Introduction

Oxidative stress has been implicated in numerous diseases. Reactive oxygen and nitrogen species, such as superoxide and peroxynitrite, play a major role in oxidative stress, as well as in signal transduction pathways under physiological and pathological conditions [1–5]. As their removal is redox-based, there is a growing interest in developing redox-active metal complexes as prospective drugs [1–3]. Among the most widely studied are Mn(II) cyclic polyamines, Mn(III) salen derivatives, and Mn(III) porphyrins. We have been studying metalloporphyrins as effective modulators of redox-sensitive pathways in vivo for more than a decade; these studies aimed not only at developing powerful catalysts of O2 •− dismutation, but also at understanding the structure–activity relationships (SAR) governing such processes in vitro and in vivo [6–10]. On the basis of SAR for this class of compounds (see below), it became apparent why Mn(III) cationic ortho-alkylpyridylporphyrins possess antioxidant capacity and are of remarkable potency in vitro and in vivo [6–10]. These same SAR studies that led to the design of some of the most potent functional superoxide dismutase (SOD) mimics described so far [11] also gave us the first indication that a widely used, negatively charged porphyrin, namely, manganese(III) 5,10,15,20-tetrakis(4-benzoic acid)porphyrin, [MnIIITBAP]3−,Footnote 1 should not be expected to be a good SOD mimic. Yet, the beneficial effects of [MnIIITBAP]3− as a mechanistic probe and therapeutic drug have been reported in hundreds of studies involving numerous models of oxidative stress and oxidative-stress-related diseases. These conflicting aspects of [MnIIITBAP]3− prompted us to undertake a thorough study on this porphyrin to clarify the nature of the effects observed in the biological systems, which have usually been attributed to a putative “SOD-like activity.” We show here that pure [MnIIITBAP]3− has no SOD-like activity in aqueous systems and discuss the impact of these findings on the biological systems.

On thermodynamic grounds, a compound that is able to dismute O2 •− in an aqueous medium must have a reduction potential in-between the one-electron reduction potential for oxygen (E°′ = −160 mV vs the normal hydrogen electrode, NHE; 1 M O2/O2 •−, pH 7) and the one-electron reduction potential of superoxide (E°′ = +890 mV vs NHE; 1 M O2 •−, 2H+/H2O2, pH 7) [12, 13]. The optimal dismuting ability is expected when the potential is midway between these two potentials, which provides an equal driving force for both half-reactions of the catalytic cycle (Fig. 1). This is indeed the case with the SOD enzymes themselves (E ½ being approximately +300 mV vs NHE [14]), for which the rate constants for the oxidation and reduction of O2 •− by the reduced and oxidized forms of SOD, respectively, are identical and reach diffusion limits (k cat = k ox = k red ∼ 2 × 109 M−1 s−1) [15, 16]. A correlation between the metal-centered reduction potential and the catalytic rate constant for the O2 •− dismutation was also found for Fe and Mn porphyrins and a SAR was established (Fig. 2); for this class of compounds, the metal reduction is considered the rate-limiting step. Mn(III) complexes of nonsubstituted tetraarylporphyrins, such as manganese(III) 5,10,15,20-tetrakis(2-pyridyl)porphyrin (MnT-2-PyP+) and manganese(III) 5,10,15,20-tetraphenylporphyrin (MnTPP+), have very negative MnIII/MnII potentials (E ½ = −280 and −270 mV vs NHE, respectively) and are, thus, not reduced by O2 •− in aqueous systems [9]. Modulation of SOD activity has been achieved by tuning the metal-centered reduction potential of metalloporphyrins via modification of the porphyrin ring and/or by tuning electrostatic facilitation. The presence of electron-withdrawing groups in the meso and/or β-pyrrole positions decreases the electron density of the metal site, increasing the MnIII/MnII potential and facilitating, therefore, its reduction. The presence of positive charges close to the metal center is also important in directing O2 •− to the catalytic site via electrostatic facilitation [8]. The alkylation of MnT-2-PyP+, for example, combines both thermodynamic and electrostatic optimizations and yielded compounds, such as manganese(III) 5,10,15,20-tetrakis(N-ethylpyridinium-2-yl)porphyrin (MnTE-2-PyP5+; E ½ = +228 mV vs NHE; log k cat = 7.76) and manganese(III) 5,10,15,20-tetrakis(N-n-hexylpyridinium-2-yl)porphyrin (MnTnHex-2-PyP5+; E ½ = +314 mV vs NHE; log k cat = 7.48), that have E ½ close to the reduction potential of the SOD enzyme itself and excellent in vitro and in vivo SOD-like activity [10, 11, 17]. Of note, MnTnHex-2-PyP5+ is active in ameliorating renal injury and mitochondrial damage during ischemia/reperfusion at single dose as low as 50 μg kg−1, which is one of the lowest (if not the lowest) doses among antioxidants commonly used in animal models of oxidative stress [11].

Fig. 1
figure 1

Design of potent superoxide dismutase (SOD) mimics based on thermodynamic grounds. Reduction potentials of selected Mn porphyrins (MnIII/MnII couple), the SOD enzymes, and relevant species for superoxide dismutation half-reactions

Full size image
Fig. 2
figure 2

Structure–activity relationship of Mn porphyrins as SOD mimics. Dependence of the SOD activity on the MnIII/II reduction potential of cationic and anionic Mn porphyrins (circles); value of the SOD enzymes is given for comparison (triangle). Data compiled from [7–10]

Full size image

Our current knowledge of the design and SAR for the SOD activity of metalloporphyrins cannot account for the in vitro and in vivo effects reportedly attributed to [MnIIITBAP]3− as an SOD mimic. On thermodynamic grounds, the reaction between [MnIIITBAP]3− (E ½ = −194 mV vs NHE [9]) and O2 •− in aqueous media is an unfavorable, nonspontaneous process ([MnIIITBAP]3− + O2 •− = [MnIITBAP]4− + O2; ΔE 0 = −34 mV, ΔG 0 = +3.3 kJ). In contrast, for example, the ΔE 0 values for the oxidation of O2 •− by [MnIIITE-2-PyP]5+ and the reduction of O2 •− by [MnIITE-2-PyP]4+ (+388 and +662 mV, respectively) indicate that the reactions are spontaneous and highly favorable. Of note, the corresponding values calculated for the SOD enzymes are approximately +460 and +590 mV, respectively. The negative charges on [MnIII/IITBAP]3−/4− would not facilitate the O2 •− dismutation on electrostatic grounds either. Given the overwhelming number of studies that have used [MnIIITBAP]3− for mechanistic purposes and as potential therapeutics, it is important to understand from where such effects arise or how could they be justified.

