|Home | About | Journals | Submit | Contact Us | Français|
The Nrf2-Keap1 signaling pathway is a protective mechanism promoting cell survival. Activation of the Nrf2 pathway by natural compounds has been proven to be an effective strategy for chemoprevention. Interestingly, a cancer-promoting function of Nrf2 has recently been observed in many types of tumors due to deregulation of the Nrf2-Keap1 axis, which leads to constitutive activation of Nrf2. Here, we report a novel mechanism of Nrf2 activation by arsenic that is distinct from that of chemopreventive compounds. Arsenic deregulates the autophagic pathway through blockage of autophagic flux, resulting in accumulation of autophagosomes and sequestration of p62, Keap1, and LC3. Thus, arsenic activates Nrf2 through a noncanonical mechanism (p62 dependent), leading to a chronic, sustained activation of Nrf2. In contrast, activation of Nrf2 by sulforaphane (SF) and tert-butylhydroquinone (tBHQ) depends upon Keap1-C151 and not p62 (the canonical mechanism). More importantly, SF and tBHQ do not have any effect on autophagy. In fact, SF and tBHQ alleviate arsenic-mediated deregulation of autophagy. Collectively, these findings provide evidence that arsenic causes prolonged activation of Nrf2 through autophagy dysfunction, possibly providing a scenario similar to that of constitutive activation of Nrf2 found in certain human cancers. This may represent a previously unrecognized mechanism underlying arsenic toxicity and carcinogenicity in humans.
The Nrf2-Keap1 pathway is the main cellular defense mechanism against oxidative and electrophilic stresses (1–3). Nrf2 is a ubiquitously expressed transcription factor that regulates the expression of genes bearing an antioxidant response element (ARE) in the promoter region (4). These genes include antioxidant proteins, phase II detoxification enzymes, transport proteins, proteasome subunits, chaperones, growth factors, receptors, and other transcription factors (4, 5). Under normal, unstressed conditions, Keap1 tightly regulates Nrf2. Keap1 is a substrate adaptor protein that associates with cullin 3 (Cul3) and Rbx1 to form an E3 ubiquitin ligase complex to ubiquitinate Nrf2, which is then degraded by the 26S proteasome to maintain low basal Nrf2 levels (6, 7). Critical cysteine residues in Keap1 act as regulatory sensors and are modified upon exposure to oxidative or electrophilic stress (8–10). It has been hypothesized that modification of critical cysteine residues causes a conformational change of the Keap1-Cul3-E3 ubiquitin ligase complex and prevents Nrf2 from being ubiquitinated. Structurally diverse small molecules, such as sulforaphane (SF), tert-butylhydroquinone (tBHQ), cinnamaldehyde, and resveratrol, as well as environmental toxicants, including the heavy metals arsenic, cadmium, and chromium, are able to activate the Nrf2-Keap1 pathway (10–12).
It is well accepted that activation of the Nrf2 pathway by certain organic compounds, or chemopreventive agents, can reduce the risk of cancer. The observation that Nrf2−/− mice are more susceptible to the deleterious effects of chemical toxicants and carcinogens, coupled with them being refractory to the protective actions of chemopreventive compounds, clearly demonstrates the role of Nrf2 in chemoprevention (1, 13). Therefore, Nrf2 has generally been considered to be a protective mechanism in normal cells. Paradoxically, the “dark side” of Nrf2 has recently been revealed. Genetic mutations (14–17) in either Keap1 or Nrf2 that disrupt the Keap1-mediated negative regulation of Nrf2, leading to constitutive activation of Nrf2, have been identified in many human tumors and cell lines. Epigenetic alteration of Keap1 resulting in reduced expression of Keap1 is yet another mechanism of Nrf2 overexpression in many cancer types (18–20). Furthermore, cancer cells have hijacked the Nrf2 response to protect themselves against cell death, resulting in intrinsic or acquired chemoresistance (21–25). This was experimentally confirmed when knockdown or inhibition of Nrf2 increased the sensitivity of cancer cells to chemotherapeutics (23, 26). Uncovering the dark side of Nrf2 in cancer has led us to question whether Nrf2 activation by arsenic is associated with the cancer-promoting role of Nrf2 and/or arsenic.
