Ann Surg Oncol (2013) 20:S684–S692 DOI 10.1245/s10434-013-3171-7

ORIGINAL ARTICLE – TRANSLATIONAL RESEARCH AND BIOMARKERS

Chromobox Homolog 4 Is Correlated with Prognosis and Tumor Cell Growth in Hepatocellular Carcinoma Boqing Wang, MD1,2, Jianjun Tang, MD, PhD2, Dan Liao, MD2, Gang Wang, MD2, Meifang Zhang, MD2, Yi Sang, PhD2, Jingying Cao, MD2, Yuanzhong Wu, PhD2, Ruhua Zhang, MD2, Shengping Li, MD2, Wei Ding, MD1, Guoqing Zhang, MD1, and Tiebang Kang, PhD2 1

Department of Hepatopancreatobiliary Surgery, Affiliated Tumor Hospital, Xinjiang Medical University, Urumqi, Xinjiang, China; 2State Key Laboratory of Oncology in South China, Sun Yat-Sen University Cancer Center, Guangzhou, Guangdong, China

ABSTRACT Background. Chromobox homolog 4 (CBX4) is a member of the chromobox family of Polycomb group proteins involved in the chromatin remodeling and transcriptional regulation. However, its clinical relevance in hepatocellular carcinoma (HCC) has not yet been explored. Methods. Immunohistochemistry was used to analyze cytoplasmic expression of CBX4 in 246 HCC specimens. The expression of CBX4 in HCC cell lines and LO2 was detected by Western blot test. Cell cycle and MTT assays were used to determine the changes of cell growth capacity. The expression of downstream genes related to proliferation was detected by Western blot test. Results. The expression of CBX4 was up-regulated in multiple HCC cell lines and clinical samples. Although the CBX4 protein was detectable in both nucleus and cytoplasm in HCC tumor tissues, the high expression of CBX4 in cytoplasm was correlated with the a-fetoprotein level in serum (P = 0.036), tumor size (P = 0.029), pathologic differentiation (P = 0.033), and tumor, node, metastasis classification

Boqing Wang and Jianjun Tang contributed equally to this study.

Electronic supplementary material The online version of this article (doi:10.1245/s10434-013-3171-7) contains supplementary material, which is available to authorized users. Ó Society of Surgical Oncology 2013 First Received: 17 April 2013; Published Online: 14 August 2013 G. Zhang, MD e-mail: [email protected] T. Kang, PhD e-mail: [email protected]

system stages (P = 0.032). Moreover, HCC patients who had a high level of CBX4 in cytoplasm had a shorter overall survival (P = 0.003) and recurrence-free survival (P = 0.012). Indeed, using HCC cell line, knockdown of CBX4 led to down-regulating proliferating cell nuclear antigen and cyclin E2 as well as up-regulating p16, followed by decreased cell proliferation and impaired cell cycle progression. Conclusions. The cytoplasmic CBX4 protein may be a useful prognostic biomarker and a potential therapeutic target for HCC. Hepatocellular carcinoma (HCC) is the third leading cause of cancer death worldwide, with increasing incidence and high mortality.1 HCC is difficult to diagnose in its early stage and has a poor survival because of its high frequency of recurrence, metastasis after hepatectomy, and resistance to common chemotherapy and radiotherapy. Therefore, numerous investigations have focused on screening for novel diagnostic and prognostic biomarkers as well as therapeutic targets in HCC.2–5 Polycomb group (PcG) proteins that often form Polycomb repressive complex (PRC) are the transcriptional regulators that involved in early embryogenesis and carcinogenesis.6 Recently, besides the typical PRC1, PcG proteins have been demonstrated to actually form different PRC1 complexes localized at different genomic regions to function differently.7,8 The chromobox family (CBXs), which includes five members (CBX2, CBX4, CBX6, CBX7, and CBX8), is an important member of PcG proteins and has different expression patterns in different tissues with distinct biological functions.9 For instance, CBX7, which is down-regulated in many human cancers, acts as a tumor suppressor.10 CBX8 plays an essential role in leukemogenesis.11 CBX4, also named hPC2, is a SUMO E3 ligase, which is different from

