Potential conflict of interest: Nothing to report.
Amplification of 1q21 has been detected in 58% to 78% of primary hepatocellular carcinoma cases, suggesting that one or more oncogenes within the amplicon play a critical role in the development of this disease. The chromodomain helicase/adenosine triphosphatase DNA binding protein 1–like gene (CHD1L) is a recently identified oncogene localized at 1q21. Our previous studies have demonstrated that CHD1L has strong tumorigenic ability and confers high susceptibility to spontaneous tumors in a CHD1L-transgenic mouse model. In this study, we demonstrate that the antiapoptotic ability of CHD1L is associated with its interaction with Nur77, a critical member of a p53-independent apoptotic pathway. As the first cellular protein identified to bind Nur77, CHD1L is able to inhibit the nucleus-to-mitochondria translocation of Nur77, which is the key step of Nur77-mediated apoptosis, resulting in the hindrance of the release of cytochrome c and the initiation of apoptosis. Knock-down of CHD1L expression by RNA interference could rescue the mitochondrial targeting of Nur77 and the subsequent apoptosis. Further studies found that the C-terminal Macro domain of CHD1L is responsible for the interaction with Nur77, and a CHD1L mutant lacking residues 600-897 failed to interact with Nur77 and prevented Nur77-mediated apoptosis. More importantly, we found that the inhibition of Nur77-mediated apoptosis by endogenous CHD1L is a critical biological cellular process in hepatocarcinogenesis. Conclusion: We demonstrate in this study that overexpression of CHD1L could sustain tumor cell survival by preventing Nur77-mediated apoptosis. (HEPATOLOGY 2009.)
Hepatocellular carcinoma (HCC) is the sixth most common human cancer in the world, and its prognosis is extremely poor.1 Amplification of 1q21 is one of the most frequent genetic alterations in HCC, which was detected in 58% to 78% of HCCs,2–5 suggesting the existence of an oncogene within this region. Recently, we isolated a candidate oncogene, the chromodomain helicase/adenosine triphosphatase DNA binding protein 1–like gene (CHD1L, previously called ALC1) within the 1q21 amplicon by hybrid selection using microdissected DNA from 1q21.6 Amplification and overexpression of CHD1L was detected in over 50% of primary HCC specimens and has strong oncogenic activities, as demonstrated by its ability to promote cell proliferation, increase colony formation in soft agar, induce tumor formation in nude mice, promote cell cycle progression, and inhibit cell apoptosis.6 To investigate the oncogenic role of CHD1L in vivo, a CHD1L ubiquitous-expression transgenic mouse model was generated. Spontaneous tumor formation was found in 10 of 41 (24.4%) transgenic mice with the formation of HCC in four mice, but not in their 39 wild-type littermates (unpublished data). These data strongly suggest that CHD1L is the target oncogene responsible for the 1q21 amplification event and that it plays a critical role in HCC development.
To further explore the molecular mechanism of CHD1L in hepatocarcinogenesis, a yeast two-hybrid assay was used to identify proteins that interact with CHD1L. The complementary DNA (cDNA) fragment of CHD1L (nt1043-2694, encoding aa349-897) was used as bait to screen a human fetal liver cDNA library. One protein, Nur77 (also called NR4A1, TR3, and NGFI-B), was isolated (unpublished data). Nur77 is a unique transcriptional factor belonging to the orphan nuclear receptor superfamily.7–9 Similar to the death signal activated when p53 interacts with antiapoptotic and proapoptotic Bcl-2 family members in mitochondria, Nur77 directly targets Bcl-2 and induces the latter to adopt a proapoptotic conformation that triggers downstream apoptotic events, including cytochrome c (Cyt c) release, activation of Apaf-1 and caspase-9, and cleavage of pro–caspase-3.10–13 The apoptosis-associated mitochondrial targeting of Nur77 has now been observed in cancer cells from the lung,14 stomach,15 ovary,16 colon,17 and prostate.12 These properties of Nur77 prompted us to investigate the molecular mechanism involved in the enhancement of tumor cell survival by CHD1L, which might suppress apoptosis through binding to Nur77 and affecting Nur77-mediated apoptotic process.
cDNA, complementary DNA; CHD1L, chromodomain helicase/ATPase DNA binding protein 1–like; CO-IP, co-immunoprecipitation; Cyt c, cytochrome c; DAPI, 4',6-diamidino-2-phenylindole; GFP, green fluorescent protein; HCC, hepatocellular carcinoma; HM, heavy membrane; PARP, poly(ADP)-ribose polymerase; siRNA, small interfering RNA; STS, staurosporine; TUNEL, terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling.
