This is a method for the assessment of reductive capacity. It's not new, basically this a titration by permanganate. It can't be used to measure TAS. Antioxidants are easily oxidizable molecules. Only relatively weak oxidants can be used for antioxidant titration. Permanganate is too strong oxidant, it reacts with almost all organic materials. The product, MnO2, can be also reduced by numerous organics. The titration end-point is not easy to detect. Hypothetically, all organics can be oxidized to CO2. In addition MnO2 is weakly, but soluble in phosphate buffer. The relations of the reductive capacity to different physiological deviation is not established. I don't understand how useful is the value of reductive capacity.

Dear Yurii V Geletii

Thank you for your valuable comments. You make some good points.

I agree that titration by permanganate is not a new method and this has been clearly emphasized in the text (it is even stated that it has been introduced by Margueritte in 1846). The novel aspect of the method is the idea of testing the samples previously fixed onto the nitrocellulose membrane. As explained in the text, this is important as the precipitate usually complicates the analysis and determination of the endpoint of titration with biological samples (eg. permanganate titrations used for determination of antioxidant capacity proposed by others). In NRP protocol, this precipitate is used to assess the reductive capacity. I am not aware of the permanganate titration being done before this way (with prefixed samples onto the nitrocellulose). If you know of a paper that used this method before I would greatly appreciate if you could share it as I've done extensive search of all scientific databases and I wasn't able to find anything similar (with the method based on the potassium permanganate agar diffusion-based determination of TAC (Zhou et al 2015; mentioned in the manuscript) being the closest thing I could find proposed for measurement of TAC based on the quantification of MnO2.

Regarding the use of reductive capacity measurements for the antioxidant capacity assessment: I agree that there should be a clear distinction between the reductive capacity and the antioxidant capacity. I often emphasize this and it was even mentioned in the first version of this manuscript (but had to be removed due to the word limit). The umbrella term "antioxidant capacity" is often used interchangeably for both biological (physiological) and chemical antioxidant capacity in the scientific literature. This is not entirely correct as biological antioxidant capacity consists of both reductive capacity of cellular antioxidants (eg. GSH) and the enzymatic capacity of cellular antioxidants), and chemical antioxidant capacity (or reductive capacity) of the chemical analytes. Although it is somewhat contextually "wrong" to consider the terms "reductive" and "antioxidant" synonyms in the biological system (due to exclusion of the enzymatic arm of the system), this is not entirely untrue as the definition of antioxidant is based on its ability to donate electrons to species of lower stability.

Now this debate opens a Pandora's box of what should be considered as a functional antioxidant in the context of the complex biological system of the cell. If I understood your point correctly, you don't see how reductive capacity is relevant for the biological system due to the fact that reductive capacity is defined in much broader range than what is considered to be relevant in the living system (eg. hypothetical oxidation of organics to CO2). I see this as a potential problem in all methods used for measurement of antioxidant capacity. The main problem is that a method that reliably reflects what happens in the biological system simply doesn't exist, so different methods rely on different strategies for approximation of the antioxidant capacity of the physiologically functional fractions. For this particular reason, different methods that are supposed to measure the same thing (antioxidant capacity) often yield different results, although well designed methods usually provide comparable measurements. Usually, limitations of such methods are imposed by the fact that redox reaction potential of the chemical substrate used for the measurements *always* differs from what is considered to be the functional physiological range of the potentially reversible electron exchange.

Due to this fact some methods overestimate a specific part of the system and underestimate others. This has been elegantly shown in the modified trihydroxybenzene superoxide-scavenging assay (Li 2012; 10.1021/jf204970r ) where slight modifications in pH and temperature clearly show over/underestimation of specific antioxidant fractions (33 antioxidants tested). In vitro conditions are different from what happens in vivo even when the system is tested in the specific conditions that resemble a physiological environment (37 C, pH 7.35-7.45 etc). So, which specific modification is best suitable for measurement of cellular antioxidant capacity? Is it more appropriate to do the trihydroxybenzene autooxidation at the pH of 8.2 (as originally proposed by Marklunds) and to underestimate antioxidants that rely on carboxylic acid, ester, or lactone groups, or should the method be modified to underestimate some other components of the system? There is no perfect method that measures all the subsystems, so all methods over/underestimate specific segments to provide the number that best approximates what might be going on in vivo.

