- Email: [email protected]
Anti malondialdehyde-adduct immunological response as a possible marker of successful aging
Experimental Gerontology 38 (2003) 1129–1135 www.elsevier.com/locate/expgero
Anti malondialdehyde-adduct immunological response as a possible marker of successful aging Nicola Traversoa,*, Stefania Patriarcaa, Emanuela Balbisa, Anna Lisa Furfaroa, Damiano Cottalassoa, Maria Adelaide Pronzatoa, Paolo Carlierb, Federica Bottac, Umberto Maria Marinaria, Luigi Fontanac a
Department of Experimental Medicine, Section of General Pathology, University of Genova, Genova, Italy b Immunohaematology Service, University Hospital S. Martino, Genova, Italy c Opera Don Orione, Centre for Medical-Scientific Studies, Genova, Italy Received 7 May 2003; received in revised form 16 July 2003; accepted 17 July 2003
Abstract Contrasting results have been obtained by various researchers about oxidative markers of aging. In this study, a healthy over-90-year-old population was examined for various plasma oxidative biomarkers and compared with a healthy population of blood donors (age range 23 – 66). Plasma malondialdehyde (MDA), evaluated by means of the thiobarbituric acid test, was significantly higher in the over-90-year-old population, confirming the presence of increased lipoperoxidation in old age. The antibody titre against MDA-protein adducts, considered a marker of lipoperoxidative protein damage in vivo, was evaluated in an ELISA test, completely home made and calibrated versus a concentrated pool of human plasma; this antibody titre was significantly higher in the over-90-year-old population. Plasma vitamin E, evaluated in RP-HPLC, was not significantly different between the two groups. Plasma protein-bound carbonyls, a marker of oxidative protein damage, were measured with the 2,4-dinitrophenylhydrazine assay; their level in the over-90-year-old population was lower than in the blood donors. The higher antibody titre against MDA-adducts may result in protection against accumulation of oxidatively damaged proteins by enhancing their removal, and, together with the preserved plasma vitamin E level, it may endow over-90-year-olds with an especially efficient antioxidant profile. The low level of protein carbonyl might reflect the more efficient removal of damaged proteins. q 2003 Elsevier Inc. All rights reserved. Keywords: Aging; Oxidation; Malondialdehyde; Antibody; Protein carbonyls; Vitamin E
1. Introduction Several intra- or extra-cellular processes can generate free radicals, especially reactive oxygen-derived species, a typical example being the mitochondrial respiratory chain (Lenaz et al., 2002). The ubiquity of oxygen and the presence of metal ions in the organisms greatly favour these phenomena, and pathological events can exacerbate them Abbreviations: AU, arbitrary units; CTR, the healthy adult population of blood donors involved in this study; HSA, human serum albumin; MDA, malondialdehyde; OVER-90, the over-90-year-old population involved in this study; PBS, phosphate buffered saline; TBA, thiobarbituric acid; TCA, trichloroacetic acid. * Corresponding author. Present address: Department of Experimental Medicine, Section of General Pathology, University of Genova, via L.B. Alberti 2, 16132 Genova, Italy Tel.: þ39-010-353-88-33; fax: þ 39-010353-88-36. E-mail address: [email protected] (N. Traverso). 0531-5565/$ - see front matter q 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0531-5565(03)00188-8
(Kehrer, 2000). Many authors maintain a fundamental role of free radicals in the pathogenesis of the morphofunctional changes that accompany and determine aging (Harman, 1992; Stadtman, 2001; Sohal, 2002). It is believed that imbalance between radical production and antioxidant barriers can occur and bring about progressive oxidative damage. Any biological structure can suffer oxidative damage, and lipid peroxidation has come under scrutiny as a marker of this damage: not only does it cause membrane disruption, it also produces aldehydic species, such as malondialdehyde (MDA), able to perpetrate further damage by binding to and modifying proteins (Esterbauer et al., 1991). Accumulation of free radical-mediated damage is the basis of the ‘wear and tear’ theory of aging (Harman, 1992). Reactive oxygen-derived species are short-lived and produce complex molecular alterations in vivo, giving rise to a wide range of molecular structures that have not yet
1130
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135
been well elucidated. The search for ideal oxidation biomarkers is still in progress: MDA evaluation, though certainly not perfect, is widely performed in biological research systems to ascertain whether lipoperoxidation has taken place (Draper et al., 1984; Knight et al., 1988; Esterbauer, 1996); protein carbonyl level is a stable and generic signal of protein oxidative damage (Stadtman, 2001). On the basis that non-enzymatic protein modifications, including aldehyde-mediated modifications, render proteins immunogenic (Palinski et al., 1996), the antibody titre against modified proteins has recently been proposed as a biomarker of oxidative damage in organisms (Salonen et al., 1992; Maggi et al., 1994; Lopes-Virella and Virella, 1996; Wu and Wu, 1997; Steinerova et al., 2001). In the present study, we evaluated plasma levels of MDA, protein carbonyls and vitamin E, and the antibody titre against MDA-modified protein in a healthy over-90year-old population compared with a healthy population of blood donors, in order to evaluate their respective lipo- and proteo-peroxidative balance and to look for possible signs of successful aging.
trichloroacetic acid (TCA), 0.375% TBA and 0.25N HCl. The mixture was kept at 100 8C for 150 and then centrifuged. The fluorescence of the supernatant was evaluated at 530 nm ex/552 nm em, and the concentration calculated on a standard curve of MDA sodium salt. This assay should measure the total (free and bound) MDA present in the samples. 2.3. Vitamin E evaluation Determination of vitamin E was performed according to the method of Lang et al. (1986). Briefly, an aliquot of plasma was added to an equal volume of ethanol and then the mixture was extracted with hexane (volume/volume). The hexane layer was dried under nitrogen and the residue re-dissolved in methanol. An aliquot was analysed in HPLC (mBondapak C18 3.9 £ 300 mm column, Waters, Milford, Massachusetts, USA; pure methanol as mobile phase, UV detection at 292 nm). Quantitation was performed with reference chromatograms of standard solution of vitamin E; tocopherol acetate was used as an internal standard. 2.4. Carbonyl evaluation
2. Materials and methods 2.1. Materials Thiobarbituric acid, trichloroacetic acid, vitamin E, tocopherol acetate, o-phenylendiamine dihydrochloride, human and bovine serum albumin, 2,4-dinitrophenylhydrazine, guanidine hydrochloride were from Sigma-Aldrich (St Louis, MO, USA); HCl and H2O2 were from Carlo Erba (Rodano, Milan, Italy); hexane and methanol were from LabScan (Dublin, Ireland); Tween 20 was from Pharmacia Biotech (Uppsala, Sweden). Plasma was obtained by immediate centrifugation of freshly drawn blood samples from 41 healthy inmates of the Paverano Institute (aged 93.5 ^ 2.4 years, range 90 – 100, OVER-90 group) and from 43 blood donors (aged 43.1 ^ 11.4 years, range 23 – 66, CTR group). People included in the OVER-90 group were free from haematological disorders, cardio-circulatory failure, diabetes mellitus, hepatic, endocrine or metabolic alterations and renal failure; moreover, nobody was suffering, at the moment of the study, from any acute diseases; blood donors are considered representative of a healthy adult population, since only healthy people can be selected as blood donors (Italian Law, 15 Jan 1991). 2.2. MDA evaluation Evaluation of plasma concentration of malondiadehyde (MDA) was performed by using thiobarbituric acid (TBA) as a revealing agent (so-called TBA test) (Esterbauer and Cheeseman, 1990). Briefly, plasma aliquots were mixed with a double volume of solution containing 15%
Protein carbonyl content was evaluated by the 2,4dinitrophenylhydrazine assay (Levine et al., 1990). Briefly, an aliquot of plasma containing one milligram of protein was mixed with an equal volume of TCA 20% (i.e. 10% TCA final concentration), in order to precipitate proteins. The pellet was re-suspended in 10 mM 2,4-dinitrophenylhydrazine in 2M HCl and left at room temperature for 60 min. After precipitation in 10% TCA and washing with ethanol/ethyl acetate 1/1 (vol/vol), the pellet was dissolved in 6M guanidine hydrochloride at 37 8C. After centrifugation, the samples were read against a complementary blank without 2,4-dinitrophenylhydrazine at 365 nm. A molar absorption coefficient of 22000 M21 cm21 for hydrazones resulting from the reaction between carbonyl moieties and 2,4-dinitrophenylhydrazine was used for the calculations. Plasma protein content was evaluated by the bicinchoninic acid method (Pierce, Prodotti Gianni, Milano, Italy). 2.5. Anti MDA-HSA antibody titre evaluation Evaluation of anti MDA-HSA (Human Serum Albumin) antibody titre was performed in ELISA. To our knowledge, no commercial kits are available for the determination of anti MDA-HSA antibody titre; so we set up a ‘home-made’ assay entirely in our laboratory. The MDA-HSA antigen was prepared in vitro by incubating HSA (1.5 mg/ml in Phosphate Buffered Saline (PBS) 10 mmol/l pH 7.4) with 20 mmol/l MDA sodium salt for 6 h at 37 8C in the dark (Chen et al., 1992). Unbound molecules of aldehyde were removed by filtration through desalting columns (molecular cut-off 5 kDa; Pharmacia, Uppsala, Sweden). In these conditions, a large fraction of epsilon amino groups of
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135
lysine residues are derivatized by the aldehyde; the presence of MDA-adducts in the modified HSA was revealed by the appearance of the specific fluorescence at 390 nm ex/ 460 nm em (Chiarpotto et al., 1997), evaluated with a Perkin – Elmer LS-5 spectrofluorometer. The ELISA assay was performed in disposable 96-well polystyrene plates (Nunc-Immunoplates MaxiSorp, Nunc, Denmark). Each well was coated with 1 mg of antigen (MDA-HSA) in PBS (10 mmol/l, pH 7.4) for 2 h at 37 8C in the dark. The remaining binding sites were blocked using 0.5% bovine serum albumin in PBS for 1 h at 37 8C. An aliquot of each serum (diluted 1:30 in blocking buffer) was added in duplicate to the wells. After incubation at 37 8C for 2 h, the wells were washed three times with 0.05% Tween 20 in PBS and then a peroxidase-conjugated antibody (Sigma) specific for human IgG (diluted 1:2000 in blocking buffer) was added. After 1 h incubation at 37 8C and extensive washing, the peroxidase activity was developed using o-phenylendiamine dihydrochloride and H2O2 as revealing agents. The absorbance was measured at 490 nm in an automatic microplate reader (Bio-Tek Instruments Inc., Winooski, Vermont, USA). Each serum was assayed at least twice in different plates (Traverso et al., 1998). In order to calculate the antibody titre, each plate was equipped with a calibration curve; as no standard material was available, we used, as a calibrator, a concentrated pool of human sera, previously prepared and stored in aliquots at 2 20 8C, to which the arbitrary value of 1 Arbitrary Unit (AU)/ml was assigned. The calibration of each plate is essential to obtain reliable values of the titres. We decided to evaluate anti MDA-HSA antibody titre in calibrated ELISA plates since this way of operating is widely used in clinical pathology: the majority of antibody titre calculations in clinical pathology are usually performed by immuno-test, by interpolating the result of the sample in a calibration curve, which is carried out for each batch of samples. 2.6. Statistics Statistical analysis included calculations of the mean and standard error of the mean (SEM), t test (or non-parametric test if necessary) and p evaluations. Significance is defined as p , 0:05: Data are expressed as mean ^ SEM.
