclin D1 expression levels of ASK1 in gastric tissue specimens. We observed that the in gastric cancer cells, whereas these effects were uncommon in levels of ASK1 were significantly elevated in GC samples com- other cancer cells. ASK1 overexpression induced the transcription of pared to nontumor gastric epithelium (Fig. 1A). Additionally, we cyclin D1, through AP-1 activation, and ASK1 levels were regulated observed that the levels of cyclin D1 and phosphorylated JNK in by cyclin D1, via the Rb–E2F pathway. Exogenous ASK1 induced GC were higher than in nontumor epithelium. We tested 66 cyclin D1 expression, followed by elevated expression of endoge- samples from nontumor or tumor tissues of patients with GC by nous ASK1. These results indicate an autoregulatory mechanism of immunoblotting and quantified ASK1 expression using densi- ASK1 in the development of gastric cancer. Targeting this positive tometry. We observed significantly increased ASK1 expression in feedback loop, ASK1 may present a potential therapeutic target for GC (P = 0.002; Fig. 1B), confirming that ASK1 was up-regulated the treatment of advanced gastric cancer. in GC cells. Next, to examine whether ASK1 expression increased in other JNK | c-Jun cancers, we compared ASK1 expression in colon cancer and nontumor colon epithelium. Unlike gastric tissues, we found no astric cancer (GC) is a common cancer worldwide, associated difference in ASK1 expression between nontumor and tumor tis- Gwith a high mortality despite its declining incidence in recent sues (Fig. 1 C and D). These results suggest that the up-regulation decades. Smoking, salted or smoked foods, and Helicobacter pylori of ASK1 is specific to GC and may play a significant role in car- appear to be major environmental inducers of GC (1–3). Al- cinogenesis. though the role of H. pylori in causing mucosal effects has been Next, we investigated the expression of ASK1 in cultured GC investigated, which molecular signal(s) initiate the program of cell lines. Among the 29 GC cell lines analyzed, ASK1 expression was detected in 23 cell lines by immunoblotting. Phosphorylation irreversible transformation remain unclear, and thus molecular of JNK was enhanced in the majority of cell lines displaying el- targeting therapies for GC have not been well established. evated ASK1 expression (Fig. 1E). Immunoprecipitation analysis Mitogen-activated protein kinase (MAPK) pathways are im- revealed constitutive ASK1 activation in several GC cell lines, portant for the development of gastric tumorigenesis (4). Apo- suggesting that ASK1 expression and its activity are elevated in ptosis signal-regulating kinase 1 (ASK1) is a ubiquitously expres- gastric tumor cells (Fig. 1F). sed MAPK kinase kinase (MAP3K), activated by various stress stimuli, including reactive oxygen species (ROS), TNF-α, and ASK1 Enhances MNU-Induced Gastric Tumorigenesis in Mice. To ex- LPS (5–7). ASK1 activates the JNK and p38 signaling pathways amine the role of ASK1 in GC, we used a mouse model of chem- and is required for both oxidative stress and cytokine-induced apoptosis (5). Furthermore, ASK1 affects multiple cellular func- tions, including survival, differentiation, and the innate immune Author contributions: Y. Hayakawa, Y. Hirata, and S.M. designed research; Y. Hayakawa, response (5, 7, 8) and has been reported to be involved in the Y. Hirata, H.N., K.S., Y. Hikiba, and M.T. performed research; Y. Hayakawa, Y. Hirata, H.N., K.S., H.K., W.N., R.T., K. Tateishi, M.T., M.A., H.Y., K. Takeda, H.I., M.O., S.M., and pathogenesis of various human diseases, including neurodegen- K.K. analyzed data; and Y. Hayakawa, Y. Hirata, and S.M. wrote the paper. fl erative (9), cardiovascular (10), and in ammatory diseases (11, The authors declare no conflict of interest. 12). Additionally, ASK1 has been shown to participate in both This article is a PNAS Direct Submission. colon (12) and skin (13) tumorigenesis through the regulation of 1To whom correspondence should be addressed. E-mail: [email protected]. fl in ammation and apoptosis. However, no reported study has This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10. demonstrated a role for ASK1 in gastric tumorigenesis. 1073/pnas.1011418108/-/DCSupplemental.

