Effect of antioxidants on the H2O2-induced premature senescence of human fibroblasts

The study was aimed at evaluation of the role of secondary oxidative stress in the stress-induced premature senescence (SIPS) of human fibroblasts induced by H2O2. Two fibroblast lines were used: lung MRC-5 and ear H8F2p25LM fibroblasts. The lines differed considerably in sensitivity to H2O2 (IC50 of 528 and 33.5 μM, respectively). The cells were exposed to H2O2 concentrations corresponding to IC50 and after 24 h supplemented with a range of antioxidants. Most of antioxidants studied slightly augmented the survival of fibroblasts at single concentrations or in a narrow concentration range, but the results were not consistent among the cell lines. Chosen antioxidants (4-amino-TEMPO, curcumin, caffeic acid and p-coumaric acid) did not restore the level of glutathione decreased by H2O2. Hydrogen peroxide treatment did not induce secondary production of H2O2 and even decreased it, decreased mitochondrial potential in both cell lines and induced changes in the mitochondrial mass inconsistent between the lines. Antioxidant protected mitochondrial potential only in H8F2p25LM cells, but attenuated changes in mitochondrial mass. These results speak against the intermediacy of secondary oxidative stress in the SIPS induced by H2O2 and suggest that the small protective action of antioxidants is due to their effects on mitochondria.


INTRODUCTION
Aging is an irreversible process affecting all higher organisms, characterized by progressive deterioration leading to a loss of function of cells, tissues, organs and finally death. Cellular aging can be accelerated by using non-lethal stresses; this phenomenon is referred to as stress induced premature senescence (SIPS). The concept of SIPS was first introduced in 2000 by Dr. Olivier Toussaint and co-workers [1,2]. Premature aging of cultured cells is usually associated with the exposure of cells to environmental stress factors. Stress induced premature senescence can be defined as the long-term effect of sub-cytotoxic stress on proliferative cell types including appearance of many features of replicative senescence. Various genotoxic agents, such as hydrogen peroxide (H2O2), tert-butyl hydroperoxide, copper sulfate, diperoxovanadate, ethanol, mitomycin C, other cytostatic drugs, heat shock or UV radiation are wellestablished inducers of SIPS. It has been suggested that SIPS can be used in toxicology to identify xenobiotics that may induce premature senescence. In each case, oxidative stress is believed to be the major cause of SIPS program activation in normal cells [2][3][4][5][6].

AGING
Up to now, among other stressors, H2O2 is perhaps the best candidate for inducing senescence, because an H2O2-induced process might mimic the oxidative environment that may occur in vivo [2,7,8]. Addition of a single bolus of H2O2 to cultured cells means a rather short exposure to an external reactive oxygen species (ROS), which is rapidly decomposed [9,10]. Hydrogen peroxide, which is plasma membrane permeable, may produce hydroxyl radical ( • OH) in the presence of Fe 2+ or Cu 2+ through the Fenton reaction. Hydroxyl radical and the superoxide anion radical (O2 •-) oxidize the unsaturated bonds of lipids to yield lipid peroxides as well as aldehydes such as 4hydroxynonenal. Hydroxyl radical and lipid-derived aldehydes react with amino acid residues in proteins to produce carbonyl proteins [11] and modify nucleic acid bases [12]. Moreover, sublethal oxidative stress was shown to arrest proliferation and promote accumulation of senescence-associated molecular hallmarks [increased activity of cyclin-dependent kinase inhibitor p21Waf1/Cip1 (p21) and of acidic βgalactosidase (SA-β-gal), as well as diminution of phosphorylated retinoblastoma gene product (ppRb)] in human fibroblasts [13].
The causative role of oxidative stress in SIPS is well established [2][3][4][5][6]14]. Nevertheless, it is of interest whether ROS play a role in secondary signaling leading to SIPS-induced cell death or if the execution of SIPS depends on the molecular machinery once triggered by oxidative stress and secondary production of ROS after initial oxidative stress is not important. One way to get an insight into this question is to examine the effect of antioxidants on human fibroblasts {on two human fibroblast lines (lung MRC-5 and H8F2p25LM fibroblasts, obtained from ear skin of an adult donor)} after H2O2 exposure and decomposition on the SIPS, which was the aim of this study. Our results speak for the second possibility.

