Microvascular Research 95 (2014) 68–75

Contents lists available at ScienceDirect

Microvascular Research journal homepage: www.elsevier.com/locate/ymvre

Oxidative stress modulates nucleobase transport in microvascular endothelial cells Derek B.J. Bone a,1, Milica Antic a, Gonzalo Vilas b, James R. Hammond a,b,⁎ a b

Department of Physiology and Pharmacology, Western University, London, ON, Canada Department of Pharmacology, University of Alberta, Edmonton, AB, Canada

a r t i c l e

i n f o

Article history: Accepted 10 June 2014 Available online 27 June 2014 Keywords: Adenosine Hypoxanthine Solute transporters Endothelium Vascular Reactive oxygen species

a b s t r a c t Purine nucleosides and nucleobases play key roles in the physiological response to vascular ischemia/reperfusion events. The intra- and extracellular concentrations of these compounds are controlled, in part, by equilibrative nucleoside transporter subtype 1 (ENT1; SLC29A1) and by equilibrative nucleobase transporter subtype 1 (ENBT1). These transporters are expressed at the membranes of numerous cell types including microvascular endothelial cells. We studied the impact of reactive oxygen species on the function of ENT1 and ENBT1 in primary (CMVEC) and immortalized (HMEC-1) human microvascular endothelial cells. Both cell types displayed similar transporter expression profiles, with the majority (N 90%) of 2-chloro[3H]adenosine (nucleoside) uptake mediated by ENT1 and [3H]hypoxanthine (nucleobase) uptake mediated by ENBT1. An in vitro mineral oil-overlay model of ischemia/reperfusion had no effect on ENT1 function, but significantly reduced ENBT1 Vmax in both cell types. This decrease in transport function was mimicked by the intracellular superoxide generator menadione and could be reversed by the superoxide dismutase mimetic MnTMPyP. In contrast, neither the extracellular peroxide donor TBHP nor the extracellular peroxynitrite donor 3-morpholinosydnonimine (SIN-1) affected ENBT1-mediated [3H]hypoxanthine uptake. SIN-1 did, however, enhance ENT1-mediated 2-chloro[3H]adenosine uptake. Our data establish HMEC-1 as an appropriate model for study of purine transport in CMVEC. Additionally, these data suggest that the generation of intracellular superoxide in ischemia/reperfusion leads to the down-regulation of ENBT1 function. Modification of purine transport by oxidant stress may contribute to ischemia/reperfusion induced vascular damage and should be considered in the development of therapeutic strategies. © 2014 Elsevier Inc. All rights reserved.

Introduction Microvascular endothelial cells (MVEC) play a critical role in the local regulation of vascular tone via the production and release of vasoactive agents such as nitric oxide and prostacyclin (Tune, 2007).

Abbreviations: CMVEC, cardiac microvascular endothelial cells (primary); CNT, concentrative nucleoside transporter; EDTA, ethylenediaminetetraacetic acid; ENBT1, equilibrative nucleobase transporter subtype 1; ENT, equilibrative nucleoside transporter; ENT1, equilibrative nucleoside transporter subtype 1, SLC29A1; ENT2, equilibrative nucleoside transporter subtype 2, SLC29A2; FeTTPs, 5,10,15,20-tetrakis (4-sulfonatophenyl) porphyrinato iron III chloride; HMEC-1, human microvascular endothelial cells type 1 (dermal, immortalized); MnTMPyP, manganese(III) tetrakis(1-methyl-4-pyridyl)porphyrin; NBMPR, nitrobenzylmercaptopurine riboside; NBTGR, nitrobenzylthioguanine riboside; NMG, N-methyl-glucamine; PBS, phosphate buffered saline; SIN-1, 3-morpholinosydnonimine; TBHP, tert-butyl hydroperoxide. ⁎ Corresponding author at: Department of Pharmacology, 9-70 Medical Sciences Building, University of Alberta, Edmonton, AB T6G 2H7, Canada. E-mail addresses: [email protected] (D.B.J. Bone), [email protected] (G. Vilas), [email protected] (J.R. Hammond). 1 Present address of DBJB is: Molecular Signaling Section, Laboratory of Bioorganic Chemistry, National Institute of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, Bethesda, MD, USA.

http://dx.doi.org/10.1016/j.mvr.2014.06.008 0026-2862/© 2014 Elsevier Inc. All rights reserved.

A key factor that stimulates the release of these agents is the endogenous nucleoside adenosine (Nyberg et al., 2010). Adenosine, a ‘retaliatory metabolite’, is released from cells under biological stress (e.g. ischemia) and stimulates a family of G-protein coupled receptors present in vascular endothelial and smooth muscle cells (Headrick et al., 2011; Ryzhov et al., 2007; Volonte and D'Ambrosi, 2009). Activation of adenosine receptors has been shown to be cardioprotective. Adenosine inhibits platelet aggregation, has anti-inflammatory activities, and is a potent vasodilator. It has also been implicated in the phenomenon of preconditioning, where a short period of ischemia/reperfusion will attenuate the damage caused by a subsequent ischemia/reperfusion event (Headrick and Lasley, 2009; Yang et al., 2010). Adenosine concentrations are regulated in the vasculature by the concerted activities of a number of enzymes and membrane transporters (Arch and Newsholme, 1978; Deussen, 2000b; Deussen et al., 2006). Under baseline physiological conditions, adenosine produced via 5′-ecto-nucleotidase activity, or arising from ATP metabolism within surrounding tissue, is rapidly taken up into cells by specific plasma membrane transporters (Loffler et al., 2007). In MVEC, adenosine is accumulated and released predominantly via the SLC29A1

