Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
Potential conflict of interest: Nothing to report.
Death evasion is crucial for both carcinogenesis and resistance to anticancer therapies. Recently, we identified nucleophosmin (NPM) as a key factor counteracting death stimuli in human hepatocellular carcinoma (HCC) cells. Here we report the identification of a novel NPM-BCL2-associated X protein (BAX) pathway orchestrating death evasion in human HCC cells. Silencing of NPM expression significantly sensitized HCC cells—particularly those bearing inactivated p53 gene (Huh7, Hep3B, and Mahlavu)—to ultraviolet irradiation, mitomycin C, doxorubicin, cisplatin, sorafenib, and lapatinib. This sensitizing effect was not changed further, as p53 expression had been simultaneously silenced. Following cell stress, NPM and BAX were induced and exported out of the nucleoli and nucleus, respectively. BAX was translocated to cytoplasm in cells with relatively high NPM level, or accumulated in the mitochondria in cells with relatively low NPM level and undergoing apoptosis. Subcellular fractionation revealed that silencing of NPM expression greatly enhanced mitochondrial translocation and oligomerization of BAX in Huh7 and Mahlavu cells. In situ proximity ligation assays and reciprocal co-immunoprecipitation revealed a direct interaction between NPM and BAX in the cytoplasm. Silencing of BAX expression abolished the sensitization effect exerted by silencing of NPM in HCC cells. Clinically, up-regulation of NPM was significantly associated with advanced tumor stage and poor prognosis. Conclusion: By directly blockading BAX mitochondrial translocation and activation, NPM helps human HCC cells evade death induction independently of p53-mediated cell death. Silencing of NPM significantly sensitized HCC cells to anticancer therapies. NPM is a potential cotarget in combination with other therapies for HCC, particularly those that harbor inactivated p53 gene. Our findings are of clinical significance because NPM up-regulation and p53 mutations are usually found in advanced human cancers, including HCC. (HEPATOLOGY 2013)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Evasion of death is a hallmark of cancer cells; it is essential for carcinogenesis and is related to resistance to anticancer therapies.1, 2 Defects or disruptions of death regulatory pathways in cancer cells contribute to resistance to anticancer therapies.3 Therefore, understanding how cancer cells evade death stimuli is critical for improvement of cancer therapies, particularly advanced hepatocellular carcinoma (HCC), for which the efficacy of all current chemotherapies and molecular target therapies remains undesirable.4-6
Recently, via induction of cellular hormetic response to ultraviolet (UV) irradiation, we identified genes/proteins that are involved in counteracting death induction in human HCC cells.7 Of them, nucleophosmin (NPM) plays a pivotal role in protecting cells from death in response to cell stress.7 These findings led us to further investigate the underlying mechanisms and its potential roles in anti-HCC therapies.
NPM is a highly conserved phosphoprotein that is located primarily in the nucleoli and shuttles between the nucleoli and cytoplasm during the cell cycle.8 Its function has been implicated in the regulation of ribosome biogenesis, centrosome duplication, and genome stability.9-11 NPM is overexpressed in many solid tumors,12, 13 while its gene, NPM1, is usually translocated, mutated, or deleted in various forms of leukemia.13, 14 The NPM1 gene is a transcription target of the proto-oncogene MYC.10, 15 High expression of NPM is seen in highly proliferating cells and cancer cells. Overexpression of NPM decreases the sensitivity of human leukemia cells to retinoic acid–induced differentiation. These findings strongly suggest that NPM1 is a proto-oncogene.14, 16 Moreover, overexpressed NPM also functions as an antiapoptosis protein.17, 18 Several mechanisms have been proposed and are always related to p53-mediated apoptosis.19 Because inactivated mutations of p53 are seen in more than half of human solid cancers and in most advanced cancers, including HCC, it is intriguing that NPM has a role in the death regulation of cancer cells harboring inactivated p53.
