Introduction
Obstructive sleep apnoea hypopnoea syndrome is the most common sleep disorder, and is related to abnormal breathing due to repetitive upper airway collapse during sleep; this results in snoring, sleep-related decreases and pauses in respiration, transient hypoxia, sleep fragmentation, and excessive daytime sleepiness.Reference Young, Peppard and Gottlieb1 Subjects with obstructive sleep apnoea hypopnoea syndrome are at high risk of developing type two diabetes mellitus, insulin resistance, central obesity and dyslipidaemia.Reference West, Nicoll and Stradling2–Reference Suzuki, Nakano, Maekawa, Okamoto, Ohnishi and Yamauchi5
A number of previous studies have proposed that several genes may be involved in the pathogenesis of obstructive sleep apnoea hypopnoea syndrome, such as those coding for tumour necrosis factor alpha, angiotensin-converting enzyme, leptin and leptin receptor.Reference Riha, Brander, Vennelle, McArdle, Kerr and Anderson6–Reference Hanaoka, Yu, Urushihata, Ota, Fujimoto and Kubo8 However, no specified genetic influence on the syndrome's development has previously been identified. Genes involved in metabolic disorders (e.g. type two diabetes mellitus, insulin resistance, obesity and dyslipidaemia) may be relevant to the development of obstructive sleep apnoea hypopnoea syndrome, because of (1) the prominence of metabolic disorders in the obstructive sleep apnoea hypopnoea syndrome phenotype, and (2) the potential impact of these genes on the expression of other traits with potential relevance to the syndrome. One study indicated that obstructive sleep apnoea hypopnoea syndrome may be significantly associated with genes affecting insulin resistance and body fat distribution.Reference Kaparianos, Sampsonas, Karkoulias and Spiropoulos9
The peroxisome proliferator activated receptor gamma gene, on chromosome 3p25, encodes a nuclear protein which is essential for adipocyte differentiation and lipid storage, and which also modulates the effect of insulin. Polymorphisms of this gene have been shown to cause functional abnormalities. For example, many studies have found that the substitution of alanine for proline (a common variant termed Pro12Ala) in exon 2 of the peroxisome proliferator activated receptor gamma gene is associated with type two diabetes mellitus, insulin resistance, obesity and dyslipidaemia.Reference Ringel, Engeli, Distler and Sharma10–Reference Masud and Ye13 These disorders are common among patients with obstructive sleep apnoea hypopnoea syndrome. Therefore, this common variant of the peroxisome proliferator activated receptor gamma gene may influence the development of obstructive sleep apnoea hypopnoea syndrome; it may also be a useful biomarker for the syndrome.
Therefore, we analysed the frequencies of genotypes and alleles of the Pro12Ala polymorphism (also known as rs1801282) of the peroxisome proliferator activated receptor gamma gene, in a total of 610 subjects of Han Chinese ethnic origin (in Shanghai, China) with and without obstructive sleep apnoea hypopnoea syndrome, in order to examine the association between this genetic variant and the syndrome.
Subjects and methods
Subjects
The study was approved by the Institutional Ethics Committee of the Hospital of Shanghai Jiao Tong University. All participants gave their informed consent to involvement in the study, together with written informed consent to phlebotomy.
This was a case–control study including a total of 420 unrelated patients (63 women and 357 men) with obstructive sleep apnoea hypopnoea syndrome who were admitted to the Sleep Centre of the Affiliated Sixth People's Hospital, Shanghai Jiao Tong University, for overnight sleep studies between January 2007 and October 2008. None of the patients had previously been treated for obstructive sleep apnoea hypopnoea syndrome. The patients were diagnosed with obstructive sleep apnoea hypopnoea syndrome based on clinical history (i.e. snoring, apnoeic periods during sleep and daytime sleepiness), otorhinolaryngological examination (including fibre-optic laryngoscopy), Muller manoeuvre and polysomnography, and were further subdivided into three subgroups as detailed below.
