JNN452890.indd © 2016 S. Karger AG, Basel 1661–6499/16/0096–0254$39.50/0 Original Paper J Nutrigenet Nutrigenomics 2016;9:254–264 The CAPN2/CAPN8 Locus on Chromosome 1q Is Associated with Variation in Serum Alpha-Carotene Concentrations Christopher R. D’Adamo a Valerie J. Dawson b Kathleen A. Ryan c Laura M. Yerges-Armstrong c Richard D. Semba d Nanette I. Steinle c, f Braxton D. Mitchell c, e Alan R. Shuldiner c Patrick F. McArdle c a Department of Family & Community Medicine, University of Maryland School of Medicine, b University of Maryland School of Medicine, c Department of Medicine, University of Maryland School of Medicine, d Wilmer Eye Institute, Johns Hopkins University School of Medicine, e Geriatrics Research and Education Clinical Center, Baltimore Veterans Administration Medical Center, and f Baltimore Veterans Affairs Medical Center, Baltimore, MD , USA Keywords Alpha-carotene · Carotenoids · Genome-wide association study · CAPN2 · CAPN8 · PRKCE · Old Order Amish Abstract Background/Aims: Alpha-carotene is a provitamin A carotenoid present in fruits and vege- tables. Higher serum concentrations of α-carotene have been associated with lower risk of cancer and all-cause mortality. Previous studies have suggested that genetic variants influ- ence serum concentrations of provitamin A carotenoids, but to date no variants have been robustly associated with serum α-carotene concentrations. The aim of this study was to iden- tify genetic associations with serum α-carotene concentrations using the genome-wide as- sociation study (GWAS) approach. Methods: A GWAS of serum α-carotene concentrations was conducted in 433 Old Order Amish adults who had consumed a 6-day controlled diet. Linear regression models adjusting for age, gender, and family structure were utilized to evaluate associations between genetic variants and serum α-carotene concentrations. Results: Ge- nome-wide significant associations with α-carotene concentrations were observed for loci on chromosome 1q41 between the genes CAPN2 and CAPN8 (rs12137025, p = 3.55 × 10 –8 ), chro- mosome 2p21 in PRKCE (rs2594495, p = 1.01 × 10 –8 ), and chromosome 4q34 (rs17830069, p = 2.89 × 10 –8 ). Conclusions: We identified 3 novel loci associated with serum α-carotene concentrations among a population that consumed a controlled diet. While replication is nec- essary, the CAPN2/CAPN8 locus provides compelling evidence for an association with serum α-carotene concentrations and may suggest a relationship with the development and pro- gression of cancers. © 2016 S. Karger AG, Basel Received: October 10, 2015 Accepted: October 27, 2016 Published online: December 22, 2016 Christopher R. D’Adamo, PhD University of Maryland School of Medicine 520 West Lombard St., East Hall Baltimore, MD 21201 (USA) E-Mail cdadamo @ som.umaryland.edu www.karger.com/jnn DOI: 10.1159/000452890 255J Nutrigenet Nutrigenomics 2016;9:254–264 DOI: 10.1159/000452890 D’Adamo et al.: The CAPN2/CAPN8 Locus on Chromosome 1q Is Associated with Variation in Serum Alpha-Carotene Concentrations www.karger.com/jnn © 2016 S. Karger AG, Basel Introduction The carotenoids are a class of pigments synthesized by plants and bacteria [1, 2] . Carot- enoids are not produced endogenously in humans and other animals and are obtained from the diet [1] . Only 40 of the 600 identified carotenoids are contained in foods, the most abundant carotenoid-containing foods being fruits and vegetables [3] . Approximately half of the dietary carotenoids can be found in human serum and tissues, with α-carotene, β-carotene, lycopene, lutein, zeaxanthin, and β-cryptoxanthin representing 90% of all the carotenoids in the human body [4] . Provitamin A carotenoids can be converted to retinol in the body and include β-carotene, α-carotene, and β-cryptoxanthin. Αlpha-carotene is obtained through the diet mainly from yellow-orange fruits and vegetables, especially carrots, pumpkin, and squash [2] . Although observational studies have consistently associated eating of carotenoid- containing fruits and vegetables with lower risk of a variety of chronic diseases, interven- tions with carotenoid-rich diets and supplementation have not shown consistent health benefits [5–7] . In fact, β-carotene supplementation has been associated with an increase in mortality among smokers, although it should be noted that a high daily dose of synthetic β-carotene was utilized in that study [8] . The discrepancy between observational and clinical studies may be due in part to genetic differences in how circulating carotenoid concentrations and their physiological effects respond to carotenoid intake. Alpha-carotene has not been studied as extensively as β-carotene, but it is believed to have numerous bene- ficial health effects. Alpha-carotene is a potent antioxidant that prevents lipid oxidation, scavenges free radicals, and enhances gap junction communication [9] . In vivo studies suggest that the ability of α-carotene to inhibit proliferation of cancer cells is