Notch1 differentially regulates oncogenesis by wildtype p53 ...

Notch1 Differentially Regulates Oncogenesis by Wildtype p53 Overexpression and p53 Mutation in Grade III Hepatocellular Carcinoma Seung-Oe Lim,1 Young Min Park,2 Hyeon Seop Kim,1 Xiaoyuan Quan,1 Jeong Eun Yoo,3 Young Nyun Park,3 Gi Hong Choi,4 and Guhung Jung1 The tumor suppressor p53 is a key prognostic factor in hepatocellular carcinoma (HCC), yet only 35% of grade III tumors exhibit mutation of p53. Several other pathways have been implicated in HCC and, among these, the role of the Notch1/Snail pathway remains unclear. Therefore, we investigated the expression of p53, Notch1, and Snail proteins in HCC with regard to both clinical grade and p53 mutational status. Immunoblotting for p53 revealed that, whereas in many tumors increased p53 was a result of p53 mutation, wildtype p53 (p53WT) expression was also frequently elevated in HCCs. Coordinated evaluation of p53, Notch1, and Snail expression suggests that grade III HCC can be subdivided based on the expression of these three proteins. We found that Notch1 expression in HCC tissues and cell lines is differentially affected by p53WT and mutant p53 (p53Mut). Notch1 expression was correlated with p53 expression in cells expressing p53WT, but was not elevated in p53Mut-expressing cells. Virally mediated expression or silencing of p53WT or p53Mut confirmed that p53WT overexpression causes Notch1 up-regulation in HCC. Surprisingly, the consequence of Notch1 overexpression for the proliferative and invasive capacity of HCC cells depends on both the p53 mutational status and activation of the Snail pathway. Conclusion: In the presence of p53WT, Snail/Notch1 activation increased the invasiveness of HCC cells. In contrast, in the absence of p53WT, Notch1 decreased the invasiveness of HCC. Taken together, these findings shed new light on the complex role of the Notch1/Snail axis in HCC and provide a framework for further classifying HCC based on the expression and mutational status of p53 and the expression of Notch1 and Snail. (HEPATOLOGY 2011;53:1352-1362)

Abbreviations: HBsAg, HBV surface antigen; HBV, hepatitis B virus; HCC, hepatocellular carcinoma; IHC, immunohistochemistry; MEF, mouse embryonic fibroblast; MMP, matrix metalloproteinase. From the 1Department of Biological Sciences, Seoul National University, Seoul, Korea; 2Hepatology Center and Laboratory of Hepatocarcinogenesis, Bundang Jesaeng General Hospital, Kyungkido, Korea; 3Department of Pathology, Center for Chronic Metabolic Disease, Brain Korea 21 Project for Medical Science, Yonsei University College of Medicine, Seoul, Korea; 4Department of Surgery, Yonsei University College of Medicine, Seoul, Korea. Received September 7, 2010; accepted January 13, 2011. Funded by: the Korea Healthcare Technology R&D Project, Ministry of Health & Welfare, Republic of Korea (A080318). Seung-Oe Lim, Hyeon Seop Kim and Xiaoyuan Quan received support from the Brain Korea 21(BK21) Research Fellowship from the Ministry of Education and Human Resources Development. Address reprint requests to: Guhung Jung, Ph.D., Department of Biological Sciences, Seoul National University, Seoul 151-747, Korea. E-mail: drjung@ snu.ac.kr; fax: (82) 2-872-1993. C 2011 by the American Association for the Study of Liver Diseases. Copyright V View this article online at wileyonlinelibrary.com. DOI 10.1002/hep.24208 Potential conflict of interest: Nothing to report. Additional Supporting Information may be found in the online version of this article. 1352

H

epatocellular carcinoma (HCC) is a highly invasive, malignant tumor with a high rate of mortality despite many scientific and clinical 1 advances. HCC is classified into four differentiation grades as suggested by Edmondson and Steiner2: welldifferentiated (grade I), moderately differentiated (grade II), poorly differentiated (grade III), and undifferentiated (grade IV). It is generally accepted that HCC undergoes stepwise hepatocarcinogenesis whereby HCC arises from dysplastic nodules, and grade II/III HCC develops from grade I HCC by way of dedifferentiation.3-5 Thus, grade I HCC represents early disease and grade III HCC constitutes advanced disease. HCC dedifferentiation correlates with increasingly aggressive behaviors including proliferation, invasion, distant metastasis, and treatment resistance.6 These oncogenic properties are determined by changes in oncogenic/antioncogenic molecules and changes in the tumor microenvironment.7

