BGJ398

Proliferation of poorly differentiated endometrial cancer cells through autocrine activation of FGF receptor and HES1 expression

Abstract

Patients with poorly differentiated endometrial cancer show poor prognosis, and effective molecular target-based therapies are needed. Endometrial cancer cells proliferate depending on the activation of HES1 (hairy and enhancer of split-1), which is induced by several pathways, such as the Notch and fibroblast growth factor receptor (FGFR) signaling pathways. In addi- tion, aberrant, ligand-free activation of the FGFR signaling pathway resulting from mutations in FGFR2 was also reported in endometrial cancer. However, a clinical trial showed that there was no difference in the effectiveness of FGFR inhibitors between patients with and without the FGFR2 mutation, suggesting a presence of another signaling pathway for the FGFR activation. Here, we investigated the signaling pathway regulating the expression of HES1 and proliferation of poorly and well-differentiated endometrial cancer cell lines Ishikawa and HEC-50B, respectively. Whereas Ishikawa cells proliferated and expressed HES1 in a Notch signaling-dependent manner, Notch signaling was not involved in HES1 and proliferation of HEC-50B cells. The FGFR inhibitor, NVP-BGJ398, decreased HES1 expression and proliferation of HEC-50B cells; however, HEC50B cells had no mutations in the FGFR2 gene. Instead, HEC-50B cells highly expressed ligands for FGFR2, suggesting that FGFR2 is activated by an autocrine manner, not by ligand-free activation. This autocrine pathway activated Akt downstream of FGFR for cell proliferation. Our findings suggest the usefulness of HES1 as a marker for the proliferation signaling and that FGFR inhibitor may be effective for poorly differentiated endometrial cancers that harbor wild-type FGFR.

Keywords : HES1 · FGF/FGFR · FGFR2 · Endometrial cancer · NVP-BGJ398

Introduction

Endometrial cancer is one of the most common gynecologi- cal malignancies, with an increasing morbidity worldwide [1]. In 2013 in United States, a total of 49,560 patients were diagnosed with endometrial cancer and 8190 patients died from endometrial cancer [2]. Most poorly differentiated endometrial cancer is insensitive to radiation therapy or cur- rent chemotherapy and thus surgical removal is currently the only effective therapy. The insensitivity of endometrial tumors to radiation and chemotherapy has led to a low 5-year survival rate for patients [3–5]. Better understanding of the mechanisms of endometrial cancer may help lead to the development of effective therapies.
Hairy and enhancer of split-1 (HES1), a basic helix-loop- helix transcriptional factor protein, plays an important role in cell differentiation, apoptosis and self-renewal of various types of cells [6, 7]. Recent studies have revealed upregu- lation of HES1, which was stimulated by aberrant Notch signaling or others, resulting in enhanced cell proliferation in several types of tumors, including pancreatic, renal, colon, ovarian and endometrial cancers [8–14]. Moreo- ver, downregulation of HES1 expression led to inhibition of cell proliferation and migration [15–18]. Notably, both Notch1 and HES1 show upregulation in endometrial can- cer tissue compared with normal endometrium [19–22]. We also previously reported that inhibition of the Notch signaling decreased cell proliferation through the decrease in the expression of HES1 in Ishikawa cells, showing the involvement of these factors in the proliferation of the well- differentiated endometrial cancer [23].

Fibroblast growth factors (FGFs) and their receptors (FGFRs) consist of 18 and 7 subtypes, respectively, which show distinct binding specificities. It is known that FGFR signaling regulates development, cell proliferation, migra- tion and angiogenesis [24, 25]. FGF binding induces FGFR receptor dimerization, which activates the intracellular kinase domain of FGFRs and phosphorylated FGFR phos- phorylated substrate 2 (FRS2). The phosphorylation of FRS2 activates at least two intracellular signaling pathways: Ras/extracellular signal-regulated kinase (RAS/ERK) and phosphoinositide 3-kinase (PI3K/AKT) signaling [24, 25]. In normal endometrium, FGF2 is mainly expressed through- out the menstrual cycle in stromal cells and blood vessels, and the FGFR2b isoform is expressed in normal, hyper- plastic and tumoral glandular epithelial cells [26, 27]. In some endometrial tumors, mutations of FGFR2 result in its constitutive activation in a ligand-free manner [24–30]. For instance, the S252W mutation, the most common mutation in the ligand-binding domain, results in increased affinity to ligands. The S373C and Y376C mutations in exon 9 in the internal domain lead to constitutive receptor dimerization without ligand binding, while the N550K mutation in exon 12 in the kinase domain leads to constitutive kinase activ- ity. These mutations result in aberrant activation of FGFR signaling and leads to enhancement of cell proliferation and migration in endometrial cancer [31]. Notably, HES1 expres- sion is also regulated by FGF and FGFR. Nakayama et al. reported that HES1 mRNA expression was regulated by the FGF signaling in mesenchymal C3H10T1/2 cells, and an inhibitor against ERK, a downstream molecule of the FGF signaling, could decrease HES1 mRNA expression [32].