Bcl2-associated X protein (BAX) is a key effector of mitochondria-mediated apoptosis. Upon significant DNA damage, BAX along with p53 is induced and targets to the mitochondrial inner membrane, where BAX is oligomerized and forms pores, with the consequence of losing the membrane potential, releasing cytochrome C into cystoplasm, and then activating cascades for apoptosis progression. Recently, NPM was found as a novel BAX binding protein with this interaction proposed to be involved in activation and translocation of BAX in mitochondrial dysfunction and apoptotic cell death.20 However, neither has this anti-apoptosis proposal for NPM been proved, nor has the role of p53 in this hypothetic NPM-mediated death evasion mechanism been examined.
In this study, we demonstrated that in response to cell stress, a set of NPM translocates from nucleolus to cytosol, binds to BAX, and blocks mitochondrial translocation, oligomerization, and activation of BAX, thereby rendering cells resistant to death induction. This novel NPM-BAX death evasion pathway is independent of p53 function. Silencing of NPM sensitizes HCC cells, particularly those with inactivated p53, to chemotherapy and targeted therapies. Our findings not only shed light on the molecular mechanisms of how cancer cells evade death stimuli, but also open an avenue for development of new anti-HCC therapies.
Human hepatoma cell lines, HepG2 (wild-type p53), and Hep3B (null-genotype p53) were purchased from American Type Culture Collection (Manassas, VA). Huh7 cells and Mahlavu were obtained from the Japanese Collection of Research Biosources and Sanofi-Synthelabo Recherche (Chilly-Mazarin, France), respectively.21 Predesigned small interfering RNAs (siRNAs) targeting against NPM and p53, and siRNAs with scrambled sequences were purchased from Ambion (Austin, TX). Transfection was performed using a commercial transfection kit (RNAiMax, Life Technologies, Invitrogen) as described.7
UV Irradiation, Drug Treatments, Cell Survival/Viability Assays, and Apoptosis Assays.
In total, 1 × 104 HCC cells were seeded onto each well of a 96-well plate 48 hours after transfection with the indicated siRNAs. Twenty-four hours later, cells were treated with the indicated dose of UV-B (290-320 nm) or specified agents. Mitomycin C (Kyowa Hakko Kogyo Co., Ltd.), cisplatin (Bristol-Myers Squibb), and doxorubicin (Pfizer Italia) were prepared in a solvent containing 3 mM NaCl/1 mM NaH2PO4 (pH 6.4) and then used to treat cells at the indicated doses. Sorafenib (kindly provided by Bayer HealthCare, Germany) and lapatinib (purchased from GlaxoSmithKline plc) were prepared in dimethyl sulfoxide. Solvent alone was added to the untreated cells as the control in each experiment. Cell viability was assayed at 24 hours and 48 hours after treatment; for UV treatment, the assays were performed 24 hours after exposure to 30, 65, or 100 mJ/cm2 of UV-B. Cell viability/survival was determined using WST1 assays (Roche Applied Science, Mannheim, Germany), while cell apoptosis was measured by detection of caspase-3 activation (Caspase 3/7 Assay, Promega) as described.7 The experiments were conducted at least twice in triplicate, and the mean of each dose was used to calculate the half maximal inhibitory concentration.
Antibodies, Immunoblotting, and Immunohistochemistry, Immunofluorescence, and Confocal Microscopy.
Details regarding antibodies, immunoblotting, and immunohistochemistry, immunofluorescence, and confocal microscopy are provided in the Supporting Information.22
In Situ Proximity Ligation Assays.
Anti-NPM mouse monoclonal antibody and anti-BAX rabbit polyclonal antibody were used as primary antibodies and anti-mouse and anti-rabbit antibodies coupled with short complementing DNA strands were used as secondary antibodies. Ligation of the DNA strands to a circularized oligomer in case of direct contact between NPM and BAX and the subsequent rolling circle amplification incorporating labeled nucleotides was performed using the Duolink II kit (Olink Bioscience, Uppsala, Sweden) according to the manufacturer's instructions. After being washed and counterstained with 4′,6-diamidino-2-phenylindole, the slides were mounted and inspected under a fluorescence microscope.
Subcellular Fractionation and Co-immunoprecipitation.
Details regarding subcellular fractionation and co-immunoprecipitation are provided in the Supporting Information.
Patients, Tissue Samples, Tissue Arrays, and Immunohistochemistry.