D.B.J. Bone et al. / Microvascular Research 95 (2014) 68–75

transporter (equilibrative nucleoside transporter subtype 1; henceforth referred to as ENT1). Once inside the cells, adenosine is immediately used to maintain the cellular adenine nucleotide pool via the adenosine kinase pathway or, once the kinase pathway is saturated, metabolized by adenosine deaminase to inosine. Hence, intracellular adenosine concentrations are typically very low, maintaining an inward adenosine gradient across the plasma membrane. Intracellular inosine arising from adenosine deaminase activity is converted to hypoxanthine, which is subsequently metabolized by xanthine oxidase to xanthine and uric acid. Xanthine oxidase mediated metabolism produces oxygen free radicals as catalytic by-products that, in turn, contribute to the reactive oxygen stress seen in ischemia/reperfusion injury (Baudry et al., 2008; Lakshmi et al., 2009; Lee et al., 2009). Xanthine oxidase-derived superoxide can also react with nitric oxide produced by endothelial cells to form peroxynitrite, which has also been implicated in the etiology of vascular dysfunction (Forstermann, 2010). The amount of inosine and hypoxanthine available to the xanthine oxidase pathway is dependent not only on the intracellular adenosine levels and adenosine deaminase activities, but also on the capacity of the cells to efflux inosine and hypoxanthine via ENT1 and ENBT1 (equilibrative nucleobase transporter subtype 1), respectively. It is also important to note that cells can replenish nucleotide pools via scavenging of extracellular adenine nucleobases released from surrounding tissue, thus limiting post-ischemic vascular injury; this cytoprotective process would also be sensitive to the activity of ENBT1. We have established previously that primary human (h) MVEC have extensive capacity for purine nucleoside and nucleobase transport via the aforementioned transporters (Bone and Hammond, 2007). The important roles of ENT1 and ENBT1 in regulating purine levels in the vasculature make them a target for pharmacological manipulation of vascular activity in ischemia/reperfusion pathologies. Recent work by Abd-Elfattah and colleagues showed the effectiveness of blockade of ENT1 in the protection of canine hearts during surgery via both the enhancement of adenosine receptor activity and the maintenance of intracellular adenosine for nucleotide pool replenishment (Abd-Elfattah et al., 2012a,b). Nucleoside transport inhibitors have also been studied as cardioprotectants for cardiac transplant procedures (Chang-Chun et al., 1992; Masuda et al., 1992). Furthermore, dipyridamole, a broad spectrum equilibrative nucleoside transport blocker, is used therapeutically to prevent postoperative thromboembolic complications after cardiac valve replacement and diagnostically to induce vasodilation for myocardial perfusion imaging (Bolognese et al., 1991; Schaper, 2005; Wackers, 1991) . Notwithstanding the therapeutic potential of ENT inhibitors, the development of cardiovascular therapies based on manipulation of adenosine concentrations has been hampered by a lack of information on the cellular regulation of these transporters in the microvasculature, as well as lack of understanding of how these transporters respond to changes in the cellular environment associated with ischemia/reperfusion. Because the expression of nucleoside transporter subtypes in endothelial cells, as well as their sensitivity to transport inhibitors, is species-related (Archer et al., 2004; Bone and Hammond, 2007; Bone et al., 2010; Hyde et al., 2001), it is imperative to assess the roles of these transporters in human cell/tissue models. However, the isolation and culture of the large amounts of human primary endothelial cells required to perform pharmacological and physiological studies have its disadvantages in terms of cost, biosafety, and limited proliferative capacity in culture. These circumstances suggest the need for an immortalized cell model that retains the endothelial phenotype and responds to regulatory factors in a manner comparable to primary MVEC. In that regard, the HMEC-1 cell line, derived from human dermal MVEC via transfection with the simian virus 40 A gene product (Ades et al., 1992; Bouis et al., 2001; Lidington et al., 1999), appears to have the greatest potential. The present study focused on the impact that reactive oxygen species, generated during ischemia/reperfusion, have on the function of ENT1 and ENBT1 in human cardiac MVEC (CMVEC; primary cultures)

69

and HMEC-1. Previous data suggest that ENT1 is down-regulated upon exposure to reactive oxygen species (Barankiewicz et al., 1995). However, no similar work has been done for ENBT1. We hypothesize that increased intracellular production of reactive oxygen species by MVEC during ischemic episodes leads to rapid changes in cellular ENT1 and ENBT1 activities which influence the biological activities of adenosine in the vasculature. To test this hypothesis we determined whether the immortalized MVEC cell line HMEC-1 can be used as a model for studying purine transporters in the microvascular endothelium and assessed the impact of simulated ischemia/reperfusion (sI/R) and oxygen free radical generation on nucleoside and nucleobase uptake by human MVEC. Materials and methods Materials 2-Chloro[8-3H]adenosine (16 Ci/mmol), [2,8-3 H]hypoxanthine (28.7 Ci/mmol), [3H]nitrobenzylmercaptopurine riboside ([3H]NBMPR, 23.8 Ci/mmol), and [3H]water (1 mCi/ml) were purchased from Moravek Biochemicals (Brea, CA). Non-radiolabeled 2-chloroadenosine, hypoxanthine, NBMPR, nitrobenzylthioguanine riboside (NBTGR), dipyridamole (2,6-bis(diethanolamino)-4,8-dipiperidinopyrimido-[5,4-d]pyrimidine), adenine, tert-butyl hydroperoxide (TBHP) and menadione were from Sigma-Aldrich. 3-Morpholinosydnonimine (SIN-1) was from Tocris Bioscience (Minneapolis, MN), and manganese(III) tetrakis(1methyl-4-pyridyl)porphyrin (MnTMPyP) and 5,10,15,20-tetrakis (4-sulfonatophenyl) porphyrinato iron III chloride (FeTTPs) were from Calbiochem (Mississauga, ON). Cell culture Primary human CMVEC were purchased from Lonza (Walkersville, MD) and were cultured in EGM-2 MV medium as supplied by the manufacturer. Cells were provided at passage 3 and used between passages 4 and 7. No morphological changes occurred during culture up to passage 7. Quality assurance from the manufacturer verified the isolation as pure endothelial cells using immunohistologic staining for the presence of acetylated LDL and von Willebrand's (factor VIII) antigen and the absence of smooth muscle α-actin. HMEC-1 were obtained from the Centers for Disease Control and Prevention (Atlanta, GA) and cultured in MCDB-131 medium supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin G, 100 μg/ml streptomycin sulfate, 10 ng/ml epidermal growth factor, and 1 μg/ml hydrocortisone. HMEC-1 were used between passages 15 and 25. It was noted in preliminary experiments that after passage 30 the amount of hypoxanthine transport by the HMEC-1 increased significantly, suggesting a change in cell characteristics at higher passage numbers. Cells were maintained in a humidified atmosphere at 37 °C with 5% CO2 in T175 culture flasks. For experimental assays, cells were trypsinized (0.05% trypsin–0.53 mM EDTA) and then diluted four-fold in 500 μg/ml trypsin inhibitor in phosphate buffered saline (PBS; in mM: 137 NaCl, 6.3 Na2HPO4, 2.7 KCl, 1.5 KH2PO4, 0.9 CaCl2–2H2O, 0.5 MgCl2–6H2 O; pH 7.4). Cells were collected by centrifugation, and pellets were washed in either PBS or a Na +-free buffer [NMG (in mM): 140 N-methyl- D -glucamine, 5 KCl, 4.2 KH2 PO 4 , 0.36 K 2HPO 4 , 10 HEPES, 1.3 CaCl2–2H2O, and 0.5 MgCl2 –6H2 O; pH 7.4] and re-suspended in the same buffer as required for immediate use. For low oxygen exposure, culture flasks were placed in hypoxia chambers (Billups-Rothenberg). Chambers were flushed with 1% O2 for 5 min and expelled air was measured with an O2 detector to confirm low oxygen levels. Chambers were then sealed and placed in a 37 °C incubator for 2 h. Following hypoxia, flasks were removed from the chambers and harvested for substrate uptake as described above.

