- Short Report
- Open Access
Variants of ESR1, APOE, LPL and IL-6 loci in young healthy subjects: association with lipid status and obesity
© Sertic et al; licensee BioMed Central Ltd. 2009
- Received: 12 June 2009
- Accepted: 5 October 2009
- Published: 5 October 2009
BMI was increased (>25) in 22% of young healthy subjects. Increased cholesterol values (>5.0 mmol/L) were found in 23% of subjects, LDL-C (>3.0 mmol/L) in 23%, triglycerides (>1.7 mmol/L) in 11% of subjects. We found statistically significant differences in subjects' weight (p = 0.015), BMI (p = 0.023), and waist-hip ratio (WHR) (p = 0.015) in regard to their diet type; subjects with Mediterranean diet had the lowest values compared to those on continental and mixed diet. Significant associations were found for: LPL genetic polymorphic variant and abdominal obesity (p = 0.013), APO epsilon4 allele and hypercholesterolemia (p = 0.003), and ESR1-TA long allele and hypercholesterolemia (p = 0.011).
Human obesity is a multifactorial syndrome influenced also by genetic factors. Among gene variants found to be involved in body weight regulation and development of obesity, particular attention has been paid to polymorphisms in genes associated with obesity-related metabolic disorders. We explored the association of genetic polymorphisms of: estrogen receptor alpha (ESR1-TA repeats); interleukin-6 (IL-6 G-174C); apolipoprotein E (APO epsilon2, epsilon3, epsilon4); lipoprotein lipase Pvu II (LPL P+/-), with clinical variables: gender, age, body mass index (BMI), diet type and biological variables: triglycerides, cholesterol, HDL-C, LDL-C, CRP, homocysteine, urate, and glucose in 105 healthy young subjects (20-35 yrs) of Croatian origin.
Genotyping of IL-6, LPL was performed by PCR-RFLP, of APOE by real-time PCR, and of ESR1 by PCR and capillary electrophoresis. Association analyses were performed of alleles and genotypes with biological variables.
ESR-1, LPL, and APO E genetic polymorphic variants could represent predictive genetic risk markers for obesity-related metabolic disorders in young healthy subjects. Mediterranean type of diet is also an important protective factor against abdominal obesity.
- Body Mass Index
- Abdominal Obesity
- Mediterranean Diet
- Young Healthy Subject
- Diet Type
Obesity has become an extensive public health concern because its prevalence has increased to epidemic proportions.
The most frequently investigated gene polymorphisms are APOE, adiponectin (ADIPOQ), angiotensin-converting enzyme (ACE), methyltetrahydrofolate reductase (MTHFR), IL-6, and estrogen receptors α and β (ESR1, ESR2) [4–7].
Our study was focused on ESR1 (L-long TA repeats; S-short TA repeats), APOE, IL-6 and LPL in young healthy population. Estrogen appears to be involved in protection against obesity and regulation of food intake. ESR1 is expressed mainly by adipocytes and it lowers the lipolytic response in subcutaneous fat by increasing the number of antilipolytic adrenoreceptors. BMI, waist circumference and HDL cholesterol levels have been associated with ESR1 polymorphisms.
A long genotype group of a common thymine-adenine (TA) dinucleotide repeat polymorphism in the regulatory region of ESR1 has been associated with coronary artery disease. There is also possibility that carriers of the long repeat are marked by decreased expression of the ESR1 gene and that they have less cardiovascular protective effects from estrogen than carriers of the short alleles [8–10].
APOE alleles and genetic variants of IL-6 have also been established to be factors of susceptibility to obesity . Data on C-174G polymorphism of IL-6 suggest that it affects the IL-6 transcription rate and plays a role in the development of obesity and insulin sensitivity .
LPL PvuII polymorphism is associated with high cholesterol levels and atherosclerotic cardiovascular diseases [12, 13]. In our previous investigations ACE  and APOE [15–17] were confirmed to be metabolic risk factors.
In this study we tried to identify some genetic variables associated with lipid status and obesity in young healthy subjects.