Despite the extensive characterization of endometrial cancer, molecular target-based therapies have not suc- cessfully been developed. In vitro studies showed higher sensitivity of mutant FGFR2 cell lines to FGFR inhibitors, e.g., dovitinib, PD173074 and ponatinib, suggesting that FGFR inhibitors would selectively inhibit the proliferation of cancer cells harboring mutations in the FGFR2 gene [31, 33]. However, in the clinical trial of dovitinib, there was no difference in the objective response rate or overall survival rates between patients with and without FGFR2 mutations [31]. These results suggest the possibility that there is another mechanism, in which FGFR2 is activated even without the mutation. In addition, there is no effec- tive marker to predict the sensitivity to FGFR inhibitors.In this report, we investigated the signaling pathway that regulates HES1 in a poorly differentiated endometrial cancer cell line, HEC-50B.

Materials and methods
Cell lines

The endometrial cancer cell lines Ishikawa 3-H-12 (well- differentiated) and HEC-50B (poorly differentiated) were purchased from Japanese Collection of Research Bioresources (JCRB1505 and JCRB1145, respectively). The cells were cultivated in Eagle’s minimum essential medium (EMEM, MP Biomedicals, Santa Ana, CA, USA) containing 10% fetal bovine serum (FBS, Biowest, Nuaillé, Maine-et-Loire, France) at 37 °C under 5% CO2. Before each experiment, cells were incubated in EMEM with 10% charcoal-dextran-treated FBS (Cell Culture Technology, Ticino, Switzerland) at 37 °C for 24 h.

Drug treatment

Basic FGF (20 ng/mL, Peprotech, Rocky Hill, NJ) was dissolved in EMEM. The selective FGFR inhibitor NVP- BGJ 398 (0.01–10 µM, Selleck Chemicals, Houston, TX, USA), the γ-secretase inhibitor DAPT (N-[N-(3, 5-dif- luorophenacetyl-L-alanyl)]-S-phenylglycine t-butyl ester; 15 µM, Sigma–Aldrich, St Louis, MO, USA), the MEK inhibitor U0126 (1–20 µM, Cell Signaling Technology, Danvers, MA, USA) and the Akt inhibitor SH-5 (1–20 µM, Abcam, Cambrigeshire, England) were initially dissolved in dimethyl sulfoxide (DMSO; Dojindo Molecular Tech- nologies, Kumamoto, Japan) at higher concentrations and then diluted with EMEM in the presence or absence of FGF. An equivalent volume of DMSO was used as control.

Cell proliferation assay

Cells were seeded into 96-well culture plates (BD Bio- sciences, Franklin Lakes, NJ, USA) at a density of 3.0 × 104/well for 24 h. The culture media were replaced with media containing the appropriate inhibitor. Cell viabilities were measured with the Cell Counting Kit-8 (Dojindo Molecular Technologies) at 0, 12, 24 and 72 h after the addition of inhibitors.

