Se is an essential trace element in human nutrition associated with the activity of the antioxidant enzyme glutathione peroxidase. It is considered to be a protective agent against free radicals through enhanced enzyme activity. Associations between low Se and increased risk in various disease indices (cancer, CVD, reproduction and neonatal health, asthma) have been reported(Reference Flatt, Pearce, Thomson, Sears, Robinson and Beasley1, Reference Rayman2).
There is increased interest in using Se as a biomarker for various cancer outcomes(Reference Niskar, Paschal, Kieszak, Flegal, Bowman, Gunter, Pirkle, Rubin, Sampson and McGeehin3, Reference Vogt, Ziegler, Patterson and Graubard4). Measuring Se intake in large cohort studies is often difficult because populations are often mobile and consume foods that were produced and prepared in different areas of the world. Se content in foods, especially grain, is highly variable depending on the Se content of the soils in which they are grown(5). Moreover, supplements containing Se are often ingested particularly by health-conscious individuals, further confounding the results. The rural elderly Chinese population represents a unique opportunity for studying Se exposure. The rural Chinese are unusually stable with most living in the same village throughout their entire life and consuming food that is locally grown. In addition, it is rare for these villagers to take dietary supplements. Chinese scientists have assembled extensive data on Se distributions in many parts of the country; hence it is possible to select sites with differing Se levels so that an extended range can be achieved to maximize statistical power for detecting potential associations. We have recruited a cohort of 2000 elderly Chinese from four rural counties to study Se level and its association with various health outcomes(Reference Gao, Jin and Hall6).
The apoE gene is involved in lipoprotein metabolism and is a risk factor for Alzheimer’s disease and cardiovascular disorders. There are also previous reports that an Se-deficient diet or targeted removal of a gene that causes complete loss of selenoprotein expression results in increased plasma cholesterol concentration with an increase in apoE protein(Reference Stone, Stewart, Nicholas and Pavuluri7–Reference Sengupta, Carlson, Hoffmann, Gladyshev and Hatfield9). Previous studies on predictors of toenail Se or serum Se have mostly concentrated on demographic factors such as age and gender. In the present paper, we examine the association between apoE genotype and Se levels measured in nail samples.
Methods
Study population
Two thousand Chinese aged 65 years and over from four counties in China were enrolled in the study. Two sites were from Sichuan Province in south-west China and the other two sites were from Shandong Province in eastern China. Since Chinese scientists have mapped Se distribution in many parts of the country, these two provinces were selected because of the varying Se levels within each province. Prior to final site selection, Chinese investigators travelled to several candidate sites collecting demographic information to ensure the local elderly population was large enough to provide a sample of 500 elderly subjects. Samples of grain (corn, rice and wheat), soil, water and nail clippings from randomly selected individuals at each candidate site were also collected and analysed for Se and other trace elements. The goal was to have two sites within one province that differed in Se levels but were similar in trace element measures and other potential confounders. Sites with known endemic diseases were excluded from consideration.
Twenty-two villages were surveyed in the present study. For each village included in the study, the Chinese investigators and a team of interviewers who were employees of provincial and county Centers for Disease Control travelled to the area, established a temporary headquarters and conducted a complete census of residents over the age 65 years in the area. They enrolled eligible residents by going door-to-door, obtaining informed consent before conducting the interview and collecting biological samples. Details of the study are published elsewhere(Reference Gao, Jin and Hall6). The study was approved by the Indiana University Institutional Review Board and the Institute for Environmental Health and Related Safety, Chinese Center for Disease Control and Prevention.
Information collected during the interview include age, gender, whether the participant attended school, alcohol consumption and smoking history, history of cancer, Parkinson’s disease, diabetes, hypertension, stroke, heart attack, head injury and bone fracture. Participants’ height, weight and blood pressure (two times) were also measured during the interview. BMI was derived from height and weight measurements. The average of the two blood pressure measures was used in our analyses.
Se measures
Nail samples from all study subjects were collected at the time of interview and stored in clean plastic bags labelled with subject identification numbers. The method of fluorometric determination of trace amounts of Se with 2,3-diaminonaphthalene, described in detail elsewhere(Reference Li, Cao and Sun10), was used to determine trace amounts of Se in the nail samples. Details on quality control measures implemented for the study have also been described previously(Reference Gao, Jin and Hall6).
apoE genotype
Blood spots on filter paper were collected from all study participants at the end of the interview. apoE genotype was determined by eluting DNA from a dried blood spot(Reference Yang, Hendrie, Hall, Oluwole, Hodes and Sahota11) followed by Hhal digestion of amplified products(Reference Hixson and Vernier12).