70

D.B.J. Bone et al. / Microvascular Research 95 (2014) 68–75

Simulated ischemia/reperfusion (sI/R) This in vitro cellular model of ischemia/reperfusion was based on a method described by Meldrum et al. (2001). Briefly, cells were harvested and suspended in fresh complete medium, placed in a 37 °C 5% CO2–95% air humidified atmosphere for 1 h to equilibrate, and collected by centrifugation, and the medium was aspirated leaving approximately 200 μl to cover the cell pellet. In some cases, the superoxide dismutase mimetic MnTMPyP or the peroxynitrite scavenger FeTTPs was added to the media at a final concentration of 100 μM; these concentrations have been shown by others to be effective in reversing free radical effects in other cell models (Day et al., 1997; Gonzalez-Polo et al., 2004). The cell pellet was re-suspended by tapping the side of the tube gently. Simulated ischemia was initiated by layering mineral oil over the cell suspension, creating a gas exchange barrier. Mineral oil over-laid cell suspensions were placed in a 37 °C incubator for 2 h. Control cells were left in complete medium with no oiloverlay and incubated in parallel at 37 °C 5% CO2–95% air. Cells were mixed gently every 30 min, taking care not to disrupt the oil barrier. Following this treatment, the oil layer was removed and cells were suspended in fresh media and incubated for a further hour in 37 °C 5% CO2–95% air to simulate ‘reperfusion’. Cells were then pelleted by centrifugation and suspended in PBS or NMG for immediate use in substrate uptake assays. Treatment with reactive oxygen species generators TBHP was diluted to 100 μM with warmed (37 °C) PBS from a concentrated stock, and menadione (±100 μM MnTMPyP) and SIN-1 (500 μM) solutions were prepared from powder in warmed (37 °C) PBS just prior to use. Immediately after harvest, cells were suspended with one of the above solutions, or PBS alone (control), and incubated for 30 min in a 37 °C 5% CO2–95% air humidified atmosphere. After the incubation period, cells were washed twice with room temperature buffer and used immediately for the substrate uptake or radioligand binding assays as shown below. [3H]NBMPR binding Cells (~ 75,000 cells/assay) were suspended in PBS and incubated with [3H]NBMPR for 45–60 min at room temperature (~ 22 °C). Cells were collected on S&S Biopath (Riviera Beach, FL) glass fiber filters using a 24-port Brandel cell harvester and washed twice with Tris buffer (10 mM Tris, pH 7.2, 4 °C) then analyzed for 3H content using standard liquid scintillation counting techniques. Specific binding was defined as total binding minus cell-associated [3H]NBMPR in the presence of 10 μM NBTGR (nonspecific binding). [3H]Nucleoside/nucleobase uptake Uptake was initiated by the addition of suspended cells (~ 750,000 cells/assay) to 2-chloro[ 3 H]adenosine (nucleoside) or [3H]hypoxanthine (nucleobase) layered over 200 μl of silicon/mineral oil (21:4 vol:vol) in 1.5-ml microcentrifuge tubes. Mediated uptake of 2-chloro[3H]adenosine was defined as that inhibited by 5 μM NBMPR/ 5 μM dipyridamole, and mediated uptake of [3H]hypoxanthine was defined as that inhibited by 1 mM adenine. We showed previously (Bone and Hammond, 2007) that 1 mM adenine is sufficient to inhibit all ENBT1-mediated [3H]hypoxanthine influx. After a defined incubation time, uptake was terminated by centrifugation for 10 s at 12,000 g. Aqueous substrate and oil layers were removed by aspiration, and pelleted cells were digested in 1 M sodium hydroxide overnight (12–16 h). A sample of the digest was removed and analyzed for 3H content using standard liquid scintillation counting techniques. Uptake data are presented as picomoles per microliter of intracellular volume after correction for the amount of extracellular 3H in the cell pellet.

Total volume was determined by incubating cells with 3H2O for 3 min and processed as above. Extracellular water space was estimated by extrapolation of the linear time course of non-mediated uptake to zero time. Data analysis and statistics All experiments were conducted a minimum of five times in duplicate. Data are expressed as mean ± SEM based on values derived from individual experimental curve fits (GraphPad Prism v5.0; Michaelis–Menten nonlinear analysis for determination of transporter kinetics). For simple two condition comparisons, statistical significance was assessed using the two-tailed Student t-test (paired or unpaired, as appropriate) with P b 0.05 considered significant. Multiple comparisons were done using two-way ANOVA followed by the Bonferroni post-test for individual comparisons within the specified data set (P b 0.05 considered significant). All statistical analyses were conducted with the aid of software routines built into GraphPad Prism v5.0. Results Nucleoside/nucleobase transport characteristics of immortalized HMEC-1 HMEC-1 bound the ENT1-selective probe [3H]NBMPR to a maximum of 87,000 sites/cell with a Kd of 0.10 nM (Fig. 1A, Table 1). These cells also accumulated 2-chloro[3 H]adenosine in a saturable, transporter-mediated fashion with a Km of 51 ± 4 μM and a Vmax of 8.4 ± 0.7 pmol μl−1 s−1. Cellular uptake of 2-chloro[3H]adenosine in the presence of supramaximal inhibitory concentrations (5 μM) of the ENT blockers dipyridamole and NBMPR led to a linear uptake:concentration profile consistent with a passive non-mediated uptake mechanism (Fig. 1B). There was no significant difference in this latter component when assays were conducted in the absence or presence of sodium (data not shown), indicating a lack of concentrative nucleoside transport activity. Incubation of cells with 100 nM NBMPR to selectively block ENT1 reduced the transporter-mediated uptake by over 90% to 0.7 pmol μl−1 s−1 with a Km of 48 μM (Fig. 1B); this likely represents ENT2 (equilibrative nucleoside transporter subtype 2; SLC29A2)-mediated uptake. Subtraction of the ENT2-mediated component from the total transporter mediated uptake led to the calculation of an ENT1-mediated uptake rate for 2-chloro[3H]adenosine of 7.5 pmol μl− 1 s− 1 (Table 1). HMEC-1 accumulated [3H]hypoxanthine via an adenine-sensitive nucleobase transport mechanism with a Km of 180 μM at a maximal rate of 7.6 pmol μl− 1 s− 1 (Fig. 1C, Table 1). Since both ENBT1 and ENT2 are known to function as hypoxanthine transporters, and ENT2 activity was identified in the 2-chloro[3H]adenosine flux studies described above, experiments were conducted in the presence and absence of 10 μM dipyridamole to block ENT2. In agreement with data previously reported for CMVEC (Bone and Hammond, 2007), [3H]hypoxanthine uptake by HMEC-1 in the presence of dipyridamole was not significantly different that observed in the absence of dipyridamole (data not shown). This suggests that ENT2 was not contributing significantly to the total hypoxanthine accumulation by HMEC-1. These characteristics of nucleoside and nucleobase transport in HMEC-1 compared well with those observed for primary CMVEC (Fig. 1, Table 1). Similar to that reported by us in a previous study (Bone and Hammond, 2007), CMVEC bound [3H]NBMPR with high affinity to a maximum of 66,000 sites/cell (Fig. 1D, Table 1), and accumulated 2-chloro[3H]adenosine predominantly via the ENT1; about 5% of the total mediated uptake was determined to be attributable to ENT2 (Fig. 1E, Table 1). CMVEC also displayed a significant amount of ENBT1-mediated (adenine-sensitive) [3H]hypoxanthine transport (Fig. 1F) with a Km (169 μM) that was not significantly different from