Herein we describe a comparative study of several commercial samples of [MnIIITBAP]3− of common sources (Alexis, Calbiochem, MidCentury Chemicals, and Porphyrin Products), along with our own [MnIIITBAP]3− preparation. A detailed chemical and biochemical characterization of these samples was undertaken and all commercial samples were found to contain varying degrees of noninnocent trace impurities. In a narrow sense, our main conclusion is that pure [MnIIITBAP]3− shows no SOD-like activity in aqueous media. We, however, introduce here the possibility that [MnIIITBAP]3− may show some O2 •− scavenging capability in organic media, such as those of cellular membranes; and this is an aspect that has been highly overlooked in the porphyrin-based studies in general (including ours). The wider implications of our results may potentially impact many studies that derived insights into the biological role of catalytic antioxidants from data using [MnIIITBAP]3−; some of these implications are addressed accordingly.

Materials and methods

General

[MnIIITBAP]3− samples were either purchased from Alexis, Calbiochem, MidCentury Chemicals, or Porphyrin Products, or prepared in-house (see later). Commercial [MnIIITBAP]3− samples were used as received. Xanthine (X) and equine ferricytochrome c (lot no. 7752) were from Sigma, whereas xanthine oxidase (XO) was prepared by R. Wiley [18] and was a gift from K.V. Rajagopalan. Pyrrole (Aldrich) was freshly distilled from CaH2 (Alfa Inorganics Ventron) in the dark under reduced pressure. Anhydrous N,N-dimethylformamide (DMF; Aldrich) was kept over 4 Å molecular sieves. Other chemicals and solvents were of reagent grade or higher from Fisher Scientific, Mallinkrodt, or Sigma-Aldrich, and were used as supplied.

UV–vis measurements were performed using a Shimadzu UV-2501PC spectrophotometer at 0.5 nm resolution. Nuclear magnetic resonance (NMR) data were recorded with a Varian Unity 500 MHz instrument (499.92 MHz for 1H) at 25 °C in CDCl3 (Acros Organics); the residual solvent proton at δ 7.26 was used as the reference for 1H NMR spectra. Elemental analyses were conducted by Atlantic Microlabs (Norcross, GA, USA).

Electrospray ionization (ESI) mass spectrometry (MS) analyses and ESI tandem MS (MS/MS) analyses were performed on using an Applied Biosystems MDS Sciex 3200 Q Trap liquid chromatography/MS/MS spectrometer at Duke Comprehensive Cancer Center, Shared Resources. All [MnIIITBAP]3− samples were prepared with the exact same solvents, analyzed in parallel under the same experimental conditions; before any [MnIIITBAP]3− analysis, the instrument was purged with water and then with a H2O/MeCN/HCO2H (40:40:1) mixture, and a baseline (control) was recorded to ensure there was no cross-contamination.

Electrochemical measurements were carried out using a model 600 voltammetric analyzer (CH Instruments), using a glassy carbon working electrode, a standard Ag/AgCl reference electrode, and a platinum wire as the auxiliary electrode. Samples (approximately 0.5 mM) were prepared in 0.05 M phosphate buffer (pH 7.8) containing 0.1 M NaCl as electrolyte (as described in [10]), or in DMF containing 0.1 M tetra(n-butyl)ammonium perchlorate as an electrolyte. Ferrocene (Fc) was used as a reference in DMF measurements, according to IUPAC recommendations for measurements in nonaqueous media [19]. No significant differences in E ½ were observed using Fc as either an internal or an external standard. Potentials against Fc+/Fc were converted to the NHE scale considering an E ½ for the couple Fc+/Fc of +400 mV versus NHE [20].

Preparation of [MnIIITBAP]3−

The synthesis of [MnIIITBAP]3− was accomplished in three steps: (1) preparation of the free-base porphyrin in the tetramethyl ester form [5,10,15,20-tetrakis(4-benzoic acid)porphyrin tetramethyl ester, H2TBAP-TME], (2) its Mn metallation, and (3) ester hydrolysis.

Preparation of H2TBAP-TME

H2TBAP-TME was prepared using an adaptation of the method of Johnstone et al. [21]. Pyrrole (0.70 mL, 10 mmol) and methyl 4-formylbenzoate (1.64 g, 10 mmol) were refluxed in a 50 mL propionic acid/PhNO2 (7:3) solution for 1 h. After the reaction mixture had been cooled to room temperature (approximately 20 °C), 50 mL of MeOH was added and the mixture was left to stand for approximately 1 h. The crude porphyrin was filtered, washed with MeOH (until the washings were colorless), and suction-dried. The solid was then dissolved in CH2Cl2 and purified by column chromatography (basic Al2O3, CH2Cl2). The first fraction was collected and evaporated to dryness. The pure porphyrin was further dried in a vacuum oven at 90 °C overnight. No chlorin was observed either spectroscopically or chromatographically. Yield: 515 mg (24%). Anal. Calcd for H2TBAP-TME, C52H38N4O8: C, 73.75; H, 4.52; N, 6.62. Found: C, 73.87; H, 4.42; N, 6.65. UV–vis (CHCl3): 376 nm (log ε/M−1 cm−1, 4.48), 400 (sh, 4.96), 421 (5.72), 485 (sh, 3.65), 516 (4.35), 552 (4.00), 590 (3.84), 645 (3.61). 1H NMR (CDCl3): δ 8.82 (s, 8 H, β-pyrrole), 8.45 (AA′BB′ pattern, 8 H, app J HH = 8.05 Hz, o-C6H4CO2Me or m-C6H4CO2Me), 8.30 (AA′BB′ pattern, 8 H, app J HH = 8.05 Hz, o-C6H4CO2Me or m-C6H4CO2Me), 4.11 (s, 12 H, p-C6H4CO2CH3), −2.81 (br s, 2 H, NH). ESI–MS (54:27:1 MeOH/CHCl3/HCO2H): m/z 847 (100%, [H2TBAP-TME + H]+), 424 (34%, [H2TBAP-TME + 2 H]2+/2). 1H NMR and UV–vis data are in agreement with those previously reported [22, 23].

