Poly(ADP-Ribose) Polymerase Is Activated in Subjects at Risk of Developing Type 2 Diabetes and Is Associated With Impaired Vascular Reactivity
Background— We have previously shown that endothelial function is impaired not only in diabetes but also in subjects at risk of developing type 2 diabetes. We hypothesized that changes in the expression or activity of the endothelial isoform of nitric oxide synthase (eNOS), the receptor for advanced glycation end products (RAGE), and poly(ADP-ribose) polymerase (PARP) are related to this impairment.
Methods and Results— We included a control group of 21 healthy subjects, a group of 22 healthy individuals with parental history of type 2 diabetes, a group of 23 subjects with impaired glucose tolerance, and a group of 21 type 2 diabetic patients. Two 2-mm forearm skin biopsies were taken from each participant and used for measurements. The percentage of PARP-positive endothelial nuclei was higher in the group with parental history of type 2 diabetes and diabetic patients compared with the controls (P<0.001). Immunoreactivity for nitrotyrosine (a marker of reactive nitrogen species) was higher in the diabetic group compared with all other groups (P<0.01). No differences in the expression of eNOS and RAGE were found among all 4 groups. The polymorphism of the eNOS gene was also studied and was not found to influence eNOS expression or microvascular functional measurements.
Conclusions— PARP activation is present in healthy subjects at risk of developing diabetes as well as in established type 2 diabetic patients, and it is associated with impairments in the vascular reactivity in the skin microcirculation.
Received July 2, 2002; revision received September 6, 2002; accepted September 6, 2002.
We have previously shown that endothelial function in the microcirculation and macrocirculation is impaired not only in diabetes but also in subjects with impaired glucose tolerance (IGT) and healthy subjects with normal glucose metabolism and parental history of type 2 diabetes.1 Other studies have also suggested an association between endothelial function and insulin resistance.2–4⇓⇓ Factors that are related to these changes include reduced expression and activity of the endothelial NO synthase (eNOS), reduced tetrahydrobiopterin availability, and reduced l-arginine availability that results in reduced bioavailability of nitric oxide.5–8⇓⇓⇓ In addition, increased production of superoxide (O−2) leads to increased degradation of NO and the formation of the oxidant peroxynitrite (ONOO−), a potent trigger of various forms of oxidative and nitrosative cell injury.9,10⇓ Peroxynitrite can trigger multiple reactions, including protein tyrosine nitration; the detection of nitrotyrosine is often used as a surrogate end point to measure the in vivo formation of peroxynitrite.11
The progressive accumulation of advanced glycation end products and the related overexpression of RAGE and nuclear factor (NF)-κB activation have also been linked to endothelial dysfunction.12,13⇓ Finally, recent studies in animal models of diabetes have suggested that activation of the poly(ADP ribose) polymerase (PARP) contributes to the development of endothelial dysfunction.14,15⇓ PARP is an abundant nuclear enzyme that recognizes oxidative DNA damage and triggers an inefficient cellular metabolic cycle, which leads to cellular dysfunction and can ultimately culminate in cell necrosis.15,16⇓
In the present study, we have hypothesized that changes in the expression or activity of eNOS, RAGE, and PARP are related to the changes in vascular reactivity in diabetic patients and healthy subjects at risk of diabetes. To test our hypothesis, we have examined the expression of these enzymes in forearm skin biopsies taken from patients with uncomplicated type 2 diabetes, subjects with IGT, healthy subjects at risk of developing diabetes, and healthy subjects without a parental history of diabetes. Furthermore, we have studied the association between the expression of these enzymes and measurements of vascular reactivity, biochemical markers of endothelial dysfunction, and the polymorphism of the eNOS gene.
Of the 143 subjects who participated in the previously mentioned study, 87 also took also part in this study. The demographics of these subjects were not different from those of the whole group and have already been published elsewhere.1 In brief, 4 groups of subjects were recruited: a control group of 21 healthy subjects with a normal oral glucose tolerance test and no history of type 2 diabetes in any first degree relative, a group of 22 healthy individuals with a normal oral glucose tolerance test and a history of type 2 diabetes in one or both parents, a group of 23 subjects with IGT, and a group of 21 subjects who had an established diagnosis of type 2 diabetes. Diabetes and IGT were defined according to the recommendations of the American Diabetes Association Expert Committee on the Classification and Diagnosis of Diabetes.17 Subjects with known factors that could affect the endothelial function were excluded from the study.1 The protocol was approved by the ethics committee at each center, and all participants gave written informed consent.