D.B.J. Bone et al. / Microvascular Research 95 (2014) 68–75

HMEC-1 80000

80000

Bound (sites/cell)

Bound (sites/cell)

A 100000 60000

Total 40000

Specific Nonspecific

20000 0 0.0

0.1

0.2

CMVEC

D

60000 Total Specific Nonspecific

40000 20000 0 0.0

0.3

0.1

8 6 4 Total +NBMPR Non-mediated

0 0

50

100

6 4 Total +NBMPR Non-mediated

2 0

150

0

Initial Rate of Uptake -1 -1 (pmol µl s )

Initial Rate of Uptake (pmol µ l-1 s-1)

100

150

F

15 Total Mediated Non-mediated

10

50

[3H]2-Chloroadenosine (µM)

C

15

0.4

8

[3H]2-Chloroadenosine (µM)

20

0.3

E

10

Initial Rate of Uptake (pmol µl-1 s-1)

Initial Rate of Uptake (pmol µl-1 s-1)

B

2

0.2 [3H]NBMPR (nM)

[3H]NBMPR (nM)

10

71

5 0

Total Mediated

10

Non-mediated

5

0 0

100

200

300

[3H]Hypoxanthine ( µM)

0

100

200

300

400

[3H]Hypoxanthine ( µ M)

Fig. 1. Nucleoside and nucleobase uptake by HMEC-1 (panels A–C) and CMVEC (panels D–F). [3H]NBMPR binding to ENT1 (panels A & D): Cells were incubated with increasing concentrations of [3H]NBMPR in the presence (nonspecific) or absence (total) of 10 μM NBTGR for 45 min. Cells were collected on glass fiber filters, and cell-associated 3H content was determined. Specific binding was determined as total bound minus nonspecific bound (n = 5). 2-Chloro[3H]adenosine uptake by ENT1/2 (panels B & E): Cells were incubated with increasing concentrations of 2-chloro[3H]adenosine for 5 s in the absence or presence of 100 nM NBMPR or 5 μM NBMPR/5 μM dipyridamole (n = 5). [3H]Hypoxanthine uptake by ENBT1 (panels C & F): Cells were incubated with increasing concentrations of [3H]hypoxanthine for 10 s in the absence (total) or presence of 1 mM adenine (non-mediated). Transportermediated uptake was calculated as the difference between the total and non-mediated uptake components (n = 14 HMEC-1, n = 5 CMVEC). Kinetic parameters derived from these experiments are shown in Table 1.

that seen previously (Bone and Hammond, 2007) and that was similar to the Km determined for hypoxanthine influx by HMEC-1 (Table 1). Effects of sI/R on nucleoside and nucleobase transport by microvascular endothelial cells To mimic the effects of ischemia/reperfusion at the cellular level, cell suspensions were overlaid with mineral oil for 2 h at 37 °C and then re-suspended in oil-free media for 1 h prior to measurement of transport. This sI/R treatment led to a dramatic decrease in the rate of ENBT1-mediated [3H]hypoxanthine uptake in both HMEC-1 and CMVEC (Fig. 2). Further analysis showed that this effect was due to a decrease in the Vmax of transport (Fig. 3) with no change in substrate Km. In contrast, sI/R had no impact on ENT1-mediated 2-chloro[3H] adenosine uptake in either cell type (Figs. 2 & 3). sI/R as conducted in the present study, like tissue ischemia, may be expected to: a) decrease oxygen tension, b) decrease pH, and c) result in the local accumulation of toxic metabolites. Incubation in a 1% O2 environment for 2 h had no effect on the capacity of the CMVEC to

accumulate [3H]hypoxanthine (data not shown). Also, no significant change was observed in the pH of the cell suspension after 2 h oil-overlay, as determined using pH indicator strips (probably due to the buffering capacity of the incubation media). These data suggest that the change in [3H]hypoxanthine uptake observed under sI/R conditions was due to an increase in metabolic by-product concentrations, specifically reactive oxygen species. To test this hypothesis, HMEC-1 and CMVEC were exposed to sI/R in the absence or presence of the superoxide dismutase mimetic MnTMPyP and their ability to transport [3H]hypoxanthine was determined. As shown in Fig. 3, addition of MnTMPyP significantly protected the cells against the effects of the sI/R. Indeed, [3H]hypoxanthine uptake by cells subjected to sI/R in the presence of MnTMPyP (Vmax = 8.6 ± 1.9 and 4.3 ± 0.5 pmol μl− 1 s− 1 for HMEC-1 and CMVEC, respectively) was not significantly different from that determined in cells that were incubated for 2 h at 37 °C in normal 5% CO2–95% air conditions (Vmax = 7.6 ± 1.1 and 5.1 ± 0.7 pmol μl−1 s−1, respectively). These data indicate that CMVEC and HMEC-1 are comparable in terms of their nucleobase and nucleoside transport mechanisms and

72

D.B.J. Bone et al. / Microvascular Research 95 (2014) 68–75

Table 1 Nucleoside and nucleobase transport by HMEC-1 and CMVEC. Kinetic parameters for 2-chloro[3H]adenosine and [3H]hypoxanthine uptake and [3H]NBMPR binding were derived from data as shown in Fig. 1. Total ENT-mediated uptake was calculated as the difference between uptake of 2-chloro[3H]adenosine in the presence or absence of 5 μM NBMPR/5 μM dipyridamole. ENT2-mediated uptake (ENT2) was defined as the difference between uptake of 2-chloro[3H]adenosine in the presence of 100 nM NBMPR and that seen after treatment of the cells with 5 μM NBMPR/5 μM dipyridamole. ENT1-mediated uptake (ENT1) was calculated as the difference between the total- and ENT2-mediated 2-chloro[3H]adenosine uptake components. ENBT1-mediated uptake was calculated as the difference between the total uptake of [3H]hypoxanthine and that observed in the presence of 1 mM adenine. Kinetic parameters are presented as mean ± SEM derived from hyperbolic curves fitted to data generated in each individual experiment (n ≥ 5; see Fig. 1). HMEC-1

