Published online before print October 20, 2005, doi: 10.1161/01.STR.0000185703.88944.7d
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
(Stroke. 2005;36:2346.)
© 2005 American Heart Association, Inc.
Original Contributions |
Evaluation of the Paraoxonases as Candidate Genes for Stroke
Gln192Arg Polymorphism in the Paraoxonase 1 Gene Is Associated With Increased Risk of Stroke
From the Pharmaceutical Research Institute (K.R., T.G.K., T.D., P.V., L.H., Z.T., P.M.S., N.C.D.), Bristol-Myers Squibb, Princeton, NJ; Celera Diagnostics (O.A.I., J.J.D., T.J.W.), Alameda, Calif; and Brigham & Women’s Hospital (F.M.S., M.S.S., E.B.), Harvard Medical School, Boston, Mass.
Correspondence to Koustubh Ranade, PhD, Pharmacogenomics & Human Genetics, PRI, Bristol-Myers Squibb, P.O. Box 5400, Princeton, NJ 08543-5400. E-mail koustubh.ranade©bms.com
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Background and Purpose— The paraoxonases are involved in protecting low-density lipoprotein (LDL) from lipid oxidation. Paraoxonase 1 (PON1) was implicated in susceptibility to coronary artery disease and stroke in previous studies. We evaluated, in a comprehensive way, all 3 paraoxonase genes for association with stroke observed in the Cholesterol and Recurrent Events (CARE) trial.
Methods— Over 2500 subjects enrolled in the CARE trial were genotyped for 14 single nucleotide polymorphisms, including 7 newly identified in this study, in the 3 paraoxonase genes.
Results— A glutamine (Gln)/arginine (Arg) polymorphism at amino acid residue 192 in PON1 was significantly associated with stroke (P=0.003 in multivariate analysis, including age, sex, LDL, hypertension, diabetes, smoking, and pravastatin treatment as covariates). The odds ratios were 2.28 (95% CI, 1.38 to 3.79) for Gln/Arg heterozygotes and 2.47 (95% CI, 1.18 to 5.19) for Arg/Arg homozygotes compared with Gln/Gln homozygotes. These results are consistent with 2 of 3 other published studies. In combined analysis of all 4 studies, the association between Gln192Arg SNP and stroke was highly significant (
28df=45.58, P<0.000001). Sequence analysis of the PON1 gene from seventy stroke cases revealed a novel nonsense mutation at codon 32 in one stroke case, which was not detected in over 2500 unaffected individuals. Polymorphisms in the PON2 and PON3 genes were not associated with stroke.
Conclusions— These results suggest that Gln192Arg genotype is an important risk factor for stroke.
Key Words: association • genetics • polymorphism • stroke
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The paraoxonase family comprises 3 related genes that share 60 to 65% similarity at the amino acid level.1 All 3 paraoxonases protect low-density lipoprotein (LDL) from oxidation, at least in vitro.2–6 Paraoxonases 1 and 3 (PON1 and PON3) reside on circulating high-density lipoprotein (HDL) particles.2–5 Paraoxonase 2 (PON2) is ubiquitously expressed and does not appear to be associated with HDL particles.6 Because of their antioxidant activity, and the longstanding hypothesis that oxidized LDL is a key mediator of atherosclerosis, the paraoxonases have been considered candidate susceptibility genes for coronary artery disease (CAD) and stroke. There have been numerous, inconsistent reports of associations between single nucleotide polymorphisms in the PON genes and susceptibility to CAD.7,8 By contrast, few studies have examined the influence of PON genotype on risk of stroke.9–12
In this study, we describe comprehensive analysis of the PON gene family for association with stroke observed in the Cholesterol and Recurrent Events (CARE) trial. Because all patients followed up in the CARE trial had a history of myocardial infarction (MI), which was a requirement for inclusion in the trial, we could not examine, in a meaningful way, genetic association with MI. For this reason, we focused on stroke.
|
Materials and Methods |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Subjects
The CARE trial and strokes observed in this trial have been described.13,14 Briefly, 4159 patients with a history of MI and cholesterol levels below 240 mg/dL received the cholesterol-lowering drug pravastatin or placebo and were followed for 5 years. In this study, 2634 white subjects, including 81 stroke cases (out of 128 in the total CARE cohort) for whom sufficient DNA was available, were evaluated. Of the 81 stroke cases analyzed in this study, 7 were classified as hemorrhagic and the others were categorized as ischemic or of unknown origin by the CARE Endpoints Committee.14 The study was approved by the appropriate Institutional Review Boards and all subjects gave written informed consent.
