Introduction

ATP is an intracellular energy source and an important extracellular signaling molecule. ATP is readily released from various cell types at certain basal rates after cell activation and after cell damage [1, 2]. Burst of ATP functions as evolutionally conserved danger signal for neighboring cells [3]. It is possible that the acute elevation of extracellular ATP is meant to protect surrounding tissue from further damage. Extracellular ATP can promote cellular survival and stimulate proliferation and migration [4, 5]. Several external factors are able to drive cells to apoptosis by destabilizing the nuclear chromatin. It remains obscure whether extracellular ATP could protect cells, such as endothelial cells (EC), against DNA-damaging conditions. Our aim in this study was to establish whether elevated extracellular ATP concentration, through CD39 inhibition, would influence DNA damage sensitivity in ECs.

Double-strand DNA breaks (DSBs) are the most hazardous form of DNA damage as they cause chromosomal rearrangements. Accumulating DSBs may induce apoptosis or cellular dysfunction. DSBs occur at basal levels due to various environmental factors, and approximately 50 endogenous DSBs occur in every cell during cell cycle [6]. Moreover, several cancer therapies such as chemotherapy and gamma irradiation induce DSBs also in non-malignant cells.

ATP and its metabolites act through several cell surface P1 and P2 receptors. Extracellular ATP is readily hydrolyzed by cell membrane-bound and soluble enzymes [7]. The most prominent ATP-hydrolyzing ectoenzyme in endothelial cell (EC) surface is ectonucleoside triphosphate diphosphohydrolase 1 (ENTPD1/CD39) [8]. It is also highly expressed by other cell types [7]. The overall concentration of extracellular ATP is regulated through ATP release and its hydrolysis by ectoenzymes such as CD39.

The effect of ATP-mediated signaling on DNA damage prevention and repair remains less studied. In this study, we show a protecting role of CD39 attenuation against DNA damage in ECs in vitro and in rat pulmonary and cardiac tissues in vivo. We suggest that targeting the ATP-mediated pathway could represent an attractive strategy for tissue protection during radiotherapy or chemotherapy.

Materials and methods

Cell culture

Human pulmonary microvascular ECs (ScienCell, Carlsbad, CA, USA, cat. 3000) were cultured in EBM-2 supplemented with EGM-2 bullet kit (Lonza Clonetics, Walkerville, MD, USA, cat. CC-3162). Human chronic myelogenous leukemia cells (K562, Sigma-Aldrich, Munich, Germany, cat. 89121407) and human diffuse large B cell lymphoma cells (SUDHL-4, ATCC, Teddington, UK, cat. CRL-2957) were cultured in RPMI-1640 (Sigma, cat. R0883) supplemented with 10 % FBS, 2 mM l-glutamine, and 1 % penicillin/streptomycin.

DNA damage was induced with γ-irradiation (4–5 Gy), methyl methanesulfonate (MMS, 500 μM, Sigma, cat. 129925), and doxorubicin hydrochloride (DOX, 1 μM, Tocris, Bristol, UK, cat. 2252). Overnight pretreatments with 10 μM ATP-γ-S (Tocris, cat. 4080) or 100 μM sodium polyoxotungstate (POM-1) (Tocris, cat. 2689) were used.

RNA interference

Small interfering RNA (siRNA) was used to silence CD39 expression (Dharmacon, cat. L-015973-00-0005). Non-target siRNA was used as a control (Dharmacon, cat. D-001810-01-05). The ECs were siRNA treated according to the previously described method [9]. Previously, it has been demonstrated that the CD39-siRNA is effective in these ECs with qRT-PCR and immunofluoresence stainings [10]. Here, the effectiveness of the CD39 siRNA is shown in protein level (Fig. S1).