Materials and Methods
HCC cell lines QGY-7703, BEL7402, PLC8024, Hep3B, Huh7, and HepG2 and two immortalized normal human liver cell lines (LO-2 and Chang liver) were obtained from the Institute of Virology, Chinese Academy of Medical Sciences, Beijing, China. Two other HCC cell lines, H2-M and H2-P, were previously established in our laboratory.18
Plasmid Constructs and Transfection.
Full-length CHD1L cDNA was amplified and cloned into expression vector pCDNA3.1 (Invitrogen) as described.6 PEGFP-CHD1L, pEGFP-Nur77, pDsRed2-Nur77, and plasmids carrying mutated forms of CHD1L were generated as described in the Supporting Materials and Methods. All expression plasmids were transfected into QGY-7703 cells using Lipofectamine (Invitrogen) according to the manufacturer's instructions.
Cells were treated with 2 μM staurosporine (STS) for different times (0, 2, 4, and 8 hours) and were collected for flow cytometry analysis after staining with Annexin-V–fluorescein isothiocyanate and propidium iodide (PI) using the Annexin-V–Fluos Staining Kit (Roche). Staining was performed as described in the Supporting Materials and Methods. Those cells stained by Annexin-V but not by propidium iodide were counted as apoptotic cells.
Apoptotic cells were determined using the In Situ Cell Death Detection Kit POD (Roche) according to the manufacturer's instructions.
Antibodies and Western Blotting.
Western blot analysis was performed as described,6 with antibodies described in the Supporting Materials and Methods.
Co-immunoprecipitation (CO-IP) assay was performed using the immunoprecipitation kit (Roche) according to the manufacturer's instructions.
Cells were transiently transfected with green fluorescent protein (GFP)-Nur77, then stained with Mitotracker-Red (Molecular Probes) to reveal mitochondria, and finally fixed and mounted with mounting medium containing 4',6-diamidino-2-phenylindole (DAPI) (Vector Laboratories). Images were captured using a fluorescence microscope (Lecia) or a confocal laser-scanning microscope (Zeiss LSM 510 META).
Separation of nuclear and cytosolic fractions was performed using the NucBuster Protein Extraction kit (Novagen) according to the manufacturer's instructions. The mitochondrial heavy membrane (HM) fraction was separated from the cytosolic fraction using the ApoAlert Cell Fractionation Kit (Clontech). An antibody against COX4 (1:200) was used to confirm the successful separation of mitochondrial fractions.
Small interfering RNA (siRNA) (20 μM) against CHD1L (Ambion) was transfected into cells in 6-well plates using Lipofectamine 2000 Reagent (Invitrogen) according to the manufacturer's instructions. At 48 hours after transfection, the effects of gene silencing were measured via northern blot analysis.
Statistical analysis was performed using an independent t test with Microsoft Office Excel software (Microsoft Corp., Redmond, WA). Statistical significance was set at P < 0.05.
CHD1L Interacts with Nur77.
Recently, the CHD1L-interacting protein Nur77 was identified via yeast two-hybrid screening (unpublished data). We confirmed the interaction between Nur77 and CHD1L in the HCC cell line QGY-7703 via (CO-IP) assay. Plasmids expressing CHD1L tagged with GFP or GFP only were transiently transfected into QGY-7703 cells to yield GFP/CHD1L-7703 or GFP-7703 cells, respectively. Cell lysates from these two cell lines were immunoprecipitated with Nur77 antibody, and western blotting was used to detect the existence of GFP/CHD1L and Nur77 protein in Nur77 immunoprecipitates. As shown in Fig. 1A, CHD1L was specifically co-immunoprecipitated by Nur77 antibody, and Nur77 protein was also specifically co-immunoprecipitated by anti-GFP antibody. The interaction of endogenous CHD1L and Nur77 was further confirmed via CO-IP experiments in HepG2 cells. The CO-IP results indicated endogenous CHD1L was detected in Nur77 immunoprecipitates (Fig. 1B). To determine the subcellular localization of CHD1L and Nur77, GFP-CHD1L and DsRed2-Nur77 were cotransfected into QGY-7703 cells. As shown in Fig. 1C, CHD1L (green) and Nur77 (red) were colocalized in the nucleus. These results strongly demonstrate the interaction of CHD1L with Nur77.
CHD1L Inhibits Apoptosis in HCC Cells.