Chronic exposure to inorganic arsenic from contaminated drinking water has long been associated with high incidences of cancer in various organs, particularly in the skin, lung, bladder, liver, and kidney (27). Arsenic affects a multitude of biological systems; however, the mechanism by which arsenic elicits its toxic and carcinogenic effects remains largely unknown. Numerous studies have been conducted to elucidate the molecular events associated with arsenic exposure, and resulting data suggest multiple mechanisms. For instance, arsenic is able to alter DNA methylation and repair, regulate gene expression through an epigenetic mechanism, affect cell proliferation, generate reactive oxygen species (ROS), and modulate various intracellular signaling pathways (28–30). Arsenic has also been shown to induce Nrf2 and the expression of Nrf2-dependent downstream genes in a variety of cell lines, including UROtsa cells, HaCaT cells, osteoblasts, and fibroblasts (31–35). Our laboratory has demonstrated that arsenic activates the Nrf2 pathway and upregulates many of its downstream genes through a Keap1 cysteine 151 (Keap1-C151)-independent mechanism (31), which is different from well-characterized Nrf2 inducers such as tBHQ and SF (3, 7, 8, 36). These data imply that different modes of Nrf2 activation (Keap1-C151 dependent versus independent) may determine whether Nrf2 is beneficial or detrimental.
Very recently, arsenic was also shown to induce autophagy, a bulk-lysosomal degradation pathway where cytoplasmic components, misfolded proteins, damaged organelles, and specifically targeted proteins are sequestered into double-membrane vesicles called autophagosomes (37–39). Autophagy was thought to be a nonselective degradation pathway; however, selective substrate adaptor proteins, such as p62, have been shown to facilitate degradation of specific proteins through autophagy (40). p62 interacts directly with microtubule-associated protein 1 light chain 3 (LC3), a component of the autophagosomal membrane, that is cleaved (LC3-I) and lipidated (LC3-II) and can be used as a marker for autophagy (41). Not only does p62 self-oligomerize and bind ubiquitin, it also is incorporated into the autophagosome and degraded via autophagy (40–42). The role of p62 in the Nrf2-Keap1 pathway was insufficiently studied until very recently, when our group and others independently demonstrated that an increase in the level of p62, either by overexpression or deregulation of autophagy, sequesters Keap1 into autophagosomes through direct interactions (43–47). p62-mediated sequestration of Keap1 stabilized Nrf2 and increased the transcription levels of its downstream genes (43). This p62-dependent mode of Nrf2 activation was termed the “noncanonical” mechanism of Nrf2 activation (43).
In the current study, we demonstrate that arsenic blocks autophagic flux, resulting in accumulation of p62 and sequestration of Keap1 into autophagosomes. This was also confirmed in human lung epithelial cells because the lung is a primary target organ of arsenic toxicity. Furthermore, this p62-mediated sequestration of Keap1 prolonged Nrf2 activation, seemingly mimicking the dark side of Nrf2 in cancer. Consistent with the notion that autophagic flux is blocked by arsenic, an overall increase in p62 protein levels was observed following low-dose arsenic treatment. More importantly, our results show that arsenic-mediated deregulation of autophagy is attenuated by SF and tBHQ, indicating that canonical Nrf2 activators (chemopreventive compounds that activate Nrf2 in a Keap1-C151-dependent manner) may be used to protect against arsenic toxicity and carcinogenicity.
Sodium arsenite, tBHQ, insulin, transferrin, dexamethasone, trypsin inhibitor, and the antibody against LC3 were all purchased from Sigma-Aldrich. l-Sulforaphane was used in all experiments and was purchased from LKT Laboratories, Inc. Primary antibodies specifically against Nrf2, Keap1, p62, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), and β-actin as well as mouse, rabbit, and goat horseradish peroxidase (HRP)-conjugated secondary antibodies were purchased from Santa Cruz Biotechnology. Alexa Fluor 488, 594, and 680 were purchased from Invitrogen.
