Microdialysis Sampling Combined with Electron ... - ACS Publications

Quantitation of superoxide radical (O2•-) production at the site of radical generation remains challenging. Microdialysis sampling is an advantageou...
0 downloads 0 Views 101KB Size
Download PDF
Anal. Chem. 2004, 76, 4734-4740

Microdialysis Sampling Combined with Electron Spin Resonance for Superoxide Radical Detection in Microliter Samples Rui Chen, Joseph T. Warden,* and Julie A. Stenken*

Department of Chemistry and Chemical Biology, Rensselaer Polytechnic Institute, 110 8th Street, Troy, New York 12180 Quantitation of superoxide radical (O2•-) production at the site of radical generation remains challenging. Microdialysis sampling is an advantageous tool for sampling from localized environments. It is difficult to combine electron spin resonance (ESR) spin traps with microdialysis because O2•- adducts with common nitrone spin traps have shorter half-lives than typical microdialysis collection times. Furthermore, typical dialysate samples (5-15 µL) suffer significant sensitivity loss when diluted for detection in a conventional ESR flat cell (200 µL). To overcome these difficulties, a cyclic hydroxylamine, 1-hydroxy-4-phosphonooxy-2,2,6,6-tetramethylpiperidine (PP-H), which produces a stable nitroxide radical (PP•) product upon reaction with O2•- was employed. Capillary cells (1.4 µL effective volume) coupled with a loop-gap resonator were utilized to measure PP• in microliter microdialysis samples (LOD 0.36 pmol). A xanthine/xanthine oxidase (X/XO) model system provided sustained O2•- production. When PP-H was included in the X/XO medium external to the microdialysis probe, a relative recovery of 22.1 ( 1.1 and 57.2 ( 5.7% for PP• was achieved at perfusion fluid flow rates of 0.5 and 1.0 µL/min, respectively. The respiratory burst in interferon-γ and zymosan-stimulated RAW 264.7 macrophages was also investigated. Reactive oxygen species such as superoxide radical (O2•-), hydrogen peroxide, and hydroxyl radical (HO•) from oxidative stress events are suspected to play important roles in various disease processes.1-3 Hydroxyl radical is the most reactive oxygen radical known and initiates lipid peroxidation as well as oxidative damage of protein and DNA. Superoxide is much less reactive with most substrates as compared to HO• but reacts rapidly with nitric oxide and transition metals. Therefore, understanding the role of O2•- during oxidative stress is of vital importance.2 Site-specific sampling methods for O2•- will be the key to understanding its in vivo functions. Since short-lived O2•- can * Corresponding authors. E-mail: [email protected]. Phone: 518-276-2045. Fax: 518-276-4887. E-mail: [email protected]. Phone: 518-276-8482. Fax: 518276-4887. (1) Halliwell, B.; Gutteridge, J. M. C. Free Radicals in Biology and Medicine, 3rd ed.; Clarendon Press: Oxford, 1999. (2) Catherine, A. R., Burdon, R. H., Eds. Free Radical Damage and its Control; Elsevier: Amsterdam, 1994. (3) Forman, H. J.; Torres, M. Mol. Aspects Med. 2001, 22, 189-216.

4734 Analytical Chemistry, Vol. 76, No. 16, August 15, 2004

behave either as a one-electron reductant or as a one-electron oxidant, a variety of indirect methods for detection of O2•- have been described in the literature. Superoxide dismutase (SOD)inhibitable reduction of ferricytochrome c (cyt c) is among the most popular approaches to use for superoxide detection due to the ease of experimental setup and inexpensive instrumentation.4 In the cyt c reduction approach, an absorbance change at 550 nm due to ferrocytochrome c production is measured. Similar spectroscopic approaches include reduction of nitroblue tetrazolium.5 Chemiluminescence assays utilizing luminol, lucigenin, and analogues have also been developed for O2•- detection.6,7 Zhao et al.8 recently reported a specific reaction between hydroethidine and O2•-. Although hydroethidine can be oxidized to ethidium by a variety of oxidants, only O2•- oxidizes hydroethidine to form a monooxygenated fluorescent product. It should be noted that care must be taken during data interpretation with all of these methods since many of the products formed can be affected by either transition metals or cellular antioxidants such as ascorbate and glutathione. Improving UV-visible or fluorescence spectroscopic methods for detecting O2•- that are not susceptible to artifacts caused by transition metals or antioxidants is an active research area.9,10 Electron spin resonance (ESR) combined with spin trapping agents has been applied to assay O2•- indirectly.1,2,11,12 5,5Dimethyl-1-pyrroline N-oxide (DMPO) is a commonly used nitrone spin trap for oxygen-centered radicals. The second-order reaction rate constants for HOO• and O2•- with DMPO are pH dependent (pKa ) 4.88) with values of 6.6 × 103 and 10 M-1 s-1, respectively.13 Therefore, the reaction of O2•- with DMPO favors acidic conditions since the rate constant is 30 M-1 s-1 at pH 7.4.14 Furthermore, (4) McCord, J. M.; Fridovich, I. J. Biol. Chem. 1969, 244, 6049-6055. (5) Bielski, B. H. J.; Shiue, G. G.; Bajuk, S. J. Phys. Chem. 1980, 84, 830-833. (6) Seitz, W. R.; Neary, M. P. Methods Biochem. Anal. 1976, 23, 161-188. (7) Campbell, A. K.; Hallett, M. B.; Weeks, I. Methods Biochem. Anal. 1985, 31, 317-416. (8) Zhao, H.; Kalivendi, S.; Zhang, H.; Joseph, J.; Nithipatikom, K.; VasquezVivar, J.; Kalyanaraman, B. Free Radical Biol. Med. 2003, 34, 1359-1368. (9) Kutala, V. K.; Parinandi, N. L.; Zweier, J. L.; Kuppusamy, P. Arch. Biochem. Biophys. 2004, 424, 81-88. (10) Barbacanne, M. A.; Souchard, J. P.; Darblade, B.; Iliou, J. P.; Nepveu, F.; Pipy, B.; Bayard, F.; Arnal, J. F. Free Radical Biol. Med. 2000, 29, 388396. (11) Rosen, G. M.; Rauckman, E. J. Methods Enzymol. 1984, 105, 198-209. (12) Rosen, G. M.; Britigan, B. E.; Halpern, H. J.; Pou, S. Free Radicals: Biology and Detection by Spin Trapping; Oxford University Press: New York, 1999. (13) Behar, D.; Czapski, G.; Rabani, J.; Dorfman, L. M.; Schwarz, H. A. J. Phys. Chem. 1974, 74, 3209-3213. 10.1021/ac035543g CCC: $27.50