Prognostic Significance of CBX4 in HCC

other members of the CBX family, and prevents human epidermal stem cells from senescence and maintains the stemness of these cells. In addition, CBX4 is also involved in the control of growth.12 Taken together, these results indicated that CBX4 may play a role in human cancers, although this has not yet been documented. We sought to identify novel prognostic biomarkers or therapeutic targets for HCC, which caused us to test whether CBX4 play pivotal roles in HCC.13,14 CBX4 could be a helpful prognostic marker and an alternative therapeutic target. MATERIALS AND METHODS Cell Lines and siRNA Transfection Normal human liver cell line (LO2) and human liver cancer cell lines (HuH7, QGY-7703, BEL-7402, SMMC7721) were obtained from the Liver Cancer Institute of Fudan University (Shanghai, China). The HepG2 cell line was purchased from the American Type Culture Collection. Both the H2P and H2M cell lines were the kind gifts of Prof. Xin-Yuan Guan (Sun Yat-Sen University Cancer Center, China). All cell lines were cultured in Dulbecco modified Eagle medium with 10 % fetal bovine serum (FBS) and incubated in a humidified atmosphere of 5 % CO2 and 95 % air at 37 °C. Transfection of siRNA was performed by RNAi-max (Invitrogen) according to the manufacturer’s protocol. The siRNA sequences are listed in the Supplementary Data. In brief, transfection with the final concentration of 50 nM siRNA was conducted when the cell density was approximately 40 % in six-well plates.

S685

surgical resection to the first recurrence. None of these patients died from perioperative complications or from other factors. (See Supplementary Data for the management of fresh tissues.) All subjects selected were required to provide written informed consent on the use of clinical specimens for medical research. The study received consent from the research medical ethics committee of Sun Yat-Sen University Cancer Center. RNA Extraction and Quantitative Real-time Polymerase Chain Reaction (qRT-PCR) These procedures were performed as described previously (Supplementary Data).14 The PCR condition was as follows: 95 °C for 10 min, followed by 40 cycles of 95 °C for 5 s and 60 °C for 1 min. Using the comparative threshold cycle (2-DDCt) method, the relative expression of the target genes was normalized to the endogenous b-actin. Protein Extraction and Western Blot Test These procedures were performed as described previously.14 Briefly, total protein was isolated from cell lines or fresh tissues including hepatic tumor tissues (TTs) and their nontumor counterparts, and was separated by 10 % SDS-PAGE gels. We then followed standard procedures using primary antibodies (Supplementary Data). The protein levels were quantified using the internal loading control of b-actin proteins. Immunohistochemistry (IHC) and IHC Evaluation

Patients and Tissue Specimens In this study, HCC and adjacent nontumor liver tissues (ANLTs) samples were obtained from 246 consecutive patients who underwent curative hepatectomy for HCC at Sun Yat-Sen University Cancer Center between 2002 and 2004. All cases were conformed to the following criteria: diagnosed by postoperative histopathology; complete follow-up data available; no extrahepatic metastasis; no other malignant diseases; and no preoperative anticancer therapy; and stage I, II, and III disease (according to the 6th edition of tumor, node, metastasis (TNM) classification system of the American Joint Committee on Cancer/ International Union Against Cancer). Tumor differentiation was assessed according to the Edmonson and Steiner grading system. Data were censored at last follow-up for patients without recurrence or death. Overall survival (OS) was calculated from the day of surgery to the day of death or to the date of the last follow-up.14 Recurrence-free survival (RFS) was defined as the interval from primary

IHC in human HCC tissue samples was performed as described previously (Supplementary Data).13–15 The staining of each specimen was evaluated independently by two pathologists blinded to origination of the features and clinical outcome. The immunoreactions for CBX4 in cytoplasm and nucleus were quantified, respectively, using the widely accepted German semiquantitative scoring system, in which 0, 1, 2, 3, or 4 was assigned as the extent score of positive cells at 0, 1–25, 26–50, 51–75, and 76–100 %, respectively.16 The staining intensity was scored as 0 (negative), 1 (weak), 2 (moderate), and 3 (strong). The immunoreactive score (IRS) was determined by the product of the extent score and the intensity score. Values of IRS ranging from 0 to 12, which were graded as follows: -(IRS 0), ?(IRS 1–4), ??(IRS 5–8), ???(IRS 9–12). The median of IRS was chosen to define the cases with high CBX4 expression (IRS [4) or low CBX4 expression (IRS B4). The staining results evaluated by the two pathologists were completely in agreement with each other in approximately 81.3 % of the cases. When a

