Hepatitis B virus X protein (HBx) enhances centrosomal P4.1-associated protein (CPAP) expression to promote hepatocarcinogenesis

Cell culture, reagents, and antibodies

HCC cell lines Huh7, SK-Hep1, Hep3B, Hep3BX, HepG2, and HepG2X were maintained in Dulbecco’s modified Eagle’s medium (DMEM, Sigma, St. Louis, MO). Hep3BX and HepG2X are HBx stably expressed Hep3B and HepG2 cells, respectively. HepAD38/Tet-off cells [14] were grown in DMEM supplemented with 200 mM GlutaMax (Gibco, Carlsbad, CA), 0.3 μg/ml doxycycline, and 400 μg/ml G418. HepG2-hNTCP-C4 cells [5] were cultured with DMEM/F-12 (Sigma) with 200 mM GlutaMax, 50 μM hydrocortisone, 5 μg/ml insulin, and 400 μg/ml G418. Media was supplemented with 10% fetal bovine serum, 100 μg/ml streptomycin, and 100 units/ml penicillin. The cells were incubated at 37 °C in a humidified chamber containing 5% CO₂. Human TNF-α, cycloheximide, and anti-α-tubulin antibodies were purchased from Sigma. The antibody against GFP was purchased from Clontech (Palo Alto, CA). Antibody against HA was purchased from Covance (Berkeley, CA). Antibodies against GAPDH and NF-κB/p65 were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The ON-TARGET plus SMARTpool siRNA for knockdown of CPAP and NF-κB/p65 were obtained from Dharmacon RNA Technologies (Lafayette, CO). shGFP and shCREB were obtained from National RNAi core facility in Taiwan.

Xenograft tumorigenicity assay

Hep3B cells (2 × 10⁶) stably expressed GFP, GFP-CPAP/WT, GFP-CPAP/MT [34] were injected subcutaneously into the right flank region of 5-week-old male NOD-SCID mice. The tumor volumes were measured using calipers every 3 days. Tumor size was measured using the formula: length X width² X 0.5. At the 28th day after injection, all mice were sacrificed and tumors were weighed and photographed.

Co-immunoprecipitation assay and Western blot analysis

Co-immunoprecipitation assay and Western blot were performed as described [34]. Lysates were analyzed by immunoblot analysis using the specific antibodies as indicated in the text. Specific bands were detected with a horseradish peroxidase-conjugated antibody and revealed by an enhanced chemiluminescence (ECL) Western blot system (PerkinElmer, Waltham, MA).

Chromatin immunoprecipitation (ChIP) and re-chromatin immunoprecipitation analysis

The ChIP assay was performed as described previously [11]. Briefly, the DNA-protein complex was eluted in elution buffer (1x TE buffer containing 1% SDS) with rotation at room temperature for 15 min, and immune complex crosslinking was reversed by heating to 65 °C overnight, followed by treating with 100 μg/ml of proteinase K at 50 °C for 1 h. DNA was extracted twice with phenol/chloroform and precipitated with ethanol. The DNA pellet was re-dissolved in H₂O and subjected to Q-PCR amplification using specific primers for the CPAP promoter (sequence information provided in Additional file 1: Table S1). For re-ChIP, chromatin immunoprecipitates from the first ChIP were desalted by re-chip buffer (50 mM Hepes, 140 mM NaCl, 1 mM EDTA, 1% Triton X-100, 0.1% NaDOC, 0.1% SDS, and 0.5 mM PMSF), incubated with the second antibody or normal-serum control, and processed as above.

Inducible expression of HBV and HBV infection

HepAD38/Tet-off cells were grown in DMEM with 0.3 μg/ml doxycycline, and the production of HBV was induced by removing doxycycline. HepG2-hNTCP-C4 cells were seeded at 5 × 10⁵ cells/well in a 6 cm collagen-coated plate. HBV derived from HepAD38/Tet-off cells was infected into HepG2-NTCP-C4 cells at 1000 genome equivalent (GEq)/cell; after 12 h infection, the media was replaced with fresh DMEM/F12 containing 2% FBS, 4% PEG 8000. The infected cells were washed three times with PBS at 0 day post infection (dpi), then incubated in fresh DMEM/F12 supplemented with 10% FBS. The cells were collected at 6 dpi for further analysis.

Luciferase reporter assay

HCC cells were grown in 24-well plate and transfected with a mixture of different vectors as indicated in the figure legends using PolyJet transfection reagent (SignaGen, Rockville, MD) or Lipofectamine reagent (Invitrogen, Grand Island, NY). Luciferase activities were measured with Briteplus (PerkinElmer) after 24 or 48 h transfection. All experiments were repeated at least three independent times in triplicate.

shRNA and RNA interference transfection

CPAP knockdown experiment is based on our previous reports [5, 34]. Briefly, CPAP siRNA duplex (5′-GGACUGACCUUGAAGAGAAUU-3′) and pSUPER-shCPAP (5′-AAUGGAAUGCACGUGACGAUG-3′) were used by transfecting into HCC cells using Lipofectamine 2000 reagent according to the manufacturer’s instructions (Invitrogen).