HEK293, NIH 3T3, and BEAS-2B cells were purchased from the American Type Culture Collection (ATCC), and enhanced green fluorescent protein (EGFP)-LC3-expressing immortalized baby mouse kidney (iBMK) cells were established in the laboratory of Eileen White (Cancer Institute of New Jersey). The 16HBE14o− (HBE) cell line was obtained from the California Pacific Medical Center, San Francisco, CA. NIH 3T3 and iBMK cell lines were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS), 1% l-glutamine, and 0.01% gentamicin. BEAS-2B cells were grown in Ham's F-12 medium (MediaTech) supplemented with 1 ml bovine hypothalamus extract (PromoCell), insulin (2 mg/ml) (Sigma), transferrin (2.5 mg/ml) (Sigma-Aldrich), dexamethasone (0.05 mM) (Sigma), cholera toxin (10 μg/ml) (List Biological Laboratories, Inc.), and epidermal growth factor (10 μg/ml) (Millipore). HBE cells were grown in modified Eagle's medium supplemented with 10% FBS, 1% l-glutamine, and 0.01% gentamicin. HEK293 cells were grown in the same medium as HBE cells and were supplemented with 0.1 mM nonessential amino acids (Cellgro) and 1 mM sodium pyruvate (Gibco). All cell lines were incubated at 37°C in a humidified incubator containing 5% CO2.
Transfection of cDNA was performed by using Lipofectamine Plus (Invitrogen), and HiPerfect was used for transfection of small interfering RNA (siRNA); both were used according to the manufacturer's instructions. p62 siRNA (SI00057596) and Keap1 siRNA (SI03246439) were purchased from Qiagen. Cells were transfected with wild-type Keap1 (Keap1-WT) and Keap1-C151S, along with plasmids encoding mouse glutathione S-transferase (mGST)–ARE–firefly luciferase and thymidine kinase (TK)-Renilla luciferase (internal control). Luciferase activities were measured by using the Promega dual-luciferase reporter assay system. The experiment was run in triplicate and repeated three times. Results are expressed as means ± standard deviations (SD). A P value of <0.05 was considered to be significant.
Cells were harvested in 1× sample buffer (50 mM Tris-HCl [pH 6.8], 2% sodium dodecyl sulfate [SDS], 10% glycerol, 100 mM dithiothreitol [DTT], 0.1% bromophenol blue) and immediately boiled for 3 to 5 min. Samples were sonicated, resolved by SDS-PAGE, and transferred onto a nitrocellulose membrane for immunoblot analysis.
For indirect immunofluorescence, cells were grown on round glass coverslips (Fisher Scientific) in 35-mm cell culture dishes. Following 20 min of fixation with prechilled methanol, coverslips were washed with phosphate-buffered saline (PBS), permeabilized with 0.2% Triton X-100–PBS for 15 min, and blocked with 2% bovine serum albumin–PBS for 30 min. Coverslips were then incubated with primary antibody for 1 h, followed by three 10-min washes in PBS. Coverslips were then incubated with secondary antibody for 50 min, followed by three 10-min washes in PBS. The coverslips were mounted onto glass slides with antifade mounting medium purchased from Invitrogen. NIH 3T3 cells were used for live-cell imaging due to the cytoplasm-to-nucleus ratio, therefore allowing the localization of proteins in the cytoplasm to be more easily distinguishable. Furthermore, the iBMK stable cell line was used to avoid induction of autophagy caused by lipid transfection of LC3 and was the only stable cell line available to assess formation of autophagosomes. For live-cell imaging, cells were grown on 35-mm glass-bottom dishes from In Vivo Scientific. Cells were gently washed once with 1× PBS and imaged in phenol red-free DMEM supplemented with 10% FBS. Images were captured with a Zeiss Observer.Z1 microscope by using the Slidebook 126.96.36.199 computer program (Intelligent Imaging Innovations, Inc.).
Electron microscopy (EM) was performed at the SWEHSC cellular imaging facility core at the University of Arizona. Briefly, HEK293 and BEAS-2B cells were treated with the indicated doses of arsenic, SF, or both for 4 h. Cells were then fixed in 2.5% glutaraldehyde in 0.1 M cacodylate buffer, followed by postfixation in 1% osmium tetroxide. Cells were then washed, pelleted, stained with 4% aqueous uranyl acetate, and dehydrated with ethanol infiltrated with Spurr's resin. Cells were incubated at 60°C to allow polymerization. Cut silver-gold sections were then mounted onto 150-mesh copper grids and stained with 2% lead acetate for 3 min. Sections were examined by using an FEI CM12 transmission electron microscope (TEM) (Philips) operated at 80 kV. Electron micrograph images were collected on an Optronics Macrofire AMT 542 digital camera at the indicated magnifications.