Preparation of manganese(III) 5,10,15,20-tetrakis(4-benzoic acid)porphyrin tetramethyl ester chloride

The Mn metallation of H2TBAP-TME was carried out using a procedure adapted from that described for related porphyrins [24]. To a refluxing solution of H2TBAP-TME (130.0 mg, 153.5 μmol) in 60 mL of a CHCl3/MeOH (2:1) mixture, 1 mL of 2,6-dimethylpyridine (2,6-lutidine) and 32 equiv of Mn(OAc)2·4 H2O (1.20 g, 4.9 mmol) were added. The progress of the reaction was monitored by UV–vis spectroscopy. After 2 h of reflux, the Mn metallation became slow and UV–vis spectroscopy indicated approximately 80% conversion. Heating was then discontinued, and the solvent was removed on a rotary evaporator. The solid (with a strong pyridine smell) was dissolved in 20 mL of CHCl3; the resulting solution was thoroughly washed free of residual 2,6-lutidine with 5 × 20 mL of 1 M HCl(aq), and then with several small portions of H2O until neutral pH was achieved. The solution in CHCl3 was dried with anhydrous Na2SO4, concentrated in a rotary evaporator, and chromatographed on a neutral Al2O3 column using CHCl3 as the eluent. Unreacted H2TBAP-TME (23.6 mg, 18%) was eluted first as a purple band, followed by the desired product, which was eluted as a brownish green band. The Mn porphyrin fractions were combined, washed first with 1 M HCl(aq) (2 × 25 mL), then with H2O (3 × 25 mL), and were evaporated to dryness on a rotary evaporator. The resulting green solid was further dried in a vacuum oven at 100 °C overnight. Yield: 107 mg (75%). Anal. Calcd for manganese(III) 5,10,15,20-tetrakis(4-benzoic acid)porphyrin tetramethyl ester chloride [Mn(TBAP-TME)Cl], C52H36N4O8ClMn: C, 66.78; H, 3.88; N, 5.99. Found: C, 66.35; H, 3.91; N, 6.01. UV–vis (CHCl3): 280 nm (log ε/M−1 cm−1, 4.43), 323 (sh, 4.48), 344 (sh, 4.68), 376 (4.83), 401 (4.73), 432 (sh, 4.34), 448 (sh, 4.20), 479 (5.13), 528 (3.81), 582 (4.07), 617 (4.07), 7.68 (2.93). UV–vis (MeOH): 232 nm (log ε/M−1 cm−1, 4.96), 291 (4.48), 318 (sh, 4.52), 352 (sh, 4.63), 378 (4.80), 399 (4.82), 418 (sh, 4.65), 439 (4.44), 466 (5.11), 515 (3.89), 563 (4.15), 596 (3.98), 683 (sh, 3.17), 708 (3.18), 774 (3.26), 807 (sh, 3.08). ESI–MS (MeOH): clusters centered at m/z 899 (100%, [Mn(TBAP-TME)Cl − Cl]+; isotope pattern confirmed by simulation).

Preparation of [MnIIITBAP]3−

The hydrolysis of the tetramethyl ester Mn complex was carried out via a modification of the procedure described by Hambright [25] for Ru porphyrin complexes. To a solution of Mn(TBAP-TME)Cl (20.0 mg, 21.4 μmol) in tetrahydrofuran (10 mL), 10 mL of 1 M NaOH(aq) solution was added. The biphasic system was magnetically stirred at room temperature for 2 days, during which time the hydrolyzed Mn porphyrin was being extracted into the aqueous phase, leaving the organic phase nearly colorless. Stirring was discontinued, the system decanted, and the organic phase collected. The dropwise addition of approximately 20 mL of 1 M HCl(aq) to the well-stirred collected aqueous phase resulted in the precipitation of crude [MnIIITBAP]3−. The precipitate was filtered off using a fine fritted funnel and washed with 5 mL of 0.1 M aqueous HCl and 10 mL of water. The resulting solid was dissolved in 0.1 M NaOH(aq) (10 mL) and reprecipitated with approximately 11 mL of 0.1 M HCl(aq). The suspension was filtered and the solid was washed with 0.1 M HCl(aq) (5 mL) and water (10 mL). This procedure of NaOH dissolution and HCl precipitation was repeated four more times. The final [MnIIITBAP]3− precipitate was washed with H2O (5 × 5 mL, until the filtrate pH was neutral), suction-dried (using paper-filtered air) for 1 day, and vacuum-oven-dried at 40 °C overnight. Yield: 15.3 mg (79%). Anal. Calcd for H3[MnIIITBAP]·1.7 H2O·0.5 NaCl, C48H30.4N4O9.7Na0.5Cl0.5Mn: C, 63.88; H, 3.40; N, 6.21. Found: C, 63.89; H, 3.44; N, 6.17. UV–vis (0.05 M phosphate buffer, pH 7.8): 230 nm (log ε/M−1 cm−1, 4.93), 290 (4.49), 381 (4.84), 401 (4.84), 420 (sh, 4.70), 468 (5.04), 515 (3.92), 566 (4.16), 599 (4.07), 684 (sh, 3.23), 712 (sh, 3.20), 780 (3.24), 811 (sh, 3.17). ESI–MS (40:40:1 H2O:MeCN:HCO2H): clusters centered at m/z 843 (100%, [MnTBAP3− + 4H+]+; isotope pattern confirmed by simulation). ESI–MS/MS analysis of the m/z 843 ion yield clusters (less than 20% abundance) centered at 786.6 (MnTBAP–CO2), 752.8 (MnTBAP–2CO2), 720.8 (MnTBAP–C6H4CO2H), 708.8 (MnTBAP–3CO2), 676.8 (MnTBAP–C6H4CO2H–CO2), 631.0 (MnTBAP–C6H4CO2H–2CO2), 586.8 (MnTBAP–C6H4CO2H–3CO2), 557.0 [MnTBAP–2(C6H4CO2H)–CO2], and 510.8 [MnTBAP–2(C6H4CO2H)–2CO2].

