15-PGDH inhibits hepatocellular carcinoma growth through 15-keto-PGE2/PPARγ-mediated activation of p21WAF1/Cip1

15-hydroxyprostaglandin dehydrogenase (15-PGDH) is a key enzyme in prostaglandin metabolism. This study provides important evidence for inhibition of hepatocellular carcinoma (HCC) growth by 15-PGDH through the 15-keto-PGE2/PPARγ/p21WAF1/Cip1 signaling pathway. Forced overexpression of 15-PGDH inhibited HCC cell growth in vitro, whereas knockdown of 15-PGDH enhanced tumor growth parameters. In a tumor xenograft model in SCID mice, inoculation of human HCC cells (Huh7) with overexpression of 15-PGDH led to significant inhibition of tumor growth, while knockdown of 15-PGDH enhanced tumor growth. In a separate tumor xenograft model in which mouse HCC cells (Hepa1-6) were inoculated into syngeneic C57BL/6 mice, intratumoral injection of adenovirus vector expressing 15-PGDH (pAd-15-PGDH) significantly inhibited xenograft tumor growth. The anti-tumor effect of 15-PGDH is mediated through its enzymatic product, 15-keto-PGE2, which serves as an endogenous PPARγ ligand. Activation of PPARγ by 15-PGDH-derived 15-keto-PGE2 enhanced the association of PPARγ with the p21WAF1/Cip1 promoter and increased p21 expression and association with CDK2, CDK4 and PCNA. Depletion of p21 by shRNA reversed 15-PGDH-induced inhibition of HCC cell growth; overexpression of p21 prevented 15-PGDH knockdown-induced tumor cell growth. These results demonstrate a key 15-PGDH/15-keto-PGE2-mediated activation of PPARγ and p21WAF1/Cip1 signaling cascade that regulates hepatocarcinogenesis and tumor progression.


INTRODUCTION
Hepatocellular carcinoma (HCC) is a primary malignancy of the liver and its incidence is rising in the United States and around the world (1)(2)(3)(4). The tumorigenic process is characterized by dysregulation of cell cycle progression and abnormal cell proliferation in the setting of chronic inflammation and fibrosis of the liver parenchyma. Consistent with the strong association between chronic inflammation and hepatocarcinogenesis, studies have shown that mediators of inflammation, such as prostaglandins (PGs), play an important role in hepatocarcinogenesis (5)(6)(7). Previous studies have been focused on defining the role of cyclooxygenase-2 (COX-2, a key enzyme that mediates prostaglandin synthesis) in HCC. Indeed, the expression of COX-2 is increased in human and animal HCCs and in dysplastic hepatocytes (5)(6)(7). In cultured HCC cells, forced overexpression of COX-2 increases tumor cell growth and invasiveness. Selective and non-selective COX-2 inhibitors prevent HCC cell growth in vitro and in animal models of hepatocarcinogenesis (5)(6)(7), although these inhibitors are known to mediate effects through both COX-dependent and -independent mechanisms. These findings suggest the possibility of targeting COX-2 for prevention and treatment of HCC in patients. This approach is expected to be safe, given that selective COX-2 inhibitors do not adversely affect renal function in cirrhosis (8,9) (in contrast to NSAID-related renal failure in decompensated cirrhosis). However, on the other hand, in light of the increased cardiovascular side effect associated with some COX-2 inhibitors (10)(11)(12)(13), it is imperative to identify specific molecular targets downstream of COX-2 for effective and safer anti-HCC therapy.
This study was designed to examine the biological function and molecular mechanism of 15-PGDH in hepatocellular carcinoma by using complementary in vitro and in vivo approaches. We show herein that the anti-tumor effect of 15-PGDH is mediated through its enzymatic product, 15-keto-PGE 2 , which activates peroxisome proliferator-activated receptor γ (PPARγ) leading to p21 WAF1/Cip1 expression and association with key downstream molecules including CDKs and PCNA. Our data shift the current paradigm and disclose an important 15-PGDH/15-keto-PGE 2 -mediated activation of PPARγ and p21 WAF1/Cip1 signaling axis that suppresses hepatocarcinogenesis and tumor progression.

