ATM protects against oxidative stress induced by oxidized low-density lipoprotein

Highlights

► OxLDL promotes induction of pATM and p21 in fibroblasts/endothelial cells. ► OxLDL activates ATM-kinase by a doubled-stranded DNA break-independent mechanism. ► ATM-deficient fibroblasts exhibit low cell viability but increased oxidative stress towards oxLDL. ► ATM expression exerts protective effects against ox-LDL-induced cellular toxicity.

Abstract

Chronic oxidative stress is involved in the pathogenesis of multiple inflammatory diseases, including cardiovascular disease and atherosclerosis. The rare autosomal recessive disorder Ataxia-telangiectasia (A-T) is characterized by progressive cerebellar ataxia secondary to Purkinje cell death, immunodeficiency, and increased cancer incidence. ATM, the protein mutated in A-T, plays a key role in cellular DNA-damage responses. A-T cells show poor cellular anti-oxidant defences and increased oxidant sensitivity compared to normal cells, and ATM functions, in part, as an oxidative stress sensor. The oxidation of low-density lipoprotein (oxLDL) and its uptake by macrophages is an initiating step in the development of atherosclerosis. We demonstrate that oxLDL activates ATM and downstream p21 expression in normal fibroblasts and endothelial cells. In ATM-deficient fibroblasts oxLDL induces DNA double-strand breaks, micronuclei formation and causes chromosome breaks. Furthermore, oxLDL decreases cell viability and inhibits colony formation in A-T fibroblasts more effectively as compared to normal controls. Formation of oxLDL-induced reactive oxygen species is significantly higher in A-T, than normal fibroblasts. Last, pre-treatment of cells with ammonium pyrrolidine dithiocarbamate, a potent antioxidant and inhibitor of transcription factor nuclear factor κB, reduces oxLDL-induced reactive oxygen species formation. Our data indicates that ATM functions in the defence against oxLDL-mediated cytotoxicity.

1. Introduction

Reactive oxygen species (ROS) are generated constantly as by-products of cellular metabolism, particularly by mitochondrial respiration [1,2]. At normal cellular concentrations, ROS play a role in regulating cell signalling pathways and gene expression [1,2]. However, when the production of ROS exceeds cellular antioxidant capacity, damage to cellular macromolecules such as lipids, proteins, and DNA may occur [2,3]. To combat such damage organisms have evolved anti-oxidant protective systems, including the glutathione/glutathione disulfide system, superoxide dismutase, catalase, metal chelation, and diverse repair systems that maintain redox homeostasis [3,4]. An imbalance between ROS-generating and scavenging systems is called “oxidative stress” and plays a crucial role in a variety of pathological disorders, among them cardiovascular and neurodegenerative diseases.

Ataxia telangiectasia (A-T) is a progressive neurodegenerative disease manifesting in early childhood. The clinical features of A-T include progressive ataxia secondary to cerebellar Purkinje cell death, premature aging, immunodeficiency, and increased cancer risk; especially for leukaemia and lymphoma [5]. Patients with A-T lack functioning A-T mutated protein (ATM), a member of the phosphatidylinositol 3-kinase like family of serine/threonine protein kinases [6]. ATM-deficient cells exhibit chromosomal instability and extreme sensitivity to DNA-double strand break (DSB)-inducing agents, such as ionizing radiation [7]. Hence, the most studied function of ATM is its role in response to DNA damage. When DNA-DSBs occur, ATM is rapidly activated by autophosphorylation at Ser¹⁹⁸¹ [8], and in turn rapidly phosphorylates a number of substrates involved in DNA replication and repair, cell cycle checkpoint control, and apoptosis [9,10]. However, there is evidence that A-T is not only due to a defect in DNA-DSB response, but also to a diminished control of ROS. Studies revealed that ATM-deficient cells are in a constant state of oxidative stress [11]. Reichenbach and co-workers [12] reported that the plasma of A-T patients exhibit a decreased antioxidant capacity. Treatment with antioxidants e.g. N-acetyl-l-cysteine and tempol, increased the lifespan of Atm^−/− mice and tempol-treatment further decreased levels of ROS and oxidative damage in thymocytes of mice [13,14]. Moreover, ATM is activated by oxidants such as t-butyl hydroperoxide and H₂O₂ [15–17]. Additionally, H₂O₂-induced phosphorylation of ATM can be blocked by N-acetyl-l-cysteine, indicating that ATM phosphorylation is responsive to redox imbalance [15].

ROS act as signalling intermediates in many normal cellular processes, and elevated ROS levels are linked to many pathological conditions including neurodegenerative diseases, diabetes, cancer, and atherosclerosis, respectively [18]. The atherosclerotic lesion is characterized by an accumulation of lipids carried by lipoproteins, such as low-density lipoprotein (LDL). LDL becomes susceptible to (non)enzymatic oxidative modification when retained in the artery wall [19]. These modifications make the LDL particle a potent affector of cellular functions. In particular, the uptake and degradation of oxidized LDL (oxLDL) by monocyte-derived macrophages is considered the leading event in the formation of cholesterol-enriched foam cells, which are the hallmark of fatty streaks, the earliest recognizable lesion of atherosclerosis [19,20].

Currently, there is no data linking ATM to the cellular responses following oxLDL exposure. However, there is indirect evidence that ATM may be involved in oxLDL-induced signalling pathways. Apparently as a consequence of increased levels of plasma cholesterol, heterozygous ATM deficiency may increase the risk of atherosclerosis-related cardiovascular disease in humans [21]. Apolipoprotein E^−/− (ApoE^−/−) mice heterozygous in Atm developed accelerated atherosclerosis and multiple features of the metabolic syndrome including glucose intolerance, hypertension, obesity and hypercholesterolemia [22,23]. Transplantation of ApoE^−/−/Atm^+/+ mice with bone marrow from ApoE^−/−/Atm^+/+ or ApoE^−/−/Atm^−/− mice revealed 80% increase in lesion severity in animals treated with Atm null bone marrow [24].

