The Detection of Hydroxyl Radicals in Vivo

Available online at www.sciencedirect.com InorganicJOURNAL OF Biochemistry

Journal of Inorganic Biochemistry 102 (2008) 1329–1333 www.elsevier.com/locate/jinorgbio

The detection of hydroxyl radicals in vivo

Wolfhardt Freinbichler b, Loria Bianchi a,1, M. Alessandra Colivicchi a, Chiara Ballini a, Keith F. Tipton a,2, Wolfgang Linert b, Laura Della Corte a,*

a Dipartimento di Farmacologia Preclinica e Clinica M. Aiazzi Mancini, Universita` degli Studi di Firenze, Viale G. Pieraccini 6, 50139 Firenze, Italy b Institute for Applied Synthetic Chemistry, Vienna University of Technology, Getreidemarkt 9/163-AC, A-1060 Vienna, Austria

Received 13 September 2007; received in revised form 12 November 2007; accepted 14 December 2007 Available online 28 December 2007

Abstract

Several indirect methods have been developed for the detection and quantification of highly reactive oxygen species (hROS), which may exist either as free hydroxyl radicals, bound ‘‘crypto” radicals or Fe(IV)-oxo species, in vivo. This review discusses the strengths and weaknesses associated with those most commonly used, which determine the hydroxylation of salicylate or phenylalanine. Chemical as well as biological arguments indicate that neither the hydroxylation of salicylate nor that of phenylalanine can guarantee an accurate hydroxyl radical quantitation in vivo. This is because not all hydroxylated product-species can be used for detection and the ratio of these species strongly depends on the chemical environment and on the reaction time. Furthermore, at least in the case of salicylate, the high concentrations of the chemical trap required (mM) are known to influence biological processes associated with oxidative stress. Two, newer, alternative methods described, the 4-hydroxy benzoic acid (4-HBA) and the terephthalate (TA) assays, do not have these drawbacks. In each case reaction with hROS leads to only one hydroxylated product. Thus, from a chemical viewpoint, they should provide a better hROS quantitation. Further work is needed to assess any possible biological effects of the required millimolar (4-HBA) and micromolar (TA) concentrations of the chemical traps. Ó 2007 Elsevier Inc. All rights reserved.

Keywords: Hydroxyl radicals; Salicylate; Phenylalanine; 4-Hydroxybenzoic acid; Terephthalate

1. Introduction neurodegenerative diseases, ischemic or traumatic brain injuries, cancer, diabetes liver injury and AIDS [4–6]. Oxidative stress is defined as an imbalance between the Moreover, the possibility that oxidative stress plays an production of reactive oxygen species (ROS) and a biolog- important role in the ageing process is under discussion ical system’s ability to readily detoxify the reactive interme- [7–9]. However, it has been difficult to distinguish whether diates and/or easily repair the resulting damage. The term oxidative stress causes the pathologies or is itself a conse- ROS covers several substances, ranging from the rather quence of them. The problematic nature of investigating À unreactive, H2O2 through O2 and singlet O2 to the highly this question may be illustrated by taking the example of reactive oxygen species (hROS), which may exist as free ageing. There is no doubt that if an organism gets older hydroxyl radicals (HOÅ), as bound (‘‘crypto”) radicals or its ability to repair DNA and protein damage decreases, as Fe(IV)-oxo species [1–3]. The occurrence of oxidative whereas the concentration of ROS increases. This observa- stress correlates with numerous pathologies including tion led to the formulation of the ‘‘radical theory of aging”, which predicts that increasing the concentrations of antioxidants or blocking ROS generation should result in * Corresponding author. Tel.: +39 055 4271226; fax: +39 055 410778. increased lifespan. Most studies have, necessarily, involved E-mail address: laura.dellacorte@unifi.it (L.D. Corte). 1 Permanent address: Azienda USL 3 di Pistoia, 51100 Pistoia, Italy. laboratory models and uncertainties in extrapolation to 2 Permanent address: Department of Biochemistry, Trinity College, human aging compound the problems. However, although Dublin 2, Ireland. there have been several reports, mostly with invertebrates,