S686

discrepancy arose for any case, the two pathologists discussed it and reached the final score by consensus. MTT Assay The MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide) assay was used to measure the cell viability. Briefly, the CBX4 siRNA-transfected BEL-7402 cells or QGY-7703 cells were seeded at the density of 2500 or 3000 per well in 96-well microplates, respectively. The cells were incubated with MTT for 4 h, and the optical density was detected at 490 nm wavelength by MicroplateReader; this was measured once a day for 5 days. The results are provided as mean ± SD of three independent experiments.

FIG. 1 CBX4 was overexpressed in HCC cell lines and clinical samples. (a, b) CBX4 was overexpressed in HCC cell lines. The CBX4 mRNA (a) and protein (b) levels were measured by qRT-PCR and Western blot test. (c) CBX4 was overexpressed in HCC clinical samples; Western blot test indicated CBX4 protein level was higher in 8 HCC tumor tissues (T) than their nontumor counterparts (N); bactin was used as the loading control. CBX4 was detectable in cytoplasm (d) and nucleus (e) in HCC by IHC. Consistent with (c), the CBX4 expression, brown immunostaining in tumor cells, was much stronger than those in nontumor liver tissue. Magnification, 9 100. (f–i) The intensity of cytoplasmic CBX4 signal in 246 HCC specimens were scored as 4 grades (3, 2, 1, 0) as indicated. The positive cells were stained brown. Magnification, 9 200

B. Wang et al.

Cell Synchronization and FACS Analysis Cells were seeded to six-well plates at 30 % confluence and synchronized at the G1/S boundary by double thymidine. These procedures have been described previously.13,17,18 Briefly, cells were treated with 2 mM thymidine for 16 h, released in fresh medium containing 10 % FBS for 9 h, and incubated with 2 mM thymidine for another 16 h. At this point, about 90 % of cells were synchronized at G1/S boundary and then released secondarily; cells were collected at 0-, 3-, 6-, and 12-h time points. The cell cycle profiles were analyzed by flow cytometry (FACS), which was conducted as described previously (Supplementary Data).13,18

Prognostic Significance of CBX4 in HCC

Statistical Analysis These procedures were described previously.13–15,18 Briefly, statistical analyses were carried out by SPSS 16.0 software (SPSS, Chicago, IL). Continuous variables were applied using the independent Student’s t test. Pearson’s v2 test was used to

FIG. 2 Overexpression of cytoplasmic CBX4 indicates a poor prognosis. (a, b) OS (a) and RFS (b) curves were generated on the basis of the CBX4 protein expression statuses from 246 HCC cases. Actuarial probabilities were calculated by the Kaplan–Meier method and compared by the log-rank test. The RFS curves were generated according to the CBX4 protein expression statuses stratified by tumor size (c), differentiation (d), serum AFP (e), and TNM stage (f)

S687

analyze the relationship between CBX4 expression and clinicopathologic characteristics. Survival curves were generated by the Kaplan–Meier method and compared by log-rank test. Cox proportional hazard regression model analyses determined independent prognostic factors. The difference was considered to be statistically significant at P \ 0.05.

