c-Jun N-terminal kinase (JNK) is a member of the mitogen-activated protein kinase (MAPK) family, and it is reportedly involved in the development of several cancers. However, the role of JNK in pancreatic cancer has not been elucidated. We assessed t he involvement of JNK in the development of pancreatic cancer and investigated the therapeutic effect of JNK inhibitors on this deadly cancer. Small interfering RNAs against JNK or the JNK inhibitor SP600125 were used to examine the role of JNK in cellular proliferation and the cell cycles of pancreatic cancer cell lines. Ptf1acre/+;LSL-KrasG12D/+;Tgfbr2flox/flox mice were treated with the JNK inhibitor to examine pancreatic histology and survival. The effect of JNK inhibition on tumor angiogenesis was also assessed using cell lines and murine pancreatic cancer specimens. JNK was frequently activated in human and murine pancreatic cancer in vitro and in vivo. Growth of human pancreatic cancer cell lines was suppressed by JNK inhibition through G1 arrest in the cell cycle with decreased cyclin D1 expression. In addition, oncogenic K-ras expression led to activation of JNK in pancreatic cancer cell lines. Treatment of Ptf1acre/+;LSL-KrasG12D/+;Tgfbr2flox/flox mice with the JNK inhibitor decreased growth of murine pancreatic cancer and prolonged survival of the mice significantly. Angiogenesis was also decreased by JNK inhibition in vitro and in vivo. In conclusion, activation of JNK promotes development of pancreatic cancer, and JNK may be a potential therapeutic target for pancreatic cancer.
c-Jun N-terminal kinase (JNK) is a member of the MAPK family, is activated by cytokines, growth factors and environmental stress, and controls cell proliferation, differentiation, apoptosis, and survival[1-4] through the regulation of various target molecules such as transcriptional factors c-Jun, activating transcription factor 2 (ATF2), E twenty-six-like transcription factor 1 (Elk1), and signal transducer and activator of transcription 3 (STAT 3).[1, 5-7]
c-Jun N-terminal kinase has been reported to play an important role in the development of various cancers.[8-11] JNK1-deficient mice exhibited a decrease in carcinogenesis of chemically-induced gastric cancer or hepatocellular carcinoma,[12-15] and JNK2-deficient mice showed reduced formation of skin tumors. Activation of JNK was reported in human pancreatic cancer specimens, and growth of a pancreatic cancer cell line was reported to be inhibited by pharmacological inhibition of JNK in vitro. However, the function of JNK in pancreatic cancer remains largely unexplored.
Ptf1acre/+;LSL-KrasG12D/+;Tgfbr2flox/flox mice (KrasG12D+Tgfbr2KO mice) are pancreas-specific Tgfbr2 knockout mice with active K-ras expression that are generated using the Cre-loxP system driven by the Ptf1a (pancreatic transcription factor-1a) promoter. These mice develop murine pancreatic intraepithelial neoplasia (mPanIN) progressively and, finally, well-differentiated pancreatic ductal adenocarcinoma (PDAC) with 100% penetrance and a median survival duration of 59 days. Because the clinical and histopathological manifestations of KrasG12D+Tgfbr2KO mice recapitulate human PDAC, these mice are considered a good model of human PDAC and are used for the evaluation of candidate treatment drugs for PDAC.
In the present study, we examined the role of JNK and the effect of JNK inhibition in pancreatic cancer using pancreatic cancer cell lines and KrasG12D+Tgfbr2KO mice. We found that inhibition of JNK can be a potential therapy for pancreatic cancer.
Materials and Methods
Ptf1acre/+;LSL-KrasG12D/+;Tgfbr2flox/flox mice (KrasG12D+Tgfbr2KO mice) and Ptf1acre/+;LSL-KrasG12D/+ mice (KrasG12D mice) were described previously.[19-21] Genotyping of Ptf1acre/+, LSL-KrasG12D/+, floxed Tgfbr2 alleles was performed using oligonucleotide primers as described previously.[22-24] All the experimental protocols were approved by the ethics committee for animal experimentation and were conducted in accordance with the Guidelines for the Care and Use of Laboratory Animals of the Graduate School of Medicine at the University of Tokyo.