0162-0134/$ - see front matter Ó 2007 Elsevier Inc. All rights reserved. doi:10.1016/j.jinorgbio.2007.12.017 1330 W. Freinbichler et al. / Journal of Inorganic Biochemistry 102 (2008) 1329–1333 that an enhanced production of antioxidants by genetic the derivatives to be reliable. However, as shown in Table manipulation significantly increases lifespan [10–12], the 1 and Fig. 1, this requirement is far from being fulfilled [20– results of similar studies with mice have been largely disap- 22]. The product ratio shows time dependence and there is pointing, and did not show a significant correlation also a strong influence of the chemical environment. As the between an enhanced production of ROS and ageing [13– ratio between 2,3-DHBA and 2,5-DHBA varies from 5:1 to 16]. Similarly, unclear results can also be found in research 1:1, the resulting quantitation error is far from being neg- reports on the possible relation between oxidative stress ligible. 2,3-DHBA is also quite unstable and, in common and neurodegenerative diseases, such as Alzheimer’s and with the other species, is quickly metabolised [18,23]. Parkinson’s diseases. Although some ROS species are believed to have specific Table 1 biological roles, many studies have been based on the Yields of 2,3-DHBA and 2,5-DHBA, and ratios of 2,3-DHBA:2,5-DHBA reported for different conditions of hROS generation extrapolation of superoxide or H2O2 data to hROS behav- iour. This may have led to contradictory results. One Method 2,3-DHBA 2,5-DHBA Ratio Reference important reason for the use of such experimental proce- (lM) (lM) dures is the lack of a simple and reliable in vivo method Radiolysisa 105 20 5.25 [22] for hROS quantitation. This account presents a brief over- Radiolysisa (1 mM 100 75 1.33 [22] view of the most common and also some newer methods Fe(III)EDTA added) Fenton systemb 22 4 5.5 [22] and their advantages and disadvantages. Fenton systemb (100 lM 25 9.5 2.6 [22] Fe(III)EDTA added) b 2. Methods for hROS quantitation Fenton system (1.3 mM 52 44 1.2 [22] Fe(III)EDTA added) Photochemical reduction of 1 [20] The only direct method of detecting hROS is electron c Fe(III) in O2 sat. solution spin resonance (ESR) spectroscopy, including the spin (0.3 mM K2C2O4) trapping technique. However, its application to in vivo Photochemical reduction of 1.7 [20] c experiments, in particular in freely moving animals, is Fe(III) in O2 sat. solution impracticable because of technical difficulties, which (3 mM K2C2O4) a include high disturbing noise levels and low sensitivity The yields were measured after a radiation dose of 600 Gy, buffered [17]. All the common methods used for in vivo experiments with 5 mM phosphate, pH 7.5, under N2O; initial concentration of salicylic acid = 1 mM. are indirect and based on the hydroxylation of aromatic b De-aerated solutions of 200 lM Fe(II)EDTA were mixed with 200 lM compounds. H2O2, buffered at pH 7.5 with 5 mM phosphate; initial concentration of salicylic acid = 1 mM. c Initial concentrations: 0.8 mM salicylic acid; 0.06 mM; Fe(III). 2.1. Salicylate

The most commonly used procedure is based on the hydroxylation of salicylic acid, which is based on the chemical reaction shown in Scheme 1. This yields three reaction products, I catechol, II 2,5-dihydroxybenzoic acid (2,5- DHBA) and III 2,3-dihydroxybenzoic acid (2,3-DHBA). By measuring all three hydroxylation products the assay may allow a simple and accurate quantitative detection of hROS by either electrochemical or photometric methods. As catechol is formed only in small amounts, only species 2,5-DHBA and 2,3-DHBA are normally used for determining the amount of hROS. However, several problems arise under in vivo conditions. 2,5-DHBA is also produced enzy- Fig. 1. Dependence of the ratio of 2,3-DHBA:2,5-DHBA on the reaction matically [18,19], leaving only 2,3-DHBA available for time during the photo-hydroxylation of 0.7 mM salicylic acid by 10 mM detection as a simple hROS product. A constant product hydrogen peroxide. The reaction was sensitised with methylene blue ratio is essential for determinations based on only one of (0.01 mM) in the presence of ferric acetate (adapted from 21).

