biomarkers of one-carbon metabolism in colorectal cancer risk1165289/... · 2017-12-13 · donors....
TRANSCRIPT
Biomarkers of One-carbon Metabolism
in Colorectal Cancer Risk
Björn Gylling
Department of Medical Biosciences
Department of Radiation Sciences
Umeå 2017
This work is protected by the Swedish Copyright Legislation (Act 1960:729)
Dissertation for PhD
ISBN: 978-91-7601-804-0
ISSN: 0346-6612
Umeå University Medical Dissertations, New Series No: 1935
Cover design by Simon Jönsson at Inhousebyrån, Umeå University
Electronic version available at: http://umu.diva-portal.org/
Printed by: UmU Print Service, Umeå University
Umeå, Sweden 2017
Tillägnad:
Alla de människor i Västerbotten som valt att delta i Northern Sweden
Health and Disease Study. Tack för er tid och ert engagemang.
i
Table of Contents:
Abstract iii Original Papers v Abbreviations vi Sammanfattning på svenska vii Background 1
Colorectal Cancer 1 Incidence and Mortality 1 Staging and Prognosis 2 Colorectal Tumorigenesis and Heterogeneity 3 Treatment 4 Screening 5 Risk Factors and Preventive Factors 6 One-Carbon Metabolism 8 Enzymatic Reactions and Metabolites 8 One-carbon Metabolism in Cancer and Genomic Instability 10 Folate Cycle Components and Folic Acid in CRC 11 Methionine and Choline Oxidation Pathway Metabolites in CRC 13 Transsulfuration Pathway Metabolites in CRC 15 B-vitamins Involved in One-carbon Metabolism and CRC Risk 16 Inflammatory Interaction with Vitamin B6 and CRC Risk 18 Summary and Overall Perspectives 20
Aims of the Thesis 22 Paper I 22 Paper II 22 Paper III 22 Paper IV 22
Materials and Methods 23 Study Population 23 Study Cohort 23 The Västerbotten Intervention Programme 23 The Mammography Screening Project 24 Blood Sampling and Analysis 24 Variables 25 Study Design 25 Selection of CRC Case and Control Participants 26 Statistical Analyses 27 Ethical Approval 29
Main Results and Discussion 30 Paper I 30
ii
Main Results 30 Interpretation 31 Paper II 35 Main Results 35 Interpretation 36 Paper III 39 Main Results 39 Interpretation 40 Paper IV 42 Main Results 42 Interpretation 42
Conclusions 44 Paper I 44 Paper II 44 Paper III 44 Paper IV 44
Future Considerations 45 Acknowledgments 46 References 48
iii
Abstract
One-carbon metabolism, a network of enzymatic reactions involving the
transfer of methyl groups, depends on B-vitamins as cofactors, folate as a
methyl group carrier, and amino acids, betaine, and choline as methyl group
donors. One-carbon metabolism influences many processes in cancer
initiation and development such as DNA synthesis, genome stability, and
histone and epigenetic methylation. To study markers of one-carbon
metabolism and inflammation in relation to colorectal cancer (CRC) risk, we
used prediagnostic plasma samples from over 600 case participants and 1200
matched control participants in the population-based Northern Sweden
Health and Disease Study cohort.
This thesis studies CRC risk with respect to the following metabolites
measured in pre-diagnostic plasma samples: 1) folate, vitamin B12, and
homocysteine; 2) components of one-carbon metabolism (choline, betaine,
dimethylglycine, sarcosine, and methionine); and 3) three markers of
different aspects of vitamin B6 status. In addition, this thesis examines three
homocysteine ratios as determinants of total B-vitamin status and their
relation to CRC risk.
In two previous studies, we observed an association between low plasma
concentrations of folate and a lower CRC risk, but we found no significant
association between plasma concentrations of homocysteine and vitamin B12
with CRC risk. We have replicated these results in a study with a larger sample
size and found that low folate can inhibit the growth of established pre-
cancerous lesions.
Using the full study cohort of over 1800 participants, we found inverse
associations between plasma concentrations of the methionine cycle
metabolites betaine and methionine and CRC risk. This risk was especially low
for participants with the combination of low folate and high methionine
versus the combination of low folate and low methionine. Well-functioning
methionine cycle lowers risk, while impaired DNA synthesis partly explains
the previous results for folate.
We used the full study cohort to study associations between CRC risk and the
most common marker of vitamin B6 status, pyridoxal' 5-phosphate (PLP), and
two metabolite ratios, PAr (4-pyridoxic acid/(PLP + pyridoxal)) estimating
vitamin B6 related inflammatory processes and the functional vitamin B6
marker 3-hydroxykynurenine to xanthurenic acid (HK:XA). Increased
iv
vitamin B6-related inflammation and vitamin B6 deficiency increase CRC
risk. Inflammation was not observed to initiate tumorigenesis.
Total B-vitamin status can be estimated by three different recently introduced
homocysteine ratios. We used the full study cohort to relate the ratios as
determinants of the total B-vitamin score in case and control participants and
estimated the CRC risk for each marker. Sufficient B-vitamin status as
estimated with homocysteine ratios was associated with a lower CRC risk.
These studies provide a deeper biochemical knowledge of the complexities
inherent in the relationship between one-carbon metabolism and colorectal
tumorigenesis.
Keywords
Colorectal cancer, one-carbon metabolism, biomarkers, folate, epidemiology.
v
Original Papers
This thesis is based on the following papers:
I. Gylling B, Van Guelpen B, Schneede J, Hultdin J, Ueland PM,
Hallmans G, Johansson I, Palmqvist R. Low Folate Levels Are
Associated with Reduced Risk of Colorectal Cancer in a
Population with Low Folate Status. Cancer Epidemiol
Biomarkers Prev, 2014; 23(10):2136-44.
II. Myte R, Gylling R, Schneede J, Ueland PM, Häggström J, Hultdin
J, Hallmans G, Johansson I, Palmqvist R, and Van Guelpen B.
Components of One-carbon Metabolism Other than Folate and
Colorectal Cancer Risk. Epidemiology, 2016; 27(6):787-96.
III. Gylling B, Myte R, Schneede J, Hallmans G, Häggström J,
Johansson I, Ulvik A, Ueland PM, Van Guelpen B, and Palmqvist
R. Vitamin B-6 and colorectal cancer risk: a prospective
population-based study using 3 distinct plasma markers of
vitamin B-6 status. Am J Clin Nutr, 2017; 105(4):897-904.
IV. Gylling B, Myte R, Ulvik A, Ueland PM, Midttun Ø, Schneede J,
Hallmans G, Häggström J, Johansson I, Van Guelpen B, and
Palmqvist R. One-carbon metabolite ratios as functional B-
vitamin markers and in relation to colorectal cancer risk.
Submitted.
The original articles were reprinted with the permission from the publishers.
vi
Abbreviations
BHMT, betaine-homocysteine methyltransferase; CIN, chromosomal
instability pathway; CUP, The Continuous Update Project; CIMP, CpG island
methylator phenotype; CRC, colorectal cancer; CRP, C-reactive protein; CBS,
cystathionine β-synthase, CSE, cystathionine γ-lyase, dTMP, deoxythymidine
monophosphate; dUMP, deoxyuridine monophosphate; DHFR, dihydrofolate
reductase; DMG, dimethylglycine; EPIC, European Prospective studies into
Cancer and Nutrition; FAP, familial adenomatous polyposis; 5-FU,
fluorouracil; f-THF, 10-formyltetrahydrofolate; HNPCC, Hereditary non-
polyposis colorectal cancer; HAA, 3-hydroxyanthranilic acid; HK, 3-
hydroxykynurenine; KYNU, kynureninase; KAT, kynurenine
aminotransferase; MSP, Mammography Screening Project; me-THF,
methylene-THF; MSI microsatellite instability; MSS, microsatellite stable
tumors; MSI-H, microsatellite instability; NSHDS, Northern Sweden Health
and Disease Study; MS, methionine synthase; MTHFR,
methylenetetrahydrofolate reductase; MMA, methylmalonic acid; m-THF, 5-
methyltetrahydrofolate; MONICA, the Northern Sweden Monitoring of
Trends and Determinants in Cardiovascular Disease cohort; NTD, neural tube
defects; PEMT, phosphatidylethanolamine N-methyltransferase; PL,
pyridoxal; PLP, pyridoxal' 5-phosphate; PA, 4-pyridoxic acid; ROC, receiver
operating characteristics; SAH, S-adenosylhomocysteine; SAM, S-
adenosylmethionine; SHMT, serine hydroxymethyltransferase; SNPs, single
nucleotide polymorphisms; SDMA, symmetric dimethylarginase; THF,
tetrahydrofolate; tHcy, total homocysteine; TYMS, thymidylate synthase;
WHI-OS, Women’s Health Initiative Observational Study; VIP, the
Västerbotten Intervention Programme; XA, xanthurenic acid.
vii
Sammanfattning på svenska
Tjocktarmscancer (Kolorektal cancer (CRC)) är en av de vanligaste
cancerformerna och också en av de vanligaste anledningarna till dödlig utgång
efter en cancerdiagnos. Den typiska patienten är diagnostiserad i 60-70
årsåldern men det finns även människor som bär på olika former av CRC med
en stark ärftlig del och dessa människor är ofta diagnostiserade tidigare i livet.
Det tar många år för CRC att utvecklas från en polyp till cancer som kan sprida
sig till andra delar av kroppen. Detta gör att screening för CRC med hjälp av
koloskopi och feces-prover har goda möjligheter att kunna hitta en tidig
cellförändring och verkar därmed kunna minska risken för dödlighet på grund
av CRC.
Det finns flera etablerade riskfaktorer för CRC, de med starkast påverkan är
ålder, manligt kön och en familjehistoria där flera andra i den närmsta släkten
blivit diagnostiserade med CRC. Flera livsstilsfaktorer spelar också roll, som
rökning, stillasittande, bukfetma, stress och sömnstörningar. En diet med
stort intag av rött kött och processat rött kött, men lågt intag av grönsaker,
fibrer, mjölkprodukter och fullkorn verkar öka risken för att drabbas av CRC.
Trots att vi vet om många riskfaktorer för CRC så är det ingen av dessa som är
väldigt starkt kopplat till risk för CRC, som rökning vid lungcancer, humant
papillomvirus vid livmoderhalscancer, eller hög alkoholkonsumtion vid
levercancer. Vi vet också att även om genetik och ärftlighet spelar roll, så
spelar livsstilsvariabler i de flesta fall en större roll. Det finns därför ett
outforskat glapp som inte kan förklaras av genetiska orsaker och som ej heller
fullt ut förklaras av de livsstilsfaktorer som identifierats som riskfaktorer.
