Leading individual features of antioxidant systematically classified by the ORAC assay and its single electron transfer and hydrogen atom transfer reactivities; analyzing ALS therapeutic drug Edaravone

Highlights

• •We attempted to classify antioxidants by separately analyzing the two steps of the SET process of peroxide radical generation in the ORAC assay and the HAT process that competes with fluorescein.
• •Analysis using cysteine, ascorbic acid, trolox, and BHT as criteria for comparison concluded that edaravone belongs to the same group as cysteine.
• •The close relationship of edaravone with cysteine provides important in the search for ALS therapeutics.

Abstract

Many natural compounds mop up radicals and limit radical reactions and may prove useful in reducing or preventing oxidative stress-related diseases in vivo. Several assays have been developed to measure antioxidant or anti-radical activity. Here, we measured the anti-radical activities of representative antioxidants using different assays. The oxygen radical absorption capacity (ORAC) assay has two mechanistic stages. We classified antioxidant behavior using two characteristic values thought to be related to the two stages — peroxyl radical formation time (lag time) and fluorescein annihilation rate (k_obs) — by applying Voronoi polyhedral division. We focused on four class-representative antioxidants, Trolox ®, vitamin C, l-cysteine, and 2,6-di‑tert‑butyl‑p-cresol, and compared their characteristic activities with those of edaravone. Our analysis indicates that edaravone is in the same group as cysteine and may function via a similar mechanism. Our results suggest that analyzing lag time and k_obs is a useful method to characterize antioxidants.

Graphic abstract

Single electron transfer and hydrogen atom transfer reactivity in the oxygen radical absorption capacity assay of class representative antioxidants and edaravone

1.Introduction

Amyotrophic lateral sclerosis (ALS) is a progressive, fatal neurodegenerative disorder [1], with a median survival time of 37 to 49 months [2]. Known prognostic factors for ALS include age at onset, site of onset, duration of debilitation, clinical impairment, and degree of respiratory function [3]. The cause of the disease is unknown, and no current treatment is known to affect survival [4].

ALS has recently been categorized as a TDP-43 proteopathy [5], [6], [7], a term comprising a group of diseases characterized by abnormal phosphorylation of TDP-43. Abnormal phosphorylation of TDP-43 is pathologically due to either loss of function or acquisition of toxic function or both and has been demonstrated to cause characteristic neurodegeneration and a clinical syndrome [8].

In 1995, the drug riluzole, a benzothiazole, was approved by the FDA for the clinical treatment of ALS [4]. Riluzole is a glutamate antagonist and is reported to block the excessive release of glutamate from motor neurons (MN) [9]. Until recently, riluzole was the only drug known to delay disease progression. In 2017, the FDA approved 3-methyl-1-phenyl-5-pyrazolone (i.e., edaravone, EDA) for the clinical treatment of ALS. EDA, developed by Mitsubishi Tanabe Pharma, was shown to inhibit ALS progression in early disease stages [10]. In addition to approval in Japan and the United States, EDA is currently approved for use in South Korea, China, Switzerland, and Canada [11]. EDA is known to act as a ROS scavenger [12], inhibiting peroxyl radical (ROO• radical)-induced peroxidation systems [13].

Many natural compounds mop up radicals and limit radical reactions, thus reducing or preventing oxidative stress-related diseases in vivo [14]. Some of these compounds are also known to stabilize food during processing and storage [15], [16], [17]. Therefore, for both health and food stability reasons, the antioxidant capacity and ingredients of the foods we consume are of great interest to the general public, medical and nutrition experts, and food science research.

Several assays have been developed to rapidly predict the antioxidant or anti-radical activity of natural compounds and extracts [18]. Among the methods developed to estimate radical-scavenging activity, assays based on the scavenging of 1,1-diphenyl-2-picrylhydrazyl (DPPH) or 2,2′-azinobis(3-ethylbenzothiazoline-6-sulfonic acid) (ABTS•⁺) are the most frequently used [19], [20], [21], [22], [23], [24]. The stable radical DPPH was first proposed for quantifying antioxidant (mostly phenol) content more than 60 years ago [25,26]. The DPPH assay is simple and can be run in most rudimentary laboratories [27]. The ABTS•⁺ assay is a development involving a stronger radical (chemically produced), and this assay is often used for screening complex antioxidant mixtures such as plant extracts, beverages, and biological fluids [28]. In addition to the food industry, the pharmaceutical and cosmetic industries are also increasingly interested in measuring antioxidant capacity [29]. Indeed, reactive oxygen species (ROS) have been implicated in the pathogenesis of several chronic diseases (and in the aging process in general), and methods for assaying antioxidant activity against ROS under pathological circumstances have proven particularly remarkable.