COOH OH COOH COOH OH OH OH OH + OH + + HO OH

III III

Scheme 1. The hydroxylation of salicylic acid by radical oxidation. W. Freinbichler et al. / Journal of Inorganic Biochemistry 102 (2008) 1329–1333 1331

Furthermore, salicylic acid inhibits cyclooxygenase (EC photometric), it seems unlikely that a coeluted unknown 1.14.99.1), a key enzyme in the pathway of prostaglandin compound was the sole explanation for these results. An synthesis. Products of this pathway have many physiologi- alternative explanation would be that, as in the case of sal- cal functions in both normal and disease conditions [24,25] icylic acid, there is a change in the ratio of hydroxylated and are known to influence inflammatory processes associ- products due to a change in the chemical environment, ated with oxidative stress [26–28]. Thus the biochemical which favours m-tyrosine formation in comparison to o- actions of salicylate itself may perturb the hROS results. tyrosine. Since p-tyrosine cannot be considered for hROS All these considerations raise serious questions about the detection and the above considerations suggest that it is validity of in vivo results obtained using this method [17– doubtful whether if the sum of o-andm-tyrosine can be 19]. taken for an artefact-free hROS quantitation in vivo see [29,34,35], results obtained with this procedure should be 2.2. Phenylalanine treated with caution. Nevertheless, this method has the potential advantage that nitration can be detected simulta- The hydroxylation reaction of phenylalanine, shown in neously, allowing concurrent monitoring of peroxynitrite Scheme 2, has been proposed as an alternative to salicylate, formation [35]. mainly because the administration of phenylalanine has fewer identified side effects [19,29]. 2.3. 4-Hydroxybenzoic acid (4-HBA) The disadvantage of multiple hydroxylation products remains, since reaction of phenylalanine (I) with hROS Detection of hROS with 4-hydroxybenzoic acid (4- yields a mixture of the 2-, 3- and 4-hydroxylated products, HBA) is less complicated, because only one hydroxylation o-tyrosine (II), m-tyrosine (III), and p-tyrosine; (IV). How- product, 3,4-hydroxybenzoic acid (3,4-DHBA), is formed ever, p-tyrosine is an endogenous compound, which is in significant amounts (Scheme 3). Thus, from a chemical formed from L-phenylalanine by the enzyme phenylalanine viewpoint, this method should provide a much more reli- hydroxylase (EC 1.14.16.1) and, therefore, cannot be used able possibility for hROS quantitation in vitro. It has been for quantitation of hROS. Since D-phenylalanine is not a used in microdialysis experiments, using HPLC separation substrate for this enzymatic hydroxylation, it has been pro- with electrochemical detection [36–38]. Although 4-HBA posed as a more suitable compound for hROS detection can be hydroxylated by monooxygenases from some micro- [30,31], although p-tyrosine still cannot be used for hROS organisms, this apparently does not occur in mammals, determination unless a detection procedure that can distin- suggesting its suitability as that it is a suitable chemical trap guish between the D- and L-enantiomers is used. An addi- for hROS determination in vivo. As 4-HBA is an endoge- tional potential problem is that D-phenylalanine is a nous compound that shows little or no apparent toxicity, substrate for D-amino acid oxidase (EC 1.4.3.3), which is it has been claimed that it could be also used for human present in brain [32] and produces H2O2 as a product. studies [37]. However, it is necessary to use high (mM Chemokinetic studies have shown hydroxylation of range) concentrations of 4-HBA and further work is neces- phenylalanine to be rather slow [19,29], and the amount sary to assess whether any biochemical processes are of hROS required to overcome the detection limit, about affected by 4-HBA, or its metabolites, at such levels. 50 nM [19,33], is rarely present in tissues, even in abnormal conditions. Furthermore, one, or more, unknown compound(s) were reported to coelute with m-tyrosine in ische- O O mia-reperfusion experiments that used HPLC analysis with OH OH electrochemical detection [29,34]. This results in an appar- +OH ently greater increase of m-tyrosine, in comparison to o- HO HO tyrosine, which has also been reported by others, with e.g. 6-hydroxydopamine as hROS generator [35]. Since OH similar results were obtained by using different experimen- Scheme 3. The hydroxylation of 4-hydroxybenzoic acid by radical tal procedures and detection methods (electrochemical and oxidation.