SOD-like activity assay and XO inhibition assay

The catalytic rate constants for the O2 •− dismutation were determined by cytochrome c assay as previously described [26]. X/XO (40 μM X, approximately 2 nM XO) was the source of O2 •−, and ferricytochrome c was used as the indicating scavenger of O2 •−. The reduction of cytochrome c was followed at 550 nm. Assays were conducted at 25 ± 1 °C in 0.05 M phosphate buffer, pH 7.8, 0.1 mM sodium salt of ethylenediaminetetraacetic acid (EDTA) in the presence or absence of 15 μg mL−1 catalase. Rate constants for the reaction of MnTBAP with O2 •− were based upon the competition with 10 μM cytochrome c (k cyt c  = 2.6 × 105 M−1 s−1), as described elsewhere [26]. The O2 •− was produced at 1.2 μM min−1. The k cat value was calculated based on the concentration that caused 50% inhibition of the cytochrome c reduction, which was obtained from the plot of v 0/v i − 1 vs Mn porphyrin concentration (described in detail in [26]). As the X/XO system is the generator of superoxide, it is imperative to check the effect of any compound on the ability to interfere with superoxide production. The inhibition of O2 •− production by X/XO was examined following urate formation at 295 nm in the absence of cytochrome c, all other conditions being the same. None of the MnTBAP preparations caused the reoxidation of cytochrome c with H2O2. This was assessed (1) by adding catalase or (2) in the following way: cytochrome c was fully reduced with sodium dithionite in 0.05 M phosphate buffer, pH 7.8, whereupon Mn porphyrin was added. We have shown previously for three compounds, namely, MnTE-2-PyP5+, manganese biliverdin dimethylester, and MnCl2, that identical results are obtained by pulse radiolysis and cytochrome c assay [26]; additionally, the catalytic rate constant reported for manganese(III) 5,10,15,20-tetrakis(N-methylpyridinium-4-yl)porphyrin (MnTM-4-PyP5+) by stopped flow [27] is in agreement with that determined by the cytochrome c assay [10].

Results and discussion

A total of six [MnIIITBAP]3− samples were analyzed in this work: five commercial ones (from Alexis, Calbiochem, MidCentury Chemicals, and Porphyrin Products—randomly assigned as A, B, C-I, C-II, and D) and one prepared as described herein All samples were analyzed by UV–vis spectroscopy, elemental analysis, MS, and cyclic voltammetry. They were also assayed on SOD-like activity (cytochrome c assay). Since the X/XO system is used to produce superoxide by urate formation, we routinely checked the effect of compounds on their ability to interfere with the X/XO system. Table 1 summarizes the elemental analysis, UV–vis, and MS data, as well as the biochemical parameters. Because very discrepant data were observed with a sample from commercial source C (labeled C-I), the supplier was contacted and promptly provided us with another sample (labeled C-II). Nonetheless, the analyses of sample C-I were included here as an “extreme” case.

Table 1 Chemical and biochemical properties of negatively charged manganese(III) 5,10,15,20-tetrakis(4-benzoic acid)porphyrin ([Mn III TBAP] 3−) samples of different sources
Full size table

Chemical analyses and biochemical assays of the commercial samples of [MnIIITBAP]3− indicated that they contain noninnocent trace impurities that may affect considerably some biological studies if the samples are to be use “as received.” As the definite nature/source of some of the impurities and the synthetic/purification steps to which the samples had been subjected during manufacturing/processing are not available, the design of a thorough purification strategy was hampered. Alternatively, we concentrated efforts on the preparation (and thorough characterization) of a high-purity [MnIIITBAP]3− sample, which is discussed later. It is worth noting that although [MnIIITBAP]3− is one of the most used porphyrin-based mimetics in therapeutics and mechanistic biological studies, most of these utilize commercial samples. The studies in which MnTBAP was prepared “in-house” do not provide nor refer to sources where undisputable chemical characterization data (if any) are presented to support the purity of the isolated compound.

Comparative evaluation of [MnIIITBAP]3− samples; chemical analyses

The UV–vis analyses of the [MnIIITBAP]3− samples showed that in terms of porphyrin-based materials all of them are mostly [MnIIITBAP]3−, as the apparent molar absorptivity values (at 468 nm) for the commercial samples were roughly close to that calculated for the pure [MnIIITBAP]3− prepared in this work and in agreement with those of other well-characterized Mn porphyrins, with the exception of the data for sample C-I (Table 1). UV–vis data for samples A, B, and D revealed, however, contamination with various amounts of the free-base 5,10,15,20-tetrakis(4-benzoic acid)porphyrin (H2TBAP). Comparison between these samples and the spectrum of pure [MnIIITBAP]3− prepared in this work allowed a crude estimation of the free-base contamination in the order of 3, 1, and 4% for samples A, B, and D, respectively (considering a molar absorptivity of the free-base porphyrin of 3.86 × 105 M−1 cm−1 at the Soret band [28]). The UV–vis spectrum for C-I showed a very high, ill-defined baseline, which increased toward the UV region and does not correspond to porphyrin-based material. Conversely, the spectrum of sample C-II was essentially superimposable on that of pure [MnIIITBAP]3−. Difference spectra of the commercial samples relative to that of the one prepared in this work are presented in Fig. 3. Given that free-base porphyrins are well-known photosensitizers, experiments (particularly those involving very sensitive cell cultures) using [MnIIITBAP]3− contaminated with [H2TBAP]4− should have been preferably carried out in the dark to avoid cell sensitization (provided that the free-base porphyrin has no other biological effect in its own right).

Fig. 3
figure 3

Difference UV–vis spectra of the commercial negatively charged manganese(III) 5,10,15,20-tetrakis(4-benzoic acid)porphyrin ([MnIIITBAP]3−) samples relative to that of the pure [MnIIITBAP]3−. All spectra were normalized (prior to subtraction) to the 468 nm band, which corresponds to the Soret band of the [MnIIITBAP]3− sample (2.8 µM [MnIIITBAP]3−, log ε 468 nm = 5.04). Filled circles sample A, inverted triangles sample B, open circles sample C-I (right-hand-side scale), squares sample C-II, diamonds sample D

Full size image

All samples were also subjected to elemental analysis. Interpretation of those data is partially hampered by the fact that we do not have access to the synthetic and purification procedures the commercial samples went through; i.e., any formulation would be purely speculative. However, if one considers that the final step in the purification of [MnIIITBAP]3− may involve some acid precipitation from the aqueous media, the formulation may likely contain some water or small amounts of inorganic salt, as observed in the [MnIIITBAP]3− prepared in this work. Water and inorganic salts, such as NaCl, although changing the absolute percentage of C, H, and N, do not modify the C/N mole ratio of the sample. As long as the commercial samples were indeed precipitated from water, and no organic acids (e.g., acetic acid) or C- or N-containing solvent/acid were used in this last step, the C/N mole ratio may provide a rough estimate of the purity of the sample. There is no correlation between the deviations of the C/N mole ratio of the samples (Table 1) and their levels of contamination with free-base [H2TBAP]4−, indicating that other C- or N-containing impurities/solvents may be present. In fact, even a formulation accounting for the [H2TBAP]4− contamination, as estimated by UV–vis spectroscopy, could not explain the deviation observed for the commercial samples (with the exception of C-II, which showed little C/N deviation and no [H2TBAP]4− contamination).