HEK293 cells were treated with low doses of arsenic for 4 h. Electron microscopy (EM) showed that 250 nM, 500 nM, and 1 μM arsenic induces formation of autophagosomes (seen as double-membrane vesicles) in HEK293 cells (Fig. 1). Consistent with the results of EM indicating an upregulation of autophagosomes, immortalized baby mouse kidney (iBMK) cells stably expressing EGFP-LC3 also showed an increase in the number of LC3 puncta when cells were exposed to 125 nM, 250 nM, 500 nM, or 1 μM arsenic for 4 h, which remained for up to 24 h (Fig. 2). In contrast, 1.25 μM SF did not induce LC3 puncta (Fig. 2), indicating that SF has no effect on autophagy. Although it has been well established that both arsenic and SF are Nrf2 inducers, these results indicate that only arsenic, and not SF, is able to enhance the number of autophagosomes.
An increase in the number of autophagosomes not only is indicative of autophagy induction but also could be the result of an inhibition of autophagosome clearance. To differentiate these two phenomena experimentally, a tandem mouse red fluorescent protein (mRFP)-GFP-LC3 construct was transfected into cells to determine which stage of the autophagic process is affected by arsenic. GFP fluorescence is quenched in acidic environments such as that of the lysosome or when an autophagosome fuses with the lysosome to form an autolysosome, whereas mRFP is more stable. Therefore, colocalization of both GFP and RFP fluorescence (yellow/orange puncta in the merged image) indicates an autophagosome that has not yet fused with a lysosome or where acidification of the lysosome is disrupted. In contrast, RFP alone (without GFP) corresponds to an autolysosome. Two positive controls were used to differentiate increased formation of autophagosomes (starvation) from blockage of autophagosome clearance (bafilomycin A1 [Baf A1], which has been reported to prevent fusion of autophagosomes with lysosomes or inhibit acidification of autolysosomes). NIH 3T3 cells transfected with the tandem mRFP-GFP-LC3 construct were either starved in Hanks balanced salt solution (HBSS) or treated with 100 nM Baf A1 for 4 h. Starved cells had both yellow and red puncta, indicating autophagosomes and autolysosomes, respectively (Fig. 3A). Cells treated with Baf A1, however, had mostly yellow/orange puncta (Fig. 3A). Interestingly, only yellow/orange puncta were observed in cells treated with 500 nM or 1 μM arsenic as early as 2 h and persisted for the latest time point measured, 12 h (Fig. 3B to toE).E). Cells at every time point treated with 1.25 μM SF resembled untreated cells, such that no or minimal puncta were observed (Fig. 3A to toE).E). These results support that the mechanism by which arsenic affects autophagy is similar to that of Baf A1, preventing autophagic flux either by inhibiting fusion of autophagosomes with lysosomes or by affecting acidification of autolysosomes.
Nrf2 activators, such as SF and tBHQ, have previously been reported to protect against arsenic toxicity; however, how they impose this protective mechanism has yet to be elucidated. Therefore, whether these Keap1-C151-dependent canonical Nrf2 activators were able to alleviate arsenic-mediated effects on autophagosomes was determined. Live-cell imaging of the EGFP-LC3-expressing stable iBMK cell line treated with 500 nM arsenic alone showed an increase in the number of LC3 puncta compared to untreated or SF- or tBHQ-treated cells (Fig. 4A). Interestingly, SF and tBHQ treatment in combination with arsenic showed a reversal or inhibition of arsenic-induced LC3 puncta (Fig. 4A). Consistent with these results, live-cell imaging of NIH 3T3 cells transfected with the tandem mRFP-GFP-LC3 construct and treated with either 1 μM arsenic, 1.25 μM SF, 25 μM tBHQ, or the indicated combinations also demonstrated that canonical Nrf2 activators reduced numbers of arsenic-induced RFP- and GFP-LC3 puncta (Fig. 4B). Taken together, these results indicate that canonical Nrf2 activators are able to suppress arsenic-mediated autophagic deregulation.
