α-Lipoic Acid Inhibits TNF-α-Induced Apoptosis in Human Bone Marrow Stromal Cells

Authors

  • Chang-Hyun Byun,

    1. Asan Institute for Life Sciences, Seoul, Republic of Korea
    Search for more papers by this author
  • Jung-Min Koh,

    1. Division of Endocrinology and Metabolism, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea
    Search for more papers by this author
  • Dong Kwan Kim,

    1. Department of Thoracic and Cardiovascular Surgery, University of Ulsan College of Medicine, Asan Medical Center, Seoul, Republic of Korea
    Search for more papers by this author
  • Seung-Il Park,

    1. Department of Thoracic and Cardiovascular Surgery, University of Ulsan College of Medicine, Asan Medical Center, Seoul, Republic of Korea
    Search for more papers by this author
  • Ki-Up Lee,

    1. Division of Endocrinology and Metabolism, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea
    Search for more papers by this author
  • Ghi Su Kim MD

    Corresponding author
    1. Division of Endocrinology and Metabolism, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea
    • Division of Endocrinology and Metabolism, Asan Medical Center, University of Ulsan College of Medicine, 388-1 Poongnap-Dong, Songpa-Gu, Seoul 138-736, Korea
    Search for more papers by this author

  • The authors have no conflict of interest.

Abstract

TNF-α is an important mediator of bone loss. In the HS-5 hBMSC, TNF-α and H2O2 increased intracellular ROS levels and induced cell apoptosis through activation of caspases, JNK and NF-κB. α-Lipoic acid prevented these changes induced by TNF-α and H2O2, suggesting its potential therapeutic applications in attenuating bone loss.

Introduction: Oxidative stress is an important mediator of bone loss. TNF-α, which plays a critical role in the bone loss after menopause, has been shown to increase intracellular oxidative stress. Because oxidative stress is associated with cell death, we analyzed the apoptotic effects of TNF-α and H2O2 on human bone marrow stromal cells (hBMSCs). We also examined the protective effects of an important biological thiol antioxidant, α-lipoic acid (α-LA), against TNF-α- and H2O2-induced apoptosis.

Materials and Methods: Using the HS-5 hBMSC cell line, we tested whether TNF-α-induced apoptosis was mediated by the generation of excessive reactive oxygen species (ROS). Apoptosis was determined by 3-(4,5-dimethylthiazol-2-yl)2,5-diphenyltetrazolium bromide (MTT) assay, trypan blue exclusion assay, quantitation of histone-associated DNA fragments in cytosol, and the activation of caspases. The mechanisms mediating these apoptotic effects were determined by Western blotting and enzyme immunoassay.

Results: Both TNF-α and H2O2 increased intracellular ROS levels, reduced total cellular glutathione levels, activated caspases-3, -9, and -8, and enhanced hBMSC apoptosis. The activation of c-jun N-terminal kinase (JNK) and NF-κB mediated these apoptotic effects. Pretreatment of cells with α-LA prevented these changes induced by TNF-α and H2O2.

Conclusions: Our data show that TNF-α increases intracellular ROS in hBMSC and that TNF-α and H2O2 induce apoptosis in hBMSC through the activation of JNK and NF-κB. Our findings also suggest that α-LA may have therapeutic applications in halting or attenuating bone loss associated with increased oxidative stress.

INTRODUCTION

OXIDATIVE STRESS PLAYS an important role in the pathogenesis of various human diseases, including aging, cancer, atherosclerosis, inflammation, diabetes, and Parkinson's disease.(1) In the past decade or so, several lines of evidence have suggested a possible link between oxidative stress and osteoporosis. Epidemiological studies have found an association between dietary vitamin C and E intake and BMD(2,3) on risk of hip fractures.(4) Antioxidant defenses are markedly decreased in osteoporotic women,(5) and administration of antioxidants such as vitamins C and E and N-acetylcysteine has beneficial effects in individuals with osteoporosis.(6–9) In addition, reactive oxygen species (ROS) are involved in bone resorption, with osteoclast-generated superoxide having a direct contribution to bone degradation.(10,11)

Substantial in vitro and in vivo evidence suggests that TNF-α plays a central role in the pathophysiology of bone loss after menopause.(12,13) TNF-α increases bone loss by multiple actions.(14) The induction of osteoblast apoptosis by TNF-α as shown in cell culture systems may contribute, in part, to the bone loss effect.(15–18) Although TNF-α is generally thought to induce apoptosis by a death receptor-mediated pathway,(19–23) it has also been shown to increase ROS levels in cells expressing cell surface TNF receptors.(24–26) Thus, it is quite possible that TNF-α-induced apoptosis in osteoblast-lineage cells is mediated by excessive generation of ROS.

