Insights Into the Mode of Action of the Anti-Candida Activity of 1,10-Phenanthroline and its Metal Chelates

Metal complexes of malonie acid (metal = Mn(II), Co(II), Ni(II), Cu(II), Zn(II), Ag(I)) were prepared and only the Ag(I) complex inhibited the growth of Candida albicans. Malonate complexes incorporating the chelating 1,10-phenanthroline (1,10-phen) ligand showed a range of activities: good (Mn(II), Cu(II), Ag(I)); moderate (Zn(II)); poor (Co(II), Ni(II)). Metal-free 1,10-phen and Ag(CH3CO2) were also highly active. The metal-free non-chelating ligands 1,7- phenanthroline and 4,7-phenanthroline were inactive and the Cu(II), Mn(II) and Zn(II) complexs of 1,7-phen displayed only marginal activity. Whereas the Cu(II) malonate/1,10-phen complex induces significant cellular oxidative stress the Zn(II) analogue does not.


Introduction
1,10-Phenanthroline (1,10-phen), 2,2'-bipyridine (2,2'-bipy) (Fig. 1) and their substituted derivatives, both in the metal-free state and as ligands coordinated to transition metals, disturb the functioning of a wide variety of biological systems. When the metal-free N,N-chelating bases are found to be bioactive it is usually assumed that the sequestering of trace metals is involved, and that the resulting metal complexes are the actual active species. 2 4 Administering the metal-free bases is known to affect succinic oxidase, triphosphopyridine nucleotide nitrate reductase, zymase 6 and the reduction of fumerate and malate by molecular hydrogen in the presence of washed suspensions of Escherichia coli. The action of liver and intestine phosphomonoesterases on [-glycerophosphate and of intestine pyrophosphotase on pyrophosphate is also considerably accelerated by solutions of the bases. '9 Whilst 1,10-phen 2+ slows cysteine oxidation 2,2'-bipy and the Fe(II), complex [Fe(2,2-bipy)3] activate cathepsin and also catalyse the autoxidation of linoleic acid. '2 Whereas the oxygen uptake of green leaves 3 and the endogeneous respiration of anabaema 14 are inhibited by the bases, both 1,10-phen and 2,2'bipy accelerate the oxygen consumption of homogenised brain brei. Highly potent anthelmintic action has been observed with both bases. 5 Furthermore, a solution of 10-phen suppresses the chemotactic power of"guinea pig" leucocytes without destroying the cells. 6' The in vitro antibacterial action of 1,10-phen has been demonstrated on several species of 2O 3172 bacteria.' Whereas phenanthroline metal complexes can be bacteriostatic and bacteriocidal towards many Gram-positive bacteria they are relatively ineffective against Gram-negative organisms. Similar reactivity trends against Gram-positive and Gram-negative bacteria were found for Cu(II) and Fe(II/III) complexes of oxine. 2 Whereas m-and p-substituted phenanthrolines were less effective than 1,10-phen at preventing fungal growth, 2,9-dimethyl-l,10-phenanthroline (dmphen) was the most potent Insights into the Mode of Action of the Anti-Candida Activity of 1,10-Phenanathroline and its Metal Chelates inhibitor. 22 Recent in vitro studies have shown that although 2,2'-bipy and its metal complexes did not suppress the growth of clinical isolates of Candida species dilute aqueous solutions of 10phen and its Cu(II) and Mn(II) complexes were extremely toxic to the cells. 3'24 Metal complexes of phenanthroline and bipyridine are toxic to mice and rats 25 and to frogs and rabbits26, and their action on enzymes 1'25'27 and on neuromuscular transmission 28 25 unts were lethal when introduced not metabolised in rats and mice.
Very small amo intraperitoneally, intravenously or subcutaneously, but even large oral doses did not cause death. Absorption through the ntestine was slow (this has also been observed on humans33) and, owing to rapid elimination through the kidney, it was not possible to build up the lethal concentration in the blood. The symptoms caused by the complexes, i.e., torpor, tremor, paralysis of the limbs, chronic convulsions, an unaffected heart, and death due to respiratory failure, indicated an attack on the central nervous system. These conclusions were supported by findings that the complexes are potent inhibitors of acetylcholinesterase and that they have a curare-like action on neuromuscular transmission. Normal activity was restored by washing out the complexes.
