Nutrition, Metabolism & Cardiovascular Diseases (2015) 25, 418e425
Available online at www.sciencedirect.com
Nutrition, Metabolism & Cardiovascular Diseases journal homepage: www.elsevier.com/locate/nmcd
High-density lipoprotein oxidation in type 2 diabetic patients and young patients with premature myocardial infarction G. Sartore a, R. Seraglia b, S. Burlina a,*, A. Bolis a, R. Marin a, E. Manzato a, E. Ragazzi c, P. Traldi b, A. Lapolla a a
Department of Medicine e DIMED, University of Padova, Italy CNR-IENI, Padova, Italy c Department of Pharmaceutical and Pharmacological Sciences, University of Padova, Italy b
Received 10 July 2014; received in revised form 10 December 2014; accepted 12 December 2014 Available online 19 December 2014
KEYWORDS HDL oxidation; Type 2 diabetes; Apolipoprotein A-I
Abstract Background and aims: ApoA-I can undergo oxidative changes that reduce antiatherogenic role of HDL. The aim of this study was to seek any significant differences in methionine sulfoxide (MetO) content in the ApoA-I of HDL isolated from young patients with coronary heart disease (CHD), type 2 diabetics and healthy subjects. Methods and results: We evaluated the lipid profile of 21 type 2 diabetic patients, 23 young patients with premature MI and 21 healthy volunteers; we determined in all patients the MetO content of ApoA-I in by MALDI/TOF/TOF technique. The typical MALDI spectra of the tryptic digest obtained from HDL plasma fractions all patients showed a relative abundance of peptides containing Met112O in ApoA-I in type 2 diabetic and CHD patients. This relative abundance is given as percentages of oxidized ApoA-I (OxApoA-I). OxApoA-I showed no significant correlations with lipoproteins in all patients studied, while a strong correlation emerged between the duration of diabetic disease and OxApoA-I levels in type 2 diabetic patients. Conclusions: The most remarkable finding of our study lies in the evidence it produced of an increased HDL oxidation in patients highly susceptible to CHD. Levels of MetO residues in plasma ApoA-I, measured using an accurate, specific method, should be investigated and considered in prospective future studies to assess their role in CHD. ª 2014 Elsevier B.V. All rights reserved.
Introduction No more than 25% of the risk of coronary heart disease (CHD) can be explained by known risk factors, despite their high prevalence [1]. High-density lipoprotein (HDL) protects artery wall from atherosclerosis, in particular they remove excess cholesterol from artery wall macrophages and carries it
* Corresponding author. Department of Medicine, University of Padova, Via dei Colli 4, 35143 Padova, Italy. Tel.: þ39 0498216848; fax: þ39 0498216838. E-mail address: [email protected] (S. Burlina). http://dx.doi.org/10.1016/j.numecd.2014.12.004 0939-4753/ª 2014 Elsevier B.V. All rights reserved.
back to the liver for excretion in bile [2]. Apolipoprotein AI (ApoA-I) is the main protein of HDL and it plays a crucial part in the first cholesterol transport reversal step by enhancing sterol efflux from macrophages [3]. Epidemiological studies have demonstrated that plasma HDL independently predict the risk of developing atherosclerosis and cardiovascular disease [4]. More recently, however, it has emerged that HDL quality also seems to be an important parameter in atheroprotection, though there is little data in the literature to support it [5]. An increasing body of evidence shows that HDL isolated from atheromas and the plasma of patients with established CHD lacks these anti-atherogenic properties [6]. HDL can be functionally deficient in populations at high
High-density lipoprotein oxidation
risk of CHD, as in type 2 diabetes mellitus, due to glycation and oxidative changes in their HDL, apolipoproteins, and/ or enzymes [7]. ApoA-I in particular can undergo oxidative changes that reduce its anti-atherogenic role [8]. Oxidation of the Tyr and Met residues in ApoA-I by myeloperoxidase drastically impairs the protein’s ability to promote cholesterol efflux via the ABCA1 pathway [9]. Levine and colleagues [10] suggested that Met residues in protein serve as endogenous antioxidants, protecting functionally important amino acids against oxidation. In ApoA-I in particular, Met86 and Met112 are thought to be important for cholesterol efflux, and Met148 is believed to be involved in LCAT activation [11]. Brock et al. recently examined the extent and sites of methionine sulfoxide (MetO) formation in the ApoA-I of HDL isolated from the plasma of healthy controls and type 1 diabetic subjects, demonstrating that MetO formation was significantly greater in diabetic patients than in a control group at all three sites considered (Met86, Met112, and Met148) [12]. Considering the relevant role of HDL oxidation in the onset of atherosclerotic processes, we ran a pilot study on a small group of type 2 diabetic patients and young people prematurely experiencing acute myocardial infarction (MI): in both these groups we found higher levels of Met112O than in healthy controls [13]. That investigation was carried out by microfluidic-LC/ESI-MS measurements. In a further study the determination of MetO content of ApoA-I in type 2 diabetic patients was performed by MALDI/MS [14] and the results obtained perfectly overlap those achieved in the previous LC/MS investigation. These results proved that possible oxidation phenomena, sometimes observed in MALDI conditions [15], are in this case absent. The aim of this study was to assess larger study groups to seek any significant differences in MetO between patients with premature MI, type 2 diabetics and healthy subjects, and to identify any correlations with these individuals’ lipoproteins. A secondary aim was to see whether the duration of the diabetic patients’ disease correlated with HDL oxidation. Methods The study involved 21 consecutive type 2 diabetic patients (10 men and 11 women) attending our outpatient clinic from July 2012 to December 2012, with no history of coronary artery or valve disease. We also studied 23 young patients (22 men and 1 woman) diagnosed with acute MI at 30 min; (2) ST elevation >0.1 mV on at least 2 adjacent electrocardiographic leads; and (3) an increase in troponine I levels to at least the upper limit of normal values. Twenty-one healthy volunteers with no cardiovascular disease and no personal or family history of illness, were also recruited as
419
controls. The study complied with the ethical guidelines of the 1975 Declaration of Helsinki and met with the approval of the local institutional review boards, and informed consent was obtained from each patient. In the study protocol assessment, the sample size was determined a priori, taking into account previous data from a pilot investigation [13] in which peptides related to ApoA-I oxidation resulted about 2.3- to 2.5-fold higher in diabetic and CHD patients. Considering a resulting effect size expressed as Cohen’s f of 0.59, for a one-way ANOVA test, at an alpha error level of 0.05 to achieve a power of 0.95, the estimated total sample size, including the three groups, was 60 patients, which has been respected with the performed patient recruitment procedure. All patients were assessed in terms of body mass index (BMI), diastolic and systolic pressure, any hypertension and/or use of antihypertensive drugs and lipid-lowering drugs. Blood was obtained from fasting venous samples for biochemical analysis. Fasting plasma glucose (FPG) was measured using a glucose-oxidase method [16]. HbA1c was measured by liquid chromatography aligned with IFCC standardization [17] (Adams HA-8180, Arkrary, Kyoto, Japan). Total cholesterol, low-denisty lipoprotein (LDL) and HDL cholesterol were measured using enzymatic analytical chemistry [18] (CHOP-PAP method; Roche, Milan, Italy), as were triglycerides [19] (GPO-PAP colorimetric enzyme test; Roche Diagnostic System). HDL fraction preparation Blood samples were collected in vacutainers after 12e14 h of overnight fasting. To protect methionine residues from oxidation, serum was prepared immediately by low-speed centrifugation at 2500 rpm for 15 min at 10 C, adding Na2-EDTA (0.04% w/v) and HDLs were isolated from the serum by sequential ultracentrifugation on the day when the blood was collected, using the Beckman 50 Ti fixedangle rotor in an Optima XL90 ultracentrifuge (Beckman Instruments, Palo Alto, California, USA) [20]. The density of 6 mL of serum was adjusted to 1.21 kg/L by adding solid KBr (Carlo Erba reagents, Milan, Italy). The resulting solution was placed in polyallomer quick-seal centrifuge tubes (Beckman Instruments, Palo Alto, California, USA) and over-layered with a density solution of KBr, d Z 1.21 kg/L, pH 7.4. Densities were adjusted and checked on a Densito 30P density meter (Mettler Toledo, Switzerland). Centrifugation was done at 40,000 rpm, at 6 C for 30 h. All the lipoproteins were removed in 4 mL of supernatant by tube slicing. The density of 4 mL of supernatant was adjusted to 1.063 (kg/L). Centrifugation was repeated as described above, at 40,000 rpm and a temperature of 6 C, for 24 h. HDLs (d Z 1.063e1.210 kg/L) were removed in 4 ml of bottom fraction and used in the subsequent analysis. Recovery of the 3 fractions of cholesterol was 90%. The HDL fraction was exhaustively dialyzed and concentrated with an Amicon Ultracel 10K centrifugal filter (Millipore Corporation, Billerica USA). Apo-AI and cholesterol concentrations were measured in serum and in the fractions
420
isolated by ultracentrifugation using an automatic Cobas Mira plus analyzer (Horiba ABX, Montpellier, France). ApoA-I was measured with the immune turbidimetric method (Horiba ABX, Montpellier, France). HDL cholesterol was measured after precipitation of the apo-B-containing lipoproteins with Phosphotungstate Precipitant (Roche Diagnostics, Mannheim, Germany) [18] using the same enzymatic-colorimetric method (Cholesterol CHOD-PAP, Roche-Diagnostics, Mannheim, Germany) [21] as for the total cholesterol. Tryptic digestion of HDL fractions The HDL fractions were digested using trypsin according to the following procedure: 1 mg of the ApoA-I fraction was dissolved in 1 mL of 50 mM NH4HCO3 buffer solution (pH Z 8.5) and 100 mL of trypsin solution (100 ng/mL) were added (substrate to enzyme ratio Z 40:1 w/w). The final solutions were left to react at 37 C overnight. The reaction was stopped by adding 50 mL of 10% trifluoroacetic acid in aqueous solution. The digestion mixture was desalted and purified with ZipTip-C18 pipette tips (Millipore, Bedford, USA), following the procedure described in the ZipTip user’s guide.
G. Sartore et al.
was attempted by comparing couples of variables. For the purpose of obtaining a more integrated analysis on selected variables of interest, the non-parametric multivariate method of Principal Component Analysis (PCA) was also used. PCA permits to reduce the dimensions of the dataset, helping to identify new, meaningful underlying variables. The reduced set contains what are called ‘principal factors’, i.e. linear combinations of the original variables. The first principal component accounts for as much of the variability in the data possible, with each successive component accounting for the remaining variability. Biplot software was used as an Excel add-in. The first three components were considered for data classification purposes. A biplot graphic display was used to present the variables’ behavior in order to examine their correlation. The most useful variable is the cosine of the eigenvectors suggesting correlations between different variables [22]. When the angle between eigenvectors nears 0 , the variables are positively correlated, while the angle for negative correlations approaches 180 and angles of 90 indicate no correlation. Results Clinical characteristics
MALDI/MS MALDI/time of flight (TOF) and MALDI/TOF/TOF measurements were performed using a MALDI/TOF/TOF UltrafleXtreme instrument (Bruker Daltonics, Bremen, Germany), equipped with a 1 kHz smartbeam II laser (l Z 355 nm) and operating in the positive reflectron ion mode. The instrumental conditions were: IS1 Z 25 kV; IS2 Z 21.65 kV; reflectron potential Z 26.3 kV; delay time Z 0 nsec. The matrix was a-cyano-4-hydroxycinnamic acid (HCCA) (saturated solution in H2O/acetonitrile (50:50; v/v) containing 0.1% TFA). Five mL of purified tryptic digest and 5 mL of matrix solution were mixed together, then 1 mL of the resulting mixture was deposited on the stainless steel sample holder and allowed to dry before placing it in the mass spectrometer. External mass calibration (Peptide Calibration Standard) was based on monoisotopic values of [MþH]þ of angiotensin II, angiotensin I, substance P, bombesin, ACTH clip [1e17], ACTH clip (18e39), somatostatin 28 at m/z 1046.5420, 1296.6853, 1347.7361, 1619.8230, 2093.0868, 2465.1990 and 3147.4714. TOF/TOF experiments were performed using the LIFT device in the following experimental conditions: IS1: 7.5 kV; IS2: 6.75 kV; Lift1: 19 kV; Lift2: 3.7 kV; Reflector1: 29.5 kV; delay time: 70 ns. Statistical analysis Values are expressed as means standard deviations. To assess the statistical differences between groups, ANOVA was used followed by Tukey’s post-hoc test. Differences were deemed statistically significant when p < 0.05. In order to evaluate the presence of any interrelationships between the measured variables, a linear correlation
Table 1 shows the demographic and clinical characteristics of patients and controls. The three groups were matched for age and smoke; the controls and diabetics were also matched for gender, while the premature MI group consisted almost entirely of men, with only one female patient. Type 2 diabetic patients were not in a situation of good metabolic control, their HbA1c levels being a mean 8.22 0.84% and their FPG 156.7 29.7 mg/dl. The three groups had similar total cholesterol levels. The group of patients with a premature MI had the highest levels of LDL cholesterol and the lowest levels of HDL cholesterol. Their triglycerides were also higher than in the healthy controls, but lower than in the diabetic patients. Characterization of Met112 and Met112-O containing peptides The typical MALDI spectra of the tryptic digest obtained from HDL plasma fractions of healthy subjects, diabetics and CHD patients are given in Fig. 1. MS/MS experiments performed on the two ions at m/z 1283.6 and 2645.4 showed that the sequences of the corresponding peptides are W108QEEM112ELYR and V97QPYLDDFQKKWQEEM112ELYR, both of which contain the methionine residue in position 112 (Met112). Looking at selected regions of the spectra related to the two above-mentioned ions, some differences appear between the healthy controls vs the diabetic patients and CHD patients. In the case of the diabetics and CHD patients, the two peaks at m/z 1299 and 2661 become more evident than those detected in the case of healthy subjects. These two peptides, differing from the above-described species by 16 Da, can be justified by the presence of the previously-mentioned peptides containing
High-density lipoprotein oxidation
421
Table 1 Clinical characteristics of type 2 diabetic patients, young patients with premature CHD and controls. Data are expressed as mean standard deviation. To assess statistical differences between groups, ANOVA followed by Tukey’spost-hoc test was used. CCC p < 0.001; CC p < 0.01; C p < 0.05; ns: not significant;e: not applied. Abbreviations: FPG Z fasting plasma glucose; TC Z total cholesterol; LDL Z low-density lipoprotein; HDL Z high-density lipoprotein; OxApoAI Z oxidized Apolipoprotein AI; MI Z myocardial infarction.
