Reactivity of redox sensitive paramagnetic nitroxyl contrast agents with reactive oxygen species

The reactivity of nitroxyl free radicals, 4-hydroxyl-2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPOL) and 3-carbamoyl-2,2,5,5-tetramethylpyrrolidine-N-oxyl (CmP), with reactive oxygen species (ROS) were compared as typical 6-membered and 5-membered ring nitroxyl compounds, respectively. The reactivity of the hydroxylamine forms of both these nitroxyl radicals (TEMPOL-H and CmP-H) was also assessed. Two free radical species of ROS, hydroxyl radical (•OH) and superoxide (O2•−), were subjected to a competing reaction. •OH was generated by UV irradiation from an aqueous H2O2 solution (H2O2-UV system), and O2•− was generated by a reaction between hypoxanthine and xanthine oxidase (HX-XO system). •OH and O2•− generated by the H2O2-UV and HX-XO systems, respectively, were measured by electron paramagnetic resonance (EPR) spin-trapping, and the amount of spin adducts generated by each system was adjusted to be equal. The time courses of the one-electron oxidation of TEMPOL, CmP, TEMPOL-H, and CmP-H in each ROS generation system were compared. A greater amount of TEMPOL was oxidized in the HX-XO system compared with the H2O2-UV system, whereas the reverse was observed for CmP. Although the hydroxylamine forms of the tested nitroxyl radicals were oxidized evenly in the H2O2-UV and HX-XO systems, the amount of oxidized CmP-H was approximately 3 times greater compared with TEMPOL-H.

Introduction N itroxyl radicals can undergo redox transformations and form hydroxylamine via oxoammonium cations or one-electron reductions (Fig. 1). The paramagnetic nitroxyl radical can be converted to form diamagnetic hydroxylamine by reduction or via the oxoammonium cation in vivo. Relatively strong reductants, such as ascorbic acid, can directly reduce nitroxyl radicals by one electron reductions to form hydroxylamine. The nitroxyl radical is commonly oxidized by oxidants, such as a hydroxyl radical ( • OH), hydroperoxyl radical (HO 2 • ), and/or Fe 3+ ion, to form the oxoammonium cation. The oxoammonium cation can then sequentially receive a hydrogen atom from a coexisting hydrogen-donor (H-donor), such as reduced glutathione (GSH), reduced bnicotinamide adenine dinucleotide (NADH), or reduced bnicotinamide adenine dinucleotide phosphate (NADPH) to form hydroxylamine. This two-step reduction of nitroxyl radicals to hydroxyl amine, i.e., one-electron oxidation followed by twoelectron reduction, can be processed enzymatically in vivo. Hydroxylamines can be slowly oxidized by oxidants, such as • OH, HO 2 • , and/or Fe 3+ ions, to re-form nitroxyl radicals. In addition, GSH may also directly and irreversibly react with oxoammonium cations to form a stable complex. (1) Superoxide dismutase (SOD) mimetic reactions have been reported for 6-membered ring-type nitroxyl radicals. (2,3) The redox pair of nitroxyl radicals and oxoammonium cation states underlie its SOD mimic effects. (4) The nitroxyl radical can be one-electronoxidized by HO 2 • , which exists in equilibrium with superoxide (O 2 •-) in an aqueous environment. Furthermore, the oxoammonium cation can be one-electron-reduced by O 2 •-. This cycling redox reaction is able to function as a SOD mimic.
