SINGLET OXYGEN QUENCHING

Carotenoids "quench" singlet oxygen primarily by a physical mechanism, in which the excess energy of singlet oxygen is transferred to the carotenoid's electron-rich structure. The carotenoid is excited by this added energy into a "triplet" state (³Car*), and then relaxes into its ground state (¹Car) by losing the extra energy as heat. Because this is a physical mechanism (as opposed to a chemical reaction), the carotenoid structure is unchanged, and the carotenoid remains to protect against further singlet oxygen.

The speed or rate of any given chemical reaction can be described by a rate constant, abbreviated k. Simply put, the faster a reaction, the larger is its rate constant. A carotenoid's singlet oxygen-quenching rate (the rate at which the carotenoid reacts with singlet oxygen) can be characterized by a rate constant, usually abbreviated k_q. A carotenoid's k_q is one indication of its potential effectiveness as an antioxidant--the larger the k_q value, the faster the carotenoid reacts with singlet oxygen. Since a carotenoid is not destroyed by physically quenching singlet oxygen, the greater the quenching rate (i. e., the larger the k_q), the more singlet oxygen can be quenched by a given amount of carotenoid. It is important to remember that k_q values vary according to the experimental conditions (e. g., temperature, solvent, type of photosensitizer, method of measuring quenching), and if k_q values are to be compared, the experimental conditions must be identical or very similar.

In a series of publications, one research group compared singlet oxygen-quenching rates for several carotenoids using steady-state measurements of the reactions (Di Mascio et al. 1989; Di Mascio et al. 1990; Di Mascio et al. 1991). Singlet oxygen was generated chemically in homogenous chloroform-ethanol solution; a known amount of the quenching compound was added and the quenching rate constant calculated from the decrease in singlet oxygen-generated infrared chemiluminescence. The authors reported the following rate constants:

Based on the above data, the authors concluded that relative singlet oxygen quenching ability of a given carotenoid is based on the number of conjugated double bonds, and that the terminal ionone ring of beta-carotene (and related carotenoids) has no effect. The carotenoids with 4 or 4' carbonyl (oxo) substituents--astaxanthin and canthaxanthin--had a slightly increased quenching ability over the corresponding alkyl carotenoid (beta-carotene), which the authors attributed to the extension of the conjugated double bond system.

The majority of carotenoids tested above had singlet oxygen-quenching rate constants in homogenous solution in the order of ~10¹⁰ L mol^-1 s^-1, that is to say, the speed of these reactions approaches the limit set by diffusion. Under these experimental chemical conditions, the carotenoids appeared to quench singlet oxygen more effectively than alpha-tocopherol, with the most effective quencher being lycopene (~100-fold more effective than alpha-tocopherol). Biological thiols are even less reactive under these experimental conditions, with the most important cellular thiol, glutathione, having a rate constant only 1/125^th that of alpha-tocopherol.

How relevant to biological activity are such chemical studies as the ones described above? Chemical studies done in homogenous solution (especially of organic solvent) do not closely resemble the situation in a living eukaryotic cell, which is made up of an essentially aqueous interior that is compartmentalized and enclosed by lipid membranes. Attempts to mimic a biological environment for experimental purposes can be done by means of membrane models, typically liposomes (vesicles made up of lipid bilayers). These are also considered chemical systems, but should more closely resemble conditions found in a living cell and thus are probably more relevant for lipid-soluble antioxidant studies. Not all analytical techniques are useable in liposome systems, however. The next step of experimental complexity is to use in vitro systems (typically cultures of mammalian cells grown in monolayers). Somewhat similar are ex vivo systems, where cells are removed from an organism for at least part of the study--an example would be where an animal is dosed with a test compound, after which a blood sample is taken and experiments carried out on the blood cells. Finally, in vivo experiments are carried out on the entire living animal; these are obviously the most complex and challenging experiments to successfully accomplish because of the additional factors that may influence the experiment's outcome.

The relative in vivo importance of a given antioxidant is likely to be based on a combination of factors. For example, tocopherols and thiols may have slower singlet oxygen-quenching rate constants (measured in a chemical experiment such as those described above), but they are naturally present in biological tissues at much higher concentrations than are carotenoids (Di Mascio et al. 1989; Di Mascio et al. 1990; Di Mascio et al. 1991). The solubility of an antioxidant is important-hydrophilic antioxidants may be expected to exert their effects in the aqueous environment of the cell interior and extracellular milieu (e. g, plasma, cerebrospinal fluid), whereas lipophilic antioxidants are probably most important in lipid-rich cell components (e. g., membranes). Other parameters (e. g., presence of other redox-capable molecules or ions, oxygen partial pressure, ionic strength, viscosity, cell structural complexity) probably also affect the in vivo activity. The in vivo importance of an antioxidant will also depend on the reactive species with which it reacts; for example, a lipophilic radical scavenger may have little effect on hydroxyl radical damage of hydrophilic cell components such as DNA.

In biological systems, carotenoids might be expected to exert most of their antioxidant effects in lipid environments, since carotenoids themselves are lipophilic. Such environments are most commonly found in cell membranes, which are made up predominantly of lipids. One study examined the ability of carotenoids and alpha-tocopherol to protect liposomes (in this case, unilammelar or single-bilayer phospholipid vesicles) from singlet-oxygen oxidation (Oshima et al. 1993). Large unilammelar vesicles were prepared from egg yolk phosphatidylcholine (PC) or dimyristoyl PC with carotenoids incorporated. A final concentration of 5 µM of antioxidant was used in all experiments. Singlet oxygen was generated by photoirradiation with methylene blue (water-soluble) or 12-(1-pyrene)dodecanoic acid ("P-12", lipid-soluble) as photosensitizing agents. Carotenoid and phosphatidylcholine hydroperoxide (PC-OOH) levels were measured by HPLC.