Gender (M/F) Age (yrs) Diabetes duration (yrs) FPG (mg/dl) HbA1c (%) TC (mg/dl) LDL (mg/dl) HDL (mg/dl) Triglycerides (mg/dl) ApoA1 (mg/dl) OX ApoA1 (%) MI (no/yes) Statin therapy (yes/no) Anti-platelet agents (yes/no) Antihypertensive drugs (yes/no) Smokers (yes/no/ex)
Control subjects (C) (n Z 21)
CHD patients (CHD) (n Z 23)
Type 2 diabetic patients (D) (n Z 21)
P C vs CHD
C vs D
CHD vs D
6/15 41.4 2.8 e 83.5 4.8 5.2 0.2 203.2 33.3 117.5 30.6 68.7 11.2 85.4 21.6 148.2 31.3 1.7 1.3 21/0 0/21 0/21 0/21 4/15/2
22/1 40.7 3.4 e 89.1 7.4 5.3 0.2 205.3 26.3 143.7 24.7 38.3 10.4 146.3 55.9 117.2 14.5 4.8 2.6 0/23 23/0 23/0 23/0 5/16/2
10/11 51.8 3.5 8.5 3.9 156.7 29.7 8.2 0.8 203.2 37.0 118.3 30.6 49.7 14.0 178.0 91.8 128.9 14.7 10.6 5.3 21/0 18/3 17/4 20/1 6/14/1
e ns e ns ns ns CC CCC CC C C e e e e ns
e ns e CCC CCC ns ns CCC CCC C CCC e e e e ns
e ns e CCC CCC ns C CC ns ns CCC e ns ns ns ns
a Met112O moiety (see Fig. 2). MS/MS experiments performed on these two ions confirms this hypothesis, based on the presence of a fragment ion due to the loss of CH3SOH. This result confirms that oxidation occurs at Met112 in both the peptides. The above-described relative abundance of peptides containing Met112O- and Met112 was ascertained for all samples. The percentages of
Figure 1
OxApoA-I were calculated dividing the sum of the abundances of the peaks at m/z 1299 and 2661 (originating from oxidation of Met112) to the sum of the abundances of the four peaks of interest: the results so obtained are shown in Table 1. Both the diabetic and the CHD patients showed significantly higher OxApoA-I levels than controls. We did not observe any significant correlation between
The typical MALDI spectra of the tryptic digest obtained from HDL plasma fractions of healthy subjects, diabetic and CHD patients.
422
G. Sartore et al.
Figure 2 Expanded view (A: m/z 1283e1299; B: m/z 2635e2690) of the MALDI mass spectra of tryptic digests from healthy subjects, diabetic patients and CHD patients.
the levels of ApoA-I and OxApoA-I in all groups (controls: r Z 0.031; diabetics: r Z 0.092; CHD patients: r Z 0.20, respectively). It is to underline that the possible ex-vivo oxidation of methionine residue was checked analyzing the lyophilized HDL samples after two and four months of storage at 30 C. No significant variation in the content of Met112O was observed, indicating that ex-vivo oxidation is inhibited at storage temperature.
Correlations OxApoA-I showed no significant correlations with lipoproteins, while there were inverse significant correlations between HDL cholesterol and triglycerides in both the diabetic and the CHD patients (p < 0.02), but not in the healthy controls, as shown in Table 2. No correlation emerged between the OxApoA-I and HbA1c levels in the diabetic patients (r Z 0.0344).
High-density lipoprotein oxidation
423
Table 2 Linear correlation between oxidized ApoA-I (Ox-ApoA-I) and serum cholesterol in the three groups of patients. Data are the Pearson productemoment correlation coefficient (Pearson’s r) with the lower and upper 95% confidence intervals (in parentheses).*p < 0.02. Correlation between variables
Control subjects, n Z 21 (Lower and upper 95% C.I.)
CHD patients, n Z 23 (Lower and upper 95% C.I.)
Diabetic patients, n Z 21 (Lower and upper 95% C.I.)
Ox-ApoA-1 vs total cholesterol
0.3757 (0.0669 0.3557 (0.0897 0.0745 (0.3690 0.1965 (0.2570 0.3973 (0.7076
0.0519 (0.3681 O 0.1688 (0.5432 O 0.3130 (0.1139 O 0.0434 (0.3755 O 0.4898* (0.7505 O
0.3287 (0.6659 O 0.1200) 0.3193 (0.6600 O 0.1303) 0.0839 (0.3608 O 0.4976) 0.1953 (0.5782O0.2581) 0.5413* (0.7887 O 0.1431)
Ox-ApoA-1 vs LDL-cholesterol Ox-ApoA-1 vs HDL-cholesterol Ox-ApoA-1vs triglycerides Triglycerides vs HDL-cholesterol
O 0.6947) O 0.6826) O 0.4904) O 0.5790) O 0.0415)
In order to evaluate with a more integrated approach the presence of interrelationships among variables, the non-parametric technique of PCA was considered. The analysis was extended to the three groups as a whole, in order to check any distribution among the individuals, and the respective role of the considered variables. As the biplot of Fig. 3 shows, it was confirmed the previously found lack of any relationship between the OxApoA-I levels and HDL cholesterol or triglycerides, and it was confirmed also the presence of an inverse correlation between HDL cholesterol or triglycerides; moreover, from this analysis a strong direct correlation between the duration of diabetic disease and OxApoA-I levels emerged. Discussion In the present, small cross-sectional study, our data analyses support the impression that the atheroprotective effect of HDL may be deficient in patients experiencing a premature MI and in cases of type 2 DM, both models of accelerated atherosclerosis [23]. This HDL dysfunction could be due to an increase in MetO levels in ApoA-I. We
0.4544) 0.2616) 0.6423) 0.4476) 0.0972)
demonstrated, not only that type 2 diabetic patients and young patients with premature acute MI share the same ApoA-I oxidation, but also and more importantly they both have a greater HDL oxidation than controls, irrespective of their HDL levels. This feature was recently observed in type 1 diabetic patients compared with healthy controls, and it may contribute to an accelerated atherosclerosis [12]. These findings provide a new clinical perspective compared to preliminary results obtained by microfluidicLC/ESI-MS [13], this time using an alternative technique (MALDI/MS), that makes the analysis far less timeconsuming, as we previously showed in type 2 diabetic patients and healthy controls [14]. Our group of type 2 diabetic patients showed no signs of CHD despite their more severely oxidized HDL. We surmise that they offset the higher levels of oxidized HDL with higher levels of HDL, so the ratio of HDL to oxidized HDL might be a better marker of CHD than low HDL levels. Unfortunately, since our method only allowed for a semiquantitative assessment of the oxidation of the above-described peptides, these data cannot be used to calculate the HDL/oxidizedHDL ratio.
Figure 3 Biplot of the first two principal components (PC1 and PC2) obtained by PCA conducted on the most representative variables from diabetic patients, CHD patients and controls.
424
It is worth noting that no correlation emerged between MetO levels in ApoA-I and HbA1c, indicating that ApoA-I oxidation appears unrelated to the degree of glycemic control. This finding is in agreement with previous observations that have shown no correlation between glycooxidation products, such as glyoxal and methylglyoxal, which better represent the real glyco-oxidative stress experienced by patients [24]. On the other hand, our data suggest that duration of disease might be the parameter most closely related to MetO levels in ApoA-I in type 2 diabetes. In this contest, the antioxidant system could play an important part in the onset of cardiovascular complications by counterregulating the increased oxidative stress, as we found in various phenotypes of type 2 diabetic patients with and without micro- and macrovascular complications [25,26]. Several studies have also demonstrated that decreased levels of antioxidants favor cardiovascular disease in subjects without diabetes [27]. As regards our data on HDL oxidation, we hypothesized that the increase of Apo-AI oxidation could be due to the decreased levels of antioxidant defenses that characterized type 2 diabetic patients with long duration of disease and patient with premature MI. Recent observations, in fact, showed that serum myeloperoxidase/paraoxonase 1 ratio is a potential indicator of dysfunctional HDL and risk stratification in CHD [28]. At the end HDL oxidation process could be partially independent from oxidative stress burden, but affected by decreased antioxidant capacity. As for the higher triglyceride levels found in our type 2 DM and CHD patients, we surmise that hypertriglyceridemia could be a prognostic marker even in young patients with premature MI, irrespective of other cardiovascular risk factors, as previously reported [29]. Both our groups of patients showed a strong inverse relationship between their HDL and triglyceride levels, a situation typical of insulin resistance and found associated with MI occurring before 40 years of age [30]. As regards LDL cholesterol, we found the highest level in young CHD patients who were all in statin therapy. Considering the very short period of statin therapy and knowing that to reach the full effect it needs at still a month, CHD patients showed LDL cholesterol levels not still at target. On the other side, statin therapy hardly have had an impact on the oxidation of HDL. In any case statin protective effect on the oxidation strengthen our conclusions. All these quantitative and qualitative lipoprotein features (higher oxidized HDL, higher triglycerides and lower HDL levels) suggest the feasibility of characterizing patients at high risk of CHD in terms of their lipid profile, as illustrated in the integrated biplot of Fig. 3. In conclusion, the most remarkable finding of our study lies in the evidence it produced of an increased HDL oxidation in patients highly susceptible to CHD. Levels of MetO residues in plasma ApoA-I, measured using an accurate, specific method, should be investigated and considered in prospective future studies in order to assess their possible role as a novel risk factor e and eventually as
