Anticancer therapeutic potential of Mn porphyrin/ascorbate system

Abstract

Ascorbate (Asc) as a single agent suppressed growth of several tumor cell lines in a mouse model. It has been tested in a Phase I Clinical Trial on pancreatic cancer patients where it exhibited no toxicity to normal tissue yet was of only marginal efficacy. The mechanism of its anticancer effect was attributed to the production of tumoricidal hydrogen peroxide (H₂O₂) during ascorbate oxidation catalyzed by endogenous metalloproteins. The amount of H₂O₂ could be maximized with exogenous catalyst that has optimized properties for such function and is localized within tumor. Herein we studied 14 Mn porphyrins (MnPs) which differ vastly with regards to their redox properties, charge, size/bulkiness and lipophilicity. Such properties affect the in vitro and in vivo ability of MnPs (i) to catalyze ascorbate oxidation resulting in the production of H₂O₂; (ii) to subsequently employ H₂O₂ in the catalysis of signaling proteins oxidations affecting cellular survival pathways; and (iii) to accumulate at site(s) of interest. The metal-centered reduction potential of MnPs studied, E_1/2 of Mn^IIIP/Mn^IIP redox couple, ranged from −200 to +350 mV vs NHE. Anionic and cationic, hydrophilic and lipophilic as well as short- and long-chained and bulky compounds were explored. Their ability to catalyze ascorbate oxidation, and in turn cytotoxic H₂O₂ production, was explored via spectrophotometric and electrochemical means. Bell-shape structure-activity relationship (SAR) was found between the initial rate for the catalysis of ascorbate oxidation, v_o(Asc)_ox and E_1/2, identifying cationic Mn(III) N-substituted pyridylporphyrins with E_1/2 > 0 mV vs NHE as efficient catalysts for ascorbate oxidation. The anticancer potential of MnPs/Asc system was subsequently tested in cellular (human MCF-7, MDA-MB-231 and mouse 4T1) and animal models of breast cancer. At the concentrations where ascorbate (1 mM) and MnPs (1 or 5 μM) alone did not trigger any alteration in cell viability, combined treatment suppressed cell viability up to 95%. No toxicity was observed with normal human breast epithelial HBL100 cells. Bell-shape relationship, essentially identical to v_o(Asc)_ox vs E_1/2, was also demonstrated between MnP/Asc-controlled cellular cytotoxicity and E_1/2-controlled v_o(Asc)_ox. Magnetic resonance imaging studies were conducted to explore the impact of ascorbate on T1-relaxivity. The impact of MnP/Asc on intracellular thiols and on GSH/GSSG and Cys/CySS ratios in 4T1 cells was assessed and cellular reduction potentials were calculated. The data indicate a significant increase in cellular oxidative stress induced by MnP/Asc. Based on v_o(Asc)_ox vs E_1/2 relationships and cellular cytotoxicity, MnTE-2-PyP⁵⁺ was identified as the best catalyst among MnPs studied. Asc and MnTE-2-PyP⁵⁺ were thus tested in a 4T1 mammary mouse flank tumor model. The combination of ascorbate (4 g/kg) and MnTE-2-PyP⁵⁺ (0.2 mg/kg) showed significant suppression of tumor growth relative to either MnTE-2-PyP⁵⁺ or ascorbate alone. In addition to optimal v_o(Asc)_ox, the compound must be also bioavailable at the site of interest. About 7-fold higher accumulation of MnTE-2-PyP⁵⁺ in tumor vs normal tissue was found to contribute largely to the anticancer effect.

Graphical Abstract

INTRODUCTION

Cationic Mn(III) N-substituted pyridyl-porphyrins (MnPs) bear large therapeutic potential as tumor radio- and chemosensitizers, while radioprotectors of normal tissue [1-7]. The anticancer and radioprotective effects, demonstrated in numerous in vitro and in vivo studies, have been reported and the related mechanism(s) discussed elsewhere [1-9]. A lead compound, Mn(III) meso-tetrakis(N-butoxyethylpyridinium-2-yl)porphyrin, MnTnBuOE-2-PyP⁵⁺ (BMX-001) is entering Phase I/II Clinical Trials at Duke University as a radioprotector of normal tissue in head and neck, and glioma cancer patients. Differential effects in tumor vs normal tissue arise predominantly from their differential redox environments [10, 11]. The levels of reactive species (RS) are present at much higher levels in tumor than in normal tissue [12-14]. Differential levels of RS are at least in part due to the differential expression and/or activities of endogenous antioxidant defenses such as superoxide dismutases (SOD), glutathione peroxidases (GPx), catalase and peroxiredoxins [8, 9, 15-31]. Substantial evidence was provided that MnPs are involved in the production of H₂O₂; moreover MnP uses H₂O₂ along with GSH in the catalysis of S-glutathionylation of signaling proteins in either thiol-oxidase or GPx-like fashion [10, 11, 32, 33]. Subsequently, the activities of those signaling proteins get suppressed; the magnitude of the effect controls cell survival. The impact of MnPs on master transcription factor, NF-κB has most frequently been assessed [8-11, 21, 34-37].

Similar to radiation and chemotherapy, metal-catalyzed ascorbate oxidation is another means to inhibit tumor growth via excessive H₂O₂ production. Thus exogenous, pharmacologic ascorbate was tested as a single agent in a Phase I Clinical Trial on pancreatic cancer patients [38]. While no safety concerns were raised, only marginal anticancer effect was demonstrated [38]. The ascorbate oxidation was catalyzed by endogenous metalloproteins, presumably metal-containing porphyrins such as cyt P450 family of enzymes [39-41]. We have already indicated that cationic MnPs may be better catalysts for ascorbate oxidation than endogenous metalloproteins [33, 42-45]. Others subsequently confirmed that MnP-driven catalysis of ascorbate oxidation bears anticancer therapeutic potential [39, 46-49]. Our studies have thus far aimed at identifying those properties of MnPs which are critical for their in vivo anticancer therapeutic potential: (i) ability to catalyze ascorbate oxidation; (ii) ability to employ H₂O₂ in a subsequent catalysis of oxidation or S-glutathionylation of thiols, and (iii) bioavailability in the close vicinity or at the site of interest. Understanding the mechanism behind the effects induced by MnP/Asc system as a source of H₂O₂, will be relevant to coupling of MnP with any source of H₂O₂, including radiation and chemotherapy. It has to be pointed out that ascorbate is our major endogenous reductant. Thus, even if it is not added exogenously, MnP would couple with endogenous ascorbate, yet with a lower yield of H₂O₂ production. Different aspects of the role of ascorbate in MnP biology have been reported and summarized elsewhere [8, 9]. Herein we studied the impact of various MnPs on their ability to catalyze ascorbate oxidation leading to H₂O₂ production (Figure 1). Comprehensive aqueous chemistry and cellular and mouse studies were conducted. In a subsequent study, we have addressed the Aim (ii), i.e. the ability of different redox-active drugs, including MnPs, to oxidize protein thiols in a GPx-like fashion [32].

Figure 1. Structures of the Mn(III) porphyrins whose ability to catalyze ascorbate oxidation was assessed herein.

Except MnTBAP³⁻ and MnTE-2-PyPhP⁵⁺, all MnPs studied are SOD mimics of various potency with log k_cat(O₂^.−) ranging from 6.53 to 7.92. Their SOD-like activity is controlled by the cationic ortho, meta and para (2, 3 or 4, respectfully) electron-withdrawing pyridyl (Py) nitrogens in the vicinity of the metal site. This in turns controls thermodynamics and kinetics (electrostatics, sterics, electronics) of not only catalysis of O₂^.- dismutation, but of reactions of MnPs with other reactive species such as ONOO⁻, H₂O₂, RS⁻ and cellular reductants, GSH, ascorbate and tetrahydrobiopterin [8, 9].

EXPERIMENTAL

Chemicals

Porphyrin ligands, H₂T-2-PyP, H₂T-3-PyP and H₂T-4-PyP, were purchased from Frontier Scientific. Reagents needed for synthesis and purification of MnPs were purchased from the sources detailed in [42, 50]. Plastic-backed silica gel TLC plates (Z122777-25EA) and (+)-sodium L-ascorbate (>98 %) were from Sigma-Aldrich. Acetonitrile (CH₃CN) and KNO₃ were purchased from Mallinckrodt. All chemicals were used as received without prior purification. All chemicals used to measure cellular thiol/disulfide redox states including reduced and oxidized glutathione, cysteine, cysteine, dansylchloride, iodoacetic acid, acetic acid, perchloric acid and methanol were purchased from Sigma-Aldrich.

Synthesis of Mn porphyrins

MnPs were synthesized and purified according to the procedures described elsewhere [42, 50-52].