In the late 1980s, Glazer [30,31] performed an assay to measure hydrophilic antioxidant activity against ROS based on the detection of chemical damage due to reduced fluorescence of R-phycoerythrin (PE). Antioxidants react with free radicals through different mechanisms, with hydrogen atom transfer (HAT) or single electron transfer (SET) mechanism or the combination of both HAT and SET mechanisms being primary examples [32]. For example, Liao et al. [33]. cited the DPPH assay as an example of the SET reaction and the oxygen radical absorption capacity (ORAC) assay as an example of the HAT reaction. Among the specific examples, the ORAC assay has recently been regarded as a universal evaluation standard for functional foods. The ORAC assay was developed as one of the few methods that combine both antioxidants ROS inhibition rate and ROS inhibition duration in one quantity [34], [35], [36], [37]. In this method, the heated 2,2′-azobis(2-methylpropionamidine) dihydrochloride (AAPH) is symmetrically disintegrated into ROO• radicals, which subsequently cause fluorescence decays of the reference compound. Antioxidant activity is assessed by comparing the area under the fluorescence quenching curve (AUC) of the antioxidant to that of the blank. The protective effect of antioxidants is quantitatively represented in terms of Trolox equivalents. The ORAC assay provides a unique and complete assessment that measures the inhibition time and degree of inhibition after the reaction is complete.

Ou et al. [29]. developed and validated an improved ORAC assay using fluorescein (FL) as the fluorescent probe (ORAC-FL). FL oxidation products induced by ROO• radicals have been identified by LC-MS, confirming that the reaction mechanism proceeds as a classical HAT mechanism [29]. In the ORAC assay, we recognized that the symmetrical disintegration of the heated AAPH into double ROO• radical equivalents only occurs with ROS, and this step should be sensitive to only antioxidants that exert their effects through the SET mechanism. Therefore, we concluded that the ORAC assay has separate reactions to the hydrogen-independent SET and HAT processes. With the SET mechanism, the inhibitory effect of antioxidants is based on the production of ROO• radicals relative to the lag time before the beginning of fluorescence quenching curves, whereas with the HAT mechanism, the kinetic inhibition of antioxidants on the quenching of FL is relative to the produced ROO• radicals. Practically, as described by Ou et al., the curves with similar inclinations towards the right (namely elongation of lag time) are dependent on the concentration of Trolox, whereas curves with slight inclinations (namely decrement of kinetic constant) are dependent on the dosage of the examined grape seed extract.

With the above applications in mind, we have classified antioxidant reaction behavior during radical scavenging activity by applying the Voronoi polyhedral division [38,39]. The Voronoi polyhedral division or Voronoi tessellation decides the borders among the territories of objects sharing common properties in the event space. In the current study, since the number of antioxidant molecules existing in the space consisting of the experimental ORAC assay parameters is too large and unknown, it is impossible to directly survey clusters corresponding to practical classified antioxidants. Therefore, we focused on typical antioxidants that are biologically classifiable; these four class-representative antioxidants are tocopherol derivatives (i.e., vitamin E) with enhanced aqueous solubility (Trolox®, TRO); vitamin C (ASC, representative of hydrophilic vitamins); L-cysteine (CYS, the reactive central amino acid residue of glutathione); and 2,6-di‑tert‑butyl‑p-cresol (BHT, generally used as a synthetic antioxidant to protect industrial products from oxidative damage). The normalized variables of antioxidative indices derived by our multiplied ORAC assay were assigned to each of the Voronoi polyhedra produced by dividing the similarity range by the flexible protean region, but not a rectangle region on a graph paper. The Voronoi polyhedra consist of a set of such regions. Each of the polyhedra has one reference point (i.e., the class-representative antioxidant) with certain coordinates. Each polyhedron comprises a set of every point closer to its reference point than to any other point.

We thought that finding the unique characteristics of EDA and developing compounds enriched with those characteristics would lead to the creation of drugs with a new ALS progression inhibitory effect. Therefore, this study aimed to discover the unique characteristics of EDA, the only antioxidant that has an ALS progression inhibitory effect.