CMVEC

Vmax (pmol μl−1 s−1)

Km (μM)

Vmax (pmol μl−1 s−1)

Km (μM)

7.5 ± 0.6 0.7 ± 0.1 7.6 ± 1.1

44 ± 6 48 ± 4 180 ± 35

3.7 ± 0.5 0.2 ± 0.1 5.1 ± 0.7

51 ± 15 58 ± 32 169 ± 37

Bmax (ENT1/cell)

Kd (nM)

Bmax (ENT1/cell)

Kd (nM)

87,000 ± 6000

0.10 ± 0.03

66,000 ± 4000

0.05 ± 0.01

2-Chloro[3H]adenosine—ENT1 2-Chloro[3H]adenosine—ENT2 [3H]Hypoxanthine—ENBT1

[3H]NBMPR binding

in their response to sI/R. On the basis of these findings, subsequent experiments were conducted using the more economical and robust HMEC-1 model. To further investigate the role of superoxide radicals in the modulation of [3H]hypoxanthine uptake, HMEC-1 were treated with the intracellular superoxide generator menadione. This treatment led to a significant (67 ± 8%) decrease in Vmax for [3H]hypoxanthine uptake, with no change in Km (Figs. 4A & C). Inclusion of MnTMPyP in the menadione treatment protocol led to a partial reversal of this effect (Figs. 4A & C). Menadione treatment also decreased the rate of 2-chloro[3H]adenosine uptake by the HMEC-1 (Fig. 4B) with a 60% decline in the Vmax for ENT-1 mediated uptake (Fig. 4C). However, in contrast to what was observed with [3H]hypoxanthine influx, MnTMPyP did not reverse the effect of menadione on 2-chloro[3H] adenosine uptake (Figs. 4B & C). Incubation with the extracellular hydrogen peroxide generator TBHP, at a concentration (100 μM) shown to generate significant levels of hydrogen peroxide in other models (Dhanya et al., 2014;

We have previously shown that primary CMVEC accumulate purine nucleosides and nucleobases via the ENT1 and ENBT1, respectively (Bone and Hammond, 2007). These transporters are bi-directional and thereby also mediate the cellular release of nucleosides and nucleobases down their concentration gradients

5 4 3 2

Control sI/R

1 0

2.0 1.5 1.0 0.5

Control sI/R

0.0 0

100

200

300

400

0

50

[3H]Hypoxanthine ( µM)

Initial Rate of Uptake -1 -1 (pmol µl s )

5 4 3 2

Control

1

sI/R

100

150

200

250

[3H]Hypoxanthine (µM)

B

6 Initial Rate of Uptake -1 -1 (pmol µl s )

CMVEC

C Initial Rate of Uptake -1 -1 (pmol µl s )

Initial Rate of Uptake -1 -1 (pmol µl s )

Discussion

HMEC-1

A

6

Zhang et al., 2013), had no effect on either [3 H]hypoxanthine or 2-chloro[3H]adenosine uptake (Fig. 5). The peroxynitrite generator SIN-1 (500 μM), on the other hand, led to an enhancement of 2-chloro[3H]adenosine uptake, but had no effect on [3H]hypoxanthine uptake (Fig. 5). In addition, the peroxynitrite decomposition catalyst FeTTPs (100 μM) did not reverse the effect of sI/R on [3H]hypoxanthine influx in HMEC-1 (data not shown). These data suggest that ENT1 and ENBT1 respond differently to reactive oxygen/nitrogen radicals, and ENBT1 is uniquely sensitive to intracellular superoxide.

D

2.0 1.5 1.0

Control

0.5

sI/R

0.0

0 0

20 40 60 [3H]2-Chloroadenosine (µM)

80

0

20

40

60

80

[3H]2-Chloroadenosine (µM)

Fig. 2. Effect of sI/R on [3H]hypoxanthine and 2-chloro[3H]adenosine uptake by microvascular endothelial cells. HMEC-1 (panels A & B) or CMVEC (panels C & D) were subjected to sI/R by incubation at 37 °C under a mineral oil barrier for 2 h followed by 1 h incubation in oil-free media at 37 °C in a 5% CO2/95% air environment (sI/R). Control cells were incubated in parallel for the same time period in normal buffer at 37 °C in a 5% CO2/95% air atmosphere. Cells (control and sI/R) were then incubated with increasing concentrations of [3H]hypoxanthine for 15 s in the presence (non-mediated uptake) or absence (total uptake) of 1 mM adenine (panels A & D), or 2-chloro[3H]adenosine for 10 s in the presence (non-mediated) or absence (total) of 5 μM NBMPR/5 μM dipyridamole. Results are presented as the initial rate of transporter-mediated uptake (pmol μl−1 s−1) calculated as the difference between the total and non-mediated uptake components (n = 6).

Control

**

sI/R + MnTMPyP

**

* *

0

Menadione Menadione + MnTMPyP

6

MnTMPyP

4 2 0

HMEC-1

CMVEC

ENBT1

HMEC-1

CMVEC

0

ENT1

Fig. 3. Effect of sI/R and MnTMPyP on the maximum rate of ENBT1- and ENT1mediated transport. HMEC-1 or CMVEC were subjected to sI/R and assessed for the rate of [3 H]hypoxanthine and 2-chloro[3H]adenosine uptake as described in Fig. 2. The superoxide dismutase mimetic MnTMPyP (100 μM) was included in parallel experiment sets to assess its ability to reverse the effect of sI/R on [3H]hypoxanthine influx. Each bar represents the mean ± SEM of the Vmax values for each condition, determined by extrapolation of Michaelis–Menten curves fitted to the transportermediated data from individual experiments. *Vmax for ENBT1 obtained from the sI/R treatment group is significantly different from control Vmax; **inclusion of MnTMPyP significantly attenuates the effect of sI/R; HMEC-1 and CMVEC are not significantly different in their response to the treatments (two-way ANOVA with Bonferroni post-test, P b 0.05, n = 6).