En av de faktorer som har setts ha en potential för att förklara att vissa
människor får CRC och andra inte är folat eller folsyra (den syntetiska formen
av folat). Folat är en B-vitamin som samverkar med flera andra B-vitaminer
(vitamin B2, B6 och B12) i ett metabolt nätverk i kroppen kallat
enkolsmetabolism. Enkolsmetabolism är ett system för hantering och
överföring av enkolsgrupper (även kallade metylgrupper). Flera andra
metaboliter spelar också roll, som kolin, betain och flera sorters
proteinbyggstenar (aminosyror). Nätverket består också av flera enzymer
(proteiner som påskyndar metabola reaktioner i cellen) som kan påverka
balansen av olika metaboliter inom enkolsmetabolism. Förenklat så har
enkolsmetabolism tre huvudsakliga slutmål: byggstenar för DNA och RNA,
metylgrupper för metylgruppsdonatorn S-adenosylmetionin (SAM) som
ansvarar för huvuddelen av metyleringsreaktioner i kroppen och nybildning
viii
av aminosyran cystein som används för den kroppsegna antioxidanten
glutation.
Både metylering och DNA-syntes har en direkt påverkan på cancercellers
tillväxt. Störningar i metyleringsprocesser kan leda till att gener som ökar
cancercellers förmåga att överleva och sprida sig uttrycks mer, och att gener
som hindrar cancerutveckling tystas. God tillgång till byggstenar för DNA ger
ett mer stabilt genom, men kan även vara nödvändiga för att snabbt delande
celler, som cancerceller, ska kunna fortsätta sin tillväxt. Just folat är en av de
faktorer som har diskuterats kunna ha en dubbel roll vid cancerutveckling,
minskar risken att normala celler utvecklas till cancerceller, men öka
livskraften hos redan existerande pre-cancerösa celler.
Folat, eller folsyra, har visat sig minska risken för fostermissbildningar om de
intas innan och tidigt i graviditeten. Många kvinnor har inte planerat sin
graviditet eller vet inte om folsyras skyddande effekt. Därför har flera länder
infört obligatorisk folsyraberikning av mjöl och flingor, så att dessa kvinnor
har en god folatstatus och risken minskar för att deras barn ska födas med
fostermissbildningar. Trots detta har många länder valt att inte införa
obligatorisk folsyraberikning. Motståndet mot berikning har till viss del berott
på misstanken om att folsyra kan öka risken för vissa cancerformer i
allmänhet och CRC i synnerhet. Att studera folsyra, och andra metaboliter i
enkolsmetabolism, i förhållande till CRC-risk är därför ett viktigt
forskningsfält.
Vi har använt blodprover insamlande före diagnos av CRC hos över 600
deltagare och runt 1200 matchande friska kontrolldeltagare som deltagit i
Northern Sweden Health and Disease Study (NSHDS). Vi har jämfört
nivåerna av enkolsmetaboliter i blodet hos de som utvecklat CRC och de som
inte gjort det för att bedöma om det finns en koppling till CRC-risk. Vi gjorde
först en valideringsstudie på en tidigare studie som visade minskad risk för
CRC kopplat till låga nivåer av folat och kunde reproducera de tidigare fynden.
Därefter studerade vi ett flertal metaboliter inblandade i enkolsmetabolism
och såg en koppling mellan höga nivåer av betain och metionin till lägre risk
för CRC. I en tredje studie undersökte vi olika aspekter av vitamin B6 i relation
till CRC-risk. Vitamin B6 i blodet är inte bara en markör för intag av
födoämnen med vitamin B6 utan är även känt att minska vid inflammatoriska
processer i kroppen. Vi såg att låga nivåer av vitamin B6 var associerade med
högre CRC-risk. Vi såg också att detta samband kunde tänkas ske via
interaktioner med inflammatoriska processer, då en markör för vitamin B6
vid inflammation var kopplat till en högre risk för CRC. I vår avslutande studie
undersökte vi flera markörer för total B-vitaminstatus i relation till CRC-risk.
Vi fann att dessa markörer på ett tillfredställande sätt speglade total B-
ix
vitaminstatus och att total B-vitaminstatus samt dessa markörer var kopplade
till en lägre risk för CRC.
Sammantaget har våra studier visat att enkolsmetabolism är associerat med
CRC-risk. De flesta metaboliter inom enkolsmetabolism verkar vara
skyddande eller neutrala gentemot risk för CRC, förutom folat, där låga nivåer
verkar minska risken. Hög total B-vitaminstatus, vilket inkluderar folat,
verkar dock vara sammankopplat med en minskad risk. Våra resultat visar att
det är tänkbart att det är folats roll vid syntes av DNA som kan öka risken för
CRC. Detta överensstämmer med hypotesen om att folat skulle ha en dubbel
roll vid cancerutveckling.
1
Background
Colorectal Cancer
Incidence and Mortality
Colorectal cancers (CRC) are malignant epithelial tumors of the colon and
rectum, also called adenocarcinomas.1 Globally, colorectal cancer is the third
most commonly diagnosed cancer.2 Each year 1.4 million new cases are
diagnosed and 600 000 individuals die from CRC, making it the fourth most
common cause of cancer death.3 In Europe, CRC is the second most common
cause of cancer death, although mortality rates range from comparably low in
Scandinavia and Western Europe to more prominent in parts of Central
Europe.2, 4 The cumulative incidence and mortality of CRC in Sweden and in
the Northern Sweden Health and Disease Study (NSHDS) cohort are depicted
in Figure 1. In Sweden, as in the rest of the world, CRC incidence increases
sharply after age 50.5 However, a recent increase in CRC incidence among
younger persons has been observed, especially in rectal cancers.6 Men are
almost twice as likely as women to be diagnosed with rectal cancers and are
also more likely to be diagnosed with distal colon cancer, while both men and
women carry the same risk for proximal colon cancer diagnosis.7-11
Figure 1. Cumulative risk of incidence and mortality in men and women in Sweden and in the
NSHDS cohort across different age categories. Data from the NSHDS and NORDCAN © 2017
Association of the Nordic Cancer Registries, IARC (Assessed 8/7/17). 5
0
1
2
3
4
5
6
7
8
10 20 30 40 50 60 70 80
Age
Pe
rce
nt
(%)
Cumulative incidence of CRCa
0
1
2
3
4
5
6
7
8
10 20 30 40 50 60 70 80
Age
Pe
rce
nt
(%)
Cumulative mortality of CRCb
Men Women NSHDS Sweden
2
Staging and Prognosis
The overall five-year survival after CRC diagnosis has gradually improved over
the last several decades, especially in wealthier countries. In many parts of
Europe, Australia, and the US, the five-year survival is approximately 65%,
although in lower income countries the five-year survival is still below 50%.4,
12, 13
Prognosis after CRC diagnosis varies depending on tumor stage. Stage is
classified based on local spread (T stage), lymph node involvement (N stage),
and distant metastasis (M stage). These factors are combined into an overall
measure called the TNM stage (ranging from I to IV with several subclasses).
The TNM stage provides valuable prognostic and treatment information.14
Most importantly, survival varies according to TNM staging. Patients
diagnosed with stage I CRC have excellent survival, but only a small
percentage of patients diagnosed with stage IV tumors are still alive five years
after diagnosis.6, 15 Figure 2 shows the five-year cancer-specific survival
according to stage in over 100 000 CRC case participants diagnosed between
1991 and 2000.15 For cases diagnosed between 2006 and 2012, five-year
cancer-specific survival for all stages combined has increased from 65.2% to
66% for colon cancer and 68% for rectal cancer; however, information about
stage is less precise in later surveys.6
Several molecular features in the CRC tumors, described in more detail below,
have been associated with CRC prognosis. Better clinical outcome is
associated with a higher density of tumor-infiltrating T-cells.16 T-cell density
is also associated with microsatellite instability (MSI), a positive prognostic
marker.17 On the other hand, mutations in the BRAF and KRAS oncogenes are
associated with poor overall prognosis, but this association is confined to
tumors with microsatellite stable tumors (MSS).18, 19 Cancer-specific survival
in patients with mutations in the PIK3CA oncogene has been observed to be
longer in patients who used aspirin regularly after diagnosis although this
association was not observed in patients with wild type PIK3CA.20
Survival also varies depending on patient characteristics. For example, cancer
cachexia – muscle and fat storage wasting stimulated by tumors – is reported
to lead to death in at least 20% of cancer patients.21, 22 As cachexia often sets
in before the diagnosis, unexplained weight loss should be seen as a warning
sign of cancer and could explain the better clinical outcome associated with
being overweight (BMI 25-30) at diagnosis.23, 24 In addition, as with many
other cancers, smoking is associated with poor clinical outcome.25
3
Figure 2. Sankey graph showing the distribution of TNM stages in CRC (left) and five-year
CRC-specific survival depending on TNM staging (right). BLUE represents patients who are
alive or dead from causes other than CRC five years after diagnosis. ORANGE represents
patients who died from CRC within the same period. Tumors with unknown stage are excluded.
Adapted from O'Connell et al.15
Colorectal Tumorigenesis and Heterogeneity
In 1990, Fearon and Vogelstein published a groundbreaking article that
described specific pathways in the progress from colorectal adenoma to
carcinoma. They found that a stepwise pattern of genetic and epigenetic
alterations activating oncogenes and silencing tumor suppressor genes
ultimately leads to cancer. The great bulk of sporadic colorectal tumors was
proposed to progress through the chromosomal instability pathway (CIN),
which is characterized by mutations in oncogenes and tumor suppressor
genes.26 Furthermore, around 15-20% of tumors are categorized by the loss
4
of expression of DNA mismatch repair proteins. These mismatch repair
deficient tumors display a high level of microsatellite instability (MSI-H) due
to deletion and insertion mutations at microsatellites spread throughout the
genome.1, 17
Molecular subtyping of CRC tumors has revealed several important tumor
characteristics besides MSI, and mapping the heterogeneity of CRC tumors
has been based on mutations in the BRAF and KRAS oncogenes as well as CpG
island methylator phenotype (CIMP) status.27 CIMP tumors are characterized
by hypermethylation of cytosine in the symmetrical dinucleotide CpG,
specifically in promoter regions in the DNA, leading to silencing of tumor-
suppressor genes.28 Mutations in KRAS and BRAF are oncogenic driver
mutations that provide CRC cells with self-sufficient growth signals via the
MAPK pathway. MAPK signals regulate proliferation, differentiation, and cell
motility through phosphorylation of transcription factors.28, 29 Almost all
sporadic MSI-H tumors have mutations in the BRAF oncogene and extensive
DNA methylation, although patients with Lynch syndrome have MSI-H
tumors but no BRAF mutation. These characteristics mean that a combination
of the two tests can help identify affected patients and families.30
In 2015, the CRC Subtyping Consortium established four molecular subtypes
for CRC: MSI immune (14%), canonical (37%), metabolic (13%), and
mesenchymal (23%).) The MSI immune subgroup is characterized by MSI,
CIMP-high, and BRAF mutations, the metabolic subgroup by KRAS mutation
and CIMP-low, and the canonical and mesenchymal subgroups by frequent
somatic copy alterations and by Wnt signaling pathway activation and
amplification of the Myc transcription factor or stromal infiltration,
respectively.31 That is, compared to the classification described by Fearon and
Vogelstein, the MSI immune subgroup represents the mismatch repair
deficient subtype and the CIN is divided into three distinct groups.
