Paraoxonase-1 activities in children and adolescents with type 1 diabetes mellitus

1Department of Pharmaceutical Biochemistry and Clinical Laboratory, Iuliu Hatieganu University of Medicine and Pharmacy, Cluj-Napoca, Romania; 2Department of Medical Informatics and Biostatistics, Iuliu Hatieganu University of Medicine and Pharmacy, Cluj-Napoca, Romania; 3Department of Molecular Sciences, Iuliu Hatieganu University of Medicine and Pharmacy, Cluj-Napoca, Romania; 4Iuliu Hatieganu University of Medicine and Pharmacy, Cluj-Napoca, Romania; 5Paediatric Emergency County Hospital, 1th Paediatric Clinic, Cluj-Napoca, Romania


INTRODUCTION
The human serum paraoxonase-1 (PON1) enzymebest known for its paraoxonase (Pon-ase) and arylesterase (Ar-ase) activities -is the most investigated member of the paraoxonase family due to its antioxidant and antiatherogenic activities as a component of high density lipoproteins (HDL) (Harel et al., 2004;She et al., 2012).Structure reactivity studies suggest that the physi-ological function of PON1 is that of lactonase (Lactase), being able to hydrolyze various lactones including δ-valerolactone and homocysteine thiolactone (Khersonsky et al., 2005;James et al., 2006).Clinical implications of this finding are of a particular interest.Due to lactone structure similarities, compounds arising from the hydrolysis of oxidized phospholipids have been proposed as PON1 physiological substrates (James et al., 2006).Growing evidence proved that PON1 is a key mediator in the limitation of both low density lipoproteins (LDL) and HDL oxidation (Aviram et al., 1998;Ng et al., 2005), in the stimulation of macrophage cholesterol efflux (Rosenblat et al., 2006), and therefore in atheroprotection (Mackness et al., 2013).
Type 1 diabetes mellitus (T1DM), with or without vascular complications, is characterized by increased levels of oxidative stress, glycation and glycoxidation.As a result, diabetic patients are predisposed to a higher degree of micro-and macro-vascular complications and are more prone to develop cardiovascular events than healthy subjects (Baynes et al., 1999).In children and adolescents with T1DM, data about PON1 activities, in particular Lact-ase, are either missing or insufficient.Several studies reported low Ar-ase and/or Pon-ase activity in adults with T1DM that was associated with poor metabolic control and vascular complications (Abbott et al., 1995;Boem et al., 2001;Mackness et al., 2002;Ferretti et al., 2004;Flekac et al., 2008).In contrast, Valabhji et al. (2001) did not observe reduced Ar-ase activity in subjects with good glycaemic control and without dyslipidaemia.
Because atherosclerosis -which begins at an early age (Ross et al., 1986) -increases the morbidity and mortality through cardiovascular disease in the developed countries, it has become mandatory to investigate and comprehend the factors that impede or promote the progression of fatty streaks.
In this study we evaluated for the first time the profile of PON1 activities (Pon-ase, Ar-ase and Lact-ase) along with oxidized LDL (oxLDL), glycated haemoglobin (HbA 1c ) and other biochemical markers in a group of children and adolescents with T1DM vs. healthy subjects, and the possible associations with other biochemical markers.

