Accepted Manuscript

SIRT1 regulates oncogenesis via a mutant p53-dependent pathway in hepato-cellular carcinoma

Zheng Yun Zhang, Doopyo Hong, Seung Hoon Nam, Jong Man Kim, Yong HanPaik, Jae Won Joh, Choon Hyuck David Kwon, Jae Berm Park, Gyu-SeongChoi, Kyu Yun Jang, Cheol Keun Park, Sung Joo Kim

PII: S0168-8278(14)00546-7DOI: http://dx.doi.org/10.1016/j.jhep.2014.08.007Reference: JHEPAT 5284

To appear in: Journal of Hepatology

Received Date: 28 January 2014Revised Date: 23 July 2014Accepted Date: 7 August 2014

Please cite this article as: Zhang, Z.Y., Hong, D., Nam, S.H., Kim, J.M., Paik, Y.H., Joh, J.W., Kwon, C.H.D., Park,J.B., Choi, G-S., Jang, K.Y., Park, C.K., Kim, S.J., SIRT1 regulates oncogenesis via a mutant p53-dependentpathway in hepatocellular carcinoma, Journal of Hepatology (2014), doi: http://dx.doi.org/10.1016/j.jhep.2014.08.007

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customerswe are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, andreview of the resulting proof before it is published in its final form. Please note that during the production processerrors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

1

SIRT1 regulates oncogenesis via a mutant p53-dependent pathway in

hepatocellular carcinoma

Zheng Yun Zhang MD, PhD1,2,*; Doopyo Hong PhD2,*; Seung Hoon Nam Ms2;

Jong Man Kim MD4; Yong Han Paik MD, PhD5; Jae Won Joh MD4; Choon

Hyuck David Kwon MD4; Jae Berm Park MD4; Gyu-Seong Choi MD4; Kyu Yun

Jang MD, PhD6; Cheol Keun Park MD, PhD7; Sung Joo Kim, MD, PhD2,3,4

1Department of Surgery, Shanghai Jiao Tong University Affiliated Sixth

People's Hospital, Shanghai, China

2Transplantation Research Center, Samsung Medical Center, Samsung

Biomedical Research Institute, Sungkyunkwan University School of Medicine,

Seoul, Republic of Korea

3Sarcoma Research Center, Samsung Medical Center, Samsung Biomedical

Research Institute, Sungkyunkwan University School of Medicine, Seoul,

Republic of Korea

4Department of Surgery, Samsung Medical Center, Samsung Biomedical

Research Institute, Sungkyunkwan University School of Medicine, Seoul,

Republic of Korea

5Department of Medicine, Samsung Medical Center, Samsung Biomedical

Research Institute,, Sungkyunkwan University School of Medicine, Seoul,

Republic of Korea

6Department of Pathology, Chonbuk National University Medical School and

2

Research Institute of Clinical Medicine, Jeonju, Republic of Korea

7Department of Pathology, Samsung Medical Center, Samsung Biomedical

Research Institute, Sungkyunkwan University School of Medicine, Seoul,

Republic of Korea

*Zheng Yun Zhang and Doopyo Hong contributed equally to this work.

Key words: SIRT1; AMPK; mTOR; metformin; p53

Corresponding author: Sung Joo Kim, MD, PhD

E-mail: kmhyj111@hotmail.com

Telephone: +82-2-2148-7308

Fax: +82-2-2148-7379

Address reprint requests to Sung Joo Kim, MD, PhD. Department of Surgery,

Samsung Medical Center, Sungkyunkwan University School of Medicine, 50

Ilwon-dong, Gangnam-gu, Seoul 135-710, Republic of Korea.

Co-corresponding author: Cheol Keun Park PhD MD

E-mail: ckpark@skku.edu

Telephone: +82-2-3410-2766

Fax: +82-2-2148-7379

Address reprint requests to Sung Joo Kim, MD, PhD. Department of

Pathology,

3

Samsung Medical Center, Sungkyunkwan University School of Medicine, 50

Ilwon-dong, Gangnam-gu, Seoul 135-710, Republic of Korea.

Electronic word count: 5364 words (abstract; 247 words)

Number of figures and tables: 4 figures and 3 tables

.

The authors have no conflicts of interest.

Financial support: this study was supported by a Sungkyunkwan University

grant, Research Fund (GT10), 2012 and Shanghai Pujiang

Program (14PJ1407300).

4

Abstract

Background & Aims: SIRT1 is a class III histone deacetylase that plays

diverse roles in various cancers. However, the clinical significance of SIRT1 in

hepatocellular carcinoma (HCC) remains unknown.

Methods: We analyzed p53 mutations and the activation of SIRT1 in 252

hepatitis B virus-positive HCC cases. None of the patients had been subjected

to pre-operative treatment.

Results: We assessed 57 p53 mutations from 248 HCC tissues. Activated

SIRT1 (phosphorylated form of Ser47) predicted a longer RFS but not a

longer OS with mutant p53 (RFS: P = 0.007, OS: P=0.280). In multivariate

analysis, activated SIRT1 remained a significant predictor of longer RFS

(OR=0.307, CI: 0.143-0.660, P=0.002). Analysis of 248 paired specimens

revealed a significant correlation between activated SIRT1 (Ser47) and

activated AMPK (Thr172) in HCC tissues harboring mutant p53 (P=0.003, n =

57). The combination of these 2 parameters was a powerful predictor for a

good prognosis in such patients. In in vitro studies, SIRT1 inactivation

stimulated the growth of HCC cells bearing mutated p53 by suppressing

AMPK activity and subsequently enhancing mammalian target of rapamycin

(mTOR) activity, resulting in induction of p70S6K1 activation in HCC cells.

