Superoxide Dismutase Mimics: Chemistry,Pharmacology,and Therapeutic Potential

A. General

R edox imbalance between reactive species and endogenous antioxidants, which results in oxidative damage to biologic molecules and impairment in signaling pathways [(i.e., in oxidative stress (145)], has been widely implicated in many ailments, including central nervous system pathologies (46, 51, 61, 122, 219, 330) [e.g., amyotrophic lateral sclerosis, (46), Parkinson's disease (219), bipolar disorder (330), Alzheimer's disease (61)], cardiovascular conditions (61, 112), pulmonary conditions (65, 153), diabetes (111, 154), eye diseases (19, 235), aging (290, 323, 236), cancer (52, 70, 317), radiation injury (220), pain/chronic morphine tolerance (89), Fanconi anemia (229). Reactive species, such as nitric oxide (^·NO), superoxide (O₂ ^·−), hydrogen peroxide (H₂O₂), peroxynitrite (ONOO^–), and others have been widely recognized as signaling species that, by affecting redox-based cellular transcriptional activity, control inflammatory and immune responses and enhance secondary oxidative stress (27, 47, 96, 151, 188, 273, 276, 298, 335, 336). Mitochondria, the major producers of reactive species, are consistently found to play a critical role in oxidative stress (55, 155, 228, 312).

B. Antioxidants

The increased perception and understanding of the involvement of oxidative stress in many pathologic conditions has been accompanied by an increased search for synthetic antioxidants, as well as by further exploration of the antioxidant potential of several natural products. Recently, it also became evident that a number of drugs, such as antiinflammatory drugs, statins, and antibiotics, which supposedly aimed at different targets in unlike disorders, have the regulation of oxidative stress as a prominent mode of action, thus potentiating the widespread awareness of the role that oxidative stress plays in several diseases and injuries (3, 27, 47, 55, 64, 155, 175, 188, 209, 228, 312, 336). Superoxide dismutase is an endogenous and first-line-of-defense enzyme that eliminates superoxide by catalyzing its dismutation into O₂ and H₂O₂ (119, 120, 212, 240). Historically, most early synthetic antioxidant compounds were originally developed as SOD mimics, especially because the role of ONOO⁻ and its decomposition products in biology were, at the time, neither accepted nor well defined (106). A greater understanding of the biologic activity of SOD mimics and redox-active compounds paralleled the increased insight into the nature and the role of ROS/RNS in oxidative-stress conditions. The redox properties that allow SOD mimics to eliminate O₂ ^·− make them also potentially efficient peroxynitrite scavengers, as well as scavengers of CO₃ ^·–, ^·NO₂ radicals, and likely of peroxyl radicals and alkoxyl radicals (110, 151, 273). Therefore, most SOD mimics are not specific O₂ ^·− scavengers. Multiple strategies and controls must be used to assure which is the predominant species involved. Whatever mechanism is in action, antioxidants would also decrease the levels of oxidatively modified biologic molecules. Reactive species, such as O₂ ^·−, H₂O₂, and ^·NO, and oxidatively modified biologic molecules (e.g., nitrated lipids and nitrosated proteins) all appear to be involved in signaling events; their removal affects both primary oxidative damage and redox-based cellular transcriptional activity (27, 47, 55, 188, 273, 298, 312, 336). Therefore, antioxidants influence both inflammatory and immune pathways and also modulate secondary oxidative-stress processes.

Removal of reactive species is redox-based. Thus, it is only natural that the search for potent SOD mimics has been concentrated primarily on metal complexes that possess a redox-active metal site and rich coordination chemistry. Redox-based pathways play major role in supporting life. Nature has developed natural metalloporphyrins (e.g., heme) as major prosthetic groups embedded in a variety of biomolecules, such as hemoglobin, myoglobin, nitric oxide synthase, cytochrome oxidase, prolyl hydroxylase, cyt P450 systems (including aromatase), and cyclooxygenase. Molecules such as heme have been found to play a critical role in nearly all living organisms (145). No wonder thus that the synthetic Fe and Mn porphyrins appeared as a natural choice for developing SOD mimics: (a) they are “body-friendly” molecules; (b) they are chemically accessible, (c) they are not antigenic, (d) there are nearly limitless possibilities of modifying the porphyrin core structure; (e) porphyrin complexes are extremely stable, assuring the integrity of the metal site under biologic conditions; and finally, (f ) they are of low molecular weight and can penetrate the cellular and subcellular membranes, whereas superoxide dismutase enzymes cannot.

The pioneering work on metalloporphyrins as SOD mimics (most notably, MnTM-4-PyP⁵⁺ and FeTM-4-PyP⁵⁺) was done by Pasternack, Halliwell, Weinberg, Faraggi, and others in the late 1970s and early 1980s (104, 157, 246 –248, 252, 293, 332, 333). These early studies encompassed the rich chemistry of these metalloporphyrins toward radicals other than O₂ ^·− alone. The next milestone came from our group; we established a structure–activity relation between metal-site redox ability and catalytic rate constant for O₂ ^·− dismutation (30) that guided most of the work thereafter.

Reports on both toxic and protective effects of Fe porphyrins have been published (30, 231, 238, 313). Although the corresponding Fe and Mn porphyrins have very similar rate constants for O₂ ^·− dismutation, all Fe porphyrins studied by us thus far were toxic to Escherichia coli; no aerobic growth was detected in SOD-negative mutants with Fe porphyrins at levels at which analogous Mn porphyrins were fully protective (30). A loss of metal from the metal complexes during redox cycling could occur, whereby “free” Fe would give rise, through Fenton chemistry, to highly oxidizing ^·OH species; Fenton chemistry presumably occurs even if reduced iron is still bound to the porphyrin ligand (338). Thus, we limited our studies to Mn porphyrins as SOD mimics (Fig. 1). Although Cu porphyrins possess SOD-like activity in a simple cyt c assay (33), the ability of “free” copper(II) to produce ^·OH radical through Fenton chemistry (like Fe) disfavored exploiting Cu porphyrins for biomedical applications. Whereas Fe porphyrins were the first compounds considered as SOD mimics (246, 247), Mn porphyrins remain the most stable and most active prospective SOD mimics. The activity of some Mn porphyrins approaches that of the SOD enzymes themselves (85). Further, closely related porphyrin compounds, such as phthalocyanines (193), porphyrazines (193), biliverdins (302), corroles (144, 275), and texaphyrins (282), have been explored as SOD mimics. Although it is not an SOD mimic, a texaphyrin MGd (282) is also addressed in this review, as it appeared efficacious as an anticancer agent and produced effects similar to those of Mn porphyrins in ameliorating amyotrophic lateral sclerosis (66).

OPEN IN VIEWER

FIG. 1.

SOD mimics. Mn(III) porphyrins, Mn(II) cyclic polyamines, Mn(III) salen derivatives, nitroxides, and fullerenes were shown to possess SOD-like activity. “Free” Mn (i.e., low-molecular-weight Mn(II) species) such as aqua, oxo, hydroxo, and carboxylato species are able to dismute O₂ ^·− also. 5,10,15,20: meso positions of methine bridges between pyrrolic rings.

Other types of Mn complexes have also been considered as SOD mimics (Fig. 1). Cyclic polyamine (aza crown ethers)-based SOD mimics were characterized in details in vitro and in vivo (295). Mn salen derivatives were investigated as well (88). Along with metal-based SOD mimics, some nonmetallic and, thus far less-efficient compounds, such as nitroxides (140) and fullerenes (181, 349), also have been explored. Further, covalently bound porphyrins and nitroxides were studied (160). The increased understanding of the critical role played by mitochondria in numerous pathologic conditions (55, 155, 227, 228, 308) gave rise to the design of mitochondrially targeted systems with antioxidant properties. Among the most successful ones are monocationic MitoQ compounds, developed by Michael Murphy et al. (227, 228). These compounds possess a positively charged moiety (triphenylphosphonium cation) that drives them into mitochondria and a lipophilic alkyl chain that facilitates their transfer across the lipid bilayer. At the end of the alkyl chain, different redox-active compounds have been attached, including nitroxides (18, 90, 227, 228, 305, 308). Mitochondrially targeted oligopeptides have been attached to Mn porphyrins as well (18). Driven by the mitochondrial membrane potential, potent pentacationic Mn porphyrins with no particular targeting moiety also were found to be directly taken up by mitochondria (305). An MnSOD knockout yeast study suggested that Mn salen can also enter mitochondria (at least those of yeast) (131). In addition to synthetic antioxidants that act catalytically, natural antioxidants have been used in numerous studies and clinical trials with partially satisfactory results (270, 291, 292, 311). The lack of full success is often ascribed to a poor design, quality of the study, external and internal validity, homogeneity of the sample, baseline status, dosing, timing, interaction among nutrients, gene polymorphism, and statistical power. Debate still exists, and a detailed study is ongoing to understand which component/s of tea, olive oil, wine, and so on, are beneficial, whether it be polyphenols or something else (146). Lately, the combined therapy of synthetic and natural antioxidants has been frequently employed.

The in vivo effects of different types of compounds are influenced primarily by (a) antioxidant ability; (b) bioavailability (i.e., the ability to accumulate within a cell and its compartments); and (c) toxicity. Bioavailability is dependent on the size, charge, shape, (conformational flexibility and overall geometry), and lipophilicity and greatly affects the in vivo efficacy; data already indicate that whereas some drugs may be good in one model, they may fail or be less efficient in another one, as a consequence of differences in the extent of accumulation in subcellular compartments targeted (88, 131, 255). Detailed pharmacokinetic and toxicology data are still scarce. None of the SOD mimics thus far has been approved for clinical use. To better judge the therapeutic potential and the advantage of one over the other type of compound, comparative studies of different types of drugs in the same cellular or animal model or both are needed. Thus far, only limited data have been provided (77, 131, 225, 255). Further, few studies have shown that protective effects in vivo parallel the in vitro magnitude of the catalytic rate constant for O₂ ^·− dismutation or peroxynitrite reduction or both (241, 255). Such data justify further efforts to understand the role of structure–activity relationship in designing SOD mimics and peroxynitrite scavengers.

We here provide an overview of the chemical properties and some in vivo effects observed with different classes of compounds, with a special emphasis on porphyrin-based compounds. Of note, in many instances, clear activities toward particular reactive species (in the form of rate constants) are missing, and assumption has been often made about such action, or the lack thereof. Finally, financial interests involved in the pharmaceutical development of these compounds have often influenced the objectivity of some of the reviews published. Often, poor management of patenting and licensing rights has resulted in the “Valley of Death” status of SOD mimics, preventing them from reaching clinical trials and clinical use in a timely manner, if at all (118).

