Eur J Nutr

DOI 10.1007/s00394-016-1162-8

ORIGINAL CONTRIBUTION

Fruit and vegetable intake and vitamin C transporter gene

(SLC23A2) polymorphisms in chronic lymphocytic leukaemia
Delphine Casabonne1,2 · Esther Gracia2,3,4 · Ana Espinosa2,3,4,5 ·
Mariona Bustamante2,3,5,6 · Yolanda Benavente1,2 · Claudia Robles1 ·
Laura Costas1,2 · Esther Alonso7 · Eva Gonzalez‑Barca8 · Adonina Tardón2,9 ·
Trinidad Dierssen‑Sotos2,10 · Eva Gimeno Vázquez11,12 · Marta Aymerich13 ·
Elies Campo13 · José J. Jiménez‑Moleón2,14,15 · Rafael Marcos‑Gragera16 ·
Gemma Castaño‑Vinyals2,3,4,5 · Nuria Aragones2,17,18 · Marina Pollan2,17,18 ·
Manolis Kogevinas2,3,4,5,19 · Carmen Urtiaga20 · Pilar Amiano2,21 · Victor Moreno2,22 ·
Silvia de Sanjose1,2 

Received: 23 September 2015 / Accepted: 18 January 2016
© Springer-Verlag Berlin Heidelberg 2016
Abstract  food frequency questionnaires and six low-penetrance
Purpose  There is currently no convincing epidemio-
logical evidence that fruit and vegetable consumption, the
primary source of vitamin C, plays a role in chronic lym-
phocytic leukaemia (CLL) aetiology. We hypothesized that
variations in vitamin C dietary intake as well as in genetic
Electronic supplementary material  The online version of this
article (doi:10.1007/s00394-016-1162-8) contains supplementary
material, which is available to authorized users.
variability in vitamin C transporter gene SLC23A2 could
explain some inconsistencies in the literature.
Methods  Fruit/vegetable/vitamin C consumption from
genetic susceptibility polymorphisms in vitamin C trans-
porter gene SLC23A2 (rs1715364, rs6133175, rs1776948,
rs6139587, rs369270 and rs6052937) were examined in
434 CLL cases and 1257 randomly selected controls from
primary care centres with genetic data of whom 275 cases
and 1094 controls having both diet and genetic informa-
tion. Logistic regression models were used to estimate odds
ratio (OR) and 95 % confidence intervals (CI).

10
* Delphine Casabonne Faculty of Medicine, University of Cantabria- IDIVAL,
dcasabonne@iconcologia.net 39011 Santander, Spain
11
1 Department of Clinical Hematology, Hospital del Mar,
Unit of Infections and Cancer (UNIC), Cancer Epidemiology
08003 Barcelona, Spain
Research Programme, IDIBELL, Institut Català d’Oncologia,
12
Av. Gran Via 199 ‑ 203, 2º, 08907 L’Hospitalet De Llobregat, Grup de Recerca Aplicada en Neoplasies Hematològiques-
Barcelona, Spain PSMAR, Barcelona, Spain
2 13
CIBER Epidemiología y Salud Pública (CIBERESP), Hematopathology Unit, Pathology Department, Hospital
28029 Madrid, Spain Clínic, University of Barcelona, Institute of Biomedical
3 Research August Pi i Sunyer (IDIBAPS), 08036 Barcelona,
Centre for Research in Environmental Epidemiology
Spain
(CREAL), 08003 Barcelona, Spain
14
4 Department of Preventive Medicine and Public Health,
Universitat Pompeu Fabra (UPF), 08002 Barcelona, Spain
University of Granada, 18071 Granada, Spain
5
Hospital del Mar Medical Research Institute (IMIM), 15
Instituto de Investigación Biosanitaria de Granada, Servicio
Barcelona, Spain
Andaluz de Salud/Universidad de Granada, 18012 Granada,
6
Center for Genomic Regulation (CRG), 08003 Barcelona, Spain
Spain 16
Epidemiology Unit and Girona Cancer Registry, Oncology
7
Department of Pathology, Hospital Universitari de Bellvitge, Coordination Plan, Department of Health, Autonomous
08907 L’Hospitalet De Llobregat, Barcelona, Spain Government of Catalonia, Catalan Institute of Oncology,
8 Girona Biomedical Research Institute (IdiBGi),
Hematology, IDIBELL, Institut Català d’ Oncologia,
17007 Girona, Spain
08907 L’ Hospitalet De Llobregat, Barcelona, Spain
17
9 National Center for Epidemiology, Carlos III Institute
Oncology Institute (IUOPA), University of Oviedo,
of Health, 28029 Madrid, Spain
33006 Oviedo, Austria