Metformin, an AMPK activator, more strongly suppressed cell growth in p53-

mutant cell lines with inactive SIRT1 than in p53-mutant cell lines with active

SIRT1.

Conclusions: SIRT1 exerted anti-carcinogenic effects via the AMPK-mTOR

pathway in HCC in the context of mutant p53. Metformin could be a

therapeutic drug for HCC in patients with mutated p53, inactivated SIRT1, and

AMPK expression.

5

Introduction

Hepatocellular carcinoma (HCC) is the fifth most common malignant

tumor worldwide and the third most common cause of cancer-related deaths

[1]. Despite the establishment of liver resection as a first-line treatment for

HCC, patients with HCC still face a high incidence of postoperative recurrence

and unsatisfactory survival rates. Consequently, identification of the biological

molecules and pathways involved in the oncogenesis of HCC remains a major

focus of researchers in this field.

Silent mating-type information regulation 2 homolog 1 (SIRT1), an

NAD+-dependent histone deacetylase, mediates cell survival via the

deacetylation of key cell cycle molecules and apoptosis regulatory proteins [2,

3]. SIRT1 deacetylase is also involved in the cell stress response and can

directly deacetylate p53, a critical tumor suppressor protein associated with

the DNA damage repair system and cell cycle arrest in response to DNA

damage. SIRT1 has also been reported to promote telomere maintenance via

the regulation of telomerase and the sheltering complex (4) and plays an

important role in genome stability [4, 5, 6, 7]. Recently, SIRT1 has been

shown to act as an oncoprotein in many cancers, including B-cell lymphoma,

ovarian cancer, and breast cancer. In addition, SIRT1 has been reported to

function as a tumor promoter in HCC [8, 9, 10, 11, 12, 13, 14].

AMP-activated protein kinase (AMPK) functions as a type of biosensor

in living cells, protecting cells from stresses such as energy depletion and

6

maintaining energy homeostasis via the down-regulation of catabolism [5].

Through these activities, AMPK is generally thought to be vital for proper

metabolism in living organs and has been reported to act as a tumor

suppressor via a pathway involving liver kinase B1 (LKB1). However, the role

of SIRT1 in AMPK signaling and the clinical relevance of SIRT1/AMPK

signaling in HCC remains unknown.

This study sought to identify the relevance of SIRT1, AMPK, and p53

in HCC using in vitro, in vivo and clinical studies. Our data show that SIRT1

inhibits the progression of HCC via AMPK in the context of mutant p53 and

may therefore serve as a prognostic factor for HCC patients with mutant p53.

Materials and methods

Detailed materials and methods are described in the Supplementary

Materials and Methods.

Results

SIRT1 and AMPK activity in primary HCC and paired adjacent nontumor

liver tissue

To identify differences in the activation of SIRT1 and AMPK in primary

HCC and paired adjacent nontumor tissues, we performed western blot for the

phosphorylation of Ser47 in SIRT1, which has been suggested to mediate the

activation of SIRT1 [15, 16], and for Thr172 in AMPK. More phosphorylated

7

SIRT1 and less phosphorylated AMPK were observed in tumor tissues than in

nontumor tissues (from N2 to T; Supplementary Fig. 1). These results suggest

that SIRT1 activity may be increased, whereas AMPK activity is decreased,

during carcinogenesis.

SIRT1 expression and activity in HCC and nontumor tissues

To further establish the significance of SIRT1 expression and activity

in HCC, we measured SIRT1 expression and activity by western blotting in

primary HCC (T) and paired adjacent nontumor liver tissues (N2 tissues,

abbreviated as N below) in patient group 2 (n = 130; Supplementary Fig. 2A).

Approximately one-half of the examined cases (53.8%) had SIRT1 expression

in HCC tissue, and this level was significantly higher than that observed in

nontumor tissues (7.7%; P < 0.001). Similarly, a significantly higher proportion

of cases (38.5%) demonstrated activated SIRT1 in tumor tissues than in

nontumor tissues (6.2%; P < 0.001; Supplementary Fig. 2B). In particular, 10

of 130 nontumor tissues exhibited SIRT1 expression, among which 8 tissues

(80%) were positive for activated SIRT1. In contrast, 70 of 130 tumor tissues

exhibited SIRT1 expression, among which 50 tissues (71.4%) were positive

for activated SIRT1, suggesting that activated SIRT1 is correlated with SIRT1

expression in nontumor and HCC tissues (Supplementary Fig. 2C). In addition,

there were significantly higher expression levels of activated SIRT1 in HCC

than in nontumor tissues (Supplementary Fig. 2D).

8

Next, large-scale analysis of SIRT1 expression and activity was

performed using western blots of an additional 122 resectable HCC and

paired adjacent nontumor liver tissues from patient group 3. The

clinicopathological parameters influencing overall and relapse-free survival

(OS and RFS, respectively) in these HCC patients (n =252) are summarized

in Supplementary Table 1. No prognostic significance of SIRT1 expression or

activation was found (SIRT1 expression: P = 0.153 for OS and P = 0.932 for

RFS; activated SIRT1: P = 0.717 for OS and P = 0.795 for RFS;

Supplementary Fig. 3). Multivariate analysis of the clinicopathological

parameters influencing OS and RFS are shown in Supplementary Table 2.

These data support the results of previous studies and suggest that SIRT1

has no significant effect on survival in HCC patients.

Identification of p53 mutations in HCC and nontumor tissues

To further understand the role of SIRT1 activation in HCC, we

examined p53 mutations in our HCC samples because SIRT1 has been

shown to mediate p53 signaling via deacetylation [3].

