Design of Mn porphyrins for treating oxidative stress injuries and their redox-based regulation of cellular transcriptional activities

Abstract

The most efficacious Mn(III) porphyrinic (MnPs) scavengers of reactive species have positive charges close to the Mn site, whereby they afford thermodynamic and electrostatic facilitation for the reaction with negatively charged species such as ${\text{O}}_2^{•-}$ and ONOO⁻. Those are Mn(III) meso tetrakis(N-alkylpyridinium-2-yl)porphyrins, more specifically MnTE-2-PyP⁵⁺ (AEOL10113) and MnTnHex-2-PyP⁵⁺ (where alkyls are ethyl and n-hexyl, respectively), and their imidazolium analog, MnTDE-2-ImP⁵⁺ (AEOL10150, Mn(III) meso tetrakis(N,N′-diethylimidazolium-2-yl) porphyrin). The efficacy of MnPs in vivo is determined not only by the compound antioxidant potency, but also by its bio-availability. The former is greatly affected by the lipophilicity, size, structure, and overall shape of the compound. These porphyrins have the ability to both eliminate reactive oxygen species and impact the progression of oxidative stress-dependent signaling events. This will effectively lead to the regulation of redox-dependent transcription factors and the suppression of secondary inflammatory- and oxidative stress-mediated immune responses. We have reported on the inhibition of major transcription factors HIF-1α, AP-1, SP-1, and NF-κB by Mn porphyrins. While the prevailing mechanistic view of the suppression of transcription factors activation is via antioxidative action (presumably in cytosol), the pro-oxidative action of MnPs in suppressing NF-κB activation in nucleus has been substantiated. The magnitude of the effect is dependent upon the electrostatic (porphyrin charges) and thermodynamic factors (porphyrin redox ability). The pro-oxidative action of MnPs has been suggested to contribute at least in part to the in vitro anticancer action of MnTE-2-PyP⁵⁺ in the presence of ascorbate, and in vivo when combined with chemotherapy of lymphoma. Given the remarkable therapeutic potential of metalloporphyrins, future studies are warranted to further our understanding of in vivo action/s of Mn porphyrins, particularly with respect to their subcellular distribution.

Introduction

The increased knowledge of the impact of normal cellular redox status, i.e. the balance between the endogenous antioxidant systems and reactive oxygen (ROS) and nitrogen (RNS) species, on cell function (Halliwell and Gutteridge 2007; Jones 2008; Trachootham et al. 2009; McCord 2000) has been accompanied by the need to restore it under pathological conditions. Several types of synthetic antioxidants as well as natural antioxidants have been employed in different models of oxidative stress injuries such as ischemia/reperfusion, radiation, cancer, diabetes, aging, injuries of central nervous systems, etc. We recently summarized the ROS/RNS-related chemistry as well as in vivo effects of several major groups of synthetic SOD mimics (Batinic-Haberle et al. 2010). Herein, we addressed the design and development of the most potent and bioavailable Mn porphyrin-based therapeutics and their role in redox-based cellular signaling pathways.

The ability of Mn porphyrins and other potent synthetic and endogenous antioxidants to easily donate and accept electrons allows them to exert both anti- and pro-oxidative activities. Both the lack of ROS/RNS and their excessive formation have been known to dysregulate cellular metabolism (McCord 2008). The anti-oxidative action of Mn porphyrins has been well substantiated and proven beneficial. However, the pro-oxidative action/s may be therapeutically beneficial as well. Several published reports provided experimental evidences that imply the pro-oxidative mode of action of MnPs with respect to (1) suppression of transcription factors activation (Tse et al. 2004) and (2) tumor killing (Jaramillo et al. 2009; Ye et al. 2009; Spasojevic et al. 2006). Understanding the action/s of MnPs (and other classes of antioxidants) remains a challenge.

Design of Mn porphyrins with respect to their ability to dismute ${\text{O}}_2^{•-}$

The most obvious approach to the design of potent SOD mimics has been to mimic the electrostatic and thermodynamic properties of the SOD enzyme. Thus, such compounds should catalyze the ${\text{O}}_2^{•-}$ dismutation (Eqs. 1 and 2) around the metal-centered reduction potential of the SOD enzyme: E_1/2 of ~+300 mV versus NHE. This is the midway potential between the potential for the reduction (+890 mV vs. NHE) and oxidation of ${\text{O}}_2^{•-}$ (−160 mV vs. NHE); therefore, the enzyme reduces and oxidizes ${\text{O}}_2^{•-}$ with identical rate constants of 2 × 10⁹ M⁻¹ s⁻¹ (Fig. 1) (Klug-Roth et al. 1973; Vance and Miller 1998; Michel et al. 2005; Goldstein et al. 2006). The remarkable contribution of the electrostatics to the ${\text{O}}_2^{•-}$ dismutation that allows the enzyme to operate at diffusion limited rates will be addressed further below. The unsubstituted Mn porphyrins (such as MnT-2(or 4)-PyP⁺ and MnTPP⁺) have E_1/2 for Mn^IIIP/Mn^IIP redox couple of ~−300 mV versus NHE (Fig. 1; Table 1) (Batinic-Haberle et al. 2010), and thus have Mn stabilized in +3 oxidation state and cannot be reduced in the first step by ${\text{O}}_2^{•-}$ (Eq. 1). Moreover, the substituted porphyrins that bear negative charges (MnTBAP³⁻ and MnTSPP³⁻), and no charges at the periphery (MnhematoP⁻), with E_1/2 ~ −200 mV versus NHE (Batinic-Haberle et al. 2010) cannot be reduced with ${\text{O}}_2^{•-}$. In order to increase the reducibility of Mn site, i.e. the affinity of MnPs for electrons, the metal center must be made electron-deficient

Fig. 1.

Design of Mn porphyrins based on thermodynamic considerations

Table 1.

The k_cat for ${\text{O}}_2^{•-}$ dismutation, k_red for ONOO⁻ reduction, E_1/2, Mn^III/Mn^II, and lipophilicity expressed as partition between n-octanol and water, P_OW for Mn(III) porphyrins and few other compounds for comparison purposes

Mn compounds	log kcat ( ${\text{O}}_2^{•-}$)	log kred (ONOO−)	E1/2 (mV versus NHE)	POW (MnIIIP)
MnTE-2-PyP5+	7.76a	7.53i	+228a	−6.89n
MnTnBu-2-PyP5+	7.25b	7.11i	+254b	−5.11n
MnTnHex-2-PyP5+	7.48b	7.11i	+314b	−2.76n
MnTnOct-2-PyP5+	7.71b	7.15i	+367	−1.24n
MnTMOE-2-PyP5+	8.04c	7.36k	+251m	~−6.53o
MnTM-3-PyP5+	6.61d	6.62j	+52d	−6.96n
MnTE-3-PyP5+	6.65n		+54n	−5.98n
MnTDE-2-ImP5+	7.83c	7.43k	+346m	~−6.48o
MnBr8TM-3-PyP4+	≥8.85r		+468r
MnT(2,6-Cl2-3-SO3P)P3−	6.00a		+88a
MnBr8TSPP3−	5.56f		+209f
Mn(hematoP)−	~3.00e		−230a
MnTBAP3−	<3.50f	5.02l	−194f
MnT-2-PyP+	4.29p		−280p
MnT-4-PyP+	4.53p		−200p
Mn(salen)+ (EUK-8)	5.78g		−100g
MnCl2	6.1–7.7h		+850q
Cu,ZnSOD	~9f	3.97l	~+300f

^aBatinic-Haberle et al. (1999)

^bBatinic-Haberle et al. (2002)

^cBatinic-Haberle et al. (2004)

^dBatinic-Haberle et al. (1998)

^eEstimate based on E_1/2

^fRebouças et al. (2008a, b, c)

^gSpasojevic et al. (2001)

^hSpasojevic et al. (2001), Rebouças et al. (2008a, b, c), Barnese et al. (2008), and Archibald and Fridovich (1982)

ⁱFerrer-Sueta et al. (2003)

^jFerrer-Sueta et al. (1999)

^kDetermined at 25°C, pH 7.32, then recalculated to 37°C using MnTE-2-PyP⁵⁺ as a reference (Ferrer-Sueta et al., unpublished)

^lBatinic-Haberle et al. (2009a, b)

^mBatinic-Haberle et al. (2004)

ⁿKos et al. (2009a)

^oEstimated from P_OW versus R_f plot for ortho Mn(III) N-alkylpyridylporphyrins based on TLC R_f values from Batinic-Haberle et al. (2004)

^pSpasojevic and Batinic-Haberle (2001)

^qSpasojevic and Batinic-Haberle (2001), oxidation potential only, Mn^II/Mn^III redox couple is irreversible

^rDeFreitas-Silva et al. (2008)

$${\text{Mn}}^{\text{III}}{\text{P}}^{+}+{\text{O}}_2^{•-}\iff {\text{MN}}^{\text{II}}\text{P}+{\text{O}}_2$$ (1)

$${\text{Mn}}^{\text{II}}\text{P}+2{\text{H}}^{+}+{\text{O}}_2^{•-}\iff {\text{Mn}}^{\text{III}}{\text{P}}^{+}+{\text{H}}_2{\text{O}}_2$$ (2)

electron-withdrawing groups attached to the porphyrin meso and/or to beta pyrrolic positions. The unlimited possibilities of porphyrin modifications at meso and beta positions allowed us to modify metal-centered reduction potential within 1 V range (Fig. 1). The very first compound of reasonable SOD-like activity, FeTM-4-PyP⁵⁺, was characterized by Pasternack and Halliwell (1979). The porphyrin is substituted with electron-withdrawing positively charged pyridyl meso substituents. The cationic nitrogens are placed in para (4) positions on pyridyls with respect to meso porphyrinic positions (Figs. 1, 2).