Catalysis of ascorbate oxidation by MnPs

1) Spectrophotometric measurements of the initial rates of MnP-catalyzed ascorbate oxidation, v_o(Asc)_ox. Under our experimental conditions (pH 7.8), monodeprotonated form of ascorbic acid, HA⁻ (Asc) is a major species in aqueous solution of sodium ascorbate. In the reaction with Mn^IIIP ascorbate is oxidized one-electronically to ascorbyl radical, HA^• which is readily deprotonated to A^•. For details on ascorbic acid speciation see [8, 9, 33]. The v_o(Asc)_ox were determined with 5 μM MnP and 0.15 mM sodium ascorbate under aerobic conditions at (25 ± 1)°C and at pH 7.8 (maintained with either 0.05 M Tris or potassium phosphate buffer). The buffer was initially treated with Chelex-100 ion-exchange resin (200-400 mesh sodium form, Bio-Rad Life Science) to remove adventitious transition metals present in solution. Ascorbate oxidation was followed at 265 nm on Shimadzu UV-2550 uv/vis spectrophotometer. The molar absorptivity of ascorbate was reevaluated to be ε₂₆₅ = 14,000 M⁻¹ cm⁻¹. The initial rates, v₀(Asc)_ox, (nMs⁻¹), were calculated based on the linear kinetic traces obtained for the first 60 seconds. The background rate for non-catalyzed ascorbate oxidation was subtracted from the rates for catalyzed reaction [33]. 2) Electrochemical measurements of initial rates of oxygen consumption, v₀(O₂)_red. Oxygen was consumed via reduction, thus initial oxygen consumption rates were described throughout text as v₀(O₂)_red. The experiments were carried out with Clark electrode at (25±1)⁰C in either 0.05 M Tris or phosphate buffer, pH 7.8. Before measurement, the solutions were purged with air (~21% oxygen). The O_2⁻ sensitive Clark electrode (0.1 M KCl as filling solution) connected to a potentiostat was used. Potential of −0.8 V vs Ag/AgCl was applied to the electrode and once the initial current was stabilized (I_max, corresponding to the [O₂] = 0.255 mM [53] in air-saturated solution) ascorbate solution was added at 1 mM concentration, followed by the addition of 10 μM MnP. The decrease in current, I_obs, corresponding to the decrease in O₂ concentration, Δ[O₂], was followed for at least 300 s. The solution was then purged with N₂ until the current was again stabilized (I_min, corresponding to the [O₂] ≈ 0 mM). This allows for the calculation of O₂ concentration in the solution (in nM) as [O₂]_obs = [(I_obs – I_min) × 2.55×10⁵] / (I_max – I_min), where I_obs is the current at any given time point. The initial reaction rate, v₀ = − Δ [O₂]_obs / Δ t = −[(I_obs2 – I_obs1) × 2.55×10⁵] / [(I_max – I_min) × (t_obs2 – t_obs1)], where I_obs2 and I_obs1 are currents at t_obs2 and t_obs1 time points related to a linear 60 s interval after the addition of MnP. The initial rates of ascorbate oxidation and oxygen consumption rates are given in nM s⁻¹ (Table 1 and Figures 3 and 4).

Table 1.

Spectrophotometrical and electrochemical characterization of MnP/Asc system.

MnPs	E1/2, mV vs NHE	R f	v0(Asc)ox, nM s−1		v0(O2)red, nM s−1
			Phosphate buffer	Tris buffer	Phosphate buffer	Tris buffer
MnTBAP3−	−194	-	3.6±0.3	2.0±1.6	4.1±4.7	2.3±2.0
MnTE-2-PyPhP5+	−65	0.28	49.6±8.7	70.8±7.5	15.2±2.8	11.8±1.4
MnTE-3-PyP5+	+54	0.08	753.9±90.3	972.2±32.0	201.4±21.2	203.9±19.8
MnTnHex-3-PyP5+	+66	0.50	639.2±7.2	864.5±41.1	193.1±40.2	191.7±10.8
MnTE-4-PyP5+	+70	0.07	696.0±52.2	887.0±31.4	185.6±27.9	204.5±31.1
MnTnOct-3-PyP5+	+74	0.55	838.0±7.1	993.6±10.0	225.4±44.5	216.5±20.0
MnTnHex-4-PyP5+	+80	0.54	670.5±2.2	887.5±50.8	199.6±25.8	206.7±29.4
MnTE-2-PyP5+	+228	0.06	679.5±28.0	922.0±89.6	285.7±37.1	307.7±35.1
MnTPhE-2-PyP5+	+259	0.37	571.1±36.2	575.6±83.6	164.5±21.0	176.3±16.4
MnTnBuOE-2-PyP5+	+277	0.43	474.1±26.3	856.9±21.0	159.8±28.9	256.5±45.7
MnTnPen-2-PyP5+	+278	0.32	414.1±32.0	532.9±36.3	152.5±24.0	164.5±21.3
MnTnHexOE-2-PyP5+	+313	0.50	226.9±5.0	469.7±78.7	81.3±14.3	135.6±10.2
MnTnHex-2-PyP5+	+314	0.46	310.7±22.9	358.0±14.9	102.7±23.8	95.0±11.3
MnTnOct-2-PyP5+	+367	0.49	306.3±65.3	304.4±36.8	74.5±15.5	72.1±6.9

Studies were carried out in 0.05 M Tris and potassium buffer, pH 7.8 at (25±1)°C. The metal-centered reduction potential for Mn^IIIP/Mn^IIP redox couple, E_1/2, the chromatographic parameter, R_f (compound path over solvent path on silica plates) as a measure of MnPs lipophilicity, initial rates of ascorbate oxidation (v₀(Asc)_ox and initial rates of oxygen reduction, (v₀(O₂)_red.

Figure 3. Ability of MnPs to catalyze ascorbate oxidation as assessed spectrophotometrically via determination of initial rates of ascorbate oxidation, v₀ (Asc)_ox.

(A) The v₀(Asc)_ox was measured for various MnPs which differ vastly with respect to their redox properties, lipophilicity/bioavailability, polarity, size and bulkiness. Initial rates were determined with 5 μM MnP, 5 mM EDTA and 0.15 mM sodium ascorbate under aerobic conditions at (25 ± 1)⁰C and at pH 7.8 maintained with either 0.05 M potassium phosphate or Tris buffer. (B) Relationship between v₀(Asc)_ox and E_1/2 is of bell-shape. (C) Six cationic MnPs with the E_1/2 ranging from −65 to +340 mV vs NHE are plotted: (#1) MnTE-2-PyPhP⁵⁺, (#2) MnTE-3-PyP⁵⁺, (#3) MnTE-2-PyP⁵⁺, (#4) MnTPhE-2-PyP⁵⁺, (#5), MnTnHexOE-2-PyP⁵⁺, and (#6) MnTnOct-2-PyP⁵⁺.

Figure 4. Ability of MnP to catalyze ascorbate oxidation as assessed electrochemically via determination of initial rates of oxygen reduction.

(A) MnP/Asc cycled under aerobic conditions, whereby oxygen is consumed through reduction to superoxide (Figure 2). The v₀(O₂)_red was measured with 10 μM MnP and 1 mM Asc under aerobic conditions ([O₂] = 0.255 mM) at (25 ± 1)⁰C and at pH 7.8 maintained with either 0.05 M phosphate or Tris buffer. (B) Relationship between v₀(O₂)_red and E_1/2 is of bell-shape. (C) Six cationic MnPs with E_1/2 ranging from −65 to +340 mV vs NHE are plotted: (#1) MnTE-2-PyPhP⁵⁺, (#2) MnTE-3-PyP⁵⁺, (#3) MnTE-2-PyP⁵⁺, (#4) MnTPhE-2-PyP⁵⁺, (#5), MnTnHexOE-2-PyP⁵⁺, and (#6) MnTnOct-2-PyP⁵⁺.

Stability of MnPs towards oxidative degradation with H₂O₂ produced during MnP/ascorbate cycling

Stability of several MnPs (5 μM) used in cellular studies was determined in the presence of ascorbate (1 mM), and at pH 7.8 maintained by Tris buffer on UV-2550 PC Shimadzu spectrophotometer. The stability of other MnPs has been previously reported and discussed [42, 54, 55].

Lipophilicity determination

The TLC chromatographic retention factor, R_f (compound path/solvent path ratio) was determined for MnPs on silica gel plastic plates using 1:1:8=KNO₃(sat) H₂O:H₂O:CH₃CN as a solvent as earlier reported [55].

MRI studies

Measurements of T1 relaxivities were performed on 7.0 Tesla on a Bruker Biospec horizontal bore scanner with 0.1 mM MnP and 5 mM ascorbate in 0.05 M potassium phosphate buffer at (25±1)⁰C [56]. T1-weighted images were processed using Paravision 4.0 software. The T1 relaxivities were calculated and expressed in mM⁻¹ s⁻¹.

In vitro studies

Cell lines

Different mouse and human cell lines were used to explore the cytotoxicity of MnP/ascorbate system. 4T1 mouse mammary breast cancer cell line was purchased from Duke Cell Culture Facility. Human breast cancer cell lines MCF-7 and MDA-MB-231 as well as non-cancerous human breast epithelial cell line HBL100 were generous gift from Professor Christopher Ford and Professor Yunus Luqmani, Kuwait University. Monolayer cultures were grown at 37°C and 5% CO₂ in RPMI 1640 medium supplemented with 10% fetal bovine serum, 1% L-glutamine, 1% penicillin/streptomycin as an antibacterial agent and 0.1% amphotericin as an antifungal agent. For this experiment cultures were used at 70 – 90% confluency. Cell counting was performed with an improved Neubauer hemocytometer and trypan blue to differentiate between the viable and non-viable cells.