We excluded from the study any patients with sleep disorders other than obstructive sleep apnoea hypopnoea syndrome, such as upper airway resistance syndrome, restless leg syndrome and narcolepsy. We also excluded any patients with congestive heart failure, intrinsic pulmonary disease, or intrinsic renal or liver disease, and any taking systemic steroids or hormonal replacement therapy.
Control subjects comprised 190 randomly selected, healthy volunteers (39 women and 151 men) who had undergone annual physical examinations. These individuals obtained passing scores on either overnight polysomnography or the Epworth Sleepiness Scale and Berlin questionnaire, in order to exclude sleep-related breathing disorders.
The study and control groups were carefully matched for age (with means ± standard deviations (SDs) of 44.15 ± 13.35 and 44.54 ± 10.58 years, respectively; p > 0.05), body mass index (BMI) (with means ± SDs of 27.23 ± 3.46 and 26.89 ± 3.68 kg/m2, respectively; p > 0.05) and sex, in order to eliminate any potential biasing effect of these factors.
All subjects were of the same ethnic origin. None of the study or control group subjects exhibited mental retardation, drug dependence, alcoholism, or somatic or neurological illness, as determined by the clinical history. No other otolaryngological disorders were found during physical examination. All subjects completed a questionnaire on demographics, sleep symptoms, medical history and medications.
Patients with obstructive sleep apnoea hypopnoea syndrome were further divided into three subgroups based on syndrome severity, either mild (i.e. an apnoea–hypopnoea index (AHI) of between five and less than 15 apnoeic episodes per hour), moderate (an AHI of between 15 and less than 30 episodes per hour) or severe (an AHI of 30 episodes or more per hour). These subgroups contained 58, 66 and 296 patients, respectively.
Polysomnography and anthropometric measurements
Polysomnography (using Alice 4™ apparatus; Respironics, Pittsburgh, Pennsylvania, USA) was performed manually according to standard criteria. Respiratory events were scored according to American Academy of Sleep Medicine criteria.14 Apnoea was defined as complete cessation of airflow lasting 10 seconds or more. Hypopnoea was defined as either an airflow reduction of 50 per cent or more for 10 seconds or longer, or a discernible airflow reduction of 50 per cent or less accompanied by either an arousal or oxyhaemoglobin saturation decrease of 4 per cent or more. The AHI was defined as the number of apnoeic or hypopnoeic events per hour during sleep time, based on the results of overnight polysomnography.
Body habitus was measured in light clothing and bare feet using standard anthropometric methods. Waist circumference was measured midway between the lower costal margin and the iliac crest. Hip circumference was taken as the maximal girth at the level of the greater trochanters. The neck circumference was measured at the level of the cricothyroid membrane while the subject was standing. All data were documented as the mean of two measurements. Body mass index was calculated as the body weight divided by the height squared, expressed as kg/m2.
Laboratory tests and genetic analysis
Venous blood was obtained for DNA and biochemical analysis.
Glucose and lipid levels were measured by enzymatic assay, using the H-7600 autoanalyser (Hitachi, Tokyo, Japan).
An immunoradiology method was adapted to measure fasting serum insulin. Insulin resistance was estimated using the homeostasis model assessment method (i.e. fasting serum insulin (μU/ml) × fasting plasma glucose (mmol/l)/22.5), as previously described.Reference Bonora, Targher, Alberiche, Bonadonna, Saggiani and Zenere15
Following extraction of DNA from a 200 µl blood sample using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany), the Pro12Ala peroxisome proliferator activated receptor gamma polymorphism was determined using polymerase chain reaction based restriction fragment length polymorphism, as previously described.Reference Prasad, Saxena, Ghoshal, Bhagat and Krishnani16 Polymerase chain reaction was performed in a total volume of 15 µl containing 50 ng DNA, 200 µm Deoxynucleotide Triphosphates (dNTPs), 20 pmol of each primer, and 1.5 μL 10 × polymerase chain reaction buffer (New England Biolabs, Ipswich, Massachusetts, USA), plus 1.25 U Taq polymerase (MBI Fermentas, Glen Burnie, Maryland, USA). Amplification was performed on an automated thermal cycler (Bio-Rad, Hercules, California, USA). Polymerase chain reaction conditions were: 5 minutes for initial denaturation at 95°C; 32 cycles of 30 seconds at 95°C for denaturation, 30 seconds at 55°C for annealing, and 30 seconds at 72°C for extension; followed by a final extension of 10 minutes at 72°C. Polymerase chain reaction products were digested with 1 U of BstU I restriction enzymes (New England Biolabs) for 4 hours at 37°C, and then separated on a 2 per cent agarose gel containing 0.5 µg/ml ethidium bromide. Fragment sizes of 227 and 43 basepairs indicated the presence of a homozygous guanine–guanine (GG) genotype (Ala12Ala), a single 270 basepair fragment indicated the presence of a wild-type homozygous cytosine–cytosine (CC) genotype (Pro12Pro), and three fragments of 270, 227 and 43 basepairs indicated the presence of a heterozygous CG genotype (Pro12Ala).
