Secretory clusterin contributes to oxaliplatin resistance by activating Akt pathway in hepatocellular carcinoma

Abstract

Secretory clusterin (sCLU) is expressed in numerous cancers and is associated with the resistance to chemotherapy. However, the role of sCLU in the resistance of hepatocellular carcinoma (HCC) to oxaliplatin (OXA), a recently used third‐generation platinum agent, remains unclear. The stable transfectants that are depleted of or overexpress sCLU and OXA‐resistant cells were generated using human HCC cells. Overexpression of sCLU abrogated OXA‐induced inhibition of cell growth and cell apoptosis, but depletion of sCLU synergized with OXA to inhibit cell growth and enhance cell apoptosis, by regulating proteins involved in mitochondrial apoptosis pathways, such as Bcl‐2, Bax, Bcl‐xL and caspase‐9, and affecting phosphorylation of Akt and GSK‐3β. Overexpression of sCLU in either OXA‐resistant cells or stable transfectants that overexpress sCLU significantly increased phosphorylated Akt. However, specific inhibition of Akt enhanced sensitivity of sCLU‐overexpressing cells to OXA, but had no effect on sCLU expression, suggesting that the regulatory effects between sCLU and pAkt may be in a one‐way manner in HCC cells. The expression levels of sCLU affected the therapeutic efficacy of OXA to treat HCC tumors established in immunodeficiency mice. The results have demonstrated that sCLU contributes to OXA resistance by activating Akt pathway, indicating that sCLU may be a novel molecular target for overcoming OXA resistance in HCC.

Liver cancer is the second most frequent cause of cancer death in men worldwide, and hepatocellular carcinoma (HCC) accounts for 70–85% of all liver cancers.1 However, HCC is extremely resistant to current chemotherapeutic drugs. Sorafenib is currently the most effective drug for HCC,2 but it prolongs survival for only 2–3 months, and has not been widely adopted because it is cost‐prohibitive in Asia and sub‐Saharan Africa, which have the highest incidence of HCC.1 Therefore, novel drugs and therapeutic strategies for HCC continue to be sought.

Oxaliplatin (OXA) is a third‐generation platinum‐derived chemotherapeutic agent that triggers cell death by inducing platinum‐DNA adducts, and appears to block DNA replication more effectively than other platinum compounds.3 OXA has been used to treat colorectal4, 5 and gastric6 cancers, and has also displayed anti‐tumor activity against HCC.7 The combination of OXA and gemcitabine or bevacizumab has been investigated to treat advanced HCC in clinical trials,8, 9 but the results achieved were not comparable with the success obtained with colorectal cancer.5 Furthermore, little data is available regarding the signaling pathways activated by OXA in HCC. Therefore, it remains a priority to find effective methods to enhance the sensitivity of HCC to OXA.

Clusterin (CLU) was initially discovered as a secretory glycoprotein with cell‐aggregating activity in vitro,10 and is overexpressed in numerous advanced‐stage and metastatic human cancers.11 Overexpression of CLU has been observed in a majority of HCC cases,12, 13 and metastatic HCC tissues have been found to have a higher level of CLU expression compared with primary HCC tissues of the same patient.14 It is noteworthy that only the secretory clusterin form (sCLU), but not the nuclear form (nCLU), is expressed in aggressive late stage tumors.15 sCLU has been used as a biomarker for pancreatic,16 breast,17 colorectal,18 ovarian,19 cervical,20 gastric,21 bladder,22 esophageal23 and lung24 cancers, as well as for HCC.25, 26 Most significantly, sCLU expression is associated with broad‐based resistance to chemotherapeutic agents such as doxorubicin, cisplatin, etoposide and camphothecin.18 Downregulation of sCLU by antisense oligonucleotides enhances the activity of heat shock protein 90 inhibitor,27 or docetaxel or mitoxantrone,28 to treat drug‐resistant prostate cancer, and the efficacy of gemcitabine, cisplatin or carboplatin to treat non‐small cell lung cancer (NSCLC).29 These data indicate that sCLU may contribute to the resistance of HCC to OXA. Therefore, we designed the present study aiming to investigate this hypothesis and the underlying molecular pathways.