α-Lipoic acid (α-LA) and its reduced form, dihydrolipoic acid (DHLA), have received considerable attention because of their activities as biological thiol antioxidants.(27) α-LA is widely used clinically, including in ischemia-reperfusion injury,(28) diabetic neuropathy,(29) HIV infection,(30) and neurodegenerative diseases.(31) In this study, we tested whether TNF-α-induced apoptosis of human bone marrow stromal cells (hBMSCs) is mediated by excessive ROS generation and whether α-LA can protect hBMSCs against TNF-α-induced apoptosis.

MATERIALS AND METHODS

Chemicals and reagents

Recombinant human TNF-α was purchased from R&D Systems (Minneapolis, MN, USA), and H2O2 was obtained from Sigma-Aldrich (St Louis, MO, USA). The pan-caspase inhibitor Z-Val-Ala-Asp-fluoromethylketone (Z-VAD-fmk), α-LA, Cell-Permeable JNK inhibitor I, and Cell-Permeable NF-κB SN50 inhibitor peptide were purchased from Calbiochem (Darmstadt, Germany). The fluorescent probe, 2′,7′-dichlorofluorescein diacetate (DCFH-DA), was purchased from Molecular Probes (Seoul, Korea). Antibodies against JNK1 and p-JNK were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

hBMSC culture

The human bone marrow stromal cell line, HS-5, was purchased from American Type Culture Collection (Manassas, VA, USA) and cultured in a 37°C, 5% CO2, humidified chamber, with growth medium consisting of DMEM, 10% FBS, and 1% penicillin/streptomycin. The medium was changed twice a week. On growth to 80% confluence, the cells were subcultured using 0.01% trypsin/0.05% EDTA.

Primary hBMSCs and an immortalized osteoblast cell line were also used in the experiments. The primary hBMSCs were isolated from ribs that were discarded at the time of open thoracotomy in patients without metabolic bone disease and were cultured in α-MEM (Sigma-Aldrich) containing 10% FBS and 1% penicillin/streptomycin, as previously described.(32) The human ribs were obtained under the approved protocol by the Ethics Committee of Asan Medical Center. Previous studies have shown that these cells possess many of the phenotypic characteristics of differentiated osteoblasts and that the absence of monocytic cells and adipocyte was confirmed.(32–34) Human fetal osteoblast cells hFOB1.19 were cultured in DMEM:F12 containing 10% FBS at 34°C, as described previously.(35) Primary hBMSCs and hFOB1.19 cells were incubated at 0.1% FBS for 24 h before treatment of TNF-α or H2O2 to reduce background growth factor response in cell viability and apoptosis assays.

Characterization of bone cell phenotype of HS-5 cells

HS-5 cells were cultured in an osteogenic media (DMEM with 10% FBS, 50 μg/ml α-ascorbic acid, 10 mM β-glycerophosphate, and 100 nM dexamethasone). Alkaline phosphatase (ALP) staining was performed according to the protocol provided by Sigma-Aldrich. The stained cells were imaged using a camera (Canon, NY, USA) connected to a light microscope (Axiovert 40C; Carl Zeiss, Gottingen, Germany).

ALP and osteocalcin assays were also performed. HS-5 cells were cultured for 2 days in DMEM containing 10% FBS, and then incubated in DMEM containing 0.1% BSA (Sigma-Aldrich) and 50 nM 1,25(OH)2D3 (Calbiochem). The cell layer was washed with PBS, and ALP activity was measured using the p-nitrophenol phosphatase hydrolysis method. For osteocalcin assay, the conditioned media were centrifuged free of cellular debris, and the concentration of osteocalcin in the supernatant was measured by immunoassay using a Novocalcin commercial kit (Metra Biosystems, Mountain View, CA, USA). Both ALP activity and osteocalcin concentration were normalized with total cellular protein content, which was determined by the Lowry method.