780–785 | PNAS | January 11, 2011 | vol. 108 | no. 2 www.pnas.org/cgi/doi/10.1073/pnas.1011418108 Downloaded by guest on September 29, 2021 Fig. 2. ASK1 enhances MNU-induced gastric tumorigenesis in mice. (A) Schematic overview of the gastric tumorigenesis model. Blue areas indicate

MNU (240 ppm) administration periods, and white areas indicate H2O ad- ministration periods. (B) Typical examples of stomach tumors. (Left)WT mice. (Right) ASK1−/− mice. (Upper) Macroscopic views of the tumor-bearing stomachs. The mucosal surface was stained with indigo carmine. Arrows Fig. 1. Increased expression of ASK1 in human gastric cancer. (A) Immu- indicate the tumors. (Lower) H&E staining of antral areas of the stomach noblot analysis of the indicated proteins in the nontumor epithelium and (magnification, 40×). (C) The numbers (Left) and maximum sizes (Right)of cancer of human stomach. Δ and▴indicate samples from the same indi- the tumors in the stomachs. Stomachs were observed macroscopically and viduals. (B) Graphic representation of ASK1 expression in gastric samples. tumors >0.5 mm in diameter were counted. Values represent the means ± SE The fold expression was quantified by densitometry and normalized to the for tumors from WT mice (n = 10) and ASK1−/− mice (n = 10). *P < 0.05, levels of β-actin. The box plot indicates the median (horizontal line), inter- compared to WT mice. (D) Immunohistochemical analysis of the indicated − − quartile range (the box itself), and 10th and 90th percentiles (bars). Mean proteins in MNU-treated stomachs. (Left) WT mice. (Right) ASK1 / mice. ASK1 expression in nontumor epithelium is represented as 1.0; *P < 0.05. (C) Representative data are shown (magnification: phospho-JNK and phospho- Immunoblot analysis of the indicated proteins in the nontumor (NT) and p38, 100×; c-Jun, 200×). (E) Immunoblot analysis of the indicated proteins in human colon cancer (T) specimens. (D) Graphic representation of ASK1 ex- the nontumor epithelium and tumors of MNU-treated WT and ASK1−/− mice. pression in colon samples. The fold expression was quantified by densi- tometry and normalized to the levels of β-actin. (E) Immunoblot analysis of the indicated proteins in gastric cancer cell lines. Representative data are (12), levels of ASK1 expression were similar to those in non- shown. (F) Total cell lysates were immunoprecipitated with anti-ASK1 anti- tumor epithelium (Fig. S1B). These data indicate that ASK1 may β bodies and immunoblotted with antiphospho-ASK1 antibodies. -Actin was play an especially important role in the development of GC. used as a loading control. ASK1 Is Important for Cell Proliferation in Gastric Cancer Cells. To ically induced gastric tumorigenesis. As the administration of examine whether ASK1 affected tumor growth in vitro, we treated the GC cell lines AGS, MKN45, MKN74, and TMK1 with N-methyl-N-nitrosourea (MNU) results in efficient gastric tumor − − a control small interfering RNA (NCsiRNA) or ASK1-specific induction in mice (14), wild-type (WT) and ASK1 / mice were siRNA (ASK1siRNA; Fig. 3A). Cell proliferation was significantly administered MNU as described previously (Fig. 2A). Histological inhibited following ASK1siRNA transfection, in a dose-de- analysis showed no significant difference between control WT and − − pendent manner (Fig. 3B), and decreased proliferation caused by ASK1 / mice. After MNU treatment, the number of detectable −/− ASK1siRNA was observed in all four GC cell lines examined (Fig. gastric tumors was significantly lower in ASK1 mice (0.30 ± 0.15) C ± 3 ). Next, we investigated the role of ASK1 in the cell cycle status than in WT mice (1.40 0.34). The average size of these tumors was of GC cells using flow cytometry. ASK1siRNA transfected cells −/− ± also smaller in ASK1 mice (0.90 0.30) than those in WT mice displayed a higher proportion of cells in the G0/G1 phase than ± B C (2.20 0.40; Fig. 2 and ) and these tumors available for histo- cells transfected with NCsiRNA (Fig. 3 D and E). logical evaluation showed moderate-to-severe dysplasia. We also evaluated the proliferative effect of ASK1 in non- Then, we investigated the expression of ASK1 and its down- transformed gastric cells. In the rat normal gastric cell line

stream signaling partners using immunohistochemical and im- RGM1, ASK1 knockdown did not reduce cell proliferation (Fig. MEDICAL SCIENCES munoblot analyses. The levels of phosphorylated JNK, p38, and S2 B and C). We also analyzed the cell cycle of gastric epithelial − − total c-Jun were higher in tumor sections of WT mice than cells isolated from WT and ASK1 / mice, using FACS. Consis- −/− ASK1 mice (Fig. 2 D and E). TUNEL staining showed few tent with the normal development of gastric epithelial cells in − − − − apoptotic cells in nontumor and tumor epithelium in both gen- ASK1 / mice, epithelial cells from ASK1 / mice showed similar otypes. The expression of ASK1 was elevated in tumor sections cell cycle status to cells from WT mice (Fig. S2 A, D, and E). These of WT mice, compared to nontumor epithelium (Fig. 2E and results suggested that ASK1 silencing altered the rate of pro- Fig. S1A), as observed in human gastric tissue specimens. In liferation in GC cells, but not in normal cells without malignant contrast, in murine colitis-associated tumors analyzed previously transformation.