To select an HCC cell line for functional study, CHD1L expression in eight HCC cell lines (QGY-7703, BEL7402, PLC8024, Huh7, HepG2, Hep3B, H2-P, H2-M) and two immortalized normal liver cell lines (LO-2 and Chang liver) was tested via western blot analysis. As shown in Fig. 2A, the lowest expression of CHD1L was detected in QGY-7703 cells. CHD1L was then stably transfected into QGY-7703 and two CHD1L transfectants (CHD1L-7703-C3 and CHD1L-7703-C5) with high CHD1L expression were selected for further study (Fig. 2B). After the treatment of the universal apoptosis inducer STS, the apoptotic cells were double-stained by Annexin-V–fluorescein isothiocyanate and propidium iodide and then detected via flow cytometry. The apoptotic index, defined as the percentage of apoptotic cells (R4), was compared between CHD1L-7703 and vector only–transfected QGY-7703 (Vec-7703) cells (Fig. 2C). The results showed that the apoptotic index of Vec-7703 cells increased rapidly during the first 4 hours, reaching over 90% after 8 hours of exposure to STS. In contrast, the apoptotic index of CHD1L-7703 cells was significantly lower at every time point in the presence of STS (Fig. 2C,D). These data indicate that CHD1L dramatically inhibits the apoptotic progression of HCC cells upon STS stimulation.
CHD1L/Nur77 Interaction Inhibits the Nucleus-to-Mitochondria Translocation of Nur77 and Subsequent Initiation of Apoptosis.
In light of the recent finding that the proapoptotic activity of Nur77 is associated with its translocation from the nucleus to the mitochondria,12 we investigated whether the interaction between CHD1L and Nur77 affects this translocation. Nur77 tagged with GFP (GFP/Nur77) was transiently transfected into both Vec-7703 and CHD1L-7703 cells, and apoptosis was induced by STS treatment. In the absence of STS, the GFP/Nur77 fusion protein (green) localized in the DAPI-stained nucleus (blue) in both Vec-7703 and CHD1L-7703 cells (Fig. 3A). When exposed to STS for 2 hours, GFP/Nur77 began to migrate from the nucleus to the mitochondria, where it colocalized with Mitotracker-Red staining (red). At 4 hours, the protein had translocated completely to the mitochondria (Fig. 3A, upper panel). Notably, in CHD1L-7703 cells, GFP/Nur77 remained in the nucleus during the first 2 hours of exposure to STS (Fig. 2A, bottom panel). After the exposure to STS for 4 hours, only partial translocation of GFP/Nur77 from the nucleus to the mitochondria was detected, suggesting that binding of CHD1L to Nur77 can hinder the latter's nucleus-to-mitochondria translocation (Fig. 3A, bottom panel). The effect of CHD1L binding to Nur77 on the latter's mitochondrial translocation was further confirmed via immunoblotting analysis. As shown in Fig. 3B, in Vec-7703 cells, STS treatment caused a significant decrease of Nur77 protein level in nucleus, accompanied by a concomitant increase in mitochondria-enriched HM fraction. However, in CHD1L-7703 cells, there was little change in Nur77 expression level in nucleus and HM fraction, further supporting our notion that CHD1L was able to bind to Nur77 and inhibited its nucleus-to-mitochondria translocation.
Next, it has been reported that when Nur77 targets mitochondria, it will induce Cyt c release from the mitochondria to the cytoplasm.12 Once released, Cyt c leads to a conformational change in Apaf-1, leading in turn to the activation of pro–caspase-9 and the subsequent cleavages of pro–caspase-3 and poly(ADP)-ribose polymerase (PARP).19, 20 In view of this, we hypothesized that CHD1L may exert its antiapoptotic function through the apoptotic pathway involving Nur77–Cyt c–caspase-9–caspase-3. To address this question, the effect of CHD1L on the Cyt c–caspase-9–caspase-3 apoptotic pathway was studied in Vec-7703 and CHD1L-7703 cells. Consistent with the mitochondrial accumulation of Nur77 in response to STS, STS treatment of Vec-7703 cells led to release of Cyt c into the cytoplasm in a time-dependent manner (Fig. 3C). Subsequently, three cleaved forms of caspase-9 (37, 35, and 17 kDa), the active subunits of caspase-3 (17 kDa) and the active subunit of PARP (85 kDa) were also detected in STS-treated Vec-7703 cells (Fig. 3D). This indicates that the Cyt c–caspase-9–caspase-3 pathway and its downstream proteins were all activated in a time-dependent manner. In CHD1L-7703 cells, however, the cytoplasmic accumulation of Cyt c, the cleavage of pro–caspase-9, and the subsequent activations of caspase-3 and PARP were all suppressed (Fig. 3C,D).
