Pioglitazone Reduces Hepatocellular Carcinoma Development in Two Rodent Models of Cirrhosis

Background Hepatocellular carcinoma (HCC) is one of the deadliest malignancies worldwide due to the lack of effective treatments. Chemoprevention in high-risk patients is a promising, alternative strategy. In this study, pioglitazone was investigated for its ability to prevent hepatocarcinogenesis in two rodent models of cirrhosis. Methods In the first model, male Wistar rats were given repeated, low-dose injections of diethylnitrosamine (DEN) to accurately recapitulate the progression of fibrosis to cirrhosis and HCC. In the second model, a single dose of DEN was administered to male C57Bl/6 pups at day fifteen followed by administration of a choline-deficient, L-amino acid defined, high-fat diet (CDAHFD) at week six for 24 weeks. Pioglitazone treatment started at the first signs of fibrosis in both models. Results Pioglitazone effectively reduced fibrosis progression and HCC development in both models. Gross tumor nodules were significantly reduced after pioglitazone treatment (7.4 ± 1.6 vs. 16.6 ± 2.6 in the rat DEN model and 5.86 ± 1.82 vs. 13.2 ± 1.25 in the mouse DEN+CDAHFD model). In both models, pioglitazone reduced the activation of mitogen-activated protein kinase (MAPK) and upregulated the hepato-protective AMP-activated protein kinase (AMPK) pathway via increasing circulating adiponectin production. Conclusion Pioglitazone is an effective agent for chemoprevention in rodents and could be repurposed as a multi-targeted drug for delaying liver fibrosis and hepatocarcinogenesis. Electronic supplementary material The online version of this article (10.1007/s11605-018-4004-6) contains supplementary material, which is available to authorized users.


Introduction
Hepatocellular carcinoma (HCC) is the fifth most common cause of cancer, and the second leading cause of cancerrelated deaths globally. 1 Given the growing number of patients with chronic hepatitis C and the increasing risk of non-alcoholic fatty liver disease (NAFLD) as a result of the obesity epidemic, the incidence of HCC will continue to rise over the next few decades. 2,3 The 5-year survival rate for patients with HCC is abysmal and only 13% of diagnosed patients undergo a curative resection. 4 Given the high mortality rate associated with HCC as well as the readily available cohort of patients at risk for HCC development, specifically those with HBV, HCV, diabetes, and NAFLD, effective chemoprevention strategies could help improve survival by preventing or delaying the onset of cancer.
The ideal chemoprevention agent should be safe and readily available. There is growing evidence that anti-Electronic supplementary material The online version of this article (https://doi.org/10.1007/s11605-018-4004-6) contains supplementary material, which is available to authorized users. diabetic medications, such as metformin, are effective in halting HCC carcinogenesis in both in vitro and in vivo settings. 5,6 Thiazolidinediones (TZDs) including pioglitazone, troglitazone, and rosiglitazone are another major family of anti-diabetic medications and have also been purported to be anti-carcinogenic. TZDs are high-affinity ligands for the nuclear receptor peroxisome proliferatoractivated receptor gamma (PPARγ). 7 TZDs function to improve insulin sensitivity by regulating glucose homeostasis and fatty acid metabolism. The PIVENS trial demonstrated that pioglitazone treatment significantly reduced steatosis and inflammatory features such as hepatocellular ballooning in non-alcoholic steatohepatitis (NASH) patients. 8 PPARγ agonists have also been shown to decrease inflammation and tumorigenesis in multiple tumor models, such as colon, thyroid and lung cancers. 9 However, the role of PPARγ agonists in preventing HCC development is less clear. Yu et al. found an increased tumor burden in PPARγ deficient mice in a diethylnitrosamineinduced HCC model, while prophylactic treatment with rosiglitazone decreased the incidence of HCC. 10 In addition, a major risk factor for HCC development is ongoing fibrogenesis leading to cirrhosis, ultimately predisposing hepatocytes to malignant transformation. Pioglitazone and rosiglitazone have been shown to reduce in vitro activation of hepatic stellate cells, the cells responsible for extracellular matrix (ECM) deposition during hepatic fibrosis. 11,12 Despite pre-clinical data demonstrating a negative correlation between PPARγ expression level and tumorigenesis, few studies have evaluated the direct effects of TZDs on the fibrosis/cirrhosis/ carcinogenesis axis in the liver.
Therefore, in the following study, we explored the effects of one of the most commonly prescribed TZDs, pioglitazone, on fibrosis progression and HCC development in two rodent models of cirrhosis.