RNA interference and plasmids
JNK1 small interfering RNA (siRNA) or JNK2 siRNA, cyclin D1 siRNA, KRAS siRNA, and control siRNA were purchased from Qiagen (Hilden, Germany) and transfected into the cells using RNAiMAX (Invitrogen Life Technologies, Carlsbad, CA, USA). Wild-type Ha-ras vector and constitutively active Ha-ras (RasG12V) vector were purchased from Clontech and subcloned into pTriEx-2 (Novagen, Madison, WI, USA). Cells were seeded into 12-well plates and, after 24 h, they were transfected with 0.3 μg of expression plasmids or the control vector (pcDNA3.1) using Effectene Transfection reagent (QIAGEN).
Cell culture, cell cycle analysis, histology, immunohistochemistry, immunofluorescence, the terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assay, immunoblotting, real-time RT-PCR, cytokine array, and Enzyme-linked immunosorbent assay (ELISA)
Details are described in Suppl. Doc S1.
Drug administration and experimental design in an in vivo model
The JNK-specific inhibitor SP600125 (LC Laboratories, Woburn, MA, USA) was dissolved in DMSO, and then was diluted in PBS and injected intraperitoneally into KrasG12D+Tgfbr2KO mice. To examine the effect of SP600125 on PDAC in vivo, we injected 50 mg/kg SP600125 five times per week from postnatal day 21 for 4 weeks, and then mice were killed. To examine survival of the mice, we injected 10 mg/kg SP600125 five times per week from postnatal day 21 until the mice became moribund. D-JNKI1 (Enzo Life Sciences, Plymouth Meeting, PA, USA), a peptide that specifically inhibits JNK, was dissolved in PBS and was injected intraperitoneally into KrasG12D+Tgfbr2KO mice (5 mg/kg once in every 3 days from postnatal day 21, for 15 days).
Capillary formation of human umbilical vein endothelial cells (HUVECs) was assessed using the Fibrin Gel In Vitro Angiogenesis Assay Kit (Chemicon, Temecula, CA, USA), as described previously. Details are described in Suppl. Doc S1.
The data were expressed as the mean ± SD. Differences between means were compared using Student's t-test; a P-value <0.05 was deemed to indicate statistical significance. The survival curves were plotted according to the Kaplan–Meier method and compared by the log-rank test.
JNK activation in human pancreatic cancer
To analyze the role of JNK in pancreatic cancer, we first examined JNK phosphorylation in human pancreatic cancer specimens by immunohistochemical staining. As shown in Fig. 1A, we observed increased levels of phosphorylated JNK in 85/93 of the pancreatic cancer samples (91.4%), and none of the 18 normal samples showed positive staining for phosphorylated JNK. The rate of cases with phosphorylated JNK did not show a significant difference between pathological stage and histological grade (Fig. 1B). Next, we investigated phosphorylation of JNK in eight human pancreatic cancer cell lines. Phosphorylated JNK was observed in all the cell lines (Fig. 1C). These results indicate that the JNK pathway is activated in most cases of human pancreatic cancer.
The role of JNK in PDAC cell proliferation
To examine whether JNK affects tumor growth in vitro, we inhibited JNK in pancreatic cancer cell lines BxPC-3 and KP-4, which contain enhanced phosphorylation of JNK, using JNK1 siRNA, JNK2 siRNA, or the JNK inhibitor SP600125. Cell proliferation was significantly inhibited following JNK inhibition (Figs 2A,B and S1A,B). Next, we investigated the cell cycle status of pancreatic cancer cells by flow cytometry. Cells transfected with JNK1 or JNK2 siRNA displayed a higher proportion of cells in the G0/G1 phase, compared with control siRNA (Fig. 2C), indicating that JNK inhibition leads to G1 arrest. Then we examined the expression of genes related to cell proliferation and cell cycle regulation. The expression of cyclin D1 was decreased in cells transfected with JNK1 or JNK2 siRNA and SP600125-treated cells (Figs 2D,E and S1C,D).
