Phenylarsine Oxide Induces Apoptosis in Bax- and Bak-Deficient Cells through Upregulation of Bim

Most anticancer drugs exert their effects through the induction of intrinsic apoptosis via mitochondrial pathway characterized by the release of cytochrome c (Cyt c) and caspase activation (1). Bcl-2 and its relatives play a pivotal role in determining the apoptotic process. Members of Bcl-2 family are divided into 3 subsets: antiapoptotic (Bcl-2, Bcl-xL, Bcl-w, Bfl1/A-1, Bcl-B, and Mcl-1) proteins, proapoptotic proteins (Bax and Bak), and BH3-only proteins (Bid, Bim, Bik, Bad, Noxa, and Puma). Overwhelming evidence has shown that Bax and Bak are indispensable for mediating Cyt c release from mitochondria during apoptosis. Cells without Bax or Bak are largely resistant to apoptotic agents and drugs. Mouse embryonic fibroblasts (MEF) isolated from Bax/Bak double knockout (DKO) mice are resistant to multiple death stimuli, including DNA-damaging agents, signal transduction through death receptors, growth factor deprivation, and endoplasmic reticulum stress (2–4). On the other hand, several studies including ours suggest the existence of a novel mechanism of intrinsic apoptosis in a Bax/Bak-independent pathway, although the mechanism remains elusive (5–7). BH3-only proteins are suggested to act as proapoptotic molecules which either directly interact with and activate Bax/Bak for their membrane permeabilizing activity or bind to antiapoptotic proteins to suppress their inhibition on Bax/Bak (2, 8). Alternatively, a unified model suggests that the BH3-only proteins induced conformational alterations and oligomerization of both Bax/Bak and Bcl-2/Bcl-xL (9). The conformation-altered Bcl-2/Bcl-xL have the potential to form pores in mitochondrial outer membrane (MOM), but the pore size is not large enough to mediate Cyt c release (10, 11). Functional conversion of Bcl-2 to a proapoptotic protein has been observed in other circumstances. Upon binding to Nur77 protein or peptide, Bcl-2 undergoes conformational changes that convert it to a proapoptotic molecule and initiates apoptosis in a Bax- or Bak-dependent manner (12, 13). Our previous studies also revealed that treatment with gossypol triggered similar conformational changes in Bcl-2 which induced apoptosis in the absence of Bax and Bak.

Arsenic compounds, such as arsenic trioxide (ATO), sodium arsenate and monomethylarsonous acid, have successfully been used in clinical cancer therapies, especially in acute promyelocytic leukemia (APL) therapy. These compounds induce either differentiation or apoptosis in cancerous cells (14–16). ATO was found to reduce the expression of Bcl-2 and trigger Cyt c release in a Bax-dependent manner (17). Several other organic compounds including phenylarsine oxide (PAO) were found to be effective for APL or even solid tumors. PAO is a membrane-permeable arsenic compound that has the most potent effects on APL through induction of apoptosis in these cancerous cells (18, 19). PAO induced apoptotic cell death via activation of cysteine protease (18) and dissipation of the mitochondrial membrane potential (20), although conflicting reports showed that PAO inhibited caspase cleavage in vitro (21) or blocks TRAIL-induced cell death (19). It is yet to be determined whether and how Bax acts in PAO-induced mitochondrial apoptosis. PAO is also reported to act as a phosphotyrosine phosphatase inhibitor which induced a dramatic accumulation of phosphotyrosine in a number of cellular proteins (22), thus altering the cell signaling pathways. It thus suggests the mechanism of which PAO inhibited proliferation of acute myelocytic leukemia (AML) cell lines as well as freshly isolated cells derived from patients with AML and erythroleukemia (18).

Here, we found that PAO triggered Cyt c release in a Bax/Bak-independent manner. Our data revealed that PAO induced the increase of the expression of Bim which interacted with Bcl-2. Our studies further implicated the potential chemotherapeutic effects of PAO through modulating the interaction between Bim and Bcl-2.

