One of the controversies surrounding the use of Cr(III)-containing nutritional supplements concerns the proposed roles of such supplements as antioxidants that reduce the diabetes-related oxidative stress [93-96], or pro-oxidants that promote the oxidative stress through the formation of ROS [97-99]. Typical reported examples of the both actions of Cr(III) are described below.
In an early report on the antioxidant action of Cr(III) , pre-treatment of mice with CrCl3 (single injection of an aqueous solution, 5mgkg-1) was found to reduce the toxic effects of CCl4 (including lipid peroxidation in the liver), which are thought to be related to the formation of strongly oxidizing •CCl3 radicals. Additions of CrCl3 (1-10 ^M, alone or in combination with 17 ^-estradiol) to the cell culture medium significantly reduced the levels of oxidized lipids and proteins in a human monocyte cell line treated with H2O2 or with excessive glucose concentrations [93, 101]. Synergistic action of grape seed extract and niacin-bound Cr(III) (2.2mgkg-1 for 10 weeks) in reducing the harmful effects of high-fat diets (including high levels of lipid peroxides in the blood plasma) in Syrian hamsters has been described . Significant decreases in the plasma lipid peroxidation levels in type II diabetic patients after a course of Cr(III) supplementation were detected in two randomized, doubleblind, and placebo-controlled human studies [95, 96]. In the first of these studies , a 6-week course of 0.40 mg Cr/day (as tris(2-pyrrolidine-5-carboxylato)chromium(III), a complex similar in chemical properties to [Cr(pic)3]) was used for the patients having fasting blood glucose (FBG) levels of >8mM. In the second study , Cr(III) was administered at 1.0 mg/day for 6 months in the form of Cr(III)-enriched yeast, and this study was the only one to compare the effects of Cr(III) supplementation at various blood glucose levels. The decreases in plasma lipid peroxidation levels, caused by Cr(III) supplementation, were more pronounced in severely hyperglycemic patients (FBG >8.5 mM) than in mildly hyperglycemic patients (FBG = 7.3-8.4 mM), while significant increases in lipid peroxidation were observed in euglycemic (non-diabetic) subjects (FBG = 4.7-5.3 mM) . Notably, none of the studies [95, 96] detected significant changes in blood glucose levels following Cr(III) supplementation.
The ability of Cr(VI) to cause oxidative stress in exposed humans and in animal models is well-documented [1, 97, 102], but similar effects of Cr(III) were little known until recently. Levels of urinary lipid peroxidation in leather tanning workers (exposed to Cr(III) but not Cr(VI)) were significantly (p < 0.01) increased over the controls and only slightly lower than the levels of such peroxidation in stainless steel welders, exposed predominantly to Cr(VI) . Oral administration of high doses of Cr(III) to rats (895mgkg-1, as aqueous CrCl3) caused significant increases in several markers of oxidative stress, including urinary excretion of oxidized lipid metabolites, hepatic lipid peroxidation, and superoxide radical production . Similar but stronger effects were caused by administration of lower doses of Cr(VI) (25mgkg-1 in water) . Intravenous administration of [Cr(pic)3] to rats (0.090 ^mol/day for 60 days) caused significant increases in the levels of urinary lipid oxidation products and 8-hydroxy-2'-deoxyguanosine (8-OHdG, a marker of oxidative DNA damage) . Treatments of cultured murine macrophages with 10-50 ^gmL-1 of either [Cr(pic)3] or niacin-bound Cr(III) (a poorly characterized compound, probably a mixture of polynuclear Cr(III) car-boxylato complexes)  caused increased DNA fragmentation and ROS production, but these effects were significantly more pronounced for [Cr(pic)3] . Treatments of cultured human macrophages with Cr(III) (as CrCl3, up to 250 ppm) caused the formation of oxidized proteins (detected by the presence of carbonyl groups) . Finally, numerous in vitro studies showed the production of strongly oxidizing species, capable of causing DNA and protein damage, in the Cr(III) + H2O2 or Cr(III) + reductant + O2 reaction systems (vide supra and vide infra).
The dual action of Cr(III) as an antioxidant or a pro-oxidant, described above, can be explained based on the redox reactions in Schemes 4 and 5. The reactions of Cr(III) complexes with lipid peroxides (ROOH in Scheme 4) are probably responsible for the abilities of these compounds to reduce the levels of lipid peroxidation [93-96, 100], but these reactions produce other strong oxidants such as Cr(V) species (Scheme 4). These species, as well as the peroxyl (ROO^) radicals formed in the redox cycling reactions of certain Cr(III) complexes (Scheme 5), are probably responsible for the increases in the oxidative stress markers caused by Cr(III) administration [97-99, 103]. Thus, Cr(III) complexes used as nutritional supplements are involved in a delicate balance between the oxidation and the reduction reactions in blood plasma, as shown most clearly by a change from a mild antioxidant effect of a prolonged treatment with Cr(III) in type II diabetic patients to a mild pro-oxidant effect of the same treatment protocol in healthy individuals . The described controversy over the roles of Cr(III) in biological redox reactions does not support the suggestion [95, 105] that the antioxidant action of Cr(III) supplements may be a major reason for their anti-diabetic activities.