FFQ
An FFQ was administered asking participants their average daily intake of various grains, vegetables, meat, seafood, fruit, nuts, cooking oil, tea and water. The questionnaire had been developed and validated for use in Chinese populations(Reference Chen, Campbell, Li and Peto13–Reference Zhai, Popkin, Ma, Yu, Guo, Jin and Ge15). Food samples were collected from various locations in each village and analysed using the same method as for the nail samples, thus providing Se measures in each food item for each village. Estimated daily Se intake was derived from the FFQ using Se levels analysed from food samples in the village in which the participants resided(Reference Cao, Yan and Zhang16, Reference He17).
Statistical analysis
Univariate associations between demographic, lifestyle and medical history variables and Se measured in nail samples were examined using mixed-effect models with Se level as the dependent variable and each village as the random effect, which controls for the potential confounding effect from correlations in Se measures obtained from participants residing in the same village. Comparisons of demographic, lifestyle and medical characteristics between apoE ε4 allele carriers and non-carriers were conducted using t tests for continuous variables and Fisher’s exact tests for categorical variables. All variables that were either univariately associated with Se levels or associated with apoE status at the P < 0·10 level were included in a multivariate mixed-effect model with Se level as the dependent variable and village as the random effect to control for potential within-village correlation. A final multivariate model was identified including those predictors significant at the 0·05 level. A separate mixed-effect model using estimated Se dietary intake as dependent variable was also conducted to examine the association of apoE status with Se dietary intake. To facilitate interpretation of parameter estimates from all mixed-effect models, we standardized all continuous variables in all models. To ensure that the association between apoE status and Se level was not due to existing CVD and potential treatments, we repeated the mixed-effect model by excluding those subjects with histories of heart attack and stroke. We also repeated the same mixed-effect models in male and female participants separately.
Results
Univariate mixed-effect models with Se levels measured in nail samples as dependent variable revealed that female participants, those who had attended school, abstainers from alcohol, non-current smokers, those with a history of stroke and those with a history of fracture appeared to have higher Se levels at the P < 0·10 level.
We compared demographic, lifestyle and medical history factors by apoE ε4 allele status (Table 1). Diastolic blood pressure was significantly higher in the ε4 carriers than in the non-carriers (P = 0·0368). Non-carriers had a significantly higher rate of fracture than ε4 carriers (2·88 % v. 0·91 %, P = 0·0353). Two other factors meeting inclusion criteria for the multivariate model were alcohol consumption (P = 0·0687) and history of head injury (P = 0·0906).
Comparisons of demographic, lifestyle and medical history variables between carriers and non-carriers of the apoE ε4 allele: rural elderly (65 years and over) residents from Sichuan Province and Shandong Province, China (n 2000)
In Table 2, all variables significant at P = 0·10 from all univariate analyses are included in a multivariate mixed-effect model to identify significant predictors of Se levels measured in nail samples. In addition, we included BMI tertile as an additional predictor in the model because it had borderline significance for Se level just below our cut-off point of P = 0·10 (P = 0·1085) and it allowed us to control for the potential effect from body composition. A separate mixed-effect model with estimated dietary Se intake as the dependent variable is also included in Table 2. apoE ε4 carriers were shown to have significantly lower Se levels measured in nail samples, but they were not significantly different from non-carriers on estimated Se dietary intake (P = 0·6451). Final multivariate models, using nail Se levels and including only factors significant at the 0·05 level, are presented in Table 3. Controlling for gender, smoking status, history of stroke and estimated dietary Se intake, apoE ε4 carriers had significantly lower Se levels measured in nail samples than non-carriers (P = 0·0396). The results of the final model demonstrate that given the same Se dietary intake level, the apoE ε4 carriers had lower Se levels in their biological samples, equivalent to being two standard deviations lower on Se dietary intake. We also fitted the same model on a sub-sample of participants without a history of heart attack or stroke; the difference in Se level between the two genotype groups remained (P = 0·0512).