Chemicals
Pioglitazone was freshly dissolved in 0.5% methylcellulose before administration.

Animal Experiments
Animal experiments were approved by the Massachusetts General Hospital Institutional Animal Care and Use Committee and all animals received humane care according to the criteria outlined in the Guide for the Care and Use of Laboratory Animals of the National Academy of Sciences. Male Wistar rats (Charles River Laboratories, Wilmington, MA) were subjected to either weekly intraperitoneal injections of PBS (n = 6) or 50 mg/kg diethylnitrosamine (DEN) (n = 18) (Sigma, St. Louis, MO) for 18 weeks. DEN is an environmental carcinogen that produces mutagenic DNA adducts, leading to gene sequence mismatches, ultimately predisposing the liver to HCC formation. 13 Based on our prior studies, repeated, low-dose DEN administration accurately models chronic liver disease progression in humans with development of fibrosis at 8 weeks, early cirrhosis at 12 weeks, and fulminant cirrhosis and carcinogenesis at 18 weeks. 14 After 8 weeks of injections, DEN-injured rats began treatment via oral gavage with pioglitazone 3 mg/kg or vehicle (0.5% methylcellulose) for the remainder of the study (n = 9 for each group).
A single injection of DEN to developing mice on day 15 is a common method to generate HCCs in 9-12 months although usually in the setting of a normal liver. Recently, a choline-deficient, L-amino acid-defined, high-fat diet consisting of 60 kcal% fat and 0.1% methionine by weight (CDAHFD) has been shown to recapitulate features of NASH. 15 We combined these approaches to accelerate the process of carcinogenesis in the setting of a NASH cirrhosis background from 52 weeks in CDAHFD alone 16 to 24 weeks with DEN + CDAHFD. Thus, male C57Bl/6 mice (Charles River Laboratories) received a single dose of 35 mg/kg DEN at day 15. At 6 weeks of age, mice were subjected to either standard chow (n = 5) or CDAHFD (n = 16) for a total of 24 weeks. Oral gavage of either vehicle control (n = 8) or pioglitazone 10 mg/kg (n = 8) was initiated 6 weeks following the onset of CDAHFD for the remainder of the study.
A human pioglitazone dose of 45 mg per day was converted to a rat dose of 3 mg/kg and a mouse dose of 10 mg/kg based on a body surface area calculation. For both models, animals were anesthetized and sedated at the time of sacrifice. A terminal blood collection was performed by cardiac puncture. Livers were weighed, fixed in formalin, and snap frozen for further analysis.

Serum Laboratory Analysis
Blood was allowed to clot for 1 h at room temperature before centrifugation at 2000 rpm for 10 min at 4°C. Serum was isolated and stored at − 80°C prior to use. Serum levels of alkaline phosphatase (ALP), aspartate transaminase (AST), alanine aminotransferase (ALT), total bilirubin (TBIL), and Gamma-glutamyl transferase (GGT) were measured.

Histology, Immunohistochemistry, Immunofluorescence
Formalin-fixed samples were embedded in paraffin, cut in 5 μm thick sections and stained with hematoxylin & eosin (H&E) or Sirius red. The collagen proportional area (CPA) was morphometrically quantified on Sirius red-stained sections with image processing software (Image J, NIH). All slides were blindly reviewed by an independent liver pathologist to calculate the NAFLD activity score (NAS) and the fibrosis score. Additional sections in the DEN model were stained with an antibody specific for proliferating cell nuclear antigen (PCNA) (Cell Signaling Technology, Danvers, MA). For immunofluorescence, sections were stained for α-SMA (Abcam, Cambridge, MA) with detection by appropriate secondary antibodies labeled with either Cy3 or Alexa488 according to the manufacturer's instructions.

ELISA
Adiponectin protein expression was quantified in serum using a commercially available ELISA (Millipore, Burlington, MA) according to the manufacturer's instructions. Each sample was quantified in duplicate and the experiments were repeated in both animal models to ensure accuracy and reproducibility. 8-Hydroxydeoxyguanosine (8-OHdG) levels were quantified in tissue using a commercially available ELISA system (Cell Biolabs, San Diego, CA).

Statistical Analysis
Data are represented as mean ± standard deviation. An unpaired two-tailed t test was used to compare differences between groups.

