Free Radical Biology & Medicine 45 (2008) 1035–1044

Contents lists available at ScienceDirect

Free Radical Biology & Medicine

j ourna l homepage: www.e lsev ie r.com/ locate / f reeradb iomed

Original Contribution

Cadmium activates the mitogen-activated protein kinase (MAPK) pathway viainduction of reactive oxygen species and inhibition of protein phosphatases 2A and 5

Long Chen a, Lei Liu a, Shile Huang a,b,⁎a Department of Biochemistry and Molecular Biology, Louisiana State University Health Sciences Center, 1501 Kings Highway, Shreveport, LA 71130-3932, USAb Feist-Weiller Cancer Center, Louisiana State University Health Sciences Center, 1501 Kings Highway, Shreveport, LA 71130-3932, USA

Abbreviations: AD, Alzheimer's disease; ALS, amyotapoptosis signal-regulating kinase 1; Cd, cadmium; Comethyl-2′,7′-dichlorodihydrofluorescein diacetate; DAdole; DMEM, Dulbecco's modified Eagle medium;regulated kinase 1/2; FBS, fetal bovine serum; H2O2, hN-terminal kinase; MAPK, mitogen-activated proteinMKP1, MAP phosphatase 1; NAC, N-acetyl-L-cysteine; O(ADP-ribose) polymerase; PBS, phosphate-buffered spropidium iodide; PD, Parkinson's disease; PP2A, proteinphosphatase 5; ROS, reactive oxygen species; zVAD-fmAsp fluoromethylketone.⁎ Corresponding author. Department of Biochemi

Louisiana State University Health Sciences Center, 150LA 71130-3932, USA. Fax: +1 318 675 5180.

E-mail address: shuan1@lsuhsc.edu (S. Huang).

0891-5849/$ – see front matter © 2008 Elsevier Inc. Aldoi:10.1016/j.freeradbiomed.2008.07.011

a b s t r a c t

a r t i c l e i n f o

Article history:

Cadmium (Cd), a highly to Received 2 June 2008Revised 12 July 2008Accepted 16 July 2008Available online 26 July 2008

Keywords:CadmiumApoptosisReactive oxygen speciesProtein phosphatase 2AProtein phosphatase 5c-Jun N-terminal kinaseExtracellular signal-regulated kinase 1/2

xic environmental pollutant, induces neurodegenerative diseases. Recently wehave demonstrated that Cd may induce neuronal apoptosis in part through activation of c-Jun N-terminalkinase (JNK) and extracellular signal-regulated kinase 1/2 (Erk1/2) pathways. However, the underlyingmechanism remains enigmatic. Here we show that Cd induced generation of reactive oxygen species (ROS),leading to apoptosis of PC12 and SH-SY5Y cells. Pretreatment with N-acetyl-L-cysteine (NAC) scavenged Cd-induced ROS, and prevented cell death, suggesting that Cd-induced apoptosis is attributed to its induction ofROS. Furthermore, we found that Cd-induced ROS inhibited serine/threonine protein phosphatases 2A (PP2A)and 5 (PP5), leading to activation of Erk1/2 and JNK, which was abrogated by NAC. Overexpression of PP2A orPP5 partially prevented Cd-induced activation of Erk1/2 and JNK, as well as cell death. Cd-induced ROS wasalso linked to the activation of caspase-3. Pretreatment with inhibitors of JNK (SP600125) and Erk1/2 (U0126)partially blocked Cd-induced cleavage of caspase-3 and prevented cell death. However, zVAD-fmk, a pancaspase inhibitor, only partially prevented Cd-induced apoptosis. The results indicate that Cd induction ofROS inhibits PP2A and PP5, leading to activation of JNK and Erk1/2 pathways, and consequently resulting incaspase-dependent and -independent apoptosis of neuronal cells. The findings strongly suggest that theinhibitors of JNK, Erk1/2, or antioxidants may be exploited for prevention of Cd-induced neurodegenerativediseases.

© 2008 Elsevier Inc. All rights reserved.

Introduction

Cadmium (Cd), a highly toxic heavy metal, is mainly released fromthe smelting, burning of fossil fuels and municipal wastes, refining ofmetals, and cigarette smoking, resulting in the pollution of water, air,and soil. Exposure of human or animals to a Cd-contaminatedenvironment or food chain causes accumulation of Cd in many organs,including kidney [1,2], liver [3], lung [4,5], testis, bone, etc. [6,7], and

rophic lateral sclerosis; ASK1,M-H2DCFDA, 5-(and-6)-chlor-PI, 4′,6-diamidino-2-phenylin-Erk1/2, extracellular signal-ydrogen peroxide; JNK, c-Junkinase; MKK, MAPK kinase;D, optical density; PARP, polyaline; PDL, poly-D-lysine; PI,phosphatase 2A; PP5, proteink, benzyloxycarbonyl-Val-Ala-

stry and Molecular Biology,1 Kings Highway, Shreveport,

l rights reserved.

thereby contributes to carcinogenesis, immunodepression, and neu-rodegeneration [8,9]. Clinical data show that exposure to Cd inducesneurological disorders, such as learning disabilities and hyperactivityin children [10,11], olfactory dysfunction and neurobehavioral defectsin attention, psychomotor speed, and memory in workers [7,12].Therefore, Cd intoxication has been considered and studied as apossible etiological factor of neurodegenerative diseases.

Studies have demonstrated that the toxicity of Cd is related to itsinduction of oxidative stress, e.g., reactive oxygen species (ROS), invarious types of cells [7,12–14]. Elevated levels of ROS may causeincreased permeability of the blood–brain barrier, tubulin alterations,and perturbation in synaptic transmission [7]. Oxidative stress is aprominent feature of many neurodegenerative disorders, such asAlzheimer's disease (AD), Parkinson's disease (PD), and amyotrophiclateral sclerosis (ALS). Cd-induced oxidative stress is closely associatedwith PD and AD [13,15–18]. Accumulating data further show thatunder pathological conditions, excessive amounts of ROS induced byCd can modify proteins, lipids, and DNA, alter their functions, andactivate related signaling pathways, thereby resulting in apoptosis ofneuronal cells [12,14,16,19–21].

There is growing evidence that members of the mitogen-activatedprotein kinase (MAPK) family may play a critical role in neuronalapoptosis [22]. MAPKs comprise a highly conserved cascade of serine/

1036 L. Chen et al. / Free Radical Biology & Medicine 45 (2008) 1035–1044

threonine kinases connecting cell surface receptors to regulatorytargets in response to various stimuli [16,22,23]. Mammals express atleast three distinct groups of MAPKs, including extracellular signal-regulated protein kinase 1/2 (Erk1/2), c-Jun N-terminal kinase (JNK),and p38 MAPK, which have beenwell characterized. In neuronal cells,Erk1/2 is primarily activated by growth factors, and is involved incellular proliferation, differentiation, and development, whereas JNKand p38 signaling cascades are preferentially activated by environ-mental stress and inflammatory cytokines, and have been shown topromote neuronal cell death [24]. Phosphorylation of MAPKs isbalanced by specific MAPK kinases and phosphatases. MAPK kinase1/2, MAPK kinase 3/6, and MAPK 4/7 are the major kinases tophosphorylate Erk1/2, p38, and JNK, respectively, though there existcross talks [22]. MAP phosphatase 1 (MKP1) and serine/threonineprotein phosphatase 2A (PP2A) are the major phosphatases thatnegatively regulate phosphorylation of Erk1/2, JNK, and p38 [25,26].Protein phosphatase 5 (PP5) has been identified as a negativeregulator of JNK cascade, involved in stress responses [27,28]. PP2Ais a heterotrimeric holoenzyme composed of a catalytic subunit(PP2Ac), an A subunit (also termed PR65), and a number of B subunits,including B (PR55), B′ (PR61), B″ (PR72), and B′″ (PR93/PR110) [29].The phosphatase activity of PP2Ac or PP5 is modulated by theirassociations with PP2A-A and -B regulatory subunits [27]. Further-more, PP2A activity is also mediated by the phosphorylation ormethylation of PP2Ac [30,31]. Recently, we have found that all threeMAPK members can be activated by Cd in neuronal cells, andidentified that Cd-induced neuronal apoptosis is partially associatedwith activation of JNK, Erk1/2, but not by p38 signaling [32]. However,it is less clear how Cd activates MAPKs signaling pathways in theneuronal cells.