In the present study, we investigated the role of ATM in protection against toxicity of copper-oxLDL [25], a commonly used experimental model for oxidative modification of LDL. Here we studied the effect of oxLDL on ATM activation and downstream signalling in normal fibroblasts and endothelial cells. We also investigated DNA damage in normal and ATM-deficient fibroblasts. Third, we studied the cytotoxicity of oxLDL on normal and ATM-deficient fibroblasts and last, we examined the effect of ATM status on oxLDL-induced ROS formation in these cells.

2. Materials and methods

2.1. Materials

Cell culture dishes, flasks and microtiterplates were from Greiner Bio-One (Frickenhausen, Germany). Dulbecco's Modified Eagle Medium (DMEM), penicillin/streptomycin and hygromycin B were from Gibco Invitrogen (Lofer, Austria), foetal calf serum (FCS) was from PAA (Linz, Austria) and bovine serum albumin (BSA) was from Serva (Heidelberg, Germany). Lysis buffer components HEPES, EDTA, glycerol, Triton X-100, Na₄P₂O₇ and Na₃VO₄ were from Sigma–Aldrich (Saint Louis, MO, USA) and NaF was from Fluka (Buchs, Switzerland). Complete Mini protease inhibitor cocktail tablets were from Roche Diagnostics (Mannheim, Germany). Trypan blue stain (0.4%), NuPAGE^® 4–12% Bis-Tris Gels, NuPAGE^® LDS sample buffer, NuPAGE^® MOPS running buffer and nitrocellulose membranes were from Invitrogen™ life technologies (Carlsbad, CA, USA).

Bisbenzimide (Hoechst 33258), 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT, Thiazolyl blue), ammonium pyrrolidine dithiocarbamate (PDTC), crystal violet and Triton X-100 were from Sigma–Aldrich (Saint Louis, MO, USA). Carboxy-H₂DCFDA (5-(and-6)-carboxy-2′,7′-dichlorodihydrofluorescein diacetate) was from Gibco Invitrogen (Lofer, Austria). Staurosporine and the ATM-kinase-inhibitor (ATM-I, 2-morpholin-4-yl-6-thianthren-1-yl-pyran-4-one) were from Calbiochem (Merck, Darmstadt, Germany). BCA™ Protein Assay Kit and Super Signal West Pico Chemiluminescent substrate were from Pierce Biotechnology, Inc. (Rockford, IL, USA). Immobilon™ Western Chemiluminescent HRP Substrate was from Millipore Corporation (Billerica, MA, USA). H₂O₂ was from Herba Chemosan (Vienna, Austria); colcemid was from Irvine Scientific (Santa Ana, CA, USA). All other chemicals were from Roth (Vienna, Austria) or Sigma–Aldrich (Saint Louis, MO, USA).

The following primary antibodies were used: (i) polyclonal rabbit phospho-ATM antibody (recognizing human ATM phosphorylated at Ser¹⁹⁸¹, termed pATM, R&D Systems, Minneapolis, MN, USA); (ii) sequence-specific polyclonal rabbit anti-ATM antibodies raised against synthetic peptides corresponding to amino acids 819–844 (Calbiochem) or 2550–2600 (Abcam, Cambridge, U.K.) of human ATM; (iii) polyclonal rabbit anti-caspase-3 antibody (raised against full-length caspase-3, Santa Cruz Biotechnology, Santa Cruz, CA, USA), (iv) polyclonal anti-β-tubulin (Santa Cruz Biotechnology); (v) polyclonal phospho-histone H2AX antibody (recognizing H2AX only when phosphorylated at Ser¹³⁹, termed γ-H2AX, Cell Signaling Technology, Beverly, MA, USA); (vi) rabbit monoclonal anti-p21 Waf1/Cip1 antibody (12D1, Cell Signaling Technology, Beverly, MA, USA), (vii) monoclonal anti-β-actin antibody (Santa Cruz Biotechnology); (viii) monoclonal anti-Poly (ADP-ribose) polymerase (PARP) antibody (recognizing the 116 kDa and the 85 kDa apoptosis-related cleaved fragment, Biomol, Hamburg, Germany). The following secondary antibodies were used: (i) HRP-conjugated goat anti-mouse IgG (Rockland, Gilbertsville, PA, USA) and (ii) HRP-conjugated goat-anti-rabbit IgG (Pierce Biotechnology, Inc., Rockford, IL, USA).

2.2. Cell culture

WI-38 VA13 (VA13) is a SV-40-immortalized fibroblast cell line (ATCC, Rockville, MD). AT22IJE-T (AT22) is an ATM-deficient SV-40 immortalized fibroblast cell line, originally established from primary A-T fibroblasts [26]. VA13 and AT22 cells were grown in DMEM with 1 g/l glucose, 4 mM l-glutamine, 110 mg/l sodium pyruvate and 25 mM HEPES, supplemented with 5% (v/v) FCS and 100 U/ml penicillin/streptomycin. Human EA.hy926 endothelial cells were grown in DMEM with 4.5 g/l glucose, 3.97 mM l-glutamine and 1 mM sodium pyruvate supplemented with 10% (v:v) FCS, 1% penicillin-streptomycin and 1 × HAT supplement [27]. All three cell lines were cultured at 37 °C in a humidified atmosphere of 5% CO₂ and 37 °C

2.3. Isolation and modification of LDL

LDL was isolated by ultracentrifugation (d = 1.027–1.063 g/ml) from fresh human plasma, obtained from healthy volunteers [28,29]. LDL was sterile-filtered and stored at 4 °C. Prior to oxidation, LDL was dialyzed overnight against PBS (10 mM; pH 7.4) at 4 °C. Oxidation of 500 μg/ml LDL was performed with a final concentration of 30 μM Cu₂SO₄ for 18 h. EDTA (final concentration 2.7 mM) terminated the reaction, the samples were saturated with N₂ and stored at 4 °C. Characterization of oxLDL was performed as described [30]. Before use, oxLDL was sterile-filtered and adjusted to a final protein concentration of 1 mg/ml by dialysis under high pressure against PBS (10 mM) (to remove EDTA and unreacted Cu₂SO₄). Lipoprotein concentrations are expressed in terms of its protein concentration, determined by the Lowry method using BSA as a standard.