Most potent SOD mimics are metal complexes, which may eventually lose metal while redox cycling. Mn, in its own right, is able to catalyze O₂ ^·− dismutation at a fair rate, and thus is, in essence, an SOD mimic, too (14, 25). In some instances, Mn released from a complex, rather than the metal complex itself, could be responsible for the effects observed (262). Therefore, we addressed herein the SOD-like ability of “free” Mn²⁺ both in aqueous solutions and in vivo.

The purity of any SOD mimic should be established very carefully with extensive and multiple analyses. Even minute impurities, negligible for most chemistry-related research, might decrease/increase/modify the SOD-like activity of the material, affect its therapeutic and/or mechanistic evaluations, jeopardize the conclusions of many studies, and harm the health of the antioxidant field as a whole (31, 263 –265).

As ONOO⁻ is an adduct of O₂ ^·− and ^·NO, and some SOD mimics are potent ONOO⁻ reduction catalysts, we found it important to address herein also the ONOO⁻-scavenging abilities of the SOD mimics.

A. SOD-like activity of manganese

All Mn-based SOD mimics, but particularly those of lower metal/ligand stability, may lose Mn (to some extent) when they are in low oxidation state (such as + 2), during the Mn³⁺ and Mn²⁺ cycling in the O₂ ^·− dismutation catalysis. Furthermore, some of the biologic chelators may dechelate Mn from SOD mimics of low metal/ligand stability, such as Mn cyclic polyamines, Mn salen derivatives, and Mn β-octabrominated porphyrins. Thus, it is important to verify whether “free” Mn²⁺ (i.e., unbound from the corresponding ligand) can exert SOD-like activity in vitro/vivo. Control experiments with Mn²⁺ and nonmetallated ligands are, therefore, critical for mechanistic conclusions.

The SOD-like activity of Mn²⁺ is dependent on the type of the ligand, whether it is a hexaaqua-, carboxylato-, monohydroxo-, or oxo/hydroxo/acetato species. A few data on the SOD-like activity of MnCl₂ in medium containing phosphate buffer are available. Our group has reported k _cat, determined by pulse radiolysis in 0.05 M potassium phosphate buffer, pH 7.8, 25°C, to be k _cat = 1.3 × 10⁶ M⁻¹s⁻¹ (302). Joan Valentine's group (25) reported a higher value, k _cat = 8.9 × 10⁶ M⁻¹s⁻¹, in essentially identical 0.05 M phosphate medium, pH 7 (25), with superoxide produced through ⁶⁰Co gamma irradiation; the formation of monoformazan in the reaction of XTT or MTS with O₂ ^·− was followed. Archibald and Fridovich (14) reported the SOD-like ability of different Mn complexes prepared in situ. The ligands investigated were phosphate, pyrophosphate, formate, lactate, citrate, succinate, acetate, cacodylate, and propionate. Mn lactate was the most potent SOD mimic; its activity expressed per milligram Mn was only 65-fold lower than that of the SOD enzyme. The other complexes were several-fold less potent; Mn phosphate being ∼253-fold less potent than SOD enzyme (k _cat ∼4 × 10⁶ M⁻¹s⁻¹) (14).

The Naughton group published SOD-like activities of different Mn complexes with ligands such as EDTA, EGTA, EHPG, EBAME, and salen by using nitrobluetetrazolium (NBT) assay (114, 115). The SOD-like activities of those complexes were expressed as IC₅₀ (μM) values (114, 115), which, if converted to the corresponding rate constants by comparison with Cu,ZnSOD and the SOD mimic MnTE-2-PyP⁵⁺ data, resulted in too high k _cat values (up to 10⁷ M⁻¹s⁻¹). The k _cat values for aquated Mn(II) (115) and Mn salen were reported by Naughton's group (114, 115) as 3.6 × 10⁶ M⁻¹s⁻¹ and 8.7 × 10⁶ M⁻¹s⁻¹, respectively, which are three- and 10-fold higher than the values determined by us (302). The NBT assay has substantial disadvantages over the cytochrome c assay, as NBT itself can mediate the formation of O₂ ^·− (195). From the experimental section in references 114 and 115, the nature of the aquated ion is not obvious (whether it is a phosphate salt). The presentation of the data in different styles in different publications prevents their direct comparisons with other compounds. Furthermore, these same reports (114, 115) show the k _cat for MnEDTA to be 6 × 10⁵ M⁻¹s⁻¹, whereas we (302), Archibald and Fridovich (14), and Baudry et al. (40) were not able to detect any measurable SOD-like activity of MnEDTA.

We recently reported that nonporphyrin Mn species, tentatively formulated as Mn hydroxo/oxo/acetato species, appear as impurities in commercial MnTBAP³⁻ preparations and, although unstable, are very effective in dismuting O₂ ^·−. Consequently, the impure MnTBAP³⁻ preparations exhibit SOD-like activity (31, 264). MnTBAP³⁻ preparations from different commercial sources and different batches from the same source contained different levels of those trace Mn species and, thus, each sample showed different SOD-like activities. As these species occur in trace amounts and are not stable (SOD activity of the commercial samples decreased with the aging of the solution), they were not isolated, and their absolute SOD-like activity was, therefore, not quantified (264). It is worth noting that because such species appear in trace amounts, they must, consequently, possess high SOD activity to account for the effect observed.

In summary, the SOD-like activity of Mn²⁺ is highly dependent on the type of potential counteranion/ligand present in the medium and may be equal to, or higher than 10⁶ M⁻¹s⁻¹.

B. The effects of manganese in vitro and in vivo

Archibald and Fridovich (15) showed that Lactobacillus plantarum compensates for the lack of SOD enzyme by accumulation of manganese to millimolar levels. SOD-deficient E. coli, lacking cytosolic SOD enzymes, does not grow aerobically, but it grows equally well as wild type if an SOD mimic is supplied in the medium to substitute for the lacking SOD enzymes (9, 15, 30, 85). Aerobic growth of SOD-deficient E. coli is an O₂ ^·−-specific, in vivo system that usefully predicts which compounds may be prospective therapeutics for clinical development. Mn²⁺ protects SOD-deficient E. coli when growing aerobically, although not as efficiently as Mn porphyrins (9, 225). The effects are related to the decrease in oxidative stress, protection of aconitase activity, and decreased mutations, which result in increased growth; all effects become obvious at >0.5 mM MnCl₂ (9). We also showed that 1 μM Mn²⁺ offers some radioprotection to ataxia telangiectasia cells, but is significantly less efficient than 1 μM of a more potent SOD mimic, Mn porphyrin MnTnHex-2-PyP⁵⁺ (255). Although Mn²⁺ seems of comparable efficacy to Mn salen and Mn cyclic polyamine (255), the latter complexes were used at higher (10 or 20 μM) concentrations, which precluded a full assessment of the extent of radioprotection by MnCl₂ in comparison to all other compounds in that particular model (255). In MnSOD-knockout Cryptococcus neoformans, whose growth is susceptible to oxidative stress at elevated temperatures, Mn salen and ascorbate, but not MnCl₂ and none of several different anionic and cationic Mn porphyrins, were protective (131). Because of the low metal/ligand stability of Mn salen, it is not clear whether Mn salen remains as such, or whether the compound acts as an Mn-carrier into the mitochondria, where released Mn could act in its own right. Our data with E. coli (262) have unambiguously shown that such Mn-transporting mechanism may be relevant for certain SOD mimics in vivo: the Mn octabrominated porphyrin, MnBr₈TSPP³⁻, which has low metal/ligand stability, can transport Mn²⁺ into the E. coli cell (262); metal-free octabrominated porphyrin ligand was spectroscopically detected within the cells (262). Exogenous Mn in millimolar concentrations rescued O₂ ^·−-sensitive phenotypes of S. cerevisiae lacking Cu,ZnSOD (279). Similar findings, wherein non-SOD manganese is a backup for Cu,ZnSOD in S. cerevisiae, was later reported by Reddi et al. and Culotta et al. (72, 267). Enhancement of stress resistance and the effect of Mn²⁺ supplementation on the life span of Caenorhabditis elegans was reported (193). The role of Mn transporters also was addressed, and carboxylates rather than phosphates were suggested as possible ligand carriers for Mn²⁺ (267). Data by Reddi et al. (267) are in agreement with our study, in which Mn oxo/hydroxo/acetato complexes, present as a non-innocent impurity in ill-purified MnTBAP³⁻ preparations, are responsible for the SOD-like activity (264). The issues with respect to Mn²⁺ remain mostly unresolved, particularly the true nature of the Mn²⁺ complexes responsible for O₂ ^·− scavenging ability of Mn²⁺ in vivo. A very recent and intriguing E. coli report by the Imlay group (13) suggested that Mn substitutes for Fe in Fe enzymes vulnerable to O₂ ^·− attack (which would have otherwise resulted in deleterious effects of Fenton chemistry) rather than act by O₂ ^·−/H₂O₂ scavenging.

Because of the dismuting ability of Mn²⁺, and particularly when mechanistic purposes are the goal of the study, it is important to have Mn-based antioxidants very pure and devoid of “free”, residual Mn²⁺ in any form. Anionic porphyrins are the most difficult to purify with respect to residual manganese. For such purposes, we developed a very sensitive method for quantifying residual, nonporphyrin-bound Mn²⁺ species in Mn-based SOD mimic systems of high metal/ligand stability (263).