The methods for classifying participants in each group and measuring the vascular reactivity at the microcirculation and macrocirculation and biochemical markers of endothelial function have been described elsewhere.1
The skin biopsies were performed the same day that vascular reactivity measurements were performed, and blood specimens were taken for the measurement of biochemical markers of endothelial dysfunction. Two 2-mm skin punch biopsies were taken from the volar aspect of the forearm under local anesthesia (1% plain lidocaine). The specimens were snapped frozen in liquid nitrogen and then kept in a −70°C freezer.
The first specimen was subsequently submitted for sections and stained with a battery of immunoperoxidase stains. This battery included eNOS (both polyclonal and monoclonal, Transduction Laboratories, 1:1000 dilution), CD31, and endothelial cell marker Factor III (von Willebrand factor, Sigma 1:1000). Immunoperoxidase- stained slides were evaluated independently by a pathologist experienced in the evaluation of these stains, and a semiquantitative scale was used to rate the overall staining intensity in accordance with previously described techniques.18 In brief, the semiquantitative scale was as follows: 2+, strong staining or staining; 1+, weak staining; and 0, feeble staining (>90% reduction). The examining pathologist was unaware as to which group the participants belonged.
The second biopsy was used for the quantitative measurement of eNOS, RAGE, and CD31 by using Western blotting techniques. In brief, tissue biopsies were homogenized in extraction buffer containing 1% SDS (Sigma), 1 mmol/L sodium vanadate (Sigma), and 50 mmol/L Tris HCl, pH 7.4 (Sigma). Protein extracts were obtained by centrifugation of the lysate at 4°C, and concentration was measured with Pierce BSA Protein Assay Reagent. Protein (40 μg) was separated by SDS-PAGE, transferred to nitrocellulose membrane (Millipore) using Bio-Rad Mini Trans-Blot Cell. Membranes were blocked for 1 hour in Tris-buffered saline-Tween with 5% dry milk and incubated with monoclonal antibodies directed against human eNOS (Transduction Laboratories) or with monoclonal antibodies directed against human CD 31(DAKO, Glostrup, Denmark). To detect RAGE antigen, anti-human RAGE IgG antibodies were used. After washing, membranes were incubated with horseradish-conjugated rabbit anti-mouse polyclonal antibodies and washed again. Antigen detection was performed with a chemiluminescent detection system (NEN).
Immunohistochemical Detection of Poly(ADP-Ribose)
PARP activity in tissues was measured using an immunohistochemical method that quantifies the accumulation of poly(ADP-ribose) (PAR), the product of the PARP enzyme in tissue sections.14,19⇓ Mouse monoclonal anti-PAR antibody (Alexis, San Diego, Calif) and isotype-matched control antibody was applied in a dilution of 1:400 for 1 hour at room temperature. After extensive washing (3×10 minutes) with 0.25% Triton/PBS, immunoreactivity was detected with a biotinylated horse anti-mouse secondary antibody and the avidin-biotin-peroxidase complex, both supplied in the Vector Elite kit (Vector Laboratories). Color was developed using Ni-DAB substrate kit (Vector Laboratories). Sections were then counterstained with nuclear fast red, dehydrated, and mounted in Permount. Photomicrographs were taken with a Zeiss Axiolab microscope equipped with a Fuji HC-300C digital camera.
The percentage of PAR-positive nuclei of cells was obtained by conventional microscopy, as previously published.15 A total of 735 to 1910 nuclei profiles were examined in each condition by an investigator who was unaware of which group each participant belonged to. The results are expressed as the percent of PAR-positive nuclei of endothelial cells, relative to the number of total nuclei counted.
Immunohistochemical Detection of 3-Nitrotyrosine
The method was described previously in detail.14 Mouse monoclonal anti-nitrotyrosine antibody (Upstate Biotechnology, Lake Placid, NY) and isotype-matched control antibody were applied in a dilution of 1:200 for 1 hour at room temperature. Additional steps of staining were identical, as described for PAR detection above. All immunohistochemical samples were coded and examined and analyzed by an investigator in a blinded fashion. Extent of immunoreactivity in cells was determined by measuring the optical density of nitrotyrosine signal in the cells. Image analysis was performed as previously described.20 A total of ≈1000 cell intensities were examined in each condition.
These tests were performed in 15 healthy controls (6 male), 9 relatives (5 male), 13 subjects with IGT (7 male), and 9 diabetic patients (2 male). The main reason that not all patients were tested is that there no longer was available skin tissue from the remaining subjects. There were no major differences between the patients who were included in the subcomponent of study and the whole study group.