2.4. Cell experiments for Western Blot analysis

VA13, AT22, and EA.hy926 cells were seeded in 6-well plates. When cells reached 70% confluence, they were incubated in serum-free DMEM overnight. Cells were treated with indicated concentrations of lipoproteins for the indicated times. For blockade of the ATM-kinase signalling pathway, cells were pre-incubated with ATM-I [31] (10 or 30 μM dissolved in DMSO) for 1 h. Cells treated with PBS and/or DMSO served as controls. DMSO concentration did not exceed 0.01% (v/v). Alternatively, the cells were treated with 200 μM H₂O₂ for 15 min and after medium-exchange, the cells were incubated for further 90 min.

For protein isolation, the cells were washed twice with ice cold PBS. Cell lysis was performed on ice in 60 μl lysis buffer (50 mM HEPES, 150 mM NaCl, 1 mM EDTA, 10 mM Na₄P₂O₇, 2 mM Na₃VO₄, 10 mM NaF, 1% (v/v) Triton X-100, 10% (v/v) glycerol and Complete Mini protease inhibitor cocktail tablets; pH 7.4) for 10 min [32]. The cell lysates were scraped and insoluble cell debris was removed by centrifugation (13,000 rpm, 4 °C) for 10 min.

To follow expression of γ-H2AX, cleavage of PARP- and procaspase-3, cells were pelleted by centrifugation (1000 rpm, 5 min) and lysed. Protein content of cell lysates was determined using the BCA™ Protein Assay Kit, according to the manufacturer's instructions. Protein lysates (30–50 μg/ml) were diluted with NuPAGE^® LDS Sample Buffer and NuPAGE^® Sample Reducing Agent (10×) and were boiled for 10 min at 70 °C. Proteins were separated in NuPAGE^® 4–12% Bis-Tris Gels and electrophoretically transferred to nitrocellulose membranes [33]. Membranes were first incubated with Tris-buffered saline Tween 20 (TBS-T, containing 5% (w/v) non-fat milk) for 2 h, before incubation with polyclonal rabbit anti-pATM antibody, polyclonal rabbit anti-ATM antibody, polyclonal rabbit γ-H2AX antibody, rabbit monoclonal anti-p21 Waf1/Cip1 antibody, polyclonal rabbit anti-caspase-3 antibody, monoclonal anti-PARP antibody, monoclonal anti-β-actin antibody or polyclonal anti-β-tubulin antibody (dilution 1:500 to 1:1000 in TBS-T containing 5% (w/v) BSA) overnight at 4 °C. Immunoreactive bands were visualized using HRP-conjugated goat anti-rabbit (dilution 1:50,000 in TBS-T containing 5% (w/v) non-fat milk) or goat anti-mouse IgG (dilution 1:100,000 in TBS-T containing 5% (w/v) non-fat milk) for 2 h, and subsequently visualized with Super Signal West Pico Chemiluminescent substrate or Immobilon™ Western Chemiluminescent HRP Substrate.

2.5. MTT assay

VA13 and AT22 cells were seeded in 12-well plates in DMEM with 5% (v/v) FCS. When cells reached 50% confluence, the medium was replaced by serum-free DMEM and the cells were incubated overnight. Then the cells were treated with lipoproteins for the indicated times and at the indicated concentrations. The cells were washed with PBS (10 mM) and incubated with MTT (0.5 mg/ml; dissolved in serum-free medium) for 2 h at 37 °C [32]. The converted dye was solubilised with acidic isopropanol (0.04 M HCl in absolute isopropanol). MTT reduction was assessed by measuring absorption at 570 nm on a microplate reader (Victor Multilabel Counter, Perkin-Elmer, Waltham, MA, USA) and corrected for background absorbance at 630 nm. Each treatment was done in triplicate and values are expressed as percentages of untreated cells.

2.6. Colony forming efficiency assay

Exponentially growing VA13, AT22, and EA.hy926 cells were plated at a density of 1.5 × 10³ cells/100 mm tissue culture dish in the absence or presence of lipoproteins in normal growth medium. When indicated, the cells were preincubated with ATM-I (10 μM dissolved in DMSO) for 1 h before addition of lipoproteins. After 18 h of incubation, the plates were washed 3 times with PBS (10 mM) the medium was replaced, and the cells were cultured for 12 more days. The cells were fixed for 5 min with methanol and stained with crystal violet (1 g/l; dissolved in aqua dest.) and cell clusters of 50 or more cells were counted as “colonies” under a microscope.

2.7. Trypan blue exclusion assay

VA13 and AT22 cells were seeded in 6-well plates until 50% confluence was reached. After overnight serum-starvation, cells were incubated with indicated concentrations of lipoproteins. At the indicated times, the cells were washed with PBS (10 mM), trypsinized and resolved in serum-free DMEM. The cell suspension was mixed with 1:1 (v/v) with 0.4% (w/v) Trypan blue stain. Viable cells, identified by a clear cytoplasm, were counted using Countess™ cell counting chamber slides and the Countess™ Automated Cell Counter (Invitrogen™ life technologies, Carlsbad, CA, USA).

2.8. Micronucleus assay and staining of nuclei with bisbenzimide

VA13 and AT22 cells were seeded in 6-well plates on glass cover slips and cultured in normal growth medium. When cells reached 50% confluence, cells were serum-starved overnight and incubated with 100 μg/ml lipoprotein for 16 h. Cells were washed with PBS (10 mM) and fixed with 90% (v/v) methanol for 5 min. Staining of nuclei was performed with 0.5 μg/ml bisbenzimide (Hoechst 33258). Cells on glass cover slips were incubated with the fluorescence-dye for 30 min in the dark, washed with aqua dest., air dried and mounted with glycerol. Micronuclei were scored and cell images recorded with an FSX100 Box-Type Fluorescence Imaging Device (Olympus Corporation, Tokyo, Japan). Before scoring the micronuclei, all slides were randomised and coded. The number of micronuclei was determined by counting 500 cells/slide. The criteria for scoring micronuclei were adapted from references [34,35]; each treatment was done in triplicate. Values are expressed as percentages of the number of micronuclei in untreated cells.