© 2004 American Chemical Society Published on Web 07/03/2004

the DMPON-superoxide radical adduct has a half-life of less than 1 min and spontaneously decomposes to the DMPO-hydroxyl radical adduct.15,16 Developing spin traps with enhanced stability such as 5-(diethoxyphosphoryl)-5′-methyl-1-pyrroline N-oxide (DEPMPO, t1/2 ) 15 min)17 and 5-ethoxycarbonyl-5′-methyl-1-pyrroline N-oxide (EMPO, t1/2 ) 5 min) has been an important research area.18,19 Currently, there is an extensive interest in using the nonselective nature of microdialysis sampling as a means to collect potential radical adducts from living tissue.20-22 Rather than analyzing samples that are far removed from the site of radical generation and subjected to metabolic pathways, microdialysis sampling is advantageous because it can be applied to sample collection within a localized environment. Microdialysis sampling is nonselective and is governed by bidirectional diffusion processes across the membrane. The efficiency of microdialysis sampling for a particular analyte is determined by the ratio of inlet and outlet sample concentrations. The extraction efficiency (Ed) is shown in eq 1,

Ed )

Cinlet - Coutlet Cinlet - Csample

(1)

where Cinlet is the analyte concentration in the inlet solution, Coutlet is the analyte concentration in microdialysates exiting the probe, and Csample is the analyte concentration far away from the probe.23 Typically, a predetermined Ed is used to relate the measured analyte concentration in microdialysate samples to the analyte concentration in the external sample medium. The nonselective nature of the microdialysis sampling process allows infusion of agents such as spin traps through the microdialysis probe to give a localized delivery of the infused substance.24,25 The formed product can then diffuse back into the probe. Microdialysis sampling coupled with conventional ESR is not practical as a routine radical detection method given sample volume and sample collection time constraints. The choice of microdialysis flow rate is a tradeoff between temporal resolution and analytical sensitivity needs.26 The typical microdialysis flow rate is between 0.5 and 2.0 µL/min. Hence, the resulting low(14) Finkelstein, E.; Rosen, G. M.; Rauckman, E. J. J. Am. Chem. Soc. 1980, 102, 4994-4999. (15) Finkelstein, E.; Rosen, G. M.; Rauckman, E. J. Arch. Biochem. Biophys. 1980, 200, 1-16. (16) Buettner, G. R.; Mason, R. P. Methods Enzymol. 1990, 186, 127-133. (17) Frejaville, C.; Karoui, H.; Tuccio, B.; le Moigne, F.; Culcasi, M.; Pietri, S.; Lauricella, R.; Tordo, P. J. Chem. Soc., Chem. Commun. 1994, 1793-1794. (18) Olive, G.; Mercier, A.; Le Moigne, F.; Rockenbauer, A.; Tordo, P. Free Radical Biol. Med. 2000, 28, 403-408. (19) Khan, N.; Wilmot, C. M.; Rosen, G. M.; Demidenko, E.; Sun, J.; Joseph, J.; O’Hara, J.; Kalyanaraman, B.; Swartz, H. M. Free Radical Biol. Med. 2003, 34, 1473-1481. (20) Zini, I.; Tomasi, A.; Grimaldi, R.; Vannini, V.; Agnati, L. F. Neurosci. Lett. 1992, 38, 279-282. (21) Bogdanov, M. B.; Ferrante, R. J.; Mueller, G.; Ramos, L. E.; Martinou, J.-C.; Flint Beal, M. Neurosci. Lett. 1999, 262, 33-36. (22) Dugan, L. L.; Lin, T. S.; He, Y. Y.; Hsu, C. Y.; Choi, D. W. Free Radical Res. 1995, 23, 27-32. (23) Bungay, P. M., Morrison, P. F., Dedrick, R. L. Life Sci. 1990, 46(2): 105119. (24) Stenken, J. A.; Holunga, D. M.; Decker, S. A.; Sun, L. Anal. Biochem. 2001, 290, 314-323. (25) Chen, R.; Stenken, J. A. Anal. Biochem. 2002, 306, 40-49. (26) Davies, M. I.; Lunte, C. E. Chem. Soc. Rev. 1997, 26, 215-222.