Three polymorphisms of the eNOS gene were tested: (1) a T-786C substitution in the promoter21; (2) an insertion-deletion in intron 4 (two alleles: the a-deletion, having 4 repeats of a 27-bp consensus sequence, and the b-insertion, having 5 repeats)22; and (3) a G894T substitution in exon 7 that results in a Glu→Asp amino acid substitution at codon 298.23
Polymorphisms in the promoter and in exon 7 were determined by allele-specific oligonucleotide hybridization protocols, as previously described.24 Briefly, a 223-bp fragment containing the T-786C polymorphism and a 267-bp fragment containing the G894T substitution in exon 7 were amplified by polymerase chain reaction (PCR). For blotting, the PCR-amplified product was added to a denaturing solution and transferred onto a nylon membrane placed on a dot-blot apparatus under vacuum. After transfer, the membrane was immersed in a 2× sodium chloride/sodium citrate solution and dried at 80°C. For hybridization, 17-bp ASO probes were used as previously described.24 Each ASO probe was labeled with (32γ)-P-ATP. The membrane was put in a hybridization chamber with 25 mL of hybridization solution containing the labeled ASO and a 20- to 25-fold excess of unlabeled ASO corresponding to the other allele. The membrane was incubated overnight at 52°C in a hybridization oven and then washed for 20 minutes in a cold TMAC wash solution and 20 minutes in a 52°C warmed TMAC wash solution and then exposed for several hours to an x-ray film.
The a-deletion/b-insertion in intron 4 was determined by the following procedure. Forward and reverse primers were used as described previously.23 After PCR amplification, 4 μL of the denatured PCR product was loaded onto a 5% polyacrylamide DNA sequencing gel and resolved electrophoretically. Dried gels were exposed to x-ray film for 12 hours.
The Minitab statistical package (Minitab Inc) for personal computers was used for the statistical analysis. For parametrically distributed data, the ANOVA test was used, followed by the Fisher test to identify differences among the various groups. For nonparametrically distributed data, the Kruskal-Wallis test was used. Correlation between variables was tested using both univariate and multivariate analyses (Pearson correlation analysis and multiple stepwise regression analysis). For the multiple regression analysis, we first screened the tested variables using univariate tests for association of the PARP staining, nitrotyrosine immunoreactivity, and RAGE levels with functional measurements of vascular reactivity and biochemical measurements of endothelial function, glycemic and lipid control, and other demographic measurements using a screening probability value of 0.20. A stepwise model was then developed from among these variables.
The demographics and the biochemical results are shown in Table 1. The results of the microvascular and macrovascular reactivity are shown in Table 2. There were no differences in the measurements observed in the subgroup of the patients who participated in this study and the results of the whole group that have been described elsewhere.1 In brief, the 4 study groups were matched for age and sex. Furthermore, no differences were found between the healthy controls and the group with parental history of diabetes regarding body mass index, fasting glucose, fasting insulin, HbA1c, and lipid measurements. However, the vasodilatory response to the iontophoresis of acetylcholine (endothelium-dependent) and sodium nitroprusside (endothelium-independent) at the forearm skin level and the flow-mediated dilation of the brachial artery were lower in the groups with parental history of diabetes, patients with IGT, and patients with diabetes compared with the healthy subjects.
The percentage of PARP-positive nuclei was higher in the group of relatives and diabetic patients compared with the controls (P<0.001, Figures 1 and 2⇓). Significant correlations were observed between the percentage of PARP-positive endothelial nuclei and fasting blood glucose, resting skin blood, maximal skin vasodilatory response to the iontophoresis of acetylcholine, and sodium nitroprusside and nitrotyrosine immunostaining intensity (Table 5). On stepwise multivariate regression analysis, the two main contributors to the variation were fasting blood glucose and resting skin blood flow, accounting for 34% of the observed variation.
Nitrotyrosine immunoreactivity intensity was higher in the diabetic group compared with all other groups (P<0.01), whereas no significant differences were observed among the remaining 3 groups (Figure 3). Significant correlations were observed between nitrotyrosine immunostaining intensity and fasting blood glucose, HbA1c, intracellular adhesion molecule, and vascular cellular adhesion molecule (VCAM) (Table 5). On stepwise multivariate regression analysis, the two main contributors to the variation were fasting blood glucose and VCAM, accounting for 29% of the observed variation.
