There is substantial debate about the role of vitamin C, a powerful antioxidant, in preventing chronic disease [1]. Whilst there is evidence from observational studies that lower vitamin C plasma levels are associated with higher risks of cancer, cardiovascular disease and death, randomised controlled trials (RCTs) of vitamin supplementation have not provided evidence for the protective effects of vitamin C against these diseases [2].

The associations seen in observational studies are likely to be confounded since plasma vitamin C levels have been shown to be associated with a number of socioeconomic and behavioural factors throughout the lifecourse [2] which are also strongly related to disease outcomes eg social class, smoking, diet. While information on these confounding factors is often collected, it is difficult to ever fully capture lifestyle differences between subjects.

B-vitamin status has also been linked to cardiovascular disease and cancer [5]. B-vitamins (folate, B12 and B6) play a key role in homocysteine metabolism and studies have shown that higher levels of plasma homocysteine are associated with increased risk of cardiovascular disease [6]. Evidence from a Mendelian randomisation study supports a causal relationship between homocysteine plasma levels and stroke [5]. However, there is conflicting evidence from RCTs about the effects of B-vitamin supplementation on risk of cardiovascular disease [6;7].

Mendelian randomisation methods can be used to investigate causal effects free from confounding or reverse causality [9] by using genetic variation associated with risk factors of interest as proxies for exposures. Since genes are randomly assigned to individuals due to random assortment of alleles at meiosis, potential confounders should be distributed evenly amongst genotypes.

Recently, common variants of genes associated with Vitamin C [9] and Vitamin B12 [10] plasma levels have been identified. SLC23A1 and SLC23A2 encode sodium-dependent vitamin C transporters, which are necessary for the major pathway of vitamin C absorption in humans[10]. Pilot data from four studies has shown that single nucleotide polymorphisms (SNPs) in the SLC23A1 gene are associated with plasma vitamin C levels (see plot below).

Appendix Figure 1.

Meta-analysis of SLC23 A1 variation effect on circulating vitamin C taken from the 10 Towns Study, the British Women's Heart and Health Study, MIDSPAN and the EPIC cohort. Meta-analysis explicitly marks the effect of rs33972313 on square root transformed vitamin C, r^2=~0.002 (BWHHS & 10 Towns) p_meta=6.5x10-12, I^2=0%.

Strong associations have been found between variants of the FUT2 gene and plasma vitamin B12 levels [10]. The mechanism by which this gene may influence B12 levels is not yet fully understood but the authors hypothesise that differences in susceptibility to infection by H.pylori conferred by variants of this gene may lead to differences in B12 absorption, since H.pylori infection leads to reduced secretion of the glycoprotein intrinsic factor, which must bind to B12 before it can be absorbed.

This project aims to use these newly discovered genetic variants to investigate whether the observed relationships between plasma concentrations of vitamins C and B12 and metabolic and cardiovascular outcomes are causal. Our aim here is to relate this to early and later lifecourse indicators of chronic disease. These findings should provide important information regarding the appropriateness of vitamin supplementation as a preventive measure for chronic diseases.

During the initial phase of the proposed study, genotyping of the relevant SNPs in all ALSPAC children and mothers for whom genetic data is available (N~7-9000 for each group) will be carried out by KBioscience.

We have selected the following genetic variants for assessment:

Vitamin C

Gene: SLC23A1 SNP number: rs33972313

Vitamin B12

Gene: FUT2 SNP number: rs492602 [11]

Statistical analyses will be conducted in STATA 10. We will test to see if genotype frequencies are in Hardy-Weinberg equilibrium. The relationship between genotypes and potential confounding factors (sex, age, social class, education, exercise, diet, smoking and alcohol consumption (in the mothers)) will be investigated using linear regression methods to assess if genotype is independent of confounders. Using linear regression we will then investigate the relationship of genotypes with outcome measures: BMI, blood lipids, glucose, insulin, blood pressure.