Scheme 1. Reaction of Hydroxylamine PP-H with Superoxide Radical

volume dialysate samples (5-15 µL) incur significant sensitivity loss when diluted for detection in a standard X-band ESR flat cell (200-µL volume). Additionally, the instability of nitrone spin trap superoxide adducts makes these agents unsuitable for O2•detection during microdialysis sampling since their half-lives are shorter than the microdialysis collection time. Furthermore, the low trapping efficiency of the nitrone spin traps necessitates using high concentrations (50-100 mM) of spin traps that may be incompatible with in vivo applications. To enable sensitive ESR measurements with microliter microdialysis samples, capillary cells coupled with a loop-gap resonator (LGR) (effective volume, 1.4 µL) were employed. As demonstrated by Froncisz and Hyde, the LGR offers substantial sensitivity enhancement for microliter aqueous samples of saturable radicals.27,28 An alternative approach for O2•- detection by ESR is “spin exchange” or “spin scavenging”, in which a hydroxylamine is oxidized by superoxide to its corresponding ESR-active nitroxide.29-32 The hydroxylamine, 1-hydroxy-4-phosphonooxy2,2,6,6-tetramethylpiperidine (PP-H, Scheme 1), has been used for in vivo detection of superoxide.31 PP-H reacts with O2•- faster than the cyclic nitrones (second-order reaction rate constant, 8.4 × 102 M-1 s-1), and the resulting nitroxide radical (PP•) is stable for hours. Pyrrolidine nitroxides are generally resistant to reduction via ascorbate and thiols.33,34 In particular, the presence of bulky groups adjacent to the carbon-nitroxide bond sterically hinders chemical reduction.35 A low concentration of 0.5 mM for PP-H is sufficient for O2•- detection. Therefore, PP-H is more suitable than the normal nitrone spin traps for microdialysis sampling of O2•-. In this work, a microdialysis/ESR assay employing PP-H for O2•- detection in microliter microdialysis samples was developed. A xanthine/xanthine oxidase (X/XO) model system was used to provide sustained O2•- production for these preliminary studies. (27) Hoff, A. J., Ed. Advanced EPR Techniques: Applications in Biology and Biochemistry; Elsevier: Amsterdam, 1989; Chapter 7. (28) Froncisz, W.; Hyde, J. S. J. Magn. Reson. 1982, 47, 515-521. (29) Rosen, G. M.; Finkelstein, E.; Rauckman, E. J. Arch. Biochem. Biophys. 1982, 215, 367-378. (30) Dikalov, S.; Grigor’ev, I. A.; Voinov, M.; Bassenge, E. Biochem. Biophys. Res. Commun. 1998, 248, 211-215. (31) Fink, B.; Dikalov, S.; Bassenge, E. Free Radical Biol. Med. 2000, 28, 121128. (32) Dikalov, S. I.; Dikalova, A. E.; Mason, R. P. Arch. Biochem. Biophys. 2002, 402, 218-226. (33) Morris, S.; Sosnovsky, G.; Hui, B.; Huber, C. O.; Rao, N. U. M.; Swartz, H. M. J. Pharm. Sci. 1991, 80, 149-152. (34) Couet, W. R.; Brasch, R. C.; Sosnovsky, G.; Lukszo, J.; Prakash, I.; Gnewuch, C. T.; Tozer, T. N. Tetrahedron 1985, 41, 1165-1172. (35) Yu, Z.; Kotake, Y.; Janzen, E. G. Redox Rep. 1996, 2, 133-139.