Subclone of GFP-HBx

HBx DNA obtined from Dr. Tasi [33] and subcloned into pEGFP-C1 vector by EcoRI restriction enzyme site.

In situ proximity ligation assay (PLA)

HCC cells were grown on sterile cover slips, and after 24 h transfection with HA-CPAP/WT or MT [34] and GFP-HBx, cells were fixed in 3.7% formaldehyde for 10 min and then in situ PLA was performed according to the manufacturer’s instructions (Olink Bioscience, Uppsala, Sweden). Two primary antibodies derived from different species were used to recognize CPAP and GFP. Secondary antibodies were species-specific PLA probes. The interaction of proteins was amplified as distinct bright-red spots and detected using a fluorescence microscope (Personal DV Applied Precision, Issaquah, WA).

RNA extraction and quantitative real-time PCR

Total RNA was extracted using TRIsure reagent (Bioline, London, UK). cDNA was synthesized with the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Grand Island, NY). Quantitative real time-PCR assay to detect mRNA expression were conducted using iQ™ SYBR® Green Supermix (Bio-Rad, Hercules, CA). Actin was used as an internal control. Primers specific for human genes are described in Additional file 1: Table S1.

Colony formation assay

For clonogenicity analysis, 2 × 10³ cells/well were seeded in 6-well plate and cultured in complete medium for 10~15 days. Colonies were fixed with formaldehyde and stained with 0.1% crystal violet.

Cell proliferation assay

Cells (2 × 10³ cells/well) were seeded in 96-well plate and maintained in complete medium overnight. Cell proliferation was performed using the CCK-8 and BrdU incorporation at indicated times according to the manufacturer’s instructions.

In vitro trans-well migration assay

Cells were resuspended in serum-free medium and 400 μl of cell suspension (1.5 × 10⁵ cells) were placed into the upper chamber (Millicell chambers, Millipore, Billerica, MA). Medium containing 10% FBS were added to the lower chamber. After 20 h of incubation, the cells remaining on upper chamber were removed, and the attached cells on the lower membrane were formaldehyde-fixed and stained with 0.1% crystal violet. The migration rate was quantified by counting the migrated cells in five random fields using an inverted microscope.

The cancer genome atlas (TCGA) data set

Data from the TCGA data set was used to analyze the overall survival (OS) curves using SurvExpress biomarker validation tool [1].

Statistical analysis

Statistical differences were assessed between experimental groups using two-tailed and un-paired Student’s t tests. *, p < 0.05; **, p < 0.01; ***, p < 0.001.

Overexpression of CPAP in HBV-associated HCC

HBx transcriptionally up-regulates CPAP by interacting with CREB

By Western blot analysis, we found that HBx can increase CPAP expression in HCC cells (Fig. 1A-i). The expression level of CPAP mRNA was higher in HBx stable expression cell lines Hep3BX or HepG2X than Hep3B or HepG2 (Fig. 1A-ii). The CPAP promoter assay further demonstrated that HBx transcriptionally up-regulates the CPAP promoter (Fig. 1A-iii, and Additional file 2: Figure S1). The same result was obtained from HBV-infected HCC or HBV genome-expressing HCC cells [14]. Expression of CPAP mRNA and protein was increased in HBV-infected (Fig. 1B) or HBV genome-expressed HCC (Additional file 2: Figure S2). To identify the specific transcription factors involved in HBx-enhanced transcriptional activation of CPAP, TFSEARCH (http://cbrc3.cbrc.jp/papia/howtouse/howtouse_tfsearch.html) was used to evaluate the CPAP promoter, and two potential CREB binding sites (CBSs) were found (Fig. 1C-i, − 226 to − 215 bp and − 86 to − 79 bp). CPAP promoter activity was significantly decreased when CREB binding site 1 was mutated (M1), whereas a mutation in CREB binding site 2 (M2) only moderately decreased the promoter activity compared with the wild type control (Additional file 2: Figure S3). Given that HBx has been shown to interact with CREB-binding protein/p300 to mediate CREB-dependent transcription [20, 37], we assumed that HBx might activate the CPAP promoter through the interaction with CREB. As anticipated, overexpression of CREB or HBx increased CPAP promoter activity, which was further enhanced by co-expression of HBx and CREB (Fig. 1C-ii and Additional file 2: Figure S4A). Overexpression of the CREB dominant-negative (DN) mutant had a smaller effect on CPAP promoter activity (Additional file 2: Figure S4B and C); knocked-down expression of CREB led to decreased CPAP promoter activity (Additional file 2: Figure S5A and B). Either HBx or CREB can induce wild type CPAP promoter activity but failed to affect CPAP/M1 or M1 + 2 promoter activities (Fig. 1D and Additional file 2: Figure S4D). Overexpression of HBx retained the ability to activate CPAP promoter in CREB knocked-down cells (Additional file 2: Figure S5C), while the CREB expression is still maintained in a low level (Additional file 2: Figure S5A).

