Introduction

Arsenic trioxide (As2O3) has been used clinically to effectively treat acute promyelocytic leukemia (APL) 1, 2 and other types of cancers 3, 4. This success prompts studies to investigate the molecular mechanisms of action underlying the clinic effectiveness. As2O3 might affect multiple intracellular signaling pathways, leading to alterations of cellular function. Subsequently, this compound was found to inhibit proliferation, or stimulate differentiation of tumor cells depending on the dosage and the cell system being used 5. Increasing evidence has shown that As2O3 inhibits growth of APL cell lines and some solid tumors by inducing apoptosis 6, 7, 2. As2O3 specifically induces degradation of a number of proteins including Bcl-2 and PML 8, 9, 10, thereby overcoming drug resistance even in the relapsed cancers. Also, As2O3 might target mitochondria to induce the opening of mitochondrial permeability transition pore (mPTP) 11, 12, thus increasing intracellular reactive oxygen species (ROS) generation, which could then lead to Bax conformational changes and its translocation 13. These changes in mitochondrial physiology would result in cytochrome c (cyto c) release and subsequent activation of the apoptotic cascade.

It is known that the intracellular redox system modulates both the anti-proliferative and pro-apoptotic effects of As2O3, and that the content of cellular GSH is negatively related to the sensitivity of tumor cells to As2O3 14, 15, 16. The redox state in mammalian cells is primarily a consequence of the precise balance between the level of ROS and endogenous thiol buffers present in the cell, particularly, GSH and thioredoxin (TRX), which protect cells from oxidative damage 17. The role of GSH in protecting cancer cells from As2O3-induced apoptosis is well-documented 18. GSH binds arsenic to form a transient As (GS)3 complex, thereby preventing the inactivation of intracellular enzymes responsible for redox homeostasis. Thioredoxin-1 (TRX1) is a 12 kD ubiquitous protein having a redox-active disulfide/dithiol within the conserved active site sequence (-Cys-Gly-Pro-Cys-). Oxidized TRX1 with a disulfide on its active site is reduced by NADPH-dependent thioredoxin reductase1 (TR1) 19, 20 to restore its functions. TRX1 plays an important role in regulating cell redox homeostasis and cell growth, differentiation, and apoptosis. TRX1 inhibits cell apoptosis in a manner similar to that of Bcl-2 in the transfected cells. The mechanism for the inhibition of apoptosis by TRX1 has been suggested, whereby TRX1 could bind to and inhibit apoptosis signal-regulating kinase 1 (ASK1), thus regulating the JNK/p38 signaling pathway in response to environmental stresses such as ROS 21, 22. The dissociation between TRX1 and ASK1 induced by ROS leads to the activation of the ASK1/JNK signaling pathway, and subsequent increase of apoptosis 23, 24. Although TRX is regarded as a critical molecule for modulating cell death, its role in determining cancer cell sensitivity towards As2O3 has not been addressed. We thus attempted to investigate the role of cytoplasmic TRX1 in As2O3-induced apoptosis. Our results suggest that the redox status of TRX1 is important for modulating mitochondrial dependent apoptosis induced by As2O3. Inactivation of TRX1 either by mutation of the active site cysteine residues or by its oxidation enhances mitochondrial dependent apoptosis induced by As2O3. A better understanding of TRX1-mediated regulation of apoptosis elicited by As2O3 may offer novel targets for treating As2O3-resistant tumors and combinative drug therapy.

Results

TRX1 is involved in As2O3-induced apoptosis in HepG2 cells

We used a clinically relevant concentration of As2O3 to treat the cell line HepG2 at indicated times and then assayed apoptosis by examining Annexin V staining (data not shown) and the nuclear condensation. As2O3 induced apoptosis in a time-dependent manner in HepG2 cells (Figure 1A and 1B). The optimal effect was obtained with 5 μM As2O3, with 45% cell death after 48 h treatment. Moreover, 5 μM As2O3 induced the release of cyto c in HepG2 cells in a time-dependent manner (Figure 1C), concomitant of the onset of caspase activation (data not shown). These results suggest that As2O3 induced mitochondria-dependent apoptotic process as we described previously.

Figure 1
figure 1

As2O3 induces apoptosis and TRX1 expression in HepG2 cells. (A, B) Sensitivity of HepG2 to As2O3-mediated apoptosis. HepG2 cells were treated with 5 μM As2O3 for the indicated time, and then stained with Hoechst 33342. At least 200 apoptotic cells were counted in each experiment. Results are expressed as the mean ± SD for three independent determinations (B, panel right), *P < 0.05, compared with untreated cells. Apoptotic cells were characterized by nuclear condensation or fragmentation (A, left panel, bottom). (C) As2O3 induces cyto c release from mitochondria to cytosol in HepG2 cells in a time-dependent manner. Cells were exposed to 5 μM As2O3 for the indicated time, and then subjected to subcellular fractionation. Then, the cytosolic fraction (Cytosol) and mitochondrial fraction (Mito) were analyzed by immunoblotting (30 μg of protein per lane) with antibody specific for cyto c. (D) TRX1 expression after As2O3 treatment. Cells were treated with 5 μM As2O3 for indicated time, and then subjected to western blotting with antibodies specific for TRX1, β-actin as a loading control (panel left). The graph is the quantitative analysis of TRX1 expression with TotalLab v2.01 software (panel right).