Hydrogen peroxide sensitivity of fibroblasts
Hydrogen peroxide showed a dose-dependent cytotoxicity against normal human fibroblast line  differed considerably in the sensitivity to H2O2, with IC50 values of 528 and 33.5 µM for MRC-5 and H8F2p25LM fibroblasts, respectively, when estimated after 24 h. The more resistant MRC-5 cells more rapidly decomposed H2O2 than H8F2p25LM fibroblasts, the half-life times of 50 µM H2O2 in the presence of 5 x 10 3 cells being 8.8 minutes for MRC-5 cells and 61.5 minutes for H8F2p25LM cells ( Figure  2A, 2B). This difference was mainly due to different catalase activity, which was about 11 times higher in MRC-5 cells than in H8F2p25LM cells (28.03 and 2.56 µmol H2O2/(s*10 6 cells), respectively).

Protection of fibroblasts against the H2O2-induced cytotoxicity
24 hours after H2O2 treatment, antioxidants were added to the cells to study their effects on the processes dependent on secondary oxidant-dependent signaling leading to decrease in cell survival. The antioxidants were first checked for their cytotoxicity (data for 4 chosen antioxidants are shown in Figure 3A and 3B) and used at non-toxic concentrations. Among the antioxidants tested , 2 µM 4-amino-TEMPO, 10 µM  TEMPOL, 2-10 µM gallic acid, 10 µM caffeic acid,  50-100 µM aminoguanidine hydrochloride, 1

Hydrogen peroxide generation by the antioxidants studied
We compared generation of hydrogen peroxide by the antioxidant compounds used in this study during incubation in the culture medium. Some of these compounds generated significant amounts of hydrogen peroxide (gallic acid > caffeic acid > oleuropein), which represents a pro-oxidant effect of these antioxidants. No detectable H2O2 generation was found for 4-hydroxy-TEMPO, 4-amino-TEMPO, TEMPO, curcumin, resveratrol, ethoxyquin, melatonin, ferulic acid, pcoumaric acid and aminoguanidine hydrochloride ( Table 1). Both H2O2-generating compounds (caffeic acid, oleuropein) and compounds which did not produce H2O2 in the culture media offered small post-exposure protection against the effects of H2O2.

Glutathione content
Treatment with hydrogen peroxide decreased the content of reduced glutathione (GSH) in MRC-5 fibroblasts and did not cause a statistically significant decrease of GSH level in H8F2p25LM cells (although a tendency for decrease was visible; Figure 6A, 6B). Posttreatment exposure to the chosen antioxidants did not augment the GSH level with respect to cells treated with H2O2 only, while 4-amino-TEMPO and 50 µM pcoumaric acid evoked a further GSH depletion in H8F2p25 LM cells ( Figure 6B).

Reactive oxygen species
The level of reactive oxygen species (ROS) estimated with H2DCF-DA decreased in H2O2-treated cells after next 24 h following the day of exposure and then increased gradually, exceeding the control level in H8F2p25LM cells. Posttreatment with antioxidants induced a small decrease in the ROS level in the majority of cases ( Figure 7A, 7B).
The level of mitochondrial superoxide decreased in H2O2-treated cells and was not significantly affected by antioxidant posttreatment (Figure 8A, 8B).

Mitochondrial membrane potential and mitochondrial mass
Estimation of changes in mitochondrial membrane potential (Δψm) by JC-1 staining of cells demonstrated that the mitochondrial membrane potential was significantly reduced in H2O2 treated cells [in   AGING Mitochondrial mass decreased in H2O2-treated MRC-5 cells ( Figure 10A), but increased (except for the time of 48 h) in H8F2p25LM cells following H2O2 treatment ( Figure 10B). Antioxidants were generally protective against these changes, except for curcumin after 24 h and 96 h and p-coumaric acid after 24 h in MRC-5 cells ( Figure 10A), and 4-amino-TEMPO as well as caffeic acid after 24 h in H8F2p25LM cells( Figure 10B).