Fig. 2.

Three most often in vivo studied ortho isomeric Mn(III) meso-tetrakis(N-ethyl(or n-hexyl)pyridinium-2-yl)porphyrins, MnTE-2-PyP⁵⁺ (AEOL10113), MnTnHex-2-PyP⁵⁺, and the N,N′-diethylimi-dazolium analog, MnTDE-2-ImP⁵⁺ (AEOL10150). The meso and beta positions on porphyrin ring are indicated as well as ortho, meta, and para positions of pyridyl nitrogens with respect to porphyrin meso positions. Also shown are three Mn porphyrins that are both meso- and beta-substituted. Such compounds are particularly valuable tool for the design of perspective therapeutics

Later, Mn analog, MnTM-4-PyP⁵⁺ was also studied (Peretz et al. 1982; Lee et al. 1998). However, both compounds are of highly planar structure (Koner and Goldberg 2009). Thus their positive charges are exposed and allow the porphyrin to associate with anionic phosphates of nucleic acids in vivo, which in turn imposes toxicity and loss of SOD-like activity (Fig. 2)(Batinic-Haberle et al. 1998). To reduce their affinity for nucleic acids, their bulkiness was increased by moving methyl groups from para (4) onto ortho (2) positions (Fig. 2) (Batinic-Haberle et al. 1998). Due to the crowding between the meso N-pyridyl alkyls and β-pyrrolic hydrogens, alkylpyridyl groups of ortho isomers cannot rotate freely, but are stuck in a position vertical to the porphyrin plane. As a result of the increased bulkiness, charges are shielded and the ortho isomer does not associate with nucleic acids, and thus, does not lose SOD-like activity in vivo as was shown in E. coli studies (Batinic-Haberle et al. 1998). Importantly, placement of alkyl groups on ortho positions afforded additional increase in Mn electron-deficiency, and enhanced electrostatic attraction of superoxide and peroxynitrite to the Mn site (see below). The ortho isomer, MnTM-2-PyP⁵⁺ has the E_1/2 = +220 mV versus NHE, very similar to the E_1/2 of the enzyme. Consequently, it has high k_cat of 5.8 × 10⁷ M⁻¹ s⁻¹ (Batinic-Haberle et al. 1998). Several other meso pyridyl porphyrins were synthesized having E_1/2 ~ +300 mV versus NHE (Batinic-Haberle et al. 2002, 2004, 2006). We also synthesized Mn porphyrins that have both meso and beta electron-withdrawing substituents. In such compounds the charges on meso substituents are maintained to afford favorable electrostatics and E_1/2 was further increased; thus, they are the most potent SOD mimics, and “sit” in the plateau region of the bell-shape curve of the structure–activity relationship (Fig. 3). With eight bromines in all beta positions, and meso meta N-methylpyridyl groups, MnBr₈TM-3-PyP⁴⁺, has E_1/2 of +460 mV versus NHE, and is the most potent SOD mimic thus far synthesized, with k_cat identical to the k_cat of the SOD enzyme (DeFreitas-Silva et al. 2008) (Fig. 2; Table 1). Yet, such high positive E_1/2 stabilized the Mn +2 oxidation state to such extent that the compound exists as a Mn(II) porphyrin. In turn, it is insufficiently stable in solution, easily loses metal, and therefore, ceases to be an SOD mimic. While excellent mechanistic tools in designing SOD mimic, such MnPs are of no practical importance as therapeutics. As compared to other types of antioxidants, most potent Mn porphyrins, particularly our lead compounds, meso substituted MnTE-2-PyP⁵⁺ and MnTnHex-2-PyP⁵⁺, are extremely stable under all conditions of acidity, dilution, temperature, and light.

Fig. 3.

Structure–activity relationship between $log{k}_{\text{cat}}({\text{O}}_2^{•-})$ and E_1/2 (Mn^IIIP/Mn^IIP) in mV versus NHE for cationic Mn(III) porphyrins. Compounds at the plateau of the bell-shaped curve are those that bear both meso and beta substituents and have highest k_cat. Yet, they are stabilized in Mn +2 oxidation state, lose readily Mn and are thus are only important for mechanistic purposes. The falling limb of the bell-shaped curve is more obvious in SARs of anionic and neutral porphyrins (Batinic-Haberle et al. 2010). Compounds are: 1 MnT(TFTMA)P^5+, 2 MnBM-2-PyP⁵⁺, 3 MnTM-4-PyP⁵⁺, 4 MnTM-3-PyP⁵⁺, 5 MnTrM-2-PyP⁵⁺, 6 MnTnBu-2-PyP⁵⁺, 7 MnTnPr-2-PyP⁵⁺, 8 MnTnHex-2-PyP⁵⁺, 9 MnTDMOE-2-ImP⁵⁺, 10 MnTnOct-2-PyP⁵⁺, 11 MnClTE-2-PyP⁵⁺, 12 MnTE-2-PyP⁵⁺, 13 MnTM-2-PyP⁵⁺, 14 MnTDE-2-ImP⁵⁺, 15 MnTM,MOE-2-ImP⁵⁺, 16 MnTMOE-2-PyP⁵⁺, 17 MnTDM-2-ImP⁵⁺, 18 MnCl₂TE-2-PyP⁵⁺, 19 MnCl₃TE-2-PyP⁵⁺, 20 MnCl₅TE-2-PyP⁵⁺, 21 MnCl₄TE-2-PyP⁵⁺, 22 MnBr₈TM-4-PyP⁴⁺, and 23 MnBr₈TM-3-PyP⁴⁺. Data are taken from Batinic-Haberle et al. (2010) and Kachadourian et al. (1998). The names of porphyrins are listed in the list of abbreviations

Some of our compounds, such as MnBr₈TSPP³⁻ (Rebouças et al. 2008a, b, c) (Fig. 2; Table 1), and members of other classes of synthetic antioxidants such as Mn(salens)⁺ and Mn cyclic polyamines possibly act as transporters of Mn into the cell and its compartments. Mn low molecular weight complexes possess fair SOD-like activity, and can under certain conditions protect against ROS/RNS-mediated injury (Batinic-Haberle et al. 2010 and refs therein).

A number of Mn porphyrins were analyzed, and structure–activity relationship (SAR), which relates the log k_cat to E_1/2, was established (Fig. 3). It encompasses compounds of different electrostatics and stericity (Batinic-Haberle et al. 1999). The SAR showed that higher (the more positive) E_1/2 gives rise to the higher k_cat. At very low potentials, E_1/2 ≪ +200 mV versus NHE, the Mn^IIIPs are stabilized in their +3 oxidation state, and their reduction to Mn^IIP (eq [1]) is a rate limiting step (raising limb of SAR). With the increase in E_1/2, the reduction of Mn^IIIP becomes easier, and thus k_cat increases. Alike enzyme, with MnTE-2-PyP⁵⁺ at +228 mV versus NHE, the rate constants for both steps of dismutation process, oxidation and reduction of MnP with ${\text{O}}_2^{•-}$ are similar, 2.5 × 10⁷ M⁻¹ s⁻¹ and 8.2 × 10⁷ M⁻¹ s⁻¹, respectively (Batinic-Haberle et al. 2004). At the potentials around +500 mV, the maximal k_cat was achieved. As the E_1/2 increases further beyond +500 mV, the electron-deficiency of the metal site increases so much that eventually Mn(II) porphyrin becomes a stable species in solution, and the oxidation of Mn^IIP to Mn^IIIP (Eq. 2) becomes a rate-limiting step (falling limb of SAR). Consequently, k_cat starts to drop again (Batinic-Haberle et al. 2002, 2010).