Full size image

Human TRX1 acts as an important redox regulatory factor and plays a crucial role in drug-induced apoptosis 25. We first examined the expression of TRX1 and found that the expression of TRX1 increased in a time-dependent manner in response to As2O3 treatment (Figure 1D), suggesting that TRX1 is involved in As2O3-induced apoptosis. To confirm that TRX1 is functionally important in the sensitivity of HepG2 cells to As2O3, we performed RNA interference to specifically knock down the expression of TRX1 with oligonucleotides encoding short-hairpin RNAs (Ti108: 108-130nt and Ti214: 214-236nt). We found that Ti214 could effectively down-regulate the expression of endogenous TRX1 (Figure 2A). Treatment of Ti214-transfected HepG2 cells with As2O3 resulted in a significant increase of apoptosis compared with untreated control cells, although extensive down-regulation of TRX1 alone appeared to increase Annexin V positive cells (Figure 2B). Down-regulation of TRX1 also augmented As2O3-induced cyto c release (Figure 2C). These data suggest that TRX1 regulates the As2O3-induced apoptosis in HepG2 cells, perhaps via regulation of mitochondrial cyto c release.

Figure 2
figure 2

Vector-based RNAi specifically knocks down endogenous TRX1 expression and increases the cyto c release and apoptosis induced by As2O3. (A) HepG2 cells were transiently transfected with RNAi expression vector pSilencer-sihTRX1 (Ti108, Ti214, 1 μg) or control vector pSilencer-2.1-U6 (1 μg). After 48 h post-transfection, endogenous TRX1 expression was determined by western blotting analysis with rabbit anti-human TRX1 polyclonal antibody, β-actin as a loading control. (B) HepG2 cells were transiently transfected with pSilencer-Ti214 (1 μg) or pSilencer-2.1-U6 (1 μg), respectively. After 48 h, flow cytometric analysis of apoptosis in transfected cells treated in the presence or absence of 5 μM As2O3 for 15 h is performed by binding of Annexin V and uptake of PI. Apoptotic cells were defined as Annexin V positive. Results are expressed as mean ± SD for three independent experiments. **P < 0.01 in comparison to As2O3-treated cells with pSilencer-2.1-U6. (C) The down-regulation of TRX1 in HepG2 cells enhances the cyto c release. After 48 h, cells were exposed to 5 μM As2O3 for 15 h and subjected to subcellular fractionation. Then, the cytosolic fraction (Cytosol) was analyzed by immunoblotting (30 μg of protein per lane) with antibody specific for cyto c (upper panel); β-actin was used as a protein loading control (lower panel). Results were quantified as the ratio of cyto c to β-actin. Mean ± SD of three independent experiments was shown. **P < 0.01 in comparison to cells transfected with pSilencer-2.1-U6; #denotes P < 0.001 in comparison to cells transfected with pSilencer-2.1-U6; ##denotes P < 0.001 in comparison to cells transfected with pSilencer-2.1-U6 and treated with 5 μM As2O3.

Full size image

TRX1 over-expression inhibits As2O3-induced apoptosis in HepG2 cells

TRX1 over-expression is observed in many human primary cancers and appears to contribute to increased cell growth and a resistance to chemotherapy 26. To further examine the effects of TRX1 on As2O3-induced apoptosis, we constructed recombinant adenoviruses expressing the wild-type TRX1 and mutant TRX1 (cTmC32/35S) in which cysteine residues in the active site were changed to serine residues (see Figure 3A). We then analyzed the sensitivity of transfected HepG2 cells to As2O3-induced apoptosis and found that over-expression of wild-type TRX1 could inhibit As2O3-induced apoptosis (Figure 3B). After As2O3 treatment for 24 h, the population of apoptotic cells in AdcTRX1-infected HepG2 cells was reduced than that in AdGFP-infected cells or AdcTmC32/35S-infected cells. To our surprise, the mutant form of TRX1 lost this protective effect, and over 55% cells were apoptotic in AdcTmC32/35S-infected cells after As2O3 treatment for 12 h. These results indicate that the enzymatic activity of TRX1 was involved in susceptibility of HepG2 cells to As2O3-induced apoptosis. Moreover, we noticed that AdcTmC32/35S alone could induce apoptosis in HepG2 cells. Since it had been well established that cyto c release was a crucial event in As2O3-induced apoptosis in APL cell lines 27, 28, we measured cyto c release and found that TRX1 could prevent cyto c release induced by As2O3, whereas TmC32/35S lost its protective effect. Actually, the mutant form of TRX1 enhanced the cyto c release from mitochondria (Figure 3C).