HEPATOLOGY, Vol. 53, No. 4, 2011

The prototypic molecular change associated with cancer is mutation of tumor suppressor p53, which is inactivated in 50% of human cancers.7 The p53 protein is a key regulator of cell cycle arrest and apoptosis, and is controlled by a complex network of posttranslational modifications that regulate its activity, stability, and molecular interactions.7 In HCC, p53 mutation occurs late in hepatocarcinogenesis.8 However, the relationship between p53WT expression and differentiation grade has not been well described. One reason for the lack of information is the very short half-life of p53WT, which makes quantification by immunohistochemistry (IHC) difficult. It is thought that increased stability of p53Mut allows detection by IHC9; however IHC and p53 gene mutation have only a 40% correlation: strong IHC staining has been observed in the absence of p53 mutation.9,10 These perplexing reports suggest that the frequency of p53WT overexpression in HCC, and also the biological significance of this change, have been underestimated. It is hypothesized that molecular changes are linked in a stepwise fashion to hepatocarcinogenesis. Yet how the many changes that occur in oncogenic networks coordinate in contributing to dedifferentiation or clinical progression of HCCs is poorly understood. We have shown that changes in expression of heat shock proteins correlate with differentiation grade.11 We have also demonstrated oxidative stress correlate with dedifferentiation through H2O2-induced Snail suppression of E-cadherin.12,13 Snail is a zinc finger transcriptional repressor that is associated with epithelial-to-mesenchymal transitions (EMTs) and metastasis, and is aberrantly expressed in invasive carcinomas including HCC.12,13 In HCC, Notch1 induces the expression of Snail and EMT.14 The Notch1 receptor regulates cell fate and differentiation in embryonic and tumor cells.15 Notch1 signaling correlates with various cancers, although it is reported to have both oncogenic and tumor-suppressive roles.16 Notch1 signaling occurs when juxtacrine ligand-receptor interactions induce Notch1 cleavage to NICD, which translocates to the nucleus altering gene transcription. Notch1 has multiple nuclear targets including induction of Snail transcription.14 Notch1 may also inhibit p53 degradation in HCC, inducing p53-dependent apoptosis.17 Involvement of the Snail pathway in p53 function was demonstrated by the ability of the Snail family member Twist to inhibit p53 transcription by interacting with the p53 DNA binding domain.18 In addition, p53 induces the degradation of Snail through a direct interaction.19,20 These results suggest an untested oncogenic link between p53, Notch1, and Snail pathways in HCC. Therefore,

LIM ET AL.

1353

in this study we analyzed the expression patterns of Notch1, Snail, p53WT, and p53Mut in clinical HCC tissue specimens and investigated how the expression of these proteins affects the biological behavior of HCC cell lines.

Materials and Methods Cell Culture. Unless otherwise noted, reagents were from Sigma (St. Louis, MO). The human cell lines Hep3B (hepatoma), Huh7 (hepatoma), HepG2 (hepatoblastoma), and 293T (embryonic kidney) and primary mouse embryonic fibroblasts (MEFs) obtained from day 13.5 embryos were cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) with 10% fetal bovine serum. For small interfering RNA (siRNA) experiments, cells were transfected with Notch1 or control siRNA (ON-TARGETplus, Dharmacon, Lafayette, CO) using Oligofectamine (Invitrogen, Carlsbad, CA). For short hairpin RNA (shRNA) experiments, cells were infected by control or p53 shRNA lentiviruses (Santa Cruz Biotechnology, Santa Cruz, CA). After a 72-hour transfection or transduction, protein expression was analyzed by immunoblot. Tissue Specimens and Histopathology. In all, 158 HCCs and corresponding non-HCC tissues were collected as surgical specimens from Severance Hospital, Yonsei University College of Medicine (Seoul, Korea) (age, 52 6 11 years; range, 29-75; 125 male and 33 female; all HBV surface antigen [HBsAg]-positive). Informed consent was obtained from each patient and the Institutional Review Board of Yonsei University approved tissue collection. Tissue was snap-frozen in liquid nitrogen and stored at 70 C. All non-HCC liver tissues showed hepatitis B virus-associated chronic hepatitis or cirrhosis. Representative sections underwent routine histological evaluation for differentiation, tumor size, tumor capsule formation, vascular invasion, and intrahepatic metastasis. Differentiation was graded according to Edmondson and Steiner’s2 criteria: of the HCC samples, 33 were grade I, 50 were grade II, and 75 were grade III.

Results Immunoblot Analyses of Notch1 (NICD), Snail, and p53 Expression in HCC. To analyze the relationship between p53, Notch1, and Snail pathways, we performed immunoblotting using HCC tissue. The size of the detected Notch1 protein was 120 kDa, consistent with active NICD. The frequency of increased expression of Notch1 (NICD), Snail, and p53 in HCCs was

1354

LIM ET AL.

HEPATOLOGY, April 2011

Table 1. Frequency of Notch1, Snail, and p53 Expression Patterns According to Differentiation Grade* HCC Differentiation Grade

Protein

Notch1 (NICD) Snail P53

Expression

GI (n533)

GII (n550)

GIII (n575)

Total (n5158)

: $ : $ : $

15 18 12 21 12 21

20 30 23 27 25 25

37 38 48 27 54 21

72 86 83 75 91 67

P

0.542 0.003 0.0001

*Protein expression intensity was interrogated by immunoblotting and quantitated by densitometry; G, Edmondson-Steiner’s grade; n, number of cases; :, number of cases with a greater than 2-fold increase in expression in tumor tissue; $, number of cases with unchanged or decreased expression in tumor tissue; P, Spearman correlation.

45.6% (72/158), 52.5% (83/158), and 57.6% (91/ 158), respectively (Table 1). Although increased expression of NICD did not correlate with differentiation grade, increased Snail expression was significantly more frequent in grade III HCC (Snail, P ¼ 0.003) (Table 1). Expression of p53 was also increased in grade III HCC

(54/75) and was strongly correlated with differentiation grade (P ¼ 0.0001) (Table 1; Fig. 1A). Some p53 mutations increase protein stability, thus increasing p53 protein levels in HCC tissues.9 Therefore, we analyzed genomic sequences spanning exons 4 to 10 of the p53 gene from 110 patient HCC tissues

Fig. 1. Expression of Notch1 (NICD), Snail, and p53 in HCC. (A) The frequency of cases in which Notch1 (NICD), Snail, and p53 were increased (:) along with the differential grade. (B) The mutation rate of p53 in HCC tissues. G, Edmondson-Steiner’s grade; n, number of cases; WT, wildtype; Mut, mutation; P, Spearman correlation; *value was statistically significant. (C) The frequency of NSP category IA1, IIA1, and IIB cases increased with Edmondson-Steiner’s grade. The changes of prevalence according to the differentiation grades were statistically significant (P ¼ 0.01). :, cases with a greater than 2-fold increase in expression in tumor tissue; $, cases with unchanged or decreased expression in tumor tissue; P, Spearman correlation. (D) Expression of Notch1 (NICD), Snail, and p53 in HCC specimens was analyzed by immunoblot. The expression patterns of these three proteins are characterized into three categories. The frequency of cases in which Notch1, Snail, and p53 protein levels increased (:) (category IA1 of Table 3) is shown with the histological grade. The region highlighted with yellow represents category IA1.