Treatment
Localization in the colorectum and TNM stage are important factors when
deciding on treatment strategies. Preoperative T-stage is best determined with
MRI in rectal cancers, 32, 33 and more advanced stages are treated with
neoadjuvant radiotherapy.34 For rectal cancers, the rectum, mesorectum, and
surrounding fascia are surgically removed;35 however, for colon cancers, a
colectomy is performed where the tumor and corresponding lymph nodes are
resected.
The use of chemotherapy depends on the TNM stage of the tumor. Stage I
tumors are often completely removed and require no adjuvant chemotherapy
5
to decrease the risk of recurrence. More advanced colonic tumors have a high
chance of recurrence after surgical removal and adjuvant chemotherapy is
recommended. The antifolate agent fluorouracil (5-FU) functions by
inhibiting thymidylate synthase (TYMS) and is the basis of chemotherapeutic
practice in CRC.36, 37 Inhibition of TYMS deprives cells of the only de novo
source of the DNA nucleoside thymidine.38 TYMS uses 5,10-
methylenetetrahydrofolate (me-THF) as a methyl group donor and the
reduced me-THF derivative leucovorin (folinic acid, not to be confused with
folic acid) increases the inhibition of TYMS by 5-FU and improves clinical
outcome parameters compared to 5-FU alone.36, 37, 39 Stage IV tumors are
tumors with distant metastasis at one or more sites, most often in the liver
(~13% of all CRC tumors), lung (~5%), bone (<1%), and brain (<1%).40
Prognosis, especially for patients with bone or brain metastasis, is poor and
treatment is a combination of cytotoxic drugs.40
Future oncological treatments may target the molecular characteristics of
specific tumors.41 Already, patients with mutations in the oncogenes BRAF or
KRAS are known to be poor responders to epidermal growth factor inhibitor
therapy.42-45
Screening
The transition time through the adenoma-carcinoma sequence is slow, and
the mean sojourn times (the period from when the tumor first could be
detected by screening until diagnosis) for advanced adenoma to CRC
diagnosis has been estimated to vary between five and 20 years, as the average
annual transition rate is between 5% and 20%.46-48. Given this slow
progression and the relative ease of curative surgery, screening for CRC has
great potential compared to other cancers.48 Randomized trials on yearly
screening with fecal occult blood tests show a decrease in CRC mortality, but
not on all-cause mortality.49 Screening with flexible sigmoidoscopy also
decreases CRC incidence and mortality, specifically in men and in women
under 60-years-old.50-52 No randomized trials on screening with colonoscopy
have been published, but observational studies show promise.53, 54 Screening
with colonoscopy is common in the ages of 50-75 in the US and Germany.55 In
Sweden, two study cohorts have been established to evaluate CRC screening
with a fecal blood test and colonoscopy.56-58 Before the start of these studies,
screening for CRC was uncommon in Sweden, and opportunistic screening
was observed to be almost nonexistent.59
6
Risk Factors and Preventive Factors
There is no single environmental factor that greatly increases CRC risk.
However, several factors that modulate CRC risk have been observed. The
Continuous Update Project (CUP), which reviews the evidence from
randomized trials and cohort studies regarding different cancers, published a
report on CRC in 2010 and an updated report in 2017.60 The CUP found
convincing evidence for increased CRC risk for body fatness, adult height,
consumption of red and processed meat, and for men with a high intake of
alcohol. The CUP also found that physical activity and foods containing fiber
decreased CRC risk. The 2017 CUP update reinforced the results for red and
processed meat and alcohol and further found convincing evidence for a lower
CRC risk associated with higher intake of dairy products and whole grains.60
Furthermore, the following additional risk factors have been identified:
inflammatory bowel diseases (particularly when poorly controlled),61-63
smoking,64 and type II diabetes.65
Overweight measured by BMI is associated with increased CRC risk; when
stratified by sex and tumor localization, BMI is associated with colon cancer
in both men and women while in rectal cancer the association is only observed
in men.66, 67 Waist-to-hip ratio is a better measurement of abdominal obesity
and is more strongly associated with colon cancer risk than BMI.66, 67 In
Mendelian randomization studies, which aim to strengthen the evidence for
causality of risk relationships, gene variants associated with high BMI were
associated with overall CRC risk, colon cancer risk in women, and rectal
cancer risk in men.68-70 Red and processed red meat have a similar
heterogeneous CRC risk distribution as body fatness when examining male
and female sex separately. Dietary intake of red and processed red meat is only
significantly associated with CRC risk in men.71, 72 Moreover, the associations
to the risk of CRC and other diseases have been observed to differ for
unprocessed and processed red meat.71, 73
Additional preventive factors include estrogen therapy,74 and higher plasma
concentrations of vitamin D, but not vitamin D intake.75 Regular aspirin use is
associated with a lower CRC risk and a lower risk of distant metastasis.76, 77 An
association with a lower risk of new adenomas in colorectal cancer patients
has also been observed.78
Although no single environmental factor contributes greatly to CRC risk,
taken together environmental factors substantially contribute to CRC risk and
a large Scandinavian twin study found that only 35% of CRC incidence is due
to hereditary factors,79 including the two most common forms of hereditary
CRC – Lynch syndrome (Hereditary non-polyposis colorectal cancer, HNPCC)
7
and familial adenomatous polyposis (FAP). These two cancers represent 3-5%
of the total CRC incidence. A family history of CRC, combining both
environmental and hereditary factors, is one of the strongest risk factors for
CRC.80
Further evidence of a strong environmental component in CRC risk comes
from comparing changes in CRC incidence in different populations and over
time. While incidence is stable or declining in Europe and the United States,
a rapid increase in CRC has been observed in populations that have made a
transition from a relatively low median income to higher income, such as in
Japan and in Eastern European countries.81, 82 Populations that have
emigrated from low median income countries to more prosperous societies
have also experienced a transition to the higher incidence rates in their new
homelands.81, 82 A classic example is the increased incidence of CRC observed
in the emigrated Japanese communities in Hawaii and California compared
to the incidence rate in Japan.83 Taken together, these epidemiological
considerations point to a combination of environmental risk factors
associated with a Western lifestyle converging in an increased CRC incidence.
In addition to the previously mentioned risk factors, it has been suggested that
stress, sleep deprivation and circadian rhythm disturbances, and an energy
dense diet could explain the higher risk in Westernized societies.81, 82
8
One-Carbon Metabolism
Enzymatic Reactions and Metabolites
One-carbon metabolism, a complex network of enzymatic reactions (Figure
3), uses different methyl group sources as enzymatic cofactors, including
certain amino acids, choline, and betaine and B-vitamins riboflavin (B2),
pyridoxal phosphate (B6), folates (B9) and cobalamin (B12).84-87 One-carbon
metabolism is essential for the synthesis of nucleotides and proteins and uses
the universal activated methylator co-substrate S-adenosylmethionine (SAM)
as the methyl group donor. Methylation reactions are important for genome
stability and gene translation, and ultimately risk of tumorigenesis.84, 85, 88
Figure 3. One-carbon metabolism. PURPLE circles signify enzymes with their respective B-
vitamin co-factors in BLACK. Methylation reactions, redox reactions, and nucleotides in
ORANGE are the three primary outputs of one-carbon metabolism. Abbreviations: BHMT,
betaine-homocysteine methyltransferase; CBS, cystathionine β-synthase; CSE, cystathionine γ-
lyase; DMG, dimethylglycine; dTMP, deoxythymidine monophosphate; dUMP, deoxyuridine
monophosphate; DHFR, dihydrofolate reductase; f-THF, 10-formyltetrahydrofolate; MS,
methionine synthase; me-THF, 5,10-methylene-THF; MTHFR, methylenetetrahydrofolate
reductase; m-THF, 5-methyltetrahydrofolate; SAH, S-adenosylhomocysteine; SAM, S-
adenosylmethionine; SHMT, serine hydroxymethyltransferase; THF, tetrahydrofolate; tHcy,
total homocysteine; TYMS, thymidylate synthase.