We also propose to undertake a scheme of genotype based selection of samples for serum vitamin C measurement within the ALSPAC cohort. This will be undertaken in efforts to provide an optimal subset of the ALSPAC cohort to provide replication data for the primary association of concern (i.e. SLC23A1 variation and circulating vitamin C). The measurement of circulating vitamin C will be undertaken by Naveed Sattar (University of Glasgow) and will be in a subset chosen to provide a balanced sample of the variant rs33972313. This variant has been chosen owing to its consistent relationship with circulating vitamin C (Figure 1), and as a consequence of the low MAF of this variant (~0.03). To capture the vitamin C effect expected from previous work, power calculations based on figures from the British Women's Heart and Health Study predict that if a MAF of 0.3 can be simulated by genotype based sample selection, then as few as ~700 individuals would be required to detect the effect of rs33972313 (80% power,). These parameters are summarised in Apendix Figure 2.

This stage of data collection and analyses will allow us to confirm associaitons between circulating vitamin C and variation at the SLC23A1 variant within a select sample suitably powered to detect the expected magnitude in vitamin change.

Findings will be written up for publication in a peer reviewed journal.

References

1. Padayatty SJ, Katz A, Wang Y, Eck P, Kwon O, Lee JH, Chen S, Corpe C, Dutta A, Dutta SK, Levine M. Vitamin C as an antioxidant: evaluation of its role in disease prevention. J Am Coll Nutr 2003; 22:18-35.

2. Lawlor DA, Davey SG, Kundu D, Bruckdorfer KR, Ebrahim S. Those confounded vitamins: what can we learn from the differences between observational versus randomised trial evidence? Lancet 2004; 363:1724-7.

3. Haggarty P. B-vitamins, genotype and disease causality. Proc Nutr Soc 2007; 66:539-47.

4. Homocysteine SC. Homocysteine and Risk of Ischemic Heart Disease and Stroke: A Meta-analysis. JAMA 2002; 288:2015-22.

5. Casas JP, Bautista LE, Smeeth L, Sharma P, Hingorani AD. Homocysteine and stroke: evidence on a causal link from mendelian randomisation. The Lancet 2005; 365:224-32.

6. Ishihara J, Iso H, Inoue M, Iwasaki M, Okada K, Kita Y, Kokubo Y, Okayama A, Tsugane S, for the JPHC Study Group. Intake of Folate, Vitamin B6 and Vitamin B12 and the Risk of CHD: The Japan Public Health Center-Based Prospective Study Cohort I. J Am Coll Nutr 2008; 27:127-36.

7. Albert CM, Cook NR, Gaziano JM, Zaharris E, MacFadyen J, Danielson E, Buring JE, Manson JE. Effect of Folic Acid and B Vitamins on Risk of Cardiovascular Events and Total Mortality Among Women at High Risk for Cardiovascular Disease: A Randomized Trial. JAMA 2008; 299:2027-36.

8. Davey Smith G, Ebrahim S. What can mendelian randomisation tell us about modifiable behavioural and environmental exposures? BMJ 2005; 330:1076-9.

9. Eck P, Erichsen HC, Taylor JG, Yeager M, Hughes AL, Levine M, Chanock S. Comparison of the genomic structure and variation in the two human sodium-dependent vitamin C transporters, SLC23A1 and SLC23A2. Hum Genet 2004; 115:285-94.

10. Hazra A, Kraft P, Selhub J, Giovannucci EL, Thomas G, Hoover RN, Chanock SJ, Hunter DJ. Common variants of FUT2 are associated with plasma vitamin B12 levels. Nat Genet 2008; 40:1160-2.

11. Erichsen HC, Engel SA, Eck PK, Welch R, Yeager M, Levine M, Siega-Riz AM, Olshan AF, Chanock SJ. Genetic variation in the sodium-dependent vitamin C transporters, SLC23A1, and SLC23A2 and risk for preterm delivery. Am J Epidemiol 2006; 163:245-54.

Appendix Figure1.

Meta-analysis of SLC23A1 gene variant effect on circulating levels of vitamin C. Gene variant rs33972313 is assessed in these analyses and data is shown as square root vitamin C levels.

Appendix Figure 2.

Summary of power calculation statistics. Three-way plot of number of samples required to achieve 80% power to detect effect sizes of varying strength at differing minor allele frequencies. As stated above, a minor allele frequency of 0.3 in a sample of ~700 ALSPAC children would yield 80% power to detect a main effect equivalent to that seen between rs33972313 and circulating vitamin C in other cohorts.