approximately 10 times more potent than that of β-carotene [10] . In addition, α-carotene has been asso- ciated in observational and animal studies with a decreased risk for gastric cancers, liver and lung cancers, and other chronic diseases, including diabetes, cardiovascular disease, and chronic lower respiratory diseases [2, 11, 12] . Due in part to its antioxidant properties, low levels of α-carotene have also been associated with diseases that are worsened by excessive oxidative stress, such as glaucoma and atrophic gastritis [13, 14] . Furthermore, data from the National Health and Nutrition Examination Surveys revealed an inverse asso- ciation between serum α-carotene concentrations and the risk of all-cause mortality as well as mortality from cardiovascular disease, cancer, and all other causes not related to cardio- vascular disease or cancer [2] . Therefore, a greater understanding of genetic associations with circulating α-carotene concentrations in humans may have important public health implications. There is one previously reported genome-wide association study (GWAS) of α-carotene concentrations. In a meta-analysis that combined results across 3 diverse study populations ( n = 3,881 subjects total), a single locus was found to be associated with provitamin A carot- enoids at a genome-wide threshold for statistical significance. In this study, single nucleotide polymorphisms (SNPs) in the gene BCO1 were associated with significantly higher circulating concentrations of β-carotene, and the same locus was associated to a lesser extent with α-carotene concentrations [15] . BCO1 catalyzes the first step of provitamin A carotenoids to vitamin A (retinol) in the small intestine. BCO1 is a 15-15 ′ dioxygenase that performs a 15-15 ′ central cleavage that generates one molecule of retinol from each molecule of α-carotene. The association between SNPs in BCO1 and α-carotene was weak ( p = 0.001), relative to its asso- ciation with β-carotene ( p = 1.6 × 10 –24 ), and did not have convincing evidence of consistency in the original report ( p = 0.057 in the replication cohort). Little else is known about genetic variants that may influence serum α-carotene concentrations, and no robustly associated genetic associations with this important micronutrient have been identified to date. 256J Nutrigenet Nutrigenomics 2016;9:254–264 DOI: 10.1159/000452890 D’Adamo et al.: The CAPN2/CAPN8 Locus on Chromosome 1q Is Associated with Variation in Serum Alpha-Carotene Concentrations www.karger.com/jnn © 2016 S. Karger AG, Basel The goal of this study was to identify novel genetic associations with serum α-carotene concentrations. We studied a relatively genetically homogenous population of Old Order Amish men and women living in Lancaster County, Pennsylvania. The study participants were administered a 6-day controlled diet, which helped reduce variability in serum α-carotene concentrations associated with varying intake of α-carotene and fat as well as other dietary factors that influence its absorption. We hypothesized that the controlled diet would enable us to more closely isolate the genetic influences of serum α-carotene concentrations. While relatively brief, the controlled diet helped reduce potential confounding of the relationship between genetic variants and circulating α-carotene concentrations introduced by varying dietary intake. Methods Study Population The study sample was composed of the 433 participants from the Heredity and Phenotype Intervention (HAPI) Heart Study who completed a controlled diet and had frozen blood samples obtained during a clinic visit on the final day of the controlled diet. The design of the HAPI Heart Study has been described previously [16] . Briefly, the initial aim of the HAPI Heart Study was to identify genetic and environmental determinants of the responses to intervention affecting cardiovascular risk factors. Of the 868 Old Order Amish adults recruited into the HAPI Heart study, 469 were administered the controlled dietary intervention and subse- quently provided a blood sample at the conclusion of the 6-day diet. There were 27 samples of insufficient quality to measure serum α-carotene concentrations and 8 participants with incomplete genotype data who were excluded from analysis. The study was approved by the Institutional Review Board of the University of Maryland School of Medicine, and all participants provided written informed consent. Controlled Diet Research staff prepared the controlled diet for study participants. A registered dietitian visited several Old Order Amish households to obtain diet histories and observe meals and foods that were in their homes. All meals in the controlled diet were designed to be representative of the typical diets of Old Order Amish adults and were delivered to the homes of study participants during the 6-day controlled diet period. Study participants did not consume any prescribed or over-the-counter medications or dietary supplements during this 6-day