Pioglitazone Reduces Carcinogenesis in the DEN Rat Model
Weekly administration of DEN to rats resulted in a significant reduction in body weight (398.1 ± 22.7 vs. 598.6 ± 20.73; p < 0.01), and an increase in the liver/body weight ratio (5.6 ± 0.34 vs. 2.9 ± 0.06; p < 0.01) as a result of increased tumor burden. Pioglitazone administration led to an improved body weight (427.5 ± 18.7 vs. 398.1 ± 22.7) and a significant reduction in liver body weight ratio (4.5 ± 0.21 vs. 5.6 ± 0.34; p < 0.05) (Fig. 1a, b). Repeated DEN administration also predictably resulted in liver failure as measured by liver function tests, and pioglitazone treatment led to a significant improvement in total bilirubin (1.2 ± 0.42 vs. 3.6 ± 0.83; p < 0.05) (Supplementary Fig. 1a-e).
Macroscopically, the livers from rats treated with the PBS vehicle control appeared normal, while livers from DENinjured rats were cirrhotic and contained many HCCs ( Fig.  1c top panel). Importantly, pioglitazone treatment resulted in a 55% reduction in the number of gross tumor nodules (7.4 ± 1.6 vs. 16.6 ± 2.6; p < 0.01) (Fig. 1d). Specifically, pioglitazone treatment significantly reduced the appearance of small (< 8 mm) tumor nodules (15.1 ± 2.5 vs. 6.3 ± 1.5; p < 0.01) (Fig. 1e), but did not have a significant effect on the number of large (> 8 mm) tumor nodules (Fig. 1f). Microscopically, H&E slides from DEN-injured rats contained tumors as well as peri-tumoral extracellular matrix deposition ( Fig.  1c middle panel). Proliferating cell nuclear antigen (PCNA) is a marker for active cellular proliferation. 17 DEN-injured rats had increased PCNA staining in comparison to pioglitazone treated rats (8 ± 1.4 vs. 3.8 ± 0.5; p < 0.5) (Fig. 1c bottom panel and Fig. 1g). The antitumor effects of pioglitazone were not attributable to a drug interaction or alterations in DEN metabolism, as liver levels of 8-hydroxydeoxyguanosine (8-OHdG), a marker of oxidative stress-derived DNA damage often used as a surrogate for DEN activity, 18 were similar between DENinjured rats treated with vehicle control or pioglitazone (0.22 ± 0.02 vs. 0.22 ± 0.02) ( Supplementary Fig. 2).

Pioglitazone Reduces Proliferation via Activation of the AMPK Pathway and Downregulation of the MAPK Pathway
Serum adiponectin levels are known to decrease in patients with cirrhosis and HCC as compared to healthy controls. 19 We similarly observed decreased serum adiponectin levels in DEN-injured rats as compared to healthy PBS animals (37.3 ± 7 vs 62.9 ± 1.1; p < 0.05). Treatment with pioglitazone normalized serum adiponectin levels (74.8 ± 5.63 vs. 37.3 ± 7.3; p < 0.01) (Fig. 3a). Adiponectin is known to activate AMPactivated protein kinase (AMPK), a serine/threonine protein kinase that regulates cellular energy homeostasis. 20 Activated AMPK can act as a tumor suppressor by directly inducing cell-cycle arrest. 21 We observed decreased phosphorylated AMPK in the livers of DEN-injured rats as compared to livers from PBS animals. Pioglitazone treatment normalized liver levels of phosphorylated AMPK similar to PBS animals. In addition, treatment with pioglitazone reduced multiple mitogen-activated protein kinase (MAPK) pathways, including reductions in phosphorylated extracellular signalregulated kinase (pERK), c-JUN N-terminal kinase (pJNK), and its downstream target phosphorylated c-JUN. By comparison, no changes in the level of phosphorylated-P38 MAPK were observed between the groups (Fig. 3b).
The combination of DEN+CDAHFD resulted in lipid laden HCCs or HCC nodules surrounded by fatty deposition (Fig.  4a bottom panel). Liver/body weight ratio was used as a tumor burden surrogate. As compared to DEN+CDAHFD mice treated with vehicle, pioglitazone treatment decreased the liver/body weight ratio (6.75 ± 0.65 vs. 11.22 ± 0.49; p < 0.01). Grossly, we measured all tumors greater than 1 mm, and found that pioglitazone treatment resulted in a significant reduction in the number of surface tumors (5.86 ± 1.82 vs. 13.2 ± 1.25; p < 0.01). In addition, the entire left lobe of each animal was formalin-fixed and H&E-stained for analysis of microscopic nodules by a blinded liver pathologist. Pioglitazone treatment resulted in a significant reduction in histological tumor nodules as well (0.83 ± 0.31 vs 2.6 ± 0.5; p < 0.05) (Fig. 4c-e).