The analytical technique that provided the most valuable information regarding the overall purity of the samples was ESI-MS. [MnIIITBAP]3− can be detected in positive mode upon the addition of formic acid to the sample, after which the peak corresponding to the fully protonated species [MnTBAP3− + 4H+]+ is observed at m/z 843. Under the ESI–MS experimental conditions, the only sample that showed a peak corresponding to the free-base porphyrin at m/z 791 ([H2TBAP4− + 5H+]+) was sample D, which is incidentally the one with higher free-base porphyrin content as judged by the UV–vis data. Figure 4 shows the ESI–MS spectra of all the samples. From this figure it is evident that the [MnIIITBAP]3− prepared in this work is indeed the purest among the samples in the study here and there is a great variability of impurity content among the commercial samples. As further protonation of [MnTBAP3− + (4 + n)H+]n+ (n > 1) to yield multiply positively charged species is unlikely and not observed (at least in any appreciable amount) in the spectrum of the [MnIIITBAP]3− prepared in this work, it may be inferred, therefore, that peaks in the low m/z region are likely from impurities. Worth mentioning, however, is that the response/sensitivity of ESI–MS to the levels of impurities and that of [MnIIITBAP]3− are likely very different. Addition of HCO2H must increase the contribution of the protonated forms, but still these species are unlikely the major species in solution (otherwise precipitation of the neutral species would take place). Therefore, despite the apparent abundance of impurity peaks in the commercial samples relative to that of [MnIIITBAP]3−, such impurities may still be in trace levels.

Fig. 4
figure 4

Electrospray ionization mass spectrometry spectra of the [MnIIITBAP]3− samples in H2O/MeCN/HCO2H (40:40:1). The peak at m/z 843 corresponds to the Mn porphyrin, whereas the remaining peaks are from impurities (see text)

Full size image

The cyclic voltammograms of the samples were dominated by the MnIII/MnII redox couple of [MnIII/IITBAP]3−/4− (at E ½ = −194 mV vs NHE) and were largely insensitive to the presence/nature of the impurities in these samples. Major differences were observed, however, in the biochemical assays.

Comparative evaluation of [MnIIITBAP]3− samples; biochemical analyses

During the measurement of the SOD-like activity of [MnIIITBAP]3− samples by the cytochrome c assay, it was verified that all of them showed some degree of inhibition of X/XO (the O2 •− generator), as judged by the decrease in the rate of urate formation with increased concentration of [MnIIITBAP]3−. As the measurement of k cat requires a constant generation of superoxide, the SOD activity calculated for each sample was corrected for the corresponding inhibition of XO at each [MnIIITBAP]3− concentration.

The observed concentration that causes 20% inhibition of xanthine oxidase activity (IC20) and the apparent log k cat (uncorrected and corrected values) for the [MnIIITBAP]3− samples are presented in Table 1. Both XO IC20 and k cat values among the samples show differences that exceed an order of magnitude. Such variability is absolutely inconsistent with effects being attributed solely to MnTBAP itself. All the samples tested, despite differences in composition, are mainly MnTBAP and variation in the actual analytical concentration of this species in solutions prepared with the different samples cannot account for such a huge span in IC20 and k cat; unless the species responsible for such a dramatic effect is not [MnIIITBAP]3−, but are trace impurities in the samples. In fact, the impurity levels of the samples differ greatly as observed by ESI–MS, and the (unknown) nature of these impurities among the different preparations may also vary accordingly.

A critical piece of information supporting the argument that the effects measured in these assays may be attributed to trace impurities was the observation of the k cat dependence upon the concentration of the [MnIIITBAP]3− sample stock solution and the incubation time prior to the cytochrome c assay. The “SOD activity” of concentrated stock solutions of [MnIIITBAP]3− (millimolar) in 0.1 mM aqueous EDTA does not change considerably within at least 2 days. However, if the stock is diluted to micromolar level in 0.1 mM aqueous EDTA and left to incubate in the dark for 1 day, a substantial decrease in activity results. Short incubation time (e.g., 30 min) has no appreciable effect on the observed k cat. Importantly, the UV–vis spectra of the samples (regardless of the treatment they were subjected to) remained unaltered, which indicated that there had been no in situ modification/demetallation/decomposition of [MnIIITBAP]3− itself or change in its overall concentration. The long time required for a measurable loss of SOD activity is also relevant. Ligand exchange and reorganization within the coordination sphere of metalloporphyrins are usually fast; in fact porphyrins are known to labilize even commonly substitution-inert metals such as RhIII via an unusual cis effect [29]. Therefore, if the loss of SOD activity were related to some equilibration process involving [MnIIITBAP]3−, this would not be expected to require such a long time (hours) as that observed in the incubation tests. Such data suggest that [MnIIITBAP]3− is unlikely to be solely (if at all) responsible for the XO inhibition and the observed SOD-like activity; these effects arise presumably from the impurities present in the samples. Worth noting is that pure [MnIIITBAP]3− exhibits little XO inhibition, and the apparent SOD activity of this pure sample is essentially an artifact: upon correction for the XO inhibition, the SOD activity (if any) is too low to be measured accurately (log k cat < 3.5). Attempts to determine the exact SOD activity of the pure sample were precluded by its too low value, which implies the use of high concentrations of MnTBAP in the SOD assay and resulted in highly colored solutions of unacceptably high absorbance; this effect is even more pronounced in the determination of the concentration dependence of the XO/X inhibition at 295 nm. Of note is that the concentration values that lead to a 50% reduction of the cytochrome c reduction by O2 •− in the SOD assay are 1.3 and 45 nM for the Cu,Zn-SOD enzyme [15, 16] and MnTE-2-PyP5+ [10], respectively; a very poor compound with log k cat ∼ 4 requires approximately a 200,000- and 6,000-fold more concentrated solution to yield the same effects observed with Cu,Zn-SOD and MnTE-2-PyP5+, respectively.

Insights into the nature of the “SOD-active” impurities

A comparison between the pure [MnIIITBAP]3− sample and sample C-II indicates that the “SOD-active” impurities are not associated with free-base porphyrin. UV–vis spectroscopy shows that both samples have the same composition in terms of porphyrin-based compounds (and no evident H2TBAP), but only sample C-II has a measurable k cat, which is incidentally of similar magnitude to those measured for the other commercial samples that definitely contain some free-base porphyrin. Additionally, free-base porphyrins do not have redox-active couples consistent with superoxide redox processes under biological conditions and are not likely to be involved in very slow processes/equilibria at low concentration.

