In Vivo Electrochemical Detection of Nitric Oxide in Tumor-Bearing

Rebecca A. Hunter , Benjamin J. Privett , W. Hampton Henley , Elise R. Breed , Zhe Liang , Rohit Mittal ... Fethi Bedioui , Abdulghani Ismail , Sophie...
0 downloads 0 Views 128KB Size
Anal. Chem. 2007, 79, 1030-1033

In Vivo Electrochemical Detection of Nitric Oxide in Tumor-Bearing Mice Sophie Griveau, Charlotte Dume´zy, Johanne Se´guin, Guy G. Chabot, Daniel Scherman, and Fethi Bedioui*

INSERM, U640, CNRS, UMR8151, Universite´ Rene´ Descartes Paris 5, EÄ cole Nationale Supe´ rieure de Chimie de Paris, Faculty of Pharmacy, Chemical and Genetic Pharmacology Laboratory, Paris, France

Interest in elucidating the mechanisms of action of various classes of anticancer agents and exploring the pathways of the induced-nitric oxide (NO) release provides an impetus to conceive a better designed approach to locally detect NO in tumors, in vivo. We report here on the first use of an electrochemical sensor that allows the in vivo detection of NO in tumor-bearing mice. In a first step, we performed the electrochemical characterization of a stable electroactive probe, K4Fe(CN)6, directly injected into the liquid microenvironment especially created around the electrode in the tumor. Second, the ability of the inserted electrode system to detect the presence of NO itself in the tumoral tissue was achieved by using the chemically modified Pt/Ir electrode as NO sensor and two NO donor molecules: diethylammonium (Z)-1-(N,N-diethylamino)diazen-1-ium 1,2-diolate (DEA-NONOate) and (Z)-1-[N(2-aminopropyl)-N-(2-ammonio propyl)amino]diazen-1ium 1,2-diolate (PAPA-NONOate). These two NO donor molecules allowed proving the electrochemical detection of (i) directly injected exogenous NO phosphate buffer solution into the tumor (decomposed DEA-NONOate) and (ii) biomimetically induced endogeneous release of NO in the tumoral tissue, upon injection of PAPA-NONOate into the tumor. This approach could be applied to the in vivo study of candidate anticancer drugs acting on the NO pathways. Tumor vasculature plays an essential role in tumor survival and development, leading to the concept of antiangiogenic and antivascular therapy.1-3 While the antiangiogenic strategy targets the formation of tumor new vessels, the antivascular drugs aim at targeting the existing tumoral vasculature and at causing tumor endothelial cells death.4 A fundamental prerequisite of this approach is based on the postulate that tumor vasculature is different from the normal one, at both morphological and molec* Corresponding author. E-mail: [email protected]. Phone: 33 144 27 67 13. Fax: 33 153 10 12 95. (1) Folkman, J. Semin. Oncol. 2002, 29, 15-18. (2) Sonveaux, P.; Feron, O. In Nitric Oxide, Cell Signaling and Gene Expression; Lamas, S., Cadenas, E., Eds.; CRC Traylor & Francis: Boca Raton, FA, 2006; pp 395-420, and references cited therein. (3) Fukumura, D; Kashiwaga, S.; Jain, R. K. Nat. Rev. Cancer 2006, 6, 521534, and references cited therein. (4) Tozer, G. M.; Kanthou, C.; Baguley, B. C. Nat. Rev. Cancer 2005, 5, 423435.