9
The main constituents of one-carbon metabolism are the folate and
methionine cycles. Labile methyl groups are mainly provided through serine
in the folate cycle and to a lesser extent via the choline oxidation pathway.85,
86 Folic acid – the stable and synthetic form of folate – is commonly added to
food (folic acid fortification) to decrease the risk of neural tube defects during
pregnancy.89 Folic acid has a higher gastrointestinal bioavailability than
naturally occurring folates and needs to be reduced twice in order to enter the
folate circle as tetrahydrofolate (THF). This conversion, catalyzed by
dihydrofolate reductase (DHFR), takes place almost exclusively in the liver,
and not as previously thought in the bowel wall. Consequently, a folic acid
bolus taken up in the duodenum passes unmodified into the portal vein.90, 91
In humans, the conversion of folic acid to THF by DHFR can be slow and
DHFR is easily saturated.92 In populations that consume folic acid that has
been added to foods, unmetabolized folic acid is found in serum, maternal
blood, and in umbilical cord blood.93, 94
THF is converted by serine hydroxymethyltransferase (SHMT) into 5,10-
methylene-THF (me-THF).84 SHMT catalyzes the conversion of serine to
glycine and THF to me-THF in a vitamin B6 dependent reaction.85, 95 me-THF
can be converted back to THF by thymidine synthase (TYMS) in a reaction
where deoxythymidine monophosphate (dTMP) is formed from deoxyuridine
monophosphate (dUMP).96 There are two additional possible pathways for
me-THF: it can either be further reduced to the predominating folate form
found in blood,91 5-methyltetrahydrofolate (m-THF), by the vitamin B2-
dependent enzyme methylenetetrahydrofolate reductase (MTHFR) or
undergo a series of reactions to form 10-formyltetrahydrofolate (f-THF).84, 97
While f-THF is a precursor for nucleotide synthesis, m-THF exits the folate
cycle and enters the methionine cycle by methyl group transfer by the vitamin
B12-dependent methionine synthase (MS) enzyme. The methyl group
provided by m-THF converts homocysteine into methionine and m-THF back
into the folate cycle as unmethylated THF. Alternatively, methyl groups for
remethylation of homocysteine can be provided via the choline oxidation
pathway and the enzyme betaine-homocysteine methyltransferase (BHMT).86
Homocysteine can also be diverted through the pyridoxal' 5-phosphate-
dependent (PLP) transsulfuration pathway, where homocysteine provides
sulfur for cysteine formation. 68, 98 The methionine formed by BHMT or MS is
the substrate for synthesis of S-adenosylmethionine (SAM) carrying an
activated methyl group. SAM, involved in the majority of methylation
reactions, is transformed into S-adenosylhomocysteine (SAH).98 SAH can be
transformed into homocysteine by S-adenosylhomocysteine transferase.99
SAM carries the role of a universal reactive methylator and constitutes one of
the most common cofactors in enzymatic reactions in the body.84, 99, 100
10
The main outputs from one-carbon transfer reactions (dTMP, f-THF,
methionine, and cysteine) is involved in three distinct functions: nucleotide
synthesis (dTMP and f-THF), methylation reactions (methionine), and redox
balance and glutathione synthesis (cysteine).84-86
One-carbon Metabolism in Cancer and Genomic Instability
The history of one-carbon metabolism is intertwined with cancer research and
treatments since the late 1940s when Sidney Farber and colleagues observed
a temporary remission of acute leukemia in children treated with the folic acid
antagonist aminopterin.101, 102 Additional chemotherapies based on the
antifolate concept followed, such as methotrexate and 5-FU.96 Many of these
inhibit TYMS, resulting in decreased synthesis of thymine and a build-up of
uracil. This leads to uracil misincorporation into DNA and DNA instability,
which induces cell cycle arrest and apoptosis.38, 103 f-THF, similar to the TYMS
product dTMP, is an essential co-factor and co-substrate for nucleotide
synthesis and is necessary for sufficient nucleotide supply and increases
genome stability.104
Much of research into one-carbon metabolism has focused on polymorphisms
that impact the activity of MTHFR, the central hub linking the folate cycle to
the methionine cycle. A common polymorphism in the gene encoding
MTHFR, MTHFR 677TT, results in thermolability and lower activity of the
enzyme in the presence of low folate concentrations.97, 105 The MTHFR 677TT
polymorphism has convincingly been associated with a lower risk of CRC,106
with a possible exception in populations with a poor folate status.107
Genome instability and increased mutation rate are one of the hallmarks of
cancer, and tumors are characterized by aberrant DNA methylation and
histone modification.108 DNA methylation, one of several epigenetic
mechanisms that control gene expression, depends on the availability of the
universal methyl donor SAM. SAM and several other factors in one-carbon
metabolism are involved in DNA and histone methylation. These factors
include polymorphisms in the MS and/or MTHFR genes and dietary intake of
folate and methionine.109-113
Tumor cells are unable to proliferate when methionine is replaced by the
immediate precursor homocysteine in cell media, whereas normal cells can
synthesize necessary methionine from homocysteine, a phenomenon known
as methionine dependency.114 Methionine is an essential amino acid involved
in DNA methylation, protein synthesis, and synthesis of polyamines and
glutathione.84, 86 In animal models, low methionine diets inhibit tumor growth
and size and synergizes with cytotoxic treatments.115 However, in humans
11
dietary methionine restrictions do not cause a reliable decrease of plasma
concentrations of methionine, and the tumor cells scavenge amino acids from
the tumor stroma, potentially limiting the impact of dietary methionine
restrictions.115
Folate Cycle Components and Folic Acid in CRC
The folate cycle comprises different folate forms (Figure 4). Folate is a generic
term for several compounds, including folic acid and its derivatives at varying
reduction states. These include 5-methyltetrahydrofolate (me-THF), 10-
formylTHF (f-THF), 5,10-methyleneTHF (m-THF), and unsubstituted THF.87,
116 Although the amino acids serine and glycine are important for feeding the
folate cycle with one-carbon units,84, 117 research has mainly focused on folate
and comparatively little on the methyl donors serine and glycine.85
Figure 4. Folate cycle. PURPLE circles signify enzymes with their respective B-vitamin co-
factors in BLACK. Nucleotides in ORANGE are the primary outputs of the folate cycle, m-THF
is integral. Abbreviations: dTMP, deoxythymidine monophosphate; dUMP, deoxyuridine
monophosphate; DHFR, dihydrofolate reductase; f-THF, 10-formyltetrahydrofolate; MS,
methionine synthase; me-THF, methylene-THF; MTHFR, methylenetetrahydrofolate reductase;
m-THF, 5-methyltetrahydrofolate; SHMT, serine hydroxymethyltransferase; THF,
tetrahydrofolate; tHcy, total homocysteine; TYMS, thymidylate synthase.
Important dietary sources of naturally occurring folate are leafy greens, fruit
(including orange juice), and legumes.87 Foods from animal sources only
contain small amounts of folate, with the exception of liver and egg yolks.
Adequate folate status before and during early pregnancy protects the fetus
against birth defects such as spina bifida, encephalocele, and anencephaly
(collectively known as neural tube defects, NTD).89, 118 NTDs can result in
death or severe life-long disability and NTD prevention represents a major
12
public health challenge.118, 119 Consequently, in the early 1990s many countries
introduced mandatory folic acid fortification of common foodstuffs.120
Countries with mandatory or widespread voluntary folic acid fortification are
depicted in Figure 5. Notable exceptions are Russia, India, China, Europe
(except Moldavia), and parts of Africa.
Figure 5. ORANGE signifies countries with a mandatory folic acid fortification of grains;
BROWN signifies countries with a widespread voluntary folic acid fortification of grains. BEIGE
signifies countries with no widespread fortification of grains. Adapted from the Food Fortification
Initiative: http://www.ffinetwork.org/global_progress/ (assessed 6/20/17).
Following fortification, studies on pre-clinical models started to emerge that
indicated that folate may have a double-edged role in CRC, with a protective
role in the normal colorectal mucosa but accelerating the growth of already
established pre-cancerous lesions.116, 117 Subsequent studies on intake of folate
and folic acid seemed to indicate a weak inverse association or no association
with CRC risk,121-123 although two Norwegian randomized trials on folic acid
supplementation, conducted with the primary aim to lower homocysteine to
reduce the risk of heart disease, observed as a secondary outcome an increased
risk of some types of cancers.124 However, a later published meta-analysis of
randomized trials on folic acid supplementation only found a borderline,
although nonsignificant, increase of some types of cancers.125 Prospective
case-control studies on plasma folate concentrations and risk of CRC were
generally underpowered and taken together provide no clear picture on the
influence of folate status in CRC risk.126-131 A study from 2006, from a northern
Swedish population with low folate intakes and no mandatory fortification,
found that study participants with the lowest plasma concentrations of folate
13
also had the lowest risk association to CRC.132 A few years later, these findings
were reproduced in a large US study,133 but the largest study to date with over
1300 cases concluded that plasma concentrations of folate were not associated
with CRC risk.134 The same results were also observed concerning the risk of
distal colorectal adenoma.135 There is still some remaining discussion
regarding a temporary interruption in the trend of decreasing CRC incidence
in the US and Canada coinciding with the introduction of folic acid
fortification120, 136 and the potentially harmful role of unmetabolized folic
acid.137-140
Concerns have been raised that folate should not be studied in isolation due
to the tightly intertwined reciprocal relationship and interactions with other
one-carbon metabolites and polymorphisms.138, 139 In a recently published
article that relies on analyses of a large panel of biomarkers and single
nucleotide polymorphisms (SNPs) involved in one-carbon metabolism and
using advanced statistical methods to investigate the full picture of one-
carbon metabolism, we observed that plasma concentrations of folate (direct
association) and vitamin B2 and B6 (inverse association) were most strongly
associated with CRC risk.141
Although serine and glycine are amino acids that are abundant in plasma,
intake or circulating levels of these metabolites have not been studied in
relation to clinical cancer risk.85 However, in cell lines and in an in vivo model,
serine starvation decreased tumor growth.142, 143
Methionine and Choline Oxidation Pathway Metabolites in CRC
The enzyme BHMT catalyzes the remethylation of homocysteine into
methionine and betaine into dimethylglycine (DMG).86 Methionine is the
substrate for SAM synthase, and the product of the synthase reaction (i.e.,
SAM) contains an activated methyl group that is essential for most
methylation reactions in the body.84, 86, 144 Betaine can be provided either
through diet (the main sources are wheat and wheat bran)145 or through
oxidation of choline.86 Betaine can be further sequentially oxidized to DMG
and sarcosine (Figure 6). Choline is not only the precursor of betaine in the
choline oxidation pathway but also involved in the hepatic synthesis of the
neurotransmitter acetylcholine and structural lipoproteins.86, 146 Dietary
choline deficiency may cause DNA strand breaks and altered epigenetic
regulation of DNA expression and histone methylation.146 Choline can be
supplied either through diet (the richest sources are organ meats and egg
yolks)145 or through de novo synthesis by the SAM-dependent enzyme
phosphatidylethanolamine N-methyltransferase (PEMT).86 PEMT is
upregulated by estrogens, and pre-menopausal women are less sensitive to
14
choline deficiency than men or post-menopausal women.147 Deficiency of the
essential nutrient choline causes fatty liver disease and liver damage.86
Figure 6. The methionine cycle and the choline oxidation pathway. PURPLE circles signify
enzymes with their respective B-vitamin co-factors in BLACK. Methylation reactions in
ORANGE are the primary outcomes of the methionine cycle. Abbreviations: BHMT, betaine-
homocysteine methyltransferase; MS, methionine synthase; m-THF, 5-methyltetrahydrofolate;
SAH, S-adenosylhomocysteine; SAM, S-adenosylmethionine; THF, tetrahydrofolate; tHcy, total
homocysteine.
Study results on dietary intake of choline and betaine in relation to CRC risk
are inconsistent. Increased dietary intake of choline appeared to be linearly
associated with colorectal adenoma risk while dietary intake of betaine was
inversely associated in a study nested in the all-female Nurses’ Health Study.
Similar results for betaine, but not choline, were observed in the Norwegian
Colorectal Cancer Prevention cohort.135, 148 In the all-male Health Professional
Follow-up Study, neither dietary intake of choline or betaine was observed to
be associated with CRC risk,149 but in a Chinese case-control study higher
dietary intake of choline, but not betaine, was associated with a decreased CRC
risk in both men and women.150
High dietary intake of methionine is consistently associated with a lower CRC
risk.151 Plasma concentrations have been investigated in two studies. A study
nested in the Women’s Health Initiative Observational Study (WHI-OS) found
that plasma concentrations of betaine were associated with a lower risk of CRC
risk, while plasma concentrations of choline and dimethylglycine were not
significantly associated with CRC risk.152 Similarly, a case-control study
comprising over 1300 men and women with diagnosed CRC nested in the
European Prospective studies into Cancer and Nutrition (EPIC) found that
plasma betaine was associated with a decreased CRC risk and observed an
interaction with folate status: in participants with a low folate status, plasma
betaine was observed to be inversely associated with CRC risk.153 In addition,
plasma concentrations of choline and methionine were associated with a
15
lower overall CRC risk, but in the case of choline, the association was mainly
driven by a lower risk in women. In the same study, plasma dimethylglycine
was not associated with CRC risk. Plasma concentrations of methionine and
betaine are also associated with a lower risk of distal colorectal adenoma.135
Transsulfuration Pathway Metabolites in CRC
Over the past several decades, homocysteine emerged as a potential causal
risk factor in atherosclerosis; during the late 1990s, several randomized trials
were launched to evaluate whether a homocysteine-lowering therapy with B-
vitamins (vitamin B6, folate, and vitamin B12) could lower the risk of
cardiovascular incidents.154 Despite significant reductions in homocysteine
concentrations, the expected lower risk of cardiovascular incidents could
largely not be verified.154, 155 Circulating homocysteine is also implicated in
cognitive disease and NTDs, but only limited evidence suggests that
homocysteine could be a risk factor in itself. Instead, homocysteine could be
considered a marker of poor status of one-carbon metabolism or secondary
to underlying disease.99
Figure 7. The transsulfuration pathway. PURPLE circles signify enzymes with their respective
B-vitamin co-factors in BLACK. Redox reactions in ORANGE are the primary outcomes of the
transsulfuration pathway. Abbreviations: CBS, cystathionine β-synthase, CSE, cystathionine γ-
lyase; tHcy, total homocysteine.
