The influence of glutathione on redox regulation by antioxidant proteins and apoptosis in macrophages exposed to 2-hydroxyethyl methacrylate (HEMA)

at SciVerse ScienceDirect

Biomaterials 33 (2012) 5177e5186

Contents lists available

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

The influence of glutathione on redox regulation by antioxidant proteins andapoptosis in macrophages exposed to 2-hydroxyethyl methacrylate (HEMA)

Stephanie Krifka a, Karl-Anton Hiller a, Gianrico Spagnuolo b, Anahid Jewett c, Gottfried Schmalz a,Helmut Schweikl a,*aDepartment of Operative Dentistry and Periodontology, University Hospital Regensburg, D-93042 Regensburg, GermanybDepartment of Oral and Maxillofacial Science, University of Naples “Federico II”, Italyc The Jane and Jerry Weintraub Center for Reconstructive Biotechnology, UCLA School of Dentistry, University of California-Los Angeles, CA 90095-1668, USA

a r t i c l e i n f o

Article history:Received 23 March 2012Accepted 2 April 2012Available online 24 April 2012

Keywords:Resin monomerHEMAReactive oxygen speciesGlutathioneAntioxidant proteinApoptosis

* Corresponding author. Fax: þ49 941 944 6025.E-mail address: [email protected]

0142-9612/$ e see front matter � 2012 Elsevier Ltd.doi:10.1016/j.biomaterials.2012.04.013

a b s t r a c t

Resin monomers like 2-hydroxyethyl methacrylate (HEMA) disturb cell functions including responses ofthe innate immune system, mineralization and differentiation, or induce cell death via apoptosis. Thesephenomena are associated with oxidative stress and a reduction in the concentration of the antioxidantglutathione (GSH), resulting in imbalanced redox homeostasis. Thus far, the precise mechanism of howresin monomers interfere with cellular redox regulation is unknown. The present study provides insightinto the induction of apoptosis and the differential expression of antioxidant enzymes depending onthe availability of GSH. Buthionine sulfoximine (BSO) was used to inhibit GSH synthesis, while2-oxothiazolidine-4-carboxylate (OTC), and N-acetylcysteine (NAC) as prodrugs supported GSH synthesisin RAW264.7 mouse macrophages exposed to HEMA (0e8 mM) for 24 h. The level of GSH was signifi-cantly decreased after cells were preincubated with BSO, and the formation of reactive oxygen species(ROS) increased in cultures subsequently exposed to HEMA. Apoptosis was drastically increased by BSOin HEMA-exposed cell cultures as well, but OTC and NAC retracted HEMA-induced cell death. Theseresults show that dental monomer-induced apoptosis is causally related to the availability of GSH. Thehydrogen peroxide decomposing enzymes glutathione peroxidase (GPx1/2) and catalase were differ-entially regulated in HEMA-exposed cultures. Expression of GPx1/2 was inhibited by HEMA and furtherreduced in the presence of BSO. SOD1 (superoxide dismutase) expression was inhibited in the presenceof HEMA, and was decreased to an even greater extent by BSO, possibly due to H2O2-feedback inhibition.The expression of catalase was considerably up-regulated in HEMA-exposed cultures, implying that H2O2

is the type of ROS that is significantly increased in monomer-exposed cells. OTC and NAC counteractedthe effect of HEMA on GPx1/2, SOD1, and catalase expression. HO-1 (heme oxygenase) expression wasstrongly enhanced by HEMA, suggesting the need for further antioxidants like bilirubin to supportenzyme activities that directly regulate H2O2 equilibrium. Expression of the oxidoreductase thioredoxin(TRX1), the second major thiol-dependent antioxidant system in eukaryotic cells, was slightly reduced,while the oxygen-sensing protein HIF-1a was downregulated in HEMA-exposed cell cultures. Theseresults indicate that cells and tissues actively respond to monomer-induced oxidative stress by thedifferential expression of enzymatic antioxidants.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

Resinmonomers including 2-hydroxyethylmethacrylate (HEMA)and triethylene glycol dimethacrylate (TEGDMA) are well-knowncomponents of composite resins and adhesives in dentistry, butbone cements and scaffolds for tissue regeneration may also incor-porate methacrylates [1,2]. Due to insufficient monomer-polymerconversion, residual monomers of composite materials leach into

sburg.de (H. Schweikl).

All rights reserved.

their surrounding aqueous environment and accordingly act onadjacent tissues [3]. Contact allergies have been observed clinically,but unreacted monomers may also affect tissues of the oral cavityincluding the dental pulp, either immediately in the case of directpulp capping procedures, or eventually possibly diffuse fromrestorative composite materials through dentinal tubules. Based onexperiments from in vitro cell culture experiments ofmultiple targetcells, resin monomers specifically interfere with various cellularfunctions, for instance, by inhibiting cytokine production, minerali-zation, cell differentiation, or inducing apoptosis related to themodification of signal transduction pathways and expression of

S. Krifka et al. / Biomaterials 33 (2012) 5177e51865178

transcription factors [4e9]. Current findings strongly suggest thatthemechanismbehind these specific cell responses is the generationof oxidative stress. It has been firmly established that resin mono-mers cause a depletion of the intracellular antioxidant glutathione(GSH) level, while in parallel increasing the formation of reactiveoxygen species (ROS) [10].

ROS are generated under normal physiological conditions or asa consequence of endogenous or exogenous stimuli includingcytokines, growth factors, bacterial endotoxins, radiation, orchemicals. Besides short-lived superoxide anions (O�

2 ) and hydroxylradicals (OH$), balanced concentrations of hydrogen peroxide(H2O2) in particular are considered physiological regulators [11].Yet, high concentrations of ROS beyond the capacity of the cellularredox balance cause oxidative damage to cellular macromoleculeslike DNA, proteins, and lipids, or impair cellular signaling resultingin apoptotic cell death [12]. To maintain redox homeostasis andvital cell functions, the dynamic equilibrium of ROS is tightlyregulated by non-enzymatic and enzymatic antioxidants [13]. Theprincipal non-enzymatic antioxidant GSH is a tripeptide at whichthe sulfhydryl group (�SH) of cysteine confers its redox-relevantfunctionality. GSH serves as a substrate of glutathione peroxidaseto catalyze the reduction of H2O2 into H2O, and is indispensable insustaining the essential thiol status of proteins. Increasing GSHformation, and thereby a shift to a more reduced intracellularenvironment, is assumed to induce cell proliferation and survival[14e16].

In view of the analysis of dental resin monomer-induced cellresponses, presumably based on oxidative stress and redox regu-lation, the role of GSH content and synthesis in particular areconsidered feasible targets in the present investigation. Moreover,the GSH pathway may be relevant for clinical intervention as well.For instance, L-buthionine sulfoximine (BSO) selectively inhibits thefirst step enzyme of GSH synthesis thus leading to GSH depletion[17]. On the other hand, precursors of the amino acid cysteine like2-oxothiazolidine-4-carboxylate (OTC) or N-acetylcysteine (NAC)should fuel GSH synthesis under stress conditions [18]. In additionto being a derivate of cysteine, NAC per se is an intracellular sourceof sulfhydryl groups and, therefore, may act as a direct ROS scav-enger [19]. These functions related to redox regulation mightexplain the protective role of NAC in a variety of monomer-inducedcellular oxidative stress responses [10].