Materials and Methods

Cell lines and culture

Human HCC cell lines, HepG2, Bel7402 and Bel7404 (the American Type Culture Collection, Rockville, Maryland, USA) and SMMC7721 and SNU739, and cervical cancer Hela cells (the Type Culture Collection Cell Bank, Chinese Academy of Sciences Committee, Shanghai, China) were used in this study. HepG2, SNU739 and Hela cells were cultured in DMEM (Thermo Fisher Scientific, Shanghai, China), while Bel7402, Bel7404 and SMMC7721 cells in RPMI 1640 medium (Thermo Fisher Scientific), supplemented with 100 IU/mL penicillin, 100 mg/mL streptomycin and 10% FBS.

Antibodies and reagents

Antibodies used in this study included antibodies against clusterin, caspase‐9 (Santa Cruz Biotechnology, Santa Cruz, California, USA), and antibodies against Bcl‐2, Bax, Bcl‐xL, Akt, phosphorylated Akt (pAkt), glycogen synthase kinase (GSK)‐3β, phosphorylated GSK (pGSK)‐3β and GAPDH (Cell Signaling Technology, Danvers, Massachusetts, USA). The Akt inhibitor V, triciribine (TCN), was purchased from Merk KGaA (Merk, Darmstadt, Germany), and OXA from Sanofi China (Shanghai, China).

Expression vectors and transfection

Four pMAGic7.1‐based shRNA vectors (CLU1, CLU2, CLU3 and CLU4), a scrambled (Sc) shRNA vector (Table 1) and a full‐length human CLU cDNA (GenBank accession NM_203339) were provided by Shanghai Sunbio Medical Biotechnology (Shanghai, China). The sCLU fragment was amplified by PCR with a pair of primers (5′‐GGACTCAGATCTCGAGGCCACCATGATGAAGACTCTGCTGCTG‐3′ and 5′‐GGCGACCGGTGGATCCCGCTCCTCCCGGTGCTTTTT‐3′), and inserted into the pIRES2‐EGFP vector (TaKaRa Biomedical Technology, Beijing, China), which was termed pIRES2‐EGFP/sCLU. Transfection of cells was performed using GenJet DNA in vitro transfection reagent (SignaGen, Rockville, Maryland, USA).

Table 1.

Sequences of oligonucleotides used as CLU RNAi primers

Construct	Position	Sequence	Target sequence
CLU1	Upstream	5′‐CCGGGCTCCAGGAAATGTCCAATttcaagaga ATTGGACATTTCCTGGAGC TTTTTTg‐3′	GCTCCAGGAAATGTCCAAT
	Downstream	5′‐AATTCAAAAAAGCTCCAGGAAATGTCCAAT tctcttgaa ATTGGACATTTCCTGGAGC‐3′
CLU2	Upstream	5′‐CCGGGGTTGACCAGGAAATACAAttcaagaga TTGTATTTCCTGGTCAACC TTTTTTg‐3′	GGTTGACCAGGAAATACAA
	Downstream	5′‐AATTCAAAAAAGGTTGACCAGGAAATACAA tctcttgaaTTGTATTTCCTGGTCAACC‐3′
CLU3	Upstream	5′‐CCGGGGGATATGATGACAAGGTTctcaagaga AACCTTGTCATCATATCCC TTTTTTg‐3′	GGGATATGATGACAAGGTT
	Downstream	5′‐ AATTCAAAAAAGGGATATGATGACAAGGTT tctcttgag AACCTTGTCATCATATCCC‐3′
CLU4	Upstream	5′‐CCGGCAGGGAAGTAAGTACGTCAATctcgag ATTGACGTACTTACTTCCCTG TTTTTTg‐3′	CAGGGAAGTAAGTACGTCAAT
	Downstream	5′‐AATTCAAAAAACAGGGAAGTAAGTACGTCAAT ctcgag ATTGACGTACTTACTTCCCTG‐3′
Scramble	Upstream	5′‐CCGGTTCTCCGAACGTGTCACGTttcaagaga TTCTCCGAACGTGTCACGT TTTTTTg‐3′	TTCTCCGAACGTGTCACGT
	Downstream	5′‐AATTCAAAAAATTCTCCGAACGTGTCACGT tctcttgaaTTCTCCGAACGTGTCACGT‐3′

CLU, clusterin.