All experiments were repeated twice to confirm the restriction fragment length polymorphism results.
Statistical analysis
The chi-square test or Fisher's exact test was applied to compare genotypes and allele frequencies in patients and controls. Descriptive characteristics of group variables were expressed as mean ± SD. Spearman's rank correlation tests were also applied to test for any correlation between genotypes and clinical parameters. The Mann–Whitney test or group t-test was applied to compare the phenotypic values of patients with different genotypes (i.e. Pro/Pro and carriers of the Ala12 allele). A general linear model was used to control for age, sex, BMI and waist circumference, as appropriate. Hardy–Weinberg equilibrium was tested using the chi-square test. Two-tailed p values of less than 0.05 were considered to indicate statistical significance. Statistical analyses were performed using the Statistical Package for the Social Sciences for Windows version 13.0.0 software program (SPSS Inc, Chicago, Illinois, USA).
Results
In the control group, the CC, CG, and GG variants were found in 175 (92.1 per cent), 15 (7.9 per cent) and 0 subjects, respectively. In the obstructive sleep apnoea hypopnoea syndrome group, these same variants were found in 372 (88.6 per cent), 47 (11.2 per cent) and one (0.2 per cent) patients, respectively.
The few Ala/Ala individuals were included in the Pro/Ala group so that the genotype could be examined as a dichotomous variable. The frequencies of C and G alleles were 0.961 and 0.039 in the control group and 0.942 and 0.058 in the study group, respectively. No significant difference was observed between the genotype and allele frequencies, comparing patients and controls (p > 0.05; Table I). Gender-specific comparisons revealed no significant differences (p > 0.05; Table II). Comparison by AHI did not reveal any differences across the three syndrome severity subgroups, and no difference was seen between any of these three subgroups and the controls (p > 0.05; Table III). The percentage of the male subjects carrying the Pro/Pro (84.7 per cent, 315/372) did not differ significantly with the carrying Ala12 alleles group (87.5 per cent, 42/48) in the OSAS (p > 0.05; Table IV).