G. Sartore et al.
a therapeutic target e to reduce the burden of cardiovascular complications. Disclosures No conflict to disclose. References [1] Pyorala K, Laakso M, Uusitupa M. Diabetes and atherosclerosis: an epidemiologic view. Diabetes Metab Rev 1987;3:463e524. [2] Lewis GF, Rader DJ. New insights into the regulation of HDL metabolism and reverse cholesterol transport. Circ Res 2005;96: 1221e32. [3] Oram JF. HDL apolipoproteins and ABCA1: partners in the removal of excess cellular cholesterol. Arterioscler Thromb Vasc Biol 2003; 23:720e7. [4] Gordon DJ, Probstfield JL, Garrison RJ, Neaton JD, Castelli WP, Knoke JD, et al. High-density lipoprotein cholesterol and cardiovascular disease. Four prospective American studies. Circulation 1989;79:8e15. [5] Frikke-Schmidt R, Nordestgaard BG, Stene MC, Sethi AA, Remaley AT, Schnohr P, et al. Association of loss of function mutations in the ABCA1 gene with high-density lipoprotein cholesterol levels and risk of ischemic heart disease. JAMA 2008;299: 2524e32. [6] Dodani S, Grice DG, Joshi S. Is HDL function as important as HDL quantity in the coronary artery disease risk assessment? J Clin Lipidol 2009;3:70e7. [7] Kontush A, Chapman MJ. Why is HDL functionally deficient in type 2 diabetes? Curr Diab Rep 2008;8:51e9. [8] Lapolla A, Sartore G. The role of lipoprotein abnormalities as risk factors for macroangiopathy in type 2 diabetes. In: Handbook of lipoprotein research, editor: Jackson E. Rathbond, [chapter 14], p. 227e231, Nova Science Publishers, Inc. [9] Shao B, Oda MN, Bergt C, Fu X, Green PS, Brot N, et al. Myeloperoxidase impairs ABCA1-dependent cholesterol efflux through methionine oxidation and site-specific tyrosine chlorination of apolipoprotein A-I. J Biol Chem 2006;281:9001e4. [10] Levine RL, Mosoni L, Berlett BS, Stadtman ER. Methionine residues as endogenous antioxidants in proteins. Proc Natl Acad Sci U S A 1996;93:15036e40. [11] Panzenbock U, Stocker R. Formation of methionine sulfoxidecontaining specific forms of oxidized high-density lipoproteins. Biochim Biophys Acta 2005;1703:171e81. [12] Brock JW, Jenkins AJ, Lyons TJ, Klein RL, Yim E, Lopes-Virella M, et al. Increased methionine sulfoxide content of apoA-I in type 1 diabetes. J Lipid Res 2008;49:847e55. [13] Lapolla A, Manzato E, Sartore G, Marin R, Cosma C, Bolis A, et al. Evaluation of methionine sulphoxide content of ApoA-I in type 2 diabetic patients and young coronaropathic subjects: a preliminary study. Rapid Commun Mass Spectrom 2011;25:391e4. [14] Seraglia R, Sartore G, Marin R, Burlina S, Manzato E, Ragazzi E, et al. An effective and rapid determination by MALDI/TOF/TOF of methionine sulphoxide content of ApoA-I in type 2 diabetic patients. J Mass Spectrom 2013;48:105e10. [15] Dreisewerd K. The desorption process in MALDI. Chem Rev 2003; 103:395e425. [16] Huggett AST, Nixon DA. Use of glucose oxidase peroxidase and Odianisine in the determination of blood and urine glucose. Lancet 1957;2:368e70. [17] Thevarajah TM, Nani N, Chew YY. Performance evaluation of the Arkray Adams HA-8160 HbA1c analyser. Malays J Pathol 2008;30: 81e6. [18] Lipid research clinics program: lipid and lipoprotein analysis. Washington DC. In: Hainline A, Karon J, Lippel K, editors. Manual of Laboratory Operations; 1982. p. 63e77. [19] Fossati P, Prencipe L. Serum triglycerides determined colorimetrically with an enzyme that produces hydrogen peroxide. Clin Chem 1982;28:2077e80. [20] Schumaker VN, Puppione DL. Sequential flotation ultracentrifugation. Meth Enzymol 1986;128:155e70.
High-density lipoprotein oxidation [21] Roschlau P, Bernt E, Gruber W. Enzymatic determination of total cholesterol in serum. Z Klin Chem Klin Biochem 1974;12:403e7. [22] Aitchinson J, Greencare M. Biplots of compositional data. Appl Stat 2002;51:375e92. [23] Onat A, Can G, Ayhan E, Kaya Z, Hergenc G. Impaired protection against diabetes and coronary heart disease by high-density lipoproteins in Turks. Metabolism 2009;58:1393e9. [24] Lapolla A, Reitano R, Seraglia R, Sartore G, Ragazzi E, Traldi P. Evaluation of advanced glycation end products and carbonyl compounds in patients with different conditions of oxidative stress. Mol Nutr Food Res 2005;49:685e90. [25] Piarulli F, Sartore G, Ceriello A, Ragazzi E, Reitano R, Nollino L, et al. Relationship between glyco-oxidation, antioxidant status and microalbuminuria in type 2 diabetic patients. Diabetologia 2009; 52:1419e25. [26] Lapolla A, Piarulli F, Sartore G, Ceriello A, Ragazzi E, Reitano R, et al. Advanced glycation end products and antioxidant status in
425
[27]
[28]
[29]
[30]
type 2 diabetic patients with and without peripheral artery disease. Diabetes Care 2007;30:670e6. Shimizu H, Kiyohara Y, Kato I, Kitazono T, Tanizaki Y, Kubo M, et al. Relationship between plasma glutathione levels and cardiovascular disease in a defined population: the Hisayama study. Stroke 2004;35:2072e7. Haraguchi Y, Toh R, Hasokawa M, Nakajima H, Honjo T, Otsui K, et al. Serum myeloperoxidase/paraoxonase 1 ratio as potential indicator of dysfunctional high-density lipoprotein and risk stratification in coronary artery disease. Atherosclerosis 2014;234: 288e94. van Loon JE, de Maat MP, Deckers JW, van Domburg RT, Leebeek FW. Prognostic markers in young patients with premature coronary heart disease. Atherosclerosis 2012;224:213e7. Cavallo-Perin P, Bergerone S, Gagnor A, Comune M, Giunti S, Cassader M, et al. Myocardial infarction before the age of 40 years is associated with insulin resistance. Metabolism 2001;50:30e5.