2. Materials and methods

2.1. Materials

ASC, 1,7-bis(4‑hydroxy-3-methoxypheyl)−1,6-heptadiene-3,5‑dione (i.e., curcumin, synthetic, CUR), DPPH, FL, TRO, EDA, myricetin (MYR), and N-vanillylnonanamide (i.e., capsaicin, synthetic, CAP) were purchased from Tokyo Chemical Industry Co. (Tokyo, Japan). ABTS•⁺, AAPH, CYS, hydrocortisone (HYD), and quercetin dihydrate (QUE) were purchased from Fujifilm Wako Pure Chemical Corporation (Osaka, Japan). BHT was purchased from Nacalai Tesque, Inc. (Kyoto, Japan). All other reagents used were of the highest commercial grade.

2.2. Conventional antioxidant capacity measurement method

2.2.1. DPPH radical scavenging activity

The DPPH radical scavenging activity of the antioxidants was measured according to the modified method of Liao's group [40]. An aliquot of 100 µL of antioxidant solution at different concentrations was mixed with 100 µL of DPPH solution (100 µM in 80% ethanol). The absorbance of the resulting solution was measured for 30 min at the interval of 60 s at the wavelength of 521 nm using a spectrophotometer (UV 2550, Shimadzu, Kyoto, Japan). Ethanol was used as a control. All determinations were performed in triplicate. The radical scavenging capacity of the tested samples was then determined using the equation below:

$$\text{Radical}\text{scavenging}\text{activity}(\%)=\frac{A_{\text{control}}-A_{\text{sample}}}{A_{\text{control}}}\times 100$$

For each antioxidant concentration tested, the radical scavenging activity (%) was plotted. Percentage of DPPH radical remaining at the steady-state was determined, and the values were then transferred onto another graph showing the percentage of residual DPPH radical at the steady-state as a function of the molar ratio of antioxidant to DPPH radical. Anti-radical activity was defined as the amount of antioxidants necessary to decrease the initial concentration of DPPH radical by 50% (EC₅₀) [41]. For clarity, antioxidant efficiency is represented as the logarithm of reciprocal EC₅₀.

2.2.2. ABTS radical scavenging activity

The ABTS scavenging activity of the antioxidants was measured according to the modified method of Re’s group [42]. A 7 mM stock solution was prepared by dissolving ABTS in 80% ethanol. Potassium persulfate was then added to this stock solution (final concentration of 2.45 mM), and the mixture was allowed to stand for 12–16 h at 25 °C until a dark blue-green color developed. Fresh stock solutions of ABTS•⁺ were prepared at use. An aliquot of 100 µL of antioxidant solution at different concentrations was mixed with 100 µL of ABTS solution (100 µM in 80% ethanol). The absorbance of the resulting solution was measured for 30 min at 60 sec intervals and 744 nm using a spectrophotometer (UV 2550, Shimadzu, Kyoto, Japan). Ethanol was used as the control. All determinations were performed in triplicate. The radical scavenging capacities of the tested samples and EC₅₀ values were determined using a similar method as that described in the previous section.

2.2.3. ORAC assay and its conventional analysis

The method by Ou's group [29] was modified to carry out the reaction in 75 mM phosphate/NaOH buffer (pH 7.4) in a final reaction mixture of 200 µL. Antioxidants (20 µL) and 120 µL of FL solutions (final concentration of 120 nM) were transferred to a 96 well microplate. The mixture was preincubated for 15 min at 37 °C. An aliquot of 50 µL of AAPH solution (final concentration of 12 mM) was rapidly added. The microplate was immediately placed in the reader, and the fluorescence was recorded for 90 min at 60 sec intervals. The microplate was automatically shaken before each reading. Several blanks containing no FL and/or no AAPH were performed using phosphate/NaOH buffer instead of an antioxidant solution and antioxidant solutions of various concentrations. All measurements were performed in triplicates [43]. The antioxidant curve on the fluorescence versus time diagram was first normalized with the highest fluorescence intensity of the original data at 100%. From the normalized curves, the area under the net fluorescence decay curve (net AUC) was determined as follows:

$$\mathit{net}\mathit{AUC}=\sum _{i=0}^{90}{f}_i$$

Where f₀ is the initial fluorescence read at 0 min, and f_i is the fluorescence read at time i. The net AUC was calculated for all antioxidants. The ORAC value in the TRO equivalent was calculated as

$$\mathit{ORAC}\mathit{value}=\frac{\mathit{net}\mathit{AU}{C}_{\text{sample}}}{\mathit{net}\mathit{AU}{C}_{\text{TRO}}}\times \frac{\left[\text{TRO}\right]}{\left[\text{sample}\right]}$$

Where net AUC_TRO is the net AUC value determined for TRO, and net AUC_sample is the net AUC value determined for each antioxidant sample except for TRO.