Table I Genotypes and alleles of PPARG PRO12ALA polymorphism, by group*
* p > 0.05, chi-square test. PPARG = peroxisome proliferator activated receptor gamma; C = cytosine; G = guanine
Table II Genotypes and alleles of PPARG PRO12ALA polymorphism, by group and gender*
* p > 0.05, chi-square test or Fisher's exact test. PPARG = peroxisome proliferator activated receptor gamma; C = cytosine; G = guanine
Table III Genotypes and alleles of PPARG PRO12ALA polymorphism, by ahi and group*
* p > 0.05, chi-square test or Fisher's exact test. PPARG = peroxisome proliferator activated receptor gamma; AHI = apnoea–hypopnoea index; C = cytosine; G = guanine
Table IV PPARG PRO12ALA polymorphism genotype by measured osahs patient parameters
Data represent means ± standard deviations unless otherwise specified. *n=372; †n = 48. ‡Chi-square test; **group t-test; §Mann–Whitney test. #By general linear model adjusted for the covariates age, sex, body mass index (BMI), and waist circumference. PPARG = peroxisome proliferator activated receptor gamma; OSAHS = obstructive sleep apnoea hypopnoea syndrome; C = cytosine; G = guanine; circ = circumference; BP = blood pressure; TC = total cholesterol; TG = serum triglyceride; HDL-C = high density lipoprotein cholesterol; LDL-C = low density lipoprotein cholesterol; Apo = Apolipoprotein; HOMA-IR = insulin resistance calculated by homeostasis model assessment method; hr = hour; min SaO2 = minimum oxygen saturation; CT90% = time during which oxygen saturation is 90% or less, expressed as percentage of total sleep time
In the study group, no correlation was detected between genotype and the various polysomnographic and anthropometric parameters measured, based on Spearman's rank correlation analysis (p > 0.05), except for high density lipoprotein cholesterol (HDL-C) (p = 0.032). Study group patients showed no significant association between genotype and age, BMI, waist circumference, fasting blood glucose (FBG), total cholesterol concentration, triglyceride concentration or insulin resistance (calculated by the homeostasis model assessment method). Study group patients with the Ala12 allele had higher HDL-C levels (p = 0.031). However, when this result was adjusted for covariates (i.e. age, sex, BMI and waist circumference) using a general linear model, the association between the Ala12 allele and the HDL-C level was no longer significant (p = 0.176) (Table IV).
Subjects were further examined according to hyperlipidaemic status. Hyperlipidaemia was defined as follows: total cholesterol concentration greater than 5.2 mmol/l; low density lipoprotein (LDL) cholesterol greater than 3.5 mmol/l; and triglyceride concentration greater than 2.3 mmol/l.Reference Swarbrick, Chapman, McQuillan, Hung, Thompson and Beilby17
In the study group, 230 patients were identified as hyperlipidaemic (196 with the Pro/Pro allele and 34 with the Ala12 allele), while 190 were not (176 with the Pro/Pro allele and 14 with the Ala12 allele). The risk of being hyperlipidaemic was strongly associated with genotype (risk estimate odds ratio = 2.181; p = 0.017, 95 per cent confidence interval (CI) = 1.133–4.198; Table V).
Table V Presence of hyperlipidaemia in osahs patients, by PPARG PRO12ALA polymorphism
*n=230; †n = 190; ‡n = 420. The Ala12Pro variant was significantly associated with hyperlipidaemia (HLD) (odds ratio = 2.181, p = 0.017, 95% confidence interval = 1.133–4.198; chi-square test). OSAHS = obstructive sleep apnoea hypopnoea syndrome; PPARG = peroxisome proliferator activated receptor gamma; Ala12 = Ala12Pro + Ala12Ala
In the control group, 68 subjects were hyperlipidaemic (62 with the Pro/Pro allele and six with the Ala12 allele), while 122 were not (113 with the Pro/Pro allele and nine with the Ala12 allele). Amongst controls, there was no association between hyperlipidaemia and genotype (odds ratio = 0.823, p = 0.782, 95 per cent CI = 0.28–2.42; Table VI).
Table VI Presence of hyperlipidaemia in controls, by PPARG PRO12ALA polymorphism
*n=68; †n = 122; ‡n = 190. The Ala12Pro variant was not significantly associated with hyperlipidaemia (HLD) (odds ratio = 0.823, p = 0.782, 95% confidence interval = 0.28–2.42; Fisher's exact test). PPARG = peroxisome proliferator activated receptor gamma; Ala12 = Ala12Pro + Ala12Ala
Discussion
Candidate genes linked with metabolic disorders may also be relevant to the genetics of obstructive sleep apnoea hypopnoea syndrome, because of (1) the prominence of metabolic disorders (e.g. type two diabetes mellitus, insulin resistance, obesity and dyslipidaemia) in the syndrome phenotype, and (2) the possible impact of these genes on the expression of other traits of potential relevance to the syndrome.