Hayakawa et al. PNAS | January 11, 2011 | vol. 108 | no. 2 | 781 Downloaded by guest on September 29, 2021 Fig. 3. ASK1 is important for cell proliferation and cell cycle progression in gastric cancer cells. (A) Immunoblot analysis of ASK1 levels in AGS, MKN45, MKN74, and TMK1 cells 48 h after transfection with NCsiRNA or ASK1siRNA. (B) The number of AGS cells counted 24, 48, and 72 h following transfection with the indicated amount of ASK1siRNA. Data are plotted as means ± SD, *P < 0.05, compared to NCsiRNA. (C) Numbers of AGS, MKN45, MKN74, and TMK1 cells counted at 48 and 72 h following transfection. Cells were trans- fected with 100 nM of NC or ASK1siRNA. Data are plotted as means ± SD, *P < 0.05, compared to NCsiRNA. (D) AGS cells were transfected with NC or ASK1siRNA and cell cycle analysis performed by flow cytometry. Represen- tative data are shown and the different phases of the cell cycle are indicated. (E) Percentage of cells in each phase of the cell cycle. Results represent the mean ± SD of triplicate samples. *P < 0.05, compared to NCsiRNA. Fig. 4. ASK1 up-regulates cyclin D1 transcription via AP-1 activation. (A) Scatterplot analysis of cDNA extracted from AGS cells transfected with NCsiRNA and ASK1siRNA. The indicated lines of the graph are fold-change ASK1 Up-Regulates Cyclin D1 Transcription Through AP-1 Activation. lines (fourfold cutoff). (B) Immunoblot analysis of the indicated proteins in To investigate the mechanism of cell cycle regulation, we ex- AGS, MKN45, MKN74, and TMK1 cells following transfection with NC or amined the expression of cell cycle-associated genes in AGS cells ASK1siRNA. (C) Immunoblot analysis of the indicated proteins in AGS and transfected with NC or ASK1siRNA. Among the 84 genes ex- MKN74 cells infected with the adenoviruses encoding LacZ, ASK-WT, or ASK- amined, cyclin D1, a known regulator of the G1-to-S phase KM for 24 h. (D and E) AGS cells were transfected with PCDNA, ASK-WT, or transition (15), was significantly down-regulated by ASK1siRNA ASK-KM together with pD1luc (D) or AP-1luc (E) for 24 h. The ratio of re- transfection, whereas both HERC5 and Bcl-2 were up-regulated porter luciferase activity to the Renilla luciferase control is indicated. Results A represent the mean ± SD of triplicate samples. (F) AGS cells were transfected (Fig. 4 and Table S1; fourfold cutoff). When we examined the with wild-type pD1luc (AP-1 wt) or the site-directed mutant of the AP-1 site effect of ASK1 knockdown in GC cells, we found reduced cyclin (AP-1 mt) together with ASK-WT or PCDNA alone for 24 h. Results represent D1 protein levels in addition to reduced expression of c-Jun and the mean ± SD of triplicate samples. (G) Immunoblot analysis of the ASK1 phosphorylated JNK, with no effect on cyclins A, E, or H protein and cyclin D1 levels in the indicated cells 48 h after transfection with NC or levels (Fig. 4B). These data were consistent with the results ASK1siRNA. BxPC-3 and Capan-1 are pancreatic cancer cell lines; HCT15, obtained from the MNU-treated mouse gastric tumor that revealed HT29, and colo205 are colon cancer cell lines. decreased cyclin D1 and proliferating cell nuclear antigen (PCNA) − − levels in tumors from ASK1 / mice (Fig. S2 F and G), but not in -containing AP-1luc. As shown in Fig. 4 D and E, ASK1- control mice (Fig. S2A). To confirm the regulation of cyclin D1 WT enhanced both the cyclin D1 promoter and AP-1 transcription expression by ASK1, we overexpressed ASK1 using recombinant adenoviruses and expression plasmids (6, 16, 17). Immunoblot factor in AGS cells. Further, kinase-mutant ASK1 (ASK1-KM) analysis showed increased c-Jun and cyclin D1 expression levels in failed to enhance these luciferase activities in addition to the c-Jun C D E cells infected with the adenoviruses encoding WT ASK1 (ASK- and cyclin D1 protein levels (Fig. 4 , ,and ), indicating that WT) compared to those infected with β-galactosidase (LacZ) in ASK1 regulates both AP-1 activity and cyclin D1 expression in a the AGS and MKN74 GC cell lines (Fig. 4C). kinase-dependent manner. Mutagenesis of the cyclin D1 promoter JNK is a critical signal transducer of the AP-1 transcription at the AP-1 binding site abolished the enhanced promoter activity factor, and AP-1 activates the cyclin D1 promoter via an AP-1 induced by ASK1 overexpression (Fig. 4F), ensuring that ASK1 binding site. To examine the role of ASK1 in this cyclin D1 trans- regulates cyclin D1 transcription via AP-1 activation. activation, we performed a luciferase assay using the cyclin D1 To examine whether the regulation of cyclin D1 protein and promoter-containing construct pD1luc (18) and consensus AP-1 cell proliferation by ASK1 is specific to GC, we used several cell