The tumors and their paratumor liver tissues and related clinical data (all anonymous) from 110 patients with HCC were requested from the Taiwan Liver Cancer Network. Normal liver tissue was obtained from a patient with focal nodule hyperplasia who had undergone tumor resection. The Internal Review Board for Medical Ethics of Chang Gung Memorial Hospital approved the specimen collection procedures, and informed consent was obtained from each subject or subject's family. All HCC tissues were reviewed, and the most representative areas of embedded tissue samples were carefully selected and sampled for the tissue microarray blocks. Two core samples were selected from different areas of each HCC tissue. The immunohistochemistry (IHC) scores are defined as follows: 0, negative; 1, weakly positive or in <20% of HCC cells; 2, moderately positive or in 20% to 60% of HCC cells; 3, strongly positive or in >60% HCC cells. The IHC scores were determined by two independent observers (S. J. L. and T. C. C.), where there was disagreement, the slides were re-examined and a consensus was reached by the observers.
A Fisher's exact test was used for comparison between variables. Kaplan-Meier analysis and a log-rank test were used to illustrate differences between each potential risk factor in time-to-recurrence and overall survival probability after patients underwent a primary hepatectomy. Univariate and multivariate Cox regression were used to identify risk factors for prediction of recurrence and death in patients with HCC after initial curative hepatectomy. In the univariate analysis, the effect of each variable on the disease recurrence and survival was tested using analysis of variance. If there was a known predominant determinant (e.g., tumor stage versus vascular invasion), bivariate analysis was conducted to detect more detailed correlations independent from the predominant one that might have been unnoticed. Only variables with a P value less than 0.05 were selected for subsequent multivariate analysis. In multivariate analysis, the selected variables were subjected to a multiple linear model. Hazard ratios (HR) with 95% confidence intervals (CI) were thus obtained. Coefficients were determined via the linear discriminating function of the variables.
Sensitization to Anticancer Cytotoxic Agents by Silencing of NPM.
To examine the potential role of NPM in resistance to anticancer therapies of HCC, HCC cell lines with different p53 genetic background including HepG2 (wild-type p53), Huh7 (C200Y mutated p53), Mahlavu (R249S mutated p53), and Hep3B (null-genotyped p53),23, 24 were treated with UV-B, mitomycin C, doxorubicin, or cisplatin. Silencing of NPM expression significantly enhanced cellular susceptibility to all kinds of treatments in Huh7, Hep3B, and Mahlavu cells, while the sensitizing effect was minimal in HepG2 cells (Fig. 1A). These findings suggest that an NPM-mediated antideath mechanism is independent of p53 functions in HCC cells, which is different from that found in hematopoietic cells.25
To further inspect the role of p53, we silenced the expression of NPM, p53, or simultaneously NPM and p53 by RNA interference (Fig. 1B). Silencing of p53 expression alone did not significantly change the sensitivity to any of the treatments in Huh7, Hep3B, and Mahlavu cells (Fig. 1B, siTP53 versus siNS). Simultaneous silencing of p53 and NPM did not further alter the sensitizing effect exerted by silencing of NPM alone (Fig. 1B, siNPM versus [siNPM + siTP53]). Silencing of NPM and p53 expression by RNA interference was confirmed via immunoblotting (Fig. 1C).
Interestingly, a negative dominant mutant of p53 converted HepG2 cells from insensitivity to sensitivity to cytotoxic and molecular targeted therapies as NPM silenced (Supporting Fig. 1). NPM apparently executes its death evasion activity independently of p53 function.
Silencing of NPM Sensitizes HCC Cells to Lapatinib and Sorafenib.
We further examined the potential role of NPM in resistance to the inhibitors of oncogenic kinases in HCC, such as lapatinib and sorafenib. Silencing of NPM expression significantly sensitized Huh7, Hep3B, and Mahlavu cells to sorafenib and lapatinib (Fig. 2A). Again, simultaneous silencing of p53 and NPM expression in HCC cells did not further change the sensitizing effects exerted by silencing of NPM expression (Fig. 2B). The sensitizing effect was further evidenced by an increase of cell apoptosis in response to treatment with sorafenib and lapatinib (Fig. 2C). We thus conclude that targeting NPM sensitizes HCC cells to oncogenic kinase inhibitors, such as sorafenib and lapatinib. NPM exerts its death evasion activity via a mechanism independent of p53 function.