The peroxisome proliferator activated receptor gamma is a key regulator of lipid metabolism and glucose homeostasis, because it is involved in adipocyte differentiation, fatty acid uptake and storage, and the transcriptional regulation of many genes involved in these processes.Reference Lehrke and Lazar18, Reference Evans, Barish and Wang19
The Pro12Ala polymorphism of peroxisome proliferator activated receptor gamma has attracted much attention. Its importance is indicated by its correlation with various metabolic diseases, such as obesity, type two diabetes and dyslipidaemia.Reference Barcela, Elorza, Barbe, Santos, Mayoralas and Agusti7–Reference Kaparianos, Sampsonas, Karkoulias and Spiropoulos9 To our knowledge, the present study is the first to investigate the potential association of the Pro12Ala polymorphism with obstructive sleep apnoea hypopnoea syndrome in Chinese patients.
Our initial overall comparison did not reveal any significant differences between the obstructive sleep apnoea hypopnoea syndrome patients and the controls in terms of genotype and allele frequencies (p > 0.05); the Ala12 variant was not associated with gender or AHI. This suggests that other mechanisms may be involved in the relationship between obstructive sleep apnoea hypopnoea syndrome and metabolic disease. Our finding also (further) supports the idea that obstructive sleep apnoea hypopnoea syndrome is a complex disorder which is related to or affected by multiple genes, as well as environmental and developmental factors.
The Ala12 variant of peroxisome proliferator activated receptor gamma is associated with preferential accumulation of polyunsaturated fat in adipose tissue, in response to polyunsaturated fat intake, and also affects the regulation of fatty acid uptake and storage; thus, the Ala12 variant affects the distribution of body fat.Reference Swarbrick, Chapman, McQuillan, Hung, Thompson and Beilby17 Various relationships have been reported between the Ala12 allele and BMI, a major quantitative trait, with different research groups reporting inconsistent findings.Reference Lehrke and Lazar18, Reference Deeb, Fajas, Nemoto, Pihlajamäki, Mykkänen and Kuusisto20, Reference Gonzalez, Serrano, Fernandez, Laakso and Martínez Larrad21 A meta-analysis revealed a significant association between the Ala12 allele and increased BMI in Caucasians.Reference Tonjes, Scholz, Loeffler and Stumvoll22
In order to determine whether obstructive sleep apnoea hypopnoea syndrome is associated with the Pro12Ala polymorphism, and to assess this variant's overall prevalence independent of obesity, it would have been more appropriate for the present study to employ a matched case–control design, using BMI as the matching variable. However, our strict exclusion criteria enabled the present study to recruit only 190 controls, from 623 willing volunteers. We found no correlation between the Ala12 allele and BMI in our obstructive sleep apnoea hypopnoea syndrome patients. The mean ± SD BMI of our study group patients (27.23 ± 3.46 kg/m2) was much lower than that reported for obese subjects in a previous study (32.9 ± 2.6 kg/m2).Reference Swarbrick, Chapman, McQuillan, Hung, Thompson and Beilby17 Taking into account different populations and genotypic milieu, our results are not directly comparable with those reported by other authors for different individuals and diseases.
Swarbrick et al. reported that obese subjects with the Pro/Ala and Ala/Ala alleles exhibited lower HDL levels (p = 0.032) and showed a trend toward higher triglyceride levels (p = 0.055), compared with obese subjects with the Pro/Pro allele.Reference Swarbrick, Chapman, McQuillan, Hung, Thompson and Beilby17 In these obese subjects, the Ala12 allele was significantly associated with combined hyperlipidaemia (odds ratio = 2.33, p = 0.042).
Beamer et al. have investigated the association between the Pro12Ala allele, HDL cholesterol and triglycerides, in a group of 57 extremely obese men.Reference Beamer, Yen, Andersen, Muller, Elahi and Cheskin23 These authors found that subjects with the Ala12 allele had lower HDL and higher triglyceride levels, compared with subjects with the Pro/Pro allele. However, this effect was not observed in 112 extremely obese women, nor in a cohort of 517 lean to moderately obese individuals.
