In a previous study, endogenous catecholamines were involved in the regulation of adipose tissue lipolysis, nonesterified fatty acid distribution, lipoprotein metabolism, obesity, glucose homeostasis, diabetes mellitus, blood pressure and metabolic syndrome 1. Metabolic syndrome was defined as abdominal obesity, hyperglycemia, hypertension, hypertriglyceridemia and low high-density lipoprotein cholesterol (HDL-C). Adrenergic receptors (ARs) can regulate lipid mobilization, energy expenditure and glycogen breakdown through these endogenous catecholamines 1. Catecholamines have been found to act through both β- and α-adrenergic receptors, which mediate their effects through different receptor pathways. Besides, many studies have demonstrated associations between β2 adrenergic receptor (β2-AR) gene polymorphisms and various diseases: obesity, type 2 diabetes (T2DM) 2, metabolic syndrome, metabolic disorders 34 and hypertension 56. Therefore, the β2-AR gene may constitute a potential candidate gene to explain part of the genetic predisposition to human obesity and related traits.

Yang-Feng et al. reported that human β2-AR gene polymorphisms were located on the distal portion of the long arm of chromosome 5 (5q32-q34) 7. Several polymorphisms have been found in the coding region of the β2-AR gene (Gly16Arg, Gln27Glu and Thr164Ile) and in the promoter region (5'LC-Cys19Arg), each of them leading to amino-acid substitution 8. In their studies, Dallongeville et al. found that the Gly16Arg and Gln27Glu variants of the β2-AR gene contribute to metabolic syndrome susceptibility in men 4 Large et al. found that the Gln27Glu polymorphism of the β2-AR gene was markedly associated with obesity with a relatively higher risk for obesity and an odds ratio of approximately 10 9. The 5'LC-Cys19Arg protein modulates receptor translation 10, and Gly16Arg and Gln27Glu alterations of cellular trafficking and receptor desensitization 111213 were found. Furthermore, another study 14 showed a significant association between the three polymorphisms of the β2-AR gene, 5'LC-Cys19Arg, Gly16Arg and Gln27Glu, and triglyceride (TG) and low-density lipoprotein cholesterol (LDL-C) plasma levels. So we hypothesized that the associations of these three single nucleotide polymorphisms (SNPs): 5'LC-Cys19Arg, Gly16Arg and Gln27Glu, and haplotypes of the β2-AR gene with lipid profiles, would be present in our study. In light of these considerations, we conducted a study in an ethnic Chinese elderly population to investigate the role of the three polymorphisms, 5'LC-Cys19Arg, Gly16Arg and Gln27Glu, of the β2-AR gene and related haplotypes in lipid profiles in elderly subjects.

The relationship between elevated serum total cholesterol (TC) and LDL-C concentrations, low HDL-C concentrations, and coronary heart disease in middle age 15161718 and early old age has been established 19. In a past study 20, we established a possible module map among apolipoprotein E (Apo E), phospholipase A2 (PLA2) polymorphism, and lipid profiles in an elderly ethnic Chinese population. The presence of a physiologic balance contributed significantly to homeostasis and compensatory responses regulation of blood HDL-C and LDL-C profiles. In other past studies 2122, we found the protective effect of the APoε2 allele against Alzheimer's disease and the higher risk of Alzheimer's disease in subjects with the APoε4 allele. And in another study 23, β-blockers were found to worsen delayed memory function in people with cognitive impairment, and adrenergic signaling was important for the retrieval of intermediate-term contextual and spatial memories. Furthermore, we thought that ApoE, PLA2 and β2-AR might all affect the lipid profiles and have interactions with each other, and associations with the causes of Alzheimer's disease. This study aimed to further explore the interactions between ApoE, PLA2 and β2-AR gene polymorphisms and lipid profiles in an elderly ethnic Chinese population.

In our study, the ε2 allele and female sex had statistically significant (p = 0.032, p = 0.027) positive effects on HDL-C. The low income group also seemed to have higher HDL-C levels (P = 0.076). In LDL-C, the PLA2 A2 allele and age had significantly (p = 0.037, p = 0.039) negative effects. So, we explored the interactive effects among PLA2 and the Apo E polymorphism, sex and income on HDL-C, using multivariate regression analysis with the Jump 5.1 for Windows software package. We found that the ε2 allele combined with low income had a positive effect on HDL-C (p = 0.0011). We also found that the A2 allele combined with the ε2 allele had a positive effect on LDL-C (p = 0.0002).