Multivariate mixed-effect model to identify variables associated with selenium levels measured in nail samples (μg/g) and with estimated selenium dietary intake: rural elderly (65 years and over) residents from Sichuan Province and Shandong Province, China (n 2000). Village was included as a random effect to control for potential within-village correlation
Final multivariate mixed-effect models including significant variables associated with selenium levels in nail samples (μg/g, standardized) in the entire sample and in a sub-sample of participants without history of heart attack or stroke: rural elderly (65 years and over) residents from Sichuan Province and Shandong Province, China (n 2000)
We also fitted the model in Table 3 to male and female participants separately; the results are presented in Table 4. Interestingly, dietary Se intake, while significantly related to nail Se levels in males, was not associated with nail Se levels in females. apoE ε4 carriers had lower Se levels in their nail samples for both male and female participants. The differences between apoE ε4 carriers and non-carriers were similar between males and females (0·0734 and 0·0755 standard deviations, respectively), although neither reached significance due to the reduced sample sizes from the model for the entire sample.
Multivariate mixed-effect models including significant variables associated with selenium levels in nail samples (μg/g, standardized) in male and female participants separately: rural elderly (65 years and over) residents from Sichuan Province and Shandong Province, China (n 2000)
Discussion
In this rural elderly Chinese cohort, we found that apoE ε4 carriers had significantly lower Se levels measured in nail samples than non-carriers after adjusting for other significant covariates and controlling for estimated dietary Se intake. The difference between the two genotype groups persisted at a similar magnitude when we repeated the analysis excluding participants with histories of heart attack and stroke.
Predictors of Se levels measured in the blood, toenails or diet have been described previously(Reference Niskar, Paschal, Kieszak, Flegal, Bowman, Gunter, Pirkle, Rubin, Sampson and McGeehin3, Reference Vogt, Ziegler, Patterson and Graubard4, Reference van den Brandt, Goldbohm, van’t Veer, Bode, Hermus and Sturmans18–Reference Virtanen, van’t Veer, Kok, Kardinaal and Aro20). However, most of these studies focused on demographic variables. Our results identifying smoking and male gender as factors associated with lower Se levels are consistent with previous studies(Reference Hunter, Morris, Chute, Kushner, Colditz, Stampfer, Speizer and Willett19, Reference Virtanen, van’t Veer, Kok, Kardinaal and Aro20). However, previous results on the association between increasing age and decreasing levels of Se were inconsistent(Reference Hunter, Morris, Chute, Kushner, Colditz, Stampfer, Speizer and Willett19, Reference Akbaraly, Hininger-Favier, Carriere, Arnaud, Gourlet, Roussel and Berr21), perhaps due to differences in age range in each study population. In the present cohort of participants aged 65 years and older, we did not find a significant association between age and nail Se levels. Nor did we find a significant association between alcohol consumption and Se level in this cohort, consistent with findings in previous studies(Reference van den Brandt, Goldbohm, van’t Veer, Bode, Hermus and Sturmans18–Reference Virtanen, van’t Veer, Kok, Kardinaal and Aro20).
To our knowledge, the present study is the first to examine the association between apoE genotype and Se levels in a large cohort of elderly participants. The apoE gene has been shown to play a central role in cholesterol uptake and lipid metabolism and the ε4 allele is linked to increased risk of Alzheimer’s disease and CVD. The mechanism underlying the association between the apoE ε4 allele and Se levels is unclear, however. It is has been reported that apoE ε2 and apoE ε3 are more effective in maintaining neuronal health than is apoE ε4 (Reference Mahley22–Reference Weisgraber and Mahley24). It is possible that apoE ε4 carriers have increased levels of oxidative stress, ischaemia and inflammation and that the increased oxidative stress in turn reduces the pool of Se.
The Framingham Offspring Study proposed that apoE ε4 may be associated with reduced antioxidant capacity based on the interaction between smoking and apoE on cardiovascular risk(Reference Talmud, Stephens, Hawe, Demissie, Cupples, Hurel, Humphries and Ordovas25). This mechanism is supported by in vitro results showing that protection of recombinant apoE against oxidation is greater in ε4 non-carriers than in ε4 carriers(Reference Smith, Miyata, Poulin, Neveux and Craig26–Reference Miyata and Smith28). Results from animal studies indicate that apoE deficiency is associated with increased oxidative damage(Reference Shea, Rogers, Ashline, Ortiz and Sheu29). However, Se levels were not directly measured in these studies.