2.9. Estimation of chromosome breaks

Logarithmically growing VA13 and AT22 cells were plated in 100 mm tissue culture plates. When the cells reached 50% confluence, they were treated with 30 μg/ml lipoproteins for 8 h. To arrest the cells in metaphase, colcemid (100 ng/ml) was added for 4 h [36]. The cells were washed with PBS and trypsinized. The reaction was stopped with DMEM and cells were pelleted 5 min at 500 × g. Then, the cells were resuspended in 0.075 mM KCl and incubated for 15 min at 37 °C. Two hundred microliter of Carnoy's fixative (methanol:glacial acetic acid, 3:1 [v/v]) was added; cells were gently mixed and pelleted (5 min, 500 × g). The supernatant was removed, cells were twice gently mixed with 5 ml of Carnoy's fixative and pelleted (5 min at 500 × g) again. Cell lysates were dropped on glass slides and dried for 30 min at 90 °C. Chromosomes were stained with Giemsa. For scoring chromosome breaks, 5000 individual chromosomes/treatment were observed under oil immersion microscopy. Each treatment was done in triplicate.

2.10. Measurement of ROS

The intracellular generation of ROS was measured using carboxy-H₂DCFDA. H₂DCFDA is deacetylated by esterases to nonfluorescent dichlorofluorescein, which is converted to fluorescent dichlorofluorescein (DCF) by ROS. VA13 and AT22 cell were cultured in 6 well plates in DMEM containing 5% FCS. Fifty % confluent cells were serum-starved overnight and incubated with indicated concentrations of lipoproteins for 5, 12 or 24 h. When indicated, cells were pre-treated with PDTC for 30 min (1 mM; dissolved in aqua dest.). For inhibition of ATM, cells were preincubated with the ATM-I for 1 h before addition of lipoproteins. DMSO concentration did not exceed 0.01% (v/v). After indicated times, the medium was aspirated and 10 μM carboxy-H₂DCFDA, dissolved in PBS (10 mM), was added to the cells [37]. Cells were incubated for another 30 min at 37 °C. To terminate the reaction, cells were kept on ice and washed with ice-cold PBS. Cell lysis was performed with 3% (v/v) Triton X-100 in PBS on a rotary shaker (1350 rpm; Heidolph Instruments, Schwabach, Germany) at 4 °C (in the dark) for 30 min. To ensure complete solubilisation of DCF, 50 μl absolute ethanol was added and the plates were shaken for a further 15 min [37]. The cell lysates were transferred to microfuge tubes and cellular debris was removed by centrifugation (13,000 rpm, 4 °C, 10 min). One hundred microliter of the supernatant was transferred into 96-well microtiter plates and fluorescence was measured on a Victor Multilabel Counter (Perkin-Elmer, Waltham, MA, USA) with excitation at 485 nm and emission at 540 nm. All steps concerning carboxy-H₂DCFDA were performed under light-protected conditions. The protein content of the cell lysates was determined using an aliquot of the supernatant and the BCA™ Protein Assay Kit according to the manufacturer's instructions.

2.11. Image analysis for generation of intracellular ROS

VA13 and AT22 cells were seeded in 6-well plates, grown to 50% confluence, and incubated with serum-free DMEM overnight. Where indicated, cells were pre-treated with 1 mM PDTC (dissolved in aqua dest.) for 30 min. Cells were incubated with 100 μg/ml lipoprotein for 5 or 12 h. Carboxy-H₂DCFDA (10 μM, dissolved in 10 mM PBS) was added to the cells and plates were incubated for further 30 min at 37 °C. To terminate the reaction, dishes were put on ice and cells were washed with PBS. For observation of the cells under a fluorescence microscope, 100 μl PBS was added to each well. The cells were observed and photographed using an inverted microscope (Nikon Eclipse TE 2000-U; Nikon Instruments, Melville, NY, USA) with a green fluorescent filter and the NIS Elements BR 2.10 software for image acquisition. To allow comparison between images, all images were acquired at the same exposure time (600 ms).

2.12. Statistical analysis

Data are presented as means ± standard deviation (SD). Two-way ANOVA or t-test statistical analyses were performed using Prism 5 software (Graphpad Software, La Jolla, CA, USA). In ANOVA analysis, Bonferroni posttest was used for all pair wise comparisons of the means of all experimental groups. Values (p < 0.05) were considered significant.

3. Results

3.1. ATM is activated by oxLDL

Previous reports performed with various cell lines revealed that dependent on the stimulus, activation of ATM occurs between 15 and 480 min [38,39]. We here show that VA13 cells (ATM^+/+) exhibited either no (data not shown) or sometimes basal pATM expression (Fig. 1A). OxLDL increased pATM levels in a time-dependent manner reaching a maximum after 90 min (Fig. 1A). The immunoreactive pATM signal decreased to baseline levels after 300 min. H₂O₂ a known activator of ATM [15,40], resulted in efficient phosphorylation of ATM in VA13 cells but not in AT22 cells (termed ATM^−/−). Densitometric evaluation of immunoreactive pATM bands revealed that H₂O₂-mediated induction is approximately 25% higher after 90 min compared with oxLDL-mediated induction. Although two different polyclonal antibodies were used to follow total ATM expression, immunoreactive β-tubulin was found to be more precise and reliable as loading control. Fig. 1B demonstrates that LDL sometimes tended to phosphorylate ATM in VA13 cells, however, only to levels in between 5 and 10% compared to oxLDL. Fig. 1B further shows that oxLDL-induced phosphorylation of ATM was completely abrogated by ATM-I (Fig. 1B).

Fig. 1.

Time-dependent phosphorylation of ATM in normal fibroblasts by oxLDL and inhibition of ATM-activation: (A) VA13 cells were incubated with 100 μg/ml oxLDL at indicated times. (B) VA13 cells were preincubated with 10 μM of the ATM-kinase-inhibitor (ATM-I) for 1 h followed by treatment with 100 μg/ml lipoprotein for 90 min. As positive controls, cells were treated with 200 μM H₂O₂ for 15 min; after medium exchange the cells were incubated for another 90 min (105*). Cell lysates were collected; proteins were subjected to SDS-PAGE and transferred to nitrocellulose membranes. Western Blot experiments were performed using anti-pATM as primary antibody. β-Tubulin was used as loading control. One representative experiment out of three is shown.