1030 Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

ular levels. One of the principal actors in the antivascular approach appears to involve the action of nitric oxide (NO).2,3,5 Indeed, several anticancer antivascular agents induce the release of NO (e.g., flavones, xanthone acetic acid).4-6 Thus, it is believed that this “induced NO release” could be considered, at least partially, as a marker of the anticancer activity of this class of agents in vivo. In addition, NO, as an apoptotic marker, might also be involved in the mechanism of action of classical cytotoxic agents on tumor cells themselves. Interest in elucidating the mechanism of action of several classes of anticancer agents and also exploring the pathways of the induced NO release provides an impetus for conceiving a welldesigned approach to locally detect NO in tumors, in vivo. Measuring NO from liquid samples is now well documented,7 and the electrochemical detection systems are recognized as being the only reliable analytical method to determine NO local concentration, without disturbing its metabolism and the associated regulatory pathways.8,9 Our interest in designing electrochemical sensing devices for biological NO detection in solution encouraged us to explore the possibility of extending the use of the electroanalytical concept and adapt it to a very special and complex biological tissue such as tumors. Although NO electrochemical sensors have been reported and used in various living cell cultures,8,9 none of these systems have been utilized for tracking NO in vivo in tumors. The major challenge in achieving this goal is the positioning of the NO sensor device in tumor-bearing mouse, so that it can function accurately in order to locally detect and quantify NO release. In the present study, we present the first successful use of platinum-based NO sensor in tumor. This paper details the testing of the electrode setup and especially its positioning in vivo. (5) Kashiwagi, S.; Izumi, Y.; Gohongi, T. J. Clin. Invest. 2005, 115, 1816-1827. (6) Lancaster, J. R., Jr.; Xie, K. Cancer Res. 2006, 66, 6459-6462. (7) Feelisch, M., Stamler, J. S., Eds, Methods in Nitric Oxide Research; Wiley: Chichester, UK, 1996. (8) Bedioui, F.; Villeneuve, N. Electroanalysis 2003, 15, 5-15, and references cited therein. (9) Wadsworth, R.; Stankevicius, E.; Simonsen, U. J. Vasc. Res. 2006, 43, 7085, and references cited therein. (10) Pereira Rodrigues, N.; Zurgil, N; Chang, SG; Henderson, J; Bedioui, F.; McNeil, C.; Deutsch, M. Anal. Chem. 2005, 77, 2733-2738. (11) Pontie´, M.; Gobin, C.; Pauporte´, T.; Bedioui, F.; Devynck, J. Anal. Chim. Acta 2000, 411, 175-185, and references cited therein. (12) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd ed.; Wiley: New York, 2001. 10.1021/ac061634c CCC: $37.00