Figure 3
figure 3

TRX1 protects HepG2 cells against the As2O3-induced apoptosis through inhibiting cyto c release and caspase activation in an active-site dependent manner. (A) Diagram of wild-type, active site mutant of TRX1 with the c-terminal Myc tag represented as fancy diagonal boxes. The amino acid sequences for active site are enlarged. (B) Fluorescence micrographs of HepG2 cells stained with Hoechst 33342. Cells infected with the recombinant adenovirus particles were treated with or without 5 μM As2O3 for 24 h. Cells were stained with Hoechst 33342 and numbered by fluorescence microscope. The apoptotic cells are characterized by nuclear condensation or fragmentation (left panel). The graph showed the percentage of apoptotic cells (n = 3, mean ± SD). *P < 0.05 in comparison to cells infected with AdGFP or AdcTRX1 which were treated with As2O3 for 12 h; **P < 0.01 in comparison to cells infected with AdcTRX1 and 24 h treatment with As2O3. (C) HepG2 cells infected with the different recombinant adenovirus particles were treated with 5 μM As2O3 for 24 h before subcellular fractionation. The cytosolic fraction (Cytosol) was analyzed by immunoblotting (30 μg of protein per lane) with antibodies specific for cyto c, and using β-actin as a protein loading control. (D) Cells were transiently transfected with pCDNA3.1 (vector), pCDNA3.1-TRX1, and pCDNA3.1-TmC32/35S, respectively. After 48 h, the cells were collected and subjected to detection of TRX1 and Tm C32/35S expression with the antibody specific for Myc tag, β-actin as a loading control. (E) Cells transfected with recombinant plasmids were treated with 5 μM As2O3 for 24 h, then followed by incubation with FITC-VAD-fmk, a FITC-conjugated inhibitory substrate of caspase, and caspase activity analyzed by flow cytometry. The graph showed the statistical analysis results. *P < 0.05 in comparison to cells transfected with vector and treated with As2O3 for 24 h.

Full size image

We next transiently transfected HepG2 cells with pcDNA3.1/Myc-His, recombinant pCDNA3.1-TRX1 and pCDNA3.1-TmC32/35S, and found that over-expression of wild-type TRX1 inhibited As2O3-induced caspase activation. In sharp contrast, over-expression of the mutant form TmC32/35S enhanced caspase activation (Figure 3E). Taken together, these studies suggest that TRX1 regulates As2O3-induced apoptosis, at least in part, by a mechanism that involves its active site.

TRX1 prevents mitochondria-related apoptotic changes induced by As2O3, while TRX1 mutant promotes the changes in vitro

We next investigated the molecular mechanisms of how TRX1 regulates As2O3-induced apoptosis in its active site-dependent fashion. The recombinant human TRX1 (rTRX1) and its mutant (rTmC32/35S) protein were expressed and purified, and these proteins were then used to treat the isolated mitochondria from mouse liver in the presence or absence of As2O3. In most cases 5 μM As2O3 was used, but in certain experiments, as indicated, to avoid potential damage to the mitochondria due to the prolonged exposure, higher dose (40 μM) and shorter duration of treatments were applied. As expected, we found that rTRX1 could potently protect isolated mitochondria from swelling, an indicator of mPTP opening, induced by As2O3 (Figure 4A, traces c and d). As2O3 elicits the disruption of mitochondrial membrane potential, which is inhibited by rTRX1 (Figure 4B, trace d), similar to that of the PTP inhibitor, cyclosporine A (CsA) (Figure 4B, trace c). Surprisingly, rTmC32/35S alone could elicit mitochondrial swelling in a dose-dependent manner (Figure 4C, traces c, d, and e), and disrupt the mitochondrial membrane potential (Δψm) in a dose-dependent manner (Figure 4C, traces b, c, and d). These effects were prevented by CsA, suggesting that rTmC32/35S could target the mPTP to evoke collapse of mitochondrial membrane potential. Our western blotting analysis showed that rTRX1 inhibited the cyto c release induced by As2O3 (Figure 4E, panel right), whereas rTmC32/35S enhanced cyto c release in a dose-dependent manner (Figure 4E, panel left). Interestingly, rTmC32/35S alone, but not other mutant forms (C62S, C69S, and C73S, data not shown), induced cyto c release. These results further suggest that wild-type TRX1 may negatively regulate As2O3-induced apoptosis through the mitochondrial apoptotic pathway. Mutation of two cysteine residues to serine alters TRX1 function as assayed in mitochondria, suggesting that the redox activity of TRX1 is important for the regulation of As2O3-induced apoptosis.