Pathological Assessment of the Mouse DEN+CDAHFD Model
The NAFLD activity score (NAS) (7.3 ± 0.48 vs 8 ± 0; p < 0.05) and fibrosis scores (3.5 ± 0.22 vs 4 ± 0; p < 0.05) were significantly reduced with pioglitazone treatment. The NAS score comprises steatosis, lobular inflammation as well as hepatocyte ballooning. Notably, pioglitazone decreased steatosis without reducing inflammation or hepatocyte ballooning. To verify this former finding, the area of lipid vacuolization was morphometrically measured and a significant reduction of fat deposition was seen with pioglitazone administration (8.4 ± 0.84 vs. 11.6 vs. 0.9; p < 0.08), consistent with the reduced steatosis observed within the NAS scoring criteria (Fig. 5a-c).

Pioglitazone Increases AMPK Activation and Decreases MAPK Signaling
In this NASH model of HCC, we observed a significant reduction in serum adiponectin levels as compared to mice fed standard chow (10.5 ± 0.79 vs. 17.3 ± 0.94; p < 0.01), and pioglitazone administration increased circulating serum adiponectin in DEN+CDAHFD mice (29.7 ± 3.2 vs. 10.5 ± 0.79; p < 0.01) (Fig. 7a). In addition, DEN+ CDAHFD mice treated with pioglitazone had increased activation of AMPK, as well as decreased activation of ERK, JNK, and its downstream target c-JUN. As observed in the rat DEN model, no changes in the level of phosphorylated P38 were seen between groups in the DEN+CDAHFD model as well. Lastly, activated AMPK phosphorylates acetyl coA carboxylase (ACC), thus inactivating this rate-limiting step for fatty acid synthesis. Pioglitazone treatment increased phosphorylated ACC and this may have resulted in the observed decrease in steatosis seen within the NAS scoring criteria as well as the lipid vacuolization quantification (Fig. 7b).

Discussion
Underlying cirrhosis is associated with 80-90% of patients with primary HCC 22 and thus at risk patients are easily identifiable unlike many other malignancies. With the rising incidence of obesity and diabetes, NAFLD/NASH-related hepatic fibrosis/cirrhosis will likely become the most common cause of HCC in the future. 23 Thus, there is increased interest in the use of easily accessible and inexpensive medications, like anti-diabetic drugs, as chemopreventive strategies.
In this study, the administration of pioglitazone at the onset of fibrosis in both animal models resembles Bprimary chemoprevention^, the administration of an agent to patients without overt disease but with known risk factors. 24 The low- The non-alcoholic fatty liver disease (NAFLD) activity score (NAS) and l the NASH fibrosis score were scored by a blinded liver pathologist. m Lipid vacuolization (LV) was morphometrically calculated using image J software. # p < 0.05 and ## p < 0.01 compared to standard chow. * p < 0.05 and ** p < 0.01 compared to DEN+ CDAHFD dose, repeated DEN rat model was used given its similarity at the histologic and transcriptomic level to human cirrhosis. 25 We observed a significant reduction in tumor nodules in the rat DEN model after treatment with pioglitazone. This effect was specific to smaller nodules (< 8 mm) suggesting that pioglitazone prevented the development of new HCCs, but had no effect on the growth of established tumors. Given its use as an anti-diabetic medication, we also tested pioglitazone in a mouse NASH-HCC model. We also observed decreased tumor incidence when pioglitazone was used to treat mice subjected to a single dose of DEN followed by a long-term feeding of CDAHFD. Another piece of evidence supporting the preventive effects of pioglitazone is the significant reduction of underlying fibrosis/cirrhosis. The pathogenesis of HCC in a fibrotic/ cirrhotic background is still unclear. The discrepancy lies in the unsettled question of whether fibrogenesis promotes HCC carcinogenesis or if the fibrosis is a byproduct of chronic inflammation and liver regeneration. 26 There is growing evidence that extracellular matrix deposition promotes carcinogenesis through the phosphoinositide 3 kinase (PI3K) and mitogen-activated protein kinase (MAPK) signaling cascades. 27 The contribution of chronic inflammation in HCC development has been well examined. 28 Overall, we saw a significant reduction of fibrosis in both the rat DEN model and the mouse DEN+CDAHFD model using morphometric quantification of collagen deposition as well as the expression of several pro-fibrotic genes. Pro-Inflammatory cytokine expression was also decreased by pioglitazone in the DEN+CDAHFD mouse model. This effect was not seen in the rat DEN model. This might be due to the fact that the inflammatory insult is constant in the DEN+CDAHFD mouse model, but the inflammatory insult in the DEN model is given on a weekly basis. In the latter case the animals were sacrificed 1 week after the last DEN injection allowing time for the inflammatory response to subside on its own.