A. Metalloporphyrins

The metalloporphyrins, and preferably water-soluble Mn but not Fe complexes, have been chosen as potential SOD mimics for the reasons cited in the introduction. Two scientists greatly influenced the design and use of metalloporphyrins as SOD mimics, Irwin Fridovich, the “father” of the free radical biology and medicine, and Peter Hambright, the “father” of water-soluble porphyrins, with both of whom we have had the honor to work and to learn from. The seminal report of Irwin Fridovich group on Mn porphyrin-based SOD mimics in the 1994 J Biol Chem, included also the MnTM-4-PyP⁵⁺ and MnTBAP³⁻ (MnTCPP³⁻) complexes (105). Although MnTBAP³⁻ was not explicitly shown to be an SOD mimic in its own right in that publication, the fact that its structure and some incorrect data were reported there may have misled the biomedical audience; for example, the E_1/2 of MnTBAP³⁻ was reported as ∼ + 110 mV versus NHE, which is 304 mV more positive than the correct value published thereafter [–194 mV vs. NHE (30)] and recently was confirmed in a pure MnTBAP³⁻ sample (264). It is worth noting that were the initial value true, MnTBAP³⁻ might have functioned as an SOD mimic. Another incorrect assignment of the MnTBAP³⁻ SOD-like activity followed in the J Pharmacol Exp Ther 1995 by Day et al. (81). Soon afterward, we established the first structure–activity relationship that correlated the ability of Mn and Fe porphyrins to dismute O₂ ^·− (log k _cat) with their metal-centered reduction potentials, E_1/2 (for Mn^IIIP/Mn^IIP redox couple) (30). The most potent compound at that time, MnTE-2-PyP⁵⁺, was identified and forwarded to in vitro and in vivo studies. In 1998, Rafael Radi (105 –109) suggested, and he and his group successfully tested, the possibility that potent SOD mimics could also be powerful ONOO⁻ scavengers. A few years later, another mechanistic aspect of the in vivo efficacy of this and other Mn porphyrins emerged as a consequence of the ongoing efforts to understand the role of ROS/RNS in signaling events in oxidative stress–related conditions, disorders, and diseases as diverse as inflammatory and immune responses, cancer, radiation injury, diabetes, aging, central nervous system disorders, and so on. It became obvious that the effects observed when using Mn porphyrins were not only the consequence of mere scavenging of ROS/RNS, but that MnPs were also able to modulate ROS/RNS-based signaling pathways. Several articles that followed provided evidence that a potent SOD mimic/ONOO⁻ scavenger, such as MnTE-2-PyP⁵⁺, can strongly inhibit excessive activation of redox-sensitive cellular transcriptional activity (39, 221, 222, 259, 288, 322, 350).

Thus, over the years, our views on Mn porphyrins evolved from SOD mimics, to O₂ ^·−/ONOO⁻ scavengers, and finally to redox modulators of cellular transcriptional activity. The same is also true for other groups of synthetic SOD mimics discussed later. At this point, we do not exclude other possible roles of Mn porphyrins. The Tauskela group (314, 315, 342) suggested the action of MnP on Ca²⁺ metabolism (which may again be ROS modulated), whereas the Kalyanaraman group (176) reported on the induction of heme oxygenase by MnP. Because of the biologically accessible metal-centered reduction potential and the ability to reach four oxidation states in vivo (+2, + 3, + 4, and + 5), cationic Mn porphyrins can redox cycle with a number of biologic molecules, such as cellular reductants, flavoenzymes and cytochrome P450 reductase, and can mimic the cyt P450 family of enzymes (79, 108, 304); as a consequence of their rich chemistry and redox-cycling capabilities, these compounds may be easily involved in beneficial and in adverse pathways. The possibility that electrostatic interactions of MnPs with biologic molecules contribute to their action/s in vivo is not excluded and will be further explored (39).

B. Design of porphyrin-based SOD mimics

1. Thermodynamics

The design of porphyrinic SOD mimics has been based on the simulation of both the thermodynamic and electrostatic properties of the enzyme itself. Self-dismutation of O₂ ^·− at pH 7.4 occurs with a rate constant, k ∼ 5 × 10⁵ M⁻¹s⁻¹, and is increased more than three orders of magnitude in the presence of SOD (145) (Fig. 2). All SOD enzymes, regardless of the type of metal (Mn, Fe, Cu, Zn, Ni), have metal-centered reduction potential around + 300 mV versus NHE, which is midway between the potential for the reduction (+850 mV vs. NHE) and oxidation of O₂ ^·− (−160 mV vs. NHE). Thus, both processes are thermodynamically equally favored at ∼ + 300 mV versus NHE. In turn, both reduction and oxidation reactions in the dismutation process occur with the same rate constant of 2 × 10⁹ M⁻¹s⁻¹ (100, 174, 325). In addition to the suitable thermodynamics of the active site, the appropriate placement of positively charged amino acid residues along a tunnel leading to the metal site in the enzymes provides electrostatic guidance for the approach of O₂ ^·− to the active site (87, 128).

OPEN IN VIEWER

FIG. 2.

The O₂ ^·− dismutation process.

The O₂ ^·− dismutation mechanism catalyzed by Mn porphyrins involves two steps in which the Mn center cycles between Mn(III) and Mn(II). As most Mn porphyrins contain Mn in the + 3 oxidation state, the first step, which coincides with the rate-limiting step, corresponds to the reduction of Mn(III) by O₂ ^·− to yield Mn(II) and O₂. The second step corresponds to the oxidation of Mn(II) by O₂ ^·− to yield H₂O₂ and reestablish the Mn(III) porphyrins. This catalytic cycle is evidently modulated by the redox potential of the metal site (Fig. 2).

Mn(III) meso-tetrakisphenylporphyrin (MnTPP⁺) has E _1/2 = −280 mV versus NHE and para Mn(III) meso-tetrakis(4-pyridylporphyrin) (MnT-4-PyP⁺) has E _1/2 = −200 mV versus NHE (299) (Table 1). Both reduction potentials are outside the window for O₂ ^·− reduction and oxidation (Fig. 3). Therefore, these Mn(III) porphyrins cannot be reduced by O₂ ^·− in the first step of the catalytic cycle and were not found to be SOD mimics (30, 299). Attaching electron-withdrawing groups to the porphyrin molecule as close to the metal site as possible has been a viable strategy to increase the metal-site electron deficiency, which makes Mn more prone to accept electrons. In turn, as the reduction potential increases, the first step of the catalytic cycle is favored. Indeed, the introduction of positive charges on pyridyl nitrogens of MnT-4-PyP⁺ to yield MnTM-4-PyP⁵⁺ increased E _1/2 dramatically by 260 mV, from −200 mV to + 60 mV versus NHE, respectively. With E _1/2 of MnTM-4-PyP⁵⁺ placed between the potential for the reduction and oxidation of O₂ ^·−, the catalytic cycle of O₂ ^·− dismutation could be established on thermodynamic grounds, giving rise to a fair value for the catalytic rate constant, k _cat = 3.8 × 10⁶ M⁻¹s⁻¹ (29); the rate-limiting step still remained the reduction of Mn^IIIP to Mn^IIP. A problem associated with compounds such as MnTM-4-PyP⁵⁺, which limits their use as SOD mimics in cell/animal experiments, is their ability to adopt a near-planar structure, as pointed out by Pasternack (249, 250), and consequently to associate with and intercalate into nucleic acids (249, 250). Still, the bulkiness imposed by the water molecules axially bound to the Mn center limits the intercalation. Mn^IIITM-4-PyP⁵⁺, with the manganese in the oxidized Mn(III) form, is more electron deficient and binds axial waters more strongly than the electron-rich reduced Mn(II) site in Mn^IITM-4-PyP⁴⁺. Thus, because of the steric hindrance, the bulkier Mn^IIITM-4-PyP⁵⁺ associates with nucleic acids much less than the Mn^IITM-4-PyP⁴⁺ (249, 250). Yet, while redox cycling with O₂ ^·−, the reduced Mn^IITM-4-PyP⁴⁺ is formed, which associates with nucleic acids. We have reported that such associations with nucleic acids fully prevented MnP from dismuting O₂ ^·− (29). When nucleic acids of MnTM-4-PyP⁵⁺-treated E. coli were removed from the cell extract [by precipitation with protamine sulfate (29, 249, 250)], the SOD-like activity of the cell extract was fully restored. Furthermore, associations with nucleic acids not only affected the in vivo SOD activity of the compound but also introduced toxicity.

OPEN IN VIEWER

FIG. 3.

Redox diagram for O₂ ^·− reduction and oxidation and the placement of Mn porphyrins on it.

OPEN IN VIEWER

Table 1.

Selected Physicochemical Properties of Some SOD Mimics

				Lipophilicity
Compound	MnIII/II potential, E½/mV vs. NHE a	SOD activity log k cat (O 2 •– ) b	PN red. activity log k red (ONOO – ) c	Rel. Rf d	log P ow	Ref.
Cationic porphyrins
MnTM-2-PyP5+	+220	7.79	7.28	0.5	−7.86	29, 179
MnTE-2-PyP5+	+228	7.76 (cyt c) 7.73 (p.r.)	7.53	1	−6.89	30, 179, 302
MnTnPr-2-PyP5+	+238	7.38	7.15	1.8	−5.93	37, 179
MnTnBu-2-PyP5+	+254	7.25	7.11	3.2	−5.11	37, 179
MnTnHex-2-PyP5+	+314	7.48	7.11	6.3	−2.76	37, 179
MnTnHep-2-PyP5+	+342	7.65		7.7	−2.10	178, 179
MnTnOct-2-PyP5+	+367	7.71	7.15	8.2	−1.24	179
MnTMOE-2-PyP5+	+251	8.04 (p.r.)		1.2		38
MnTTEG-2-PyP5+	+250	8.11		1.0		36
MnTrM-2-PyP4+	+118	6.63		1.1		30, 265
MnBM-2-PyP3+	+53	6.52		2.3		30, 265
MnTrE-2-PyP4+				2.2		30, 265
MnBE-2-PyP3+				4.4		30, 265
MnTDM-2-ImP5+	+320	8.11		0.5		266
MnTDE-2-ImP5+	+346	7.83 (p.r.)		2.3	−6.48	38
MnTDnPr-2-ImP5+	+320	8.11				167
MnTM,MOE-2-ImP5+	+356	7.98 (p.r.)		1.2		38
MnTDMOE-2-ImP5+	+365	7.59 (p.r.)		1.6		38
MnTDTEG-2-ImP5+	+412	8.55		2.0		36
MnTM-3-PyP5+	+52	6.61	6.62	0.8	−6.96	29, 179
MnTE-3-PyP5+	+54	6.65		1.7	−5.98	178, 179
MnTnPr-3-PyP5+	+62	6.69		3.7	−5.00	178, 179
MnTnBu-3-PyP5+	+64	6.69		6.7	−4.03	178, 179
MnTnHex-3-PyP5+	+64	6.64		9.2	−2.06	178, 179
MnTM-4-PyP5+	+60	6.58	6.63	0.5		29, 266
MnTE-4-PyP5+	+70	6.86				299
MnTDM-4-PzP5+	−4	5.83		0.6		266
MnT(TriMA)P5+	−100	5.11		1.8		266
MnT(TFTriMA)P5+	+58	6.02				30
MnCl1TE-2-PyP5+	+293	7.75		1.1		166
MnCl2TE-2-PyP5+	+343	8.11		1.2		166
MnCl3TE-2-PyP5+	+408	8.41		1.3		166
MnCl4TE-2-PyP5+	+448	8.60		1.3		165
MnCl5TE-2-PyP5+	+560	8.41				165
MnBr8TM-3-PyP4+	+468	≥8.85				85
MnBr8TM-4-PyP4+	+480	≥8.67				33, 85
CuTM-4-PyP4+		<3.7				33
CuBr8TM-4-PyP4+		6.46				33
Neutral porphyrins
MnT-2-PyP+	−280	4.29		10.2		301
MnBr8T-2-PyP+	+219	5.63				261, 301
MnT-4-PyP+	−200	4.53				299
MnTPP+	−270	4.83				299
MnTPFPP+	−120	5.00				299
MnTBzP+	+88	5.20				182
Anionic porphyrins
MnTBAP3– (pure form)	−194	3.16	5.02			264
MnTSPP3–	−160	3.93	5.53			30
MnT(2,6-Cl2-3-SO3-P)P3–	+88	6.00				30
MnT(2,6-F2-3-SO3-P)P3–	+7	5.51				30
MnBr8TSPP3–	+209	5.56				262
MnBr8TCPP3–	+213	5.07				262
Salens
Mn(salen)+, EUK-8	−130	5.78 e		12		255, 302
EUK-134		5.78				255
EUK-189		5.78				255
Cyclic polyamines
M40403	+525 (ACN)	7.08, 6.55		13		204, 255, 269
M40404(2R,21R-Me2-M40403)	+452 (ACN)	Inactive		12		204, 269, 255
2S,21S-Me2-M40403	+464 (ACN)	8.37, 9.20				204, 26921
Fullerenes
C60-water-soluble fullerene (C3)		6.30				8
Miscellaneous
Mn2+	+850 f	6.11 (cyt c)6.28 (p.r.)6.95 (p.r.)		0		299, 255, 302, 25
Mn-EDTA		inactive (cyt c)				302, 14, 40
[MnBV2–]2	−300 (+460 g )	7.40 (cyt c)6.95 (p.r.)				302
[MnBVDME]2	−230 (+450 g )	7.70 (cyt c)6.95 (p.r.)				302
[MnMBVDME]2	−260 (+440 g )	7.36				299
[MnBVDT2−]2	−260 (+470 g )	7.40				299
4-carboxy-Tempo		7.54 (pH 5.4)				140
MnTrM-2-Corrole3+	+910 g	5.94				99
OsO4		9.14 h				135
CeO2 (3-5 nm particles)		9.55				177
Honokiol		5.50				92
Mn texaphyrin		∼4.48				289