Activation of JNK and p38 MAPK by ROS has been described [24].This prompted us to study whether Cd activates MAPKs signaling viaits induction of ROS in neuronal cells. Since fully developed neuronsare permanently amitotic (not dividable), transformed PC12 and SH-SY5Y cells, two established cell lines derived from rat adrenalmedulla and human sympathetic neurons, respectively, are widelyused as in vitro neuronal models, Here we show that Cd activationof MAPKs is indeed related to its induction of ROS generation inneuronal (PC12 and SH-SY5Y) cells. This is supported by the findingsthat Cd induced generation of ROS; N-acetyl-L-cysteine (NAC), anantioxidant and ROS scavenger, effectively blocked Cd-inducedactivation of Erk1/2, JNK, and p38 signaling network, and preventedCd-induced cell death. We also found that the activation of MAPKsis associated with inhibition of PP2A and PP5 by Cd-induced ROS.Furthermore, we observed that Cd-induced ROS activated caspase-3,which could be blocked by NAC. Treatment with zVAD-fmk, a pancaspase inhibitor, only partially protected against Cd-induced celldeath, implying that caspase-dependent and -independent apoptoticmechanisms are involved. Pretreatment with MAPK inhibitors(SP600125 and U0126) partially inhibited Cd-induced cleavage ofcaspase-3 and apoptosis, suggesting that those inhibitors, plusantioxidant (e.g., NAC), may be exploited for prevention of Cd-inducedneurodegenerative diseases.

Materials and methods

Materials

Cadmium chloride (Sigma, St. Louis, MO) was dissolved in steriledistilled water to prepare the stock solutions (0–120mM), aliquoted,and stored at room temperature. Dulbecco's modified Eagle medium(DMEM) was purchased from Mediatech (Herndon, VA). Horse serumand fetal bovine serum (FBS) were supplied by Hyclone (Logan, UT),whereas 0.05% Trypsin-EDTA was from Invitrogen (Grand Island, NY).Enhanced chemiluminescence solutionwas from Pierce (Rockford, IL).CellTiter 96 AQueous One Solution Cell Proliferation Assay kit was from

Promega (Madison, WI). The MAPK inhibitors, SB203580, U0126, andSP600125, were obtained from LC Laboratories (Woburn, MA). Thecaspase inhibitor, z-Val-Ala-Asp-CH2F (zVAD-fmk) was purchasedfrom ALEXIS Biochemicals Corporation (San Diego, CA). The followingantibodies were used: ASK1, phospho-ASK1 (Thr845), MKK4, phos-pho-MKK4 (Ser257/Thr261), phospho-Erk1/2 (Thr202/Tyr204) (CellSignaling Technology, Beverly, MA), PP2ACα, PP5 (BD Biosciences, SanJose, CA), PP2A-A subunit, PP2A-B subunit (Upstate, Lake Placid, NY),JNK1, phospho-JNK (Thr183/Tyr185), c-Jun, phospho-c-Jun (Ser63),Erk2, p38, phospho-p38 (Thr180/Tyr182), MEK1/2, phospho-MEK1/2,MEK3/6, phospho-MEK3/6, demethylated-PP2A, MKP1, FLAG, HA(Santa Cruz Biotechnology, Santa Cruz, CA), phospho-PP2A (Epitomics,Burlingame, CA), and β-tubulin (Sigma). Annexin V-FITC ApoptosisDetection Kit I was purchased from BD Biosciences. 5-(and-6)-chloromethyl-2′,7′-dichlorodihydrofluorescein diacetate (CM-H2DCFDA) and propidium iodide (PI) were from MP Biomedicals Inc.(Solon, OH), and Alfa Aesar (Ward Hill, MA), respectively. N-Acetyl-L-cysteine, poly-D-lysine (PDL), 4′,6-diamidino-2-phenylindole (DAPI),and all the other chemicals were purchased from Sigma.

Cell culture

Rat pheochromocytoma (PC12) and human neuroblastoma (SH-SY5Y) cell lines were from American Type Culture Collection (ATCC)(Manassas, VA). PC12 cells were grown in antibiotic-free DMEMsupplemented with 10% horse serum and 5% FBS, whereas SH-SY5Ycells were grown in antibiotic-free DMEM supplemented with 10%FBS. Cells were maintained in a humid incubator (37°C, 5% CO2).

Recombinant adenoviral constructs and infection of cells

Plasmids encoding N-terminal FLAG-tagged wild-type rat PP2Aαcatalytic subunit (PP2Ac-α) and hemagglutinin (HA)-tagged wild-type human PP5 [27,28] were gifts from Dr. Hitoshi Nakagama(Hokkaido University, Sapporo, Japan) and Dr. Hidenori Ichijo(University of Tokyo, Tokyo, Japan), respectively. The recombinantadenoviruses encoding FLAG-tagged PP2Ac-α (Ad-PP2A), HA-taggedPP5 (Ad-PP5), and the control virus encoding the green fluorescenceprotein (GFP) (Ad-GFP) were generated using the ViraPower Adeno-viral Expression System (Invitrogen, Carlsbad, CA), following themanufacture's instruction. The viruses were amplified and titrated asdescribed [33]. For experiments, PC12 and SH-SY5Y cells were grownin the growth medium, and infected with the individual adenovirusfor 24h at 5 of multiplicity of infection (MOI = 5). Subsequently, cellswere used for experiments. Ad-GFP served as a control. Expression ofFLAG-tagged PP2Ac-α and HA-tagged PP5 was determined byWestern blot with antibodies to FLAG and HA (Santa Cruz Biotechnol-ogy), respectively.

Analysis for generation of ROS

The production of ROS was measured by detecting the fluorescentintensity of oxidant-sensitive probe CM-H2DCFDA, which is a stablenonfluorescent molecule that passively diffuses into cells, where theacetate can be cleaved by intracellular esterases to produce a polar diolthat is well retained within the cells. PC12 cells were seeded at adensity of 1 × 104 cells/well in 96-well plate, precoated with PDL(0.2μg/ml). The next day, cells were loaded with CM-H2DCFDA as perthe manufacturer's protocol and incubated in the presence of variousconcentrations of Cd (0–120μM) for 24h or 20μMCd for different times(0–24h) with 6 replicates of each treatment. In some cases, cells werepreincubated with NAC (5mM) for 1h, and then treated with/withoutCdCl2 (10 and 20μM) for 24h, followed by loading with CM-H2DCFDAfor 40min. For experiments with inhibitors, including the MAPKinhibitors, SB203580, U0126, and SP600125 as well as the caspaseinhibitor zVAD-fmk, cells were preincubated with each inhibitor for

1037L. Chen et al. / Free Radical Biology & Medicine 45 (2008) 1035–1044

30min, and then treated with 20μM Cd for 12h followed by loadingwith CM-H2DCFDA. Fluorescent intensity was recorded by excitationat 485nm and emission at 535nm using a Wallac 1420 MultilabelCounter (Perkin-Elmer Life Sciences, Wellesley, MA).