3.2. ATM expression moderates the toxic effect of oxLDL

Cells that fail to repair damaged DNA before entering mitosis may exhibit chromosomal strand breaks, leading to disruption in subsequent cell cycles resulting in a defective colony formation [41]. As ATM plays an important role in the recognition and signalling of DNA damage [42], we studied whether the lack of ATM affects the clonogenic survival of cells. Fig. 2A shows that oxLDL, but not LDL, caused a dose-dependent inhibition of colony formation in VA13 and AT22 cells. However, at protein concentrations higher than 3-μg oxLDL/ml, colony formation in AT22 cells was significantly (p < 0.001) reduced compared to VA13 cells.

Fig. 2.

OxLDL alters colony-forming efficiency in normal and ATM-deficient fibroblasts: (A) Exponentially growing VA13 and AT22 cells were plated at a density of 1.5 × 10³ cells/100 mm dish in DMEM with increasing lipoprotein concentrations. After 18 h, the medium was replaced and cells were cultured for 12 days. (B) 1.5 × 10³ cells of VA13 and AT22 cells, plated on a 100 mm dish, were pre-incubated with 10 μM ATM-I for 1 h before addition of 30 μg/ml lipoprotein. After 18 h the medium was replaced by normal growth medium and cells were allowed to grow for 12 days. Cell colonies were fixed, stained with crystal violet and colonies consisting of groups of 50 or more cells were counted. Data are given as number of colonies (as percentage compared to untreated cells, set as 100%). Values represent the means ± SD (n = 5). UT (untreated cells). **p < 0.01 and ***p < 0.001 (significant difference between VA13 and AT22 cells).

To support our observation, that the presence of ATM affects the clonogenic survival, ATM-activation in VA13 cells was inhibited before oxLDL treatment. Fig. 2B shows that ATM-I reduced colony formation in VA13 cells to levels found in AT22 cells when treated with oxLDL. Again, LDL did not alter colony formation when compared to untreated control cells.

3.3. ATM and cell viability in the presence of oxLDL

Next, mitochondrial function and cell viability of normal and ATM-deficient cells were investigated using two different assay systems. The MTT-test forms blue formazan crystals that are reduced by mitochondrial dehydrogenase in living cells [43]. OxLDL decreased cell viability in VA13 and AT22 cells in a time- (Fig. 3A) and concentration-dependent manner (Fig. 3B). AT22 cells are more sensitive to oxLDL exposure than VA13 cells (Fig. 3). LDL had no adverse effect on the viability of either cell type. Next, cell survival was measured using the Trypan blue exclusion assay. Incubation of VA13 and AT22 cells with oxLDL up to 24 h decreased the number of living cells in a time-dependent manner up to 30% (Fig. 4A). Again, oxLDL was more toxic to AT22 cells at all times, compared to VA13 cells. LDL (used as control) had no effect on cell the survival of both cell lines. To visualize nuclear changes after treatment with lipoproteins, VA13 and AT22 cells were stained with Hoechst 33258 and fluorescence intensity was checked. Control and LDL-treated cells exhibited diffuse chromatin staining (Fig. 4B). However, exposure of VA13 cells to oxLDL led to morphological changes, such as areas of condensed chromatin and shrunken nuclei. In contrast, AT22 cells treated with oxLDL exhibited a decrease in size and number of nuclei, but no chromatin condensation.

Fig. 3.

Time- and concentration-dependent effect of lipoproteins on cell viability of in normal and ATM-deficient fibroblasts using the MTT assay: (A) VA13 and AT22 cells were treated with 100 μg/ml lipoprotein for indicated times. (B) VA13 and AT22 were exposed to 3, 30 and 100 μg/ml lipoprotein for 18 h. MTT reduction was assessed at 570 nm on a microplate reader and values were corrected for background absorbance at 630 nm. Values are expressed as percentages of untreated cells. The data are expressed as a percentage of untreated cells and represent the means ± SD (n = 6). *p < 0.05 and ***p < 0.001 (significant difference between VA13 and AT22 cells).

Fig. 4.

Time-dependent effect of lipoproteins on cell viability of normal and ATM-deficient fibroblasts using the Trypan blue exclusion assay: (A) AT22 and VA13 cells were cultured in the presence of 100 μg/ml lipoprotein at indicated times. Cells were stained with Trypan blue and viable cells were counted. Each time point represents the mean value ± SD of three experiments performed in triplicates. (B) Fluorescence images of Hoechst 33258-stained VA13 and AT22 cells after treatment with 100 μg/ml lipoprotein for 16 h. Untreated cells were used as controls. Aggregated and fragmented chromatin appears as bright areas. One representative experiment out of four is shown.

3.4. OxLDL induces DNA double strand breaks in A-T cells

ATM primarily responds to DSBs [7]. Since phosphorylation of the histone H2AX (termed γ-H2AX) is a sensitive cellular indicator for the presence of DNA-DSBs [44,45], the effect of lipoproteins on H2AX phosphorylation via ATM was studied. Fig. 5A shows that exposure of VA13 and AT22 cells to oxLDL led to formation of immunoreactive γ-H2AX only in AT22 but not in VA13 cells. Also, time-dependent incubation of both cell lines with oxLDL, but not LDL, confirmed the presence of immunoreactive γ-H2AX after 16 h only in AT22 cells (Fig. 5B). Since the MTT assay demonstrated that oxLDL is toxic to VA13 and AT22 cells (Fig. 4A), PARP cleavage and activation of procaspase 3 (two hallmarks of apoptosis) were investigated. After 16 h of oxLDL exposure neither PARP cleavage nor procaspase 3 processing was observed in either cell type (Fig. 5A). Following time-dependent incubation of cells with lipoproteins up to 24 h, neither LDL nor oxLDL promoted PARP cleavage or activation of caspase 3 (Fig. 5B).

Fig. 5.