Loss of carotenoids and alpha-tocopherol in egg yolk PC liposomes was followed over the course of 48 hours photoirradiation in the absence of any photosensitizer. The order of stability to photoirradiation (based on residual amounts of carotenoids) was found to be: astaxanthin > alpha-tocopherol > (alpha-carotene approximately equal beta-carotene) > lycopene. Lycopene was least light-stable with 50% loss after ~3 hours of irradiation; 50% of astaxanthin was lost after ~34 hours of irradiation. Alpha-tocopherol and the two carotenes were intermediate in stability. In the absence of photosensitizer, less than 20 µM PC-OOH was found after 6 hours irradiation (Oshima et al. 1993). In relation to these results, another research group reported on the stability to photooxygenation of astaxanthin and canthaxanthin in the absence of any photosensitizer (Christopherson et al. 1991). These authors found that, in general, astaxanthin is less sensitive to light than is canthaxanthin (Christopherson et al. 1991).

In the presence of methylene blue and egg yolk PC liposomes, loss of alpha-tocopherol was rapid, with 50% gone after ~30 minutes of irradiation and >90% lost after 1.5 hours of irradiation (Oshima et al. 1993). The loss of beta-carotene was slower, with 50% gone after ~1.5 hours of irradiation, whereas ~80% of astaxanthin still remained after 6 hours of irradiation. In the presence of methylene blue, PC-OOH accumulated linearly with 300 µM found after 3 hours irradiation without added antioxidant. With the addition of alpha-tocopherol, the generation of PC-OOH was delayed by ~1 hour after which the rate of PC-OOH accumulation was identical to that observed in the control; these results are similar to those reported by Terao (1989) for alpha-tocopherol's retarding of lipid peroxidation in a radical-generating system. Astaxanthin and beta-carotene decreased the rate of PC-OOH formation to about the same degree, with 300 mM PC-OOH observed after about ~5.5 hours of irradiation. Astaxanthin's inhibition of methylene blue-catalyzed oxidation was thus apparently less than that of beta-carotene, since equivalent effects were seen for much larger amounts of astaxanthin compared to beta-carotene (Oshima et al. 1993).

In the presence of P-12 and egg yolk PC liposomes, loss of alpha-tocopherol was much slower than with methylene blue as photosensitizer; 50% loss was observed after ~6 hours of irradiation. The rate of loss of beta-carotene was only a little slower. Astaxanthin was once again more resistant to irradiation, with 50% remaining after ~10 hours irradiation and ~20% remaining after 24 hours of irradiation. In the presence of P-12, PC-OOH formation also proceeded at an overall slower rate than that observed for methylene blue catalysis; 100 µM PC-OOH had accumulated after ~17 hours irradiation without added antioxidant. Alpha-tocopherol and astaxanthin retarded the generation of PC-OOH by ~3 hours but did not affect the rate of PC-OOH formation once the reaction was underway. Addition of beta-carotene resulted in the largest decrease in the rate of PC-OOH formation, with ~75 µM PC-OOH observed after 24 hours irradiation. Again, beta-carotene appeared to have the greatest inhibitory activity against P-12-catalyzed PC-OOH formation (Oshima et al. 1993).

Rates of carotenoid loss were compared for astaxanthin and beta-carotene in dimyristoyl PC liposomes with either methylene blue or P-12 as photosensitizer. Again, rates of loss were slower with P-12 catalyzed reactions. In both cases, astaxanthin was more stable than beta-carotene. However, the fact that beta-carotene was more effective as an antioxidant than was astaxanthin in both methylene blue- and P-12-catalyzed formation of PC-OOH, indicates that the site of singlet oxygen generation and the location of the carotenoid in the lipid bilayer strongly affect a carotenoid's overall antioxidant properties (Oshima et al. 1993).

Another study compared carotenoids, alpha-tocopherol, and butylhydroxytoluene (BHT, a commonly used synthetic antioxidant), both in ethanol solution and in phospholipid liposomes as a model for biological membranes (Fukuzawa et al. 1998). Singlet oxygen was generated using two photosensitizers--rose bengal (RB, water soluble) and pyrenedodecanoic acid (PDA, lipid soluble). Two indicators of singlet oxygen-quenching ability were measured--quenching rate constants (k_q) and lipid peroxidation activities ([IC₅₀]^-1, the reciprocal of the concentration required for 50% inhibition of lipid peroxidation; larger values indicate higher antioxidant activity).

In ethanol solution, k_q values varied no more than 2-fold between the carotenoids beta-carotene, astaxanthin, and canthaxanthin, tested with either RB or PDA. In ethanol solution, carotenoid k_q values were 40- to 80-fold higher than that of alpha-tocopherol. These results are similar to the previous findings described above (Di Mascio et al. 1989; Di Mascio et al. 1990; Di Mascio et al. 1991). In ethanol solution, the k_q values of the synthetic antioxidant BHT were about 40-fold lower than that of alpha-tocopherol. The [IC]₅₀^-1 values in ethanol solution correlated well with these k_q values: [IC]₅₀^-1 values for beta-carotene, for astaxanthin or canthaxanthin, and for BHT, were about 6-fold higher, 18-fold higher, and 214-fold lower, respectively, than that of alpha-tocopherol (Fukuzawa et al. 1998).

However, in liposomes this difference in k_q and [IC]₅₀^-1 values was much less pronounced. The apparent k_q and [IC]₅₀^-1 values were revised to take into account the local concentration of the reactants in the liposome. The revised k_q values in liposomes were reduced (relative to k_q values in ethanol solution) on average by 88-fold for alpha-tocopherol and 630- to 1200-fold for carotenoids. These lower values may be due to decreased diffusion rates for the reactants in the lipid environment, as well as the enforced "two-dimensional" space in which membrane-associated molecules exist. In liposomes, the revised k_q values of carotenoids were now only about 6-fold greater than that of alpha-tocopherol (Fukuzawa et al. 1998). These data should be more reflective of kinetic values in biological systems than data from reactions carried out in homogeneous solutions, at least for lipophilic antioxidants. In vivo, hydrophilic antioxidants are much more likely to exist in relatively homogenous solution and have the ability to diffuse in three dimensions (albeit perhaps limited by membrane boundaries), and thus data obtained for hydrophilic antioxidants in homogenous solution probably more closely reflect their kinetic behavior in biological systems.