Detailed information describing these p53 mutations is provided in

Supplementary Table 3. Fifty-eight of 252 (23.0%) HCC tissues harbored p53

mutations, whereas only 4 of 252 (1.6%) nontumor tissues demonstrated

mutated p53 (Supplementary Table 4), indicating that p53 mutations are more

frequent in HCC than nontumor tissues. Among the 4 cases, 3 cases had p53

9

mutation only in the nontumor tissues, but not in the HCC tissues. However

one case had p53 mutation in both of nontumor and HCC tissue, suspected to

have Li Fraumeni Syndrome with no information. All about 4 mutations, we

regarded that nontumor tissues were contaminated with tumor cells.

Therefore, we excluded these 4 cases from this study thereafter (total n = 248,

57 have p53 mutation in tumor tissues).

Identification of p53 loss of function

To identify mutations leading to the loss of p53 function, we analyzed

the identified p53 mutations using the p53.free.fr database, which is available

at http://p53.free.fr, and the MUT-TP53 program (Supplementary Table 3). To

define the effects of these mutations on the biological activity of p53, we

performed western blotting for p21 expression, the most important p53 target.

Among 57HCC tissues with p53 mutations, we failed to detect p21 expression

in all but 1 case (Supplementary Fig. 4). These data validate the notion that

tissues with p53 mutations have no p53 function or biological activity,

indicating that these 57 HCC tissues with p53 mutations were useful used for

further studies.

p53 mutations are associated with SIRT1 and predict patient survival

The clinicopathological characteristics of 248 HCC cases according to

p53 mutation status are summarized in Supplementary Table 5. The

10

clinicopathological characteristics of patients with p53 mutations did not

significantly differ from those with wild-type p53. However, a significantly

higher proportion of tissues with mutant p53 demonstrated activated SIRT1

than did tissues with wild type p53 (63.2% vs. 31.4%, respectively, P < 0.001;

Table 1), suggesting that activated SIRT1 was correlated with p53 mutation

status. Next, we performed survival analysis by stratifying the cases according

to p53 mutational status, and the results showed that activated SIRT1 did not

predict OS or RFS in cases with wild-type p53 (P = 0.565 and P = 0.330,

respectively, n = 191; Fig. 1A,B). However, activation of SIRT1 predicted

longer RFS but not OS in cases with mutant p53 (RFS: P = 0.007, n = 57; OS:

P = 0.280, n = 57; Fig. 1C, D). In univariate and multivariate Cox regression

analysis, activated SIRT1 in HCC independently predicted a lower risk of

disease recurrence in p53 mutant cases (odds ratio [OR]: 0.307; 95%

confidence interval [CI]: 0.143 – 0.660; P = 0.002; Table 2). However,

activated SIRT1 did not serve as an independent risk factor for patients with

wild-type p53 (Supplementary Table 6). Moreover, activated SIRT1 was

negatively associated with satellite nodules in HCC in the context of mutant

p53 (OR: -0.291, P=0.028; Supplementary Table 7), and neither p53 mutation

nor activated SIRT1 was correlated with disease stage (Supplementary Table

8).

AMPK expression and activity in HCC and nontumor tissues

11

AMPK has been shown to act as a tumor suppressor and be

associated with SIRT1 signaling [16]. Therefore, to analyze the correlation

between SIRT1 and AMPK in HCC formation, we examined AMPK expression

and activity in HCC tissues (n = 248). Western blotting analysis indicated that

100% of the cases had positive AMPK expression in both nontumor and HCC

tissues (Fig. 2A). Moreover, a significantly higher proportion of nontumor

tissues exhibited activated AMPK (97.0% versus 47.2% in nontumoral tissues

versus HCC tissues, respectively; P < 0.001; Fig. 2B). In addition, the relative

expression level of total AMPK in nontumor tissue was significantly higher

than that in HCC tissues (Fig. 2C).

Analysis of 248 paired specimens revealed a strong correlation between

activated SIRT1 and activated AMPK in HCC tissues harboring mutant p53 (n

= 57) but not in HCC tissues with wild-type p53 (n = 191; Table 3). Moreover,

HCC patients with mutant p53, activated SIRT1, and activated AMPK

demonstrated significantly better prognoses in terms of RFS than patients

with inactive SIRT1 and AMPK (Fig. 1E). In patients with wild-type p53,

disease outcomes were not noticeably different in pairwise comparisons (Fig.

1F). These data suggest that simple evaluation of SIRT1 and AMPK activation

may allow for more accurate predictions for the prognosis of HCC patients

with mutant p53.