A common nitroxyl radical, 4-hydroxy-2,2,6,6-tetramethylpiperidine-N-oxyl (TEMPOL), has gained attention as a new class of normal tissue-selective radio-protector. (3,(5)(6)(7)(8) This tissue selectivity is closely associated with the in vivo redox mechanisms of nitroxyl radicals. Although nitroxyl radicals may have direct radioprotective effects, hydroxylamines do not. (3,5) The addition of a free radical form of TEMPOL to cell culture showed radioprotective effects in a dose-dependent manner. (3) In contrast, the addition of hydroxylamine in the form of TEMPOL (TEMPOL-H) did not show any radioprotective effects at 100 mM. (3) However, the administration of TEMPOL-H in mice showed similar radioprotective effects to TEMPOL, (5) because TEMPOL-H is re-oxidized to form TEMPOL in vivo. The concentration of the free radical forms in the blood increased even after the administration of hydroxylamine. (5) The pharmacokinetic curves of the free radical form in blood, after the administration of TEMPOL-H and that of TEMPOL, almost overlapped, 15 min after the administration and later. (5) TEMPOL has been shown to protect C3H mice against wholebody radiation-induced bone marrow failure. (6) Tumor growth curves generated after 10 and 33.3 Gy of radiation showed no difference in growth between TEMPOL-and PBS-treated animals. (7) In another study, TEMPOL was evaluated for potential differences in radiation protection in salivary glands and in tumors using fractionated radiation. (8) Five daily 6 Gy fractions to the head region of mice reduced saliva production by 54%; however, these effects were significantly reduced by 5 daily treatments with TEMPOL. These results suggested that TEMPOL may protect the salivary tissues from radiation. Because TEMPOL treatment had no effect on radiation-induced tumor regrowth, TEMPOL may work as a selective radioprotector for normal tissue. The difference in radioprotection between normal and tumor tissues may be due to TEMPOL being reduced or cleared from the tumor tissue faster than in normal tissues.
Paramagnetic nitroxyl radicals have a proton T 1 -shortening effect and can be a T 1 -weighted magnetic resonance (MR) contrast agent. Mapping in vivo decay rates of CmP in an SCC tumor loaded on a mouse thigh was measured using MR imaging (MRI). (18) The tumor tissue showed a faster decay rate of MR intensity compared with normal tissue and was similar to the EPR results. Good temporal MRI resolution makes dynamic imaging of TEMPOL possible. (8,19) The decay rate of TEMPOL in MR intensity was similar for both normal muscle and salivary glands; however, the decay rate was significantly faster in the tumor region. (8) Takeshita et al. (20) reported that the rate constants of the nitroxyl radical reactions with • OH were in the order of 10 9 M -1 s -1 and that with O 2 •in the presence of cysteine was 10 3 -10 4 M -1 s -1 . It is unclear whether this high reaction rate constant between nitroxyl radicals and • OH is due to the reaction specificity of nitroxyl radicals or the result of a disordered reactivity of • OH. That is, a higher reaction rate constant with • OH may not be due to the efficient neutralization of • OH. Furthermore, each nitroxyl radical has a reactive specificity with O 2 •-. The normalized reduction rates of TEMPOL and CmP (0.1 mM) reacted with O 2 •generated using the hypoxanthine and xanthine oxidase (HX-XO system) (0.02 U/ml) in air-saturated phosphate buffer were 0.88 and 0.17, respectively. (21) When the 6-membered ring nitroxyl radicals were compared, the reaction specificity of the nitroxyl radicals with other reductants was found to be dependent on Gibbs energy. (22) However, studies regarding the competitive reaction specificity of nitroxyl radicals at specific doses of ROS are still in progress.
In this study, the amount of one-electron oxidation of TEMPOL, CmP, and their hydroxylamine forms (TEMPOL-H and CmP-H, respectively) in • OH and O 2 •atmospheres were compared. To allow for comparison, the amount of spin-trapped O 2 •in the HX-XO reaction system and the amount of spin-trapped • OH generated from H 2 O 2 by UV irradiation was adjusted to be equal.
Adjustment of the amount of ROS generation. O 2 •was generated from the reaction between hypoxanthine and xanthine oxidase (HX-XO system). An aliquot of the hypoxanthine solution was added (final concentration 0.5 mM) to initiate a reaction in a mixture containing 1 U/ml xanthine oxidase and 0.5-100 mM CYPMPO. An aliquot of the reaction mixture (120-130 ml) was sampled in a quartz flat cell. The flat cell containing the sample solution was then placed in a TE-mode cavity that was attached to an X-band EPR spectrometer (JEOL, Akishima, Tokyo), and the EPR spectrum was measured immediately. The X-band EPR signal derived from the O 2 •--adduct of CYPMPO (CYPMPO-OOH) was measured repeatedly using the EPR conditions described below.