Cell viability assay

Cells were seeded at a density of 1 × 104 cells/well in a flat-bottomed 96-well plate, precoated with (for PC12) or without (for SH-SY5Y) PDL (0.2μg/ml). The next day, cells were treated with/withoutCd (10 and 20μM) in the presence or absence of NAC (5mM) for 24hwith 4–6 replicates of each treatment. For cells infected with Ad-PP2A,Ad-PP5, and Ad-GFP, respectively, Cd (10 and 20μM) was added for24h. After incubation, 20μl of one solution reagent (Promega, Madison,WI) was added to each well and incubated for 4h. Cell viability wasdetermined by measuring the optical density (OD) at 490nm using aWallac 1420 Multilabel Counter (Perkin-Elmer Life Sciences).

Cell morphological analysis

Cells were seeded at a density of 1 × 106 cells/well in a 6-well plate,precoated with (for PC12) or without (for SH-SY5Y) PDL (0.2μg/ml).Next day, Cd (0–40μM) or Cd (10 and 20μM) following 1h of NAC(5mM) preincubation was added. Additionally, cells infected with Ad-PP2A, Ad-PP5, and Ad-GFP, respectively, were exposed to Cd (10 and20μM). After incubation for 24h, images were taken with an Olympus

Fig. 1. Cd-induced ROS generation is associated with neuronal cell apoptosis. Cd induced ROcells treated with different concentrations of Cd for 24 h. (B) ROS change in PC12 cells treatecells treated with different concentrations of Cd for 24 h. ROS generation was evaluated bmorphology was assessed using an Olympus inverted phase-contrast microscope (200×) equ⁎⁎Pb0.01, difference vs control group.

inverted phase-contrast microscope (Olympus Optical Co., Melville,NY) (200×) equipped with the Quick Imaging system.

Apoptosis assay using DAPI staining

Cells were seeded at a density of 5 × 105 cells/well in 6-well platecontaining a PDL-coated glass coverslip per well. Next day, the cellswere preincubated with NAC (5mM) for 1h, followed by treatmentwith/without CdCl2 (10, 20μM) for 24h. Subsequently, cells were fixedwith 4% paraformaldehyde prepared in PBS for 2h at 4°C. The cellswere washed three times with PBS, and then stained with DAPI (4μg/ml in deionized water) for 30min at room temperature in the dark.Following a brief washing with PBS, slides were mounted in glycerol/PBS (1/1, v/v) containing 2.5% 1,4-diazabiclo-(2,2,2)octane. Photo-graphs were taken with a Nikon Eclipse TE300 fluorescence micro-scope (Nikon Instruments Inc., Melville, NY) equipped with a digitalcamera. Cells with condensed nuclei were scored to be apoptotic.

Apoptosis assay using flow cytometry

Cells were seeded in 100-mm dishes, precoated with (for PC12) orwithout (for SH-SY5Y) PDL, at a density of 2 × 106 cells/dish in com-pleted growth medium. Next day, after 1h of NAC (5mM) preincuba-tion, cells were cotreated with/without CdCl2 (10 and 20μM) for 24h,followed by apoptosis assay using the Annexin V-FITC ApoptosisDetection Kit I (BD Biosciences, San Diego, CA), as described [34].

S production in a concentration- and time-dependent manner. (A) ROS change in PC12d with 20 μM Cd for indicated time. (C) Morphological alterations in PC12 and SH-SY5Yy CM-H2DCFDA (10 μM) oxidation-based fluorescence using a microplate reader. Cellipped with Quick Imaging system. Results are presented as mean±SE; n=4–6. ⁎Pb0.05,

1038 L. Chen et al. / Free Radical Biology & Medicine 45 (2008) 1035–1044

Western blot analysis

After treatment, cells were briefly washed with cold PBS. On ice,cells were lysed in RIPA buffer [50mM Tris, pH 7.2; 150mM NaCl; 1%sodium deoxycholate; 0.1% SDS; 1% Triton-X 100; 10mM NaF; 1mMNa3VO4; protease inhibitor cocktail (1:1000, Sigma). Lysates weresonicated for 10s and centrifuged at 14,000rpm for 10min at 4°C.

Fig. 2. N-acetyl-L-cysteine (NAC) strongly blocks Cd-induced ROS and abolished neuronal celCd (10 and 20 μM) for 24 h. ROS generation was assayed by CM-H2DCFDA (10 μM) oxidationinduced ROS generation in the cells. (B) Morphology of PC12 cells treated with/without Cd incontrast microscope (200×) equipped with Quick Imaging system. A, control; B, 5 mMNAC; CNAC plus 10 and 20 μM Cd, respectively, showing obvious protection of NAC. (C) Cell viabilitabsence of NAC (5 mM) for 24 h was evaluated using one solution assay. (D) Cells were harveNAC (5mM), stained with annexin-V-FITC and propidium iodide, and analyzed by the fluorescshown (Right panel). Cd-induced apoptosis was significantly blocked by NAC (Left panel). RecPb0.01, difference vs 10 μM Cd group; dPb0.01, difference vs 20 μM Cd group.

Protein concentration was determined by bicinchoninic acid assaywith bovine serum albumin as standard (Pierce). Equivalent amountsof protein were separated on 7.5–12% SDS–polyacrylamide gel andtransferred to polyvinylidene difluoride membranes (Millipore, Bed-ford, MA). Membranes were incubated with PBS containing 0.05%Tween 20 and 5% nonfat dry milk to block nonspecific binding andwere incubated with primary antibodies, then with appropriate

l apoptosis. (A) PC12 cells were pretreated with NAC (5 mM) for 1 h, and then exposed to-based fluorescence using a microplate reader, showing that NAC strongly blocked Cd-the presence or absence of NAC for 24 h was assessed using an Olympus inverted phase-and D,10 and 20 μMCd, respectively, showing a great loss of cell integrity; E and F, 5 mMy of PC12 and SH-SY5Y cells treated with/without Cd (10 and 20 μM) in the presence orsted after 24 h treatment with/without Cd (10 and 20 μM) in the presence or absence ofence-activated cell sorting using flowcytometry. A representative experimental result issults are presented as mean±SE; n=4–6. aPb0.05, bPb0.01, difference vs control group;

1039L. Chen et al. / Free Radical Biology & Medicine 45 (2008) 1035–1044

secondary antibodies conjugated to horseradish peroxidase. Immu-noreactive bands were visualized by using enhanced chemilumines-cence solution (Pierce). To check the amount of protein loaded, theimmunoblots were treated with stripping solution (62.5mM Trisbuffer, pH 6.7, containing 2% SDS and 100mM β-mercaptoethanol) for30min at 50°C and incubated with mouse monoclonal anti-β-tubulinantibody (Sigma) followed by horseradish peroxidase-coupled goatanti-mouse IgG (Pierce).