The role of SIRT1 in the p53 pathway in HCC cells

12

To confirm our clinical outcomes, we conducted in vitro and in vivo

studies using HepG2 HCC cells, which express wild-type p53, and PLC5 HCC

cells, which express mutant p53. We also used the p53-null HCC cell line

Hep3B to confirm our results following transfection with wild-type or mutant

p53. SIRT1 silencing by shRNA resulted in the acetylation of p53 in HepG2

(HepG2-shSIRT1) and PLC5 (PLC5-shSIRT1) cells. In HepG2 cells, p21 was

upregulated, demonstrating that p53 signaling was activated, whereas p21

was not induced in PLC5 cells despite the activation of p53. To further confirm

the above findings, Hep3B cells were transfected with a retrovirus expressing

wild-type or mutant (H168R) p53 (p53WT or p53MUT, respectively) and a

lentivirus containing shSIRT1. Consistent with our previous results, p53

signaling induced p21 expression in Hep3B-shSIRT1-p53WT cells but not in

Hep3B-shSIRT1-p53MUT cells despite p53 acetylation as mediated by SIRT1

knockdown (Supplementary Figs. 5A and 7A). Because knockdown of SIRT1

does not provide a full explanation for the role of SIRT1 in p53 signaling, we

over-expressed the SIRT1 gene in all of the cell lines. In contrast to SIRT1

knockdown in HepG2 cells, SIRT1 over-expression in HepG2 cells (HepG2-

hSIRT1) yielded different results: although SIRT1 was activated, p53

activation was unchanged, and p21 retained its activation compared with in

scramble transfected HepG2 cells. By contrast in PLC5 cells (PLC5-hSIRT1),

the acetylated form of p53 was present at lower levels when comparing with

scramble transfected cells (PLC5-scramble), and p21 was not detected in

13

either cells. In Hep3B (Hep3B-hSIRT1) cells, the activation of p53 and p21

was below detectable levels.(Supplementary Fig. 6B). These results show

that wild type p53 has bypass signaling from the deacetylation activity of

SIRT1, repressed conditions generated by SIRT1 over-expression, in addition

this was confirmed by the increasing level of p21 expression in HepG2 cells.

However, the results of SIRT1 overexpression in mutant p53 cell lines agree

with results from knockdown cells. From these results, SIRT1 knockdown or

over-expression results in the alteration of p53 acetylation, suggesting that

SIRT1 acts as a deacetylase regardless of the p53 genotype.

The role of SIRT1 in the AMPK-mTOR pathway

AMPK is an important protein involved in sensing and mediating

metabolic stress and has been shown to play a crucial regulatory role in

normal and malignant cells. Moreover, activation of AMPK has been shown to

suppress mammalian target of rapamycin (mTOR) signaling and negatively

control malignant transformation and cell proliferation [17, 18, 19]. We found

that silencing SIRT1 results in a decrease in phosphorylated AMPK and an

increase in mTOR and p70S6K1 phosphorylation in PLC5 cells. In contrast,

silencing SIRT1 in HepG2 cells (expressing wild-type p53) results in

increased AMPK phosphorylation and leads to a reduction in the phospho-

mTOR and phospho-p70S6K1 levels. Consistent with these findings, the

results obtained for the Hep3B-shSIRT1-p53MUT cells were similar to those

14

for PLC5 cells, whereas the results for the Hep3B-scramble-p53WT cells were

similar to those for HepG2 cells (Supplementary Figs. 5B and 7B). To identify

the cause of the reduction in mTOR phosphorylation in HepG2-shSIRT1 and

Hep3B-shSIRT1-p53WT cells, we analyzed the expression and activation of

components of the PTEN-PI3K-AKT pathway. SIRT1 knockdown led to up-

regulation of PTEN expression and down-regulation of AKT signaling, and this

effect was not observed in PLC5-shSIT1 and Hep3B-shSIRT1-p53MUT cells

(Supplementary Figs. 5C and 7C). To further confirm these results, we

observed these pathways in SIRT1 over-expressing cell lines. Similar to the

SIRT1 knockdown results, the activated AMPK level was induced by SIRT1

overexpression in HepG2-hSIRT1 cells. Moreover, despite the increased

phopho-AMPK levels, there was induction of activated mTOR and p70S6K1,

an mTOR target protein. In addition, SIRT1 over-expressing HepG2 cell lines

had increased AKT activation. In contrast, PLC5-hSIRT1 and Hep3B-hSIRT1

cells had inverted results compared with SIRT1 knockdown cells (PLC5-

shSIRT1, Hep3B-shSIRT1-p53MUT) (Supplementary Fig. 6A, B). Given the

above results, we asked whether there is an unknown factor that modulates

AMPK and SIRT1 and hypothesized that SESTRINs are the most likely

candidates in p53 wild type HCC cell lines.

SESTRIN1 and SESTRIN2 activate AMPK, suppress the activation of

mTORC1, and mediate many p53-dependent processes, including

metabolism and the generation of reactive oxygen species (ROS). Thus,

15

these factors act as linkers between p53 and AMPK signaling [20]. Therefore,

we sought to determine whether the inhibition of sestrin1 and sestrin2

mediates crosstalk between SIRT1 and AMPK in HCC cells expressing wild-

type p53. As expected, knockdown of SESTRIN1/2 resulted in a decrease in

phospho-AMPK and phospho-SIRT1 signals. Moreover, treatment with

metformin, an AMPK activator, increased SIRT1 phosphorylation in

SESTRIN1/2-knockdown cells, suggesting that there may be crosstalk

between AMPK and SIRT1 signaling in these cells (Supplementary Fig. 5D).

Effects of SIRT1 on cell proliferation and colony formation in vitro

To address the effects of SIRT1 on cell proliferation, we measured cell

proliferation and colony formation following transfection with two shSIRT1

constructs in the HepG2, PLC5, and Hep3B cell lines and SIRT1 over-

expression constructs in the same cell lines. Significant inhibition of cell

proliferation was observed in SIRT1-knockdown HepG2 cells, whereas

enhanced proliferation was observed in SIRT1-knockdown PLC5 cells (P <

0.01; Fig. 3A, B). SIRT1 over-expressing HepG2 cells did not have

augmented proliferation compared with HepG2-scramble cells. However, for

PLC5 and Hep3B cells, the proliferation rates were significantly suppressed

(P < 0.01; Supplementary Fig. 6C).