From the results shown in Fig. 2, the concentration of CYPMPO in the experiments described below was set at 50 mM, which was probably sufficient to detect most of the O 2 •generated in the HX-XO system. The experiment was repeated using 50 mM CYPMPO, and then the time courses of the EPR signal intensity derived from the O 2 •--adduct of CYPMPO (CYPMPO-OOH) and the • OH-adduct of CYPMPO (CYPMPO-OH) were obtained separately (Fig. 3A).
• OH was generated from H 2 O 2 by UVB irradiation (H 2 O 2 -UV system). An aliquot of the reaction mixture (120-130 ml) containing 1 mM H 2 O 2 and 50 mM CYPMPO was sampled in a quartz flat cell. The flat cell containing the sample solution was subsequently placed in a TE-mode cavity using a special cell holder. The reaction mixture was then irradiated by UVB (290 nm) using a UV irradiation unit (JEOL) equipped onto the TE-mode cavity. The time courses of the EPR signal intensity derived from the • OH-adduct of CYPMPO (CYPMPO-OH) were obtained (Fig. 3B). Using the diaphragm, the UV intensity was adjusted, and the H 2 O 2 -UV experiments were repeated until the concentration of CYPMPO-OH in the H 2 O 2 -UV system (circle in Fig. 3B) was similar to the concentration of CYPMPO-OOH in the HX-XO system (square in Fig. 3A) at plateau. The UV irradiation parameters were fixed during all the experiments.
Reactivity of the nitroxyl radical and ROS. An aliquot of a hypoxanthine solution was added (final concentration 0.5 mM) to initiate a reaction mixture containing 1 U/ml xanthine oxidase, 1 mM GSH, and 0.1 mM TEMPOL or CmP. An aliquot of the reaction mixture (120-130 ml) was sampled in a quartz flat cell. The flat cell containing sample solution was subsequently placed in a TE-mode cavity using a special cell holder, and the EPR spectrum was measured immediately. The time courses of EPR signal decay of TEMPOL or CmP were observed using the EPR conditions described below.
An aliquot of the reaction mixture (120-130 ml) containing 1 mM H 2 O 2 , 1 mM GSH, and 0.1 mM TEMPOL or CmP was sampled in a quartz flat cell. The flat cell containing the reaction mixture was then placed in a TE-mode cavity and subsequently irradiated by UV irradiation. The time courses of the EPR signal decay of TEMPOL or CmP was observed using the EPR conditions described below.
Reactivity of hydroxylamine and ROS. An aliquot of a hypoxanthine solution was added (final concentration 0.5 mM) to initiate a reaction mixture containing 1 U/ml xanthine oxidase and 0.1 mM TEMPOL-H or CmP-H. An aliquot of the reaction  and/or • OH generated in the reaction mixture was spin trapped using CYPMPO. The generation of CYPMPO OH in the H 2 O 2 UV system (circle) was adjusted to be similar to the generation of CYPMPO OOH in the HX XO system (square) at the plateau stage. mixture (120-130 ml) was sampled in a quartz flat cell. The flat cell containing the sample solution was then placed in a TE-mode cavity using a special cell holder, and the EPR spectrum was measured immediately. The time courses of the EPR signal for TEMPOL or CmP were observed using the EPR conditions described below.
An aliquot of the reaction mixture (120-130 ml) containing 1 mM H 2 O 2 and 0.1 mM TEMPOL-H or CmP-H was sampled in a quartz flat cell. The flat cell containing the reaction mixture was placed in a TE-mode cavity and then irradiated by UV. The time courses of EPR signal for TEMPOL or CmP were observed subsequently.