Cell viability assay

Cell viability was determined by the 3-(4,5-dimethylthiazol-2-yl)2,5-diphenyltetrazolium bromide (MTT)-dye reduction microassay(36) or by trypan blue exclusion.(37) For the MTT assay, the cells were incubated for 48 h, after which 10 μl of MTT (Sigma-Aldrich) was added for 2 h to each well of a 96-well microplate, and the absorbance was read at 450 nm using a microplate reader (SPECTRAmax 340 PC; Molecular Devices, Palo Alto, CA, USA) with a reference wavelength at 650 nm. Viable cells were also quantified by counting after harvesting the cells with trypsin digestion and adding an equivalent volume of 0.4% trypan blue solution to an aliquot of resuspended cells for 5 minutes. The stained and unstained cells were counted by means of a hemacytometer.

Apoptosis measurement

Cell death induced by treatment with TNF-α or H2O2 was measured using a Cell Death Detection ELISAPLUS kit (Roche Applied Science) following the manufacturer's instructions. Briefly, cells in a 96-well plate were incubated with 20 ng/ml of TNF-α or 300 μM H2O2 for 2 days, with or without pretreatment for 1 day with 0.5 mM α-LA, lysated (200 μl of lysis buffer), and centrifuged. The supernatant was removed for analysis of cytoplasmic histone-associated DNA fragments. The absorbance of each well was measured at 405 nm in a microplate reader.

Caspase activity assay

The activity of specific caspase (caspase-9, -3, or -8) was measured using ApoAlert Caspase Fluorescent Assay kits (Clontech) following the manufacturer's instructions. Caspase-9 activity was measured using the fluorometric substrate Ac-LEHD-AFC. In brief, cell lysates were incubated with 2.5 mM substrate in caspase assay buffer for 1 h, and fluorescence was measured on a microplate fluorometer (SPECTRAmax GEMINI-XS; Molecular Devices) with excitation at 400 nm and emission at 505 nm. The activities of caspase-3 and -8 were measured in the same way, but using Ac-DEVD-AMC and Ac-IETD-AMC as substrates, with excitation at 360 nm and emission at 460 nm.

Measurement of intracellular ROS production

A fluorescent probe, DCFH-DA, was used to measure intracellular ROS formation in hBMSCs as described.(38) Briefly, cells were incubated with 20 ng/ml of TNF-α or 300 μM H2O2 at 37°C for 6 h, followed by incubation with DCFH-DA (10 μM) at 37°C for 30 minutes. After washing and lysis, the DCF fluorescence in the supernatant was measured using a fluorometer (SPECTRAmax GEMINI-XS; Molecular Devices), with excitation at 500 nm and emission at 530 nm.

Total glutathione assay

Total Glutathione Colorimetric Assay kit (Oxford Biomedical Research) was used to measure the total glutathione content in cell lysates. Briefly, cells (5 × 103) were seeded into 96-well plates and washed with PBS/pH 7.2 after incubation. To each well, 100 μl of 5% metaphosphoric acid (MPA) was added, and the plate was frozen at −80°C and subsequently thawed at 37°C. After two freeze/thaw cycles, 50 μl of each cell lysate was transferred into new microtiter wells, and 50 μl each of 5,5′-dithiobis (2-nitrobenzoic acid; DTNB, 0.5 mg/ml) and glutathione oxidoreductase solution was added to each well. The microtiter plate was incubated for 10 minutes at room temperature, and 50 μl of the reduced form of β-nicotinamide adenine dinucleotide phosphate (β-NADPH2, 0.6 mg/ml) was added to each well. Absorbances were measured at 405 nm after incubation at room temperature for 3 minutes.

Western blot analysis

Cells were harvested by centrifugation, washed in PBS, and resuspended in ice-cold ProPrep Protein Extraction Solution (Intron Biotechnology) containing 20 mM NaF/0.2 mM Na3VO4. Protein concentration in each cell lysate was determined using Coomassie reagent (Pierce Chemical Co.). Cell lysates (20 μg of total protein per lane) were separated by SDS-PAGE and transferred to HyBond ECL nitrocellulose membranes (Amersham Biosciences). After blocking with 5% skim milk/10 mM Tris-HCl, pH 7.4/150 mM NaCl/0.1% Tween 20, the membrane was incubated for 2 h at room temperature with α-JNK1 or α-p-JNK antibodies. Specific antibody binding was detected by horseradish peroxidase-conjugated secondary antibodies (Santa Cruz Biotechnology) and visualized using enhanced chemiluminescence detection reagent (Amersham Biosciences). The band intensity was quantified by densitometric analysis using Quantity One software (VersaDoc Model 3000 Imaging System; BioRad).