SNP Discovery and Genotyping
SNPs were identified by sequencing a panel of DNA samples (M44PDR) obtained from the Coriell repository or 70 stroke cases from the CARE cohort. Genotyping was done using the TaqMan or allele-specific polymerase chain reaction methods.15–17 Probes and primers used in the genotyping are provided in the Table I (available online only at http://www.strokeaha.org).
|
Statistical Analysis
SNP genotypes and alleles were assessed for association with stroke using a chi-squared test in preliminary analysis. Significant association (P<0.05) between the PON1 glutamine/arginine (Gln192Arg) SNP and stroke was followed up by logistic regression analysis, including age, sex, body mass index, LDL, diabetes, hypertension, smoking status, aspirin, and warfarin use and treatment with pravastatin as covariates. Interactions between Gln192Arg genotype and smoking, diabetes, and pravastatin treatment were assessed by including an interaction term in the logistic regression model and by stratifying on these variables. Breslow-Day and Tarone’s tests were used to assess the homogeneity of odds ratios in the different strata. SNP genotypes were evaluated for influence on lipid and glucose levels using analysis of variance. Probability values were combined across studies using Fisher’s method.18 Homogeneity of odds ratios across studies was examined using a logit-based test as described.19 Attributable fraction was estimated as described.20 Analysis was done using S-plus (version 6; Insightful Corp.) or SPSS (version 12, SPSS Inc.).
|
Results |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A total of 2634 white patients enrolled in the CARE trial for whom sufficient DNA was available were genotyped for 14 polymorphisms in all 3 paraoxonase genes. Patient characteristics are provided in Table 1, and SNPs are shown in the Figure. In the first screen, we examined 5 SNPs in the PON1 gene, including 2 that result in amino acid substitu-tions—leucine to methionine at residue 55 (Leu55Met) and Gln192Arg.21,22 Two other SNPs, PON1G-907C and PON1C-107T, in the promoter of PON1 were also examined.23 A previously described nonsense mutation at tryptophan codon 19424 was not detected in CARE subjects. Two missense SNPs that result in alanine to glycine (Ala148Gly) and serine to cysteine (Ser311Cys) amino acid substitutions at residues 148 and 311, respectively, in the PON2 gene were evaluated.25 Four polymorphisms in the PON3 gene were examined. Three of these SNPs resulted in amino acid changes—glutamate to lysine (Glu146Lys), alanine to aspartate (Ala179Asp), and tyrosine to cysteine (Tyr233Cys) at residues 146, 179, and 233, respectively. One was a synonymous substitution at residue 99 (Ala99Ala).
|
|
The genotype distribution for each SNP is given in Table 2. The Gln192Arg SNP in the PON1 gene was significantly associated with stroke (P=0.001). The odds ratios were 2.43 (95% confidence interval [CI], 1.48 to 3.98) and 2.73 (1.32 to 5.63) for Gln/Arg heterozygotes and Arg/Arg homozygotes, respectively, compared with Gln/Gln homozygotes. The frequency of the arginine allele in stroke cases was 0.41 compared with 0.28 in unaffected subjects (P=0.0002). None of the other SNPs in PON1 or PON2 and PON3 was significantly associated with stroke in this cohort, and CARE subjects were not polymorphic for the Tyr233Cys polymorphism in PON3. Because of the large number of subjects without stroke, we had over 80% power to detect allele frequency differences of 0.1 or greater with a probability value of 0.05 for an allele with frequency 1 to 20% in the unaffected subjects. None of the SNPs was associated with plasma lipid and glucose levels (Table II available online only at http://www.strokeaha.org).