The non-enzymatic GSH is part of the cellular redox homeo-stasis regulated by tightly balanced and concerted activities ofantioxidant enzymes [20]. Superoxide dismutase (SOD1) catalyzesthe dismutation of superoxide anions (O�

2 ) to hydrogen peroxide(H2O2) and molecular oxygen (O2). The resulting H2O2 levels arecontrolled by either catalase, which degrades H2O2 into O2 andwater, or glutathione peroxidase (GPx1/2), which uses GSH as anelectron donor [21]. Thioredoxin-1 (TRX1) and TRX-dependentperoxidase are a supplementary mechanism for reducing H2O2concentrations. Moreover, it is ubiquitously present as a substrateof TRX reductase plus NADPH as a cofactor for modulating theintracellular redox state by reducing oxidized thiol groups onproteins [22]. NADPH is a link to inducible heme oxygenase-1(HO-1), which finally degrades heme, a stable prosthetic groupin hemoproteins including cytochromes P450 or catalase, into theantioxidant bilirubin [23]. Therefore, it is possible that HO-1expression, a phase II enzyme not directly related to ROSmetabolism, might increase in cases of monomer-induced oxida-tive stress. Furthermore, the commonly expressed hypoxia-inducible factor-1 (HIF-1a) is an oxygen-sensing protein, whichis either stabilized during hypoxia or rapidly degraded under non-hypoxic conditions [24]. Noteworthy is that the promoter regionof human HO-1 contains AREs (antioxidant response elements) aswell as hypoxia-response elements (HREs) motifs which allow for

the binding of HIF-1a and suggests a role of HIF-1a in HO-1expression [25].

Our recent finding that a dental resin monomer triggered theexpression of stress-responsive genes at the transcriptional levelthrough the accumulation of ROS was a major step forward in theanalysis and understanding of the adaptive responses of potentialtarget cells to monomer-induced oxidative stress. We furtherconcluded that monomer-induced apoptosis was an active cellresponse to levels of ROS exceeding the cells abilities to maintainredox homeostasis [28]. Consequently, we hypothesized in thepresent study that, first, cells differentially activate a balancednetwork of enzymatic cellular antioxidants as a response to controlthe intracellular oxidative state after exposure to the HEMAmonomer. To this end we analyzed the adaptive expression ofcytosolic enzymes either directly involved in the metabolism ofROS like SOD1, catalase, and GPx1/2, or else related to mechanismswhich increase the cells’ antioxidant state like TRX1, HO-1, or HIF-1a. Second, we assumed that the major non-enzymatic antioxidantGSH influences the expression of antioxidant enzymes as well asthe apoptotic cell response. Therefore, we modified the amounts ofintracellular GSH using BSO, OTC, or NAC to specifically inhibit orsupport GSH synthesis in RAW264.7 mouse macrophages used asa model to study cell responses of the innate immune system [6].

2. Materials and methods

2.1. Chemicals and reagents

2-Hydroxyethyl methacrylate (HEMA; CAS-No. 868-779) was purchased fromMerck (Darmstadt, Germany). L-buthionine sulfoximine (BSO; CAS-No. 83730-53-4),(R)-(-)-2-oxothiazolidine-4-carboxylic acid (OTC; CAS-No. 19771-63-2), N-ace-tylcysteine (NAC; CAS-No. 616-91-1), C12E10 (decaethylene glycol monododecylether (CAS-No. 6540-99-4), and a glutathione assay kit (no. CS0260) came fromSigmaeAldrich (Taufkirchen, Germany). RPMI 1640 medium containing L-glutamineand 2.0 g/l NaHCO3 was from PAN Biotech (Aidenbach, Germany). Fetal bovineserum (FBS), penicillin/streptomycin, and phosphate-buffered saline supplementedwith 5 mM EDTA (PBS-EDTA) were purchased from Life Technologies, Gibco BRL(Eggenstein, Germany). The FACS Annexin V-FITC apoptosis detection kit wasobtained from R&D Systems (Minneapolis, MN, USA). Crystal violet came from GibcoInvitrogen (Karlsruhe, Germany) and 2070-dichlorodihydrofluorescin diacetate(H2DCF-DA) was purchased from MoBiTec (Göttingen, Germany).

Anti-catalase (H-300, sc-50508), anti-heme oxygenase-1 (HO-1, M-19, sc-1797)and anti-HIF-1a (H-206, sc-10790) polyclonal antibodies as well as anti-Cu-Znsuperoxide dismutase (SOD1, B-1, sc-271014) and anti-glutathione peroxidase 1/2(GPx1/2, D-12, sc-133152) monoclonal antibodies came from Santa Cruz Biotech-nology (Santa Cruz, CA, USA). Anti-thioredoxin 1 (TRX1) (no. 2298) and anti-rabbitIgG HRP-linked antibody (no. 7074) were obtained from Cell Signaling (NEBFrankfurt, Germany), goat anti-mouse IgG (Hþ L)-HRP conjugate was obtained fromBio-Rad Laboratories (Munich, Germany), and Amersham hyperfilm ECL came fromGE Healthcare (Munich, Germany). The protease inhibitor cocktail (complete mini)was obtained from Roche Diagnostics (Mannheim, Germany), anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) monoclonal antibody (clone 6C5) andCHEMICON re-blot plus mild antibody stripping solution came from Millipore(Schwalbach, Germany). All other chemicals used in the present study were at leastchemical grade.

2.2. Cell culture

RAW264.7 mouse macrophages (ATCC TIB71) were incubated with RPMI 1640medium containing L-glutamine, sodium-pyruvate and 2.0 g/l NaHCO3 supple-mented with 10% fetal bovine serum (FBS) and penicillin-streptomycin at 37 �C and5% CO2.

2.3. Determination of glutathione (GSH) content

RAW264.7 mouse macrophages (2 � 106 cells) were cultured in cell cultureplates (150 mm in diameter) at 37 �C for 24 h and then preincubated with 50 mM BSOor 5mM OTC for 20 h. The cells were exposed to NAC in a concentration of 10mM (pH7.2) for 1 h. Four replicate cell cultures were analyzed in repeated independentexperiments. The concentrations of BSO, OTC, and NAC were selected after testinga wide concentration range in preliminary experiments (not shown). After dis-carding the exposuremedia, the cells were washed with ice-cold PBS, detached withPBS/5 mM EDTA and collected by centrifugation. After being resuspended andwashed twice in ice-cold PBS, 1�106 cells were lysed in 50 mM TriseCl (pH 7.4), 0.1%

S. Krifka et al. / Biomaterials 33 (2012) 5177e5186 5179

C12E10, 150 mM NaCl and a protease inhibitor cocktail (complete mini) for 10 min onice. After centrifugation at 14,000 � g for 10 min the supernatant was collected. Aglutathione assay kit (Sigma) was used to measure the level of total glutathione.Serially diluted GSH standards and 10 ml aliquots of cell lysates were pipetted intoseparate wells of a 96-well plate and 150 ml working mixture (100 mM potassiumphosphate buffer, pH 7.0, with 1 mM EDTA, 6 units/ml glutathione reductase,1.5 mg/ml DNTB (5,50dithiobis[2-nitrobenzoic acid]) was added. After a 5 minincubation period at room temperature, 50 ml NADPH (0.16 mg/ml) was added toeach well and optical densities were determined after 30 min at 405 nm in a multi-well spectrophotometer (TECAN, Infinite 200 PRO, Männedorf, Switzerland). Thisincubation period was found to be optimal in preliminary experiments (not shown).In parallel, the amount of protein in each cell lysatewas determined by a BCA proteinassay (Sigma) using bovine serum albumin as a standard, and optical densities werespectrophotometrically analyzed at 540 nm (TECAN, Infinite 200 PRO). GSHconcentrations and amounts of total protein were calculated from individual valuesof optical density readings using GSH and protein standards, respectively.