S688

RESULTS CBX4 Is Up-Regulated in HCC Cell Lines and Clinical Samples To investigate whether CBX4 plays an important role in HCC, we initially used multiple HCC cell lines to measure CBX4 at its mRNA and protein levels by real-time RT-PCR and Western blot test, respectively. As shown in Fig. 1a, b, both the mRNA and protein levels of CBX4 were higher in the seven HCC cell lines including BEL-7402, QGY-7703, SMMC-7721, HepG2, HuH7, H2P, and H2 M than that in LO2. More importantly, by using fresh HCC specimens, we also demonstrated that CBX4 proteins were up-regulated in all of eight HCCs compared with matched ANLTs (Fig. 1c). After that, we performed IHC using 246 HCC tissues. The CBX4 protein was detectable in 194 of 246 (78.9 %) cases (Supplementary Fig. S2). The staining of CBX4 was mainly in cytoplasm of HCC TTs (Fig. 1d). Images are shown as graded intensity of immunostaining (Fig. 1f–i). In contrast, CBX4 was barely detected in cytoplasm in matched ANLTs. Nuclear staining was also observed (Fig. 1e). The median and mean of IRS in TT cytoplasm were 4 and 4.28 (standard error [SE] 0.258), respectively. According to the cutoff value, 116 of 246 (47.2 %) TTs had a high cytoplasmic expression, and 130 of 246 (52.8 %) TTs had a low cytoplasmic expression. In matched cytoplasm of ANLTs, median and mean of IRS were 0 and 0.052 (SE 0.025), respectively. Cytoplasmic IRS of CBX4 in TTs were significantly higher compared with those in the corresponding ANLTs (Wilcoxon exact test, Z = -11.4, P \ 0.001, Fig. 1d). For CBX4 nuclear expression, the mean of IRS in TT was 1.46 (SE 0.186), which was significantly higher than those in the matched ANLTs, at 0.13 (SE 0.040, Wilcoxon exact test, Z = -6.470, P \ 0.001, Fig. 1e). Medians of nuclear IRS in TTs and ANLTs were 0. Taken together, this evidence supports the notion that CBX4 was up-regulated in HCC. Cytoplasmic CBX4 Protein Is a Prognostic Factor for Surgical HCC Patients The OS and RFS at 5 years were, respectively, 34.8 and 23.5 % for high cytoplasmic CBX4 expression and 58.7 and 38.0 % for low cytoplasmic CBX4 expression (both P \ 0.05, Fig. 2a, b). Univariate analysis revealed there were 6 and 7 risk factors associated with OS and RFS, respectively (Supplementary Table S1). The cytoplasmic, but not nuclear, CBX4 protein level was significantly correlated with OS (P = 0.003, Fig. 2a) and RFS (RFS, P = 0.012, Fig. 2b). The patients with low cytoplasmic CBX4 expression in HCC TT had better OS and RFS rates than those with high cytoplasmic CBX4 expression.

B. Wang et al.

The expression of cytoplasmic CBX4 was significantly correlated with a-fetoprotein (AFP) level (P = 0.036), tumor size (P = 0.029), differentiation (P = 0.033), and TNM stage (P = 0.032) (Table 1). To further investigate the relationship between the CBX4 expression and RFS, TABLE 1 Correlation of cytoplasmic CBX4 expression with clinicopathologic features in surgical HCC patients Variable

n

HCC cytoplasmic expression, n (%) Low

All cases

High

246 130 (52.8)

116 (47.2)

Female

29

14 (48.3)

Male

217 115 (53.0)

102 (47.0)

132 62 (47.0) 114 64 (56.1)

70 (53.0) 50 (43.9)

Negative

36

16 (44.4)

Positive

210 110 (52.4)

100 (47.6)

B25 ng/ml

90

54 (60.0)

36 (40.0)

[25 ng/ml

156 72 (46.2)

84 (53.8)

Absent

151 85 (56.3)

66 (43.7)

Present

95

45 (47.4)

50 (52.6)

Well (I)

25

19 (76.0)

6 (24.0)

Moderate (II)

109 60 (55.0)

49 (45.0)

Poor (III)

107 48 (44.9)

59 (55.1)

Gender

Age

0.898 15 (51.7)

b

B50 y [50 y

0.151

HBsAg

0.724 20 (55.6)

Serum AFP

0.036*

Liver cirrhosis

0.172

Differentiation

0.033*

Undifferentiated (IV) 5 No. of tumors Single Multiple

P-valuea

3 (60.0)

2 (40.0) 0.469

216 116 (53.7)

100 (46.3)

30

16 (53.3)

14 (46.7)

Tumor size

0.029*

B5 cm

198 109 (55.1)

89 (44.9)

[5 cm

48

30 (62.5)

18 (37.5)

Vascular invasion

0.579

Absent

240 128 (53.3)

112 (46.7)

Present

6

4 (66.7)

2 (33.3)

TNM stagec

0.032*

I

167 93 (55.7)

74 (44.3)

II

31

11 (35.5)

20 (64.5)

III

48

19 (39.6)