Another potential mechanism for the observed association between apoE genotype and Se levels is through reduced lipid metabolism in the ε4 carriers. Animal experiments have demonstrated that Se deficiency can lead to increased levels of plasma cholesterol, apoE and oxidized LDL in rats and mice(Reference Rosenblat and Aviram30, Reference Mazur, Nassir, Gueux, Moundras, Bellanger, Grolier, Rock and Rayssiguier31). Oxidized LDL accumulates in atherosclerotic lesions in man(Reference Aviram, Maor, Keidar, Hayek, Oiknine, Bar-El, Adler, Kertzman and Milo32) and increased LDL susceptibility to oxidation was shown in patients with hypercholesterolaemia(Reference Lavy, Brook, Dankner, Ben Amotz and Aviram33), hypertension(Reference Keidar, Kaplan, Shapira, Brook and Aviram34) and diabetes(Reference Beaudeux, Guillausseau, Peynet, Flourie, Assayag, Tielmans, Warnet and Rousselet35).
There seems to be a correlation between Se levels and metabolism of lipids and lipoproteins. Lipid metabolism may be affected by the oxidative state of the lipid molecule, on which Se would have a direct effect. A Finnish population with a higher proportion of apoE ε4 homozygous carriers and high serum cholesterol was associated with lower CHD mortality than in other communities. The lower mortality was attributed to higher antioxidant status, in particular Se intake in this population(Reference Luoma, Nayha, Sikkila and Hassi36). An interactive relationship between Se, lipids and the apoE genotype may also be worth investigating.
Our result, showing that participants with a history of stroke had increased Se levels, was surprising given the previous report on the protective effects of Se on CHD. A few studies have reported lower Se levels in acute stroke patients compared with healthy controls(Reference Zimmermann, Winnefeld, Streck, Roskos and Haberl37), although others found no significant difference(Reference Chang, Lai, Cheng, Lau and Hu38). However, in one study, lower Se was found to be associated with stroke mortality, suggesting a potential explanation that the effects of Se on CHD in relation to mortality might leave fewer participants living with stroke(Reference Virtamo, Valkeila, Alfthan, Punsar, Huttunen and Karvonen39). This possibility will be examined further in our planned follow-up of this cohort.
The strength of the present study includes the extended range of Se distribution in the cohort, allowing the exploration of other factors, and the large cohort size for the investigation. An additional strength is that we also collected dietary data, so that the association between apoE ε4 and Se could be examined controlling for dietary Se intake.
A limitation of the study is that the association between the apoE ε4 allele and Se levels is found in an elderly Chinese population, which may limit the generalizability of the findings. Independent confirmation in other populations is needed. The study also lacked cholesterol and other lipid measures in the cohort, so that it is not possible to examine any potential interactive role played by the metabolism of lipids and lipoproteins.
Acknowledgements
The research was supported by NIH grants R01 AG019181, R01 AG09956 and P30 AG10133. None of the authors has any conflicts of interest. The authors contributed as follows. Conception and design: S.G., Y.J., K.S.H., C.L., F.W.U., J.R.M. and H.C.H. Acquisition of data: S.G., Y.J., K.S.H., C.L., F.W.U., J.R.M., F.M., J.M., Y.C., J.-C.B., P.L. and H.C.H. Analysis and interpretation of data: S.G., H.C.H. and J.-Z.S. Drafting of manuscript: S.G., Y.J., C.L. and H.C.H. Critical revision of manuscript for important intellectual content: S.G., Y.J., K.S.H., C.L., F.W.U., J.R.M., F.M., J.-Z.S., J.M., Y.C., J.-C.B., P.L. and H.C.H. Statistical expertise: S.G. and J.-Z.S. Obtaining funding: S.G., Y.J., K.S.H., C.L., F.W.U., J.R.M. and H.C.H. Administrative, technical or material support: Y.J., C.L., F.M., Y.C., P.L., J.-C.B. and J.M. Supervision: S.G., Y.J., C.L. and H.C.H. We thank Dr Linda Williams for helpful comments on results reported in this manuscript.