OxLDL induces phosphorylation of H2AX in ATM-deficient fibroblasts: AT22 and VA13 cells were treated (A) with 100 μg/ml oxLDL for 16 h or (B) with 100 μg/ml lipoprotein at indicated times. VA13 cells were treated with 1 μM staurosporine (St) for 4 h to induce phosphorylation of H2AX and apoptosis. Cells were lysed and protein lysates were subjected to Western Blot analysis following SDS-PAGE. Polyclonal anti-γ-H2AX was used as primary antibody. Membranes were then stripped and incubated with anti-PARP and anti-caspase-3 antibody as primary antibodies. β-Actin served as a loading control.

To assess whether the immunoreactive γ-H2AX signal correlates with micronucleus formation following oxLDL exposure, and to investigate a possible clastogenic effect of oxLDL, the in vitro micronucleus technique was employed. Micronuclei arise during cell division and contain chromosome breaks lacking centromeres and/or whole chromosomes, and are unable to travel to the spindle poles during mitosis [34]. Our findings show that oxLDL-treated AT22 cells displayed a significantly higher micronuclei number compared to similarly treated VA13 cells (Fig. 6A). Treatment of both cell lines with LDL did not alter the micronuclei number when compared to untreated controls. Since micronuclei formation is a sign of chromosomal damage, the number of chromosomal breaks was further counted in VA13 and AT22 cells in the absence or presence of lipoproteins. Cells lacking ATM exhibited a slightly higher number (although statistically not significant) of chromosomal breaks in untreated cells compared to VA13 (Fig. 6B). However, oxLDL significantly (p < 0.001) increased chromosomal breaks in both cell lines. In VA13 cells, the number of chromosomal breaks after 8 h increased up to 30. In AT22 cells the number of chromosomal breaks increased up to 42. Fig. 6B further shows that the number of oxLDL-induced chromosomal breaks in AT22 cells are significantly (p < 0.001) higher when compared to VA13 cells. Treatment of VA13 and AT22 cells with LDL was without effects on chromosomal breaks when compared to untreated cells (Fig. 6B).

Fig. 6.

OxLDL promotes formation of micronuclei and chromosomal breaks in AT22 fibroblasts: (A) VA13 and AT22 cells were incubated with 100 μg/ml lipoprotein for 16 h. Cells were fixed with methanol and stained with Hoechst 33285. Numbers of micronuclei/500 cells were counted using a fluorescence microscope. The arrow indicates a typical micronucleus formation in AT22 cells. Data are expressed as percentage of untreated cells and represent the means ± SD (n = 6). (B) VA13 and AT22 cells were incubated with 30 μg/ml lipoprotein for 8 h. Metaphase preparations were made and chromosomal breaks were counted. Each data point represents 5000 individual chromosome observations performed under oil immersion microscopy done in triplicate. UT (untreated cells). *p < 0.05 and ***p < 0.001 (significant difference between VA13 and AT22 cells).

3.5. PDTC scavenges oxLDL-induced elevated ROS levels in A-T cells

ATM-deficient cells are in a constant state of oxidative stress and may exhibit diminished antioxidant capacity [11,46]. We show that AT22 cells exhibited approx. 1.5-fold higher ROS levels when compared to VA13 cells (Fig. 7A). Incubation of cells with oxLDL further increased ROS levels in VA13 and AT22 in a time-dependent manner. ROS formation induced by oxLDL was significantly (p < 0.001) higher in AT22 cells at 5 and 12 h compared to VA13 cells. After 24 h, ROS levels were also higher in AT22 cells, although not statistically significant. LDL did not affect ROS levels in VA13 or AT22 cells (Fig. 7B). Treatment of cells with increasing concentrations of oxLDL for 5 h led to a dose-dependent increase of ROS, which is significantly higher (up to 3-fold compared with untreated cells) in AT22 cells compared to VA13 cells (Fig. 7C). Findings obtained with the DCFDA/DCF assay (Fig. 7), i.e. incubation of cells with lipoproteins and subsequent ROS measurements, were confirmed using fluorescence microscopy (Fig. 7D). AT22 cells exposed to oxLDL exhibited higher fluorescence intensity when compared to untreated or LDL-treated cells. In line with data shown in Fig. 7B, a slight increase in fluorescence-intensity could also be observed in oxLDL-treated VA13 cells when compared to untreated or LDL-treated cells.

Fig. 7.

Time- and concentration-dependent intracellular ROS levels in normal and ATM-deficient fibroblasts: (A) To measure basal intracellular ROS levels, AT223 and VA13 cells were loaded for 30 min with 10 μM carboxy-H₂DCFDA in the absence of agonists. The DCF fluorescence counts detected with VA13 cells were set 100%. (B) Cells were incubated with 100 μg/ml lipoprotein for indicated times and ROS levels were measured as described above. The DCF fluorescence counts in untreated cells were set 100%. (C) Cells were incubated with indicated concentrations of oxLDL for 5 h before ROS measurements. The DCF fluorescence counts in untreated cells were set 100%. (D) Cells were incubated with 100 μg/ml lipoprotein for 12 h and loaded with 10 μM carboxy-H₂DCFDA for 30 min. DCF fluorescence was observed under fluorescence microscope. (E) Cells were pre-treated with 10 μM ATM-I for 1 h prior to incubation with 100 μg/ml lipoprotein for 5 h. Then ROS levels were measured. The DCF fluorescence counts in untreated cells were set 100%. (A, B, C, and E) DCF fluorescence values were normalized to cell protein content. Data are expressed as mean ± SD (n = 6). UT (untreated cells). **p < 0.01 and ***p < 0.001 (significant difference between VA13 and AT22 cells).

To confirm, that ATM regulates ROS formation, cells were pre-treated with ATM-I before incubation with oxLDL. DCF fluorescence measurements revealed that inhibition of ATM led to significantly (p < 0.01) higher levels of basal ROS in VA13 cells but also when cells were treated with oxLDL (Fig. 7E). No significant difference in ROS levels were found in oxLDL-treated AT22 cells in the absence or presence of ATM-I indicating that the compound per se did not alter ROS formation.

To scavenge ROS, cells were pre-incubated with PDTC, a potent antioxidant and suppressor of transcription factor nuclear factor κB [47], prior to incubation with oxLDL. PDTC effectively reduced oxLDL induced ROS-formation in AT22 and VA13 cells to basal levels (Fig. 8A). Also fluorescence microscopy technique showed less fluorescence intensity in oxLDL-treated cells after preincubation with PDTC for 1 h (Fig. 8B).