X band EPR measurement. The X-band EPR conditions were as follows: microwave frequency, 9.4 GHz; microwave power, 4 mW; center field, 334 mT; sweep width, 10 mT; sweep speed, 5 mT/min; modulation frequency, 100 kHz; modulation amplitude, 0.079 mT; and time constant, 0.03 s. Figure 2A shows the accumulation profiles of CYPMPO-OOH in the HX-XO reaction system obtained using several different concentrations of CYPMPO. When the CYPMPO concentration in the reaction mixture was increased, the amount of CYPMPO-OOH detected increased with the CYPMPO concentration. The maximum EPR signal intensities of CYPMPO-OOH obtained at 60 min after starting reaction were plotted versus the molecular density of CYPMPO (Fig. 2B). The molecular density was defined as the reciprocal of the molecule-to-molecule distance in the previous paper. (24,25) Detected CYPMPO-OOH increased linearly with increasing CYPMPO density, and was saturated when the CYPMPO density was more than 209 mm -1 , which corresponded to a concentration of 15 mM. This result suggested that most O 2

Results and Discussion
•generated in this reaction mixture could be detected using CYPMPO at a concentration of more than 15 mM. Because the inflection point of the plot was obtained at 209 mm -1 , localized generation of O 2 •at the corresponding concentration of 15 mM was expected. The molecular density of 209 mm -1 , i.e., concentration of 15 mM, corresponds to a molecule-to-molecule distance of 5 nm. This was the roughly the same order as the size of a subunit of xanthine oxidase reported previously, 10´9´7 nm. (26) Therefore, the generation of O 2 •may be localized on the xanthine oxidase enzyme. In addition, the concentration of 50 mM used in later experiments was enough to detect most of the O 2 •generated. Figure 3A shows the accumulation profile of CYPMPO-OOH in the HX-XO reaction system. The simultaneous accumulation of CYPMPO-OH with CYPMPO-OOH was also observed in this system. The concentration of CYPMPO-OOH increased gradually with time and reached a concentration of 0.4 mM after 60 min from the start of the reaction before gradually decaying. Similarly, the concentration of CYPMPO-OH increased gradually until 90 min after the start of the reaction and reached a plateau at 0.1 mM. The accumulation rates for CYPMPO-OH and CYPMPO-OOH in the HX-XO reaction system were 0.0207 min -1 and 0.0562 min -1 , respectively. Figure 3B shows the accumulation profile of CYPMPO-OH in the H 2 O 2 -UV reaction system. The concentration at which CYPMPO-OH plateaued was positively correlated with the level of UV intensity (data not shown). Using the diaphragm of the UV irradiation unit, the plateau level of CYPMPO-OH accumulation was adjusted to approximately 0.4 mM. The accumulation rate of CYPMPO-OH in the H 2 O 2 -UV reaction system was 0.0409 min -1 . This UV condition was used for all subsequent experiments.
The instantaneous generation of CYPMPO-OOH and CYPMPO-OH in the very early moment of the reaction can be calculated from generation rates of 0.377 and 0.275 mM/s, respectively. Therefore, 37% more O 2 •compared with • OH can be reacted with CYPMPO in the early stage of the reactions. The time course of instantaneous generation of CYPMPO-ROS adducts, i.e., differentiation of the accumulation profile, decreased gradually with time, and the lead switched after 20 min (Fig. 3C).
The accumulation rates of CYPMPO-OOH and CYPMPO-OH in the reaction system should be proportional to the generation rates of O 2 •and • OH; however, this did not exactly mean the true generation rates of O 2 •and • OH. Such reactive species, especially • OH, can react with other molecules found in the reaction mixture and part of them could be canceled before they reached CYPMPO. The generations and reactions of O 2 •and • OH with CYPMPO can be indicated as Eq. 1-4.
[1] [2] [3] [4] Where, k s1 and k h1 are the generation rates of O 2 •and • OH, but these are difficult to estimate in this experiment; k s2 and k h2 are the generation and k s3 and k h3 are the degradation rates of CYPMPO-OOH and CYPMPO-OH, respectively.
The accumulation rate of CYPMPO-OH and CYPMPO-OOH can be indicated as shown below.
