Antibody against Granulin–Epithelin Precursor Sensitizes Hepatocellular Carcinoma to Chemotherapeutic Agents

Hepatocellular carcinoma (HCC) is the sixth most common cancer in the world, with an estimation of 748,000 new cases annually (1). Curative treatments, including partial hepatectomy or liver transplantation, are limited to early-stage patients (2, 3). Nonetheless, patients receiving curative surgery still have high recurrence rate (2, 3). There is currently no effective treatment for patients with advanced tumors (4). Chemotherapy is widely used to treat advanced HCC but chemoresistance is observed in the majority of the patients (5, 6). Therefore, there is an urgent demand of better therapeutic targets and treatment approaches for this aggressive disease.

Sorafenib, a multikinase inhibitor, has demonstrated survival benefits in two double blinded randomized phase III clinical trials in patients with advanced HCC (7, 8). Sorafenib alone stabilized disease in a subset of patients but did not result in tumor regression. The use of sorafenib in combination with conventional chemotherapy has been studied extensively and some promising results have been demonstrated in combination with doxorubicin, tegufar/uracil, or everolimus (9–11). The data suggest that the use of molecular-targeted therapy with conventional chemotherapy is feasible and could be an alternative for patients with advanced HCC.

Granulin–epithelin precursor (GEP; also named progranulin, acrogranin, or PC-derived growth factor) is a novel growth factor with functional role in fetal development, inflammation, and neuronal cell survival (12, 13). We have previously reported the expression patterns of liver cancers (14, 15) with overexpression of GEP in the majority of HCCs controlling proliferation, invasion, and tumorigenicity (16, 17). GEP overexpression has also been demonstrated in various human cancers including prostate, breast, and ovarian cancers (18–20). GEP was considered primarily as a growth factor; however, evidence is emerging that GEP has crucial role on regulating cancer stem cell (CSC) properties and chemoresistance (21, 22). We have previously demonstrated that GEP antibody inhibited the growth of subcutaneous xenografts (23). Here, we investigated the combined effect of GEP antibody and chemotherapeutic agents using orthotopic model systems, and examined the efficacy on targeting chemoresistance properties.

Human liver cancer cell line Hep3B was obtained from the American Type Culture Collection and authenticated by the company using short tandem repeat (STR) profiles. Huh7 was acquired from Health Science Research Resources Bank and characterized by the company using genotypes. The cell lines were immediately expanded, multiple aliquots were cryopreserved, and cells were used within 6 months of resuscitation. Both cell lines were maintained as recommended by the respective sources. The selections of cisplatin- and doxorubicin-resistant subpopulations have been described previously (21).

Hep3B, Huh7, and chemoresistant subpopulations were incubated with GEP monoclonal antibody (Versitech Ltd; ref. 23), ABCB5 polyclonal antibody (Rockland Immunochemical), chemotherapeutic drugs, Wortmannin (Cell Signaling Technology), or in combination for 24 hours and harvested for subsequent analysis.

Apoptosis was determined by flow cytometry with the Annexin V–Propidium Iodide Apoptosis Detection Kit (BD Biosciences). The percentage of apoptotic cells was quantified by flow cytometer (FACSCalibur, BD Biosciences).

The studies were approved by the Committee on the Use of Live Animals in Teaching and Research at the University of Hong Kong (Hong Kong). Hep3B and doxorubicin-resistant cells were transfected with firefly luciferase gene and implanted into the left liver lobe of nude mice (4–5 weeks old) under anesthesia. The size of tumor was monitored by the in vivo imaging system, IVIS100 (Caliper Life Sciences) after intraperitoneal (i.p.) injection of 150 mg/kg of d-Luciferin (Gold Biotechnology) under anesthesia.

There were two sets of animal experiments to investigate the therapeutic effects of chemotherapeutic agents alone. In the first set of animal experiment (Fig. 1), mice with Hep3B intrahepatic xenografts were randomized to three different groups receiving drugs through i.p. injections: control (saline), cisplatin (3.75 mg/kg/wk), or doxorubicin (3.75 mg/kg/wk; n = 8 for each group). In the second set of animal experiment (Fig. 2), mice with doxorubicin-resistant intrahepatic xenografts received control (saline) or doxorubicin (3.75 mg/kg/wk; n = 11 for both groups). In both the first and second sets of animal experiments, mice received treatments for 3 weeks and were sacrificed at the fourth week.