Mn(III)/Mn(II) reduction potential (E_½); SOD activity (O₂ ^·– dismuting catalytic rate constant, log k _cat); peroxynitrite (PN) reducing activity (ONOO^– reduction rate constant, log k _red); lipophilicity of Mn^IIIP (chromatographic retention time, R _f; octanol-water partition coefficient, log P _ow).

a

E_½ data measured either directly in 0.05 M phosphate buffer, pH 7.8, 0.1 M NaCl, or converted accordingly to this medium, unless noted otherwise.

b

SOD activity measured by the cyt c assay in 0.05 M phosphate buffer, pH 7.8, 25 ± 1°C, unless noted otherwise.

c

Measurements in 0.05 M phosphate buffer, pH 7.4, 37 ± 0.1°C.

d

Data relative to the R _f value of MnTE-2-PyP⁵⁺ in plastic-backed silica-gel thin-layer chromatography plates eluted with 1:1:8 KNO_3(sat):H₂O:MeCN.

e

No SOD-like activity was observed in the presence of EDTA (283).

f

Oxidation potential only, Mn^III/Mn^II redox couple is irreversible.

g

E_½ data associated with the Mn^IV/III reduction potential.

h

pH 5.1–8.7.

p.r., pulse radiolysis.

To overcome such problems and further to enhance SOD-like activity, we placed the electron-withdrawing groups closer to the Mn site, into the ortho positions, to yield MnTM-2-PyP⁵⁺ (AEOL10112). The E _1/2 value of MnTM-2-PyP⁵⁺ was increased by 160 mV relative to MnTM-4-PyP⁺, resulting in a potential of + 220 mV versus NHE, which was very close to the E _1/2 of the enzyme itself. Further, because of the steric hindrance between the methyl groups in the ortho positions of the pyridyl rings and the protons at the β-pyrrolic carbons, the pyridyl moiety remains relatively perpendicular to the porphyrin plane, and MnTM-2-PyP⁵⁺ (and related compounds) can no longer adopt a near-planar conformation. The overall bulkiness diminishes the interactions with nucleic acids and toxicity. Retrospectively, moving the positive charges from distant para into closer ortho positions afforded also a large enhancement in the electrostatic facilitation for the approach of O₂ ^·− to the Mn site (266, 301) (see later).

Dramatic effects also were achieved by introducing electron-withdrawing bromines or chlorines onto β-pyrrolic positions of MnTM-4-PyP⁵⁺ or MnTE-2-PyP⁵⁺. Such a maneuver shifted the reduction potential 420 mV and 220 mV more positively in octabrominated MnBr₈TM-4-PyP⁴⁺ (E _1/2 = + 480 mV vs. NHE), and tetrachlorinated MnCl₄TE-2-PyP⁵⁺ (E _1/2 = + 448 mV vs. NHE) versus nonhalogenated analogues (Table 1) (33, 166). The MnBr₈TM-4-PyP⁴⁺ has a very high k _cat of 2.2 × 10⁸ M⁻¹s⁻¹. Yet, a huge shift in E_1/2 (i.e., a dramatic increase in electron-deficiency) stabilized Mn in its + 2 oxidation state. (When cationic porphyrins bear 5⁺ charge they have Mn in + 3 oxidation state [Mn^IIIP⁵⁺], and with 4⁺ total charge, Mn is in + 2 [Mn^IIP⁴⁺] or + 4 oxidation states [O = Mn^IVP⁴⁺].) Consequently, MnBr₈TM-4-PyP⁴⁺ has a metal/ligand stability constant of only 10^8.08 M (33), and loses Mn readily upon dilution at pH 7.8. Although its characterization was important for the design of Mn porphyrins, as a proof of concept, the compound itself was not of practical importance. With the availability of a wide spectrum of porphyrins (many provided by MidCentury Chemicals, Chicago, IL), their physicochemical and SOD-like activities were explored, and the very first structure–activity relationship was established for a variety of water-soluble Fe and Mn porphyrins possessing different charges and geometry (Fig. 4A) (30). We found that the higher the E _1/2, the more electron deficient the porphyrin is (as witnessed by the protonation of the pyrrolic nitrogens), and thus the higher the k _cat. The most potent SOD mimics are the cationic Mn(III) N-alkylpyridylporphyrins, in particular, the ortho isomers. Further increase in E_1/2 stabilizes Mn²⁺ so much that the Mn porphyrins exist predominantly in the reduced Mn(II) form; the Mn^IIP oxidation step then becomes rate limiting, and the SOD-like activity decreases (Fig. 4B) (37). It must be emphasized that the validity of k _cat data obtained with cytochrome c assay was confirmed by both pulse-radiolysis study (302) and stopped-flow technique (184, 185). The same agreement between these two methods was further confirmed by the Zeev Gross group (99).

OPEN IN VIEWER

FIG. 4.

Struture-activity relationships. (A) The very first structure–activity relationship between log k_cat (O₂ ^·−) and E_1/2 (Mn^IIIP/Mn^IIP) included porphyrins of different charge, different stericity, and different electrostatics for O₂ ^·− dismutation. (B) As the E_1/2 increases, the Mn⁺² oxidation state gets stabilized, and eventually oxidation of porphyrin becomes the rate-limiting step, and k _cat starts to decrease again. Only water-soluble Mn(III) porphyrins are given in the left, linear section of the curve that obeys the Marcus equation, and data (circles) are from ref. 30: (1) Mn^IIITCPP³⁻, (2) Mn^IIIT(TMAP)⁵⁺, (3) Mn^IIIT(2,6-F₂-3-SO₃-P)P³⁻, (4) Mn^IIIT(TFTMAP)P⁵⁺, (5) Mn^IIIT(2,6-Cl₂-3-SO₃-P)P³⁻, (6) Mn^IIIBM-2-PyP³⁺, (7) Mn^IIITM-3-PyP⁵⁺, (8) Mn^IIITM-4-PyP⁵⁺, (9) Mn^IIITrM-2-PyP⁴⁺, (10) Mn^IIITM-2-PyP⁵⁺, and (11) Mn^IIITE-2-PyP⁵⁺. Data for EUK-8 are from ref. 302, and for MnCl₂, from refs. 299 and 302; data for Mn(II) cyclic polyamine M40403 are from ref. 21. Data for SOD are from ref. 218. Data for Mn^IIICl_1-4MnTE-2-PyP⁵⁺ (12–15) are from ref. 166, data for Mn^IIBr₈TM-4-PyP⁴⁺ (16) are from ref. 33, and for Mn^IICl₅TE-2-PyP⁴⁺ (#17), from ref. 165 (triangles). Data for Mn^IIITnBu-2-PyP⁵⁺ (a) are from ref. 37; Mn^IIITMOE-2-PyP⁵⁺ (b) from ref. 38; Mn^IIITD(M)E-2-ImP⁵⁺ (c, d) from refs. 38 and 167; Mn^IIITM,MOE-2-PyP⁵⁺ (e), and for Mn^IIITDMOE-2-ImP⁵⁺ (f ) from ref. 38 (squares). Data points 18 and 19 (diamonds) belong to MnTTEG-2-PyP⁵⁺ and MnTDTEG-2-ImP⁵⁺ and are from ref. 36.

Over the years, given the lipophilic nature of the major drugs, the biomedical community has been concerned that excessively charged, pentacationic Mn porphyrins would fail to accumulate within the cell at levels sufficient to provide therapeutic effects. To account for such reasonable objections, the MnTalkyl-2-PyP⁵⁺ series was synthesized (alkyl moieties ranging from methyl to octyl) and their k _cat and E _1/2 determined (37). The lipophilicity described firstly by thin-layer chromatography retention factor, R _f and later by partition between n-octanol and water (P _OW) (179), increased from methyl to n-octyl in a linear fashion (Table 1). All of the analogues, bearing alkyl chains of different length, have high SOD activity. Small fluctuations observed were attributed to the interplay of the hydration and steric effects (37).