782 | www.pnas.org/cgi/doi/10.1073/pnas.1011418108 Hayakawa et al. Downloaded by guest on September 29, 2021 lines derived from other tumor types. In contrast to gastric ylation of Rb by CDKs inhibits Rb-E2F binding, allowing E2F cancer cells, most of the colon and pancreatic cancer cell lines activation and cell cycle progression (27), whereas the phos- did not show reduced cyclin D1 expression by ASK1 knockdown phorylation of Rb at Ser780 requires cyclin D1 (28, 29). (Fig. 4G). We found that ASK1 silencing reduced cyclin D1 Because ASK1 has been reported to be a direct target of protein levels in some cancer cell lines, such as ASPC-1 and the E2F transcription factor (30, 31), we investigated whether DLD-1 (Fig. S3A), but these cell lines did not show enhanced ASK1 expression was itself regulated by cyclin D1. We observed cyclin D1 levels on ASK1 overexpression (Fig. S3B). Addition- that following stimulation with EGF, ERK phosphorylation was ally, ASK1 silencing did not affect cell proliferation in these induced in both NCsiRNA- and cyclin D1-specific siRNA colon, pancreatic, or lung cancer cell lines (Fig. S3C), suggesting (CyD1siRNA)-transfected cells, but the increase in cyclin D1 and that the ASK1-dependent regulation of the cyclin D1 level and Ser780 phosphorylation of Rb was induced only in control cells cellular proliferation is specific to GC. (Fig. 5C). Additionally, we found that the expression of ASK1 was elevated following EGF treatment in control cells, but not in cyclin Positive Feedback Regulation of ASK1 by Cyclin D1. Epidermal D1-silenced cells (Fig. 5C and Fig. S4A), whereas the expression of growth factor (EGF) regulates cell growth, proliferation, and another major MAP3K, TGF-β–activated kinase 1 (TAK1), was differentiation, and plays an important role in GC or precan- not affected. To further confirm this observation, we overex- cerous lesions (19–21). EGF activates both the extracellular pressed cyclin D1 in these cells. Overexpression of cyclin D1 en- signal-regulated (ERKs) and JNKs (22–24) and induces hanced the phosphorylation of Rb, followed by increased ASK1 cyclin D1 transcription through AP-1 activation (25, 26). When expression (Fig. 5D), suggesting that the regulation of ASK1 ex- AGS cells were treated with EGF, ASK1 phosphorylation was pression was cyclin D1 dependent. enhanced, suggesting that EGF stimulates ASK1 activation (Fig. Finally, we investigated the role of ASK1 and cyclin D1 in ASK1 5A). EGF stimulation also induced JNK phosphorylation and promoter activation. As expected, cyclin D1 overexpression in- increased cyclin D1 expression and phosphorylation of the reti- duced E2F transcriptional activation (Fig. S4B). A chromatin im- noblastoma protein, Rb, in NCsiRNA transfected cells, but not munoprecipitation (ChIP) assay showed that cyclin D1 expression in ASK1siRNA transfected cells (Fig. 5B). increased E2F1 binding to the ASK1 promoter (Fig. 5E). Fur- Rb regulates cell cycle progression, cooperating with the cyclin thermore, treatment of cells with an adenovirus encoding a con- and the cyclin-dependent kinase (CDK) family (15). Phosphor- stitutively active mutant of ASK1 (ASK-ΔN) (16) enhanced JNK phosphorylation and cyclin D1 expression, resulting in an increase in endogenous ASK1 (Fig. 5F). In the ChIP assay, E2F1 binding to the ASK1 promoter was increased in cells expressing exogenous ASK1 (Fig. 5G). Collectively, these data indicate that ASK1 reg- ulates the expression of cyclin D1, which, in turn, regulates the expression of ASK1 in GC through the Rb–E2F pathway (Fig. 5H). Discussion Although the role of ASK1 in carcinogenesis has been reported, most studies have focused on ASK1-mediated apoptosis or in- flammatory responses, rather than cellular proliferation. Indeed, several in vivo analyses revealed that ASK1 is a critical regulator of apoptosis in liver, heart, and nerve tissue and a regulator of in- flammation in the skin and colon (10–13, 32). This study showed that the expression of ASK1 was elevated in human GC tissue. Furthermore, we demonstrated that ASK1 is important for gastric tumorigenesis in mice and regulates cell proliferation and cell cycle progression through controlling cyclin D1 expression in GC cells. These results suggest that ASK1 plays an important role in GC that is not mediated through either apoptosis or inflammatory effects. MAPKs are important signaling components that convert ex- tracellular stimuli into a wide variety of cellular responses in- cluding proliferation, differentiation, and apoptosis (33). MAPK pathways consist of three classes of protein kinases: MAPK, MAP2K, and MAP3K. MAP3Ks