Statistical analysis

Results were expressed as mean values ± standard error (mean ±SE). Statistical analysis was performed by Student's t test (STATISTICA,Statsoft Inc., Tulsa, OK). A level of P b 0.05 was considered to besignificant.

Results

Cd-induced neuronal apoptosis is associated with its induction of ROS

Recently, we have demonstrated that Cd induces apoptosis of PC12and SH-SY5Y cells in a time- and concentration-dependent manner[32]. To determine whether this is associated with Cd induction ofROS, we measured ROS generation after PC12 and SH-SY5Y cells wereexposed to Cd. As shown in Fig. 1A, treatment with Cd for 24h resultedin a concentration-dependent increase of ROS production at concen-trations of 0–20μM in PC12 cells. Starting at 40μM, ROS productiongradually declined, and dropped to a level below basal conditions at120μM. This is probably due to extremely high toxicity of Cd atconcentrations of N80μM, which within hours resulted in death(necrosis) of too many cells that were not able to produce ROS. Cd alsoinduced a time-dependent elevation of cellular ROS within 24h (Fig.1B). After 4h treatment, Cd (20μM) significantly increased cellular ROSlevels, which is consistent with the finding of a decreased cell viabilityobserved in our previous studies [32]. Similar results were seen in SH-SY5Y cells (data not shown). By phase-contrast microscopic observa-tion, more round or shrunken cells appeared, when exposed toincreasing concentrations of Cd (2.5–40μM) (Fig. 1C).

NAC blocks Cd-induced ROS and neuronal apoptosis

Free radical scavenger or antioxidant NAC, a thiol-containingcompound, has been shown to directly reduce the levels of ROS

Fig. 3. Cd-induction of ROS activates MAPK pathway in neuronal cells. PC12 cells treated withwere subjected to Western blot analysis using indicated antibodies. The blots were probindependent experiments. (A–C) NAC blocked Cd-activated JNK, Erk1/2, and p38 as well as

[35–37]. To confirm that Cd-induced neuronal apoptosis is indeed dueto its induction of ROS generation, PC12 and SH-SY5Y cells werepretreated with NAC (5mM) for 1h, and then exposed to Cd (10 and20μM) for 24h. We found that NAC dramatically blocked Cd-inducedROS generation in PC12 cells (Fig. 2A) and SH-SY5Y cells (data notshown). Morphological analysis (Fig. 2B) reveals that NAC itself did notalter PC12 cell shape (B vs A). Cd alone (10 and 20μM) induced cellroundup and shrinkage (C, D vs A). However, NAC almost completelyabolished Cd-induced morphological change (E, F vs A). Similar datawere also seen in SH-SY5Y cells (data not shown). Results from onesolution assay (Fig. 2C) further demonstrate that NAC potentlysuppressed Cd-induced loss of cell viability in Cd-exposed PC12 andSH-SY5Y cells. In addition, to further quantify the protective effect ofNAC on Cd-induced apoptosis via blockage of ROS in a larger cellpopulation, we performed annexin-V-FITC and propidium iodidestaining followed by flow cytometry. As shown in Fig. 2D, NAC alonedid not affect cell viability. However, it significantly blocked Cd-induced apoptosis. Similar results were also seen using DAPI staining(data not shown).

Cd-induction of ROS activates MAPK pathway in neuronal cells

Recently we have demonstrated that activation of the MAPKpathway in part contributes to Cd-induced apoptosis of the neuronalcells [32]. In the present study, we have shown that Cd-inducedneuronal apoptosis is attributed to its induction of ROS. Therefore, wenext investigated whether Cd-induced ROS activates the MAPKpathway, leading to cell death. To this end, PC12 and SH-SY5Y cellswere pretreated with or without NAC (5mM), followed by exposure toCd (10–20μM). We found that pretreatment with NAC remarkablyinhibited Cd-induced phosphorylation of JNK, Erk1/2, and p38 as wellas upstream kinases, such as ASK1, MKK4, MEK1/2, and MEK3/6 inPC12 cells (Figs. 3A, B, and C) and SH-SY5Y cells (data not shown).Collectively, the findings support the notion that Cd induction of ROSactivates the MAPK pathway, triggering apoptosis of the neuronalcells.

Cd induction of ROS downregulates PP2A and PP5, resulting in activationof Erk1/2 and JNK

Previous studies have shown that Cd may elicit cell apoptosis byactivation of Erk1/2 and JNK as well as the upstream kinases, such asASK1, MEK3/6, and MKK4 in neuronal (PC12 and SH-SY5Y) cells

/without Cd in the presence or absence of NAC for 24 h were harvested. The cell lysatesed for β-tubulin as a loading control. Similar results were observed in at least threethe upstream ASK1, MKK4, MEK1/2, and MEK3/6.

Fig. 4. Cd-induction of ROS does not alter cellular protein level of MPK1, butdownregulates PP2A and PP5. PC12 cells, treated with 0–20 μM Cd for 24 h, orcotreated with/without Cd and NAC for the indicated concentration and time, wereharvested, and the cell lysates were subjected toWestern blot analysis using antibodies.The blots were probed for β-tubulin as a loading control. Similar results were observedin at least three independent experiments. Cd did not alter cellular protein level ofMPK1, PP2Ac, and PP2A-B, but decreased PP2A-A and PP5 in a dose-dependent manner.Cd also obviously increased expression of demethylated- and phospho-PP2A (Tyr307)(A). The above effects were strongly blocked by NAC (B), suggesting that Cd-inducedROS downregulates PP2A and PP5.

1040 L. Chen et al. / Free Radical Biology & Medicine 45 (2008) 1035–1044

[12,32]. Since phosphorylation of Erk1/2 and JNK is not only regulatedby the upstream kinases but also regulated by a series of phospha-tases, such asMKP1, PP2A, and PP5 [25,26], we therefore hypothesizedthat Cd activation of Erk1/2 and JNK pathways could be related todownregulation of protein levels/activities of these phosphatases. Totest this hypothesis, PC12 cells were exposed with 0–20μM Cd for upto 24h, followed by Western blotting. We observed that Cd did notalter cellular protein levels of MKP1, PP2Ac, and PP2A-B, but markedlyreduced protein levels of PP2A-A and PP5 in a concentration-dependent manner (Fig. 4A). In addition, exposure of PC12 cells toCd resulted in dose-dependent increase of demethylated- andphospho-PP2Ac (Fig. 4A). Similar data were seen in SH-SY5Y cells(data not shown). Collectively, our data reveal that Cd did notobviously affect MKP1 protein levels, but might inhibit PP2A activity,

Fig. 5. Overexpression of PP2A partially prevents Cd-induced activation of Erk1/2 and JNK, aexposed with Cd (10 and 20 μM) for 4 h and then harvested. The cell lysates were subjected toa loading control. Similar results were observed in at least three independent experiments. (Bcells treatedwith/without indicated Cd dose for 24 hwere assessed using an Olympus invertesolution assay, respectively. Results are presented as mean±SE; n=4–6. ⁎⁎Pb0.01, differenc

by reducing PP2A-A protein levels, and increasing demethylation andphosphorylation of PP2Ac, and inhibit PP5, by reducing protein levelsof PP5 and PP2A-A. The findings suggest that Cd may downregulatePP2A and PP5, resulting in activation of Erk1/2 and JNK pathways.