Establishment of stable transfectants

Cells were seeded in 10‐cm plastic dishes and grown to approximately 67% confluence when Bel7404 cells were transfected with 4 μg of pIRES2‐EGFP/sCLU or pIRES2‐EGFP vectors, and Bel7402 and SMMC7721 cells with 4 μg of shRNA vectors. Cells were detached with trypsin after transfection for 48 h, and seeded in selection medium containing 400 μg/mL of geneticin (for pIRES2‐EGFP vectors) or 3 μg/mL of puromycin (for shRNA vectors). Stable transfectants were selected 4 weeks later. Bel7404 cells stably transfected with pIRES2‐EGFP/sCLU or pIRES2‐EGFP vector were termed Bel7404‐sCLU^high or Bel7404‐vec. Bel7402 or SMMC7721 cells stably transfected with CLU4 or Sc shRNA vectors were termed Bel7402‐sCLU^low, Bel7402‐vec, SMMC7721‐sCLU^low or SMMC7721‐vec, respectively.

Quantitative real‐time PCR analysis

Total RNA was extracted from cells, and cDNA was synthesized using a cDNA synthesis kit (Invitrogen, Shanghai, China) with a pair of primers (5′‐ATGATGAAGACTCTGCTGCTGTT‐3′ and 5′‐TTCCTGGAGCTCATTGTCTG‐3′). The reaction mixtures for quantitative real‐time PCR were prepared (Shanghai Sunbio Medical Biotechnology, Shanghai, China) and analyzed using MX3000P Real‐time PCR systems (Stratagen, Wilmington, DE, USA). Experiments were performed in triplicate, and the data were calculated using ∆∆C _t methods.

Cell viability and apoptosis assays

The methods have been described previously.30, 31, 32

Immunoblotting

Protein concentrations in cell lysates were determined (Bio‐Rad, Richmond, California, USA). Lysates were resolved on SDS‐PAGE gels, transferred to PVDF membranes, and immunoblotted as previously described.30, 31, 32 The density of each band was measured using a densitometric analysis program (FR200, Shanghai, China). The relative level of each protein was normalized with respect to GAPDH band density. In preliminary experiments, serial dilutions of lysates (containing 2.5, 5, 10, 20, 40 or 80 μg protein) were immunoblotted, band intensities were measured and plotted against protein amounts to generate a standard curve, and the amount of protein for each immunoblot was determined.

Establishment of oxaliplatin‐resistant cells

The OXA‐resistant cells, termed Bel7404‐OR, were established by incubating Bel7404 cells with OXA with gradient increment of concentrations from 1 to 60 μM. Briefly, cells were cultured in medium containing 1 μM of OXA for 1 week, and then 2 μM of OXA for 1 week, followed by 4 μM of OXA for 1 week until 60 μM was reached. The cells that were able to survive in the medium containing 60 μM of OXA were considered OXA‐resistant cells.

Animal experiments

All surgical procedures and care administered to the animals were in accordance with institutional guidelines. Six‐week‐old male athymic BALB/c nu/nu mice were purchased from the Shanghai Institute of Materia Medica, Chinese Academy of Science, China. The experimental protocol has been described previously.30, 31, 32 Briefly, 1 × 10⁶ of cells were subcutaneously injected into the back of the mice. Tumor volumes were estimated: π/6 × a ² × b, where a is the short axis and b the long axis. Two weeks later, the mice were assigned to respective groups receiving weekly intraperitoneal injection of 200 μL of either PBS or OXA at a dose of 10 mg/kg. Mice were killed on day 42.

Statistical analysis

The data are expressed as mean values ± SD. Comparisons were made using a one‐way anova followed by Dunnett's test with spss software (version 17.0, SPSS China, Shanghai, China). P < 0.05 was considered statistically significant.

Results

Expression of secretory clusterin in hepatocellular carcinoma cells

The expression of pre‐sCLU and sCLU was examined in human HCC cells. Hela cells served as a positive control.33 HCC cells expressed different levels of sCLU in the following order (from high to low): Bel7402, SMMC7721, HepG2, SNU739 and Bel7404 (Fig. 1A,B).

Figure 1.

Expression of secretory clusterin (sCLU) in hepatocellular carcinoma (HCC) cells. Lysates (A, B) from Hela, Bel7402, SMMC7721, HepG2, SNU739 and Bel7404 cells, or (C, D) from parental Bel7404 cells (control), or Bel7404 cells stably transfected with pIRES2‐EGFP vector (Bel7404‐vec) or pIRES2‐EGFP/sCLU (Bel7404‐sCLU ^high), or (E, F) from parental Bel7402 cells (control), or Bel7402 cells transfected with scrambled (Sc) or clusterin (CLU) 1, CLU2, CLU3 and CLU4 shRNA vectors were immunoblotted. (B, D, F) The density of each band in (A, C, E) was measured and normalized to that of GAPDH, respectively. (G) The expression of sCLU mRNA in the above cells in (E) was analyzed by quantitative real‐time PCR, and normalized to that of β‐actin. A significant (P < 0.05) difference from the respective control is denoted by “*,” and a highly significant (P < 0.001) difference by “**.”