provide the specificity for stim- ulus-dependent activation of downstream kinases (34). To date, 21 MAP3Ks are known to activate MAP2Ks, but these have rarely Fig. 5. Positive feedback regulation of ASK1 by cyclin D1. (A) AGS cells were been investigated in the pathogenesis of GC. BRAF, which infected with ASK-WT adenovirus for 24 h and stimulated with EGF (100 nM). Immunoblot analysis of phospho-ASK1 and total ASK1 are shown. (B and C) transduces mitogenic stimuli from Ras leading to ERK1/2 acti- Immunoblot analysis of the indicated proteins in AGS cells transfected with vation, is reportedly mutated in 2% of GC cases (35). JNKs reg- NC or ASK1siRNA (B) or AGS cells transfected with NC or CyD1siRNA (C). Cells ulate multiple cellular functions including carcinogenesis, through were stimulated with EGF (100 nM) for the indicated times. (D) Immunoblot activation of the AP-1 transcription factor. We previously repor- analysis of the indicated proteins in AGS cells transfected with PCDNA or ted that the JNK signaling pathway is critical in the development of cyclin D1-expressing vector. (E) ChIP assay of AGS cells transfected with GCs (4), whereas its mechanism of activation remains unclear. MEDICAL SCIENCES PCDNA or cyclin D1-expressing vector. PCR amplification products of E2F1- Thus, we examined the function of ASK1 in GC cells as one of the ChIP and input DNA using ASK1 promoter primers are shown. (F) Immunoblot MAP3Ks that possesses the ability to activate the JNK pathway. analysis of the indicated proteins in AGS cells infected with adenoviruses In this study, we found that ASK1 silencing reduced JNK phos- encoding LacZ or ASK-ΔN for 24 h. (G) ChIP assay of AGS cells infected with adenoviruses encoding LacZ, ASK-WT, or ASK-ΔN for 24 h. The PCR amplifi- phorylation, indicating that ASK1 is a critical regulator of JNK cation products of E2F1-ChIP and input DNA using ASK1 promoter primers signaling in GC. When we compared the effects of ASK1 inhibition are shown. ASK1 overexpression was confirmed by the PCR products of input with the JNK1- or JNK2-mediated inhibition of cellular pro- DNA using ASK1 cDNA primers. (H) Proposed model of the feedback regu- liferation, ASK1 silencing resulted in greater inhibition of cellular lation of ASK1 through the cyclin D1-dependent pathway in gastric cancer. proliferation than either JNK1 or JNK2 silencing (Fig. S5A). In

Hayakawa et al. PNAS | January 11, 2011 | vol. 108 | no. 2 | 783 Downloaded by guest on September 29, 2021 GC cells, ASK1siRNA was able to reduce both JNK1 and JNK2 pylori-induced gastric tumorigenesis, a long-term infection model activation (Fig. S5B), suggesting that ASK1 mediates proliferation may be required. through both these kinases. In summary, we demonstrated that ASK1 regulated the ex- There is a wealth of evidence showing how the disturbance of pression of cyclin D1, which, in turn, regulated ASK1 expression specific cyclins plays an important role in many types of human in a positive feedback system in GC cells. These results indicate cancer. In GC, the cyclin E gene is amplified in 15–20% of an essential role of ASK1 in gastric carcinogenesis and suggest gastric cancers, whereas cyclin B2 or cyclin C is also up-regulated the potential of specific ASK1-targeting therapies for GC. (36). Amplification or mutation of the cyclin D1 gene has not been reported in GC, although elevated cyclin D1 expression has Materials and Methods − − been observed in ≈50% of GC cases (37). In GC, several factors Mice and Gastric Tumorigenesis Model. The generation of ASK1 / mice has − − are reportedly involved in cyclin D1 overexpression, including been described previously (8). ASK1 / mice were backcrossed into the −/− the RAF/MEK/ERK pathways (38), abrogation of TGF-β sig- C57BL/6 strain at least 14 times and both ASK1 mice and C57BL/6 WT mice naling (39), gastrin (40, 41), Epstein-Barr virus (42), and H. pylori (Clea Japan) were used in the study. (43). However, the precise mechanism that regulates cyclin D1 All of the experimental protocols were approved by the ethics committee for animal experimentation and conducted in accordance with the Guidelines expression in GC remains unclear. In this report, we describe a – for the Care and Use of Laboratory Animals of the Graduate School of unique mechanism of cyclin D1 overexpression through an ASK1 Medicine, the University of Tokyo, and the Institute for Adult Diseases, Asahi cyclin D1 feedback loop. Life Foundation. − − The associations between ASK1, cyclin D1, and cellular pro- For the gastric tumorigenesis model, 6-wk-old WT and ASK1 / mice were liferation seemed specific to GC cells, because ASK1 knockdown given