Currently, the main concern over Cr(III)-induced DNA damage is focused on the use of [Cr(pic)3] as a nutritional supplement (see Chapters 9 and 10 of this book). Several authors suggested that the deleterious actions of [Cr(pic)3] are entirely due to the picolinato ligand and that the use of alternative Cr(III) compounds, such as trinuclear Cr(III) propionate [58, 106], niacin-bound Cr(III) , or [CrIIIL3] complexes (where LH = histidine or phenylalanine) [107, 108], is safe. However, other studies have shown a genotoxic potential of Cr(III) compounds other than [Cr(pic)3]. For instance, a preconception exposure of male mice to aqueous CrCl3 (1.0mmolkg-1 as a single intraperitoneal injection) did not cause an acute toxicity but led to multiple neoplastic transformations in the progeny . Exposure to Cr(III) (as a mixture of Cr(III) aqua and hydroxo complexes) [63, 65] in leather tannery workers led to a pattern of DNA damage in lymphocytes, including DNA-protein cross-links and an increased incidence of micronuclei, which was similar to genotoxicity markers found in stainless steel welders exposed to Cr(VI)-containing fumes . These results point to a similarity in the mechanisms of DNA damage, promoted by either Cr(III) or Cr(VI) compounds.
Over the last 20 years, many authors [110-113] claimed that Cr(III) is the ultimate genotoxic form of Cr based on the ability of aquated CrCl3 (mainly trans-[CrIIICl2(OH2)4]+) or Cr(NO3)3 ([CrIII(OH2)6]3+)  to bind to isolated DNA (causing changes in DNA conformation), and on the absence of such binding for Cr(VI). Others  found that DNA lesions produced by CrCl3 were non-mutagenic in mammalian cell assays. Indeed, the use of CrCl3 or Cr(NO3)3 as "representative" Cr(III) compounds is the most persistent mistake in studies of biological roles of Cr(III), since it does not take into account their incompatibility with neutral aqueous solutions . Some of the likely transformations of these compounds in biological media are illustrated in Scheme 7a (for a review, see Ref. ).
Dissolution of CrCl36H2O (trans-[CrIIICl2(OH2)4]+, 5 in Scheme 7a) in water leads to rapid deprotonation of 5 (which is a stronger acid than H2O) with the formation of an uncharged species 6, which is then gradually converted into a mixture of water-insoluble polynuclear hydroxo complexes (7 in Scheme 7a) . Biological media contain many types of organic bases, including deprotonated phosphato groups of phospholipids or DNA and carboxylato residues of proteins ((RO)2P(O)O- and RCOO-, respectively, in Scheme 7a), that will react with 5 more readily than H2O. These reactions proceed through an initial rapid formation of ion pairs between the positively charged species 5 and the negatively charged functional groups of biomolecules (8a, b in Scheme 7a), followed by slower formation of covalently bound Cr(III)-biomolecule adducts (9a, b in Scheme 7a) . Binding of Cr(III) to the phospholipids of cellular or nuclear membranes, caused by the additions of freshly prepared aqueous solutions of CrCl3, is probably responsible for the reported "efficient uptake" of this compound by isolated cell nuclei  or by whole cells . Pre-incubation of CrCl3 with cell culture media, which probably leads to the formation of Cr(III) complexes with amino acids and proteins [1, 63], dramatically decreases the cellular binding of Cr(III) . Generally, Cr(III) complexes are very slow to penetrate cell membranes, due to their octahedral structures and kinetic inertness [1, 30]. Nevertheless, a significant accumulation of Cr(III) in cell compartments, including the nuclei, can occur as a result of occupational exposure to Cr(III)  or a long-term excessive use of Cr(III)-containing nutritional supplements .
In a way that is similar to the processes described in Scheme 7a, stable cationic Cr(III) complexes with polypyridyl ligands, such as [Cr(phen)3]3+ and its analogs, strongly bind to isolated DNA in vitro due to the electrostatic attraction of the positively charged complex ions to the negatively charged DNA backbone . Binding of these photo-sensitive Cr(III) complexes makes DNA susceptible to photo-induced oxidation by O2, probably through the formation of singlet oxygen (:O2 in Scheme 1) . These reactions are unlikely to have major implications in the in vivo systems, since: (i) highly
Was this article helpful?