Overexpression of secretory clusterin in hepatocellular carcinoma cells

The stable transfectants, Bel7404‐sCLU^high, expressed distinctly higher levels of sCLU than parental Bel7404 cells, while the levels of sCLU in control stable transfectants, Bel7404‐vec, remained unchanged (Fig. 1C,D).

Silencing secretory clusterin in hepatocellular carcinoma cells

Transfection of CLU1, CLU2, CLU3 and CLU4 resulted in reductions of sCLU protein expression in Bel7402 cells by 14.6%, 47.4%, 22.0% and 73.3%, respectively, compared with Sc controls (Fig. 1E,F). Similarly, quantitative real‐time PCR analysis showed that CLU1, CLU2, CLU3 or CLU4‐transfected Bel7402 cells had notable reductions in sCLU mRNA expression, by 42.1%, 51.7%, 46.7% and 55.4%, respectively, compared with Sc controls (Fig. 1G). Among the four shRNA vectors, CLU4 displayed the strongest gene‐silencing ability, and was used in the following experiments. The gene‐silencing efficiency of the shRNA vectors were also evaluated in SMMC7721 and HepG2 cells (data not shown).

Secretory clusterin affects chemosensitivity to oxaliplatin

Bel7402‐sCLU^low, SMMC7721‐sCLU^low and Bel7404‐sCLU^high cells had similar proliferation rates, compared with the respective parental cells, or Sc shRNA or empty vector‐transfected controls (Fig. 2A–C). The above cells were incubated in the absence or the presence of OXA (16 μM) for 48 h, and cell viability was measured. OXA alone significantly reduced the viability of Bel‐7402, SMMC7721 and Bel7404 cells (Fig. 2D–F). Depletion of sCLU significantly reduced, while overexpression of sCLU significantly increased, the viability of cells treated with OXA (Fig. 2D–F). Bel7402‐sCLU^low and SMMC7721‐sCLU^low cells had a significant reduction in IC₅₀ of OXA by 68.9% and 67.9%, compared to their respective parental cells (Table 2). However, Bel7404‐sCLU^high cells had a significantly higher IC₅₀ of OXA, by approximately threefold, compared with parental Bel7404 cells (Table 2).

Figure 2.

Secretory clusterin (sCLU) affects chemosensitivity to oxaliplatin (OXA). (A–C) Parental Bel7402, SMMC7721 or Bel7404 cells, or the cells stably transfected with scrambled (Sc) (Bel7402‐vec or SMMC7721‐vec) or clusterin (CLU) 4 shRNA vector (Bel7402‐sCLU ^low or SMMC7721‐sCLU ^low), or empty vector or sCLU expression vector (Bel7404‐vec or Bel7404‐sCLU ^high) were cultured for 7 days, and the viability was measured at indicated time points. (D–F) The above cells cultured in the presence or absence of OXA (16 μM) for 48 h, and the viability was measured. A significant (P < 0.05) difference from respective untreated parental cells is denoted by “*.” A significant (P < 0.05) reduction from OXA‐treated parental cells is denoted by “‡” and a significant (P < 0.05) increase by “†.”

Table 2.

Effect of sCLU on IC ₅₀ to oxaliplatin in HCC cells

Cells	Relative sCLU	IC50 (μM)	95% confidence interval (μM)
SMMC7721	0.77	40.159	0.211–3.156
SMMC7721‐sCLUlow	0.17	12.891	0.519–2.934
Bel7402	1.08	54.349	0.028–3.677
Bel7402‐sCLUlow	0.25	16.888	0.398–3.016
Bel7404	0.27	22.043	0.325–3.019
Bel7404‐sCLUhigh	1.56	69.053	0.327–3.118

Cells were incubated with various concentrations of oxaliplatin for 48 h, and the cell viability was determined using a CCK‐8 kit. HCC, hepatocellular carcinoma; sCLU, secretory clusterin.