There were another two sets of animal experiments to assess the therapeutic efficacy of combination treatment. Mice bearing Hep3B intrahepatic xenografts were randomized to four treatment groups, with at least 5 animals per group, receiving drugs i.p.: control (saline), GEP monoclonal antibody A23 (10 mg/kg, twice weekly; Versitech Ltd.; ref. 23), cisplatin (3.75 and 5 mg/kg/wk for the third and fourth sets, respectively), or combination of the latter two regimens. In the third set of animal experiment (Figs. 5 and 6), treatment was continued for 6 weeks and the tumor was harvested at the seventh week for further analysis. In the fourth set of animal experiment (Fig. 7), mice received higher dose of cisplatin (5 mg/kg/wk). Treatment was continued for 3 weeks and animals were monitored continuously. Control and GEP antibody treatment groups were sacrificed 3 weeks after treatment cessation because of tumor burden. Cisplatin and combination treatment groups were monitored for another 6 weeks for signs of tumor recurrence/tumor growth.

Total protein was extracted with cell lysis buffer (Cell Signaling Technology) in the presence of complete protease inhibitor cocktail (Roche) and separated in 8% to 10% SDS-PAGE gel. Proteins were then electrotransferred onto polyvinylidene difluoride membranes and subsequently incubated with primary anti-human antibody at 4°C for overnight. Detection was performed by horseradish peroxidase–labeled secondary antibodies at room temperature for 1 hour and visualized with the Enhanced Chemiluminescence Western Blotting Detection Kit (Amersham Biosciencse). These experiments were repeated at least twice.

Freshly isolated tumor cells were seeded at a density of 1,000 cells per well in 6-well plates. For spheroid formation, cells were cultured in serum-free medium supplemented with human recombinant EGF, insulin (Sigma-Aldrich), human recombinant basic fibroblast growth factor (bFGF), and B27 (Invitrogen) for 28 days in ultra-low attachment plate. Spheroids were counted at magnification of ×100 under light microscopy when they reached diameter of 50 μm. For colony formation, cells were allowed to grow for 14 days and colonies were stained by crystal violet (Sigma-Aldrich) and counted.

For expression of GEP and ABCB5, single-cell suspensions were permeabilized with ice-cold 0.1% saponin and then incubated with FITC-conjugated mouse anti-human GEP, unconjugated goat anti-human ABCB5, or equal amounts of corresponding isotype controls (22). Single-cell suspensions were then incubated with PerCP-conjugated donkey anti-goat IgG secondary antibody (R&D Systems) if the primary antibody was not conjugated with fluorochrome. For surface expression of CD133, cells were incubated with allophycocyanin (APC)-conjugated mouse anti-human CD133 (Miltenyi Biotec), or equal amount of corresponding isotype controls, before permeabilization and staining with GEP. The percentage of positive staining cells was determined by flow cytometer (FACSCalibur, BD Biosciences).

The level of GEP in mouse serum (dilution 1:100) or culture medium (dilution 1:3) was quantified with the Progranulin (human) ELISA Kit (Adipogen International) according to the manufacturer's instructions. All samples were examined in duplicate and the experiments were repeated at least three times.

Immunohistochemistry for apoptosis and cell proliferation was performed on formalin-fixed paraffin-embedded sections with in situ cell death detection kit peroxidase (POD) (Roche Diagnostics) and Ki-67 staining (BD Bioscience) according to the manufacturer's protocol. Detection was visualized with Dako Envision Plus System (Dako) following the manufacturer's instruction. Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) and proliferation-positive cells were calculated by averaging the number of stained cells for the five randomly selected fields at the magnification of ×200 and ×400, respectively, under light microscopy.

All analyses were performed using the statistical software GraphPad Prism version 3.00 for Windows (GraphPad Software) or SPSS version 16.0 for Windows (SPSS Inc.). Continuous variables were assessed by the Pearson correlation analyses. One-way analysis of variance (ANOVA) or the Student t test was used to compare between groups. A P value less than 0.05 was considered statistical significant.

Materials and Methods details have been described in Supplementary Data.

Chemotherapy with either cisplatin or doxorubicin delayed the growth of human Hep3B intrahepatic tumors in the orthotopic xenografts nude mice model. Tumors were implanted into the liver, and allowed to grow into significant size (greater than 10⁶ photons/s/cm²/sr). Animals were randomized into three groups to receive systemic drug administration through i.p. injection of either saline (control), cisplatin (3.75 mg/kg), or doxorubicin (3.75 mg/kg) once weekly for 3 weeks. The tumor-bearing animals revealed no signs of disability or behavior abnormalities. Growth inhibition of the intrahepatic tumors was observed as reduced bioluminescence signals in mice received chemotherapeutic agents throughout the treatment course (Fig. 1A and B; ANOVA, P = 0.090) and also reduction of tumor volume at treatment endpoint (Fig. 1C; ANOVA, P = 0.028).