Alpha-tocopherol's k_q and [IC]₅₀^-1 values in liposomes were both 5-fold greater for singlet oxygen generated with RB than with PDA, indicating that alpha-tocopherol is more active near the polar surface of the liposome. No such effects of photosensitizer were seen for the three carotenoids tested, supporting the idea that the site for singlet oxygen quenching by a carotenoid is in the hydrophobic polyunsaturated chain. These results demonstrate that the singlet oxygen-quenching ability of a given lipid-soluble antioxidant is dependent on various factors, including the antioxidant's concentration and localization in the membrane (Fukuzawa et al. 1998).

Another study examined the singlet oxygen-quenching effect of carotenoids in an in vitro system consisting of human blood lymphoid cells treated with water-soluble solutions of synthetic beta-carotene, astaxanthin, lycopene, or canthaxanthin (Tinkler et al. 1994). The authors estimated about 1% of the cell surfaces was covered by carotenoid, but no quantification of cell-bound carotenoid was presented. Rose bengal and meso-tetra(4-sulfonatophenyl)porphine were used as photosensitizers. All cells treated with carotenoids showed some singlet oxygen quenching ability whereas untreated cells showed less quenching. In an accompanying ex vivo experiment, lymphoid cells were isolated from human volunteers who had been administered a beta-carotene supplement. The authors were unable to quantify the amount of beta-carotene in the cells but attributed the greater singlet oxygen quenching ability of these cells to beta-carotene bound to the cells. In a third series of experiments, the authors treated lymphoid cells with carotenoid, exposed the cells to photosensitizing agents (which generate singlet oxygen and possibly other oxidizing species), and estimated cell viability by a dye-exclusion assay. The following protection factor (ratio of cells destroyed in the absence or presence of carotenoid) values were found:

With either photosensitizer tested, the most effective protector against singlet oxygen-induced cell death was found to be lycopene, followed by astaxanthin, beta-carotene, and canthaxanthin (which conferred little protection to the porphine-generated oxidative damage) (Tinkler et al. 1994). This ranking is in agreement with previously published singlet oxygen-quenching rate constant data (Di Mascio et al. 1989; Di Mascio et al. 1990; Di Mascio et al. 1991; Conn et al. 1991)

RADICAL SCAVENGING

Carotenoids can react with ("scavenge") radicals in several ways. One way is for the radical to obtain its "missing" electron by removing an electron from another molecule. Another way is for the radical to add itself to another molecule in its attempt to pair its single electron, forming an adduct. In either case, the electron-rich character of carotenoids make them attractive to radicals, thus sparing other cell components (lipids, proteins, DNA) from radical damage.

A detailed and elegant study of carotenoid-radical reaction kinetics was reported by Mortensen et al. (1997). Using a high-speed technique known as pulse radiolysis, the reactions of carotenoids with different types of radicals was monitored (Mortensen et al. 1997). The authors reported the following rate constants for radical scavenging by carotenoids, measured in homogenous tert-butanol/water solutions:

These rate constants can be summarizes as follows: mercaptoethanol thiyl radical (HOCH₂CH₂S ^.) > methanesulfonyl radical (CH₃SO₂^.) > glutathione thiyl radical (GS^.) > nitrogen dioxide radical (NO₂^.).

The reactions also followed different mechanisms. It was shown, for example, that the nitrogen dioxide radical (NO₂^.) may obtain an electron from a carotenoid, resulting in the carotenoid forming a cationic (positively charged) radical (Mortensen et al. 1997).

The observed order of reactivity with NO₂^. was:

In the same study, it was shown that carotenoids react in a different manner with thiyl (sulfur) radicals (RS^.). The mercaptoethanol thiyl radical (HOCH₂CH₂S^.) adds itself to the electron-rich carotenoid, forming an adduct with an unpaired electron located on the carotenoid portion (Mortensen et al. 1997).

All 5 carotenoids studied reacted with the mercaptoethanol thiyl radical very rapidly. The observed order of reactivity with RS^. was:

Lutein has one less conjugated double bond than does the other 4 carotenoids, which may account for its somewhat slower reactivity.

Carotenoids react with another thiyl radical, the glutathione thiyl radical (GS^.), by forming an adduct by a similar radical addition process (Mortensen et al. 1997). In this case, the observed order of reactivity was:

With a third thiyl radical studied, the methanesulfonyl thiyl radical (CH3SO2^.), a more complex reaction was observed, with the thiyl radical forming an intermediate adduct, which then decomposes into the carotenoid radical cation (Mortensen et al. 1997):

The observed order of reactivity with CH₃SO₂^. was:

In all of the chemical systems used in the above study, there was a narrow range of reactivity observed for the 5 carotenoids studied, with the most reactive and least reactive carotenoids differing in reactivity rates by no more than a factor of 2.5 (Mortensen et al. 1997).

Another study examined the antioxidant effects of beta-carotene, zeaxanthin, canthaxanthin, and astaxanthin on the prevention of hydroperoxide formation from a fatty acid methyl ester (Terao 1989). The chemical reactions were carried out in non-aqueous (hexane-isopropanol-tetrahydrofuran) solution, using a diazo radical initiator to induce a radical chain reaction. The resulting formation of hydroperoxides was monitored by HPLC analysis of the reaction mixture. With individual carotenoids added at a level of 1.0 mole percent relative to the fatty acid, hydroperoxide formation rates compared to the control (no antioxidant added) were ~2-fold slower for beta-carotene and zeaxanthin and ~3-fold slower for canthaxanthin and astaxanthin, respectively. The author speculated that the presence of the electron-withdrawing 4(4')-carbonyl oxygens was responsible for the greater antioxidant effects of canthaxanthin and astaxanthin. At lower concentrations (0.2 mole percent relative to the fatty acid), canthaxanthin or astaxanthin had little effect on oxidation; under the same conditions, alpha-tocopherol was able to retard the beginning of oxidation for several hours, although it too had no effect on oxidation rates once oxidation began (after the observed disappearance of all alpha-tocopherol). The author also examined the autoxidation of carotenoids initiated by the radical generator, and found that the autoxidation rate of canthaxanthin and astaxanthin was ~2-fold slower than that of beta-carotene and zeaxanthin. Based on the above, the author proposed that 4(4')-oxo carotenoids can serve as more effective antioxidants than beta-carotene in peroxyl radical-dependent lipid peroxidation (Terao 1989).