Available online at www.sciencedirect.com
Nutrition, Metabolism & Cardiovascular Diseases journal homepage: www.elsevier.com/locate/nmcd
High-density lipoprotein oxidation in type 2 diabetic patients and young patients with premature myocardial infarction G. Sartore a, R. Seraglia b, S. Burlina a,*, A. Bolis a, R. Marin a, E. Manzato a, E. Ragazzi c, P. Traldi b, A. Lapolla a a
Department of Medicine e DIMED, University of Padova, Italy CNR-IENI, Padova, Italy c Department of Pharmaceutical and Pharmacological Sciences, University of Padova, Italy b
Received 10 July 2014; received in revised form 10 December 2014; accepted 12 December 2014 Available online 19 December 2014
KEYWORDS HDL oxidation; Type 2 diabetes; Apolipoprotein A-I
Abstract Background and aims: ApoA-I can undergo oxidative changes that reduce antiatherogenic role of HDL. The aim of this study was to seek any significant differences in methionine sulfoxide (MetO) content in the ApoA-I of HDL isolated from young patients with coronary heart disease (CHD), type 2 diabetics and healthy subjects. Methods and results: We evaluated the lipid profile of 21 type 2 diabetic patients, 23 young patients with premature MI and 21 healthy volunteers; we determined in all patients the MetO content of ApoA-I in by MALDI/TOF/TOF technique. The typical MALDI spectra of the tryptic digest obtained from HDL plasma fractions all patients showed a relative abundance of peptides containing Met112O in ApoA-I in type 2 diabetic and CHD patients. This relative abundance is given as percentages of oxidized ApoA-I (OxApoA-I). OxApoA-I showed no significant correlations with lipoproteins in all patients studied, while a strong correlation emerged between the duration of diabetic disease and OxApoA-I levels in type 2 diabetic patients. Conclusions: The most remarkable finding of our study lies in the evidence it produced of an increased HDL oxidation in patients highly susceptible to CHD. Levels of MetO residues in plasma ApoA-I, measured using an accurate, specific method, should be investigated and considered in prospective future studies to assess their role in CHD. ª 2014 Elsevier B.V. All rights reserved.
Introduction No more than 25% of the risk of coronary heart disease (CHD) can be explained by known risk factors, despite their high prevalence [1]. High-density lipoprotein (HDL) protects artery wall from atherosclerosis, in particular they remove excess cholesterol from artery wall macrophages and carries it
* Corresponding author. Department of Medicine, University of Padova, Via dei Colli 4, 35143 Padova, Italy. Tel.: þ39 0498216848; fax: þ39 0498216838. E-mail address: [email protected] (S. Burlina). http://dx.doi.org/10.1016/j.numecd.2014.12.004 0939-4753/ª 2014 Elsevier B.V. All rights reserved.
back to the liver for excretion in bile [2]. Apolipoprotein AI (ApoA-I) is the main protein of HDL and it plays a crucial part in the first cholesterol transport reversal step by enhancing sterol efflux from macrophages [3]. Epidemiological studies have demonstrated that plasma HDL independently predict the risk of developing atherosclerosis and cardiovascular disease [4]. More recently, however, it has emerged that HDL quality also seems to be an important parameter in atheroprotection, though there is little data in the literature to support it [5]. An increasing body of evidence shows that HDL isolated from atheromas and the plasma of patients with established CHD lacks these anti-atherogenic properties [6]. HDL can be functionally deficient in populations at high
High-density lipoprotein oxidation
risk of CHD, as in type 2 diabetes mellitus, due to glycation and oxidative changes in their HDL, apolipoproteins, and/ or enzymes [7]. ApoA-I in particular can undergo oxidative changes that reduce its anti-atherogenic role [8]. Oxidation of the Tyr and Met residues in ApoA-I by myeloperoxidase drastically impairs the protein’s ability to promote cholesterol efflux via the ABCA1 pathway [9]. Levine and colleagues [10] suggested that Met residues in protein serve as endogenous antioxidants, protecting functionally important amino acids against oxidation. In ApoA-I in particular, Met86 and Met112 are thought to be important for cholesterol efflux, and Met148 is believed to be involved in LCAT activation [11]. Brock et al. recently examined the extent and sites of methionine sulfoxide (MetO) formation in the ApoA-I of HDL isolated from the plasma of healthy controls and type 1 diabetic subjects, demonstrating that MetO formation was significantly greater in diabetic patients than in a control group at all three sites considered (Met86, Met112, and Met148) [12]. Considering the relevant role of HDL oxidation in the onset of atherosclerotic processes, we ran a pilot study on a small group of type 2 diabetic patients and young people prematurely experiencing acute myocardial infarction (MI): in both these groups we found higher levels of Met112O than in healthy controls [13]. That investigation was carried out by microfluidic-LC/ESI-MS measurements. In a further study the determination of MetO content of ApoA-I in type 2 diabetic patients was performed by MALDI/MS [14] and the results obtained perfectly overlap those achieved in the previous LC/MS investigation. These results proved that possible oxidation phenomena, sometimes observed in MALDI conditions [15], are in this case absent. The aim of this study was to assess larger study groups to seek any significant differences in MetO between patients with premature MI, type 2 diabetics and healthy subjects, and to identify any correlations with these individuals’ lipoproteins. A secondary aim was to see whether the duration of the diabetic patients’ disease correlated with HDL oxidation. Methods The study involved 21 consecutive type 2 diabetic patients (10 men and 11 women) attending our outpatient clinic from July 2012 to December 2012, with no history of coronary artery or valve disease. We also studied 23 young patients (22 men and 1 woman) diagnosed with acute MI at 30 min; (2) ST elevation >0.1 mV on at least 2 adjacent electrocardiographic leads; and (3) an increase in troponine I levels to at least the upper limit of normal values. Twenty-one healthy volunteers with no cardiovascular disease and no personal or family history of illness, were also recruited as
419
controls. The study complied with the ethical guidelines of the 1975 Declaration of Helsinki and met with the approval of the local institutional review boards, and informed consent was obtained from each patient. In the study protocol assessment, the sample size was determined a priori, taking into account previous data from a pilot investigation [13] in which peptides related to ApoA-I oxidation resulted about 2.3- to 2.5-fold higher in diabetic and CHD patients. Considering a resulting effect size expressed as Cohen’s f of 0.59, for a one-way ANOVA test, at an alpha error level of 0.05 to achieve a power of 0.95, the estimated total sample size, including the three groups, was 60 patients, which has been respected with the performed patient recruitment procedure. All patients were assessed in terms of body mass index (BMI), diastolic and systolic pressure, any hypertension and/or use of antihypertensive drugs and lipid-lowering drugs. Blood was obtained from fasting venous samples for biochemical analysis. Fasting plasma glucose (FPG) was measured using a glucose-oxidase method [16]. HbA1c was measured by liquid chromatography aligned with IFCC standardization [17] (Adams HA-8180, Arkrary, Kyoto, Japan). Total cholesterol, low-denisty lipoprotein (LDL) and HDL cholesterol were measured using enzymatic analytical chemistry [18] (CHOP-PAP method; Roche, Milan, Italy), as were triglycerides [19] (GPO-PAP colorimetric