Activated NF-κB measurement

Nuclear extracts were prepared as previously described,(39) and protein concentration in the nuclear extracts was determined using Coomassie reagent. Activated NF-κB in nuclear extracts was measured using an Enzyme Immunoassay kit for NF-κB (Oxford Biomedical Research), following the manufacturer's instructions. Briefly, 60 μl of the diluted activated NF-κB binding buffer was added to each well, which had been coated with an oligonucleotide containing the NF-κB binding consensus sequence. To each well was added 60 μl nuclear extract containing 2 μg total protein, and the plate was incubated at room temperature with gentle orbital shaking for 2 h. Each well was washed four times with 350 μl of Washing Buffer, and 100 μl of diluted α-p50/p105 primary antibody was added to each well. After incubation for 1 h at room temperature, the wells were washed three times, and specific primary antibody binding was detected by ALP-conjugated secondary antibody. The relative luciferase units (RLUs), correlated with the amount of activated NF-κB in the nuclear extracts, were measured using a chemiluminescence detector (SPECTRAmax GEMINI-XS; Molecular Devices).

Statistics

All data are expressed as means ± SD. Between-group differences were assessed using the Mann-Whitney U-test, and differences among three or more groups were assessed by ANOVA with posthoc analysis by Duncan's multiple range test. A p value <0.05 was considered statistically significant. SPSS 10.0 package (SPSS, Chicago, IL, USA) was used for statistical procedures.

RESULTS

Characteristics of bone cell phenotype of HS-5 cells

To test effects of TNF-α and H2O2 on hBMSCs, we elected to use the readily available HS-5 cell line derived from normal bone marrow for these studies. The HS-5 cells possessed osteoblast-lineage characteristics as shown by ALP staining after incubation for 3 days in osteogenic media, and the staining was further increased after 10 days (Fig. 1A). Consistent with this result, ALP activity and osteocalcin secretion were detected after incubation for 3 days and increased after 7 days (Figs. 1B and 1C). Additionally, we could detect RANKL in the conditioned media and cellular extracts of HS-5 cells (data not shown). These data supported that the HS-5 cells were capable of osteoblastic differentiation.

Figure FIG. 1..

Characterization of bone cell phenotype of HS-5 cells. Cells were cultured for the indicated times. ALP staining (A), ALP activity (B), and osteocalcin secretion (C) were measured. (A) Representative stained cells (magnification, ×300). (B and C) Averages and SDs of three independent experiments.

Cytotoxicity of TNF-α or H2O2 in hBMSCs

To examine the effect of TNF-α or H2O2 on HS-5 cells, we first measured the effect of each on cell viability. Using the MTT assay, we found that incubation with 20 ng/ml TNF-α or 300 μM H2O2 for 2 days significantly reduced cell counts (Fig. 2A; p < 0.01). Reduced survival was also confirmed by trypan blue exclusion, in that treatment with TNF-α or H2O2 reduced viable cells to 48.9 ± 6.5% and 35.8 ± 15.2%, respectively, of the control level (Fig. 2B; p < 0.01).

Figure FIG. 2..

TNF-α- or H2O2-induced cytotoxicity in hBMSCs. hBMSCs were treated with 20 ng/ml TNF-α or 300 μM H2O2 for 2 days, and cell viability was measured by the MTT assay (A) or trypan blue exclusion (B). Shown are the averages and SDs of six independent experiments. *p < 0.01 compared with untreated control.

Caspase-dependent cell death by TNF-α or H2O2 in hBMSCs

Caspase-dependent apoptosis of HS-5 cells was confirmed by Cell Death Detection ELISAPLUS kit assay. TNF-α and H2O2 significantly increased cell apoptosis (Fig. 3A; p < 0.01). Preincubation of the cells with 50 μM Z-VAD-fmk, a pan-caspase inhibitor, nearly completely blocked the cytotoxic effects of TNF-α and H2O2, both in the cell death detection ELISA (Fig. 3A) and MTT assays (Fig. 3B). Thus, induction of hBMSC apoptosis by TNF-α or H2O2 seems to be through a caspase-dependent pathway.

Figure FIG. 3..

Caspase-dependent cell death by TNF-α or H2O2 in hBMSCs. hBMSCs were pretreated with 50 μM Z-VAD-fmk for 2 h followed by treatment with 20 ng/ml TNF-α or 300 μM H2O2 for 2 days. Cell viability was measured by the Cell Death Detection ELISAPLUS assay (A) or the MTT assay (B). Shown are the averages and SDs of four independent experiments. *p < 0.01 compared with untreated control.