For the HCC cells isolated from the treated xenografts, the residual cells that survived after chemotherapeutic treatment showed enhanced ability on colony formation (Fig. 1D; ANOVA, P = 0.413). In addition, cisplatin-treated cells possessed greater ability of spheroid formation (Fig. 1E; ANOVA, P = 0.140). Flow cytometry analysis for hepatic CSC marker demonstrated that the HCC cells selected under chemotherapeutic agents showed increased expression of CD133, GEP, and ABCB5 (Fig. 1F; ANOVA, P = 0.029, 0.058 and 0.298, respectively). These results demonstrated that the residual cancer cells that survived under chemotherapy showed enhanced expressions on CD133, GEP, and ABCB5, and improved ability on colony and spheroid formations. Note that some of the differences were not statistically significant and only showed the trend of enhancement on cancer-aggressive properties.

To further investigate the in vivo response of chemoresistant HCC cells, we have inoculated the HCC cells that have been selected under doxorubicin over extended period of time [doxorubicin-resistant cells, details described previously (21)] into the liver organs of nude mice. Mice that established intrahepatic tumors were randomized and subjected to either saline (control) or doxorubicin (3.75 mg/kg) treatment once weekly for 3 weeks. The animals showed no signs of disability or behavior abnormalities. However, approximately 10% loss of body weight was observed in both the control and treatment groups of animals by the experimental endpoint (third week). Notably, the cachexia syndrome was only observed in the mice bearing the doxorubicin-resistant xenografts, but not the Hep3B parental xenografts. To examine the effect of combination therapy, the subsequent animal experiments therefore focused on the Hep3B parental xenografts, else the extra cachexia burden of the chemoresistant cell populations on the mice might interfere with data interpretation.

Nonetheless, though doxorubicin-resistant cells showed doxorubicin resistance in vitro, systemic doxorubicin treatment is still capable of inhibiting intrahepatic tumor growth as reflected by the bioluminescence signals throughout the treatment course (Fig. 2A and B) and the tumor volume recorded at the endpoint (Fig. 2C). Importantly, cells isolated from doxorubicin-treated doxorubicin-resistant xenografts demonstrated increased ability on both colony formation (Fig. 2D) and spheroid formation (Fig. 2E). Notably, spheroids were only observed in doxorubicin-resistant xenografts treated with doxorubicin, but not in the xenografts that received control treatment and Hep3B parental cells that received doxorubicin treatment (Fig. 1E). Subsequent flow cytometry analysis demonstrated an increased expressions of CD133, GEP, and ABCB5 in the cells isolated from the doxorubicin-treated xenografts (Fig. 2F). Both HCC xenograft models, established from Hep3B and doxorubicin-resistant cells, exhibited consistent data that treatment of chemotherapeutic agents led to increased hepatic stem cell marker levels CD133, GEP, and ABCB5, and enhanced cancer-aggressive properties including colony and sphere formation abilities.

Because chemotherapy increased GEP expressions in parental and chemoresistant HCC cells, GEP antibody was used to investigate the role of GEP on apoptosis induction. GEP antibody alone did not induce apoptosis, but sensitized the cells to apoptosis when combined with cisplatin or doxorubicin compared with single-agent chemotherapy (Fig. 3A) while control mouse IgG showed no effect (Supplementary Fig. S1). This sensitization effect on apoptosis induction by GEP antibody was observed in Hep3B and Huh7 HCC cells, as well as the chemoresistant cells derived from Hep3B [cisplatin-resistant cells and doxorubicin-resistant cells, as described previously (21)].

The two common survival pathways Akt–PKB and MEK–ERK were investigated for their involvements in the synergistic functions of GEP antibody in chemotherapy. Treatment of both cisplatin and doxorubicin increased the cellular expression levels of GEP and secretion of GEP, and the expressions of ABCB5 in both HCC cell lines (Fig. 3B–C; Supplementary Fig. S2). These data echoed with the observation from the xenografts (Fig. 1F and Fig. 2F). Importantly, combined GEP antibody and chemotherapeutic agent could counteract the chemodrug-alone–induced effects, including the upregulated GEP cellular and secretory levels, upregulated ABCB5 cellular level, and the activation of the PDK, Akt, MEK, and ERK signaling molecules. In addition, both chemodrugs induced Bcl-2 (antiapoptotic) upregulation, while combination with GEP antibody suppressed this induction and reverted to the balance of the Bax (proapoptotic) to Bcl-2 (antiapoptotic) ratio. Similarly, combined ABCB5 antibody and chemotherapeutic agent could suppress the induction of hepatic stem cell markers by chemotherapeutic agent treatment (Fig. 4).