Cytotoxicity assay

Cytotoxicity of MnP/Asc system was assessed with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl-tetrazolium bromide (MTT) assay as previously described [57]. Cells were seeded into a flat-bottom 96-well microplate at a concentration of 5 × 10⁴ cells/well and incubated overnight to adhere. Tested compounds were added at indicated concentrations in quadruplicate wells. Cells were incubated for 24 hours with either MnP or ascorbate or MnP/Asc. Control wells without MnP and ascorbate were incubated under same conditions. After incubation, the wells were washed with PBS, fresh medium was added followed by the addition of MTT reagent (10 μL per well). MTT reagent was prepared by dissolving 5 mg MTT in 1 mL PBS. Cultures were incubated for three hours at 37°C and then 10% SDS in 0.01 M HCl was added followed by overnight incubation. Absorbance was measured at 560 nm (formazan product) and 650 nm (background) using a microplate reader (Tecan Sunrise, Tecan Instruments). Results are presented as mean ± S.E. Differences were considered significant when p < 0.05 was achieved.

Cell proliferation assay

Effect of MnTE-2-PyP⁵⁺/Asc on cell proliferation was investigated using the sulforhodamine B (SRB) assay on mouse breast cancer 4T1 and human breast cancer MDA-MB-231 cells. In brief, 15,000 cells/well were plated in quadruplicates onto a 96-well plate and left overnight to adhere. Five μM MnP was added concurrently with 1 mM ascorbate onto plate, and plates were kept in CO₂ incubator. Cell number was determined at 0, 24, 48, and 72 hours after the addition of MnP/Asc. Cells were fixed with ice cold trichloroacetic acid (TCA) at a final concentration of 10% and plates were kept in a refrigerator for an hour. Wells were then washed 5 times with deionized water. Fixed cells were stained with 0.4% sulforhodamine B dissolved in 1% acetic acid and dark-incubated for 25-30 minutes. Plates were carefully washed 5 times with 1% acetic acid to remove the excess dye and left to air dry. Bound dye was solubilized with 100 μL of 10 mM Tris. Absorbance was measured at 690 nm for the background and 510 nm for the dye.

Accumulation of MnPs in 4T1 cells in the presence and absence of ascorbate

To determine the MnP uptake, cells were seeded at a density of 3 × 10⁵ cells per well in a 6-well plate. Cells were left for 24 h in CO₂ incubator to adhere. Then 5 μM MnP alone or in combination with 1 mM ascorbate were added. Catalase was added to each well in the amount of 1,000 U/mL to avoid cell damage by MnP/Asc-generated H₂O₂. Controls containing MnP plus catalase, and ascorbate alone, were tested in parallel. After incubation in CO₂ incubator, the wells were washed with PBS. The 1 mL of 2% SDS was then added to each well to solubilize the cells. After 24 h the content of each well was centrifuged for 10 minutes at 75,000 × g. The clear supernatant was used to measure the MnP content as previously described [58]. In brief, spectra were recorded and the area under Soret band quantified. Standard curve was used to assess the concentration of MnP in each sample. Standard curve was constructed by adding known concentration of MnP to solubilized, untreated cells. Separate standard curves were prepared for each compound. Protein levels were measured by Lowry method [59].

Measurements of redox states of GSH/GSSG and Cys/CySS in 4T1 cells

The 4T1 cells were seeded in 6-well plates at the density of 0.5 × 10⁵ cells/well and allowed to reach ~80% confluence. They were then treated for 4 hours in a growth medium with MnTE-2-PyP⁵⁺ (5 μM or 15 μM) or ascorbate (1 mM or 3 mM) alone or in combination. Cells were washed 3 times with ice-cold PBS. Then 0.5 mL of the solution - consisting of 5% perchloric acid, 0.2 M boric acid and 10 μM γ-glutamylglutamate, γ-EE - was added, cells were scraped and transferred to microfuge tubes and centrifuged at 13,500 rpm for 1 min. The supernatants were transferred into new tubes for further derivatization as described elsewhere [60]. Concentrations of thiols and disulfides were determined by integration relative to internal γ-EE standard [61]. Reduction potential (Eh) was calculated using GSH and GSSG, or Cys and CySS concentrations by Nernst equation as described previously [62]. Protein levels were determined by Thermo Scientific Pierce BCA Protein Assay Kit (cat# 23225).

Measurements of intracellular total thiol content

The 4T1 cells were treated as described above for GSH/GSSG and Cys/CySS measurements. They were harvested and lysed by ultrasonication on ice. The intracellular thiol (RSH) content was determined as described in [63].

In vivo study

Mouse 4T1 breast cancer model

A total of hundred fifty-two 6-8 week-old female Balb/c mice weighing on average 20 g (Jackson Laboratory) were used. Two experiments were performed where either the impact of the dose of MnP (0.2 or 2 mg/kg/day), or the stage in the development of tumor (volume at the moment the treatments were initiated), on the tumor growth suppression were explored. In both studies, tumors were established via subcutaneous (sc) injection of 100 μL suspension of 10⁶ cells into the mouse flank. In a 1^st experiment, eighty mice were randomized into four treatment groups: (1) PBS (200 μL, intraperitoneally, ip); (2) Asc (4 g/kg/day, ip); (3) MnTE-2-PyP⁵⁺ (0.2 mg/kg/day, sc); and (4) MnTE-2-PyP⁵⁺ (0.2 mg/kg/day, sc) + Asc (4 g/kg/day, ip). Treatment started 24 hours post injection of tumor cell suspensions. In a 2^nd experiment, once the tumor reached volume of ~80 mm³, seventy-two mice were randomized into six groups: (1) PBS (200 μL, ip); (2) Asc (4 g/kg/day, ip); (3) MnTE-2-PyP⁵⁺ (0.2 mg/kg/day, sc); (4) MnTE-2-PyP⁵⁺ (2 mg/kg/day, sc); (5) MnTE-2-PyP⁵⁺ (0.2 mg/kg/day, sc) + Asc (4 g/kg/day, ip); and (6) MnTE-2-PyP⁵⁺ (2 mg/kg/day, sc) + Asc (4 g/kg/day, ip). Tumor volumes were measured daily with vernier digital caliper and calculated with formula: V = length × width × width × 1/2. The doses of MnP and ascorbate were based on those previously reported [45, 64]. Mice were treated for the duration of the study (15 days and 18 days for the 1^st and 2^nd experiment, respectively). The humane end point was set at an average tumor volume of 1,500 mm³. To allow for the significant clearance of MnP from plasma [65], mice were sacrificed at 24 h post last treatment and tumors and muscles from the opposite leg excised and snap frozen in liquid nitrogen. No significant differences in mouse weights were found among treatment groups during the course of both experiments. Animal handlings and procedures were approved by Duke University Institutional Animal Care and Use Committee.

Statistical analysis

It was performed using a mixed-effects repeated measures model and a mixed-effects parametric survival-time model. First, the mixed-effects repeated measures model for ln(tumor volume) was estimated, including linear and quadratic time-trends, as well as interactions of time-trends with three treatment groups (Asc, MnP, and MnP/Asc, with baseline being the control group). A test conducted to evaluate the incremental effect of adding quadratic factors to the model was statistically significant (p-value is 0.0000). This test confirms a nonlinear relationship between time and ln(tumor volume) in the control and treatment groups. The model assumes mouse-specific random effects and an autoregressive correlation structure, of order one, within a mouse. In particular, it includes two mouse-specific random effects: random intercept and a random slope on the time variable at a mouse level (the model thus assumes that the effect of time could be mouse-specific). This model is estimated using the maximum-likelihood method. Additionally, pairwise comparisons between treatment groups were conducted (p-values were adjusted using Sidak correction in order to account for multiple comparisons). Second, the mixed-effects parametric survival-time model for the time the tumor volume reaches the endpoint at 1,500 mm³ was applied. In particular, the random-effects Weibull model, with normally distributed random effects within each mouse, was estimated using the maximum likelihood method with mode-curvature adaptive Gauss-Hermite quadrature integration method. All statistical analyses were carried out in STATA 14.0 Data Analysis and Statistical Software package.

Accumulation of MnTE-2-PyP⁵⁺ in tumor and muscle

The MnP levels in tumor and muscle were measured in the study where MnP was administered at 0.2 mg/kg/day. Tumors and muscles from opposite legs were homogenized and proteins removed from tissue with 1% acetic acid in methanol, solvent evaporated and residue reconstituted in LCMS/MS mobile phase. Analyses were performed on an Applied Biosystems MDS Sciex 4000 Q Trap LC/MS/MS spectrometer at Duke Cancer Institute, Pharmaceutical Research Shared Resource, PK/PD Core laboratory as detailed elsewhere [65].