We have recently showed that there is a remarkable ~100-fold increase in the k_cat of cationic versus those porphyrins that have either no charges or negative charges on the periphery (Batinic-Haberle et al. 2010). As expected, in the latter case, due to the repulsion of anionic superoxide from anionic porphyrin, the effect was larger than with neutral compounds. The effect of the charges on the k_cat of MnPs is similar in magnitude to the effect established with the enzyme itself (Spasojevic et al. 2003 and refs therein). The high k_cat of the enzyme suggests a diffusion-limited process; the crystal structure showed that the positively charged amino acids provide a tunnel to guide anionic superoxide to the metal site. Recently, with the wealth of data on numerous MnPs we established the individual SARs for cationic, anionic, and neutral porphyrins (Rebouças et al. 2009), accounting for the electrostatic effects. Such SARs show that, for any given E_1/2, the k_cat is the highest for the cationic porphyrins and most so for those compounds which have positive charges closer to the Mn site, i.e. ortho isomeric substituted pyridylporphyrins (Rebouças et al. 2009). Based on SAR, the ortho isomeric Mn(III) tetrakis(N-alkylpyridinium-2-yl)porphyrins (Figs. 2, 3; Table 1) were identified as the most potent compounds. The ethyl analog, MnTE-2-PyP⁵⁺ (Fig. 2), was forwarded to in vitro and in vivo studies.

The reactivity toward other ROS/RNS

In collaboration with Rafael Radi and Gerardo Ferrer-Sueta, the reactivity of ortho Mn(III) N-alkylpyridylporphyrins toward ONOO⁻ was determined to parallel their ability to dismute ${\text{O}}_2^{•-}$, making them the most potent synthetic scavengers of peroxynitrite also (Table 1) (Ferrer-Sueta et al. 2003; Ferrer-Sueta and Radi 2009). Carnieri et al. have reported the high potency of para isomer, MnTM-4-PyP⁵⁺, in reducing hypochlorous acid, HClO, which is produced by myeloperoxidase during inflammation (Halliwell and Gutteridge 2007). Given the fact that HClO is able to oxidize MnP in a manner similar to ONOO⁻ (Carnieri et al. 1982; Ferrer-Sueta et al. 2003), and that ortho isomer of Mn(III) alkylpyridylporphyrins are more efficacious in reducing ONOO⁻ than para analogs, we would expect MnTM-2-PyP⁵⁺ to reduce HClO more readily than MnTM-4-PyP⁵⁺. The reactivity toward HClO just adds to the wide spectrum of the species that MnPs may eliminate, which in turn challenges us with the complexity of reactions and mechanisms of action(s) of this and other synthetic antioxidants. Further, MnPs react readily with cellular reductants: ascorbate, glutathione, uric acid, and tetrahydrobiopterin (Ferrer-Sueta et al. 1999, 2006; Batinic-Haberle et al. 2004). They can be also reduced with cellular flavoenzymes (Ferrer-Sueta et al. 2006; Kachadourian et al. 2004) and thioredoxin reductase (Faulkner et al. 1994) as well as by nitric oxide (Spasojevic et al. 2000; Pfeiffer et al. 1998). Thus, in vivo, in a first step of MnP-based catalysis of ${\text{O}}_2^{•-}$ dismutation, Mn^IIIP would likely be reduced to Mn^IIP with cellular reductants, and not with ${\text{O}}_2^{•-}$. Also the reduction of ONOO⁻ by MnP may be coupled to cellular reductants. In a first step, Mn^IIIP could be reduced to Mn^IIP. In a second step, Mn^IIP would react two-electronically with ONOO⁻, whereby O=Mn^IVP and benign NO₂⁻ would be formed. The highly reactive, high-valent Mn(IV) species, O=Mn^IVP could then be reduced back to a stable Mn +3 oxidation state with either uric acid, or ascorbic acid or glutathione, whereby sparing biological molecules (Ferrer-Sueta et al. 1999).

Design of Mn porphyrins with respect to their bioavailability

For a while the prevailed opinion had been that the excessively charged compounds cannot accumulate to a high extent in important cellular compartments, particularly in mitochondria. Most recently, several groups have designed compounds that target mitochondria due to their critical role in normal cell functioning. Murphy (2008) developed a concept that in addition to a redox-able unit (ubiquinone, nitroxide, vitamin E, ebselen, etc.), the compound must be positively charged and of appropriate lipophilicity in order to reach mitochondria. Based on such concept, we hypothesized that our cationic MnPs would accumulate in mitochondria. For that purpose we developed HPLC/fluorescence and LC-MS/MS methods to determine MnTE-2-PyP⁵⁺ and its analogs in vivo (Spasojevic et al. 2007, 2008). In a study by Ferrer-Sueta et al. (2006), ≥3 μM MnTE-2-PyP⁵⁺ was able to protect submitochondrial particles against ONOO⁻ flux. In a following study we found that MnTE-2-PyP⁵⁺ reaches mouse heart mitochondria at levels high enough (5.1 μM) to protect it against ONOO⁻-mediated damage (Spasojevic et al. 2007). Preliminary data already indicate higher mouse heart mitochondrial accumulation of MnTnHex-2-PyP⁵⁺ as compared to hydrophilic MnTE-2-PyP⁵⁺ (Spasojevic et al., unpublished). We further found that MnTE-2-PyP⁵⁺ accumulates in nucleus of macrophages at threefold higher levels than in cytosol, presumably driven there by the negatively charged phosphates of nucleic acids (Fig. 8b).

Fig. 8.

a The electrophoretic mobility shift assay (EMSA) of NF-κB in the presence of several different anionic and cationic Mn porphyrins, Mn(salen)⁺ (EUK-8) and MnCl₂. The procedure was done as previously described (Tse et al. 2004). In brief, EMSAs with recombinant human NF-κB p50 (Promega) were performed in a cell-free system with the addition of anionic and cationic Mn porphyrins directly in the binding reaction for 15 min prior to the addition of ³²P-labeled NF-κB consensus oligo. The following final concentrations of Mn porphyrins were added to the binding reaction: 13.5 μM Mn(salen)⁺, 11 μm MnTMOE-2-PyP⁵⁺, 21.3 μM MnhematoP⁻, 10 μM MnCl₂, 9.7 μM MnTM-3-PyP⁵⁺, 11.7 μM MnTnHex-2-PyP⁵⁺, 11 μM MnTnBu-2-PyP⁵⁺, 7.6 μM MnT (2,6-Cl₂-3-SO₃P)P³⁻, 11 μM MnTBAP³⁻, and 14.3 μM MnTE-2-PyP⁵⁺. Protein and DNA complexes were separated on a non-denaturing polyacrylamide gel and identified by autoradiography. b Unstimulated and LPS-stimulated macrophages were treated with 34 μM MnTE-2-PyP⁵⁺ for 1.25 h as described elsewhere (Tse et al. 2002). Using HPLC/fluorescence methodology (Spasojevic et al. 2008), the levels of MnP in the cytosol of unstimulated and LPS-stimulated macrophages were 35 and 44 ng/mg protein, respectively, while in the nucleus, the levels of MnP observed were 99 and 156 ng/mg protein, respectively. The study proved that MnTE-2-PyP⁵⁺ accumulates more readily in nucleus than in cytosol during macrophage activation

We have obviously achieved the maximal antioxidant potency of MnP in aqueous systems (k_cat) as discussed above (plateau in Fig. 3). We can further enhance the in vivo effects by increasing the availability of potent Mn(III) N-substituted pyridylporphyrins, primarily by reducing their excessive hydrophilicity. To do so, we recently developed a method to quantify the MnP lipophilicity by its partition between n-octanol and water (P_OW) (Table 1) (Kos et al. 2009b). Such measurements allow us to assess more precisely the MnPs lipophilicity and compare them to other drugs of similar targets. At the present state of our insight into the structure and in vivo actions of MnPs, the lipophilicity of Mn(III) N-alkylpyridylporphyrins can be increased tenfold: (1) by lengthening the meso alkylpyridyl chains by each additional CH₂ group, or (2) by shifting the alkyl groups from ortho onto meta (indicated by number 3 in formulas) pyridyl positions (Fig. 2). The first modification does not affect the k_cat as the charges are still positioned at the same distance with respect to the Mn site. Thus, MnTM-2-PyP⁵⁺ and MnTnOct-2-PyP⁵⁺ are of nearly identical k_cat ( ${\text{O}}_2^{•-}$) and k_red (ONOO⁻) (Table 1). Second modification places positive charges further away from the Mn site, and thus, causes the decrease in SOD-like activity. The meta isomeric compounds are still bulky enough, and thus, do not associate significantly with nucleic acids and are not toxic when compared to para analogs.