13

Results  CLL patients were more likely to have a higher
fruit consumption than controls (highest versus low-
est quartile in g/day OR: 1.48; 95 % CI: 1.00 to 2.18;
P  = 0.03), whereas no associations were found with veg-
etable or total vitamin C intake. Based on log-additive mod-
els, rs6133175_A > G (OR: 1.19, 95 % CI: 1.00 to 1.41;
P = 0.05) and rs1776948_T > A (OR: 1.20; 95 %CI: 1.01 to
1.41; P = 0.04) were associated with CLL. The haplogeno-
type analysis (rs1715364, rs6133175) supported the geno-
type results. No gene–diet interactions in CLL remained
statistically significant after correction for multiple testing.
Conclusions  These data suggest that both fruit intake and
genetic marker in SLC23A2 may play an independent role
in CLL biology.
Keywords  Fruit intake · Vegetable intake · Vitamin C
intake · SLC23A2 · Polymorphism · Chronic lymphocytic
leukaemia
Introduction
Vitamin C, also known as ascorbic acid, is a water-soluble
vitamin which cannot be synthesized by humans and must
be obtained from the diet, mainly from fresh fruits and
vegetables. Vitamin C is a strong antioxidant that regulates
the synthesis of collagen, carnitine, catecholamine and the
neurotransmitter norepinephrine [1]. Despite being widely
believed, the beneficial role of the antioxidant activity of
vitamin C on cancer is controversial [2]. Nutritional experts
from the World Cancer Research Fund concluded that there
is currently no convincing epidemiological evidence that
fruits and vegetables play a role in cancer aetiology with
the exception of colorectal cancer and fibre intake [3]. In
particular, nutritional epidemiological data on a potential
effect of vitamin C in the pathogenesis of chronic lym-
phocytic leukaemia (CLL) are sparse and conflicting. Two
European cohort studies have pointed towards an increased
risk of CLL with higher intake of citrus [4] and fruit [5].
However, overall results from cohort [4–7] and case–con-
trol [8–10] studies are mainly inconclusive, which might be
18
Instituto de Investigación Sanitaria (IIS), 28222 Puerta De
Hierro Majadahonda, Spain
19
National School of Public Health, 115 21 Athens, Greece
20
Public Health Division of Gipuzkoa, Basque Health
Department, 20013 San Sebastian, Spain
21
Public Health Division of Gipuzkoa, BioDonostia Research
Institute, Basque Health Department, 20013 San Sebastian,
Spain
22
Cancer Prevention and Control Program, IDIBELL,
Institut Català d’Oncologia, University of Barcelona,
08907 L’Hospitalet De Llobregat, Barcelona, Spain
13
Eur J Nutr
partly due to sample size issues but also to large differences
in fruit and vegetable intake across countries.
Variations in genetic susceptibility of the studied popu-
lation could explain some of the inconsistencies reported
in the published literature in relation to fruit and vegeta-
ble intake and CLL risk. Polymorphisms in the vitamin C
receptor, solute carrier family 23 member 2 (SLC23A2),
have been reported to be associated with non-Hodgkin
lymphoma, colorectal adenocarcinoma, gastric cancer and
HPV-associated head and neck squamous cell carcinoma
[11–14]. In a case–control study [15] examining the asso-
ciation between CLL and single nucleotide polymorphisms
(SNPs) in SLC23A2, three variants and one haplotype in
SLC23A2 were related to CLL, but no information on nutri-
ents intake was reported. Many research studies focus on
either nutrition or genetic factors independently of each
other. Here, we hypothesized that the host response to die-
tary vitamin C intake could be modulated by the genetic
variants in vitamin C transporter gene SLC23A2.
Using data from the Spanish multi-case–control study
(MCC-Spain), we examined main and combined effect
of six genetic variants in the Na+-dependent L-ascorbic
acid transporters 2 (encoded by the SLC23A2 gene) and
fruit and vegetable intake, primary source of vitamin C in
humans, among cases with CLL and their controls.
Methods
Study participants
Cases were recruited within the MCC-Spain study and in
collaboration with the international cancer genome consor-
tium on chronic lymphocytic leukaemia project (ICGC–
CLL, www.cllgenome.es and www.icgc.org). The main
objective of the MCC-Spain study is to investigate lifetime
environmental, infectious, medical and occupational expo-
sures, and genetic factors associated with 5 cancer sites.
Further information can be found elsewhere [16]. In brief,
the recruitment of CLL cases took place between March
2010 and July 2012 and included CLL cases enrolled in
11 hospitals of 5 Spanish areas (Austria, Barcelona, Canta-
bria, Girona and Granada). Frequency-matched on age and
sex controls were randomly selected from the administra-
tive records of selected primary care health centres located
within these hospitals’ catchment area. All participants had
to be 20–85 years, to have resided in the catchment area
for at least 6 months prior to recruitment, and to be able
to answer the epidemiological questionnaire. Response
rates were 87 and 53 % for cases and controls, respectively.
Information was requested through a computerized face-to-
face interview (the questionnaire in Spanish is available at
www.mccspain.org). Prior to interview, a signed consent
Eur J Nutr
form was requested to be incorporated into the study. This
form included acceptance of analysis of biological material
and verification of clinical information.
Outcome definition
CLL cases were diagnosed according to the criteria of the
International Workshop on Chronic Lymphocytic Leukae-
mia [17]. All diagnoses were morphologically and immu-
nologically confirmed using flow cytometry immunophe-
notype and complete blood count. Disease severity was
evaluated using the Rai staging system obtained at the
time of interview from medical records and verified by
local haematologists. For this study, Rai stages were cat-
egorized into two groups: A) low-risk category including
asymptomatic patients with lymphocytosis only (Rai 0)
and B) intermediate/high-risk category including patients
with lymphocytosis with or without lymphadenopathy,
hepatomegaly, splenomegaly, anaemia and/or thrombocyto-
penia (Rai I–IV). CLL and small lymphocytic lymphoma
were considered the same underlying disease [18]. Given
the indolent course of the disease, both newly diagnosed
and prevalent cases were included in the study and the
impact of the inclusion of cases with longer disease time
prevalence was examined through sensitivity analyses.
The median time between CLL diagnosis and interview
was 2.3 years with interquartile range of 0.8 and 4.9 years
(maximum: 29.1 years).
Genotyping
DNA was extracted from whole blood or saliva. Genome-
wide genotyping was performed using a customized version
of the Infinium Human Exome BeadChip Kit v.1.1 (Illumina)
that has 242 901 SNPs and allows the inclusion of approxi-
mately 6000 customized SNPs. A custom content was also
added to include additional SNPs in candidate genes. The
genotype calling was done with the GeneTrain2.0 algorithm
(GenomeStudio software) based on CHARGE clusters [19].
PLINK was used for the genetic data quality control [20],
and the standard checks were done: sample call rate, sex con-
cordance, heterozygosity, relatedness, population stratifica-
tion, minimum SNPs call rate of 95 % and departure from
Hardy–Weinberg equilibrium (HWE) for each SNP. The final
database included 232 396 SNPs with 94 659 of them mono-
morphics and 99 411 with a MAF lower than 1 %, 8853 with
MAF 1–5 %, and 29 473 with MAF > 5 %.
Single nucleotide polymorphisms in SLC23A2
The gene SLC23A2 with approximately 150 kbp is located
in human chromosome 20p12 and encodes the sodium-
dependent vitamin C transporter 2 (SVCT2) proteins.
Four intronic SNPs (rs1715364_A > G, rs6133175_A > G,
rs1776948_T > A and rs6139587_A > T) previously associ-
ated with CLL [15] as well as two SNPs from the genome-
wide genotyping (rs369270_A > G and rs6052937_C > A
at 3′ UTR SLC23A2) were included in the analysis. The six
SNPs passed Hardy–Weinberg equilibrium (P value >0.05).
Information for 2 SNPs: rs1776948 and rs6139587 was
missing for 1 individual. The genotype and allele frequency
of the six SNPs examined in control subjects is shown
in Online Resource 1. The minor allele frequency for
rs1715364, rs6133175, rs1776948, rs6139587, rs369270
and rs6052937 was, respectively, allele G: 45 %; allele G:
32 %; allele A: 45 %; allele T: 41 %; allele G: 39 % and
allele A: 18 %. The SNPs rs1715364 and rs6133175 were
in linkage disequilibrium (Online Resource 2).
Diet exposure assessment
The fruit and vegetable consumption was measured using a
validated semi-quantitative Spanish Food Frequency Ques-
tionnaire (FFQ) which was modified to include regional
products [21] and was used to assess usual dietary intake
during the last year [16] (available in Spanish at http://
www.mccspain.org). It was self-completed at home and
posted back using a stamped envelope. The consumption
of fruit and vegetables referred to the year preceding the
interview, and the question was phrased as: “How often
did you eat this food item?” The answers included eight
options: never or less than once a month, 1–3 times per
month, 1–2 times per week, 3–4 times per week, 5–6 times
per week, once per day, 2–3 times a day, 4 or more times
per day. Exposure was based on the reported intake of 17
fruits (orange/grapefruit/mandarin, banana, apple, pear,
grape, kiwi, strawberry, cherry, peach/apricot/nectarine,
figs, water melon/melon, plums, mango/papaya, avocado,
other fresh fruit, fresh orange juice and fresh tomato juice)
and 22 vegetables (lettuce, green leafy vegetable, tomatoes,
cucumber, onion, radish, asparagus, carrot, green beans,
spinach/chard/watercress, cauliflowers/broccoli/cabbage/
Brussels sprouts, aubergine/courgette, red peppers, green
peppers, mushrooms, peas, garlic, corn, artichoke, pump-
kin, sweet potatoes and other vegetables). One unit was
defined differently for each food item: for instance, one
fruit unit was defined as one orange or one apple or ten
olives or one cup of strawberries. A detailed definition of
a unit for each food item can be found in the food ques-
tionnaire at http://www.mccspain.org. Standard portions
were used for each item to calculate g/d intake. Total vita-
min C (mg/d) level and energy intake (kcal/d) were calcu-
lated using a food composition table (FCT) compiled from
the Centre de Enseñanza Superior de Nutrición y Dieté-
tica (CESNID) and other specific sources. Information on
the use of vitamin C supplement was not available. Total
13