2.4. Determination of apoptosis

RAW264.7 mouse macrophages (1 � 105 cells/well) were cultivated in 6-wellplates at 37 �C for 24 h. Next, the culture medium was removed and cell cultureswere subsequently preincubated with either 50 mM BSO or 5 mM OTC for 20 h,whereas NAC (10 mM), adjusted to pH 7.2, was added to cell cultures 1 h previous toexposure to HEMA. Then, preincubation media were removed and the cells weretreated with HEMA (0e6e8 mM) in the presence or absence of BSO, OTC, or NAC.Exposure was stopped by discarding the medium after 24 h, and cell cultures werewashedwith phosphate-buffered saline (PBS) at room temperature. Accordingly, thecells were detached with PBS/5 mM EDTA, washed in PBS, and finally collected bycentrifugation. Apoptotic cell death was analyzed after staining cells with annexinV-FITC and propidium iodide (PI) as previously described [27]. Therefore,1.5 � 105e1 �106 cells per cell culture were incubated in 100 ml binding buffer withannexin V-FITC and PI, and determined by flow cytometry (FACSCanto, BectonDickinson, San Jose, CA, USA). FITC fluorescence (FL-1) was analyzed by a 530/30band pass filter, and PI fluorescence (Fl-3) by a 650 nm long pass filter. Dataacquisition (at least 2 � 104 events for each sample) was performed with theFACSDiva� 5.0.2 software which was also used for analysis. The numbers of viable(annexin V�; PI�) cells were detected in the lower left quadrant (unstained) ofdensity plots, and the percentages of cells in apoptosis (annexin Vþ; PI�; lower rightquadrant), late apoptosis (annexin Vþ; PIþ; upper right quadrant), and necrosis(annexin V�; PIþ; upper left quadrant) were determined as well.

2.5. Detection of reactive oxygen species (ROS)

The intracellular level of ROS in RAW264.7mouse macrophages was determinedusing the oxidation-sensitive fluorescent probe 2070-dichlorodihydrofluorescindiacetate (H2DCF-DA) as described earlier [26]. The cells were cultured in 6-wellplates (5 � 104/well) for 24 h at 37 �C. Next, cells were preincubated with 50 mMBSO or 5 mM OTC for 20 h or with 10 mM NAC for 1 h, and then exposed to HEMA(0e6e8 mM) in the presence or absence of 50 mM BSO, 5 mM OTC, or 10 mM NAC for4 h at 37 �C. This exposure period was selected as being optimal after kinetic studies(30 mine24 h) in preliminary experiments (not shown). Next, the cells were incu-bated with 2.5 ml H2DCF-DA (10 mmol/l) in complete medium prior to harvestingusing PBS/5 mM EDTA. Subsequently, cells were collected by centrifugation, washedwith PBS and resuspended in 200 ml CMFePBS. Finally, DCF fluorescence intensitywas measured by flow cytometry (FACSCanto, Becton Dickinson) at an excitationwavelength of 495 nm and an emission wavelength of 530 nm (Fl-1). Mean fluo-rescence intensities were obtained by histogram statistics using the FACSDiva�5.0.2 software. Individual values of fluorescence intensities were normalized tofluorescence detected in untreated control cultures (¼1.0).

2.6. Analysis of the expression of antioxidant enzymes

RAW264.7 mouse macrophages (1.5 � 106 cells) were cultured in cell cultureplates (150 mm in diameter) at 37 �C for 24 h. The cells were preincubated with50 mM BSO, 5 mM OTC, or 10 mM NAC, and subsequently exposed to HEMA(0e6e8 mM) for 24 h as described above. Then, exposure was stopped by collectingthe exposure media and floating cells. Adherent cells were washed with ice-coldPBS, detached with PBS/5 mM EDTA and collected by centrifugation. The cell pelletwas resuspended and washed twice in ice-cold PBS. Next, the cells were lysed in50 mM Tris-Cl (pH 7.4), 0.1% Nonidet NP-40, 0.1% C12E10, 150 mM NaCl, 2 mM EDTA,5 mM NaF, 1 mM NaVO3, and a protease inhibitor cocktail (complete mini) for 10 minon ice. After centrifugation at 14,000 � g for 4 min, the supernatant was collectedand the amount of protein present in each cell lysate was determined by a BCAprotein assay (Sigma) using bovine serum albumin as a standard.

2.7. Western blot analysis

Proteins (20 mg per lane) were first separated on a 12% sodium dodecyl sulfate-polyacrylamide gel by electrophoresis (SDS-PAGE), and transferred to a nitrocellulose

membrane in SDSeelectroblot buffer (25 mM TrisCl, 192 mM glycin, 20% methanol, pH8.3) at 350 mA for 60 min. The membrane was then washed twice in TBS (25 mM

TriseCl, 150 mM NaCl, pH 7.4) and blocked with 5% nonfat milk in TBST (TBS plus 0.1%Tween 20, pH7.4) or bovine serumalbumin (detectionofGPx1/2) at room temperaturefor 60 min. Accordingly, the membrane was incubated with antibodies specific for thedetection of SOD1, catalase, GPx1/2, HO-1, TRX1, or HIF-1a over night at cold temper-atures. Themembranewas thenwashedwithTBSTthree timesat roomtemperature for10 min and primary antibodies were detected by horseradish peroxidase-conjugatedsecondary antibodies in TBST for 60 min. Bound secondary antibodies were visual-ized by enhanced chemiluminescence (ECL) after washing for 20 min in TBST and10min in PBS. Then, themembraneswere stripped for reprobing at room temperature(15min) using an antibody stripping solution. Finally, the membraneswerewashed inPBS at room temperature for 30min, reprobedwith anti-glyceraldehyde-3-phosphatedehydrogenase (GAPDH) antibody, and the bound secondary antibody was visualizedby ECL. To further illustrate distinct expression of antioxidant proteins, immunoblotswere analyzed by densitometry using image analysis software (Optimas Corp., version6.2, Bothell, WA, USA). Positive signals of antioxidant proteins were related to corre-spondingGAPDHsignals. This relationwasset to1 foruntreatedcell cultures (medium),and the fold change of antioxidant protein expression in treated cell cultures wascalculated accordingly.

2.8. Statistical analyses

Data are presented as medians (25e75% quartiles) summarized from individualvalues in series of independent experiments (SigmaPlot 12.0, Systat Software, SanJose, CA, USA) as indicated in detail in the figure legends. The statistical analyses ofthe differences between median values were performed using the ManneWhitneyU-test (SPSS 19.0, SPSS, Chicago, IL, USA) for pairwise comparisons among groups.The level of significance was set at a ¼ 0.05.

3. Results

3.1. The role of glutathione in HEMA-induced apoptosis

According to the current paradigm, resin monomer-induced celldeath via apoptosis is associated with an increase in ROS produc-tion and depletion of the intracellular GSH content. To furtherelucidate the precise role of GSH, HEMA-induced apoptosis inRAW264.7 mouse macrophages was analyzed here in the presenceand absence of GSH synthesis-modulating substances such asbuthionine sulfoximine (BSO), an inhibitor of the GSH synthesis, 2-oxo-4-thiazolidine-carboxylic acid (OTC), which supports GSHsynthesis, as well as N-acetylcysteine (NAC). To this end, theintracellular GSH level was determined directly after preincubatingcell cultures to analyze for the effects of the GSH synthesis-modulating substances BSO, OTC, and NAC (Fig. 1). Whileamounts of protein remained constant, the level of total GSHdrastically decreased after cell cultures were preincubated withBSO for 20 h. About 9 nmol GSH/mg protein was found in BSO-exposed macrophages compared to 66 nmol GSH/mg proteindetected in untreated cultures (p ¼ 0.000). As expected, no signif-icant changes in total GSH levels were observed in cell culturesexposed to OTC or NAC (Fig. 1).