3. Results Evaluation of plasma MDA showed a significant difference between OVER-90 and CTR: the OVER-90 group had a mean value of 1260 ^ 38 pmoles/ml, while the CTR group had a mean value of 1113 ^ 21 pmoles/ml ðp , 0:01: Fig. 1). The OVER-90 group also showed a higher mean level of anti-MDA-modified protein antibody titre than the CTR group (732 ^ 144 vs 291 ^ 42 mAU/ml; p , 0:01:
1131
Fig. 1. Plasma MDA levels in CTR (healthy population of blood donors) and OVER-90 (over-90-year-old population). Data are expressed as mean ^ SEM. * p , 0:01:
Fig. 2(a)). Analysis of the frequency distribution seems to indicate the existence of a minor subpopulation within the OVER-90 with a higher anti-MDA antibody titre (Fig. 2(b), see the arrow). No significant difference in plasma vitamin E level was observed between the two groups, though the mean value in the OVER-90 group was slightly lower than in the CTR group (27.26 ^ 1.60 vs 30.58 ^ 1.42 nmoles/ml, p . 0:05; not shown). Plasma protein carbonyl levels were significantly lower in the OVER-90 group than in the CTR group (0.68 ^ 0.04 vs 0.90 ^ 0.05 nmol/mg protein, p , 0:01: Fig. 3). MDA levels, anti-MDA antibody titres and protein carbonyl content correlated significantly with age when both the populations examined were considered together (respectively: p , 0:001; p , 0:05; p , 0:01; the last one with negative slope); the correlations were not significant and no trends existed within each group: this indicates that the correlations depend only on the differences between the two groups. Vitamin E plasma levels did not correlate with age. No significant correlations were found with any combination of parameters in either or both of the study populations, except the correlation between carbonyl and MDA plasma levels in the OVER-90 group (p , 0:01; positive slope); this seems to indicate parallel patterns of lipo- and proteo-peroxidation in old age, which we did not observe in the control subjects. Fig. 4 shows the receiver operating characteristic (ROC) curve for the ability of anti-MDA antibody titre to distinguish OVER-90 from CTR individuals: it evaluates sensitivity and specificity of that parameter at the various cut-off values indicated. The curve indicates that when specificity was 75%, sensitivity was 58%, and when specificity was as high as 95%, sensitivity was still 27%: more than one fourth of the OVER-90 population had titre levels over the 95th percentile of the CTR group. The area under the curve, which indicates the test accuracy, resulted to be 0.701.
1132
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135
Fig. 2. (a) Level of anti MDA-protein adduct antibody titre in CTR (healthy population of blood donors) and OVER-90 (over-90-year-old population). Data are expressed as mean ^ SEM in milli-Arbitrary Unit/ml (mAU/ml). * p , 0:01: (b) Relative frequency distribution of the anti-MDA-protein adduct antibody titre in CTR (healthy population of blood donors) and OVER-90 (over-90-year-old population). Bin width: 360; centre of first bin: 0. The arrow indicates a secondary subpopulation within the OVER-90 with higher anti-MDA antibody titre.
4. Discussion Contrasting results have been published in the literature on the relationship between oxidative damage and aging. As far as the variability of published data is concerned, we can suppose that differences might arise from differently selected populations, different methodology, different robustness of methodology, pre-analytical or analytical variability: it should be kept in mind that the evaluation of the oxidative equilibrium in vivo is a delicate and critical procedure. In a population aged 20– 70 years, Kasapoglu and Ozben (2001) observed an age-dependent increase in plasma MDA and protein carbonyls, which indicate oxidative damage to lipids and proteins, respectively; a decrease in sulfhydryl groups was also noted, which can certainly be considered a further signal of oxidative damage. They also reported an increase in erythrocyte superoxide dismutase and catalase activity, while glutathione peroxidase was seen to decrease
Fig. 3. Plasma protein carbonyl levels in CTR (healthy population of blood donors) and OVER-90 (over-90-year-old population). * p , 0:01:
with aging. Vitamin C and E levels are reported as not changing with age. On studying a population aged 6 months to 69 years, Inal et al. (2001) also found increasing plasma MDA levels and catalase activity, but reported decreased superoxide dismutase activity and increased glutathione peroxidase activity with aging. Garibaldi et al. (2001) evidenced no increase in plasma protein carbonyls with age in a healthy population aged 22– 89 years; moreover, they noted age-dependent increases in the plasma levels of some lipophilic antioxidants, such as alpha- and gamma-tocopherol and retinol. Mecocci et al. (2000) observed age-dependent decreases in vitamin C, vitamin E and vitamin A plasma levels, together with age-dependent increases in superoxide dismutase and glutathione peroxidase activity. However,
Fig. 4. ROC curve of anti-MDA antibody titre. The ability to distinguish OVER-90 from CTR individuals is evaluated. The bisector is also shown to help visual perception of the ROC curve pattern. The following cut-off values were chosen: 113, 167.5, 233, 317, 680, 2079 mAU/ml, which corresponded, respectively, to the 5th, the 25th, the 50th, the 75th, the 95th percentile of CTR and to the 95th percentile of OVER-90. The area under the curve is 0.701.