Analytical Chemistry, Vol. 76, No. 16, August 15, 2004

4735

Sensitive measurements of PP• in 5-µL dialysates were performed using capillary cells coupled with a LGR. Additionally, O2•production from stimulated RAW 264.7 macrophages was investigated. EXPERIMENTAL SECTION Reagents. PP-H was purchased from Alexis Corp. Superoxide dismutase, ferricytochrome c, xanthine, and the cell culture reagents were obtained from Sigma Chemical (St. Louis, MO). Xanthine oxidase suspension of 20 units/mL was purchased from Roche Molecular Biochemicals (Indianapolis, IN). All other chemicals were obtained from Fisher Scientific (Pittsburgh, PA) and were reagent grade or better. Nanopure water was prepared using a Barnstead Nanopure water purification system (Dubuque, IA). Two separate diethylenetriaminepentaacetic acid (DTPA)containing buffers were used, denoted as sodium phosphate buffers A and B. Sodium phosphate buffer A contained 50 mM formal concentration of phosphate, pH 7.4, 0.9 wt % NaCl, and 1 mM DTPA. Sodium phosphate buffer B contained 50 mM formal concentration phosphate, pH 7.4, 0.9 wt % NaCl, and 0.25 mM DTPA. The third buffer was phosphate-buffered saline (PBS) and contained 50 mM formal concentration of phosphate, pH 7.4, and 0.9 wt % NaCl. The RAW 264.7 macrophage cell line is a murine macrophage cell line and was obtained as a gift from Professors Daniel Loegering and Michelle Lennartz, Albany Medical College. This cell line was originally purchased from American Type Culture Collection (ATCC, Manassas, VA). Trypan blue solution (0.4%), interferon-γ (IFN-γ) from mouse, zymosan A from Saccharomyces cerevisiae, and sterile Dulbecco’s phosphate-buffered saline solution were obtained from Sigma Chemical. Dulbecco’s modification of Eagle’s medium, penicillin (100 units/mL), and streptomycin (100 µg/mL) were purchased from Fisher Scientific (Pittsburgh, PA). Fetal bovine serum was purchased from Biowhittaker (Walkersville, MD). Preparation of Nitroxide Products. A fresh 5 mM stock solution of PP-H was prepared daily by dissolving the solid in oxygen-free (bubbled for 20 min with argon) sodium phosphate buffer A. This 5 mM PP-H stock solution was stored in an airtight brown vial at 4 °C until use. PP-H/X/XO samples were prepared by mixing 45 µL of sodium phosphate buffer B, 15 µL of xanthine in PBS (500 µM-5 mM), 65 µL of water, and 15 µL of 5 mM PP-H; then the reaction was initiated by addition of 10 µL of XO solution (0.2-2 units/mL). The final concentration of DTPA in this mixture was 0.2 mM. The solution was gently agitated before analysis. Superoxide Production. To ensure that X/XO had reproducible kinetics, uric acid production was monitored at 295 nm.36 To measure this activity, a mixture was prepared using 300 µL of sodium phosphate buffer A, 300 µL of PBS, 150 µL of xanthine (500 µM in PBS), and 740 µL of water. This reaction was initiated by addition of 10 µL of XO solution (2 units/mL). The reference cell contained the same solution with xanthine being replaced by PBS. The absorbance increase at 295 nm due to uric acid production was monitored using a Hitachi U-2000 spectrophotometer. Superoxide production was confirmed using a cyt c reduction assay.4 A mixture was prepared using 300 µL of sodium phosphate (36) Kalckar, H. M. J. Biol. Chem. 1947, 167, 429-443.

4736

Analytical Chemistry, Vol. 76, No. 16, August 15, 2004

buffer A, 270 µL of PBS, 150 µL of xanthine (500 µM) in PBS, 740 µL of water, and 30 µL of cyt c (1 mM in PBS). Addition of 10 µL of XO solution (2 units/mL) initiated the reaction. The reference cell contained the above-mentioned solution and 10 µL of SOD (10 000 units/mL). The increase in absorbance at 550 nm due to cyt c reduction was monitored using a Hitachi U-2000 spectrophotometer. Xanthine oxidase production of uric acid from xanthine can be stopped using an appropriate enzyme denaturing agent such as acetic acid or allopurinol, a XO inhibitor. The enzymedenaturing effectiveness of either acetic acid or allopurinol was determined by mixing different concentrations of these inhibitors with the X/XO solution coupled with monitoring the decrease in uric acid generation at 295 nm. For the LC/MS study of PP• production, 100-µL samples were withdrawn at different time points from the PP-H/XO/X mixture and mixed immediately with 2 µL of 10 M acetic acid to denature the xanthine oxidase. These acetic acid-treated samples were centrifuged, and the supernatants were stored at 4 °C prior to same-day LC/MS analysis. For ESR time course studies, 15-µL aliquots were withdrawn from the PP-H/XO/X mixture per minute and mixed immediately with 1 µL of allopurinol (3 mM, prepared in 0.1 M HCl). LC/MS Experiments. Mass spectrometry was performed using an Agilent 1100 series LC/MSD instrument (Agilent, Palo Alto, CA) with electrospray (ESI) ionization in both the positiveand negative-ion modes. A Hypersil C18 column (5 µm, 250 × 2.00 mm) with an Ultracarb ODS 20 (4 µm, 30 × 2.00 mm) guard column (Phenomenex, Torrance, CA) was used for separation. Ammonium acetate (10 mM) was added to the mobile phase (95/ 5, v/v, water/methanol) to enhance the production of positively charged ions. The mobile phase was delivered at 0.2 mL/min, and the sample loop volume was 10 µL. A photodiode array detector was used to obtain absorbance spectra for each peak between 200 and 600 nm. Microdialysis Sampling. Microdialysis experiments were performed using a nonmetal syringe (Hamilton Syringes, Reno, NV) and nonmetal microdialysis probes with 4 mm of exposed membrane (BR-4 probe, Bioanalytical Systems, Inc., West Lafayette, IN). A CMA-102 syringe pump (CMA/Microdialysis, North Chelmsford, MA) was used to perfuse the probe. A diagram depicting the microdialysis sampling setup for both delivery and recovery experiments is shown in Figure 1. In all time course studies, dialysates collected every 10 min were either subjected to ESR analysis immediately or stored on ice for less than 30 min prior to analysis. Aliquots (5 µL) from samples external to the probe were withdrawn every 10 min and mixed immediately with 1 µL of 1 M allopurinol. For PP-H recovery experiments, 0.5 mM PP-H was included in the O2•- generating system external to the probe. The external sample medium (PP-H/X/XO) was prepared by mixing 100 µL of phosphate buffer A, 200 µL of PBS, 100 µL of 500 µM xanthine in PBS, 495 µL of water, and 100 µL of 5 mM PP-H; then the reaction was initiated by addition of 5 µL of XO solution (2 units/ mL). The final concentration of DTPA in the mixture was 0.2 mM. The probe was perfused with a saline (0.9% NaCl) solution at flow rates of 0.5 and 1.0 µL/min.