enzyme test; Roche Diagnostic System). HDL fraction preparation Blood samples were collected in vacutainers after 12e14 h of overnight fasting. To protect methionine residues from oxidation, serum was prepared immediately by low-speed centrifugation at 2500 rpm for 15 min at 10 C, adding Na2-EDTA (0.04% w/v) and HDLs were isolated from the serum by sequential ultracentrifugation on the day when the blood was collected, using the Beckman 50 Ti fixedangle rotor in an Optima XL90 ultracentrifuge (Beckman Instruments, Palo Alto, California, USA) [20]. The density of 6 mL of serum was adjusted to 1.21 kg/L by adding solid KBr (Carlo Erba reagents, Milan, Italy). The resulting solution was placed in polyallomer quick-seal centrifuge tubes (Beckman Instruments, Palo Alto, California, USA) and over-layered with a density solution of KBr, d Z 1.21 kg/L, pH 7.4. Densities were adjusted and checked on a Densito 30P density meter (Mettler Toledo, Switzerland). Centrifugation was done at 40,000 rpm, at 6 C for 30 h. All the lipoproteins were removed in 4 mL of supernatant by tube slicing. The density of 4 mL of supernatant was adjusted to 1.063 (kg/L). Centrifugation was repeated as described above, at 40,000 rpm and a temperature of 6 C, for 24 h. HDLs (d Z 1.063e1.210 kg/L) were removed in 4 ml of bottom fraction and used in the subsequent analysis. Recovery of the 3 fractions of cholesterol was 90%. The HDL fraction was exhaustively dialyzed and concentrated with an Amicon Ultracel 10K centrifugal filter (Millipore Corporation, Billerica USA). Apo-AI and cholesterol concentrations were measured in serum and in the fractions
420
isolated by ultracentrifugation using an automatic Cobas Mira plus analyzer (Horiba ABX, Montpellier, France). ApoA-I was measured with the immune turbidimetric method (Horiba ABX, Montpellier, France). HDL cholesterol was measured after precipitation of the apo-B-containing lipoproteins with Phosphotungstate Precipitant (Roche Diagnostics, Mannheim, Germany) [18] using the same enzymatic-colorimetric method (Cholesterol CHOD-PAP, Roche-Diagnostics, Mannheim, Germany) [21] as for the total cholesterol. Tryptic digestion of HDL fractions The HDL fractions were digested using trypsin according to the following procedure: 1 mg of the ApoA-I fraction was dissolved in 1 mL of 50 mM NH4HCO3 buffer solution (pH Z 8.5) and 100 mL of trypsin solution (100 ng/mL) were added (substrate to enzyme ratio Z 40:1 w/w). The final solutions were left to react at 37 C overnight. The reaction was stopped by adding 50 mL of 10% trifluoroacetic acid in aqueous solution. The digestion mixture was desalted and purified with ZipTip-C18 pipette tips (Millipore, Bedford, USA), following the procedure described in the ZipTip user’s guide.
G. Sartore et al.
was attempted by comparing couples of variables. For the purpose of obtaining a more integrated analysis on selected variables of interest, the non-parametric multivariate method of Principal Component Analysis (PCA) was also used. PCA permits to reduce the dimensions of the dataset, helping to identify new, meaningful underlying variables. The reduced set contains what are called ‘principal factors’, i.e. linear combinations of the original variables. The first principal component accounts for as much of the variability in the data possible, with each successive component accounting for the remaining variability. Biplot software was used as an Excel add-in. The first three components were considered for data classification purposes. A biplot graphic display was used to present the variables’ behavior in order to examine their correlation. The most useful variable is the cosine of the eigenvectors suggesting correlations between different variables [22]. When the angle between eigenvectors nears 0 , the variables are positively correlated, while the angle for negative correlations approaches 180 and angles of 90 indicate no correlation. Results Clinical characteristics
MALDI/MS MALDI/time of flight (TOF) and MALDI/TOF/TOF measurements were performed using a MALDI/TOF/TOF UltrafleXtreme instrument (Bruker Daltonics, Bremen, Germany), equipped with a 1 kHz smartbeam II laser (l Z 355 nm) and operating in the positive reflectron ion mode. The instrumental conditions were: IS1 Z 25 kV; IS2 Z 21.65 kV; reflectron potential Z 26.3 kV; delay time Z 0 nsec. The matrix was a-cyano-4-hydroxycinnamic acid (HCCA) (saturated solution in H2O/acetonitrile (50:50; v/v) containing 0.1% TFA). Five mL of purified tryptic digest and 5 mL of matrix solution were mixed together, then 1 mL of the resulting mixture was deposited on the stainless steel sample holder and allowed to dry before placing it in the mass spectrometer. External mass calibration (Peptide Calibration Standard) was based on monoisotopic values of [MþH]þ of angiotensin II, angiotensin I, substance P, bombesin, ACTH clip [1e17], ACTH clip (18e39), somatostatin 28 at m/z 1046.5420, 1296.6853, 1347.7361, 1619.8230, 2093.0868, 2465.1990 and 3147.4714. TOF/TOF experiments were performed using the LIFT device in the following experimental conditions: IS1: 7.5 kV; IS2: 6.75 kV; Lift1: 19 kV; Lift2: 3.7 kV; Reflector1: 29.5 kV; delay time: 70 ns. Statistical analysis Values are expressed as means standard deviations. To assess the statistical differences between groups, ANOVA was used followed by Tukey’s post-hoc test. Differences were deemed statistically significant when p < 0.05. In order to evaluate the presence of any interrelationships between the measured variables, a linear correlation
Table 1 shows the demographic and clinical characteristics of patients and controls. The three groups were matched for age and smoke; the controls and diabetics were also matched for gender, while the premature MI group consisted almost entirely of men, with only one female patient. Type 2 diabetic patients were not in a situation of good metabolic control, their HbA1c levels being a mean 8.22 0.84% and their FPG 156.7 29.7 mg/dl. The three groups had similar total cholesterol levels. The group of patients with a premature MI had the highest levels of LDL cholesterol and the lowest levels of HDL cholesterol. Their triglycerides were also higher than in the healthy controls, but lower than in the diabetic patients. Characterization of Met112 and Met112-O containing peptides The typical MALDI spectra of the tryptic digest obtained from HDL plasma fractions of healthy subjects, diabetics and CHD patients are given in Fig. 1. MS/MS experiments performed on the two ions at m/z 1283.6 and 2645.4 showed that the sequences of the corresponding peptides are W108QEEM112ELYR and V97QPYLDDFQKKWQEEM112ELYR, both of which contain the methionine residue in position 112 (Met112). Looking at selected regions of the spectra related to the two above-mentioned ions, some differences appear between the healthy controls vs the diabetic patients and CHD patients. In the case of the diabetics and CHD patients, the two peaks at m/z 1299 and 2661 become more evident than those detected in the case of healthy subjects. These two peptides, differing from the above-described species by 16 Da, can be justified by the presence of the previously-mentioned peptides containing
High-density lipoprotein oxidation
421
Table 1 Clinical characteristics of type 2 diabetic patients, young patients with premature CHD and controls. Data are expressed as mean standard deviation. To assess statistical differences between groups, ANOVA followed by Tukey’spost-hoc test was used. CCC p < 0.001; CC p < 0.01; C p < 0.05; ns: not significant;e: not applied. Abbreviations: FPG Z fasting plasma glucose; TC Z total cholesterol; LDL Z low-density lipoprotein; HDL Z high-density lipoprotein; OxApoAI Z oxidized Apolipoprotein AI; MI Z myocardial infarction.