Induction of NPM and BAX Expression by Cellular Stresses.
NPM was up-regulated in Huh7, Hep3B, and Mahlavu cells following treatment with UV, cisplatin, and doxorubicin (Fig. 3A). Meanwhile, we also observed the induction of BAX expression, a key effector initiating mitochondria-mediated cell death, in all three cell lines (Fig. 3A). Simultaneous induction of NPM (antiapoptosis) and BAX (proapoptosis) represents hormetic mechanisms regulating cell survival versus death in response to stress.7, 26
Nucleoplasmic and Cytoplasmic Translocation of NPM Following Death Stimulation.
To understand how NPM helps HCC cells evade death, we first inspected subcellular distribution of NPM and BAX in response to cell stress. Before UV irradiation, NPM was mainly located in the nucleoli and partially in the nucleoplasm (Fig. 3B, left), whereas BAX was primarily located in nucleoplasm and some in the cytoplasm (Fig. 3C, upper left). Following UV irradiation, NPM was translocated from nucleoli to nucleoplasm, and a set of NPM was further translocated to the cytoplasm (Fig. 3B, right). On the other hand, BAX was translocated to the cytosol and accumulated in the mitochondria (Fig. 3C), particularly in cells undergoing apoptosis (Fig. 3C, right).
Notably, as being transfected with siRNA to down-regulate the expression of NPM, cells with relatively low NPM expression usually presented with aggregation of BAX in the mitochondria and undergoing apoptosis (Fig. 4A), whereas cells with relatively high NPM level presented with less degree of mitochondrial accumulation of BAX and more resistance to apoptosis induction (Fig. 4A). Interestingly, colocalization of NPM and BAX in the cytoplasm was noted in some cells presenting with cytoplasmic NPM (Fig. 4B). These findings suggest that the antiapoptosis activity of NPM is involved in blockade of mitochondrial translocation of BAX.
Blockade of BAX Mitochondria Translocation and Oligomerization by NPM.
The observation that colocalization of a set of NPM with BAX in cytoplasm in HCC cells with relative resistance to death stimuli intrigued us to examine the role of NPM in mitochondrial translocation of BAX in HCC cells. As shown in Fig. 5A, the NPM level in the cytosol of Mahlavu cells was increased after UV irradiation, while NPM was not detected in the mitochondria either before or after UV irradiation. On the other hand, the amount of BAX was increased in both the cytosol and the mitochondria after UV irradiation. Silencing of NPM expression decreased the cytosolic BAX, but increased the mitochondrial BAX, suggestive of blockade of BAX mitochondrial translocation by NPM in response to UV treatment. Prohibitin and glyceraldehyde 3-phosphate dehydrogenase were used as the markers for mitochondrial and cytosolic components, respectively. Similar results of enhancement of mitochondrial accumulation of BAX by silencing of NPM following treatment with other cytotoxic agents, such as staurosporin, were observed in Huh7 cells (data not shown).
To validate the role of NPM in BAX mitochondrial translocation and activation, we used a nonreducing condition for preparation of cellular protein, we found that silencing of NPM expression greatly increased the dimmers and oligomers of the mitochondrial BAX following UV irradiation (Fig. 5B, lane 2), whereas the dimmers and oligomers were barely detected before UV irradiation (Fig. 5B, lane 1) or without silencing of NPM (Fig. 5B, lane 3). Therefore, following death stimuli, NPM blockaded the mitochondrial translocation and oligomerization of BAX in HCC cells.
Directly Binding of NPM to BAX in Cytoplasm.
To further confirm blockade of BAX mitochondrial translocation via direct binding of NPM to BAX, and to know where it occurs, we used an in situ proximity ligation assay, a highly sensitive and specific method to detect protein-protein interaction and its subcellular localization.27 We found that the positive proximity ligation assay signals were greatly increased only after UV irradiation and were majorly localized in the cytosol of Huh7 cells (Fig. 5C, middle versus lower). Consistent results were also obtained in Mahlavu cells. Therefore, there is a direct binding between NPM and BAX (primarily in the cytosol) in response to death stimulation.