O OH O O O

OH OH OH OH + +

NH NH2 NH NH2 . 2 2 OH HO

OH III III IV

Scheme 2. The hydroxylation of L-phenylalanine by radical oxidation. 1332 W. Freinbichler et al. / Journal of Inorganic Biochemistry 102 (2008) 1329–1333

COO- COO- phenylalanine can provide an artefact-free quantitation of hROS. The two newer methods, which use either 4-HBA OH or TA as the trap, do not have many of these drawbacks. + OH The sensitivity, selectivity and ease of detection, together with its apparent biochemical inertness, make terephthalate COO- COO- an attractive choice for hROS detection in vivo. However, further studies will be necessary to ensure that determina- Scheme 4. The hydroxylation of terephthalic acid by radical oxidation. tions with terephthalate are, indeed, free from artefacts. As human brain microdialysis has become acceptable diag- 2.4. Terephthalic acid (TA) nostic studies (see e.g. [47,48]), the availability of a safe and reliable procedure for monitoring hROS production may The hydroxylation reaction (Scheme 4) of TA is specific have great potential value, especially for investigating mod- for hROS, as it requires the high redox potentials of OH- els of neurodegenerative diseases involving hROS forma- radicals or ferryl species (>1.6 V) [39] and has no signifi- tion [49,50]. cant reactivity with other ROS, such as superoxide, ROOÅÀ 4. Abbreviations or H2O2 [17,18,40–42]. This reaction has been used for several years as a dosim- eter in radiolysis experiments [18,40–42,43] because of its DHBA dihydroxybenzoic acid ease of use and its good performance, especially in the 4-HBA 4-hydroxybenzoic acid low dose range [42]. A major advantage of this system is HPLC high-performance liquid chromatography that, apart from some minor fragmentation products, hROS highly-reactive oxygen species which occur in all aromatic hydroxylation processes, the OH-TA 2-hydroxyterephthalate symmetry of the molecule leads to only one hydroxylated ROS reactive oxygen species product, 2-hydroxyterephthalate (OH-TA). This hydroxyl- TA terephthalic acid ated product, which is stable, is highly fluorescent, whereas TA is practically non-fluorescent. This has led to the sug- gestion that this system could be used in vivo [40] and Acknowledgements Yan et al. reported its application for detecting hROS during brain microdialysis in sheep [19]. In contrast to the Thanks for financial support are due to the ‘‘Fonds zur other systems described above, neither the hROS trap Fo¨rderung der Wissenschaftlichen Forschung in O¨ sterr- (TA) nor the hydroxylated product (OH-TA) is endoge- eich” (Project 19335-N17), Ente Cassa di Risparmio di nously present or is metabolized. Moreover, TA has been Firenze (Firenze, Italy), Foundations MPS (Siena, Italy) reported to be non-toxic and not accumulative [44,45]. and ERAB (Brussels, Belgium) and to the EU COST ac- The fluorescence of OH-TA (excitation wavelength tion D34 supporting our international cooperation. 