SIRT1 knockdown reduced the number and size of HepG2 colonies

but increased the number and size of PLC5 colonies as determined by colony

16

formation assays (P < 0.01; Fig. 3C, D). In HepG2-hSIRT1 cells, SIRT1 had

no effect in colony formation assays, but results from p53 mutant cells

revealed its inhibitory effects on colony formation ability (PLC5-hSIRT1, P <

0.01: Hep3B-hSIRT1, P < 0.01; Supplementary Fig. 6D). Similarly, Hep3B-

shSIRT1-p53MUT cells had a higher proliferation rate and better colony-

forming ability than Hep3B-scramble-p53MUT cells, whereas Hep3B-

shSIRT1-p53WT cells had a lower proliferation rate and worse colony-forming

ability compared with Hep3B-scramble-p53WT cells (P < 0.01). In a control

experiment, Hep3B-scramble-p53MUT cells demonstrated a higher

proliferation rate and better colony-forming ability than Hep3B-scramble-

p53WT cells (P < 0.01). Taken together, these results suggest that SIRT1

knockdown by shRNA promotes carcinogenesis in mutant p53 HCC cells

(Supplementary Fig. 7D, E).

Effects of SIRT1 on tumor growth in vivo

To test the effects of SIRT1 on tumor formation in mice, HepG2 and

PLC5 cells were transfected with a shSIRT1 lentivirus vector and then

subcutaneously injected into right flanks of NOG mice. HepG2-shSIRT1 cells

formed tumors that were significantly smaller than those of control cells (P <

0.01), whereas PLC5-shSIRT1 cells formed tumors that were significantly

larger than those of control cells (P < 0.01; Fig. 3E). Interestingly, tumors

derived from HepG2-shSIRT1 cells exhibited similar AMPK expression to the

17

control, whereas tumors derived from PLC5-shSIRT1 cells exhibited lower

AMPK expression than the control (P < 0.01; Fig. 3F). These data confirm the

effects of SIRT1 on tumorigenesis in vivo.

Metformin suppresses the growth of HCC cells in vitro

To evaluate the effects of AMPK activation on cell proliferation in p53-

mutant HCC cells, HepG2-scramble, HepG2-shSIRT1, PLC5-scramble, and

PLC5-shSIRT1 cells were treated with different doses of metformin, an AMPK

activator, for 0, 24, and 48 h. Metformin was effective in inhibiting cell growth

when used at different concentrations and different times in HepG2-scramble

and PLC5-shSIRT1 cells. Interestingly, metformin treatment resulted in a

greater reduction in the proliferation of PLC5-shSIRT1 cells compared with

PLC5-scramble cells (Fig. 4A, B). To explore the mechanisms underlying the

effects of metformin, we analyzed the activation and expression of various

pathway components following treatment of cells with 100 mM metformin for

24 h. These results showed that metformin induced the phosphorylation of

acetyl CoA carboxylase (ACC), a specific AMPK target, in all cell lines and

induced an increase in AMPK phosphorylation in all cell lines except HepG2-

shSIRT1 cells. Notably, within 24 h, metformin also inhibited the

phosphorylation of mTOR in HepG2-scramble, PLC5-scramble, and PLC5-

shSIRT1 cells (Fig. 4C, D).

18

Metformin suppressed the growth of HCC tumors in mice

As shown above, metformin demonstrated a strong influence on the

suppression of PLC5-shSIRT1 cells in vitro. Therefore, we hypothesized that

AMPK activation via metformin treatment may suppress the in vivo growth of

PLC5-shSIRT1 tumors in mice. Two weeks after the inoculation of PLC5-

scramble and PLC5-shSIRT1 cells in mice, metformin treatment (150

mg/kg/day) was initiated. As expected, there was a significant reduction in the

growth of tumors in mice after 6 weeks of treatment with metformin (P < 0.01)

compared with the untreated control group. Furthermore, metformin therapy

led to an even greater reduction in the growth of tumors generated from

PLC5-hSIRT1 cells than those generated from PLC5-scramble cells (Fig. 4E).

Additionally, tumor tissues derived from PLC5-shSIRT1 cells exhibited a

higher induction of phospho-AMPK and phospho-mTOR compared with those

derived from PLC5-scramble cells (P < 0.01; Fig. 4F). These data indicate that

metformin specifically inhibits the growth of tumors derived from cells

harboring mutant p53 and inactive SIRT1.

Discussion

The role of SIRT1 in tumorigenesis is controversial, and previous studies

have indicated that the function of SIRT1 may be tumor-type specific and

depend on the stage of oncogenesis. In this report, we demonstrated

significant roles for SIRT1, AMPK, and p53 in HCC using in vitro, in vivo, and

19

clinical studies. Overall, our data demonstrate that the role of SIRT1 depends

on the p53 mutation status of HCC cells and tumor tissues, which supports

the need for biomarker evaluation in the prognosis and treatment evaluation

of HCC patients. SIRT1 is generally thought to disrupt cell cycle control and

promote tumor progression through p53 deacetylation and inactivation,

suggesting that p53 is a vital intermediate between SIRT1 and SIRT1’s

oncogenic activity [4, 5, 6, 7]. Recently, dynamic interactions have been

reported between AMPK and SIRT1 via the cellular NAD+ levels [21]. In

addition, SIRT1 deacetylates LKB1, inducing augmented AMPK

phosphorylation and activation [22, 23].

Based on our results and previous data, we hypothesize that SIRT1 is

required to induce the phosphorylation of AMPK and the inhibition of the

mTOR pathway [24, 25]. Indeed, SIRT1 inhibition by shRNA activated the

phosphorylation of mTOR (Ser2448) via decreased phospho-AMPK in the

mutant p53 cells, PLC5-shSIRT1.

The details of the mechanism involving p53, phospho-AMPK (Thr172) and

phospho-SIRT1 (Ser47) remains unclear. Our current findings provide some

understanding for the connection between the p53-SESTRIN1/2-AMPK

pathways [20]. Baik et al. argued that mTOR inhibits SIRT1 via the

phosphorylation of SIRT1 (Ser47) while promoting A431 cell survival, and

these results are different from those of previous studies [26, 27]. The

importance of individual SIRT1 targets may depend on the cell process and

20

cell type studied. However, it has been shown that phospho-SIRT1 can adjust

for cell survival by directly interacting with indispensable factors (i.e., mTOR)

[28]. In this context, our results indicate that phospho-SIRT1 (Ser47) remains

responsive to stress-induced activation in mutant p53 HCC. Furthermore,

phospho-SIRT1 inhibits HCC growth via a certain tumor-suppressive pathway.