Gender (M/F) Age (yrs) Diabetes duration (yrs) FPG (mg/dl) HbA1c (%) TC (mg/dl) LDL (mg/dl) HDL (mg/dl) Triglycerides (mg/dl) ApoA1 (mg/dl) OX ApoA1 (%) MI (no/yes) Statin therapy (yes/no) Anti-platelet agents (yes/no) Antihypertensive drugs (yes/no) Smokers (yes/no/ex)
Control subjects (C) (n Z 21)
CHD patients (CHD) (n Z 23)
Type 2 diabetic patients (D) (n Z 21)
P C vs CHD
C vs D
CHD vs D
6/15 41.4 2.8 e 83.5 4.8 5.2 0.2 203.2 33.3 117.5 30.6 68.7 11.2 85.4 21.6 148.2 31.3 1.7 1.3 21/0 0/21 0/21 0/21 4/15/2
22/1 40.7 3.4 e 89.1 7.4 5.3 0.2 205.3 26.3 143.7 24.7 38.3 10.4 146.3 55.9 117.2 14.5 4.8 2.6 0/23 23/0 23/0 23/0 5/16/2
10/11 51.8 3.5 8.5 3.9 156.7 29.7 8.2 0.8 203.2 37.0 118.3 30.6 49.7 14.0 178.0 91.8 128.9 14.7 10.6 5.3 21/0 18/3 17/4 20/1 6/14/1
e ns e ns ns ns CC CCC CC C C e e e e ns
e ns e CCC CCC ns ns CCC CCC C CCC e e e e ns
e ns e CCC CCC ns C CC ns ns CCC e ns ns ns ns
a Met112O moiety (see Fig. 2). MS/MS experiments performed on these two ions confirms this hypothesis, based on the presence of a fragment ion due to the loss of CH3SOH. This result confirms that oxidation occurs at Met112 in both the peptides. The above-described relative abundance of peptides containing Met112O- and Met112 was ascertained for all samples. The percentages of
Figure 1
OxApoA-I were calculated dividing the sum of the abundances of the peaks at m/z 1299 and 2661 (originating from oxidation of Met112) to the sum of the abundances of the four peaks of interest: the results so obtained are shown in Table 1. Both the diabetic and the CHD patients showed significantly higher OxApoA-I levels than controls. We did not observe any significant correlation between
The typical MALDI spectra of the tryptic digest obtained from HDL plasma fractions of healthy subjects, diabetic and CHD patients.
422
G. Sartore et al.
Figure 2 Expanded view (A: m/z 1283e1299; B: m/z 2635e2690) of the MALDI mass spectra of tryptic digests from healthy subjects, diabetic patients and CHD patients.
the levels of ApoA-I and OxApoA-I in all groups (controls: r Z 0.031; diabetics: r Z 0.092; CHD patients: r Z 0.20, respectively). It is to underline that the possible ex-vivo oxidation of methionine residue was checked analyzing the lyophilized HDL samples after two and four months of storage at 30 C. No significant variation in the content of Met112O was observed, indicating that ex-vivo oxidation is inhibited at storage temperature.
Correlations OxApoA-I showed no significant correlations with lipoproteins, while there were inverse significant correlations between HDL cholesterol and triglycerides in both the diabetic and the CHD patients (p < 0.02), but not in the healthy controls, as shown in Table 2. No correlation emerged between the OxApoA-I and HbA1c levels in the diabetic patients (r Z 0.0344).
High-density lipoprotein oxidation
423
Table 2 Linear correlation between oxidized ApoA-I (Ox-ApoA-I) and serum cholesterol in the three groups of patients. Data are the Pearson productemoment correlation coefficient (Pearson’s r) with the lower and upper 95% confidence intervals (in parentheses).*p < 0.02. Correlation between variables
Control subjects, n Z 21 (Lower and upper 95% C.I.)
CHD patients, n Z 23 (Lower and upper 95% C.I.)
Diabetic patients, n Z 21 (Lower and upper 95% C.I.)
Ox-ApoA-1 vs total cholesterol
0.3757 (0.0669 0.3557 (0.0897 0.0745 (0.3690 0.1965 (0.2570 0.3973 (0.7076
0.0519 (0.3681 O 0.1688 (0.5432 O 0.3130 (0.1139 O 0.0434 (0.3755 O 0.4898* (0.7505 O
0.3287 (0.6659 O 0.1200) 0.3193 (0.6600 O 0.1303) 0.0839 (0.3608 O 0.4976) 0.1953 (0.5782O0.2581) 0.5413* (0.7887 O 0.1431)
Ox-ApoA-1 vs LDL-cholesterol Ox-ApoA-1 vs HDL-cholesterol Ox-ApoA-1vs triglycerides Triglycerides vs HDL-cholesterol
O 0.6947) O 0.6826) O 0.4904) O 0.5790) O 0.0415)
In order to evaluate with a more integrated approach the presence of interrelationships among variables, the non-parametric technique of PCA was considered. The analysis was extended to the three groups as a whole, in order to check any distribution among the individuals, and the respective role of the considered variables. As the biplot of Fig. 3 shows, it was confirmed the previously found lack of any relationship between the OxApoA-I levels and HDL cholesterol or triglycerides, and it was confirmed also the presence of an inverse correlation between HDL cholesterol or triglycerides; moreover, from this analysis a strong direct correlation between the duration of diabetic disease and OxApoA-I levels emerged. Discussion In the present, small cross-sectional study, our data analyses support the impression that the atheroprotective effect of HDL may be deficient in patients experiencing a premature MI and in cases of type 2 DM, both models of accelerated atherosclerosis [23]. This HDL dysfunction could be due to an increase in MetO levels in ApoA-I. We
0.4544) 0.2616) 0.6423) 0.4476) 0.0972)
demonstrated, not only that type 2 diabetic patients and young patients with premature acute MI share the same ApoA-I oxidation, but also and more importantly they both have a greater HDL oxidation than controls, irrespective of their HDL levels. This feature was recently observed in type 1 diabetic patients compared with healthy controls, and it may contribute to an accelerated atherosclerosis [12]. These findings provide a new clinical perspective compared to preliminary results obtained by microfluidicLC/ESI-MS [13], this time using an alternative technique (MALDI/MS), that makes the analysis far less timeconsuming, as we previously showed in type 2 diabetic patients and healthy controls [14]. Our group of type 2 diabetic patients showed no signs of CHD despite their more severely oxidized HDL. We surmise that they offset the higher levels of oxidized HDL with higher levels of HDL, so the ratio of HDL to oxidized HDL might be a better marker of CHD than low HDL levels. Unfortunately, since our method only allowed for a semiquantitative assessment of the oxidation of the above-described peptides, these data cannot be used to calculate the HDL/oxidizedHDL ratio.
Figure 3 Biplot of the first two principal components (PC1 and PC2) obtained by PCA conducted on the most representative variables from diabetic patients, CHD patients and controls.
424
It is worth noting that no correlation emerged between MetO levels in ApoA-I and HbA1c, indicating that ApoA-I oxidation appears unrelated to the degree of glycemic control. This finding is in agreement with previous observations that have shown no correlation between glycooxidation products, such as glyoxal and methylglyoxal, which better represent the real glyco-oxidative stress experienced by patients [24]. On the other hand, our data suggest that duration of disease might be the parameter most closely related to MetO levels in ApoA-I in type 2 diabetes. In this contest, the antioxidant system could play an important part in the onset of cardiovascular complications by counterregulating the increased oxidative stress, as we found in various phenotypes of type 2 diabetic patients with and without micro- and macrovascular complications [25,26]. Several studies have also demonstrated that decreased levels of antioxidants favor cardiovascular disease in subjects without diabetes [27]. As regards our data on HDL oxidation, we hypothesized that the increase of Apo-AI oxidation could be due to the decreased levels of antioxidant defenses that characterized type 2 diabetic patients with long duration of disease and patient with premature MI. Recent observations, in fact, showed that serum myeloperoxidase/paraoxonase 1 ratio is a potential indicator of dysfunctional HDL and risk stratification in CHD [28]. At the end HDL oxidation process could be partially independent from oxidative stress burden, but affected by decreased antioxidant capacity. As for the higher triglyceride levels found in our type 2 DM and CHD patients, we surmise that hypertriglyceridemia could be a prognostic marker even in young patients with premature MI, irrespective of other cardiovascular risk factors, as previously reported [29]. Both our groups of patients showed a strong inverse relationship between their HDL and triglyceride levels, a situation typical of insulin resistance and found associated with MI occurring before 40 years of age [30]. As regards LDL cholesterol, we found the highest level in young CHD patients who were all in statin therapy. Considering the very short period of statin therapy and knowing that to reach the full effect it needs at still a month, CHD patients showed LDL cholesterol levels not still at target. On the other side, statin therapy hardly have had an impact on the oxidation of HDL. In any case statin protective effect on the oxidation strengthen our conclusions. All these quantitative and qualitative lipoprotein features (higher oxidized HDL, higher triglycerides and lower HDL levels) suggest the feasibility of characterizing patients at high risk of CHD in terms of their lipid profile, as illustrated in the integrated biplot of Fig. 3. In conclusion, the most remarkable finding of our study lies in the evidence it produced of an increased HDL oxidation in patients highly susceptible to CHD. Levels of MetO residues in plasma ApoA-I, measured using an accurate, specific method, should be investigated and considered in prospective future studies in order to assess their possible role as a novel risk factor e and eventually as