To study if another apoptotic pathway (such as the death receptor pathway) is involved in the antiapoptotic effects of CHD1L, CHD1L-7703 and Vec-7703 cells were treated with 5-fluorouracil, which is able to trigger the death receptor pathway. The result showed that 5-fluorouracil could induce apoptosis through the activation of death receptor pathway (cleavage of pro–caspase-8). However, activation of the death receptor pathway was not associated with CHD1L, because no difference was observed in the level of cleaved caspase-8 (43/41 kDa) between Vec-7703 and CHD1L-7703 cells (Supporting Fig. 1).
Silencing CHD1L Expression by siRNA Restores STS-Induced Apoptosis.
To further confirm our findings, RNA interference was used to knock down the expression of CHD1L in CHD1L-7703 cells. An siRNA against CHD1L effectively down-regulated CHD1L expression, whereas a negative control siRNA against GFP showed no effect (Supporting Fig. 2A). Silencing of CHD1L using siRNA not only removed the suppression of the nucleus-to-mitochondria translocation of Nur77 (Fig. 4A), but also restored STS-induced apoptosis detected via terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling (TUNEL) assay (Supporting Fig. 2B). These results were further confirmed via immunoblotting analysis. In contrast to a negative control siRNA, an siRNA against CHD1L promoted the mitochondrial accumulation of Nur77 following STS treatment (Fig. 4B), thereby triggering release of Cyt c into cytoplasm (Fig. 4C) and the cleavage of caspase-9 and caspase-3 (Fig. 4D). These data demonstrate that knock-down of CHD1L expression restores Nur77-mediated apoptotic process.
The C-Terminal Macro Domain of CHD1L Is Responsible for the CHD1L/Nur77 Interaction.
CHD1L contains a conserved SNF2-N domain, a helicase superfamily domain (HELICc), and a C-terminal Macro domain.6 To identify the CHD1L domains responsible for interaction with Nur77, different constructs of CHD1L fused to GFP were generated, including CHD1L/1-600 (encoding aa1-600), CHD1L/300-500 (aa300-500), and CHD1L/300-897 (aa300-897) (Fig. 5A). CO-IP assays showed that only CHD1L/300-897 was precipitated by the Nur77 antibody (Fig. 5B), indicating that the C-terminal domain (aa600-897) of CHD1L was responsible for interaction with Nur77. To confirm this observation, CHD1L/1-600 and CHD1L/300-897 were transiently transfected into QGY-7703 cells. Confocal image analysis indicated that CHD1L/300-897, but not CHD1L/1-600, impeded the mitochondrial translocation of Nur77 in response to STS (Fig. 5C). Results of the TUNEL assay revealed that the apoptotic index was significantly higher in QGY-7703 cells expressing CHD1L/1-600 than the counterparts in cells expressing CHD1L/300-897 (P < 0.001) (Fig. 5D). Western blotting results demonstrated that the cleavages of caspase-9, caspase-3, and PARP were all suppressed under STS stimulation in QGY-7703 cells expressing CHD1L/300-897 compared with cells expressing CHD1L/1-600 (Supporting Fig. 3). These data suggest that the C-terminal domain of CHD1L (aa600-897) is responsible for the protein's antiapoptotic activity.
Blockade of Nur77-Mediated Apoptosis by Endogenous CHD1L Is a Pivotal Biological Process in Hepatocarcinogenesis.
In order to determine how inhibition of Nur77-mediated apoptosis by CHD1L may account for uncontrolled cell survival during HCC progression, we investigated the correlation between the expression level of endogenous CHD1L and the apoptotic index among eight HCC cell lines and two normal liver cell lines. Notably, the expression level of CHD1L was negatively associated with the apoptotic index in these 10 cell lines (Figs. 2A, 6A). The apoptotic indices in five cell lines with low-level expression of CHD1L (H2-P, QGY-7703, H2-M, Chang liver, and LO-2) were approximately two-fold higher than that in two cell lines (HepG2 and BEL7402) with high-level expression of CHD1L (Fig. 6A).
To confirm if Nur77-mediated apoptosis was inhibited by endogenous CHD1L, LO-2 (with low-level expression of CHD1L) and HepG2 cells (with high-level expression of CHD1L) were selected for further study. After STS treatment, the mitochondrial translocation of Nur77 could be clearly observed in LO-2 cells but not in HepG2 cells (Fig. 6B). Western blot analysis confirmed that the mitochondrial accumulation of Nur77 and initiation of apoptosis were detected in LO-2 cells but inhibited in HepG2 cells (Fig. 6C-E), suggesting that the endogenous CHD1L plays an important role in the enhancement of HCC cell survival through the inhibition of Nur77-mediated apoptosis.