Oxaliplatin and depletion of sCLU induced apoptosis of Bel7402 cells at rates of 9.2% and 4.7%, respectively, which were significantly higher than that of untreated Bel7402 cells (0.6%); and Bel7402‐sCLU^low cells treated with OXA had an even higher apoptosis rate (20.0%; Fig. 3A,B). Similar results were obtained with SMMC7221 cells. The untreated Bel7404‐sCLU^high cells had almost the same apoptosis rate as parental Bel7404 cells. OXA treatment increased apoptosis of parental Bel7404 cells, with a rate of 31.5%, which was significantly higher than that of Bel7404‐sCLU^high cells treated with OXA (10.5%; Fig. 3A,B).

Figure 3.

Cell apoptosis. (A) Representative dot plots were from cytometrically‐analyzed parental Bel7402, SMMC7721 or Bel7404 cells (control), and their stable transfectants depleted of secretory clusterin (sCLU) by clusterin (CLU) 4, or overexpressing sCLU (sCLU ^high), which were incubated in the absence or the presence of OXA (16 μM) for 24 h, and cell apoptosis rates (%) were measured (B). A significant (P < 0.05) difference from respective control is denoted by “*,” and a highly significant (P < 0.001) difference by “**.” A significant (P < 0.05) increase from respective OXA‐treated parental cells is denoted by “†,” and a significant (P < 0.05) reduction by “‡.”

Secretory clusterin affects cell mitochondrial apoptosis pathways

Bel7402‐sCLU^low and SMMC7721‐sCLU^low cells were cultured in the absence or the presence of OXA (16 μM) for 24 h. Both OXA and sCLU knockdown upregulated Bax and downregulated Bcl‐2, thus reducing the Bcl‐2/Bax ratio; and the combination of OXA and sCLU knockdown further decreased the Bcl‐2/Bax ratio (Fig. 4A,B). Overexpression of sCLU alone had almost no effect on the expression of Bcl‐2 and Bax, but OXA significantly reduced the Bcl‐2/Bax ratio; the combination of OXA and sCLU overexpression significantly increased the Bcl‐2/Bax ratio, compared with OXA alone (Fig. 4A,B). In addition, OXA alone significantly downregulated Bcl‐xL expression and increased caspase‐9 activation, and the combination of sCLU depletion and OXA further downregulated Bcl‐xL expression and increased caspase‐9 activation in Bel7402 cells (Fig. 4C,D).

Figure 4.

Mitochondrial apoptosis pathway. (A) Parental Bel7402, SMMC7721 or Bel7404 cells, and their stable transfectants depleted of secretory clusterin (sCLU) by clusterin (CLU) 4, or overexpressing sCLU, were incubated in the absence or presence of oxaliplatin (OXA) (16 μM) for 24 h, and immunoblotted. (B) Each band density in (A) was measured and normalized to that of GAPDH, respectively. (C) Parental Bel7402 cells (control) and their stable transfectants transfected with scrambled (Sc) or CLU4 shRNA vectors were incubated in the presence or absence of OXA (16 μM) for 24 h, and immunoblotted. (D) Each band density in (C) was measured and normalized to that of GAPDH. “*” indicates a significant (P < 0.05) difference from the control, and “**,” a highly significant (P < 0.001) difference. A significant (P < 0.05) increase from OXA‐treated parental cells is denoted by “†,” and a significant (P < 0.05) reduction by “‡.”

Secretory clusterin and phosphorylation of Akt

To further investigate whether sCLU participates in OXA resistance, we established OXA‐resistant Bel7404 cell lines (Bel7404‐OR). Both Bel7404‐OR and Bel7404‐sCLU^high cells had higher levels of sCLU than their parental cells (Fig. 5A,B). pAkt has been considered to be a cancer multidrug resistance locus34; thus, its expression was detected. Both Bel7404‐OR and Bel7404‐sCLU^high cells had higher levels of pAkt, while they had similar levels of Akt, compared with their parentals (Fig. 5A,B). The data drove us to detect expression of pAkt in the HCC cells in Figure 1(A), but a correlation between endogenous expression of sCLU and Akt was not observed (data not shown). Furthermore, specific inhibition of pAkt by TCN downregulated pAkt but had no effect on sCLU expression in Bel7404‐OR cells (Fig. 5C,D). In addition, TCN was highly efficient in suppressing Akt phosphorylation in a dose‐dependent manner, but did not affect sCLU expression in Bel7402 cells (Fig. 5E,F). TCN was originally dissolved in DMSO at a concentration of 10 mg/mL; thus, medium containing DMSO also served as a control.