|
|
The PON1 association described here was followed up with multivariate logistic regression to determine if the association between the Gln192Arg SNP and stroke was independent of other known risk factors for stroke such as diabetes, hypertension, and smoking (Table 3). In the multivariate analysis, age, diabetes, hypertension, smoking, and LDL levels were significant predictors of stroke. The association between the Gln192Arg SNP and stroke persisted in this analysis. The odds ratios were 2.28 (95% CI, 1.38 to 3.79) and 2.47 (95% CI, 1.18 to 5.19) for Gln/Arg heterozygotes and Arg/Arg homozygotes, respectively. When the 7 cases of hemorrhagic stroke were excluded, the strength of the association was largely unchanged (P=0.002), and the odds ratios for Gln/Arg and Arg/Arg increased to 2.41 (95% CI, 1.42 to 4.10) and 2.84 (95% CI, 1.33 to 6.01), respectively. The increase in risk conferred by the Arg allele was comparable to that conferred by diabetes (2.54; 95% CI, 1.48 to 4.34), smoking (2.16; 95% CI, 1.26 to 3.71), and hypertension (1.95; 95% CI, 1.22 to 3.13).
|
Because previous studies suggested that smoking and diabetes can modify the effect of this SNP,26–28 we examined if there was interaction between these traits and Gln192Arg genotype. There was insignificant evidence for interaction between the SNP and smoking (P=0.264) and diabetes (P=0.379) or treatment with pravastatin (P=0.084) with regard to susceptibility to stroke.
We estimated the attributable fraction (AF), ie, the fraction of stroke cases that would not have occurred had the high-risk arginine allele been absent in this population, to be approximately 39%. In contrast, the AFs for diabetes, smoking, and hypertension were lower—17%, 21%, and 29%, respectively. Thus, although the odds ratios associated with these 3 risk factors were comparable to those of the high-risk genotypes of the Gln192Arg SNP, these risk factors individually accounted for a smaller proportion of strokes because of their lower prevalence in this high-risk cohort.
We next examined the frequency of the Arg192 allele in different subgroups within the CARE cohort (Table 4). The frequency of the high-risk Arg192 allele was higher in stroke cases compared with unaffecteds across all subgroups tested. This consistency indicates that the association with stroke is not primarily the result of a single subgroup within the CARE cohort.
|
Sequence analysis of the entire PON1 gene failed to reveal another SNP that was also significantly associated with stroke in the CARE cohort. We sequenced all exons, 100 base pairs of the flanking intron sequence for each exon, and 1000 base pairs of the PON1 promoter from 70 stroke cases and, for comparison, from 32 healthy individuals of diverse ethnic origins. Twenty-three SNPs were identified. CARE subjects were genotyped for 3 SNPs that changed the coding sequence and one that affected a putative NF-1-binding site in the promoter. The other SNPs were in introns or were silent and were not investigated further. An arginine-to-glycine substitution at residue 160 was detected in one individual from the human diversity panel but was absent in CARE subjects. An alanine/valine substitution at residue 201 and the A-161G SNP in the putative NF-1-binding site in the promoter region were not associated with stroke (Table 2). A nonsense mutation at codon 32 (Arg32Stop) was detected in only one individual who had a stroke but was not detected in the remaining subjects (Table 2).
Three other independent studies investigated whether PON1 Gln192Arg genotype was associated with stroke. The frequency of the high-risk Arg192 allele was higher in stroke cases in 2 of these studies (Table 5), including one that examined a different ethnic group (Japanese). One group examined the same cohort twice with different PON1 SNPs, and only their first study was considered in this analysis.11,12 In a small study from Croatia that examined 56 stroke cases and 124 controls, there was no significant difference in the frequency of the Arg192 allele between the 2 groups.10 There was no evidence of heterogeneity of odds ratios across the 4 studies (23df=4.76, P=0.19), however, and the combined odds ratio for the Arg192 allele versus Gln192 allele was 1.64 (95% CI, 1.39 to 1.94). When probability values for allele frequency differences were combined across the 4 studies, the association between the Gln192Arg SNP and stroke was highly significant (28df=45.58, P<0.000001).