Since Cd-induced ROS activated MAPK pathways in neuronal cells(Fig. 3), we next examined whether Cd downregulation of PP2A andPP5 is related to its ROS induction. Western blot analysis shows thatNAC completely blocked Cd-induced increase of demethylated- andphospho-PP2Ac, and rescued downregulation of PP2A-A and PP5protein levels in PC12 cells (Fig. 4B) and SH-SY5Y cells (data notshown).

Overexpression of PP2A or PP5 partially prevents Cd-induced activationof Erk1/2 and JNK, as well as cell death

To confirm the role of PP2A and PP5 in Cd-induced activation ofMAPKs and neuronal apoptosis, PC12 cells, infectedwith Ad-PP2A, Ad-PP5, and Ad-GFP (as control), were exposed to Cd (10 and 20μM) for4h, followed by Western blot analysis. We observed that over-expression of PP2Ac partially suppressed Cd-induced phosphorylationof Erk1/2 and JNK, but not that of p38 MAPK (Fig. 5A), whereasoverexpression of PP5 inhibited Cd-induced activation of JNK, but notErk1/2 and p38 MAPK (Fig. 6A). Of importance, overexpression ofPP2Ac or PP5 also in part prevented Cd-induced cell death in PC12 andSH-SY5Y cells (Figs. 5B and C and Figs. 6B and C).

Cd induces caspase-dependent and -independent apoptosis ofneuronal cells

Studies have shown that Cd may trigger cell death by caspase-dependent and/or -independent apoptoticmechanisms, depending oncell types [12,16,38–45]. It remains to be defined whether Cd inducesneuronal apoptosis through activation of caspase-dependent path-ways. To address this question, we determined proteolytic cleavages ofcaspase-3 and poly (ADP-ribose) polymerase (PARP) in PC12 and inSH-SY5Y. Our Western blot results show that treatment of PC12 cellswith Cd for 24h resulted in robust activation of caspase-3 and PARP ina concentration-dependent manner, as detected by increased clea-vages of caspase-3 and PARP (Fig. 7A). Further, treatment of PC12 cells

s well as cell death. (A) PC12 cells, infected with Ad-PP2A and Ad-GFP (as control), wereWestern blot analysis using indicated antibodies. The blots were probed for β-tubulin asand C) Morphology and cell viability of Ad-PP2A- and Ad-GFP-infected PC12 or SH-SY5Yd phase-contrast microscope (200×) equippedwith Quick Imaging system and using onee with control group; ##Pb0.01, Ad-PP2A group vs Ad-GFP group.

Fig. 6.Overexpression of PP5 partially prevents Cd-induced activation of JNK, as well as cell death. (A) PC12 cells, infected with Ad-PP5 and Ad-GFP (as control), were exposedwith Cd(10 and 20 μM) for 4 h and then harvested. The cell lysates were subjected to Western blot analysis using indicated antibodies. The blots were probed for β-tubulin as a loadingcontrol. Similar results were observed in at least three independent experiments. (B and C) Morphology and cell viability of Ad-PP5- and Ad-GFP-infected PC12 or SH-SY5Ycellstreated with/without indicated Cd dose for 24 h were assessed using an Olympus inverted phase-contrast microscope (200×) equipped with Quick Imaging system and using onesolution assay, respectively. Results are presented as mean±SE; n=4–6. ⁎⁎Pb0.01, difference with control group; ##Pb0.01, Ad-PP5 group vs Ad-GFP group.

1041L. Chen et al. / Free Radical Biology & Medicine 45 (2008) 1035–1044

with Cd (20μM) also increased cleavages of caspase-3 and PARP in atime-dependent manner (Fig. 7B). Similar data were seen in SH-SY5Ycells (data not shown).

Fig. 7. Cd activates caspase-3 signaling related to ROS induction. PC12 cells, treated with0–20 μMCd for 24 h or with 20 μMCd for 0–12 h, or cotreated with/without Cd and NACfor the indicated concentration and time, were harvested. The cell lysates weresubjected toWestern blot analysis using antibodies. The blots were probed for β-tubulinas a loading control. Similar results were observed in at least three independentexperiments. Cd obviously increased cleavage of caspase-3 and PARP in a concentration-dependent (A) and time-dependent (B) manner. (C) NAC potently blocked the cleavageof caspase-3 or PARP, indicating that Cd-induced ROS may mediate activation ofcaspase-3 pathway.

Fig. 8. Cd induces neuronal apoptosis by caspase-dependent and -independentmechanisms. (A) PC12 cells, treated with 20 μM Cd for 4 h and 12 h followingpretreatment with a pan caspase inhibitor, zVAD-fmk (100 μM) for 2 h, were harvested.The cell lysates were subjected toWestern blot analysis using antibodies. The blots wereprobed for β-tubulin as a loading control. Similar results were observed in at least threeindependent experiments. (B) Morphology of PC12 cells, treated with/without 20 μMCdin the presence or absence of 100 μM zVAD-fmk for 24 h, was assessed using anOlympus inverted phase-contrast microscope (200 ×) equipped with Quick Imagingsystem.

1042 L. Chen et al. / Free Radical Biology & Medicine 45 (2008) 1035–1044

Since Cd induced apoptosis of PC12 and SH-SY5Y cells due toits induction of ROS (Fig. 1), we next tested whether Cd activates thecaspase-3 pathway related to its ROS induction. The results indicatethat NAC almost completely blocked Cd-induced cleavage ofcaspase-3 and PARP in PC12 cells (Fig. 7C) and SH-SY5Y cells (datanot shown).

To unveil whether there exists a caspase-independent mechanisminvolved in Cd-induced neuronal apoptosis, PC12 cells were exposedto 20μM Cd for 4 and 12h after pretreatment with zVAD-fmk, a pancaspase inhibitor, for 2h. As shown in Fig. 8A, Cd-activated caspase-3was obviously attenuated by zVAD-fmk (100μM). However, whenPC12 cells were pretreated with this inhibitor for 2h, followed byexposure to 20μM Cd for 24h, morphological analysis revealed thatzVAD-fmk itself did not obviously alter cell shape, but only partiallyrescued cells from Cd-induced apoptosis (Fig. 8B).

Inhibitors of JNK and Erk1/2 partially inhibit Cd-activated caspasepathway, preventing neuronal apoptosis

Recently we have shown that inhibitors of JNK and Erk1/2 partiallyprevent Cd-induced neuronal apoptosis [32]. To investigate whetherthis prevention is associated with inhibition of caspase activation,PC12 cells were exposed to 20μM Cd for 12h after pretreatment withthe MEK1/2 (upstream of Erk1/2) inhibitor U0126, JNK inhibitorSP600125, or p38 inhibitor SB203580 for 30min, respectively. Asshown in Fig. 9A, pretreatment with SP600125 (20μM) and U0126(5μM), but not SP203580 (10μM), obviously reduced Cd-inducedcleavage of caspase-3. Additionally, we also noted that SP600125 andU0126, but not SB203580, could partially attenuate Cd-induced ROS(Fig. 9B).