The in vitro antitumour activities of Cu(II) complexes of 3,5-disubstituted salicylates increased by an order of magnitude by incorporation of a phenanthroline ligand into the complex. 34 These ternary complexes had cytotoxicities comparable to that of the anticancer drug cisplatin, cis-[PtCI(NH3)2]. However, although [Cu(dmphen)(Hdips)2] (Hdips 3,5-diisopropylsalicylate) had potent in vitro cytotoxicity it was inactive when tested on mice bearing tumour, suggesting that it is destroyed by interaction with biofluids and is not stable enough to reach the tumour cells in vivo.
in vitro experiments revealed that metal complexes of 3,5,6,8or 3,4,7,8-tetramethyl-l,10phen (metal Ni(II), Fe(II), Co(II), Cu(II), Zn(II), Cd(II), Mn(II), Ru(II)), the metal-free hydrochlorides and quaternary salts of the bases were bactericidal to veterinary samples of Erysipelothrix rhusiopathiae and Fusiformis nodosus. 35 '36 On the basis of these experiments it was suggested that the complexes could possibly be used to treat erysipelas in pigs, and may also be beneficial in the treatment of other topical infection of bacteria, mycotic or even viral origin. Furthermore, the latter Cu(II) complex was shown to be at least as effective as solutions of formalin and dioxyethyl laural ammonium chloride in the treatment of foot-rot in sheep. While useful clinically as topical antimicrobials, a selection of phenanthroline complexes, when administered parenterally, were found to be chemptherapeutically ineffective against experimental infection of mice with S. aureus, Streptococcus pneumoniae and M. tuberculosis. Also, the complexes did not decrease ulcer growt-h rate in guinea-pigs infected with M. tuberculosis. Extensive microbiological and pharmacological investigations with phenanthrolines and related chelates 3'36 led to clinical studies with the highly stable Ni(II) and Fe(II) complexes of 3,4,7,8-tetramethyl-l,10-phen. These complexes were known to have a wide spectrum of antimicrobial actions and to produce negligible toxicity to skin, subcutaneous tissues and mucous membranes. 36 Preliminary studies had shown that the complexes were effective in controlling infections due to S. aureus, C. albicans and selected dermatophytes found in hospital casualty, dermatology, ear, nose and throat and gynaecology departments and in surgical wards. The Ni(II) complex was equally as effective as hexachlorophene, when used as an immediate preoperative skin preparation, in decreasing the incidence of postoperative staphylococcal wound infection in elective abdominal surgery. Further trials established that the same complex was again as effective as hexachlorophene in the prophylaxis of staphylococcal infection in the newly born and also in 37 patients undergoing elective obstetric or gynaecological surgery.
The complex was also beneficial in controlling secondary infection in adolescents with acne vulgaris of long standing.
The Ni(II) and the Cu(iI) complexes provided rapid relief of symptons in patients with either chronic monilial infection of the nail or trichomonal vaginitis, attributable to C. albicans and Trichophyton vaginalis, respectively. Furthermore, no toxic manifestations were observed with either substance.
Mn(II) complexes of 3,4,7,8-tetramethyl-l,10-phen were used topically to treat patients suffering from a variety of skin conditions, many of whom had chronic dermatological infections due to dermatophytes (e.g. Malassezia furfur, Trichophyton rubrum) or Candida species. 38 The complexes produced a significant decrease in microbial infection in approximately 50% of cases, with infection due to Gram-positive bacteria generally responding much more rapidly and readily Metal Based Drugs Vol. 7, No. 4,2000 to treatment than that due to Gram-negative bacteria. The complexes did not produce noteworthy irritation or sensitization of the underlying dermatosis or dermatomycosis and, furthermore, significant microbial resistance did not develop.

Materials and Methods
All chemicals were reagent grade and used without further purification. Reactions involving silver salts we[e carried out in the dark. Infrared spectra were recorded as KBr discs in the region 4000-400 cmon a Nicolet Impact 400D FT-IR Spectrometer. The spectra of the metal carboxylate complexe contained prQninent VasymOCO and VsymOCO stretching bands at ca. 1600 cm and ca. 1400 cm-, reslectively. Additional characteristic bands were exhibited by the 1,10phen ligand (ca. 855  [N.i(mal)(HxO)z] and [Ag2(mal)] (1) were prepared from Ni(CH3COz)z.4H20 and Ag(CH3CO2) and using water as the reaction solvent.
[Mn(1,7-phen)2(CH3CO23)2 (14). To a solution of 1,7-phenanthroline (1,7-phen) (0.5 g, 2.8 mmol) in ethanol (40 cm was added Mn(CH3CO2)e.4H20 (0.15 g 0.61 mmol) and th resulting yellow solution was refluxed for h The solution was then evaporated down to 5 cm by rotary eva3Poration, and then acetone (20 cm was added. The resulting solution was then evaporated to 3 cm and diethyl ether (7 cm added. This solution was allowed to evaporate in air over a period of 24 h and the resulting brown crystals were filtered off, washed with diethyl ether and ethanol, and then air-dried.