Homocysteine can either be remethylated to methionine by vitamin B12
dependent MS using m-THF as a cosubstrate or by the betaine-dependent
BHMT.84, 99, 156 Higher dietary methionine intake increases homocysteine
concentrations and the transformation of excess homocysteine to cysteine by
the two vitamin B6 dependent enzymes cystathionine β-synthase (CBS,
homocysteine to the intermediate cystathionine) and cystathionine γ-lyase
(CSE, cystathionine to cysteine).98, 99, 156 CBS and CSE are part of the
transsulfuration pathway depicted in Figure 7. Elevated plasma
16
concentrations of homocysteine mainly reflect deficiencies in folate and/or
vitamin B12, but could also reflect impaired clearance due to insufficient
vitamin B6 concentrations or genetic variants of genes involved in
homocysteine metabolism (MTHFR, MS, CBS, and CSE).97, 99, 100, 156
Supplementation with folic acid, vitamin B6, vitamin B12, and the methyl
group donor betaine can decrease plasma homocysteine concentrations,157
and for vitamin B2 the same decrease has been found in individuals with the
MTHFR 677TT polymorphism.158 Recently, three homocysteine ratios were
introduced that could more accurately than homocysteine determine
deficiencies in total B-vitamin status.159 Increases in the ratios of
homocysteine to cysteine (hcy:cys) and to creatinine (hcy:cre) and hcy:cys to
creatinine (hcy:cys:cre) partly reflect disturbances in the transsulfuration
pathway (hcy:cys), the methionine cycle (hcy:cre), and/or both (hcy:cys:cre).
Compared to homocysteine, all three ratios had higher sensitivity and
specificity in estimating total B-vitamin status.159
Plasma concentrations of homocysteine have not been observed to be
significantly associated with CRC risk,132, 160, 161 with one exception: a large
study nested in the WHI-OS reported a linear association between
homocysteine concentrations and CRC risk.162 The authors hypothesized that
the results could be due to the larger all-female cohort with overall lower
plasma concentration of homocysteine compared to previous studies. In the
same study, higher plasma cysteine was associated with a decreased CRC risk.
Because vitamin B6 deficiency may impair homocysteine catabolism, low
plasma concentrations of cysteine could be an expression of poor vitamin B6-
status,156 and since plasma concentrations of pyridoxal' 5-phosphate (PLP),
the most commonly used marker of vitamin B6-status, is consistently
associated with a decreased CRC risk,163 the association between cysteine and
CRC could reflect a low vitamin B6 status. Alternatively, since the
transsulfuration pathway supplies approximately half of the cysteine needed
for synthesis of the major redox buffer glutathione,95 reactive oxygen species
could hypothetically lie behind the increased CRC risk. Another study found
that plasma concentrations of cysteine have a similar trend, although
statistically nonsignificant; the study’s small sample size (n = 118) may explain
the lack of a significant association.160
B-vitamins Involved in One-carbon Metabolism and CRC Risk
One-carbon metabolism depends on B-vitamins as essential co-factors.
Vitamin B2 is a cofactor for MTHFR, vitamin B12 and m-THF (as a co-
substrate) is a cofactor for MS, and vitamin B6 is a cofactor for SHMT, CBS,
and CSE.84, 164 Thus, vitamin B6 modulates the influx of methyl groups from
serine into the folate cycle by converting THF to me-THF as well as the efflux
17
of homocysteine through the transsulfuration pathway. Vitamin B2 is a co-
factor for the enzymatic reaction catalyzed by MTHFR that regulates the
balance between me-THF and m-THF and consequently the balance between
the folate cycle (me-THF) and the methionine cycle (m-THF). Finally, vitamin
B12 is necessary for the only reaction m-THF can take part in, remethylation
of homocysteine to methionine transferring the methyl groups donated from
serine to the universal methyl donor SAM.84 Poor vitamin B12 availability can
result in an accumulation of me-THF, a phenomenon known as the folate
trap,165 and inhibition of thymidine synthesis and increased genome
instability.166 Excess homocysteine can be transformed to the amino acid
cysteine by the vitamin B6 dependent enzymes of the transsulfuration
pathway.84, 98, 99
Overall, associations between dietary intake of these B-vitamins and CRC risk
are inconsistent. Presently, there are no reports on statistically significant
associations to CRC risk with either micronutrient.163, 167, 168
The most commonly used plasma marker of vitamin B6 status is pyridoxal' 5-
phosphate (PLP).164 In contrast to the studies on estimated dietary intake of
vitamin B6, high plasma concentrations of PLP are consistently associated
with a lower CRC risk.163, 167-170
Vitamin B2 status in plasma can be measured either as riboflavin alone or a
combination of riboflavin and flavin mononucleotide (FMN).171 Riboflavin is
the precursor of FMN,87 and both vitamers are useful indicators of vitamin B2
status in population studies.171 Two previous studies have been conducted on
plasma markers of vitamin B2: in a study from 2008, no association to CRC
risk was observed for plasma concentrations of riboflavin,161 and in a study
nested in the EPIC-cohort (1300 case participants) total plasma vitamin B2
concentrations were associated with a decreased CRC risk.170
Vitamin B12 status is generally measured in plasma as total cobalamin – all
forms of vitamin B12 bound to different binding proteins.172 The serum
marker methylmalonic acid (MMA) and total homocysteine are also routinely
used as functional vitamin B12 markers; in cancer patients, a combination of
plasma cobalamin, MMA, and homocysteine has been proposed to be better
for a total assessment of vitamin B12 status.173 Four prospective studies on
plasma markers of vitamin B12 status and CRC risk have been conducted so
far and no significant associations have been found.160, 161, 170, 174 One of these
studies, nested in the NSHDS and based on a portion of the CRC cases and
controls in this thesis, found that plasma concentrations of cobalamin were
associated with a lower risk of rectal cancer, but not CRC.174 It was suggested
that this was due to fewer rectal tumors with microsatellite instability (MSI),
18
which is associated with hypermethylation of the mismatch repair gene
MLH1.174 However, similar associations between plasma cobalamin and rectal
cancer risk were not observed in the larger EPIC study or in a cohort of male
smokers.161, 170
A recently published comprehensive investigation of all one-carbon
metabolites using Bayesian network learning found that vitamin B2 and B6
had strong independent associations with lower CRC risk.141 Plasma
concentrations of vitamin B2 and B6 are also associated with a lower risk of
distal colorectal adenoma.135
Inflammatory Interaction with Vitamin B6 and CRC Risk
Vitamin B6 is most commonly measured in plasma as PLP.164 PLP is the active
coenzyme form of vitamin B6 and functions in over 100 different enzymatic
reactions in the human body.175 PLP is a cofactor for enzymes in the
kynurenine pathway,176 and in one-carbon metabolism, PLP is involved in the
transsulfuration pathway and the folate cycle.84 Free PLP in plasma can be
converted to pyridoxal (PL) by alkaline phosphatase, an enzyme located on
the membrane of all cell types.164 PL is dephosphorylated PLP that is capable
of crossing the cell membrane and can therefore be categorized as the
transport form of vitamin B6.164 Excess PL is catabolized to 4-pyridoxic acid
(PA) in the liver and excreted by the kidneys.164
Conditions associated with inflammation – e.g., cancer, heart disease, stroke,
diabetes, rheumatoid arthritis, and inflammatory bowel diseases – are
associated with low plasma PLP.164 Furthermore, Plasma PLP is inversely
associated with inflammatory biomarkers, including C-reactive protein (CRP)
and neopterin.177 The decrease of plasma PLP during inflammation is due to
both increased catabolism and increased tissue uptake at the site of
inflammation.177
To further study the inflammatory aspects of vitamin B6, a ratio of the
different species is used. PAr is the ratio of PA/(PLP+PL) – i.e., the catabolite
divided by the sum of the active and the transport forms of vitamin B6.164, 177
This ratio mirrors the increased catabolism observed in inflammation as well
as the redistribution of plasma PLP and PL into tissues with inflammatory
activity. Consequently, variance in PAr was observed in multiple linear
regression models to be almost entirely determined by a combination of four
other inflammatory markers, although the common biomarker confounders
smoking and kidney function did not influence PAr. PAr, unlike PLP, was not
influenced by vitamin B6 supplementation.177 Taken together, PAr shows
19
promise as a sensitive marker of vitamin B6 mediated systemic inflammation,
and PAr is robust in regards to potential confounders.
NAD+ synthesis from tryptophan via the kynurenine pathway mainly takes
place in the liver.176 Two of the enzymes in this pathway are strictly PLP-
dependent: the kynurenine aminotransferase (KAT) enzyme that converts 3-
hydroxykynurenine (HK) to xanthurenic acid (XA) and kynureninase (KYNU)
that converts the same substrate to 3-hydroxyanthranilic acid (HAA).176
Consequently, increased concentrations of HK in plasma and urine are
observed in vitamin B6-deficiency.164, 176 The ratios of HK to XA and HAA
(HK:XA and HK:HAA, respectively) are therefore increased in PLP-deficiency
and inversely associated with plasma PLP.178 HK:XA has a slightly stronger
association with plasma PLP than HK alone and is less influenced by
confounders such as BMI, kidney function, and inflammation.178 Therefore,
HK:XA is a potential marker of functional intracellular vitamin B6 status. It is
currently unknown if this ratio also reflects aberrations in the kynurenine
pathway and if these potential aberrations might influence carcinogenesis.
However, XA has been observed to be negatively associated with the
inflammatory marker CRP and cancer mortality.179
In the Hordaland Health Study cohort, PAr was associated with cancer risk
after adjusting for confounders although HK:XA only reached borderline
significance. 180 When stratifying for cancer site, PAr was mainly associated
with lung cancer risk, but also weakly associated with an increased CRC
risk.180
In summary, vitamin B6 status in plasma is commonly estimated as PLP, but
the influence of inflammation on PLP concentrations makes the association
between decreased CRC risk and plasma PLP observed across many studies
difficult to interpret.
20
Summary and Overall Perspectives
CRC is one of the most commonly diagnosed cancers and has a high societal
and economic cost2 as many people will either be diagnosed or have a friend
or family member diagnosed with CRC. Although treatment is slowly
improving, CRC is still one of the most common causes of cancer death.3
Epidemiological studies on emigrated populations and changes in populations
over time suggest that there is a strong environmental aspect in CRC
incidence.81-83 Many of the suggested environmental risk factors are grouped
together as factors associated with a Western lifestyle (i.e., stress, circadian
rhythm disturbances, smoking, an energy dense diet lacking in
micronutrients, and lack of physical activity).81, 82 Many of the environmental
risk factors are still unknown or not sufficiently researched.