Fig. 9. Inhibitors of JNK and Erk1/2 partially inhibit the Cd-activated caspase pathway,preventing neuronal apoptosis. (A) PC12 cells were treated with 20 μM Cd for 12 hfollowing preincubationwith 20 μM JNK inhibitor SP600125, 5 μMMEK1/2 (upstream ofErk1/2) inhibitor U0126, or 10 μM p38 inhibitor SB203580 for 30 min, respectively.Western blot analysis was performed using indicated antibodies. The blots were probedfor β-tubulin as a loading control. Similar results were observed in at least threeindependent experiments. (B) PC12 cells were pretreated with SP600125 (20 μM),U0126 (5 μM), or SB203580 (10 μM) for 30 min, respectively, and then exposed to Cd(10 μM) for 12 h. ROS generation was evaluated by CM-H2DCFDA (10 μM) oxidation-based fluorescence using a microplate reader. Results are presented as mean±SE;n=4–6. Different letters above bars indicate significant difference at Pb0.05.

Discussion

Recently we have demonstrated that Cd induces apoptosis ofneuronal cells in part by activation of Erk1/2 and JNK signalingpathways [32]. However, the underlying mechanism remains enig-matic. Here, we provide evidence that Cd induced ROS generation,resulting in activation of JNK, Erk1/2, and p38 MAPK, and theirupstream kinases, such as ASK1, MKK4, MKK1/2, and MEK3/6 [22],and concurrently inhibition of PP2A and PP5, two phosphatases thatnegatively regulate Erk1/2, p38, and JNK, respectively [25–28], in PC12and SH-SY5Y cells. Of importance, a ROS scavenger, NAC, couldcompletely block these events. Overexpression of PP2A or PP5 wasable to partially inhibit Cd-induced activation of Erk1/2 and/or JNK,and apoptosis of the cells. Our data clearly indicate that Cd inducesapoptosis of the neuronal cells, at least in part by inhibition of PP2Aand PP5, leading to activation of Erk1/2 and JNK pathways.

Here, for the first time, we show that Cd induction of ROSdownregulates the activities of PP2A and PP5 in the cells. This isstrongly supported by the findings that: (i) Cd treatment reduced thecellular protein level of PP2A-A, which is essential for activation ofPP2A or PP5 [27]; (ii) Cd treatment decreased the cellular protein levelof PP5, which directly determines the catalytic activity of PP5; (iii) Cdtreatment increased demethylation of PP2Ac and phosphorylation ofPP2Ac (Tyr 307), which are the indicators for inhibition of PP2Aactivity [30,31]; (iv) all above events were blocked by pretreatmentwith NAC, a ROS scavenger, in PC12 and SH-SY5Y cells. Our results arein agreement with the finding that hydrogen peroxide, a source of ROSgenerator, inhibits protein phosphatase 1, PP2A, and calcineurin(protein phosphatase 2B) in SK-N-SH neuroblastoma cells [46], as wellas protein tyrosine phosphatase 1B in HEK 293 cells [47]. We did notfind that Cd altered the cellular protein level of MKP1 (Fig. 4), a majorphosphatase that dephosphorylates Erk1/2, JNK, and p38 MAPK.However, we could not exclude the possibility that Cd may actuallyinhibit MKP1 activity by unidentified mechanisms. It would be ofgreat importance to investigate whether ROS inhibition of phospha-tases is a general mechanism by which corresponding kinases areactivated in the cells.

We found that overexpression of PP2A and PP5 inhibitedphosphorylation of Erk1/2 and JNK, but only partially prevented Cd-induced apoptosis. This suggests that other signaling pathways,related to the neuronal survival, could also be targeted by Cd.Recently, we have demonstrated that mTOR signaling was activatedby Cd in PC12 and SH-SY5Y cells [32]. Inhibition of mTOR byrapamycin or downregulation of mTOR by RNAi in part attenuatedCd-induced apoptosis of the neuronal cells [32], implying that Cd-activation of the mTOR pathway could contribute to its induction ofneuronal cell death as well.

Recently we have shown that the apoptosis induced by Cd, in PC12and SH-SY5Y cells, was partially rescued by SP600125 (JNK inhibitor)and U0126 (inhibitor of MEK1/2, upstream kinases of Erk1/2), but notby SB203580 (p38 MAPK inhibitor) [32]. Consistently, here we notedthat SP600125 and U0126, but not SB203580, were able to partiallyattenuate Cd-induced ROS and cleavage of caspase-3 in PC12 and SH-SY5Y cells. Our data suggest that these inhibitors may preventneuronal cell death, not only by directly inhibiting phosphorylationof the kinases, but also by inhibiting Cd induction of ROS, which, inturn, results in less activation of MAPKs due to less ROS production. Itis unclear how SP600125 and U0126 inhibit Cd induction of ROS.Further studies are needed to address this issue.

Caspase-3 appears to play an important role in the regulation ofcell death in the brain [48–50]. Activation of caspase-3 has beenimplicated in PD and AD [51–55], implying that activation of thecaspase pathway may be a common mechanism of neuronal celldeath. Because Cdmay trigger cell death by caspase-dependent and/or-independent apoptotic mechanisms, depending on cell types[12,16,38–45], we investigated the role of caspase-3 in Cd-induced

1043L. Chen et al. / Free Radical Biology & Medicine 45 (2008) 1035–1044

neuronal apoptosis. Our results show that Cd significantly increasedproteolytic cleavages of caspase-3 and PARP in PC12 cells in aconcentration- and time-dependent manner. Pretreatment of thecells with NAC almost completely blocked increase of cleaved caspase-3 and PARP induced by Cd, suggesting that induction of ROS by Cd isassociated with activation of caspase-3 signaling pathway, whichcontributes to neuronal cell death. Pretreatment of the cells with abroad-spectrum caspase inhibitor, zVAD-fmk, did prevent cleavage ofcaspase-3, but only partially prevented Cd-induced apoptosis, reveal-ing that Cd induces neuronal apoptosis through caspase-dependentand -independent mechanisms. Similar findings have been reportedin other types of cells, such as lymphoblastoid cells [56], humanembryonic kidney 293 cells [41], and cultured kidney proximal tubulecells [40]. We also found that U0126 (5μM) and SP600125 (20μM), butnot SP203580 (10μM), respectively, caused the partial inhibition ofcleaved caspase-3 with a concomitant reduction of ROS generated byCd. This may explain why U0126 and SP600125 only partially protectagainst Cd-induced neuronal cell death, as shown in our recentreport [32].

In conclusion, we have identified that Cd activates the MAPKpathway by induction of reactive oxygen species, which not onlyactivate the upstream kinases of Erk1/2 and JNK, but also inhibitnegative regulators, PP2A and PP5, leading to caspase-dependent and-independent apoptosis of neuronal cells. Our findings support thenotion that the inhibitors of JNK, Erk1/2, or antioxidants may beexploited for prevention of Cd-induced neurodegenerative diseases.

Acknowledgments

We thank Drs. Hitoshi Nakagama and Hidenori Ichijo forgenerously providing PP2A and PP5 constructs. This work wassupported in part by a Feist-Weiller Cancer Research Award (S.H.), anEdward P. Stiles Award (S.H.), and a Start-up Fund (S.H.) jointly fromLouisiana State University Health Sciences Center in Shreveport, LA.

References

[1] Baker, J. R.; Edwards, R. J.; Lasker, J. M.; Moore, M. R.; Satarug, S. Renal and hepaticaccumulation of cadmium and lead in the expression of CYP4F2 and CYP2E1.Toxicol. Lett. 159:182–191; 2005.