The influence of HEMA and BSO on the induction of apoptosis asdetected by flow cytometry is shown in representative density blots(Fig. 2). Untreated cell cultures and cultures exposed to BSO alonehardly showed any signs of apoptotic cell death since more than90% of the cells remained viable (unstained). On the other hand,exposure of cell cultures to 8 mM HEMA reduced the percentage ofviable cells to about 61%, and percentages of cells in the variousphases of cell death, mostly in late apoptosis, were analogouslyincreased. HEMA-induced cell death was drastically amplified inthe presence of BSO, and the percentage of viable cells declined byhalf in parallel (Fig. 2).

Fig. 3 represents a summary of the analyses of cell cultures fromrepeated experiments. A dose-dependent decrease in thepercentage of viable cells was observed in RAW264.7 mousemacrophages exposed to HEMA (Fig. 3AeC). The number of viablecells was reduced from about 98% in untreated cultures to about90% in cultures exposed to 6 mM HEMA and about 60% after

Fig. 2. Induction of apoptosis and necrosis in RAW264.7 mouse macrophages. Representative dual-parameter fluorescence density blots were derived from untreated cell culturesand cell cultures exposed to 8 mM HEMA in the presence or absence of 50 mM BSO. After a 24 h exposure period, cell cultures were stained with annexin V-FITC (Annexin)/propidiumiodide (PI) and then analyzed by flow cytometry. Percentages of viable cells (unstained), and cells in apoptosis (Annexin), late apoptosis (Annexin þ PI) and necrosis (PI) of onetypical experiment are denoted in the quadrants of each density blot.

Fig. 1. Glutathione levels in RAW264.7 mouse macrophages. Cell cultures were exposed to medium (UC), 50 mM BSO or 5 mM OTC for 20 h or 10 mM NAC for 1 h. Bars representmedian values (plus 25% and 75% percentiles) of protein concentration (mg/ml) and glutathione levels (nmol/mg). Data were calculated from nine independent experiments (n ¼ 9).* ¼ significant differences between protein concentrations found in untreated controls and cell cultures exposed to BSO, OTC, or NAC; a ¼ significant differences between GSH levelsin the presence or absence of BSO, OTC, or NAC.

S. Krifka et al. / Biomaterials 33 (2012) 5177e51865180

Fig. 3. Induction of apoptosis in RAW264.7 mouse macrophages. Cell cultures wereexposed to HEMA (0e6e8 mM) in the presence or absence of 50 mM BSO (A), 5 mM

OTC (B), or 10 mM NAC (C) for 24 h and analyzed by flow cytometry after stainingwith annexin V-FITC and propidium iodide (PI). Bars represent percentages of viablecells (unstained), cells in apoptosis (Annexin), late apoptosis (Annexin þ PI)and necrosis (PI) as medians (25% and 75% percentiles) calculated from five(n ¼ 5) independent experiments. * ¼ significant differences between median values

S. Krifka et al. / Biomaterials 33 (2012) 5177e5186 5181

exposure to 8 mM HEMA, whereas the percentage of cells mostly inlate apoptosis (annexin þ PI) increased simultaneously (Fig. 3A).Moreover, the presence of BSO further increased the number ofcells in late apoptosis 2-fold to about 60% when cultures weretreated with 8 mM HEMA, and the percentage of viable cells wasreduced to 32% (Fig. 3A). Quite in contrast, the presence of OTC in8 mM HEMA-treated cell cultures significantly (p¼0.026) enhancedthe percentage of viable cells from about 60% tomore than 70%, andminimized the percentage of RAW264.7 mouse macrophages inlate apoptosis in parallel (Fig. 3B). Exposure of cell cultures to NACalone neither influenced the percentage of viable cells nor reducedcells in apoptosis and late apoptosis (Fig. 3c). However, the pres-ence of NAC in cell cultures exposed to HEMA significantly atten-uated the decrease in cell viability. Cell viability was considerablyincreased from about 60% in cultures exposed to 8 mM HEMA aloneto more than 90% in the presence of NAC (p¼ 0.008). Consequently,NAC significantly reduced the percentage of cells, for instance, inlate apoptosis from almost 40%e8% (p ¼ 0.008).

3.2. Generation of reactive oxygen species (ROS)

The formation of ROS in RAWmousemacrophages was analyzedafter cell cultures were exposed to HEMA in the presence andabsence of BSO, OTC, or NAC for 4 h to identify differential effects ofthe GSH-modifying substances as well (Fig. 4). First, BSO alonesignificantly increased the level of ROS more than 2-fold(p ¼ 0.000), whereas OTC slightly enhanced ROS productioncompared to untreated cell cultures. In contrast, the amount of ROSwas reduced to less than 50% in cell cultures exposed to NAC alone(p ¼ 0.000). Both 6 and 8 mM HEMA significantly increased ROSproduction about 1.5 and 1.8-fold (p ¼ 0.000). Moreover, theformation of ROS by 6 mM HEMAwas even further enhanced in thepresence of BSO, and had most likely reached saturation since noadditional increase was detected with 8 mM HEMA (Fig. 4).Increased ROS formation in cells exposed to HEMAwas observed inthe presence of OTC. Although not completely inhibited, theamount of ROS remained at low levels in cultures treated withHEMA in the presence of NAC.

3.3. HEMA-induced expression of antioxidant proteins

The induction of apoptosis by HEMA as shown here wassignificantly influenced by substances which modify the amount ofthe intracellular non-enzymatic antioxidant GSH. Since GSH servesas an electron donor to reduce hydrogen peroxide by glutathioneperoxidase (GPx1/2) we further analyzed the adaptive and coor-dinated expression of various directly and indirectly acting anti-oxidant enzymes in RAW264.7 mouse macrophages after exposureto HEMA. The substances which modulate GSH synthesis and alsochange the intracellular redox state, such as NAC, clearly influencedprotein expression as detected by Western blotting (Fig. 5A). Thesechanges in the differential protein expression are presented forclarity reasons as well (Fig. 5B).

GPx1/2, which decomposes H2O2, was constitutively expressedin untreated RAW264.7mousemacrophages (Fig. 5A). However, thelevel of GPx1/2 was not increased but dose-dependently reducedin cell cultures exposed toHEMA. Likewise, preincubation of HEMA-treated cultureswithBSO lowered the amount of GPx1/2 evenmore,

obtained in cultures exposed to HEMA (0e6e8 mM) in the presence of BSO, OTC, orNAC. a ¼ significant differences between median values obtained in control cultures(0 mM HEMA) and cultures exposed to 6 or 8 mM HEMA; b ¼ significant differencesbetween median values obtained in cultures exposed to HEMA in the presence ofBSO, OTC, or NAC.

Fig. 4. Generation of reactive oxygen species (ROS) in RAW264.7 mouse macrophages.Cells were exposed to HEMA alone or in the presence of 50 mM BSO, 5 mM OTC, or10 mM NAC for 4 h. Bars represent median values (plus 25% and 75% percentiles)calculated from seven independent experiments (n ¼ 7). Significant differencesbetween median values obtained in control cultures and cultures exposed to 6 mM

HEMA (*) or 8 mM HEMA (**) in the presence or absence of BSO, OTC, or NAC.Significant differences between median values obtained in untreated cell cultures andcultures exposed to HEMA alone (a), 6 mM HEMA (b) or 8 mM HEMA (c) in the presenceor absence of BSO, OTC, or NAC.