Figure 1. Diagram of microdialysis sampling setup. (A) 5 mM PP-H solution was perfused inside of the microdialysis probe. X/XO mixture was placed outside of the microdialysis probe. (B) Saline solution was perfused inside of the microdialysis probe. 0.5 mM PP-H solution was mixed with X/XO and placed external to the microdialysis probe.

For delivery experiments, the same conditions were used as described for the recovery experiments except that 100 µL of sodium phosphate buffer A was substituted for 100 µL of PP-H. A 5 mM PP-H solution in argon-treated sodium phosphate buffer A was perfused through the probe at 0.5 µL/min, and the solution external to the probe was kept well stirred. ESR Experiments. ESR measurements were obtained at room temperature using a computer-controlled Varian E-9 spectrometer at X-band. Suprasil quartz capillary cells (0.6-mm i.d., Vitrocom Inc., Moutain Lakes, NJ) were mounted in a loop-gap resonator (model XP-0201, Molecular Specialties, Milwaukee, WI). Typical ESR spectrometer settings were as follows: modulation amplitude, 1 G; time constant, 0.3 s; microwave power, 2 mW; receiver gain, 2.5 × 103; scan time, 256 s. Quantitation of PP• was performed by comparison of its baseline-corrected second integral with that of a 10 µM nitroxide standard, 4-hydroxy-2,2,6,6-tetramethyl-piperidinooxy (TEMPOL). Integration was performed using WINEPR (Bruker Instruments, Inc., Billerica, MA). To enable accurate quantitation of signals with low S/N, ESR spectra were simulated using the optimization routines in WinSim (available from the NIEHS37) prior to integration.38 The limit of detection (LOD) and limit of quantitation (LOQ) were obtained by performing replicate measurements (n ) 7) of a 3 µM TEMPOL standard as outlined in a standard reference text.39 Cell Culture Experiments. RAW 264.7 cells were cultured in Dulbecco’s modification of Eagle’s medium supplemented with 10% fetal bovine serum, penicillin (100 units/mL), and streptomycin (100 µg/mL). Cells were incubated at 37 °C with humidified 5% CO2. Cells were plated at 2 × 106/well in 24-well plates and incubated overnight with IFN-γ (100 units/mL). Then cells were washed with sterile Dulbecco’s phosphate-buffered saline solution (Dulbecco’s PBS), and 800 µL of Dulbecco’s PBS solution and 100 µL of 5 mM PP-H solution were added to each well. For SOD controls, 790 µL of Dulbecco’s PBS solution, 10 µL of SOD (37) http://epr.niehs.nih.gov/pest.html. (38) Duling, D. R. J. Magn. Reson. Ser. B 1994, 104, 105-110. (39) Harris, D. C. Quantitative Chemical Analysis, 6th ed.; W. H. Freeman & Co.: New York, 2003; pp 726-727.

Figure 2. ESR spectra of representative samples observed with the loop-gap resonator: (a) 500 µM PP-H, 500 µM xanthine with 0.02 unit of XO and 50 unit of SOD; (b) 500 µM PP-H with 0.02 unit of XO; (c) 500 µM PP-H; (d) 500 µM PP-H, 500 µM xanthine with 0.02 unit of XO. The hyperfine ESR splitting constant of PP• was aN ) 16.8 G with a line width (peak-to-peak) of 1.9 G.

(10 000 units/mL), and 100 µL of PP-H solution were added. The respiratory burst was introduced by an addition of 100 µL of zymosan (1 mg/mL). Prior to addition to the cell culture wells, the freshly prepared 5 mM PP-H solution was sterile filtered through 0.2-µm filters (Corning Glass Works, Corning, NY). Aliquots (5 µL) were withdrawn from the culture wells at different time points and analyzed by ESR. Cyt c was used to validate superoxide production and was performed using similar sample collection procedures, in which only the PP-H solution was replaced with a solution containing 640 µM cyt c. One-milliliter cell culture media samples containing cyt c were obtained for UV analysis. RESULTS LGR Performance. Less than 5 µL of dialysate sample was sufficient to fill the 1.4-µL volume of the quartz capillary cells used for these studies. The LOD for this microliter ESR assay was determined using TEMPOL standards. TEMPOL is an appropriate standard for PP• measurements since both radicals exhibit threeline spectra with similar hyperfine splittings. The LOD and LOQ were determined to be 0.26 and 0.82 µM, respectively, at the 98% confidence level. Thus, a minimum of 0.36-pmol spins are detectable in a single scan in the 1.4-µL volume. PP• Nitroxide Production from PP-H/X/XO Mixture. An in vitro X/XO system provided sustained O2•- generation via oxidation of xanthine to uric acid.40 Figure 2 shows the characteristic three-line ESR signal observed using the loop-gap resonator for PP• nitroxide after the reaction of PP-H with O2•generated from the X/XO system. The absence of the PP• signal in various controls including the PP-H and PP-H/XO mixture indicates that PP• was only formed from the reaction between PP-H and XO-produced superoxide. Addition of SOD to the PPH/XO mixture quenched the PP• signal, demonstrating that the observed PP• signal was O2•- dependent. LC/MS experiments confirmed the production of PP• from PP-H/X/XO mixtures. PP• has a molecular weight of 252, and the expected m/z for the ESI positive and negative modes is 253 (40) Fridovich, I. In CRC Handbook of Methods for Oxygen Radical Research; Greenwald, R. A., Ed.; CRC Press: Boca Raton, FL, 1985; pp 51-53.