AMPK activation via SIRT1 (phospho-SIRT1) may play an important role

in preventing mTOR activation in p53 mutant HCC. Despite the complex

regulation of mTOR in HCC cells, metformin data show that enhanced AMPK

phosphorylation is sufficient to decrease p70S6K translational activity via

mTOR inactivation in p53 mutant HCC cells and mice. Therefore, AMPK

activation by metformin is a potential strategy for enhancing HCC targeting

(Fig. 4G). Metformin has been investigated as an anticancer drug for

decreasing the HCC risk of diabetes patients. We observed a tumor growth

suppressive effect in mice during metformin treatment. Our results suggest

further investigation of this drug as an approach for targeting HCC. In addition,

these may aid in the development of more selective therapeutic approaches

for HCC patients in the future.

Acknowledgments

We thank Prof. Ben C.B. Ko (Department of Anatomical and Cellular

Pathology, the Chinese University of Hong Kong) and Prof. Yick-Pang Ching

(Department of Anatomy, the University of Hong Kong) for advice during the

21

preparation of this manuscript and critically reviewing the manuscript.

22

References

1. Parkin, D.M. Global cancer statistics in the year 2000. Lancet Oncol. 2001;

2; 533-543.

2. Luo J., Nikolaev A.Y., Imai S., Chen D., Su F., Shiloh A., Guarente L., Gu

W. Negative control of P53 by Sir2alpha promotes cell survival under stress.

Cell. 2001; 107; 137-148.

3. Vaziri H., Dessain S.K., Ng Eaton E., Imai S.I., Frye R.A., Pandita T.K.

Guarente L., Weinberg R.A.. hSIR2 (SIRT1) functions as an NAD-dependent

P53 deacetylase. Cell. 2001; 107;149-159.

4. Chen W.Y., Wang D.H., Yen R.C., Luo J., Gu W., Baylin S.B. Tumor

suppressor HIC1 directly regulates SIRT1 to modulate P53-dependent DNA-

damage responses. Cell. 2005; 123; 437-448.

5. Baylin S.B., Ohm J.E. Epigenetic gene silencing in cancer-a mechanism for

early oncogenic pathway addiction? Nat Rev Cancer. 2006; 6; 107-116.

6. Lim C.S. Human SIRT1: a potentialbiomarker for tumorigenesis? Cell Biol

Int. 2007; 31; 636-637.

7. Saunders L.R., Verdin E. Sirtuins: critical regulators at the crossroads

between cancer and aging. Oncogene. 2007; 26; 5489-5504.

8. Cha EJ, Noh SJ, Kwon KS, Kim CY, Park BH, Park HS, Lee H, Chung

MJ, Kang MJ, Lee DG, Moon WS, Jang KY. Expression and prognostic

significance of SIRT1 in ovarian epithelial tumors. Pathology .2009; 41; 366-

371.

23

9. Lee H, Kim KR, Noh SJ, Park HS, Kwon KS, Park BH, Jung SH, Youn

HJ, Lee BK, Chung MJ, Koh DH, Moon WS, Jang KY. Expression of DBC1

and SIRT1 is associated with poor prognosis for breast carcinoma. Hum

Pathol. 2011; 42; 204-213.

10. Jang KY, Hwang SH, Kwon KS, Kim KR, Choi HN, Lee NR, Kwak JY, Park

BH, Park HS, Chung MJ, Kang MJ, Lee DG, Kim HS, Shim H, Moon WS.

SIRT1 expression is associated with poor prognosis of diffuse large B-cell

lymphoma. Am J Surg Pathol. 2008; 32; 1523-1531.

11. Chen J, Zhang B, Wong N, Lo AW, To KF, Chan AW, Ng MH, Ho

CY, Cheng SH, Lai PB, Yu J, Ng HK, Ling MT, Huang AL, Cai XF, Ko BC.

Sirtuin 1 is upregulated in a subset of hepatocellular carcinomas where it is

essential for telomere maintenance and tumor cell growth. Cancer Res.

2011;71; 4138-4149.

12. Choi H.N., Bae J.S., Jamiyandorj U., Noh S.J., Park H.S., Jang K.Y.,

Chung M.J., Kang M.J., Lee D.G., Moon W.S. Expression and role of SIRT1

in hepatocellular carcinoma. Oncol Rep. 2011; 26; 503-510.

13. Kabra N., Li Z., Chen L., Li B., Zhang X., Wang C., Yeatman T., Coppola

D., Chen J. SirT1 is an inhibitor of proliferation and tumor formation in colon

cancer. J Biol Chem. 2009; 28; 18210-18217.

14. Chen H.C., Jeng Y.M., Yuan R.H., Hsu H.C., Chen Y.L. SIRT1 Promotes

Tumorigene- sis and Resistance to Chemotherapy in Hepatocellular

Carcinoma and its Expression Predicts Poor Prognosis. Ann Surg Oncol.

24

2012; 19; 2011-2019.

15. Revollo J.R., Grimm A.A., Imai S. The NAD biosynthesis pathway

mediated by nicotinamide phosphoribosyltransferase regulates Sir2 activity in

mammalian cells. J Biol Chem. 2004; 279; 50754-50763.