Chow
Pioglitazone targets PPARγ which belongs to a family of nuclear hormone receptors, with three different isoforms, PPAR alpha, gamma, and beta/delta. 29 TZDs, such as pioglitazone, activate PPARγ which is predominately expressed in adipose tissue. Activated PPARγ mediates the release of adiponectin from adipocytes which has several important functions including improving insulin-sensitization and reducing inflammation. 30 We observed that pioglitazone increased adiponectin release which activated AMPK in the liver leading to a downregulation of the MAPK pathways (mainly ERK/JNK/cJUN). These results were remarkably consistent in both the rat DEN and mouse DEN+CDAHFD models. In our NASH cirrhosis model, we also observed that CDAHFD activated Acetyl-CoA carboxylase (ACC), the rate-  Fig. 7 Pioglitazone increases serum adiponectin and liver AMPK activation and decreases liver MAPK signaling in the mouse DEN + CDAHFD model. a Serum adiponectin was measured in the mouse DEN+CDAHFD model. b Western blot analysis of phosphorylated (Ser79) Acetyl-CoA Carboxylase (pACC)/total ACC, phosphorylated (Thr172) 5′ adenosine monophosphate-activated protein kinase (pAMPK)/total AMPK, phosphorylated (Thr202/Tyr204)-p44/42 mitogen-activated protein kinase MAPK (pErk1/2)/ total Erk1/2, phosphorylated (Thr183/Tyr185) c-Jun N-terminal kinase (pSAPK/ JNK)/ total SAPK/JNK, phosphorylated (Ser73) c-Jun/total c-Jun, phosphorylated (Thr180/Tyr182) p38 mitogen-activated protein kinase (pP38)/total P38. Actin was used as a loading control. # p < 0.05 and ## p < 0.01 compared to PBS or standard chow. * p < 0.05 and ** p < 0.01 compared to DEN or DEN+CDAHFD limiting step of fatty acid synthesis, and treatment with pioglitazone decreased ACC activation and steatosis as seen with the reduction in the NAS score and quantification of lipid vacuolization.
One limitation of this study is the use of only male animals. It is well documented that female rodents are relatively protected from both fibrosis progression and HCC development. 31,32 The gender-related incidence of HCC carcinogenesis has been reported to be largely due to the inhibitory effect of estrogen-dependent IL-6 production. 33 Globally, HCC occurs more frequently in men than in women, with a male to female ratio ranging between 8:1 and 2:1. 34 Women have also been shown to have less aggressive HCC tumors than men at initial diagnosis. 35 However, underrepresentation of females in preclinical studies is a major concern with respect to better understanding the role of female biology on cancer development. Future endeavors are needed to better address gender discrepancy and develop female animal models of HCC.
In pre-diabetes or patients with type-2 diabetes and biopsy-proven NASH, long-term pioglitazone treatment resulted in a 51% resolution of NASH and improvement in fibrosis score as well as insulin sensitivity. 36 More recent results suggest that pioglitazone reduces advanced fibrosis in NASH even in patients without diabetes. 37 However, its use has been limited at least in part due to concerns about increased bladder cancer risk in patients who receive long-term pioglitazone treatment. 38 These concerns may be overstated given the overall low risk of bladder cancer in patients with diabetes as well as the much greater beneficial effects of pioglitazone on cardiovascular disease and NASH. 39 Our results in two animal models are consistent with the data from human studies demonstrating reduced fibrosis after pioglitazone treatment and suggest that pioglitazone should be resurrected as a treatment for NASH and even further should be investigated as an HCC chemopreventive strategy in prospective trials, especially given the approximately 3-13% annual incidence for HCC development in patients with NASH-related cirrhosis. 40

Conclusion
This study is one of the first to examine the effects of pioglitazone as a chemopreventive drug in animal models. Our results suggest that besides for its insulin sensitivity functions, pioglitazone reduces fibrosis and could be repurposed as an agent for HCC chemoprevention especially in the setting of NASH. Given its good safety profile and its frequent use, pioglitazone could be further explored in the setting of clinical trials in combination with other HCC chemopreventive agents.
Funding This work was supported by the National Institutes of Health (grant numbers T32CA071345 to SL and CA140861 and DK104956 to BCF).
Open Access This article is distributed under the terms of the Creative Comm ons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.