PAO and ATO were purchased from Sigma. Antibodies (Abs) used for actin, Bim, Bcl-2, Bad, Puma, Bid, Bik, Mcl-1, Cyt c, PARP, and Bcl-xL were purchased from BD. Secondary Abs were purchased from Pharmingen. Annexin V kit of was obtained from Pharmingen. CaspACE In Situ kit was purchased from Promega. MitoTracker Red was purchased from Invitrogen. All other chemicals were purchased from Sigma unless otherwise specified.

SV40 T antigen–immortalized Bax/Bak DKO mouse embryonic fibroblasts (MEF), wild-type MEFs, bim^−/− MEFs, and HCT116 bax^−/− cells were cultured in Dulbecco's Modified Eagle's Medium supplemented with 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin. Cells were maintained in a humidified 5% CO₂ atmosphere at 37°C.

For Hoechst staining, cells were grown on glass cover slides at a density of 5 × 10⁴ cells/mL in a 6-well plate. After treatments, cells were stained with Hoechst in PBS for 15 minutes at room temperature in the dark. Cells were then washed 3 times with PBS and analyzed with a fluorescence microscope. At least 200 cells were counted.

For terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) staining, cells were grown on glass coverslips at a density of 5 × 10⁴ cells/mL in a six-well plate. After treatments, cells were stained by the In Situ Cell Death Detection Kit according to manufacturer's instructions. Apoptotic cells were visualized under fluorescence microscopy.

Quantitative apoptosis was detected by flow cytometric analysis after staining with fluorescein isothiocyanate (FITC)-conjugated Annexin V and propidium iodide (PI) by a commercially available kit according to manufacturer's instructions. All data were analyzed with Cell Quest software (BD).

Cells were treated as indicated, then collected and washed with PBS. CaspACE In Situ Marker was added to the cells to a final concentration of 10 μmol/L and incubated for 20 minutes in the dark. Cells were then washed 3 times with PBS and resuspended in 400 μL PBS and analyzed with a FACScan.

Cells were homogenized with a Dounce homogenizer, and the homogenate was centrifuged at 1,000 × g for 5 minutes to remove unbroken cells and nuclei. The cytosolic and mitochondrial fractions were obtained by further centrifugation as follows. Collect and spin the supernatant at 10,000 × g for 15 minutes. Collect the pellet which contains heavy membrane including mitochondria. Collect supernatant and spin at 100,000 × g for 1 hour to obtain cytosolic fraction and light membrane fraction.

The cells were grown on glass cover slips, treated with PAO and stained with 50 nmol/L MitoTracker Red for 30 minutes at 37°C. After washed with PBS, fixed in 3.7% formaldehyde/PBS solution, the cells were incubated in 0.1% Triton X-100/PBS to be permeabilized. Primary Abs were diluted in 1% bovine serum albumin (BSA)/PBS and incubated with the cells at 4°C for 12 hours. The FITC-conjugated secondary Abs were diluted in 1% BSA/PBS and incubated at room temperature for 2 hours. The cells were examined by confocal microscopy using the Zeiss LSM 510 META.

Total RNA was isolated from the cells with PAO treatment or not for reverse transcription and generation of cDNA. Primers for all the 3 isoforms of mouse Bim were designed as follows: forward (ATGGCCAAGCAACCTTCTGA) and reverse (TCAATGCATTCTCCACACCA). Primers for mouse glyceraldehyde-3-phosphate dehydrogenase (GAPDH): forward (ACCACAGTCCATGCCATCAC) and reverse (TCCACCACCCTGTTGCTGTA).

The mouse Bim short hairpin RNA (shRNA) sequence (TGATGTAAGTTCTGAGTGTG) was inserted into pSilencer 2.1-CMV-Hygro cut by BamHI and HindIII. Control scrambled and shRNA plasmids (1.5 mg) were transfected into DKO MEFs plated at 50% confluence in six-well plates using Lipofectamine 2000 (Invitrogen). Transfectants were selected in culture medium containing 100 μg/mL hygromycin to enrich the culture for cells that were successfully transfected.

Cells were lysed with 1% NP-40 lysis buffer (10 mmol/LHEPES, 150 mmol/L NaCl, 1% NP-40, pH 7.4) containing protease inhibitors. Total protein (500 μg) was incubated with 2 μg of anti-Bim antibody (Ab) in 500 μL of NP-40 lysis buffer at 4°C overnight on a rotator. Immunoprecipitates were collected by incubating with 20 μL protein A agarose for 2 hours at 4°C, followed by centrifugation for 1 minute. The pellets were washed 3 times with NP-40 lysis buffer, and beads were boiled in loading buffer and analyzed by Western blotting using the anti-Bcl-2 Ab.