In vivo and in vitro effects

The critical role of the MnP bioavailability

Most of the in vitro and in vivo effects have been recently summarized in Batinic-Haberle et al. 2010. Herein, we pointed out in particular those studies which clearly addressed the critical importance of lipophilicity for in vivo efficacy of MnPs.

The effect of the length of alkyl chains on meso pyridyl substituents

The up to 120-fold increased efficacy of MnTnHex-2-PyP⁵⁺ in vivo when compared to MnTE-2-PyP⁵⁺ (Saba et al. 2007; Batinic-Haberle et al. 2009b) (Table 1) is due to its 13,500-fold increased lipophilicity (Kos et al. 2009b) (Table 1) and 12-fold higher ability to cross blood brain barrier (Doyle et al. 2009). In a rat renal ischemia/reper-fusion model, single iv injection of only 50 μg of MnTn-Hex-2-PyP⁵⁺ protected against ATP depletion, MnSOD inactivation, protein 3-nitrotyrosine formation, and renal dysfunction (Saba et al. 2007). In a rat stroke 90-min middle cerebral artery occlusion model, due to the 12-fold higher ability to cross blood brain barrier (Doyle et al. 2009), MnTnHex-2-PyP⁵⁺ decreased stroke volume and neurologic deficit when 450 μg/kg/day was given intravenously for a week, starting at 90 min after reperfusion (Warner et al. 2010) (of note, previous data show that MnP delivery as late as 6 h after reperfusion still reduces infarct size significantly; Mackensen et al. 2001). In a rabbit cerebral palsy model, MnTnHex-2-PyP⁵⁺ rescued rabbit puppies, but MnTE-2-PyP⁵⁺ did not (Yu et al. 2010). Mn porphyrins were given to a rabbit dam twice, 30 min before and 30 min after the 40-min uterus ischemia at 0.1 and 1 mg/kg (MnTnHex-2-PyP⁵⁺) and 6 mg/kg (MnTE-2-PyP⁵⁺). MnTnHex-2-PyP⁵⁺ groups had 14 normal, 2 mild, 3 severely affected P1 kits (32 days of gestation) and 0 stillbirths, compared to saline with 4 normal, 3 mild, 5 severe, and 6 stillbirths from 2 dams/group. The enhanced ability to cross blood brain barrier contributed to the increased efficacy of MnTnHex-2-PyP⁵⁺ to prevent chronic morphine tolerance (anti-inflammatory condition) in a mouse pain model given intraperitoneally at 0.1 mg/kg/day for 4 days as compared to 3 mg/kg of MnTE-2-PyP⁵⁺ (Batinic-Haberle et al. 2009b). In a G93A mouse mutant amyotrophic lateral sclerosis study, three metalloporphyrins were thus far tested, FeTBAP³⁻, MnTDE-2-ImP⁵⁺, and MnTnHex-2-PyP⁵⁺; the lipophilic MnTnHex-2-PyP⁵⁺ was efficacious at 5–10-fold lower dose than MnTDE-2-ImP⁵⁺ (Crow 2005, 2006; Crow et al. 2005). In a rat pulmonary radioprotection study, again, hexyl was more efficacious than ethyl porphyrin. In a radioprotection of ataxia telangiectasia (AT), among six Mn porphyrins, three Mn salens and two Mn cyclic polyamines tested, only a very lipophilic and cationic compound that possesses high antioxidant capacity, MnTnHex-2-PyP⁵⁺, was protective (Pollard et al. 2009). Equally potent, but 13,500 more hydrophilic MnTE-2-PyP⁵⁺ was ineffective. Potent and lipophilic SOD mimic, Mn(II) cyclic polyamine lacks positive charges which likely prevents its mitochondrial accumulation and therefore presumably its efficacy. Thus, the data indicate that mitochondrial accumulation, driven by positive charges and the lipophilicity of the compound, may play a critical role in AT model. Another support for the impact of mitochondrial accumulation on the in vivo efficacy of MnPs was provided from (1) our ongoing Saccharomyces cerevisiae studies where MnTnHex-2-PyP⁵⁺ when compared to MnTE-2-PyP⁵⁺ accumulates in mitochondria at significantly higher levels than in cytosol (Gralla et al., unpublished) as well as from (2) our mouse study where hexyl analog accumulates more in heart mitochondria than does ethyl porphyrin (Spasojevic et al. 2006, unpublished). The mouse ongoing MRI experiments at 7T show that following a single mouse injection, the relaxation changes of MnTnHex-2-PyP⁵⁺ in prostate tumors measure 10–11-fold greater than in normal prostate gland (Lascola et al. 2010).

The effect of the position of alkyl chains on meso pyridyl substituents

In the aerobic growth of SOD-deficient E. coli, the effect of the ortho versus meta position on meso pyridyls was established (Fig. 4); the tenfold increased lipophilicity of the meta isomer, MnTE-3-PyP⁵⁺, with respect to the ortho analog MnTE-2-PyP⁵⁺ (Table 1) leads to its tenfold higher accumulation within E. coli which in turn fully compensates for the tenfold lower k_cat of the meta isomer. Consequently, meta isomer was equally efficient as ortho in rescuing SOD-deficient E. coli when it grew aerobically (Kos et al. 2009a).

Fig. 4.

Lipophilicity of MnPs increases tenfold by either (1) lengthening alkyl chains by each additional carbon atom, or (2) shifting alkyl groups from ortho (2) to meta 5 positions. The tenfold increased lipohilicity of meta ethyl analog, MnTE-3-PyP⁵⁺, resulted in its tenfold higher accumulation in the cytosol of E. coli as compared to ortho isomer, MnTE-2-PyP⁵⁺. Such enhanced accumulation compensated for a tenfold lower ability of MnTE-3-PyP⁵⁺ to dismute ${\text{O}}_2^{•-}$. In turn both isomers were equally able to substitute for the lack of cytosolic Cu,ZnSOD when E. coli grew in aerobic medium (Kos et al. 2009a, b)

Toxicity

The TD₅₀ (the dose at which toxic effects were observed with 50% of mice) of MnTnHex-2-PyP⁵⁺ is 12.5 mg/kg when given subcutaneously (Pollard et al. 2009). Given its significantly increased efficacy, the therapeutic window, i.e. the ratio of toxic (12.5 mg/kg) to effective dose (0.4–0.05 mg/kg), is 31–250 for MnTnHex-2-PyP⁵⁺, while it is 15 for MnTE-2-PyP⁵⁺ (91.5 mg/kg over 6 mg/kg) (Batinic-Haberle et al. 2010).

The impact of MnPs on cellular redox-based transcriptional activity

The decrease in oxidative stress injuries via scavenging ROS/RNS by Mn porphyrins have been shown in numerous studies, some reviewed in Batinic-Haberle et al. (2010). Among them are: (1) superoxide-specific effects exerted in E. coli study where SOD deficient E. coli grew aerobically only if medium was substituted for the cytoplasmic SOD enzymes; (2) radioprotective effects where MnP decreased oxidative damage by scavenging ROS/RNS; (3) ischemia/reperfusion injuries which involve burst in ROS/RNS upon reperfusion, etc. Yet, if a single dose of Mn porphyrins was given only at the moment of primary oxidative stress or prior to injury, the short-term protective effects were observed, but protection was not sustained over a long period of time (Sheng et al. 2009). Another interesting observation was recently made in a radioprotection (Gauter-Fleckenstein et al. 2010) and stroke study (Sheng et al. 2009): if the MnP was absent at the moment of primary oxidative stress but was given hours or weeks after it, the remarkable protective effects could still be achieved. Such observations indicate that Mn porphyrins are not merely scavenging ROS/RNS produced at the moment of initial oxidative insult. Several years ago two independent, diabetes and radiation/cancer, studies were the first to offer an explanation (Moeller et al. 2004; Piganelli et al. 2002; Bottino et al. 2004; Tse et al. 2004). In a diabetes-related study, MnTDE-2-ImP⁵⁺ prevented NF-κB activation (Tse et al. 2004; Bottino et al. 2004) and impacted NF-κB-dependent gene transcription. MnTDE-2-ImP⁵⁺ suppressed pro-inflammatory cytokine production and secondary inflammation as a result of oxidative stress and the ensuing cyclic inflammatory immune responses (Fig. 5). Radiation/cancer-related study on a 4T1 breast mouse carcinoma model (Moeller et al. 2004) assessed the effects of MnTE-2-PyP⁵⁺ on HIF-1α. Soon afterward, additional studies on a murine skin carcinogenesis model demonstrated that MnTE-2-PyP⁵⁺ inhibited AP-1 transcriptional activity (Zhao et al. 2005). Suppression of HIF-1α, VEGF, and TGF-β in rodent pulmonary radioprotection studies and LPS-stimulated macrophages were also shown (Gauter-Fleckenstein et al. 2008, 2010). Below is the summary of the effects observed along with related mechanistic considerations in particular with respect to NF-κB. The data indicate that the inhibition of HIF-1α and AP-1 via antioxidant action likely occurs in the cytosol, while the impact upon NF-κB via pro-oxidative action seems to happen predominantly in nucleus.