Another group reported on the relationship between carotenoid structure and reactivity in a comparative study of free-radical mediated oxidation of carotenoids (Woodall et al. 1997). Chemical reactions were carried out in homogenous solution, with radicals generated by either of two methods: (1) a modified Fenton reaction, a non-specific oxidative system that generates the hydroxyl radical HO^. among other oxidants, or (2) by the diazo radical initiator compounds 2,2'-azobis-isobutyronitrile (AIBN) and 2,2'-azobis(2,4'-dimethylvaleronitrile) (AMVN), which undergo unimolecular decomposition to produce peroxyl radicals.

Solutions of individual carotenoids were subjected to oxidation by addition of a solution of Fe⁺³ and hydrogen peroxide, or by addition of solutions of a diazo radical initiator. Carotenoid oxidation was monitored by loss of absorption at the wavelength of maximum absorption of the carotenoid; no assumptions about the order of the reaction(s) were made and thus initial reaction rates were measured (Woodall et al. 1997).

The modified Fenton reaction produces a rapid burst of oxidant generation and a corresponding rapid initial decrease in carotenoid absorbance, which slowed after approximately 1 minute. Apparent initial rates varied about 22-fold between the fastest (lycopene) and slowest (canthaxanthin and astaxanthin) reactions, but there was significant loss of all carotenoids by 5 minutes (Woodall et al. 1997).

The diazo radical initiators AIBN and AMVN decompose at a constant rate (at a given temperature) to give a steady source of peroxyl radicals. Carotenoids reacted under these conditions at an approximately linear rate with time. With AIBN as the initiator, no more than a 2-fold difference was seen between the fastest (lycopene) and slowest (isozeaxanthin) reactions. With AMVN as initiator, a slightly larger range in activities in hexane solutions was seen with a 4-fold difference between the fastest (again, lycopene) and slowest (canthaxanthin) reactions. The authors attributed the slower initial reaction rate of zeaxanthin compared to beta-carotene in this system to the poor solubility of zeaxanthin in hexane. When the reactions were done in a solvent where both carotenoids were soluble, the initial rates were approximately identical (Woodall et al. 1997).

In all cases lycopene was the most reactive compound, followed by beta-carotene. The diketocarotenoids astaxanthin and canthaxanthin were, on average, the least reactive. These results are similar to those previously reported for carotenoid autoxidation with AMVN by Terao (1989), who postulated that the polyene chain of astaxanthin and canthaxanthin were relatively electron deficient due to the electron-withdrawing potential of the 4(4') carbonyls. However, molecular orbital calculations show that the electron distribution of astaxanthin does not differ substantially from beta-carotene except near the 4(4') carbonyls, suggesting that any difference in reactivity attributed to electron distribution must be due to primarily to reactions that occur near these carbons and not along the polyene chain (Woodall et al. 1997). The potential importance of reactions that do not involve the polyene chain is supported by the difference in initial rates between isozeaxanthin and zeaxanthin (which differ only in the position of the hydroxyl groups on the terminal rings). An example of reactions that occur at the terminal ring would be alkoxy substitution at the reactive allylic carbons (e. g., the 4 or 4' carbons found in beta-carotene or zeaxanthin but absent in astaxanthin and canthaxanthin), for which mass spectral evidence was presented (Woodall et al. 1997)

Again, the results of chemical studies such as those described above cannot be extrapolated to biological systems without consideration of conditions found in living cells. A number of studies using more complex experimental systems have been reported.

One such study compared the antioxidant properties of astaxanthin, canthaxanthin, beta-carotene, and alpha-tocopherol in rat liver microsomes as a membrane model (Palozza and Krinksy 1992). Microsomes are vesicles formed from the endoplasmic reticulum (an extensive system of membranes within a cell) when cells are disrupted. Solutions of the individual antioxidants were added to the microsomes, and lipid peroxidation initiated by either of two methods: (1) addition of the water-soluble compound 2,2'-azobis(2-amidinopropane) (AAPH), which like other diazo radical initiator generates peroxyl radicals, or (2) addition of chelated iron (Fe³⁺/adenosine 5'-diphosphate) and a reductant (beta-nicotinamide adenine dinucleotide phosphate, NADPH)--this system generates oxygen radicals. The lipid oxidation products, malondialdehyde or lipid hydroperoxides, were measured spectrophotometrically.

In the AAPH system, both astaxanthin and canthaxanthin (but not beta-carotene) retarded the beginning of malondialdehyde formation, though no effect on the actual oxidation rates was observed once oxidation began (Palozza and Krinksy 1992). Thus, in this membrane model, 4(4')-ketocarotenoids are more efficient peroxyl radical trappers than the corresponding hydrocarbon carotenoid. These results agree with those reported by Terao (1989) for peroxyl radical reactions carried out in homogenous solution. Although astaxanthin and canthaxanthin are more resistant to autoxidation (Terao 1989), this does not explain the greater antioxidant effects seen in this experiment since high levels of beta-carotene still remained (indicating it had not been consumed and thus ineffective).

In this membrane model, astaxanthin, canthaxanthin, and alpha-tocopherol had approximately equal antioxidant activities under AAPH radical initiation (Palozza and Krinksy 1992). This is in contradiction to other authors (Kurashige et al. 1990; Miki 1991) who report that astaxanthin exhibits greater antioxidant activity against lipid peroxidation than does alpha-tocopherol in ex vivo experiments (discussed below); however it should be pointed out that in neither ex vivo study were actual antioxidant concentrations in membranes verified.

In the Fe³⁺/ADP/NADPH system, astaxanthin was shown to inhibit both malondialdehyde and lipide hydroperoxide formation at a level comparable to that of alpha-tocopherol (Palozza and Krinksy 1992). No other carotenoid was compared.