A total of 105 young healthy subjects (58 female, 47 male), aged 20-35 years, participated in the study. They completed a questionnaire on family history of cardiovascular diseases, hypertension, obesity, taking medications, smoking and diet. All participants signed informed consent forms, and the study protocol was approved by Ethics Committee. BMI was calculated as weight (kg)/height (square meter), and WHR was calculated as the ratio of abdomen or waist to hip circumference. Blood samples for biochemical analyses (total cholesterol, LDL- and HDL-cholesterol, glucose, CRP, urate) were collected after overnight fasting and analyzed.
DNA extraction and genotyping
Dinucleotide polymorphism of ESR1 was analyzed using PCR amplification with labeled primer 5'-6-FAM-GAC GCA TGA TAT ACT TCA CC-3' and 5' - GCA GAA TCA AAT ATC CAG ATG-3'. PCR was performed with 25 cycles that consisted of 2 min at 94°C, 1 min at 58°C, and 1 min at 72°C, followed by 30 min at 60°C after the last cycle. The alleles were size-separated by capillary gel electrophoresis using Gene Scan Fragment Analysis Software 4.0. In short, 1 μl of the product was diluted with 12 μl of deionized formamide containing 0.5 μl GeneScan-500 ROX internal lane standard for sizing DNA fragments. Capillary electrophoresis was then carried out using ABI PRISM 310 Genetic Analyzer and POP-6 that had been formulated for applications requiring high resolution under denaturing electrophoretic conditions .
The primers used in the PCR were: 5' TGACTTCAGCTTTACTCTTTGT 3' and 5' CTGATTGGAAACCTTAT TAAG 3'. The reaction was carried out in a final volume of 50 μl containing 1.5 mmol/l of MgCl2, 0.2 mmol/l of each dNTP, 0.2 mmol/l of each primer, and 2.5 unit of Taq polymerase. DNA was amplified during 35 cycles with an initial denaturation of 10 minutes at 94°C and a final extension of 10 min at 72°C. The cycle program consisted of 1 minute denaturation at 94°C, 1 minute and 35 s annealing at 55°C and 1 min extension at 72°C. PCR products were digested with Nla III restriction enzyme at 37°C overnight and electrophoresed on a 2% agarose gel with ethidium bromide staining.
The identified genotypes were named according to the presence or absence of the enzyme restriction sites, so Nla III (GG), (GC), and (CC) are homozygotes for the presence of the site (140/58 bp), heterozygotes for the presence and absence of the site (198/140/58 bp), and homozygotes for the absence of the site (198 bp), respectively .
LPL P+/- and APOE epsilon2, epsilon3, epsilon4
LPL P+/- was determined using PCR-RFLP and agarose gel electrophoresis technique, as previously described in 
APOE genotyping was performed by applying DNA assay that uses rapid-cycle PCR and fluorescence resonance energy transfer (FRET) with the LightCycler. In the assay, a 265-bp fragment of exon 4 of the APOE gene was amplified from genomic DNA. The detection probes that cover codons 112 and 158 were 5' labeled with LC-Red 640 and LC-Red 705, respectively .
Study the associations of sociodemographics and all clinical features (Table 1) with metabolic parameters. The differences between mean values of metabolic parameters (BMI, abdominal obesity, cholesterol, LDL, HDL, triglycerides, glucose, blood pressure, CRP and urate) among different groups according to clinical features and sociodemographics were determined using t-tests for independent samples or analysis of variance, as appropriate. In case of significant differences between groups, logistic regression was performed to test for the predictors of BMI, abdominal obesity, high cholesterol, high LDL, low HDL, high triglycerides, glucose intolerance, hypertension, abnormal CRP and urate. For that purpose variables describing metabolic status (BMI, abdominal obesity, cholesterol, LDL, HDL, triglycerides, glucose, CRP and urate) were grouped in categories according to their normal values - described in Table 2.
Study the association of genetic variants with lipid status and BMI. For that purpose variables describing metabolic status (BMI, abdominal obesity, cholesterol, LDL, HDL, triglycerides, glucose, blood pressure, CRP and urate) were again grouped in categories as described in Table 2. Associations of alleles, genotypes and haplotypes with biological variables were established using UNPHASED-3.0.10. Sex and age were used as covariates [19, 20].