Fig. 5.

MnP prevents activation of redox-active cellular transcription factors. In an in vitro 4T1 mouse breast cancer cell study, HIF-1α was activated by H₂O₂ and nitric oxide but blocked with equimolar concentrations of MnTE-2-PyP⁵⁺, indicated in I as SOD (Moeller et al. 2004). In II, the AP-1 activation in a mouse skin cancerogenesis model was inhibited with MnTE-2-PyP⁵⁺ given at 5 ng daily for 5 days a week for 4 weeks (Zhao et al. 2005), while in III, in a rat stroke 90-min middle cerebral artery (MCAO) occlusion model, the NF-κB activation at 6 h post-MCAO was inhibited with imidazolyl analog, MnTDE-2-ImP⁵⁺, given for 1 week intracerebroventricularly (ICV) at 900 ng bolus dose + 56 ng/h ICV infusion for a week starting at 90 min after 90-min MCAO (Sheng et al. 2009)

HIF-1α

HIF-1α is a pro-angiogenic transcription factor that is activated under conditions of intermittent hypoxia and upregulates a number of genes including VEGF, EGF, and TGF-β to drive angiogenesis and fibrinogenesis. Recent studies have indicated how critical NF-κB signaling is to inflammatory gene expression and to the regulation of HIF-1α (Oliver et al. 2009). The very first data published on the inhibitory effect of MnTE-2-PyP⁵⁺ on HIF-1α activation in 4T1 mouse breast cancer cells that were exposed to H₂O₂ and ^•NO (Fig. 5) (Moeller et al. 2004) were subsequently confirmed in a mouse breast cancer study where treatment of mice bearing 4T1 cells led to a significant inhibition of hypoxia, HIF-1α, VEGF, vascular density, and suppressed tumor growth (Rabbani et al. 2009). Importantly, oxidative stress, assessed through measuring 8-OHdG, nitrotyrosine, NADPH oxidase, and macrophages infiltration, was significantly reduced, leading to a conclusion that anti-angiogenic effects presumably arise from the decreased levels of ROS/RNS which in turn prevents activation of HIF-1α and expression of its downstream product, VEGF (Rabbani et al. 2009). Further support also comes from another recent study with N-acetylcysteine presumably suppressing tumor growth through inhibition of HIF-1α activation (Gao et al. 2007).

The effect of MnP on HIF-1α and VEGF was also observed in a radiosensitization study where MnTE-2-PyP⁵⁺ was administered immediately after each of three doses of 5 Gy radiation, separated by 12 h. Tumor vasculature density was decreased by 79% at 48 and 72 h after radiation, and tumor growth significantly suppressed; amifostine produced no effect (Moeller et al. 2005). In a parallel in vitro experiment, tumor cells were grown under hypoxic conditions (0.5% O₂), whereby releasing cytokines into medium (Moeller et al. 2005). If the growing medium, which lacked MnTE-2-PyP⁵⁺, was collected and later used to grow endothelial cells for 2 h before radiation, the endothelial cells were fully viable as compared to the irradiated cells being in medium which was collected from tumor cells growing under hypoxic conditions with MnP added. MnTE-2-PyP⁵⁺ inhibited the release of cytokines (VEGF, EGF, etc.) from hypoxic tumor cells, which would have otherwise protected endothelial cells from irradiation.

In the same radiosensitization study (Moeller et al. 2005), no cytotoxicity of MnP to tumor cells (4T1 and R3232) was observed. Further, four different tumor cell lines, Hela cells, CaCo-2 cells, 4T1, and HCT116, were grown with MnTE-2-PyP⁵⁺ and MnTnHex-2-PyP⁵⁺ (Ye et al. 2009). No toxicity was observed with 10,000 cells/well at 30 μM MnP (Ye et al. 2009). At a higher cellular density of 50,000 cells/well, up to 1 mM both Mn porphyrin solutions were not cytotoxic.

In an in vitro study, alveolar macrophages under hypoxic conditions (0.5% of O₂) upregulated TGF-β and VEGF production, while MnTE-2-PyP⁵⁺ reduced their levels to basal (normoxic) values. Production of superoxide was also greatly enhanced under hypoxic conditions and normalized in the presence of MnTE-2-PyP⁵⁺ indicating the involvement of ROS and their scavengers in modulation of the cytokine expression (Jackson et al. 2007).

In rodent pulmonary radioprotection studies, Mn(III) N-alkylpyridyl- and N,N′-dialkylimidazolylporphyrins were effective when given before, immediately after injury, and even days or weeks after injury (Vujaskovic et al. 2002; Rabbani et al. 2007a, b). The protective effect was at the level of primary radiation injury and/or secondary oxidative events. In the most recent studies, histopathological and functional lung damage was dramatically improved when 6 mg/kg/day MnTE-2-PyP⁵⁺ and 0.05–1 mg/kg/day MnTnHex-2-PyP⁵⁺ were given for 2 weeks started at different time points after the irradiation (Gauter-Fleckenstein et al. 2008, 2010). HIF-1α, VEGF, macrophage accumulation, oxidative stress (8-OHdG), TGF-β, and hypoxia were greatly reduced by the treatment. In a study with MnTE-2-PyP⁵⁺, we have shown for the first time that it is able to reverse the overall rat lung damage even when its 2-week delivery started at the time of established lung injury, 8 weeks after the irradiation (Gauter-Fleckenstein et al. 2010).

AP-1

AP-1 transcription factor regulates gene expression and controls a number of cellular processes such as cell survival, proliferation, and death (Shaulian and Karin 2001). At the same time the inhibitory effect of MnP on HIF-1α was reported, the effects upon AP-1 in skin carcinogenesis were also noted (Zhao et al. 2005). Tumor was induced by the mutagenic initiator 7,12-dimethylbenz(α)-anthracene (DMBA) and promoted by 12-O-tetra-decanoylphorbol-13-acetate (TPA). MnTE-2-PyP⁵⁺ was applied to skin (5 ng/mouse/day) either at 30 min prior to TPA (MnP/TPA) or at 12 h following TPA application (TPA/MnP) 5 days/week for 14 weeks. The suppressive effect was much more significant when MnTE-2-PyP⁵⁺ was given 12 h following TPA treatment as cells undergo apoptosis but prior to proliferation. The decrease in oxidative stress by MnP paralleled AP-1 inhibition (Fig. 5). MnTE-2-PyP⁵⁺ further suppressed PCNA (proliferating cellular nuclear antigen, a well-recognized proliferative marker) and mitosis but not apoptosis, because it was given 12 h after TPA treatment as cells underwent apoptosis. The effect of MnTE-2-PyP⁵⁺ on tumor incidence and multiplicity was much greater with timed administration of SOD mimic than with MnSOD-overexpressed mice (where timely administration is impossible). Near full suppression of papillomas with MnTE-2-PyP⁵⁺ (31 vs. 5 in control group) as compared to 31 versus 17 in MnSOD-overexpressed mice was found. In a latter case, MnSOD affected both cell proliferation and apoptosis which diminished the final outcome (Zhao et al. 2001).

NF-kB

NF-κB family of transcription factors [p65 (RelA), RelB, c-Rel, NF-κB1 (p50 and its precursor p105), and NF-κB2 (p52 and its precursor p100)] controls hundreds of genes involved in regulation of immunity, development, apoptosis, cell proliferation, and inflammation (Gloire and Piette 2009; Oliveira-Marques et al. 2009). Its misregulation in cells is linked with several diseases including cancer, autoimmunity, and chronic inflammation. Therapies that specifically target this signaling pathway may prove to be efficacious in the treatment of various pro-inflammatory-mediated diseases since NF-κB activation is prevalent in arthritis, gastritis, inflammatory bowel disease, and sepsis. In resting cells, NF-κB proteins are sequestered in cytoplasm through their tight association with IκB proteins. The inhibition of NF-κB by MnP was first shown by the Piganelli group with nuclear extracts from LPS-treated bone marrow-derived macrophages (Tse et al. 2004; Bottino et al. 2004). Treatment with MnTE-2-PyP⁵⁺ and MnTDE-2-ImP⁵⁺ prevented the binding of the p50 subunit of the canonical NF-κB moiety to the cognate DNA response element. This inhibition of NF-κB DNA-binding abrogated the initiation of the innate immune response as evidenced by suppression of pro-inflammatory cytokine TNF-α and IL-1β. The authors showed that all other facets of NF-κB pathways were unaffected: IKKα/β phosphorylation, IκB-α phosphorylation/degradation, and NF-κB p50/p60 nuclear translocation. Furthermore, these results were recapitulated in human pancreatic islets under normal conditions and when pancreatic cells were cultured for 30 min with medium containing proteolytic enzymes and byproducts generated by pancreatic cells during isolation (Bottino et al. 2004). In the presence of 34 μM MnTDE-2-ImP⁵⁺, a reduction in NF-κB-dependent gene transcription of cytokines and chemokines, such as IL-6, IL-8, and MCP-1 (macrophage chemoattractant porphyrin-1), was observed in addition to PARP [poly(ADP-ribose)polymerase] suppression.