Analytical Chemistry, Vol. 76, No. 16, August 15, 2004

4737

Figure 3. Uric acid produced (9), cytochrome c reduced (O), and PP• accumulated (2) as a function of time after initiation of xanthine oxidation by XO. In all cases, the samples contained 50 µM xanthine and 0.02 unit of XO. 20 µM cyt c or 500 µM PP-H was added to the XO samples.

and 251, respectively. Using ESI in the negative mode, mass peaks of m/z 252 (PP-H) and 251 (PP•) were identified in the LC/MS measurements of the PP-H/X/XO mixture. The PP• ESR signal increase as a function of X/XO reaction time was compared to cyt c reduction experiments. Figure 3 shows the time-dependent uric acid generation, cyt c reduction, and PP• production. The absorbance change resulting from uric acid generation indicates a continuous xanthine turnover coupled with superoxide and hydrogen peroxide generation. Cyt c reduction monitored at 550 nm confirmed O2•- generation. To confirm PP• production, aliquots were withdrawn at different times and measured by ESR. The accumulation kinetics of PP• is consistent with cyt c reduction kinetics. During the first 3 min, uric acid was produced at a rate of 3.50 µM/min, whereas the O2•- generation rate determined from both the cyt c reduction and PP• accumulation was 1.64 µM/min. Uric acid is produced during the addition and mixing process and thus has a nonzero concentration when placed into the spectrometer. These production rates are consistent with those reported by Fridovich.41 Microdialysis Sampling. Figure 4 shows the detected PP• concentrations when 5 mM PP-H was delivered through the microdialysis probe at 0.5 µL/min to a well-stirred X/XO medium. Aliquots containing PP• removed from the X/XO medium (Figure 1A) at specified times were compared to dialysates as shown in Figure 4. Approximately 2 µM PP• nitroxide background signal was observed in the 5 mM PP-H perfusion fluid. This background signal represents ∼0.04 mol %, which is not atypical for an ESR spin trap. The PP• concentration in the outside medium increased during the first 20 min and plateaued. The PP• concentration in the collected dialysates also increased for the first 20 min. However, after the superoxide generation in the external medium reached its plateau, the dialysate PP• concentration continued to increase. Control experiments were performed with a mixture containing only xanthine oxidase external to the probe. The results of the controls are plotted as open symbols in Figure 4. In these control experiments, the dialysate PP• concentrations increased after 30 min. This correlated well with the concentration increases (41) Fridovich, I. J. Biol. Chem. 1970, 245, 4053-4057.

4738 Analytical Chemistry, Vol. 76, No. 16, August 15, 2004

Figure 4. PP• collected in dialysates and outside medium samples when 5 mM PP-H was delivered through a microdialysis probe at 0.5 µL/min. The external solution was well stirred. Closed symbols represent the results for three replicate runs when outside solution contained X/XO mixture. Open symbols represent the control experiments results when xanthine was absent. (9) PP• collected in the dialysates; (2) PP• in outside medium. (0) PP• collected in the control dialysates; (4) PP• in the control outside medium.

Figure 5. PP• collected in dialysates and outside medium samples for three replicate runs when 0.5 mM PP-H/X/XO mixture was external to a microdialysis probe and well stirred. Saline solution was perfused at 1.0 (closed symbols) and 0.5 µL/min (open symbols). (9) 1.0 µL/min dialysates; (2) outside medium for 1.0 µL/min perfusion experiments; (0) 0.5 µL/min dialysates; (4) outside medium for 0.5 µL/min perfusion experiments.

observed in the previous experiment. A possible explanation for these observations is that PP-H either reacts with oxygen that may have entered the airtight syringe or is reacting with other components of the microdialysis system. PP-H can also be included in the X/XO medium external to the microdialysis probe as shown in Figure 1. In these experiments, a 10-fold lower concentration of PP-H is sufficient since the PP• diffusion path is only back to the probe. Figure 5 shows the dialysate PP• concentrations from two different flow rates as compared to the external PP• concentration. With 0.5 mM PP-H included in the external medium, the detectable background PP• concentration was less than the LOD. When the probe was immersed in the PP-H/X/XO medium, the dialysate PP• concentrations increased with time and stabilized. These data indicate that O2•- generation occurred external to the probe. As expected, the slower perfusion flow rate resulted in a higher relative recovery. The final dialysate PP• concentration collected was

Table 1. PP• Concentration in 5-µL Aliquots Taken from the Cell Medium of Stimulated Macrophages at Different Times vs SOD Control cell treatment

1h

1.5 h

RAW cell + IFN-g/zymosan +PP-H (µM) SOD control (µM)

0.68 nda

1.09 0.68

a

Not detected.