Figure 5.

Akt phosphorylation. (A, B) Parental Bel7404, Bel7404‐OR or Bel7404‐sCLU ^high cells were cultured for 24 h. (C, D) Bel7404‐OR cells were cultured in medium, or media containing DMSO or 3 μM of triciribine (TCN), for 4 h. (E, F) Bel7402 cells were cultured in medium, or media containing DMSO or 1 and 3 μM of TCN for 4 h. (G, H) Parental Bel7402 cells or their stable transfectants depleted of secretory clusterin (sCLU) by clusterin (CLU) 4 were cultured in medium in the absence or presence of 16 μM of oxaliplatin (OXA) for 24 h. The above cells were immunoblotted (A, C, E, G), and each band density was measured and normalized to that of GAPDH (B, D, F, H), respectively. “*” indicates a significant (P < 0.05) difference.

However, both depletion of sCLU and OXA treatment reduced Akt phosphorylation, and the level of pAkt was further reduced in Bel7402‐sCLU^low cells treated with OXA (Fig. 5G,H). In addition, both depletion of sCLU and OXA reduced phosphorylation of GSK‐3β, a downstream factor of Akt (Fig. 5G,H).

Secretory clusterin inhibits cell apoptosis induced by Akt inhibition

Akt inhibitors have been shown to induce cell apoptosis.35 Here, we further confirmed that TCN highly significantly reduced the level of pAkt (Fig. 6A,B), and increased the apoptosis rates of both parental Bel7404 and Bel7404‐sCLU^high cells (Fig. 6C). Overexpression of sCLU itself had almost no effect on apoptosis rates but increased Akt phosphorylation and suppressed TCN‐induced inhibition of Akt phosphorylation (Fig. 6A,B), and inhibited the increased apoptosis of Bel7404 cells induced by TCN (Fig. 6C). The data drove us to investigate whether TCN could increase the sensitivity of sCLU‐overexpressing HCC cells to OXA. As shown in Figure 6(D), OXA had only a non‐significant effect on, but TCN highly significant reduced, the viability of Bel7404‐sCLU^high cells. The combination of TCN and OXA resulted in even lower viability of the cells, compared with TCN alone.

Figure 6.

Akt inhibition counteracts secretory clusterin (sCLU) overexpression on cell apoptosis and overcomes resistance to oxaliplatin (OXA). Parental Bel7404 or Bel7404‐sCLU ^high cells were cultured in medium in the absence or presence of TCN (3 μM) for 4 h. (A) The expression of Akt/pAkt was detected. (B) Each band density in (A) was measured and normalized to that of GAPDH. (C) The apoptosis rates (%) were measured. (D) Bel7404‐sCLU ^high cells were incubated with OXA (16 μM), triciribine (TCN) (3 μM), or the combination for 8 h, and cell viability was measured. Untreated parental Bel7404 cells served as controls. “**” indicates a highly significant (P < 0.001) difference. A significant (P < 0.05) increase from untreated parental Bel7404 cells is denoted by “†,” and a significant (P < 0.05) reduction by “‡.”

Secretory clusterin affects the efficacy of oxaliplatin to inhibit tumors

To investigate the influence of sCLU on the efficacy of OXA to inhibit tumor growth in vivo, we designed two sets of experiments using parental Bel7402 and Bel7402‐sCLU^low cells, and parental Bel7404 and Bel7404‐sCLU^high cells. As shown in Figure 7(A), OXA had an obvious antitumor effect by reducing tumor growth by 45.1% compared to PBS in parental Bel7402 tumors on day 42. There was no significant difference in parental Bel7402 and Bel7402‐sCLU^low tumors treated with PBS, but Bel7402‐sCLU^low tumors treated with OXA grew significantly slower than OXA‐treated parental Bel7402 tumors (Fig. 7A). In contrast, OXA demonstrated even stronger anti‐tumor activity in parental Bel7404 tumors by reducing tumor growth by 61.6% compared with that in PBS‐treated parental Bel7404 tumors (Fig. 7B). There was no significant difference in Bel7404 tumors and Bel7404‐sCLU^high tumors treated with PBS, but Bel7404‐sCLU^high tumors treated with OXA grew significantly faster than OXA‐treated parental Bel7404 tumors (Fig. 7B). These results indicate that the expression levels of sCLU may determine the efficacy of OXA in treating HCC tumors.