Fig. 8.

PDTC protects against ROS-generation induced by oxLDL in ATM-deficient fibroblasts: (A) AT22 and VA13 cells were incubated with PDTC for 1 h and treated with 100 μg/ml oxLDL for 5 h. Levels of endogenous ROS were calculated by measuring DCF fluorescence. Results are expressed as fluorescence intensity as % of control and represent the means ± SD (n = 6) ***p < 0.001. (B) Fluorescence microscopy of AT22 and VA13 cells stained with carboxy-H₂DCFDA. Cells were pre-treated with 1 mM PDTC for 1 h and incubated with 100 μg/ml oxLDL for 5 h.

3.6. OxLDL induces p21 via an ATM-dependent pathway

Activation of the ATM-kinase may promote induction of p53 [10]; stabilized p53 serves as a transcription factor and stimulates expression of the cyclin-dependent kinase inhibitor p21 [48]. Fig. 9 shows oxLDL-mediated induction of p21 in VA13 (but not in AT22) cells. Inhibition of the ATM-kinase activity in VA13 cells reduced oxLDL-induced expression of immunoreactive p21 to baseline levels.

Fig. 9.

ATM-kinase-dependent activation of p21 in VA13 fibroblasts by oxLDL: AT22 and VA13 cells were incubated with 100 μg/ml oxLDL at indicated times in the absence or presence of 10 μM ATM-I (preincubated for 1 h). Cell lysates were subjected to SDS-PAGE, proteins were transferred to nitrocellulose and membranes were incubated with anti-p21 Waf1/Cip1 as the primary antibody. β-Actin was used as loading control. One representative experiment out of two is shown.

In the series of experiments, we have tested whether oxLDL-mediated expression of pATM and subsequent induction of p21 is also operative in cells other than fibroblast. These data indicate that induction of pATM by oxLDL in endothelial cells occurs in a time-dependent manner (Fig. 10A) similar as found in VA13 fibroblasts (Fig. 1A); densitometric evaluation of immunoreactive pATM bands revealed a 1.7-fold induction after 90 min (Fig. 10A). Furthermore, pre-incubation of endothelial cells with ATM-I did not only inhibit phosphorylation of the ATM-kinase (Fig. 10B) but also down regulated time-dependent expression of p21 (Fig. 10C) as well as colony formation of oxLDL-treated cells (Fig. 10D).

Fig. 10.

ATM-kinase activation and induction of p21 in endothelial cells by oxLDL: EA.hy926 cells were incubated with 100 μg/ml oxLDL (A, B, C) or LDL (B) at indicated times or for 90 min in the absence or presence of 10 μM (C) or 10 and 30 μM (B) ATM-I (preincubated for 1 h). As positive controls, cells were treated with 200 μM H₂O₂ for 15 min; after medium exchange the cells were incubated for another 90 min (105*) (A). Cell lysates were subjected to SDS-PAGE, proteins were transferred to nitrocellulose and membranes were incubated with anti-pATM (A, B) or anti-p21 (C) as primary antibody. β-Tubulin (A, B) or β-actin (C) was used as loading control. One representative experiment out of two is shown. (D) Colony forming efficiency of EA.hy926 cells treated with 30 μg/ml lipoprotein in the absence or presence of 10 μM ATM-I was performed exactly as described in Fig. 2B. Data are given as number of colonies (as percentage compared to untreated cells, set as 100%). Values represent the means ± SD (n = 3). UT (untreated cells). **p < 0.01 and ***p < 0.001.

4. Discussion

A-T, an autosomal recessive disorder resulting from ATM gene mutation, is characterized by a high incidence of lymphoid malignancies, neurodegeneration, immunodeficiency, premature aging, elevated radiosensitivity, and genomic instability. Genomic instability is characterized by chromosome breaks, chromosome gaps, translocations, and aneuploidy [42,49]. Recent findings suggested that DNA damage links mitochondrial dysfunction to the metabolic syndrome and atherosclerosis, indicating that prevention of mitochondrial dysfunction may represent a novel target of cardiovascular disease [23]. Basically, mitochondrial dysfunction is linked to ATM-heterozygosity [50,51] and results in an imbalance of ROS [23]. As ROS levels are tightly coupled with inflammatory diseases e.g. atherosclerosis, increased ROS levels in ATM^−/− and ATM^+/− cells may be due to alterations in cellular defence mechanisms and possibly due to cellular dysfunction induced by modified/oxidized (lipo)proteins. Among different lipoprotein modifications, the oxidation of LDL by transition metals such as copper ions represents a suitable experimental approach to mimic oxidative modifications of LDL in vivo [19,25,52,53]. OxLDL has been reported to participate in the development of atherosclerosis primarily by promoting vascular cell growth [54–57]. OxLDL is a potent proinflammatory chemoattractant for macrophages and T-lymphocytes. OxLDL is also cytotoxic for endothelial cells and stimulates them to release soluble inflammatory molecules. In addition, oxLDL has turned out to be highly immunogenic and promotes changes in cell cycle protein expression, and subsequent translocation and activation of transcription factors [58,59]. These events help to perpetuate a cycle of vascular inflammation and lipid/(lipo)protein dysregulation within the artery wall and also may create a cellular pro-thrombotic state that complicates later stages of atherosclerosis [59].

In the present study, we demonstrated that oxLDL, known to generate oxidative stress in the vascular system [19,25], induced phosphorylation of ATM and downstream activation of p21 in fibroblasts and endothelial cells. The immunoreactive pATM signal induced by oxLDL was almost comparable to levels induced by H₂O₂. ATM-deficient cells are extremely sensitive to the toxic effects of H₂O₂, nitric oxide radical, and t-butyl hydroperoxide, respectively [38,60]. To obtain information on sensitivity of ATM-null fibroblasts to oxLDL, several different cytotoxicity assays were employed (Figs. 2–4). All three assays demonstrated that in comparison to wild-type cells, ATM-deficient fibroblasts are more sensitive to oxLDL treatment – indicating that ATM-expression lessens oxLDL-mediated toxicity. However, fibroblasts lacking ATM were more sensitive to oxLDL-treatment in the colony-forming assay, than was observed in the short-term culture assays (MTT- and Trypan blue exclusion assay). This is probably due to defective cell cycle response in A-T cells, as these cells may replicate their DNA despite having unrepaired DNA breaks. Both, the MTT and the Trypan blue exclusion assay, and the appearance of condensed chromatin, demonstrated that oxLDL exhibited mild toxic effects on VA13 cells, with PARP cleavage and caspase 3 activation not being detected. We assume that due to the mild toxic effects of oxLDL in normal fibroblasts, ATM induction triggers an activation of cell cycle checkpoints and not apoptotic cascade activation.