[5] [6] The half-lives of CYPMPO-OOH and CYPMPO-OH, were reported as 51 min and more than 60 min, respectively. (23) The decay constants k s3 and k h3 can be calculated as 0.23´10 -3 and 0.19´10 -3 s -1 from the half-lifes, when the half-life of CYPMPO-OH was as 60 min for simplification. The decay of CYPMPO-OOH and CYPMPO-OH can be negligible, since the decay rates k s3 and k h3 are sufficiently slow. If the accumulation rates of CYPMPO-OOH and CYPMPO-OH were ascribed as being the same, k s2 [O 2 •-] must be equal to k h2 [ • OH]. The second-order rate constants for trapping O 2 •and • OH by CYPMPO were reported as 48 M -1 s -1 and >10 9 M -1 s -1 , respectively. (23) When the trapping rate of • OH was 1´10 9 M -1 s -1 for simplification, [7] Then, [O 2 •-] should be roughly 20,000,000 times more than [ • OH]. This fact suggests that only 0.000005% or less of total • OH could react with CYPMPO in each moment in this reaction system. In a practical sense, however, the actual concentration difference between [O 2 •-] and [ • OH] might be much smaller, since the ratecontrolling factors for accumulating CYPMPO-OOH and CYPMPO-OH in the sequential reactions are definitely k s1 and k h1 , respectively. It was due to that the trapping rates k s2 and k h2 were predominantly faster than the generation rates k s1 and k h1 .
In 50 mM CYPMPO, the molecular distance between CYPMPO molecules was calculated as 3.2 nm, which is a longer distance than • OH can travel in water. There can be 10 or more water molecules present within a 3.2 nm linear space. • OH was generated from 1 mM H 2 O 2 in the H 2 O 2 -UV reaction system in our study. The concentration of 1 mM H 2 O 2 resulted in a molecular distance of 11.8 nm. The molecular distance between CYPMPO molecules, the detector of • OH, was sufficiently shorter than the distance of • OH generated, whereas a concentration of 50 mM CYPMPO was not sufficient for detecting all • OH molecules generated.
This level, however, would probably be sufficient for detecting most of the O 2 •generated in the reaction mixture, as described above. The O 2 •generated in the HX-XO reaction system was derived from dissolving oxygen in the reaction mixture. The maximum possible oxygen concentration in air-saturated water at 25°C is 0.26 mM, resulting in an average molecular distance of 18.7 nm between oxygen molecules. However, O 2 •generation may be localized on the enzyme and it may corresponding to 15 mM at a local concentration.
The amount of total adduct formation (0.4 mM), which corresponded to the amount of ROS able to reach the target, was almost the same between the HX-XO and H 2 O 2 -UV systems used in this study ( Fig. 3A and B). The spin-adduct formation at each instantaneous moment in the O 2 •atmosphere in the HX-XO system and • OH atmosphere in the H 2 O 2 -UV system used in this paper were approximately the same (Fig. 3C). Available oxidation reactions at each instantaneous moment may be similar between these reaction systems. The use of 50 mM CYPMPO as the target molecules for oxidation is probably suitable when a congested intracellular environment is considered. (27) Therefore, this experiment was actually adjusting the total oxidation reaction by ROS to target molecules in the HX-XO and H 2 O 2 -UV systems.
CYPMPO could provide more stable adducts of ROS compared with DMPO. DMPO adducts of O 2 •-(DMPO-OOH) have a relatively short half-life (≈ 1 min), whereas the CYPMPO-OOH adduct has a half-life of 51 min in the HX-XO system. (23) CYPMPO-OH is more stable and has a longer half-life (>60 min). (23) These are sufficiently long half-lives to follow the reactions caused gradually over 60-120 min.