One of the primary target organs for arsenic exposure is the lung. In order to confirm the above-mentioned findings in a relevant in vitro model for arsenic toxicity, two types of immortalized but nontransformed lung epithelial cell lines, BEAS-2B and HBE, were utilized. EM was conducted on BEAS-2B cells left untreated or treated with 500 nM arsenic. Consistent with the results seen with HEK293 cells (48), arsenic-treated BEAS-2B cells showed an increase in the number of autophagosomes, whereas 1.25 μM SF had no effect. SF in fact reduced the number of autophagosomes when cells were treated with the combination of SF and arsenic (Fig. 5A). Furthermore, indirect immunofluorescence analysis of Keap1, LC3, and p62 in BEAS-2B and HBE cells (HBE cell data not shown) treated with 1 μM arsenic, 1.25 μM SF, or the combination was conducted. The results further verified that arsenic, and not SF, increases the formation of autophagosomes (Fig. 5B). Arsenic induced puncta that were positive for Keap1, LC3, and p62 (Fig. 5B), whereas cells treated with 1.25 μM SF were similar to untreated cells (Fig. 5B). More importantly, SF reversed the effects of arsenic by decreasing the number of Keap1-, LC3-, and p62-positive puncta (Fig. 5B). These results further suggest that canonical Nrf2 activators, such as SF, alleviate arsenic-mediated autophagosomal upregulation and prevent p62-mediated sequestration of Keap1 into autophagosomes.
Thus far, arsenic has been shown to deregulate autophagy, leading to upregulation of autophagosomes in which p62 and Keap1 accumulate. Previously, it was reported that p62 contains an ARE in its promoter region, creating a positive-feedback loop (47). Based on that report, there are two possible mechanisms mediating accumulation of p62 in autophagosomes in response to arsenic: p62 can accumulate either by activating Nrf2 to increase p62 transcription or by blocking autophagic flux, since p62 is itself a substrate of autophagy. Therefore, the p62 or Nrf2 dependence of the observed arsenic-mediated autophagosomal upregulation and the alleviation conferred by SF or tBHQ was explored by using indirect immunofluorescence analysis. BEAS-2B cells transfected with control siRNA, p62 siRNA, or Nrf2 siRNA were treated with arsenic, SF, tBHQ, or the indicated combination. Similar to nontransfected cells, arsenic, but not SF or tBHQ, induced puncta in which Keap1, LC3, and p62 colocalized in control-siRNA-transfected cells (Fig. 6A). In cells cotreated with arsenic plus SF or tBHQ, there were minimal (similar to basal) amounts of puncta comparable to those in untreated cells (Fig. 6A). However, when p62 expression was blocked by p62 siRNA, even arsenic was unable to induce puncta (Fig. 6B), indicating that arsenic-mediated autophagosomal upregulation is p62 dependent. On the other hand, arsenic still induced puncta in cells transfected with Nrf2 siRNA (Fig. 6C). More importantly, upon knockdown of Nrf2, SF and tBHQ could no longer alleviate arsenic-induced puncta (Fig. 6C). These data indicate that arsenic-mediated autophagosomal upregulation and accumulation of p62 and Keap1 in the autophagosome are due to blockage of autophagic flux. It is unlikely that arsenic first induces Nrf2 that feeds back to further enhance the transcription of p62, since upregulation/accumulation of p62 in autophagosomes was still observed when Nrf2 expression was suppressed by Nrf2 siRNA.