The DKO MEFs were transformed with KRasV12 and E1A. The cells were harvested by centrifugation and suspended in Hank's balanced salt solution (HBSS). A tumor cell suspension (5 × 10⁵ KRasV12/E1A DKO MEFs or 3 × 10⁶ HCT116 bax^−/− cells in 100-μL PBS, respectively) was injected into the oxter of 4 to 5 weeks old nude nu/nu mice with a 27-gauge needle. Two groups of mice (5 animals each group) were used and maintained in specific pathogen-free (SPF) conditions. One week after implantation, animals were intraperitoneally treated with PAO (1 mg/kg/2 days) or vehicle control [dimethyl sulfoxide (DMSO)], respectively. The mice were euthanized after PAO treatment for 2 weeks. Tumor size was monitored by measurement of the length (a) and width (b) with a slide gauge. Tumor volumes (V) were calculated according to the formula: V = a × b²/2. Tumors were excised and stained by TUNEL staining. Apoptotic cells were visualized under fluorescence microscopy.

Four- to 5-week-old nude nu/nu mice were anaesthetized with phenobarbital during the surgical procedure. The caecum was exteriorized through a small midline laparotomy and 3 × 10⁶ HCT116 bax^−/− cells in 50-μL PBS were injected into the cecal wall with a 27-gauge needle. A cotton swab was held cautiously for 1 minute over the injection site to prevent leakage, and the abdominal wall was closed and sutured. One week after implantation, mice were randomly assigned into 2 groups (5 animals each group) and were intraperitoneally treated with PAO (1 mg/kg/2 days) or vehicle control (DMSO), respectively. The mice were euthanized after PAO treatment for 2 weeks. Primary tumors in the cecum were excised, the final tumor volumes (V) were calculated according to the formula: V = a × b²/2 (a, length; b, width), and the number of metastasis colonies was counted in the intestines. Each tumor tissue excised from mice was snap frozen in liquid nitrogen and divided into 2 parts. The first part was fixed in formalin and embedded in paraffin. Serial sections of the intestine were prepared and stained by hematoxylin and eosin (H&E) staining for histopathologic evaluation. The second part was stained by TUNEL staining. Apoptotic cells were visualized under fluorescence microscopy. All protocols were approved by the Institutional Animal Care and Use Committee of Institute of Zoology accredited by AAALAC International.

Data are expressed as mean ± SEM. Significant differences between values under different experimental conditions were determined by paired or unpaired Student t test and ANOVA with repeated measurements, and P < 0.05 was considered statistically significant.

We screened a number of compounds for their capacity to induce cell death in SV40 T antigen–immortalized DKO MEFs by MTT assay and found that PAO induced a significant reduction of viability in these cells (data not shown). However, other agents or drugs including ATO had no effect on the viability of DKO MEFs. To further examine PAO-induced apoptosis, we used Annexin V/PI double staining assay. The results showed that the increase of Annexin V–positive population was in both time- and concentration-dependent manners after PAO treatment (Fig. 1A). The typical apoptotic hallmark, the exposure of phosphatidylserine on cell surface, was observed after 1.5 μmol/L PAO treatment for 24 hours (Fig. 1A).

To show that PAO could induce typical intrinsic apoptotic cell death in DKO MEFs, we analyzed intracellular caspase activity with CaspACE In Situ Marker. Caspase was activated by PAO both in time- and concentration-dependent manners (Fig. 1B). PAO also induced the proteolytic cleavage of PARP which is the substrate of caspase-3 in DKO MEFs (Fig. 1C). The caspase inhibitor (z-VAD-fmk) inhibited PAO-induced phosphatidylserine exposure and caspase activation (data not shown). In addition, nucleus condensation and fragmentation, which is a defining characteristic for apoptosis, was detected by Hoechst staining after PAO treatment (Fig. 1D). Collectively, the results showed that PAO induced intrinsic apoptosis in Bax- and Bak-deficient cells.