carbohydrate intake (in g/day) was calculated and further
divided into total sugar intake (in g/day), total complex
carbohydrate intake (in g/day) and total fibre intake (in g/
day). A question on changes in diet in the last 5 years prior
to interview was phrased as: “Have you made important
changes in your diet due to health issues or other reasons
in the last five years?” Participants were then asked to
describe their changes related to meat, egg, fish and sea-
food, vegetable, legume, fresh fruit, pasta, bread, olives, oil
and/or butter intake.
Statistical analysis
The analysis of the genetic main effects of the six SNPs
and CLL was performed using logistic regression models
adjusted for age (based on the tertile distribution of con-
trols: ≤60, 61–70, ≥71), sex and region (Asturias, Barce-
lona, Cantabria, Girona and Granada). To minimize type
I error inflation, log-additive models were used as genetic
models.
All data involving nutritional data from question-
naire were further adjusted for education (lower level: no
school, incomplete or complete primary school; higher
level: higher than primary school), family history of hae-
matological cancer (no/yes), body mass index (<25, 25–29,
≥30), history of diabetes (no/yes), total energy intake (con-
tinuous) [22] current alcohol intake (no, low, high; cut-off
based on median distribution of controls) and smoking sta-
tus (never, past, current). Fruit, vegetable and total vitamin
C intake were examined as continuous (1 unit/day increase
in vegetable or fruit intake; 100 g/day increase in vegeta-
ble or fruit intake; 100 mg/day increase in total vitamin C
intake) as well as categorical using the quartile distribution
of the controls. Total carbohydrate, sugar, complex car-
bohydrate and fibre intake were categorized based on the
quartile distribution of the controls. Missing data repre-
sented less than 5 % for each variable and were grouped in
a new category for all analyses. The comparison of means
of fruit, vegetable and vitamin C intake across risk factor
categories was analysed using one-way analysis of variance
(ANOVA). Sensitivity analyses restricted to complete case
analysis were performed. Multiplicative interactions were
assessed between fruit intake, vegetable intake, total vita-
min C intake, specific food items (using a continuous scale)
and the six polymorphisms (using co-dominant models), by
comparing models with and without interaction term using
log-likelihood ratio tests. While examining detailed food
items, multiple testing was taken into account by means of
false discovery rate (q-values).
Sensitivity analyses were performed to examine how
the inclusion of the cases with longer period of time from
diagnosis to recruitment (≤6 vs. >6 months), the inclusion
of treated patients for CLL and the impact of Rai stages
13
Eur J Nutr
affect the overall estimates of CLL for the genes and the
diet analysis. For the sensitivity analysis, odds ratios were
obtained using polytomous unconditional logistic regres-
sion and P-values for heterogeneity using case–case analy-
sis. The change in diet in the last 5 years between cases and
controls was examined using Chi-square tests for independ-
ence. All P-values were two-sided, and data analyses were
performed using STATA computer (version 10.1), the R
(version 3.2.2) and the Haploview packages [23].
Results
In Table 1, selected demographic and nutritional charac-
teristics of cases and controls were compared. Genetic
data were available for 434 cases and 1257 controls and
data from food frequency questionnaire for 345 cases
and 1602 controls. Of these participants, 275 cases and
1094 controls had both genetic and nutritional data. Over-
all, CLL cases were slightly older than controls (on aver-
age 66 vs. 63 years, respectively), more likely to report a
family history of haematological cancer, to consume more
fruit, whereas cases were less likely to consume alcohol
and to smoke compared to controls. CLL cases and con-
trols were similar in terms of distribution of sex, educa-
tion level, body mass index, diabetes, vegetable intake,
total vitamin C intake and total energy intake. Higher mean
intake of fruits was observed in cases compared to controls
(364.7 g/day, standard deviation SD 212.7 and 324.7 g/
day, SD 224.2 respectively; P  = 0.008), whereas no sta-
tistically significant differences were found for total veg-
etable intake [172.9 g/day (SD 100.1) vs. 173.2 g/day (SD
126.9); P = 0.98, respectively] and total vitamin C intake
[165.4 mg/day (SD 91.9) vs. 156.4 mg/day (SD 100),
P = 0.17, respectively].
Among controls, lower intake in vitamin C-rich food
was strongly associated with current smoking, males,
higher alcohol intake and lower total energy intake (Online
resource 3). Following full adjustment for basic variables
(age, sex, region and education), smoking was not associ-
ated with CLL (P-value for smoking categories ≥0.10,
data not shown). Consequently, smoking did not modify
strongly the association between CLL and fruit, vegeta-
ble and vitamin C intake (change in odds ratios for trend
<1 %). Also, when the interaction for smoking was evalu-
ated, the P-values for interaction between fruit, vegetable
and vitamin C intake and smoking in relation to CLL were
not statistically significant (P > 0.65). In particular, fruit
intake was less frequent in current smokers, in males, in
the youngest individuals, in participants with higher educa-
tion and those with higher alcohol intake. However, none
of these factors confounded or modified the association
between CLL and fruit intake. The association between
 Eur J Nutr

Table 1  Characteristics of cases and controls

Participants with genetic data Participants with nutritional data Participants with both genetic and nutritional data
Controls Cases P-valuea Controls Cases P-valuea Controls Cases P-valuea