Sociodemographic and clinical data of subjects (N = 105)
Mean +/- SD
0.81 +/- 0.8
Clinical and metabolic parameters of subjects (N = 105)
With known effects on investigated features (Contraceptives)
Elevated hepatic enzymes
Positive family history
More than one
Hypertension (>135/85 mmHg)
Abdominal obesity (WHR >90 (M), >85 (F)
Hypertriglyceridemia (>1.7 mmol/L)
Hypercholesterolemia (>5 mmol/L)
HDL-C (<1.0 mmol/L)
LDL-C (>3 mmol/L)
Glucose levels (>6.4 mmol/L)
CRP levels (mg/L)
Elevated urate levels (>383 mmol/L)
The sequential Bonferroni adjustments (Holm, 1979; Rice, 1989) were applied to correct for the effect of multiple tests using SAS Release 8.02 (SAS Institute, 1999).
Basic characteristics of 105 study participants are shown in Table 1 and Table 2. Age of the subjects was 25.82 +/- 3.28 years. BMI values ranging between 25 - 30 were found in 21.9% of subjects. Positive family history for obesity was established in 32.4% of subjects.
Hypertriglyceridemia was detected in 11.4% of subjects. 12.4% of subjects were on Mediterranean diet. In addition, increased CRP levels were found in 9.5% of our subjects, and elevated urate levels in 11.4%. Out of 105 subjects, 9 were taking medication for cardiovascular diseases, 4 were taking other medications, for 5 subjects data were missing and the rest (N = 87) were not using medications at the time of the study.
Differences in weight, BMI, waist and WHR of subjects with regard to their diet type
LSD post hoc,
p values (only significant)
F = 4.365, p = 0.015
Mixed vs. continental, MD = 12.18, p = 0.025
Mixed vs. Mediterranean, MD = 19.58, p = 0.004
F = 3.915, p = 0.023
Mixed vs. continental, MD = 2.65, p = 0.039
Mixed vs. Mediterranean, MD = 4.42, p = 0.006
F = 4.803, p = 0.01
Mixed vs. continental, MD = 8.45, p = 0.036
Mixed vs. Mediterranean, MD = 15.33, p = 0.003
F = 4.421, p = 0.015
Mediterranean vs. continental, MD = -0.05, p = 0.042
Mediterranean vs. mixed, MD = -0.095, p = 0.004
Association of ESR1 and APOE genetic polymorphisms and hypercholesterolemia
Cholesterol < 5 mmol/L
Cholesterol > 5 mmol/L
Test p value
X2 = 6.391
df = 1
p = 0.011
X2 = 7.147
df = 2
p = 0.028
X2 = 11.46
df = 2
p = 0.003
X2 = 14.98
df = 4
p = 0.004
LPL genotyping results that indicate a risk associated with abdominal obesity
<0.9 (M); <0.85 (F)
>0.9 (M); >0.85 (F)
X2 = 1.158
df = 1
p = 0.272
X2 = 7.794
df = 2
p = 0.013
Association of IL-6 genetic polymorphisms and blood pressure
BP < 135/85
BP > 135/85
X 2 = 0.952
df = 1
p = 0.329
X 2 = 5.326
df = 2
p = 0.069
Genetic biomarkers have an increasing impact in laboratory medicine. They are currently used to predict disease onset or disease recurrence, and to individualize treatment and assess response to treatment.
We found an association between BMI and diet type. Differences in weight, BMI, waist measures, and WHR of subjects with regard to diet type (continental, Mediterranean, mixed type) were indicative. Subjects on Mediterranean diet had the best results with regard to BMI and WHR, which confirmed the assumption that the type of diet could have a substantial effect on the health of young subjects. Further, association was also established between ESR1 and APOE genetic polymorphisms and hypercholesterolemia. Actually, our finding was confirmed by results of other studies  who found positive correlation between APOE polymorphism and hypercholesterolemia. Association between estrogen receptor-alpha gene polymorphisms and coronary artery disease with familial hypercholesterolemia was previously reported in other studies .
Complex approach includes multiple genetic markers, gene expression, and functional protein. Such complex approach is needed for understanding physiopathology and etiology of metabolic disorders and cardiovascular disease, and leads to development of biomarkers for prevention of disease. Possible metabolic biomarkers that we examined in our study are CRP and urate: we found elevated levels (9.5%; 11.4%, respectively) of these metabolic parameters, and similar results in young adults were reported previously [22, 23]. Recent data also suggest that elevated serum CRP and urate levels may contribute to the development of coronary artery disease . It has been established that men with long alleles at the ESR1 promoter TA repeat have significantly more narrowed coronaries, larger areas of complicated lesions, and more calcification of coronary arteries than men with short alleles. A recent Finnish study proposed that long-repeat carriers have lower expression of the ESR1 gene and less cardiovascular protective effects from estrogen than carriers of short alleles .