LIM ET AL.

Table 2. Correlation Between Up-regulation of Notch1 and Snail Proteins and p53 Gene Mutation in Grade III HCCs* p53 gene status Protein Expression Patterns

p53 Notch1 (NICD) Snail

: $ : $ : $

Mutant (n521)

Wild (n539)

Total (n560)

20 1 6 15 10 11

25 14 22 17 26 13

45 15 28 32 36 24

P

0.007 0.040 0.156

*Protein expression was examined by immunoblotting and quantitated by densitometry; G, Edmondson-Steiner’s grade; n, number of cases; :, number of cases with a greater than 2-fold increase in expression in tumor tissue; $, number of cases with unchanged or decreased expression in tumor tissue; P, Spearman correlation.

with available genomic DNA. Mutations were found in 25 of 110 HCC patient tissues (22.7%), and 84% (21/25) of these mutations occurred in grade III HCCs (Fig. 1B). Most p53 mutations were in exons 5-9, which encode the DNA-binding domain (Table S1). Sequence analysis revealed several mutations, including the six well-established hot spot mutations, were predicted to result in p53 loss of function.20 In grade III HCCs, increased p53 protein expression was significantly correlated with p53 mutation: 44.4% (20/45) of the cases with increased p53 showed a p53 mutation, whereas in those cases in which p53 was not increased only 6.7% (1/15) showed mutations (P ¼ 0.007) (Table 2). Moreover, p53WT overexpression was not significantly different according to the differentiation grade (Supporting Information Table S2). Therefore, we suspect that the strong correlation of p53 increase and differentiation grade is due to the high prevalence of p53 mutations in grade III HCCs. Additionally, p53 mutation is correlated with increased p53 expression, and in most cases both elevated p53 expression and progression to grade III HCC occurs. Although NICD expression was increased in nearly half of HCC tumors, suggestive of a pro-oncogenic role, expression did not correlate with HCC grade. Surprisingly, when we examined the relationship between p53 mutation and NICD expression, a significant inverse correlation was noted (P ¼ 0.040) (Table 2). The mutation rate of p53 gene was 46.9% (15/32) in cases without increased NICD expression, but only 21.4% (6/28) in cases with increased NICD expression (P ¼ 0.040) (Table 2). Furthermore, of the 21 grade III HCCs with p53 mutations, 15 (71.4%) lacked increased NICD (Table 2). Snail expression was not correlated with p53 mutational status; however, a strongly positive correlation was seen between Snail and

1355

NICD expression in grade III HCCs (P ¼ 0.0001) (Supporting Information Tables S3-S5). These relationships between p53WT/Mut and Notch1 expression patterns were observed only in grade III HCCs and not in grade I/II HCCs. This is because p53 mutation is most frequently observed in grade III HCCs and rarely in grade I/II HCCs. Therefore, we performed the statistical analysis only on grade III HCCs. These results indicate that, whereas NICD expression is linked to Snail expression and oncogenesis in the context of p53WT, in the presence of p53Mut elevated NICD may inversely correlate with oncogenesis. Classification of HCC According to NICD, Snail, and p53 Expression. Immunoblot analysis of NICD, Snail, and p53 tissue expression patterns established correlations between each protein and differentiation grade (Table 1). However, these data do not address whether functional interactions between these three proteins might contribute to HCC. The clear interrelationship of Snail and NICD as well as p53 mutation and NICD suggests a functional relationship between these three proteins and HCC. In order to investigate how patterns of p53, NICD, and Snail expression relate to HCC, we categorized tissue expression of p53 and Snail in relation to NICD (Table 3). When HCC tissues were categorized by NICD, Snail, and p53 expression, for convenience the NSP (NICD, Snail, and p53) classification, several patterns were observed (Table 3). Surprisingly, although there was no significant correlation between NICD expression and differentiation grade (Table 1), the most common expression pattern for differentiation grade III HCC was category NSP-IA1, defined by elevated expression of all three proteins (Table 3). Three categories, NSP-IA1, IIA1, and IIB, were increased in grade III HCC (P ¼ 0.01; Fig. 1C) and the percentage of tissues classified as NSPIA1 increased with grade level (Fig. 1D). Representative expression patterns of NICD, Snail, and p53 in NSPIA1 tissues are shown by immunoblotting (Fig. 2A) and immunohistochemistry (Fig. 2B). Differences in p53 Mutation Rates According to the Proposed NSP Classification. Sequencing of p53 from HCC tissues (Fig. 1B; Supporting Information Table S1) revealed a mutation frequency similar to previous reports.9,10 Surprisingly, however, tissue immunoblotting of HCC revealed that p53 expression was elevated in 57.8% (25/45) of cases in which the p53 gene was wildtype (Table 2). The frequency of enhanced p53WT expression in HCC suggests compensatory mechanisms of oncogenesis and led us to investigate whether p53 mutational status correlated with certain grade III HCCs. As p53Mut was inversely

1356

LIM ET AL.