Figure 4
figure 4

Purified rTRX1 regulates the apoptotic changes on the mitochondria level in an active site-dependent manner. Mitochondria swelling and Δψm disruption induced by As2O3 were inhibited by PTP inhibitor and rTRX1. Isolated mitochondria (0.1 mg protein/ml) were incubated at 25 oC in 3 ml PT-1 medium and were pretreated with different concentrations of rTRX1. (A) Trace a is control, CaCl2 (added at 1 200 s) as positive control for mitochondrial swelling; Trace b is 40 μM As2O3; Trace c is 2 μg/ml rTRX1 plus 40 μM As2O3; Trace d is 4 μg/ml rTRX1 plus 40 μM As2O3. (B) Trace a is control, CCCP (added at 1 500 s) as positive control for the loss of mitochondrial membrane potential; Trace b is 40 μM As2O3; Trace c, 5 μM CsA plus 40 μM As2O3; Trace d, 4 μg/ml rTRX1 protein plus 40 μM As2O3. (C) Trace a, Control, CaCl2 (added at 1800 s); Trace b, 8 μg/ml rTRX1; Trace c, 4 μg/ml rTmC32/35S; Trace d, 8 μg/ml rTmC32/35S; Trace e, 16 μg/ml rTmC32/35S; Trace f, 40 μM As2O3. (D) Trace a, control, CCCP (added at 1600 s) as positive control; Trace b, 4 μg/ml rTmC32/35S; Trace c, 8 μg/ml rTmC32/35S; Trace d, 16 μg/ml rTmC32/35S. Mitochondrial swelling (A, C) and Δψm (B, D) were monitored by the decrease of 90° light scatter at 520 nm using a Jobin-Yvon FluoroMax-2 spectrofluorimeter or assessed by measuring the fluorescence intensity of the membrane potential dependent dye rhodamine 123 (30 nM) using a Jobin-Yvon FluoroMax-2 (excitation λ = 505 nm and emission λ = 534 nm), respectively. (E) Wild-type rTRX1 inhibited cyto c release from mitochondria induced by As2O3, rTmC32/35S enhanced the inductive effect of As2O3. Isolated mitochondria (1 mg protein/ml) were incubated for 60 min in 50 μl PT-1 medium with different concentrations of rTmC32/35S as indicated (panel left), isolated mitochondria were treated with As2O3, the combination of As2O3 and rTmC32/35S or rTRX1, CaCl2 (0.4 μM) used as a positive control (panel right). The samples were then centrifuged at 12 000×g for 15 min at 4 °C. The levels of cyto c were measured by western blotting with an anti-cyto c antibody. Equal protein loading was confirmed by western blotting using the COX-IV antibody in the mitochondrial pellets.

Full size image

Recombinant TmC32/35S enhances ROS production and NAD(P)H oxidation in mitochondria

We next addressed how rTmC32/35S promotes mitochondrial changes. One possibility is that rTmC32/35S could translocate to mitochondria and perturb mitochondrial physiology, leading to mitochondrial dysfunction. It is well known that mitochondria are the major sources of superoxide and hydrogen peroxide in cells. Hence we examined whether rTRX1 and rTmC32/35S could target mitochondria. Following the incubation of isolated mitochondria with recombinant TRX1 and rTmC32/35S, both rTRX1 and rTmC32/35S could be detected in mitochondrial membrane fractions (Figure 5A). We further examined the effects of rTmC32/35S on ROS production in isolated mitochondria. We used dischlorofluorescein (DCFH) to measure ROS production in mitochondrial matrix and found that rTmC32/35S increased the rate of ROS generation in isolated mitochondria, which was inhibited by recombinant Bcl-xL protein (Figure 5B) and CsA (data not shown). The levels of ROS were measured in the presence of succinate, so it is possible that ROS were produced by complex III in the respiratory chain, where the ratio of the reduced form versus oxidized form of NAD(P)H is an indicator of the redox potential in mitochondria. We thus examined the effects of rTmC32/35S on oxidation of NAD(P)H and found that rTmC32/35S induced NAD(P)H oxidation in mitochondria. Interestingly, we found that Bcl-xL potently inhibited the oxidation reaction (Figure 5C). These results suggest that rTmC32/35S may promote apoptosis through changing the mitochondrial redox balance, resulting in mPTP opening and cyto c release.

Figure 5
figure 5

Recombinant TmC32/35S enhances ROS production and NAD(P)H oxidation in mitochondria. (A) Membrane insertion of wild and mutant his-tagged rTRX1 onto mitochondria. Purified wild or mutant forms of rTRX1 were incubated with the isolated mouse liver mitochondria in respiration buffer for 30 min on ice, then centrifuged at 10 000×g for 15 min at 4 °C. Mitochondria pellets were washed with respiration buffer for three times, and then subjected to western blot analysis using anti-his tag monoclonal antibody. (B) rTmC32/35S induces an increase of mitochondrial ROS. ROS level was evaluated as described in Materials and Methods. Trace a is control; Trace b, 16 μg/ml rTmC32/35S was added at indicated time (arrow, 200 s); Trace c, Bcl-xL (20 μg/ml) was pre-incubated with the mitochondria before rTmC32/35S (16 μg/ml) was added as indicated. (C) The rTmC32/35S induces oxidation of NAD(P)H in mitochondria. The oxidation of NAD (P) H in mitochondria was measured as described in Materials and Methods. Trace a is control; Trace c, rTmC32/35S was added at indicated time (arrow, 200 s); Trace b, Bcl-xL (20 μg/ml) was preincubated with mitochondria before rTmC32/35S (16 μg/ml) was added (arrow).