S. Krifka et al. / Biomaterials 33 (2012) 5177e51865182

whereas BSO alone increased GPx1/2 expression. Moreover, anincrease in GPx1/2 expression induced by OTC in untreated cellcultures was again reduced by HEMA. Similar but weaker effectswere detected in macrophages exposed to NAC (Fig. 5A,B). Inter-estingly, the expression of catalase was not detectable in untreatedcell cultures, and clearly measurable amounts of catalase were alsoabsent in cultures exposed to OTC, BSO, or NAC (Fig. 5A). In contrast,HEMA alone drastically increased the expression of catalase morethan 3-fold. Although the presence of BSO had only little effect oncatalase expression, the amount of catalase was slightly reduced byOTC in cells treatedwithHEMA, and considerably lowered inHEMA-exposed cultures after preincubation with NAC (Fig. 5A,B). Super-oxide dismutase (SOD1) was found in untreated cell cultures, andthe expression was reduced by half after exposure to 8 mM HEMA(Fig. 5A,B). Although SOD1 expression increased in cultures exposedtoBSOalone, onlymarginal amounts of SOD1were found in culturesexposed toHEMAafter preincubationwith BSO.While OTC had onlylittle influence on SOD1 expression, NAC drastically increased thelevels of SOD1 in untreated cultures as well as in cells exposed toHEMA (Fig. 5A,B).

Enzymes, in particular heme oxygenase (HO-1), indirectlyrelated to ROS metabolism were differentially expressed in HEMA-treated cell cultures as well. The expression of HO-1, which wasabsent in untreated cells, was drastically increased in culturesexposed to HEMA. BSO alone enhanced the expression of HO-1, andthis amount increased even further when cell cultures were addi-tionally exposed to HEMA. In contrast, HO-1 expression wasmarkedly reduced in cultures preincubated with OTC and NAC(Fig. 5A,B). In addition to GSH, the thioredoxin system is the secondmajor thiol-dependent antioxidant system in eukaryotic cells. TRX1expression detectable in untreated cell cultures did not consider-ably change in the presence of HEMA, and BSO only slightlydecreased TRX1 (Fig. 5A,B). A slight decrease in TRX1 expression inthe presence of OTC or NAC alone was absent in HEMA-treated cellcultures (Fig. 5A,B). Furthermore, exposure of cell cultures to HEMAreduced the expression of HIF-1a detected in untreated cells, andthe inhibitory effect of HEMAwas enhanced in the presence of BSO.In contrast, OTC and NAC counterbalanced HEMA-induced inhibi-tion of HIF-1a expression (Fig. 5A,B).

4. Discussion

The cellular response in the state of oxidative stress is crucial tomaintain redox functions, thus the balance is carefully regulated byvarious protective mechanisms consisting of non-enzymatic andenzymatic antioxidants [21,28]. Dental resin monomers, includingthe dimethacrylate TEGDMA or HEMA as a monofunctional meth-acrylate, are reported to be external sources of oxidative stress,among others noted by an increase in the formation of ROS and thedepletion of GSH [10]. In conjunction with the disturbed intracel-lular redox environment, a differential activation of signalingcascades through mitogen-activated protein kinases (MAPK) andtranscription factors downstream could be associated withapoptotic cell death, although a causal relationship was notdetected [9]. Here, we hypothesized that resin monomers mightalso activate a flexible system of cellular antioxidants as a responseto reestablish the intracellular redox equilibrium, with GSH beingthe major buffering system. It has been previously suggested thatthe redox potential of the GSH/GSSG system between �260and �200 mV differentially influences cell functions. While redoxvalues below �200 mV are related to cell proliferation and differ-entiation, cell death via apoptosis is probably the consequence of anelevated redox potential [16]. Thus the synthesis of GSH wasmodulated here to clarify its effect upon monomer-inducedapoptosis and the expression of antioxidant proteins.

4.1. The effect of GSH synthesis modulators on HEMA-inducedapoptosis

In the present study the resin monomer HEMA causeda decrease in the number of viable RAW264.7 mouse macrophages,while the number of apoptotic and necrotic cells simultaneouslyincreased in a dose-dependent manner similar to previous results[29]. Most relevant, we present experimental evidence that theavailability of GSH is causally related to HEMA-induced apoptosis.The total amounts of GSH detected in RAW264.7 mouse macro-phages here are present in similar amounts in other cell lines [30].By using the inhibitor L-buthionine sulfoximine (BSO), the GSHlevel was extremely reduced at the beginning of HEMA exposure.BSO specifically inhibits GSH synthesis and, accordingly, effectsobserved during GSH deficiency are due to ROS [17]. Likewise,increased formation of ROS was detected here in mouse macro-phages in the presence of BSO. Noteworthy, BSO considerablyincreased the percentage of cells in late apoptosis and necrosisindicating that HEMA-mediated apoptosis is a result of GSHdepletion and related to ROS formation. Consequently and mostrelevant, ROS production in the presence of HEMA probablyexceeded the capacities of other components of the intracellularantioxidant protective system to compensate for GSH function. Incontrast to the effectiveness of BSO, 2-oxothiazolidine-4-carboxylate (OTC) supports GSH production by providing theusually limiting amino acid cysteine for GSH synthesis [18].However, an increase in GSH levels was not observed when cellcultures were preincubated with OTC in the present study. Thisfinding suggests that intracellular GSH levels in mouse macro-phages were in a functional steady-state regulated by a feedbackinhibition of de novo synthesis by GSH [31]. The present findingsalso indicated that the additional supply of cysteine by OTC wasused for GSH synthesis when cells were exposed to HEMA, becausethe cysteine prodrug protected mouse macrophages from HEMA-induced apoptosis. Yet, the precise changes in the amounts ofreduced GSH and GSSG disulfide, as well as in the ratios ofGSH/GSSG in cell cultures preincubated with BSO, OTC, NAC, andresin monomers like HEMA, is still under current investigation. Incontrast to OTC, NAC is a known antioxidant for scavenging ROS,

Fig. 5. Expression of antioxidant enzymes in RAW264.7 mouse macrophages. (A) Western blotting. Cell cultures were exposed to HEMA in the presence or absence of 50 mM BSO, 5 mM OTC, or 10 mM NAC for 24 h. Proteins were detectedby immunoblotting using specific antibodies as described in Materials and methods. Expression of GPx1/2 (glutathione peroxidase 1/2), SOD1 (superoxide dismutase 1), catalase, HO-1 (heme oxygenase-1), TRX1 (thioredoxin 1), andHIF-1a (hypoxia-inducible factor-1a, indicated by an arrow) is shown in one of at least two independent experiments. The expression of GAPDH (glyceraldehyde-3-phosphate dehydrogenase) was also monitored on the samemembranes after stripping and reprobing with a specific antibody. (B) The expression of proteins is shown as bars which represent the fold change for ratios between antioxidant proteins and GAPDH as calculated after densitometry ofthe immunoblot shown in (A). The ratio in untreated cultures (medium) was set to 1.

S.Krifka

etal./

Biomaterials

33(2012)

5177e5186

5183

S. Krifka et al. / Biomaterials 33 (2012) 5177e51865184

but NAC and others can act like OTC as a surrogate for cysteine afterdeacetylation to facilitate GSH production [32]. NAC was alsoobserved to prevent HEMA-induced apoptosis here, similar toa previous report which showed that cell death induced by HEMAwas significantly inhibited by NAC [29,33]. However, diversemechanisms of NAC function have been discussed to date. Forinstance, experimental evidence of extra- and intracellular NAC-HEMA adduct formation has been recently shown, such thatbioavailability of HEMA might be constricted in the present study[34]. Moreover, NAC is known to influence several signal trans-duction pathways resulting in promotion of cell survival anddifferentiation, reduction of proliferation and inflammation, orinduction of apoptosis [19].