Study groups.
The study was conducted in accordance with the Declaration of Helsinki and approved by the University Ethics Committee.Written informed consent was obtained from the parents of all the participants before enrolling in the study.
The diabetic group consisted of 82 children and adolescents receiving insulin treatment and free of other acute or chronic conditions.We divided the diabetic subjects in two subgroups, one with good glycaemic control (GG; n = 45; HbA 1c < 9%) and the other one with poor glycaemic control (PG, n = 37; HbA 1c ≥ 9%).
The control group consisted of 41 healthy volunteers with similar age and gender distribution, without inflammatory or other pathologic conditions.
Venous blood samples were collected after overnight fasting on a clot activator and were centrifuged at 590 × g for 10 minutes; serum samples were stored at -80ºC until assayed.
PON1 lactonase activity (kU/L).Lact-ase activity, expressed as kU/L, was measured using a pH-sensitive colorimetric assay (Khersonsky et al., 2005), with minor variations.Briefly, 3 µl of serum were incubated with 1 mM δ-valerolactone in 1.2 ml of 2.5 mM bicine buffer (pH 8.3), containing 0.15 M NaCl, 1 mM CaCl 2 and 0.2 mM m-cresol purple.The bleaching rate, resulting from the carboxylic acid formation, was monitored at 577 nm for 90 seconds.For the calculation of the rate factor, a standard calibration curve was performed with 10 mM acetic acid.
PON1 arylesterase activity (kU/L).Ar-ase activity was measured spectrophotometrically according to Eckerson et al. (1983), with minor variations.Briefly, 3 µl of serum were incubated in 1.2 ml of 20 mM Tris-HCl (pH 8.0) containing 1 mM phenyl acetate and 1 mM CaCl 2 .The rate of phenyl acetate hydrolysis was monitored for 90 seconds at 270 nm.Blank samples without serum were analyzed at the same time to correct for spontaneous hydrolysis of the substrate.For activity calculation, expressed as kU/L, a molar extinction coefficient (1310 M -1 cm -1 ) was used.
Salt-stimulated PON1 paraoxonase (ssPon-ase) activity (U/L).ssPon-ase activity was measured spectrophotometrically according to Eckerson et al. (1983), with minor variations.Briefly, 25 µl of serum were incubated with 1 mM paraoxon in a total volume of 1.2 ml of 50 mM glycine-NaOH buffer (pH 10.5) containing 2 M NaCl and 1 mM CaCl 2 .The rate of p-nitrophenol hydrolysis was monitored for 90 seconds at 412 nm.Blank samples without serum were assayed simultaneously to correct for spontaneous hydrolysis of the substrate.For activity calculation, a molar extinction coefficient (18290 M -1 cm -1 ) was used.
All enzymatic activities were measured at 25°C, with a UV-VIS double-beam spectrophotometer (V-530 Jasco) equipped with temperature controller (ETC 505T Jasco).
Statistical analysis.Qualitative data are presented as counts and percentages.The association between qualitative variables was assessed with Pearson's chisquare test.Quantitative data are presented as mean ± standard deviation (SD) for normally distributed data or otherwise as median and interquartile range (IQR).Normality of the data was checked with strip-chart, quantile-quantile plot and Shapiro-Wilk test.Student's t-test for independent samples was used to check for differences between two independent groups of normally distributed data and otherwise by the Mann-Whitney U test.Pairwise comparisons between multiple groups regarding non-normally distributed quantitative variables were performed with Tukey-Kramer test.
Spearman correlation coefficient was used to assess the relations between quantitative variables not following the normal distribution.
Two multiple linear regression analyses for the Lactase activity were performed: one on subjects with diabetes, including HbA 1c and diabetes duration as covariates; the other on all the subjects, including diabetes as a grouping variable (yes vs. no) instead of HbA 1c .Beside these variables, the full model contained: age, gender, TG, TC, LDL, HDL, TC/HDL ratio and oxLDL.Then, a best subset variable selection procedure was employed using Bayesian information criterion to identify the best model.Regression diagnostics procedures were employed verifying linearity, residuals' normality, multicolinearity and heteroskedasticity.The regression coefficients along with 95% confidence intervals (CI) were presented.For all statistical tests used, the significance level α was 0.05 and the two-tailed P value was computed.The statistical analysis was made in R environment for statistical computing and graphics, version 1.15.1 (R Development Core Team, 2010).

RESULTS
The main characteristics of the studied groups are summarized in Table 1.
Age and gender distributions were similar in the diabetic and control groups.As expected, both the glycaemic control (reflected by the HbA 1c level) and the lipid profile were significantly altered in diabetic patients.Compared to controls, TC and LDL levels were significantly increased and the HDL level was significantly decreased in diabetics.
Regarding PON1 activities, we found lower values for diabetics when compared to controls.The decrease was statistically significant only for the Lact-ase activity, the difference being higher when referring to the subgroup with poor glycaemic control.Also, the Lact-ase/HDL ratio was lower in diabetics vs. controls, but not statistically significant.
The linear regression analysis performed on all the subjects gave a final model predicting lactonase that contained the presence of diabetes as the only variable.The coefficient for diabetes presence was -6.50 (95% CI -9.75 --3.25),P < 0.001 and the determination coefficient was 0.11.The presence of diabetes was associated with a low Lact-ase level.The final linear regression model predicting lactonase, performed only on diabetic subjects after the selection procedure, contained only the HbA 1c variable and although not statistically significant, it was close to the significance level.The coefficient for HbA 1c presence was -0.83 (-1.68 -0.03), P = 0.063 and the determination coefficient was 0.04.