The reducing environment in the nucleus is critical for NF-κB binding (Toledano et al. 1993; Gloire and Piette 2009) (Fig. 6). There are seven cysteine amino acids in NF-κB p50, and cysteine 62 is in a well-conserved Rel homology domain, characteristic of NF-κB family members and necessary for DNA-binding (Toledano et al. 1993). Point mutations of all seven cysteine amino acids in NF-κB p50 to serine did not abolish DNA-binding, but DNA-binding was prevented when cysteine 62 was oxidized with diamide or N-ethylmaleimide except for the C62S point mutant (Toledano et al. 1993). Therefore, cysteine 62 is the only amino acid that is redox-sensitive in NF-κB p50 (Toledano et al. 1993; Nishi et al. 2002; Matthews et al. 1992). In order to efficiently bind DNA, cysteine 62 of the NF-κB p50 subunit must be reduced. Had MnP exhibited antioxidative activity in the nucleus and further enhanced the reducing environment, NF-κB DNA binding would not have been suppressed. Therefore, the pro-oxidative action of MnPs has been proposed. Several quinones were reported to act similar to MnP in a pro-oxidative fashion (Gloire and Piette 2009). Quinones have been known to undergo one-electron reduction to semiquinone which in turn can be oxidized back to quinone, whereby reducing oxygen to superoxide (alike the ubiquinone in mitochondrial respiration). Several enzymes have been reported to control the redox state of cysteine 62 in NF-κB p50: thioredoxin (TRX) (Matthews et al. 1992) and AP endonuclease 1/redox factor 1 (APE1/Ref-1) (Jeon and Irani 2009), the latter acting either as a redox cofactor or redox chaperone (Fig. 6). Glutatione (GSH) is also transported into nucleus in early phases of cell growth when most of cells are in active division phase to assure reducing conditions (Markovic et al. 2007). GSH may directly act as a transcriptional regulator of NF-κB, AP-1, and p53 by altering nuclear redox state (Jang and Surh 2003). S-nitrosation of cysteines (Matthews et al. 1996) or nitration of tyrosines would also unfavor DNA binding (Fig. 6) (Gloire and Piette 2009). Taken together, our data suggest that MnP either directly oxidizes cysteine 62 of NF-kB p50, or prevents AEP1/Ref-1 and/or thioredoxin to secure the reduction of cysteine 62 and facilitate DNA binding (Gloire and Piette 2009).

Fig. 6.

Redox-regulation of NF-κB DNA-binding. In resting cells, NF-κB is predominantly found in the cytoplasm of the cell, with an oxidized p50 cysteine 62. Upon activation, NF-κB translocates into the nucleus where p50 cysteine 62 is reduced by APE1/Ref-1, thereby allowing NF-κB DNA binding. APE1/Ref-1 can both act as a redox factor by directly reducing p50 or as a redox chaperone by promoting the reduction of p50 by TRX or GSH. NF-κB DNA binding leads to ^•NO synthase (NOS) expression, thereby leading to nitric oxide (^•NO) production. ^•NO can in turn modify p50 cysteine 62 and p65 cysteine 38 by S-nitrosylation, which unbinds NF-κB from the DNA and contributes to the resolution of inflammation. ^•NO can also react with ${\text{O}}_2^{•-}$ to generate ONOO⁻. ONOO⁻ induces tyrosine nitration of p65 on tyrosine 66 and tyrosine 152, thereby leading to its association with IκBα for nuclear export. Copy rights obtained to use Fig. 7 from Gloire and Piette (Antioxid Redox Signal, 11:2209–2222, 2009)

Mn^IIIP can oxidize glutathione (pK_a = 9.2) (Halliwell and Gutteridge 2007) as depicted in Fig. 7. The reaction is pH dependent, and is slow at pH 7.8 where the protonated thiol is a major species. At higher pH (≥9), where there is a significant contribution of reactive thiolate in solution, a reaction is much faster. The average pK_a of protein cysteine thiols is ~8.5. However, many proteins have domains that lower the pK_a significantly, so that under physiological conditions thiolate anion is a reactive species (Hill et al. 2009). Under such conditions (which would resemble the oxidation of GSH at higher pH; Fig. 7, spectrum 3), Mn^IIIP would readily oxidize cysteines in vivo; Mn^IIIP/Mn^IIP redox couple would be employed. Further, under in vivo oxidative stress with increased levels of peroxide and/or peroxynitrite, Mn^IIIP would likely be oxidized to some extent to a strongly oxidizing high-valent species O=Mn^IVP. We reported that O=Mn^IVP rapidly oxidizes GSH (Ferrer-Sueta et al. 1999), and could thus efficiently oxidize cysteine 62; in such case O=Mn^IVP/Mn^IIIP redox couple would be utilized.

Fig. 7.

The pH-dependent reduction of MnTE-2-PyP⁵⁺ with gluta-thione. Spectra were taken in 0.1 M tris buffer with 10 μM MnTE-2-PyP⁵⁺ and 0.7 mM glutathione. The pH values were: 7.63 (1), 6.67 (2), and 8.14 (3). Spectra were taken at 1 min after the reaction had started. The Mn^IIITE-2-PyP⁵⁺ has the absorbance at 454 nm (spectrum 1) (Batinic-Haberle et al. 2002) and the reduced porphyrin Mn^IITE-2-PyP⁴⁺ at 438.4 nm (spectrum 3) (Spasojevic et al. 2000). The appearance of the shoulder at ~415 nm of the metal-free ligand (Batinic-Haberle et al. 2002) in spectrum 3 suggests that upon reduction a loss of Mn occurs to some extent. At higher pH > 9, GSH is predominantly in a form of more reactive thiolate species in solution. The reaction of MnP with GSH at higher pH may resemble the in vivo conditions where protein cysteines have presumably lower pK_a values (Hill et al. 2009)

A study (depicted in Fig. 8a) aimed to distinguish the effect of charges and the redox properties of Mn porphyrins and compounds of similar action on the inhibition of p50 DNA binding. The study was performed in a cell-free system where the likely scenario is the reduction of Mn^IIIP to Mn^IIP with cysteine. In such a system the presence of ROS/RNS to promote the formation of O–Mn^IVP and its consequent reduction to Mn^IIIP with cysteine is unlikely. As noted above, the thiolate is presumably the reactive species of p50 cysteine (Hill et al. 2009). Thus, the positive charges on the porphyrins would favor the approach, and thus, the oxidation of cysteine, while the negatively charged compounds would be repelled from p50. Consequently, the proper electrostatics for the approach of the MnP to p50 protein and a proper reduction potential E_1/2 for Mn^IIIP/Mn^IIP redox couple are critical. The data in Fig. 8a clearly showed that compounds that are anionic or neutral at the periphery (MnTBAP³⁻, MnhematoP⁻, Mn(salen)⁺—EUK-8 and MnCl₂) do not prevent NF-kB p50 DNA binding, while all cationic porphyrins are suppressive (MnTM-3-PyP⁵⁺, MnTnHex-2-PyP⁵⁺, MnTE-2-PyP⁵⁺, MnTnBu-2-PyP⁵⁺, and MnTMOE-2-PyP⁵⁺) (Fig. 8a). Among anionic porphyrins, MnTBAP³⁻ and MnhematoP⁻, with very negative E_1/2 of ≤−200 mV versus NHE, are very redox-inert (only a strong reductant, sodium dithionite can reduce MnTBAP³⁻) (Batinic-Haberle et al., unpublished), while MnT(2,6-Cl₂-3-SO₃P)P³⁻ has ~300 mV more positive potential (+88 mV vs. NHE) which allows for a marginal effect, despite its anionic charges (Table 1).