Another study examined the effects of beta-carotene, canthaxanthin, zeaxanthin, astaxanthin, or alpha-tocopherol on peroxyl radical initiated peroxidation of egg yolk phosphatidylcholine (PC) liposomes (Lim et al. 1992). Liposomes were prepared with 5 or 50 µM carotenoid. Liposome suspensions were subjected to oxidation initiated by the diazo radical generators 2,2'-azobis(2,4-dimethylvaleronitrile) (AMVN, lipid-soluble) or 2,2'-azobis(2-amidinopropane)hydrochloride (AAPH, water-soluble), both of which produce peroxyl radicals at a constant rate. At 50 µM, all carotenoids delayed the initiation of phosphatidylcholine hydroperoxide (PC-OOH) formation catalyzed by 5 µM AMVN. The order of delaying PC-OOH formation was: astaxanthin > zeaxanthin > canthaxanthin > beta-carotene. From the data presented it appears that none of the carotenoids had little or no effect on the rate of PC-OOH formation once peroxidation was underway.

A similar experiment was done with 5 µM beta-carotene, astaxanthin, or alpha-tocopherol. With these lower levels, PC-OOH formation initiated by 4 µM AMVN was delayed equally (~3 hours) by astaxanthin and beta-carotene, whereas alpha-tocopherol retarded peroxidation by a much greater rate (~15 hours). Again, no effect by any antioxidant was apparent once peroxidation began, in agreement with findings by Terao (1989) for lipid peroxide formation in homogenous solution. Loss of antioxidants was monitored during peroxidation; both carotenoids were lost at similar rates with ~50% gone after 3 hours, whereas alpha-tocopherol was more stable with ~50% lost after 8 hours. The above data indicate that xanthophylls have greater antioxidant activity than beta-carotene against AMVN-induced peroxidation of liposomes, but this ability to trap lipid peroxyl radicals is still much less than that of alpha-tocopherol.

With 10 mM AAPH, all carotenoids decreased the rate of PC-OOH formation with no clear delay of the reaction (unlike AMVN-catalyzed oxidations). At 50 µM of carotenoid, the reaction rates found were: control (no carotenoid), 100%; canthaxanthin, 82%; beta-carotene, 45%; zeaxanthin, 22%; and astaxanthin, 18%. No data were shown for alpha-tocopherol. The xanthophylls containing hydroxyl groups (zeaxanthin and astaxanthin) appear better able to trap aqueous-phase peroxyl radicals (presumably at the surface of liposomes) than canthaxanthin (with carbonyls but no hydroxyl groups) or beta-carotene (alkane carotenoid). This may be partly due to the orientation of the xanthophyll molecule in the lipid bilayer, or to differences in the carotenoid-radical reactions (particularly reactions at the terminal rings).

Additional experiments were performed to test dietary carotenoid effects on plasma lipid peroxidation. Birds typically accumulate xanthophylls at levels higher than usually seen in human plasma, where alpha-tocopherol is normally the dominant lipid-soluble antioxidant. The chicken was chosen as a known xanthophyll-accumulating animal. Two-day-old male white Leghorn chicks were divided into three groups of six chicks each, and fed a control or supplemented diet for 4 weeks. The basal diet included both 10.6 mg/kg alpha-tocopherol and 13.3 mg/kg lutein/zeaxanthin. Supplemented diets contained in addition either 500 mg/kg beta-carotene or astaxanthin. After 4 weeks of feeding, the chicks were fasted 24 hours, sacrificed, and plasma collected. Carotenoids and alpha-tocopherol were extracted and analyzed by HPLC; these results are as follows:

No significant difference in body weights was observed for the three groups. Chicks on the control and astaxanthin-supplemented diets had approximately equal amounts of total plasma xanthophylls, with astaxanthin apparently decreasing both plasma lutein/zeaxanthin and alpha-tocopherol levels in the astaxanthin-supplemented chicks. Total plasma xanthophylls and alpha-tocopherol were also significantly decreased from control levels in the beta-carotene-supplemented chicks. No difference in plasma PC-OOH levels was seen between the chicks fed the basal or astaxanthin-supplemented diets. Plasma PC-OOH levels in chicks fed the beta-carotene-supplemented diet were significantly higher than in chicks fed the basal diet. Dietary astaxanthin decreased the level of plasma alpha-tocopherol, but apparently compensated for this in terms of plasma lipid resistance to oxidation since PC-OOH levels did not differ between astaxanthin-supplemented and control chicks.

Plasma from chicks fed the basal diet was treated with 100 mM AAPH and both PC-OOH and malondialdehyde (MDA) were measured as indicators of lipid peroxidation. Alpha- tocopherol was completely depleted during the first hour of incubation. Xanthophylls were almost completely depleted by 2 hours of incubation, after which an increase in both PC-OOH and MDA rates of formation was observed.

The 4(4')-diketocarotenoid, astaxanthin, was tested for antioxidant properties in a tissue culture in vitro model (Lawlor and O'Brien 1995). Primary cell cultures were prepared from fibroblast cells harvested from chicken embryos. The herbicide paraquat (methyl viologen), which generates a wide variety of reactive oxygen species, was used to induce oxidative stress; in this cell culture system, 0.25 mM paraquat upregulates superoxide dismutase (SOD) and catalase (CAT), and downregulates glutathione peroxidase (GSH-Px). Lactate dehydrogenase (LDH) release expressed as a percent of the total LDH released from Triton-treated cells is used as an estimate of cytotoxicity. Cell cultures were treated in either of two ways. In the first, cells were incubated for 18 hours with or without paraquat and astaxanthin (from 0 to 10 nM). In the second, cells were incubated continuously from the initial cell seeding in medium containing a fixed amount of astaxanthin (from 0 to 10 nM); 18 hours prior to the end of the experiment, the astaxanthin-containing medium was removed and replaced with medium with or without paraquat.