15-PGDH inhibits HCC growth in SCID mice
To examine the effect of 15-PGDH signaling on HCC growth in vivo, the above four stable Huh7 cell lines were inoculated into SCID mice and the animals were closely monitored for tumor development. As shown in Figure 2A, 15-PGDH overexpression inhibited tumor growth whereas 15-PGDH depletion accelerated growth. When 15-PGDH was overexpressed, the tumor weight decreased to approximately one-third of the control group (0.208±0.057g versus 0.748±0.153g, p < 0.01). Conversely, when 15-PGDH was knocked down, the tumor weight increased approximately three fold compared to the control group (2.311±0.498g versus 0.681±0.124g, p < 0.01). The tumor appearance time in 15-PGDH overexpressed group was prolonged (15.13±4.27 days versus 8.12±3.24 days in the control group, p < 0.01); in contrast the tumor appearance time in 15-PGDH knockdown group was shortened (5.98±2.13 days versus 9.03±3.19 days in the control group, p < 0.01). Immunohistochemical staining for the proliferating cell nuclear antigen (PCNA) showed that the percentage of PCNA-positive cells was significantly lower in 15-PGDH overexpressed tumors (18.24±2.99%) compared to the control group (36.83±9.51%, p < 0.01). In contrast, the percentage of PCNA-positive cells was significant higher in 15-PGDH knockdown tumors (91.22±11.86%) compared to the control group (41.32±6.31%, p < 0.01) ( Figure 2B). Accordingly, western blotting confirmed that the level of PCNA was lower in 15-PGDH overexpressed tumors and higher in 15-PGDH knockdown tumors ( Figure 2C). Furthermore, 15-PGDH overexpression increased the expression level of the cyclin kinase inhibitor p21 WAF1/Cip1 in xenograft tumors, whereas 15-PGDH knockdown reduced it ( Figure 2C).

15-PGDH inhibits HCC growth in C57BL/6 mice
We next utilized a complementary syngeneic HCC xenograft model in which a murine HCC cell line originated from C57BL/6 mouse strain (Hepa1-6) was inoculated subcutaneously at armpit in C57BL/6J mice. When the tumors become palpable, an adenoviral vector expressing 15-PGDH (pAd-15-PGDH) or the pAd control vector was directly injected into the tumor modules (at three day interval, starting 10 days after inoculation till the end of the experiment). The tumor size in the pAd-15-PGDH injected group was significantly smaller compared to the pAd-control group (p < 0.01) ( Figure 3A). The average tumor weight in pAd-15-PGDH treated group was also significantly lower compared to the pAd control group (0.57±0.12 grams versus 2.51±0.58 grams p < 0.01) ( Figure 3B). Increased 15-PGDH protein in pAd-15PGDH treated tumors was confirmed by Western blotting analysis ( Figure  3C). pAd-15PGDH treatment decreased PCNA expression in xenograft tumor cells ( Figure  3D-E). Furthermore, treatment with pAd-15PGDH increased the level of p21 WAF1/Cip1 (predominantly in the nuclei) ( Figure 3E-F). These findings further demonstrate that 15-PGDH inhibits the progression of HCC in vivo and suggest that a potential role of p21 WAF1/Cip1 .

The effect of 15-PGDH on p21 WAF1/Cip1 association with PCNA and CDKs in HCC cells
Consistent with the notion that p21 inhibits PCNA and CDKs in the nucleus, Western blotting analysis showed that 15-PGDH overexpression caused accumulation of p21 in the nucleus and decrease of p21 in the cytoplasm ( Figure 6A). In contrast, 15-PGDH knockdown led to reduction of p21 in the nucleus and increase of p21 in the cytoplasm. Coimmunoprecipitation analysis showed that 15-PGDH overexpression increased the interaction between p21 and PCNA, whereas 15-PGDH knockdown inhibited their interaction ( Figure 6B). In parallel, 15-PGDH overexpression also increased p21 interaction with CDK2, CyclinE, CDK4 and CyclinD1, whereas 15-PGDH knockdown inhibited these interactions ( Figure 6C and 6D). Furthermore, 15-PGDH overexpression enhanced E2F1 binding to RB and decreased E2F1 binding to C-myc, while an opposite pattern of interactions was observed in cells with 15-PGDH knockdown ( Figure 6E). These findings demonstrate that 15-PGDH signaling influences p21 interaction with key cell cycleregulatory molecules.