[Cu(l,7-phen)(CH3CO2)2].2H20 (15)  Cells growing in the mid exponential phase (9 h) were collected by centrjfugation (3,000 rpm for 10/nin), washed twice with sterile PBS and suspended in PBS (9 m). To this was added cm of the stock complex solution (complex 4 or 6; 0.2 g in 100 cm of water) to yield working solutions of the test complexes of concentration 200 gg cm. The cell suspension was then incubated at 37 C for h with gentle shaking. Cells were collected by centrifugation (3,Q00 rpm for 10 min), washed three times with PBS and then suspended in ice-cold PBS (5 Cell viability_ studies were conducted by diluting the cells 1,000-fold with sterile distilled water. Serial dilutions were performed so that approximately 100 cells were plated onto individual SDA plates. The plates were incubated at 37 C for 24 h and the number of colonies counted. Cell viability is expressed as % cells surviving after treatment with drug compared with an untreated control. Spheroplasts were prepared as follows. Glass beads were added to test and control cell suspensions and the cells disrupted using a whirlimixer (30 cycles each of 2 min duration). In between cycles cells were placed on ice. Cell wall removal was observed by light microscopy. The spheroplasts were separated from the glass beads by aspiration and collected by centrifqgation (3,000 rpm for 15 min). The resulting spheroplast pellet was resuspended in PBS (10 cm) and samples of this were used separately for proten estrnation, lipid peroxidation and glutathione assays. Protein estimation on the spheroplasts was carried out using the Protein Assay Kit, Sigma P 5656.
Lipid peroxidation. To ascertain if there were differences in lipid peroxidation in the absence and in the presence of a metal complex the formation of lipid peroxiqlCs was indirectly determined spectrophotometrically by a modification of the published method?" Malonaldehyde (MDA) is a secondary product of lipid peroxidation. Thiobarbituric acid reacts with linoleic acid hydroperoxide (a product of lipid peroxidation) in a heated reaction to form MDA and other compounds. Since MDA is not the only product ofperoxidation the assay is not a direct measurem_ent of MDA, although this is used as a standard. A pink colour is formed by this reaction (, 540 nm) and an increased absorbance value is indicative of more extensive lipid peroxidation and oxidative damage.
GSH content alone was determined using the same procedure as that described above for total glutathione with the exception that 2-vinyl-pyridine (10 gl) was added prior to the addition of the glutathione reductase.

Results and Discussion
At a concentration of 20 gg cm 3 malonic acid has previously been reported to have negligible inhibitory effect on C. albicans. 23 In addition, the present study has shown that the simple metal malonates (metal Mn(II), Co(II), Ni(II), Cu(II), Zn(II)) and Ni(CH3CO2)2.4H20 also have negligible activity against the yeast.
It is significant that Ag(CH3CO2) is the only complex not to contain 1,10-phenanthroline but to exhibit activity similar to that of metal-free 1,10-phen and other metal complexes containing the 1,10-phen ligand. Previous studies have shown that the MIC values of simple metal acetates (metal Mn(II), Cu(II) and Co(II)) 44 is >20 gg cm-3. In addition, by studying the anti-Candida activity of a series of complexes (concentration 20 gg cm"3) of general formula [M(1,10phen)n]Xy (M Mn(II), Cu(II), Zn(II); X malonate, acetate, chloride) it was established that the nature of the counter anion X does not influence the inhibitory effect (Table 2).
In contrast to the performance of the highly active metal-free 1,10-phen ligand metal-free 1,7-and 4,7-phenanthroline (Fig. 2) were both inactive and, furthermore, Cu(II), Mn(II) and Zn(II) complexs of 1,7-phen displayed only marginal activity ( Table 2). Although all of the phenanthroline isomers can coordinate to metal centres 1,10-phen is the only ligand capable of actually chelating the metal and forming an extremely stable metal-phen entity in solution. These observations would appear to substantiate the hypothesis that the bioactivity of N,N-chelating bases is attributed to their ability to sequester specific transition metals and that it is the resulting metal chelate complex that are the active species. Table 2.
Anti Complexes were tested at a concentration of 20 l.tg cm "3 of aqueous RPMI medium. Yeast cells were grown for 24 h at 37 C. Results are presented as % cell growth and the effectiveness of the compounds are compared to the growth of the control (no added complex).