[2] Torra, M.; To-Figueras, J.; Rodamilans, M.; Brunet, M.; Corbella, J. Cadmium andzinc relationships in the liver and kidney of humans exposed to environmentalcadmium. Sci. Total Environ. 170:53–57; 1995.

[3] Goering, P. L.; Fisher, B. R.; Kish, C. L. Stress protein synthesis induced in rat liver bycadmium precedes hepatotoxicity. Toxicol. Appl. Pharmacol. 122:139–148; 1993.

[4] Manca, D.; Ricard, A. C.; Tra, H. V.; Chevalier, G. Relation between lipidperoxidation and inflammation in the pulmonary toxicity of cadmium. Arch.Toxicol. 68:364–369; 1994.

[5] Shukla, G. S.; Chiu, J.; Hart, B. A. Cadmium-induced elevations in the geneexpression of the regulatory subunit of gamma-glutamylcysteine synthetase in ratlung and alveolar epithelial cells. Toxicology 151:45–54; 2000.

[6] Sarkar, S.; Yadav, P.; Bhatnagar, D. Cadmium-induced lipid peroxidation and theantioxidant system in rat erythrocytes: the role of antioxidants. J. Trace Elem. Med.11:8–13; 1997.

[7] Baxter, L. C.; Sparks, D. L.; Johnson, S. C.; Lenoski, B.; Lopez, J. E.; Connor, D. J.;Sabbagh, M. N. Relationship of cognitive measures and gray and white matter inAlzheimer's disease. J. Alzheimers Dis. 9:253–260; 2006.

[8] Chuang, S. -M.;Wang, I. -C.; Yang, J. -L. Roles of JNK, p38 and ERKmitogen-activatedprotein kinases in the growth inhibition and apoptosis induced by cadmium.Carcinogenesis 21:1423–1432; 2000.

[9] Lopez, E.; Figueroa, S.; Oset-Gasque, M. J.; Gonzalez, M. P. Apoptosis and necrosis:two distinct events induced by cadmium in cortical neurons in culture. Br. J.Pharmacol. 138:901–911; 2003.

[10] Marlowe, M.; Cossairt, A.; Moon, C.; Errera, J.; MacNeel, A.; Peak, R.; Ray, J.;Schroeder, C. Main and interaction effects of metallic toxins on classroombehavior. J. Abnorm. Child Psychol. 13:185–198; 1985.

[11] Pihl, R.; Parkes, M. Hair element content in learning disabled children. Science198:204–206; 1977.

[12] Kim, S.; Moon, C.; Eun, S.; Ryu, P.; Jo, S. Identification of ASK1, MKK4, JNK, c-Jun,and caspase-3 as a signaling cascade involved in cadmium-induced neuronal cellapoptosis. Biochem. Biophys. Res. Commun. 328:326–334; 2005.

[13] Monroe, R. K.; Halvorsen, S. W. Cadmium blocks receptor-mediated Jak/STATsignaling in neurons by oxidative stress. Free Radic. Biol. Med. 41:493–502; 2006.

[14] Stohs, S. J.; Bagchi, D. Oxidative mechanisms in the toxicity of metal ions. FreeRadic. Biol. Med. 18:321–336; 1995.

[15] Johnson, S. Gradual micronutrient accumulation and depletion in Alzheimer'sdisease. Med. Hypotheses 56:595–597; 2001.

[16] Li, Z.; Arnaud, L.; Rockwell, P.; Figueiredo-Pereira, M. E. A single amino acidsubstitution in a proteasome subunit triggers aggregation of ubiquitinatedproteins in stressed neuronal cells. J. Neurochem. 90:19–28; 2004.

[17] Okuda, B.; Iwamoto, Y.; Tachibana, H.; Sugita, M. Parkinsonism after acutecadmium poisoning. Clin. Neurol. Neurosurg. 99:263–265; 1997.

[18] Panayi, A. E.; Spyrou, N. M.; Iversen, B. S.; White, M. A.; Part, P. Determination ofcadmium and zinc in Alzheimer's brain tissue using inductively coupled plasmamass spectrometry. J. Neurol. Sci. 195:1–10; 2002.

[19] Figueiredo-Pereira, M. E.; Yakushin, S.; Cohen, G. Disruption of the intracellularsulfhydryl homeostasis by cadmium-induced oxidative stress leads to proteinthiolation and ubiquitination in neuronal cells. J. Biol. Chem. 273:12703–12709;1998.

[20] Stadtman, E. Protein oxidation and aging. Science 257:1220–1224; 1992.[21] Green, K. N.; Peers, C. Divergent pathways account for two distinct effects of

amyloid beta peptides on exocytosis and Ca(2+) currents: involvement of ROS andNF-kappaB. J. Neurochem. 81:1043–1051; 2002.

[22] Kyriakis, J. M.; Avruch, J. Mammalian mitogen-activated protein kinase signaltransduction pathways activated by stress and inflammation. Physiol. Rev.81:807–869; 2001.

[23] Pearson, G.; Robinson, F.; Beers Gibson, T.; Xu, B. -e.; Karandikar, M.; Berman, K.;Cobb, M. H. Mitogen-activated protein (MAP) kinase pathways: regulation andphysiological functions. Endocr. Rev. 22:153–183; 2001.

[24] Rockwell, P.; Martinez, J.; Papa, L.; Gomes, E. Redox regulates COX-2 upregulationand cell death in the neuronal response to cadmium. Cell Signal. 16:343–353;2004.

[25] Liu, Y.; Shepherd, E. G.; Nelin, L. D. MAPK phosphatases—regulating the immuneresponse. Nat. Rev. Immunol. 7:202–212; 2007.

[26] Van Kanegan, M. J.; Adams, D. G.; Wadzinski, B. E.; Strack, S. Distinct proteinphosphatase 2A heterotrimers modulate growth factor signaling to extracellularsignal-regulated kinases and Akt. J. Biol. Chem. 280:36029–36036; 2005.

[27] Huang, S.; Shu, L.; Easton, J.; Harwood, F. C.; Germain, G. S.; Ichijo, H.; Houghton, P.J. Inhibition of mammalian target of rapamycin activates apoptosis signal-regulating kinase 1 signaling by suppressing protein phosphatase 5 activity. J.Biol. Chem. 279:36490–36496; 2004.

[28] Morita, K.; Saitoh, M.; Tobiume, K.; Matsuura, H.; Enomoto, S.; Nishitoh, H.; Ichijo,H. Negative feedback regulation of ASK1 by protein phosphatase 5 (PP5) inresponse to oxidative stress. EMBO J. 20:6028–6036; 2001.

[29] Janssens, V.; Goris, J. Protein phosphatase 2A: a highly regulated family of serine/threonine phosphatases implicated in cell growth and signalling. Biochem. J.353:417–439; 2001.

[30] Kowluru, A.; Seavey, S. E.; Rabaglia, M. E.; Nesher, R.; Metz, S. A. Carboxylmethyla-tion of the catalytic subunit of protein phosphatase 2A in insulin-secreting cells:evidence for functional consequences on enzyme activity and insulin secretion.Endocrinology 137:2315–2323; 1996.

[31] Chen, J.; Martin, B. L.; Brautigan, D. L. Regulation of protein serine-threoninephosphatase type-2A by tyrosine phosphorylation. Science 257:1261–1264;1992.