2.3. Analyses of measurement of the lag time and determination of the kinetic rate of FL quenching reaction in ORAC assay

From the results obtained in the ORAC assay, the first derivative of the fluorescence intensity over time was calculated and graphed (Fig. 3). The differential value of fluorescence intensity to time with hour unit was defined as the observed reaction rate constant (k_obs), and the period until the differential fluorescence intensity started increasing was defined as a lag time.

Fig. 1.

Results of DPPH assay. DPPH radical scavenging activity (Δ DPPH /%) was plotted for EDA concentrations of 1.0 μM to 25 μM.

Fig. 2.

Results of ABTS assay. ABTS radical scavenging activity (Δ ABTS /%) was plotted for EDA concentrations of 1.0 μM to 25 μM.

Fig. 3.

Results of ORAC assay. (a) EDA; (b) TRO; (c) ASC; (d) CYS; (e) BHT; (f) QUE; (g) MYR. The top row shows the change in fluorescence intensity due to each antioxidant recorded at 37 °C for 90 min every 60 s. The bottom row is a graph of the first derivative of fluorescence intensity over time.

2.4. Classification of reaction modes of radical scavenging activity of antioxidants using the Voronoi polyhedral division

A Voronoi diagram of a set of points given on a plane is a division of the plane based on the closest points. This is a concept that has been used in various fields even before the beginning of computational geometry [44].

As described in the Introduction section, With the SET mechanism, the inhibitory effect of antioxidants is based on the production of ROO• radicals relative to the lag time before the beginning of fluorescence quenching curves, whereas with the HAT mechanism, the kinetic inhibition of antioxidants on the quenching of FL is relative to the produced ROO• radicals. Therefore, we determined two parameters: the differential elongation of lag time to the logarithm of reciprocal [antioxidant] and the differential inclination of quenching fluorescence curve to the logarithm of reciprocal [antioxidant]. Using the average and standard deviation, we standardized these parameters into the orthogonal space, including equivalent deviations toward both axes. Two-dimensional Voronoi polyhedra for four representative antioxidants were divided by the perpendicular bisecting lines between each pair of these antioxidants in this orthogonal space. These tasks were done manually using a ruler and a calculator.

3. Results and discussion

3.1. DPPH radical scavenging activity of model antioxidants and sample antioxidants

The radical scavenging activity of EDA was measured using the DPPH assay. When DPPH radical consumption after 30 min was plotted on the vertical axis against added EDA concentration, the absorption of DPPH radicals at 521 nm was initially (<7.5 μM) observed to decrease in proportion to EDA concentration. At 7.5 µM, DPPH radical absorption disappeared, and DPPH radical consumption increased linearly. At EDA concentrations in the 7.5 – 25 µM range, 100% of the DPPH radicals were consumed. Using the DPPH assay, the observed activity of EDA (EC₅₀) was 4.21 µM. Osamu Tokumaru et al. [45]. previously reported the concentration at which 50% of the DPPH radicals are scavenged by EDA as 4.7 ± 0.3 μM. Kraus et al. [46]. previously reported the concentration at which 50% of the DPPH radicals are scavenged by EDA as 6.5 ± 1.3 μM. The reason why the EC₅₀ obtained in this study was low compared to these results may be related to the difference in the composition of the solvents used, which also competes with antioxidants in radical removal.

Next, the radical scavenging activities of TRO, ASC, CYS, and BHT were also measured using the DPPH assay (Table 1). The DPPH radical scavenging activities of the flavonoids (QUE and MYR), a steroid (HYD), and spice components (CUR and CAP) were also measured. While the flavonoids showed very high activity using the DPPH assay, no significant activity was observed in the steroid. Under these conditions, the radical scavenging activities of EDA were more aggressive. Thus, EDA showed the third-highest activity after flavonoids.

Table 1.

Antioxidant concentration that can reduce the concentration of radicals at the end of the reaction to 50% of the initial concentration (in the DPPH and ABTS assays).