© 2007 American Chemical Society Published on Web 12/20/2006

EXPERIMENTAL SECTION Chemicals. Nickel(II) tetrasulfonated phthalocyanine (NiTSPc), sodium phosphate salt, potassium chloride, sodium hydroxide, o-phenylenediamine salt, and Nafion solution (5% in aliphatic alcohols) were reagent grade and provided from Aldrich. Quadruple Teflon insulated platinum/iridium alloy (90/10) wire was purchased from Advent Research Materials Ltd., and silver (99.9%) wire was from Goodfellow. Diethylammonium (Z)-1-(N,N-diethylamino)diazen-1-ium 1,2-diolate (DEA-NONOate) and (Z)-1-[N-(3aminopropyl)-N-(n-propyl)amino]diazen-1-ium 1,2-diolate (PAPANONOate) were from Cayman Chemical and kept at -80 °C. These NO donors are also called NONOates. Tumor-Bearing Mice. Lewis lung carcinoma tumors were implanted subcutaneously in the right flank of 5-week-old female C57Bl/6 mice and 10-14 days later, studies were undertaken. Mice were anesthetized by intraperitoneal injection of a mix of ketamine (85.8 mg/kg; Centravet) and xylazine (3.1 mg/kg; Bayer) diluted in 150 mM NaCl. Experiments were conducted following the NIH recommendations and in agreement with a regional ethic committee for animal experimentation. NONOate Solution Preparation. NONOate stock solution (∼50 mM) was prepared by dissolving 1 mg of the NONOate salt in 100 µL of 0.01 M NaOH solution. The stock solution was prepared daily and kept at 0 °C in the dark. NO stock solutions injected in the tumors from DEA-NONOate were prepared as follows: the release of NO was initiated by adding an aliquot of the NONOate stock solution in 2 mL of desoxygenated 0.1 M phosphate buffer solution (PBS, pH 7.4). The half-life of DEANONOate at pH 7.4 and 26 °C is 9 min.13 The solution was injected in the tumor after 15 min (NO concentration, 1.1 mM). PAPANO solution injected in the tumor was the NONOate stock solution in 0.01 M NaOH. Electrochemical Apparatus. Electrochemical experiments were carried out at room temperature, in aerobic conditions, using a Picostat model potentiostat (eDAQ Pty Ltd.) for cyclic voltammetry and a BAS Petit Ampere for chronoamperometry. The pseudoreference Ag/AgCl electrode, serving also as the counter electrode, was a homemade silver/silver chloride wire (diameter 50 µm). NO Sensor. A length of ∼5 cm of a Teflon-PTFE insulated Pt/Ir wire (125-µm diameter) was sharply cut to form a neat diskshaped area at its end. This disk area (electrode) was then modified by electrodepositing a thin film of nickel phthalocyanine from 2 mM nickel tetrasulfonated phthalocyanine in 0.1 M NaOH aqueous solution, by cyclic voltammetry (30 successive scans at 0.1 V s-1).10 Following this step, the electrode was rinsed, dried, and then immerged in Nafion alcoholic solution during 15 s. To ensure a good adhesion of the Nafion layer to the electrode surface, the electrode was placed at 80 °C for 5 min. This protocol was repeated 4 times. Finally, o-phenylenediamine (o-PD) thin film was electrochemically deposited by controlled potential electrolysis at 0.9 V versus Ag/AgCl in 6.5 mM o-PD in PBS during 20 min. This outermost layer was aimed at enhancing the selectivity of the electrode against possible interfering molecules.8,11 This in-house-constructed electrode was used as the working electrode (NO sensor) and was associated with an Ag/AgCl wire (50)µm diameter) as a reference electrode. In order to minimize the size (13) www.caymanchem.com.

Figure 1. Schematic representation of the platinum/iridium-based electrochemical NO sensor and Ag/AgCl reference electrode.

of the electrode setup, the Ag/AgCl wire was twisted around the Teflon-PTFE insulated NO sensor (Figure 1). The electrochemical signals related to NO were measured amperometrically at 0.8 V versus Ag/AgCl.8,10 RESULTS AND DISCUSSION The study in the tumor-bearing mouse requires a precise positioning of the electrodes in order to avoid deteriorating the surface modifications of the sensor and to minimize the disturbance of the electrical signals due to the mouse breathing. To do so, the tumor was slightly perfored using a 12-gauge needle in order to create a liquid microenvironment, which would allow the electrochemical detection. The electrode device was then guided by a plastic capillary and inserted inside the tumor (Figure 2). In a first step, we performed the electrochemical characterization of a stable electroactive probe, K4Fe(CN)6, directly injected in the liquid microenvironment especially created around the electrode in the tumor. This was aimed at assessing the possibility of electrodetecting redox moieties present within the tumor tissue, at the vicinity of the Pt/Ir and Ag/AgCl set of electrodes. The needle of the syringe was introduced 1 mm distant parallel to the electrode device. Figure 3 displays the cyclic voltammograms recorded at the inserted electrode system in the absence of K4Fe(CN)6 (curve 1) and upon injection of 0.6 × 10-6 (curve 2) and 0.9 × 10-6 mol (curve 3) of K4Fe(CN)6 in PBS. The total injected volumes were 60 and 90 µL, respectively. The cyclic voltammograms were recorded less than 1 min after each injection. A quasi-reversible pair of voltammetric peaks, related to the redox couple FeIII/FeII,12 is clearly observed at 0.14 V versus Ag/AgCl. Thus, it can be clearly stated that the electrode system is functional in the tumoral tissue, without any significant complication (such as passivation, fouling etc.). Note that, due to the leaking of the liquid solution from the drilled cavity containing the electrode setup, the intensity of the shown cyclic voltammograms evolved with time. Subsequent experiments probed the ability of the inserted electrode system to electrodetect the presence of NO itself in the tumor. This was accomplished by using the chemically modified Pt/Ir electrode as NO sensor and two NO donor molecules: DEANONOate and PAPA-NONOate. Two experimental protocols were performed: (i) the direct injection of exogenous NO into the tumor (through decomposed DEA-NONOate in PBS) and (ii) the direct injection of PAPA-NONOate directly into the tumor, in order to mimic an endogenous release of NO. The first experiment was performed by injecting locally, in the microenvironment created Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