Number 1257 434 1602 345 1094 275

Region of residence 560 (45) 307 (71) <0.001 862 (54) 217 (63) <0.001 502 (46) 164 (60) <0.001
Barcelona (%)
Mean age, year (SD) 63.3 (11.4) 66.3 (10.1) <0.001 63.5 (11.4) 66.2 (10.2)  <0.001 63.2 (11.4) 66.7 (9.8) <0.001
Sex, male number (%) 730 (58) 269 (62) 0.15 914 (57) 204 (59) 0.48 644 (59) 167 (61) 0.58
Education,  % higher 555 (44) 181 (42) 0.38 756 (47) 157 (46) 0.57 497 (45) 115 (42) 0.28
education
Smoking consumption, 253 (20) 52 (12) <0.001 275 (17) 42 (12) 0.02 207 (19) 32 (12) 0.004
current smoker number (%)
Family history of haemato- 54 (4) 52 (12) <0.001 80 (5) 44 (13) <0.001 49 (4) 32 (12) <0.001
logical cancer, number (%)
Body mass index (kg/m2), 27.1 (4.5) 27.3 (4.4) 0.43 27.0 (4.5) 27.4 (4.5) 0.16 27.1 (4.5) 27.6 (4.3) 0.06
mean (SD)
Diabetes, ever N (%) 205 (16) 74 (17) 0.72 249 (16) 55 (16) 0.87 176 (16) 50 (18) 0.41
Ethanol intake (g/day), – – – 11.8 (16.7) 9.5 (14.3) 0.02 11.9 (16.7) 9.6 (14.3) 0.04
mean (SD)
Fruit intake (g/day), – – – 328.5 (224.6) 358.6 (210.7) 0.02 324.7 (224.2) 364.7 (212.7) 0.008
mean (SD)
Fruit intake (unit/day), – – – 2.5 (1.6) 2.7 (1.6) 0.01 2.4 (1.6) 2.8 (1.6) 0.003
mean (SD)
Vegetable intake (g/day), – – – 178 (123.7) 176.5 (108.2) 0.83 173.2 (126.9) 172.9 (100.1) 0.98
mean (SD)
Vegetable intake (units/day), – – – 2.6 (1.8) 2.6 (1.7) 0.76 2.6 (1.9) 2.6 (1.7) 0.81
mean (SD)
Total vitamin C intake – – – 157.5 (99.3) 165.3 (94.9) 0.18 156.4 (100) 165.4 (91.9) 0.17
(mg/day), mean (SD)
Total energy intake – – – 1888.8 (649.6) 1939 (685.0) 0.20 1879.4 (630.1) 1914.1 (614.7) 0.41
(kca/day), mean (SD)

N number, SD standard deviation

a
 Unadjusted P-values

13
 Eur J Nutr

Table 2  Association between fruit/vegetables/total vitamin C intake and CLL using questionnaire data
Controls Cases
N (%) N (%) Basic adjustmenta Full adjustmentb
OR & 95 % CI OR & 95 % CI
1602 345

Fruit intake (units/day)

<1.2 370 (23 %) 55 (16 %) REF REF
1.2– 398 (25 %) 87 (25 %) 1.29 (0.89–1.87) 1.21 (0.83–1.77)
2.2– 424 (26 %) 93 (27 %) 1.32 (0.91–1.91) 1.22 (0.83–1.79)
>3.2 410 (26 %) 110 (32 %) 1.67 (1.16–2.40) 1.50 (1.01–2.23)
P-trend 0.008 0.05
1 unit increase/day 1.08 (1.00–1.16) 1.06 (0.98–1.14)
P-trend (g/day) 0.04 0.15
<162 370 (23 %) 57 (17 %) REF REF
162– 415 (26 %) 83 (24 %) 1.17 (0.81–1.70) 1.10 (0.75–1.61)
296.8– 409 (26 %) 95 (28 %) 1.33 (0.92–1.92) 1.26 (0.86–1.84)
>431.1 408 (25 %) 110 (32 %) 1.63 (1.13–2.34) 1.48 (1.00–2.18)
P-trend 0.005 0.03
100 g increase/day 1.05 (1.00–1.10) 1.04 (0.98–1.10)
P-trend 0.06 0.23
Vegetable intake (units/day)
<1.4 385 (24 %) 84 (24 %) REF REF
1.4– 362 (23 %) 87 (25 %) 1.06 (.76–1.49) 1.00 (0.71–1.42)
2.2– 399 (25 %) 80 (23 %) 0.88 (0.62–1.24) 0.82 (0.57–1.17)
>3.3 456 (28 %) 94 (27 %) 0.88 (0.63–1.23) 0.77 (0.54–1.09)
P-trend 0.28 0.08
1 unit increase/day 0.98 (0.91–1.05) 0.95 (0.88–1.02)
P-trend (g/day) 0.50 0.13
<94.2 376 (23 %) 69 (20 %) REF REF
94.2– 377 (24 %) 96 (28 %) 1.24 (0.88–1.76) 1.23 (0.86–1.75)
149.3– 399 (25 %) 85 (25 %) 0.98 (0.68–1.40) 0.95 (0.66–1.37)
>218.7 450 (28 %) 95 (28 %) 0.94 (0.66–1.33) 0.83 (0.66–1.21)
P-trend 0.36 0.13
100-g increase/day 0.94 (0.85–1.04) 0.90 (0.80–1.01)
P-trend 0.24 0.06
Total vitamin C (mg/day)
<86.1 362 (23 %) 62 (18 %) REF REF
86.1– 435 (27 %) 95 (28 %) 1.14 (0.80–1.63) 1.09 (0.75–1.57)
138.4– 389 (24 %) 85 (25 %) 1.13 (0.78–1.63) 1.06 (0.73–1.56)
>199 416 (26 %) 103 (30 %) 1.35 (0.94–1.92) 1.23 (0.83–1.82)
P-trend 0.12 0.33
100-mg increase/day 1.06 (0.95–1.19) 1.02 (0.90–1.17)
P-trend 0.30 0.72
a
  Age, sex, region and education
b
  Age, sex, region, education, smoking, diabetes, current alcohol intake, total energy intake, body mass index and family history of haematologi-
cal cancer

fruit intake and CLL remained statistically significant No statistically significant association was obtained
(highest versus lowest quartile in g/day intake: 1.48; 95 % with vegetable intake nor with total vitamin C intake.
CI 1.00–2.18; P-trend = 0.03) after adjusting for multiple Further adjustment by any of the six genetic variants
potential confounding factors (Table 2). did not materially change the association between fruit,