	TRO	ASC	CYS	BHT	EDA	QUE	MYR	HYD	CUR	CAP
log (1/ECDPPH50)	4.97	5.06	4.90	4.28	5.47	5.67	5.72	ND	5.08	–
	±0.00502	±0.0236	±0.0168	±0.0145	±0.00880	±0.0229	±0.0230		±0.00800
log(1/ECABTS50)	4.92	4.81	5.22	4.90	5.20	5.33	5.45	ND	5.10	5.10
	±0.0128	±0.00991	±0.0660	±0.0207	±0.0277	±0.0475	±0.0323		±0.0109	±0.0285
ORAC value	1.00	1.11	0.400	0.203	5.65	10.9	5.08	0.00541	1.14	5.03

The ORAC value calculated from the ORAC assay is also reported.

3.2. ABTS radical scavenging activity of model antioxidants and sample antioxidants

The radical scavenging activity of EDA was measured using the ABTS method. Our understanding of the mechanism of the chemical reaction utilized in the ABTS assay is clearer than that utilized in the DPPH assay. Moreover, the results are considered to be faithful to the SET reaction [47]. At EDA concentrations below 12.5 µM, the absorption of ABTS radicals at 744 nm decreased in proportion to the EDA concentration (and the absorption of ABTS also decreased). At 12.5 µM, ABTS radical absorption disappeared and ABTS radical consumption increased linearly. At EDA concentrations in the 12.5 – 25 µM range, 100% of ABTS radicals were consumed. The EC₅₀ of EDA in the ABTS assay, the radical scavenging activity of EDA (defined in exactly the same way as in the DPPH assay), was 5.52 µM. Kraus et al. [46]. reported an IC₅₀ of 24 ± 5 μM (the concentration at which EDA scavenges 50% of the ABTS radicals), which was increased compared with our results. The dissimilar solvent and pH conditions used in the two studies may account for the observed difference.

Next, the ABTS radical scavenging activities of TRO, ASC, CYS, and BHT were measured (Table 1). The radical scavenging activities of EDA were more aggressive than those of TRO, ASC, and BHT, although less aggressive than that of CYS. As antioxidant concentration decreased, the ABTS radical scavenging rates of QUE, MYR, CAP, and CUR samples gradually decreased. This was considered to be due to the radical scavenging action of the two-step reaction. According to measurements, EDA showed the fourth-highest activity after flavonoids and CYS. The activities of TRO, ASC, and EDA were weaker than the corresponding activities obtained using the DPPH assay; i.e., the log (1/EC₅₀) values were smaller. In contrast, CYP and BHT achieved 50% elimination at a lower concentration than that achieved with the DPPH assay and showed slightly stronger activities. Overall, the activities of the model antioxidants were not substantially different between the DPPH and ABTS assays. The same was true for the sample antioxidants QUE, MYR, CUR, and CAP. The results are presented in Table 1. The flavonoids QUE and MYR possessed high DPPH and ABTS radical scavenging activities (Table 1).

3.3. Conventional ORAC analysis of model antioxidants and sample antioxidants

When the radical scavenging activity of EDA was measured using the ORAC assay, net AUC increased as the EDA concentration increased (Fig. 3). The observed ORAC value of EDA was 5.65 µM of Trolox equivalent/ µM (Table 1). There was a slight positive correlation between the ORAC value and log (1/EC^DPPH₅₀) (R² = 0.63). There was little correlation between the ORAC value and log (1/EC^ABTS₅₀) (R² = 0.44). Qiang et al. [48]. reported an ORAC value for EDA of 0.72 ± 0.03 µM of Trolox equivalent/ µM, which is considerably smaller than the value reported here. The activities of the model antioxidants TRO, ASC, CYS, and BHT were also measured using the ORAC assay (Fig. 3). According to the ORAC assay measurements, EDA was the second most active compound after the flavonoid QUE. TRO and ASC both showed concentration-dependent lag times that facilitated a comparison of antioxidant activities under the same conditions. However, no lag time was observed with BHT, where the FL fluorescence quenching rate was observed to decrease in a concentration-dependent manner. Thus, in this case, the reaction mechanism was different from that of the above compounds.