16. Nin V, Escande C, Chini CC, Giri S, Camacho-Pereira J, Matalonga J, Lou

Z, Chini EN. Role of Deleted in Breast Cancer 1 (DBC1) Protein in SIRT1

Deacetylase Activation Induced by Protein Kinase A and AMP-activated

Protein Kinase. J Biol Chem.2012; 287; 23489-23501.

17. Lin J.N., Lin V.C., Rau K.M., Shieh P.C., Kuo D.H., Shieh J.C., Chen W.J.,

Tsai S.C., Way T.D. Resveratrol modulates tumor cell proliferation and protein

translation via SIRT1-dependent AMPK activation. J Agric Food Chem. 2010;

58;1584-1592.

18. Vakana E., Altman J.K., Platanias L.C. Targeting AMPK in the treatment

of malignancies. J Cell Biochem. 2012;113; 404-409.

19. Woodard J., Platanias L.C. AMP-activated kinase (AMPK)-generated

signals in malignant melanoma cell growth and survival. Biochem Biophys

Res Commun. 2010; 398; 135-139.

20. Budanov A.V., Karin M. p53 target genes sestrin1 and sestrin2 connect

genotoxic stress and mTOR signaling. Cell. 2008; 134; 451-460.

21. Canto, C., Gerhart-Hines, Z., Feige, J.N., Lagouge, M., Noriega, L., Milne,

J.C., Elliott, P.J., Puigserver, P., and Auwerx, J. AMPK regulates energy

expenditure by modulating NAD+ metabolism and SIRT1 activity. Nature.

25

2009; 458; 1056–1060.

22. Ivanov, V.N., Partridge, M.A., Johnson, G.E., Huang, S.X., Zhou, H., and

Hei, T.K. Resveratrol sensitizes melanomas to TRAIL through modulation of

antiapoptotic gene expression. Exp. Cell Res. 2008; 314; 1163–1176

23. Lan, F., Cacicedo, J.M., Ruderman, N., and Ido, Y. SIRT1 modulation of

the acetylation status, cytosolic localization, and activity of LKB1. Possible

role in AMP-activated protein kinase activation. J. Biol. Chem. 2008; 283;

27628–27635.

24. Price NL, Gomes AP, Ling AJ, Duarte FV, Martin-Montalvo A, North BJ,

Agarwal B, Ye L, Ramadori G, Teodoro JS, Hubbard BP, Varela AT, Davis JG,

Varamini B, Hafner A, Moaddel R, Rolo AP, Coppari R, Palmeira CM, de

Cabo R, Baur JA, Sinclair DA. SIRT1 is required for AMPK activation and the

beneficial effects of resveratrol on mitochondrial function. Cell Metab.

2012;15; 675-690

25. Gwinn DM, Shackelford DB, Egan DF, Mihaylova MM, Mery A, Vasquez

DS, Turk BE, Shaw RJ. AMPK phosphorylation of raptor mediates a metabolic

checkpoint. Mol Cell. 2008;30; 214-226

26. Juan ME, Alfaras I, Planas JM. Colorectal cancer chemoprevention by

trans resveratrol. Pharmacol Res.2012; 65; 584-591

27. Calay D, Vind-Kezunovic D, Frankart A, Lambert S, Poumay Y, Gniadecki

R. Inhibition of Akt Signaling by Exclusion from Lipid Rafts in Normal and

Transformed Epidermal Keratinocytes. J invest Dermatol. 2010;130; 1136-

26

1145

28. Back JH, Rezvani HR, Zhu Y, Guyonnet-Duperat V, Athar M, Ratner D,

Kim AL. Cancer cell survival following DNA damage-mediated premature

senescence is regulated by mammalian target of rapamycin (mTOR)-

dependent Inhibition of sirtuin 1. J Biol Chem. 2010; 286; 19100-19108

27

Table 1 Correlation between activated SIRT1 and p53 mutation in HCC

(n=248)

Activated SIRT1 P

+ -

p53 mutation + 36 (63.2%) 21 (36.8%)

- 60 (31.4%) 131 (68.6%) <0.001

28

Table 2 Univariate and multivariate analysis of factors influencing relapse-free survival (RFS) in HCC with mutant p53 (n=57)

Independent factor

Univariate multivariate

Exp(B) 95% CI for Exp(B) P Exp(B) 95% CI for Exp(B) P

Gender 0.597 0.316-1.128 0.112

Age 1.624 0.863-3.053 0.132

AFP 1.654 0.874-3.130 0.122

Tumor number 0.554 0.295-1.041 0.066 0.267 0.130-0.549 <0.001

Tumor size 1.540 0.826-2.869 0.174

Edmondson grade 1.327 0.471-3.739 0.592

Fibrous capsule formation 0.597 0.289-1.230 0.162

Microvascular invasion 1.367 0.679-2.752 0.381

Portal vein invasion 1.077 0.331-3.500 0.902

Serosa invasion 1.029 0.248-4.276 0.969

Satellite nodule 3.199 1.658-6.174 0.001 3.479 1.606-7.534 0.002

AJCC staging 1.343 0.714-2.527 0.360

29

Activated SIRT1 0.432 0.229-0.814 0.009 0.307 0.143-0.660 0.002

AFP, alpha-fetoprotein; AJCC, American Joint Committee on Cancer; CI, confidence interval

30

Table 3 The correlation between activated SIRT1 and activated AMPK in HCC

with wild type and mutant p53 (n=248)

p53 WT Activated AMPK P

+ -

Activated SIRT1 + 26 (43.3%) 34 (56.7%)

- 71 (54.2%) 60 (45.8%) 0.212

p53 MUT Activated AMPK P

+ -

Activated SIRT1 + 18 (50.0%) 18 (50.0%)