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
In a comprehensive analysis of the paraoxonases as candidate stroke genes, we found that the Gln192Arg SNP of the PON1 gene significantly increased risk of stroke. These results are consistent with 2 other published studies. This reproducibility across 3 independent studies and in 2 different ethnic groups—white and Japanese—is reassuring and suggests that the association is unlikely to be spurious. However, in a fourth study, the difference in frequency of the high-risk Arg192 allele between stroke cases and controls was not significant.10 This lack of association could be the result of differences in diagnostic criteria for stroke or population differences.
Two different experimental approaches indicate that the Gln192Arg SNP is functional. It affects the substrate specificity of PON1 in vitro. Gln192-bearing PON1 has substantially greater specificity for paraoxon, whereas Arg192-bearing PON1 has a higher specificity for diazooxon.24 In vitro studies showed also that Gln192-bearing PON1 was better than Arg192-bearing PON1 at protecting LDL from oxidation.29,30 Together with our results, these functional studies suggest that the Gln192Arg SNP may be causative. However, we cannot exclude the possibility that this SNP is in linkage disequilibrium with another causative but as yet undetected variant.
Several studies found an association between the Gln192Arg SNP and increased risk of MI.7 Because all subjects enrolled in the CARE trial had a history of MI, one might expect that the frequency of the Arg192 allele would be higher in this cohort than in the controls free from MI that were used in the other studies that also examined whites (Table 5). In fact, the frequency of the Arg192 allele in the overall CARE trial was higher (0.31 vs 0.28). However, among subjects without stroke but with a history of MI in the CARE trial, the frequency of Arg192 is almost identical to the controls in the other studies (0.28 to 0.29). One speculative explanation for this observation is that the controls used in the other studies were too young to have had an MI or had undiagnosed CAD.
Our study has several limitations. First, we genotyped a high-risk cohort in which all individuals had a history of MI. Second, there was a relatively small number of stroke cases and women. Finally, only whites were analyzed in this study. For these reasons, additional studies using large population-based cohorts will be needed to determine the general applicability of our results.
|
Acknowledgments |
|---|
This study was funded by Bristol-Myers Squibb.
Received June 30, 2005; revision received July 19, 2005; accepted July 27, 2005.
|
References |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
- Primo-Parmo SL, Sorenson RC, Teiber J, La Du BN. The human serum paraoxonase/arylesterase gene (PON1) is one member of a multigene family. Genomics. 1996; 33: 498–507.[CrossRef][Medline] [Order article via Infotrieve]
- Blatter MC, James RW, Messmer S, Barja F, Pometta D. Identification of a distinct human high-density lipoprotein subspecies defined by a lipoprotein-associated protein, K-45. Identity of K-45 with paraoxonase. Eur J Biochem. 1993; 211: 871–879.[Medline] [Order article via Infotrieve]
- Mackness MI, Arrol S, Abbott C, Durrington PN. Protection of low-density lipoprotein against oxidative modification by high-density lipoprotein associated paraoxonase. Atherosclerosis. 1993; 104: 129–135.[CrossRef][Medline] [Order article via Infotrieve]
- Watson AD, Berliner JA, Hama SY, La Du BN, Faull KF, Fogelman AM, Navab M. Protective effect of high density lipoprotein associated paraoxonase. Inhibition of the biological activity of minimally oxidized low density lipoprotein. J Clin Invest. 1995; 96: 2882–2891.[Medline] [Order article via Infotrieve]
- Reddy ST, Wadleigh DJ, Grijalva V, Ng C, Hama S, Gangopadhyay A, Shih DM, Lusis AJ, Navab M, Fogelman AM. Human paraoxonase-3 is an HDL-associated enzyme with biological activity similar to paraoxonase-1 protein but is not regulated by oxidized lipids. Arterioscler Thromb Vasc Biol. 2001; 21: 542–547.
[Abstract/Free Full Text] - Ng CJ, Wadleigh DJ, Gangopadhyay A, Hama S, Grijalva VR, Navab M, Fogelman AM, Reddy ST. Paraoxonase-2 is a ubiquitously expressed protein with antioxidant properties and is capable of preventing cell-mediated oxidative modification of low density lipoprotein. J Biol Chem. 2001; 276: 44444–44449.