Senescence-associated β-galactosidase
Staining for acidic β-galactosidase activity showed a gradual increase in the staining of H2O2-treated MRC-5 cells, which was not significantly affected by treatment with antioxidants. In H8F2p25LM cells, the staining was not increasing after 48 h and even decreased after 96 h. Again, no significant effects of the antioxidants was evident ( Figure 11A, 11B).

DISCUSSION
Dermal fibroblasts are long lived cells undergoing ageassociated damage accumulation underlying the agerelated changes in the skin. Skin wrinkles and sagging are important factors defining skin aging. Fibroblasts play a main role in the production of extracellular AGING matrix components in the skin. Skin aging is the consequence of reduced numbers of fibroblasts, decreased skin elasticity and tonus as well as lower levels of extracellular matrix proteins, thus resulting in the formation of wrinkles [11,15]. Fibroblasts play also a role in the accelerated aging of the skin induced by UV radiation and other environmental factors [16][17].
Fibroblasts are also model cells in studies of in vitro cellular aging and SIPS is an acknowledged model of accelerated aging in vitro, mimicking some aspects of accelerated skin aging in vivo. Our results indicate that there may be significant differences between different fibroblast lines in the sensitivity to H2O2 ( Figure 1A, 1B) linked to the capacity to decompose this compound (Figure 2A, 2B).
Hydrogen peroxide is naturally produced in the human cells during many physiologic and pathological processes and has been widely used as a model prooxidant in the study of oxidative stress. It has been reported that ROS, especially H2O2 and superoxide anions, are associated with the cellular proliferation of many cell types including fibroblasts [18]. However, elevated levels of H2O2 induce premature senescence and H2O2 treatment is a standard procedure used for this purpose [2,7,8].
Both cell lines studied by us were treated with H2O2 at concentrations corresponding to their IC50 values. Oxidative stress has been repeatedly proposed to be the factor responsible for the induction of fibroblast SIPS and antioxidants have been reported to be protective. AGING For example, pretreatment with rapamycin decreased the extent of SIPS induced by UV exposure and this effect was ascribed to the modification of antioxidant defense by rapamycin and suppression of free radical production by irradiation [19]. We were interested in checking whether the mechanism of SIPS-induced cell death is dependent on the secondary production of ROS and if antioxidants can modulate the course of cellular changes following exposure to H2O2. Literature data do not allow for reaching an unanimous answer to this question.
Gamma-tocotrienol, biodynes, tocotrienol-rich fraction (TRF) and tocopherol were found to be protective against changes in collagen synthesis and degradation caused by H2O2-induced SIPS, nonetheless the cells were pretreated with these compounds prior to exposure to H2O2 [20,21]. Centella asiatica herb extracts containing various pentacyclic triterpenes protected fibroblasts against the effects of H2O2-induced SIPS but, again, the cells were pre-exposed to the extracts before exposure to H2O2 [22]. Malvidin [23], cyanidin [24] and phloroglucinol [25] were reported to protect fibroblasts from H2O2-induced SIPS and decrease the level of lipid peroxidation enhanced by H2O2 exposure. Nevertheless, in these experiments the compounds were introduced soon after exposure to H2O2, so the results do not allow for distinguishing between effects of primary and secondary oxidative stress, if the latter was involved in the signaling mediating the development of SIPS.