We further investigated whether the transcriptional up-regulation of CPAP by HBx was due to HBx interacting with CREB and binding to the CPAP promoter. ChIP assay confirmed that overexpressed CREB can enhance the association of HBx with the CPAP promoter, re-Chip assay further demonstrated that CREB can form a complex with HBx to bind to the CPAP promoter (Fig. 1E). By contrast, CREB and HBx did not associate with the CPAP/M1 promoter (Fig. 1F). These results indicated that the CPAP promoter − 86 to − 79 bp is a cis-regulatory element for HBx-mediated transcriptional activation; HBx enhances CPAP expression by binding onto the CPAP promoter through the association with CREB.

CPAP and HBx cooperatively enhance NF-κB activation

It has been well recognized that NF-κB is a transcription factor regulated by HBx, and contributes to tumorigenesis by activating the expression of several genes involved in the immune response, inflammation, cell proliferation and the prevention of apoptosis [22]. On the other hand, it has been reported that CPAP is essential for the transcriptional activation of NF-κB and synergistically increases HBx-mediated NF-κB activation [13, 34]. Considering the data from clinical specimens of HBV-HCCs and the critical role of CPAP in NF-κB activation, we investigated the molecular functions of CPAP and HBx in NF-κB signaling. The results showed that the either overexpression of CPAP or HBx both increases the transcriptional activity of NF-κB, and the co-expression of CPAP and HBx induced a substantial increase in NF-κB-responsive transcriptional activity (Additional file 2: Figure S6A, left panel). Knockdown of CPAP by siRNA diminished the HBx-induced NF-κB activation (Additional file 2: Figure S6A, right panel), confirming that CPAP is essential for HBx-enhanced NF-κB activation. The same effects were found using the IL-8 promoter, activation of which is dependent on NF-κB activity (Additional file 2: Figure S6B and C). Individual expression or co-expression of CPAP and HBx increased the expression of the NF-κB downstream targets TNF-α, IL-8 and ICAM-1 (Fig. 2a), whereas CPAP knockdown reduced the induction of NF-κB-mediated gene expression by HBx (Fig. 2b).

To further investigate how CPAP and HBx cooperate to enhance NF-κB activity, we first checked the association between CPAP and HBx in the NF-κB pathway. Co-immunoprecipitation analysis and in situ PLA indicated that CPAP interacts with HBx (Fig. 2 c, d and Additional file 2: Figure S7A). Our previous report indicated that TNF-α-mediated SUMO-1 modification of CPAP is important for NF-κB activation [34]. Here, we investigated whether SUMO modification of CPAP is required for HBx-induced NF-κB activation. The results showed that SUMO-deficient CPAP did not interaction with HBx (Fig. 2 c, d and Additional file 2: Figure S7A), and decreased the HBx-enhanced NF-κB transcriptional activity (Fig. 2e). Furthermore, the interaction between CPAP and HBx was increased upon TNF-stimulation (Additional file 2: Figure S7B). These findings suggest that SUMO modification is essential for CPAP to associate with HBx and cooperatively enhance NF-κB activity.

SUMO modification is required for CPAP to promote HCC proliferation and tumorigenicity

Next, we examined the role of CPAP in HCC development. HepG2 cells that stably expressed GFP-CPAP/WT exhibited a higher proliferative rate and increased colony formation compared with GFP-CPAP/MT and control cells (Additional file 2: Figure S8A). Moreover, NF-κB/p65 knockdown decreases CPAP-enhanced colony formation in Hep3B cells (Additional file 2: Figure S9), suggesting that the CPAP-enhanced proliferation of HCC is dependent on NF-κB activity. To further determine the effect of CPAP on hepatoma cell growth in vivo, GFP-CPAP/WT or GFP-CPAP/DM stably expressed HCC cells were subcutaneously injected into NOD/SCID mice. The results showed that overexpression of CPAP/WT significantly increased tumor growth compared with CPAP/DM (Additional file 2: Figures S8B and S10). These results suggest that CPAP promotes HCC growth both in vitro and in vivo.