29 (60.4)

HCC hepatocellular carcinoma, HBsAg hepatitis B surface antigen, AFP a-fetoprotein, TNM tumor, node, metastasis classification system a

Chi square tests were performed for all the analyses

b

Median age

c

No patients had stage IV disease

* P \ 0.05

Prognostic Significance of CBX4 in HCC

S689

TABLE 2 Multivariate Cox regression analyses on overall survival and recurrence-free survival Variable

Overall survival

Recurrence-free survival HR (95 % CI)

P-valuea

NA

1.260 (0.903–1.759)

0.174

0.098

1.346 (0.982–1.844)

0.065

0.000*

3.145 (1.876–5.273)

0.000*

HR (95 % CI)

P-value

Serum AFP (B25 ng/ml vs. [25 ng/ml)

NA

Differentiation (I/II vs. III/IV)

1.339 (0.948–1.890)

No. of tumors (single vs. multiple)

3.833 (2.320–6.334)

a

Tumor size (B5 cm vs. [5 cm)

1.997 (0.672–5.937)

0.213

1.225 (0.727–2.063)

0.446

Vascular invasion (absent vs. present)

1.348 (0.496–3.663)

0.558

1.800 (0.721–4.494)

0.208

TNM stage (I/II vs. III)

0.578 (0.188–1.782)

0.340

0.977 (0.599–1.592)

0.925

CBX4 in cytoplasm (low vs. high)

1.662 (1.176–2.348)

0.004*

1.428 (1.051–1.940)

0.023*

HR hazard ratio, CI confidence interval, AFP a-fetoprotein, NA not assessed, TNM tumor, node, metastasis classification system * P \ 0.05 a

Cox proportional hazard regression model

patients were divided to subgroups by these factors. Patients in subgroups with high CBX4 expression had lower RFS rate than those with low CBX4 expression (P = 0.036, tumor size B5 cm, Fig. 2c; P = 0.028, differentiation I to II, Fig. 2d; P = 0.014, AFP [25 ng/ml, Fig. 2e; P = 0.007, TNM stage II to III, Fig. 2f). According to multivariate survival analyses (Table 2), CBX4 in cytoplasm and number of tumors were determined as independent prognostic factors for OS and RFS in this HCC cohort. We proposed a prognostic model that combined the two factors (Supplementary Fig. S4). These results demonstrated that high cytoplasmic CBX4 protein level was a poor prognostic factor for surgical HCC patients. Knockdown of CBX4 Inhibits Cell Proliferation and Delays Cell Cycle Progression Because we demonstrated that the cytoplasmic CBX4 protein was a prognostic factor for surgical HCC patients, we sought to test the functions of endogenous CBX4 in two HCC cell lines, BEL-7402, and QGY-7703.19,20 Using two different siRNAs that specifically target CBX4, we found that viabilities of QGY-7703 and BEL-7402 cells were obviously impaired by MTT assays when CBX4 was knocked down (Fig. 3a, b). Furthermore, knockdown of CBX4 delayed the cell cycle progression from the G1/S boundary synchronized by double thymidine blocks in both QGY-7703 and BEL-7402 cells (Fig. 3c–f; Supplementary Fig. S5). When we sought the potential mechanism by which CBX4 affects the cell cycle progression and the cell proliferation, we found that knockdown of CBX4 resulted in the decrease of both proliferating cell nuclear antigen (PCNA) and cyclin E2, as well as an increase in p16 (Fig. 3g, h).