period. The complete menus for the 6-day controlled diet that the study participants consumed are provided in the online supplementary material (see www.karger.com/doi/10.1159/000452890). The controlled diet contained an average of 3,277 kilocalories per day, with 49% from carbohydrate, 36% from fat, and 15% from protein. There was an average of 525 mg of cholesterol per day in the diet. The diet contained approximately 1,724 μg of α-carotene per day, coming primarily from carrots, green beans, and lettuce. Compliance with the controlled diet was assessed by comparing sodium, potassium, and creatinine levels from first morning urine samples obtained (1) prior to consuming the 6-day controlled diet, (2) on the final day of the 6-day controlled diet, and (3) on the final day of a second 6-day controlled diet that was low in salt and consumed after the blood draw that was used to conduct the GWAS. The compliance data have been reported in detail previously [17] . In brief, while not a direct measure of carotenoid intake, the excreted sodium, potassium, and creatinine levels reflecting the varying salt content in the different diets suggest that compliance with the controlled diet in this study was excellent. Serum Micronutrient Measurement Frozen blood samples taken on the final day of the 6-day controlled diet were assayed for serum concen- trations of α-carotene, other key carotenoids (β-carotene, lycopene, lutein, zeaxanthin, cryptoxanthin), vitamin E (γ-tocopherol, α-tocopherol), and retinol (preformed vitamin A) at the Johns Hopkins University Nutritional Biochemistry Laboratory. Reverse-phase high-pressure liquid chromatography was utilized to assess serum α-carotene concentrations from each 200 μL frozen blood sample [18] . The intra- and inter- assay coefficients of variability for α-carotene were 8.2 and 19.4%, respectively. 257J Nutrigenet Nutrigenomics 2016;9:254–264 DOI: 10.1159/000452890 D’Adamo et al.: The CAPN2/CAPN8 Locus on Chromosome 1q Is Associated with Variation in Serum Alpha-Carotene Concentrations www.karger.com/jnn © 2016 S. Karger AG, Basel Genotyping Genotypes were obtained using either the Affymetrix 500K or the Affymetrix 1M SNP chip v6.0 by the Genomics Core Laboratory at the University of Maryland. Genotyping calls were made separately for the 2 chips using BRLMM (500K array) and Birdseed version 2 (1M array), which is part of the Birdsuite tools [19, 20] . Called genotypes were then synthesized and filtered by excluding markers with >5% missing and extreme Hardy-Weinberg equilibrium ( p < 5 × 10 –6 ). A total of 397,704 SNPs passed quality control stan- dards, were called per each calling algorithm, were in common on both arrays, and had a minor allele frequency (MAF) ≥ 1%. These SNPs were then passed to the imputation phase. Imputation was conducted with MACH using the HapMap CEU reference sample [21] . An imputation quality score of ≥ 30% was used as a final quality control filter. Variants with an imputed MAF ≥ 2% were used in association analyses for a final analyzed SNP count of 2,193,082. HaploReg version 4.1 [22] was utilized to pull data from ENCODE and the Roadmap Epigenomic projects to annotate the functions of any variants associated with α-carotene concentrations in our analyses. The ENCODE and the Roadmap Epigenomic projects provide data across a variety of tissue and cell types. For each associated variant, we looked for its predicted chromatin state across multiple cell types, its effect on regulatory motifs, and potential enrichment of cell type-specific enhancers. Statistical Methods Descriptive statistics were computed to characterize the study sample at baseline and determine the mean serum α-carotene concentrations. We estimated the effect of genotype on α-carotene levels at each SNP, adjusting for the effects of age and sex by utilizing a general linear model. Genotype was coded as the number of copies of the reference allele (0, 1, or 2), thereby corresponding to an additive genetic model. GWAS analyses were performed using the MMAP software, which accounts for family structure [23] . Statis- tical analysis was performed using a variance component approach to account for relatedness among study participants. This approach has been shown to provide valid estimates of regression parameters [24] . We estimated that our sample provided 80% power to detect SNPs accounting for 9–10% of trait variation at genome-wide thresholds for statistical significance ( p < 5 × 10 –8 ). Our secondary aim of the statistical analysis was to perform replicative association testing of previously reported associations. Ferrucci et al. [15] previously reported associations with serum α-carotene, with one locus achieving genome-wide significance. We here report replicative association tests for rs6564851 at the BCO1 locus. Table 1. Characteristics of our Old Order Amish study population from Lancaster County, PA after consuming a 6-day controlled diet Characteristic All (n = 433) Female (n = 181) Male (n = 252) p value1 Age, years 43.1 ± 12.3 45.6 ± 13.1 41.4 ± 12.5 <0.0001 BMI 26.4 ± 4.22 27.8 ± 5.06 25.4 ± 3.13 <0.0001 Lycopene, μmol/L 0.73 ± 0.37 0.71 ± 0.33 0.74 ± 0.40 0.2 Lutein, μmol/L 0.25 ± 0.10 0.24 ± 0.09 0.27 ± 0.10 0.004 Zeaxanthin, μmol/L 0.12 ± 0.06 0.11 ± 0.05 0.13 ± 0.06 0.01 Cryptoxanthin, μmol/L 0.16 ± 0.07 0.16 ± 0.07 0.16 ± 0.06 0.9 Αlpha-carotene, μmol/L 0.29 ± 0.23 0.32 ± 0.25 0.27 ± 0.21 0.04 Βeta-carotene, μmol/L 0.70 ± 0.50 0.78 ± 0.58 0.64 ± 0.42 0.006 Retinol, μmol/L 1.53 ± 0.37 1.51 ± 0.37 1.55 ± 0.36 0.3 Gamma-tocopherol, μmol/L 4.68 ± 1.68 4.71 ± 1.81 4.66 ± 1.58 0.6 Alpha-tocopherol, μmol/L 30.5 ± 7.79 31.4 ± 8.02 29.8 ± 7.57 0.07 1 The Student t test was utilized to calculate p values for differences in all variables between females and males. 258J Nutrigenet Nutrigenomics 2016;9:254–264 DOI: 10.1159/000452890 D’Adamo et al.: The CAPN2/CAPN8 Locus on Chromosome 1q Is Associated with Variation in Serum Alpha-Carotene Concentrations www.karger.com/jnn © 2016 S. Karger AG, Basel Results The baseline characteristics of the study sample are provided in Table 1 . There were 252 men and 181 women in the study. The participants had a mean age of 43.1 years and a mean BMI of 26.4; mean α-carotene concentrations were 0.29 μmol/L. The residual heritability of α-carotene concentrations after accounting for age and sex was estimated to be 0.23 ± 0.11. A Manhattan plot summarizing the results of the GWAS is provided in Figure 1 . There was little evidence for genomic inflation (lambda = 1.00), as illustrated in Figure 2 . Details of the genome-wide significant ( p < 5 × 10 –8 ) associations are given in Table 2 . 8 6 4 2 0 1 2 3 4 5 6 7 8 Chromosome –l o g 10 (p ) 9 10 11 12 13 14 15 16 18 20 22 8 10 GC lambda = 1 6 4 2 0 0 2 4 6 8 10 Theoretical p values (–log10) O b se rv ed p v al ue s (– lo g 10 ) Fig. 1. Manhattan plot for the genome-wide association study of serum α-carotene concentrations in the study population following a 6-day controlled diet. The x axis represents the chromosomal position along the genome. The y axis shows the p value for the association test at each locus on the log scale. Fig. 2. Quantile-quantile plot of the genome-wide association study of serum α-carotene con- centrations. The axes plot the ob- served ( y axis) versus theoretical ( x axis) association p values on the log scale for all single nucleo- tide polymorphisms with a minor allele frequency >1%. The Old Or- der Amish are a closed founder population with little admixture expected. The genomic control (GC) lambda is estimated to be 1.00, indicating little bias due to population stratification. C o lo r ve rs io n av ai la b le o nl in e C o lo r ve rs io n av ai la b le o nl in e 259J Nutrigenet Nutrigenomics 2016;9:254–264 DOI: 10.1159/000452890 D’Adamo et al.: The CAPN2/CAPN8 Locus on Chromosome 1q Is Associated with Variation in Serum Alpha-Carotene Concentrations www.karger.com/jnn © 2016 S. Karger AG, Basel Three novel loci with genome-wide significant associations with α-carotene concentra- tions were detected on chromosomes 1, 2, and 4. The strongest genetic evidence of an associ- ation exists on chromosome 1q41. Figure 3 provides a regional association plot of the genome- wide significant association on chromosome 1. Several SNPs in linkage disequilibrium tag this association. The minor allele at the lead SNP (rs12137025, p = 3.55 × 10 –8 ) was associated with a 0.19 μmol/L increase in serum α-carotene, and this locus accounted for 7.1% of the variation in α-carotene levels. The minor allele at the locus C is present in the Old Order Amish (MAF = 0.069) at frequencies similar to those of other populations (HapMap CEU = 0.062). Two other genome-wide significant associations were identified, though there was little corroborating evidence from nearby SNPs at those loci. The strongest of the associations with α-carotene concentrations was a locus on chromosome 2p21 (SNP = rs2594495; p = 1.01 × 10 –8 ). Each copy of the A allele was associated with a 0.37 μmol/L increase in serum α-carotene, and this locus accounted for 7.2% of the variation in α-carotene levels. Figure 4 provides a regional association plot of the genome-wide significant association on chromosome 2. The MAF at the locus A is present in the Old Order Amish (MAF = 0.039), though it is more frequent in other populations (HapMap CEU = 0.11). The rs2594495 variant is situated in an intronic region of the PRKCE gene. Finally, we identified a genome-wide significant association with a locus on chromosome 4q34 (lead SNP = rs17830069; p = 2.89 × 10 –8 ). Each copy of the G allele was associated with a 0.38 μmol/L increase in serum α-carotene, and this locus accounted for 7.5% of the vari- ation in α-carotene levels. The minor allele at the locus G is present in the Old Order Amish (MAF = 0.023) as well as other populations (HapMap CEU = 0.050). Figure 5 provides a regional association plot of the genome-wide significant association on chromosome 4; rs17830069 