Figure 6. PP• collected in dialysates and outside medium samples for three replicate runs when 0.5 mM PP-H/X/XO was external to a microdialysis probe and not stirred. Saline solution was perfused at 0.5 µL/min. (9) dialysates; (2) outside medium.

control was probably due to oxidation by solution oxygen since peroxynitrite is not produced with this stimulation protocol.44 When the produced PP• in the cell culture medium was monitored at defined time points, the PP• was stable for hours under these experimental conditions and exhibited concentrations a concentration of 1.95 ( 0.09 µM, with an RSD of 4.5% (n ) 7) over a 3.5-h time period.

compared with the outside sample PP• concentration to calculate the relative recovery (RR). The RR was 22.1 ( 1.1 and 57.2 ( 5.7% at perfusion flow rates of 1.0 and 0.5 µL/min, respectively. Figure 6 shows the PP• concentrations at selected time points for three replicate runs with the probe immersed in a quiescent PP-H/X/XO solution. In the quiescent medium, the PP• production rates were decreased resulting in lower PP• concentrations as compared to those obtained from the well-stirred medium. Lower microdialysis PP• relative recovery would be expected for an external quiescent medium, since a quiescent medium provides a greater mass transport resistance than a well-stirred medium.23 Similarly, due to the mass transport resistances from tissue, a lower Ed would be expected for in vivo studies with PP-H. However, the PP• concentration detected in the outside medium should be similar to that of the well-stirred case. To ensure changes in the enzyme activity did not occur during the course of the experiment, a freshly prepared 2 units/mL XO was used in the third run; the results were similar to those for the first and second runs. Since the dialysis-detected PP• production rate for the quiescent medium is reduced relative to that for the wellstirred medium, it is expected that the limiting PP• concentration may be similar if a longer monitoring time were used. However, conversion of xanthine to uric acid by xanthine oxidase requires oxygen as a substrate. Thus, differences in the solution oxygen concentrations between the stirred and unstirred experiments may account for the differences in the PP• production rates. O2•- Generation in Stimulated Macrophages. A cell culture system is an appropriate model for transition in complexity from the in vitro, well-characterized enzyme mixture. When treated with IFN-γ/zymosan, RAW 264.7 macrophages are known to give a respiratory burst.42,43 In a preliminary experiment, PP-H was added together with zymosan to IFN-γ-primed macrophages. PP• accumulation was monitored at various time points from 30 min to 3 h (Table 1). The 2-fold increase in detected PP• from a 5-µL aliquot versus a SOD-containing control at 1 h demonstrated that O2•- was generated. The subsequent production of PP• in the SOD

DISCUSSION Microdialysis sampling combined with ESR O2•- detection with improved temporal resolution has been achieved by using a loopgap resonator in conjunction with reaction with PP-H. Using low molecular weight spin scavengers such as PP-H or the nitrone spin traps (DMPO, DEMPO, etc.) has significant advantages over the standard cyt c approach for microdialysis collection of O2•-. Cyt c has a MW greater than 12 000 and therefore a low diffusion coefficient,45 which limits its diffusion out of the probe and subsequent reduction by O2•- external to the probe. As a result, there is extraneous O2•- dismutation accompanying the slower diffusion through the sample medium and dialysis membrane. Microdialysis sampling combined with ESR detection provides direct proof of free radicals in the localized environment surrounding the microdialysis probe. As demonstrated above, superoxide generation by the X/XO system has been monitored by microdialysis sampling using capillary ESR cells and the PP-H scavenging agent. Because of the importance of superoxide in oxidative stress, it is necessary to have the tools available that will allow in vivo measurements. The stability of PP• not only makes microdialysis sampling experiments feasible but also facilitates the data interpretation. Accumulation of PP• reflects the actual production of radicals during the sampling interval because PP• is stable and not prone to degradation. However, if conventional nitrone spin traps, e.g., DEPMPO, are used for O2•- trapping, the DEPMPO-OOH adduct production kinetics are coupled to its subsequent decay process. Kinetic modeling is then required for accurate interpretation using the nitrone spin traps. ESR measurements of free-radical-containing microdialysis samples are not common. The volume requirement for ESR measurement is generally satisfied with high perfusion flow rates (µL/min), which results in low microdialysis relative recovery.23 Alternatively, microdialysis samples collected at lower perfusion flow rates can be diluted, which results in ESR signals with limited signal-to-noise ratio.20,22 The temporal resolution for both approaches is typically 20-30 min. However, less than 5 µL of

(42) Pfeiffer, S.; Lass, A.; Schmidt, K.; Mayer, B. J. Biol. Chem. 2001, 276, 34051-34058. (43) Loegering, D. J.; Lennartz, M. R. Inflammation 2004, 28, 23-31.