[Abstract/Free Full Text] - Wheeler JG, Keavney BD, Watkins H, Collins R, Danesh J. Four paraoxonase gene polymorphisms in 11212 cases of coronary heart disease and 12786 controls: meta-analysis of 43 studies. Lancet. 2004; 363: 689–695.[CrossRef][Medline] [Order article via Infotrieve]
- Robertson KS, Hawe E, Miller GJ, Talmud PJ, Humphries SE, Northwick Park Heart Study II. Human paraoxonase gene cluster polymorphisms as predictors of coronary heart disease risk in the prospective Northwick Park Heart Study II. Biochim Biophys Acta. 2003; 1639: 203–212.[Medline] [Order article via Infotrieve]
- Imai Y, Morita H, Kurihara H, Sugiyama T, Kato N, Ebihara A, Hamada C, Kurihara Y, Shindo T, Oh-hashi Y, Yazaki Y. Evidence for association between paraoxonase gene polymorphisms and atherosclerotic diseases. Atherosclerosis. 2000; 149: 435–442.[CrossRef][Medline] [Order article via Infotrieve]
- Topic E, Timundic AM, Ttefanovic M, Demarin V, Vukovic V, Lovrencic-Huzjan A, Zuntar I. Polymorphism of apoprotein E (APOE), methylenetetrahydrofolate reductase (MTHFR) and paraoxonase (PON1) genes in patients with cerebrovascular disease. Clin Chem Lab Med. 2001; 39: 346–350.[CrossRef][Medline] [Order article via Infotrieve]
- Voetsch B, Benke KS, Damasceno BP, Siqueira LH, Loscalzo J. Paraoxonase 192 Gln/Arg polymorphism: an independent risk factor for nonfatal arterial ischemic stroke among young adults. Stroke. 2002; 33: 1459–1464.
[Abstract/Free Full Text] - Voetsch B, Benke KS, Panhuysen CI, Damasceno BP, Loscalzo J. The combined effect of paraoxonase promoter and coding region polymorphisms on the risk of arterial ischemic stroke among young adults. Arch Neurol. 2004; 61: 351–366.
[Abstract/Free Full Text] - Sacks FM, Pfeffer MA, Moye LA, Rouleau JL, Rutherford JD, Cole TG, Brown L, Warnica JW, Arnold JM, Wun CC, Davis BR, Braunwald E. The effect of pravastatin on coronary events after myocardial infarction in patients with average cholesterol levels. Cholesterol and Recurrent Events Trial investigators. N Engl J Med. 1996; 335: 1001–1009.
[Abstract/Free Full Text] - Plehn JF, Davis BR, Sacks FM, Rouleau JL, Pfeffer MA, Bernstein V, Cuddy TE, Moye LA, Piller LB, Rutherford J, Simpson LM, Braunwald E. Reduction of stroke incidence after myocardial infarction with pravastatin: the Cholesterol and Recurrent Events (CARE) study. The Care Investigators. Circulation. 1999; 99: 216–223.
[Abstract/Free Full Text] - Ranade K, Chang MS, Ting CT, Pei D, Hsiao CF, Olivier M, Pesich R, Hebert J, Chen YD, Dzau VJ, Curb D, Olshen R, Risch N, Cox DR, Botstein D. High-throughput genotyping with single nucleotide polymorphisms. Genome Res. 2001; 11: 1262–1268.
[Abstract/Free Full Text] - Germer S, Holland MJ, Higuchi R. High-throughput SNP allele-frequency determination in pooled DNA samples by kinetic PCR. Genome Res. 2000; 10: 258–266.
[Abstract/Free Full Text] - Begovich AB, Carlton VE, Honigberg LA, Schrodi SJ, Chokkalingam AP, Alexander HC, Ardlie KG, Huang Q, Smith AM, Spoerke JM, Conn MT, Chang M, Chang SY, Saiki RK, Catanese JJ, Leong DU, Garcia VE, McAllister LB, Jeffery DA, Lee AT, Batliwalla F, Remmers E, Criswell LA, Seldin MF, Kastner DL, Amos CI, Sninsky JJ, Gregersen PK. A missense single-nucleotide polymorphism in a gene encoding a protein tyrosine phosphatase (PTPN22) is associated with rheumatoid arthritis. Am J Hum Genet. 2004; 75: 330–337.[CrossRef][Medline] [Order article via Infotrieve]
- Fisher RA. Statistical Methods for Research Workers, 13th ed. London: Oliver and Lloyd, Ltd; 1925
- Fisher LD, van Belle G. Biostatistics: A Methodology for the Health Sciences. New York: Wiley-Interscience; 1993.