Fig. 2. The example of expression of Notch1 (NICD), Snail, and p53 in HCC. (A) Immunoblot of grade III NSP-IA1 (#1 to #5) and NSP-IIA1 (#6, #7) HCC tissues. (B) Immunohistochemistry of tissue samples representative of category IA1, IIA1, and III as defined in Table 3 (original magnification, 400).

correlated with NICD expression, and NSP-IA1 (NICD:/Snail:/p53:) HCC comprised the dominant NSP classification of differentiation grade III HCCs, we examined the mutational status of p53 within this category. Strikingly, 77.8% (14/18) of differentiation grade III NSP-IA1 cases were shown to be p53WT overexpression (Supporting Information Tables S6, S7). Conversely, in NSP-IIA1 and NSP-IIB cases, which like NSP-IA1 correlated with differentiation grade, but in which NICD was not increased, more than half was p53Mut (55.6% and 62.5%, respectively; Supporting Information Table S6 and Fig. S1). These results suggest that NSP-IA1 might represent a distinct mechanism of oncogenesis and a distinct class of grade III HCC. Indeed, satellite nodules were more frequent in NSP-II cases (NICD $) compared to

NSP-I cases (NICD:) (46.2% versus 23.3%, respectively) (P ¼ 0.020) (Supporting Information Table S8), further suggesting that NSP-IA1 defines a subset of HCC with characteristic behavior. p53WT Increases NICD by Way of Transcriptional Regulation of Notch1. Given the prevalence of NSP-IA1 among grade III HCC, we wanted to explore the mechanism(s) responsible for the correlation of p53WT and NICD expression. To determine whether NICD affects p53 expression, we infected HepG2 cells, which express p53WT, using an MSCV retroviral system to express NICD (MSCV-NICD). In NICD-overexpressing cells, there was no change in p53 expression (Supporting Information Fig. S2A). We also treated HepG2 cells with Notch1 siRNA to down-regulate expression of NICD and, likewise, found that


LIM ET AL.

1357

Fig. 3. p53WT, but not p53Mut, increases NICD by up-regulating transcription of Notch1. (A) Hep3B cells were infected by control MSCV, MSCV-p53WT, and MSCV-p53Mut viruses. After puromycin selection, endogenous NICD and p53 protein levels were analyzed by immunoblot with the indicated antibodies. (B) HepG2 cells were infected by control and p53 shRNA lentiviruses. After puromycin selection, endogenous Notch1 (NICD) and p53 protein levels were analyzed by immunoblot with the indicated antibodies. (C) p53WT up-regulates Notch1 promoter activity. Hep3B cells were transfected with p53WT and p53Mut expression vectors and a vector containing a Notch1 wildtype (WT) or p53 binding site mutated (Mut) promoter fused to luciferase for 48 hours and analyzed for luciferase activity. Pro, promoter; Control, control vector (pCMV/ Myc). (D) HepG2 cells were treated with UV radiation. After 12 hours, the expression of Notch1 (NICD), p53, and p21 were analyzed by immunoblot. (E) MEF cells were obtained from p53þ/þ and p53/ mice, treated with UV, and analyzed for expression of Notch1 (NICD), p53, and p21 by immunoblot. (F) HepG2 cells were treated for 6 hours with 10 lM nutlin-3, and expression of Notch1 (NICD), p53, and p21 was assessed by immunoblot analysis.

endogenous p53 expression levels were unchanged (Supporting Information Fig. S2B). These data suggest that NICD does not affect expression of p53 in HCC cells. Although the regulation of Notch1 by p53 had not previously been investigated in HCC, one report had shown that p53 can up-regulate Notch1 transcription in epithelial cells.21 Therefore, to test whether p53 regulated NICD expression in HCC, we analyzed NICD levels following p53 expression in p53 null Hep3B cells. MSCV retroviral constructs encoding p53WT (MSCVp53WT) or p53Mut (MSCV-p53Mut; R249S hot spot mutation) were used to stably transduce Hep3B cells. Consistent with our analyses of primary tissues, p53WT but not p53Mut increased NICD protein expression (Fig. 3A). As a control, levels of p21, which are regulated by p53,22 were analyzed (Fig. 3A). To determine if the change in NICD protein levels reflected transcriptional regulation of Notch1 we performed real-time reverse-transcription polymerase chain reaction (RTPCR) analyses, demonstrating that p53WT, but not

p53Mut, induced Notch1 mRNA expression (Supporting Information Fig. S3). Therefore, in the context of a p53 null cell line, p53 expression is sufficient to drive Notch1 transcription and NICD expression. To determine whether endogenous p53 constitutively regulates the level of NICD, we used p53 shRNA to silence the expression of p53 in p53WT HepG2 cells. In p53 shRNA-transfected HepG2 cells, endogenous NICD protein and Notch1 messenger RNA (mRNA) expression levels were decreased compare to control shRNAtransfected cells (Fig. 3B; Supporting Information Fig. S4), demonstrating that p53WT is necessary for constitutive up-regulation of NICD in HepG2 cells. Additionally, in an analysis of wildtype Notch1 promoter activity (using Notch1WT-Luc), p53WT, but not p53Mut, increased Notch1 promoter activity (Fig. 3C). Consistent with a direct effect on Notch1 transcription, p53WT did not increase transcription of a Notch1 promoter containing a mutated p53 binding site (Notch1Mut-Luc) (Fig. 3C). These data imply that p53WT

1358

LIM ET AL.