The kinetic growth profile of the yeast when it was exposed to sub-inhibitory concentration of [Mn(1,10-phen)2(mal)].2H20 (2) (0.156 l.tg cm"3) is shown in Fig. 3. Control cells and cells treated with complex 2 enter the early exponential phase at the same time (4 h). However, after this stage growth of the Mn-treated cells is significantly reduced (doubling time approximately 600 min) when compared to the control (120 min). This result suggests that the cells have to metabolically active before the administered metal complexes take effect.
The MIC studies revealed two distinct groups of complexes exhibiting extremes of growth inhibition. The Mn(II), Cu(II) and Ag(I) complexes (2, 4 and 7, respectively) were highly toxic, whereas the Co(II), Ni(II) and Zn(II) complexes (3, 5 and 6, respectively) were ineffective. On the basis of these results a representative complex was selected from each group ([Cu(1,10phen)2(mal)].2H20 (4) and [Zn(1,10-phen)2(mal)].2H20 (6)) for oxidative stress studies. Cells were grown to mid-exponential phase and separate samples of these cells (each batch containing approximately 2 x 10 cells) were treated with 4 and 6 (200 _g cm-3) for h at 37 C. The viability of cells treated with 4 (9 _+ 2%) and 6 (80 +_ 9%) highlights the pronounced toxicity of the cop.per z15,46 complex (Table 3). GSH xs known to be an xmportant antloxdant molecule in yeast cells. GSH:GSSG ratios determined for the control (17:1), [Zn(1,10-phen)_(mal)].2HaO (6) (22:1) and [Cu(1,10-phen)a(mal)].2HaO (4) (5:1) clearly demonstrates that the copper complex induces significant cellular oxidative stress. Furthermore, whereas cells treated with the Zn(II) complex 6 showed no elevation in the level of lipid peroxides (compared to the control) those exposed to the Cu(II) complex 4 had a 7-fold increase. In relation to this result it is known that, as well as targeting ergosterol in the fungal cell membrane and causing pore formation, the polvene antifungal amphotericin B also promotes peroxidation of membrane lipids in Candida cells. 47 C. albicans when exposed to a low dose of gamma-radiation is also known to undergo lipid peroxidation. 4 The kinetics of a sub-inhibitory concentration (0.156 Bg cm3) of complex 2 was followed spectrophotometrically (abs. 540 nm). The cells were grown in RPMI aqueous medium and readings were taken every 4 h over a 24 h period. The results presented are a representative of a total of three independent experiments. Viability is expressed_ as the % cellso ..survivin-(mean _.+ se (n 9)) after treatment with drug comparedto an untreated control. Lipid peroxide formation was measured at an absorbance at 540 nm. The level of lipid peroxidati-on was obtained by comparing the absorption of the test drugs compared to the absotion of the control. The results are expressed as the mean of three independent experiments. The ratio of reduced glutathione (GSH) and oxidized glutathione (GSSG) was calculated from the absorbance at 412 nm using the rate of reduction of DTNB as an external standard, and expressed as mean of three independent experiments.
In conclusion, it seems possible that the heightened anti-Candida activity of the bis(1,10-phen) Mn(II) and Cu(II) complexes, in contrast to that of the Ni(II) and Zn(II) analogous, may be attributed to the fact that the metal centers in the former can have variable oxidation states and are thus potential Fenton reagents. Indeed, the ability of redox-active phenanthroline and Insights into the Mode of Action of the Anti-Candida Activity of 1,10-Phenanathroline and its Metal Chelates bipyridine metal complexes to catalyze the oxidative modification of proteins, nucleic acids and lipids has previously been attributed to the interactions of the substrate with H202 and the metal; i.e. to simple Fenton-type chemistry. 48 The inability of the metal centers in the present Ni(II) and Zn(II) 1,10-phen complexes to redox-cycle would render them incapable of assuming the role of a Fenton reagent. Although the tris(1,10-phen)Co(II) complex (3) would be expected to have a readily accessible and reversible Co(III) oxidation state the metal is coordinatively saturated with respect to 1,10-phen and the complex is thus likely to be extremely kinetically inert. The consistently high activity displayed by the three different Ag(I) complexes, Ag(CH3CO2), [Ag2(mal)] (1) and [Ag2(1,10-phen)3(mal)].2H20 (7), indicates that in these instances it is the metal center and not the ligands that dictates complex performance. As Ag(I) is renowned for its ability to bind strongly to S-donor ligands 25 it is likely that the silver complexes will have a high affinity for the sulphur atoms of thiol proteins and may, ultimately, induce protein inactivation.