Western immunoblotting

Whole cell or tissue lysates were prepared as previously described [9]. Primary antibodies used are the following: anti-phospho-histone H2A.X (Ser139) 1:1500 (Merck Life Science, Millipore, Espoo, Finland, cat. MABE205), p53 (ser15) 1:1000 (Cell Signaling, Leiden, The Netherlands, cat. 9284), and β-actin 1:2000 (Santa Cruz Biotechnology, Heidelberg, Germany, cat. sc-1615). Secondary antibodies used are the following: goat anti-rabbit IgG-HRP (Santa Cruz Biotechnology, cat. sc-2004) and donkey anti-goat IgG-HRP (Santa Cruz Biotechnology, cat. sc-2020).

Immunocytochemistry

After 4-h recovery from γ-irradiation, the cells were fixed with 4 % paraformaldehyde and permeabilized with 0.1 % Triton-X-100. Primary antibody was γH2AX (Ser139) 1:100 (Millipore, cat. MABE205). Secondary antibody was Alexa Fluor 488 1:100 (Life technologies, cat. A-21206). The cells were stained with DAPI prior to mounting. The slides were evaluated and photographed under fluorescent microscope, and γH2AX foci in single cells were calculated with ImageJ software (40 cells per condition) as previously described [11].

Immunohistochemistry

The collected rat heart and lung tissues were fixed with 10 % formalin for 24 h and then transferred to 70 % ethanol. After fixation, the tissues were paraffin embedded and sectioned to microscopy slides. The primary antibody was γH2AX (Ser139) 1:1000 (Millipore), and the secondary antibody was in Rabbit-on-Rodent HRP-Polymer kit (Biocare Medical, Concord, CA, USA, Cat. RMR622), which was used according to the manufacturer’s instructions. Tissue samples were analyzed under a microscope, and quantification was done with ImageJ software with IHC Toolbox plugin [12] to determine the ratio between positive and total nuclei.

Caspase assay

K562 and SUDHL-4 cells were seeded in 96-well plate, 104 cells per well, in supplemented growth media and let to recover overnight. Next, the cells were treated with 100 μM POM-1 or vehicle, let to recover overnight, and then treated with 1 μM DOX or vehicle for 24 h before caspase 3/7 activity measurement (Caspase Glo 3/7 assay, Promega, Nacka, Sweden, cat. G8091).

Animals

Male Sprague-Dawley (SD) rats (170–200 g, N = 3) were used in experiments, which were done with the permission of the National Animal Experiment Board. The rats were given POM-1 (Tocris) (10 mg/kg, intraperitoneal (i.p.)) or PBS in three consecutive days. At the third day, the rats were further treated with monocrotaline (MCT) (60 mg/kg, s.c., Sigma, Cat. 2401), DOX (Tocris) (6 mg/kg, i.p.), or PBS. In MCT group, lungs were collected 24 h after injection, and in the DOX group, hearts were collected 8 h after injection. In all groups, the rats were euthanatized with CO2.

Statistics

Statistical analysis was done with Prism GraphPad 6 (La Jolla, CA, USA). Unpaired t test was used for comparing groups. Results are expressed in mean ± SEM from at least three independent experiments. P < 0.05 was considered as significant, and in the figures, p values are expressed with stars: *P < 0.05, **P < 0.01, and ***P < 0.001.

Results

Silencing of CD39 protects ECs from DNA damage

Suppression of CD39 expression in ECs significantly decreased γH2AX expression in both control and γ-irradiated cells when compared to non-target (NT) siRNA-treated cells (Fig. 1a). Compared to control, the γH2AX protein expression was 56 % lower after 45-min recovery (p = 0.0090) and 69 % lower after 4-h recovery (p = 0.0009) in CD39-deficient cells (Fig. 1a). Supporting results were obtained from immunocytochemistry experiments where the number of γH2AX foci in individual cells (Fig. 1b) was quantified in γ-irradiated cells (p = 0.0008). Similarly to irradiation, the expression level of γH2AX in CD39-siRNA silenced cells was 60 % lower after 4-h MMS treatment (p = 0.0143), compared to control siRNA-treated cells (Fig. 1c). Results from CD39-deficient cells treated with DOX for 4 h supported the hypothesis (p = 0.0524) (Fig. 1c). Similarly, the CD39-deficient ECs treated with MMS had significantly decreased (59 %) expression of phosphorylated p53 protein, a marker of activated DNA damage pathway, when compared to control (p = 0.0023, Fig. 1d), while the total p53 protein expression remained unaltered (Fig. S3).