Fig. 4. Analysis of the cellular function of p53WT/Mut and NICD. Hep3B cells were infected with MSCV-p53WT, p53Mut, and/or MSCV-NICD and selected in puromycin. (A) The selected cells were analyzed for proliferation using a CCK-8 proliferation assay. (B) The selected cells were assayed for cell invasion using an Oris cell invasion and detection assay. (C) Hep3B cells were infected by MSCV, MSCV-p53WT, or p53Mut and selected in puromycin. The selected cells were analyzed for expression of invasion- or metastasis-related genes using a PCR array. The list shows the significantly up-regulated metastasis suppressor genes. (D) In the same sample as (C), we analyzed the mRNA expression levels of MMPs using real-time RT-PCR.

up-regulates NICD by increasing the transcription of Notch1. We next asked whether induction of endogenous p53 would also increase NICD. In HepG2 cells, ultraviolet (UV) treatment increased both p53 and NICD expression (Fig. 3D). However, as UV irradiation activates multiple signaling pathways, we wished to determine whether the effect on NICD specifically required p53. Therefore, MEFs were obtained from p53þ/þ and p53/ mice. In p53þ/þ MEFs, UV treatment increased NICD expression, but this effect was not observed in p53/ MEFs (Fig. 3E). Finally, a specific activator of p53, nutlin-3, which inhibits the p53/ MDM2 interaction leading to p53 accumulation and activation, was used. When HepG2 cells were treated with nutlin-3, NICD protein levels were increased (Fig. 3F). Taken with our tissue immunoblot analyses, these data demonstrate that p53WT can drive Notch1 and NICD expression in HCC cells, and suggest that elevated NICD in NSP-IA1 HCC is a result of elevated p53WT expression.

p53WT Reduces HCC Cell Proliferation and Invasion. Given the high representation of NSP-IA1 in grade III HCC, we investigated whether NICD contributed to the proliferative and invasive capacity of p53WT HCC. The lack of NICD overexpression in p53Mut cells also suggests that NICD may inhibit the proliferative and invasive capacity of p53Mut HCC. Therefore, to understand the function of p53 and NICD in cancer cell behavior, we infected MEF, HepG2, or Hep3B cells with MSCV driving expression of p53WT, p53Mut, and/or Myc-NICD and performed proliferation, apoptosis, migration, and invasion assays. The efficiency of cotransfection was equivalent in Notch1 and/or Snail expressing MSCV-infected cells (Supporting Information Fig. S5). As expected, MSCVp53WT-infected cells showed decreased and MSCVp53Mut-infected cells showed increased proliferation compared to control MSCV-infected cells (Fig. 4A). Surprisingly, NICD coexpression did not alter the proliferation of p53WT cells. However, consistent with the inverse correlation between NICD and p53Mut in


HCC tissues, decreased proliferation was observed in p53Mut/NICD coinfected cells (Fig. 4A). We also assessed the effect of NICD and p53 expression on the invasive capacity of Hep3B cells. In MSCV-p53WTinfected cells, the number of invading cells was less than control MSCV- or p53Mut-infected cells (Fig. 4B). NICD expression did not have a significant effect on either Hep3B invasion or the suppression of invasion by p53WT (Fig. 4B). To understand why p53WT induced these changes in Hep3B cells, a PCR array for metastasis-related genes was performed. Consistent with the functional characterization of these cells, there was a significant increase in the expression of the metastasis suppressor genes, KISS1R and CD82, in p53WT overexpressing cells (Fig. 4C). The examination of matrix metalloproteinase (MMP) expression was, likewise, consistent with the consequences of p53 expression in that there was a significant decrease in MMP1, MMP3, and MMP9 in p53WT-infected cells (Fig. 4D). Thus, as expected, isolated expression of p53WT inhibits the proliferative and invasive capacity of HCC cells. Unexpectedly, NICD coexpression failed to alter the suppressive effects of p53WT, suggesting additional changes must contribute to the frequency of grade III HCC with elevated p53WT and NICD (NSP-IA1). NICD and Snail Differentially Regulate Proliferation and Invasion Depending on p53 Status. Notch1 and Snail signaling are known to regulate cell proliferation and invasion13,23 and given that Snail and NICD were coordinately up-regulated in grade III HCC, we sought to determine whether overexpression of both might be needed to augment p53WT HCC invasiveness. Thus, we performed proliferation and invasion assays in p53WT or null MEFs transduced with MSCV-Myc-NICD and/or Flag-Snail constructs. In p53þ/þ MEFs, Snail and NICD coinfected cells showed decreased proliferation compared to control MSCVinfected cells; however, in p53/ MEFs, Snail and NICD coinfection increased proliferation (Fig. 5A). Interestingly, cells coinfected with Snail and NICD showed a p53-dependent invasiveness. In p53þ/þ MEFs, the number of invading Snail/NICD coinfected cells was increased. However, in p53/ MEFs, invasiveness of Snail/NICD coexpressing cells was decreased compared with control MSCV- or Snail-infected cells (Fig. 5B). These data suggest the coexpression of Snail and NICD can promote some aspects of cancer cell behavior in p53WT cells. Furthermore, the correlation between the ability of NICD/Snail to promote invasion in p53WT but not p53 null cells and our observations that increased NICD expression was associated with tumor grade primarily in the context of Snail and

LIM ET AL.