RESULTS AND DISCUSSION

Over decades we have synthesized numerous Mn porphyrins [8, 9, 39, 66, 67]. We initially modified redox properties of Mn site (described by metal-centered reduction potential, E_1/2, for Mn^IIIP/Mn^IIP redox couple) to optimize MnPs as catalysts for O₂^.− dismutation [8, 9, 39, 66, 67]. It was subsequently shown that ability of MnPs to catalyze O₂^.− dismutation parallels their ability to reduce ONOO⁻ [8, 9, 68]. It became soon obvious that bioavailability is another major factor that affects MnP therapeutic potential and can compensate for somewhat inferior redox properties [42, 54, 58]. Bioavailability is affected by the charge, size, shape, polarity and lipophilicity of MnPs. Our studies and studies of others demonstrate that, while SOD-like and ONOO⁻ reduction activities are not excluded, some other processes may predominate in vivo. Such are those related to the coupling of MnPs with major cellular reductants, ascorbate and small molecular weight thiols (GSH, cysteine), as well as protein thiols [8, 9]. We have already indicated that cationic ortho Mn(III) N-substituted pyridyl-porphyrins bear therapeutic potential as catalysts for ascorbate oxidation with concomitant production of cytotoxic H₂O₂ [33, 45, 69]. H₂O₂ can subsequently be employed by MnP to oxidize or S-glutathionylate critical protein thiols, thereby affecting cell survival pathways [8-11, 32, 70] (Figure 2).

Figure 2. Catalysis of ascorbate oxidation by MnP leads to H₂O₂ production.

MnP cycles between +3 (Mn^IIIP) and +2 (Mn^IIP) oxidation states while oxidizing ascorbate to ascorbyl radical, Asc•. The re-oxidation of Mn^IIP with oxygen closes the catalytic cycle, giving rise to superoxide, O₂^.− which readily undergoes dismutation to H₂O₂. MnP subsequently uses H₂O₂ to suppress cellular transcriptional pathways such as NF-κB, AP-1 and SP-1. The magnitude of the outcome is dependent upon the levels of MnPs and H₂O₂. Based on our data and those of others, both Mn porphyrin (this work and ref [45]) and H₂O₂ [10-14] are at higher levels in tumor than in normal tissue. Consequently, suppression of tumor growth accompanied by normal tissue healing was seen herein and elsewhere [1, 2, 8-11, 39, 66, 67]. Asc and Asc^• describe the predominant monodeprotonated ascorbic acid (HA⁻) and deprotonated ascorbyl radical at pH 7.8, respectively. The catalysis of ascorbate oxidation is characterized in this work with initial rates of ascorbate oxidation (v₀(Asc)_ox) and oxygen consumption/reduction (v₀(O₂)_red).

H₂O₂-mediated therapeutic effects of MnPs

Work done by Tome's group [10, 11] was instrumental in providing undoubted evidence that cytotoxicity of MnP/dexamethasone to lymphoma cell (but not to normal lymphocytes) is due to the production of high levels of H₂O₂. H₂O₂ was subsequently used by MnP/GSH to catalyze S-glutathionylation of p65 subunit of NF-κB. In turn, levels of H₂O₂ remained unchanged. When either H₂O₂ or GSH is removed no S-glutathionylation was observed [10].

Tome's group subsequently showed that other protein thiols have been modified also with subsequent loss of protein activity. Complexes I, III and IV of mitochondrial respiration were S-glutathionylated with MnP/H₂O₂/GSH which led to the inhibition of complexes I and III and suppression of ATP production. When MnP was coupled with 2-deoxyglucose, the glycolysis was suppressed accompanied with suppression of ATP [10]. S-glutathionylation of glycolysis proteins likely plays role also. In turn MnP/H₂O₂ suppressed both mitochondrial and glycolytic energy sources of cancer cell, while not being harmful to normal lymphocytes [10]. Both cytosolic and mitochondrial actions of MnP/H₂O₂ play role in MnP therapeutic effects. While S-glutathionylation of p65 subunit of NF-κB and glycolysis, as well as phosphorylation of ERK [33] happen in cytosol ([11] and refs therein), the impact of MnP on electron transport chain is a clear mitochondrial event. Nucleus may be involved also in the actions of MnPs: Piganelli's group suggested (based on an indirect evidence) that MnTE-2-PyP⁵⁺ oxidizes p50 subunit of NF-κB at the level of nucleus preventing in turn the DNA binding of NF-κB - such suggestion was supported by 3-fold higher levels of MnP in nucleus than in cytosol [21, 36]. Several studies by St. Clair's group showed the effect of MnPs at mitochondrial level also [71-73].

Once the key impact of H₂O₂ in the actions of MnP was identified, we started understanding the mechanism of MnP radio- and chemosensitizing effects [2, 8], We also realized the therapeutic potential of MnP/Asc system as a source of H₂O₂ in its own right, but also when combined with additional sources of H₂O₂ - radiation or chemotherapy. Radio- and chemotherapy have always been standard of care with patients who are recruited for prospective Clinical Trials. We undertook this study to identify those properties of MnP that affect its ability to cycle with ascorbate thereby producing H₂O₂ and to correlate such ability with cytotoxicity to cells and tumors. Studies are also in progress to understand the impact of the chemical and physical properties of MnP on a subsequent step - MnP/H₂O₂/GSH-driven oxidation of protein thiols [32, 74].

Cancer-related pathways, affected specifically by the MnP/Asc H₂O₂-producing system, were studied on inflammatory breast cancer cell SUM149. The data correlate well with lymphoma studies [10, 11]. The study indicated that both cytosolic and mitochondrial pathways are involved, and that at least in part NF-κB was a key player involved in MnP/Asc toxicity [33]. GSH was found to play major role also, and catalase eliminated the toxicity. MnP/Asc system was able to reduce GSH levels and phosphorylate NF-κB and ERK (extracellular signal-regulated kinase). The X-linked inhibitor of apoptosis protein, XIAP, was decreased and annexin V increased indicating apoptosis-related cytotoxicity imposed by MnP/Asc system. In addition, the translocation of apoptosisinducible factor, AIF, into nucleus implicates AIF-mediated and caspase-independent SUM 149 cell death pathway.

The interaction of ascorbate with other redox-active compound, menadione, has been explored and clinically tested. An oral combination of ascorbate and menadione (Apatone) has been evaluated as an anticancer agent in a clinical study in prostate cancer patients who had failed standard therapy. Promising delay in the biochemical progression of the disease was observed in Apatone-treated patients. Indeed, a significant increase in the PSA (Prostate Serum Antigen) doubling time was induced by Apatone treatment without dose-limiting adverse effects [75-77].

Fourteen MnPs (Figure 1) were evaluated in this work with the goal to: (1) identify the best catalyst(s) of ascorbate oxidation among existing Mn porphyrins and to correlate such data with their thermodynamic and kinetic properties most so with their SOD-like potency; (2) show that the magnitude of the catalysis of ascorbate oxidation parallels the magnitude of the cellular cytotoxicity of MnP/Asc system; and (3) evaluate the tumor cytoxicity of best catalyst in a mouse model. These MnPs studied differ largely with respect to E_1/2, charge, size, shape, bulkiness, polarity and lipophilicity. Among these compounds, MnTBAP³⁻ is anionic and disfavors reaction with negatively charged ascorbate on both thermodynamic and electrostatic grounds. All other porphyrins are positively charged. Among them there are MnPs with short and long linear pyridyl substituents. Some bear polar oxygen atoms embedded within linear substituents. Two MnPs have cyclic substituents and are therefore very bulky. Few MnPs are hydrophilic, but many are lipophilic. Among lipophilic MnPs, some are polar due to the presence of oxygens within pyridyl side-chain substituents. We have described the ability of MnPs to catalyze ascorbate oxidation with initial rates of ascorbate oxidation (v₀(Asc)_ox), and initial rates of oxygen consumption/reduction (v₀(O₂)_red). We have discussed kinetic data with respect to the thermodynamic properties of Mn site (E_1/2 of Mn^IIIP/Mn^IIP). We then explored the ability of seven most characteristic MnPs to induce cytotoxicity to three breast cancer (human MCF-7 and MDA-MB-231 and mouse 4T1) and non-cancerous epithelial breast cell (human HBL100). We finally conducted in vivo study on the suppression of 4T1 tumor growth in a mouse sc flank model.

Aqueous chemistry of MnP/Asc system - Reduction potential of MnP controls Asc oxidation rate

The ability of MnPs to catalyze ascorbate oxidation was assessed by direct (spectrophotometrical) and indirect (electrochemical) means (Figure 2) via measuring initial rates of ascorbate oxidation (v₀(Asc)_ox, Table 1 and Figure 3) and initial rates of oxygen reduction, (v₀(O₂)_red, Table 1 and Figure 4). The metal-centered reduction potential, E_1/2, as a measure of MnP redox properties and R_f, as a measure of their lipophilicity are also listed in Table 1.

Initial rates, v₀(Asc)_ox and (v₀(O₂)_red, are related to E_1/2 in a bell shape fashion (Figure 3B and 4B). MnPs with negative E_1/2 (such as MnTBAP³⁻ and MnTE-2-PyPhP⁵⁺) have very low initial rates. Such compounds cannot be easily reduced with ascorbate in a 1^st step of catalytic cycle (Figure 2). In addition, anionic MnTBAP³⁻ would have been repelled from negatively charged monodeprotonated ascorbate [8, 9]. Conversely, those MnPs with too positive potential are readily reduced to the Mn +2 oxidation state but are difficult to be oxidized back to the Mn +3 state to close the catalytic cycle (Figures 2, 3, and 4). Under aerobic conditions, due to high oxygen levels, the reoxidation of MnP will likely occur with O₂ [33]. The steric hindrance towards the approach of ascorbate to Mn site of bulky molecules, such as MnTPhE-2-PyP⁵⁺ (Figs. 3C and 4B), MnTnHexOE-2-PyP⁵⁺, and MnTnOct-2-PyP⁵⁺, contributes also to the lower values of v₀(Asc)_ox relative to MnTE-2-PyP⁵⁺.