RNA extraction and reverse‑transcription PCR

Total RNA was extracted using TRIZOL reagent (Thermo Fisher Scientific, Waltham, MA, USA). Reverse transcrip- tion was performed using the SuperScriptIII Reverse Tran- scriptase (Thermo Fisher Scientific), according to the manu- facturer’s instructions. The expressions of FGFR2, FGF1, FGF2, Notch1, Notch4 and GAPDH mRNAs were analyzed using EmeraldAmp™ PCR Master Mix (Takara Bio, Shiga, Japan) and HES1 mRNA was evaluated using real-time PCR with LightCycler FastStart DNA MasterPLUS SYBR Green I (Roche, Basel, Switzerland), according to the manufactur- ers’ instructions. The primer pairs used for each gene were as follows: human HES1, 5′-GAA GAA AGA TAG CTC GCG GCA TTC CAA G -3′ (forward) and 5′-GTC ACC TCG TTC ATG CAC TCG CTG A-3′ (reverse); human FGFR2, 5′-CTT CCT CTC GTT CCC CAA AT-3′ (forward) and 5′-GAC CAG GCA GAT GAA ACG AC-3′ (reverse); human FGF1, 5′ -CCA GCA CAT TCA GCT GCA GCTCAG -3′ (forward) and 5′-CTT TCT GGC CAT AGT GAG TC-3′ (reverse); human FGF2, 5′-GAG GAG TTG TGT CTA TCA AAG-3′ (forward) and 5′-GTT CGT TTC AGT GCC ACA TAC C-3′ (reverse); human Notch1, 5′-ACG AGG ACC TGG AGA CCA AGA AGT TC-3′ (forward) and 5′-GAT CAG GAT CTG GAA GAC ACC TTG TG-3′ (reverse); human Notch4, 5′-CTG AGC CAA GGC ATA GAC GTC TCT TC-3′ (forward) and 5′-CAC ACT GGC AGA GAT ACC CAC TG-3′ (reverse); human GAPDH, 5′-GCT TGT CAT CAA TGG AAA TCC C-3′ (forward) and 5′-TTC ACA CCC ATG ACG AAC ATG-3′ (reverse). The PCR protocol to evaluate mRNA expression was 35 cycles of denaturation at 95 °C for 30 s, annealing at 60 °C for 30 s, and extension at 72 °C for 60 s. The real-time PCR protocol was denaturation at 96 °C for 3 m, annealing at 60 °C for 15 s, and extension at 72 °C for 60 s, respectively. GAPDH was used as the internal control. Image J was used for estimation of intensity [34].

Immunoblotting

Cell homogenates (50 µg) prepared in HEPES buffer (Nacalai Tesque, Kyoto, Japan) were separated on an acrylamide gel (12%). Proteins were then transferred to PVDF membranes (GE Healthcare Lifescience, Little Istanbul, UK), and membranes were incubated with the fol- lowing primary antibodies at room temperature for 24 h: anti-FGFR2 (1:1000; 11835, Cell Signaling Technology),
anti-FGF1 (1:1000; 3139, Cell Signaling Technology), anti-FGF2 (1:1000; 3196, Cell Signaling Technology), anti- FRS2 (1:1000, abcam), anti-phospho-FRS2-α (phospho- FRS2, 1:1000; 3861, Cell Signaling Technology), anti-Akt (1:1000; 9272, Cell Signaling Technology), anti-phospho- Akt (1:1000; 4051, Cell Signaling Technology), anti-Erk1/2 (1:1000; 9102, Cell Signaling Technology), anti-phospho- Erk1/2 (1:1000; 9101, Cell Signaling Technology) and anti- β-actin (1:5000; A5441, Sigma–Aldrich). The immunoreac- tion was visualized with alkaline phosphatase-conjugated antibodies (1:7500; 250793 and 249300, Promega, Madison, WI, USA) and BCIP-NBT Solution Kit (Nacalai Tesque). β-actin was used as the internal control. Image J was used for the estimation of band intensity.

ELISA

To evaluate the ability of FGF secretion in HEC-50B and Ishikawa cells, the concentration of FGF in culture medium were analyzed using Human FGF basic Quantikine ELISA Kit (R&D systems, Inc., Minneapolis, MN, USA). The cells were cultivated in EMEM with 10% charcoal-dextran-treated FBS at 37 °C for 72 h. Then, culture mediums were collected for ELISA analysis. The protocol according to the manufac- turer’s instructions was conducted.

Cell cycle analysis

Cells treated with NVP-BGJ 398 for 48 h were fixed with 70% ethanol at 4 °C for 4 h and then incubated with 100 µM RNase (Qiagen, Hilden, Germany) and propidium iodide (1 ng/mL, Sigma–Aldrich) for 30 min at 37 °C in dark. Cell cycle distributions were analyzed with a FACSAria III cell sorter (BD Biosciences).