When TEMPOL or CmP was exposed to a reaction with ROS generation, the EPR signal intensity of TEMPOL or CmP decreased in the presence of GSH (Fig. 4). This reduction reaction of nitroxyl radicals was not observed in the absence of GSH (data not shown). Therefore, as shown in Fig. 1, the nitroxyl radicals were one-electron oxidized to form an oxoammonium cation and subsequently two-electron reduced to form hydroxylamine or another stable molecule. The one-electron oxidation of nitroxyl radicals in the aqueous environment of the HX-XO reaction system is most likely due to HO 2 • , an alternate form of O 2 •-. However, its population is small compared with that of O 2 •-. There was a greater reduction in TEMPOL in the O 2 •atmosphere (HX-XO reaction system) than in the • OH atmosphere (H 2 O 2 -UV reaction system) (Fig. 4A). In other words, TEMPOL is more oxidized in environments containing HO 2 • than those with • OH. When the ROS generation was adjusted to reach a plateau at approximately 0.4 mM after 60 min, approximately 0.1 mM TEMPOL could be reduced during this time in the HX-XO reaction system, whereas 0.08 mM TEMPOL could be reduced in the H 2 O 2 -UV system. A greater amount of CmP was reduced in the • OH atmosphere (H 2 O 2 -UV reaction system) than in the O 2 •atmosphere (HX-XO reaction system) (Fig. 4B). Therefore, CmP is oxidized to a greater extent in a reaction with • OH than with HO 2 • . Considering the population of HO 2 • , the nitroxyl radicals, especially TEMPOL, preferably react with HO 2 • . In general, CmP is more redox stable compared with TEMPOL. The redox potentials of TEMPOL and CmP reported by Krishna et al. (21) and Kinoshita et al. (22) suggested that CmP is more resistant against reduction. Another reason for the stability of CmP is the stearic hindrance of the nitroxyl radical moiety. The 5-membered ring structure of CmP is fixed and kept in a plane ring, whereas the 6-membered ring of TEMPOL can bend and exchange between boat and chair forms. The free radical on the pie-orbital of the N-O bond of CmP was always hidden by 4 methyl moieties on neighbor carbons, whereas the free radicals on TEMPOL can peep from the hindrance by methyl moieties due to the bending of the ring formation. Therefore, the free radical on TEMPOL has more chance for a reaction. However, the exact mechanism of the faster reaction of TEMPOL with HX-XO is unclear. Figure 5 shows the reactivity between hydroxylamines and ROS. The one-electron oxidation of hydroxylamines to the corresponding nitroxyl radical occurs almost evenly in both the • OH (H 2 O 2 -UV reaction system) and O 2 •atmospheres (HX-XO reaction system). Reactions between TEMPOL-H and ROS occurred at a rate of less than 1/10 of those that occurred between TEMPOL and ROS. There was no difference in the oxidation reactivity of CmP-H in the • OH or O 2 •atmospheres. The amount of oxidized CmP-H was approximately 3 times greater than that of TEMPOL-H.
TEMPOL could be oxidized to the corresponding oxoammonium cation form relatively quickly, and then sequentially reduced to a diamagnetic species with coexisting GSH. However, the oxidation of TEMPOL-H, which is a one-electron reduced form of TEMPOL, was substantially slower, when similar amounts of oxidizing ROS were provided. Moreover, TEMPOL could be more sensitively oxidized in the O 2 •-/HO 2 • atmosphere compared with the • OH atmosphere. Therefore, the accelerated reduction of TEMPOL in vivo may be a good candidate as a redox probe for O 2 •--induced oxidative stress. Because O 2 •--induced oxidative stress may be caused by an error during electron transfer in mitochondrial energy production, the amphiphilicity of TEMPOL is convenient as a probe in such intracellular patho-physiological conditions. On the other hand, CmP as an in vivo redox probe is probably sensitive enough to show the existence or absence of the reoxidation process of CmP-H to CmP, which may be caused in extracellular conditions. Investigators should choose a suitable nitroxyl redox probe for the purpose of the investigation according to its characteristic redox behavior.

Conclusion
By generating similar amounts of • OH or O 2 •in reaction mixtures containing different types of nitroxyl radicals, the reaction specificity of nitroxyl radicals for a particular ROS can be estimated. TEMPOL was more easily oxidized in an O 2 •atmosphere than in an • OH atmosphere, whereas the opposite was observed for CmP. Although the hydroxylamine forms of TEMPOL-H and CmP-H were oxidized evenly in the • OH or O 2 •atmospheres, the amount of oxidized CmP-H was 3 times greater than that of TEMPOL-H.