CPAP and HBx cooperatively enhance HCC proliferation, migration and tumorigenesis

Next, we evaluated the contribution of CPAP in HBx-mediated hepatocarcinogenesis by investigating the effects of CPAP on the tumorigenic properties of HCC cells. The results indicated that Hep3B cells expressing either CPAP or HBx showed a higher proliferative rate and colony number than Hep3B cells; co-expression of CPAP significantly increased HBx-enhanced cell growth (Fig. 3 a, b and c/upper panel). Moreover, expression of CPAP in Hep3BX cells enhanced HBx-induced cell migration ability (~ 1.5 fold, Fig. 3c/lower panel). To further validate the effect of CPAP in these observations, CPAP was knocked down in Hep3BX cells, and the results showed an impairment of HBx-induced cell growth and migration (Fig. 3 d). These results support that CPAP is crucial for HBx-induced tumorigenesis in HCC.

CPAP is required for maintaining TNF-α-mediated HBx protein stability through enhancing the NF-κB activity

Previous study demonstrated that TNF-α can induce a notable accumulation of HBx by increasing protein stability due to reduced proteasomal degradation through NF-κB signaling. Interestingly, we found that the protein expression, but not the mRNA, of GFP-HBx was increased in HA-CPAP-overexpressing cells (Fig. 4a), indicating that the prolonged protein stability of GFP-HBx may under the regulation of HA-CPAP. Since CPAP is a co-activator of IKK-mediated NF-κB activation in response to TNF-α treatment [34], we further clarified whether CPAP can increase the protein stability of HBx through enhancing TNF-α-induced NF-κB activation. The results showed that the highest level of HBx protein was observed 1 h after TNF-α treatment, and then slowly decreased in a time-dependent manner (Additional file 2: Figure S11A). Overexpressed CPAP significantly augmented HBx protein expression but not mRNA expression in response to TNF-α treatment (Fig. 4b). To verify whether the accumulation of HBx protein induced by CPAP is associated with the increased stability of HBx, HCC cells with HA-CPAP and GFP-HBx expression were treated with the protein synthesis inhibitor cycloheximide, and the level of GFP-HBx in the absence or presence of TNF-α was determined at various time points. The results showed that CPAP significantly increased TNF-α-induced HBx stabilization (Additional file 2: Figure S11B).

Our previous study indicated that SUMOylated CPAP is important for TNF-α-mediated NF-κB activation [34], therefore, we investigated whether SUMOylation of CPAP is necessary for HBx protein stability. The results showed that CPAP/WT increased HBx protein expression, whereas the CPAP SUMO-deficient mutant had only a minor effect on maintaining HBx protein expression even in the presence of TNF-α (Fig. 4c, compare lanes 4 and 6). Next, we knocked down NF-κB/p65 and evaluated the effect of CPAP-enhanced TNF-α-induced HBx stabilization. The results showed that the expression of HA-HBx was markedly reduced in NF-κB/p65 knocked down cells even in the presence of TNF-α (Fig. 4d, compare lanes 2 and 6). Moreover, the overexpression of CPAP could not enhance TNF-α-induced HBx stabilization in p65-knocked down cells (Fig. 4d, compare lanes 4 and 8). In addition, after the knockdown of CPAP, HBx expression was decreased regardless of TNF-α treatment (Fig. 4e, compare lanes 2 and 4). These results suggest that CPAP is required for TNF-α-induced HBx stabilization.

Overexpressed CREB/CPAP positively correlates with a poor survival rate in HBx-positive HCC

In order to give insight into the clinical impact of the interaction between HBx and CPAP, we checked the activation status of NF-κB in CPAP-overexpressing HBx-positive HCC. The result indicated that HCC tissues with overexpressed HBx and CPAP has an increased activation of NF-κB (Fig. 5a), whereas HCC tissues without HBx and CPAP overexpression present no enhanced NF-κB activity (Fig. 5b). The correlation between enhanced NF-κB phosphorylation with overexpressed HBx and CPAP is statistically significant (Fig. 5c). The correlation between overexpressed HBx and CPAP with activated NF-κB is positively correlated with DFS (p = 0.0086) in HCC patients (Additional file 2: Figure S12). Overexpressed CREB is positively correlated with an augmented CPAP level in HBx-positive HCC (Fig. 5d, e). Importantly, overexpressed CPAP/CREB showed a lower DFS than instances with a lower expression level of CPAP/CREB (Additional file 2: Figure S13A). We also analyzed the TCGA data set to check the correlation between the overall survival rate (OS) and overexpressed CPAP/CREB using SurvExpress biomarker validation tool [1], and the results indicated that an increased expression of CREB and CPAP mRNAs exists in the high-risk group with a lower survival rate (Additional file 2: Figure S13B).