AGING
The SIPS induced by exposure to H2O2 is known to proceed in a different way and to involve different pathways than replicative senescence. In our study, the senescence-associated β-galactosidase activity, one of indices of replicative aging, observed in H2O2-treated cells, was increased in MRC-5 cells, but its increase was not modified by antioxidants ( Figure 11A). No increase in the acidic β-galactosidase activity was seen in H8F2p25LM cells, which are more sensitive to H2O2 and were treated with a lower concentration of H2O2 ( Figure  11B). Interestingly, no increased staining for the senescence-associated β-galactosidase was also reported for Werner syndrome fibroblasts, which are more sensitive to H2O2 than normal fibroblasts [26]. Perhaps, higher H2O2 concentrations are required for induction of increase in acid β-galactosidase activity than for triggering the main signaling pathways inducing the SIPS program.
Results from other studies indicate that these pathways include NF-κB, iNOS, p53, COX-2, the caspase-3/keratin-18 pathway and serine/threonine kinase Aurora A/MDM2 pathway as well as proteasome/ubiquitin ligase pathways of protein degradation [23,27]. Hydrogen peroxide-induced SIPS was also found to involve pancreatic ER kinase (PERK)-mediated upregulation of CCAAT/enhancer-binding protein homologous protein and activation of unfolded protein response [28]. AGING Posttreatment with antioxidants provided small and inconsistent protection against the toxicity of H2O2 ( Figures 4A-4D and 5A-5D). Antioxidants did not affect considerably changes in GSH level induced by H2O2 treatment (Figure 6A, 6B).
The treatment with H2O2 did not induce secondary production of ROS but, at least initially, rather decreased it. The effect of antioxidants on the ROS level in H2O2-treated cells was slight or none ( Figures  7A, 7B and 8A, 8B).
Hydrogen peroxide treatment decreased mitochondrial membrane potential in both cell lines studied; antioxidants provided some protection only in H8F2p25LM cells, but not in MRC-5 cells ( Figure 9A, 9B).
Changes in the mitochondrial mass were inconsistent between the lines, a decrease being observed in MRC-5 cells and mostly an increase in H8F2p25LM cells. In both lines antioxidants were generally protective with respect to these changes ( Figure 10A, 10B).
In summary, these results speak against the intermediacy of secondary ROS formation in the SIPS induced by H2O2. In line with our results, Le Boulch et al. analyzed protein carbonylation in human fibroblasts in the course of SIPS induced by H2O2 exposure identifying an "Oxi-proteome", i.e. a set of proteins that are building up as oxidized. However, they did not observe an increase in the overall level of protein oxidation during SIPS; rather, this level was decreasing during several days after exposure [29]. These results AGING speak against increase in ROS production after H2O2 exposure, which should lead to an increase in the overall protein carbonylation.
The small protective action of antioxidants observed in our study is most probably due to their effects on mitochondria of H2O2-treated cells. The critical role of mitochondrial changes involving release of cytochrome c in the development of SIPS has been documented by others [20].

Assay of hydrogen peroxide generation
Protocol for evaluation of H2O2 generation by antioxidants consisted in addition of 18 μl of 10 mM antioxidant to 162 μl of DMEM/DMEM + glutaMAX + serum. The samples were incubated for 3 h at 37±1°C with shaking and the peroxide content was estimated before and after incubation by the ferric-Xylenol Orange method [30].

The kinetics of hydrogen peroxide decomposition by fibroblasts
The fibroblasts were plated in wells of a multi-well plate. After 24 h the medium was removed and a new one containing H2O2 at a concentration of 50 μM was added. After 0, 10, 30, 45, 60, 120 and 180 minutes 90 μl of medium was collected from each well and the concentration of remaining H2O2 was determined.
To determine hydrogen peroxide, 10 μl of Xylenol Orange reagent was added to the collected 90 μl aliquots. After 30-min incubation at room temperature, absorbance of the samples was measured at 560 nm.

Determination of hydrogen peroxide inhibitory concentration (IC50)
The cell were seeded in 96-well clear plate at a density of 2.5×10 3 cells/well in 100 µl culture medium and allowed to attach for 24 h at 37°C. After incubation cells were treated with hydrogen peroxide at concentrations ranging from 0-500 µM (MRC-5) or 0-100 µM (H8F2p25LM). Working solution of hydrogen peroxide was prepared in culture medium suitable for appropriate fibroblast line. After 24 h exposure to hydrogen peroxide medium was removed, replaced with 100 µl of 2% Neutral Red solution and incubated for 1 h at 37°C. Then the cells were washed with PBS, fixed with 100 µl/well 50% ethanol, 49% H2O and 1% glacial acetic acid, and shaken for 20 min (700 rpm) at room temperature. Absorbance was measured at 540 nm against 620 nm.