13
Eur J Nutr

Table 3  Association between Polymorphism Controls Cases ORa & 95 % CI P-value

six vitamin C transporter gene
(SLC23A2) polymorphisms and N = 1257 N = 434
CLL using log-additive models
rs1715364
AA 381 (30 %) 117 (27 %)
AG 627 (50 %) 221 (51 %) per G allele
GG 249 (20 %) 96 (22 %) 1.11 (0.94–1.30) 0.22
rs6133175
AA 581 (46 %) 172 (40 %)
AG 542 (43 %) 207 (48 %) per G allele
GG 134 (11 %) 55 (13 %) 1.19 (1.00–1.41) 0.05
rs1776948
TT 379 (30 %) 108 (25 %)
AT 632 (50 %) 229 (53 %) per A allele
AA 246 (20 %) 96 (22 %) 1.20 (1.01–1.41) 0.04
rs6139587
AA 427 (34 %) 148 (34 %)
AT 630 (50 %) 209 (48 %) per T allele
TT 200 (16 %) 76 (18 %) 1.00 (0.85–1.18) 1.00
rs369270
AA 464 (37 %) 148 (34 %)
AG 600 (48 %) 220 (51 %) per G allele
GG 193 (15 %) 66 (15 %) 1.02 (0.86–1.20) 0.82
rs6052937
CC 839 (67 %) 309 (71 %)
AC 384 (31 %) 112 (26 %) per A allele
AA 34 (3 %) 13 (3 %) 0.84 (0.67–1.04) 0.11
a
  Adjusted for age, sex and region

Table 4  Association between Bloc 1 Frequency Frequency Frequency Adjusted ORa & P-value
six selected vitamin C (rs6133175, (%) controls (%) cases (%) 95 % CI
transporter gene (SLC23A2) rs1715364)
polymorphisms and CLL—Pair-
haplotype analysis Haplotype analysis
Hap1 (A–A) 55 55 52 REF
Hap2 (A–G) 12 13 11 0.94 (0.72–1.22) 0.62
Hap3 (G–G) 33 32 37 1.17 (0.98–1.40) 0.08
Pair-haplotype analysis
Hap1/Hap1 29 30 27 REF
Hap2/Hap2 1 1 1 NE –
Hap3/Hap3 11 11 13 1.31 (0.88–1.95) 0.18
Hap1/Hap2 14 15 12 0.96 (0.65–1.42) 0.83
Hap1/Hap3 36 35 39 1.27 (0.95–1.69) 0.10
Hap2/Hap3 8 8 8 1.14 (0.73–1.80) 0.56