The notion that arsenic treatment increases the number of autophagosomes and leads to the accumulation of p62 and Keap1 in autophagosomes prompted us to test if arsenic-mediated induction of Nrf2 is p62 dependent (noncanonical mechanism). Therefore, a dual-luciferase assay was conducted to determine whether arsenic-mediated Nrf2 activation is indeed p62 and/or Keap1-C151 dependent. HBE cells were first transfected with an siRNA specific for the 3′ untranslated region of Keap1 to suppress endogenous Keap1 expression in combination with control siRNA or p62-siRNA. Cells were subsequently transfected with either Keap1-WT or Keap1-C151S. After transfection, cells were treated with 10 μM arsenic, 5 μM SF, or 25 μM tBHQ for an additional 16 h. Relative firefly and Renilla luciferase activities were then measured. As expected, treatment with arsenic, SF, or tBHQ significantly induced mGST-ARE-mediated luciferase activity in the Keap1-WT/control siRNA group (Fig. 7, top). Furthermore, when Keap1-C151S was cotransfected with control siRNA, luciferase activity was still significantly induced by arsenic but not by SF or tBHQ (Fig. 7, top). Although SF and tBHQ were still able to induce mGST-ARE luciferase activity, arsenic was no longer able to induce activity in the Keap1-WT/p62 siRNA group. None of the compounds were able to induce luciferase activity when p62 was knocked down and Keap1-C151S was expressed. Immunoblot analysis was also conducted in parallel. The results were consistent with the results of the dual-luciferase assay (Fig. 7, bottom). These results demonstrate a distinct mechanism of Nrf2 activation: SF or tBHQ activates Nrf2 in a Keap1-C151-dependent, p62-independent manner, whereas arsenic activates the Nrf2 pathway in a p62-dependent, Keap1-C151-independent manner.
Next, the p62 dependence of Nrf2 induction in response to arsenic was further compared to that in response to SF and tBHQ. BEAS-2B and HBE cells transfected with either control siRNA or p62 siRNA were treated with different doses of arsenic or SF for 4 h. Immunoblot analysis of Nrf2 and p62 showed a dose-dependent increase in the control siRNA group in both cell lines when treated with arsenic (Fig. 8A). When p62 was knocked down in either cell line, Nrf2 induction by arsenic was completely abrogated (Fig. 8A). A slight dose-dependent increase in the amount of LC3-II was observed in cells transfected with control siRNA and treated with arsenic, indicating an increase in the number of autophagosomes, which was unchanged in the p62 siRNA group (Fig. 8A). In contrast, SF increased the amount of Nrf2 in a dose-dependent manner in cells transfected with either control siRNA or p62 siRNA, although the fold induction of Nrf2 with p62 siRNA was slightly decreased (Fig. 8B). Furthermore, treatment of cells with SF elicited no change in p62 or LC3 in either cell lines (Fig. 8B). When BEAS-2B and HBE cells were treated with 25 μM tBHQ for 4 h, a dose-dependent increase in the amount Nrf2 was observed in both the control siRNA and p62 siRNA groups (Fig. 8C). Interestingly, there was also a dose-dependent increase in the amount of p62 in the control-siRNA-transfected cells, but there was no significant changes in the levels of Keap1 or LC3 (Fig. 8C). Also notable was an increase in the amount Keap1 when p62 was knocked down in any of the treatment groups. This may explain the blunted induction of Nrf2 in response to canonical Nrf2 activators, especially SF, in cells transfected with p62 siRNA (Fig. 8B). Together, these data confirm that arsenic-mediated induction of Nrf2 is p62 dependent, whereas induction mediated by SF and tBHQ is p62 independent.
We hypothesized that the noncanonical mechanism of Nrf2 activation by arsenic requires a longer time to resolve p62-mediated sequestration of Keap1 in autophagosomes and thus causes prolonged Nrf2 activation. Therefore, BEAS-2B and HBE cells were treated with arsenic, SF, or tBHQ for the indicated duration, and Nrf2 protein levels were analyzed. In both BEAS-2B and HBE cells, arsenic induced Nrf2 within 4 h yet persisted for as long as 48 h (Fig. 9A and andB).B). Treatment with SF or tBHQ increased Nrf2 protein levels by as early as 4 h, and these levels returned to basal levels by approximately 36 h in both cell lines (Fig. 9A and andB).B). Furthermore, a time-dependent increase in the level of p62 was observed in arsenic- or tBHQ-treated cells, whereas SF did not change the p62 level (Fig. 9A and andB).B). The Keap1 level did not change upon treatment with any compound (Fig. 9A and andB).B). To further confirm that Nrf2 activation by arsenic is prolonged compared to activation by SF, the half-life of Nrf2 under different treatment conditions was measured. The results suggest that the half-life of Nrf2 in the presence of arsenic is 44.9 min, whereas the half-life in the presence of SF-treated cells is only 18.6 min. These data support a scenario whereby arsenic-mediated activation of Nrf2 is prolonged, resembling the constitutive activation of Nrf2 found in certain types of cancers, which may contribute to arsenic carcinogenicity.