Such impact of E_1/2 on ascorbate oxidation coincides with the optimal E_1/2 for the catalysis of O₂^.− dismutation; the reasoning being similar. Like SOD enzyme, its mimic must be able to equally efficiently oxidize and reduce O₂^.−, thus operating at E_1/2 which is right in between the reduction and oxidation potential of O₂^.−₂, in the range of ~ +200 to ~ +400 mV vs NHE [42]. It is important to note that SOD-like ability of MnPs parallels their ability to undergo all reactions thus far studied, such as reduction of ONOO⁻ [68], catalase-like activity [54], ability to catalyze ascorbate oxidation (this work) and ability to mimic GPx enzyme [32].

The ascorbate oxidation was followed in Tris and phosphate buffers at pH 7.8. Those buffers differ with regards to hydrogen bonding. The difference between the Mn/ascorbate kinetics in those two buffers was prominent with those MnPs which have the largest affinity for hydrogen bonding: MnTnBuOE-2-PyP⁵⁺ and MnTnHexOE-2-PyP⁵⁺ (Figures 3 and 4). The hydrogen of the protonated Tris buffer could form hydrogen bonding with the oxygen atom of the oxygenated butoxyethyl or hexoxyethyl side chains of MnTnBuOE-2-PyP⁵⁺ and MnTnHexOE-2-PyP⁵⁺. This would increase the net positive charge of MnP, which in turn could enhance the guidance of the negatively charged ascorbate towards Mn site. The strong impact of Tris relative to phosphate buffer on the kinetics of ascorbate oxidation by alkoxyalkyl derivatives suggests that it is still difficult to fully predict the magnitude of the therapeutic effects of MnPs in a very complex milieu of a cell. This is particularly true in the case of those MnPs that bear different amphiphilic features which allow diverse interactions with various biomolecules: 5 positive charges, alkyl chains, oxygen atoms, cationic electron-deficient Mn site favoring axial ligation of different species.

Lipophilicity of MnPs

The lipophilicity of MnPs, described with chromatographic retention factor R_f, along with their E_1/2 values are plotted in Figure 5. We have explored MnPs of diverse properties: those that are hydrophilic (such as MnTE-2-PyP⁵⁺) and lipophilic (MnTnOct-2-PyP⁵⁺) but have favorable E_1/2 that allows them to cycle with ascorbate. We have also studied those that are lipophilic but whose E_1/2 limits their cycling with ascorbate (such as MnTE-2-PyPhP⁵⁺). Among lipophilic compounds MnTnHexOE-2-PyP⁵⁺ is polar as it bears oxygen atoms in pyridyl substituents, while MnTnOct-2-PyP⁵⁺ bears apolar long alkyl chains. Polar oxygen atoms largely influence the reactivity of MnTnHexOE-2-PyP⁵⁺ towards ascorbate (see discussion in Aqueous Chemistry of MnPs). Some lipophilic compounds bear linear substituents such as MnTnOct-2-PyP⁵⁺, while a lipophilic MnTPhE-2-PyP⁵⁺ has aromatic rings as N-substituents. These rings affect the lipophilicity of MnTPhE-2-PyP⁵⁺ and make it bulkier, and in turn restrict its accumulation within cell (see discussion under Accumulation MnPs in 4T1 cells in the absence and presence of ascorbate. The interplay of those properties affect reaction of MnPs with ascorbate and their accumulation within cell where MnP/H₂O₂ driven oxidation of signaling proteins, affecting cell survival, takes place.

Figure 5. Lipophilicity of MnPs expressed in terms of chromatographic retention factor, R_f (blue bars).

R_f presents ratio of compound path to solvent path on silica gel plastic plates with 1:1:8=KNO_3(sat) H₂O:H₂O:CH₃CN as solvent. MnPs with low, medium, and high lipophilicity were included in this study. Their E_1/2 values (marked in red) range from −65 to +340 mV vs NHE, and (illustrate the diversity of MnPs investigated in this study.

Magnetic resonance imaging studies

Mn porphyrins

Mouse prostate cancer study indicated that MnTE-2-PyP⁵⁺ and MnTnHex-2-PyP⁵⁺ have potential as MR imaging agents being superior to commercially available gadolinium chelates [56]. We have therefore explored different MnPs on their MRI properties in the presence and absence of ascorbate, our major endogenous antioxidative defense. The difference in the relaxivities of paramagnetic MnPs stems mainly from the difference in the degree of interaction of the paramagnetic Mn with water protons in the inner coordination sphere of its porphyrin complex [78-80]. In Mn^IIIP, the Mn +3 has 4 unpaired d electrons and two water molecules in axial positions that are available for exchange with bulk waters. Our MRI aqueous studies demonstrated that T1 relaxivity depends upon MnP ability to exchange axial water with bulk waters. T1 relaxivity is thus controlled by the strength of axial Mn-H₂O bond, which is best described by the deprotonation constant for 1^st axial water, pK_a(ax). Steric factors were also found to play a role, giving rise to a bell-shaped relationship between the T1 relaxivity and pK_a(ax) (Figure 6). We have reported that pK_a(ax) is linearly but inversely correlated with E_1/2 [54, 81]. When the pK_a(ax) decreases (from ~ 12 to ~11), as anticipated, the E_1/2 increases. The drop in pK_a(ax) (increase in E_1/2) from MnTBAP³⁻ and MnTE-2-PyPhP⁵⁺(#1) to MnTE-3-PyP⁵⁺(#2) and to MnTE-2-PyP⁵⁺(#3) is accompanied by strengthening of Mn-H₂O bond, which lowers the water-exchange rates and causes the T1 relaxivity to drop (Figure 6B and 6C). Yet T1 relaxivity increases as MnPs [MnTPhE-2-PyP⁵⁺(#4), MnTnHexOE-2-PyP⁵⁺(#5) and MnTnOct-2-PyP⁵⁺ (#6)] become bulky (regardless of further drop in their pK_a(ax) values), giving rise to a bell shape relationship (Figure 6B and 6C).

Figure 6. MRI studies of MnPs.

MnPs were 0.1 mM and ascorbate was 5 mM in 0.05 M phosphate buffer, pH 7.8, (25±1)⁰C. The T1 relaxivity measurements were carried out immediately after ascorbate addition. (A) T1 relaxivity values in the absence and presence of ascorbate, Asc, are plotted for different MnPs. (B) T1 relaxivity values in the presence and absence of ascorbate are plotted versus metal-centered reduction potential, E_1/2. (C) T1 relaxivity values are plotted versus E_1/2 and pK_a(ax). The following porphyrins are plotted in (B) and (C): (#1) MnTE-2-PyPhP⁵⁺, (#2) MnTE-3-PyP⁵⁺, (#3) MnTE-2-PyP⁵⁺, (#4) MnTPhE-2-PyP⁵⁺, (#5), MnTnHexOE-2-PyP⁵⁺, and (#6) MnTnOct-2-PyP⁵⁺.

Our MRI results are in agreement with data published on the water exchange rates of several MnPs [82]. Anionic MnTSPP³⁻ (E_1/2 = −160 mV vs NHE) and MnTBAP³⁻ (E_1/2 = −194 mV vs NHE) have similar redox properties, overall charge and stericity. MnTSPP³⁻ has pK_a(ax) of 12.3, while the pK_a(ax) of MnTBAP³⁻ is 12.6. Thus, relative to MnTE-2-PyP⁵⁺ (pK_a(ax) =10.7), with less stronger binding of axial water, MnTSPP³⁻ has >10-fold higher k_ex(H₂O). While of similar pK_a1 values (10.7 vs 11), larger stericity of MnTnHex-2-PyP⁵⁺ relative to MnTE-2-PyP⁵⁺ resulted in ~40% larger k_ex(H₂O). Most likely the same steric factors valid for MnTnHex-2-PyP⁵⁺ contribute to the increase in T1 relaxivity of even bulkier molecules, MnTPhE-2-PyP⁵⁺(#4), MnTnHexOE-2-PyP⁵⁺(#5) and MnTnOct-2-PyP⁵⁺(#6), relative to MnTE-2-PyP⁵⁺ despite similar pK_a(ax) values [54].

Mn porphyrins + ascorbate

Ascorbate affects T1 relaxivity of only those MnPs which are easily reducible (Figure 6A and B). The T1 relaxivity drops as Mn^IIIP gets reduced. It has been argued that the elongation of the effective electron spin density of the manganese center in the direction of the bound water molecules is mostly responsible for the large increase in the relaxivity of Mn^IIIP (with Mn in +3 oxidation state) when compared to other potential contrast agents such as Gd³⁺ [Gd(DTPA)²⁻] and Mn²⁺ chelates [Mn(EDTA)²⁻] [56, 83, 84]. The point dipole approximation that predicts a spherical distribution of the electron cloud of manganese in S-state ions such as Gd³⁺ and Mn²⁺, does not hold for Mn^IIIP, since the electron spin density of Mn +3 in Mn^IIIP complex is closer to the protons of the bound waters [83-85]. The increased dipolar interaction results in a shorter distance (r) between the metal center and the water protons. The paramagnetic effect is inversely proportional to r, resulting in higher relaxivity of Mn^IIIP than for Mn²⁺ with S-state electron configurations.