[32] Chen, L.; Liu, L.; Luo, Y.; Huang, S. MAPK and mTOR pathways are involved incadmium-induced neuronal apoptosis. J. Neurochem. 105:251–261; 2008.

[33] Liu, L., Li, F., Cardelli, J. A., Martin, K. A., Blenis, J., Huang, S., Rapamycin inhibits cellmotility by suppression of mTOR-mediated S6K1 and 4E-BP1 pathways. Oncogene25:7029-7040; 2006.

[34] Beevers, C. S.; Li, F.; Liu, L.; Huang, S. Curcumin inhibits the mammalian target ofrapamycin-mediated signaling pathways in cancer cells. Int. J. Cancer 119:757–764;2006.

[35] Aruoma, O. I.; Halliwell, B.; Hoey, B. M.; Butler, J. The antioxidant action of N-acetylcysteine: its reaction with hydrogen peroxide, hydroxyl radical, superoxide,and hypochlorous acid. Free Radic. Biol. Med. 6:593–597; 1989.

[36] Kim, J.; Sharma, R. P. Calcium-mediated activation of c-Jun NH2-terminal kinase(JNK) and apoptosis in response to cadmium in murine macrophages. Toxicol. Sci.81:518–527; 2004.

[37] Poliandri, A. H.; Cabilla, J. P.; Velardez, M. O.; Bodo, C. C.; Duvilanski, B. H.Cadmium induces apoptosis in anterior pituitary cells that can be reversed bytreatment with antioxidants. Toxicol. Appl. Pharmacol. 190:17–24; 2003.

[38] Kim, J.; Sharma, R. P. Cadmium-induced apoptosis in murine macrophages isantagonized by antioxidants and caspase inhibitors. J. Toxicol. Environ. Health A69:1181–1201; 2006.

[39] Kim, M. S.; Kim, B. J.; Woo, H. N.; Kim, K. W.; Kim, K. B.; Kim, I. K.; Jung, Y. K.Cadmium induces caspase-mediated cell death: suppression by Bcl-2. Toxicol.145:27–37; 2000.

[40] Lee, W. K.; Abouhamed, M.; Thevenod, F. Caspase-dependent and -independentpathways for cadmium-induced apoptosis in cultured kidney proximal tubulecells. Am. J. Physiol. Renal Physiol. 291:F823–F832; 2006.

[41] Mao, W. P.; Ye, J. L.; Guan, Z. B.; Zhao, J. M.; Zhang, C.; Zhang, N. N.; Jiang, P.; Tian, T.Cadmium induces apoptosis in human embryonic kidney (HEK) 293 cells bycaspase-dependent and -independent pathways acting on mitochondria. Toxicol.In Vitro 21:343–354; 2007.

[42] Arriazu, R.; Pozuelo, J. M.; Henriques-Gil, N.; Perucho, T.; Martin, R.; Rodriguez, R.;Santamaria, L. Immunohistochemical study of cell proliferation, Bcl-2, p53, andcaspase-3 expression on preneoplastic changes induced by cadmium and zincchloride in the ventral rat prostate. J. Histochem. Cytochem. 54:981–990; 2006.

[43] Pham, T. N.; Marion, M.; Denizeau, F.; Jumarie, C. Cadmium-induced apoptosis inrat hepatocytes does not necessarily involve caspase-dependent pathways. Toxi-col. In Vitro 20:1331–1342; 2006.

1044 L. Chen et al. / Free Radical Biology & Medicine 45 (2008) 1035–1044

[44] Shih, Y. L.; Lin, C. J.; Hsu, S. W.; Wang, S. H.; Chen, W. L.; Lee, M. T.; Wei, Y. H.; Shih,C. M. Cadmium toxicity toward caspase-independent apoptosis through themitochondria-calcium pathway in mtDNA-depleted cells. Ann. N. Y. Acad. Sci.1042:497–505; 2005.

[45] Yuan, C.; Kadiiska, M.; Achanzar, W. E.; Mason, R. P.; Waalkes, M. P. Possible role ofcaspase-3 inhibition in cadmium-induced blockage of apoptosis. Toxicol. Appl.Pharmacol. 164:321–329; 2000.

[46] Sommer, D.; Coleman, S.; Swanson, S. A.; Stemmer, P. M. Differential suscept-ibilities of serine/threonine phosphatases to oxidative and nitrosative stress. Arch.Biochem. Biophys. 404:271–278; 2002.

[47] Bogeski, I.; Bozem, M.; Sternfeld, L.; Hofer, H. W.; Schulz, I. Inhibition of proteintyrosine phosphatase 1B by reactive oxygen species leads to maintenance ofCa2+ influx following store depletion in HEK 293 cells. Cell Calcium 40:1–10;2006.

[48] Shimohama, S.; Tanino, H.; Fujimoto, S. Differential expression of rat brain caspasefamily proteins during development and aging. Biochem. Biophys. Res. Commun.289:1063–1066; 2001.

[49] Yakovlev, A. G.; Ota, K.;Wang, G.; Movsesyan, V.; Bao,W. L.; Yoshihara, K.; Faden, A.I. Differential expression of apoptotic protease-activating factor-1 and caspase-3genes and susceptibility to apoptosis during brain development and aftertraumatic brain injury. J. Neurosci. 21:7439–7446; 2001.

[50] De Bilbao, F.; Guarin, E.; Nef, P.; Vallet, P.; Giannakopoulos, P.; Dubois-Dauphin, M.Postnatal distribution of cpp32/caspase 3 mRNA in the mouse central nervoussystem: an in situ hybridization study. J. Comp. Neurol. 409:339–357; 1999.

[51] Jellinger, K. A. Cell death mechanisms in Parkinson's disease. J. Neural. Transm.107:1–29; 2000.

[52] Jordan, J.; Galindo, M. F.; Cena, V.; Gonzalez-Garcia, C. Cysteine proteinase andneurodegeneration. Rev. Neurol. 31:333–340; 2000.

[53] Rideout, H. J.; Stefanis, L. Caspase inhibition: a potential therapeutic strategy inneurological diseases. Histol. Histopathol. 16:895–908; 2001.

[54] Stadelmann, C.; Deckwerth, T. L.; Srinivasan, A.; Bancher, C.; Bruck, W.; Jellinger,K.; Lassmann, H. Activation of caspase-3 in single neurons and autophagicgranules of granulovacuolar degeneration in Alzheimer's disease. Evidence forapoptotic cell death. Am. J. Pathol. 155:1459–1466; 1999.

[55] Hartmann, A.; Hunot, S.; Michel, P. P.; Muriel, M. P.; Vyas, S.; Faucheux, B. A.;Mouatt-Prigent, A.; Turmel, H.; Srinivasan, A.; Ruberg, M.; Evan, G. I.; Agid, Y.;Hirsch, E. C. Caspase-3: a vulnerability factor and final effector in apoptotic deathof dopaminergic neurons in Parkinson's disease. Proc. Natl. Acad. Sci. USA97:2875–2880; 2000.

[56] Coutant, A.; Lebeau, J.; Bidon-Wagner, N.; Levalois, C.; Lectard, B.; Chevillard, S.Cadmium-induced apoptosis in lymphoblastoid cell line: involvement of caspase-dependent and -independent pathways. Biochimie 88:1815–1822; 2006.