4.2. Generation of reactive oxygen species (ROS)

Increased ROS formation induced by dental resin monomers hasbeen associated with the process of cell death via apoptosis onvarious occasions [10]. Accordingly, substances for scavenging ROSand antioxidants including NAC, ascorbate (vitamin C) or vitamin Eare noted to protect cells from monomer-induced cell damage[35,36]. In accordance with previous studies, the resin monomerHEMA enhanced the amount of ROS in a dose-dependent mannerand supports the current hypothesis of ROS relating to cytotoxicityby resin monomers [29,37,38]. Here, the presence of the GSHsynthesis inhibitor BSO alone resulted in a significant increase inROS formation, which was drastically enhanced in cell culturestreated with HEMA. First of all, these findings underscore thecrucial function of GSH in counteracting ROS formation to maintaincellular redox balance under physiological conditions and evenmore so in HEMA-exposed cultures. Secondly, they prove thatHEMA-induced apoptosis is a consequence of reduced GSH levels.Yet, exactly how a subsequent accumulation of ROS is linked to theactivation of signal transduction pathways and pro-apoptoticproteins still needs to be established. Thus far, the presence ofthe antioxidant NAC prevented the monomer-induced activation ofMAPK and influenced the expression of transcription factorsdownstream. However, it appears as if the activation of MAPK asa consequence of monomer-induced ROS formation only reflectsthe activation and balance of general redoxsensitive pro-survivalversus pro-apoptotic pathways [9].

Exposure of RAW264.7 mouse macrophages to OTC also causeda slight additional accumulation of ROS under the current experi-mental conditions. This result was unforeseen since OTC indirectlyreplenishes intracellular GSH by supplying cysteine and conse-quently a reduced level of ROS was expected. The reason for thispresent effect of OTC is still unclear, and is in contrast to the stablelevel of ROS reported in human kidney epithelial cells in thepresence and absence of OTC [39]. Yet, we observed the identicalformation of ROS after exposure to OTC in odontoblast-like MDPC-23 cells, a cell line of different origin than mouse macrophages(unpublished observations). Although the total amount of GSHdetected in cells preincubated with OTC remained constant asdiscussed above, we cannot rule out that, instead of a de novosynthesis, OTC was used as a cysteine supply for the synthesis ofGSH. In this case, GSH synthesis needed more metabolic energy,which may have led to intensified energy metabolism andincreased ROS formation [18]. In addition, an enhanced expressionof GPx1/2 as discussed below was presently noticed, probably dueto elevated levels of H2O2 or hydroperoxides and amplified levels ofGSH in the presence of OTC. Increased ROS formation in the pres-ence of BSO and, at least to some extent, OTC has been observedhere but a protective effect of OTC on cell viability in contrast to BSOwas detected as well. This seeming paradox suggests that, based ondecreased levels of GSH in BSO-treated cultures and constant

amounts of GSH in cells exposed to OTC, the equilibrium state of theentire cellular antioxidant system including GSH/GSSG is moreaccountable than ROS production in the maintenance of cellviability, and possibly other functions critically regulated by thecellular redox state. Thus, it seems that variations in ROS formationare compensated, at least in part, by the coordination of GSH andantioxidant enzyme activities as discussed below.

4.3. Influence of glutathione on the expression of antioxidantproteins in HEMA-exposed cells

The enzymatic antioxidant defense system controls the intra-cellular oxidative state by metabolizing and scavenging diverseROS, including superoxide anion (O�

2 ), hydroxyl radical (OH$), and

hydrogen peroxide (H2O2), with the objective of a balanced redoxsystem. Here we analyzed the adaptive expression of cytosolicenzymes directly regulating levels of ROS caused by HEMA, as wellas stress-responsive proteins related to general protective functionsunder conditions of oxidative stress. The role of the non-enzymaticGSH in the expression of the antioxidant enzymes was of particularrelevance. Expression of glutathione peroxidase (GPx1/2) andsuperoxide dismutase (SOD1) was detected in untreated RAW264.7mouse macrophage cultures. GPx1/2 is a member of theselenocysteine-containing proteins, which reduce hydroperoxidesincluding H2O2 using GSH as a co-substrate [15]. SOD1, on the otherhand, is a cytosolic enzyme which catalyzes the conversion ofsuperoxide (O$�

2 ) to H2O2 [40]. Since expression of catalase, anenzyme which converts H2O2 to O2 and H2O was not detected, onlyGPx1/2 and SOD1 appear to regulate steady-state levels ofhydrogen peroxide in RAW264.7 mouse macrophages underphysiological conditions. Moreover, the equilibrium between GPx1/2 and SOD1 inducible expression is most likely controlled by GSHsince GPx1/2 uses reduced GSH as an electron donor in thedegradation of H2O2. The results presented here also suggest thatGPx1/2 and SOD1 expression are shifted to a higher leveldepending on the availability of GSH. Cysteine is the limiting factorin the first step of the de novoGSH synthesis catalyzed by glutamatecysteine ligase (GCL) [30]. Here, the additional supply of cysteine bythe presence of OTC or NAC may constantly keep GSH amounts athigh levels leading to increased GPx1/2 expression. In contrast,upregulation of GPx1/2 expression under conditions of reducedGSH levels in the presence of BSO alone appears at first glance to bea paradox. However, the present results suggest that increasedexpression of GPx1/2 was caused by high amounts of H2O2 asa consequence of GSH depletion in the presence of BSO. Apparently,the formation of ROS levels in the presence of BSO was still belowa threshold necessary to induce catalase expression. Noteworthy isthat it has been previously suggested that GPx1/2 was active in thedegradation of lower H2O2 levels, while catalase was more impor-tant at higher concentrations of ROS [41].

HEMAobviously interferedwith equilibrium in the expression ofGPx1/2, SOD1, and catalase. It is likely that thedecreasingexpressionof GPx1/2 in the presence of increasing concentrations of HEMAwasa result of GSH depletion, almost certainly caused by the recentlyidentified adduct formation of GSH with resin monomers [42,43].We speculate that HEMA-GSH formation might be catalyzed byglutathione-S-transferase activities, detoxifying phase II enzymesinducible by ROS [44]. It appears as if depletion of GSH increased theamounts of ROS far beyond the capacities of the triad SOD1, GPx1/2,and GSH, which regulates steady-state hydrogen peroxide turnover.Consequently, catalase expression was up-regulated as a line ofdefense against oxidative stress inHEMA-exposed cells. Under theseconditions, high levels of H2O2 probably led to a feedback inhibitionof SOD1 expression. Moreover, the increased expression of theinducible stress-responsive heme oxygenase (HO-1) enzyme

S. Krifka et al. / Biomaterials 33 (2012) 5177e5186 5185

depending on HEMA concentrations might indicate the need forantioxidants in addition to those functioning in a balanced redoxenvironment. HO-1 transforms the protoporphyrin IX ring of hemein heme proteins into Fe, CO, and biliverdin, which is finally con-verted into the antioxidant bilirubin. These products of hemecatabolism are in total cytoprotective and purported to underlie thesalutary effects of HO-1 activity [23]. The expression of GPx1/2 inHEMA-exposed cell cultures was further reduced in the presence ofBSO, apparently due to the lack of GSH as a substrate as discussedabove, but catalase and HO-1 expression was kept at high levels tocounteract oxidative stress. In contrast, the expression of GPx1/2,catalase, and HO-1 was simultaneously reduced in HEMA-exposedcells pre-treated with OTC and NAC. This concurrent down-regulation indicated a reduced level of oxidative stress most likelydue to a constant supply of cysteine, which could act as an antiox-idant by itself or support GSH synthesis. The conclusions from thepresent findings are summarized in a model on resin monomer-induced expression of antioxidant enzymes (Fig. 6). Noteworthy, isthat the expression of superoxide dismutase, catalase, glutathioneperoxidase, and heme oxygenase is under the control of Nrf2(nuclear factor E2-related factor 2) [45].