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135
they noted that centenarians showed particularly high levels of vitamin A and E compared with elderly subjects who were less than 100 years old; healthy centenarians could be particularly protected by high levels of lipophilic antioxidants. However, contrasting results have emerged with regard to the global plasma antioxidant status in centenarians, which has been reported as reduced (Pinzani et al., 1997) or increased (Paolisso et al., 1998). Our study indicates a higher level of lipid peroxidation in old people: the increase of plasma MDA level in OVER-90 in comparison with CTR is small (around 10%) but significant. This is consistent with many experimental observations and with the free-radical theory of aging: this hypothesis claims that free radicals, either generated by endogenous mechanisms such as mitochondrial respiration, or exogenously induced e.g. by pollution, would wear and tear the organisms through progressive accumulation of damage, mainly oxidative in nature; such accumulation should derive also from a progressive decrease of antioxidative mechanisms or reduced ability to remove the oxidatively damaged molecules (Harman, 1992, 1998a,b). We would like to underline that the MDA levels we reported in our manuscript for CTR are in the range suggested by Nielsen and co-workers (Nielsen et al., 1997), who performed a very accurate research for the evaluation of the reference value of plasmatic MDA, while other papers report quite different values (Kasapoglu and Ozben, 2001; Inal et al., 2001). Increased plasma MDA levels comparable with those we report are indicated as noteworthy in various prestigious journals (Nielsen et al., 1997; Jain et al., 1998). However, the MDA level found in OVER-90 is certainly not to be expected similar to the MDA concentrations caused by experimental induction of oxidative stress, intoxication or vitamin depletion (Bagchi et al., 1993; Cottalasso et al., 2002; Halsted et al., 2002; Vendemmiale et al., 1989), which are usually conditions much more severe than physiological aging can be reasonably imagined; moreover, depletion of antioxidant vitamin E was not present in the OVER-90 population. If 10% increased MDA levels can be of any pathophysiological consequence is probably a matter of discussion: many authors report in vitro effects of MDA at concentrations far above those detected in vivo (Vohringer et al., 1998; Olynyk et al., 2002; Chen and Yu, 1994; Tesoriere et al., 2002); other authors signal possible MDA-mediated damage even at physiological concentrations (Rodriguez-Martinez et al., 1998; Draper et al., 1986). Finally, it is to be kept in mind that the plasma level of MDA is the result of a steady state between production, metabolism, consumption and excretion, so it can be regarded as a useful marker of this equilibrium rather than as an indication of possible further damage. On the other hand, the preserved level of plasma vitamin E in the OVER-90 group could provide this population with an advantageous antioxidant profile. However, we do not think that the simple evaluation of antioxidant levels has so far shed light on how very old people escape major
1133
age-related diseases and attain successful aging. We suppose that some other characteristics may be critical for successful aging. We focused our attention on the immunological response against oxidatively modified proteins, in particular MDA-modified proteins. Many authors have revealed the presence of anti-oxidatively modified protein antibody titres in various pathological conditions (Mottaran et al., 2002; Orem et al., 2002; Fornasieri et al., 2002) and even in healthy individuals (Steinerova et al., 2001; Vay et al., 2001). However, it should be underlined that we calibrated the evaluation of this titre: to our knowledge, this approach has not been undertaken in any other published study. Calibration, which was performed in each ELISA plate by means of a home-made high-titre human standard (see Section 2), is essential in order to compare different plates; indeed, it allows a much more precise interpretation of the results and a more objective evaluation of the variances. Moreover, it should be borne in mind that the majority of antibody titre calculations in clinical pathology are performed by calibrated immunotests, such as ELISA; we think this is the way of operating to obtain reliable results. Some authors have suggested using the ratio (or the difference) between reactivity against the modified protein and reactivity against the native protein (Bellomo et al., 1995). In our view, this ‘modified/native reactivity’ ratio (or difference), although in theory useful in order to rule out reactivity towards the native proteins, can lead to misinterpretations, since it does not take into consideration the total reactivity against the modified protein; indeed, samples with very different reactivity towards MDA-protein may have similar ‘modified/native reactivity’ ratios (or differences). In this study, we chose to evaluate total reactivity towards MDA-proteins, calibrated versus a home-made standard, because we think it represents a more significant signal of oxidative damage and of the immunological response towards it. We are not aware of any other study about the correlation between age and anti-MDA protein antibodies, in particular as far as over-90-year old people are concerned; however, Vay and co-workers (Vay et al., 2001) reported no influence of age on anti MDA-HSA antibody titre in a healthy population with age range 25– 63. Higher antibody titres to oxidatively modified proteins are often interpreted as the sign that an increased oxidative damage has taken place in the organisms (Salonen et al., 1992; Maggi et al., 1994); this may apply to the OVER-90 population examined. However, the higher anti-MDA protein antibody titre in the OVER-90 population may indicate a different immunological responsiveness to oxidatively modified proteins; evidence exists that the immunological response can play a protective role against non-enzymatically modified proteins by enhancing their immuno-mediated removal (Palinski et al., 1995; Zhou et al.,
1134
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135
2001). The old people examined might therefore have developed an immunological response particularly efficient in removing MDA-modified proteins or, more generally, oxidatively modified proteins; this could have provided the OVER-90 population with protection from the accumulation of modified protein and their deleterious effects; indeed, the level of plasma protein carbonyls, signal of accumulation of oxidatively damaged proteins, was lower in the OVER-90 population, which might reflect higher efficiency in immuno-mediated removal of oxidatively damaged proteins. Special immunological characteristics have been described in centenarians (Sansoni et al., 1993; Franceschi et al., 2000), and a paradoxical increased tendency towards autoimmunity while lymphocytes lose their ability to respond to antigenic stimulation has been noted, even if specific pathologies have not been attributed to these antibodies (Stacy et al., 2002). Since immune defence is a two-edged sword, the prevalence of ‘physiological autoimmunity’ (i.e. immuno-mediated clearance) over pathological autoimmunity (Urban et al., 2002) may provide the OVER-90 population with a selective advantage and lead to successful aging. In the OVER-90 group, the frequency distribution pattern indicates the existence of a minor subpopulation with a higher anti-MDA antibody titre mode; this seems to indicate that some individuals develop a particularly efficient immunological response, and suggests that a high antimodified protein antibody titre may be a sufficient, but not a necessary, condition for successful aging. In order to better understand the meaning of the antiMDA antibody titre as a marker of aging, we evaluated the ability of this test to distinguish OVER-90 from CTR individuals by studying the ROC curve (Fig. 4). It should be noted that, with a cut-off of 317 mAU/ml (corresponding to the 75th percentile of the CTR group), a sensitivity of 58% is obtained, while with a cut-off of 680 mAU/ml (corresponding to the 95th percentile of the CTR group), a sensitivity of 27% is still obtained. The area under the curve is 0.701, which means the test accuracy is barely acceptable in distinguishing between the two groups: the ROC curve does not indicate anti-MDA antibody titre as a powerful means of singling out all OVER-90 individuals, but it does stress the existence of a considerable proportion of OVER90 individuals (more than one fourth) with titre levels over the 95th percentile of the CTR group. This is also evident in the frequency distribution pattern, which shows an OVER90 subpopulation clustering in the secondary peak mode. The development of a particularly efficient immunological response against oxidatively modified proteins may be one of the mechanisms leading to successful aging, and a particularly high anti-MDA antibody titre may be regarded as a marker of the corresponding subpopulation of successfully aged people. A longitudinal study on healthy adults showing high anti-MDA antibody titres should be performed in order to determine whether they are more likely to age safely.
Acknowledgements This work was supported by grants from the University of Genova, MIUR Cofin 2001 #2001064293_005 (Coord. Scient. Naz. Prof. E. Bergamini), and FIRB 2002 (Coord. Scient. Naz. Prof G. Poli). We thank Dr I. Vazzana and Dr B. Tasso (Department of Pharmaceutical Sciences, University of Genova, Italy) for the synthesis of MDA sodium salt.
References Bagchi, D., Bagchi, M., Hassoun, E., Moser, J., Stohs, S.J., 1993. Effects of carbon tetrachloride, menadione, and paraquat on the urinary excretion of malondialdehyde, formaldehyde, acetaldehyde, and acetone in rats. J. Biochem. Toxicol. 8 (2), 101 –106. Bellomo, G., Maggi, E., Poli, M., Agosta, F.G., Bollati, P., Finardi, G., 1995. Autoantibodies against oxidatively modified low-density lipoproteins in NIDDM. Diabetes 44 (1), 60–66. Chen, J.J., Yu, B.P., 1994. Alterations in mitochondrial membrane fluidity by lipid peroxidation products. Free Radic. Biol. Med. 17 (5), 411 –418. Chen, Q., Esterbauer, H., Jurgens, G., 1992. Studies on epitopes on lowdensity lipoprotein modified by 4-hydroxynonenal, Biochemical characterization and determination. Biochem. J. 288, 249–254. Chiarpotto, E., Scavazza, A., Leonarduzzi, G., Camandola, S., Biasi, F., Teggia, P.M., Garavoglia, M., Robecchi, A., Roncari, A., Poli, G., 1997. Oxidative damage and transforming growth factor beta 1 expression in pretumoral and tumoral lesions of human intestine. Free Radic. Biol. Med. 22 (5), 889–894. Cottalasso, D., Domenicotti, C., Traverso, N., Pronzato, M., Nanni, G., 2002. Influence of chronic ethanol consumption on toxic effects of 1,2dichloroethane: glycolipoprotein retention and impairment of dolichol concentration in rat liver microsomes and Golgi apparatus. Toxicology 178 (3), 229–240. Draper, H.H., Polensek, L., Hadley, M., McGirr, L.G., 1984. Urinary malondialdehyde as an indicator of lipid peroxidation in the diet and in the tissues. Lipids 19 (11), 836 –843. Draper, H.H., McGirr, L.G., Hadley, M., 1986. The metabolism of malondialdehyde. Lipids 21 (4), 305–307. Esterbauer, H., 1996. Estimation of peroxidative damage. A critical review. Pathol. Biol. (Paris) 44 (1), 25–28. Esterbauer, H., Cheeseman, K.H., 1990. Determination of aldehydic lipid peroxidation products: malonaldehyde and 4-hydroxynonenal. Methods Enzymol. 186, 407 –421. Esterbauer, H., Schaur, R.J., Zollner, H., 1991. Chemistry and biochemistry of 4-hydroxynonenal, malonaldehyde and related aldehydes. Free Radic. Biol. Med. 11 (1), 81–128. Fornasieri, A., Perugini, C., Seccia, M., Napodano, P., D’Amico, G., Bellomo, G., 2002. Circulating antibodies recognising oxidativelymodified low-density lipoproteins in patients with IgA nephropathy, membranous glomerulonephritis and focal glomerulosclerosis. J. Nephrol. 15 (4), 349–357. Franceschi, C., Bonafe´, M., Valensin, S., 2000. Human immunosenescence: the prevailing of innate immunity, the failing of clonotypic immunity, and the filling of immunological space. Vaccine 18 (16), 1717–1720. Garibaldi, S., Valentini, S., Aragno, I., Pronzato, M.A., Traverso, N., Odetti, P., 2001. Plasma protein oxidation and antioxidant defense during aging. Int. J. Vitam. Nutr. Res. 71 (6), 332 –338. Halsted, C.H., Villanueva, J.A., Devlin, A.M., Niemela, O., Parkkila, S., Garrow, T.A., Wallock, L.M., Shigenaga, M.K., Melnyk, S., James, S.J., 2002. Folate deficiency disturbs hepatic methionine metabolism and promotes liver injury in the ethanol-fed micropig. Proc. Natl. Acad. Sci. USA 99 (15), 10072–10077.