High positive oxidation potential of MnCl₂ (+850 mV vs. NHE) allows it to act as reductant only, thus preventing it to oxidize SH groups. Among cationic compounds, the meta isomer, MnTM-3-PyP⁵⁺, that is tenfold less potent antioxidant than MnTE-2-PyP⁵⁺ and has E_1/2 inferior to that of MnT(2,6-Cl₂-3-SO₃P)P³⁻ (+52 mV vs. NHE), is still able to fully suppress p50 DNA binding. Such data support the notion that electrostatics plays major role; however, the favorable thermodynamics could outbalance the unfavorable electrostatics to some extent. The commercial MnTBAP³⁻ used in this study had Mn oxo/hydroxo/acetato impurities that possess high SOD-like activity (Rebouças et al. 2008a, b, c). Those impurities are negatively charged at the periphery and did not inhibit NF-κB p50 DNA binding. Further, MnCl₂ and Mn(salen)⁺ that bear positive charge on a metal site only, are stabilized in their respective (+2 and +3) oxidation states. Yet, in vivo, in addition to the interaction of MnP with p50, the accumulation of MnP would play a major role. The anionic compounds may not enter nucleus to any significant degree. If the anionic charges allowed their nuclear localization, in an oxidizing environment, anionic compounds such as MnTBAP³⁻ could be oxidized to a high-valent, oxidizing Mn(IV) species with either peroxide or peroxynitrite (Ferrer-Sueta et al. 2003), which in turn could oxidize cysteine SH groups to some extent despite the unfavorable electrostatics.

Additional experiments were designed to provide evidence that the cationic porphyrin, MnTE-2-PyP⁵⁺, accumulates in the nucleus. Bone marrow-derived macrophages were stimulated with LPS and treated with 34 μM MnTE-2-PyP⁵⁺ for 1.25 h. As expected, the accumulation of cationic MnTE-2-PyP⁵⁺ in nucleus was threefold higher than in the cytosol (Fig. 8b). The cationic compounds appear to be attracted toward the nucleus by the negatively charged phosphate groups of nucleic acids.

Warner et al. have recently shown in a 90-min middle cerebral artery occlusion stroke model that suppression of NF-κB activation mediated by MnTDE-2-ImP⁵⁺ contributes in part to the improvement in neurologic and histologic outcome (including decrease in cortical and subcortical infarct volume) at 8 weeks after reperfusion, if the MnP was given intracerebroventricularly to rats for 1 week starting at 90 min after reperfusion (Sheng et al. 2009). This was demonstrated by NF-κB electrophoretic mobility shift assay performed at 6 h after reperfusion (Fig. 5). In ischemic hemisphere the p50 NF-κB subunit DNA binding was fully suppressed, but was not prevented in contralateral hemisphere. The same effect on NF-κB and neurologic and histologic outcome has recently been observed with MnTnHex-2-PyP⁵⁺ (Warner et al. 2010). However, due to the 13,500-fold increased lipophilicity and 12-fold increased distribution in brain versus plasma when compared to hydrophilic analogs such as MnTE-2-PyP⁵⁺ (Doyle et al. 2009), MnTnHex-2-PyP⁵⁺ provided remarkable protection when given iv; in addition, it has significantly smaller effect on decreasing blood pressure than MnTE-2-PyP⁵⁺ and MnTDE-2-ImP⁵⁺ (Warner et al. 2010).

It is still possible that interactions of porphyrins with p50 subunit other than pro-oxidative and possibly electrostatic in nature (based on porphyrin cationic character) are operative. Cationic MnP could compete with p50 binding to the negatively charged DNA binding sites. As already noted, it is also possible that MnP could oxidize cysteines of Ref-1 or glutathione, rather than of p50 subunit. The use of redox-inactive metal-free ligand or Zn porphyrin to distinguish between redox and electrostatic effects has thus far proved difficult as those compounds are potent photosensitizers. The use of p50 C62S mutant (Matthews et al. 1992) that binds to DNA while not possessing a redox-sensitive cysteine p62 would perhaps provide additional evidences for the nature of events occurring in the nucleus.

Based on published reports, HIF-1α seems to be under the control of NF-κB via cyclooxygenase/prostaglandin (Bonello et al. 2007; Sun et al. 2007) and canonical signaling pathways [activated through the IκB kinase (IKK) complex consisting of IKKα, IKKβ, IKKγ, which usually results in nuclear localization of p65/p50 dimers] (Oliver et al. 2009). Thus, the effects of Mn porphyrins on HIF-1α may be at least in part influenced by the impact of MnP on NF-κB. Further studies would aim to distinguish between NF-κB and HIF-1α-based effects.

SP-1

This transcription factor plays multiple and sophisticated roles as a modulator of tissue-specific transcription. Activity of SP-1 shows complex regulation by glycosylation, phosphorylation, oxidative stress, proteolytic cleavage, and other transcription factors including NF-κB (Fountain et al. 2007). In a manner similar to NF-κB, MnTDE-2-ImP⁵⁺ was shown to inhibit SP-1 DNA binding (Tse et al. 2004). The ability of MnTM-2-PyP⁵⁺ to prevent the induction of the Na⁺/H⁺ exchangers NHE-1 and NHE-3 and to reduce the activity of the Na⁺/K⁺-ATPase in a rat diabetes model is presumably due to the interference of MnP with SP-1 activation (Khan et al. 2009). NHE together with ubiquitously expressed Na⁺/K⁺-ATPase maintains Na⁺ and fluid homeostasis, the latter being also regulated by NF-κB and AP-1 transcription factors.

In summary, the inhibition of NF-κB activation in activated macrophages (Tse et al. 2004) was the very first example demonstrating that MnP could exert beneficial effects on cellular transcriptional activity via pro-oxidative action. As shown above, ROS/RNS activate transcription factors and their removal (presumably in cytosol) fully suppressed their activation (Fig. 5). Modulation of cellular and subcellular redox status by MnP compounds can affect gene transcription by either reducing or increasing levels of oxidants as dictated by the charge and subcellular compartmentalization of MnP compounds. Further studies are needed to elucidate if the pro-oxidative action of MnPs involved in suppressing NF-κB activation in nucleus may also be operative for other redox-sensitive transcription factors.

Cancer studies

Recently, it was shown that at least in part, MnTE-2-PyP⁵⁺ sensitized murine thymic lymphoma cells, WEHI7.2, to glucocorticoid and cyclophosphamide treatment, suppressing their growth via pro-oxidative action, presumably via H₂O₂ production (Jaramillo et al. 2009). Similarly, the Li et al. (2000) results (reviewed in Holley et al. 2010) on human glioma cells showed that GPX overexpression not only reversed the tumor cell growth inhibition caused by MnSOD overexpression, but also altered the cellular contents of total glutathione, reduced glutathione, oxidized glutathione, and intracellular reactive oxygen species. The results suggest that hydrogen peroxide or other hydroper-oxides appear to be key reactants in the tumor suppression by MnSOD overexpression and that growth inhibition correlates with the intracellular redox status.

As already noted and reported, Mn porphyrin with E_1/2 ~ +200 mV versus NHE can easily donate and accept electrons from redox able compounds, such as are cellular reductants. As stated above, cellular reductants readily reduce porphyrin to Mn^IIP (Batinic-Haberle et al. 2004), which in turn can react both with superoxide and oxygen. Though the reaction with ${\text{O}}_2^{•-}$ is much faster than with oxygen, the higher levels of oxygen may favor reaction of Mn^IIP with O₂. Binding of oxygen to Mn center would lead to cyt P450-like chemistry where highly oxidizing species H₂O₂ and O=Mn^IVP would be formed (Spasojevic et al. 2006). Also, were the cellular reductants depleted and levels of H₂O₂ increased, the O=Mn^IVP would be formed (from Mn^IIIP), and if not reduced back to Mn^IIIP with cellular reductants, it would oxidize biological targets such as cysteines of signaling proteins (Ferrer-Sueta et al. 1999). We have performed the in vitro studies on four cancer cell lines in the presence of MnP and ascorbate alone and in their combination. When combined, H₂O₂ was produced. Tumor cell survival data from the clonogenic assay agree well with the tetrazolium salt WST-1 data (the latter assay assesses the metabolically active cells). The combination of ascorbate and MnP is highly cytotoxic to tumor cells; the cells were not able to establish colonies anymore (Ye et al. 2009).

Concluding remarks

When compared to other types of synthetic antioxidants, Mn porphyrins possess: (1) highest k_cat ( ${\text{O}}_2^{•-}$) and k_red (ONOO⁻); (2) extreme metal/ligand stability to assure the integrity of the redox active Mn site, and (3) countless possibilities of derivatization to optimize the potency, bioavailability, and the toxicity. Though such properties make them advantageous over other types of synthetic antioxidants for treating diseases that have oxidative stress in common, comparative studies are still needed to fully assess the impact of different compounds in the same and different models of oxidative stress disorders.