Based on the same thermodynamic and electrostatic premises, the imidazolyl derivatives, MnTD-alkyl-2-ImP⁵⁺, were then synthesized, where alkyls were methyl, ethyl (AEOL10150), and propyl (38, 167). These porphyrins again bear five positive charges close to the metal site that allowed both thermodynamic and electrostatic facilitation; thus the compounds possessed SOD-like activity similar to that of their ortho pyridyl analogues. The synthesis of methoxyethyl derivatives, MnTMOE-2-PyP⁵⁺ (38) and the N-triethyleneglycolated pyridylporphrin, MnTTEG-2-PyP⁵⁺ (36) and the corresponding imidazolium derivative, MnTDTEG-2-ImP⁵⁺(36), followed; all three compounds showed as high or higher k _cat than the previous complexes of the pyridyl and imidazolyl series (Table 1). These structures are bulkier, particularly the imidazolium series, in which alkyl or triethyleneglycols are located on both sides of the porphyrin plane. The goal of introducing triethyleneglycol moieties was to increase the blood-circulation lifetime (36, 313). Yet bulkiness may limit to some extent their access to cells and in turn their efficacy, as observed with protection of SOD-deficient E. coli (36, 38, 241).

With our ongoing goals to improve SOD-like activity in vitro and the efficacy in vivo, we recently synthesized and characterized the β-brominated meta isomer, MnBr₈TM-3-PyP⁴⁺; with an E _1/2 of  + 468 mV versus NHE and log k _cat >8.85, this complex has, thus far, the highest dismuting ability among metalloporphyrins (85), which approximates that of the enzyme itself (21, 26, 100, 136, 218, 269, 325). As observed for the para analogue, MnBr₈TM-4-PyP⁴⁺, the meta isomer has Mn in its + 2 oxidation state and, thus, has insufficient metal/ligand stability for in vivo studies.

2. Electrostatics

To quantify the electrostatic effects, we initially compared the related porphyrins, the monocationic MnBr₈T-2-PyP⁺ and the pentacationic MnTE-2-PyP⁵⁺ (301); of note, the former porphyrin is neutral on the periphery. Whereas the E _1/2 values for these Mn porphyrins were nearly identical (219 mV for MnBr₈T-2-PyP⁺ and 228 mV vs. NHE for MnTE-2-PyP⁵⁺), their k _cat values differed by almost two orders of magnitude (log k _cat = 5.63 for MnBr₈T-2-PyP⁺; log k _cat = 7.76 for MnTE-2-PyP⁵⁺) (Fig. 5A). The remarkable contribution of the electrostatics seen in these Mn porphyrins parallels the effect observed in the SOD enzyme catalysis and was confirmed by kinetic salt-effect measurements (301), and further substantiated in other studies.

OPEN IN VIEWER

FIG. 5.

The effect of charges on k _cat (O ₂ ^.− ). Mono- vs. pentacationic porphyrins differ in k _cat for 2 log units, whereas cationic vs. anionic porphyrins differ in k _cat for more than two orders of magnitude (A), like imidazolium vs. pyrazoliumporphyrins (B).

A second study was designed to investigate the impact of spatial charge distribution on the SOD catalysis, which also included the imidazolium and pyrazolium porphyrins (Fig. 5B) (266). Both compounds have been viewed as having delocalized charges. Yet, as the imidazolium compound has charges closer to the Mn site than does the pyrazolium porphyrin, the former had a k _cat value more than two orders of magnitude higher than the latter. Whereas the charges in imidazolium, pyrazolium, and MnTM-4-PyP⁵⁺ compounds are distributed in plane with the porphyrin ring, the charges of MnTM-2-PyP⁵⁺ are either above or below the plane, which results in a more efficient channeling of the negatively charged superoxide toward the axial positions of the Mn porphyrin, as revealed by kinetic salt-effect measurements.

In a third study, we compared negatively and positively charged porphyrins of the same E _1/2, such as MnBr₈TSPP³⁻ to MnTE-2-PyP⁵⁺. The difference in k _cat, as expected, was much bigger than that when MnBr₈T-2-PyP⁺ and MnTE-2-PyP⁵⁺ were compared (Fig. 5A; Table 1). The overall negative charge of the anionic porphyrins hampered the approach of the negatively charged superoxide, as additionally supported by kinetic salt-effect measurements (262).

With such strong electrostatic effects, the original structure–activity relationship (30) was revised (262), and three separate relations were established to account for the electrostatics of Mn compounds derived from neutral, positively, or negatively charged porphyrins (Fig. 6) (262). With potentials close to the optimum, the pentacationic Mn porphyrins are more than two orders of magnitude more potent SOD mimics than the Mn complexes derived from anionic or neutral porphyrins (Fig. 6). The design of the potent SOD mimics based on anionic and neutral Mn porphyrins is, thus, severely limited by the lack of appropriate electrostatic facilitation, even when the thermodynamics is suitably tuned (262).

OPEN IN VIEWER

FIG. 6.

Structure–activity relations between log k _cat (O₂ ^·−) and E_1/2 (Mn^IIIP/Mn^IIP) for porphyrins that have negative charges (lower curve), no charges (middle curve), and positive charges on the periphery (upper curve).

C. Stability of metalloporphyrins

Because of the macrocyclic effect, all undistorted Mn porphyrins are extremely stable with respect to the metal loss, even in concentrated acids. MnTnHex-2-PyP⁵⁺ undergoes no demetallation for 3 months in 36% HCl. Under such conditions, only 50% of MnTM-2-PyP⁵⁺ loses Mn within a month. As expected, EDTA is not able to demetallate Mn porphyrins under all concentration conditions (37).

D. Aerobic growth of SOD-deficient Escherichia coli

Since the early 1990s, the aerobic growth of the SOD-deficient E. coli strain provided by J. Imlay (JI132), was used as O₂ ^·− specific in vivo assay, and as a first step to identify prospective SOD mimics in vivo. Based on E. coli studies, ortho isomeric Mn(III) N-alkylpyridylporphyrins were forwarded to in vivo mammalian models. In all cases thus far studied, the E. coli model unambiguously and correctly identified compounds that proved efficacious in mammalian studies (241). In addition, the E. coli studies helped us to understand which factors, other than k _cat, contribute to the in vivo efficacy of MnPs. Thus, with the E. coli model, we recently started to comprehend fully the impact of lipophilicity, size, charges, bulkiness, and substituents on the in vivo cellular accumulation and efficacy of MnP (179).

E. Bioavailability of Mn porphyrins

Our growing insight into the in vivo action of SOD mimics taught us that both antioxidant capacity (as a result of thermodynamics and electrostatics of the metal site) and bioavailability of a compound determine its in vivo efficacy. The lack of either of these properties will lead to the absence of efficacy. Quantification of the lipophilicity of SOD mimics has been a challenge until recently. For years we used the thin-layer chromatography retention factor, R _f to assess porphyrin lipophilicity. We recorded very small, severalfold differences only between the R _f values of MnTE-2-PyP⁵⁺ and MnTnHex-2-PyP⁵⁺, whereas the latter was up to 120-fold more potent in vivo, and the former, in some models, was ineffective (see later under the in vivo effects of Mn porphyrins). Recently, we were able to overcome the methodologic difficulties associated with the determination of the partition coefficient of MnPs between n-octanol and water, P _OW (179). Whereas R _f is linearly related to log P _OW, small differences in R _f translate into considerable differences in log P _OW. The P _OW, as opposed to R _f, is a common and practical indicator of drug lipophilicity that allows comparison of MnPs with other drugs (Table 1). By using P _OW, we showed that a ∼10-fold gain in lipophilicity is achieved by either (a) moving the alkyl groups from ortho to meta positions of meso pyridyl substituents, or (b) by increasing the length of alkyl chains by one CH₂ group (Table 1). Because of a significant increase in the lipophilicity (∼13,500-fold MnTnHex-2-PyP⁵⁺ vs. MnTE-2-PyP⁵⁺, and ∼450,000-fold MnTnOct-2-PyP⁵⁺ vs. MnTE-2-PyP⁵⁺), an up to 3,000-fold increase in in vivo efficacy occurs, going from ethyl (MnTE-2-PyP⁵⁺) to hexyl (MnTnHex-2-PyP⁵⁺) to octyl porphyrin (MnTnOct-2-PyP⁵⁺) in different models of oxidative stress (see later under in vivo effects).

F. The effect of the length of the N-alkylpyridyl chains on in vivo efficacy of ortho isomers

With aerobic growth of SOD-deficient E. coli, higher accumulation of lipophilic MnTnHex-2-PyP⁵⁺ within the cell paralleled high SOD-like activity of the cell extract, which in turn resulted in a 30-fold higher efficacy when compared with MnTE-2-PyP⁵⁺ (241). The second study, radioprotection of ataxia telangiectasia cells, showed that compounds that either lack appropriate bioavailability, or possess low or no antioxidant capacity, exert low or no efficacy (255). MnTnHex-2-PyP⁵⁺, but not MnTE-2-PyP⁵⁺, was effective; both compounds have nearly identical abilities to dismute O₂ ^·− and to reduce ONOO⁻ in aqueous solutions. Lipophilic Mn salen compounds and Mn cyclic polyamine of fair antioxidant potency but without positive charges to attract O₂ ^·− or to drive their accumulation in mitochondria, or both, were not efficacious. In a rabbit cerebral palsy study (Tan et al., unpublished data), MnTnHex-2-PyP⁵⁺, but not MnTE-2-PyP⁵⁺, was effective. Preliminary data on the efficacy of MnTnHex-2-PyP⁵⁺ in a rat stroke (MCAO) model are highly encouraging (305).

G. The effect of the location of pyridinium nitrogens with respect to porphyrin meso position: meta vs. ortho vs. para isomeric Mn(III) N-alkylpyridylporphyrins

The effect of the location of alkyl groups on the pyridyl rings with respect to the porphyrin core meso positions is schematically shown in Fig. 7. Although the first evidence of their in vivo effects was published in J Biol Chem 1998 (29), meta isomers have been overlooked for decade. They are 3.6- to 15-fold less-potent SOD mimics, but are 10-fold more lipophilic and accumulate more in E. coli than ortho analogues (Table 1) (178). Figure 7 depicts the most obvious case; meta MnTE-3-PyP⁵⁺ is an ∼10-fold less potent SOD mimic but is ∼10-fold more lipophilic than MnTE-2-PyP⁵⁺. Because of higher lipophilicity and greater planarity plus conformational flexibility, the meta isomer crosses cell wall more easily, which leads to ∼10-fold higher cytosolic accumulation (Fig. 7). Higher accumulation in cytosol overcomes its inferior thermodynamics for O₂ ^·− dismutation; in turn, both isomers exert identical ability to compensate for the lack of cytosolic SOD in SOD-deficient E. coli (18).