DISCUSSION In present study, the up-regulation of CBX4 was determined in multiple HCC cell lines and clinical samples. Moreover, the high cytoplasmic level of CBX4 was correlated with multiple clinicopathologic features and was a new biomarker for poor prognosis in the patients with HCC. Finally, CBX4 affected cell proliferation, probably involving PCNA, cyclin E2 and p16. CBX4 is a member of CBXs and a relatively specialized PcG protein as a result of its small ubiquitin-related modifier (SUMO) E3 ligase activity, and it plays important roles in biological functions by affecting multiple key proteins, such as HIPK2, CtBP, and Bmi1.21,22 However, the roles of CBX4 in human cancers have not been explored. Using HCC cell lines and clinical samples, we demonstrated here that CBX4 was overexpressed in HCC. CBX7 could inhibit the transcription of other CBX family members including CBX4, whereas the loss of CBX7 protein has been documented in many cancers; this may contribute to the overexpression of CBX4 in HCC.23,24 In addition, HCC is characterized by aberrant viral protein expression and genomic instability, which may also account for the overexpression of CBX4 in HCC.25 Uncontrolled proliferation is a feature of cancer cells. Cyclin E binds to CDK2 and promotes cell cycle progression, and the overexpression of cyclin E has been reported to be a therapeutic target by affecting the cell proliferation and migration in HCC.26,27 PCNA participates in DNA synthesis and repair, and high expression of PCNA was associated with poor survival in patients with HCC.28 p16 often regulated by PRC and plays an important role in controlling G1/S transition by suppressing the CDK4/6 kinase activity.26,29 In this report, we found that both cyclin E2 and PCNA were decreased whereas p16 was increased

S690 FIG. 3 a, b Silencing CBX4 inhibits HCC cell growth. The viability of the 2 cells, BEL7402 (a) and QGY-7703 (b), was examined by MTT assay at different time points as indicated. The results were expressed as mean ± SD of 3 independent experiments, *P \ 0.05, **P \ 0.01, by independent Student’s t-test. c, d Silencing CBX4 delays G1/S transition in HCC cells. Cells were synchronized at G1/S boundary by double thymidine block and released, and the cell cycle profiles were measured by FACS. Endogenous CBX4 was silenced in BEL7402 (c) and QGY-7703 (d) cells, and the cell cycle profiles were measured by FACS at 12-h and 6-h time points after release from G1/S boundary, respectively. e, f Summary of 3 independent FACS tests for the BEL7402 (e) and QGY-7703 (f) cell lines, *P \ 0.05, **P \ 0.01. CBX4 siRNAs were transiently transfected into BEL-7402 cells; 6 h later, medium was replaced with fresh serum-free media and the cells were starved for another 48 h, followed by 10 % FBS medium stimulated for 12 h. g, h Western blot test (g) and qRTPCR (h) showing changes of downstream genes after knockdown of CBX4

B. Wang et al.

(a)

(b)

7402 cell viability 1.5

7703 cell viability 1.5

1.2

Scramble siCBX4-1 siCBX4-2

** 1.2

Scramble siCBX4-1 siCBX4-2

*

*

0.9

0.9 * 0.6

0.6

0.3

0.3 2

1

3 Time (day)

4

(c)

2

1

5

3 Time (day)

4

5

(d)

(e)

(f)

7402 cell percentage 100

Scramble siCBX4-1 siCBX4-2

7703 cell percentage 100

Scramble siCBX4-1 siCBX4-2

**

80

80

** ** **

60

60 *

40

40

20

20 G1

(g)

S G2/M G1

S G2/M G1

0h

6h

S G2/M

* *

G1

12h

*

S G2/M G1

S G2/M G1

S G2/M

0h

3h

6h

(h)