In the NSHDS cohort, low plasma folate status is associated with a lower CRC
risk.132 In 2007, these results were taken into account when the Swedish
Agency for Health and Disease assessed the potential risks and benefits of folic
acid fortification.181 The complexity of one-carbon metabolism is mirrored in
the differing opinions researchers have on mandatory folic acid fortification
of grains.117, 119 While many researchers stress the importance of lowering the
incidence of birth defects, others stress the need for more research, especially
on cohorts with an overall low folate status since folic acid fortification and
supplementation often results in supraphysiological folate concentrations.91,
119 The population-based NSHDS cohort provides an overall low folate status
and is sufficiently large to allow for subgroup analysis.182
One-carbon metabolism influences both nucleotide synthesis and methylation
reactions, not only providing fodder for rapidly dividing cells but also
increasing genome stability.84, 117 Folate acts as a carrier of the methyl groups
needed for both methylation and nucleotide synthesis, while other one-carbon
metabolites are only involved in methylation. A comparison of these one-
carbon metabolites and folate in relation to CRC risk makes it possible to
distinguish between the two main outputs of one-carbon metabolism and
relate these outputs to carcinogenesis.
Previous research into one-carbon metabolism has treated the many factors
as independent entities when relating them to risk estimates, despite the
interdependence of the metabolites and enzymes involved. This can be
mitigated by calculating interactions between metabolites, or by estimating
the total status of the major pathways of one-carbon metabolism, and relating
these estimates to CRC risk. Interactions between folate and cobalamin have
been observed in cancer, cognitive decline, and anemia.166, 183 Could a possible
deleterious role for folic acid be attenuated by fortifying foodstuffs with both
21
folic acid and vitamin B12?184 Moreover, metabolites in the choline oxidation
pathway (mainly betaine and choline) are particularly important in
individuals with low folate status, and has, similarly to folate, implications for
epigenetic alterations and plasma homocysteine concentrations.86, 185, 186
Should these metabolites also be considered to be included in a fortification
program? A broader understanding of the impact of one-carbon metabolism
as a whole could have implications for public health initiatives.
While the rationale and motives for multi-metabolite fortification can be well-
meaning and scientifically sound, there is always a risk of negative unforeseen
public health outcomes. In 1994 evidence of harmful effects of vitamin
supplements started to emerge in the Alpha-Tocopherol Beta-Carotene
Cancer Study. Almost 30000 male Finnish smokers took part in the
randomized controlled trial with lung cancer as the primary outcome.
Contrary to expectations, beta-carotene increased deaths from lung cancer
and ischemic heart disease. Similarly, the Selenium and Vitamin E Cancer
Prevention Trial was a randomized controlled trial involving over 35000 men
initiated to investigate if vitamin E and selenium could influence the risk of
prostate cancer. The trial had to be stopped prematurely in 2011 when a
significant increase in prostate cancer incidence was observed in the
treatment arm.187 In the Norwegian Aspirin/Folate Polyp Prevention Study,
over 1000 participants with a previous history of colorectal adenoma took part
in a randomized controlled trial with folic acid and aspirin as the intervention
and recurrence of colorectal adenoma as the primary outcome. Folic acid
supplementation resulted in an increased risk of three or more colorectal
adenomas and non-significant increases in overall colorectal adenoma
recurrence.188
A thorough biochemical knowledge of one-carbon metabolism is necessary in
order to understand the relation to CRC risk. The interdependence and
interaction between the many factors involved calls for improved methods to
validate the previous results based on single factors in relation to risk, and to
gain new insights. Studies carried out in populations with a high portion of
total folate intake derived from naturally occurring folates gives the
opportunity to study one-carbon metabolism in a setting where the lower
spectrum of physiological folate status is represented.
22
Aims of the Thesis
This thesis investigates the associations of one-carbon metabolites to CRC risk
in a mature, relatively large population-based cohort. We used prospective
blood samples collected from over 600 case participants and approximately
1200 matched control participants in prospective case-control studies nested
in the NSHDS.
This thesis consists of four papers with the following specific aims:
Paper I
To investigate plasma concentrations of folate, vitamin B12, and
homocysteine in relation to CRC risk.
To stratify the study cohort according to tumor and case participant
characteristics and to investigate further the associations of the
included plasma markers to CRC risk.
Paper II
To investigate plasma concentrations of one-carbon metabolites
predominantly involved in remethylation of methionine from
homocysteine in relation to CRC risk.
To divide the full study cohort according to tumor and case
participant characteristics and to investigate the interactions between
one-carbon metabolism factors in relation to CRC risk.
To calculate marginal absolute risk estimates and risk difference,
expressed as changes in CRC incidence per 100 000 individuals.
Paper III
To investigate different aspects of vitamin B6 status in relation to
CRC risk using three different markers.
To specifically investigate timing to exposure using the long follow-
up from screening to diagnosis.
Paper IV
To investigate the performance of three different ratio-based B-
vitamin markers in relation to CRC risk.
To investigate the ability of the ratios to determine low total B-vitamin
score in both case and control participants.
23
Materials and Methods
Study Population
The studies in this thesis are based on the inhabitants of Västerbotten County.
Västerbotten lies in the northern part of Sweden stretching from the Baltic Sea
to the Norwegian border. Västerbotten is a large county, slightly larger than
Denmark, but sparsely populated, with a total population of 260,000, of
which 120,000 live in the coastal county capital, Umeå.189 Umeå is an
expanding university city with a large proportion of students, so the
populations have a relatively low mean age. The rural western areas of the
county have a low population density and a higher median age.189
Study Cohort
The studies in this thesis are nested in the Northern Sweden Health and
Disease Study (NSHDS). The NSHDS consists of three subcohorts, the
Northern Sweden Monitoring of Trends and Determinants in Cardiovascular
Disease cohort (MONICA), the Västerbotten Intervention Programme cohort
(VIP), and the Mammography Screening Project cohort (MSP).190 The VIP and
MSP cohorts are described in further detail below. Because the MONICA
cohort is mainly used for cardiovascular disease research, it is not part of the
investigations in this thesis. As the NSHDS is a population-based study
cohort, the CRC incidence and mortality in the cohort is essentially identical
as in Sweden as a whole (Figure 1).
The Västerbotten Intervention Programme
The VIP was initiated in 1985 as the Norsjö Project, a response to the high
mortality in cardiovascular disease in Västerbotten County and particularly in
the Norsjö Municipality. This population-based cohort aimed to include all
residents in Västerbotten County. Residents are invited to take part in a health
examination when turning 30 (1985-1996 only), 40, 50, and 60 years old. The
health examination consists of a medical examination and laboratory tests,
collection of a fasting blood sample, and an extensive diet and lifestyle
questionnaire.190, 191
The participation rate in the VIP is approximately 60% of the total invited
population and in the most recent survey had increased to 66% between 2005
and 2010.191 A comparison of the rate of cancer incidence in the VIP cohort to
the expected incidence supports the population-based nature of the cohort,
and selection bias has been observed to be low.190 As of December 31, 2009,
24
the stop date for case identification for the studies in this thesis, the VIP
included 85,877 individuals and 115,147 blood samples.
The Mammography Screening Project
From 1996 until 2006, all women in Västerbotten County between the ages of
50 and 70 years old who underwent mammography screening were invited to
participate in the MSP. In addition to the screening, the women were invited
to donate a blood sample for future research and to complete a questionnaire
on reproductive history, diet, and lifestyle factors.190, 192 The participation in
the screening program was 85% and participation both the screening and the
donation of a blood sample was 33%.192 The included women had a higher
incidence of breast cancer than expected one year after mammography, but
this was compensated by a lower incidence in the following years. These
women also had a slightly lower risk of lung cancer.190 At its conclusion, the
MSP included 54,401 blood samples from 28,802 women.
Blood Sampling and Analysis
In the NSHDS, plasma from venous blood samples is collected in sample tubes
that are separated into buffy coat, plasma, and erythrocyte fractions. The
plasma samples are then aliquoted and frozen within one hour of collection
either at -80°C or -20°C for up to one week before transfer to a -80°C freezer
at a central location at Umeå University Hospital for long-term storage. The
VIP blood samples are fasting samples collected in the morning. The MSP
blood samples were collected throughout the day, and are almost exclusively
non-fasting blood samples.
The concentration of one-carbon metabolites and polymorphisms involved in
one-carbon metabolism were analyzed in EDTA plasma at Bevital AS (Bergen,
Norway, www.bevital.no). Plasma concentrations of tHcy and cysteine were
measured using an isotope dilution gas chromatography-mass spectrometry
method (between-day CV: 2–8%).193 Vitamin B12 (cobalamin) and folate
concentrations were determined with a microbiological method using
Lactobacillus leichmannii and Lactobacillus casei, respectively, which were
adapted to a microtiter plate format and carried out by a robotic workstation
(between-day CV: 5%).194, 195 Plasma concentrations of methionine,
methionine sulfoxide, choline, betaine, dimethylglycine, creatinine,
neopterin, kynurenines (kynurenin, HK, and XA), vitamin B2 (riboflavin),
symmetric dimethylarginine (SDMA), and vitamin B6-species (PLP, PA, and
PL) were measured with liquid chromatography–mass spectrometry methods
(between-day coefficient of variation (CV): 3–13%).196, 197 Single nucleotide
polymorphisms (SNPs) were determined using MALDI-TOF mass
25
spectrometry (average error rate of <0.1% estimated by analysis of sample
duplicates).198 The samples were analyzed in matched case-control sets with
the position of the case in each set assigned randomly. The laboratory staff
and investigators were blinded to case and control status.
Variables
Plasma concentrations of metabolites and SNPs involved in or related to one-
carbon metabolism were analyzed in the articles presented in this thesis. The
metabolites and inflammatory markers were folate, vitamin B6-species (PLP,
PA, and PL), vitamin B2 (riboflavin), vitamin B12 (cobalamin), homocysteine
(tHcy), cysteine, methionine, methionine sulfoxide, choline, betaine,
dimethylglycine, kynurenine, and sarcosine. In Paper III, two ratios were
calculated from metabolite concentrations: HK:XA (HK/XA) and PAr
(PA/(PL+PLP)). In Paper IV, a B-vitamin score and three homocysteine ratios
were calculated from plasma concentrations. The B-vitamin score was
calculated as the sum of the standardized concentrations of riboflavin, PLP,
folate, cobalamin, and betaine. Standardization of the plasma concentrations
was performed by subtracting the mean and dividing by the standard
deviation. Furthermore, the ratio-based B-vitamin markers hcy:cys
(tHcy/cysteine), hcy:cre (tHcy/creatinine), and hcy:cys:cre
(tHcy/cysteine/creatinine) were also calculated. Lastly, the sensitive kidney
function marker SDMA, and the inflammatory markers neopterin and
kynurenine-to-tryptophan-ratio (KTR, kynurenine/tryptophan) were
calculated.