OPEN IN VIEWER

FIG. 7.

Higher lipophilicity of meta Mn(III) N -alkylpyridylporphyrins drives their higher accumulation inside E. coli and compensates for lower antioxidant potency when compared with ortho analogues. Consequently, meta and ortho isomers are similarly efficacious in protecting SOD-deficient E. coli that lacks cytosolic SOD (178). Here, the most obvious case with ortho and meta N-ethylpyridylporpyrin is illustrated: meta isomer is ∼10-fold less SOD-active than the ortho species, but is ∼10-fold more lipophilic and accumulates ∼10-fold more in E. coli. In turn, both compounds are equally efficient in substituting for cytosolic superoxide dismutases.

Para isomers appear more lipophilic than their ortho analogue (with the exception of methyl porphyrin). The in vivo studies with the shorter methyl analogue, MnTM-4-PyP⁵⁺, were reported, presumably because of its commercial availability (191, 223, 224). With longer alkyl chains, bulkiness restrain toxic interactions with nucleic acids, whereas lipophilicity may compensate for the lower SOD-like activity. Such analogues may thus be prospective therapeutics.

H. Mitochondrial accumulation of Mn porphyrins

As the awareness of the importance of mitochondria grows, so grows the interest in compounds that may be both mechanistic tools to increase our insight into mitochondrial function and potential therapeutics in mitochondrially based disorders. Michael Murphy (227, 228) advanced the field by showing that redox-able compounds possessing both positive charge and appropriate lipophilicity would enter mitochondria driven by mitochondrial potential. Roberston and Hartley (271) reported a similar design to target mitochondria with a molecule in which cationic N-arylpyridyl (instead of triphenylphosphonium cation) is coupled with nitrone and with a lipophilic moiety. We and others using pentacationic Mn porphyrins wondered what is the intracellular site of accumulation of these excessively charged and thus very hydrophilic compounds. We first aimed to see whether they can enter mitochondria (305). The study was preceded by the Ferrer-Sueta work (108) in which it was shown that, if submitochondrial particles are exposed to ONOO⁻ fluxes, the components of the mitochondrial electron-transport chain were protected with >3 μM MnTE-2-PyP⁵⁺. Our subsequent study, in which C57BL/6 mice were injected with a single IP dose of 10 mg/kg of MnTE-2-PyP⁵⁺, showed that heart mitochondria contained 5.1 μM MnTE-2-PyP⁵⁺; based on the Ferrer-Sueta study, such levels are high enough to protect mitochondria against peroxynitrite-mediated damage (305). Preliminary data from a collaborative study with Edith Gralla (UCLA) (Gralla et al, unpublished data) suggest that all ortho Mn(IIII) N-alkylpyridylporphyrins accumulate in yeast mitochondria at levels which are dependent upon the length of the alkyl chains.

I. Nuclear and cytosolic accumulation of Mn porphyrins

Macrophages and lipopolysaccharide (LPS)-stimulated macrophages were cultured with 34 μM MnTE-2-PyP⁵⁺ for 1.25 h (39). Threefold higher levels of MnTE-2-PyP⁵⁺ were found in nucleus than in cytosol: 35 and 44 ng/mg of cytosolic protein and 99 and 156 ng/mg of nuclear protein when macrophages and LPS-stimulated macrophages were treated (39). It is obvious that positively charged porphyrin favors environments with the abundance of anionic polymers, such as nucleic acids.

J. Pharmacokinetics

K. Other modes of action

Still only limited knowledge exists about the action of synthetic antioxidants/redox modulators in vivo. Even if they possess high k _cat (O₂ ^·−) in vivo, they likely exert more, rather than a single action because of their multiple redox states and varied axial coordination. Therefore, other possible “chemistries,” which are likely dependent on the thermodynamics and electrostatics of the metal site discussed previously, are given here in brief.

L. The effects of Mn porphyrins in suppressing oxidative-stress injuries in vitro and in vivo

M. Fe porphyrins

We and several other groups worked extensively on in vitro and in vivo studies of Fe porphyrins as SOD mimics and ONOO⁻⁻ scavengers (17, 30, 45, 76, 91, 126, 169, 170, 184, 185, 197, 211, 229, 231, 232, 238, 239, 309, 310, 313). The k _cat for O₂ ^·− dismutation, as well as for ONOO⁻ reduction, is very similar for Fe and Mn analogues (184, 185, 310) (Table 1). Beneficial effects in vivo and their impact on transcriptional pathways were reported (17, 91, 169, 170, 309), yet toxic effects also were published (234, 239). When in a reduced state, both Mn and Fe porphyrins could release metals; with H₂O₂ present, Fe, but not Mn, would produce highly damaging ^·OH radicals (234). Also, like cyt P450 enzymes, metalloporphyrins in their own right, when metal site is reduced, can bind oxygen and reduce it to superoxide and hydrogen peroxide, leading eventually to the oxidation of other biomolecules. Fe porphyrins are more successful than Mn analogues in doing so, as we reported with hydroxylation of cyclophosphamide (304). Further, a suggestion was made by Ohse et al. (239) that, like “free” iron, the Fe(II) site of FeTM-4-PyP⁵⁺ reacts with H₂O₂ (formed through O₂ ^·− dismutation), giving rise to ^·OH radicals. Finally, richer coordination chemistry of Fe than of Mn porphyrins (50, 53) may make difficult the mechanistic studies on FePs.

N. Cu porphyrins

Copper complexes have not been extensively studied. We have shown that CuBr₈TM-4-PyP⁴⁺ has a significant SOD-like activity, log k _cat = 6.46 (33). Moreover, when compared with the Mn(II) analogue, Mn^IIBr₈TM-4-PyP⁴⁺, Cu porphyrin is significantly more stable and undergoes demetallation only in concentrated sulfuric acid. Whereas SOD activity was achieved by choosing highly electron-deficient porphyrins, such as the β-octabrominated derivatives, simpler Cu porphyrins, such as CuTM-4-PyP⁴⁺, are not SOD mimics (log k _cat < 3.7) (33). Possible Fenton chemistry on Cu(II) site within porphyrin has not been explored.

O. Co and Ni porphyrins

Co porphyrin is not a strong SOD mimic. Pasternack and Skowronek (247) reported the rate constants for the reduction of O^· ₂ ⁻ of 3 × 10⁷ M⁻¹s⁻¹ for FeTM-4-PyP⁵⁺ and 1 × 10⁵ M⁻¹s⁻¹ for CoTM-4-PyP⁺ in 0.05 M carbonate buffer, pH 10.1 (via NBT /XO/X assay) (247). Because NiSOD exists in nature, we were tempted to synthesize the Ni analogue of MnTE-2-PyP⁵⁺, but to our disappointment and despite favorable electrostatics and electron-deficiency of porphyrin ligand, the log k _cat for NiTE-2-PyP⁴⁺ is only 5.43; it is more than two orders of magnitude less active than MnTE-2-PyP⁵⁺. Its k _cat is around the rate constant for O₂ ^·− self-dismutation (5 × 10⁵ M/s, pH 7.0). Contrary to the porphyrin ligand, the ligand field around Ni in NiSOD enzyme allows it to cycle easily between + 2 and + 3 oxidation state with O₂ ^·−.

A. Mn(III) biliverdin and its analogues

We also studied Mn(III) biliverdin [MnBV²⁻]₂ and its analogues with respect to O₂ ^·− dismutation (299, 302). They are dimmers, with each trivalent Mn bound to four pyrrolic nitrogens of one biliverdin molecule and to the enolic oxygen of another molecule. They are also the first compounds shown to dismute O₂ ^·− by using the Mn(IV)/Mn(III) redox couple, which has the E_1/2 = + 450 mV versus NHE. This potential is similar to the potential of the Mn(III)/Mn(II) couple of Mn(III) N-alkylpyridylporphyrins and of the SOD enzyme (302). The complexes exhibit a high k _cat ∼ 5 × 10⁷ M⁻¹s⁻¹ (302) (Table 1). The most recent data indicate that corroles, which may be considered modified porphyrins, dismute superoxide efficiently by using also the Mn(IV)/Mn(III) redox couple (99) (see under Corroles). The usefulness of Mn biliverdins is hampered by their water insolubility.

B. Texaphyrins

Texaphyrins and their lanthanide complexes (Fig. 8) are porphyrin-like compounds (282). Motexafin gadolinium (MGd³, Xcytrin^R) has been used as a cancer chemotherapeutic in Phase III clinical trials. A prooxidative mechanism of tumor killing, through increased ROS production, was proposed. At the expense of NADPH or ascorbate, thioredoxin reductase TrxR would reduce MGd (E_1/2 ∼ −40 mV vs. NHE in N,N'-dimethylformamide), which would in turn transfer electrons to oxygen, producing superoxide and eventually H₂O₂ (148). The noncompetitive inhibition of thioredoxin reductase and ribonucleotide reductase, resulting in increased levels of ROS, may also play a role. Only modest peroxynitrite scavenging ability has been evaluated with an Mn analogue (289). The rate constant for Mn(II)/Mn(III) oxidation by ONOO⁻ is estimated at 3 × 10⁴ M⁻¹s⁻¹, whereby ^·NO₂ is formed. It has been proposed that the Mn(III) compound is reduced back to Mn(II) texaphyrin either by nitrite or by ascorbate in vivo. With fairly negative E_1/2 (though obtained in dimethylformamide, and thus not readily projected onto aqueous systems), and no electrostatic facilitation for O₂ ^·− dismutation, a significant SOD-like activity in aqueous systems is unlikely.

OPEN IN VIEWER

FIG. 8.

Mn(III) 5,10,15 tris ( N -methylpyridinium-2-yl)corrole and Gd(III) texaphyrin.

In addition to anticancer activity, MGd also was proposed to enhance tumor-radiation therapy by enhancing the anaerobic production of ^·OH and aerobic formation of O₂ ^·−. In clinical trials, it did not have sufficient anticancer effect as a single agent (10). It has been tested in combination with radiotherapy in brain metastases and primary brain tumors with varying success. Although promising results were obtained in some clinical trials in combination with radiation [non–small cell lung cancer with brain metastasis (215)], most recently, no effect on overall survival was reported with whole-brain radiotherapy of metastases from solid tumors (326).

Both MGd and MnTnHex-2-PyP⁵⁺ (68) have been tested in an ALS model and exerted similar efficacy; the data still need explanation, given that production of ROS/RNS is reportedly a major pathway for MGd, and the opposite still holds true for MnTnHex-2-PyP⁵⁺ (see also Prooxidative action of metalloporphyrins).