1031

Figure 2. Schematic drawing of the positioning of the electrochemical sensor in the tumor-bearing mouse.

Figure 3. Intratumoral cyclic voltammograms recorded at the inserted Pt/Ir (125 µm) and Ag/AgCl (50 µm) electrode system in the absence of K4Fe(CN)6 (curve 1) and upon injection of 0.6 × 10-6 (injected volume, 60 µL; curve 2) and 0.9 × 10-6 mol (injected volume, 90 µL; curve 3) of K4Fe(CN)6. Scan rate, 50 mV/s.

for inserting the electrode device, aliquots of 30 µL of NO phosphate buffer solution (pH 7.4) prepared at room temperature (26 °C) from totally decomposed DEA-NONOate.13-15 The amperometric response of the inserted electrode was measured at 0.8 V versus Ag/AgCl. Figure 4 shows the obtained amperogram upon three successive boluses of 33 × 10-9 moles of NO (injected volume, 30 µL). It displays a sharp increase in intensity immediately after each injection of NO (less than 10 s), followed by a relatively slow decrease in the amplitude of the current to reach the baseline 100 s later. The shape of the amperometric signals is similar to typical amperometric responses related to instantaneous NO production in various biological solutions.8,9 Thus, the evolution of the amperometric response reflects (i) the introduction of exogenous NO in the tumor (transcribed in the amperogram by the sharp increase in the measured current) and (ii) its spreading and reactivity (with oxygen and other biological species (14) Horstmann, A.; Menzel, L.; Ga¨bler, R.; Jentsch, A.; Urban, W.; Lehman, J. Nitric Oxide 2002, 6, 135-141. (15) Feelisch, M. Naunym Schimedeberg’s Arch. Pharmacol. 1998, 358, 113122, and references cited therein.

1032 Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

Figure 4. Intratumoral chronoamperograms recorded at the inserted NO sensor upon repeated injection of 33 × 10-9 mol of NO (injected volume, 30 µL; arrows indicated the injections of the solution aliquots). Operating potential, 0.8 V vs Ag/AgCl.

present in the tumor) through the tumoral tissue (transcribed in the amperogram by the decrease of the measured current). These data clearly indicate that, upon each addition, NO can be detected and the evolution of its local presence can be temporarily observed and analyzed. It should be noted that the second and third injections of NO were intentionally performed at a distance closer to the electrode setup. This explains the large intensity of the measured current, immediately after these injections, due to the higher availability of NO at the sensor proximity, before it started to dissipate. Finally, injections of 30 µL of pure PBS aliquots (exempt of NO) did not provoke any change in the baseline current (data not shown). Also, the NO sensor showed no interference from totally decomposed DEA-NONOate solution, containing only the ultimate byproducts of the decomposition of the NONOate. In a second series of experiments, alkaline PAPA-NONOate solution was directly injected in the tumor, at the vicinity of the sensor (the needle of the syringe was introduced 1-2 mm distant parallel to the electrode). PAPA-NONOate does not decompose in the alkaline solution,13-15 and it is reported that its half-life is 15 min at pH 7.4 and 37 °C.13 Thus, it is expected that PAPANONOate starts releasing NO only upon its introduction in the liquid microenvironment created in the tumoral tissue. Also, it is