Furthermore, we tested the effect of PAO on cancerous cells with HCT116 and HCT116 bax^−/− cell lines, respectively. Dramatically, PAO rather than ATO induced apoptosis which was in a time-dependent manner in these colorectal cancer cells, regardless of Bax absence (Fig. 1E). Data of TUNEL staining (Fig. 1F) and caspase activation assay (Supplementary Fig. S1) also showed that PAO could induce apoptotic cell death in Bax-deficient cancer cells, which is resistant to multiple death stimuli.

The release of Cyt c from mitochondria into cytosol is a defining event which triggers intrinsic mitochondrial apoptotic signals (23). Previous reports have suggested the existence of Cyt c release without Bax and Bak (5, 6). We thus studied Cyt c release during PAO-induced apoptosis in the absence of Bax and Bak. We first exposed DKO MEFs to 1.5 μmol/L PAO, then fractionated and extracted mitochondria to detect the protein levels of Cyt c in the cytosolic and mitochondrial fractions, respectively. The Western blotting data revealed that PAO induced Cyt c release in a time-dependent manner, as the amount of Cyt c increased in the cytosolic fraction with an accompanied decrease in the mitochondrial fraction (Fig. 2A). We also examined intracellular distribution of endogenous Cyt c by immunostaining. Cyt c showed a diffused distribution in cytosol after 1.5 μmol/L PAO treatment for 12 hours, whereas it was well colocalized with mitochondria in untreated cells (Fig. 2B). Consistent with the fractionation data, these results indicate that PAO induced Cyt c release in a Bax/Bak-independent manner. Cyt c release was also detected in Bax-deficient colorectal cancer cells (Fig. 2C), which further indicates that PAO-induced apoptosis is through a mitochondrial signaling pathway and PAO-induced Cyt c release is Bax independent.

Next, we sought to understand the mechanism of PAO-induced Bax/Bak-independent apoptosis. Despite the absence of Bax and Bak, other Bcl-2 family members could be activated to mediate Cyt c release as we previously suggested (7). We thus tested the expressions of several members of Bcl-2 family (Puma, Bik, Bim, Bid, Bad, Bcl-2, Bcl-xL, and Mcl-1) during PAO-induced apoptosis. Interestingly, we found the protein expressions of 3 isoforms of Bim including BimEL, BimL, and BimS were all upregulated in DKO MEFs after the treatment with 1.5 μmol/L PAO for 6 hours, whereas the expressions of other Bcl-2 family proteins had no significant changes upon PAO treatment (Fig. 3A). Similarly, the protein expressions of all 3 isoforms of Bim were upregulated in HCT116 bax^−/− cells after PAO treatment for 12 hours (Fig. 3C). We further analyzed mRNA expression of Bim by reverse transcriptase PCR (RT-PCR) and found that the amount of mRNA of these isoforms was increased after PAO treatment (Supplementary Fig. S2A). More interestingly, fractionation followed by Western blotting analysis showed the protein levels of these isoforms of Bim were increased both in cytosolic and mitochondrial fractions during PAO-induced apoptosis (Fig. 3B). Immunostaining analysis by confocal microscopy also showed that the expression of endogenous Bim was induced both in cytosol and in mitochondria in the cells treated by PAO, compared with control ones (Supplementary Fig. S2B).

Bim has been characterized as serving as a death ligand, which regulates other members of the Bcl-2 family as a proapoptotic molecule once it has been activated (2). To find out whether the upregulation of Bim induced by PAO is required for apoptosis in DKO MEFs, we knocked down the expression of Bim. A shRNA which is specific to all the 3 isoforms of Bim was stably transfected into the DKO MEFs (Fig. 4A). As we expected, knockdown of Bim significantly inhibited PAO-induced apoptosis in DKO MEFs, compared with those stably transfected with scrambled shRNA (Fig. 4B). To further show the role of Bim in PAO-induced apoptosis, we treated bim^−/− MEFs and the wild-type MEFs with PAO, respectively. Annexin V/PI double staining assay showed that PAO-induced apoptotic cells were decreased significantly in the absence of Bim, regardless of the presence of Bax/Bak (Fig. 4C). All these data revealed that Bim was involved in and initiated PAO-induced apoptosis in DKO MEFs.