Full size image

Oxidation of TRX1 abolishes the inhibitory effects of TRX1 on As2O3-induced apoptosis

Our above results suggest that the redox status of TRX1 is important to its function. It had been reported that, in the presence of oxidant, the active site loop forms part of the dimer interface and the activity of TRX1 may thus be blocked 29. To confirm this, firstly, we used diamide as an oxidant to treat purified rTRX1 and observed the existent forms of rTRX1 through non-reducing SDS-PAGE electrophoresis. Diamide induced the oligomerization of rTRX1 in a dose-dependent manner (Figure 6A). Moreover, we found that the oligomerization of rTRX1 not only altered its redox status (Figure. 6B), resulting in a decrease of the reduced form of rTRX1, but also led to the loss of its ability to inhibit As2O3-induced swelling in isolated mitochondrial (Figure 6C, trace c), and the failure to protect against As2O3-induced mitochondrial cyto c release (Figure 6D). These results further support our notion that the regulation of As2O3-induced apoptosis by TRX1 is dependent on its active site. An inactive form elicited by oxidant fails to protect mitochondria from swelling and cyto c release initiated by As2O3.

Figure 6
figure 6

rTRX1 oxidization elicited by diamide fails to protect against As2O3-induced mitochondrial swelling and cyto c release. (A) Purified rTRX1 was treated with diamide as the indicated concentration, and then subjected to SDS-PAGE under non-reducing condition and stained with the Coomassie brilliant blue. (B) Determination of the redox status of rTRX1 56. Briefly, purified rTRX1 protein (1 μg/ml) was treated with or without 1 mM diamide, and was directly precipitated in 10% trichloroacetic acid (TCA). Precipitates were washed in 70% acetone and resuspended in 40 μl of reaction buffer (80 mM Tris-HCl, pH 6.8, 2% SDS) supplemented with a cocktail of protease inhibitors (Roche) and 15 mM freshly prepared AMS. Samples were incubated for 30 min at RT and for 10 min at 37 °C and then resolved by 15% SDS-PAGE. The redox state of TRX1 was detected by western blotting with anti-TRX1 antibody. (C) Isolated mitochondria were pre-incubated with 4 μg/ml oxidized rTRX1, then As2O3 was added at the indicated time (200 s). Trace a, Control (CaCl2 added at 1 800 s); Trace b, 40 μM As2O3; Trace c, 8 μg/ml oxidized rTRX1 plus 40 μM As2O3; Trace d, 4 μg/ml rTRX1 plus 40 μM As2O3. Mitochondrial swelling was monitored as described in Materials and Methods. Data are representative of three independent experiments. (D) Isolated mitochondria (1 mg protein/ml) were pre-incubated with the oxidized rTRX1 or rTRX1, and then treated with the indicated concentration of As2O3. The release of cyto c was measured by western blotting with an anti-cyto c antibody, COX-IV in the mitochondrial pellets as a loading control.

Full size image

2,4-Dinitrochlorobenzene (DNCB) elicits the oxidization of TRX1, induces apoptosis and increases the sensitivity of HepG2 cells to As2O3-induced apoptosis

Thioredoxin reductase (TR) is important for maintaining the reduced form of TRX1 in cells, and the inhibition of TR prevents the conversion of oxidized TRX1 into reduced form 30, 31. We found that As2O3 could inhibit TR activity (data not shown), which is consistent with a previous report 32. We treated HepG2 cells with As2O3, DNCB (a specific inhibitor of TRX reductase) alone or the combination of both, then analyzed the redox status of TRX1 through non-reducing SDS-PAGE electrophoresis. Our results showed that TRX1 was oxidized by As2O3 and DNCB treatments in a time-dependent manner. The combination of As2O3 and DNCB induced significant increase in the oxidized form of TRX1 (Figure 7A). To further confirm whether the oxidization of TRX1 enhances As2O3-induced apoptosis in HepG2 cells, we used a clinically relevant concentration of As2O3 (5 μM) to treat HepG2 cells for 12 and 24 h in the presence or absence of DNCB, and then measured apoptosis with Annexin V and propidium iodide (PI) staining by flow cytometry. Our results showed that both As2O3 and DNCB induced an increase of Annexin V positive HepG2 cells compared with control cells. The combination of both led to the greatest increase (11.5-fold) compared with control cells (Figure 7B). Moreover, the combination of As2O3 and DNCB significantly increased cyto c release from mitochondria (Figure 7C), consistent with caspase activation in a time-dependent manner (Figure 7D). These data indicate that the inhibition of TR activity by DNCB results in the oxidization of TRX1, and subsequently facilitates As2O3-induced apoptosis.

Figure 7
figure 7

The alteration of TRX1 redox status increases the sensitivity of HepG2 cells to As2O3-induced apoptosis. (A) As2O3 and DNCB could elicit the oxidation of TRX1. HepG2 cells were treated with or without indicated concentrations of DNCB and As2O3 or the combination of both for the indicated time, then were directly precipitated in 10% TCA. Precipitates were resuspended in 40 μl reaction buffer and modified with 15 mM AMS at RT for 30 min, and then incubated at 37 °C for 10 min after washed twice in 70% acetone. Samples were resolved by 15% SDS-PAGE and detected by western blotting with TRX1 polyclonal antibody. (B) HepG2 cells were treated with 5 μM As2O3 and 15 μM DNCB respectively or combinatively for 0, 12, 24 h; apoptotic cells were defined as Annexin V positive. The graph shows the statistical analysis results (n = 3, mean ± SD). **P < 0.01 in comparison with untreated cells, *P < 0.05 in comparison to cells treated with As2O3 or DNCB for 24 h. (C) HepG2 cells were treated as indicated in B, and subjected to subcellular fractionation. Then, the cytosolic fraction (Cytosol) was analyzed by immunoblotting (30 μg of protein per lane) with antibodies specific for cyto c, β-actin as a loading control. (D) The caspase activity was detected as described in Materials and Methods. HepG2 cells were treated as indicated, and then followed by incubation with FITC-VAD-fmk, a FITC-conjugated inhibitory substrate of caspase, and analyzed by flow cytometry. The graph showed the statistical analysis results (n = 3, mean ± SD). **P < 0.01 in comparison to untreated cells; *P < 0.05 in comparison to cells treated with As2O3 or DNCB for 24 h.