The lutetium analogue (Lu-Tex) is a photosensitizer developed for photodynamic tumor therapy (282). It is retained selectively in tumors, presumably because of the association with low-density lipoproteins. As it localizes in plaques, it has been tested for photoangioplastic treatment of atherosclerotic plaques in peripheral arteries. The efficacy has been demonstrated in various models (282).

C. Corroles

As compared with porphyrins, corroles contain one less meso bridge and therefore have only up to three meso substituents (Fig. 8). They are also tri-anionic compared with di-anionic porphyrin ligands (with respect to pyrrolic nitrogen deprotonation) (285). Thus, they tend to form high-oxidation-state air-stable Mn(IV) complexes (23). The synthesis of water-soluble derivatives led to their increased development for biomedicinal purposes (23). The first step in Mn(III) reduction is thermodynamically very unfavorable, as it occurs at very negative potentials (<−1 V vs. NHE in different solvents) (285). The oxidation of Mn(III) to Mn(IV) corroles is fairly facile and has been first examined in conjunction with the much stronger oxidant, ONOO⁻. The potency of metal corroles (Mn, Fe, and Ga), as well as the ability to attenuate atherosclerosis (144), has thus far been ascribed to peroxynitrite or H₂O₂ scavenging; catalytic action was provided by the reduction of Mn(IV) corrole to Mn(III) corrole with either nitrite or ascorbate. The k _cat ranged from ∼4 × 10⁴ (negatively charged corrole with three pentafluorophenyl meso groups and two sulfonated pyrrolic positions) to 4 × 10⁵ M⁻¹s⁻¹ (positively charged corrole with pentafluorophenyl group on one meso bridge and two para methylpyridyl groups on two opposing meso bridges) (127). The values for Fe corroles are much higher: ≤ 2 × 10⁶ M⁻¹s⁻¹ for a positively charged corrole (200). Most recently, the synthesis of ortho isomeric N-methylpyridylcorrole, MnTrM-2-corrole³⁺ and analogue of a potent SOD mimic and ONOO⁻ scavenger, MnTM(or E)-2-PyP⁵⁺, was reported (Fig. 8) (275). The SOD-like activity of this corrole and of other Fe and Mn anionic and cationic corroles, along with Mn(IV)/Mn(III) reduction potentials, has just been reported (99). The first step in a dismutation process, the oxidation of a metal center from + 3 to + 4 oxidation state is rate limiting, as opposed to most of the porphyrins, other than biliverdins. As a result of the thermodynamic and electrostatic effects, the log k _cat varies from 5.68 to 6.34. The E_1/2 for the Mn(IV)/Mn(III) redox couple varies between + 760 and + 910 mV versus NHE. In agreement with a bell-shaped curve that we have established for porphyrins (37, 38, 262, 266), as the reduction potential increases, the oxidation of porphyrins (and in this case, a corrole) eventually becomes a rate-limiting step, and the further enhancement of k _cat may be achieved only by decreasing the E_1/2 (introducing the electron-donating groups), which would in turn favor the oxidation of metal corrole (reduction of O₂ ^·−). MnTrM-2-corrole³⁺ bears 2 less charges than analogous porphyrin, MnTE-2-PyP⁵⁺, and is thus more lipophilic. Enhanced lipophilicity may in vivo compensate for lower k_cat. Of note, whereas structural characteristics of new corroles were reported, no elemental analysis data were provided (22, 143, 275).

The antitumor effects of a 5,10,15-tris[2,3,5,6-tetrafluoro-4-(N-methyl-2-pyridinium)]corrole was reported by Aviezer et al. (22, 143) in vitro and in vivo. At 1–10 μg/ml, corrole inhibited binding of FGF2 to the FGF receptor. When Lewis lung carcinoma D122 cells (200,000 cells/mouse) were injected into foot pads of the 10-week-old C57 black mouse, mice were allowed to develop primary tumors for 4 weeks. Tumors were then amputated, and metastases were allowed to develop for 4 weeks before mice were killed. Mice were treated with corrole at 5 mg/kg twice per week. A significant suppression of tumor metastasis was observed. Peroxynitrite-based protection of rat β cells with cationic corroles also was reported (242).

The significant tumor growth suppression and tumor-imaging properties of the fluorescent, negatively charged gallium sulfonated corrole noncovalently associated with breast cancer–targeted cell-penetration protein (HerOBK10), which accumulates in HER2⁺ tumors, was recently reported; the complex retains integrity in human serum and accumulates in tumor (4, 5, 43). Nude mice bearing HER2⁺ received daily IV injections of 0.008 mg/kg of HerGa complex for 7 days, starting at the time tumors reach 250–300 mm³ volume; tumor size was measured until 25 days of growth. Sulfonated gallium corrole has a modest effect on tumor growth in its own right. The mechanism of action is not fully understood (4, 5).

A. SOD-like activity of Mn salens

Mn salen compounds are complexes of Mn with N,N'-bis(salicylidene)ethylenediamine (Fig. 1). Although salen compounds have been of interest in catalysis since the beginning of the last century, the possible role of their Mn complexes in decreasing oxidative-stress injuries arose quite recently and parallels studies on another two groups of metal complexes, Mn cyclic polyamines and Mn porphyrins (40, 93). EUK-8, the Mn salen (Mn(III) complex with prototypical N,N'-bis(salicylidene)ethylenediamine) (Fig. 1), has a fairly negative reduction potential for the Mn(III)/Mn(II) redox couple, being −130 mV versus NHE, which is insufficient to provide high efficacy in catalyzing O₂ ^·− dismutation (302). Moreover, with only one positive charge on the Mn site, electrostatic guidance of O₂ ^·− to the Mn site is missing. The SOD-like activity of a EUK-8 (Fig. 1) is close to the activity of MnCl₂ in 0.05 M phosphate buffer, k _cat = 6.0 × 10⁵ M⁻¹s⁻¹ (302), which is further similar to 1.3 × 10⁶ M⁻¹s⁻¹ determined by nitrobluetetrazolium assay by Baudry et al. (40), and represents only ∼0.1% of the k _cat of the SOD enzymes (24, 171). Because of the lack of the macrocyclic effect, both we and Baudry et al. (40) reported the loss of SOD-like activity of EUK-8 in the presence of EDTA due to the formation of the SOD-inactive, Mn complex with EDTA (14, 302). This may hold true for EUK-134 and EUK-189. Therefore, such derivatives are likely to lose Mn in vivo in the presence of other carboxylate- and phosphate-based cellular chelators. In EUK-207 (Fig. 1), the structural modifications have been made to enhance the stability by cyclizing the ligand with an eight-membered crown ether moiety, thus introducing the macrocyclic effects similar to the porphyrin (94). The catalase-like activity and the cytoprotective effects in vivo have been preserved in comparison with the EUK-8, EUK-134, and EUK-189 series. The in vivo effects of the latter three compounds have been continuously reported. Our data on C. neoformans (see later) suggest that EUK-8 may possess sufficient stability to reach, in an intact form, the targeted subcellular compartment and release Mn there (131). Doctrow et al. (94) reported that 220 μM Mn salens [EUK-113 (similar to EUK-134 but with axial acetate instead of chloride); EUK-189; EUK-178 (similar to EUK-113 but with an upper aromatic bridge); EUK-207] (Fig. 1) are stable for hours with 25 mM EDTA at room temperature and at pH 7.4. The most stable EUK-207 was found largely intact even after 70 h under same conditions (94). The EUK-134 (Fig. 1) analogue possesses two methoxy groups on phenyl rings instead of hydrogens; when compared with EUK-8; it has only slightly enhanced SOD-like activity, but alkoxy groups appear to enhance catalase-like activity (1, 130, 245).

B. Catalase-like activity of Mn salens

Mn salen derivatives reportedly possess significant catalase- and peroxidase-like activity. It has been reported that EUK-8 and EUK-134 have 0.6 and 0.5 U/mg (171), whereas the enzyme has ∼5,000 U/mg. If calculated per milligram basis, the EUK-8 would have 0.01% of catalase activity. If one does the calculation per molar basis, assuming the MW of the enzyme to be 250,000, and considering its tetramer structure (MW = 62,500), and the MW = 357 for EUK-8, EUK-8 has 7 × 10⁻⁵% of catalase activity (171). In agreement with these calculations, Sharpe et al. (284) considered Mn salens to be poor catalysts of H₂O₂ breakdown (284). Although SOD-like activity is reportedly not dependent on the structure, the structure–activity relationship has been established for catalase-like activity (93, 94). Alkoxy groups in position 3 (EUK-134, EUK-189) enhanced, whereas in position 5, decreased the catalase activity when compared with EUK-8 compound. Although less stable (EUK-178 decomposes in aqueous solution within days), compounds containing six-membered aromatic bridge possess up to sevenfold higher catalase-like activity when compared with unsubstituted EUK-8 (93).

C. Reactivity toward other ROS/RNS

Besides O₂ ^·−- and H₂O₂-related activity, Sharpe et al. (284) reported the reactivity of EUK-8 and EUK-134 toward OCl⁻ and ONOO⁻, in which Mn(V) oxo species were formed by a two-electron process, thus giving rise to the benign species, H₂O, O₂, and NO₂ ^·−. The Mn(V) oxo species can react with ^·NO, giving rise to ^·NO₂, which will combine with ^·NO, giving rise to N₂O₃, which, with water, will produce HNO₂. PEG-ylated compounds were also prepared (245). We provided data that although PEG-ylated porphyrins possess higher SOD-like activity than the non–PEG-ylated analogues (36), their cellular localization is hampered by their increased size (36). Thus, PEG-ylated salen compounds, although two- to threefold more potent than EUK-134, may also be of excessive size to enter the cell or mitochondria at sufficient levels (245).