It has been reported that Bim could suppress Bcl-2 which is a primary inhibitor of apoptosis and constitutively localizes on outer mitochondrial membrane (2). As shown in Fig. 3B, the protein expression of Bim on mitochondria was increased after PAO treatment, and immunostaining data confirmed the localization of Bim on mitochondria (Supplementary Fig. S2B). We thus supposed that PAO induced apoptosis in Bax- and Bak-deficient cells by augmenting the expression of Bim on mitochondria where it inhibited Bcl-2. To test this hypothesis, we detected the interaction between endogenous Bim and Bcl-2 by immunoprecipitation. The amount of Bim/Bcl-2 complex was increased after PAO treatment for 6 hours, indicating that Bim interacted with Bcl-2 and the interaction between them was strengthened during PAO-induced apoptosis (Fig. 4D). To further investigate the role of Bcl-2 in Bax/Bak-independent apoptosis, we used YC137 as a specific inhibitor to Bcl-2 (24). PAO-induced apoptotic cells were significantly increased in YC137-pretreated DKO MEFs, compared with control ones in the absence of YC137 (Fig. 4E). This Bcl-2 inhibitor augmented PAO-induced apoptosis, suggesting that Bcl-2 was involved in this process.

We next tested the effects of PAO on the growth of xenograft tumors in vivo. The immunodeficient nude mice were implanted with KRasV12/E1A-transformed DKO MEFs in the oxter and underwent intraperitoneal treatment with PAO (1 mg/kg/2 days) or DMSO (the vehicle control) one week later. The tumor volume was measured every other day, and it was found that PAO inhibited tumor growth (Fig. 5A). Two weeks after PAO treatments, the mice were euthanized, and the xenograft tumors were isolated. The average tumor weight of the treated group was significantly reduced, compared with that of the control group (Fig. 5B and C). Importantly, this dose of PAO inhibited tumor growth in the mice without significant converse effects on body weight (Fig. 5D). Moreover, we detected PAO-induced apoptosis in vivo in tumors formed with DKO MEFs by TUNEL assay in situ. Fluorescence microscopy images showed an increase of apoptotic cells in tumors from the group treated with PAO, compared with those from control group (Fig. 5E).

Both subcutaneous and orthotopic xenograft tumor implantation models were used here to test the effects of PAO on the growth of Bax-deficient cancer cells. In the subcutaneous tumor model, HCT116 bax^−/− cells were injected in the oxter of nude mice. One week after implantation, mice were treated with PAO or DMSO as the vehicle control, respectively. Two weeks after PAO treatments, the mice were euthanized and the xenograft tumors were isolated. PAO attenuated the growth of xenograft tumors (Fig. 6A) and reduced the tumor weight dramatically (Fig. 6B and C). Consistent with the data from DKO MEFs in vivo assay, the treatments of PAO in the dose of 1 mg/kg/2 days have no significant converse effects on the mice (Fig. 6D). TUNEL assay in situ further revealed that PAO induced apoptotic cell death in HCT116 bax^−/− cells in vivo (Fig. 6E).

In the orthotopic tumor model, the treatment started one week after HCT116 bax^−/− cells implantation and was continued for 2 weeks. The mice were euthanized, and the tumors were isolated, dissected, and measured. The results showed that PAO decreased orthotopic tumor growth (Fig. 6F) and the tumor volume in PAO treatment group was significantly lower than that in the control group (Fig. 6G). Furthermore, the presence of orthotopic xenograft tumors in the cecum was confirmed by H&E staining (Fig. 6H). PAO-induced apoptotic cell death was confirmed by TUNEL staining (Fig. 6I), and PAO-induced apoptosis was not detected with TUNEL assay in normal colorectal tissues (Supplementary Fig. S3).