Regulation of p21 WAF1/Cip1 by 15-PGDH is independent of p53 in HCC cells
Given that the expression of p21 in human cells is regulated by wild type p53, we performed further experiments to determine whether p53 is implicated in 15-PGDH-mediated regulation of p21 in HCC cells. To this end, we utilized HepG2 cell line that expresses wild type p53 (Huh7 cells have p53 mutation). Western blotting analysis showed that shRNA depletion of p53 did not alter 15-PGDH-induced p21 expression (nuclear accumulation and cytoplasmic reduction), despite that p53 depletion reduced p21 protein in cells without 15-PGDH overexpression ( Figure 7A). Similarly, luciferase reporter activity assay showed that shRNA depletion of p53 did not alter 15-PGDH-induced p21 promoter activity, although p53 depletion reduced p21 promoter reporter activity in cells without 15-PGDH overexpression ( Figure 7B). Furthermore, p53 depletion did not influence 15-PGDHinduced interaction of p21 with CDK2, CDK4, PCNA and the interaction between E2F1 and RB, although p53 depletion decreased the interactions of these molecules in cells without 15-PGDH overexpression ( Figure 7C). Finally, CHIP assay showed that p53 depletion did not alter 15-PGDH-induced PPARγ binding to the p21 promoter DNA, although p53 depletion reduced their binding in cells without 15-PGDH overexpression ( Figure 7D). These results suggest that 15-PGDH signaling upregulates p21 expression in HCC cells through mechanisms independent of p53.

p21 WAF1/Cip1 mediates the anti-tumor effect of 15-PGDH
To further evaluate the role of p21 in 15-PGDH-mediated anti-tumor effect, additional experiments were performed to determine whether overexpression or knockdown of p21 would influence 15-PGDH-regulated cell growth. As shown in Figure 8A, knockdown of p21 reversed 15-PGDH-induced inhibition of HCC cell proliferation and clonogeneic growth. On the other hand, overexpression of p21 prevented cell proliferation and clonogenic growth induced by 15-PGDH knockdown ( Figure 8B). These results demonstrate a key role of p21 in 15-PGDH-mediated inhibition of HCC cell growth.

DISCUSSION
Recent studies suggest an emerging role of 15-PGDH in several non-hepatic cancers (15)(16)(17)(18)(19)(20)(21)(22)(23)(25)(26)(27)(28)(29). Reduction of 15-PGDH is associated with enhanced cell proliferation and is an independent predictor for poor survival in gastric adenocarcinoma (27). A haplotype in the 15-PGDH gene is positively associated with colorectal cancer risk (20). In mouse models of colonic carcinogenesis, overexpression of 15-PGDH decreases cancer cell growth or delays tumor formation, whereas deletion of 15-PGDH increases susceptibility to chemically or genetically induced colon tumors (16). Targeted adenovirus-mediated delivery of 15-PGDH gene inhibited colon cancer growth in a mouse xenograft model (30). The hepatocyte growth factor (HGF) and its receptor c-Met signaling promotes PGE 2 biogenesis in colorectal cancer cells via up-regulation of COX-2 and down-regulation of 15-PGDH (31). Reciprocal regulation between COX-2 and 15-PGDH expression has been documented in several cancers (32). Omega-3 polyunsaturated fatty acids reduce the level of PGE 2 in hepatocellular carcinoma and cholangiocarcinoma cells through down-regulation of COX-2 and induction of 15-PGDH (33,34). Anti-cancer therapeutics, such as transforming growth factor (TGF)-β1, glucocorticoids and histone deacetylase (HDAC) inhibitors, have been shown to exert their anti-carcinogenic activity in part through induction of 15-PGDH expression (26,35). All of these findings suggest a tumor suppressive function of 15-PGDH. However, to date, the action of 15-PGDH is largely attributable to its degradation of biologically active PGE 2 , with its 15-keto metabolite being considered largely inactive. The current study provides paradigm-shifting evidence for an active role of 15-keto-PGE 2 in 15-PGDH-mediated inhibition of cancer cell growth. Our results reveal that 15-PGDH-derived 15-keto-PGE 2 is a natural PPARγ ligand which induces PPARγ association with p21 WAF1/Cip1 promoter and enhances p21 gene expression leading to inhibition of HCC growth (illustrated in Figure 9). p21 WAF1/Cip1 is a potent inhibitor of cyclin-dependent kinases (CDKs). It inhibits cell cycle progression through binding to cyclin-cdk complexes (36)(37)(38). Association of p21 to cyclincdk complexes also prevents phosphorylation of the retinoblastoma (RB) protein thus preventing the release of E2F transcription factor (39). In addition, p21 also binds to proliferating cell nuclear antigen (PCNA) and interferes with PCNA-dependent DNA polymerase activity leading to inhibition of DNA replication (40). The growth-inhibitory action of p21 is attributed to the functions of the carboxy-terminal PCNA-binding domain as well as the amino-terminal CDK-cyclin inhibitory domain (41,42). Consistent with the growth-inhibitory effect of p21, we have shown that 15-PGDH-derived 15-keto-PGE 2 induces p21 expression in HCC cells and this signaling pathway suppresses HCC cell growth. The role of p21 in 15-PGDH/15-keto-PGE 2 -mediated inhibition of HCC cell growth is attested by the observations that p21 knockdown reversed 15-PGDH-induced inhibition of tumor cell growth and that overexpression of p21 prevented the cell growth induced by 15-PGDH knockdown.
PPARγ is a ligand-activated nuclear transcription factor regulating the expression of target genes by binding to PPRE in target genes (43,44). The activity of PPARγ is regulated by several ligands, including thiazolidinediones (such as ciglitazone), 15-deoxy-Δ 12,14 prostaglandin J 2 (15d-PGJ 2 ), and other fatty acid derivatives. Our results in the current study document a key role of PPARγ in mediating 15-PGDH/15-keto-PGE 2 actions in HCC cells.
The results of this study depict a key 15-PGDH/15-keto-PGE 2 /PPARγ/p21 signaling axis that suppresses hepatocarcinogenesis and tumor progression. Since 15-PGDH converts oncogenic PGE 2 to tumor suppressive 15-keto-PGE 2 , induction of endogenous 15-PGDH expression or delivery of exogenous 15-PGDH/15-keto-PGE 2 may represent promising future therapeutic interventions. It is conceivable that this approach may provide more effective anti-tumor therapy with fewer side effects compared to the selective COX-2 inhibitors.