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135 Harman, D., 1992. Free radical theory of aging. Mutat. Res. 275 (3-6), 257–266. Harman, D., 1998a. Aging: phenomena and theories. Ann. N Y Acad. Sci. 854, 1 –7. Harman, D., 1998b. Extending functional life span. Exp. Gerontol. 33 (1-2), 95–112. Inal, M.E., Kanbak, G., Sunal, E., 2001. Antioxidant enzyme activities and malondialdehyde levels related to aging. Clin. Chim. Acta. 305 (1-2), 75–80. Jain, S.K., Krueger, K.S., McVie, R., Jaramillo, J.J., Palmer, M., Smith, T., 1998. Relationship of blood thromboxane-B2 (TxB2) with lipid peroxides and effect of vitamin E and placebo supplementation on TxB2 and lipid peroxide levels in type 1 diabetic patients. Diabetes Care 21 (9), 1511–1516. Kasapoglu, M., Ozben, T., 2001. Alterations of antioxidant enzymes and oxidative stress markers in aging. Exp. Gerontol. 36 (2), 209–220. Kehrer, J.P., 2000. The Haber-Weiss reaction and mechanisms of toxicity. Toxicology 149 (1), 43–50. Knight, J.A., Pieper, R.K., McClellan, L., 1988. Specificity of the thiobarbituric acid reaction: its use in studies of lipid peroxidation. Clin. Chem. 34 (12), 2433–2438. Lang, J.K., Gohil, K., Packer, L., 1986. Simultaneous determination of tocopherols, ubiquinols, and ubiquinones in blood, plasma, tissue homogenates, and subcellular fractions. Anal. Biochem. 157 (1), 106–116. Lenaz, G., Bovina, C., D’Aurelio, M., Fato, R., Formiggini, G., Genova, M.L., Giuliano, G., Pich, M.M., Paolucci, U., Castelli, G.P., Ventura, B., 2002. Role of mitochondria in oxidative stress and aging. Ann. N Y Acad. Sci. 959, 199–213. Levine, R.L., Garland, D., Oliver, C.N., Amici, A., Climent, I., Lenz, A.G., Ahn, B.W., Shaltiel, S., Stadtman, E.R., 1990. Determination of carbonyl content in oxidatively modified proteins. Methods Enzymol. 186, 464 –478. Lopes-Virella, M.F., Virella, G., 1996. Cytokines, modified lipoproteins, and arteriosclerosis in diabetes. Diabetes 45 (3), S40–S44.Suppl. Maggi, E., Chiesa, R., Melissano, G., Castellano, R., Astore, D., Grossi, A., Finardi, G., Bellomo, G., 1994. LDL oxidation in patients with severe carotid atherosclerosis. A study of in vitro and in vivo oxidation markers. Arterioscler. Thromb. 14 (12), 1892–1899. Mecocci, P., Polidori, M.C., Troiano, L., Cherubini, A., Cecchetti, R., Pini, G., Straatman, M., Monti, D., Stahl, W., Sies, H., Franceschi, C., Senin, U., 2000. Plasma antioxidants and longevity: a study on healthy centenarians. Free Radic. Biol. Med. 28 (8), 1243–1248. Mottaran, E., Stewart, S.F., Rolla, R., Vay, D., Cipriani, V., Moretti, M., Vidali, M., Sartori, M., Rigamonti, C., Day, C.P., Albano, E., 2002. Lipid peroxidation contributes to immune reactions associated with alcoholic liver disease. Free Radic. Biol. Med. 32 (1), 38 –45. Nielsen, F., Mikkelsen, B.B., Nielsen, J.B., Andersen, H.R., Grandjean, P., 1997. Plasma malondialdehyde as biomarker for oxidative stress: reference interval and effects of life-style factors. Clin. Chem. 43 (7), 1209–1214. Olynyk, J.K., Khan, N.A., Ramm, G.A., Brown, K.E., O’Neill, R., Britton, R.S., Bacon, B.R., 2002. Aldehydic products of lipid peroxidation do not directly activate rat hepatic stellate cells. J. Gastroenterol. Hepatol. 17 (7), 785–790. Orem, A., Yandi, Y.E., Vanizor, B., Cimsit, G., Uydu, H.A., Malkoc, M., 2002. The evaluation of autoantibodies against oxidatively modified low-density lipoprotein (LDL), susceptibility of LDL to oxidation, serum lipids and lipid hydroperoxide levels, total antioxidant status, antioxidant enzyme activities, and endothelial dysfunction in patients with Behcet’s disease. Clin. Biochem. 35 (3), 217–224.
1135
Palinski, W., Miller, E., Witztum, J.L., 1995. Immunization of low density lipoprotein (LDL) receptor-deficient rabbits with homologous malondialdehyde-modified LDL reduces atherogenesis. Proc. Natl. Acad. Sci. USA 92 (3), 821– 825. Palinski, W., Horkko, S., Miller, E., Steinbrecher, U.P., Powell, H.C., Curtiss, L.K., Witztum, J.L., 1996. Cloning of monoclonal autoantibodies to epitopes of oxidized lipoproteins from apolipoprotein Edeficient mice. Demonstration of epitopes of oxidized low density lipoprotein in human plasma. J. Clin. Invest. 98 (3), 800–814. Paolisso, G., Tagliamonte, M.R., Rizzo, M.R., Manzella, D., Gambardella, A., Varricchio, M., 1998. Oxidative stress and advancing age: results in healthy centenarians. J. Am. Geriatr. Soc. 46 (7), 833 –838. Pinzani, P., Petruzzi, E., Orlando, C., Stefanescu, A., Antonini, M.F., Serio, M., Pazzagli, M., 1997. Reduced serum antioxidant capacity in healthy centenarians. Clin. Chem. 43 (5), 855–856. Rodriguez-Martinez, M.A., Alonso, M.J., Redondo, J., Salaices, M., Marin, J., 1998. Role of lipid peroxidation and the glutathione-dependent antioxidant system in the impairment of endothelium-dependent relaxations with age. Br. J. Pharmacol. 123 (1), 113 –121. Salonen, J.T., Yla-Herttuala, S., Yamamoto, R., Butler, S., Korpela, H., Salonen, R., Nyyssonen, K., Palinski, W., Witztum, J.L., 1992. Autoantibody against oxidised LDL and progression of carotid atherosclerosis. Lancet 339 (8798), 883 –887. Sansoni, P., Cossarizza, A., Brianti, V., Fagnoni, F., Snelli, G., Monti, D., Marcato, A., Passeri, G., Ortolani, C., Forti, E., et al., 1993. Lymphocyte subsets and natural killer cell activity in healthy old people and centenarians. Blood 82 (9), 2767–2773. Sohal, R.S., 2002. Role of oxidative stress and protein oxidation in the aging process. Free Radic. Biol. Med. 33 (1), 37– 44. Stacy, S., Krolick, K.A., Infante, A.J., Kraig, E., 2002. Immunological memory and late onset autoimmunity. Mech. Ageing Dev. 123 (8), 975– 985. Stadtman, E.R., 2001. Protein oxidation in aging and age-related diseases. Ann. N Y Acad. Sci. 928, 22–38. Steinerova, A., Racek, J., Stozicky, F., Zima, T., Fialova, L., Lapin, A., 2001. Antibodies against oxidized LDL—theory and clinical use. Physiol. Res. 50 (2), 131–141. Tesoriere, L., D’Arpa, D., Bufera, D., Pintaudi, A.M., Allegra, M., Livrea, M.A., 2002. Exposure to malondialdehyde induces an early redox unbalance preceding membrane toxicity in human erythrocytes. Free Radic. Res. 36 (1), 89–97. Traverso, N., Menini, S., Cosso, L., Odetti, P., Albano, E., Pronzato, M.A., Marinari, U.M., 1998. Immunological evidence for increased oxidative stress in diabetic rats. Diabetologia 41 (3), 265–270. Urban, L., Bessenyei, B., Marka, M., Semsei, I., 2002. On the role of aging in the etiology of autoimmunity. Gerontology 48 (3), 179– 184. Vay, D., Parodi, M., Rolla, R., Mottaran, E., Vidali, M., Bellomo, G., Albano, E., 2001. Circulating antibodies recognizing malondialdehydemodified proteins in healthy subjects. Free Radic. Biol. Med. 30 (3), 277– 286. Vendemmiale, G., Altomare, E., Grattagliano, I., Albano, O., 1989. Increased plasma levels of glutathione and malondialdehyde after acute ethanol ingestion in humans. J. Hepatol. 9 (3), 359–365. Vohringer, M.L., Becker, T.W., Krieger, G., Jacobi, H., Witte, I., 1998. Synergistic DNA damaging effects of malondialdehyde/Cu(II) in PM2 DNA and in human fibroblasts. Toxicol. Lett. 94 (3), 159–166. Wu, J.T., Wu, L.L., 1997. Autoantibodies against oxidized LDL. A potential marker for atherosclerosis. Clin. Lab. Med. 17 (3), 595–604. Zhou, X., Caligiuri, G., Hamsten, A., Lefvert, A.K., Hansson, G.K., 2001. LDL immunization induces T-cell-dependent antibody formation and protection against atherosclerosis. Arterioscler. Thromb. Vasc. Biol. 21 (1), 108–114.