Recently, we and other groups independently reported a noncanonical mechanism of Nrf2 activation where direct interaction of p62 inactivates Keap1, causing stabilization of Nrf2 protein and increasing the transcription of ARE-bearing genes (43–47, 49). It has been well established by previous studies conducted in our laboratory that arsenic activates Nrf2 in a Keap1-C151-independent manner, unlike SF and tBHQ, which are common Nrf2 activators that activate the pathway through the canonical Keap1-C151 mechanism (8, 31). Furthermore, arsenic stabilizes Nrf2 by compromising Keap1-Cul3-dependent ubiquitination (31, 50); however, the exact mechanism(s) by which arsenic activates the Nrf2 pathway has yet to be elucidated. In this investigation, we demonstrate for the first time that arsenic-mediated activation of Nrf2 is through the noncanonical mechanism (p62 dependent). We report that acute low-dose arsenic blocks autophagic flux and enhances the number of autophagosomes, causing accumulation of p62 and sequestration of Keap1 in autophagosomes. This process inactivates Keap1, leading to stabilization of Nrf2. Interestingly, arsenic-mediated effects on autophagy were prevented or reversed by cotreatment with SF or tBHQ, indicating that canonical Nrf2 activators can prevent the detrimental effects of arsenic.
The mode of action by which arsenic causes cancer is not yet known. Enhanced ROS production and oxidative stress may play cardinal roles in arsenic toxicity and carcinogenicity, since generation of ROS can lead to genotoxicity, alter signal transduction pathways, inhibit DNA repair, and increase cell proliferation (28–30). Acute high-dose arsenic has been shown to increase ROS production as well as upregulate a number of oxidative stress-related genes, such as heme-oxygenase 1, NAD(P)H quinone oxidoreductase, and metallothionein (31, 51). However, low-dose arsenic has been shown to reduce ROS levels. Snow et al. demonstrated that human fibroblast cell lines treated with 500 nM arsenic for 24 h significantly decreased the amount of ROS present in cells by inducing a series of stress response genes compared to untreated cells (48). The 500 nM arsenic dose is considered to be in a subtoxic range and induces changes in overall gene expression (48). In our current study, treatment with 1 μM arsenic for 4 h did not changes ROS levels in comparison with levels in untreated cells (data not shown). Therefore, it is unlikely that activation of Nrf2 by low-dose arsenic is through generation of ROS. Furthermore, previous studies conducted by Bolt et al. with lymphoblastoid cell lines showed that arsenic induced autophagy and not apoptosis (38, 39). Also, human urothelial cells exposed to 1 or 4 μM arsenic for 48 h exhibited increases in the number of autophagosomes as well as protein expression levels of LC3 and Beclin-1, critical regulators for the formation of autophagosomes (37). These studies established an association between arsenic exposure and autophagy; however, the detailed mechanism by which low-dose arsenic causes pathological changes has remained largely unknown until now. Here, we demonstrate that low, environmentally relevant doses of arsenic block autophagic flux, causing an increase in the number of autophagosomes and activating the Nrf2 pathway via the noncanonical mechanism, which is independent of ROS as well.
It has been shown that a lack of essential autophagy genes, such as Atg5 and Atg7, leads to the formation of p62-positive aggregates and accumulation of ubiquitin inclusions in the liver (44, 52). Recently, Inami et al. reported that persistent Nrf2 activation through p62 in liver-specific-autophagy-deficient mice contributes to the development of hepatocellular carcinoma (53). An accumulation of p62 and elevated levels of Nrf2 downstream genes were also observed in bronchial epithelial cells of autophagy-deficient mice (54). Simultaneous knockout of either p62 or Nrf2 in autophagy-deficient mice suppressed ubiquitin accumulation and protein aggregation in the liver and brain and alleviated liver injury (44, 45). Here, we demonstrate that arsenic-mediated activation of Nrf2 is not only p62 dependent but also prolonged. Taken together, these studies provide evidence that deregulation of autophagy resulting in the accumulation of p62 and prolonged Nrf2 activation plays a critical role in the pathogenesis of many human diseases, which may include arsenic-induced cancers.