We next investigated the role of Ras in JNK activation in pancreatic cancer because oncogenic K-ras mutation is known to be highly prevalent in pancreatic cancer, and Ras has been reported to cause activation of JNK.[1, 27, 28] We overexpressed wild-type Ras or oncogenic RasG12V in BxPC-3 and KP-4 cells using expression plasmids. The expression levels of phosphorylated JNK were higher in RasG12V cells than in wild-type Ras cells after normalization to Ras expression (Figs 2F and S1E). Extracellular signal-regulated kinase (ERK) was also activated, which was considered to be dependent on the Ras/Raf/MEK/ERK pathway. We also knocked down KRAS in KP-4 cells using KRAS siRNA. The level of phosphorylated JNK was decreased in KRAS knocked-down cells (Fig. 2G), and the growth of KP-4 cells was significantly inhibited by KRAS knockdown (Fig. 2H).
Cyclin D1 is a well-established cell cycle regulator of proliferation, and we confirmed that growth of cell lines was inhibited by transfection of cyclin D1 siRNA (Fig. 3A). To investigate the role of Ras activation in cyclin D1 transactivation, we performed a luciferase assay using the cyclin D1 promoter-containing construct pD1luc and consensus AP-1 binding site-containing AP-1luc. Overexpression of RasG12V enhanced both the cyclin D1 promoter and AP-1 transcription factor in KP-4 cells, whereas treatment with SP600125 abolished the enhanced promoter activities (Fig. 3B). We also observed that cyclin D1 mRNA was decreased by the JNK inhibitor SP600125 as well as siRNAs against JNK1 and JNK2 in KP-4 cells (Fig. 3C). These results ensured that oncogenic Ras regulates cyclin D1 transcription via JNK activation. Next, we examined the expression of cyclin D1 in human pancreatic cancer tissues. Human pancreatic cancer tissue array showed positive staining in pancreatic cancer tissue in 15/20 cases and 0/4 in normal cases (Fig. 3D). These results indicate that JNK plays an important role in the proliferation of pancreatic cancer through cyclin D1 regulation, and mutation of K-ras seems to be one of the mechanisms involved in JNK activation of pancreatic cancer.
JNK activation in murine PDAC tissue and the effect of JNK inhibition on PDAC development and survival of the mice
Although KrasG12D+Tgfbr2KO mice develop PDAC, KrasG12D mice develop mPanIN and rarely develop pancreatic carcinoma within 1 year.[19, 21] We examined JNK activation in normal, mPanIN, and PDAC tissue, obtained from wild-type, KrasG12D, and KrasG12D+Tgfbr2KO mice, respectively. Immunohistochemical staining showed increased expression of phosphorylated JNK according to the progression of PDAC, in accordance with cyclin D1 expression (Fig. 4A). A similar result was obtained in immunoblotting analysis (Fig. 4B). The levels of phosphorylated c-Jun were also elevated in accordance with histological progression; however, phosphorylated ERK was activated in mPanIN and PDAC to the same extent (Fig. 4B). These results have indicated that the JNK pathway is activated in the mouse model of PDAC with mutated K-ras and transforming growth factor (TGF)β signaling, and loss of TGFβ signaling seems to have an additional effect on JNK activity mediated by K-ras mutation in vivo.
To investigate the therapeutic effect of JNK inhibition in vivo, we treated KrasG12D+Tgfbr2KO mice with JNK inhibitors. Histological analysis showed significantly decreased tumor areas in the pancreas of the mice treated with SP600125 (n = 9; 55.5 ± 20.2%) compared with the control mice (n = 8; 86.6 ± 8.9%) (Fig. 4C). We observed similar results using another JNK inhibitor, D-JNKI1, which has a mechanism of JNK inhibition distinct from that of SP600125 (Fig. S2A). Next, we examined whether survival is prolonged by JNK inhibitor treatment. As shown in Fig. 4D, the survival time was significantly improved in mice treated with SP600125 (median survival time, 63 days), compared with the control vehicle (57.5 days; P = 0.036, log-rank test).