Given the wealth of data, the present prevailing understanding is that potent catalytic antioxidants Mn(III) N-alkylpyridylporphyrins scavenge ROS/RNS ( ${\text{O}}_2^{•-}$, ONOO⁻, ${\text{CO}}_3^{•-}$, HClO and others) and decrease primary oxidative stress to biological molecules. By doing this, they eliminate the signal for upregulation of inflammatory-mediated pathways via suppression of redox-sensitive transcription factors (Fig. 9). The site of action is yet to be fully recognized, but is likely the cytosol.

Fig. 9.

Mn porphyrins protect biomolecules by scavenging a variety of ROS/RNS, whereby eliminating a signal for transcription factors activation; such actions occur likely in cytosol. In addition to this antioxidant action, the data by Piganelli group strongly supports the notion that Mn porphyrins in nucleus act as pro-oxidants preventing p50 DNA binding. The drawing was inspired by McCord (2000) publication

Yet, the NF-κB data indicate that at least in part MnPs may provide beneficial effects through pro-oxidative action in the nucleus by oxidizing redox-sensitive transcription factors and thus preventing their DNA binding. Consequently, the excessive upregulation of inflammatory and immune responses is suppressed. Lymphoma cancer (Jaramillo et al. 2009) and our 4T1 mouse breast cancer studies (Ye et al. 2009) also demonstrate that at least in part the pro-oxidative action of MnPs may contribute to antitumorigenesis.

Finally, not to detail here, as this is the subject of other manuscript within the same issue (Benov et al. 2010), photodynamic therapy with redox-inactive, metal-free ligands and Zn porphyrins utilizes pro-oxidative action of porphyrins to inactivate critical proteins which in turn lead to cell killing (Al-Mutairi et al. 2007; Benov et al. 2010).

In addition to the above-mentioned actions of MnPs, Tauskela and Brunette (2009) have repeatedly suggested that action(s) other than antioxidative may be involved in neuroprotection, such as impact of porphyrins on Ca²⁺ metabolism. Their data on redox-active and redox-inactive compounds are not fully consistent (Tauskela et al. 2005), suggesting that the compounds might not have been sufficiently characterized with respect to identity and purity to fully support their conclusions. For example, the possibility that porphyrins interact with proteins through electrostatic and/or hydrophobic interactions and not via pro-oxidative action may not be fully excluded. Multiple actions of a single MnP are highly likely in vivo. Further studies would hopefully increase our insight into Mn porphyrins action/s in vivo.

It is important to point out that either for mechanistic studies or for development of MnPs as therapeutics their highest possible purity is of key importance. In several occasions thus far, the preparations that contained as little as ~20% of the MnTE-2-PyP⁵⁺ were used (Munroe et al. 2007; Warner et al., unpublished; Rebouças et al. 2008c). Also routinely, commercial preparations of MnTBAP³⁻ were used in published studies. We have shown that such MnTBAP³⁻ preparations not only differ among commercial sources, but also differ from batch to batch within one commercial source and may contain high levels of “free” Mn and photosensitizer, metal-free porphyrin (Rebouças et al. 2008a, b, c; Batinic-Haberle et al. 2009a, 2010).

All the data thus far published on MnPs are intriguing. Given the remarkable potential of metalloporphyrins as therapeutics, the elucidation of their role(s) in vivo is important and is of interest not only for clinicians, but for mechanistic studies as well.

Abbreviations

ROS

Reactive oxygen species

RNS

Reactive nitrogen species

Peroxynitrite

ONOO⁻ + H⁺ → ONOOH (pK_a= 6.6, at pH 7.8 peroxynitrite exists predominantly as ONOO⁻)

${\text{O}}_2^{•-}$

Superoxide anion radical

^•NO

Nitric oxide

${\text{CO}}_3^{•-}$

Carbonate anion radical

MnP

Mn porphyrin in +3 (Mn^IIIP) and +2 (Mn^IIP) oxidation states; when porphyrins bear 5+ charge; they have Mn in +3 oxidation state; with 4+total charge, Mn is in +2 oxidation state, except in O=Mn^IVP⁴⁺ where it is in +4 oxidation state

5, 10, 15, and 20

meso positions of the porphyrin ring (Fig. 2)

MnT-2(or 3 or 4)-PyP⁺

Mn(III) meso-tetrakis[2(or 3 or 4)-pyridyl]porphyrin; 2, 3 and 4 relate to ortho, meta and para isomers, respectively (Fig. 2)

MnTPP⁺

Mn(III) meso-tetraphenylporphyrin

MnTalkyl-2 (or 3 or 4)-PyP⁵⁺

Mn(III) meso-tetrakis(N-alkylpyri-dinium-2(or 3 or 4)-yl)porphyrin, alkyl being methyl (M in ortho, 2 position, AEOL10112), ethyl (E in ortho, 2 position, AEOL10113, FBC-007), n-propyl (nPr), n-butyl (nBu), n-hexyl (nHex), n-heptyl (nHep), n-octyl (nOct)

MnBM-2-PyP⁵⁺

5,10-bis(2-pyridyl)-15,20-bis (N-methylpyridinium-2-yl)porphyrin

MnTrM-2-PyP⁵⁺

5-(2-pyridyl)-10,15,20-tris (N-methylpyridinium-2-yl)porphyrin

MnT(TFTMA)P⁵⁺

5,10,15,20-tetrakis (2,3,5,6-tetrafluoro-N,N,N-trimethylanilinium-4-yl)porphyrin

MnBr₈TM-3-PyP⁴⁺

Mn(II) β-octabromo-meso-tetrakis (N-methylpyridinium-3-yl) porphyrin

MnBr₈TM-4-PyP⁴⁺

Mn(II) β-octabromo-meso-tetrakis (N-methylpyridinium-4-yl) porphyrin

MnCl_1-4TE-2-PyP⁵⁺

Mn(III) β-(mono-tetra) chloro-meso-tetrakis (N-ethylpyridinium-2-yl)porphyrin

MnCl₅TE-2-PyP⁴⁺

Mn(II) β-pentachloro-meso-tetrakis (N-ethylpyridinium-2-yl)porphyrin

MnTSPP³⁻

Mn(III) meso-tetrakis (4-sulfonatophenyl)porphyrin

MnTBAP³⁻

Mn(III) meso-tetrakis (4-carboxylatophenyl)porphyrin (also abbreviated as MnTCPP³⁻)

MnBr₈TSPP³⁻

Mn(III) β-octabromo-meso-tetrakis (4-sulfonatophenyl)porphyrin

MnT(2,6-Cl₂-3-SO₃P)P³⁻

Mn(III) meso-tetrakis (2,6-dichloro-3-sulfonatophenyl) porphyrin

MnhematoP⁻

Mn(III) hematoporphyrin IX

FeTBAP³⁻

Fe(III) meso-tetrakis (4-carboxylatophenyl)porphyrin (also abbreviated as FeTCPP³⁻)

MnTDE(M)-2-ImP⁵⁺

Mn(III) meso-tetrakis[N,N′-diethyl (or dimethyl)imidazolium-2-yl) porphyrin; diethyl analog is AEOL10150

MnTMOE-2-PyP⁵⁺

Mn(III) meso-tetrakis (N-methoxyethylpyridinium-2-yl)porphyrin

MnTDMOE-2-ImP⁵⁺

Mn(III) meso-tetrakis[N,N′-di (2-methoxyethyl) imidazolium-2-yl]porphyrin

MnTM,MOE-2-ImP⁵⁺

Mn(III) tetrakis [N-methyl-N′-methoxethyl) imidazolium-2-yl]porphyrin

Salen

N,N′-bis-(salicylideneamino) ethane

Mn(salen)⁺

Known as EUK-8

P_OW

Partition coefficient between n-octanol and water

TLC

Thin-layer chromatography

R_f

Thin-layer chromatographic retention factor that presents the ratio between the compound and solvent path in 10:10:80 = sat. KNO₃(aq): H₂O:acetonitrile

E_1/2

Half-wave reduction potential

SOD

Superoxide dismutase

NHE

Normal hydrogen electrode

HIF-1α

Hypoxia inducible factor-1α

NF-κB

Nuclear factor κB

AP-1

Activator protein-1

SP-1

Specificity protein-1

TF

Transcription factor

TPA

12-O-tetradecanoylphorbol-13-acetate

MCAO

Middle cerebral artery occlusion

NOS

Nitric oxide synthase

TRX

Thioredoxin

GSH

Glutathione

p⁶⁵ and p⁵⁰

NF-κB subunits