Anti malondialdehyde-adduct immunological response as a possible marker of successful aging Nicola Traversoa,*, Stefania Patriarcaa, Emanuela Balbisa, Anna Lisa Furfaroa, Damiano Cottalassoa, Maria Adelaide Pronzatoa, Paolo Carlierb, Federica Bottac, Umberto Maria Marinaria, Luigi Fontanac a
Department of Experimental Medicine, Section of General Pathology, University of Genova, Genova, Italy b Immunohaematology Service, University Hospital S. Martino, Genova, Italy c Opera Don Orione, Centre for Medical-Scientific Studies, Genova, Italy Received 7 May 2003; received in revised form 16 July 2003; accepted 17 July 2003
Abstract Contrasting results have been obtained by various researchers about oxidative markers of aging. In this study, a healthy over-90-year-old population was examined for various plasma oxidative biomarkers and compared with a healthy population of blood donors (age range 23 – 66). Plasma malondialdehyde (MDA), evaluated by means of the thiobarbituric acid test, was significantly higher in the over-90-year-old population, confirming the presence of increased lipoperoxidation in old age. The antibody titre against MDA-protein adducts, considered a marker of lipoperoxidative protein damage in vivo, was evaluated in an ELISA test, completely home made and calibrated versus a concentrated pool of human plasma; this antibody titre was significantly higher in the over-90-year-old population. Plasma vitamin E, evaluated in RP-HPLC, was not significantly different between the two groups. Plasma protein-bound carbonyls, a marker of oxidative protein damage, were measured with the 2,4-dinitrophenylhydrazine assay; their level in the over-90-year-old population was lower than in the blood donors. The higher antibody titre against MDA-adducts may result in protection against accumulation of oxidatively damaged proteins by enhancing their removal, and, together with the preserved plasma vitamin E level, it may endow over-90-year-olds with an especially efficient antioxidant profile. The low level of protein carbonyl might reflect the more efficient removal of damaged proteins. q 2003 Elsevier Inc. All rights reserved. Keywords: Aging; Oxidation; Malondialdehyde; Antibody; Protein carbonyls; Vitamin E
1. Introduction Several intra- or extra-cellular processes can generate free radicals, especially reactive oxygen-derived species, a typical example being the mitochondrial respiratory chain (Lenaz et al., 2002). The ubiquity of oxygen and the presence of metal ions in the organisms greatly favour these phenomena, and pathological events can exacerbate them Abbreviations: AU, arbitrary units; CTR, the healthy adult population of blood donors involved in this study; HSA, human serum albumin; MDA, malondialdehyde; OVER-90, the over-90-year-old population involved in this study; PBS, phosphate buffered saline; TBA, thiobarbituric acid; TCA, trichloroacetic acid. * Corresponding author. Present address: Department of Experimental Medicine, Section of General Pathology, University of Genova, via L.B. Alberti 2, 16132 Genova, Italy Tel.: þ39-010-353-88-33; fax: þ 39-010353-88-36. E-mail address: [email protected] (N. Traverso). 0531-5565/$ - see front matter q 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0531-5565(03)00188-8
(Kehrer, 2000). Many authors maintain a fundamental role of free radicals in the pathogenesis of the morphofunctional changes that accompany and determine aging (Harman, 1992; Stadtman, 2001; Sohal, 2002). It is believed that imbalance between radical production and antioxidant barriers can occur and bring about progressive oxidative damage. Any biological structure can suffer oxidative damage, and lipid peroxidation has come under scrutiny as a marker of this damage: not only does it cause membrane disruption, it also produces aldehydic species, such as malondialdehyde (MDA), able to perpetrate further damage by binding to and modifying proteins (Esterbauer et al., 1991). Accumulation of free radical-mediated damage is the basis of the ‘wear and tear’ theory of aging (Harman, 1992). Reactive oxygen-derived species are short-lived and produce complex molecular alterations in vivo, giving rise to a wide range of molecular structures that have not yet
1130
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135
been well elucidated. The search for ideal oxidation biomarkers is still in progress: MDA evaluation, though certainly not perfect, is widely performed in biological research systems to ascertain whether lipoperoxidation has taken place (Draper et al., 1984; Knight et al., 1988; Esterbauer, 1996); protein carbonyl level is a stable and generic signal of protein oxidative damage (Stadtman, 2001). On the basis that non-enzymatic protein modifications, including aldehyde-mediated modifications, render proteins immunogenic (Palinski et al., 1996), the antibody titre against modified proteins has recently been proposed as a biomarker of oxidative damage in organisms (Salonen et al., 1992; Maggi et al., 1994; Lopes-Virella and Virella, 1996; Wu and Wu, 1997; Steinerova et al., 2001). In the present study, we evaluated plasma levels of MDA, protein carbonyls and vitamin E, and the antibody titre against MDA-modified protein in a healthy over-90year-old population compared with a healthy population of blood donors, in order to evaluate their respective lipo- and proteo-peroxidative balance and to look for possible signs of successful aging.
trichloroacetic acid (TCA), 0.375% TBA and 0.25N HCl. The mixture was kept at 100 8C for 150 and then centrifuged. The fluorescence of the supernatant was evaluated at 530 nm ex/552 nm em, and the concentration calculated on a standard curve of MDA sodium salt. This assay should measure the total (free and bound) MDA present in the samples. 2.3. Vitamin E evaluation Determination of vitamin E was performed according to the method of Lang et al. (1986). Briefly, an aliquot of plasma was added to an equal volume of ethanol and then the mixture was extracted with hexane (volume/volume). The hexane layer was dried under nitrogen and the residue re-dissolved in methanol. An aliquot was analysed in HPLC (mBondapak C18 3.9 £ 300 mm column, Waters, Milford, Massachusetts, USA; pure methanol as mobile phase, UV detection at 292 nm). Quantitation was performed with reference chromatograms of standard solution of vitamin E; tocopherol acetate was used as an internal standard. 2.4. Carbonyl evaluation
2. Materials and methods 2.1. Materials Thiobarbituric acid, trichloroacetic acid, vitamin E, tocopherol acetate, o-phenylendiamine dihydrochloride, human and bovine serum albumin, 2,4-dinitrophenylhydrazine, guanidine hydrochloride were from Sigma-Aldrich (St Louis, MO, USA); HCl and H2O2 were from Carlo Erba (Rodano, Milan, Italy); hexane and methanol were from LabScan (Dublin, Ireland); Tween 20 was from Pharmacia Biotech (Uppsala, Sweden). Plasma was obtained by immediate centrifugation of freshly drawn blood samples from 41 healthy inmates of the Paverano Institute (aged 93.5 ^ 2.4 years, range 90 – 100, OVER-90 group) and from 43 blood donors (aged 43.1 ^ 11.4 years, range 23 – 66, CTR group). People included in the OVER-90 group were free from haematological disorders, cardio-circulatory failure, diabetes mellitus, hepatic, endocrine or metabolic alterations and renal failure; moreover, nobody was suffering, at the moment of the study, from any acute diseases; blood donors are considered representative of a healthy adult population, since only healthy people can be selected as blood donors (Italian Law, 15 Jan 1991). 2.2. MDA evaluation Evaluation of plasma concentration of malondiadehyde (MDA) was performed by using thiobarbituric acid (TBA) as a revealing agent (so-called TBA test) (Esterbauer and Cheeseman, 1990). Briefly, plasma aliquots were mixed with a double volume of solution containing 15%
Protein carbonyl content was evaluated by the 2,4dinitrophenylhydrazine assay (Levine et al., 1990). Briefly, an aliquot of plasma containing one milligram of protein was mixed with an equal volume of TCA 20% (i.e. 10% TCA final concentration), in order to precipitate proteins. The pellet was re-suspended in 10 mM 2,4-dinitrophenylhydrazine in 2M HCl and left at room temperature for 60 min. After precipitation in 10% TCA and washing with ethanol/ethyl acetate 1/1 (vol/vol), the pellet was dissolved in 6M guanidine hydrochloride at 37 8C. After centrifugation, the samples were read against a complementary blank without 2,4-dinitrophenylhydrazine at 365 nm. A molar absorption coefficient of 22000 M21 cm21 for hydrazones resulting from the reaction between carbonyl moieties and 2,4-dinitrophenylhydrazine was used for the calculations. Plasma protein content was evaluated by the bicinchoninic acid method (Pierce, Prodotti Gianni, Milano, Italy). 2.5. Anti MDA-HSA antibody titre evaluation Evaluation of anti MDA-HSA (Human Serum Albumin) antibody titre was performed in ELISA. To our knowledge, no commercial kits are available for the determination of anti MDA-HSA antibody titre; so we set up a ‘home-made’ assay entirely in our laboratory. The MDA-HSA antigen was prepared in vitro by incubating HSA (1.5 mg/ml in Phosphate Buffered Saline (PBS) 10 mmol/l pH 7.4) with 20 mmol/l MDA sodium salt for 6 h at 37 8C in the dark (Chen et al., 1992). Unbound molecules of aldehyde were removed by filtration through desalting columns (molecular cut-off 5 kDa; Pharmacia, Uppsala, Sweden). In these conditions, a large fraction of epsilon amino groups of
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135
lysine residues are derivatized by the aldehyde; the presence of MDA-adducts in the modified HSA was revealed by the appearance of the specific fluorescence at 390 nm ex/ 460 nm em (Chiarpotto et al., 1997), evaluated with a Perkin – Elmer LS-5 spectrofluorometer. The ELISA assay was performed in disposable 96-well polystyrene plates (Nunc-Immunoplates MaxiSorp, Nunc, Denmark). Each well was coated with 1 mg of antigen (MDA-HSA) in PBS (10 mmol/l, pH 7.4) for 2 h at 37 8C in the dark. The remaining binding sites were blocked using 0.5% bovine serum albumin in PBS for 1 h at 37 8C. An aliquot of each serum (diluted 1:30 in blocking buffer) was added in duplicate to the wells. After incubation at 37 8C for 2 h, the wells were washed three times with 0.05% Tween 20 in PBS and then a peroxidase-conjugated antibody (Sigma) specific for human IgG (diluted 1:2000 in blocking buffer) was added. After 1 h incubation at 37 8C and extensive washing, the peroxidase activity was developed using o-phenylendiamine dihydrochloride and H2O2 as revealing agents. The absorbance was measured at 490 nm in an automatic microplate reader (Bio-Tek Instruments Inc., Winooski, Vermont, USA). Each serum was assayed at least twice in different plates (Traverso et al., 1998). In order to calculate the antibody titre, each plate was equipped with a calibration curve; as no standard material was available, we used, as a calibrator, a concentrated pool of human sera, previously prepared and stored in aliquots at 2 20 8C, to which the arbitrary value of 1 Arbitrary Unit (AU)/ml was assigned. The calibration of each plate is essential to obtain reliable values of the titres. We decided to evaluate anti MDA-HSA antibody titre in calibrated ELISA plates since this way of operating is widely used in clinical pathology: the majority of antibody titre calculations in clinical pathology are usually performed by immuno-test, by interpolating the result of the sample in a calibration curve, which is carried out for each batch of samples. 2.6. Statistics Statistical analysis included calculations of the mean and standard error of the mean (SEM), t test (or non-parametric test if necessary) and p evaluations. Significance is defined as p , 0:05: Data are expressed as mean ^ SEM.