NE not estimated due to low frequency

a
  Adjusted for age, sex and region

13

vegetable and total vitamin C intake and CLL (data not
shown).
Multivariate logistic regression analyses examin-
ing the association between the six SNPs of SLC23A2
and CLL (Table 3) showed statistically significant
associations for two SNPs using log-additive mod-
els [rs6133175_A > G: OR (95 % CI): 1.19 (1.00–1.41),
P  = 0.05; rs1776948_T > A: OR (95 % CI): 1.20 (1.01–
1.41); P = 0.04]. A non-statistically significant association
was found for the SNP rs1715364, in linkage disequilib-
rium with rs6133175 (Online resource 2) (OR 1.11, 95 %
CI 0.94–1.30, P = 0.22). Supporting the genotype results,
associations with CLL were also found for the two-SNP
GG haplotype (rs6133175 and rs1715364) in both haplo-
type and haplotype pair analyses; however, these results did
not reach statistical significance (Table 4). No main effect
was found for the SNPs rs6139587 and rs369270. Further
adjustment for fruit, vegetable and total vitamin C intake
did not materially change the association between the
genetic variants and CLL nor did the restricted analysis to
those individuals with both diet and genetic data (data not
shown).
Overall, the association between self-reported veg-
etable, fruit and total vitamin C intake and CLL was not
modified by genetic variation in the six examined SNPs in
SLC23A2. By specific vegetable and fruit items, no inter-
actions between specific food items and genetic suscepti-
bility in relation to CLL remained statistically significant
after correction for multiple testing (q-values >0.20; data
not shown).
In addition, sensitivity analyses were performed (Online
resource 4) stratifying cases with time from diagnosis
to recruitment (≤6 vs. >6 months), by Rai stages (Rai
0 vs. Rai I–IV) and by use of CLL treatment (ever being
treated for CLL versus no treatment) using log-additive
models. No effect modifications in relation to the asso-
ciation between the genetic main effects of rs1715364,
rs6133175, rs1776948, rs369270 and rs6052937 and CLL
were observed. In contrast, the association between the
genetic variant rs6139587 and CLL was slightly modified
by treatment (P-heterogeneity  = 0.03) and by Rai stages
(P-heterogeneity  = 0.08), suggesting a positive associa-
tion between CLL and rs6139587 with the presence of the
rare allele in patients with more aggressive CLL. Further
sensitivity analyses were performed to examine the impact
of these three factors (disease stages, prevalence time
and treatment) on the association between CLL and the
diet variables, and no interactions were detected (Online
resource 5). Three-way interactions on sex, genetic varia-
tion and vitamin C-rich food items were performed, but no
interactions were statistically significant at the 5 % level.
Since diet can be influenced by many external factors
associated with survival period, we examined whether
13
Eur J Nutr
cases and controls have substantially modified their diet
in the last 5 years before. No statistical significant differ-
ences in changes of fruit and vegetable intake were found
between controls and cases (Online resource 6). Since
patients with longer survival as well as newly diagnosed
patients might have modified their dietary habits, changes
in fruit and vegetable intake were further examined by
stratifying the cases by prevalence time (time from diag-
nosis to interview) such as <6 months, 6 months–4.9 years
and 5 years or more. No statistical differences were found
between the different groups (Online resource 6). Further
analysis was performed to examine the association between
total carbohydrate intake, total sugar intake, total complex
carbohydrate intake and total fibre intake and CLL. A bor-
derline statistical significant (P-trend  = 0.08) was found
for total sugar intake, and no associations were observed
for the other carbohydrate groups (Online resource 7).
Discussion
This study reports on the independent and combined effect
of six genetic variants in the Na+-dependent ascorbic acid
transporters 2 SVCT2 (encoded by the SLC23A2 gene) and
fruit/vegetable/total vitamin C intake in relation to CLL.
Overall, higher fruit consumption was positively associ-
ated with CLL, whereas total vitamin C and vegetable
intake showed no consistent association. Moreover, CLL
was associated with two SNPs located in SLC23A2, but the
relationship between CLL and dietary intake of fruit/vege-
table, primary source of vitamin C, as well as total vitamin
C intake was not found to be modified by genetic variants
in vitamin C transporter.
In a previous study based on a large case–control
study in the San Francisco Bay Area examining candidate
genes, two out of our six genetic variants (rs6133175 and
rs1776948) in SLC23A2 were also associated with CLL
[15]. Rs1715364 was also found to be related to CLL in
this previous report [15], but this result was confined to
male individuals in our study (data not shown). The weak
association observed with the SNP rs6052937 located at
3′ UTR SLC23A2 will need further investigation. Unfor-
tunately, no diet information was reported in this earlier
study and, to our knowledge, no other studies have been
published in relation to this gene and CLL. Previous epi-
demiological evidence for an association between CLL and
fruit/vegetable/total vitamin C intake has shown inconsist-
ent results [4–10]. Differences in genetic backgrounds of
studied populations and fruit/vegetable consumption pat-
terns across countries could dilute the strength of the pos-
sible association. In turn, CLL might be related to multi-
ple mutated genes interacting with dietary intake. Despite
having poor vitamin C contents, other nutrients enclosed
Eur J Nutr
in diet could work in synergy with vitamin C transport
and absorption. However, following multiple testing crite-
ria there was no evidence of interactions between genetic
susceptibility, specific fruit and vegetable items and CLL
in our study.
The overall difference in mean fruit consumption
between cases and controls was statistically significant
but represents a small nutritional variation (+11 % g/day
increase in cases compared to controls). However, the quar-
tile distributions based on the controls might reflect better
the variation in fruit consumption from very low (<1.2 unit
per day) to high consumption (>3.2 units per day) between
cases and controls. It is well established that smokers con-
sume less vitamin C-rich food, such as fruits, and that levels
of circulating ascorbic acid are dropped in smokers inde-
pendently of vitamin C intake [24]. A modest protective
effect of smoking has been reported in the INTERLYMPH
consortium [25]; however, in our analysis, smoking did not
confound or modify the association between fruit, vegeta-
ble or vitamin C intake and CLL. The strongest associa-
tion observed for the highest fruit quartile might be con-
founded by other unmeasured factors such as pesticide in
food. In a recent report from the European Safety Author-
ity, the highest maximum-residue-level exceedance rate as
well as multiple residue levels were observed for several
fruits among twelve food items tested [26]. Pesticide expo-
sure has been associated with CLL in occupational studies
[27], but the evaluation of lifetime exposure to pesticide
through food consumption is a challenging task. The two
main mechanisms related to the transport of ascorbic acid
and its oxidized form (dehydroascorbic acid, DHAA) from
dietary intake to cells uptake depend on proteins from the
human SLC23 and SLC2 families, consisting mainly of
sodium-dependent ascorbate co-transporters (SVCTs) and
of DHAA hexose transporters (GLUT), respectively [28].
The SVCT1 (encoded by the SLC23A1 gene) is thought to
be associated with vitamin C transport across membranes
(mainly epithelial cells in the intestine, kidney and liver),
whereas SVCT2 (encoded by SLC23A2) is thought to be
responsible for the absorption of circulating ascorbic acid
and for the regulation of the accumulation of vitamin C
in tissue, in particular brain, eyes and adrenals [28, 29].
SVCT2 but not SVCT1 is expressed in almost all tissues
except erythrocytes [30]. If the presence of rs6133175
homozygous GG is associated with higher vitamin C con-
centrations in plasma, as previously reported [13], as well
as a higher likelihood of CLL, it could be speculated that
individuals with this genetic variant despite accumulating
high circulating levels of ascorbic acid might not express
fully functioning proteins for SVCT2 for rapid cell absorp-
tion [29]. Thus, individuals carrying these genetic vari-
ants might have a decrease in vitamin C cellular uptake
even with a high dietary vitamin C intake and might have
therefore an increase in urine vitamin C excretion from the
body. However, the functional role of this polymorphism is
not known, and since this SNP as well as rs1776948 and
1715364 lies in the intronic regions of the gene SLC23A2,
it makes the interpretation of the results difficult. From
HaploReg 3, rs6052937 is located at 3′ UTR SLC23A2 and
rs369270 is in LD with the SNP rs261360 that has been
associated with expression quantitative trait loci (eQTL)
on the liver gene expression traits [31]. The independent
association of total fruit intake with CLL would suggest
that nutrients contained in fruit other than vitamin C might
be involved in CLL aetiology. In particular, sugar (espe-
cially glucose) has been put forward as an important fac-
tor in CLL aetiology [32]. In this report, total sugar intake
(including monosaccharides and disaccharides) was weakly
associated with CLL.
The inclusion of prevalent cases might be a cause for
concerns since the individuals who survived might have a
very different aetiology than those who died rapidly after
diagnosis. The risk factors identified in a study including
prevalent cases might simply reflect factors associated with
survival and not causal factors associated with the disease
of interest. In particular, diet can be influenced by many
external factors and patients with longer survival period
might have substantially modified their diet. However, no
statistical differences in changes of fruit and vegetable
intake in the last 5 years between cases and controls were
observed. Furthermore, the association between CLL and
the three diet variables were not found to be modified by
factors related to time from diagnosis to recruitment, dis-
ease stages or use of treatment.
One of the main limitations is the study design since
case–control studies are prone to selection and recall biases.
Measurement errors in the estimation of fruit and vegetable
intake due to the FFQ self-report use are also of some con-
cern. However, the FFQ was validated in Spanish popula-
tion and included regional products. Vitamin C evaluation
from food frequency questionnaire has been assessed with
confidence in other studies [33]. Nevertheless, it should be
kept in mind that self-reported vitamin C intake at recruit-
ment might not reflect the current or usual lifetime plasma
or body status of vitamin C, and hence it might explain the
null results observed in the interaction analysis. Unfortu-
nately, data on the use of vitamin C supplement were not
available. The results of the other sensitivity analysis sug-
gested that the use of prevalent cases might have introduced
selective survival bias for the polymorphism rs6139587,
and this finding needs further research. The assessment of
many potential confounding variables gives strength to the
current study.
In summary, this study provides further evidence for a
role of both genetic variants (rs6133175 and rs1776948) in
vitamin C transporter gene and total fruit intake in CLL.
13