In addition to serving as an electron donor for diverse cellularprocesses, the oxidoreductase thioredoxin (TRX1) is highly involvedin maintaining a balanced redox environment and the TRX systemcomplements GSH [46]. Furthermore, the TRX system is known tocontrol the activity of several transcription factors including NF-kB,p53, and HIF and thus regulate DNA synthesis or apoptosis as well[47]. In the present investigation, the expression of TRX1was visiblein general, but differed to a certain extent. Expression of TRX1 wasslightly reduced in cell cultures exposed to the high HEMAconcentration as well as the combination of BSO and HEMA.Moreover, it seemed that the expression of TRX1 was slightlyenhanced in cell cultures exposed to 8 mM HEMA in the presence ofNAC. Comparing these findings with the present results on theinduction of apoptosis, the expression of TRX1 might be inverselyrelated, assuming that TRX1 acts as an anti-apoptotic signalingmolecule. It has been described that oxidation of TRX1 led to theactivation of the apoptosis signal-regulating kinase (ASK1) resultingin apoptotic cell death [46]. The discrimination between oxidizedand reduced TRX1 via redox western methodology, which was notused here, will provide more insight into the specific role TRX1 in

cell membrane

H2ORM-GSH

O2-

oxidative burst

environment

SODH2O

Catalase

heme HO-1 biliverdin

bilirubinNADPHBVR

H2O2

Resin Monomers (RM)

GSHGPx1/2

Fig. 6. Model of redox regulation in cells exposed to dental resin monomers (RM). Thecurrent results suggest that HEMA, as a representative of resin monomers, binds toglutathione (GSH), and a decrease in GSH levels causes downregulation of theexpression of glutathione peroxidase (GPx1/2). As a consequence, increased H2O2

formation leads to the induction of catalase and a feedback inhibition of superoxidedismutase (SOD) expression. Enhanced oxidative stress also increases the expression ofheme oxygenase (HO-1) and finally results in the generation of the antioxidantbilirubin.

monomer-induced redox regulation [48]. In contrast to the enzymesdirectly involved in ROS regulation, hypoxia-inducible transcriptionfactors (HIFs) regulate the cellular response to changes in oxygentension. The oxygen-sensing protein HIF1 adjusts cellular metabo-lism under hypoxic conditions to promote cell survival by rapidstabilization of its HIF-1a subunit, which is degraded under nor-moxia [49]. It has been suggested that hypoxia prevented accumu-lation of HIF-1a, but, quite in contrast, under normoxia increasedlevels have been detected in H2O2-treated cells. It was assumed thata threshold concentration ofH2O2mayexist, and lowconcentrationsof hydrogen peroxide may increase expression of HIF-1a, but highconcentrations of H2O2 prevented HIF-1a expression [50]. Thesesuggestionswere in linewith ourfindings on downregulatedHIF-1aexpression in cell cultures exposed toHEMAandBSO inparticular, aswell as restoredHIF-1a expression inNAC-treated cells. It appears asif HIF-1a was a very sensitive detector of oxygen tension inRAW264.7mousemacrophages, indicating that even slight changesin the intracellular redox state in monomer-treated cells may besufficient to shift HIF-1a expression.

5. Conclusion

The present findings suggest that GSH is the primary antioxi-dant central to the regulation of cell response towards oxidativestress induced by resin monomers like HEMA. We present directevidence that a reduction in the amount of GSH by BSO leads toamajor increase in the percentage of cells in apoptosis and necrosis,while HEMA-induced cell death via apoptosis is counteracted bythe presence of GSH precursors like OTC or NAC. Furthermore, theexpression of the antioxidant enzyme GPx1/2 depends on theavailability of GSH, and GPx1/2 is the main enzyme responsible fordegradation of H2O2. Since both GPx1/2 and SOD1 were expressedin untreated cell cultures, these enzymes must regulate endoge-nous H2O2 homeostasis. We also showed that the expression of ROSmetabolizing enzymes was inducible, and expression was differ-entially modified in cells exposed to HEMA. Expression of catalasewas considerably induced in HEMA-exposed cultures, indicatingfirst that H2O2 is the type of ROS preferentially increased inmonomer-exposed cells, and second, large amounts of H2O2 arenecessary for triggering catalase expression. Moreover, oxidativestress induced by HEMA is severe since redox-regulating enzymeslike HO-1, which are beyond enzyme activities directly regulatingH2O2 equilibrium, were activated. These findings imply that evensmall variations in the oxidative state of cells as detected by ROSformation differentially influence cell functions related to thepresence of monomers. Finally, the present investigation providesexperimental evidence that monomer-induced apoptosis is directlyand causally related to the availability of GSH, and thus connectedto a shift in the intracellular redox potential.

Acknowledgments

The skillful technical assistance of Christine Petzel, Carola Bolay,and Claudia Waha is thankfully acknowledged. The authors aregrateful for the expert advice on the manuscript by ChristineRoss-Cavanna. The project was supported by a grant from theUniversity Hospital Regensburg (S.K.) and by the DeutscheForschungsgemeinschaft (DFG, Schw 431/13-1).

References

[1] Peutzfeldt A. Resin composites in dentistry: the monomer systems. Eur J OralSci 1997;105:97e116.

[2] Navarro M, Michiardi A, Castaño O, Planell JA. Biomaterials in orthopaedics. J RSoc Interface 2008;5:1137e58.

S. Krifka et al. / Biomaterials 33 (2012) 5177e51865186

[3] Durner J, Spahl W, Zaspel J, Schweikl H, Hickel R, Reichl FX. Eluted substancesfrom unpolymerized and polymerized dental restorative materials and theirNernst partition coefficient. Dent Mater 2010;26:91e9.

[4] Drucker AM, Pratt MD. Acrylate contact allergy: patient characteristics andevaluation of screening allergens. Dermatitis 2011;22:98e101.

[5] Tsukimura N, Yamada M, Aita H, Hori N, Yoshino F, Chang-Il Lee M, et al.N-acetyl cysteine (NAC)-mediated detoxification and functionalization ofpoly(methyl methacrylate) bone cement. Biomaterials 2009;30:3378e89.

[6] Eckhardt A, Harorli T, Limtanyakul J, Hiller KA, Bosl C, Bolay C, et al. Inhibitionof cytokine and surface antigen expression in LPS-stimulated murinemacrophages by triethylene glycol dimethacrylate. Biomaterials 2009;30:1665e74.

[7] GallerKM,SchweiklH,HillerKA, CavenderAC, BolayC,D’SouzaRN, et al. TEGDMAreduces mineralization in dental pulp cells. J Dent Res 2011;90:257e62.

[8] Bakopoulou A, Leyhausen G, Volk J, Tsiftsoglou A, Garefis P, Koidis P, et al.Effects of HEMA and TEDGMA on the in vitro odontogenic differentiationpotential of human pulp stem/progenitor cells derived from deciduous teeth.Dent Mater 2011;27:608e17.

[9] Krifka S, Hiller KA, Bolay C, Petzel C, Spagnuolo G, Reichl FX, et al. Function ofMAPK and downstream transcription factors in monomer-induced apoptosis.Biomaterials 2012;33:740e50.

[10] Schweikl H, Spagnuolo G, Schmalz G. Genetic and cellular toxicology of dentalresin monomers. J Dent Res 2006;85:870e7.

[11] Forman HJ, Maiorino M, Ursini F. Signaling functions of reactive oxygenspecies. Biochemistry 2010;49:835e42.

[12] Ryter SW, Kim HP, Hoetzel A, Park JW, Nakahira K, Wang X, et al. Mechanismsof cell death in oxidative stress. Antioxid Redox Signal 2007;9:49e89.

[13] Halliwell B. Antioxidant defence mechanisms: from the beginning to the end(of the beginning). Free Radic Res 1999;31:261e72.