3. Results Evaluation of plasma MDA showed a significant difference between OVER-90 and CTR: the OVER-90 group had a mean value of 1260 ^ 38 pmoles/ml, while the CTR group had a mean value of 1113 ^ 21 pmoles/ml ðp , 0:01: Fig. 1). The OVER-90 group also showed a higher mean level of anti-MDA-modified protein antibody titre than the CTR group (732 ^ 144 vs 291 ^ 42 mAU/ml; p , 0:01:
1131
Fig. 1. Plasma MDA levels in CTR (healthy population of blood donors) and OVER-90 (over-90-year-old population). Data are expressed as mean ^ SEM. * p , 0:01:
Fig. 2(a)). Analysis of the frequency distribution seems to indicate the existence of a minor subpopulation within the OVER-90 with a higher anti-MDA antibody titre (Fig. 2(b), see the arrow). No significant difference in plasma vitamin E level was observed between the two groups, though the mean value in the OVER-90 group was slightly lower than in the CTR group (27.26 ^ 1.60 vs 30.58 ^ 1.42 nmoles/ml, p . 0:05; not shown). Plasma protein carbonyl levels were significantly lower in the OVER-90 group than in the CTR group (0.68 ^ 0.04 vs 0.90 ^ 0.05 nmol/mg protein, p , 0:01: Fig. 3). MDA levels, anti-MDA antibody titres and protein carbonyl content correlated significantly with age when both the populations examined were considered together (respectively: p , 0:001; p , 0:05; p , 0:01; the last one with negative slope); the correlations were not significant and no trends existed within each group: this indicates that the correlations depend only on the differences between the two groups. Vitamin E plasma levels did not correlate with age. No significant correlations were found with any combination of parameters in either or both of the study populations, except the correlation between carbonyl and MDA plasma levels in the OVER-90 group (p , 0:01; positive slope); this seems to indicate parallel patterns of lipo- and proteo-peroxidation in old age, which we did not observe in the control subjects. Fig. 4 shows the receiver operating characteristic (ROC) curve for the ability of anti-MDA antibody titre to distinguish OVER-90 from CTR individuals: it evaluates sensitivity and specificity of that parameter at the various cut-off values indicated. The curve indicates that when specificity was 75%, sensitivity was 58%, and when specificity was as high as 95%, sensitivity was still 27%: more than one fourth of the OVER-90 population had titre levels over the 95th percentile of the CTR group. The area under the curve, which indicates the test accuracy, resulted to be 0.701.
1132
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135
Fig. 2. (a) Level of anti MDA-protein adduct antibody titre in CTR (healthy population of blood donors) and OVER-90 (over-90-year-old population). Data are expressed as mean ^ SEM in milli-Arbitrary Unit/ml (mAU/ml). * p , 0:01: (b) Relative frequency distribution of the anti-MDA-protein adduct antibody titre in CTR (healthy population of blood donors) and OVER-90 (over-90-year-old population). Bin width: 360; centre of first bin: 0. The arrow indicates a secondary subpopulation within the OVER-90 with higher anti-MDA antibody titre.
4. Discussion Contrasting results have been published in the literature on the relationship between oxidative damage and aging. As far as the variability of published data is concerned, we can suppose that differences might arise from differently selected populations, different methodology, different robustness of methodology, pre-analytical or analytical variability: it should be kept in mind that the evaluation of the oxidative equilibrium in vivo is a delicate and critical procedure. In a population aged 20– 70 years, Kasapoglu and Ozben (2001) observed an age-dependent increase in plasma MDA and protein carbonyls, which indicate oxidative damage to lipids and proteins, respectively; a decrease in sulfhydryl groups was also noted, which can certainly be considered a further signal of oxidative damage. They also reported an increase in erythrocyte superoxide dismutase and catalase activity, while glutathione peroxidase was seen to decrease
Fig. 3. Plasma protein carbonyl levels in CTR (healthy population of blood donors) and OVER-90 (over-90-year-old population). * p , 0:01:
with aging. Vitamin C and E levels are reported as not changing with age. On studying a population aged 6 months to 69 years, Inal et al. (2001) also found increasing plasma MDA levels and catalase activity, but reported decreased superoxide dismutase activity and increased glutathione peroxidase activity with aging. Garibaldi et al. (2001) evidenced no increase in plasma protein carbonyls with age in a healthy population aged 22– 89 years; moreover, they noted age-dependent increases in the plasma levels of some lipophilic antioxidants, such as alpha- and gamma-tocopherol and retinol. Mecocci et al. (2000) observed age-dependent decreases in vitamin C, vitamin E and vitamin A plasma levels, together with age-dependent increases in superoxide dismutase and glutathione peroxidase activity. However,
Fig. 4. ROC curve of anti-MDA antibody titre. The ability to distinguish OVER-90 from CTR individuals is evaluated. The bisector is also shown to help visual perception of the ROC curve pattern. The following cut-off values were chosen: 113, 167.5, 233, 317, 680, 2079 mAU/ml, which corresponded, respectively, to the 5th, the 25th, the 50th, the 75th, the 95th percentile of CTR and to the 95th percentile of OVER-90. The area under the curve is 0.701.