Our previous study found that the oncogenic function of CHD1L was associated with its antiapoptotic ability.6 The inhibition of apoptosis is one of the major mechanisms in cancer development, which ultimately extends cell survival and allows for the accumulation of genetic instability and mutations. Based on our observation in which CHD1L interacts with Nur77, we investigated the molecular mechanism of the antiapoptotic role of CHD1L. In the present study, we demonstrated that CHD1L could bind Nur77 and retain the latter in the nucleus, and subsequently inhibit the nucleus-to-mitochondria translocation of Nur77. In response to apoptotic stimuli, Nur77 translocates from the nucleus to the mitochondria and subsequently triggers the initiation of apoptosis. This paradigm in cell apoptosis has been described in various cell types, including prostate, breast, colon, lung, ovary, and gastric cancer cells.12–17, 21 However, the mechanism by which the nucleus-to-mitochondria translocation of Nur77 is manipulated is still unclear. A previous study showed that Epstein-Barr virus nuclear antigen EBNA2 can bind to Nur77 and retain it in the nucleus, thereby hindering Nur77-mediated apoptosis.22 To our knowledge, CHD1L is the first known cellular protein that is able to bind Nur77 and inhibit its nucleus-to-mitochondria translocation under apoptotic stimuli.
In this study, STS, a potent inducer of apoptosis, was used to trigger the nucleus-to-mitochondria translocation of Nur77. Under normal circumstances, in response to apoptotic stimuli, Nur77 translocates from the nucleus to the mitochondria, where it interacts with Bcl-2. The interaction of Nur77 with Bcl-2 induces a conformational change in Bcl-2, converting it into a proapoptotic molecule that triggers the release of Cyt c and the Cyt c–caspase-9–caspase-3 apoptotic pathway (Fig. 7).11–13 However, when CHD1L interacts with Nur77, it inhibits the nucleus-to-mitochondria translocation of Nur77, thereby antagonizing the subsequent release of Cyt c, the cleavage of caspase-9, and the activation of caspase-3 (Fig. 7). Further studies have indicated that the C-terminal Macro domain of CHD1L (aa600-897) is responsible for this interaction with Nur77. A CHD1L mutant lacking residues 600-897 did not interact with Nur77 and failed to prevent Nur77 nuclear export and subsequent apoptosis. In addition to the mitochondrial pathway (involving caspase-9), the effect of CHD1L on death receptor pathway (involving caspase-8) was also investigated in this study. The result indicated that the antiapoptotic effect of CHD1L was not associated with the death receptor pathway.
The apoptosis-associated nuclear export of Nur77 has been detected in several cancers. In ovarian cancer cells, nucleus-to-mitochondria translocation of Nur77 was detected after AHPN/CD437 treatment.23, 24 A similar phenomenon was also detected in the prostate cancer cell line LNCaP and in the lung cancer cell line H460.12, 13, 25 In this study, we demonstrated that the mitochondrial targeting of Nur77 was a crucial component of apoptotic response in HCC cells, and the blockade of Nur77-mediated apoptosis was attributed to the high expression level of endogenous CHD1L. Consistently, expression levels of the endogenous CHD1L were negatively correlated with the apoptotic index among HCC cell lines, suggesting the antiapoptotic effect of CHD1L through preventing the nucleus-to-mitochondria translocation of Nur77 plays a critical role in hepatocarcinogenesis and HCC cell survival.
During cancer progression, avoiding apoptosis is one of the major mechanisms, because it extends cell survival and allows for the accumulation of genetic instability and mutations.26, 27 Thus, exploring the mechanisms underlying the Nur77-mediated apoptotic pathway will greatly facilitate our understanding of cancer development and progression. It has been reported that high mechanistic similarities exist between the pro-death paradigm of p53 and Nur77 in mitochondria.10 Although both p53 and Nur77 can directly target mitochondria to induce apoptosis, the Nur77-mediated apoptotic pathway is p53-independent.10 Recent evidence provides sufficient rationale to further investigate how this mitochondrial Nur77 pathway could become an exploitable target for new cancer therapeutics.10 In this study, we have shown that the inhibition of Nur77-mediated apoptosis is a biological process in hepatocarcinogenesis and have elucidated the critical role of CHD1L in preventing a Nur77-mediated apoptotic pathway. Given that CHD1L is frequently amplified and overexpressed in HCC, a better understanding of the oncogenic mechanisms of CHD1L in HCC development and progression may lead to a much more effective strategy for the management of HCC with amplification of CHD1L via early detection, precise prognostication, and molecularly targeted treatment.