Direct interaction between NPM and BAX was further demonstrated using reciprocal co-immunoprecipitation in Mahlavu and Huh7 cells. Following UV irradiation, BAX and NPM were co-immunoprecipitated with NPM and BAX, respectively (Fig. 5D, lanes 2 and 3, respectively). In addition, BAX was not co-immunoprecipitated with NPM before UV irradiation (Fig. 5D, lane 5).
The Essential Role of BAX in the NPM-Mediated Death Evasion Pathway.
To confirm the essential role of BAX in this NPM-mediated death evasion pathway, the expression of NPM and BAX was silenced respectively or simultaneously by RNA interference in Huh7 and Mahlavu cells (Fig. 6A). Silencing of NPM significantly sensitized Huh7 and Mahlavu cells to both sorafenib and lapatinib (Fig. 6B, siNS versus siNPM), whereas this sensitization effect was lost, as BAX had been simultaneously silenced (Fig. 6B, siNPM versus siNPM + siBAX). We thus conclude that BAX plays an essential role in the NPM-mediated death evasion pathway (Fig. 6C).
Up-regulation of NPM in Human HCC Was Associated with Hepatitis B, Portal Vein Invasion, and Poor Prognosis.
Immunoblotting assays revealed that NPM level was increased in four out of six examined HCC tissues compared with that in the matched paratumor liver tissues and a normal liver control (Fig. 7A). We further examined the expression of NPM in 110 pairs of tumor and paratumor liver samples on tissue arrays using IHC (Fig. 7B). NPM expression was detected in 40% HCC (43/107, Supporting Table 1) and its overexpression (strongly positive for NPM in >60% hepatoma cells, IHC score = 3) was strongly associated with younger age (P < 0.001), chronic hepatitis B (P = 0.0056), advanced tumor stages (P = 0.0001), portal vein invasion (P = 0.0003), tumor recurrence (P = 0.004), and lower overall survival (P = 0.002) (Table 1; Fig. 7C,D). Association of NPM overexpression with higher tumor recurrence and higher mortality was further demonstrated via univariate Cox regression (recurrence: HR 2.35, 95% CI 1.18-4.71, P = 0.0156; death: HR 2.69, 95% CI 1.23-5.91, P = 0.0135; Supporting Table 2) and multivariate Cox regression analyses (recurrence: HR 1.68, 95% CI 0.81-3.51, P = 0.164; death: HR 1.92, 95% CI 0.92-4.02, P = 0.082; Supporting Table 3), and Kaplan-Meier analyses and log-rank test (time-to-recurrence, P =.0.004; overall survival, P = 0.002; Fig. 7C,D). Interestingly, NPM overexpression in HCC was particularly associated with higher early recurrence (within 24 months after initial hepatectomy; P = 0.007) and higher early mortality (P = 0.003; Supporting Fig. 2).
Table 1. Correlation of NPM Expression with Clinical Manifestations of HCC
No. of Cases
Score <3 (n = 93)
Score = 3 (n = 17)
A total of 110 cases with HCC are included. NPM expression was assayed via tissue array by IHC. IHC score [0, negative; 1, weakly positive; 2, positive; 3, strongly positive (high intensity in >60% HCC cells] was determined by two independent pathologists. Histology grade: Edmonson and Steiner's grading system. Tumor stage: AJCC/UICC, 7th edition. Statistical analysis was performed with a two-sided Fisher's exact test or Kruskal-Wallis test (for histological grade).
Abbreviations: AFP, alpha-fetoprotein; AJCC, American Joint Committee on Cancer; HBV, chronic hepatitis B; HCV, chronic hepatitis C; UICC, Union for International Cancer Control.