Probable redox-active impurities are perhaps those associated with residual Mn species reminiscent of the Mn metallation step and/or inadequate/insufficient purification procedures. The time- and concentration-dependence for the loss of SOD activity within the commercial samples suggests that these active impurities may be some Mn clusters/aggregates, whose likely ligands are supposedly oxo, hydroxo, and/or carboxylato (e.g., acetate from the Mn precursor, or benzoate from the porphyrin moiety). Declustering/deaggregation of clusters and aggregates in general is facilitated by dilution. This process would eventually lead to simpler Mn species, which may then be scavenged by EDTA; this is in agreement with the experimental observations for the commercial [MnIIITBAP]3− samples. Of note is that control experiments with Mn-EDTA showed no SOD activity [30]. In practical sense, a direct implication of these findings is that the use of Chelex (EDTA-like resins) [26] for “purification” of MnTBAP samples may not be suitable at millimolar concentrations; conversely, the use of micromolar concentrations may eliminate some of the SOD-active contaminations but this would require both an unreasonable amount of time (days) of incubation and a cumbersome workup procedure to yield stock solutions at appropriate concentration for SOD assaying and general use (e.g., for the preparation of 10 mg of MnTBAP, one would need to work up volumes of approximately 12 L at 1 μM concentration).

MS analyses also provided some evidence for the presence of residual Mn cluster/aggregate species in the commercial samples. Two recurring, prominent peaks in the ESI–MS spectra of the commercial [MnIIITBAP]3− samples were that of m/z 381 and 341, which could be tentatively assigned to [Mn2(OAc)4(OH)2 + H]+ and [Mn2(OAc)3(H2O)(OH) + H]+ fragments, respectively. ESI-MS/MS analyses of these ions revealed that the m/z 341 ion-product does not yield ion-daughters, while the ion m/z 381 fragments further to give ion-daughter clusters at m/z 89.2, 91.1, 189.9–193.1, and 241.1–245.3, which may be associated with species such as Mn(OH)2, Mn(OAc)(OH)3(H2O), Mn(OAc)2(H2O), and Mn(OAc)2(H2O)4, respectively. ESI–MS/MS analysis of the m/z 381 ion-product under mother-scan mode indicates that this species comes from an m/z 739 ion. Optimization of the m/z 739 ion and MS/MS shows a prominent m/z 381 peak, but the remaining mass loss (of 358) is not observed, which suggests that this mass corresponds to either a neutral species (and, therefore, is not detected) or fragments of m/z smaller than 50 (which are below the minimum mass limit of the instrument). A neutral loss analysis for an m/z 358 fragment indicates the m/z 1,201 ion as the mother ion, which upon intensity optimization and MS/MS analysis yields the [MnTBAP3− + 4H+]+ ion at m/z 843 as the prominent peak and low-abundance clusters corresponding to the fragmentation of the porphyrin moiety. If the tentative assignment of the m/z 381 and 341 ions as manganese acetate/oxo/hydroxo clusters is correct, the association of the Mn cluster impurities with [MnIIITBAP]3− (which is suggested by the m/z 1,201 ion) may happen via interactions with the porphyrin benzoate moieties, as they may be able to replace some of the acetate groups of the putative Mn clusters. Carboxylate groups are ubiquitous ligands in chemistry and biology. Although the peaks at m/z 381 and 341 may be assigned to homobimetallic Mn species, the presence of high molecular mass clusters (that may yield these smaller fragments under ESI–MS conditions) cannot be ruled out at this point.

Overall, these data suggest that the commercial samples contain [MnIIITBAP]3− as the major component, and likely “SOD-active” Mn clusters, such as manganese oxo/hydroxoacetate and/or manganese oxo/hydroxoacetate–MnTBAP species, as trace impurities. Bimetallic and polymetallic Mn clusters are plausible candidates and are of chemical, physical, and biological interest. Mn clusters are known to be involved in crucial electron-transfer steps in photosynthesis [31] and oxo/hydroxoacetate compounds containing up to 12 Mn atoms (such as Mn12(RCOO)16(H2O)4O12; R is methyl or phenyl) have remarkable magnetic properties [32]. Commercially available [MnIIITBAP]3− does not have the same chemical and biological characteristics; such differences are likely a consequence of different synthetic and purification procedures. Attempts to isolate and fully characterize the impurities of commercial [MnIIITBAP]3− were not considered, as they lose activity with time in solution; instead, great effort was concentrated on preparing a high-purity [MnIIITBAP]3− sample and this is described next.

Synthesis of a pure [MnIIITBAP]3− sample

Two strategies were envisioned for preparing [MnIIITBAP]3−: (1) direct Mn metallation of the free-base [H2TBAP]4−; and (2) Mn metallation of the ester form of [H2TBAP]4− (H2TBAP-TME), followed by hydrolysis of the ester moieties [33]. All attempts to obtain a pure sample of [MnIIITBAP]3− (free from noninnocent trace impurities) using route 1 were unsuccessful. Mn metallation of porphyrins in general can be accomplished in DMF, but because free-base porphyrins that bear some reactive substituents are reportedly sensitive to this solvent at refluxing conditions [34, 35], and the possibility of side reactions between DMF and [H2TBAP]4− could not be neglected (e.g., amidation of the carboxylate groups by HNMe2; as some DMF invariably decomposes to CO and HNMe2 at high temperatures), metallation via the traditional DMF route was avoided. Attempts to metallate [H2TBAP]4− in water [36] resulted in samples that contained free-base porphyrin (approximately 6%, as revealed by UV–vis spectroscopy) and low molar absorptivity (comparable to those of the commercial samples). A convenient route for separation of [H2TBAP]4− and [MnIIITBAP]3− has not been established.

The general route 2 (Scheme 1) has been suggested by Bommer and Hambright [33] as the cleanest one to prepare [H2TBAP]4− derivatives and was successfully adapted to the synthesis of [MnIIITBAP]3−. The preparation of H2TBAP-TME via condensation of equimolar amounts of pyrrole and 4-carbomethoxyphenyl aldehyde was accomplished in high purity using a modification of the method of Johnstone et al. [21], which yielded a chlorin-free porphyrin. A higher-yield synthesis with no loss in purity is also achievable under the conditions of Lindsey et al. [22], but the method of Johnstone et al. was chosen for its generally easier experimental set-up. The Mn metallation of H2TBAP-TME with Mn(OAc)2 in a chloroform/methanol mixture containing 2,6-lutidine as an acid scavenger [24] yielded the corresponding Mn complex, Mn(TBAP-TME)+, which was stripped from the excess 2,6-lutidine with HCl extraction and purified by chromatography on alumina. Because some ligand exchange (Cl− to OH−) may take place on alumina, the fraction containing the Mn(TBAP-TME)+ was washed again with HCl; this afforded a material of known composition to facilitate the analysis.