drinking water containing 240 ppm of MNU (Sigma Chemicals) on in other cell lines did not inhibit proliferation and ASK1 over- alternate weeks for a total of 5 wk of exposure, as described previously (53). expression did not increase cyclin D1 protein levels. We found Forty weeks after beginning MNU administration, the stomach was re- increased levels of ASK1 caused by cyclin D1 overexpression in moved, opened along the greater curvature, and stained with indigo car- A mine. The number and long diameter of tumors in the stomach were many cell types (Fig. S6 ), which could be explained by the ef- > fect of E2F on the ASK1 promoter, as reported previously in measured under a dissecting microscope. Tumors 0.5 mm in diameter were mapped and counted in a blinded fashion. colon and melanoma cell lines (30, 31). Thus, the enhancement fi of cyclin D1 and cell proliferation by ASK1 is speci c to GC. The Human Tissue Samples. Clinical tissue specimens from patients were obtained discrepancies between the MNU-induced gastric tumorigenesis from the archives of University of Tokyo Hospital and Motojima General model in this study and the colitis-associated colon cancer model Hospital with the approval of the medical ethics committee and with informed in which we previously demonstrated that ASK1 deficiency pro- consent. moted tumorigenesis may also demonstrate the specificity of the ASK1–cyclin D1 feedback loop in gastric tumorigenesis. In other Cell Lines. The human cancer cell lines were cultured in Ham’s F-12, DMEM, or types of cancer, the mutation of specific genes, such as APC or RPMI medium, supplemented with 10–20% FBS. Cell numbers were de- β-catenin in colon cancer or K-ras in pancreatic cancer, has been termined using a Cell Counting Kit-8 according to the manufacturer’s pro- reported to lead to cyclin D1 overexpression and carcinogenesis tocol (Dojindo Laboratories). (44–47). Although no major mutation leading to cyclin D1 Cell Cycle Analysis by Flow Cytometry. The proportion of cells in either G0/G1, overexpression has been reported in GC, the ASK1-mediated fl H S, or G2/M phase was determined by ow cytometric analysis of DNA positive feedback loop containing cyclin D1 (Fig. 5 ) may play content. Briefly, cells obtained by trypsinization were washed with PBS and a role in tumorigenesis equivalent to the genetic alteration ob- incubated in propidium iodide solution (50 mg/mL in PBS) for 30 min. The served in other types of cancer, suggesting that ASK1 may be cells were then analyzed for cell cycle status using the Guava EasyCyte Plus a potential therapeutic target for GC. (Guava Technologies). We used a MNU model of murine gastric tumorigenesis in this study, although there are many other models that can be used to ChIP Assays. Chromatin immunoprecipitation was performed essentially as evaluate the role of host genetics. For example, some studies described (54). Antibodies to E2F1 (sc-193; Santa Cruz) were used to pre- used Helicobacter as a driver of inflammation and found impor- cipitate chromosomal DNA, using cross-linked chromatin prepared from exponentially growing cells. The immunoprecipitated DNA was analyzed by tant cytokines, signaling molecules, and potential stem cells that ′ act in gastric tumorigenesis (48–52). Because the MNU model PCR using the ASK1 promoter region primers 5 -GAGTGGGTGGCCAGAAGC and 5′-CGGAGCTTCCTTTTCTTGGC. induces little inflammation, we think that this model is useful for analyzing the pure proliferative effect of ASK1 in gastric epi- Statistical Analyses. The differences between means were compared using thelium. We also investigated the role of ASK1 in cell pro- Student’s t-testortheWilcoxontest.P values < 0.05 were considered liferation caused by H. pylori infection in a short-term model. H. significant. pylori infection induced gastric cellular hyperplasia in infected Please see SI Materials and Methods for details. WT mice. Increased ASK1 and cyclin D1 levels were also ob- −/− served following infection. In ASK1 mice, gastric hyperplasia ACKNOWLEDGMENTS. We thank Dr. Takaaki Sano and Dr. Teiji Motojima and cyclin D1 up-regulation were also observed in infected mice, (Division of Pathology and Abdominal Surgery, Motojima General Hospital, but these changes were not as prominent as in WT mice (Fig. S6 B Gumma, Japan) for providing the gastric and colorectal carcinoma speci- C fi H mens. We thank Mitsuko Tsubouchi for technical assistance. S.M. was and ). These ndings suggest the involvement of ASK1 in . supported by a grant-in-aid from the Japanese Ministry of Education, pylori-induced gastritis. To determine the role of ASK1 in H. Culture, Sports, Science, and Technology (19390205).