In this study, we found that PAO could induce typical intrinsic apoptosis in Bax/Bak DKO MEFs, which are largely resistant to multiple chemotherapeutical drugs and apoptosis-inducing agents. PAO induced dramatic increase of Bim expression which is involved in Bax/Bak-independent apoptosis. PAO also induced apoptotic cell death in chemotherapy-resistant colorectal cancer cells, such as HCT116 bax^−/− cells. The results contrast with our previous report which showed that ATO induced apoptosis in a Bax-dependent manner (22). Our data also showed that Bim was not essential for ATO-induced apoptosis in MEFs as ATO could induce apoptosis in the absence of Bim (Supplementary Fig. S6D), although Bim was reported to be transcriptional upregulated and activated by ATO in neuronal cell line (25, 26). Clearly, PAO is much more powerful in overcoming drug resistant cancers that are refractory to conventional chemotherapeutical agents. Our data suggested that PAO treatment had advantages over the conventional anticancer therapies that rely on the Bax or Bak to kill cancer cells. Moreover, the orthotopic tumor implantation model in nude mice clearly showed that PAO inhibited the growth of Bax-deficient cancer cells in vivo without significant side effects on mice. It has been well known that both bax and bak genes are frequent targets for mutation in a subset of human tumors, including hematopoietic malignancies, breast cancers, gastric cancers, colorectal cancers, and prostate cancers (27–31). Loss or mutational inactivation of Bax or Bak is often reported and considered as a mechanism of protection against apoptosis in cancer cells, and these cancers are largely resistant to chemotherapy, including arsenic compounds treatments (32–35). Therefore, our study implicates that PAO has potential value for overcoming chemotherapy resistance in the treatments of cancers, especially in the treatments of Bax or Bak inactivation cancers.

Our results provide new evidence for the existence of a Bax/Bak-independent apoptosis pathway to mediate Cyt c release. It is recently reported that apoptosis could be induced through a mitochondrial permeability transition pore (mPTP)—not Bax/Bak-dependent manner. However, here, specific mPTP inhibitor cyclosporine A (CsA) or bongkrekic acid (BA) fails to inhibit PAO-induced apoptosis (Supplementary Figs. S4 and S6C). Prior studies showed that PAO increased the concentration of intracellular calcium (20), which was indicated to be involved in Bax/Bak-independent mechanism of Cyt c release (5). Here, we found that EGTA-am had no effect on PAO-induced apoptosis (Supplementary Fig. S5), suggesting that intracellular calcium is not involved in this process. Ionomycin could indeed induce cell death in HCT116 bax^−/− cells in an EGTA-am–inhibitable manner. However, activated caspase and other typical intrinsic apoptotic hallmarks were not detected in ionomycin-induced cell death (data not shown). Intriguingly, we showed that PAO upregulated the expression of Bim which is involved in PAO-induced apoptosis in the absence of Bax/Bak. PAO strengthened the interaction between Bim and Bcl-2. Moreover, specific suppression of Bcl-2 by YC137 can enhance PAO-induced apoptosis. This is in line with our previous finding that Bcl-2 underwent structural change during Bax/Bak-independent apoptosis, which might convert the protective molecule into a killer (7). Our recent results showed that Bim binding to Bcl-2 could trigger conformational and functional conversion of Bcl-2, which permeabilized the MOM, thereby triggering Cyt c release during the apoptosis process in the absence of Bax and Bak. The exact mechanism of how Bim interacts with Bcl-2 for the proapoptotic action is still a subject in future study. Nevertheless, the unexpected effects of Bim and Bcl-2 on PAO-induced apoptosis provide a novel view to explore the mechanism of Bax/Bak-independent apoptosis and imply new targets in the treatments of Bax or Bak inactivation cancers which are chemotherapy resistant.

No potential conflicts of interest were disclosed.

Q. Ma and Q. Chen designed the experiments. B. Ni, Q. Ma, B. Li, L. Zhao, Y. Liu, and Y. Zhu carried out the experiments. Q. Ma made the data statistics and analysis. Q. Ma and Q. Chen drafted the manuscript. All authors read and approved the final manuscript.

The authors thank Weixing Zong for generously sharing bax^−/− bak^−/− MEFs and David Huang for sharing wild-type and bim^−/− MEFs.

This work is supported by grants from the National Key Basic Research Program of China to Q. Chen (2009CB521802, 2011CB910903) and Y. Zhu (2010CB912200), and the National Science Foundation of China to Q. Ma, Y. Liu, and Q. Chen (31000614, 31000615, 81130045).

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