Full size image

Discussion

In this study, we showed the critical role of TRX1 in regulating As2O3-induced apoptosis. First, the expression level of TRX1 in cells is important for regulating cell death and drug sensitivity. Increased expression of wild-type TRX1 prevents HepG2 cells from As2O3-induced apoptosis, while down-regulation of TRX1 with RNAi sensitizes the cells towards As2O3-induced apoptosis. Second, we found that the redox status of TRX1 determines the sensitivity of HepG2 cells to As2O3 treatment. Mutation of the active site cysteines in TRX1 enhanced As2O3-induced apoptosis. In vitro oxidization of TRX1 also caused a lost of protective function. To further support this notion, we found that DNCB, a specific inhibitor of TR, could induce TRX1 oxidation and significantly sensitized cells to As2O3-induced apoptosis. Previous studies have shown that arsenic can react with cysteine residues of GSH or TR to inactivate intracellular enzymes responsible for redox homeostasis 33, 34. Our results also suggest a novel mechanism for sensitizing cancer cells to As2O3-induced apoptosis by modulating TRX1 redox status. This can be achieved either by inhibiting TR activity (Figure 7) or by direct reactions with cysteine residues within the TRX1 active site.

The redox status of TRX1 could also affect its interaction with ASK1, or other important molecules which mediate cell apoptosis 35. However, our findings of this study suggest that mutation of the TRX1 active site cysteines could directly elicit destabilization of mitochondrial bioenergetics as well as cyto c release in the cells and isolated mitochondria, while wild-type TRX1 could prevent the apoptotic effects induced by both mutant TRX1 and As2O3. Direct application of the mutant form of TRX1 (C32S/35S), but not other mutant forms, to isolated mitochondria elicited the opening of mPTP, the loss of Δψm, and subsequently the release of cyto c. Similarly, the oxidized form of TRX1 lost its function as an anti-apoptosis molecule and was able to enhance the apoptotic catastrophe of mitochondria. These effects appear to be direct in our in vitro system, without the involvement of cytosolic factors or signaling pathways as reported previously. Since TRX1 acts as one of the key components in the antioxidant system which antagonizes oxidative stress-induced apoptosis 35, 36, our work supports that inactivation of TRX1 would confer a therapeutic advantage to cancer treatment by targeting mitochondria.

Questions remain regarding how inactivation or mutation of TRX1 directly elicits a mitochondria-related apoptotic response. We found that mutant TRX1 could disturb the redox potential of mitochondria as it caused the oxidation of mitochondrial NADH/NADPH and increased ROS production. This may subsequently perturb the mitochondrial redox pool of GSH and mtTRX2 (mitochondrial thioredoxin-2 (TRX2)), which are substrates of mitochondrial glutathione peroxidase and thioredoxin peroxidase, and thus decrease the capacity of these enzymes to remove H2O2. The perturbation of the redox pool may then globally modulate the mitochondria physiology, resulting in the opening of mPTP and increase of ROS production 37. In addition, it has also been shown that cyto c that is released from mitochondria is a potent catalyst of DCFH oxidation 38, and DCF fluorescence observed in the presence of rTmC32S/35S may be a reflection of increased cyto c content in the incubation buffer. This is unlikely in our system since this dye measured the ROS in the mitochondrial matrix. To support these notions, we found that DTT, NADH, and ROS scavengers prevented rTmC32S/35S-induced mitochondria-related apoptotic changes (unpublished data). Another possibility is that the insertion of recombinant TRX1 proteins into the mitochondria and interactions with VDAC1 or other molecules modulate PTP activity for cyto c release. Our results were consistent with a previous report which proved the interaction between TRX1 and VDAC1 39, since our pull-down assay showed that TRX1 could interact with VDAC1 in our system (data not shown). One report indicates that mitochondrial TRX2 might directly regulate cyto c release through the physical interaction with cyto c in mitochondria 40. It remains to be determined if TRX1 can interact with cyto c.

In summary, our results indicate that TRX1 regulates As2O3-induced apoptosis by preventing mitochondrial cyto c release. Inactivation of TRX1 either by mutation of the active cysteine residues or by oxidation enhances mitochondria-dependent apoptosis induced by As2O3. Further studies are required to develop strategies to enhance the effect of As2O3 therapy by modulating TRX1 activity in vivo.