1. Tsugane S, Sasazuki S (2007) Diet and the risk of gastric cancer: Review of epide- 6. Nishitoh H, et al. (1998) ASK1 is essential for JNK/SAPK activation by TRAF2. Mol Cell miological evidence. Gastric Cancer 10:75–83. 2:389–395. 2. Jakszyn P, et al. (2006) Endogenous versus exogenous exposure to N-nitroso 7. Matsuzawa A, et al. (2005) ROS-dependent activation of the TRAF6-ASK1-p38 compounds and gastric cancer risk in the European Prospective Investigation into pathway is selectively required for TLR4-mediated innate immunity. Nat Immunol 6: Cancer and Nutrition (EPIC-EURGAST) study. Carcinogenesis 27:1497–1501. 587–592. 3. Uemura N, et al. (2001) Helicobacter pylori infection and the development of gastric 8. Tobiume K, et al. (2001) ASK1 is required for sustained activations of JNK/p38 MAP cancer. N Engl J Med 345:784–789. kinases and apoptosis. EMBO Rep 2:222–228. 4. Shibata W, et al. (2008) c-Jun NH2-terminal kinase 1 is a critical regulator for the 9. Nishitoh H, et al. (2002) ASK1 is essential for endoplasmic reticulum stress-induced neuronal development of gastric cancer in mice. Cancer Res 68:5031–5039. cell death triggered by expanded polyglutamine repeats. Genes Dev 16:1345–1355. 5. Ichijo H, et al. (1997) Induction of apoptosis by ASK1, a mammalian MAPKKK that 10. Yamaguchi O, et al. (2003) Targeted deletion of apoptosis signal-regulating kinase 1 activates SAPK/JNK and p38 signaling pathways. Science 275:90–94. attenuates left ventricular remodeling. Proc Natl Acad Sci USA 100:15883–15888.

784 | www.pnas.org/cgi/doi/10.1073/pnas.1011418108 Hayakawa et al. Downloaded by guest on September 29, 2021 11. Nakagawa H, et al. (2008) Deletion of apoptosis signal-regulating kinase 1 attenuates 32. Harada C, et al. (2006) Role of apoptosis signal-regulating kinase 1 in stress-induced acetaminophen-induced liver injury by inhibiting c-Jun N-terminal kinase activation. neural cell apoptosis in vivo. Am J Pathol 168:261–269. Gastroenterology 135:1311–1321. 33. Wagner EF, Nebreda AR (2009) Signal integration by JNK and p38 MAPK pathways in 12. Hayakawa Y, et al. (2010) Apoptosis signal-regulating kinase 1 regulates colitis and cancer development. Nat Rev Cancer 9:537–549. colitis-associated tumorigenesis by the innate immune responses. Gastroenterology 34. Craig EA, Stevens MV, Vaillancourt RR, Camenisch TD (2008) MAP3Ks as central 138:1055–1067. regulators of cell fate during development. Dev Dyn 237:3102–3114. 13. Iriyama T, et al. (2009) ASK1 and ASK2 differentially regulate the counteracting roles 35. Lee SH, et al. (2003) BRAF and KRAS mutations in stomach cancer. Oncogene 22: of apoptosis and inflammation in tumorigenesis. EMBO J 28:843–853. 6942–6945. 14. Tatematsu M, et al. (1993) Induction of glandular stomach cancers in C3H mice 36. Stock M, Otto F (2005) Gene deregulation in gastric cancer. Gene 360:1–19. treated with N-methyl-N-nitrosourea in the drinking water. Jpn J Cancer Res 84: 37. Arber N, et al. (1999) Overexpression of cyclin D1 occurs in both squamous carcinomas 1258–1264. and adenocarcinomas of the esophagus and in adenocarcinomas of the stomach. 15. Sherr CJ (1996) Cancer cell cycles. Science 274:1672–1677. Hum Pathol 30:1087–1092. 16. Saitoh M, et al. (1998) Mammalian is a direct signal- 38. Wang X, et al. (2009) RAF may induce cell proliferation through hypermethylation of regulating kinase (ASK) 1. EMBO J 17:2596–2606. tumor suppressor gene promoter in gastric epithelial cells. Cancer Sci 100:117–125. 17. Takeda K, et al. (2007) Apoptosis signal-regulating kinase (ASK) 2 functions as 39. Han SU, et al. (2004) Loss of the Smad3 expression increases susceptibility to a mitogen-activated protein kinase kinase kinase in a heteromeric complex with tumorigenicity in human gastric cancer. Oncogene 23:1333–1341. ASK1. J Biol Chem 282:7522–7531. 40. Pradeep A, et al. (2004) Gastrin-mediated activation of cyclin D1 transcription involves 18. Müller H, et al. (1994) Cyclin D1 expression is regulated by the retinoblastoma beta-catenin and CREB pathways in gastric cancer cells. Oncogene 23:3689–3699. protein. Proc Natl Acad Sci USA 91:2945–2949. 