Included SNPs were MTHFR 677 C>T, MTR 2756 A>G, and BHMT 742 G>A.
Other factors considered were smoking status (current, ex-smoker, and never
smoker), body mass index (BMI) (< 25, 25–30, ≥ 30 kg/m2), alcohol intake
(zero intake, above/below sex-specific median of self-reported grams of
alcohol/day), dietary fiber intake (above/below sex and age-specific (10-year
groups) median of self-reported intake in grams/2000 kcal), recreational- and
occupational physical activity (self-reported on a scale from 1-5, ranging from
1 for sedentary and 5 for highly active), and estimated glomerular filtration
rate (eGFR) calculated by the Cockcroft-Gault formula (based on age, plasma
creatinine levels, sex, and body weight).199
Study Design
As all the included studies have a prospective design, ideally the blood samples
were collected from the case participants when they were still healthy.
However, CRC is a slowly developing disease and could take up to 20 years to
develop from the period from when the tumor first could be detected by
26
screening until diagnosis.46-48 One-carbon metabolism and inflammation may
be involved in the promotion of CRC but not necessarily in the initiation of
CRC.138 The timing of the exposure of folate and other one-carbon metabolites
is therefore important to consider.
There are also several opportunities and pitfalls when assessing associations
between biomarkers of inflammation and cancer. Associations between
inflammatory biomarkers and cancer may be explained by other factors that
increase both biomarker concentrations and the risk of cancer (i.e.,
confounding) or by the tumor itself increasing biomarker concentrations (i.e.,
reverse causality). To address this issue, several procedures has been used: a
sensitivity analysis where samples collected close to diagnosis are censored;200
Mendelian randomization studies where random polymorphisms in genes
associated with increased biomarker concentrations are a proxy for increased
biomarker concentration enabling a randomized trial like set-up;201 and
repeat sampling where biomarker concentrations are followed over time.202
Ideally, studies should have samples collected long before diagnosis.203 In our
studies, we have a long median follow-up time from sampling to diagnosis and
a sufficiently large number of case participants - conditions that allowed us to
stratify the cohort according to follow-up time.
The studies in this thesis are nested in the NSHDS, and the case participants
are matched with control participants taken from the full study cohort
depending on matching criteria, such as sex and age. In biobank research, a
nested case-control design also allows for matching according to when a
sample is collected, effectively matching for storage time.204 Matching for
storage time is important because biological samples may degrade over time.
However, matching on several criteria means that the control participants do
not represent the whole cohort, but rather a cohort with the same distribution
of the matching criteria variables in the case and control participants. For
example, our studies are skewed towards women because of the inclusion of
the all-female MSP cohort. This sex difference is not seen in the NSHDS.
Selection of CRC Case and Control Participants
CRC case participants diagnosed between October 17, 1986, and March 31,
2009, were identified by linkage with the essentially complete Cancer Registry
of Northern Sweden (ICD-10 C18.0 and C18.2–C18.9 for colon and C19.9 and
C20.9 for rectum). All cases and tumor data were verified by a single
pathologist who specialized in gastrointestinal pathology. Patient records
were used to verify the tumor site. Exclusion criteria were previous cancer
diagnosis (other than non-melanoma skin cancer), primary tumor location
outside of the colorectum, prioritization to other studies, insufficient volume
27
of stored plasma sample, serious infectious disease (for the protection of the
laboratory staff), or no matching control available. For each case participant
in our studies, two controls were selected, matched by age (±6 months), sex,
cohort, fasting status, and year of blood sampling and data collection. The
exclusion criteria for the control participants were the same as for the case
participants. Furthermore, five case-control sets had a high proportion of
methionine sulfoxide, indicating the full set had been thawed for a prolonged
time. See Paper I for further discussion.205 After exclusions (81 cases and 41
controls), 613 cases and 1190 controls were included for data analyses, with
the exception of Paper I where 331 case and 662 control participants were
included. In Paper I the analyzed metabolites have previously been
investigated in two studies nested within the same cohort.132, 174 The selection
of case and control participants is depicted in Figure 8.
Figure 8. Cases and control participants in Papers II, III, and IV. In Paper I, 16 cases were
excluded (11 cases because of tumor outside of the colorectum or not verifiable and five cases and
their matched controls because of high methionine sulfoxide). a excluding non-melanoma skin
cancer. b high methionine sulfoxide indicates degradation of samples.
Statistical Analyses
We used Mann-Whitney U and Chi-square tests to test for differences in
baseline characteristics between the case and control participants. Linear
regression was used to evaluate associations between exposure variables and
baseline characteristics. Multivariable-adjusted odds ratios (ORs) for CRC
risk were calculated using conditional logistic regression models stratified on
the matched case sets. The exposure variables were divided into ordinal
variables using percentile cut-offs (tertiles or quartiles) depending on the
28
distribution in the control participants. In Paper III, we also divided the study
cohort as vitamin B6 deficient or sufficient depending on plasma PLP
concentrations. Trends across the ordinal variables were tested by modeling
the variables as continuous variables. Interactions between exposure variables
were tested with likelihood ratio tests comparing an interaction model to an
additive model. We estimated subgroup-specific ORs for CRC depending on
tumor and case participant characteristics (with controls given the same value
as their index case to retain matching) by conditional logistic regression. In
Paper I, heterogeneity between results of the subgroup analysis was
investigated by a Chi-squared based test. In Paper III and IV, heterogeneity
was tested with likelihood ratio tests, comparing a model where risk
associations varied across tumor subtypes to a model where all associations
were held constant.206
In order to improve the interpretability and validity of the analyses, in Paper
II and III we also computed absolute risk estimates in the form of marginal
risk differences (presented as estimated changes in the number of cases per
100 000 cohort participants). We used a weighted maximum likelihood
estimator using cumulative incidence data from the whole NSHDS and within
groups defined by age, cohort, sex, and year of sampling.207, 208 Cross-
validation of 95% CIs for the risk differences were calculated by normal
approximation based on 1000 nonparametric bootstrap replications
resampled from within the matched case set.209
In Paper III, we evaluated dose-response and potential nonlinear relations by
modeling continuous log-transformed plasma biomarkers in relation to CRC
risk with the use of restricted cubic splines in logistic regression models (with
5 knots at the 5th, 25th, 50th, 75th, and 95th percentiles of the plasma
biomarker distributions). We tested for nonlinearity with the Wald test, which
compared the model with spline terms to a linear model.
In Paper IV, the influence of the B-vitamin score and other variables on tHcy
and B-vitamin markers were evaluated in multiple linear regression models.
We also evaluated the contribution of individual B-vitamins by estimating the
same models but with log-transformed standardized B-vitamins included as
separate variables. To compare associations between B-vitamin score and the
ratio-based B-vitamin markers, we evaluated regression-based sensitivity (B-
vitamin marker variance explained by total B-vitamin score), regression-
based specificity (B-vitamin marker variance explained by total B-vitamin
score divided by total explained variance), and performance (defined as
sensitivity multiplied by specificity). Furthermore, we complemented these
analyses with conventional receiver operating characteristics (ROC) analysis
to estimate sensitivity, specificity, and area under the ROC curve (AUC) for
29
the markers in identifying individuals with a B-vitamin score lower than the
5th percentile in controls.
In Paper I, all calculations were made in IBM SPSS Statistics version 21.0
(IBM Corporation). In Papers II-IV, calculations were made in R v.3.3.2 (R
Foundation for Statistical Computing, Vienna, Austria). Statistical tests and
corresponding P values were two-sided and P values <0.05 were considered
statistically significant.
Ethical Approval
Written informed consent was obtained from all study participants at the time
of recruitment to the NSHDS. The study protocol and data handling
procedures were approved by the Umeå University Research Ethics
Committee (Dnr 03-186 and 2015/167-32).
30
Main Results and Discussion
Paper I
Main Results
Paper I includes cases diagnosed between 2003 and 2009 and is a follow-up
study to two previous studies in the same cohort containing cases diagnosed
between 1986 and 2002.132, 174 The previous studies analyzed plasma
concentrations of folate and homocysteine and vitamin B12, respectively. The
results of the follow-up study on the same metabolites are strikingly similar to
the two previous studies.
The study contains participants from both the VIP (85% of the participants)
and the all-female MSP (15% of the participants). This composition is reflected
in the percentage of women in the study (59.2 % women). There were no
statistically significant differences in baseline characteristics between the case
and control participants. The multivariable ORs for CRC according to tertiles
of plasma concentrations of folate, homocysteine, and vitamin B12 are shown
in Figure 9. We found the lowest CRC risk in the first tertile of folate
concentrations, with ORs of 1.62 (95% CI, 1.08-2.42) and 1.42 (95% CI, 0.94-
2.21) for the second and third tertile versus the first tertile of plasma
concentrations of folate, respectively. Plasma concentrations of homocysteine
or vitamin B12 were not significantly associated with CRC risk. We also
stratified the full study cohort into subgroups based on sex, age, follow-up
time between screening and diagnosis, tumor site and stage, and cohort. The
plasma concentrations of folate were positively associated with CRC risk in
most subgroups except for follow-up time above the median of 10.8 years,
lower tumor stage, and rectal cancers. For the second versus the first tertile of
plasma folate concentrations, significant associations were observed in
subgroups based on female sex, age over the median, follow-up time under the
median, tumors with higher stage or localization in the right colon, and the
MSP cohort. Furthermore, for vitamin B12 we observed a decreased risk of
rectal cancer risk for the second and the third tertile versus the first. The third
tertile of vitamin B12 versus the first was also associated with a decreased CRC
risk in men, but not in women.
31
Figure 9. Odds ratios (ORs) of CRC risk for tertiles of plasma concentrations of folate, vitamin
B12 (cobalamin), and homocysteine. Adjusted OR adjusted for BMI, current smoking,
recreational and occupational physical activity, and alcohol intake.