Mutation analysis of FGFR2

The genomic DNA for FGFR2 was prepared from HEC- 50B cells and cloned into pUC19 plasmid. The nucleotide sequences were analyzed with the BigDye Terminator v1.1 Cycle Sequencing Kit and ABI PRISM 310 Genetic Ana- lyzer (Thermo Fisher Scientific) using the following primer pairs: FGFR2 exon 7, 5′-TCT CTC ATT CTC CCA TCC-3′ (forward) and 5′-AAA CCC ATG AAG GAG ACC-3′ (reverse); exon 9, 5′-TGC TAA CTC TAT GGC CTG C-3′ (forward) and 5′-ACA AGA TCC ACA AGC TGG C-3′ (reverse); and exon 12, 5′-TGT AGG CCT TTG TCC CTT C-3′ (forward) and 5′-AGG AAC ATC TTC CAA TGG GG-3′ (reverse).

Statistical analysis

All experiments were performed three times, however, muta- tion analysis was performed five times. When comparing data between two groups, Student’s t test or correlation coefficient (cc) was used for statistical analysis. The data obtained from various groups were analyzed statistically by one-way analysis of variance (ANOVA) followed by Student’s t test. A p value of < 0.05 and cc > 0.75 [35] were considered significant.

Results
Notch‑independent activation of HES1 in HEC‑50B endometrial cancer cells

Previous studies showed that cell proliferation and HES1 expression were induced by activation of the Delta/Notch pathway in Ishikawa cells [23]. We examined if this induc- tion pathway existed in a poorly differentiated endometrial cancer cell line, HEC-50B cells, using DAPT, a Notch sign- aling inhibitor. DAPT significantly decreased the viability of Ishikawa cells to 77% ± 9% compared with controls (n = 3, p < 0.05) (Fig. 1a). However, DAPT only slightly decreased the viability of HEC-50B cells (91% ± 9%) (n = 3, p = 0.5127). We next examined the effect of DAPT on HES1 mRNA using RT-PCR. Both cells expressed HES1 mRNA (Fig. 1b). However, while DAPT had no effect on HES1 mRNA expression in HEC-50B cells, Ishikawa cells treated with DAPT showed significantly reduced HES1 mRNA levels compared with controls (n = 3, p < 0.05) (Fig. 1b). This result is consistent with a previous study that showed that HES1 mRNA expression was decreased in Ishikawa cells by DAPT [23]. These findings suggest that Notch signaling is not active in HEC-50B cells. To confirm this possibility, we examined the expression of Notch1 and Notch4 mRNAs. As expected, both were at low or below detectable levels in HEC-50B cells (Fig. 1c), suggesting that HEC-50B cells express HES1 mRNA and proliferate in a Notch-independent way. Autocrine activation of FGFR in HEC‑50B endometrial cancer cells We next speculated on how HES1 expression was expressed and cells proliferated in HEC-50B cells in a Notch-independent manner. One possibility is that these cells harbor the mutation on FGFR [24, 28–30], leading to stimulation of HES1 expression and cell proliferation. Alternatively, expression of the FGF1 and FGF2 mRNA was reported in endometrial cancer [36], suggesting that it may stimulate FGFR in an autocrine manner. We thus cloned and analyzed the FGFR2 gene from endometrial cancer cells. We found no hotspot mutations in exons 7, 9 or 12 of FGFR2 cDNA prepared from HEC- 50B or Ishikawa cells (Fig. 2) [37]. We next examined the expression of FGFs and FGFRs in them. Both cells expressed FGF1, FGF2 and FGFR2 mRNA (Fig. 3a), and FGF1, FGF2 and FGFR2 protein expressions were confirmed with immunoblot analysis (Fig. 3b). Notably, the levels of FGF1, FGF2 and FGFR2 mRNA and protein were higher in HEC-50B cells than those in Ishikawa cells (n = 3, p < 0.05). To confirm the autocrine secretion of FGF, we measured FGF2 concentration in culture medium. ELISA assay revealed that FGF2 concentration in culture medium prepared from HEC-50B cells was higher than that from Ishikawa cells (n = 3, p < 0.05) (Fig. 3c). These findings suggest that FGFR is activated by FGF expressed by HEC-50B cells in an autocrine manner. The NVP‑BGJ 398 FGFR inhibitor suppressed cell proliferation and HES1 expression in HEC‑50B cells To confirm the involvement of FGF/FGFR signaling in cell proliferation in HEC-50B cells, we first tested whether FGFs stimulates cell proliferation of HEC-50B cells. Addition of FGFs stimulated cell proliferation compared with that without FGF (24 h incubation, n = 3, p < 0.05) (Fig. 4a). An FGFR inhibitor NVP-BGJ 398 suppressed cell proliferation of