Using the first treatment method, paraquat induced SOD and CAT enzyme activity by about 2-fold over that of controls, and depressed GSH-Px enzyme activity to about half that of the control. Paraquat exposure nearly doubled LDH release in cells not treated with astaxanthin. Treatment with astaxanthin resulted in a less pronounced but still significant induction of SOD and CAT in paraquat-treated cells over that of the controls, and had no effect on LDH release. The presence of astaxanthin at all concentrations tested maintained GSH-Px expression in paraquat-treated cells at the same level at that of the control (Lawlor and O'Brien 1995).

Using the second treatment method, astaxanthin at all concentrations tested maintained SOD and CAT enzyme activities in paraquat-treated cells at or below the level of controls. GSH-Px activity was also maintained at the level of the control by all concentrations of astaxanthin tested. LDH release increased slightly in paraquat-treated cells over that of the control, but decreased progressively to levels below the control with increasing amounts of astaxanthin (Lawlor and O'Brien 1995).

Astaxanthin uptake by the cultured cells was determined by HPLC analysis of an organic solvent extract of the cells. The nanomolar concentrations of astaxanthin with which the cells were treated in these studies was too small to be detected by the authors' HPLC method, so another experiment was conducted with higher (1.0 to 10.0 µM) levels of astaxanthin. Astaxanthin content of cells increased with increasing astaxanthin concentration in the medium, indicating that cells took up the carotenoid (Lawlor and O'Brien 1995).

The authors previously demonstrated that both alpha-tocopherol and beta-carotene had antioxidant properties in this cell culture system. Astaxanthin was found to be superior to alpha-tocopherol and beta-carotene in protecting against paraquat-induced oxidative stress in this in vitro model (Lawlor and O'Brien 1995).

An animal model was used to determine the effect of beta-carotene, astaxanthin, and lycopene on UV-induced carcinogenesis (Black 1998). Adult female SKh-Hr-1 hairless mice (a mouse strain susceptible to UV irradiation-induced skin tumors) were fed semi-synthetic diets containing 0.07% (700 ppm) by weight of the respective carotenoids. This level of carotenoid was chosen as it was the lowest concentration at which beta-carotene had demonstrated significant protection against UV-induced skin tumor development, and thus should allow discrimination between beneficial and detrimental effects of carotenoid supplementation. This level (700 ppm) is equivalent to a consumption that is ~160-fold higher per unit body weight in the mouse than those employed in the clinical trials (50 mg day for a 75-kg human). The mice were not allowed food ad libitum but instead each animal received 4 g of the appropriate diet daily.

After a 2-week pre-feeding period, animals received daily UV irradiation at a level about 80% of that required to produce erythema (abnormal flushing of the skin), 5 days per week for 11 weeks. Animals were identified individually by abdominal tattoo, and evaluated weekly for the appearance of tumors, with a 1-mm diameter lesion as an endpoint. At the end of the study (week 28, 17 weeks after cessation of UV treatments), animals were sacrificed. Epidermal carotenoid levels were determined by spectrophotometric analysis of extracts of excised epidermis. Results are given in the following table (Black 1998).

Long term feeding of any of the carotenoids had no effect on mortality or final body weights. Liver enlargement, chosen as a sign of toxicity, was not observed in any of the groups. Carotenoid levels in epidermis (measured non-specifically by spectrophotometry) were highest for mice supplemented with astaxanthin (~25-fold and ~125-fold higher than for mice supplemented with beta-carotene or lycopene, respectively). This could be attributed to a higher rate of uptake (i. e., higher bioavailability), preferential accumulation in epidermis, lower rate of catabolism, or other factors. The mean level of carotenoid (presumably astaxanthin) in epidermis in astaxanthin-supplemented mice was 125 µg/g (=125 mg/kg) wet weight (Black 1998), three times the highest level reported in wild sockeye salmon (Turujman et al. 1997).

Dietary supplementation with beta-carotene and astaxanthin resulted in a statistically significant exacerbation of UV-induced carcinogenesis. The tumor latent period, expressed as the median time to a cumulative tumor probability of 0.5, was shortened by nearly 2 weeks compared to controls. The number of tumors per animal was also statistically greater in animals on the beta-carotene- and astaxanthin-supplemented diets. Lycopene had no statistically significant effect on either parameter (Black 1998). These results were in contradiction to those published by others (see review by Black & Mathews-Roth 1991; Mathews-Roth and Krinsky 1985), but experimental design (closed-formula vs. semi-synthetic diets, ad libitum vs. fixed-dose feeding, commercial carotenoid beadlets vs. crystalline carotenoids) varied between the studies. This suggests that responses to carotenoid supplementation depend on the presence of and interaction with other dietary factors. The author postulated that the exacerbation of carcinogenesis observed for beta-carotene and astaxanthin might have resulted from unrepaired carotenoid radicals produced under the experimental UV stress (Black 1998).

This report also described studies on carotenoid stability and reaction kinetics. The stability of carotenoids in hexane solution to autoxidation by oxygen bubbled in the dark was measured by monitoring the decrease in absorbance. Stability was found to be in the order: beta-carotene > astaxanthin > lycopene (Black 1998). These data contrast with those published by Terao (1989) who reported that astaxanthin is more resistant than is beta-carotene to autoxidation in the presence of radical initiators.

In the same paper, Black (1998) describes an investigation of carotenoid reaction kinetics and gives some results but neither describes the experiments in detail nor shows actual data. The original work appears to have been published by Böhm et al. (1997). Reaction kinetics were investigated by time-resolved spectroscopy in homogenous non-aqueous solution. Alpha-tocopherol (TOH) in hexane solution was converted by pulse radiolysis to the corresponding radical cation (TOH^.+); the test carotenoid was added and the reaction leading to the carotenoid radical cation (CAR^.+) monitored (reaction 1).