Stability of MnPs towards Asc-mediated oxidative degradation

Stabilities of several MnPs explored in cellular studies are shown in Figure 7. The most stable compounds are those (MnTBAP³⁻ and MnTE-2-PyPhP⁵⁺) which are either not at all, or only slightly reducible with ascorbate (exemplified with negative values of E_1/2). Due to the lack (MnTBAP³⁻) or very slow (MnTE-2-PyPhP⁵⁺) rate of ascorbate oxidation either no or very small amount of H₂O₂ is produced in the medium. This results in their negligible degradation (Figure 7, shown for MnTE-2-PyPhP⁵⁺). MnTE-2-PyP⁵⁺ demonstrates higher stability than its meta(3) analog, MnTE-3-PyP⁵⁺. Our data suggest that oxidative degradation proceeds via Mn +3 oxidation state. In turn meta(3) analog, MnTE-3-PyP⁵⁺ which rapidly re-oxidizes to Mn +3 oxidation state is the least stable among MnPs studied (Figure 7). The ortho(2) MnPs (MnTE-2-PyP⁵⁺, and MnTPhE-2-PyP⁵⁺) with ~200 mV higher E_1/2 than MnTE-3-PyP⁵⁺ are stabilized in Mn +2 oxidation state, resist re-oxidation and in turn oxidative degradation (Figure 7). The data agree well with reported data on H₂O₂-driven oxidative degradation of methyl isomers, MnTM-2(or 3 or 4)-PyP⁵⁺ [82]. The E_1/2 of MnTnHexOE-2-PyP⁵⁺ and MnTnOct-2-PyP⁵⁺ is 139 or 85 mV higher than that of MnTE-2-PyP⁵⁺. Therefore they are further stabilitized in Mn +2 state and are thus less prone to close the catalytic cycle. The consequence of such behavior is (i) lower ability to catalyze ascorbate oxidation and (ii) lower affinity towards degradation. Additionally both are bulky molecules which further reduces their catalytic potency (see Table 1, Figure 3, Figure 4, and Cytotoxicity of MnP/Asc data below).

Figure 7. Stabilities of MnPs towards Asc-mediated oxidative degradation.

5 μM MnPs were incubated with 1 mM ascorbate in the 0.05 M Tris buffer at pH 7.8. The H₂O₂, which is produced during the MnP/Asc cycling, accumulates in the solution and leads to MnP degradation The stability of MnPs was followed spectrophotometrically at Soret band which corresponds either to Mn +3 (Mn^IIITE-3-PyP⁵⁺ and Mn^IITE-2-PyPhP⁴⁺), or +2 oxidation state (Mn^IITE-2-PyP⁴⁺, MnTnOct-2-PyP⁵⁺, MnTnHexOE-2-PyP⁵⁺ Mn^IITPhE-2-PyP⁴⁺) oxidation states. Mn^IIITE-3-PyP⁵⁺ is hardly reducible under aerobic conditions and once reduced re-oxidizes to Mn +3 oxidation state order of magnitude faster than MnTE-2-PyP⁵⁺ [33].

In vitro studies

Cytotoxicity of MnP/Asc

MnP/Asc is cytotoxic to breast cancer cells

Seven MnPs, with E_1/2 ranging from −194 to +340 mV vs NHE (Figure 3C and 4C), were selected to explore the impact of ascorbate oxidation rate on the cytotoxic potential of MnP/Asc system. The choice of compounds exemplifies diverse properties of MnPs catalytic potency in aqueous solutions.

Bell-shape relationship between E_1/2-controlled v_o(Asc)_ox and MnP/Asc-mediated cellular cytotoxicity

Redox property of MnP, described by E_1/2, which controls the catalysis of ascorbate oxidation has also major impact on MnP/Asc-induced cellular cytotoxicity (Figures 7-9). Steric and electrostatic factors (see Aqueous Chemistry of MnP/Asc), as well as stability of MnPs towards H₂O₂-driven oxidative degradation (see below) play also a role (Figure 7). The highly negative E_1/2 of −194 mV vs NHE precludes reduction of MnTBAP³⁻ with ascorbate (Figures 3 and 4), and in turn deemed it nontoxic to cells (Figures 8 and 9). With E_1/2 = −65 mV vs NHE, MnTE-2-PyPhP⁵⁺ has already modest ability to catalyze ascorbate oxidation. Thus it induced cell death at higher concentrations than those MnPs with higher E_1/2; its modest lipophilicity contributes favorably to its modest effect (Figures 8 and 9). Its catalytic potency is though much lower than that of two lipophilic MnPs of favorable E_1/2 values: MnTnHexOE-2-PyP⁵⁺ (E_1/2 = +313 mV vs NHE) and MnTPhE-2-PyP⁵⁺ (E_1/2 = +259 mV vs NHE). The bulkiness of MnTnOct-2-PyP⁵⁺ precludes an easy access of ascorbate to Mn site deeply buried within cavity encircled with long alkyl chains. In addition its high E_1/2 of +367 mV vs NHE stabilizes it in +2 oxidation state disfavoring thus closing a catalytic cycle. The lipophilic octyl side-chains contribute to MnTnOct-2-PyP⁵⁺ toxicity as a single agent (Figure 10). Though MnTnHexOE-2-PyP⁵⁺ is also bulky, the presence of oxygen atoms in its pyridyl side-chain substituents facilitate the approach of ascorbate. It also has 50 mV less positive E_1/2 and would thus favor re-oxidation more than does MnTnOct-2-PyP⁵⁺. In turn, MnTnHexOE-2-PyP⁵⁺ is a better catalyst than MnTnOct-2-PyP⁵⁺ (Figures 3 and 4) and when combined with ascorbate it induces higher cellular cytotoxicity MnTnOct-2-PyP⁵⁺/Asc (Figure 8 and 10). A very similar pattern demonstrated with 4T1 was seen with human breast cancer cell line MCF-7 and MDA-MB-231 (Figure 10). With all cell lines and among those MnPs studied, MnTE-2-PyP⁵⁺ bears the highest potential as an anticancer drug when combined with ascorbate. Importantly, MnPs alone exhibit either slight or no toxicity towards cancer cells. The toxicity of MnTnOct-2-PyP⁵⁺ is primarily due to its long lipophilic octyl chains.

Figure 9. Role of the redox properties of MnPs (described by metal-centered reduction potential, E_1/2) in the cytotoxicity of MnP/Asc to 4T1 mouse breast cancer cell line.

Cell viability data are taken from Figure 8. The bell-shape relationship between the cell viability and E_1/2 correlates well with the structure-activity relationship observed between v₀(Asc)_ox and E_1/2 (Figure 3): the higher is E_1/2, the higher is the v₀(Asc)_ox. In turn the higher is the rate of H₂O₂ production and the higher is the cytotoxicity induced by MnP/Asc system. While the bulky compounds (MnTPhE-2-PyP⁵⁺, MnTnHexOE-2-PyP⁵⁺ and MnTnOct-2-PyP⁵⁺) have favorable redox properties, the approach of ascorbate towards their Mn site is hindered. In turn the production of H₂O₂ and the cytotoxicity is diminished relative to a smaller molecule of MnTE-2-PyP⁵⁺, giving rise to bell shape behavior. Reduced cellular uptake of such bulky MnPs may also play a role. The following porphyrins are plotted: (#1) MnTE-2-PyPhP⁵⁺, (#2) MnTE-3-PyP⁵⁺, (#3) MnTE-2-PyP⁵⁺, (#4) MnTPhE-2-PyP⁵⁺, (#5), MnTnHexOE-2-PyP⁵⁺ and (#6) MnTnOct-2-PyP⁵⁺.

Figure 8. Cytotoxicity of MnP/Asc to 4T1 mouse breast cancer cell line.

Cells were incubated with MnP at various concentrations in the absence and presence of 1 mM ascorbate for 24 hours, followed by MTT-based assessment of cell viability.

Figure 10. Cytotoxicity of MnP/Asc to human breast cancer MCF-7 (A) and MDA-MB-231 (B) cell lines.

Cells were incubated with MnP +/− ascorbate at given concentrations for 24 hours. Viability was determined by the MTT assay. The effect of catalase was exemplified with MDA-MB-231. Exogenous catalase (1000 units/mL) fully blocked the cytotoxic effect of MnTE-2-PyP⁵⁺/Asc (B), indicating that H₂O₂ produced in medium is primarily responsible for the observed cytotoxicity.

Cytotoxicity is primarily due to the extracellular H₂O₂ production

Our earlier observations [8, 9, 44, 45], and studies of others [49, 86], as well as the impact that catalase has in preventing cytotoxicity of MnP/Asc (exemplified on MDA-MB-231 in Figure 10), demonstrated that cytotoxicity is primarily due to the MnP cycling with ascorbate outside of cell. The production of H₂O₂ within cell, though, cannot be fully excluded; the minor impact of lipophilicity has been observed. When cells (MCF-7) were pre-incubated with MnPs, and washed afterwards, slight sensitivity towards ascorbate was observed (data not shown).