In the present study, obstructive sleep apnoea hypopnoea syndrome patients with the Ala12 allele were found to have higher HDL-C levels, compared with patients with the Pro/Pro allele (p = 0.031), whereas total cholesterol, triglyceride and LDL levels did not show the same trend (p > 0.05). When these results were adjusted for covariates (i.e. age, sex, BMI and waist circumference) using a general linear model, the association between the Ala12 allele and HDL cholesterol levels became insignificant (p = 0.176).
These findings suggest that the Pro12Ala polymorphism of peroxisome proliferator activated receptor gamma does not affect HDL cholesterol levels in patients with obstructive sleep apnoea hypopnoea syndrome. Lower HDL cholesterol levels and higher triglyceride levels are both known risk factors for coronary artery disease, and the presence of the Ala12 allele may provide some defence against this condition by altering the lipid profile. However, in patients with obstructive sleep apnoea hypopnoea syndrome, HDL cholesterol and triglyceride levels did not differ significantly, comparing patients with the Pro/Pro and Pro/Ala alleles.
Dyslipidaemia can be responsible, at least in part, for metabolic alterations which are commonly associated with obstructive sleep apnoea hypopnoea syndrome and which increase the risk of cardiovascular events and death. Effective assessment and management of obstructive sleep apnoea hypopnoea syndrome may correct endocrine changes, improve quality of life, and prevent associated morbidity or death.
Swarbrick et al. found that the Pro12Ala polymorphism of peroxisome proliferator activated receptor gamma was significantly associated with an altered lipid profile in obese carriers.Reference Swarbrick, Chapman, McQuillan, Hung, Thompson and Beilby17 For such patients, this common polymorphism may constitute a novel genetic risk factor for combined hyperlipidaemia and cardiovascular disease.
We investigated the potential connection between this polymorphism and obstructive sleep apnoea hypopnoea syndrome. The odds ratio for the risk of hyperlipidaemia in the presence of the Pro12Ala allele indicated an association in the obstructive sleep apnoea hypopnoea syndrome group (odds ratio = 2.181, p = 0.017, 95 per cent CI = 1.133–4.198) but not in the control group (odds ratio = 0.823, p = 0.782, 95 per cent CI = 0.28–2.42).
• This study assessed the association between the Pro12Ala polymorphism of peroxisome proliferator activated receptor gamma and obstructive sleep apnoea hypopnoea syndrome and hyperlipidaemia in a Chinese population
• Genetic analysis was performed for 420 patients and 190 healthy controls
• Three distinct genotype variants (Pro12Pro, Pro12Ala and Ala12Ala) were assessed
• The Pro12Ala variant was significantly associated with hyperlipidaemia in obstructive sleep apnoea hypopnoea syndrome patients; no other significant associations were found
These results suggest that the Pro12Ala polymorphism of peroxisome proliferator activated receptor gamma may be a predisposing factor for, and predictor of, hyperlipidaemia in patients with obstructive sleep apnoea hypopnoea syndrome. Genetic factors probably interact to produce the obstructive sleep apnoea hypopnoea syndrome phenotype, particularly in middle-aged subjects. The hypoxia, insulin resistance and obesity caused by obstructive sleep apnoea hypopnoea syndrome may have an adverse influence on lipid metabolism, and may also interact with each other. However, more research is needed to clarify the actual mechanism leading to an increased prevalence of dyslipidaemia in patients with obstructive sleep apnoea hypopnoea syndrome.
Conclusion
The Pro12Ala polymorphism of peroxisome proliferator activated receptor gamma is not clearly associated with obstructive sleep apnoea hypopnoea syndrome. However, the Ala12 allele was significantly associated with the presence of hyperlipidaemia in patients with this syndrome. We assume these findings to be valid, as all our patients were assessed by overnight polysomnography, an established diagnostic method for obstructive sleep apnoea hypopnoea syndrome. In obstructive sleep apnoea hypopnoea syndrome patients, this polymorphism may constitute a novel genetic risk factor for hyperlipidaemia. It is hoped that these findings will benefit future genetic studies in humans.
Acknowledgements
We are grateful to Mr Chen Haoyan for his assistance in study design and statistical analysis. This study was supported by grants-in-aid from the Science and Technical Committee of Shanghai in China (07JC14029).