- 2 (9.5%) 19 (90.5%) 0.003

WT, wild type; MUT, mutant

31

Figure Legends

Fig. 1 Stage-specific postoperative survival as calculated by the Kaplan-Meier

method. Activated SIRT1 was not a prognostic indicator for overall survival

(OS) or relapse-free survival (RFS) in cases with wild-type. A P-value less

than 0.05 was considered significant (A, B). (C) Activated SIRT1 was not a

prognostic indicator for overall survival (OS) but was a prognostic indicator for

relapse-free survival (RFS) in cases with mutant p53 (D). (E) In HCC patients

with mutant p53, activated SIRT1 and activated AMPK demonstrated a

significantly better prognosis in terms of RFS than in patients with inactive

SIRT1 and AMPK. (F) In patients with wild-type p53, neither activated SIRT1

nor activated AMPK predicted a better RFS in pair wise comparisons. A P-

value less than 0.05 was considered significant. *the number at risk on the

last day of contact

Fig. 2 AMPK expression and activation in nontumor and HCC tissues. (A) The

total and phosphorylated AMPK protein pattern in nontumor and HCC tissues

(B) Quantification of the percentage of AMPK and activated AMPK-positive

cases according to the results in (A). (C) Relative expression of total AMPK (t-

AMPK) and phosphorylated AMPK (p-AMPK) in HCC and nontumor tissues.

Fig. 3 Cell growth and proliferation assays and the effect of SIRT1 knockdown

on tumor growth in vivo. The proliferation and colony-forming ability of HepG2

cells (A, C) and PLC5 cells (B, D) were analyzed after SIRT1 knockdown (*, p

< 0.01). SIRT1 knockdown and control cells (both HepG2 and PLC5) were

subcutaneously injected in the right flanks of NOG mice (E). Tumor volumes

were measured after 6–8 weeks. (*, P < 0.01) Immunohistochemical analysis

of AMPK activity in tumors derived from SIRT1-knockdown and scramble-

transfected HepG2 and PLC5 cells (F). (*, P < 0.01)

Fig. 4 Restriction of cancer cell growth in vitro and reduced HCC xenograft

tumor growth in vivo by metformin. Cell proliferation was measured in (A)

HepG2 and (B) PLC5 cells. Phosphorylation of the indicated targets was

32

analyzed in (C) HepG2 and HepG2-shSIRT1 and (D) PLC5 and PLC5-

shSIRT1 cells. Western blot data are representative of at least 3 independent

experiments. *, P < 0.01. (E) PLC5 and PLC5-shSIRT1 cells were

subcutaneously injected into mice. Metformin treatment was started 2 weeks

later. Tumor volumes were determined every 2 weeks. (F)

Immunohistochemical analysis and quantification of AMPK and mTOR

phosphorylation in mouse tumors. (*P < 0.01) and (G) schematic summary of

the different roles of SIRT1 in HCC as suggested by our current data.

Metformin may be more effective at treating a specific class of HCC in which

p53 is mutated and SIRT1 and AMPK are inactive.

Figure 1

F

Figure

Figure 2

A

B

C

Figure 2

0

0.5

1

1.5

2

2.5

3

0h 24h 48h 72h

Ab

so

rb

an

ce O

D 4

50(m

m)

HepG2-scramble

HepG2-shSIRT1-1

HepG2-shSIRT1-2

* *

SIRT1

Actin

Figure 3

A

0

0.5

1

1.5

2

2.5

3

0h 24h 48h 72h

Ab

so

rban

ce O

D 4

50(m

m)

PLC5-scramble

PLC5-shSIRT1-1

PLC5-shSIRT1-2

*

*

SIRT1

Actin

Figure 3

B

0

1

2

3

4

5

6

7

HepG2-scramble HepG2-shSIRT1-1 HepG2-shSIRT1-2

co

lob

y n

um

ber(

>=

1m

m)

*

*

Figure 3

C

0

1

2

3

4

5

6

7

PLC5-scramble PLC5-shSIRT1-1 PLC5-shSIRT1-2

co

lob

y n

um

ber(

>=

1m

m)

*

*

Figure 3

D

0

0.5

1

1.5

2

2.5

3

HepG2 PLC5

Tu

mo

r vo

lum

e(c

m3)

scramble

shSIRT1* *

scramble scramble

shSIRT shSIRT

Figure 3

E

scramble scramble shSIRT1 shSIRT1

HepG2 PLC5

SIRT1

phosphor-AMPK

Figure 3

F

0

20

40

60

80

100

120

SIRT1 phosphor-AMPK

Po

sti

ve s

tain

ing

(%)

HepG2-scramble

HepG2-shSIRT1

PLC5-scramble

PLC5-shSIRT1

* * *

Figure 3

F

Figure 4

Figure 4

Metformin

(100 mmol/l, 24h) - +

HepG2

t-SIRT1

p-ACC

t-ACC

β-actin

t-AMPK

P-SIRT1

P-AMPK

- +

HepG2-shSIRT1

t-mTOR

p-mTOR

Figure 4

C

- +

PLC5

Metformin

(100 mmol/l, 24h)

t-SIRT1

p-ACC

t-ACC

β-actin

t-AMPK

P-SIRT1

P-AMPK

- +

PLC5-shSIRT1

t-mTOR

p-mTOR

Figure 4

D

0

0.5

1

1.5

2

2.5

3

0W 2W 4W 6W 8W

Tu

mo

r vo

lum

e(c

m3)

PLC5-scramble:control,n=4

PLC5-scramble:metformin 150mg/kg/Day,n=6

PLC5-shSIRT1:control,n=5

PLC5-shSIRT1:metformin 150mg/kg/Day,n=8

*

*

*

Metformin treatment

Figure 4

E

Figure 4

F

Figure 4

G