The PDAC specimens of the mice were immunohistochemically stained, and we observed decreased expression of phosphorylated JNK, c-Jun, cyclin D1, and PCNA in the PDAC tissues treated with SP600125 (Fig. 4E), suggesting that JNK inhibition suppressed cell proliferation. We next performed the TUNEL assay for PDAC sections. The number of TUNEL-positive cells in PDAC tissue treated with SP600125 was similar to those in control sections, showing that the anti-tumor effect of JNK inhibitors is not dependent on apoptosis (Fig. S2B). These results indicate that treatment with JNK inhibitors had therapeutic effects on PDAC in KrasG12D+Tgfbr2KO mice.
The mechanism of the anti-tumor effect by JNK inhibition
To further investigate the mechanism of the anti-tumor effect of JNK inhibition, we evaluated the involvement of tumor angiogenesis. We found fewer blood vessels in PDAC specimens treated with SP600125 compared with those in control specimens (Fig. 5A). We examined the levels of cytokines in the supernatant of KP-4 cells treated with SP600125 or control vehicle using cytokine array. We found that some angiogenic cytokines, such as interleukin-8 (IL-8), growth-related oncogene α (GROα), vascular endothelial growth factor (VEGF), and monocyte chemotactic protein-1 (MCP-1),[32-35] were reduced by SP600125 treatment (Fig. S3A). We quantitated the levels of these cytokines in the supernatant of KP-4 cells treated with SP600125 or siRNA against JNK1 and 2 by ELISA, and confirmed that these cytokines are decreased in the supernatant of KP-4 cells with JNK inhibition (Figs 5B and S3B).
Next, we incubated HUVECs with the supernatant of KP-4 cells transfected with siRNAs against JNK1 and 2, and found that capillary formation by HUVECs was significantly reduced (Fig. 5C). Similar results were obtained with the supernatant of KP-4 cells treated with SP600125 (Fig. S3C). Incubating HUVECs with the supernatant of BxPC-3 cells also showed similar results (data not shown). These results indicate that JNK inhibition has anti-angiogenic effects on pancreatic cancer in vitro and in vivo, which may be associated with therapeutic effect observed in vivo.
CD44 is an adhesion molecule known to be upregulated in human PDACs, and putative pancreatic cancer stem cells have been identified based on the expression of CD44, CD24, and epithelial specific antigen. CD44-positive cells were found in PDAC of KrasG12D+Tgfbr2KO mice, whereas normal acinar and ductal cells showed no CD44 staining (Fig. 5D). We observed higher expression of CD44 in PDAC tissue of KrasG12D+Tgfbr2KO mice compared with mPanIN and normal pancreatic tissue (Fig. 5E). CD44-positive cells in PDAC tissue were decreased by SP600125 treatment (Fig. 5F). The KP-4 cell line showed decreased CD44 expression when treated with SP600125 (Fig. 5G). Thus, it is possible that the JNK inhibitor exerts an anti-tumor effect by decreasing CD44-positive cells.
JNK1 and JNK2 may play different roles depending on the cell type and condition. For example, JNK1-deficient mice spontaneously develop intestinal tumors, but loss of JNK2 shows no phenotype in the intestine. In the present study, silencing JNK1 or JNK2 in vitro both showed an inhibitory effect on the growth of pancreatic cancer cells (Fig. 2A), indicating that JNK1 and JNK2 may be cooperatively involved in the proliferation of pancreatic cancer cell lines.
In the pancreatic cancer tissue of KrasG12D+Tgfbr2KO mice, JNK1 was strongly activated, whereas JNK2 was barely activated (Fig. 4B). In human PDAC cell lines, mainly JNK2 was activated (Fig. 1C), and JNK1 was activated by K-ras induction (Fig. 2F). Conversely, histologically normal acinar cells of KrasG12D mice, which also contain the mutant K-ras gene knocked in by Ptf1acre, did not show phosphorylated JNK staining (Fig. 4A middle). Thus, K-ras mutation is required for JNK activation but may not be sufficient for JNK activation in vivo. Presently, we do not have sufficient data to discuss the differences in JNK1 and JNK2 activation status between mouse and human cells or between the in vitro and in vivo model. Because JNK is known to be activated by various molecules, such as MAPK kinase 4 (MKK4) and MKK7, the difference in JNK activation between in vitro and in vivo may originate from the balance of these signals. Experiments using JNK1 or JNK2 knockout mice may help investigate the difference in JNK activation by mutant K-ras in vitro and in vivo.