In our study, we found an association between ESR1 long TA allele and hypercholesterolemia in young healthy subjects and it could represent a possible risk factor for developing metabolic disorder and cardiovascular disease.
For genetic prediction of disease risk, it is important to identify functional gene polymorphism in order to establish genotype-phenotype correlation. Polymorphisms can be associated with either health (low risk predictors) or possible development of disease (high risk predictors; hypercholesterolemia, hypertension, obesity, metabolic syndrome, cardiovascular disease). Adipogenesis-related genes such as PPAR gamma and genes related to cytokines such as IL-6 and lipid metabolism including LPL have also been associated to the weight lowering outcome induced by hypocaloric diets . Efforts to reduce cholesterol level in order to prevent coronary heart disease have led to declined mortality. However, new information in this regard have recently been recorded in western countries: a decreasing trend in HDL cholesterol and increasing trends in obesity and triglycerides .
In literature, correlation between LPL polymorphisms and obesity has remained controversial; however, we found evidence of this association in our study. Actually, presence of LPL P+/-, P+/+ genotypes that we investigated in our study might, as a predictive risk factor related to obesity-related metabolic disorder, indicate the need for individualized diet adjustments.
A study reported previously  evaluated the role of IL-6-174G>C polymorphism in the risk of developing metabolic alterations in people with excessive body weight. Their data demonstrated that the occurrence of C allele of IL-6-174G>C gene polymorphism in people with excessive body weight was accompanied by increased risk of developing obesity-related metabolic disorders, especially insulin resistance.
Regardless of age and gender, subjects with IL-6 C and PPAR gamma variants had lower BMI and plasma triglyceride levels than those carrying only one of these variants. Thus, it might be possible that IL-6 gene polymorphism alone has none or only slight effects; its effects may be apparent only in the presence of a factor like PPAR gamma . The influence of interleukin-6 G-174C gene polymorphism on coronary artery disease, cardiovascular complications and mortality in dialysis patients was described in a previously reported study. Their results suggest that IL-6 gene G-174C polymorphism is associated with the incidence of cardiovascular events and mortality in chronic dialysis patients . Although this indicates an important correlation between high blood pressure and IL-6 genotype, larger studies are needed to confirm this significance.
Results of examining correlation between weight, BMI and WHR and continental, Mediterranean and mixed diet showed optimal values in subjects on Mediterranean diet.
Statistically significant correlation was detected between ESR1 long allele and total cholesterol. We also found statistically significant associations of APOE 2/3 and APOE3/4 genotypes with lowered and increased cholesterol levels, respectively. Statistically significant association was observed between LPL P+/- polymorphism and abdominal obesity. These three polymorphisms indicate an association with traits that characterize lipid status and obesity.
This study has been carried out as part of Project No. 108-1080134-0136, Croatian Ministry of Science, Education and Sports.