is in an intergenic region near the non-protein-coding RNA LINC00290 . There are no obvious candidate genes in the region. We also performed look-ups for SNPs at the BCO1 locus previously reported to be asso- ciated with serum β-carotene levels in the GWAS meta-analysis reported by Ferrucci et al. [15] . There was no evidence of an association between this locus and α-carotene concentra- tions in our study population (lead SNP: rs6564851, p = 0.28). GWAS was also performed for the other carotenoids (β-carotene, lycopene, lutein, zeaxanthin, cryptoxanthin), vitamin E (α-tocopherol, γ-tocopherol), and retinol that were assayed. In brief, there were no meaningful associations noted between retinol, vitamin E, or any of the other carotenoids and any other genetic loci ( p > 5 × 10 –8 , data not shown), with the exception of lycopene. Serum lycopene concentrations among our study population were associated ( p = 3.41 × 10 –9 ) with the variant rs7680948 on chromosome 4, located in the intron region of the SETD7 gene [25] . In addition, our results offered nominal support ( p = 3.79 × 10 –4 ) for the association previously noted between SCARB1 and serum lycopene levels, albeit with a different variant (rs11057841) in the region. Table 2. Genome-wide (p < 5 × 10 – 8) associations with serum α-carotene concentrations SNP Chromo- some Position Gene MAF Coded allele Beta (SE) p value rs2594495 2 46262970 PRKCE 0.04 A 0.37 (0.06) 1.01 × 10 – 8 rs17830069 4 182415300 0.02 A –0.38 (0.07) 2.89 × 10 – 8 rs12137025 1 221938674 CAPN8, CAPN2 0.07 C 0.19 (0.03) 3.55 × 10 – 8 MAF, minor allele frequency; SE, standard error. 260J Nutrigenet Nutrigenomics 2016;9:254–264 DOI: 10.1159/000452890 D’Adamo et al.: The CAPN2/CAPN8 Locus on Chromosome 1q Is Associated with Variation in Serum Alpha-Carotene Concentrations www.karger.com/jnn © 2016 S. Karger AG, Basel Discussion We report the results of a GWAS of serum α-carotene concentrations among a study sample that consumed a controlled diet. Three genome-wide significant associations were identified, though all three require further replication from independent studies and fine mapping to help determine the potential functional effects of these associations. The CAPN2/ CAPN8 locus on chromosome 1 has the strongest evidence for association due to the associ- ation being identified with relatively common genetic markers (MAF ∼ 7%) and multiple SNPs. The other two loci, the PRKCE locus on chromosome 2 and the chromosome 4q34 locus, lack strong collaborating evidence from nearby SNPs. While having associated SNPs in strong linkage disequilibrium is not a requirement of a causal genetic determinant, it lends supporting statistical evidence for a meaningful association and lowers the probability of an association being identified due to genotype error. We were unable to replicate the association previ- ously noted between a variant on BCO1 and α-carotene concentrations. However, this is not surprising because this SNP was more strongly associated with β-carotene than α-carotene concentrations and did not have strong replicative evidence in the initial publication [15] . GWAS performed on retinol, vitamin E (α-tocopherol, γ-tocopherol), and the other carot- 8 10 Plotted SNPs 0.8 r2 0.6 0.4 0.2 rs12137025 6 4 2 0 SUSD4 CAPN8 CAPN2 FBXO28 DEGS1C1orf65 TP53BP2 –l o g 10 (p ) 80 100 60 40 20 0 Re co m b in at io n ra te (c M /M b ) 222.4222.2222.0221.8221.6 Position on chromosome 1 (Mb) Fig. 3. Regional association plot of chromosome 1q41. The x axis represents the chromosomal position with the location of genes at the locus annotated. The left y axis shows the p value for association tests at each lo- cus (dot) on the log scale. The right y axis provides recombination rates in centimorgans per megabase (cM/ Mb) in the chromosomal region identifying recombination hotspots in the region (blue line). The most sig- nificantly associated variant in the region is indicated by a purple diamond. All other variants in the region are indicated with colored solid dots. The color reflects the linkage disequilibrium (measured in r 2 ) between the variant and the top hit. Variants in high linkage disequilibrium ( r 2 > 0.8) are in red, variants in low link- age disequilibrium are in dark blue ( r 2 < 0.2), with all other colors representing linkage disequilibrium in between (0.2 ≤ r 2 ≤ 0.8), as described in the color legend in the upper left hand corner of the figure. C o lo r ve rs io n av ai la b le o nl in e 261J Nutrigenet Nutrigenomics 2016;9:254–264 DOI: 10.1159/000452890 D’Adamo et al.: The CAPN2/CAPN8 Locus on Chromosome 1q Is Associated with Variation in Serum Alpha-Carotene Concentrations www.karger.com/jnn © 2016 S. Karger AG, Basel 8 10 Plotted SNPs 0.8 r2 0.6 0.4 0.2 rs2594495 6 4 2 0 PIGFEPAS1 LOC388946 RHOQ CRIPT PRKCE ATP6V1E2 –l o g 10 (p ) 80 100 60 40 20 0 Re co m b in at io n ra te (c M /M b ) 46.646.446.246.045.8 Position on chromosome 2 (Mb) 8 10 Plotted SNPs 0.8 r2 0.6 0.4 0.2 rs17830069 6 4 2 0 LINC00290 –l o g 10 (p ) 80 100 60 40 20 0 Re co m b in at io n ra te (c M /M b ) 182.4182.2182.0181.8 Position on chromosome 4 (Mb) Fig. 5. Regional association plot of chromosome 4q34. See Figure 3 for details. Fig. 4. Regional association plot of chromosome 2p2. See Figure 3 for details. C o lo r ve rs io n av ai la b le o nl in e C o lo r ve rs io n av ai la b le o nl in e 262J Nutrigenet Nutrigenomics 2016;9:254–264 DOI: 10.1159/000452890 D’Adamo et al.: The CAPN2/CAPN8 Locus on Chromosome 1q Is Associated with Variation in Serum Alpha-Carotene Concentrations www.karger.com/jnn © 2016 S. Karger AG, Basel enoids (β-carotene, lycopene, lutein, zeaxanthin, cryptoxanthin) assayed in this study revealed no other meaningful genetic associations, with the exception of lycopene [25] . Our results introduce a potentially interesting relationship between CAPN2/CAPN8 , α-carotene, and gastric cancer. Calpains are a group of calcium-sensitive cysteine proteases that are ubiquitously expressed in mammals. The calpain system has been shown to be dysregulated in cancer [26] . CAPN8 is a stomach-specific calpain whose expression is localized to gastric pit cells. Functioning CAPN8 genes produce nCL-2, a cysteine protease essential for mucosal defense [27] . Dysfunction of this protease and the pit cells in which they are expressed can lead to atrophic gastritis and metastatic gastric cancers [28] . We determined that the genetic variant rs12137025, found near CAPN8 , was associated with higher serum α-carotene concentrations. Higher levels of α-carotene have been shown to reduce the risk of atrophic gastritis, a precancerous condition associated with gastric cancer [13] . Alpha-carotene protects the gastric mucosa by scavenging free radicals and preventing the initiation or prop- agation of lipid peroxidation reactions [29] . While our findings suggest that there could potentially be a relationship between dysfunction of CAPN8 , lower circulating α-carotene concentrations, and gastric cancer mediated by reduced protection of the gastric mucosa from lipid peroxidation, this potential relationship will need to be evaluated in future studies. Interestingly, though the relationship noted between the PRKCE locus and α-carotene concentrations relies on a single SNP, PRKCE may also have implications in cancer as uncon- trolled chronic activation of PRKCE has been shown to lead to malignant tumor development [30] . Additional studies are clearly needed to confirm whether α-carotene concentrations play a role in mediating the oncogenic activity of these genes. The significant association on chromosome 4q34 is in a region with no obvious candidate genes. More work is required, and replication evidence needed, to verify this and all other associations reported here. There are several notable strengths of the study which may have resulted in successful identification of associated loci. This was the first GWAS aimed at identifying genetic associa- tions with serum α-carotene concentrations that was conducted among participants who had consumed a controlled diet. While we did not assess serum micronutrient concentrations prior to the controlled diet, as determining the specific effects of this diet on α-carotene concentrations was beyond the scope of the current study, the controlled diet minimized the potential for confounding of the relationship between serum α-carotene and genetic variants due to differences in dietary intake among people consuming variable diets. The controlled diet enabled us to more closely isolate the genetic contributions to the variance in serum α-carotene. The diet was designed to be culturally appropriate, representative of the typical diet of our study population, and was delivered to study participants at their homes to support compliance. It should be noted that the controlled diet contained more α-carotene (1,724 μg), as is typical of the diet of our Old Order Amish study population, than that of the average adult population of the United States (451 μg), as expressed in the most recent publicly available National Health and Nutrition Examination Surveys data [31] . Urinary excretion tests suggested that adherence to the controlled diet was excellent. The Old Order Amish study population also provided unique advantages in a study of this nature. The relationship structure of the Old Order Amish enabled us to perform the first estimate of the heritability of serum α-carotene concentrations in humans. The Old Order Amish are also a relatively homogenous population with respect to both genetics and lifestyle. The genetic homogeneity provided increased power to detect genetic variants associated with α-carotene, and the similar lifestyles further minimized potential sources of confounding of the study findings. There were also a number of key limitations to this study. The relatively small sample size ( n = 433) provided power to detect associations of genetic variants with relatively large effect sizes only. Despite our relatively