Our results suggest that both environmental and genetic
factors affect CLL independently. The association between
SLC23A2 genetic variants and CLL should be validated in
large prospective studies with nutritional information along
with genetic variations and gene expression of SLC23A1
and other genes such as SVCT, GLUT, SOD, NOS or glu-
tathione S-transferases as well as tissue and circulating
(blood and urine) levels of ascorbic acid [34]. No gene-diet
interactions in CLL remained statistically significant after
correction for multiple testing, and further investigations
will have to be undertaken in large database.
Acknowledgments  The authors would like to thank all the subjects
for their contribution to the study and the MCC-Spain-CLL collabo-
rators from: Catalan Institute of Oncologia, L’Hospitalet de Llobre-
gat, Barcelona, Spain (Teresa Alonso, Vanesa Camon, Eva Domingo-
Domenech, Dolores Dot, Anna Esteban, Elisabeth Guinó, Yolanda
Florencia, Joellen Klaustermeier, Santi Mercadal, Ana Oliveira, Isabel
Padrol, Paloma Quesada, Victor Moreno, Josep Sarra, Yasmin Sabaté,
Marleny Vergara); Centre for Research in Environmental Epidemi-
olog (CREAL), Barcelona, Spain (Mireia García, Cecília Persavento);
Hospital Clínic, Barcelona, Spain (Cristina Capdevila Lozar, Ainara
Expósito, Silvia Martin Román, Amparo Muñoz, Yolanda Torralba);
Hospital del Mar, Barcelona, Spain (Eugènia Abella, Estela Car-
rasco, Judith Cirac, Francesc Garcia, Antonio Salar); Hospital Bell-
vitge, L’Hospitalet de Llobregat, Barcelona, Spain (Esmeralda de
la Banda); Institut de Prestacions d’Assistència Mèdica al Personal
Municipal (PAMEM) & Centros de Atención Primaria (CAP), Bar-
celona, Spain (Jesús Almeda, Marifé Alvarez Rodriguez, Alex Bassa
Massanas, Albert Boada Valmaseda, Enric Duran, Olga Gonzalez
Ferrer, Clara Izard, Manoli Liceran, Carmen López, Josep Manuel
Benítez, Dolors Petitbó, Angelina Potrony, Sònia Sarret, Laura
Sebastián, Josep M. Vilaseca); Cantabria group (Inés Gómez-Acebo,
Pilar González Echezarreta, Maria del Mar González Martínez, Javier
Llorca, Luis Mariano López López, Almudena de la Pedraja Pavón,
Paula Picón Sedano); Instituto Universitario de Oncología, Universi-
dad de Oviedo, Asturias, Spain (Cristina Arias, Guillermo Fernández,
María José Fernández González, Ana Fernández Somoano, Carlos
López-Otín, Ana Souto); Servicio de Salud del Principado de Astu-
rias, SESPA, Asturias, Spain (Enrique Colado, Begoña Martínez-
Argüelles, Manuel Rivas del Fresno, Marta María Rodríguez-Suárez);
Departemento de Medicina Preventiva y Salud Pública, Instituto de
Investigación Biosanitaria de Granada, Hospitales Universitarios de
Granada/Universidad de Granada (Jose Juan Jiménez-Moleón, Aurora
Bueno Cavanillas, Obdulia Moreno Abril); Ciber de Epidemiología
y Salud Pública (CIBEResp) (Rocío Olmedo Requena); Servicio
de Hematología, Hospital Universitario Virgen de las Nieves, José
Luis Gastón Morata Centro de Salud Zaidín Sur, Servicio Anda-
luz de Salud (Paloma García-Martín, Eva Garrido Morales); Unitat
d’Epidemiologia i Registre de Cancer de Girona (Maria Buxo, M
Carme Carmona-Garcia, Angel Izquierdo, Rafael Marcos-Gragera,
Patricia Martí Bargalló, Gemma Osca-Gelis, Montse Puig-Vives,
Rocio Rodriguez Romanos, Esther Rodriguez Sanchez, Marc Saez,
Carlota Torner Galindo, Loreto Vilardell); Hospital Universitari de
Girona Dr. Josep Trueta (Elena Alvarez Castaño, Ignacio Blanco,
Rosa Coll, David Gallardo, Josep Maria Roncero, Luis Miguel
Alonso Ruano); Hospital Santa Caterina (Rocio Jurado Perez, Isaura
Marcé Pujol, Joan Melendez Rusiñol); CAP de Santa Clara (Conxa
Bou); CAP de Angles (Gabriel Coll de Tuero, Alba Coll Negre).
Financial support  This work was supported by the “Acción Trans-
versal del Cancer”, approved on the Spanish Ministry Council on
the 11th October 2007, by the Instituto de Salud Carlos III (Spanish
13
Eur J Nutr
Government) (grants PI08/1770, PI08/0533, PI08/1359, PS09/00773-
Cantabria, PI11/01810, PI11/02213, RCESP C03/09, RTICESP
C03/10, RTIC RD06/0020/0095, RTIC RD12/0036/0056, Rio Hort-
ega CM13/00232, SV-09-CLINIC-1 and CIBERESP), by the Fun-
dación Marques de Valdecilla (API 10/09), by Obra Social CAJAS-
TUR (SV-CAJASTUR-1), by the Recercaixa (2010ACUP 00310), by
the Spanish Association Against Cancer (AECC) Scientific Founda-
tion and by the Agència de Gestió d’Ajuts Universitaris i de Recerca
(AGAUR)–Generalitat de Catalunya (Catalonian Government)
(grants AGAUR 2009SGR1026, 2014SGR756 and 2009SGR1465).
The ICGC CLL-Genome Project is funded by Spanish Ministerio
de Economía y Competitividad (MINECO) through the Instituto
de Salud Carlos III (ISCIII) and Red Temática de Investigación del
Cáncer (RTIC RD12/0036/0036) del ISCIII. Sample collection and
storage was partially supported by the Instituto de Salud Carlos III
FEDER (RD09/0076/00036), Xarxa de Bancs de Tumors de Catalu-
nya sponsored by Pla Director d’Oncologia de Catalunya (XBTC).
Compliance with ethical standards 
Conflict of interest  All authors have no conflict of interest.
Ethical considerations  Ethical approval was granted for each par-
ticipating centre of the study. All procedures followed were in accord-
ance with the ethical standards of the responsible committee on human
experimentation (institutional and national) and with the Helsinki Dec-
laration of 1975, as revised in 2000. Informed written consent was
obtained from all patients for being included in the study. None of
the sources of funding had any influence over the design, execution,
analyses, interpretation or reporting of data of the study.
References
1. Bürzle M, Hediger MA (2012) Functional and physiological
role of vitamin C transporters. Curr Top Membr 70:357–375.
doi:10.1016/B978-0-12-394316-3.00011-9
2. Mendelsohn AR, Larrick JW (2014) Paradoxical effects of anti-
oxidants on cancer. Rejuvenation Res 17:306–311. doi:10.1089/
rej.2014.1577
3. Norat T, Aune D, Chan D, Romaguera D (2014) Fruits and veg-
etables: updating the epidemiologic evidence for the WCRF/
AICR lifestyle recommendations for cancer prevention. Cancer
Treat Res 159:35–50. doi:10.1007/978-3-642-38007-5_3
4. Chang ET, Smedby KE, Zhang SM et al (2005) Dietary factors