G. Sartore et al.
a therapeutic target e to reduce the burden of cardiovascular complications. Disclosures No conflict to disclose. References [1] Pyorala K, Laakso M, Uusitupa M. Diabetes and atherosclerosis: an epidemiologic view. Diabetes Metab Rev 1987;3:463e524. [2] Lewis GF, Rader DJ. New insights into the regulation of HDL metabolism and reverse cholesterol transport. Circ Res 2005;96: 1221e32. [3] Oram JF. HDL apolipoproteins and ABCA1: partners in the removal of excess cellular cholesterol. Arterioscler Thromb Vasc Biol 2003; 23:720e7. [4] Gordon DJ, Probstfield JL, Garrison RJ, Neaton JD, Castelli WP, Knoke JD, et al. High-density lipoprotein cholesterol and cardiovascular disease. Four prospective American studies. Circulation 1989;79:8e15. [5] Frikke-Schmidt R, Nordestgaard BG, Stene MC, Sethi AA, Remaley AT, Schnohr P, et al. Association of loss of function mutations in the ABCA1 gene with high-density lipoprotein cholesterol levels and risk of ischemic heart disease. JAMA 2008;299: 2524e32. [6] Dodani S, Grice DG, Joshi S. Is HDL function as important as HDL quantity in the coronary artery disease risk assessment? J Clin Lipidol 2009;3:70e7. [7] Kontush A, Chapman MJ. Why is HDL functionally deficient in type 2 diabetes? Curr Diab Rep 2008;8:51e9. [8] Lapolla A, Sartore G. The role of lipoprotein abnormalities as risk factors for macroangiopathy in type 2 diabetes. In: Handbook of lipoprotein research, editor: Jackson E. Rathbond, [chapter 14], p. 227e231, Nova Science Publishers, Inc. [9] Shao B, Oda MN, Bergt C, Fu X, Green PS, Brot N, et al. Myeloperoxidase impairs ABCA1-dependent cholesterol efflux through methionine oxidation and site-specific tyrosine chlorination of apolipoprotein A-I. J Biol Chem 2006;281:9001e4. [10] Levine RL, Mosoni L, Berlett BS, Stadtman ER. Methionine residues as endogenous antioxidants in proteins. Proc Natl Acad Sci U S A 1996;93:15036e40. [11] Panzenbock U, Stocker R. Formation of methionine sulfoxidecontaining specific forms of oxidized high-density lipoproteins. Biochim Biophys Acta 2005;1703:171e81. [12] Brock JW, Jenkins AJ, Lyons TJ, Klein RL, Yim E, Lopes-Virella M, et al. Increased methionine sulfoxide content of apoA-I in type 1 diabetes. J Lipid Res 2008;49:847e55. [13] Lapolla A, Manzato E, Sartore G, Marin R, Cosma C, Bolis A, et al. Evaluation of methionine sulphoxide content of ApoA-I in type 2 diabetic patients and young coronaropathic subjects: a preliminary study. Rapid Commun Mass Spectrom 2011;25:391e4. [14] Seraglia R, Sartore G, Marin R, Burlina S, Manzato E, Ragazzi E, et al. An effective and rapid determination by MALDI/TOF/TOF of methionine sulphoxide content of ApoA-I in type 2 diabetic patients. J Mass Spectrom 2013;48:105e10. [15] Dreisewerd K. The desorption process in MALDI. Chem Rev 2003; 103:395e425. [16] Huggett AST, Nixon DA. Use of glucose oxidase peroxidase and Odianisine in the determination of blood and urine glucose. Lancet 1957;2:368e70. [17] Thevarajah TM, Nani N, Chew YY. Performance evaluation of the Arkray Adams HA-8160 HbA1c analyser. Malays J Pathol 2008;30: 81e6. [18] Lipid research clinics program: lipid and lipoprotein analysis. Washington DC. In: Hainline A, Karon J, Lippel K, editors. Manual of Laboratory Operations; 1982. p. 63e77. [19] Fossati P, Prencipe L. Serum triglycerides determined colorimetrically with an enzyme that produces hydrogen peroxide. Clin Chem 1982;28:2077e80. [20] Schumaker VN, Puppione DL. Sequential flotation ultracentrifugation. Meth Enzymol 1986;128:155e70.
High-density lipoprotein oxidation [21] Roschlau P, Bernt E, Gruber W. Enzymatic determination of total cholesterol in serum. Z Klin Chem Klin Biochem 1974;12:403e7. [22] Aitchinson J, Greencare M. Biplots of compositional data. Appl Stat 2002;51:375e92. [23] Onat A, Can G, Ayhan E, Kaya Z, Hergenc G. Impaired protection against diabetes and coronary heart disease by high-density lipoproteins in Turks. Metabolism 2009;58:1393e9. [24] Lapolla A, Reitano R, Seraglia R, Sartore G, Ragazzi E, Traldi P. Evaluation of advanced glycation end products and carbonyl compounds in patients with different conditions of oxidative stress. Mol Nutr Food Res 2005;49:685e90. [25] Piarulli F, Sartore G, Ceriello A, Ragazzi E, Reitano R, Nollino L, et al. Relationship between glyco-oxidation, antioxidant status and microalbuminuria in type 2 diabetic patients. Diabetologia 2009; 52:1419e25. [26] Lapolla A, Piarulli F, Sartore G, Ceriello A, Ragazzi E, Reitano R, et al. Advanced glycation end products and antioxidant status in
425
[27]
[28]
[29]
[30]
type 2 diabetic patients with and without peripheral artery disease. Diabetes Care 2007;30:670e6. Shimizu H, Kiyohara Y, Kato I, Kitazono T, Tanizaki Y, Kubo M, et al. Relationship between plasma glutathione levels and cardiovascular disease in a defined population: the Hisayama study. Stroke 2004;35:2072e7. Haraguchi Y, Toh R, Hasokawa M, Nakajima H, Honjo T, Otsui K, et al. Serum myeloperoxidase/paraoxonase 1 ratio as potential indicator of dysfunctional high-density lipoprotein and risk stratification in coronary artery disease. Atherosclerosis 2014;234: 288e94. van Loon JE, de Maat MP, Deckers JW, van Domburg RT, Leebeek FW. Prognostic markers in young patients with premature coronary heart disease. Atherosclerosis 2012;224:213e7. Cavallo-Perin P, Bergerone S, Gagnor A, Comune M, Giunti S, Cassader M, et al. Myocardial infarction before the age of 40 years is associated with insulin resistance. Metabolism 2001;50:30e5.