Cell lines
Human HCC cell lines (Huh7 and HepG2) and murine HCC cell line (Hepa1-6) were obtained from ATCC. The cells were maintained in Minimum Essential Medium (MEM) (Gibco BRL Life Technologies) supplemented with 10% heat-inactivated (56°C, 30 minutes) fetal bovine serum (Sigma) in a humidified atmosphere of 5% CO 2 incubator at 37°C.

Cell proliferation WST-1 assay
The cells were synchronized in G0 phase by serum deprivation and then released from growth arrest by reexposure to complete medium with serum. Cell proliferation was detected by reagent WST-1 kit (Roche) according to the manufacturer instruction. Cell growth curve was based on the normalized values of OD450 and each point represents the mean of three independent samples.

Cell cycle analysis
Cell cycle distribution was determined by using flow cytometry. 5×10 6 cells in a 10 cm dish were grown overnight. Then the cells were then kept in serum-deprived medium for 48 h for synchronization to G0 phase. The cells were released from growth arrest by reexposure to 10% fetal bovine serum for 24 h. Cells were collected by trypsinization followed by centrifugation, washed once with PBS, and resuspended in 0.2 ml of ice-cold PBS. The collected cells were fixed in 70% cold ethanol [in 50 mM glycine buffer (pH 2.0)] overnight at −20°C. 100 µg/ml RNase A (Qiagen) was added to the cells with incubation for 30 minutes at 37°C. The cells were resuspended in 0.5 ml 100µg/ml propidium iodide (PI) solution (Borhoringer Nannheim Co.) for staining. The stained cells were analyzed by a FACScan Flow Cytometer at the LCRC FACS core facility. The percentage of cells in S, G0/G1, and G2/M phases of the cell cycle was determined using EXPO32 Cell Quest software. All experiments were conducted in triplicate.
Soft agar colony formation assay 10 3 cells were plated on a 10-cm dish containing 0.5% (down) and 0.35% (up) double layer soft-agar. The dishes were incubated at 37°C in humidified incubator for 21 days. The cells were fed 1-2 times per week with DMEM (Gibco BRL Life Technologies). The colonies were stained with 2.5 ml of 0.005% Crystal Violet (Sigma) for more than 1 hour and the numbers of colonies were tallied.