- Visvikis-Siest S, Siest G: The Stanislas Cohort: a 10-year follow-up of supposed healthy families. Gene-environment interactions, reference values and evaluation of biomarkers in prevention of cardiovascular diseases. Clin Chem Lab Med. 2008, 46 (6): 733-47. 10.1515/CCLM.2008.178.View ArticlePubMedGoogle Scholar
- Maumus S, Marie B, Vincent-Viry M, Siest G, Visvikis-Siest S: Analysis of the effect of multiple genetic variants of cardiovascular disease risk on insulin concentration variability in healthy adults of the STANISLAS cohort; the role of FGB - 455 G/A polymorphism. Atherosclerosis. 2007, 191 (2): 369-76. 10.1016/j.atherosclerosis.2006.04.003.View ArticlePubMedGoogle Scholar
- Rampersaud E, Mitchell BD, Pollin TI, Fu M, Shen H, O'Connell JR, Ducharme JL, Hines S, Sack P, Naglieri R, Shuldiner AR, Snitker S: Physical activity and the association of common FTO gene variants with body mass index and obesity. Arch Intern Med. 2008, 168 (16): 1791-7. 10.1001/archinte.168.16.1791.PubMed CentralView ArticlePubMedGoogle Scholar
- Barbieri M, Rizzo MR, Papa M, Acampora R, De Angelis L, Olivieri F, Marchegiani , Franceschi C, Paolisso G: Role of interaction between variants in the PPARG and interleukin-6 genes on obesity related metabolic risk factors. Gerontology. 2005, 40: 599-604. 10.1016/j.exger.2005.05.004.View ArticleGoogle Scholar
- Humphries SE, Morgan L: Genetic risk factors for stroke and carotid atherosclerosis: insights into pathophysiology from candidate gene approaches. The Lancet Neurology. 2004, 3: 227-236. 10.1016/S1474-4422(04)00708-2.View ArticlePubMedGoogle Scholar
- Morgan L, Humphries SE: The genetics of stroke. Current Opinion in Lipidology. 2005, 16: 193-199. 10.1097/01.mol.0000162324.85290.ac.View ArticlePubMedGoogle Scholar
- Sanghera DK, Ortega L, Han S, Singh J, Ralhan SK, Wander GS, Mehra NK, Mulvihill JJ, Ferrell RE, Nath SK, Kamboh MI: Impact of nine common type 2 diabetes risk polymorphisms in Asian Indian Sikhs: PPARG2 (Pro12Ala), IGF2BP2, TCF7L2 and FTO variants confer a significant risk. BMC Medical Genetics. 2008, 9: 59-10.1186/1471-2350-9-59.PubMed CentralView ArticlePubMedGoogle Scholar
- Kunnas TA, Lehtimäki T, Karhunen PJ, Laaksonen R, Janatuinen T, Vesalainen R, Nuutila P, Knuuti J, Nikkari ST: Estrogen receptor genotype modulates myocardial perfusion in young men. J Mol Med. 2004, 82: 821-5. 10.1007/s00109-004-0594-4.View ArticlePubMedGoogle Scholar
- Nilsson M, Dahlman I, Jiao H, Gustafsson J-A, Arner P, Dahlman-Wright K: Impact of estrogen receptor gene polymorphisms and mRNA levels on obesity and lipolysis - a cohort study. BMC Medical Genetics. 2007, 8: 73-10.1186/1471-2350-8-73. doi: 10.1186/1471-2350-8-73PubMed CentralView ArticlePubMedGoogle Scholar
- Kunnas TA, Holmberg-Marttila D, Karhunen PJ: Analysis of estrogen receptor dinucleotide polymorphism by capillary gel electrophoresis with a population genetic study in 180 Finns. Hum Hered. 1999, 49: 142-145. 10.1159/000022862.View ArticlePubMedGoogle Scholar
- Boiardi L, Casali B, Farnetti E, Pipitone N, Nicoli D, Cantini F, Macchioni P, Bajocchi G, Catanoso MG, Pulsatelli L, Consonni D, Salvarani C: Relationship between interleukin 6 promoter polymorphism at position -174, IL-6 serum levels, and the risk of relapse/recurrence in polymyalgia rheumatica. The Journal of Rheumatology. 2006, 33 (4): 703-708.PubMedGoogle Scholar
- Kurt O, Yilmaz H, Isbir T, Berkkan H: Effects of lipoprotein lipase Pvu II gene polymorphism on serum lipoprotein levels in Turkish patients with type 2 diabetes mellitus. Adv Mol Med. 2006, 2 (1): 41-46.Google Scholar
- Pašalić D, Sertić J, Kunović B, Miličević Z, Pašić A, Zrinski-Topić R, Ferenčak G, Stavljenić Rukavina A: Lipoprotein lipase gene polymorphism and lipid profile in patients with hypertriglyceridemia. Croat Med J. 2001, 42 (5): 517-22.PubMedGoogle Scholar