41. Song DH, et al. (2003) Gastrin-induced gastric adenocarcinoma growth is mediated 19. Sugiyama K, Yonemura Y, Miyazaki I (1989) Immunohistochemical study of epidermal through cyclin D1. Am J Physiol Gastrointest Liver Physiol 285:G217–G222. growth factor and epidermal growth factor receptor in gastric carcinoma. Cancer 63: 42. Nakagawa H, et al. (1997) The targeting of the cyclin D1 oncogene by an Epstein-Barr 1557–1561. virus promoter in transgenic mice causes dysplasia in the tongue, esophagus and 20. Filipe MI, et al. (1995) Expression of transforming growth factor alpha, epidermal forestomach. Oncogene 14:1185–1190. growth factor receptor and epidermal growth factor in precursor lesions to gastric 43. Hirata Y, et al. (2001) Helicobacter pylori activates the cyclin D1 gene through carcinoma. Br J Cancer 71:30–36. mitogen-activated protein kinase pathway in gastric cancer cells. Infect Immun 69: 21. Messa C, Russo F, Pricci M, Di Leo A (2000) Epidermal growth factor and 17beta- 3965–3971. estradiol effects on proliferation of a human gastric cancer cell line (AGS). Scand J 44. Kinzler KW, Vogelstein B (1996) Lessons from hereditary colorectal cancer. Cell 87: Gastroenterol 35:753–758. 159–170. 22. Minden A, et al. (1994) Differential activation of ERK and JNK mitogen-activated 45. Morin PJ, et al. (1997) Activation of beta-catenin-Tcf signaling in colon cancer by protein kinases by Raf-1 and MEKK. Science 266:1719–1723. mutations in beta-catenin or APC. Science 275:1787–1790. 23. Westwick JK, Weitzel C, Minden A, Karin M, Brenner DA (1994) 46. Hahn SA, Schmiegel WH (1998) Recent discoveries in cancer genetics of exocrine alpha stimulates AP-1 activity through prolonged activation of the c-Jun kinase. J Biol pancreatic neoplasia. Digestion 59:493–501. Chem 269:26396–26401. 47. Ijichi H, et al. (2006) Aggressive pancreatic ductal adenocarcinoma in mice caused by 24. Minden A, et al. (1994) c-Jun N-terminal phosphorylation correlates with activation of pancreas-specific blockade of transforming growth factor-beta signaling in cooperation the JNK subgroup but not the ERK subgroup of mitogen-activated protein kinases. with active Kras expression. Genes Dev 20:3147–3160. Mol Cell Biol 14:6683–6688. 48. Houghton J, et al. (2004) Gastric cancer originating from bone marrow-derived cells. 25. Herber B, Truss M, Beato M, Müller R (1994) Inducible regulatory elements in the Science 306:1568–1571. human cyclin D1 promoter. Oncogene 9:1295–1304. 49. Lee CW, et al. (2009) Combination of sulindac and antimicrobial eradication of 26. Albanese C, et al. (1995) Transforming p21ras mutants and c-Ets-2 activate the cyclin Helicobacter pylori prevents progression of gastric cancer in hypergastrinemic INS- D1 promoter through distinguishable regions. J Biol Chem 270:23589–23597. GAS mice. Cancer Res 69:8166–8174. 27. Knudsen ES, Wang JY (1997) Dual mechanisms for the inhibition of E2F binding to RB 50. Shibata W, et al. (2010) Conditional deletion of IkappaB-kinase-beta accelerates by cyclin-dependent kinase-mediated RB phosphorylation. Mol Cell Biol 17:5771– helicobacter-dependent gastric apoptosis, proliferation, and preneoplasia. 5783. Gastroenterology 138:1022–1034. 28. Lundberg AS, Weinberg RA (1998) Functional inactivation of the retinoblastoma 51. Tu S, et al. (2008) Overexpression of interleukin-1beta induces gastric inflammation protein requires sequential modification by at least two distinct cyclin-cdk complexes. and cancer and mobilizes myeloid-derived suppressor cells in mice. Cancer Cell 14: Mol Cell Biol 18:753–761. 408–419. 29. Geng Y, et al. (2001) Deletion of the p27Kip1 gene restores normal development in 52. Goldenring JR, Nam KT, Wang TC, Mills JC, Wright NA (2010) Spasmolytic cyclin D1-deficient mice. Proc Natl Acad Sci USA 98:194–199. polypeptide-expressing metaplasia and intestinal metaplasia: Time for reevaluation 30. Dong YB, et al. (2007) Induction of apoptosis signal-regulating Kinase 1 by E2F-1 may of metaplasias and the origins of gastric cancer. Gastroenterology 138:2207–2210. not be essential for E2F-1-mediated apoptosis in melanoma cells. Tumour Biol 28: 53. Sakamoto K, et al. (2010) Inhibitor of kappaB kinase beta regulates gastric 111–122. carcinogenesis via interleukin-1alpha expression. Gastroenterology 139:226–238. 31. Tan J, et al. (2006) Apoptosis signal-regulating kinase 1 is a direct target of E2F1 and 54. Hershko T, Korotayev K, Polager S, Ginsberg D (2006) E2F1 modulates p38 MAPK contributes to histone deacetylase inhibitor-induced apoptosis through positive phosphorylation via transcriptional regulation of ASK1 and Wip1. J Biol Chem 281: feedback regulation of E2F1 apoptotic activity. J Biol Chem 281:10508–10515. 31309–31316. MEDICAL SCIENCES

Hayakawa et al. PNAS | January 11, 2011 | vol. 108 | no. 2 | 785 Downloaded by guest on September 29, 2021