small sample size, our study was able to identify 3 novel loci associated with serum α-carotene concentrations. We believe that these successes 263J Nutrigenet Nutrigenomics 2016;9:254–264 DOI: 10.1159/000452890 D’Adamo et al.: The CAPN2/CAPN8 Locus on Chromosome 1q Is Associated with Variation in Serum Alpha-Carotene Concentrations www.karger.com/jnn © 2016 S. Karger AG, Basel may be attributable in part to the controlled diet administered to study participants prior to the blood draw and the relatively similar lifestyle habits of the Old Order Amish study popu- lation which enabled us to more closely isolate the genetic contributions to serum α-carotene concentrations. While the novel association between a variant near CAPN8 and serum α-carotene concentrations, both of which have been associated with gastritis and gastric cancer, may provide the rationale for further study into the specific mechanisms of this rela- tionship, this study did not collect data on family history of gastric cancer or other markers of the disease, and thus no direct inference can be made. Bioinformatics analysis utilizing HaploReg demonstrated that some of the variants associated with α-carotene concentrations, or other variants in high linkage disequilibrium with them, have potential regulatory function by virtue of being associated with enhancer or promoter activity across multiple tissues. However, without more direct experimental evidence, it is not possible to ascribe a regulatory function of these SNPs to nearby genes (e.g., CAPN2 or CAPN8 ), nor the nearby genes to regu- lation of α-carotene concentrations. In summary, this was the first study to identify genome-wide significant associations with serum α-carotene concentrations. Three novel genome-wide associations were noted with serum α-carotene levels: the CAPN2/CAPN8 locus, PRKCE , and the chromosome 1q41 locus. These findings suggest that genetics may influence serum concentrations of α-carotene. Further studies are needed to replicate these findings and to identify potential mechanisms underlying the relationships between the identified genetic loci, α-carotene concentrations, and clinical endpoints such as gastric cancer. Acknowledgments This work was supported by the National Institutes of Health (U01 HL072515, P30 DK072488, R01 AG27012, and T35 DK095737). The authors would also like to acknowledge the study participants for their partnership in research, the staff at the Amish Research Clinic for their collaboration, and Nga Hong Brereton, MS, RD, at Johns Hopkins University School of Medicine for designing the controlled diet in this study and evaluating its nutrient content. Disclosure Statement The authors do not have any conflicts of interest to disclose. References  1 Á lvarez R, Vaz B, Gronemeyer H, de Lera AR: Functions, therapeutic applications, and synthesis of retinoids and carotenoids. Chem Rev 2014; 114: 1–125.  2 Li C, Ford ES, Zhao G, Balluz LS, Giles WH, Liu S: Serum α-carotene concentrations and risk of death among US adults: the Third National Health and Nutrition Examination Survey Follow-up Study. Arch Intern Med 2011; 171: 507–515.  3 Chen H, Yueh T, Chen Y, Huang C, Yang C, Hu M: Antimetastatic effects of α-carotene and possible mechanisms of action in human hepatocarcinoma SK-hep-1 cells. J Agric Food Chem 2013; 61: 10368–10376.  4 Rao AV, Rao LG: Carotenoids and human health. Pharmacol Res 2007; 55: 207–216.  5 Hennekens CH, Buring JE, Manson JE, Stampfer M, Rosner B, Cook NR, Belanger C, LaMotte F, Gaziano JM, Ridker PM: Lack of effect of long-term supplementation with beta carotene on the incidence of malignant neoplasms and cardiovascular disease. N Engl J Med 1996; 334: 1145–1149.  6 Ried K, Fakler P: Protective effect of lycopene on serum cholesterol and blood pressure: meta-analyses of intervention trials. Maturitas 2011; 68: 299–310.  7 Writing Group for the AREDS2 Research Group, Bonds DE, Harrington M, Worrall BB, Bertoni AG, Eaton CB, Hsia J, Robinson J, Clemons TE, Fine LJ, Chew EY: Effect of long-chain ω-3 fatty acids and lutein + zeaxanthin supplements on cardiovascular outcomes: results of the Age-Related Eye Disease Study 2 (AREDS2) ran- domized clinical trial. JAMA Intern Med 2014; 174: 763–771. 264J Nutrigenet Nutrigenomics 2016;9:254–264 DOI: 10.1159/000452890 D’Adamo et al.: The CAPN2/CAPN8 Locus on Chromosome 1q Is Associated with Variation in Serum Alpha-Carotene Concentrations www.karger.com/jnn © 2016 S. Karger AG, Basel  8 Heinonen OP, Albanes D: The effect of vitamin E and beta carotene on the incidence of lung cancer and other cancers in male smokers. The Alpha-Tocopherol, Beta Carotene Cancer Prevention Study Group. N Engl J Med 1994; 330: 1029–1035.  9 Ziegler RG, Colavito E, Hartge P, McAdams MJ, Schoenberg J, Mason T, Fraumeni JF Jr: Importance of alpha- carotene, beta-carotene, and other phytochemicals in the etiology of lung cancer. J Natl Cancer Inst 1996; 88: 612–615. 