To address the role of Akt in GEP-regulated signaling pathways, Wortmannin, the PI3K/Akt inhibitor, was used alone or in combination with chemotherapeutic agents (Fig. 3B). In the presence of Wortmannin, chemotherapy induced expressions of ABCB5 and Bcl-2 were suppressed, but the GEP upregulation induced by chemotherapeutic agents was independent from Akt inhibition. This part of the study indicated that GEP antibody treatment enhanced chemotherapy response through suppression of ABCB5 levels, and reduced cell survival signaling through Akt and Bcl-2.

We have demonstrated previously that GEP antibody treatment decreased the serum GEP levels and inhibited the growth of subcutaneous xenografts (23). In this study, we investigated the combined effect of GEP antibody with conventional chemotherapeutic agent cisplatin on HCC orthotopic xenografts. Immunocompromised mice with established intrahepatic human HCC xenografts were randomized to receive various systemic treatments through i.p. injections, including saline (control), GEP antibody (10 mg/kg, twice weekly), cisplatin (3.75 mg/kg, once weekly), or combination of the latter two regimens for 6 weeks. No signs of disability, behavior abnormalities, or significant changes on body weight were noted during treatment. No impairment on the vital organs, including hearts, lungs, kidneys, spleens, or pancreas, and no tissue damage, such as fibrosis, in normal livers was observed. GEP antibody treatment alone was able to inhibit the growth of intrahepatic HCC compared with control treatment. Similarly, cisplatin also delayed tumor growth. Notably, combined treatment demonstrated prominent inhibitory effect on intrahepatic HCC growth revealed by the reduced bioluminescence signals (Fig. 5A and B; ANOVA, P < 0.001), and the decreased tumor volume (Fig. 5C; ANOVA, P < 0.001). Systemic GEP antibody treatment decreased the serum GEP levels, and combination therapy with cisplatin resulted in further reduction of serum GEP levels (Fig. 5D; ANOVA, P < 0.001). In addition, the serum GEP levels were significantly associated with tumor burden as demonstrated by the correlation with tumor bioluminescence signals (Fig. 5E; r = 0.719, P < 0.001) and tumor volume (Fig. 5F; r = 0.718, P < 0.001). This part of the study showed that combined GEP antibody and cisplatin treatment further inhibited tumor growth when compared with single-agent treatment, and that serum GEP levels reflected tumor burden.

The treatment effects on apoptosis and proliferation have been investigated by TUNEL assay and Ki-67 immunohistochemistry staining, respectively. Either GEP antibody or cisplatin treatment alone already induced apoptosis and inhibited proliferation (Fig. 6A and B). Combination therapy demonstrated synergistic effect on apoptosis induction as reflected by the increment of TUNEL-positive cells (ANOVA, P < 0.001) and suppression of proliferation as indicated by the reduction of Ki-67-positive cells (ANOVA, P = 0.020).

The tumors after treatment were further analyzed by flow cytometry for hepatic stem cell marker expressions (Fig. 6C). Reduced CD133-positive populations were observed in GEP antibody treatment (24.26%) and combination with cisplatin treatment (21.74%) compared with the control saline–treated group (33.01%) and cisplatin-alone treatment (31.45%). Notably, GEP and ABCB5 have been demonstrated previously with fundamental function on HCC chemoresistance and CSC traits (21). Importantly, cisplatin treatment selected a subpopulation of HCC cells enriched with GEP and ABCB5-positive cells (9.64% and 37.01%, respectively) compared with control (5.92% and 21.32%, respectively) and GEP antibody treatment (5.18% and 22.79%, respectively). Nonetheless, combined GEP antibody and cisplatin treatment resulted in a reduced expression of GEP and ABCB5 (6.57% and 26.55%, respectively) compared with the cisplatin-alone treatment.