315 nm; emission wavelength 435 nm) was found to be lin- References early dependent on the concentration of H2O2 added to a model Fenton-reaction systems, and independent of the initial TA concentration (range 250–1000 lM). The high [1] J.D. Rush, W.H. Koppenol, J. Am. Chem. Soc. 110 (1988) 4957– 4963. fluorescence of OH-TA allows its detection at concentra- [2] C. Walling, Acc. Chem. Res. 8 (1975) 125–131. tions as low as 0.5 nM in vitro. This high sensitivity and [3] R.V. Lloyd, Free Radical Biol. Med. 22 (1977) 885–888. the reactivity of TA with hROS allow relatively low con- [4] B. Halliwell, Drugs Aging 18 (2001) 685–716. centrations of TA to be used for reliable detection of hROS [5] J. Moskovitz, M.B. Yim, P.B. Chock, Arch. Biochem. Biophys. 397 formation under in vivo conditions. In microdialysis exper- (2002) 354–359. [6] G.W. Pace, C.D. Leaf, Free Radical Biol. Med. 19 (1995) 523–528. iments in vivo, where 250 lM TA was perfused and the [7] D. Harman, J. Gerontol. 11 (1956) 298–300. OH-TA formed was resolved by HPLC, the detection limit [8] D. Harman, Proc. Natl. Acad. Sci. USA 78 (1981) 7124–7128. was 0.5 fmol/ll of effluent sample. At this concentration [9] K.B. Beckman, B.N. Ames, Physiol. Rev. 78 (1998) 547–581. there was no evidence for toxicity and it did not evoke [10] W.C. Orr, R.S. Sohal, Science 263 (1994) 1128–1130. amino-acid neurotransmitter release [46]. [11] W.C. Orr, R.J. Mockett, J.J. Benes, R.S. Sohal, J. Biol. Chem. 278 (2003) 26418–26422. [12] T.L. Parkes, A.J. Elia, D. Dickinson, A.J. Hilliker, J.P. Phillips, G.L. Boulianne, Nat. Genet. 19 (1998) 171–174. 3. Conclusions [13] R.S. Sohal, R.J. Mockett, W.C. Orr, Free Radical Biol. Med. 33 (2002) 575–586. Several trapping compounds have been developed for [14] J.P. de Magalhaes, A. Sandberg, Mech. Ageing Dev. 126 (2005) 1026– determining hROS production in vivo microdialysis 1033. [15] E.S. Schriner, N.J. Linford, G.M. Martin, P. Treuting, E. Charles, systems. These have sometimes been used with insufficient C.E. Ogburn, M. Emond, P.E. Coskun, W. Ladiges, N. Wolf, H.v. experimental validation. Both chemical and biochemical Remmen, D.C. Wallace, P.S. Rabinovitch, Science 308 (2005) 1909– considerations indicate that neither salicylate nor the 1911. W. Freinbichler et al. / Journal of Inorganic Biochemistry 102 (2008) 1329–1333 1333