To delineate mechanisms leading to caspase-dependent apoptosis in TNF-α- or H2O2-treated HBMSC, we first measured activities of caspases using fluorescent probes. Incubation of HS-5 cells for 4 h with TNF-α or H2O2 increased caspase-3 activity, peaking after 12 h (Fig. 4A). Similarly, TNF-α and H2O2 also elevated activities of caspase-9 (Fig. 4B) and caspase-8 (Fig. 4C).

Figure FIG. 4..

Activation of caspases by TNF-α or H2O2 in hBMSCs. hBMSCs were treated with 20 ng/ml TNF-α or 300 μM H2O2 for the indicated times. Cellular extracts were prepared, and caspase-3 activity was measured using Ac-DEVD-AMC (A), caspase-9 activity was measured using Ac-LEHD-AFC (B), and caspase-8 activity was measured using Ac-IETD-AMC (C). Results are expressed as fluorescence ratios of the treated over the untreated control. Shown are the averages and SDs of five independent experiments. *p < 0.05; **p < 0.01 compared with untreated control.

Increase of intracellular ROS levels and decrease of total glutathione levels in hBMSCs treated with TNF-α or H2O2

To determine whether treatment of HS-5 cells with TNF-α or H2O2 causes a change in cellular redox status, we measured intracellular ROS and total glutathione (GSH). Dichlorofluorescein (DCF) fluorescence, an indicator for intracellular ROS levels, was increased after treatment for 5 h with TNF-α or H2O2, to 162.3 ± 7.5% or 177.5 ± 4.1% of control, respectively (Fig. 5A; p < 0.01). Figure 5B shows that the treatment with TNF-α or H2O2 for 4 or 12 h also decreased the total cellular GSH level (p < 0.01). These data indicate that both TNF-α and H2O2 induced oxidative stress in hBMSCs.

Figure FIG. 5..

Increase of intracellular ROS levels and decrease of GSH levels in hBMSCs treated with TNF-α or H2O2. (A) hBMSCs were treated with 20 ng/ml TNF-α or 300 μM H2O2 for 6 h. Cell lysates were prepared, and intracellular ROS was measured using the fluorescent probe, DCFH-DA. Results are expressed as fluorescence ratios of the treated over the untreated control (cells incubated with DCFH-DA only). Shown are the averages and SDs of five independent experiments. (B) hBMSCs were treated with 20 ng/ml TNF-α or 300 μM H2O2 for 4 and 12 h. Cell lysates were prepared, and GSH content was determined by the dithionitrobenzoic acid-glutathione disulfide (DTNB-GSSG) reductase recycling colorimetric assay. Shown are the averages and SDs of four independent experiments. *p < 0.01 compared with untreated control.

α-LA inhibition of cell death mediated by TNF-α- or H2O2-induced oxidative stress

To test if the oxidative stress induced by TNF-α and H2O2 was the culprit in hBMSC death, we used α-LA, a potent biological thiol antioxidant. Preincubation of HS-5 cells with α-LA prevented the decrease in GSH (Fig. 6A), as well as protecting these cells against apoptosis as shown by trypan blue exclusion (Fig. 6B; p < 0.01) and by cell death detection ELISA assay (Fig. 6C; p < 0.01) after TNF-α or H2O2 treatment. Moreover, α-LA inhibited the increases in caspase-3 (Fig. 6D; p < 0.01), caspase-9 (Fig. 6E; p < 0.01), and caspase-8 (Fig. 6F; p < 0.01) activities induced by TNF-α or H2O2.

Figure FIG. 6..

Inhibition of TNF-α- or H2O2-induced decrease in GSH levels, caspases activation, and apoptosis by α-LA. hBMSCs were pretreated with 0.5 mM α-LA for 1 day before the addition of 20 ng/ml TNF-α or 300 μM H2O2. (A) Cell lysates were prepared after incubation with TNF-α or H2O2 for 12 h, and GSH levels were determined. Shown are the averages and SDs of four independent experiments. (B and C) Cell viability was measured after incubation with TNF-α or H2O2 for 2 days by trypan blue exclusion (B) or Cell Death Detection ELISAPLUS assay (C). Shown are the averages and SDs of six independent experiments. (D-F) Cell lysates were prepared after incubation with TNF-α or H2O2 for 6 h, and caspase-3 (D), caspase-9 (E), and caspase-8 (F) activities were measured. Results are expressed as fluorescence ratios of the treated over the untreated control. Shown are the averages and SDs of four independent experiments. *p < 0.01 compared with untreated control.