Materials and Methods

Materials

Arsenic trioxide (A-1010) was purchased from Sigma (St. Louis, MO); 4-acetamido-4′-maleimidylstilbene-2,2′-disulfonic acid, disodium salt (AMS) (A-485), and anti-cyto c oxidase monoclonal antibody (subunit IV) were from Molecular Probes (Eugene, OR). ProBondTM metal affinity Resin was purchased from Invitrogen (San Diego, CA), and Amicon®Ultra-15 concentrator was from Millipore (Billerica, MA). Anti-cyto c mouse monoclonal antibody was from BD Biosciences Pharmingen (San Jose, CA). Rabbit anti-human thioredoxin polyclonal IgG1 (sc-20146) was from Santa Cruz Biotechnology (Santa Cruz, CA). All other chemicals were purchased from Sigma (St. Louis, MO).

Cell culture and transfection

Human embryonic kidney 293A (HEK293A) cells, purchased from the Cell Bank of the Chinese Academic of Sciences (Shanghai, China), were maintained in Dulbecco's modified Eagle's medium (GIBCO); human hepatoma HepG2 cells (from ATCC) were maintained in 1640 medium, both supplemented with 10% (v/v) fetal bovine serum (HyClone), 100 μg/ml streptomycin and 100 U/ml penicillin (HyClone) at 37 °C in humidified 5% CO2 atmosphere. Cell transfection was performed as described in the LIPOFECAMINE™ Reagent (GIBCO) manual.

Detection of cell death by Hoechst 33342 and Annexin-V-FITC

For Hoechst 33342 staining, cells were plated at low density on glass coverslides in a six-well plate and treated with 5 μM As2O3. After indicated time, cells were stained with Hoechst 33342 in PBS (15 min at room temperature in the dark). Cells were then washed three times with PBS and analyzed under a fluorescence microscope, and at least 200 cells were counted. Apoptotic cells were detected by flow cytometry. Phosphatidylserine exposed on the outside of the cells was determined by Annexin V-FITC kit. Briefly, HepG2 cells were plated in six-well plates and treated with As2O3 for the indicated time. At the end of the experiment, cells were digested with trypsin-EDTA solution, then collected by centrifugation and washed twice with ice-cold PBS. After being washed one more time with the incubation buffer (10 mM HEPES/NaOH, pH 7.4, 140 mM NaCl, and 5 mM CaCl2), cells were stained with Annexin V-FITC and PI for 20 min in dark at room temperature; then an additional 400 μl binding buffer was added before FACS analysis 41. Flow cytometric analysis was performed to monitor the green fluorescence of the FITC-conjugated Annexin V (530±30 nm) and the red fluorescence of DNA-bound PI (630±22 nm). All data were analyzed with Cell Quest software (BD).

SDS-PAGE and immunoblotting

SDS-PAGE and immunoblotting were performed as described elsewhere 42. Briefly, the cells or membrane fractions were resuspended in NP-40 containing lysis buffer (10 mM HEPES (pH 7.4), 2 mM EGTA, 0.5% NP-40, 1 mM NaF, 1 mM NaVO4, 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM DTT, 50 mg/ml trypsin inhibitor, 10 mg/ml aprotinin, and leupeptin) and placed on ice for 30 min. The lysates were centrifuged at 12 000×g for 12 min at 4 °C, and the protein concentration was determined with BSA as a standard. Equivalent samples (30 μg protein) were subjected to SDS-PAGE on 12% gel. The proteins were then transferred onto nitrocellulose membranes, and probed with the indicated antibodies followed by the appropriate secondary antibodies conjugated to horseradish peroxidase (KPL, Gaithersburg, MD, USA). Immunoreactive bands were visualized using enhanced chemiluminescence (Pierce). The molecular sizes of the developed proteins were determined by comparison with prestained protein markers (Invitrogen, Carlsbad, CA, USA).

Cell fractionation

As2O3-treated cells were fractionated by differential centrifugation as described previously 43, 44, 45. Briefly, cells were harvested through trypsin digestion, and then centrifuged and resuspended in three volumes of hypotonic buffer (210 mM sucrose, 70 mM mannitol, 10 mM HEPES (pH 7.4), 1 mM EDTA) containing 1 mM PMSF, 50 mg/ml trypsin inhibitor, 10 mg/ml leupeptin, 5 mg/ml aprotinin and 10 mg/ml pepstatin. After gentle homogenization with a Dounce homogenizer, cell lysates were centrifuged at 1 000×g for 5 min to remove unbroken cells and nuclei and the cytosolic fractions were obtained by further centrifugation at 10 000×g for 30 min.

Generation of adenoviral recombinants and plasmids construction

The adenoviral recombinants of Ad-cTRX1 and Ad-cTmC32/35S were generated according to previous methods 46. Briefly, PCR products (Bgl II + Not I) of TRX1 and its mutant were subcloned into pAdTrack-CMV shuttle vector. Recombinant adenoviral plasmids were generated by homologous recombination of recombinant shuttle vector linearized with PmeI and AdEasy-1 in BJ5183 Escherichia coli cells. Then, the recombinant adenovirus plasmids were linearized with PacI and then transfected into HEK293A cells to package recombinant virus. Virus propagation and amplification were then achieved by standard techniques.