Scramble siCBX4-1 siCBX4-2

Relative mRNA level 1.2 0.9 0.6 0.3 CBX4

PCNA

Cyclin E2

p16

Prognostic Significance of CBX4 in HCC

when CBX4 was knocked down, indicating that CBX4 may regulate the proliferation and the cell cycle progression through cyclin E2, PCNA, and p16, and that CBX4 may be a potential therapeutic target for HCC. It has been reported that dynamic localization of the nuclear–cytoplasmic and/or subnuclear distribution of members of the CBX family occurs during the maternal-toembryonic transition.30 In addition, nuclear proteins in normal conditions are frequently overexpressed in cytoplasm in various human cancers, including HCC, melanoma, papillary thyroid carcinoma, ductal breast carcinoma, and acute lymphoblastic leukemia.31,32 In fact, it has been reported that HIPK2 phosphorylated CBX4 to control its intranuclear localization in response to DNA damage.33 CBX4 could associate with long noncoding RNAs (lincRNAs), which control the cellular localization of transcription factors.12,34 In this study, we demonstrated that CBX4 was mainly localized in the nucleus in HCC cell lines (data not shown), and that the majority of CBX4 were observed in the cytoplasm in most of HCC samples. More interestingly, the cytoplasmic CBX4 may be a useful prognostic biomarker for HCC. The reasons for this cytoplasmic location of CBX4 in HCC need to be further investigated in the future. Subgroup analyses revealed that the high cytoplasmic CBX4 expression predicts higher recurrent rate among patients with small tumor size (B5 cm) or well-differentiated disease (Edmonson I to II). These findings provide evidences for clinicians to distinguish cases with a high recurrence rate from the two subgroups, which are commonly regarded as having a better outcome. On the other hand, higher serum AFP levels ([25 ng/ml) and advanced disease stage (TNM II to III) generally indicate worse survival; the high expression of CBX4 also exerts predictive power for higher recurrence rate. Therefore, we speculate that the combination of cytoplasmic CBX4 expression levels with the above four factors may be more accurate in determining the clinical outcome and optimal treatment strategies for HCC patients after hepatectomy. In conclusion, cytoplasmic CBX4 protein levels are inversely associated with prognosis in surgical HCC patients, which could be explained by the fact that CBX4 affects cell proliferation capacity. These data provide compelling evidence that CBX4 serve as a novel prognostic predictor and may provide a promising therapeutic target for HCC. ACKNOWLEDGMENT Supported in part by key Grant from NSFC (30930045 to T. Kang), 973 projects (2010CB912201 and 2012CB967000 to T. Kang), and Key Subjects of Xinjiang Medical University–High Incidence Rate Oncology of Minority in Xinjiang (XYDXK50780306). DISCLOSURE

The authors declare no conflict of interest.

S691

REFERENCES 1. El-Serag HB. Epidemiology of viral hepatitis and hepatocellular carcinoma. Gastroenterology. 2012;142:1264–73.e1. 2. Chen D, Xing W, Hong J, et al. The beta2-adrenergic receptor is a potential prognostic biomarker for human hepatocellular carcinoma after curative resection. Ann Surg Oncol. 2012;19:3556–65. 3. Li JJ, Pan K, Gu MF, et al. Prognostic value of soluble MICA levels in the serum of patients with advanced hepatocellular carcinoma. Chin J Cancer. 2013;32:141–8. 4. Yi Y, Wu H, Gao Q, et al. Interferon regulatory factor (IRF)-1 and IRF-2 are associated with prognosis and tumor invasion in HCC. Ann Surg Oncol. 2013;20:267–76. 5. Song TJ, Fong Y, Cho SJ, et al. Comparison of hepatocellular carcinoma in American and Asian patients by tissue array analysis. J Surg Oncol. 2012;106:84–8. 6. Richly H, Aloia L, Di Croce L. Roles of the Polycomb group proteins in stem cells and cancer. Cell Death Dis. 2011;2:e204. 7. Vandamme J, Volkel P, Rosnoblet C, Le Faou P, Angrand PO. Interaction proteomics analysis of polycomb proteins defines distinct PRC1 complexes in mammalian cells. Mol Cell Proteomics. 2011;10:M110 002642. 8. Gao Z, Zhang J, Bonasio R, et al. PCGF homologs, CBX proteins, and RYBP define functionally distinct PRC1 family complexes. Mol Cell. 2012;45:344–56. 9. Bracken AP, Helin K. Polycomb group proteins: navigators of lineage pathways led astray in cancer. Nat Rev Cancer. 2009; 9:773–84. 10. Forzati F, Federico A, Pallante P, et al. CBX7 is a tumor suppressor in mice and humans. J Clin Invest. 2012;122:612–23. 11. Tan J, Jones M, Koseki H, et al. CBX8, a polycomb group protein, is essential for MLL-AF9-induced leukemogenesis. Cancer Cell. 2011;20:563–75. 12. Yang L, Lin C, Liu W, et al. ncRNA- and Pc2 methylationdependent gene relocation between nuclear structures mediates gene activation programs. Cell. 2011;147:773–88. 13. Xu S, Feng Z, Zhang M, et al. hSSB1 binds and protects p21 from ubiquitin-mediated degradation and positively correlates with p21 in human hepatocellular carcinomas. Oncogene. 2011;30:2219–29. 14. Hong J, Hu K, Yuan Y, et al. CHK1 targets spleen tyrosine kinase (L) for proteolysis in hepatocellular carcinoma. J Clin Invest. 2012;122:2165–75. 15. Tang QL, Xie XB, Wang J, et al. Glycogen synthase kinase3beta, NF-kappaB signaling, and tumorigenesis of human osteosarcoma. J Natl Cancer Inst. 2012;104:749–63. 16. Remmele W, Schicketanz KH. Immunohistochemical determination of estrogen and progesterone receptor content in human breast cancer. Computer-assisted image analysis (QIC score) vs. subjective grading (IRS). Pathol Res Pract. 1993;189:862–6. 17. Xu S, Wu Y, Chen Q, et al. hSSB1 regulates both the stability and the transcriptional activity of p53. Cell Res. 2013;23:423–35. 18. Kang T, Wei Y, Honaker Y, et al. GSK-3 beta targets Cdc25A for ubiquitin-mediated proteolysis, and GSK-3 beta inactivation correlates with Cdc25A overproduction in human cancers. Cancer Cell. 2008;13:36–47. 19. Chen R, Zhu D, Ye X, Shen D, Lu R. Establishment of three human liver carcinoma cell lines and some of their biological characteristics in vitro. Sci Sin. 1980;23:236–47. 20. Wang JB. Establishment and some characteristics of a hepatoma cell line (QGY-7703). Zhonghua Zhong Liu Za Zhi. 1981;3:241–4. 21. Kagey MH, Melhuish TA, Wotton D. The polycomb protein Pc2 is a SUMO E3. Cell. 2003;113:127–37. 22. Ismail IH, Gagne JP, Caron MC, et al. CBX4-mediated SUMO modification regulates BMI1 recruitment at sites of DNA damage. Nucleic Acids Res. 2012;40:5497–510.