Fig. 1
figure 1

CD39 siRNA silenced ECs have decreased γH2AX protein expression after irradiation or chemically induced DNA damage. a γH2AX protein expression from CD39 siRNA silenced (CD39) and non-target siRNA (NT)-treated ECs at basal level and after irradiation followed by 45-min or 4-h recovery time. The bars represent the optical density (OD) of the γH2AX/β-actin ratio from three independent western immunoblot (WB) experiments. b Immunocytochemistry from γH2AX protein expression after irradiation from CD39 siRNA silenced (CD39) and non-target siRNA-treated (NT) ECs. Histogram represents a quantification of average number of γH2AX foci/cell; 40 cells per condition were analyzed with ImageJ. c γH2AX protein expression and quantification after MMS- or DOX-induced DNA damage from CD39 siRNA silenced (CD39) and non-target siRNA (NT)-treated ECs. d Serine 15 phosphorylated p53 protein expression and quantification after MMS-induced DNA damage from CD39 siRNA silenced (CD39) and non-target siRNA (NT)-treated ECs. β-Actin was used as a loading control in all experiments

Full size image

Both ATP and POM-1 protect ECs from DNA damage

ATP analogue, ATP-γ-S, pretreatment markedly decreased γH2AX expression after irradiation (p = 0.0592)- or MMS (p = 0.0461)-induced DNA damage (Fig. 2a) compared to control cells. Similarly, CD39 inhibitor POM-1 pretreatment significantly decreased γH2AX expression at basal level (p = 0.0136) and after MMS (p = 0.0036)-induced DNA damage (up to 80 %, Fig. 2b) compared to control cells. The expression level of serine 15-phosphorylated p53 was suppressed 47 % in POM-1-treated ECs compared to control after exposure to MMS-induced DNA damage (p = 0.0163, Fig. 2c), while the total p53 protein expression remained unaltered (Fig. S4).

Fig. 2
figure 2

ATP stimulation and suppression of CD39 activity decrease γH2AX protein expression after irradiation or chemically induced DNA damage. a γH2AX protein expression at basal rate and after irradiation or MMS-induced DNA damage with or without ATPγS pretreatment. Histograms represent quantifications from three independent WB experiments. The bars indicate the optical density (OD) of the γH2AX/β-actin ratio. b γH2AX protein expression at basal rate and after MMS-induced DNA damage with or without POM-1 treatment. Histograms represent quantifications from three independent WB experiments. The bars indicate the optical density (OD) of the γH2AX/β-actin ratio. c Serine 15 phosphorylated p53 protein expression at basal rate and after MMS-induced DNA damage with or without POM-1 treatment. Histogram represents quantifications from three independent WB experiments. The bars indicate the optical density (OD) of the p53/β-actin ratio. β-Actin was used as a loading control in all experiments

Full size image

ATPase activity is decreased in CD39 siRNA and POM-1-treated cells

Specific ATPase activity was significantly decreased in POM-1-treated and CD39-deficient ECs compared to control cells (Fig. S2). Moreover, POM-1 treatment had no additional inhibitory effects on CD39 siRNA silenced cell ATPase activity.

POM-1 treatment enhanced DNA damage resistance in vivo

To test whether the observed effects of CD39 inhibition and ATP accumulation would apply also in vivo, we used SD rats. The γH2AX expression in lung lysates was significantly lower in MCT animals pretreated with POM-1 compared to control animals (48 %, p = 0.0109, Fig. 3a). Similarly, we discovered significantly lower γH2AX expression in heart lysates in DOX-treated rats that were pretreated with POM-1 (53 %, p = 0.0371, Fig. 3b). Semiquantitative immunohistochemistry analysis confirmed the result in cardiac cells. The γH2AX expression in cardiac tissue was significantly lower in POM-1-pretreated rats compared to animals, which received DOX without POM-1 pretreatment (p = 0.0013, Fig. 3c).