FGFs treated cell to 88% ± 7% (n = 3, p < 0.05) (Fig. 4a). Importantly, NVP-BGJ 398 suppressed cell proliferation even in the absence of exogenous FGF (n = 3, p < 0.05) (Fig. 4a). This finding again confirmed that proliferation is dependent on the autocrine of FGF in HEC-50B cells. We next examined the time course of NVP-BGJ 398 on cell proliferation. We found a gradual suppression of the cell viabilities of HEC-50B cells in response to 5 µM NVP-BGJ 398: the normalized cell pro- liferation ratio at 0, 12, 24 and 72 h were 103% ± 11%, 77% ± 6%, 56% ± 3% and 34% ± 2%, respectively (n = 3, p < 0.05) (Fig. 4b). The suppression of proliferation was also dependent on the concentration of NVP-BGJ 398; the proliferation ratios of cells treated with 0.01, 0.1, 1 and 10 µM NVP-BGJ for 72 h were 99% ± 2%, 90% ± 10%, 85% ± 10% and 70% ± 2%, respectively (n = 3, p < 0.05) (Fig. 4c). Since NVP-BGJ 398 was reported to arrest the cell cycle at G1 phase [38], we next analyzed the effects of the inhibitor on the HEC-50B cell cycle. Treatment with 5 µM NVP-BGJ 398 for 72 h increased the percentage of cells in G1/G0 phase from 45.0% ± 2.0% to 57.2% ± 4.0% and decreased cells in S and G2/M phase from 26.8% ± 3.0% and 25.6% ± 1.0–19.6% ± 1.0% and 20.4% ± 2.0%,respectively (n = 3, p < 0.05) (Fig. 4d). These findings show that the FGFR inhibitor suppressed proliferation of HEC-50B cells along with a G1 phase cell cycle arrest. Activation of the FGFR leads to the phosphorylation of FRS2 for the transmission of extracellular signals from FGF/FGFR [25]. To confirm NVP-BGJ 398 indeed inhibit FGFR signaling, we measured the protein levels of FRS2 and phosphorylated FRS2 with immunoblot. As expected, NVP-BGJ 398 decreased the expression of phosphorylated FRS2 protein (Fig. 4e). Our previous study revealed the pivotal role of HES1 in the proliferation of Ishikawa cells [23]. To test whether this is the case in HEC-50B cells, we measured HES1 mRNA expression in HEC-50B cells treated with FGF and NVP-BGJ 398. NVP-BGJ 398 treatment of HEC-50B cells decreased HES1 mRNA levels to 62% ± 2% at 3 h and 27% ± 4% at 12 h (n = 3, p < 0.05) (Fig. 5), suggesting an involvement of HES1 in the FGFR signaling for cell proliferation. FGFR activated HEC‑50B cell proliferation through Akt and HES1 To determine if either pathway mediates FGF/FGFR stimu- lation for HES1 expression and proliferation in HEC-50B cells, we used inhibitors against ERK (U0126) and Akt (SH- 5), respectively. Both 20 µM U0126 and SH-5 treatment decreased cell proliferation to 73% ± 9% and 19% ± 2%, respectively (n = 3, p < 0.05) (Fig. 6a, b), and the decrease was larger in SH-5-treated cells (n = 3, p < 0.05) (Fig. 6c). Furthermore, both inhibitors suppressed cell proliferation even in the absence of exogenous FGF in HEC-50B cells (n = 3, p < 0.05) (Fig. 6d–g). These results indicated the involvement of ERK and Akt signaling pathways in cell pro- liferation induced by autocrine FGF. Similar results were observed in HES1 mRNA expression. U0126 treatment sub- tly decreased the expression of HES1 mRNA to 99% ± 3% and 85% ± 5% at 3 h and 12 h, respectively (n = 3, p < 0.05) (Fig. 7a). In contrast, SH-5 treatment had a bigger impact on reducing HES1 mRNA levels (71% ± 8% and 36% ± 4% at 3 h and 12 h, respectively) (n = 3, p < 0.05) (Fig. 7b). The extent of decrease was significantly larger in SH-5-treated cells compared with U0126-treated cells (64% ± 4% and 15% ± 5% at 12 h) (n = 3, p < 0.05) (Fig. 7c). To confirm that U0126 and SH-5 blocked their signaling pathway, we exam- ined the levels of phosphorylated Akt and ERK proteins. As expected, each inhibitor decreased the levels of phosphoryl- ated Akt and ERK protein (Fig. 7d, e). These results sug- gest that FGFR signaling regulation of HES1 expression and proliferation is largely mediated by Akt in HEC-50B cells. Correlation between the sensitivities of HES1 expression and cell proliferation to various drugs The data presented above indicate the similarities in the extents of inhibition of HES1 expression and cell prolif- eration by the drugs we used. For instance, NVP-BGJ 398 largely inhibited them, and SH-5 modestly inhibited them. In contrast, DAPT did neither. We, therefore, analyzed correla- tion between the extents of inhibition in the HES1 