(44) Alvarez, M. N.; Trujillo, M.; Radi, R. Methods Enzymol. 2002, 359, 353366. (45) Gupte, S. S.; Hackenbrock, C. R. J. Biol. Chem. 1988, 263, 5241-5247.

Analytical Chemistry, Vol. 76, No. 16, August 15, 2004

4739

undiluted dialysate has been sensitively measured using the LGR. Therefore, there is no sensitivity loss due to dilution and the temporal resolution is decreased to 10 min for the 0.5 µL/min perfusion rate. Capillary cells coupled with the LGR allow sensitive detection of microliter samples collected by microdialysis with an LOD of 0.36-pmol spins in the 1.4-µL samples. This demonstrates that highly sensitive ESR assays for microliter samples are feasible utilizing the loop-gap resonator. In addition to the sensitive detection of microliter samples, the low quality factor of LGR makes it less sensitive to noise, vibration, and mechanical movement of samples during fluid flow. These features make the LGR ideal for the implementation of continuous-flow microdialysis ESR measurements.46 Potentially, other free radicals can be measured even if the radical adducts are not as stable as PP•. A 10-fold higher concentration of PP-H was needed for the delivery experiments described above as compared to the recovery experiments. This larger PP-H concentration is required to obtain quantifiable PP• because during the delivery experiments formed PP• can either diffuse away from the probe or into the probe.24 The adventitious oxidation of the PP-H to PP• is a concern for microdialysis delivery experiments. However, PP-H has been reported to show much less background oxidation than a similar hydroxylamine spin scavenging agent, 1-hydroxy-3-carboxyoxopyrrolidine, in blood samples.30 Rosen et al. have reported adding chelating agents such as DTPA to the hydroxylamine, 2-ethyl-1hydroxy-2,5,5-trimethyl-3-oxazolidine (OXANOH) to prevent OXANOH from being oxidized to its corresponding nitroxide, 2-ethyl-2,5,5-trimethyl-3-oxazolidinoxyl.29 It should be noted that free radical sampling with microdialysis sampling is not trivial, and nonspecific production of free radicals has been previously reported.47 In the microdialysis delivery experiments, the necessary precautions were taken to reduce PP-H oxidation. The negligible presence of transition metal in the buffers was determined using a well-known ascorbic acid assay that we have previously described in our work with hydroxyl radical.25,48 A chelating agent (200 µM DTPA) was always present in the X/XO mixture. When 5 mM PP-H was delivered through the microdialysis probe, the (46) Hubbell, W. L.; Froncisz, W.; Hyde, J. S. Rev. Sci. Instrum. 1987, 58, 18791886. (47) Montgomery, J.; Ste-Marie, L.; Boismenu, D.; Vachon, L. Free Radical Biol. Med. 1995, 19, 927-933. (48) Buettner, G. R. Methods Enzymol. 1990, 186, 125-127. (49) Dikalov, S.; Skatchkov, M.; Bassenge, E. Biochem. Biophys. Res. Commun. 1997, 231, 701-704.

4740

Analytical Chemistry, Vol. 76, No. 16, August 15, 2004

oxidation was observed initially at 30 min and the rate for the following 30 min was ∼0.03 µM/min, which is less than that reported by Rosen for a similar hydroxylamine.29 These results suggest that microdialysis sampling can be used to deliver the PP-H to detect superoxide in a localized environment surrounding the probe, but caution is warranted since the high concentration of PP-H required can result in adventitious oxidation resulting in an increased ESR signal. Comparison of the nitroxide production with SOD-containing controls can be used to compensate for nonspecific PP-H oxidation. However, this option for reducing nonspecific PP• production is not possible during in vivo microdialysis experiments. Localized delivery of agents from a microdialysis probe followed by collection of products is a complex mass transport process.24 The recovery and delivery experiments with PP-H were originally performed in stirred media since our previous experience HO• trapping using 4-hydroxybenzoic acid and the X/XO enzyme system in a quiescent medium resulted in poor reproducibility among three replicate experiments.25 The reproducibility for O2•- detection using PP-H was significantly improved as compared to HO• trapping. The reasons for these differences are likely caused by the lower reactivity of O2•- as compared to HO• toward different components in the in vitro system. This greatly improved reproducibility bodes well for studies in cell culture or in vivo experiments since stirring is impossible in these sample media. However, it is important to note that while PP-H does not react with hydrogen peroxide and is resistant to ascorbic acid and thiols, there is a specificity concern for this methodology since peroxynitrite may be produced in activated macrophages.49 ACKNOWLEDGMENT The National Science Foundation (CHE-9984150), National Institutes of Health (EB001441), and the RPI Office of Research funded this work (J.A.S.). The LC/MS was purchased by a grant from the National Science Foundation (CRIF-0091892). Modification to the electron spin resonance spectrometer was supported by NIH Grant NS042915 (J.T.W.). We acknowledge our colleagues, Professors Daniel Loegering and Michelle Lennartz, Albany Medical College, for providing the macrophage cell line and for helpful comments regarding this paper. Received for review December 31, 2003. Accepted May 25, 2004. AC035543G