JNK and NF-κB signaling involved in TNF-α- or H2O2-induced cell death in hBMSCs

Oxidants have been shown to regulate the activation of JNK, and oxidant-induced JNK activation mediates apoptosis.(40,41) In addition, oxidative stress has been implicated in the activation of NF-κB.(42,43) Therefore, to elucidate the mechanisms mediating the apoptotic effects of TNF-α- or H2O2-induced oxidative stress, we assayed their effects on JNK and NF-κB signaling. We found that JNK activity, measured as an increase in phosphorylated JNK (p-JNK) band density, was stimulated by TNF-α and H2O2 in HS-5 cells after 15 minutes of incubation, to 240.6 ± 8.4% and 139.4 ± 7.3%, respectively, of untreated control (Figs. 7A and 7B; p < 0.05 each). JNK activity continued to increase, reaching 487.8 ± 19.6% and 236.0 ± 16.3%, respectively, at 2 h (p < 0.01 each). Pretreatment of cells with α-LA for 1 day almost completely blocked the activation of JNK induced by TNF-α or H2O2 for 2 h (Figs. 7A and 7B). We also found that treatment of cells with TNF-α or H2O2 for 4 h induced 325.5 ± 29.6% or 224.5 ± 41.2%, respectively, increases in NF-κB activity as shown by the increase in relative luciferase units (Fig. 7C; p < 0.01). More importantly, pretreatment of cells with α-LA for 1 day completely inhibited NF-κB activation by TNF-α and H2O2 (Fig. 7C).

Figure FIG. 7..

Activation of JNK and NF-κB by TNF-α or H2O2 is blocked by α-LA in hBMSCs. (A and B) hBMSCs were treated with 20 ng/ml TNF-α (A) or 300 μM H2O2 (B) for 15 minutes, 1 h, and 2 h with or without pretreatment with 0.5 mM α-LA for 1 day. Cells were harvested at the indicated times, and levels of JNK and p-JNK were detected by Western blot analysis. Lane 1, untreated control; lanes 2-4, 15 minutes, 1 h, and 2 h incubation with TNF-α or H2O2; lane 5, α-LA only; lane 6, 2 h of incubation with TNF-α or H2O2 after 1 day pretreatment with α-LA. The p-JNK band intensity was quantified by densitometric analysis and expressed as percentage of the treated over the untreated control. Shown are the averages and SDs of three independent experiments. (C) hBMSCs were treated with 20 ng/ml TNF-α or 300 μM H2O2 for 4 h, with or without pretreatment with 0.5 mM α-LA for 1 day. Nuclear extracts were prepared, and activated NF-κB was measured by enzyme immunoassay. Results are expressed as change in relative luciferase units (RLUs) of the treated over the untreated control. Shown are the averages and SDs of four independent experiments. *p < 0.05; **p < 0.01 compared with untreated control.

To further examine the direct involvement of JNK and NF-κB activation in hBMSC apoptosis induced by TNF-α or H2O2, we measured caspase-3 activities after treatment with cell-permeable inhibitors of JNK and NF-κB. When pretreated for 1 h with 1 μM of the JNK inhibitor or 18 μM of the NF-κB inhibitor, the increase in caspase-3 activity induced by TNF-α (Fig. 8A; p < 0.01) or H2O2 (Fig. 8B; p < 0.01) was blocked. Apoptotic cell death induced by TNF-α and H2O2 was also inhibited by these inhibitors (Figs. 8C and 8D; p < 0.01 each). These data indicate that TNF-α and H2O2 each activate JNK and KF-κB signals in hBMSCs and that these pathways may play an important role in TNF-α- or H2O2-induced hBMSC apoptosis.

Figure FIG. 8..