DISCUSSION
Data about PON1 in diabetic subjects are controversial.Several studies reported low Ar-ase and/or Pon-ase activity in adults with T1DM that had poor metabolic control and vascular complications (Abbott et al., 1995;Boem et al., 2001;Mackness et al., 2002;Ferretti et al., 2004;Flekac et al., 2008), but Valabhji et al. (2001) did not observe reduced Ar-ase activity in subjects with good glycaemic control and without dyslipidaemia.In diabetic children and adolescents (duration of diabetes 8.8 years), Kordonouri et al. (2001) found low Ar-ase only in patients with blood glucose > 10 mmol/l.However, the results reported by Nair et al. (2011) showed normal Ar-ase and Pon-ase activities in diabetic children (approximately 10 years of age).We also did not find significantly differences regarding Ar-ase and Pon-ase in diabetics vs. controls.Some of the above PON1-related studies did not provide data about the distribution of PON1 genotypes, so we could not exclude the hypothesis that the controversial results might be due to significant differences in PON1 genotypes distribution between the diabetic and control groups.We have identified the individual genotype of our subjects, according to Eckerson et al. (1983) (results not included) and we have observed a similar PON1 Q192R genotype distribution in both diabetic and control groups.
Growing evidence suggests that Lact-ase is the physiological activity of PON1, having at least three antiatherogenic effects: protection of LDL, HDL and macrophages against peroxidation, promotion of macrophage cholesterol efflux and reduction of macrophage-mediated oxLDL uptake (Shih et al., 2000;Rozenberg et al., 2003;Rozenberg et al., 2008;Aharoni et al., 2013).The study presented here shows for the first time that the Lact-ase activity is significantly reduced in children and adolescents with T1DM, but when dividing Lact-ase by HDL, the difference between groups became statistically not significant (p = 0.08 instead of p < 0.001).This might suggest that the low PON1 carrier concentration (observed in the diabetic children) could explain, at least partially, the decrease in Lact-ase activity.If the low level of HDL would be the only source of this decrease, there ought to exist a good correlation among these parameters, which we did not notice (Spearman correlation coefficient = 0.11 (95% CI 0.08-0.32,p = 0.214).
As a result, we thought that other causes have to be taken into account, such as quantitative and/or qualitative alterations in HDL and thus in PON1.Rosenblat et al. (2006) reported that, in diabetic subjects, significant amount of serum PON1 is dissociated from HDL to the lipoprotein-deficient serum (LPDS).Therefore, PON1 conformation is changed and consequently the Pon-ase activity is stimulated, but Lact-ase and Ar-ase activities drastically decrease.Furthermore, they observed that PON1 in LPDS is not able to protect against lipid peroxidation and to stimulate macrophage cholesterol efflux and concluded that abnormal serum PON1 distribution could be responsible for the accelerated atherosclerosis in these patients.Consistent with this data, purified Lactase activity -but not HDL-bound Lact-ase -showed an increased susceptibility to be inactivated under oxidative stress conditions (ascorbate/Cu 2+ oxidation), when compared to the Ar-ase and Pon-ase activities (Nguyen et al., 2009).More recent findings suggest that PON1 has a catalytic promiscuity with different active sites configurations for Pon-ase vs. Ar-ase and Lact-ase activities, which could be responsible for the preserved Pon-ase activity vs. altered Lact-ase and Ar-ase activities in diabetic patients (Ben-David et al., 2012).Besides that, in diabetic patients, HDL and/or PON1 could be glycated and glycoxidated, leading to impaired PON1 stability and/or activities (Kordonouri et al., 2001;Mastorikou et al., 2008).
The results presented here also showed that the diabetic subjects were dyslipidaemic and had poor glycaemic control, but had normal levels of oxLDL.A higher level of LDL was associated with a lower Ar-ase activity in the diabetic group vs. control group.LDL is considered to be a substrate for the antioxidant activity of PON1, and the Ar-ase activity (which is not influenced by Q192R polymorphism) is a substitute of PON1 serum concentration.We have observed statistically higher values for the LDL/Ar-ase ratio in diabetic subjects, a result that could suggest a reduced antioxidant protection of LDL.However, the level of oxLDL was similar in the investigated groups.Compared to our results, higher levels, but not statistically significant, of oxLDL have been reported in T1DM children (9-16 years of age) with short duration of the disease (Hamad et al., 2010).Since the oxidized phospholipids from LDL are the physiological substrates of PON1 Lact-ase, we expected that reduced levels of Lact-ase activity must be associated with high levels of serum oxLDL in the diabetic subjects.Conversely, we did not observe any differences when compared to the control group, probably due to either an efficient antioxidant mechanism (involving other enzymes) or due to rapid removal of oxLDL from the blood (due to immune complexes and macrophage scavenger receptors) (Li et al., 2011).
In our study, both Lact-ase and Ar-ase activities were inversely correlated with HbA 1c , although only the latter was statistically significant.The correlation between Lactase and HbA 1c and also the regression between the two were very close to the statistical significance level.Previous studies reported low Pon-ase and Ar-ase activities related to high blood glucose levels, but not to HbA 1c levels (Kordonouri et al., 2001).A possible explanation for the significant correlation of Ar-ase with HbA 1c could be a higher HbA 1c concentration in our diabetic group (almost half of the subjects had HbA 1c > 9%) suggesting they constantly had elevated blood glucose levels that could directly or indirectly affect PON1 by glycation and/or glycoxidation.
The small sample size was a limitation of this study, however our results underline the importance of further analyzing the relationship between PON1 and childhood diabetes from the perspective of the Lact-ase activity.

CONCLUSIONS
All PON1 activities were lower in children and adolescents with T1DM, but only the decrease in the Lactase activity was statistically significant.The observed negative correlation between HbA 1c and Lact-ase or Arase was close to the statistical significance level for the first one and significant for the second one.Although the Lact-ase activity, lipid profile and the glycaemic control have been altered in diabetic subjects, we did not observe any differences between groups regarding the oxLDL level.