The results of the immunostaining of the skin biopsies are shown in Table 3. No differences in the intensity of staining of von Willebrand factor, CD31, and eNOS were found among all 4 groups. The results of Western blotting for CD31, eNOS, and RAGE are shown in Table 4. No differences were found among the 4 groups. No differences were observed when the eNOS and RAGE measurements were corrected for the quantity of CD31 (data not shown). A moderate but statistically significant correlation was found between RAGE and fasting blood glucose and HbA1c (Table 5). On stepwise multivariate analysis, HbA1c was the only significant contributor, accounting for 20% of the RAGE variation. No correlations were observed between eNOS levels and any measurement of the vascular reactivity or biochemical markers of endothelial dysfunction. There were no differences in the eNOS expression among the tested polymorphisms (data not shown). Regression analysis also failed to show any influence of the genotype in the eNOS expression. No correlations were observed between eNOS and RAGE levels. No differences were found in microvascular and macrovascular reactivity among the various genotypes of eNOS (data not shown).
The results of the eNOS gene polymorphism are shown in Table 6. No differences were found among the 4 groups in all tested polymorphisms. Of note when grouping relatives, patients with IGT, and diabetic patients, there was a lower frequency of a carriers in controls [total χ2(1 degree of freedom)=4.9, P=0.026].
In the present study, we have shown that PARP is activated in human dermal microvascular tissues not only in type 2 diabetes but also in healthy subjects with parental history of type 2 diabetes. These changes were also associated with fasting blood glucose and HbA1c and the maximal skin vasodilatory response to the iontophoresis of acetylcholine (which indicates endothelium-dependent vasodilation) and sodium nitroprusside (which indicates endothelium-independent vasodilation). Nitrotyrosine immunoreactivity was also increased in type 2 diabetic patients alone. No differences in the eNOS and RAGE levels were observed in any of the 4 study groups, but a significant association was observed between RAGE, fasting blood glucose, and HbA1c.
Previous studies have shown PARP activation in experimental diabetes that results in impaired vascular relaxant function and can be prevented or reversed by PARP inhibitors.14,15,25,26⇓⇓⇓ This is the first study to report increased activation of PARP in human diabetes or any other disease. Furthermore, our results indicate (to a lesser degree) the activation of PARP in healthy subjects with parental history of type 2 diabetes. The activation of PARP was associated with changes in the vascular reactivity of the skin microcirculation at the same area of the body, the forearm. Therefore, these data are consistent with the hypothesis that the activation of PARP contributes to the observed changes in the microvascular reactivity. Similar changes to that in the group of healthy subjects with parental history of type 2 diabetes were observed in the subjects with IGT. The most likely reason that there was no statistical difference between this group and the controls is the small number of tested subjects in each group. In animal models, pharmacological inhibition or genetic absence of PARP restores normal vasodilatory function in established diabetes, but no data are available for human diabetes.14,15,26⇓⇓
Increased nitrotyrosine levels have been previously reported in cardiac myocytes and endothelial cells from type 2 diabetic patients, suggesting a causative link between the formation of peroxynitrite and myocyte and endothelial dysfunction.27 Nitrotyrosine levels were increased in the diabetic patients, but there were no changes in the subjects with parental history of type diabetes or IGT. Thus, it seems that changes in the peroxynitrite formation represent a late event that occurs only in diabetes and not in the prediabetic stage. In addition to the peroxynitrite formation, other pathways can also lead to the formation of nitrotyrosine; therefore, nitrotyrosine is a marker of nitrating species rather than a specific indicator of peroxynitrite formation.28 Thus, our results can be interpreted as an indication of increased nitrosative stress in diabetes.29
Diabetes has been associated with increased glycoxidation stress, the production of advanced glycation end products, and overexpression of RAGE.30,31⇓ Although no difference was found in the RAGE expression among the 4 tested groups, it is interesting that RAGE levels were associated with both fasting blood glucose and HbA1c. These findings additionally support the concept that RAGE is overexpressed under conditions of hyperglycemia. In addition, no changes were observed in the eNOS expression. However, the activity of eNOS can be impaired in diabetes despite a normal expression, and this can be caused by the PARP activation-related depletion of NADPH.14,15⇓ Finally, we have found no association between eNOS polymorphism, eNOS expression, and endothelial function at both the microcirculation and macrocirculation level. The eNOS polymorphism has been found in some, but not all, studies to be associated with hypertension, vascular disease, and diabetic nephropathy.22,24,32,33⇓⇓⇓ The lower frequency of a-carriers in the control group is an interesting observation and should be confirmed in larger studies.
Previous studies have shown that the skin microcirculation is impaired in diabetes, whereas the present study also indicates PARP activation in the endothelial cells.1,2⇓ In addition, preliminary studies in our unit have suggested an association between the changes in skin microcirculation and the cortical oxygenation of the kidney, a primary target of microvascular disease in diabetes.34 These findings emphasize the potential importance of the skin microcirculation as a surrogate end point and as a target organ of diabetic vascular disease, and additional work should be done toward this direction.