[14] Lu SC. Regulation of glutathione synthesis. Mol Aspects Med 2009;30:42e59.[15] Brigelius-Flohé R. Tissue-specific functions of individual glutathione peroxi-

dases. Free Radic Biol Med 1999;27:951e65.[16] Schafer FQ, Buettner GR. Redox environment of the cell as viewed through the

redox state of the glutathione disulfide/glutathione couple. Free Radic BiolMed; 2001:1191e212.

[17] Griffith OW, Meister A. Potent and specific inhibition of glutathione synthesisby buthionine sulfoximine (S-n-butyl homocysteine sulfoximine). J Biol Chem1979;254:7558e60.

[18] Williamson JM, Meister A. Stimulation of hepatic glutathione formation byadministration of L-2-oxothiazolidine-4-carboxylate, a 5-oxo-L-prolinasesubstrate. Proc Natl Acad Sci U S A 1981;78:936e9.

[19] Zafarullah M, Li WQ, Sylvester J, Ahmad M. Molecular mechanisms ofN-acetylcysteine actions. Cell Mol Life Sci 2003;60:6e20.

[20] Valko M, Leibfritz D, Moncol J, Cronin MT, Mazur M, Telser J. Free radicals andantioxidants in normal physiological functions and human disease. Int J Bio-chem Cell Biol 2007;39:44e84.

[21] Imlay JA. Cellular defenses against superoxide and hydrogen peroxide. AnnuRev Biochem 2008;77:755e76.

[22] Kondo N, Nakamura H, Masutani H, Yodoi J. Redox regulation of human thi-oredoxin network. Antioxid Redox Signal 2006;8:1881e90.

[23] Gozzelino R, Jeney V, Soares MP. Mechanisms of cell protection by hemeoxygenase-1. Annu Rev Pharmacol Toxicol 2010;50:323e54.

[24] Semenza GL. Oxygen homeostasis. Wiley Interdiscip Rev Syst Biol Med 2010;2:336e61.

[25] Yeligar SM, Machida K, Kalra VK. Ethanol-induced HO-1 and NQO1 aredifferentially regulated by HIF-1alpha and Nrf2 to attenuate inflammatorycytokine expression. J Biol Chem 2010;285:35359e73.

[26] Schweikl H, Hiller KA, Eckhardt A, Bolay C, Spagnuolo G, Stempfl T, et al.Differential gene expression involved in oxidative stress response caused bytriethylene glycol dimethacrylate. Biomaterials 2008;29:1377e87.

[27] Vermes I, Haanen C, Steffens-Nakken H, Reutelingsperger C. A novel assay forapoptosis. Flow cytometric detection of phosphatidylserine expression onearly apoptotic cells using fluorescein labeled Annexin V. J Immunol Methods1995;184:39e51.

[28] Forman HJ, Torres M. Redox signaling in macrophages. Mol Aspects Med2001;22:189e216.

[29] Spagnuolo G, Mauro C, Leonardi A, Santillo M, Paternò R, Schweikl H, et al.NF-kappaB protection against apoptosis induced by HEMA. J Dent Res 2004;83:837e42.

[30] Forman HJ, Zhang H, Rinna A. Glutathione: overview of its protective roles,measurement, and biosynthesis. Mol Aspects Med 2009;30:1e12.

[31] Huang CS, Moore WR, Meister A. On the active site thiol of gamma-glutamylcysteine synthetase: relationships to catalysis, inhibition, and regu-lation. Proc Natl Acad Sci U S A 1988;85:2464e8.

[32] De Vries N, De Flora S. N-acetyl-l-cysteine. J Cell Biochem Suppl 1993;17F:270e7.

[33] Paranjpe A, Cacalano NA, Hume WR, Jewett A. N-acetyl cysteine mediatesprotection from 2-hydroxyethyl methacrylate induced apoptosis via nuclearfactor kappa B-dependent and independent pathways: potential involvementof JNK. Toxicol Sci 2009;108:356e66.

[34] Nocca G, D’Antò V, Desiderio C, Rossetti DV, Valletta R, Baquala AM, et al.N-acetyl cysteine directed detoxification of 2-hydroxyethyl methacrylate byadduct formation. Biomaterials 2010;31:2508e16.

[35] Stanislawski L, Lefeuvre M, Bourd K, Soheili-Majd E, Goldberg M, Périanin A.TEGDMA-induced toxicity in human fibroblasts is associated with early anddrastic glutathione depletion with subsequent production of oxygen reactivespecies. J Biomed Mater Res A 2003;66:476e82.

[36] Walther UI, Siagian II , Walther SC, Reichl FX, Hickel R. Antioxidative vitaminsdecrease cytotoxicity of HEMA and TEGDMA in cultured cell lines. Arch OralBiol 2004;49:125e31.

[37] Chang HH, Guo MK, Kasten FH, Chang MC, Huang GF, Wang YL, et al. Stim-ulation of glutathione depletion, ROS production and cell cycle arrest of dentalpulp cells and gingival epithelial cells by HEMA. Biomaterials 2005;26:745e53.

[38] Samuelsen JT, Dahl JE, Karlsson S, Morisbak E, Becher R. Apoptosis induced bythe monomers HEMA and TEGDMA involves formation of ROS and differentialactivation of the MAP-kinases p38, JNK and ERK. Dent Mater 2007;23:34e9.

[39] Lee S, Moon SO, Kim W, Sung MJ, Kim DH, Kang KP, et al. Protective role ofL-2-oxothiazolidine-4-carboxylic acid in cisplatin-induced renal injury.Nephrol Dial Transplant 2006;21:2085e95.

[40] McCord JM, Fridovich I. The utility of superoxide dismutase in studying freeradical reactions. I. Radicals generated by the interaction of sulfite, dimethylsulfoxide, and oxygen. J Biol Chem 1969;244:6056e63.

[41] Makino N, Sasaki K, Hashida K, Sakakura Y. A metabolic model describing theH2O2 elimination by mammalian cells including H2O2 permeation throughcytoplasmic and peroxisomal membranes: comparison with experimentaldata. Biochim Biophys Acta 2004;1673:149e59.

[42] Nocca G, Ragno R, Carbone V, Martorana GE, Rossetti DV, Gambarini G, et al.Identification of glutathione-methacrylates adducts in gingival fibroblasts anderythrocytes by HPLC-MS and capillary electrophoresis. Dent Mater 2011;27:e87e98.

[43] Samuelsen JT, Kopperud HM, Holme JA, Dragland IS, Christensen T, Dahl JE.Role of thiol-complex formation in 2-hydroxyethyl- methacrylate-inducedtoxicity in vitro. J Biomed Mater Res A 2011;96:395e401.

[44] Whalen R, Boyer TD. Human glutathione S-transferases. Semin Liver Dis 1998;18:345e58.

[45] de Vries HE, Witte M, Hondius D, Rozemuller AJ, Drukarch B, Hoozemans J,et al. J. Nrf2-induced antioxidant protection: a promising target to counteractROS-mediated damage in neurodegenerative disease? Free Radic Biol Med2008;45:1375e83.

[46] Lillig CH, Holmgren A. Thioredoxin and related moleculesefrom biology tohealth and disease. Antioxid Redox Signal 2007;9:25e47.

[47] Holmgren A, Lu J. Thioredoxin and thioredoxin reductase: current researchwith special reference to human disease. Biochem Biophys Res Commun2010;396:120e4.

[48] Imhoff BR, Hansen JM. Differential redox potential profiles during adipo-genesis and osteogenesis. Cell Mol Biol Lett 2011;16:149e61.

[49] Chandel NS, Budinger GR. The cellular basis for diverse responses to oxygen.Free Radic Biol Med 2007;42:165e74.

[50] Kietzmann T, Görlach A. Reactive oxygen species in the control of hypoxia-inducible factor-mediated gene expression. Semin Cell Dev Biol 2005;16:474e86.