Luciferase reporter assay
Cells (1 × 10 5 per well in six-well plate) were transiently transfected with 1 µg of luciferase construct (pGL3-PPRE or pGL3-WAF1/Cip1/p21 promoter) and 0.1 µg of pRL-Tk (Addgene) using Lipofectiamine TM 2000 (Invitrogen) (with additionally plasmids as indicated). 36 hours after transfection, the cells were harvested with lysis buffer and luciferase activities of the cell extracts were measured using the dual luciferase assay system (Promega). The luciferase activity was normalized for transfection efficiency with Renilla luciferase activity.

Xenograft tumor study in SCID mice
Four-week male athymic NOD CB17-prkdc/SCID (severe combined immunodeficiency) mice were purchased from Jackson laboratory and maintained in the animal facilities Lu et al.

Page 10
Oncogene. Author manuscript; available in PMC 2014 August 27. according to the protocol approved by the American Association for Accreditation of Laboratory Animal Care. Six athymic SCID mice per group were injected subcutaneously at the armpit area with Huh7 cells stably transfected with pCMV6-AC-GFP, pCMV6-AV-GFP-15PGDH, pGFP-V-RS, pGFP-V-RS-15PGDH, respectively (1×10 8 cells in 100µl of PBS). The mice were observed over 4 weeks and then sacrificed to recover the tumors. The wet weight of each tumor was determined. Portion of the tissue from each tumor was snapfrozen. Additional portion of each tumor was fixed in 4% paraformaldehyde and embedded in paraffin for hematoxylin and eosin (H&E) stain and for PCNA immunostain.

Xenograft tumor study in C57BL/6J mice
Mouse HCC cell line (Hepa1-6) (1×10 8 cells in 0.2 ml of PBS) was inoculated subcutaneously at armpit into syngenetic C57BL/6J mice. The mice were divided into two groups and subjected to intratumoral injection of pAd or pAd-15-PGDH (10 10 pfu). The diameters of the tumors were measured every three days and the tumor volume was calculated using the formula V=L/2*w 2 . The mice were sacrificed 31 days after inoculation to recover the tumor tissue. The wet weight of each tumor was determined for each mouse. A portion of the tissue from each tumor was snap-frozen. Additional portion of each tumor was fixed in 4% paraformaldehyde and embedded in paraffin for hematoxylin and eosin (H&E) stain and for PCNA immunostain.

Statistical analysis
The values are presented as mean±standard error of the mean (SEM) unless otherwise noted, with a minimum of three replicates. The results were evaluated by SPSS12.0 statistical soft and Student's t-test was used for comparisons, with p < 0.05 considered significant.        Western blotting for p21 in the nuclear fraction (nP21) and p21 in the cytoplasmic fraction (cP21) from Huh7 stable cell lines with altered 15-PGDH expression. Histone and β-actin were used as the internal control, respectively.
B. Co-immunoprecipitation and western blotting analysis using indicated antibodies in Huh7 stable cell lines with altered 15-PGDH expression. IgG IP was used as the negative control. PCNA western blotting was used as input control. C. Co-immunoprecipitation and western blotting analysis using indicated antibodies in Huh7 stable cell lines with altered 15-PGDH expression. IgG IP was used as the negative control. rIP denotes repeat co-immunoprecipitation. D. Co-immunoprecipitation and western blotting analysis using indicated antibodies in Huh7 stable cell lines with altered 15-PGDH expression. IgG IP was used as the negative control. rIP denotes repeat co-immunoprecipitation. E. Co-immunoprecipitation and western blotting analysis using indicated antibodies in Huh7 stable cell lines with altered 15-PGDH expression. IgG IP was used as the negative control.
Lu et al.

Figure 8. 15-PGDH inhibits HCC cell growth through p21 WAF1/Cip1
A. Huh7 cells transfected with 15-PGDH overexpression vector with or without cotransfection of p21 RNAi. a Western blotting for 15-PGDH and p21 (β-actin was used as loading control). b WST cell proliferation assay. Each sample was assayed in triplicate for 6 consecutive days. Data are means±SEM from three independent experiments (**p < 0.01). c Soft-agar colony formation assay. The results are representative of three independent experiments (**p < 0.01). B. Huh7 stable cell transfected with 15-PGDH RNAi vector with or without co-transfection of p21 expression vector. a Western blotting for 15-PGDH and p21 (β-actin was used as loading control). b WST cell proliferation assay. Each sample was assayed in triplicate for 6 days consecutively. Data are mean±SEM from three independent experiments (**p < 0.01). c Soft-agar colony formation assay. The results are representative of three independent experiments (**p < 0.01).