Fig. 3
figure 3

CD39 activity inhibition decreases γH2AX protein expression in vivo after chemically induced DNA damage. a γH2AX protein expression from lung tissue lysates with control (C) and after MCT-induced DNA damage with or without POM-1 treatment. Histogram represents quantifications from three independent WB experiments. The bars indicate the optical density (OD) of the γH2AX/β-actin ratio. b γH2AX protein expression from heart tissue lysates with control (C) and after DOX-induced DNA damage with or without POM-1 treatment. Histogram represents quantifications from three independent WB experiments. The bars indicate the optical density (OD) of the γH2AX/β-actin ratio. c Heart tissue γH2AX protein expression in IHC with control (C) and after DOX-induced DNA damage with or without POM-1 treatment. The foci and the density of foci represent the amount of DSB in single cell, and the histogram represents the ratio of γH2AX nuclei to all nuclei. Scale bar 200 μm. d γH2AX protein expression from two cancer cell line (K562 and SUDHL-4) lysates with control (C) and after DOX-induced DNA damage with or without POM-1 treatment. β-Actin was used as a loading control in all WB experiments. e Caspase 3/7 activity in two cancer cell lines pretreated with 100 μM POM-1 prior to 1-μM DOX treatment for 24 h

Full size image

POM-1 treatment does not rescue cancer cells against DOX treatment

Human leukemia- and lymphoma-derived K562 and SUDHL-4 cells, respectively, were treated with POM-1 and subsequently exposed to DOX for 24 h. Immunoblot analysis of γH2AX expression showed no difference between the groups (Fig. 3d). To test whether CD39 inhibition could protect these cancer cells against DOX-induced apoptosis, we evaluated caspase 3/7 activity after 24-h DOX treatment with or without POM-1 pretreatment. We did not observe any attenuation of caspase 3/7 activity in POM-1 pretreated cells (Fig. 3e). The POM-1 pretreatment significantly increased caspase 3/7 activity after DOX treatment in K562 cells and had no effect in SUDHL-4 cells.

Discussion

In this study, we show for the first time how suppression of CD39 and resulting elevated extracellular ATP niche [10] can promote resistance to DNA damage under various DNA-damaging conditions in vitro and in vivo. In addition, we demonstrate that inhibition of CD39 does not promote DNA damage repair or apoptosis resistance in transformed cancer cells.

Homologous recombination (HR) and non-homologous end joining (NHEJ) are the two main DNA repair mechanisms used in cells to repair DSBs. The cell cycle phase is the major determinant of the used mechanism. The more effective and less error-prone HR is active only in G2/S phase when the homologous DNA strand can be used as a template [13] while NHEJ operates in all cell cycle phases [14]. In this study, DNA damage was induced with four different DNA DSB-producing mechanisms. The high-energy γ-irradiation breaks the DNA strands directly and mainly involves the NHEJ repair pathway [15, 16]. It is not fully confirmed whether MMS directly induces DSBs, but it stalls DNA replication and HR is involved in the repair of stalled replication forks [17]. DOX induces DNA damage through increased oxidative stress and by intercalating the DNA strands [18]. MCT is metabolized to genotoxic MCT pyrrole and to (+/−)6,7-dihydro-7-hydroxy-1-hydroxymethyl-5H-pyrrolizine (DHP) in vivo, which leads to DNA crosslink and DHP-DNA adduct formation, which are repaired by Fanconi anemia pathway following HR [19, 20].

Cells are constantly exposed to DSBs that need to be effectively repaired to ensure the normal cell function [6]. DNA damages stall the DNA replication and cell cycle and inhibit cell proliferation. Moreover, accumulated DNA damages can drive cells to apoptosis. As extracellular ATP has been shown to induce DNA replication and cell proliferation in ECs and in smooth muscle cells [4, 21], it could be plausible that these effects are partly mediated through enhanced DNA damage repair. ATP-induced cell proliferation might have a connection to enhanced DNA damage repair as our results demonstrate that basal levels of γH2AX were lower in ATP-activated cells.