Figure 5. (a) Intratumoral chronoamperogram recorded at NO sensor upon injection of 5 µL of PAPA-NONOate alkaline solution (300 × 10-9 mol). Arrow indicates the injection time. Operating potential, 0.8 V vs Ag/AgCl. (b) Typical amperometric curve illustrating the kinetics of production/ degradation of NO in aerobic phosphate buffer solution (pH 7.4) at 26 °C upon injection of PAPA-NONOate stock 0.01 M NaOH solution (final PAPA-NONOate concentration in PBS, 20 µM). Operating potential, +0.8 V.

important to note that the microcavity where the sensor device was inserted was rinsed twice by physiological PBS solution (150 µL in total) before injecting the NO donor. Figure 5a depicts the amperogram, which displays an increase in the intensity of the measured current less than 15 s after injection of 300 × 10-9 mol of PAPA-NONOate (injected volume, 5 µL). The current reached a maximum after 2 min and then slowly decreased during the following 15 min. The shape of the obtained amperogram is similar to a typical amperometric response related to NO release from PAPA-NONOate in PBS upon injection of the parent alkaline solution containing the NO donor (Figure 5b). It should be noted that no interfering signal was observed at the used sensor from totally decomposed PAPANONOate in PBS solution containing only the ultimate byproducts (data not shown). Thus, the maximal current measured 2 min after the injection of PAPA-NONOate is indicative of a NO release process, which is known to occur more rapidly at low pH (which is the case of tumor pH).2 Indeed, due to their high metabolism and relative hypoxic status, tumors are recognized to show 0.5-1 pH unit decrease relative to physiological value. Thus, the shape of the amperogram and its evolution are clearly indicative of the release of NO from PAPA-NONOate in the tumor, followed by its consecutive diffusion and degradation (mainly by reaction with oxygen). Finally, it is important to note that, in the reported experiment, only 5 µL of alkaline solution (containing PAPANONOate) was injected in the tumor, after its rinsing with PBS. In fact, the volume of the NO donor alkaline solution should be minimized to a certain extent. Indeed, one can expect that exposure to alkaline solution may result in stressing tumoral and endothelial cells, leading to spontaneous release of various cytotoxic reagents and their apoptosis.3,16 Such a situation may in (16) Tolias, C. M.; McNeil, C.; Kazlauskaite, J.; Hillhouse, E. W. Free Radical Biol. Med. 1999, 26, 99-106.

turn induce interfering electrochemical signals. In a separate experiment, injections of large volumes of alkaline solution (>50 µL), with or without NO donor, induced the apparition of spikes on the amperogram (data not shown). As a final remark, given the difficulties of estimating the precise local concentration of the injected compounds, due to unknown real volume of the cavity, the leaking of the solution from the tumor, the high reactivity of NO in the complex medium, etc., the obtained data are only indicative of the possibility of tracking and monitoring mimicked endogenous NO release in the tumor. Also, due to the fact that both inducible and constitutive nitric oxide synthase enzymes are involved in NO production by the tumoral system,3 the expected local NO concentration would be within the micromolar range, which is within the detection range of the above-described electrochemical detection system.8 CONCLUSION We have demonstrated that a miniaturized electrochemical sensor can be used in vivo for probing electroactive species in tumor-bearing mice. This was first achieved by performing for the first time in vivo cyclic voltammetry of K4Fe(CN)6. In addition, the use of two NO donors allowed us, for the first time, to demonstrate the feasibility of the electrochemical detection of directly injected exogenous NO in live tumor-bearing mice and to induce in vivo biomimetic endogeneous release of NO in tumoral tissue. Although the performances of the sensor need to be optimized, this study clearly shows the usefulness of this practical approach that could be applied to study the mechanism of action and monitoring the efficiency of anticancer drugs acting on NO pathways. Received for review August 31, 2006. Accepted November 7, 2006. AC061634C

Analytical Chemistry, Vol. 79, No. 3, February 1, 2007

1033