expres- sion and cell proliferation by the drugs we used (Fig. 8). The inhibition in the HES1 expression was significantly corre- lated with that in the cell proliferation (r = 0.94). Discussion The unregulated proliferation of cancer cells is attributable to the aberrant activation of growth signaling pathways. However, the specific signaling pathways seem to differ depending on endometrial cancer types [39, 40]. Therefore, it is important to identify the specific intracellular signal- ing pathway that drives uncontrolled proliferation of each cancer type for the accurate selection of an appropriate targeted therapy, and effective markers are needed for this purpose. In this report, we demonstrated that an autocrine FGF/FGFR signaling, not Notch signaling, plays an essential role in HES1 expression and cell proliferation of poorly dif- ferentiated endometrial cancer cells. This finding might be useful to distinguish patients who would respond to FGFR inhibitors (see below). Furthermore, HES1 expression might be useful as a marker for the activation of this pathway. Lack of involvement of Notch in HEC‑50B cells Our previous study reported that inhibition of the Notch signaling decreased cell proliferation through downregula- tion of HES1 expression in Ishikawa cells [23]. However, in the current research, HEC-50B cells did not respond to Notch inhibition in terms of HES1 expression or cell prolif- eration. This was further confirmed by the lack of Notch1 or Notch4 mRNA expression. These differences in Notch expressions between HEC-50B and Ishikawa cells might be due to the carcinogenesis process. Endometrial cancers are categorized into two types based on clinical behavior and carcinogenesis process; Ishikawa and HEC-50B cells are classified as type 1 and 2, respectively [1, 39]. Type 1 carcinomas, which consist of well and moderately differ- entiated endometrioid carcinoma, arise from a background of unopposed estrogen stimulation and express estrogen receptor (ER) and progesterone receptor (PgR). In contrast, type 2 carcinoma, which contained poorly differentiated endometrioid carcinoma, serous carcinoma and clear cell carcinoma, does not express ER or PgR and thereby prolif- erate in an estrogen-independent manner. Wei et al. dem- onstrated the essential role of estrogen and ER stimulation in the expression of Notch1 in type 1 endometrial cancer cells [41]. Therefore, the lack of Notch1 or Notch4 mRNA expression in type 2 HEC-50B cells might be attributable to the lack of ER expression, and this lack of Notch expression might be common among type 2 cells, which do not express ER. If this is the case, poorly differentiated type 2 endo- metrial cancer cells should be resistant to inhibitors for the Notch signaling, because they do not express ER or Notch. Although we did not examine the causative role of HES1 in cell proliferation of endometrial cancer cells, several studies clearly showed the essential role for HES1 in various can- cers, e.g., glioblastoma, colon and breast cancer [16, 42, 43]. Our present results also demonstrated that correlation between the sensitivities of HES1 expression and cell pro- liferation to drugs. Namely, in HEC-50B cells, cell prolifera- tion and HES1 expression were resistant to DAPT, sensitive to NVP-BGJ 398, slightly sensitive to U0126 and sensitive to SH-5. Accordingly, HES1 gene expression significantly correlated with the proliferation ability of cancer cells, sug- gesting the essential role of HES1 in cell proliferation. Thus, HES1 expression might be a useful marker to predict the proliferation ability and sensitivity to anticancer drugs. The expression level of HES1 varied even among poorly differ- entiated endometrial cancer cells and some cells expressed only low levels of HES1 [44]. FGFR signaling pathway in HEC‑50B cells We showed that HES1 expression was induced by the expression of FGFs through the autocrine activation of wild- type FGFR in HEC-50B cells, rather than Notch1 signaling. We considered two possibilities for these observations: (1) FGFR may harbor mutations leading to ligand-free activa- tion, which was reported in gastric, lung and endometrial cancer [45, 46]; or (2) autocrine activation of wild-type FGFR, in which HEC-50B cells expressed and secreted FGFs, and