In the NRP this error is also present, and its source is, as you say, the potential to enter redox reactions with more substrates than could happen in the living system (so the noise from reductive capacity with little or no physiological significance is also present in the readings). In other words, KMnO4 will definitely oxidize not only cellular antioxidants like GSH, but also other organic material present inside the cell some of which would not donate electrons for stabilization of unstable chemical species in vivo. Nevertheless, the difference in the capacity in this scenario is derived from the difference in residual antioxidant capacity of two biological systems (eg. control and the test sample). In chemical experiments this might be a problem in the specific context as NRP might give false positives if you have more organic matter in one sample. However, in the experiment with biological samples you have several steps for correction of this potential bias. First, you always have appropriate control samples of the same biological origin and chemical complexity (for example ipsilateral and contralateral side of the brain as shown in the manuscript; or hippocampus of the control animal and the animal previously treated with a drug). This already assures you have the same amount of oxidizable organic material. Secondly, as the change in the overall organic matter can occur in specific cases (but is not very likely in most of the experiments), and this is a well known and significant bias in biological experiments (eg. the cells form your cell culture died, and you sampled only the medium), the control for this is a standard step in all experiments (eg. After making sure you have a linear relationship between the amount of your analyte and the overall concentration of your housekeeping variable, you report ratiometric values to provide additional correction for the potential bias introduced either in the biological experiment or during the tissue processing). As you can see from the manuscript (and the provided Supplements) the NRP of tissue samples (hippocampal homogenates of control rats and rats treated intracerebroventricularly with streptozotocin, as well as wild type mice hippocampi and hippocampi of the Tg2576 transgenic animals) is reported in a ratiometric manner (with the amount of protein used as the housekeeping variable (measured by two different methods, to make sure we also controlled for this)). Moreover, a potential bias of ratiometric correction was analyzed separately to determine the concentration drift that enables this approach to be used with minimal introduction of processing confounders. So the fact that KMnO4 is a strong oxidant and that it would oxidize much more than is expected to be oxidized by unstable radicals in the physiological system consequently providing a reductive defense was taken into account and appropriately addressed. Furthermore, a very wide reaction potential of KMnO4 is better suited for detection of defective antioxidant systems, as in comparison with other oxidants KMnO4 enters more redox reactions. This chemical "unspecificity" allows a greater sensitivity to detect a defective antioxidant defense in comparison with other methods used for TAC evaluation. I hope this makes it a bit clearer why the KMnO4 was used even though it is a strong oxidant, and although it demonstrates the ability to enter redox reactions with various compounds.

Noteworthy, this "drift" in the measured reductive range is present in other methods used for determination of antioxidant capacity, so this is something that is universally present as the problem in antioxidant capacity measurements.

Additionally, the potential of KMnO4 for the purpose of measurement of total antioxidant capacity has also been recognized by other researchers (eg. as mentioned above in the potassium permanganate agar method for measurement of total antioxidant capacity proposed by Zhou et al. 2015, for the determination of total mass of antioxidant substances and antioxidant capacity per unit mass in serum by potassium permanganate redox titration proposed by Zhang et al 2014; 10.1155/2014/928595 etc.).

Regarding the problem of physiological significance of reductive capacity. Is it really something that matters in the biological system? In my opinion yes as the literature suggests it is hard to assess and dissect all the antioxidant systems. For example, is glyceraldehyde 3-phosphate dehydrogenase (GAPDH; EC 1.2.1.12) an antioxidant? In the context of the current narrow definition of the cellular antioxidants one would not consider GAPDH to be an antioxidant. Nevertheless, like many other enzymes, cysteine residues in the active site of the enzyme sense redox balance of the cellular environment and donate their electrons to the functional stabilization pool with consequent formation of the intramolecular disulfides and reversible protective inactivation of the enzyme, metabolic shift towards the pentose phosphate pathway for replenishment of NADPH, and transcriptional regulation of compensatory antioxidant response. There are many other similar functional regulatory “antioxidants” and I don’t see why we would want to exclude their reductive capacity from the measurements of the overall redox balance if we want to gain contextual information critical for understanding the response of the individual cellular redox regulation subsystems. For example, how is measuring the amount of phenolic compounds commonly used as a proxy for antioxidant capacity more informative of the overall redox balance of the sample? Finally, I didn’t address the MnO2 and the phosphate buffer comment. First, MnO2 is never in contact with the phosphate buffer in the proposed method so I don’t see how this is relevant? Second, the whole idea of using the nitrocellulose was to trap the MnO2 precipitate. The precipitate doesn’t go away at all. When I first started to play around with the concept of NRP I left the membrane under a running tap and the precipitate still didn’t go away after extensive forceful washing. Even if the membrane with the precipitate was washed in the phosphate buffer I doubt that the precipitate would go away. Proteins are also soluble in different buffers you emerge your membrane following the SDS-PAGE and subsequent techniques (eg. Western blot). Nevertheless, after proper fixation they don’t really float away just like that. I’ll try to do the NRP and wash away the MnO2 precipitate tomorrow and let you know the results. Most of what I wrote is stated in the manuscript and the supplements, however, I hope I made the whole idea a bit more clear. Extensive validation tests suggest that the method is indicative of the change in the antioxidant capacity of biological samples. Nevertheless if you have the idea what tests might provide an important additional layer of validation for the proposed method, I’d be happy to do them to either support or dismiss the current findings and validity of the proposed method. I’d like the method to be the best possible and validated and verified by others (this is why I started this discussion in the first place) so constructive criticism and ideas are very welcome. If you feel that I didn’t understand your comments properly, please don’t hesitate to explain them and I will give my best to try to explain the idea of the proposed concept of NRP.