Scheme 1
scheme 1

Preparation of negatively charged manganese(III) 5,10,15,20-tetrakis(4-benzoic acid)porphyrin

Full size image

The preparation of [MnIIITBAP]3− was accomplished via hydrolysis of its ester precursor with 1 M NaOH(aq). Mn(TBAP-TME)Cl was dissolved in tetrahydrofuran and to this solution 1 M NaOH was layered. The vessel was capped and left under vigorous stirring for 2 days, during which period the hydrolyzed Mn porphyrin transferred from the organic to the aqueous phase. At the end of the reaction, after phase separation, the organic phase collected was essentially colorless. Purification of the aqueous-phase [MnIIITBAP]3− was carried out by successive precipitation/solubilization/reprecipitation cycles. Protonation of the carboxylic acid moieties at low pH leads to precipitation of the Mn porphyrin from aqueous media and the precipitate can be readily solubilized in aqueous NaOH, where the sodium salt of [MnIIITBAP]3− is prevalent.

All three compounds prepared within route 2 (i.e., H2TBAP-TME, Mn(TBAP-TME)Cl, and [MnIIITBAP]3−) gave satisfactory spectroscopic, spectrometric, and elemental analyses. This route has yielded the purest [MnIIITBAP]3− described so far. The UV–vis spectrum of this sample is illustrated in Fig. 5, while the ESI–MS spectrum is presented in Fig. 4 and the cyclic voltammogram is given in Fig. 6.

Fig. 5
figure 5

UV–vis spectrum of MnTBAP in 0.05 M phosphate buffer (pH 7.8). The scale for the expanded visible region is given on the right-hand side

Full size image
Fig. 6
figure 6

Cyclic voltammograms for the redox processes of the MnIII/MnII couple of [MnIIITBAP]3− and Mn(TBAP-TME)Cl and the Fc+/Fc couple in N,N-dimethylformamide; scan rate 0.1 V s−1. Fc ferrocene

Full size image

Biological repercussions

The results of the biochemical assays vary greatly with the source of [MnIIITBAP]3− used. Our pure [MnIIITBAP]3− preparation does not show detectable SOD activity and has the least inhibitory effect on the XO enzyme. As a pure sample has no ability to dismute O2 •−, which would be thermodynamically expected (given the low reduction potential of [MnIIITBAP]3−), [MnIIITBAP]3− is not a functional SOD mimic. Of note, as the SOD activity of Mn porphyrins parallels their peroxynitrite reducing ability [37, 38], is that this further casts doubts on [MnIIITBAP]3− as a good ONOO− scavenger.

The coupling of reductants to the Mn porphyrin system, if possible, would allow for a subsequent reduction step of O2 •− by Mn(II) porphyrin in a superoxide reductase fashion [39]. While this cannot be rule out in the case of Mn porphyrins that exhibit high MnIII/MnII reduction potential [40], such reactivity pattern in the [MnIIITBAP]3− case is least likely under biological settings. All attempts to reduce [MnIIITBAP]3− with common cellular reductants, such as ascorbic acid, glutathione, tetrahydrobiopterin, or NAD(P)H (anaerobically), were unsuccessful (unpublished data). Moreover, submitochondrial particles, i.e., components of the mitochondrial respiratory chain, are also not able to reduce [MnIIITBAP]3− [41]; only stronger reductants, such as sodium dithionite (Na2S2O4), were able to easily effect such a reduction.

Despite these observations, the amount of in vivo and in vitro data on the so-called SOD activity of (commercial) [MnIIITBAP]3− is overwhelming. Whether these data relate to a true [MnIIITBAP]3− effect via some nonclassical mechanism (porphyrins are involved in a multitude of biological processes) or via the effect of impurities in the samples remains an open issue. Considerable SOD-like activity of non-porphyrin Mn complexes containing carboxylic acids as ancillary ligands have been reported [29], and even ordinary Mn2+ salts, such as MnCl2, are able to rescue SOD-deficient and oxidative-stress models [42–45] at high concentration (consistent with the in vitro SOD activity of MnCl2; log k cat ∼ 6.2) [26]. Worth mentioning is also the remarkable SOD activity of some transition-metal oxides/clusters, such as that of cerium nanoparticules [46] and osmium tetroxide solutions [47], the former showing promising in vivo antioxidant properties [46], and the latter being of clinical use for almost half a century for treating arthritis [47].

Whereas the [MnIIITBAP]3− impurities may have some biologically relevant SOD-like activity, their action and that of [MnIIITBAP]3− may also involve the inhibition of enzymes, just as was observed with XO in this work. Although some XO inhibition by [MnIIITBAP]3− itself cannot be completely ruled out at this stage, the high inhibitory effect observed with the use of some commercial samples is unlikely to be associated solely with the porphyrin complex. If the XO inhibition is largely due to impurities, this implies that, whatever these impurities are, they are indeed good inhibitors of XO. The IC20 values reported in Table 1 are relative to the concentration of [MnIIITBAP]3− and if the impurities were in a concentration of at least an order of magnitude lower than that of [MnIIITBAP]3− itself, the actual IC20 values relative to the impurities should also be (at least) an order of magnitude lower. The implication of these findings is that the use of commercial [MnIIITBAP]3− samples (which have low apparent IC20) in biological studies might possibly alter the oxidative stress in vitro and in vivo via inhibition of (for example, but not limited to) XO, decreasing therefore the levels of O2 •− and its progeny. The effect of these impurities on other enzymes remains unknown and the impact of the XO inhibition on the overall conclusions of any particular study using these commercial [MnIIITBAP]3− samples would need to be assessed/investigated very carefully.