the secreted FGFs activated FGFR. We exam- ined the hotspot mutations in FGFR2 cloned from HEC-50B cells, but failed to detect any mutations. In contrast, FGF-1 and FGF-2 mRNA and protein were robustly expressed in HEC-50B cells to significantly higher levels than those in Ishikawa cells. Furthermore, the concentration of FGF was high in culture medium prepared from HEC-50B cells. The addition of FGF further enhanced the cell proliferation. FGFR inhibitor suppressed cell proliferation in both the presence and absence of additional FGF treatment. These results suggest the autocrine activation of cell proliferation of HEC-50B cells. Contrastingly, Ishikawa cells, classified as type 1 carcinoma, could hardly express FGF-1, FGF-2 and FGFR2. Our previous report showed that Estrogen/Notch pathway play a crucial role in cell proliferation of Ishikawa cells [23]. Therefore, type 1 and 2 endometrial cancer cells seem to use different signaling pathway. So, moderately dif- ferentiated endometrial cancer cells are classified to type 1, might not use FGF autocrine activation mechanism. Indeed, NVP-BGJ 398 decreased HES1 expression and cell prolif- eration, confirming the involvement of FGFR without any mutations. FGF-2 was expressed in stromal cells and blood vessels in normal endometrium regardless of the menstrual cycle [26]. FGF expression might be maintained even after the carcinogenesis process. Similar autocrine and paracrine stimulations of FGFRs have been identified in melanomas and non-small cell lung cancer [47, 48]. Our results provide additional evidence of the autocrine activation of FGFR, which leads to the proliferation of endo- metrial cancers. We also showed a major role for Akt in the intracellular signaling downstream of FGFR. Both the U0126 MEK inhibitor and SH-5 Akt inhibitor decreased HES1 expression and cell proliferation in HEC-50B cells. However, the extent of decrease was significantly larger in SH-5-treated cells than U0126-treated cells, suggesting a major role for Akt signaling downstream of FGFR activation in HEC-50B cells. Consistent with these results, an essential role of PI3K/AKT signaling in cell proliferation and HES1 expression and Akt phosphorylation were reported in vari- ous endometrial cancer cell lines, including HEC-50B cells [49]. Most (93%) endometrioid tumors had mutations in PI3K/AKT pathway-related genes, leading to aberrant acti- vation of PI3K/AKT signaling [50]. Indeed, HEC-50B cells contain a mutation in the PIK3R1 gene, which potentially facilitates oncogenesis together with phosphorylation of Akt [51]. In addition, the RAS/MAPK pathway interacts with the PI3K/AKT pathway [52], and HEC-50B cells were reported to have mutations in the KRAS gene. Akt phosphorylation was not affected by RAS mutation in endometrial cancer [49, 53], suggesting this crosstalk is not involved in Akt phospho- rylation. Similarly, in this study, U0126 slightly decreased cell proliferation and HES1 expression in HEC-50B cells, again suggesting the unessential role of ERK and RAS sign- aling in HEC-50B cells [49]. Previous studies reported com- plete inhibitions on phosphorylation of ERK1/2 with 10 µM U0126 [54, 55], and, therefore, this slight inhibition was not attributable to the insufficient concentration of U0126. Clinical significance We showed that the autocrine activation of FGFR, rather than Notch1, stimulated HES1 expression and prolifera- tion of HEC-50B cells. Our study provides an insight into the choice of anticancer medicine depending on types of endometrial cancer. We previously showed that the prolifera- tion of Ishikawa cells was dependent on both estrogen and Notch signaling. This suggests that Notch inhibitors should be only effective for patients with estrogen-dependent endo- metrial cancer. Furthermore, FGFR inhibitors are postulated to only be used for patients with mutated FGFRs [31, 33]. However, our study demonstrated the efficacy of NVP-BGJ 398 on HEC-50B cells with wild-type FGFR2. Thus, we propose that FGFR inhibitors should be also used for endo- metrial cancers that do not have mutations in FGFR2. We also expect that HES1 expression is useful to distinguish the sensitivity of Notch-negative poorly differentiated endome- trial cancer harboring wild-type FGFR2 to FGFR inhibitors.