Figure 7.

The status of secretory clusterin (sCLU) affects sensitivity to oxaliplatin (OXA) in vivo. (A) Parental Bel7402 or Bel7402‐sCLU ^low cells, or (B) parental Bel7404 or Bel7404‐sCLU ^high cells, were implanted into groups of mice. Two weeks later, the mice received weekly injection of either OXA or PBS. “n” indicates the number of mice in each group. A significant (P < 0.05) difference from parental tumors treated with PBS is denoted by “*,” and a highly significant (P < 0.001) difference by “**.” A significant (P < 0.05) reduction from parental tumors treated with OXA is denoted by “‡,” and a significant (P < 0.05) increase by “†.”

Discussion

Secretory clusterin has exhibited inhibitory activities against many therapeutic approaches, such as ionizing radiation,36 and cytotoxic chemotherapy.37 However, it remains unknown whether sCLU participates in the resistance of HCC to OXA. The present study has for the first time demonstrated that sCLU contributes to OXA resistance by activating Akt pathways in HCC. OXA‐resistant cells displayed increased sCLU expression accompanied by increased phosphorylation of Akt. Overexpression of sCLU abrogated the inhibition of cell growth and the induction of cell apoptosis by OXA, and depletion of sCLU synergized with OXA to inhibit cell growth and enhance cell apoptosis. The status of sCLU influenced the efficacy of OXA in treating HCC tumors in vivo. sCLU regulated expression of proteins in mitochondrial apoptosis pathways, such as Bcl‐2, Bax, Bcl‐xL and caspase‐9, and phosphorylation of Akt and GSK‐3β. In accordance, it has been reported that CLU inhibits apoptosis by interfering Bax.38 Although not investigated herein, sCLU may also regulate death receptor‐mediated apoptosis pathways, as sCLU stimulates secretion of TNF‐α through activation of PI3K/Akt pathway.39 The results indicate that sCLU contributes to OXA resistance by activating Akt pathway, and sCLU may be a novel molecular target for overcoming OXA resistance in HCC.

The present study has identified a previously unrecognized pathway linking sCLU overexpression to Akt phosphorylation induced by OXA‐resistance, and the subsequent molecular signaling pathway of cell apoptosis. Interestingly, a correlation between endogenous expression of sCLU and pAkt was not observed in untreated HCC cells, and the regulation between sCLU and pAkt is in a one‐way manner. Thus, overexpression of sCLU by OXA treatment or exogenous gene transfection increased, while depletion of sCLU by shRNA inhibited phosphorylation of Akt, but specific inhibition of Akt did not affect the expression of sCLU in HCC cells. The effect of sCLU on pAkt pathway is also supported by its regulatory effects on pGSK‐3β and Bcl‐2 family members, the downstream factors of pAkt. It has been reported that high activation of the PI3K/Akt pathway is implicated in HCC carcinogenesis,40 and targeting pAkt converts the positive feedback loops into negative ones, leading to the disappearance of multidrug resistance.34 In accordance, sCLU activates PI3K/Akt pathways to induce matrix metalloproteinase‐9 expression,41 and inhibition of GSK‐3 promotes apoptosis of human HCC cells.42

As a third‐generation platinum‐derived compound, OXA has displayed stronger anti‐cancer activities than other platinum drugs. Its main mechanism is to trigger cell death by forming adducts with DNA and block DNA replication.3 The present study has also demonstrated that OXA induced apoptosis of HCC cells by upregulating expression of Bax, downregulating Bcl‐2 and Bcl‐xL, and leading activation of caspase‐9, in accordance with a report by El Fajoui et al.43

Downregulation of sCLU by siRNA or antisense oligonucleotides resensitizes resistant prostate cancer cells to docetaxel,44, 45 and overexpression of sCLU in prostate cancer cells renders them resistant to paclitaxel.46, 47 Custirsen (OGX‐011), an antisense oligonucleotide drug, has been applied to treat prostate cancer and NSCLC in clinical trials.27, 28, 29 The present results warrant future investigation by using this drug, particularly in combination with OXA, to treat HCC.

Disclosure Statement

The authors have no conflict of interest to declare.

(Cancer Sci, doi: 10.1111/cas.12088, 2013)