1359

p53WT overexpression suggested that these results may have direct implications for the oncogenic potential of p53WT HCC. To confirm that these differences in invasiveness depend on the p53 status of HCC cells, we again made use of HCC cell lines with different p53 mutational states. In HepG2 cells (p53WT), coexpression of Snail and NICD resulted in increased invasiveness (Fig. 5C). Thus, in cell lines, as in NSP-IA1, coordinate expression of Snail and NICD can augment the invasive behavior of p53WT cells. However, in Hep3B cells (p53 null), Snail and NICD coinfected cells exhibited decreased invasion (Fig. 5D). Furthermore, when p53WT was exogenously expressed in Hep3B cells, coexpression of Snail and NICD increased invasiveness as it had in HepG2 cells. Consistent with these in vitro findings, in cases of encapsulated grade III HCC, capsular invasion occurred in more than half of HCCs with elevated p53WT (NSP-IA1), but none in which p53 was not elevated (NSP-IA2) (Supporting Information Table S9). Together, these observations support the idea that NICD and Snail differentially regulate HCC cell proliferation and invasion depending on p53 status.

Discussion The studies we have described here provide several novel insights into the pathobiology of HCC. First, increased expression of p53WT occurs frequently in HCCs, more frequently than conventionally known prevalence. Second, increased expression of p53WT in HCC is correlated with increased expression of NICD and Snail. Third, p53 mutation is not associated with NICD up-regulation. Fourth, classification of grade III HCCs by Notch1, Snail, and p53 expression reveals that the most common expression pattern is defined by upregulation p53WT, NICD, and Snail. Finally, Notch1/ NICD and Snail coordinately promote the invasiveness of HCC cells in the context of p53WT overexpression. These results suggest that Notch1-Snail signaling differentially regulates cancer cell behavior in grade III HCCs, depending on the mutational status of p53. The relatively high frequency of p53WT overexpression in HCC was surprising, and suggests, in contrast to the idea that 5% immunoreactivity is a reliable IHC threshold value to detect p53 mutation in HCC, p53 immunoreactivity does not strictly correlate with mutation.9,10,24 Instead, immunoblotting suggests IHC underestimates the frequency p53WT elevation in HCCs. The mechanism that regulates p53WT accumulation in grade III HCCs is currently not known, although one recent report suggests that granulin-epithelial precursor (GEP)

1360

LIM ET AL.


Fig. 5. Snail and NICD regulation of invasion depends on p53 status. (A) p53þ/þ, and p53/ MEFs were infected by MSCV-NICD and/or MSCV-Snail, selected in puromycin, and analyzed for proliferation rate using a CCK-8 proliferation assay. (B) p53þ/þ and p53/ MEFs were infected with MSCV-NICD and/or MSCV-Snail, selected in puromycin, and assayed for cell invasion using an Oris assay. (C) HepG2 cells were infected with MSCV-NICD and/or MSCV-Snail, selected in puromycin, and assayed for cell invasion as in (B). (D) Hep3B cells were infected with MSCV, MSCV-p53WT, MSCV-NICD, and/or MSCV-Snail, selected in puromycin and assayed for cell invasion as in (B).

can increase p53WT expression in HCC.25 Even less clear are the molecular changes that confer oncogenic potential to cells overexpressing p53WT. However, the data provided here suggest that up-regulation of the NICD/Snail axis may be one mechanism that allows p53WT HCC to progress to differentiation grade III. Our classification of grade III HCCs by relative expression levels of Notch1, Snail, and p53WT/Mut (Table 3, Supporting Information Table S6) revealed a nonrandom expression pattern for these proteins. For example, NICD expression was increased in the majority of HCC tissues in which p53WT was elevated, but only in a minority of p53Mut HCC (Supporting Information Tables S6, S7). Indeed, we found that p53WT, but not p53Mut, induces Notch1 transcription. We also observed that overexpressed p53WT inhibited cell proliferation and invasion in HCC

cells.13,23 These results are consistent with the known role of p53 as a tumor suppressor and p53 mutation as an indicator of poor prognosis.26,27 Yet we also demonstrate that overexpression of p53WT is not an insurmountable obstacle in HCC progression. Up-regulation of Notch1 by p53, in combination with Snail up-regulation, likely by Notch1,14 promotes HCC invasion and correlates with increasing Edmondson and Steiner grade. Depending on p53 status, the pattern of HCC cell proliferation and invasion, as determined by Notch1 and Snail, was different (Fig. 5). In order to investigate the mechanism underlying these results, we performed an analysis of p21 promoter activity using three different HCC cell lines that express Notch1 and Snail at endogenous levels and have well documented p53 status, i.e., HepG2, p53WT; Hep3B, p53 null; Huh7, p53Mut. When p21 promoter activity


LIM ET AL.

1361

Table 3. NSP Classification of HCC and Prevalence of Each Pattern According to Differentiation Grade* I

Notch1 Snail p53 G I (n¼33) G II (n¼50) G III (n¼75) Total (n¼158)

II

A1

A2

B1

B2

A1

A2

B

III

: : : 3 9.1% 7 14.0% 22 29.3% 32 20.3%

: : $ 4 12.1% 8 16.0% 10 13.3% 22 13.9%

: $ : 3 9.1% 3 6.0% 4 5.3% 10 6.3%

: $ $ 5 15.2% 2 4.0% 1 1.3% 8 5.1%

$ : : 2 6.1% 6 12.0% 11 14.7% 19 12.0%

$ : $ 3 9.1% 2 4.0% 5 6.7% 10 6.3%

$ $ : 4 12.1% 9 18.0% 17 22.7% 30 19.0%

$ $ $ 9 27.3% 13 26.0% 5 6.7% 27 17.1%

*Protein expression was examined by immunoblotting and quantitated by densitometry; G, Edmondson-Steiner’s grade; n, number of cases; :, number of cases with a greater than 2-fold increase in expression in tumor tissue; $, number of cases with unchanged or decreased expression in tumor tissue; %, the number of cases with the given expression pattern/total cases for each grade x 100; Table data are summarized graphically in Fig. 1C.