100

200

300

400

[3H]Hypoxanthine ( µM)

ENT1

B 10 Initial Rate of Uptake -1 -1 (pmol µ l s )

5

ENBT1 Control

8 Initial Rate of Uptake -1 -1 (pmol µ l s )

10

73

A

sI/R

-1

-1

Vmax (pmol µ l s )

D.B.J. Bone et al. / Microvascular Research 95 (2014) 68–75

Control Menadione

8

Menadione + MnTMPyP MnTMPyP

6 4 2 0 0

50

100

150

[3H]2-Chloroadenosine (µM)

C Control

10

MnTMPyP Menadione

-1

-1

Vmax (pmol µ l s )

(Bone and Hammond, 2007; Robillard et al., 2008). These previous works also showed that there was no evidence for the operation of ENT2, nor was there evidence for any sodium-coupled concentrative nucleoside transporter (CNT) activity. In the present study, the parameters obtained for 2-chloro[ 3H] adenosine and [ 3 H]hypoxanthine uptake by immortalized dermal HMEC-1 compared well with the ENT1 and ENBT1 characteristics of primary CMVEC determined in parallel (Fig. 1, Table 1). The data obtained in this set of experiments using CMVEC are also similar to those obtained in a previous, more extensive, study from our laboratory on this cell model (Bone and Hammond, 2007). The HMEC-1 also responded to sI/R in a manner similar to that in the CMVEC (although the primary cultures were more sensitive in this regard than were the immortalized HMEC-1). Previous published work has established that HMEC-1 are comparable to primary CMVEC with respect to morphology, phenotype and function (Ades et al., 1992; Bouis et al., 2001; Lidington et al., 1999). Therefore, HMEC-1 appear to be an economical and physiologically relevant model for the study of purine transport in human microvascular endothelium. We employed a mineral oil overlay approach to establish conditions that mimic vascular ischemia/reperfusion in microvascular endothelial cells. Meldrum and colleagues used this approach successfully to mimic the effects of in vivo ischemia/reperfusion injury in cultured LLC-PK1 renal tubular cells (Meldrum et al., 2001). In the present study, this treatment did not affect intracellular water volume of the cells suggesting that there was no significant cell lysis or change in general membrane permeability. Likewise, the function of the ENT1 was not affected, indicating that the loss of ENBT1 function seen with sI/R did not reflect a nonspecific loss of cellular transport activity. Like true vascular ischemia, the cell suspension under the mineral oil layer would be expected to become relatively hypoxic and there would be an increase in the local concentration of cellular metabolites relative to the control conditions. Indeed, a similar procedure has been shown to result in nutrient deprivation and restriction in metabolite washout in isolated guinea pig ventricular muscle (Vanheel et al., 1989) and cultured rat ventricular myocytes (Henry et al., 1996). We established that a 2 h incubation of primary CMVEC in a 1% oxygen environment

Menadione + MnTMPyP

5

** *

*

ENBT1

ENT1

*

0

Fig. 4. Effect of menadione on nucleoside and nucleobase uptake by HMEC-1. Cells were incubated in parallel in the absence (control) and presence of 100 μM menadione or with 100 μM MnTMPyP in the presence and absence of 100 μM menadione for 30 min at 37 °C. Cells were then washed once with room temperature buffer and immediately used for the analysis of the rate of uptake of [3H]hypoxanthine (15 s incubation, ENBT1) in the presence and absence of 1 mM adenine (panel A, n = 6), or 2-chloro[3H]adenosine (10 s incubation, ENT1) in the presence and absence of 5 μM NBMPR/5 μM dipyridamole (panel B; n = 5). Data are shown as the mean ± SEM of the initial rate (pmol μl−1 s−1) of inhibitor-sensitive (adenine or NBMPR/dipyridamole) substrate uptake. Panel C shows the Vmax values (mean ± SEM) derived by extrapolation of Michaelis–Menten curves fitted to the transporter-mediated data calculated for each independent experiment. *Significantly different from control; **significantly different from menadione treatment alone (two-way ANOVA with Bonferroni post-test, ENBT1 and ENT1 data sets analyzed independently, P b 0.05, n = 6).

did not affect either ENT1 or ENBT1 activity, suggesting that the loss of ENBT1 with sI/R was likely due to elevated levels of cellular metabolites. One possibility that needs to be considered is that sI/R resulted in an increase in endogenous intracellular hypoxanthine levels which would attenuate the observed rate of [3H]hypoxanthine influx by ENBT1 due to a reduction in the inwardly-directed transmembrane hypoxanthine concentration gradient. However, one would not expect this to be

74

D.B.J. Bone et al. / Microvascular Research 95 (2014) 68–75

*

Control 15

-1

-1

Vmax (pmol µ l s )

20 TBHP SIN1 10 5 0 ENBT1

ENT1

Fig. 5. Effect of TBHP and SIN-1 on nucleoside (ENT1) and nucleobase (ENBT1) uptake by HMEC-1. Cells were incubated at 37 °C in the absence (control) or presence of 100 μM TBHP or 500 μM SIN-1 for 30 min. Following incubation, cells were washed with room temperature buffer and then assessed for their capacity to accumulate 2-chloro[3H]adenosine (10 s incubation, ENT1) in the presence and absence of 5 μM NBMPR/5 μM dipyridamole, or [ 3 H]hypoxanthine (15 s incubation, ENBT1) in the presence and absence of 1 mM adenine. Data are shown as Vmax values (pmol μl−1 s−1, mean ± SEM, n = 5) calculated from plots of the initial rate of ENBT1- or ENT1-mediated substrate uptake versus substrate concentration. *Significantly different from control (two-way ANOVA with Bonferroni post-test, ENBT1 and ENT1 data sets analyzed independently, P b 0.05, n = 5).

reversed by the superoxide dismutase mimetic MnTMPyP or mimicked by the intracellular superoxide generator menadione, as was seen in this study. Thus it is reasonable to conclude that ENBT1-mediated transport of hypoxanthine by CMVEC was down-regulated under conditions that simulate vascular ischemia/reperfusion due to the generation of intracellular superoxide radicals. ENT1-mediated 2-chloro[3H]adenosine transport was also decreased with menadione treatment. However, this could not be reversed by MnTMPyP, and sI/R had no effect of ENT1 activity. Therefore, the effect of menadione on ENT1, unlike ENBT1, may be unrelated to superoxide generation. It is noteworthy that, in endothelial cells, menadione has been reported to deplete glutathione (GSH) and inhibit key intracellular metabolic enzymes such as glucose-6-phosphate dehydrogenase and glyceraldehyde-3-phosphate dehydrogenase independent of oxygen free radical production (McAmis et al., 2003). It is well established that vascular ischemia followed by reperfusion results in the formation of reactive oxygen species, which contribute to the endothelial dysfunction associated with this pathology (Forstermann, 2010; Munzel et al., 2010). Some of these reactive oxygen species arise from the activity of xanthine oxidase (Baudry et al., 2008; Lee et al., 2009). Under ischemic conditions, adenosine is generated from ATP metabolism in smooth muscle cells as well as myocytes and released into the interstitial space and vasculature where it is rapidly taken up by the vascular endothelial cells (Deussen, 2000a; Deussen et al., 1999). Under baseline physiological conditions, the adenosine accumulated by endothelial cells is metabolized to AMP, re-entering the cellular nucleotide pool. However, under the higher adenosine concentrations associated with ischemia, adenosine kinase is saturated and the excess adenosine is metabolized by adenosine deaminase to inosine. Inosine is then metabolized by purine nucleoside phosphorylase to hypoxanthine which is a substrate for xanthine oxidase (Arch and Newsholme, 1978). Therefore, it appears that down-regulation of ENBT1 during ischemia/ reperfusion has two potential deleterious consequences for endothelial cell function: 1) a reduced capacity of these cells to release hypoxanthine via ENBT1 that would lead to elevated intracellular hypoxanthine concentrations and thus enhanced generation of reactive oxygen species via xanthine oxidase, and 2) a decreased ability of cells to scavenge nucleobases from the extracellular milieu via ENBT1 that would attenuate the rate at which cells can replenish intracellular nucleotide pools via HGPRT.