The residual tumors posttreatments were also examined by Western blot analysis (Fig. 6D). Chemotherapy treatment activated the Akt and ERK and upregulated bcl-2. The combination therapy inhibited the cell survival pathways of Akt and ERK when compared with chemotherapy alone. Combination therapy also disturbed the balance between proapoptotic (Bax) and antiapoptotic (Bcl-2) members through the downregulation of Bcl-2. These data indicated that GEP antibody therapy suppressed the aggressive phenotype and molecular changes of the residual HCC cells (chemoresistant cells) that survived under chemotherapy.

High dose of cisplatin (5 mg/kg) was used to further examine the therapeutic effect of combination therapy on the same mice model with intrahepatic human HCC (Fig. 7). Combined GEP antibody and high dose of cisplatin treatment eliminated one of the intrahepatic HCC after 1 week of treatment (1 of 7, 14%), and eradicated all the established tumors by the third week (7 of 7, 100%). High-dose cisplatin-alone treatment could not eliminate any established tumor in 1 week, and two tumors showed complete regression by the third week (2 of 6, 33%). Because of the concern on the drop of body weight under high-dose chemotherapy, and that the combined treatment regimen had already eradicated all tumors by the third week, the treatment was discontinued for all therapeutic groups after 3 weeks of treatment. There were significant tumor signals differences among the control and treatment groups (ANOVA, P = 0.001). The animals were monitored continuously for tumor recurrence and growth.

GEP antibody treatment alone for 3 weeks was able to inhibit tumor growth compared with control treatment (2.22 vs. 3.39 ×10⁸ photons/s/cm²/sr; inhibition = −35%; Fig. 7A and B). Notably, after cessation of treatment from the fourth to fifth to sixth week, GEP-treated tumors grew less rapid compared with control (4.17, 5.04, and 8.59 vs. 5.35, 12.41, and 13.71 × 10⁸ photons/s/cm²/sr; thus growth by the fifth week = +21% vs. +132%, and by the sixth week = +107% vs. +156%). By the sixth week, the differences on tumor signals remained significant among the control and treatment groups (ANOVA, P < 0.001). As there was significant tumor burden and with the concern on drop of body weight, the control and GEP antibody treatment groups were terminated.

Monitoring continued for the high-dose cisplatin group and combination treatment group. Throughout the monitoring period without treatment, all the animals remained tumor-free and there was no tumor recurrence in any animals in the combination group that had received GEP antibody plus high-dose cisplatin (Fig. 7; Supplementary Fig. S3). For the group that was treated with high-dose cisplatin alone, there were two more complete regression (4 of 6, 67%) by the fourth week, and since then no further change on the incidence of complete tumor regression was seen. Nonetheless, the two residual tumors after cisplatin treatment grew rapidly (fourth to fifth to sixth week = 3.44, 5.57, and 15.56 × 10⁵ photons/s/cm²/sr; thus growth by the fifth week = +62%, and by the sixth week = +352%). By the 12th week, there was significant tumor burden in the two mice bearing the cisplatin-treated tumor, and with the concern on drop in body weight, all the remaining animals were sacrificed. For the combined treatment group that received GEP antibody plus high-dose cisplatin, histologic examination of the liver tissues revealed no residual tumor cells, and these data corroborated the bioluminescence signal analysis. These results further consolidated the therapeutic potential of combination approach on GEP antibody therapy plus conventional chemotherapy for the treatment of HCC.

The work in this study is the first report to show that chemotherapy enriched a subpopulation of HCC cells with increased GEP/ABCB5/CD133 expressions, and the enriched aggressive cell populations could be targeted by GEP antibody. Notably, both the Hep3B and chemoresistant xenografts demonstrated the enrichment of CD133 cells postchemotherapy (Figs. 1F and 2F). We and others have shown that HCC cells with CD133 expressions defined the CSCs populations with enhanced chemoresistance (21, 22, 24, 25). In addition, our group previously showed that GEP and ABCB5 double-positive cells demonstrated enhanced self-renewal abilities and chemoresistance in hepatic CSCs (21, 22). ABCB5 expressions were enhanced in response to chemotherapeutic treatments (Figs. 3 and 4) and also in chemoresistant cells (21). ABCB5 expressions have been demonstrated to associate with CSCs properties in melanoma (26) and oral squamous cell carcinoma (27). However, reimplantation of the chemotreated residual cells to test the tumor-forming abilities should be attempted if there have enough cells for the experiment after chemotherapy. Although enhanced GEP and ABCB5 expressions have been demonstrated in a number of human CSCs, whether other cell types would express GEP and ABCB5 remain to be elucidated. Nonetheless, the presence of CSCs has been demonstrated in a number of solid tumors, including brain, colon, breast, prostate, pancreatic, and liver cancers (24–32). The enrichment of CSCs after chemotherapy has recently been revealed in breast, lung, and liver cancers (33–35). The present study showed that the HCC xenografts survived postchemotherapy had enhanced colony and sphere formation abilities, and with increased GEP/ABCB5/CD133 expressions. Importantly, GEP antibody could sensitize the naïve HCC cells and the chemoresistant subpopulations to chemotherapy, and combination of GEP antibody with chemotherapeutic agent could eradicate the intrahepatic xenografts.