Taken together, our data are consistent with the proposal that early activation of the PARP, at least in the microcirculation, leads to impaired vasorelaxant function not only in diabetes but also in healthy subjects at risk of developing of diabetes. Increased oxidative and nitrosative stress seems to be present only in diabetic patients and is associated with increased production of the proatherogenic molecules intracellular adhesion molecule and VCAM. Finally, hyperglycemia may also be related to the overexpression of RAGE. Additional studies are required to investigate whether PARP inhibition or pharmacological reversal of some of the other noted abnormalities in humans can beneficially affect the development of vascular dysfunction.
This study was supported by a clinical research grant to Dr Horton from the American Diabetes Association and by a grant to Dr Szabo (No. 1R01HL/DK71215-01) from the National Institutes of Health. It was also supported in part by grant RR 01032 to the Beth Israel Deaconess Medical Center General Clinical Research Center from the National Institutes of Health. Anti-human RAGE IgG was kindly provided by Ann Marie Schmidt, MD, Columbia University, New York, NY.
- ↵Caballero AE, Arora S, Saouaf R, et al. Micro- and macro-vascular reactivity is impaired in subjects at risk for type 2 diabetes. Diabetes. 1999; 48: 1863–1867.
- ↵Vehkavaara S, Seppala-Lindroos A, Westerbacka J, et al. In vivo endothelial dysfunction characterizes patients with impaired fasting glucose. Diabetes Care. 1999; 22: 2055–2060.
- ↵Cosentino F, Hishikawa K, Katusic ZS, et al. High glucose increases nitric oxide synthase expression and superoxide anion generation in human aortic endothelial cells. Circulation. 1997; 96: 25–28.
- ↵White CR, Brock TA, Chang LY, et al. Superoxide and peroxynitrite in atherosclerosis. Proc Natl Acad Sci U S A. 1994; 91: 1044–1048.
- ↵Kislinger T, Fu C, Huber B, et al. Nε-(carboxymethyl)lysine adducts of proteins are ligands for receptor for advanced glycation end products that activate cell signaling pathways and modulate gene expression. J Biol Chem. 1999; 274: 31740–31749.
- ↵Pacher P, Liaudet L, Soriano FG, et al. The role of poly(ADP-ribose) polymerase activation in the development of myocardial and endothelial dysfunction in diabetes. Diabetes. 2002; 51: 514–521.
- ↵Szabo C, Zingarelli B, O’Connor M, et al. DNA strand breakage, activation of poly (ADP-ribose) synthetase, and cellular energy depletion are involved in the cytotoxicity of macrophages and smooth muscle cells exposed to peroxynitrite. Proc Natl Acad Sci U S A. 1996; 93: 1753–1758.
- ↵American Diabetes Association. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care. 1997; 20: 1183–1197.
- ↵Bakondi E, Bai P, Szabo E, et al. Detection of poly(ADP-ribose) polymerase activation in oxidatively stressed cells and tissues using biotinylated NAD substrate. J Histochem Cytochem. 2002; 50: 91–98.
- ↵Pacher P, Hajnoczky G. Propagation of the apoptotic signal by mitochondrial waves. EMBO J. 2001; 20: 4107–4121.
- ↵Nakayama M, Yasue H, Yoshimura M, et al. T-786C mutation in the 5′-flanking region of the endothelial nitric oxide synthase gene is associated with coronary spasm. Circulation. 1999; 99: 2864–2870.
- ↵Soriano FG, Pacher P, Mabley J, et al. Rapid reversal of the diabetic endothelial dysfunction by pharmacological inhibition of poly(ADP-ribose) polymerase. Circ Res. 2001; 89: 684–691.
- ↵Frustaci A, Kajstura J, Chimenti C, et al. Myocardial cell death in human diabetes. Circ Res. 2000; 87: 1123–1132.
- ↵Schmidt AM, Hori O, Chen JX, et al. Advanced glycation endproducts interacting with their endothelial receptor induce expression of vascular cell adhesion molecule-1 (VCAM-1) in cultured human endothelial cells and in mice: a potential mechanism for the accelerated vasculopathy of diabetes. J Clin Invest. 1995; 96: 1395–1403.
- ↵Hingorani AD, Liang CF, Fatibene J, et al. A common variant of the endothelial nitric oxide synthase (Glu298->Asp) is a major risk factor for coronary artery disease in the UK. Circulation. 1999; 100: 1515–1520.