Thank you once again for the comment!

Best, Jan

Dear Jan Homolak , thanks for your comprehensive respond. Please, keep in mind that I'm commonly responding in a very straightforward way. I really appreciate what you have done and my comments are friendly. I went to the "antioxidant" area as a chemist with a deep knowledge in chemistry of free radicals. I believe that any antioxidant assay should be done under physiological conditions (this is the reason why I mentioned phosphate buffer). Before any discussion we have to clarify the terminology. The list of species damaging cell constituents is pretty short and called NORS (nitrogen and oxygen reactive species). It includes not only radicals but also peroxides. For "antioxidant assay", preferably NORS should be used. The "antioxidant" is a molecule, which stoichiometricaly reacts with NORS and forms harmless product. The system protecting against NORS consists of several subsystems. 1. the antioxidants themselves; 2. enzymes, which catalytically (meaning not consumed in the reaction) remove NORS, such as superoxide dismutase; 3. The systems, which remove pro-oxidants, such as ceruloplasmine; 4. the systems, which repairs damages caused by NORS. There are two different types of antioxidant assays: 1) total amounts of antioxidants; 2) the rates of the reactions between NORS and antioxidants. For the first type of assay you need to choose an oxidant with a "cut-off" reactivity, which reacts only with antioxidants. The good choice is ABTS. For the second type, you need to use one of NORS, which are difficult to generate cleanly. Also, there are water and lipid soluble antioxidants. As mentioned above, the antioxidant test must be performed under physiological condition, e.g. in water at pH around 7 and temperature between 20 and 44C. Most of known assays ignore this requirement. There exist the tests, which measure the antioxidant status. In other words, the overall performance of all antioxidant systems, e.g. the ratio of R-S/RSSR, or the level of ascorbic acid in blood.

Based on the above, I don't think that permanganate is a right choice for antioxidants. The concept of the "reductive capacity" is not well developed. I don't see the link between the reductive capacity with a health status of living cells.

This is not an answer to your post or to your comments. just my personal interpretation of common terms,

Yurii

In my opinion, there are no simple methods to evaluate antioxidizing capacities of potential antioxidants. General tests like DPPH, ABTS... only (roughly) measure the ability of the antioxidants to scavenge free radicals. An antioxidant is much more than that; it can act as a scavenger of free radicals, repair of the oxidized species (reduction by one electron transfer) or by cascade (cumulative) effect . Furthermore, the antioxidant capacity can also depend on the agent responsible for the oxidative stress conditions (selective antioxidants).I recommend the evaluation of the mutual behavior of both antioxidant and the target to be protected in every specific oxidizing environment.For details see e.g.:

• Santos, P.M.P.; Telo, J.P.; Vieira, A.J.S.C. "Structure and Redox Properties of Radicals Derived from One-electron Oxidised Methylxanthines", Redox Report (2008), 13(3), 123; DOI: 10.1179/135100008X259231
• Santos, P.M.P.; Antunes, A.M.M.; Noronha, J.; Fernandes, E.; Vieira, A.J.S.C. "Scavenging activity of aminoantipyrines against hydroxyl radical", Eur. J. Med. Chem. (2010), 45, 2258-2264; DOI: 10.1016/j.ejmech.2010.01.071
• Santos, P.M.P.; SILVA, S.A.G., JUSTINO, G.C.; VIEIRA, A.J.S.C. "Demethylation of Theophylline (1,3-Dimethylxanthine) to 1-Methylxanthine: the First Step of an Antioxidising Cascade", Redox Report (2010), 15(3), 138; DOI: 10.1179/174329210X12650506623726
• SANTOS, P.M.P.; VIEIRA, A.J.S.C. “Antioxidising activity of cinnamic acid derivatives against oxidative stress induced by oxidising radicals”, J. Phys. Org. Chem. (2013), 26, 432; DOI: 10.1002/poc.3104
• VIEIRA, A.J.S.C.; SANTOS, P.M.P. “A Tentative Classification of Antioxidants: Which Role They Play when Protecting Biological Targets from Oxidative Stress Induced Damage?”, J Med Chem Drug Des (2017), 1, 1-3; DOI: 10.16966/jmcdd.101
• VIEIRA, A.J.S.C.; GASPAR, E.M.; SANTOS, P.M.P. “Mechanisms of potential antioxidant activity of caffeine”, Rad. Phys. Chem. (2020), 174, 108968; DOI 10.1016/j.radphyschem.2020.108968