The cytotoxicity is also dependent upon the stability of MnP catalyst towards oxidative degradation

Due to ~180 mV less positive E_1/2, which facilitates its re-oxidation (from Mn^IIP to Mn^IIIP), meta MnTE-3-PyP⁵⁺ is more prone to oxidative degradation with H₂O₂ than are ortho MnPs studied (Figure 7). Lower stability likely accounts for its lower cytotoxicity at higher concentrations, where MnTE-3-PyP⁵⁺ is rapidly degraded with H₂O₂ in the medium due to its generation and accumulation (Figure 8). The 10-fold higher lipophilicity of MnTE-3-PyP⁵⁺, however, likely compensate for its lower stability relative to MnTE-2-PyP⁵⁺ resulting in similar cytotoxicities at lower concentrations (Figures 8 and 9). Once within cell, H₂O₂ is used by MnP in the catalysis of protein oxidation which ultimately results in suppression of tumor growth [8, 9]. Thus the lipophilicity of MnPs, affecting their cellular uptake, plays also a role.

In summary, the interplay between the redox property, lipophilicity, and stericity/bulkiness of MnPs, which affect MnP/Asc-induced H₂O₂ production, its utilization in oxidative damage of proteins and MnP cellular uptake, controls the magnitude of in vitro and in vivo anticancer effects.

Being identified as the best catalyst among MnPs studies, MnTE-2-PyP⁵⁺ was tested additionally in a cell proliferation assay. The data clearly show that MnP/Asc completely blocked cell proliferation (Figure 12).

Figure 12. MnP/Asc system is not toxic to non-cancerous human epithelial breast HBL-100 cell line.

Cells were incubated with MnP (10 μM) +/− ascorbate (1 mM) for 24 hours, followed by MTT-based assessment of cell viability.

MnP/Asc is not cytotoxic to non-cancerous breast cell line

The sensitivity of non-cancerous cell to MnP/Asc treatment was also studied. At the concentrations used, and under the conditions of this study, none of MnPs was toxic to non-cancerous human breast epithelial cell line HBL-100 in the presence of ascorbate (Figure 12). Only MnTnOct-2-PyP⁵⁺ when applied alone inflicted toxicity due to its excessive lipophilicity and likely surfactant effect arising from its long lipophilic alkyl chains.

Our data on cell cytotoxicity agree well with the data published earlier on several other cancer and normal cell lines. Thus, MnP/Asc system was toxic to 3 different human glioma cell lines, but not to normal astrocytes [47]. It was also toxic to MCF-7 and HeLa cells, but not to normal human dermal fibroblasts, NHDF [8].

MnPs accumulation in 4T1 cells in the absence and presence of ascorbate

Based on our and data of others, under in vitro conditions, cytotoxic H₂O₂, produced in the medium (in the reaction of MnP with ascorbate) enters the cell where it exhibits cytotoxicity via reactions catalyzed by MnPs [33, 41, 54]. In turn accumulation of MnPs is anticipated to play a role in anticancer effects of MnPs. The study was therefore conducted to determine the accumulation of MnPs and the impact ascorbate has on it. Modest increase in MnP accumulation in 4T1 cells has been observed in the presence of ascorbate and was pronounced with those MnPs that could be reduced with ascorbate (Figure 13). Thus, no impact of ascorbate was observed on cellular uptake of MnTBAP³⁻. The accumulation of MnTE-2-PyP⁵⁺ in 4T1 cells agrees well with the data on its accumulation in tumors derived from 4T1 cells (Figure 15B) where no major effect of ascorbate was also seen. As anticipated, a more lipophilic MnPs, such as MnTPhE-2-PyP⁵⁺, MnTnHexOE-2-PyP⁵⁺ and MnTnOct-2-PyP⁵⁺, accumulate in 4T1 cells to a larger extent than more hydrophilic MnPs, such as MnTE-2-PyP⁵⁺ and MnTE-3-PyP⁵⁺ (Figure 13).

Figure 13. Cellular uptake of MnPs.

4T1 cells were incubated with 5 μM MnP in the absence and presence of 1 mM ascorbate. Catalase (1000 unit/mL) was added to the medium to prevent cell damage by H₂O₂ produced via MnP/Asc coupling.

Figure 15.

(A) Anticancer effect of MnTE-2-PyP⁵⁺/Asc on a growth of 4T1 mammary carcinoma in a mouse flank. 10⁶ cells were implanted into the flank of Balb/c mouse. Treatment started 24 hours after the injection of 4T1 tumor cells. MnTE-2-PyP⁵⁺ was administered sc at 0.2 mg/kg/day and ascorbate was given ip at 4g/kg/day throughout the study. 20 mice were used per group. (B) At day 15^th, 95% of mice treated with MnP/Asc have tumors whose volumes were less than 1,500 mm³ (C) ~7-fold higher accumulation of MnTE-2-PyP⁵⁺ was found in tumor than in muscle and was not affected by ascorbate. Of note, a 2^nd study was done where treatment started once tumors on average reached 80 mm³. As similar tumor suppression was observed in either study, early start of treatment when tumor was not established has no additional impact on tumor growth suppression by MnP/Asc.

MnP/Asc increases 4T1 cell oxidative burden

Earlier studies clearly indicated that MnP/Asc is a source of H₂O₂ which MnP employs for oxidative modifications of signaling proteins inducing in turn cytotoxicity [33, 45]. Herein we demonstrated that MnP/Asc system deprives cells from thiol-derived reducing equivalents. The MnP/Asc-driven oxidation of glutathione and cysteine in 4T1 cells allowed us to quantify the changes in cellular reduction potential towards more oxidative environment (Figure 14). Levels of reduced global thiols, RSH, (Figure 14A) and glutathione, GSH, (Figure 14B) were significantly lower in the cells treated with the MnP/Asc relative to those cells that received single treatments. Based on the ratios of oxidized and reduced glutathione (GSSG/GSH) and cysteine (CySS/Cys), the cellular reduction potentials were calculated and are shown in Figure 14C and D. Our results demonstrate that MnP/Asc imposes higher oxidative burden on cell redox environment than either treatment alone. Data agree well with the results on redox proteomics to be addressed in a subsequent publication.

Figure 14. Effect of MnP/Asc treatment on intracellular thiol redox status.

The 4T1 cells were treated for 4 hours with MnTE-2-PyP⁵⁺ (5 μM or 15 μM) or ascorbate (1 mM or 3 mM) alone or in combination. (A) Intracellular concentrations of reduced total thiols were measured spectrophotometrically in the cell lysates using 5,5'-dithiobis-(2-nitrobenzoic acid) (Ellman's reagent, DTNB). (B) Significant reduction of intracellular GSH concentration is observed in 4T1 cells as a result of MnP/Asc treatment. (C) Steady-state reduction potentials (E_h) were determined for intracellular GSSG/GSH redox couple in various treatment cohorts. (D) Steady-state cellular reduction potentials (E_h) were determined for intracellular CySS/Cys redox couple in various treatment cohorts. The E_h values were calculated using Nernst equation as described previously [62]. Bars represent the mean ± SEM (*p < 0.05).

In vivo study

Mouse flank tumor study

Tumor growth suppression by MnTE-2-PyP/Asc in a sc mouse flank model

Our data on MnTnHex-2-PyP⁵⁺/Asc system in a 4T1 mouse flank tumor model suggested its anticancer potential [45]. Only trend towards significance was, though, demonstrated because (i) the treatment started when tumor was already in aggressive phase of growth (volumes averaging ~250 mm³), and (ii) MnTnHex-2-PyP⁵⁺ is not a very good catalyst of ascorbate oxidation (Figures 3 and 4). Also the toxicity, due to fairly high dosing of a very lipophilic MnP, might have outbalanced its efficacy [8, 9].

Among MnPs characterized, MnTE-2-PyP⁵⁺ has the highest ability to catalyze ascorbate oxidation in aqueous system and the highest anticancer efficacy in cellular studies exemplified herein with 3 breast cancer cell lines (Figures 8-10). The proliferation assay demonstrated that MnTE-2-PyP⁵⁺/Asc treatment inhibits cancer cell proliferation (Figure 11). It was therefore forwarded to a mouse flank 4T1 breast cancer model. Two studies were done. In a 1^st study either 2 or 0.2 mg/kg/day dose of MnP was tested in combination with pharmacological dose (4g/kg/day) of ascorbate, and treatment started once tumors on average reached 80 mm³. In a 2^nd study MnP/Asc treatment started 24 h after the injections of tumor cells (Figure 15A). As similar tumor suppression was observed in either study, early start of treatment appears not to be necessary. Early in development tumor has neither developed its vasculature, nor are the supporting cells, such as macrophages, infiltrated into its mass. Those events likely play role in the actions of MnPs [6, 45, 87].

Figure 11. MnTE-2-PyP⁵⁺/Asc inhibits proliferation of human breast cancer cells.