Previously, our laboratory demonstrated the protective role of Nrf2 against the cytotoxic effects induced by acute arsenic exposure both in cultured cells and in mice (55–57). Specifically, UROtsa cells with reduced Nrf2 expression were more susceptible to arsenic-induced toxicity than control cells (55). Pretreatment or cotreatment with an Nrf2 inducer, SF or tBHQ, rendered UROtsa cells more resistant to arsenic and monomethylarsonous acid toxicity (55). We also demonstrated the Nrf2 dependence of SF- and tBHQ-mediated protection through the use of Nrf2−/− mouse embryonic fibroblast (MEF) cells (55). Another group also showed that SF suppresses the cellular accumulation of arsenic and decreases toxicity in primary mouse hepatocytes (58). Our previous work in mice demonstrated that Nrf2−/− mice are more sensitive than their wild-type counterparts to arsenic-induced DNA hypomethylation, oxidative DNA damage, and apoptotic cell death in the liver and bladder when exposed to arsenic in their drinking water for 6 weeks (56). Our current study demonstrates that canonical Nrf2 activators prevent or reverse arsenic-mediated deregulation of autophagy, which may be partly responsible for the Nrf2-mediated protection against arsenic toxicity observed in previous studies. Further investigation is necessary to determine the biochemistry of how SF- or tBHQ-mediated activation of Nrf2 in the canonical manner supersedes arsenic-mediated deregulation of autophagy. One possible mechanism is that activation of Nrf2 alleviates arsenic-mediated effects through a reduction of the intracellular arsenic concentration, based on previous findings (54). Collectively, these studies clearly demonstrate that activation of the Nrf2 pathway through the canonical Keap1-C151 mechanism is important in protecting against the carcinogenic effects of arsenic and remains an effective strategy for chemoprevention.
One interesting observation is that tBHQ itself, similar to arsenic, was able to induce p62 at the protein level in a dose-dependent and time-dependent manner (Fig. 8C and and9A9A and andB).B). Nevertheless, tBHQ did not result in upregulation of autophagosomes. tBHQ may enhance the transcription of p62 though its reported ARE in the promoter, leading to an increase in the level of p62 (47). Furthermore, since autophagic flux is not affected, clearance of autophagosomes remains intact, and proteins such as LC3 and p62 continue to be degraded. Conversely, SF had no effect on the protein level of p62. It is interesting to note that different Nrf2 activators can upregulate distinct sets of Nrf2 target genes. During this study, we observed that arsenic, SF, and tBHQ had different effects on the induction of these well-studied Nrf2 target genes, including heme-oxygenase, γ-glutamylcysteine synthetase, aldose reductase, and NAD(P)H dehydrogenase 1 (data not shown). It would be interesting to understand how activation of Nrf2 by different compounds results in upregulation of distinct Nrf2 downstream genes.
In conclusion, we propose a model where arsenic induces persistent or prolonged Nrf2 activation, mimicking constitutive activation of Nrf2 caused by the dysfunctional Keap1-Nrf2 axis seen in many types of tumors. This study supports the notion that the Keap1-C151-dependent mechanism of Nrf2 activation is protective, while p62-mediated activation or somatic mutations leading to persistent or prolonged Nrf2 activation comprise the “dark side” of Nrf2 that is expected to promote tumor growth and survival. We believe that our study lays the groundwork for future investigations both in vitro and in vivo to explore the interplay between Nrf2, Keap1, p62, and autophagy as well as their roles in arsenic carcinogenicity.
We thank A. S. McElhinny for critical reading and editing of the manuscript.
This study was supported by the following grants: grants R01ES015010 (NIEHS) and R01CA154377 (NCI), which were awarded to D.D.Z.; an SOT-Novartis graduate student fellowship awarded to A.L.; and grant ES006694 (NIH), a center grant.
Published ahead of print 15 April 2013