OxLDL-mediated toxicity was significantly higher in ATM-deficient fibroblasts. We assume that these cells (defective in G₁, S and G₂ checkpoint functions [16,61]) are unable to respond adequately to oxLDL-induced oxidative stress and/or DNA damage. The result is oxLDL hypersensitivity and eventual cell death. To confirm this hypothesis the effect of oxLDL on DNA damage was investigated. A very early step in the response to DNA-DSBs is the appearance of immunoreactive γ-H2AX [44]. γ-H2AX is an essential component for the recruitment and accumulation of DNA repair proteins to sites of DSB damage, including 53BP1, BRCA1, RAD51 and MDC1 and the MRE11/RAD50/NBS1 complex [62]. In the presence of DNA-DSBs, H2AX is rapidly phosporylated by ATM [63]. However, H2AX can also be phosphorylated by other members of the phosphatidylinositol 3-kinase family, including DNA-dependent protein kinase and the ATM-and Rad3-related protein kinase [64,65]. We found that following oxLDL exposure immunoreactive γ-H2AX was present only in ATM-deficient AT22, but not in VA13 cells. As oxLDL leads to ATM-phosphorylation in VA13 cells, this data indicates that ATM is activated by oxLDL in the absence of DNA-DSBs. ATM is a key player in DSBs responses, being activated by these breaks and phosphorylating key down-stream proteins, leading to cell cycle checkpoint arrest and/or apoptosis [66]. However, lack of ATM causes not only a defective response to DNA-DSBs, but also a defect in regulating cellular responses to oxidative stress [7,17]. Our findings are consistent with a recent study [17], demonstrating that ATM activation induced by H₂O₂ occurs in the absence of DNA damage. The observation that oxLDL-dependent H2AX phosphorylation was only observed in ATM^−/− cells suggested that another member of the phosphatidylinositol 3-kinase family is likely to be involved in this pathway. Furthermore, the appearance of γ-H2AX in ATM-deficient cells makes it reasonable to assume that ATM protects against oxLDL-induction of DNA-DSBs. Increased formation of micronuclei and a higher number of chromosomal breaks in oxLDL-treated AT22 cells (as compared to VA13 cells) gives further support to this hypothesis.

Accumulating evidence suggests that oxidative stress is involved in the pathogenesis of A-T. Loss of ATM leads to increased oxidative damage to proteins and lipids and many cell-types, such as bone marrow stem cells and thymocytes of mice, exhibit elevated levels of ROS [7,67–69]. In line with these observations, we detected increased basal levels of ROS in ATM-deficient fibroblasts. Treatment with oxLDL further amplified ROS formation in ATM-deficient and normal fibroblasts. Also, oxLDL-induced ROS formation was significantly higher in ATM-deficient AT22 cells and in response to pharmacological inhibition of ATM in VA13 cells. This indicates that ATM protects from oxLDL-induced intracellular ROS production and that ATM expression may play a critical role in cell function and survival in atherosclerosis. Most importantly, cellular and molecular responses of fibroblasts from atherosclerosis patients towards ionizing radiation, activating the ATM-stress response, are similar to those observed from cells obtained from A-T patients [70]. The oxLDL-induced elevation of ROS, but no signs of DNA damage, in normal fibroblasts, confirmed the hypothesis, that not DNA-DSBs but ROS triggers oxLDL-induced activation of ATM. These observations parallel recent data [71] where ROS potently and rapidly activates ATM in the cytoplasm suggesting that mechanisms other than DNA-DSBs in the nucleus are operative to promote activation of ATM.

Administration of antioxidants to Atm^−/− mice exhibited a variety of beneficial effects, including extended lifespan, reduced tumorigenesis and improvement of motor deficits [13,14,72]. Pre-treatment of ATM-deficient cells with N-acetyl-l-cysteine attenuated ROS formation and blocked activation of ATM [15,17,67]. Due to redox cycling, N-acetyl-l-cysteine is able to reduce Cu²⁺- to Cu⁺-ions that can promote metal-catalyzed lipid peroxidation in vitro [73]. However, we here used PDTC to scavenge oxLDL-induced formation of ROS. PDTC induces glutathione synthesis in endothelial cells and suppresses the activation of transcription factor nuclear factor κB [73,74]. Most importantly, PDTC possesses metal-chelating properties and therefore, generation of free Cu²⁺-ions, recently reported to activate ATM in murine neuroblastoma cells and human HeLa cells [75], can be excluded under our experimental conditions. Here we show that PDTC is able to diminish formation of intracellular ROS induced by oxLDL in both, normal and ATM-deficient cells almost to basal levels.

In conclusion, we demonstrated that ATM is involved in oxLDL-mediated signalling. OxLDL-mediated activation of ATM occurs via intracellular formation of ROS and not via induction of DNA-DSBs. We propose that under conditions of ATM deficiency, oxLDL-dependent ROS production induces DNA damage chromosomal instability and cell death. As a consequence, H2AX, required for the repair mechanisms of ROS-induced DNA damage in ATM-deficient cells [76], is phosphorylated. Furthermore, we demonstrated that PDTC acts as a potent antioxidant against oxLDL-induced ROS formation. Our data enforce the role of ATM as a sensor of oxidative stress that could be important for protection against oxLDL-mediated cellular toxicity. Therefore, the ability of oxLDL to activate the ATM pathway may represent a critical adaptive response to maintain cell viability at sites of vascular inflammation and atherosclerosis.

Conflict of interest

The authors state that there is no conflict of interest.