D. Mn salens in suppressing oxidative-stress injuries in vivo

The in vivo effects of EUK compounds, as already noted, reflect their ability to scavenge different reactive species and not specifically superoxide or peroxide. Further, like porphyrins, such activities result in modulating cellular transcriptional activity, affecting NF-κB and MAPK pathways (260, 339). EUK-134 was first reported by Melov et al. (216) to increase life span of the nematode Caenorhabditis elegans by 50% (216), which was later contradicted by Keaney et al. (171) [i.e., the effect was seen only if superoxide generators (paraquat and plumbagin) were used]. The data were further complicated with the effect on the decreasing life span of domestic fly, Musca domestica, under hypoxic conditions (41). EUK-8, EUK-134, and EUK-189 were shown to affect activation of microglia with the 42-amino-acid form of β-amyloid peptide (A β 42) when insulted with H₂O₂ or ONOO⁻ (11). A modest but significant effect was observed in extending the survival in a mouse model of prion disease with IP injections of 30 mg/kg of EUK-189, 7 days per week, starting at day 7 after the inoculation of the diseased brain homogenate (51). EUK-134 was able to decrease infarct size in a rat stroke model when given in a single dose of 2.5 mg/kg IV at 3 h after MCAO, and assessed at 21 h after MCAO (24). We showed with MnTDE-2-ImP⁵⁺ that such protective action in single treatment, on the level of primary oxidative stress as a direct consequence of ischemia–reperfusion, fades away later as a result of excessive upregulation of cellular transcriptional activity/inflammation (288). Prolonged treatment with an antioxidant that would modulate secondary oxidative stress is thus necessary; in turn, both delayed-treatment and long-term effects must be evaluated to address the clinical utility of any antioxidant for stroke treatment. In apoptosis inducing factor–deficient (harlequin) mouse, EUK-8 reduces oxidative stress after 4 weeks of pressure-overload–induced heart failure if given IP at 25 mg/kg/day, 3 times per week, for the duration of the study (324). Along with Mn cyclic polyamine M40403, EUK-134 protected mouse heart from superoxide-induced reperfusion injury (343). EUK-8 diminished the consequences of multiple organ failure in endotoxic shock caused by E. coli lipopolysaccharide (6 mg/kg) if given continuously IV at 0.3 and 1 mg/kg/h for 6 h (213). Along with Mn porphyrins, Mn salen derivatives attenuated lung inflammation in a porcine model of LPS-induced acute respiratory syndrome (260), supposedly through decreasing oxidative stress, which in turn suppresses NF-κB activation and pro-inflammatory gene expression (260). A mild effect on the survival of the G93A mouse in an amyotrophic lateral sclerosis model (eight low copies model) was observed with EUK-134 and EUK-8 (33 mg/kg, 3 times per week for the duration of the study), the former being more effective (164). The effect of Mn salens in neurodegenerative disorders was summarized in a review by Patel (133). The effect on the inhibition of MAPK signaling pathways and decrease in p53 accumulation in ultraviolet B–exposed primary human keratinocytes, leading to increased cell survival, was reported with EUK-134 (84). A rather unexpected effect of EUK-8 as well as of the NADPH oxidase inhibitor, diphenyleneiodonium, on promoting the induction of TGF-β, was observed (339). EUK-8 prolonged the survival of islet allografts (from BALB/c mice) in newly diabetic female NOD mice after treatment with 100 mg/kg for 50 days (starting at 6 weeks of age), and prevented the spontaneous type 1 diabetes with NOD mice for up to 1 year, if given at 100 mg/kg every other day for only 35 weeks (243). At 5 mg/kg/day for 50 days, EUK-8 prevented adoptive transfer of type 1 diabetes. EUK-189, given as a single SC injection of 70 mg/kg, 24 h before the whole-body radiation, exerted radioprotection of two mouse strains (C3H/HeN and CD2F1) (307). ⁶⁰Cobalt gamma irradiation was applied (midline tissue dose was 1–10 Gy) (307). Protection also was seen if the drug was given at 6 h after radiation but with one strain only (C3H/HeN). The LD_50/30 values were 7.96 and 9.13 Gy for saline- and EUK-189–treated groups. A study is in progress to fully evaluate EUK compounds as potential mitigators. Doctrow et al. (94) reported an increased survival of MnSOD^−/− mice, improved neurologic deficits, and protected mitochondrial aconitase and complexes I–IV with EUK-207 and EUK-189. MnTBAP³⁻ was less effective, which the authors attribute to its lower ability to cross an intact blood–brain barrier and protect brain mitochondria (94). M40403 also was inefficient, presumably because of the inability to cross the blood–brain barrier or to enter mitochondria or both. In MnSOD-knockout yeast, C. neoformans, Mn salen (EUK-8) and ascorbate were able to offer protection, but not MnCl₂, or Tempol, or any of the five Mn porphyrins (cationic and anionic) (133). In aerobic growth of SOD-deficient E. coli and radioprotection of ataxia telangiectasia, Mn salen was, however, of inferior efficacy, as compared with Mn porphyrin, MnTnHex-2-PyP⁵⁺ (225, 241, 255). A limited effect of Mn salen was observed in a later study, and none, in the former. As in other redox-able metal complexes, Matthijssens et al. (210) reported that EUK-8 has prooxidant activity also, increasing the levels of ROS in E. coli; no protection of the SOD-deficient E. coli strain, when growing aerobically, was detected (210). When Mn salen and MnTM-2-PyP⁵⁺ were compared in a small bowel ischemia–reperfusion injury (given to rats into the portal vein before and after ischemia at 1 mg/kg), similar abilities to scavenge reactive species were observed (331). Prevention of cognitive deficits and brain oxidative stress was reported with EUK-189 and EUK-207 when C57/BL6 mice were injected with Mn salens via osmotic pumps for 6 months, starting at 17 months of age (62).

With EUK compounds, control experiments involving Mn²⁺ may be beneficial primarily for mechanistic purposes, to distinguish whether the effects observed originate from Mn ligated in Mn salen complex, or from Mn²⁺ in its own right; in other words, is Mn salen transporting Mn into the cell?

A. SOD-like activity

Riley et al. (21, 204, 269, 277) developed Mn complexes with cyclic polyamines/tetraaza crown ethers as catalysts for O₂ ^·− dismutation (Fig. 9). Structural modifications led to compounds of increased activity; complexes with k _cat as high as 10⁹ Ms have been reported (21). The detailed kinetic study published by Maroz et al. (204) found that the reaction of a transient adduct (formed with the first O₂ ^·− molecule) with the second molecule of O₂ ^·− is particularly dependent on the structure of the polyamine.

OPEN IN VIEWER

FIG. 9.

Mn(II) cyclic polyamines: the SOD-active M40403 (2S,21S nonmethylated analogue) and SOD inactive M40404 (2R,21R dimethyl derivative).

Thus, among three compounds studied, the so-called M40403 (SODm1), which is the 2S,21S nonmethylated analogue, has a k _cat of 3.55 × 10⁶ M⁻¹s⁻¹ (Fig. 9). Its 2S,21S dimethyl derivative (SODm2) is 100-fold more active, k _cat = 2.35 × 10⁸ M⁻¹s⁻¹ at pH 7.4, whereas its 2R,21R dimethyl derivative (M40404, SODm3) is inactive. A sevenfold higher rate constant for the 2S,21S dimethyl derivative was reported by Aston et al. (21) (k _cat = 1.6 × 10⁹ M/s at pH 7.4). The E_1/2 values for all three complexes are fairly similar at +525 (SODm1), +464 (SODm2), and +452 mV versus NHE (SOD m3) in acetonitrile (204, 269) (Table 1), and cannot justify the total lack of activity of the 2R,21R derivative. Such values suggest that other effects, particularly conformational flexibility of pseudooctahedral geometry of oxidized species (as determined by X-ray data and may not necessarily describe the geometry in aqueous solution) presumably govern the differences in the SOD-like activity of those three compounds. The catalysis involves a fast oxidation of [Mn^II(L)]²⁺ with O₂ ^·− to [Mn^III(L)O₂]⁺, followed by the second rate-determining reduction of [Mn^III(L)O₂]⁺ to Mn^II(L)O₂], which rapidly equilibrates with [Mn^II(L)]²⁺. Temperature studies revealed major differences in the thermodynamics of the catalytic cycles involving SODm2 or SODm1. In the case of SODm2, the observed high entropic contribution to the activation energy is indicative of ligand conformational changes during the catalytic cycle (21). The k _cat also may be increased if the charge distribution and substituents are changed. The most active compound (SODm2) has not been used in any in vivo studies. These complexes contain Mn in a +2 oxidation state and are thus of limited metal/ligand stability, with log K = 13.6 for the mostly in vivo studied polyamine 2S,21S-nonmethyl analogue, M40403 (269). Riley et al. (230) found the intact complex down to pH 6.1 only, which may be insufficient to account for the in vivo efficacy. Based on the reported in vivo efficacy, Mn cyclic polyamines may act as Mn transporters. Once Mn is lost from the complex, the ligand, which is in essence an aza crown ether with high affinity toward metals (particularly mono and divalent), may bind other biologic cations, therefore exerting toxicity. The specificity of Mn cyclic polyamines toward O₂ ^·− was claimed (277); activities toward ONOO⁻, ^·NO and H₂O₂ were assumed absent. Such specificity may be desirable for mechanistic studies, but not for therapeutic purposes.

B. Mn(II) cyclic polyamines in suppressing oxidative stress in vivo and in vitro

M40403 is thus far the most studied cyclic polyamine in vivo. It suppressed inflammation in a variety of models: carrageenan-induced pleurisy mouse model (75), rat model of septic shock (198), E. coli lipopolysaccharide serotype 0111:B4 (LPS)-induced cytokine production by cultured rat alveolar macrophages (233), inflammatory pain in a carrageenan model of paw edema (230, 277, 329), guinea-pig model of allergic asthma-like reactions (207), rodent model of colitis (74), myocardial ischemia–reperfusion injury (208, 343), chronic hypoxia-induced pulmonary hypertension (86), and so on. M40403 also was reported to provide protection against radiation-induced mucositis in the Syrian golden hamsters. The optimal regiment was 30 mg/kg, given IP twice daily for 3 days, starting at 30 min before 40-Gy focus radiation (226). The radiation was given at 40 Gy to a cheek pouch (226). The action on NF-κB signaling pathways has been indicated in the carrageenan-induced mouse pleurisy model (316). The strong enhancement of the antitumor effect of interleukin-2 (IL-2) by M40403, ascribed to the ROS-scavenging ability of the SOD mimic, has been reported (278). In addition, M40403 prevented IL-2 from causing dose-limiting hypotension. Since ActivBiotics, Inc., acquired the rights to M40403 (previously Metaphore) a few years ago, only a single report has been published. Recently, the protective effects of tetraaza polyamine (derivatized with carboxylates) with activated macrophages at nanomolar levels and in an animal acute and chronic inflammation model (5–15 mg/kg, single injection) were reported. The metal/ligand stability (log K = 14.73), and the k _cat (O₂ ^·−) = 1.1 × 10⁶ M⁻¹ s⁻¹ were given (103).

A. Fullerenes

B. Nitroxides