GEP antibody neutralized the tumor-secreted GEP in the cell culture supernatant in vitro and in the serum in vivo, and also suppressed the cellular expression of GEP and ABCB5 induced by chemotherapy. The effect of GEP antibody could be explained by the autocrine and paracrine nature of the growth factor GEP (12), thereby neutralizing the secreted GEP would disrupt the feedback loop, and therefore decrease the intracellular GEP levels. We have reported the regulatory role of GEP on ABCB5, and cells with coexpressions of GEP/ABCB5/CD133 demonstrated enhanced CSCs properties (17, 18). Therefore, GEP antibody would decrease the extracellular GEP, subsequently intracellular GEP/ABCB5, thus the GEP antibody treatment targeting the aggressive cell populations. Furthermore, GEP antibody inhibited the prosurvival pathways Akt–PKB and MEK–ERK induced by chemotherapy. The combination treatment also disturbed the balance between proapoptotic (Bax) and antiapoptotic (Bcl-2) members through the inhibition of Bcl-2. Other research groups have reported that GEP overexpression led to the development of tamoxifen and letrozole resistance through the protection of Bcl-2 in breast cancers (36, 37). The roles of Akt–PKB and MEK–ERK in apoptosis induced by cytotoxic drugs, and the interactions between the Akt–PKB pathways and Bcl-2 family members in the regulation of apoptosis have been described extensively (38–40). PI3K–AKT inhibitor Wortmannin was used to evaluate the role of Akt–PKB in GEP antibody–enhanced apoptosis induced by chemotherapy. Wortmannin inhibited the expression of Bcl-2 induced by chemotherapeutic agents. There were reports on GEP binding to tumor necrosis factor receptor (TNFR; ref. 41) and the potential role of GEP antibody in patients with arthritis (42). Nonetheless, the binding of GEP and TNFR in different cell types and the functional roles remain to be elucidated (43, 44). This study demonstrated that GEP antibody sensitized HCC cells to chemotherapeutic agents through disruption of Akt/Bcl-2 apoptosis signaling, inhibition of drug efflux through suppression of GEP/ABCB5.

In summary, the study presented here demonstrated the significance of GEP in HCC chemoresistance and the therapeutic potential of combination therapy. Antibodies have already been demonstrated to be effective therapeutic agents for various cancers (45, 46). We demonstrated that GEP antibody in combination with chemotherapeutic agent was able to eradicate all intrahepatic xenografts in orthotopic mice models. Combination treatment approach might serve as therapeutic options for the patients with chemoresistant HCC.

The University of Hong Kong has filed patent applications for the described works. The authors P.F.Y. Cheung, S.T. Fan and S.T. Cheung are inventors of these patents. S.T. Cheung has received Pfizer collaborative research grants. I.O.-L. Ng reports receiving other commercial research support from a contract research with Pfizer Inc. No potential conflicts of interest were disclosed by the other authors.

Conception and design: N.C.L. Wong, S.T. Cheung

Development of methodology: N.C.L. Wong, S.T. Cheung

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): N.C.L. Wong, P.F.Y. Cheung, C.W. Yip, S.T. Cheung

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): N.C.L. Wong, P.F.Y. Cheung, K.F. Chan, S.T. Cheung

Writing, review, and/or revision of the manuscript: N.C.L. Wong, P.F.Y. Cheung, C.W. Yip, I.O.-L. Ng, S.T. Fan, S.T. Cheung

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): P.F.Y. Cheung, S.T. Cheung

Study supervision: S.T. Cheung

This study was supported, in part, by Hong Kong Research Grants Council (GRF 764111 and 764112 to S.T. Cheung, HKU7/CRG/09 to I.O.-L. Ng), Sun C.Y. Research Foundation for Hepatobiliary and Pancreatic Surgery (to S.T. Fan), and Seed Funding Program of the University of Hong Kong (to S.T. Cheung).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.