MCF-7 cells were incubated with either 1 (data not shown) or 5 μM MnTE-2-PyP⁵⁺ and 1 mM ascorbate. Complete inhibition of proliferation by MnP/Asc under identical conditions was also seen with 4T1 mouse breast cancer cells. MnTE-2-PyP⁵⁺ was chosen for cell proliferation and mouse study. When combined with ascorbate (i) it has the largest ability to catalyze ascorbate oxidation, and (ii) exhibited the largest cytotoxicity towards 3 different cancer cell lines.

Statistical analysis, namely the estimated mixed-effects repeated measures model, demonstrated that MnP/Asc treatment has significant tumor growth suppression compared to PBS-treatment group (p<0.01). As tumor continued growing, the impact of MnP/Asc became more pronounced (Figure 15A). In particular, for days 12-15, pairwise comparisons show that differences between MnP/Asc group vs each of PBS, Asc and MnP groups were statistically significant, p<0.05. On day 15th, the majority of mice in PBS, Asc and MnP treatment groups (>60%) had a tumor volume larger than 1,500 mm³, while in the MnP/Asc treatment group 95% of animals remained under this threshold (namely, 19 out of 20 mice), as shown in Figure 15B. Furthermore, the results of the survival-time model estimation suggest that the MnP/Asc treatment has a statistically significant negative effect on the tumor reaching 1,500 mm³ volume, compared to PBS-treatment group (p=0.0008). The differences between the effects of the MnP/Asc treatment and that of Asc and MnP treatments are also statistically significant at a level lower than 1% (p-values are 0.0010 and 0.0017, respectively). These results thus suggest that mice under the MnP/Asc treatment have a significantly higher probability of not reaching the tumor volume of 1,500 mm³ than other three groups (PBS, Asc, and MnP) within the time-period studied.

Several different MnPs have recently been explored on their in vivo anticancer potential in presence of combined sources of H₂O₂ - MnP/Asc and radiation [46]. The addition of radiation profoundly amplified the anticancer effect of MnP/Asc [46].

Some of the cancer-related pathways associated with treatment, where Mn was coupled with RT-driven H₂O₂ production, have been assessed. The gene profiling of human D 245-MG glioma growing on Balb/c nu/nu mouse flank showed that MnTnBuOE-2-PyP⁵⁺/RT, relative to RT, down-regulated metastatic pathways (ctss, catepsin L, becn1, beclin1) as well as anti-apoptotic and NF-κB pathways (Nfkb1, Bcl211, Bcl2) and PI3kinase and mTOR (Rsp6kb1). The protein translation changes were implicated also (EIF5b and Rsp6kb1) [8]. The collaborative efforts of ours and Dean Jones's group on MnP/Asc-driven impact on redox proteomics support such observations [manuscript in preparation].

Preferential accumulation of MnP in tumor relative to normal tissue

Ascorbate does not significantly affect accumulation of MnTE-2-PyP⁵⁺ in either tumor or muscle. However large difference in MnP accumulation (of ~7-fold) was found in tumor vs normal tissue (Figure 15B). Such selectivity in MnP accumulation, along with high ascorbate accumulation in tumor via Glut transporters [88], substantiates differential impact of MnP, in the presence of exogenous source(s) of H₂O₂, on tumor vs normal tissue.

Concluding remarks

It has been well documented that, when compared to normal tissue, tumor is frequently under oxidative stress, i. e. has high levels of H₂O₂ [8, 9, 15-31]. The magnitude of oxidative stress depends upon the stage of tumor development and is frequently due to the lower expression and/or activity of endogenous antioxidative defenses as well as to the perturbed ratio of superoxide dismutases (most so MnSOD) to enzymes that controls the H₂O₂ levels, such as catalases, GPx and peroxiredoxins [8, 9, 15, 17-31]. Similar to MnP/radiation and/or MnP/chemotherapy, MnP/Asc system was explored as an anticancer strategy to increase tumor H₂O₂ levels [2, 6, 8-11, 33, 45-47, 64]. Cycling of MnP with ascorbate primarily happens in extracellular space [33, 40, 41, 64]. Once produced, H₂O₂ enters cancer cell. There it gets utilized by MnP in the catalysis of H₂O₂-driven protein thiol oxidation or S-glutathionylation in a GPx-like fashion [8-11, 21, 23, 36, 67, 89]. In turn, pro-apoptotic pathways might be activated. Along with high tumor H₂O₂ levels and high ability of MnP to catalyze ascorbate oxidation, ~7-fold higher accumulation of MnP in tumor vs normal tissue as well as preferential tumor accumulation of ascorbate [88, 90-92], dictate differential impact of MnP/Asc treatment seen elsewhere: suppression of tumor growth, while healing of normal tissue injury inflicted during cancer chemo- or radiotherapy [1, 2, 8]. The differential effects upon tumor and normal tissue agree well with the data from cellular studies where cytotoxicity was seen with different cancer but not normal cells (this work and [8, 47]). In the presence of exogenous H₂O₂, tumor protection by MnP has never been observed [1, 2, 8, 9, 33, 45-47]. Our data on H₂O₂-related chemistry and biology of MnPs (this work and [54, 74]) contribute to the understanding of differential effects of MnPs [1-4, 6-11, 33, 45-47, 64, 93-95].

The ability of MnP to amplify anticancer effect, in the environment where H₂O₂ is produced, carries large therapeutic potential. The production of H₂O₂ via cycling of MnP with ascorbate, and in turn the resulting anticancer effect could be further enhanced with additional source(s) of peroxide such as chemotherapy [48, 93, 94] and radiation [8, 46, 95].

HIGHLIGHTS.

• Mn porphyrins (MnPs) catalyze ascorbate (Asc) oxidation, thereby producing H₂O₂

• MnP-driven H₂O₂ production depends on the Mn^III/Mn^II reduction potential

• The potency of MnP in catalyzing Asc oxidation correlates with the cytotoxicity induced by MnP/Asc

• MnP/Asc is cytotoxic to 3 breast cancer cell lines, but not to non-cancerous breast epithelial cells

• MnP/Asc suppresses tumor growth in a mouse flank model

• MnP accumulates to much higher levels in tumor than normal tissue

Figure 16. Cooperative actions of MnP/Asc/H₂O₂ system result in suppression of tumor growth.

enhanced tumor H₂O₂ levels via MnP/Asc coupling, high rate of MnP-catalyzed ascorbate oxidation, increased tumor MnP and ascorbate accumulation and MnP/H₂O₂-driven oxidation of signaling proteins [8, 9].

ABBREVIATIONS

O₂^•–

superoxide

ONOO⁻

peroxynitrite

γ-EE

γ-glutamylglutamate

MnTBAP³⁻

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

MnTE-2-PyP⁵⁺

AEOL10113, BMX-010, Mn(III) meso-tetrakis(N-ethylpyridinium-2-yl)porphyrin

MnTE-3-PyP⁵⁺

Mn(III) meso-tetrakis(N-ethylpyridinium-3-yl)porphyrin; MnTE-3-PyP⁵⁺

MnTE-4-PyP⁵⁺Mn(III)

meso-tetrakis(N-ethylpyridinium-4-yl)porphyrin

MnTnPen-4-PyP⁵⁺

Mn(III) meso-tetrakis(N-npentylpyridinium-4-yl)porphyrin

MnTnHex-2-PyP⁵⁺

Mn(III) meso-tetrakis(N-n-hexylpyridinium-2-yl)porphyrin

MnTnHex-3-PyP⁵⁺

Mn(III) meso-tetrakis(N-n-hexylpyridinium-3-yl)porphyrin

MnTnHex-4-PyP⁵⁺

Mn(III) mesotetrakis(N-n-hexylpyridinium-4-yl)porphyrin

MnTnOct-2-PyP⁵⁺

Mn(III) meso-tetrakis(N-n-octylpyridinium-2-yl)porphyrin

MnTnOct-3-PyP⁵⁺

Mn(III) meso-tetrakis(N-n-octylpyridinium-3-yl)porphyrin

MnTPhE-2-PyP⁵⁺

Mn(III) meso-tetrakis (N-(2’-phenylethyl)pyridinium-2-yl)porphyrin

MnTnBuOE-2-PyP⁵⁺

Mn(III) mesotetrakis(N-(2’-n-butoxyethyl)pyridinium-2-yl)porphyrin (BMX-001)

MnTnHexOE-2-PyP⁵⁺,

n(III) mesotetrakis(N-(2’-n-hexoxyethyl)pyridinium-2-yl)porphyrin

MnTE-2-PyPhP⁵⁺

Mn(III) meso-tetrakis(phenyl-4’-(N-ethylpyridinium-2-yl))porphyrin

MnTPhE-2-PyP⁵⁺

Mn(III) meso-tetrakis(N-(2′-phenylethyl)-pyridinium-2-yl)porphyrin

DMF

N,N-dimethylformamide

HFBA

heptafluorobutyric acid

R_f

thin-layer chromatographic retention factor that presents the ratio between the solvent and compound path in 1:1:8=KNO_3(sat) H₂O:H₂O:CH₃CN solvent system

E_1/2

half-wave reduction potential

SOD

superoxide dismutase

NHE

normal hydrogen electrode; monodeprotonated form of ascorbic acid, HA⁻ (Asc) is a major species in aqueous solution at pH 7.8

Asc^•

monodeprotonated ascorbyl radical; charges of MnPs are omitted in some Figures for simplicity