in the presence of overexpressed Notch1 and Snail was examined it was found that it depends on the p53 status of the HCC cell line (Supporting Information Fig. S6). Thus, in cells expressing p53WT (HepG2), NICD and Snail have little effect on p21 promoter activity, whereas in cells lacking or with mutant p53, NICD increases activity. To confirm the contribution of p53, we examined the effects of NICD/Snail on transcription from a p21 promoter with a mutated p53 site and found that NICD now augmented p21 promoter activity in HepG2 cells. These data further suggest that NICD-mediated induction of p21 expression could preclude NICD overexpression in p53Mut HCC. These results suggest that changes in p53 expression and p53 mutations may regulate cancer cell behavior by modulating Notch1-Snail signaling. There are two opposing theories concerning Notch1 in hepatocarcinogenesis: Notch1 is pro- or antioncogenic.23 Up-regulated Notch1 signaling increases oncogenic potential by preventing differentiation and inhibiting apoptosis; accordingly, inhibiting Notch1 signaling would be an attractive therapeutic strategy for HCC.28 We have previously shown Snail up-regulation can down-regulate E-cadherin by way of epigenetic changes in HCCs.12 Therefore, Notch1 up-regulation of Snail both reduces E-cadherin and induces EMT and the conversion of polarized epithelial cells into motile, invasive cells.15,29,30 Consistent with this oncogenic role, we found a significant correlation between Notch1 and Snail expression in grade III HCCs and demonstrated that NICD/Snail coexpression increases invasiveness of p53WT HCC. In contrast, it is suggested that up-regulation of Notch1 signaling may be a treatment strategy for HCC.31 One recently described mechanism of Notch1 antioncogenic activity is through sensitizing HCC cells to apoptosis.32 Indeed, we found that in the context of p53Mut, elevated Notch1 activation may,

indeed, have anti-oncogenic effects. Thus, not only are both theories of Notch1 in HCC possible, but both are likely functioning in HCC. In summary, we have shown that in HCC, p53 mutation is not required for tumor progression. Instead, p53WT modulates the activity of the Notch1-Snail axis, allowing HCC progression to differentiation grade III. Although no significant differences were observed in the incidence of several pathological features of grade III HCCs and the NSP classification strategy, satellite nodules tended to be more frequent in NSP-II cases (NICD $) compared to NSP-I cases (NICD:) (P ¼ 0.020) (Supporting Information Table S8). Furthermore, although fibrous capsule formation in grade III HCCs occurs frequently independently of NSP classification, capsular invasion by tumor cells was highly associated with expression of p53 (Supporting Information Table S9). This is distinctly evident in the comparison between NSP IA1 and IA2, in which p53 expression is the only distinguishing feature. This correlation between our proposed system of categorization and a prognostic factor33 suggests both that the NSP scheme may be clinically useful and that the relationship of p53, Snail, and NICD plays an important role in HCC pathobiology. Taken together, the combinatorial expression pattern of Notch1, Snail, and p53WT proteins (NSP category) and the mutational status of p53 are likely to be useful in predicting the cancerous behavior of HCC and should be taken into consideration as efforts are made to target the Notch1/Snail axis therapeutically.

References 1. Spangenberg HC, Thimme R, Blum HE. Targeted therapy for hepatocellular carcinoma. Nat Rev Gastroenterol Hepatol 2009;6:423-432. 2. Edmondson HA, Steiner PE. Primary carcinoma of the liver: a study of 100 cases among 48,900 necropsies. Cancer 1954;7:462-503.

1362

LIM ET AL.