Direct involvement of JNK and NF-κB activation in hBMSC apoptosis induced by TNF-α or H2O2. (A and B) hBMSCs were treated with 20 ng/ml TNF-α (A) or 300 μM H2O2 (B) for 6 h, with or without pretreatment with 1 μM of cell-permeable JNK inhibitor I (JNK i) or 18 μM of cell-permeable NF-κB SN50 inhibitor peptide (NF-κB i) for 1 h. Cell lysates were prepared, and caspase-3 activity was determined. Results are expressed as fluorescence ratios of the treated over the untreated control. Shown are the averages and SDs of four independent experiments. (C and D) hBMSCs were treated with 20 ng/ml TNF-α or 300 μM H2O2 for 2 days, with or without pretreatment with 1 μM of cell-permeable JNK inhibitor I or 18 μM of cell-permeable NF-κB SN50 inhibitor peptide for 1 h. Cell viability was measured after incubation with TNF-α (C) or H2O2 (D) by Cell Death Detection ELISAPLUS assay. Shown are the averages and SDs of six independent experiments. *p < 0.05; **p < 0.01 compared with untreated control.

Prevention of TNF-α- or H2O2-induced cell death by α-LA in primary hBMSCs and an osteoblast cell line

To confirm our results, we further tested the effects of TNF-α and H2O2 using primary hBMSCs and the hFOB1.19 osteoblast cell line. Like HS-5 cells, primary hBMSCs and hFOB1.19 also showed similar response; TNF-α or H2O2 induced apoptosis, which was prevented by α-LA pretreatment (Figs. 9A-9D).

Figure FIG. 9..

α-LA prevents TNF-α- or H2O2-induced apoptosis of primary hBMSCs and the hFOB1.19 osteoblast cell line. Primary hBMSCs (A and B) and hFOB1.19 osteoblast cells (C and D) were pretreated with 0.5 mM α-LA for 1 day, followed by treatment with 20 ng/ml TNF-α or 300 μM H2O2 for 2 days. Cell viability was measured by the MTT assay (A and C) and Cell Death Detection ELISAPLUS assay (B and D). Shown are the averages and SDs of three independent experiments. *p < 0.01 compared with untreated control.

DISCUSSION

TNF-α affects bone by multiple actions, including apoptosis of cells of the osteoblast lineage. However, the mechanism by which TNF-α induces apoptosis in these cells has not been fully characterized. Here we show that TNF-α-induced apoptosis in hBMSCs is mediated largely by ROS-mediated mitochondrial activation, involving the activation of JNK and NF-κB. We also show that α-LA, a potent thiol antioxidant, nearly rescued cells from apoptosis induced by TNF-α or H2O2, by reducing oxidative stress and by inhibiting the JNK and NF-κB signaling pathways. To our knowledge, this is the first demonstration of a mechanism of TNF-α-induced apoptosis in hBMSCs and of the protective role of α-LA in hBMSC apoptosis.

TNF-α, a major mediator of inflammation, may act as a low-grade stimulus after menopause and during aging.(14) Substantial in vitro and in vivo evidence suggests that TNF-α plays a central role in the pathophysiology of skeletal loss after menopause.(12,13) This was shown most definitively in experiments showing that ovariectomy-induced bone loss was prevented by blockade of TNF-α action and that TNF-α-deficient and TNF-α receptor knockout mice were resistant to ovariectomy-induced bone loss.(44–46) TNF-α increases bone resorption and simultaneously inhibits new bone formation. TNF-α has been found to inhibit osteoblast activity by suppressing mature osteoblast function, by blocking the differentiation of new osteoblasts from their progenitors, and by inducing osteoblast resistance to 1,25-dihydroxyvitamin D3.(14) TNF-α has also been shown to induce apoptosis in rodent osteoblasts.(15–18) In agreement with previous studies, we have shown that TNF-α induced apoptosis in HS-5 cells, primary hBMSCs, and hFOB1.19 cells, suggesting that apoptosis of osteoblast-lineage cells is important in the decreased bone formation induced by TNF-α in humans.

Caspases play a crucial role in apoptotic cell death. Two different pathways lead to the activation of executioner caspases, which are responsible for the morphological features of apoptosis.(18) One pathway involves the activation of initial caspases-8, -10, and -2 as a result of their binding to TNF-α receptor-associated adaptor proteins. Alternatively, death signaling to the mitochondria leads to the formation of a functional apoptosome, resulting in the activation of another initiator caspase, caspase-9. ROS is thought to induce a drop in mitochondrial membrane potential,(47) to release cytochrome c from mitochondria, and to induce apoptosis cascade (activating caspases-9 and -3), which ultimately leads to apoptosis.(48,49) We have shown here that TNF-α decreased cellular total glutathione, increased intracellular ROS, and activated caspases-9 and -3, indicating that excessive ROS generation and the mitochondrial pathway are important in TNF-α-induced apoptosis in hBMSCs.(50–53) In addition, TNF-α may induce apoptosis of hBMSCs by the death receptor-mediated pathway, through TNF-α-activated caspase-8. Procaspase-8 was recently shown to be localized in the mitochondria of human fibroblasts and mouse clonal striated cells and to be released from mitochondria by TNF-α.(54) These results suggest that during TNF-α-induced apoptosis in hBMSCs, the mitochondrial pathway may also contribute to the activation of caspase-8.