While prolonged lag time reflects the process of producing ROO• radical from AAPH, the reaction rate constant k_obs reflects the ROO• radical and FL reaction process. The action differs depending on the drug, but the ORAC assay adopts the policy of using AUC so that the antioxidant action can be compared collectively [49]. However, we believe that changes in lag time and changes in k_obs also contain important information that reflects qualitative differences in antioxidants. Because of the lack of stoichiometric equivalence, simply comparing the magnitude of action of both drug doses using the area below the curve was not a useful analytical approach. For CYS, the delay time was extended and the quenching rate of FL fluorescence was reduced. Similarly, EDA prolonged lag time and reduced the extinction rate of FL fluorescence in a concentration-dependent manner. The ratio of the strength of action to the AAPH radical reaction in TRO and ASC was considerably different from the ratio of the strength of action to ROO• radical in BHT. The ratio of CYS to EDA was between these two values. With the flavonoids QUE and MYR, the transition from fluorescence to quenching of FL continued from the end of the lag time, and the intensity of action on AAPH and ROO• radical did not appear to change significantly. The spices CAP and CUR showed similar behavior to the flavonoids described above in that the transition from extended lag time to slower ROO• radical kinetics changed continuously and gradually.

From the above results, two properties were unique to each antioxidant. These were the ratio of the concentration-dependent strength of action to the strength of competition for FL with fluorescence quenching and the ratio of the strength of suppression of the AAPH reaction to the strength of competition for FL with fluorescence quenching. The MYR, which showed high activity in the DPPH and ABTS assays, was less active in the conventional ORAC analysis. With the flavonoids QUE and MYR, the fluorescence intensity gradually decreased over time. This may be because the gradual change in the gradient due to the change in QUE concentration was smaller than the corresponding change in MYR, CAP, and CUR, and thus the area was overestimated.

No EDA-specific functionality was obtained following testing using the DPPH assay, ABTS assay, or traditional ORAC analysis. Therefore, the reaction curves obtained by the ORAC assay were classified into lag time extension type, reaction rate constant k_obs reduction type, and intermediate type, and the characteristics peculiar to EDA were searched by focusing on the typing of these reactions.

3.4. ORAC assay analysis focusing on the combination of lag time and reaction rate constant

The ORAC assay analysis used focused on the reaction rate constant k_obs and lag time. In the differential kinetics plot, the magnitude of the kinetics of TRO and ASC were constant with respect to antioxidant concentration, and the lag time could be defined as the time from the start of the reaction to the base of this peak. No non-reaction period corresponding to lag time was observed with BHT, and peak height decreased in a concentration-dependent manner. Therefore, in the case of BHT, the difference between the reaction rate constant and the gradient of FL alone was defined as the intensity of activity. Since CYS demonstrated a lag time, it was analyzed in the same way as TRO and ASC. However, since the peak height became smaller in a concentration-dependent manner, CYS was also analyzed similarly to BHT. EDA, the flavonoids QUE and MYR, and the spice components CAP and CUR were all analyzed similarly to CYS.

The values of k_obs and lag time were plotted separately against the reciprocal of the antioxidant concentration (Fig. 4). For each antioxidant, the resulting plots demonstrated a nearly constant slope. From these graphs, a delay time extension due to concentration was observed due to the steep gradients of TRO, ASC, CYS, QUE, MYR, CAP, and CUR. Moreover, CYS, EDA, QUE, MYR, and CAP were observed to have a large slope value, and the change in k_obs due to the change in concentration was large. However, the ASC graph was associated with large errors, and the accuracy of this data could not be confirmed. ASC and Vitamin E both remove active oxygen in the cell membrane. Vitamin E moves around in the membrane, and ASC restores vitamin E in the aqueous phase, expelling radicals to oxygen-rich areas. Therefore, ASC is involved in the redox reaction at the gas phase interface, so it is more susceptible to air oxidation than hydrophobic drugs unless oxygen is expelled from the experimental system. Thus, we thought that ASC should be especially considered for air oxidation. As a reaction mechanism, lag time reflects a delay in the process from AAPH to ROO• radical formation, consistent with a reaction involving a one-electron transfer mechanism (SET reaction) to active oxygen. The observed change in k_obs with antioxidant concentration suggests that the competitive antagonistic effect of FL fluorescence on quenching reflected the strength of the hydrogen atom abstraction reaction (HAT reaction) from ROO• radical. We were able to quantify the difference between QUE, which has a high lag time extension effect, and MYR, which has a medium lag time. QUE and CUR contribute significantly to the extension of the lag time. In this respect, QUE has a stronger influence on the SET reaction than MYR, and CUR has a stronger influence on the SET reaction than CAP. Here, it has been reported that demethoxycurcumin, an analog of CUR, may be useful for neurodegenerative diseases associated with mutant TDP43, and it is discussed that it may affect the redox state of proteins [50]. Consequently, ORAC analysis of the combination of lag time and reaction rate constant provided additional information not displayed in the conventional ORAC analysis.