The obtained second-order rate constants for reaction 1 indicate that beta-carotene and lycopene react at about the same rate (~10¹⁰ L mol^-1 s^-1) with the tocopherol radical cation. These rate constants are extremely fast and are close to the diffusion-controlled limit. However, for astaxanthin, "the reverse of reaction … was indicated with a rate constant of
8 x 10⁷ M^-1 s^-1" (Black 1998), which presumably means that the reverse reaction was observed, i. e., that the astaxanthin radical cation was found to oxidize alpha-tocopherol. In the original paper by Böhm et al. (1997), the authors report observing that electron transfer occurs from various carotenoids (lutein, canthaxanthin , beta-carotene, septapreno-beta-carotene, lycopene, 7,7'-dihydro-beta-carotene, and zeaxanthin) to the alpha-tocopherol radical (i. e., these carotenoids can repair alpha-tocopherol) (reaction 1, above). Astaxanthin was the sole exception to this pattern (Böhm et al. 1997).

Other investigators have reported a detailed study of the reaction kinetics of beta-carotene and alpha-tocopherol, where they found no evidence supporting reaction 1, but in fact found the reverse reaction (the oxidation of alpha-tocopherol by the beta-carotene radical cation) to occur (Mortensen et al. 1998).

In this report, a system to specifically produce the phenoxy radical was used. Rates of reaction by this radical with beta-carotene and alpha-tocopherol were comparable (~10⁹ L mol^-1 s^-1). In the presence of a large excess of alpha-tocopherol, most phenoxy radicals are scavenged by alpha-tocopherol, and only a small amount of beta-carotene radical cations are formed; under these conditions, the beta-carotene radical cation has a much shorter lifetime, indicating that it is scavenged by alpha-tocopherol (reaction 2).

Reaction 2 was found to be first order with alpha-tocopherol concentration, and the rate constant of this reaction was determined to be 1.7 x 10⁷ L mol^-1 s^-1, in the same order as the rate reported for astaxanthin (Black 1998). In the absence of alpha-tocopherol, the phenoxy radical reacts rapidly with beta-carotene (with concomitant bleaching of beta-carotene) to form the beta-carotene radical cation, which then undergoes bimolecular decay. In the presence of alpha-tocopherol, no bleaching of beta-carotene is observed but only formation of the alpha-tocopherol radical. The complete inhibition of beta-carotene degradation by the phenoxy radical in the presence of alpha-tocopherol indicates that the alpha-tocopherol radicals resulting from reaction with the phenoxy radical do not react with beta-carotene. Further experiments led the authors to conclude that reaction 2 appears to be the rule, at least for beta-carotene, i. e., the beta-carotene radical cation oxidizes alpha-tocopherol with a rate of ~10⁷ L mol^-1 s^-1 in homogenous solution, a rate sufficiently fast to implicate the same reaction in carotenoid-alpha-tocopherol interactions in biological environments (Mortensen et al. 1998). These results are also similar to those reported by Haila (1999), who found that in lipid environments even a minor amount of alpha-tocopherol protected carotenoids from destruction and thus inhibits potential pro-oxidant activity of carotenoids.

Astaxanthin occurs as the major carotenoid in encysted cells of the green alga Haematococcus pluvialis. Astaxanthin biosynthesis is upregulated by environmental stresses, and it has been speculated that the accumulated carotenoid may serve as an antioxidant. One study used the herbicide paraquat (methyl viologen) to induce oxidative stress in either immature (10 pg astaxanthin/cell) or mature (50 pg astaxanthin/cell) H. pluvialis cysts (Kobayashi et al. 1997). Immature cysts were induced by the addition of sodium acetate, mature cysts by the addition of both ferrous sulfate and sodium acetate (addition of ferrous sulfate alone resulted in the algal cells maintaining vegetative growth until stationary phase). Photosynthesis and respiration rates were estimated by measuring the increase in light or decrease in the dark, respectively, of dissolved oxygen concentration. An enzyme assay for superoxide dismutase (SOD) was used to assess antioxidant activity against superoxide (O₂^.-) in both permeabilized whole cells and cell-free extracts. This assay measures the production of nitrite from oxidation of hydroxylamine, and permits discrimination between SOD sensitivity to its specific inhibitors, potassium cyanide and hydrogen peroxide. Pigments (chlorophyll and carotenoids) were extracted and measured; carotenoid composition and concentrations were determined by densitometric scanning of a thin-layer chromatogram. All experiments were done in triplicate, with means (but no standard deviations) reported. Chlorophyll content was about 10 pg/cell in both immature and mature cysts. Addition of paraquat to immature cysts resulted in a decrease in chlorophyll content that was dose-related; a decrease was seen even at the lowest paraquat dose (1 µM) tested. Chlorophyll was not degraded in mature cysts with paraquat added up to 0.1 mM. At the highest dose tested (1 mM) chlorophyll degradation after 3 days was similar in both immature and mature cysts, though the rate of loss was higher in immature cysts. In mature cysts, astaxanthin did not degrade at up to 0.1 mM paraquat; after 3 days exposure to 1 mM paraquat, astaxanthin decreased to about 50% of the initial levels.

After 3 days exposure to various levels of paraquat, photosynthetic and respiratory rates were measured in immature and mature cysts. In mature cysts, photosynthesis ceased with 0.1 mM paraquat (even though chlorophyll content had not been affected), but respiration was maintained. At 1 mM paraquat, respiration in mature cysts was reduced by more than half that of the control. In immature cysts, 3 days of exposure to paraquat at 1 or 10 µM had resulted in decreased levels of chlorophyll, but no effect on photosynthesis rates was seen at these doses; as in mature cysts, photosynthesis ceased at 0.1 mM paraquat. In immature cysts, respiratory activity was reduced by about half at the lowest paraquat dose (1 µM). Thus mature cysts appear to be more tolerant to paraquat-induced stress on respiration than are immature cysts.

Antioxidant activity was reported (as units of SOD-like activity) in units per 10⁷ cells as well as the customary units per mg protein since cysts were low in protein content. Results are given in the following table:

In vegetative cells, antioxidant activity was similar between whole cells and cell-free extracts. This activity was destroyed by heat (suggesting protein denaturation) and was sensitive to both potassium cyanide and hydrogen peroxide, suggesting that this activity was enzymatic (probably copper/zinc-dependent SOD).