Peroxynitrite is another toxic metabolite that is produced in the microvasculature via the interaction of nitric oxide and superoxide during ischemia/reperfusion (Muller-Delp et al., 2012; Munzel et al., 2010). Increased superoxide production due to hypoxanthine metabolism might be expected to lead to significant peroxynitrite production in this model. To test whether the ENT1 and/or ENBT1 activity of MVEC is sensitive to peroxynitrite, HMEC-1 were incubated with 500 μM SIN-1, a generator of peroxynitrite, for 30 min. These conditions have been shown to decrease choline uptake by the sodium-dependent choline transporter in human neuronal SH-SY5Y cells (Cuddy et al., 2012). ENBT1 activity was not affected by SIN-1, and the peroxynitrite decomposition catalyst FeTTPs did not reverse the deleterious effect of sI/R on ENBT1 activity. However, ENT1-mediated uptake of 2-chloroadenosine was significantly elevated by SIN-1 treatment. This suggests that peroxynitrite formation in the vasculature during ischemic stress may enhance the cellular uptake and release of adenosine via ENT1, but does not affect ENBT1-mediated nucleobase flux. In summary, our results support the use of the HMEC-1 cell line as a model for the study of purine transport and metabolism in MVEC. We have also shown that ENBT1-mediated uptake of hypoxanthine by MVEC is decreased upon exposure of the cells to conditions that simulate vascular ischemia/reperfusion and that this effect may be mediated by the intracellular production of superoxide. Peroxynitrite exposure, on the other hand, enhanced ENT1 activity but had no effect on ENBT1. These ischemia/reperfusion-associated changes in purine handling by MVEC are important factors to consider in the development of therapeutic protocols involving manipulation of purinergic metabolism aimed at reducing vascular damage in ischemia/reperfusion injury. Acknowledgments This work was supported by a grant (T-7275) to JRH from the Heart and Stroke Foundation of Canada. DBJB would like to thank the Schulich School of Medicine and Dentistry at Western University for financial support during the course of this work. DBJB was also the recipient of a PhD Studentship Award from the Canadian Institutes for Health Research. We wish to thank Dr Tim Regnault (Physiology and Pharmacology, Western University, London, Canada) for the use of his hypoxic cell culture chambers. We also wish to acknowledge the technical support of Diana Quinonez. JRH and GV also acknowledge the Faculty of Medicine and Dentistry, University of Alberta for support during the latter stages of this study. References Abd-Elfattah, A.S., Aly, H., Hanan, S., Wechsler, A.S., 2012a. Myocardial protection in beating heart cardiac surgery: I: pre- or postconditioning with inhibition of es-ENT1 nucleoside transporter and adenosine deaminase attenuates post-MI reperfusion-mediated ventricular fibrillation and regional contractile dysfunction. J. Thorac. Cardiovasc. Surg. 144, 250–255. Abd-Elfattah, A.S., Ding, M., Jessen, M.E., Wechsler, A.S., 2012b. On-pump inhibition of es-ENT1 nucleoside transporter and adenosine deaminase during aortic crossclamping entraps intracellular adenosine and protects against reperfusion injury: role of adenosine A1 receptor. J. Thorac. Cardiovasc. Surg. 144, 243–249. Ades, E.W., Candal, F.J., Swerlick, R.A., George, V.G., Summers, S., Bosse, D.C., Lawley, T.J., 1992. HMEC-1: establishment of an immortalized human microvascular endothelial cell line. J. Investig. Dermatol. 99, 683–690. Arch, J.R.S., Newsholme, E.A., 1978. The control of the metabolism and the hormonal role of adenosine. Essays Biochem. 14, 82–123. Archer, R.G., Pitelka, V., Hammond, J.R., 2004. Nucleoside transporter subtype expression and function in rat skeletal muscle microvascular endothelial cells. Br. J. Pharmacol. 143, 202–214. Barankiewicz, J., Uyesaka, J., Kossenjans, W., Rymaszewski, Z., 1995. Inhibition of nucleoside transport by reactive oxygen species in bovine heart microvascular endothelial cells. Adv. Exp. Med. Biol. 370 (PT 2), 775–778. Baudry, N., Laemmel, E., Vicaut, E., 2008. In vivo reactive oxygen species production induced by ischemia in muscle arterioles of mice: involvement of xanthine oxidase and mitochondria. Am. J. Physiol. Heart Circ. Physiol. 294, H821–H828. Bolognese, L., Sarasso, G., Bongo, A.S., Rossi, L., Aralda, D., Piccinino, C., Rossi, P., 1991. Dipyridamole echocardiography test: a new tool for detecting jeopardized myocardium after thrombolytic therapy. Circulation 84, 1100–1106.