In another study 14, Petrone et al. found that 5'LC-Cys¹⁹ homozygous showed higher TG and LDL-C levels compared to 5'LC-Arg¹⁹ homozygous. A similar increase in TG and LDL-C levels was observed for the Gly16Arg polymorphism and the Gln27Glu polymorphism. And, they found that the Cys¹⁹Arg¹⁶Gln²⁷ haplotype determined a significant increase in TG and LDL-C levels compared to the Arg¹⁹Gly¹⁶Glu²⁷ haplotype. After performing the final parsimonious model of multiple linear regression coefficients between the plasma lipid levels, the Gln27Glu polymorphism of the β2-AR gene, Apo E and PLA2 genotypes, and other possibly related factors, the Gln27Glu polymorphism Glu/Glu or Gln/Glu subjects had lower TC levels compared to the Gln27Glu polymorphism Gln/Gln subjects, but no statistically significant difference was noted (p = 0.092) (Table 3). And in different haplotypes of the β2-AR gene associated with plasma lipid levels, we found significantly lower TC levels (p = 0.082) and lower LDL-C levels (p = 0.045) in subjects with the Cys¹⁹Arg¹⁶Glu²⁷ haplotype, and significant lower TC levels (p = 0.06) in subjects with the Cys¹⁹Gly ¹⁶Glu²⁷ haplotype (Table 4). But there were too few subjects with these two haplotypes. We also found higher VLDL-C levels (p = 0.185) and higher TG levels (p = 0.190) in subjects with the Cys¹⁹Gly ¹⁶Gln²⁷ haplotype. Altogether, none of our results showed a statistically significant difference.

However, we believe that the difference in results between our study and the study by Petrone et al.¹⁵ regarding the β2-AR gene may be due to the following reasons. Firstly, their subjects were overweight/obese subjects (body mass index [BMI] >25 kg/m²), who were consecutively recruited from the metabolic day hospital of the Department of Clinical Sciences of their university. Our subjects were selected from southern Taiwan, and were older than 65 years when they were recruited for this study. The difference in subjects between their study and our study may have led to the different results. And we had not let our subjects to check BMI in our study. So we can not discuss the impact of BMI on lipid profiles. On the other hand, we had not use case-control study to evaluate the different lipid profiles between obesity subjects and control cases. Secondly, our study had fewer subjects than theirs, and there were fewer frequencies of some polymorphisms of the β2-AR gene. The fewer subjects may have led to subjects with fewer of some polymorphisms of the β2-AR gene, thus leading to different results. In our study, the lipid profiles of the individual homozygous for the 5'LC-Arg19 allele and the lipid profile of the individual homozygous for the Glu27 allele are identical maybe due to this is the same individual. Thirdly, in their study, they only defined the frequency of each haplotype, and not the haplotypes of every subject. In our study, we used Phase Version 2.1 with Baysin analysis to define the haplotypes of all subjects to reduce the false positive rate 23.

On the other hand, we were also concerned that epistasis may present between multiple genes in lipid profiles. The low income group seemed to have a higher HDL-C level (p = 0.076) in our study, but in spite of this result not reaching statistically significant difference, we still wanted to add the income factor. After adding the income factor in the study, we found that the ε2 allele combined with low income had a positive effect on HDL-C (p = 0.0011). In our past study 20, we also found the opposite effects of the ε2 allele and the PLA2 A2 allele on HDL-C and LDL-C. The presence of the ε2 allele with the PLA2 A2 allele appeared to significantly decrease HDL-C, and the PLA2 A2 allele with the presence of the ε2 allele increased LDL-C.

We concluded that not only do biological factors like Apo E, PLA2 and β2-AR affect lipid profiles, but social factors like income may also affect lipid profiles. We need to explore more of the effects that can affect lipid profiles to establish a clearer model to explain the interaction between genetic factors, social factors, and lipid profiles in the future. Using the model, if we can find high risk groups for hyperlipidemia and related disorders like metabolic syndrome, diabetes, cardiovascular disease, and obesity, and adopt adequate solutions, the prevalence of these disorders may be reduced. On the other hand, we may also have a clearer understanding of the causes of Alzheimer's disease using the model.

FW had full access to all the design for this study, take responsibility for the integrity of the data and accuracy of the data analysis, and drafted the final manuscript, WT have carried out the molecular genetic studies, performed the statistical analysis, and drafted the manuscript, YC have conceived of the study, participated in its design and performed the statistical analysis. All authors read and approved the final manuscript.