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135
they noted that centenarians showed particularly high levels of vitamin A and E compared with elderly subjects who were less than 100 years old; healthy centenarians could be particularly protected by high levels of lipophilic antioxidants. However, contrasting results have emerged with regard to the global plasma antioxidant status in centenarians, which has been reported as reduced (Pinzani et al., 1997) or increased (Paolisso et al., 1998). Our study indicates a higher level of lipid peroxidation in old people: the increase of plasma MDA level in OVER-90 in comparison with CTR is small (around 10%) but significant. This is consistent with many experimental observations and with the free-radical theory of aging: this hypothesis claims that free radicals, either generated by endogenous mechanisms such as mitochondrial respiration, or exogenously induced e.g. by pollution, would wear and tear the organisms through progressive accumulation of damage, mainly oxidative in nature; such accumulation should derive also from a progressive decrease of antioxidative mechanisms or reduced ability to remove the oxidatively damaged molecules (Harman, 1992, 1998a,b). We would like to underline that the MDA levels we reported in our manuscript for CTR are in the range suggested by Nielsen and co-workers (Nielsen et al., 1997), who performed a very accurate research for the evaluation of the reference value of plasmatic MDA, while other papers report quite different values (Kasapoglu and Ozben, 2001; Inal et al., 2001). Increased plasma MDA levels comparable with those we report are indicated as noteworthy in various prestigious journals (Nielsen et al., 1997; Jain et al., 1998). However, the MDA level found in OVER-90 is certainly not to be expected similar to the MDA concentrations caused by experimental induction of oxidative stress, intoxication or vitamin depletion (Bagchi et al., 1993; Cottalasso et al., 2002; Halsted et al., 2002; Vendemmiale et al., 1989), which are usually conditions much more severe than physiological aging can be reasonably imagined; moreover, depletion of antioxidant vitamin E was not present in the OVER-90 population. If 10% increased MDA levels can be of any pathophysiological consequence is probably a matter of discussion: many authors report in vitro effects of MDA at concentrations far above those detected in vivo (Vohringer et al., 1998; Olynyk et al., 2002; Chen and Yu, 1994; Tesoriere et al., 2002); other authors signal possible MDA-mediated damage even at physiological concentrations (Rodriguez-Martinez et al., 1998; Draper et al., 1986). Finally, it is to be kept in mind that the plasma level of MDA is the result of a steady state between production, metabolism, consumption and excretion, so it can be regarded as a useful marker of this equilibrium rather than as an indication of possible further damage. On the other hand, the preserved level of plasma vitamin E in the OVER-90 group could provide this population with an advantageous antioxidant profile. However, we do not think that the simple evaluation of antioxidant levels has so far shed light on how very old people escape major
1133
age-related diseases and attain successful aging. We suppose that some other characteristics may be critical for successful aging. We focused our attention on the immunological response against oxidatively modified proteins, in particular MDA-modified proteins. Many authors have revealed the presence of anti-oxidatively modified protein antibody titres in various pathological conditions (Mottaran et al., 2002; Orem et al., 2002; Fornasieri et al., 2002) and even in healthy individuals (Steinerova et al., 2001; Vay et al., 2001). However, it should be underlined that we calibrated the evaluation of this titre: to our knowledge, this approach has not been undertaken in any other published study. Calibration, which was performed in each ELISA plate by means of a home-made high-titre human standard (see Section 2), is essential in order to compare different plates; indeed, it allows a much more precise interpretation of the results and a more objective evaluation of the variances. Moreover, it should be borne in mind that the majority of antibody titre calculations in clinical pathology are performed by calibrated immunotests, such as ELISA; we think this is the way of operating to obtain reliable results. Some authors have suggested using the ratio (or the difference) between reactivity against the modified protein and reactivity against the native protein (Bellomo et al., 1995). In our view, this ‘modified/native reactivity’ ratio (or difference), although in theory useful in order to rule out reactivity towards the native proteins, can lead to misinterpretations, since it does not take into consideration the total reactivity against the modified protein; indeed, samples with very different reactivity towards MDA-protein may have similar ‘modified/native reactivity’ ratios (or differences). In this study, we chose to evaluate total reactivity towards MDA-proteins, calibrated versus a home-made standard, because we think it represents a more significant signal of oxidative damage and of the immunological response towards it. We are not aware of any other study about the correlation between age and anti-MDA protein antibodies, in particular as far as over-90-year old people are concerned; however, Vay and co-workers (Vay et al., 2001) reported no influence of age on anti MDA-HSA antibody titre in a healthy population with age range 25– 63. Higher antibody titres to oxidatively modified proteins are often interpreted as the sign that an increased oxidative damage has taken place in the organisms (Salonen et al., 1992; Maggi et al., 1994); this may apply to the OVER-90 population examined. However, the higher anti-MDA protein antibody titre in the OVER-90 population may indicate a different immunological responsiveness to oxidatively modified proteins; evidence exists that the immunological response can play a protective role against non-enzymatically modified proteins by enhancing their immuno-mediated removal (Palinski et al., 1995; Zhou et al.,
1134
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135
2001). The old people examined might therefore have developed an immunological response particularly efficient in removing MDA-modified proteins or, more generally, oxidatively modified proteins; this could have provided the OVER-90 population with protection from the accumulation of modified protein and their deleterious effects; indeed, the level of plasma protein carbonyls, signal of accumulation of oxidatively damaged proteins, was lower in the OVER-90 population, which might reflect higher efficiency in immuno-mediated removal of oxidatively damaged proteins. Special immunological characteristics have been described in centenarians (Sansoni et al., 1993; Franceschi et al., 2000), and a paradoxical increased tendency towards autoimmunity while lymphocytes lose their ability to respond to antigenic stimulation has been noted, even if specific pathologies have not been attributed to these antibodies (Stacy et al., 2002). Since immune defence is a two-edged sword, the prevalence of ‘physiological autoimmunity’ (i.e. immuno-mediated clearance) over pathological autoimmunity (Urban et al., 2002) may provide the OVER-90 population with a selective advantage and lead to successful aging. In the OVER-90 group, the frequency distribution pattern indicates the existence of a minor subpopulation with a higher anti-MDA antibody titre mode; this seems to indicate that some individuals develop a particularly efficient immunological response, and suggests that a high antimodified protein antibody titre may be a sufficient, but not a necessary, condition for successful aging. In order to better understand the meaning of the antiMDA antibody titre as a marker of aging, we evaluated the ability of this test to distinguish OVER-90 from CTR individuals by studying the ROC curve (Fig. 4). It should be noted that, with a cut-off of 317 mAU/ml (corresponding to the 75th percentile of the CTR group), a sensitivity of 58% is obtained, while with a cut-off of 680 mAU/ml (corresponding to the 95th percentile of the CTR group), a sensitivity of 27% is still obtained. The area under the curve is 0.701, which means the test accuracy is barely acceptable in distinguishing between the two groups: the ROC curve does not indicate anti-MDA antibody titre as a powerful means of singling out all OVER-90 individuals, but it does stress the existence of a considerable proportion of OVER90 individuals (more than one fourth) with titre levels over the 95th percentile of the CTR group. This is also evident in the frequency distribution pattern, which shows an OVER90 subpopulation clustering in the secondary peak mode. The development of a particularly efficient immunological response against oxidatively modified proteins may be one of the mechanisms leading to successful aging, and a particularly high anti-MDA antibody titre may be regarded as a marker of the corresponding subpopulation of successfully aged people. A longitudinal study on healthy adults showing high anti-MDA antibody titres should be performed in order to determine whether they are more likely to age safely.
Acknowledgements This work was supported by grants from the University of Genova, MIUR Cofin 2001 #2001064293_005 (Coord. Scient. Naz. Prof. E. Bergamini), and FIRB 2002 (Coord. Scient. Naz. Prof G. Poli). We thank Dr I. Vazzana and Dr B. Tasso (Department of Pharmaceutical Sciences, University of Genova, Italy) for the synthesis of MDA sodium salt.
References Bagchi, D., Bagchi, M., Hassoun, E., Moser, J., Stohs, S.J., 1993. Effects of carbon tetrachloride, menadione, and paraquat on the urinary excretion of malondialdehyde, formaldehyde, acetaldehyde, and acetone in rats. J. Biochem. Toxicol. 8 (2), 101 –106. Bellomo, G., Maggi, E., Poli, M., Agosta, F.G., Bollati, P., Finardi, G., 1995. Autoantibodies against oxidatively modified low-density lipoproteins in NIDDM. Diabetes 44 (1), 60–66. Chen, J.J., Yu, B.P., 1994. Alterations in mitochondrial membrane fluidity by lipid peroxidation products. Free Radic. Biol. Med. 17 (5), 411 –418. Chen, Q., Esterbauer, H., Jurgens, G., 1992. Studies on epitopes on lowdensity lipoprotein modified by 4-hydroxynonenal, Biochemical characterization and determination. Biochem. J. 288, 249–254. Chiarpotto, E., Scavazza, A., Leonarduzzi, G., Camandola, S., Biasi, F., Teggia, P.M., Garavoglia, M., Robecchi, A., Roncari, A., Poli, G., 1997. Oxidative damage and transforming growth factor beta 1 expression in pretumoral and tumoral lesions of human intestine. Free Radic. Biol. Med. 22 (5), 889–894. Cottalasso, D., Domenicotti, C., Traverso, N., Pronzato, M., Nanni, G., 2002. Influence of chronic ethanol consumption on toxic effects of 1,2dichloroethane: glycolipoprotein retention and impairment of dolichol concentration in rat liver microsomes and Golgi apparatus. Toxicology 178 (3), 229–240. Draper, H.H., Polensek, L., Hadley, M., McGirr, L.G., 1984. Urinary malondialdehyde as an indicator of lipid peroxidation in the diet and in the tissues. Lipids 19 (11), 836 –843. Draper, H.H., McGirr, L.G., Hadley, M., 1986. The metabolism of malondialdehyde. Lipids 21 (4), 305–307. Esterbauer, H., 1996. Estimation of peroxidative damage. A critical review. Pathol. Biol. (Paris) 44 (1), 25–28. Esterbauer, H., Cheeseman, K.H., 1990. Determination of aldehydic lipid peroxidation products: malonaldehyde and 4-hydroxynonenal. Methods Enzymol. 186, 407 –421. Esterbauer, H., Schaur, R.J., Zollner, H., 1991. Chemistry and biochemistry of 4-hydroxynonenal, malonaldehyde and related aldehydes. Free Radic. Biol. Med. 11 (1), 81–128. Fornasieri, A., Perugini, C., Seccia, M., Napodano, P., D’Amico, G., Bellomo, G., 2002. Circulating antibodies recognising oxidativelymodified low-density lipoproteins in patients with IgA nephropathy, membranous glomerulonephritis and focal glomerulosclerosis. J. Nephrol. 15 (4), 349–357. Franceschi, C., Bonafe´, M., Valensin, S., 2000. Human immunosenescence: the prevailing of innate immunity, the failing of clonotypic immunity, and the filling of immunological space. Vaccine 18 (16), 1717–1720. Garibaldi, S., Valentini, S., Aragno, I., Pronzato, M.A., Traverso, N., Odetti, P., 2001. Plasma protein oxidation and antioxidant defense during aging. Int. J. Vitam. Nutr. Res. 71 (6), 332 –338. Halsted, C.H., Villanueva, J.A., Devlin, A.M., Niemela, O., Parkkila, S., Garrow, T.A., Wallock, L.M., Shigenaga, M.K., Melnyk, S., James, S.J., 2002. Folate deficiency disturbs hepatic methionine metabolism and promotes liver injury in the ethanol-fed micropig. Proc. Natl. Acad. Sci. USA 99 (15), 10072–10077.