Antioxidant cytotoxicity
The cells were seeded in a transparent 96-well plate at a density of 5×10 3 cells/well in 100 µl culture medium and allowed to attach for 24 h at 37°C. After incubation the cells were treated with different antioxidants (TEMPO, 4-amino TEMPO, 4-hydroxy TEMPO, melatonin, ethoxyquin, p-coumaric acid, ferulic acid, gallic acid, aminoguanidine hydrochloride, oleuropein, resveratrol, curcumin, caffeic acid) at 1, 10 and 100 µM concentrations. Working solutions of antioxidants were prepared in culture medium.
After 24 h exposure to antioxidants medium was removed and replaced with 100 µl of 1% Neutral Red solution and incubated for 1 h at 37°C. Than the cells were washed with PBS, fixed with 100 µl/well 50% ethanol, 49% H2O and 1% glacial acetic acid and shaken for 20 min (700 rpm) at room temperature. Absorbance was measured at 540 nm against 620 nm.

Stress-induced premature senescence (SIPS)
SIPS studies were performed on cells grown at a density of 5 x 10 3 cell/well (both MRC-5 and H8F2p25LM), 24 h post-seeding. After incubation cells were treated with H2O2 at 50% inhibitory concentration (IC50) for 24 h at 37°C to induce senescence. Working solutions of H2O2 were prepared in culture medium (DMEM or DMEM+GlutaMax). After 24 h the cells were posttreated with studied antioxidants at various concentrations depending of substances (ranging from 0 to 100 µM) for 24, 48 or 96 h. Cells treated with H2O2, but not posttreated with antioxidants were used as a control. Subsequently, fibroblast viability was tested using NR assay, as a described above.

Catalase activity
Fibroblast cells (MRC-5, H8F2p25LM) were seeded in transparent 6-well plate and allowed to attach for 24 h at 37 °C. After incubation the cells were trypsinized and their number was counted. Then the cells were transferred to a 15 ml falcon, centrifuged at 900 rpm for 5 minutes, washed with PBS and centrifuged once again. After these steps, supernatant was removed and 50 mM phosphate buffer (pH 7) with 0.1% Triton X-AGING 100 was added. The samples were centrifuged at 14000 g for 25 minutes at 0°C and supernatant was collected and used for further determinations. 0.036% solution of H2O2 in 50 mM phosphate buffer was introduced into a quartz cuvette (1 ml). Then the supernatant from cell lysis was added and absorbance was measured at 240 nm for 5 minutes (kinetic measurement).

Content of reduced glutathione
To investigate the content of reduced glutathione in the fibroblasts, the method with ortho-phtal-dialdehyde was used. The cells were seeded in a transparent 96-well plate at an amount 7. Subsequently, 160 µl/well of cold 0.1 M phosphate buffer (pH 6.8) and 25 µl/well of freshly prepared 0.5% OPA in methanol were added into both plates. Both plates were incubated for 30 minutes at room temperature under constant stirring. Fluorescence was measured at 355/430 nm [31]. The content of reduced glutathione was determined by subtracting the fluorescence of the '-NEM' plate from the fluorescence '+ NEM' plate and calculated with respect to the protein content. Protein content in cell lysates was determined according to Lowry et al. [32]. The results were expressed as a percentage of untreated control.