Suppression of CD39 not only increases the extracellular ATP but also leads to decreased extracellular adenosine levels in ECs [4]. Adenosine is widely considered to be anti-inflammatory and protective toward vasculature. On the other hand, sustained high adenosine concentration can also be harmful and pro-apoptotic to lung ECs [22]. In CD39-siRNA and POM-1-treated ECs, the decreased adenosine could contribute to DNA damage sensitivity. However, ATP-γ-S treatment, considered not affecting significantly adenosine levels, showed results consistent with CD39 suppressed cells. Other growth factor signaling pathways have been shown to enhance DSB repair after γ-irradiation. Epidermal growth factor receptor variant III (EGFRvIII) signaling has been shown to have a key role in the radioresistance in glioblastoma. EGFRvIII is known to activate downstream effectors such as phosphatidylinositol 3-kinase (PI3K), Akt-1, Ras, and mitogen-activated protein kinase (MAPK) [23–25]. This downstream signaling eventually leads to DNA-dependent protein kinase (DNA-PKcs) hyperactivation and enhanced DSB repair [25]. Interestingly, ATP has been shown to activate the same downstream signaling pathways and even function in synergy with EGF receptor signaling [26–28].

The few previous studies connecting purinergic signaling to DNA damage repair have been mainly done with cancer cell lines, such as lung cancer. Previous study demonstrated that ATP sensitizes cancer cells to irradiation-induced DNA damage through P2Y6 and P2Y12 receptor activation [29, 30]. In a wider study with six different cancer cell lines, ATP treatment protected against DOX-induced cytotoxicity only in non-metastatic CL1.0 lung cancer cells [31]. Few studies have shown that extracellular ATP promotes survival in non-small cell lung cancer A549 cell line [32, 33]. However, this has not been shown to be a cause of enhanced DNA damage resistance. The distinct purine receptor representation in cancer- and non-malignant cells could explain the difference in cellular response [34]. Other explanation could be difference in purine-inactivating cell surface enzymes between ECs and cancer cells. While CD39 is the main ATP-hydrolyzing enzyme in ECs, cancer cells have additional phosphatases, which could explain the differential response to CD39 inhibition with POM-1 [35]. Our results with cancer cells, where γH2AX expression in POM-1-pretreated cells was increased after DOX-induced DNA damage, fit well to these previous findings. In addition, CD39 inhibition with POM-1 was not able to rescue the cancer cells from DOX-induced cytotoxicity (Fig. 3e). As opposite to cancer cells, in circulating blood cells, others have described that ATP inhibits the radiation-induced DNA damage ex vivo [36].

The great improvements in early cancer detection and cancer treatment strategies have decreased the rate of cancer-related deaths over the last decades [37]. Unfortunately, at the same time, the risk of late-onset cardiovascular complications is increased due to chemotherapy and radiotherapy [38]. Currently, cardiovascular morbidity is the most common nonmalignant cause of death among the cancer survivors [39].

Our study, together with previous observations, indicates that there is a profound difference in ATP signaling between cancer cells and non-cancer cells. Considering the protective actions of transient elevation of extracellular ATP in non-cancer cells, we reason this as an attractive strategy for tissue protection during cancer treatments. Additional research is still needed to discover the full mechanism of ATP signaling-mediated resistance to DNA damage in quiescent differentiated cells. Future studies are also needed to better understand the differences in ATP responses between cancer and non-cancer cells and whether certain cancer cells are responsive to ATP similarly than differentiated cells. We propose that targeting and inhibiting CD39 activity could be an attractive strategy to suppress especially cardiovascular injury associated with cancer treatments. Larger in vivo experimental series are now needed to further evaluate the clinical utility of this observation.