- Sertić J, Hebrang D, Januš D, Salzer B, Nikšić M, Čvorišćec D, Stavljenić Rukavina A: Association between deletion polymorphism of the angiotensin-converting enzyme gene and cerebral atherosclerosis. Eur J Clin Chem Clin Biochem. 1996, 34: 301-304.PubMedGoogle Scholar
- Stavljenić Rukavina A, Sertić J, Salzer B, Dumić M, Radica A, Fumić K, Krajina A: Apolipoprotein E phenotypes and genotypes as determined by polymerase chain reaction using allele specific oligonucleotide probes and amplification refractory mutation system in children with insulin-dependent diabetes mellitus. Clin Chem Aca. 1993, 216: 191-198. 10.1016/0009-8981(93)90152-T.View ArticleGoogle Scholar
- Čubrilo-Turek M, Stavljenić Rukavina A, Sertić J, Zrinski R, Turek S, Grgac G, Ljubojević N: Apolipoprotein E genotypes and metabolic risk factors for coronary heart disease in middle-aged women. Coll Antropol. 1998, 22: 149-155.PubMedGoogle Scholar
- Babić T, Mahović Lakušić D, Sertić J, Petrovečki M, Stavljenić Rukavina A: Apo E genotyping and response to galanthamine in Alzheimer's disease - a real life retrospective study. Coll Antropol. 2004, 28 (1): 199-204.PubMedGoogle Scholar
- Wittwer CT, Ririe KM, Andrew RV, David DA, Gundry RA, Balis UJ: The LightCycler: a microvolume multisample fluorimeter with rapid temperature control. Biotechniques. 1997, 22: 176-181.PubMedGoogle Scholar
- Holm S: A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics. 1979, 6: 65-70.Google Scholar
- Rice WR: Analyzing tables of statistical tests. Evolution. 1989, 43: 223-225. 10.2307/2409177.View ArticleGoogle Scholar
- Lu H, Higashikata T, Inazu A, Nohara A, Yu W, Shimizu M, Mabuchi H: Association of estrogen receptor-alpha gene polymorphisms with coronary artery disease in patients with familial hypercholesterolemia. Arterioscler Thromb Vasc Biol. 2002, 22 (5): 817-23. 10.1161/01.ATV.0000014424.18209.21.View ArticlePubMedGoogle Scholar
- McDade TW, Rutherford JN, Adair L, Kuzawa C: Population differences in associations between C-reactive protein concentration and adiposity: comparison of young adults in the Philippines and the United States. Am J Clin Nutr. 2009, 89 (4): 1237-45. 10.3945/ajcn.2008.27080.PubMed CentralView ArticlePubMedGoogle Scholar
- Lu JJ, Jiang DD, Chou SM, Hor CB, Lay JD, Wang HL: Prevalence of obesity and its association wtih cardiovascular disease risk factors in adolescent girls from a college in central Taiwan. Kaohsiung J Med Sci. 2008, 24 (3): 144-51. 10.1016/S1607-551X(08)70142-6.View ArticlePubMedGoogle Scholar
- Eroglu S, Elif Sade L, Yildirir A, Demir O, Bozbas H, Muderrisoglu H: Serum levels of C-reactive protein and uric acid in patients with cardiac syndrome X. Acta Cardiol. 2009, 64 (2): 207-11. 10.2143/AC.64.2.2036139.View ArticlePubMedGoogle Scholar
- Martinez JA, Parra MD, Santos JL, Moreno-Aliaga MJ, Marti A, Martinez-Gonzalez MA: Genotype-dependent response to energy-restricted diets in obese subjects: towards personalized nutrition. Asia Pac J Clin Nutr. 2008, 17 (1): 119-122.PubMedGoogle Scholar
- Gupta R, Guptha S, Agrawal A, Kaul V, Gaur K, Gupta VP: Secular trends in cholesterol lipoproteins and triglycerides and prevalence of dyslipidemias in an urban Indian population. Lipids in Health and Disease. 2008, 24: 7-40.Google Scholar
- Goyenechea E, Parra D, Martinez JA: Impact of interleukin 6-174G>C polymorphism on obesity-related metabolic disorders in people with excess in body weight. Metabolism. 2007, 56 (12): 1643-8. 10.1016/j.metabol.2007.07.005.View ArticlePubMedGoogle Scholar
- Aker S, Bantis C, Reis P, Kuhr N, Schwandt C, Grabensee B, Heering P, Ivens K: Influence of interleukin-6 G-174C gene polymorphism on coronary artery disease, cardiovascular complications and mortality in dialysis patients. Nephrol Dial Transplant. 2009, 24 (9): 2847-2851. 10.1093/ndt/gfp141.View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.