α-LA in the diet is transported to tissues and incorporated into cells, where it is translocated into the mitochondria, which contain α-LA-requiring enzyme complexes. In mammals, α-LA is not sufficiently supplied by diet, and de novo synthesis takes place in the heart, liver, and testis.(55) Hence, the concentration of free α-LA in the circulation is very low, and supplementation is necessary to reach potential therapeutic levels. In our experiments, 0.5 mM α-LA nearly rescued apoptosis induced by H2O2 and TNF-α. At concentrations >1.0 mM, however, α-LA stimulated apoptosis (data not shown), indicating that α-LA exerts both antioxidant and prooxidant properties in osteoblast-lineage cells depending on its concentration. These data are consistent with previous observations in other systems.(56) For example, in human lymphocytes, low concentrations (0.01-0.1 mM) of α-LA showed antioxidant activity, increasing intracellular thiol concentrations, whereas higher concentrations (2-5 mM) had prooxidant activity. Thus, it is important to pinpoint a dose of α-LA that shows optimal beneficial effects on bone. Our finding that pretreatment with α-LA prevented all the changes induced by TNF-α and H2O2 suggests that the anti-apoptotic effects of α-LA are largely the result of their antioxidant properties, although other protective mechanisms have been suggested.(57) Recently, antioxidants have been shown to prevent bone loss induced by estrogen deficiency.(58) Because TNF-α plays an important role mediating bone loss induced by estrogen deficiency,(12,13) these data support the antagonistic effects of antioxidants on TNF-α activities.

Oxidants have been shown to regulate the activation of JNK(59,60) as well as the transcription factor NF-κB.(42) The contribution of JNK to many phenotypic outcomes, including survival(61–64) and apoptosis,(65,66) seems to depend on cell type, stimulus, and the duration of JNK activation, as well as the engagement of other signaling modules.(64) Oxidant-induced JNK activation has generally been linked to apoptosis,(40,41) and JNK is strongly activated during TNF-α- and H2O2-induced apoptosis in hepatocytes(59) and lung fibroblasts.(60) Transcription factor NF-κB is also activated by oxidative stress,(42) but there is uncertainty about the role of redox events in NF-κB activation.(43,67) For example, NF-κB can be pro- or anti-apoptotic, depending on the timing of modulating NF-κB activity relative to the death stimulus.(68) Soy isoflavone supplementation has been shown to protect blood lymphocytes from oxidative stress-inducing agents by inhibiting NF-κB activation and by decreasing DNA adduct levels.(69) NF-κB activation has also been found to mediate TNF-α-induced apoptosis in murine clonal osteoblasts.(15) Our study showed that TNF-α and H2O2 activated JNK and NF-κB in hBMSCs and that inhibitors of JNK or NF-κB prevented TNF-α- or H2O2-induced caspase activation and subsequent cell death. These results suggest the importance of JNK and NF-κB signaling in the apoptotic mechanisms of hBMSCs. Furthermore, we have found that α-LA exerts its anti-apoptotic effects by inhibiting JNK and NF-κB signaling upstream of these signaling modules.

In conclusion, we have shown that TNF-α-induced apoptosis of hBMSCs is mediated largely by excessive generation of ROS. TNF-α and H2O2 share common signaling modules, including JNK and NF-κB, in the apoptotic machinery. In addition, α-LA effectively protects against oxidative stress-induced apoptosis of hBMSCs by reducing oxidative stress and by inhibiting activation of JNK and NF-κB. Treatment with α-LA may serve in the future as a strategy against bone loss associated with increased oxidative stress, including estrogen deficiency and aging.

Acknowledgements

This work was supported by the Korea Health 21 R&D Project form Korean Ministry of Health and Welfare (01-PJ3-PG6-01GN11-0002) and 21C Frontier Proteomics Project from Korean Ministry of Science and Technology (FPR02A3-5-130).

Ancillary