Compared with vegetative cells, immature cysts were low in antioxidant activity (expressed as units per 10⁷ cells) both as cell-free extracts and as whole cells. About two-thirds of this activity was lost in cell-free extracts compared to whole cells. Mature cysts had 3-fold higher antioxidant activity (expressed as units per 10⁷ cells) than vegetative cells when assayed as whole cells; this activity was largely lost when assayed as cell-free extracts This activity was not destroyed by heat, nor was it sensitive to potassium cyanide or hydrogen peroxide, suggesting that the antioxidant activity was not due to an enzyme or protein. This antioxidant activity was directly correlated to astaxanthin content in whole cells. These results suggest that the vegetative cells resist the oxidative stress generated by paraquat by enzymatic means whereas mature cysts resist oxidative stress by accumulating astaxanthin.

Black, H. S. and Mathews-Roth, M. M. (1991) Protective role of butylated hydroxytoluene and certain carotenoids in photocarcinogenesis. Photochem. Photobiol., 53:707-716.

Böhm F., Edge R., Land E. J., McGarvey D. J., and Truscott T. G. (1997) Carotenoids enhance vitamin E antioxidant efficiency. J. Am. Chem. Soc., 119:621-622.

Christophersen, A. G., Jun, H., Jørgensen, K., and Skibsted, L. H. (1991) Photobleaching of astaxanthin and canthaxanthin: quantum-yields dependence of solvent, temperature, and wavelength of irradiation in relation to packageing and storage of carotenoid pigmented salmonoids. Z. Lebensm. Unters. Forsch., 192:433-439.

Conn, P. F., Schalch, W., and Truscott, T. G. (1991) The singlet oxygen and carotenoid interaction. J. Photochem. Photobiol. B, 11:41-47.

Di Mascio, P., Devasagayam, T. P. A., Kaiser, S., and Sies, H. (1990) Carotenoids, tocopherols, and thiols and biological singlet molecular oxygen quenchers. Biochem. Soc. Trans., 18:1054-1056.

Di Mascio, P., Kaiser, S., and Sies, H. (1989) Lycopene as the most efficient biological carotenoid singlet oxygen quencher. Arch. Biochem. Biophys., 274:532-538.

Di Mascio, P., M. E. Murphy, and H. Sies. (1991) Antioxidant defense systems: the role of carotenoids, tocopherols, and thiols. Am. J. Clin. Nutr., 53:194S-200S.

Haila, K. (1999) Effects of carotenoids and carotenoid-tocopherol interaction on lipid oxidation in vitro. Dissertation, EKT series 1165. University of Helsinki, Finland. Electronic versions available at: http://ethesis.helsinki.fi/julkaisut/maa/skemi/vk/haila/

Kobayashi, M., Kakizono, T., Hishio, N., Nagai, S., Kurimura, Y., and Tsuji, Y. (1997) Antioxidant role of astaxanthin in the green alga Haematococcus pluvialis. Appl. Microbiol. Biotechnol., 48:351-356.

Fukuzawa, K., Y. Inokami, A. Tokumura, J. Terao, and A. Suzuki. (1998) Rate constants for quenching singlet oxygen and activities for inhibiting lipid peroxidation of carotenoids and alpha-tocopherol in liposomes. Lipids, 33:751-756.

Kurashige, M., Okimasu, E., Inoue, M., and Utsumi, K. (1990) Inhibition of oxidative injury of biological membranes by astaxanthin. Physiol. Chem. Pys. & Med. NMR, 22:27-38.

Lawlor, S. M. and O'Brien, N. M. (1995) Astaxanthin: antioxidant effects in chicken embryo fibroblasts. Nutr. Res., 15:1695-1704.

Lim, B. P., A. Nagao, J. Terao, K. Tanaka, T. Suzuki, and K. Takama. (1992) Antioxidant activity of xanthophylls on peroxyl radical-mediated phospholipid peroxidation. Biochim. Biophys. Acta, 1126:178-184.

Mathews-Roth, M.M., and N.I. Krinsky. (1985) Carotenoid dose level and protection against UV-B induced skin tumors. Photochem. Photobiol., 42:35-38.

Miki, W. (1991) Biological functions and activities of animal carotenoids. Pure Appl. Chem., 63(1):141-146.

Mortensen, A., Skibsted, L. H., Sampson, J., Rice-Evans, C., and Everett, S. A. (1997) Comparative mechanisms and rates of free radical scavenging by carotenoid antioxidants. FEBS Letters, 418:91-97.

Mortensen, A., Skibsted, L. H., Willnow, A., and Everett, S. A. (1998) Re-appraisal of the tocopheroxyl radical reaction with beta-carotene: evidence for oxidation of vitamin E by the b-carotene radical cation. Free Rad. Res., 28:69-80.

Oshima, S., Ojima, F., Sakamoto, H., Ishiguro, Y., and Terao, J. (1993) Inhibitory effect of b-carotene and astaxanthin on photosensitized oxidation of phospholipid bilayers. J. Nutr. Sci. Vitaminol., 39:607-615.

Palozza, P. and Krinsky, N. I. (1992) Astaxanthin and canthaxanthin are potent antioxidants in a membrane model. Arch. Biochem. Biophys., 297:291-295

Terao, J. (1989) Antioxidant activity of b-carotene-related carotenoids in solution. Lipids, 24: 659-661.

Tinkler, J. H., Böhm, F., Schalch, W., and Truscott, T. G. (1994) Dietary carotenoids protect human cells from damage. J. Photochem. Photobiol. B, 26:283-285.

Turujman, S. A., Wamer, W. G., Wei, R. R., and Albert, R. H. (1997) Rapid liquid chromatographic method to distinguish wild salmon from aquacultured salmon fed synthetic astaxanthin. J. AOAC Int., 80(3):622-632.

Woodall, A. A., S. W. Lee, R. J. Weesie, M. J. Jackson, and G. Britton. (1997) Oxidation of carotenoids by free radicals: relationship between structure and reactivity. Biochim. Biophys. Acta, 1336:33-42.