If we speculatively assume that most of the reported in vivo effects of “MnTBAP” do not arise from its contaminants (and such a hypothesis is not amenable to checking with current literature data), alternatives to account for such effects need to be explored. We first considered the possibility of in situ derivatization of MnTBAP via reactions readily accessible biologically, such as protonation, esterification, or amidation of the carboxylic moieties to yield the acid, ester, or peptide derivatives, respectively. Hammett constant values [48] for protonated carboxylic acid (σ p = 0.45), ester (σ p = 0.45), and amide (σ p = 0.36) relative to the carboxylate substituent (σ p = 0.00) indicate that an increase of the MnIII/MnII reduction potential should be expected. In fact, Mn(TBAP-TME)Cl shows a metal-centered process at E ½ = −201 mV versus NHE (measured in DMF, owing to the insolubility of this tetramethyl ester in water) that is 24 mV more positive than that of [MnIIITBAP]3− under the same conditions (Fig. 6). This small shift in reduction potential, which is coupled with an increased electrostatic facilitation (by going from a trianionic species to a monocationic Mn complex), should be of little impact on the dismutation of O2 •−. This assumption is supported by the fact that an “improved” Mn porphyrin analogue (containing trimethyl anilinium groups instead of the carboxylic ones, E ½ = −0.100 V vs NHE), which has a reduction potential 94 mV more positive than that of [MnIIITBAP]3− and is pentacationic, shows only marginal SOD activity (log k cat = 5.11) [10].

Although of little effect on the reduction potential of [MnIIITBAP]3− itself, in situ derivatization via protonation, esterification, or amidation of the carboxylic moieties changes dramatically the lipophilicity of the compound, which may result in a different tuning of cellular localization. One plausible option [24], which is introduced here and may explain the alleged [MnIIITBAP]3− effect on the overall levels of oxidative stress (if not from the impurities in the samples), arises from shifting from the classic aqueous systems point of view to a nonaqueous one, considering the localization of derivatized [MnIIITBAP]3− in the lipid membranes, for example. In an aprotic solvent such as DMF (which may be a model for cellular lipid membranes) O2 •− is oxidized to O2 at −600 mV versus NHE [13], i.e., 440 mV more easily than in water. Consequently, in a lipid environment, despite the lack of electrostatic facilitation, O2 •− may be able to effectively reduce [MnIIITBAP]3− (likely protonated, esterified, or amidated) to yield O2 and the MnII counterpart complex in a stoichiometric fashion. Reoxidation of [MnIITBAP]4− by O2 •− to yield [MnIIITBAP]3− is not feasible in aprotic media [13] given the instability of “naked” O2 2−. A catalytic dismutation of O2 •− in the lipid environment may be possible only with the participation of a suitable proton donor [49–51] within the membrane (or in the aqueous medium, if the reaction takes place at/near the water–lipid interface). The scavenging of O2 •− by [MnIIITBAP]3− may also be made catalytic if an oxidant, other than O2, is coupled to the MnIII/MnII reduction to oxidize [MnIITBAP]4− back to [MnIIITBAP]3−, via a superoxide oxidase mechanism instead of an SOD one.

The biological activity of [MnIIITBAP]3−, without directly modulating reactive oxygen species (ROS) and reactive nitrogen species (RNS), has been considered recently by Tauskela et al. [52, 53]. They studied a variety of metalloporphyrins derived from [H2TBAP]4− and proposed an alternative Ca2+-dependent mechanism for the action for this class of compounds; it is worth noting that the presence of redox-active impurities in their [ZnIITBAP]4− preparation (for example) is unlikely, but even in this case an “antioxidant”-like activity was observed. Also Konorev et al. [54] suggested that the possible effects of negatively charged Mn porphyrins, including [MnIIITBAP]3−, may be related to the induction of heme oxygenase-1, which would eventually modulate ROS and RNS indirectly.

The activity of [MnIIITBAP]3− as “SOD mimics” has also been questioned by Szabó et al. [2], Groves [55], and Munroe et al. [56], which further justifies the clarification of the properties of this compound.

Concluding remarks

[MnIIITBAP]3− is the most studied porphyrin-based “antioxidant.” Some sources unsupportedly state that MnTBAP is as potent an SOD scavenger as the SOD enzymes themselves. Moreover, it is even claimed to be an efficient ROS/RNS scavenger of broad spectrum. In light of the findings described herein, misguided conclusions on the physiological and pathological roles of superoxide may have been derived from studies with [MnIIITBAP]3−. A thorough (re)investigation of the literature [MnIIITBAP]3− data with a pure compound is, therefore, essential to unambiguously concluding on the role of [MnIIITBAP]3− (if any) in the modulation of oxidative stress or enzymatic/signaling pathway, and to rule out artifacts that may arise from the “SOD-active” trace impurities of commercial or insufficiently purified samples.

Regardless of the commercial source or the method used to prepare [MnIIITBAP]3− in-house, every batch intended to be used in biological or mechanistic studies needs to be thoroughly characterized and checked for noninnocent impurities. Among the analytical methods for assessing the level of contamination, ESI–MS and cytochrome c assay were the most sensitive ones, whereas cyclic voltammetry findings were unaltered by the presence of trace impurities; UV–vis spectroscopy is suitable for detecting contaminations with free-base porphyrin.

As our knowledge about metalloporphyrins in biological settings grew, it became apparent that those Mn porphyrins of favorable and biologically compatible redox potentials may act not only as SOD mimics but also as, for example, superoxide reductases (possibly via peroxo complexes [57–59]), or finely tune signaling pathways either by eliminating signaling ROS and RNS that can activate transcription factors (NF-κB, AP-1, HIF-1) [60–62], or by directly interacting with transcription factors (NF-κB) [62]; possible effects on protein expression (whether or not via ROS/RNS modulation) have also been considered recently [11]. These studies, however, have all been done with porphyrins of highly positive MnIII/MnII reduction potential and very high superoxide dismuting ability (as indicated by their high log k cat values approaching those of the SOD enzymes themselves). By establishing that pure [MnIIITBAP]3− does not possesses the property of dismuting superoxide in aqueous media, we provided here the foundations to support the hypothesis that if any biological effect is observed using pure [MnIIITBAP]3−, then the in vivo effect of this porphyrin should take place through some mechanism other than superoxide dismuting itself. This would bring a new dimension to the development and mechanistic studies of synthetic antioxidants as a whole. Such an important conclusion cannot be reached, however, if some “elusive” SOD activity is a priori attributed to a compound (whether [MnIIITBAP]3− or not) or if impure samples containing noninnocent impurities are used in vivo.

Additionally, this work nicely illustrates the need for thorough consideration of SAR studies not only for drug design and optimization, but also as a “watchdog” mechanism for checking/spotting eventual incongruence of drug activity in (bio)chemical and biological settings.