- Khoury M, Beaty TH, Cohen B. Fundamentals of Genetic Epidemiology. New York: Oxford University Press; 1993.
- Adkins S, Gan KN, Mody M, La Du BN. Molecular basis for the polymorphic forms of human serum paraoxonase/arylesterase: glutamine or arginine at position 191, for the respective A or B allozymes. Am J Hum Genet. 1993; 52: 598–608.[Medline] [Order article via Infotrieve]
- Humbert R, Adler DA, Disteche CM, Hassett C, Omiecinski CJ, Furlong CE. The molecular basis of the human serum paraoxonase activity polymorphism. Nat Genet. 1993; 3: 73–76.[CrossRef][Medline] [Order article via Infotrieve]
- Leviev I, James RW. Promoter polymorphisms of human paraoxonase PON1 gene and serum paraoxonase activities and concentrations. Arterioscler Thromb Vasc Biol. 2000; 20: 516–521.
[Abstract/Free Full Text] - Jarvik GP, Jampsa R, Richter RJ, Carlson CS, Rieder MJ, Nickerson DA, Furlong CE. Novel paraoxonase (PON1) nonsense and missense mutations predicted by functional genomic assay of PON1 status. Pharmacogenetics. 2003; 13: 291–295.[CrossRef][Medline] [Order article via Infotrieve]
- Sanghera DK, Aston CE, Saha N, Kamboh MI. DNA polymorphisms in two paraoxonase genes (PON1 and PON2) are associated with the risk of coronary heart disease. Am J Hum Genet. 1998; 62: 36–44.[CrossRef][Medline] [Order article via Infotrieve]
- Sen-Banerjee S, Siles X, Campos H. Tobacco smoking modifies association between Gln-Arg192 polymorphism of human paraoxonase gene and risk of myocardial infarction. Arterioscler Thromb Vasc Biol. 2000; 20: 2120–2126.
[Abstract/Free Full Text] - Senti M, Aubo C, Tomas M. Differential effects of smoking on myocardial infarction risk according to the Gln/Arg 192 variants of the human paraoxonase gene. Metabolism. 2000; 49: 557–559.[CrossRef][Medline] [Order article via Infotrieve]
- Koch M, Hering S, Barth C, Ehren M, Enderle MD, Pfohl M. Paraoxonase 1 192 Gln/Arg gene polymorphism and cerebrovascular disease: interaction with type 2 diabetes. Exp Clin Endocrinol Diabetes. 2001; 109: 141–145.[CrossRef][Medline] [Order article via Infotrieve]
- Mackness B, Mackness MI, Arrol S, Turkie W, Durrington PN. Effect of the human serum paraoxonase 55 and 192 genetic polymorphisms on the protection by high density lipoprotein against low density lipoprotein oxidative modification. FEBS Lett. 1998; 423: 57–60.[CrossRef][Medline] [Order article via Infotrieve]
- Aviram M, Billecke S, Sorenson R, Bisgaier C, Newton R, Rosenblat M, Erogul J, Hsu C, Dunlop C, La Du B. Paraoxonase active site required for protection against LDL oxidation involves its free sulfhydryl group and is different from that required for its arylesterase/paraoxonase activities: selective action of human paraoxonase allozymes Q and R. Arterioscler Thromb Vasc Biol. 1998; 18: 1617–1624.
[Abstract/Free Full Text]
This article has been cited by other articles:
 |
|
 |
|
  M. Saeed, N. Siddique, W. Y. Hung, E. Usacheva, E. Liu, R. L. Sufit, S. L. Heller, J. L. Haines, M. Pericak-Vance, and T. Siddique Paraoxonase cluster polymorphisms are associated with sporadic ALS Neurology, September 12, 2006; 67(5): 771 - 776. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||