3. Maeda T, Adachi E, Kajiyama K, Takenaka K, Honda H, Sugimachi K, et al. CD34 expression in endothelial cells of small hepatocellular carcinoma: its correlation with tumour progression and angiographic findings. J Gastroenterol Hepatol 1995;10:650-654. 4. Matsui O. Imaging of multistep human hepatocarcinogenesis by CT during intra-arterial contrast injection. Intervirology 2004;47:271-276. 5. Midorikawa Y, Yamamoto S, Tsuji S, Kamimura N, Ishikawa S, Igarashi H, et al. Allelic imbalances and homozygous deletion on 8p23.2 for stepwise progression of hepatocarcinogenesis. HEPATOLOGY 2009;49:513-522. 6. Zhou L, Rui JA, Ye DX, Wang SB, Chen SG, Qu Q. EdmondsonSteiner grading increases the predictive efficiency of TNM staging for long-term survival of patients with hepatocellular carcinoma after curative resection. World J Surg 2008;32:1748-1756. 7. Vousden KH, Lu X. Live or let die: the cell’s response to p53. Nat Rev Cancer 2002;2:594-604. 8. Ng IO, Lai EC, Chan AS, So MK. Overexpression of p53 in hepatocellular carcinomas: a clinicopathological and prognostic correlation. J Gastroenterol Hepatol 1995;10:250-255. 9. Lee SN, Park CK, Sung CO, Choi JS, Oh YL, Cho JW, et al. Correlation of mutation and immunohistochemistry of p53 in hepatocellular carcinomas in Korean people. J Korean Med Sci 2002;17:801-805. 10. Anzola M, Saiz A, Cuevas N, Lopez-Martinez M, Martinez de Pancorbo MA, Burgos JJ. High levels of p53 protein expression do not correlate with p53 mutations in hepatocellular carcinoma. J Viral Hepat 2004;11:502-510. 11. Lim SO, Park SG, Yoo JH, Park YM, Kim HJ, Jang KT, et al. Expression of heat shock proteins (HSP27, HSP60, HSP70, HSP90, GRP78, GRP94) in hepatitis B virus-related hepatocellular carcinomas and dysplastic nodules. World J Gastroenterol 2005;11:2072-2079. 12. Lim SO, Gu JM, Kim MS, Kim HS, Park YN, Park CK, et al. Epigenetic changes induced by reactive oxygen species in hepatocellular carcinoma: methylation of the E-cadherin promoter. Gastroenterology 2008;135:2128-2140, 2140 e2121-2128. 13. Peinado H, Olmeda D, Cano A. Snail, Zeb and bHLH factors in tumour progression: an alliance against the epithelial phenotype? Nat Rev Cancer 2007;7:415-428. 14. Sahlgren C, Gustafsson MV, Jin S, Poellinger L, Lendahl U. Notch signaling mediates hypoxia-induced tumor cell migration and invasion. Proc Natl Acad Sci U S A 2008;105:6392-6397. 15. Grego-Bessa J, Diez J, Timmerman L, de la Pompa JL. Notch and epithelial-mesenchyme transition in development and tumor progression: another turn of the screw. Cell Cycle 2004;3:718-721. 16. Bresnick EH, Chu J, Christensen HM, Lin B, Norton J. Linking Notch signaling, chromatin remodeling, and T-cell leukemogenesis. J Cell Biochem Suppl 2000;(Suppl 35):46-53. 17. Wang C, Qi R, Li N, Wang Z, An H, Zhang Q, et al. Notch1 signaling sensitizes TRAIL-induced apoptosis in human hepatocellular carcinoma cells by inhibiting Akt/Hdm2-mediated p53 degradation and up-regulating p53-dependent DR5 expression. J Biol Chem 2009;284: 16183-16190.


18. Shiota M, Izumi H, Onitsuka T, Miyamoto N, Kashiwagi E, Kidani A, et al. Twist and p53 reciprocally regulate target genes via direct interaction. Oncogene 2008;27:5543-5553. 19. Wang SP, Wang WL, Chang YL, Wu CT, Chao YC, Kao SH, et al. p53 controls cancer cell invasion by inducing the MDM2-mediated degradation of Slug. Nat Cell Biol 2009;11:694-704. 20. Lim SO, Kim H, Jung G. p53 inhibits tumor cell invasion via the degradation of snail protein in hepatocellular carcinoma. FEBS Lett 2010; 584:2231-2236. 21. Yugawa T, Handa K, Narisawa-Saito M, Ohno S, Fujita M, Kiyono T. Regulation of Notch1 gene expression by p53 in epithelial cells. Mol Cell Biol 2007;27:3732-3742. 22. el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Trent JM, et al. WAF1, a potential mediator of p53 tumor suppression. Cell 1993;75:817-825. 23. Dotto GP. Crosstalk of Notch with p53 and p63 in cancer growth control. Nat Rev Cancer 2009;9:587-595. 24. Cesarman E, Inghirami G, Chadburn A, Knowles DM. High levels of p53 protein expression do not correlate with p53 gene mutations in anaplastic large cell lymphoma. Am J Pathol 1993;143:845-856. 25. Cheung ST, Wong SY, Lee YT, Fan ST. GEP associates with wild-type p53 in hepatocellular carcinoma. Oncol Rep 2006;15:1507-1511. 26. Hayashi H, Sugio K, Matsumata T, Adachi E, Takenaka K, Sugimachi K. The clinical significance of p53 gene mutation in hepatocellular carcinomas from Japan. HEPATOLOGY 1995;22:1702-1707. 27. Sugo H, Takamori S, Kojima K, Beppu T, Futagawa S. The significance of p53 mutations as an indicator of the biological behavior of recurrent hepatocellular carcinomas. Surg Today 1999;29:849-855. 28. Ning L, Wentworth L, Chen H, Weber SM. Down-regulation of Notch1 signaling inhibits tumor growth in human hepatocellular carcinoma. Am J Transl Res 2009;1:358-366. 29. Zavadil J, Cermak L, Soto-Nieves N, Bottinger EP. Integration of TGF-beta/Smad and Jagged1/Notch signalling in epithelial-to-mesenchymal transition. EMBO J 2004;23:1155-1165. 30. Yang MH, Chen CL, Chau GY, Chiou SH, Su CW, Chou TY, et al. Comprehensive analysis of the independent effect of twist and snail in promoting metastasis of hepatocellular carcinoma. HEPATOLOGY 2009; 50:1464-1474. 31. Qi R, An H, Yu Y, Zhang M, Liu S, Xu H, et al. Notch1 signaling inhibits growth of human hepatocellular carcinoma through induction of cell cycle arrest and apoptosis. Cancer Res 2003;63:8323-8329. 32. Wang C, Qi R, Li N, Wang Z, An H, Zhang Q, et al. Notch1 signaling sensitizes tumor necrosis factor-related apoptosis-inducing ligand-induced apoptosis in human hepatocellular carcinoma cells by inhibiting Akt/Hdm2-mediated p53 degradation and up-regulating p53-dependent DR5 expression. J Biol Chem 2009;284:16183-16190. 33. Choi SB, Lee JG, Kim KS, Yoon DS, Choi JS, Lee WJ, et al. The prognosis and survival analysis according to seven staging systems of hepatocellular carcinoma following curative resection. Hepatogastroenterology 2008;55:2140-2145.