Tumor size, cm
Portal vein invasion
NPM is localized primarily in the nucleolus and shuttles between the nucleoli and cytoplasm during the cell cycle. However, little is known about the biological significance of cytoplasmic translocation of NPM mutations. We report here a novel NPM-BAX pathway in human HCC cells whereby cytoplasmic translocation of NPM plays a pivotal role in death evasion of HCC cells. In response to death stimuli, BAX is activated and translocated out of the nucleus and targets the mitochondria, where it oligomerizes on the mitochondrial membranes, thereby initiating mitochondria-mediated apoptosis (Fig. 6C, upper panel). By contrast, in cancer cells with NPM overexpression, a set of NPM is translocated from the nucleolus to the cytoplasm in response to cell stress, where it binds to BAX and blocks the mitochondrial translocation and oligomerization of BAX, so as to render cancer cells resistant to death stimulation (Fig. 6C, lower panel). Notably, NPM binds to BAX in the cytosol after stress stimulation, suggesting that an activated conformationally changed BAX is required for the binding. Indeed, NPM has been found to be a chaperone of BAX.28 A BAX C-terminal antibody specifically inhibited the BAX-NPM interaction indicates a specific interaction between the C-terminal of BAX and NPM.20
Loss of p53 functions by mutations of the TP53 gene plays a crucial role in tumorigenesis and is a great hurdle for anticancer therapy.29, 30 Interestingly, we found that sensitizing to anticancer therapies by silencing of NPM was more prominent in HCC cells harboring inactivated p53. Simultaneous silencing of p53 in these HCC cells did not further change the sensitizing effects by silencing of NPM alone. Obviously, this sensitization to anticancer therapies by silencing of NPM expression is different from reported p53-dependent NPM-mediated antiapoptosis mechanisms in malignant hematopoietic cells, whereby NPM regulates the stability and function of p53 to modulate resistance to cell death and mutations at the p53 interacting domain in NPM abrogate the antiapoptosis activity of NPM.19, 25, 31 Our findings are of particularly clinical significance, since inactivated mutations of p53 are found in more than half of HCC, particularly in advanced stage, for whom chemotherapy and target therapy are usually a last resort to control disease progression.32 Moreover, our findings that cotargeting NPM therapies are more effective in HCC harboring inactivated p53 imply that cotargeting NPM increases therapeutic specificity and efficacy in tumor cells harboring inactivated p53, but not nontumor cells whose TP53 genes usually remain not mutated. We thus speculate that cotargeting NPM with other anti-HCC therapies including molecular target therapies will not only increase therapeutic efficacy and specificity, but also lower therapeutic dosages, so as to reduce side effects accompanied by anticancer therapies. It is also intriguing to speculate that p53 mutations and NPM overexpression can predict the therapeutic efficacy of the NPM cotargeted therapies.
Noticeably, silencing of NPM greatly sensitizes HCC cells to lapatinib more than to sorafenib (Fig. 2). Lapatinib is a dual kinase inhibitor simultaneously suppressing epidermal growth factor receptor and HER2 signaling. Recently, we reported that HER2/ERBB3 signaling plays a crucial role in HCC progression and recurrence, suggestive of therapeutic benefits by targeting HER2/ERBB3 signaling pathways for HCC.22 However, clinical trials showed only modest effects of lapatinib in patients with advanced HCC.6, 33 Our current findings indicate that simultaneously targeting NPM and HER2/ERBB3 signaling might significantly attain therapeutic benefits in patients with advanced HCC, but further studies are warranted.
In conclusion, we have identified a novel NPM-BAX pathway orchestrating death evasion and sensitivity to anticancer therapies independently of p53 function in HCC cells. Following cell stress, NPM is induced and translocated from nucleolus to cytosol, where it directly binds to BAX and blocks its mitochondrial translocation and oligomerization, thereby rendering HCC cells resistant to death stimuli. Silencing of NPM expression greatly sensitizes HCC cells to anti-HCC therapies, particularly in those harboring inactivated p53. NPM is frequently overexpressed in HCC and is associated with more advanced stage and worse prognosis. NPM is a promising cotarget in combination with chemotherapy or target therapies for HCC. Our findings are of broad clinical significance because NPM up-regulation and inactivated mutations of p53 are usually found in advanced human cancers.
We thank the Taiwan Liver Cancer Network for providing the liver tumor tissue samples, tissue arrays, and related clinical data.