N. Traverso et al. / Experimental Gerontology 38 (2003) 1129–1135 Harman, D., 1992. Free radical theory of aging. Mutat. Res. 275 (3-6), 257–266. Harman, D., 1998a. Aging: phenomena and theories. Ann. N Y Acad. Sci. 854, 1 –7. Harman, D., 1998b. Extending functional life span. Exp. Gerontol. 33 (1-2), 95–112. Inal, M.E., Kanbak, G., Sunal, E., 2001. Antioxidant enzyme activities and malondialdehyde levels related to aging. Clin. Chim. Acta. 305 (1-2), 75–80. Jain, S.K., Krueger, K.S., McVie, R., Jaramillo, J.J., Palmer, M., Smith, T., 1998. Relationship of blood thromboxane-B2 (TxB2) with lipid peroxides and effect of vitamin E and placebo supplementation on TxB2 and lipid peroxide levels in type 1 diabetic patients. Diabetes Care 21 (9), 1511–1516. Kasapoglu, M., Ozben, T., 2001. Alterations of antioxidant enzymes and oxidative stress markers in aging. Exp. Gerontol. 36 (2), 209–220. Kehrer, J.P., 2000. The Haber-Weiss reaction and mechanisms of toxicity. Toxicology 149 (1), 43–50. Knight, J.A., Pieper, R.K., McClellan, L., 1988. Specificity of the thiobarbituric acid reaction: its use in studies of lipid peroxidation. Clin. Chem. 34 (12), 2433–2438. Lang, J.K., Gohil, K., Packer, L., 1986. Simultaneous determination of tocopherols, ubiquinols, and ubiquinones in blood, plasma, tissue homogenates, and subcellular fractions. Anal. Biochem. 157 (1), 106–116. Lenaz, G., Bovina, C., D’Aurelio, M., Fato, R., Formiggini, G., Genova, M.L., Giuliano, G., Pich, M.M., Paolucci, U., Castelli, G.P., Ventura, B., 2002. Role of mitochondria in oxidative stress and aging. Ann. N Y Acad. Sci. 959, 199–213. Levine, R.L., Garland, D., Oliver, C.N., Amici, A., Climent, I., Lenz, A.G., Ahn, B.W., Shaltiel, S., Stadtman, E.R., 1990. Determination of carbonyl content in oxidatively modified proteins. Methods Enzymol. 186, 464 –478. Lopes-Virella, M.F., Virella, G., 1996. Cytokines, modified lipoproteins, and arteriosclerosis in diabetes. Diabetes 45 (3), S40–S44.Suppl. Maggi, E., Chiesa, R., Melissano, G., Castellano, R., Astore, D., Grossi, A., Finardi, G., Bellomo, G., 1994. LDL oxidation in patients with severe carotid atherosclerosis. A study of in vitro and in vivo oxidation markers. Arterioscler. Thromb. 14 (12), 1892–1899. Mecocci, P., Polidori, M.C., Troiano, L., Cherubini, A., Cecchetti, R., Pini, G., Straatman, M., Monti, D., Stahl, W., Sies, H., Franceschi, C., Senin, U., 2000. Plasma antioxidants and longevity: a study on healthy centenarians. Free Radic. Biol. Med. 28 (8), 1243–1248. Mottaran, E., Stewart, S.F., Rolla, R., Vay, D., Cipriani, V., Moretti, M., Vidali, M., Sartori, M., Rigamonti, C., Day, C.P., Albano, E., 2002. Lipid peroxidation contributes to immune reactions associated with alcoholic liver disease. Free Radic. Biol. Med. 32 (1), 38 –45. Nielsen, F., Mikkelsen, B.B., Nielsen, J.B., Andersen, H.R., Grandjean, P., 1997. Plasma malondialdehyde as biomarker for oxidative stress: reference interval and effects of life-style factors. Clin. Chem. 43 (7), 1209–1214. Olynyk, J.K., Khan, N.A., Ramm, G.A., Brown, K.E., O’Neill, R., Britton, R.S., Bacon, B.R., 2002. Aldehydic products of lipid peroxidation do not directly activate rat hepatic stellate cells. J. Gastroenterol. Hepatol. 17 (7), 785–790. Orem, A., Yandi, Y.E., Vanizor, B., Cimsit, G., Uydu, H.A., Malkoc, M., 2002. The evaluation of autoantibodies against oxidatively modified low-density lipoprotein (LDL), susceptibility of LDL to oxidation, serum lipids and lipid hydroperoxide levels, total antioxidant status, antioxidant enzyme activities, and endothelial dysfunction in patients with Behcet’s disease. Clin. Biochem. 35 (3), 217–224.
1135
Palinski, W., Miller, E., Witztum, J.L., 1995. Immunization of low density lipoprotein (LDL) receptor-deficient rabbits with homologous malondialdehyde-modified LDL reduces atherogenesis. Proc. Natl. Acad. Sci. USA 92 (3), 821– 825. Palinski, W., Horkko, S., Miller, E., Steinbrecher, U.P., Powell, H.C., Curtiss, L.K., Witztum, J.L., 1996. Cloning of monoclonal autoantibodies to epitopes of oxidized lipoproteins from apolipoprotein Edeficient mice. Demonstration of epitopes of oxidized low density lipoprotein in human plasma. J. Clin. Invest. 98 (3), 800–814. Paolisso, G., Tagliamonte, M.R., Rizzo, M.R., Manzella, D., Gambardella, A., Varricchio, M., 1998. Oxidative stress and advancing age: results in healthy centenarians. J. Am. Geriatr. Soc. 46 (7), 833 –838. Pinzani, P., Petruzzi, E., Orlando, C., Stefanescu, A., Antonini, M.F., Serio, M., Pazzagli, M., 1997. Reduced serum antioxidant capacity in healthy centenarians. Clin. Chem. 43 (5), 855–856. Rodriguez-Martinez, M.A., Alonso, M.J., Redondo, J., Salaices, M., Marin, J., 1998. Role of lipid peroxidation and the glutathione-dependent antioxidant system in the impairment of endothelium-dependent relaxations with age. Br. J. Pharmacol. 123 (1), 113 –121. Salonen, J.T., Yla-Herttuala, S., Yamamoto, R., Butler, S., Korpela, H., Salonen, R., Nyyssonen, K., Palinski, W., Witztum, J.L., 1992. Autoantibody against oxidised LDL and progression of carotid atherosclerosis. Lancet 339 (8798), 883 –887. Sansoni, P., Cossarizza, A., Brianti, V., Fagnoni, F., Snelli, G., Monti, D., Marcato, A., Passeri, G., Ortolani, C., Forti, E., et al., 1993. Lymphocyte subsets and natural killer cell activity in healthy old people and centenarians. Blood 82 (9), 2767–2773. Sohal, R.S., 2002. Role of oxidative stress and protein oxidation in the aging process. Free Radic. Biol. Med. 33 (1), 37– 44. Stacy, S., Krolick, K.A., Infante, A.J., Kraig, E., 2002. Immunological memory and late onset autoimmunity. Mech. Ageing Dev. 123 (8), 975– 985. Stadtman, E.R., 2001. Protein oxidation in aging and age-related diseases. Ann. N Y Acad. Sci. 928, 22–38. Steinerova, A., Racek, J., Stozicky, F., Zima, T., Fialova, L., Lapin, A., 2001. Antibodies against oxidized LDL—theory and clinical use. Physiol. Res. 50 (2), 131–141. Tesoriere, L., D’Arpa, D., Bufera, D., Pintaudi, A.M., Allegra, M., Livrea, M.A., 2002. Exposure to malondialdehyde induces an early redox unbalance preceding membrane toxicity in human erythrocytes. Free Radic. Res. 36 (1), 89–97. Traverso, N., Menini, S., Cosso, L., Odetti, P., Albano, E., Pronzato, M.A., Marinari, U.M., 1998. Immunological evidence for increased oxidative stress in diabetic rats. Diabetologia 41 (3), 265–270. Urban, L., Bessenyei, B., Marka, M., Semsei, I., 2002. On the role of aging in the etiology of autoimmunity. Gerontology 48 (3), 179– 184. Vay, D., Parodi, M., Rolla, R., Mottaran, E., Vidali, M., Bellomo, G., Albano, E., 2001. Circulating antibodies recognizing malondialdehydemodified proteins in healthy subjects. Free Radic. Biol. Med. 30 (3), 277– 286. Vendemmiale, G., Altomare, E., Grattagliano, I., Albano, O., 1989. Increased plasma levels of glutathione and malondialdehyde after acute ethanol ingestion in humans. J. Hepatol. 9 (3), 359–365. Vohringer, M.L., Becker, T.W., Krieger, G., Jacobi, H., Witte, I., 1998. Synergistic DNA damaging effects of malondialdehyde/Cu(II) in PM2 DNA and in human fibroblasts. Toxicol. Lett. 94 (3), 159–166. Wu, J.T., Wu, L.L., 1997. Autoantibodies against oxidized LDL. A potential marker for atherosclerosis. Clin. Lab. Med. 17 (3), 595–604. Zhou, X., Caligiuri, G., Hamsten, A., Lefvert, A.K., Hansson, G.K., 2001. LDL immunization induces T-cell-dependent antibody formation and protection against atherosclerosis. Arterioscler. Thromb. Vasc. Biol. 21 (1), 108–114.