Fig. 4.

Results of ORAC assay. (a): Relationship between antioxidant concentration and apparent k_obs. (b): Relationship between antioxidant concentration and lag time.

3.5. Classification of radical scavenging activity reaction modes of antioxidants using the Voronoi polyhedral division

Using the vertical and horizontal axes of the slopes obtained in Fig. 4, the Voronoi division was performed utilizing non-hierarchical cluster analysis. In this manner, the antioxidants used in this study were classified into four model antioxidants (Fig. 5). The high reactivities of QUE and CUR are consumed in reactions at the biological membrane involving ROS in the body. However, MYR and CAP, which are delivered intracellularly, may maintain their activities. A similar situation is observed with CYS, which demonstrates a bimolecular reaction. EDA, which is in the same group as CYS, is also involved in radical reactions with intracellular proteins.

Fig. 5.

Reaction classification of antioxidants. The larger the value on the vertical axis, the higher the reactivity with ROO• radicals. The smaller the value on the horizontal axis, the longer the lag time.

EDA demonstrated almost no prolongation of lag time with concentration, and the reactivity with ROO• radical was improved. As mentioned above, the delay time reflects the delay from AAPH to ROO• radical formation, suggesting a single electron transfer (SET reaction) to ROS. It was suggested that k_obs competitively antagonized the quenching of FL fluorescence because of the attack on ROO• radical and the subsequent transfer of a hydrogen atom (HAT reaction) to ROO• radical. While conventional ORAC analysis is a consensus method for comparing these processes with a single numerical value simultaneously, comparing it in a vector space as shown here makes it schematically understandable. Therefore, we reasoned that ORAC analysis of the combination enables the observation of both the SET reaction and the HAT reaction distinguishably.

Our analysis indicates that EDA is situated in the same group as the model antioxidant CYS. As CYS suppresses protein oxidation (a property shared with glutathione), EDA may function similarly. This is because if the mutation in the TDP-43 protein that causes the mutation in ALS is due to oxidation, it can be considered that EDA suppresses this and therefore has an effect of suppressing ALS progression. In addition, EDA presented a small delay-prolonging effect similar to that observed with water-soluble ASC and hydrophobic and membranous TRO groups. However, EDA was more reactive towards ROO• radicals than the TRO groups. Indeed, EDA is characterized by a relatively high reactivity towards ROO• radicals. Therefore, it is concluded that EDA may have a strong influence on the HAT reaction. Focusing on the results of the ORAC assay and analyzing the hydroxyl radical formation time (lag time) and FL annihilation rate (k_obs), as in this study, is a useful way to characterize antioxidants. The effect of CUR on ALS is being investigated, but no success has been recorded yet. Since glutathione transferase plays an important role in the mechanism, protein-acting redox modification appears to be important. However, EDA, CAP, and MYR are added to the CYS class, suggesting that MYR may be more suitable than QUE for flavonoids and CAP than CUR for spices. However, these SET reactions are weaker than EDA, and this seems to be a challenging factor in the search for new candidate compounds.

4. Conclusions

In the ORAC assay, the heated AAPH is symmetrically disintegrated into ROO• radicals, after which FL fluorescence is attenuated. The lag time that occurs when antioxidants are added reflects the delay in the process from AAPH to ROO• radical formation, suggesting a single electron transfer (SET reaction) to ROS. It was suggested that Cobb competitively antagonizes the quenching of FL fluorescence due to the attack on ROO• radicals followed by the transfer of hydrogen atoms to ROO• radicals (HAT reaction). The balance between this HAT reaction and the SET reaction is unique to each antioxidant. When this was classified by TRO, ASC, CYS, and BHT, it was in the same group as CYS. (Fig. 5) EDA, similar to CYS, may function through a mechanism that suppresses protein oxidation (a property shared with glutathione). Although this result considers only the radical scavenging reaction, the effect of the drug is not limited to the strength at the site where the pharmacological action is expressed. Proper delivery to the affected area is important, and consideration must be given to the water octanol partition coefficient of the drug and the effects of metabolic enzymes. Still, we expect it to be worth looking at as an indicator for selecting candidates.

Funding sources

This work was partially supported by JSPS KAKENHI Grant Number 17K05366.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.