and risk of non-hodgkin lymphoma in men and women. Cancer
Epidemiol Biomark Prev 14:512–520. doi:10.1158/1055-9965.
EPI-04-0451
5. Rohrmann S, Becker N, Linseisen J et al (2007) Fruit and veg-
etable consumption and lymphoma risk in the European Pro-
spective Investigation into Cancer and nutrition (EPIC). Cancer
Causes Control 18:537–549. doi:10.1007/s10552-007-0125-z
6. Ross JA, Kasum CM, Davies SM et al (2002) Diet and risk of
leukemia in the Iowa women’s health study. Cancer Epidemiol
Biomark Prev 11:777–781
7. Tsai H-T, Cross AJ, Graubard BI et al (2010) Dietary factors and
risk of chronic lymphocytic leukemia and small lymphocytic
lymphoma: a pooled analysis of two prospective studies. Can-
cer Epidemiol Biomark Prev 19:2680–2684. doi:10.1158/1055-
9965.EPI-10-0585
8. Chiu BC-H, Kwon S, Evens AM et al (2011) Dietary intake of
fruit and vegetables and risk of non-Hodgkin lymphoma. Cancer
Causes Control 22:1183–1195. doi:10.1007/s10552-011-9797-5
9. Purdue MP, Bassani DG, Klar NS et al (2004) Dietary fac-
tors and risk of non-Hodgkin lymphoma by histologic
Eur J Nutr
subtype: a case-control analysis. Cancer Epidemiol Biomark
Prev 13:1665–1676
10. Holtan SG, O’Connor HM, Fredericksen ZS et al (2012) Food-
frequency questionnaire-based estimates of total antioxidant
capacity and risk of non-Hodgkin lymphoma. Int J Cancer
131:1158–1168. doi:10.1002/ijc.26491
11. Erichsen HC, Peters U, Eck P et al (2008) Genetic variation
in sodium-dependent vitamin C transporters SLC23A1 and
SLC23A2 and risk of advanced colorectal adenoma. Nutr Cancer
60:652–659. doi:10.1080/01635580802033110
12. Wright ME, Andreotti G, Lissowska J et al (2009) Genetic
variation in sodium-dependent ascorbic acid transporters and
risk of gastric cancer in Poland. Eur J Cancer 45:1824–1830.
doi:10.1016/j.ejca.2009.01.027
13. Duell EJ, Lujan-Barroso L, Llivina C et al (2013) Vitamin C
transporter gene (SLC23A1 and SLC23A2) polymorphisms,
plasma vitamin C levels, and gastric cancer risk in the EPIC
cohort. Genes Nutr 8:549–560. doi:10.1007/s12263-013-0346-6
14. Chen AA, Marsit CJ, Christensen BC et al (2009) Genetic vari-
ation in the vitamin C transporter, SLC23A2, modifies the risk
of HPV16-associated head and neck cancer. Carcinogenesis
30:977–981. doi:10.1093/carcin/bgp076
15. Skibola CF, Bracci PM, Halperin E et al (2008) Polymorphisms
in the estrogen receptor 1 and vitamin C and matrix metallopro-
teinase gene families are associated with susceptibility to lym-
phoma. PLoS One 3:e2816. doi:10.1371/journal.pone.0002816
16. Castaño-Vinyals G, Aragonés N, Pérez-Gómez B et al (2015)
Population-based multicase-control study in common tumors
in Spain (MCC-Spain): rationale and study design. Gac Sanit.
doi:10.1016/j.gaceta.2014.12.003
17. Hallek M, Cheson BD, Catovsky D et al (2008) Guidelines for
the diagnosis and treatment of chronic lymphocytic leukemia:
a report from the International Workshop on chronic lympho-
cytic leukemia updating the National Cancer Institute—Work-
ing Group 1996 guidelines. Blood 111:5446–5456. doi:10.1182/
blood-2007-06-093906
18. Swerdlow SH, Campo E, Harris NL et al WHO classification of
tumours of haematopoietic and lymphoid tissues, Vol 2, 4th edn.
IARC Scientific Publications, WHO Classification of Tumours.
Lyon, France
19. Grove ML, Yu B, Cochran BJ et al (2013) Best practices and
joint calling of the HumanExome BeadChip: the CHARGE Con-
sortium. PLoS One 8:e68095. doi:10.1371/journal.pone.0068095
20. Purcell S, Neale B, Todd-Brown K et al (2007) PLINK: a tool
set for whole-genome association and population-based linkage
analyses. Am J Hum Genet 81:559–575. doi:10.1086/519795
21. García-Closas R, García-Closas M, Kogevinas M et al (2007)
Food, nutrient and heterocyclic amine intake and the risk of
bladder cancer. Eur J Cancer 43:1731–1740. doi:10.1016/j.
ejca.2007.05.007
22. Willett WC, Howe GR, Kushi LH (1997) Adjustment for
total energy intake in epidemiologic studies. Am J Clin Nutr
65:1220S–1228S discussion 1229S–1231S
23. Barrett JC, Fry B, Maller J, Daly MJ (2005) Haploview: analy-
sis and visualization of LD and haplotype maps. Bioinformatics
21:263–265. doi:10.1093/bioinformatics/bth457
24. Schectman G, Byrd JC, Gruchow HW (1989) The influence
of smoking on vitamin C status in adults. Am J Public Health
79:158–162
25. Slager SL, Benavente Y, Blair A et al (2014) Medical history,
lifestyle, family history, and occupational risk factors for chronic
lymphocytic leukemia/small lymphocytic lymphoma: the Int-
erLymph Non-Hodgkin Lymphoma Subtypes Project. J Natl
Cancer Inst Monogr 2014:41–51. doi:10.1093/jncimonographs/
lgu001
26. The European Food Safety Authority (2015) The 2013 European
Union report on pesticide residues in food
27. Cocco P, Satta G, Dubois S et al (2013) Lymphoma risk
and occupational exposure to pesticides: results of the Epi-
lymph study. Occup Environ Med 70:91–98. doi:10.1136/
oemed-2012-100845
28. Wilson JX (2005) Regulation of vitamin C transport. Annu Rev
Nutr 25:105–125. doi:10.1146/annurev.nutr.25.050304.092647
29. Michels AJ, Hagen TM, Frei B (2013) Human genetic variation
influences vitamin C homeostasis by altering vitamin C trans-
port and antioxidant enzyme function. Annu Rev Nutr 33:45–70.
doi:10.1146/annurev-nutr-071812-161246
30. May JM, Qu Z, Qiao H, Koury MJ (2007) Maturational loss of
the vitamin C transporter in erythrocytes. Biochem Biophys Res
Commun 360:295–298. doi:10.1016/j.bbrc.2007.06.072
31. Schadt EE, Molony C, Chudin E et al (2008) Mapping the
genetic architecture of gene expression in human liver. PLoS
Biol 6:e107. doi:10.1371/journal.pbio.0060107
32. Jitschin R, Braun M, Qorraj M et al (2015) Stromal cell-medi-
ated glycolytic switch in CLL-cells involves Notch-c-Myc sign-
aling. Blood. doi:10.1182/blood-2014-10-607036
33. Henríquez-Sánchez P, Sánchez-Villegas A, Doreste-Alonso J
et al (2009) Dietary assessment methods for micronutrient intake:
a systematic review on vitamins. Br J Nutr 102(Suppl):S10–S37.
doi:10.1017/S0007114509993126
34. Michels AJ, Frei B (2013) Myths, artifacts, and fatal flaws:
identifying limitations and opportunities in vitamin C research.
Nutrients 5:5161–5192. doi:10.3390/nu5125161
13