To generate the over-expression of cTRX1 and cTmC32/35S vector, their cDNA fragments were subcloned into pCDNA3.1/myc-his vector and were confirmed by sequencing. In addition, to generate TRX1 RNAi vector, one annealed set of oligonucleotides encoding short-hairpin transcripts corresponding to nt 108-130 (named Ti108) and 214-236 (named Ti214) of human TRX1 mRNA (GenBank accession no. AF276919) were cloned into pSilencer-2.1-U6 (Ambion; hereafter abbreviated to pSilencer). In brief, the short-hairpin-RNA-encoding complementary single-stranded oligonucleotides, which hybridized to give overhangs compatible with BamHI and HindIII, were designed with a computer program available on the internet http://www.ambion.com/techlib/misc/psilencer_converter.html. Oligonucleotides encoding short-hairpin RNAs were then ligated into pSilencer. The positive clones were confirmed by sequencing.

Expression and purification of recombinant proteins

The recombinant proteins were expressed and purified as described previously 47. Briefly, His-tagged human TRX1 and TmC32/35S were obtained through subcloning into the pET-28a (+). The resulting plasmids were transformed into E. coli BL21 (DE3). The positive transformants were induced for 4 h at 25 °C by the addition of 1 mM isopropyl-b-D-thiogalactopyranoside (IPTG; Sigma). The induced cells were pelleted, resuspended in native binding buffer (20 mM sodium phosphate, 500 mM sodium chloride, pH 7.8), sonicated on ice, and centrifuged at 4 °C for 30 min at 30 000×g to remove cell debris. The supernatants containing fusion proteins were purified with ProBondTM metal affinity resin and concentrated with Amicon® Ultra-15 concentrators. The protein purity was verified by 12% SDS-PAGE and stained with Coomassie brillant blue R-250. The protein concentration was determined with BSA as a standard.

Isolation of mouse liver mitochondria and measurement of PTP opening and mitochondrial membrane potential (Δψm)

Liver mitochondria from Balb/c mice were isolated by method as described previously 48, 49, 50. Briefly, mice livers were homogenized with a glass-Teflon Potter homogenizer. Samples were centrifuged at 4 °C, 1 000×g for 10 min. Then the supernatant was transferred to another tube and centrifuged at 4 °C, 10 000×g for 10 min. Mitochondria were washed twice, then resuspended in the same medium. Isolated mitochondria (5 mg protein/ml) were kept in MT buffer containing 250 mM sucrose, 2 mM HEPES, pH 7.4, 0.1 mM EDTA, and 0.1% fatty acid-free BSA. PTP opening was monitored by the decrease of 90° light scattering at 520 nm at 25 °C in medium PT-1 containing 250 mM sucrose, 2 mM HEPES, pH 7.4, 0.5 mM KH2PO4, 2 mM rotenone, and 4.2 mM potassium succinate, using a Jobin Yvon FluoroMax-2 spectrofluorimeter as described 51.

In addition, determination of mitochondrial membrane potential (Δψm) was performed as described previously 52, 53. Briefly, after addition of 30 nM Rh123 to a mitochondria suspension, Dym was assessed at 25 °C in medium PT-1 by measuring the uptake of Rh123 using a spectrofluorimeter (Jobin Yvon FluoroMax-2).

Assay of ROS production and determination of the NAD(P)H redox state

The generation of mitochondrial ROS was evaluated in isolated mitochondria in medium PT-1 using DCFH as a probe as described previously 54. DCF formation was monitored using a Jobin Yvon FluoroMax-2 spectrofluorimeter. The oxidation of NAD(P)H in the mitochondria was measured at excitation and emission wavelengths of 350 and 450 nm as described 55. All assays were carried out in triplicate.

Cyto c release and western blot analysis

Mouse liver mitochondria were isolated following the above method and protein content of isolated mitochondria was determined by the micro-biuret method using BSA as a standard. Equal mitochondria fractions were treated with different concentrations of rTmC32/35S at 25 °C in PT-1 buffer for 60 min; in addition, 12 μg/ml rTRX1 was pre-incubated with mitochondria for 10 min, then 12 μg/ml rTmC32/35S and 40 μM arsenic trioxide was added and incubated for the same time. The samples were then centrifuged at 12 000×g for 15 min at 4 °C. Cyto c resulting from the supernatant was detected by western blotting using anti-cyto c monoclonal antibody and visualized by ECL Supersignal system (Pierce), and equal protein loading was confirmed by immunodetection of cyto c oxidase subunit IV (COX-IV) in the mitochondrial pellets. The sample treated with 0.4 μM CaCl2 was a positive control as indicated.

Detection of caspase activation in situ

The method was performed following the CaspACETM FITC-VAD-FMK in situ Marker kit manual. Briefly, cells were digested with trypsin-EDTA solution, then collected and washed with PBS. CaspACETM FITC-VAD-FMK In situ Marker was added to the cells at a final concentration of 10 μM, and then incubated for 20 min in the dark. Cells were washed three times with PBS, and resuspended in 400 μl PBS and measured with a FACScan.

Statistical analysis

Significant differences between values under different experimental conditions were determined by paired Student t test analyses. A value of P <0.05 was considered to be significant.