[16] J.E. Fleming, I. Reveillaud, A. Niedzwiecki, Mutat. Res. 275 (1992) [32] S. Moreno, R. Nardacci, A. Cimini, M.P. Ceru, J. Neurocytol. 28 267–279. (1999) 169–185. [17] L. Linxiang, Y. Abe, Y. Nagasawa, R. Kudo, N. Usui, K. Imai, T. [33] J.Z. Sun, H. Kaur, B. Halliwell, X.Y. Li, R. Bolli, Circ. Res. 73 (1993) Mashino, M. Mochizuki, N. Miyata, Biomed. Chromatogr. 18 (2004) 534–549. 470–474. [34] S. Reddy, B. Halliwell, A.D. Jones, J.C. Longhurst, Free Radical [18] V.M. Mishin, P.E. Thomas, Biochem. Pharmacol. 68 (2004) 747–752. Biol. Med. 27 (1999) 1465. [19] E.B. Yan, J.K. Unthank, M. Castillo-Melendez, S.L. Miller, S.J. [35] B. Ferger, C. Themann, S. Rose, B. Halliwell, P. Jenner, J. Langford, D.W. Walker, J. Appl. Physiol. 98 (2005) 2304–2310. Neurochem. 78 (2001) 509–514. [20] K. Lang, D.M. Wagnerova, J. Brodilova, Collect. Czech. Chem. [36] M. Liu, S. Liu, S. Peterson, M. Miyake, K.J. Liu, Mol. Cell. Biochem. Commun. 59 (1994) 2447–2453. 234 (2002) 379–385. [21] S. Lunak, J. Muzart, J. Brodilova, Collect. Czech. Chem. Commun. [37] N. Marklund, F. Clausen, T. Lewander, L. Hillered, J. Neurotraum. 59 (1994) 905–912. 18 (11) (2001) 1217–1227. [22] Z. Maskos, J.D. Rush, W.H. Koppenol, Free Radical Biol. Med. 8 [38] D.V. Matteoa, M. Pieruccia, D.G. Giovanni, A.D. Santoa, A. Poggia, (1990) 153–162. A. Benigno, E. Esposito, Brain Res. 1095 (2006) 167–177. [23] M. Strolin Benedetti, G. Brogin, M. Bani, F. Oesch, J.G. Hengstler, [39] W.H. Koppenol, J.F. Liebman, J. Phys. Chem. 88 (1984) 99–101. Xenobiotica 29 (1999) 1171–1180. [40] M. Saran, K.H. Summer, Free Radical Res. 31 (1999) 429–436. [24] H.C. Salzberg-Brenhouse, E.Y. Chen, D.F. Emerich, S. Baldwin, K. [41] X. Fang, G. Mark, C. von Sonntag, Ultrason. Sonochem. 3 (1996) Hogeland, S. Ranelli, D. Lafreniere, B. Perdomo, L. Novak, T. 57–63. Kladis, K. Fu, A.S. Basile, J.H. Kordower, R.T. Bartus, J. Pharma- [42] R.W. Matthews, Radiat. Res. 83 (1980) 27–41. col. Exp. Ther. 306 (2003) 218–228. [43] X.-F. Yang, X.-Q. Guo, Analyst 126 (2001) 928–932. [25] T. Wang, L. Qin, B. Liu, Y. Liu, B. Wilson, T.E. Eling, R. [44] W.T. Muse, J.S. Anthony, J.D. Bergmann, D.C. Burnett, C.L. Langenbach, S. Taniura, J.S. Hong, J. Neurochem. 88 (2004) 939– Crouse, B.P. Gaviola, S.A. Thomson, Drug Chem. Toxicol. 20 (1997) 947. 293–302. [26] A. Catania, J. Arnold, A. Macaluso, M.E. Hiltz, J.M. Lipton, Proc. [45] OECD SIDS, Terephthalic acid (TPA): Initial Assessment Report; Natl. Acad. Sci. USA 88 (1991) 8544–8547. For 12th SIAM (Paris, France June 2001), United Nations Environ- [27] R.J. Dinis-Oliveira, C. Sousa, F. Remia˜o, J.A. Duarte, A. Sa´nchez ment Programme (UNEP) Publications, Nairobi, Kenya, 2001. Navarro, M.L. Bastos, F. Carvalho, Free Radical Biol. Med. 42 [46] W. Freinbichler, M.A. Colivicchi, M. Fattori, K.F. Tipton, W. (2007) 1017–1028. Linert, L. Della Corte, J. Neurochem., doi:10.1111/j.1471- [28] M. Grilli, M. Pizzi, M. Memo, P.F. Spano, Science 274 (1996) 1383– 4159.2007.05168.x. 1385. [47] M. Raiteri, Pharmacol. Rev. 58 (2006) 162–193. [29] C. Themann, P. Teismann, K. Kuschinsky, B. Ferger, J. Neurosci. [48] O. Rooyackers, A. Thorell, J. Nygren, O. Ljungqvist, Curr. Opin. Meth. 108 (2001) 57–64. Clin. Nutr. Metab. Care 7 (2004) 515–521. [30] H. Kaur, B. Halliwell, Anal. Biochem. 220 (1994) 11–15. [49] W. Linert, M.H. Bridge, M. Huber, K.B. Bjugstad, S. Grossman, [31] R. Biondi, Y. Xia, R. Rossi, N. Paolocci, G. Ambrosio, J.L. Zweier, G.W. Arendash, Biochim. Biophys. Acta 1454 (1999) 143–152. Anal. Biochem. 290 (2001) 138–145. [50] W. Linert, G.N.L. Jameson, J. Inorg. Biochem. 79 (2000) 1–4.