S692 23. O’Loghlen A, Munoz-Cabello AM, Gaspar-Maia A, et al. MicroRNA regulation of Cbx7 mediates a switch of Polycomb orthologs during ESC differentiation. Cell Stem Cell. 2012;10:33–46. 24. Pallante P, Federico A, Berlingieri MT, et al. Loss of the CBX7 gene expression correlates with a highly malignant phenotype in thyroid cancer. Cancer Res. 2008;68:6770–8. 25. McGivern DR, Lemon SM. Tumor suppressors, chromosomal instability, and hepatitis C virus-associated liver cancer. Annu Rev Pathol. 2009;4:399–415. 26. Malumbres M, Barbacid M. Cell cycle, CDKs and cancer: a changing paradigm. Nat Rev Cancer. 2009;9:153–66. 27. Wang X, Meng X, Sun X, et al. Wnt/beta-catenin signaling pathway may regulate cell cycle and expression of cyclin A and cyclin E protein in hepatocellular carcinoma cells. Cell Cycle. 2009;8:1567–70. 28. Stroescu C, Dragnea A, Ivanov B, et al. Expression of p53, Bcl-2, VEGF, Ki67 and PCNA and prognostic significance in hepatocellular carcinoma. J Gastrointestin Liver Dis. 2008;17:411–7.

B. Wang et al. 29. Sparmann A, van Lohuizen M. Polycomb silencers control cell fate, development and cancer. Nat Rev Cancer. 2006;6:846–56. 30. Ruddock-D’Cruz NT, Prashadkumar S, Wilson KJ, et al. Dynamic changes in localization of Chromobox (Cbx) family members during the maternal to embryonic transition. Mol Reprod Dev. 2008;75:477–88. 31. Huynh H. Overexpression of tumour suppressor retinoblastoma 2 protein (pRb2/p130) in hepatocellular carcinoma. Carcinogenesis. 2004;25:1485–94. 32. Reno EM, Haughian JM, Dimitrova IK, Jackson TA, Shroyer KR, Bradford AP. Analysis of protein kinase C delta (PKC delta) expression in endometrial tumors. Hum Pathol. 2008; 39:21–9. 33. Roscic A, Moller A, Calzado MA, et al. Phosphorylationdependent control of Pc2 SUMO E3 ligase activity by its substrate protein HIPK2. Mol Cell. 2006;24:77–89. 34. Ponting CP, Oliver PL, Reik W. Evolution and functions of long noncoding RNAs. Cell. 2009;136:629–41.