D.B.J. Bone et al. / Microvascular Research 95 (2014) 68–75 Bone, D.B., Hammond, J.R., 2007. Nucleoside and nucleobase transporters of primary human cardiac microvascular endothelial cells: characterization of a novel nucleobase transporter. Am. J. Physiol. Heart Circ. Physiol. 293, H3325–H3332. Bone, D.B., Choi, D.S., Coe, I.R., Hammond, J.R., 2010. Nucleoside/nucleobase transport and metabolism by microvascular endothelial cells isolated from ENT1−/− mice. Am. J. Physiol. Heart Circ. Physiol. 299, H847–H856. Bouis, D., Hospers, G.A., Meijer, C., Molema, G., Mulder, N.H., 2001. Endothelium in vitro: a review of human vascular endothelial cell lines for blood vessel-related research. Angiogenesis 4, 91–102. Chang-Chun, C., Masuda, M., Szabò, Z., Szerafin, T., Szècsi, J., Van Belle, H., Flameng, W., 1992. Nucleoside transport inhibition mediates lidoflazine-induced cardioprotection during intermittent aortic crossclamping. J. Thorac. Cardiovasc. Surg. 104, 1602–1609. Cuddy, L.K., Gordon, A.C., Black, S.A., Jaworski, E., Ferguson, S.S., Rylett, R.J., 2012. Peroxynitrite donor SIN-1 alters high-affinity choline transporter activity by modifying its intracellular trafficking. J. Neurosci. 32, 5573–5584. Day, B.J., Fridovich, I., Crapo, J.D., 1997. Manganic porphyrins possess catalase activity and protect endothelial cells against hydrogen peroxide-mediated injury. Arch. Biochem. Biophys. 347, 256–262. Deussen, A., 2000a. Metabolic flux rates of adenosine in the heart. Naunyn Schmiedeberg's Arch. Pharmacol. 362, 351–363. Deussen, A., 2000b. Quantitative integration of different sites of adenosine metabolism in the heart. Ann. Biomed. Eng. 28, 877–883. Deussen, A., Stappert, M., Schafer, S., Kelm, M., 1999. Quantification of extracellular and intracellular adenosine production: understanding the transmembranous concentration gradient. Circulation 99, 2041–2047. Deussen, A., Brand, M., Pexa, A., Weichsel, J., 2006. Metabolic coronary flow regulation— current concepts. Basic Res. Cardiol. 101, 453–464. Dhanya, R., Arun, K.B., Syama, H.P., Nisha, P., Sundaresan, A., Santhosh Kumar, T.R., Jayamurthy, P., 2014. Rutin and quercetin enhance glucose uptake in L6 myotubes under oxidative stress induced by tertiary butyl hydrogen peroxide. Food Chem. 158, 546–554. Forstermann, U., 2010. Nitric oxide and oxidative stress in vascular disease. Pflugers Arch. 459, 923–939. Gonzalez-Polo, R.A., Soler, G., Rodriguezmartin, A., Moran, J.M., Fuentes, J.M., 2004. Protection against MPP+ neurotoxicity in cerebellar granule cells by antioxidants. Cell Biol. Int. 28, 373–380. Headrick, J.P., Lasley, R.D., 2009. Adenosine receptors and reperfusion injury of the heart. Handb. Exp. Pharmacol. 189–214. Headrick, J.P., Peart, J.N., Reichelt, M.E., Haseler, L.J., 2011. Adenosine and its receptors in the heart: regulation, retaliation and adaptation. Biochim. Biophys. Acta 1808, 1413–1428. Henry, P., Popescu, A., Puceat, M., Hinescu, M.E., Escande, D., 1996. Acute simulated ischaemia produces both inhibition and activation of K + currents in isolated ventricular myocytes. Cardiovasc. Res. 32, 930–939. Hyde, R.J., Cass, C.E., Young, J.D., Baldwin, S.A., 2001. The ENT family of eukaryote nucleoside and nucleobase transporters: recent advances in the investigation of structure/function relationships and the identification of novel isoforms. Mol. Membr. Biol. 18, 53–63.

75

Lakshmi, S.V., Padmaja, G., Kuppusamy, P., Kutala, V.K., 2009. Oxidative stress in cardiovascular disease. Indian J. Biochem. Biophys. 46, 421–440. Lee, B.E., Toledo, A.H., Anaya-Prado, R., Roach, R.R., Toledo-Pereyra, L.H., 2009. Allopurinol, xanthine oxidase, and cardiac ischemia. J. Investig. Med. 57, 902–909. Lidington, E.A., Moyes, D.L., McCormack, A.M., Rose, M.L., 1999. A comparison of primary endothelial cells and endothelial cell lines for studies of immune interactions. Transpl. Immunol. 7, 239–246. Loffler, M., Morote-Garcia, J.C., Eltzschig, S.A., Coe, I.R., Eltzschig, H.K., 2007. Physiological roles of vascular nucleoside transporters. Arterioscler. Thromb. Vasc. Biol. 27, 1004–1013. Masuda, M., Chang-Chun, C., Möllhoff, T., Van Belle, H., Flameng, W., 1992. Effects of nucleoside transport inhibition on long-term ex vivo preservation of canine hearts. J. Thorac. Cardiovasc. Surg. 104, 1610–1617. McAmis, W.C., Schaeffer Jr., R.C., Baynes, J.W., Wolf, M.B., 2003. Menadione causes endothelial barrier failure by a direct effect on intracellular thiols, independent of reactive oxidant production. Biochim. Biophys. Acta 1641, 43–53. Meldrum, K.K., Meldrum, D.R., Hile, K.L., Burnett, A.L., Harken, A.H., 2001. A novel model of ischemia in renal tubular cells which closely parallels in vivo injury. J. Surg. Res. 99, 288–293. Muller-Delp, J.M., Gurovich, A.N., Christou, D.D., Leeuwenburgh, C., 2012. Redox balance in the aging microcirculation: new friends, new foes, and new clinical directions. Microcirculation 19, 19–28. Munzel, T., Gori, T., Bruno, R.M., Taddei, S., 2010. Is oxidative stress a therapeutic target in cardiovascular disease? Eur. Heart J. 31, 2741–2748. Nyberg, M., Mortensen, S.P., Thaning, P., Saltin, B., Hellsten, Y., 2010. Interstitial and plasma adenosine stimulate nitric oxide and prostacyclin formation in human skeletal muscle. Hypertension 56, 1102–1108. Robillard, K.R., Bone, D.B., Hammond, J.R., 2008. Hypoxanthine uptake and release by equilibrative nucleoside transporter 2 (ENT2) of rat microvascular endothelial cells. Microvasc. Res. 75, 351–357. Ryzhov, S., McCaleb, J.L., Goldstein, A.E., Biaggioni, I., Feoktistov, I., 2007. Role of adenosine receptors in the regulation of angiogenic factors and neovascularization in hypoxia. J. Pharmacol. Exp. Ther. 320, 565–572. Schaper, W., 2005. Dipyridamole, an underestimated vascular protective drug. Cardiovasc. Drugs Ther. 19, 357–363. Tune, J.D., 2007. Control of coronary blood flow during hypoxemia. Adv. Exp. Med. Biol. 618, 25–39. Vanheel, B., Leybaert, L., De, H.A., Leusen, I., 1989. Simulated ischemia and intracellular pH in isolated ventricular muscle. Am. J. Physiol. 257, C365–C376. Volonte, C., D'Ambrosi, N., 2009. Membrane compartments and purinergic signalling: the purinome, a complex interplay among ligands, degrading enzymes, receptors and transporters. FEBS J. 276, 318–329. Wackers, F.J.T., 1991. Adenosine or dipyridamole: which is preferred for myocardial perfusion imaging. J. Am. Coll. Cardiol. 17, 1295–1296. Yang, X., Cohen, M.V., Downey, J.M., 2010. Mechanism of cardioprotection by early ischemic preconditioning. Cardiovasc. Drugs Ther. 24, 225–234. Zhang, T., Liang, X., Shi, L., Wang, L., Chen, J., Kang, C., Zhu, J., Mi, M., 2013. Estrogen receptor and PI3K/Akt signaling pathway involvement in S-(−)equol-induced activation of Nrf2/ARE in endothelial cells. PLoS ONE 8, e79075.