c-Jun N-terminal kinase phosphorylation is stronger in PDAC than in mPanIN, whereas ERK activation was almost the same in both tissue types (Fig. 4B). This result also suggests the critical role of JNK in pancreatic carcinogenesis, consistent with the report that Ras-induced transformation of primary murine embryonic fibroblasts (MEFs) requires JNK. Additionally, JNK signaling might be more important than ERK signaling in the late phase of carcinogenesis.
Cyclin D1 is one of the molecules that regulates the cell cycle and is important in many types of human cancer, including gastric cancer. Cyclin D1 is overexpressed in a significant proportion of human pancreatic cancers, and elevated cyclin D1 levels correlate with abnormal growth and tumorigenicity in human pancreatic cancer.[42, 43] Our results that JNK inhibition leads to decreased cyclin D1 expression and cell cycle retardation (Fig. 2C–E) are consistent with previous reports,[44-47] suggesting that the therapeutic effect of JNK inhibition is dependent on suppression of cyclin D1 expression.
c-Jun N-terminal kinase-targeted cancer therapy has been investigated mainly in vitro using antisense oligonucleotides and pharmacological inhibitors.[9, 18, 48] Regarding pancreatic cancer, SP600125 inhibits the growth of the pancreatic cancer cell line MIAPaCa2. We demonstrated that JNK inhibition suppressed progression of PDAC and prolonged the survival of mice with PDAC (Fig. 4A–D). As the result of TUNEL staining showed that apoptosis was not increased in the tumors of SP600125-treated mice (Fig. S2B), the anti-tumor effect by JNK inhibition is not dependent on apoptosis in our in vivo model.
The growth and progression of solid tumors are affected by angiogenesis, and angiogenesis has been reported to have a positive correlation with disease progression of pancreatic cancer.[50, 51] Furthermore, JNK and p38 are essential for VEGF mRNA stabilization in vitro. In the present study, we observed decreased angiogenesis by treatment with SP600125 in vitro and in vivo. Therefore, suppression of vascularization is probably one of the mechanisms of the anti-tumor effect by the JNK inhibitor in vivo. Conversely, stromal cells surrounding tumors can secrete angiogenic cytokines. We found that some stromal cells are positive for phosphorylated JNK in immunohistochemical staining (Figs 1A and 4A). The effect of the JNK inhibitor on fibroblasts or macrophages in the PDAC remains to be investigated.
We also found that the JNK inhibitor decreased CD44 expression in PDAC of KrasG12D+Tgfbr2KO mice. The CD44 expression level is associated with poor prognosis and gemcitabine resistance in pancreatic cancer. Our results indicate that JNK inhibition may affect CD44-positive cells (Fig. 5D–G). Because CD44-positive pancreatic cancer cells were previously shown to increase when treated with gemcitabine, a combination of JNK inhibition with gemcitabine treatment may be more effective, by overcoming gemcitabine resistance.
In summary, we showed that JNK has a cancer-promoting effect in pancreatic cancer, and inhibiting JNK leads to growth inhibition, cell cycle arrest, and decreased angiogenesis, resulting in prolonged survival. Thus, JNK is a potential therapeutic target for pancreatic cancer.
We thank Chris Wright for providing the Ptf1acre mice, Tyler Jacks for providing the LSL-KrasG12D mice, and Mitsuko Tsubouchi for technical assistance. This work was supported by a grant-in-aid from the Japanese Ministry of Education, Culture, Sports, Science, and Technology to S.M. (#22300317) and a grant from the Japanese Society of Gastroenterology to H.I. as well as Japan Society for the Promotion of Science (JSPS) Core-to-Core Program “Cooperative International Framework in TGF-β Family Signaling”.
All authors declare that they have no conflict of interest.