Mitochondrial membrane potential (Δψm)
To estimated changes in the mitochondrial membrane potential after the use of selected antioxidants, JC-1 (5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolylcarbocyanine iodide) with a Mitochondrial Membrane Potential Assay kit from Abnova was used. JC-1 is a lipophilic, cationic dye that can selectively enter the mitochondria and reversibly change color from green to red as the membrane potential increases.
In cells with high Δψm, JC-1 spontaneously forms complexes known as J-aggregates with intense red fluorescence. However, in injured cells with low Δψm, JC-1 remains in the monomeric form and shows only green fluorescence.
Briefly, fibroblasts were seeded in a 96-well flat clearbottom black plate at a density of 7.5×10 3 /well and allowed to attach at 37°C for 24 h. After incubation the cells were treated with H2O2 (24 h) and posttreated with selected antioxidants as described earlier. Then 10 µl of AGING JC-1 staining solution were added into plate wells and the plate was incubated at 37°C for 30 min. After this time the cells were centrifuged at 4000 rpm for 5 min and the supernatant was gently removed. The plate was washed twice using the buffer included in the kit and centrifuged at 4000 rpm for 5 min. The supernatant was replaced by the buffer (100 µl/well) and fluorescence was measured at 535/595 nm (J-aggregates) and 485/535 nm (J-monomers). Data are shown as a ratio of fluorescence of J-aggregates to that of J-monomers.

Mitochondrial mass
MRC-5 and H8F2p25LM cells were seeded at an amount of 1×10 5 cells/well into a 6-well plate and allowed to attach at 37°C. After incubation the cells were treated with H2O2 (24 h) and posttreated with selected antioxidants (curcumin, 4-amino-TEMPO, caffeic acid, p-coumaric acid) for 24, 48, 96 h. Subsequently the cells were trypsinized, counted, transferred to Eppendorf tubes and centrifuged for 5 minutes at 2000 rpm, then washed with 1 ml of PBS and centrifuged again. Subsequently, 1 ml of 10 μM Nnonyl acridine orange (NAO) in PBS was added and the cells were incubated at 37°C for 10 min. After incubation with NAO cells were centrifuged and washed with PBS (1 ml) and then the cell pellet was resuspended in 300 μl of PBS. Each sample was transferred into a 96-well black plate (100 μl/well, 3 repetitions). Fluorescence was measured at 435/535 nm. The results were calculated respectively to cell number.

Staining cells for senescence-associated βgalactosidase
2×10 5 MRC-5 cells and 1.4×10 5 H8F2p25LM cells were seeded into wells of 6-well plates and allowed to grow for 24 hours. Then the medium was exchanged for a fresh one containing H2O2 in a final concentration of 600 µM or 35 µM, respectively (except for control wells) and the cells were incubated for further 24 hours. The medium was again replaced by a fresh one containing 1 µM curcumin, 2 µM 4-amino-TEMPO, 10 µM caffeic acid or 50 µM coumaric acid. Cells were harvested after 24, 48 or 96 hours since the last medium exchange, centrifuged (100×g for 10 minutes, room temperature), resuspended in 200 µL of 1 µM 4-methylumbelliferyl β-D-galactopyranoside dissolved in fresh medium, incubated for 1 hour at 37°C and then analysed by flow cytometry (excitation: 355 nm, emission: 425-475 nm). Median fluorescence of control samples at 24-hour posttreatment was assumed as 100%.

AUTHOR CONTRIBUTIONS
I. S.-B. was responsible for the concept of the study, design of experiments and supervision of experimental work, performed part of experiments as well as had a leading role in the analysis of the results and preparation of the manuscript. N. P. performed main part of experiments in the cellular system and their statistical evaluation as well as contributed to reagents/materials/ analysis tools. She further contributed also to data acquisition and interpretation and wrote parts of the manuscript. G. B. participated in the revision of the manuscript and was also responsible for providing the funding for the study. M. P. took part in the execution of experiments. M. G.-P. performed part of experiments in cell-free systems as well as helped with study of H2O2 sensitivity of fibroblasts and their statistical evaluation. M. G. carried out the assay of acid lysosomal βgalactosidase activity and statistical evaluation of its results. All authors have read and approved the final manuscript.

ACKNOWLEDGMENTS
We would like to express our special appreciation and thanks to Edyta Bieszczad-Bedrejczuk, M.Sc., (Department of Analytical Biochemistry, University of Rzeszów, Poland) for the excellent technical assistance as