10 Murakoshi M, Takayasu J, Kimura O, Kohmura E, Nishino H, Iwashima A, Okuzumi J, Sakai T, Sugimoto T, Imanishi J: Inhibitory effects of alpha-carotene on proliferation of the human neuroblastoma cell line GOTO. J Natl Cancer Inst 1989; 81: 1649–1652. 11 Murakoshi M, Nishino H, Satomi Y, Takayasu J, Hasegawa T, Tokuda H, Iwashima A, Okuzumi J, Okabe H, Kitano H: Potent preventive action of alpha-carotene against carcinogenesis: spontaneous liver carcinogenesis and promoting stage of lung and skin carcinogenesis in mice are suppressed more effectively by alpha-carotene than by beta-carotene. Cancer Res 1992; 52: 6583–6587. 12 Nomura AM, Ziegler RG, Stemmermann GN, Chyou PH, Craft NE: Serum micronutrients and upper aerodi- gestive tract cancer. Cancer Epidemiol Biomarkers Prev 1997; 6: 407–412. 13 Ito Y, Suzuki K: The effect of serum carotenoids on atrophic gastritis among the inhabitants of a rural area in Hokkaido, Japan. Environ Health Prev Med 2001; 6: 184–188. 14 Giaconi JA, Yu F, Stone KL, Pedula KL, Ensrud KE, Cauley JA, Hochberg MC, Coleman AL; Study of Osteoporotic Fractures Research Group: The association of consumption of fruits/vegetables with decreased risk of glaucoma among older African-American women in the study of osteoporotic fractures. Am J Ophthalmol 2012; 154: 635–644. 15 Ferrucci L, Perry JR, Matteini A, Perola M, Tanaka T, Silander K, Rice N, Melzer D, Murray A, Cluett C, Fried LP, Albanes D, Corsi AM, Cherubini A, Guralnik J, Bandinelli S, Singleton A, Virtamo J, Walston J, Semba RD, Frayling TM: Common variation in the beta-carotene 15,15 ′ -monooxygenase 1 gene affects circulating levels of carot- enoids: a genome-wide association study. Am J Hum Genet 2009; 84: 123–133. 16 Mitchell BD, McArdle PF, Shen H, Rampersaud E, Pollin TI, Bielak LF, Jaquish C, Douglas JA, Roy-Gagnon M, Sack P: The genetic response to short-term interventions affecting cardiovascular function: rationale and design of the Heredity and Phenotype Intervention (HAPI) Heart Study. Am Heart J 2008; 155: 823–828. 17 Montasser ME, Douglas JA, Roy-Gagnon MH, Van Hout CV, Weir MR, Vogel R, Parsa A, Steinle NI, Snitker S, Brereton NH, Chang YP, Shuldiner AR, Mitchell BD: Determinants of blood pressure response to low-salt intake in a healthy adult population. J Clin Hypertens (Greenwich) 2011; 13: 795–800. 18 Sowell AL, Huff DL, Yeager PR, Caudill SP, Gunter EW: Retinol, alpha-tocopherol, lutein/zeaxanthin, beta-cryp- toxanthin, lycopene, alpha-carotene, trans-beta-carotene, and four retinyl esters in serum determined simul- taneously by reversed-phase HPLC with multiwavelength detection. Clin Chem 1994; 40: 411–416. 19 Korn JM, Kuruvilla FG, McCarroll SA, Wysoker A, Nemesh J, Cawley S, Hubbell E, Veitch J, Collins PJ, Darvishi K, Lee C, Nizzari MM, Gabriel SB, Purcell S, Daly MJ, Altshuler D: Integrated genotype calling and association analysis of SNPs, common copy number polymorphisms and rare CNVs. Nat Genet 2008; 40: 1253–1260. 20 Affymetrix. BRLMM: an improved genotype calling method for the GeneChip ® human mapping 500K array set. 2006. 21 Li Y, Abecasis G: MACH 1.0. Ann Arbor, University of Michigan School of Public Health, http://csg.sph.umich. edu/abecasis/MACH/ (accessed May 7, 2016). 22 HaploReg, Version 4.1. Broad Institute. 23 MMAP documentation. University of Maryland School of Medicine. 24 McArdle PF, O’Connell JR, Pollin TI, Baumgarten M, Shuldiner AR, Peyser PA, Mitchell BD: Accounting for relat- edness in family based genetic association studies. Hum Hered 2007; 64: 234–242. 25 D’Adamo CR, D’Urso A, Ryan KA, Yerges-Armstrong LM, Semba RD, Steinle NI, Mitchell BD, Shuldiner AR, McArdle PF: A common variant in the SETD7 gene predicts serum lycopene concentrations. Nutrients 2016; 8: 82. 26 Ho WC, Pikor L, Gao Y, Elliott BE, Greer PA: Calpain 2 regulates Akt-FoxO-p27(Kip1) protein signaling pathway in mammary carcinoma. J Biol Chem 2012; 287: 15458–15465. 27 Hata S, Abe M, Suzuki H, Kitamura F, Toyama-Sorimachi N, Abe K, Sakimura K, Sorimachi H: Calpain 8/nCL-2 and calpain 9/nCL-4 constitute an active protease complex, G-calpain, involved in gastric mucosal defense. PLoS Genet 2010; 6:e1001040. 28 Zhao Z, Sun Y, Hou N, Teng Y, Wang Y, Yang X: Capn8 promoter directs the expression of Cre recombinase in gastric pit cells of transgenic mice. Genesis 2009; 47: 674–679. 29 Nair S, Norkus EP, Hertan H, Pitchumoni C: Micronutrient antioxidants in gastric mucosa and serum in patients with gastritis and gastric ulcer: does Helicobacter pylori infection affect the mucosal levels? J Clin Gastroen- terol 2000; 30: 381–385. 30 Akita Y: Protein kinase C-epsilon (PKC-epsilon): its unique structure and function. J Biochem 2002; 132: 847– 852. 31 U.S. Department of Agriculture, Agricultural Research Service: What We Eat in America, NHANES 2011–2012. Nutrient intakes from food and beverages: mean amounts consumed per individual, by gender and age, in the United States, 2011–2012. Available from http://www.ars.usda.gov/SP2UserFiles/Place/80400530/pdf/ 1112/Table_1_NIN_GEN_11.pdf (accessed May 7, 2016).