Interpretation
Folate and vitamin B12 are part of one-carbon metabolism, an intracellular
network of enzymatic reactions involving the transfer of methyl groups
(Figure 3).84 The main outputs of one-carbon metabolism are nucleotides for
DNA and the labile methyl groups for the universal methylator SAM. Folate in
different forms acts as a carrier of one-carbon groups and vitamin B12 is an
essential co-factor for MS, an enzyme that takes a methyl group from folate
and remethylates homocysteine to form methionine. Methionine, in turn,
donates the methyl group to form SAM and reverts to homocysteine.84
Therefore, low availability of either vitamin B12 or folate can increase the
concentration of homocysteine. Homocysteine exists only as a byproduct of
one-carbon metabolism and cannot be supplied from dietary sources,99
although plasma concentrations increase after ingestion of a methionine
load.210
Intake of folate and vitamin B12 differ in that folate is often found naturally in
foods that are considered healthy (e.g., fruits, leafy vegetables, and legumes)
and vitamin B12 is found in meat, processed meat, and organ meats.211 In
many countries, the main source of folate is the synthetic form, folic acid,
found in supplements or added to grains to decrease the incidence of NTDs
(Figure 5).118, 119 Sweden has not implemented mandatory fortification of
grains,181 and the folate status in the NSHDS cohort is low.212 This situation
gives us the opportunity to evaluate the lower spectrum of folate
concentrations, whereas in many other populations, folate concentrations are
not easily studied since even the groups with the lowest folate concentrations
are higher than what would be plausibly achievable by consuming naturally
Variable
Folate (tertiles)
1
2
3
Vitamin B12 (tertiles)
1
2
3
Homocysteine (ter tiles)
1
2
3
Cases/controls
94/221
129/221
108/220
117/220
120/221
94/220
112/221
100/221
119/220
Unadjusted OR (95% CI)
Ref
1.55 (1.05–2.28)
1.31 (0.87–1.98)
Ref
1.00 (0.73–1.38)
0.77 (0.55–1.11)
Ref
0.90 (0.65−1.25)
1.07 (0.77−1.48)
Adjusted OR (95% CI)
Ref
1.62 (1.08–2.42)
1.42 (0.94–2.21)
Ref
0.98 (0.73–1.36)
0.74 (0.52–1.07)
Ref
0.87 (0.62−1.24)
1.01 (0.72−1.43)
0.50 0.70 1.0 1.5 2.0 2.5
Odds ratio
32
occurring folates.213 A special case is a study of almost 1000 case participants
nested in the WHI-OS in which screening was performed before mandatory
fortification and diagnosis several years after fortification, essentially
ensuring that the blood samples did not reflect the folic acid exposure the
participants had been subjected to.131 In another study in the same cohort,
plasma concentrations of homocysteine were associated with CRC risk.162 As
folic acid is known to decrease homocysteine concentrations,99 this study
shares the same issues with blood samples not reflecting exposure over time.
Other studies on the associations of homocysteine to CRC risk,160, 161 including
two from the NSHDS,132, 205 did not find any significant associations. Increases
in homocysteine concentrations can depend on factors that have opposing
associations to CRC risk,97, 99, 214 such as the increased risk associated with low
vitamin B6 and methionine status,163 the lower risk associated with the
MTHFR 677TT polymorphism,107, 133 and the lower risk for CRC associated
with low folate status in the NSHDS.132, 205 Therefore, we believe that the null
association between homocysteine and CRC risk in the NSHDS is biologically
plausible.
Many studies on plasma concentrations of folate in relation to CRC have been
published observing no significant association,126, 129-131, 134, 161 with the
exception of one small study where plasma concentrations of folate were
associated with a lower CRC risk.127 However, several studies in populations
with a low folate status and no mandatory folic acid fortification have found
an association between folate and a higher risk of either CRC or colon
cancer.132, 133, 161, 205 As with the studies nested in the NSHDS, Takata et al.
found an association between higher CRC risk in the middle quartile versus
the lowest quartile of plasma folate concentrations in subjects with a follow-
up time closer to diagnosis.130 Lee et al. found an association between lower
plasma folate concentrations and a lower CRC risk in a study of over 600 case
participants in a study on pre-fortification blood samples nested in the Nurses’
Health Study, the Health Professionals Follow-up Study, and the Physicians’
Health Study.133 However, the largest study to date, comprising over 1300 case
participants in the EPIC cohort, found no association with CRC risk.134 In that
study, mean folate concentrations were relatively high, which could be a result
of differences in diet and voluntary folic acid fortification in different parts of
Europe.212
In accordance with our previous study on vitamin B12 in relation to CRC, we
found an association between plasma concentrations of vitamin B12 and a
lower risk for rectal cancer, but not CRC risk. We also observed a linear
decrease in CRC risk in men, but not in women. In a study comprising over
1300 case participants nested in the EPIC cohort, Eussen et al. observed no
association between plasma concentrations of vitamin B12 and CRC
33
irrespective of sex or subsite in the colon. The reason for the discrepancy
between the studies is currently unknown. The southern and central European
parts of the EPIC cohort have considerably higher plasma concentrations of
folate than in the northern region, where the NSHDS is situated.212 In the
NSHDS and other populations with low folate, the combination of low vitamin
B12 and the overall low folate status could lead to lower activity of methionine
synthase (MS), an enzyme that remethylates homocysteine to methionine for
the universal methylator SAM in a reaction where vitamin B12 is a cofactor
and me-THF donates methyl groups. Lower MS activity leads to aberrations
in DNA and histone methylation, processes that influence tumorigenesis.88 To
some extent, women are protected from the deleterious effect of decreased MS
activity since the alternative path of remethylation of homocysteine through
BHMT is more active in premenopausal women.147 Furthermore, compared to
women, men are almost twice as likely to develop rectal cancer.7-9
Consequently, the association between vitamin B12 and a lower rectal cancer
risk may partly be explained by the association between plasma
concentrations of vitamin B12 and a lower CRC risk overall found in men.
When examining causation in prospective nested case-control studies, it is
necessary that both the prospective case and control participants are healthy
at baseline. That is, the exposure must precede events related to the
outcome.215 However, this is difficult to attain in slowly developing diseases
and chronic disease such as CRC.215 The time period from when a CRC tumor
first could be detected by screening until diagnosis has been estimated to
between five and 20 years.46-48 Most nested case-control studies on folate have
a median follow-up time from screening to diagnosis of well below ten years.
This short follow-up time indicates that some of the prospective case
participants have established pre-cancerous lesions when their blood sample
is collected. In both our study and in the previously mentioned study by
Takata et al., the associations were significant in samples collected closer to
diagnosis.130, 205 In addition, plasma concentrations of folate are not
associated with the precursor to CRC, colorectal adenoma.135 Consistent with
animal model research, these prospective studies suggest that folic acid
supplementation is protective in the normal mucosa, but may accelerate the
growth of established pre-cancerous lesions.116 A considerable part of the
population will at some point develop CRC (the cumulative incidence of CRC
is over 5% at 80 years of age (Figure 1)),6 and many people with precancerous
lesions would be affected by mandatory folic acid fortification. In comparison,
NTDs are rare, but similar to CRC represents a considerable public health
issue.118, 119 Different measures have been proposed to mitigate this situation.
Public health efforts to increase use of folic acid supplements during
pregnancy have largely been unsuccessful.216 Some alternatives have been
proposed instead of mandatory folic acid fortification, such as packaging folic
34
acid pills with feminine hygiene products.119 One could also argue that a
comprehensive CRC screening program may attenuate a potential increase of
CRC progression from folic acid fortification. In the US where mandatory folic
acid fortification was implemented in 1998, the CRC incidence and mortality
are steadily decreasing, which could be a result of widespread screening with
colonoscopy.55, 217 While the pros and cons of folic acid fortification are still
debated, it is important to stress the importance of folate in preventing NTDs.
Since CRC is rarely present in the young (Figure 1), temporary folic acid
supplementation is unlikely to influence CRC development in this population.
We, the authors of Paper I, strongly recommend folic acid supplementation
for women of childbearing age.
35
Paper II
Main Results
This prospective case-control study nested in the NSHDS Paper II contains all
case participants diagnosed between 1985 and 2009 and their matched
control participants. The selection of case and control participants is depicted
in Figure 8.
The study contains participants from both the VIP (78% of the participants)
and the all-female MSP (22% of the participants). Consequently, the study has
a higher percentage of women than men (59% women). Figure 10 depicts the
multivariable ORs for CRC according to tertiles of plasma concentrations of
betaine, choline, dimethylglycine, sarcosine, and methionine. We found
significant associations between a lower CRC risk and the third versus the first
tertile of betaine and methionine concentrations, with ORs of 0.76 (95% CI,
0.59-0.99) and 0.72 (95% CI, 0.55-0.94), respectively. We calculated absolute
risk difference as incidence per 100 000 individuals for both betaine and
methionine. When comparing the third and the first tertiles of plasma
concentrations, the risk difference for both metabolites was approximately
200 fewer diagnosed CRC cases per 100 000 participants. Plasma
concentrations of choline, dimethylglycine, or sarcosine were not significantly
associated with CRC risk. We also stratified the full study group based on the
median follow-up time between screening and diagnosis. The plasma
concentrations of betaine were inversely linearly associated with CRC trend in
the subgroup above the median follow-up time, whereas for methionine a
similar risk distribution was observed in the subgroup below the median
follow up-time.
We explored biologically plausible interactions between one-carbon factors.
Low folate combined with high methionine status was associated with a low
CRC risk, OR 0.39 (95% CI, 0.24-0.64). Methionine was mainly associated
with a lower CRC risk in study participants with the variant MTR 2756 AG/GG
genotypes.
36
Figure 10. Odds ratios (ORs) and Risk differences (RDs) of CRC risk for plasma concentrations
of choline, betaine, dimethylglycine, sarcosine, and methionine. ORs are unadjusted since no
potential confounder substantially changed the ORs. RDs are adjusted for the matching variables,
BMI, smoking status, occupational physical activity, and plasma folate, riboflavin (vitamin B2),
cobalamin (vitamin B12), and methionine concentrations.
Interpretation
The one-carbon metabolism factors included in Paper II are all involved in the
methionine cycle and the transfer of methyl groups carried by folates or from
the choline oxidation pathway for the universal methylator SAM.84 In the
choline oxidation pathway, BHMT catalyzes the transfer of a methyl group
from betaine to homocysteine to produce methionine and dimethylglycine.
Betaine can be supplied either through diet or through oxidation of choline.
Dimethylglycine, in turn, can be oxidized to sarcosine.84, 86
Methionine in one-carbon metabolism is the product of the vitamin B12
dependent enzyme MS and used for remethylation of S-
adenosylhomocysteine (SAH) to S-adenosylmethionine (SAM),
simultaneously reverting methionine back to homocysteine (Figure 3).84
Excess homocysteine can also be diverted through the PLP-dependent
transsulfuration pathway where homocysteine provides sulfur for cysteine
formation.99 Methyl groups for MS are provided by me-THF, a form of folate
formed by MTHFR that has no other use than donating methyl groups for
MS.84 Consequently, vitamin B12 deficiency decreases the activity of MS and
Variable
Choline (tertiles)
1
2
3
Betaine (tertiles)
1
2
3
Dimethylglycine (tertiles)
1
2
3
Sarcosine (tertiles)
1
2
3
Methionine (tertiles)
1
2
3
Cases/controls (n)
213/397
196/396
197/396
221/397
216/404
169/388
216/397
172/396
218/396
225/400
186/394
202/396
219/397
223/396
171/397
Odds ratios (95% CI)
Ref
0.91 (0.71−1.2)
0.90 (0.69−1.2)
Ref
0.96 (0.75−1.2)
0.76 (0.59−0.99)
Ref
0.80 (0.62−1.0)
1.0 (0.79−1.3)
Ref
0.84 (0.66−1.2)
0.88 (0.67−1.2)
Ref
1.0 (0.79−1.3)
0.72 (0.55−0.94)
Risk differences
(cases/100 000)
Ref
−38 (−220−160)
−39 (−235−162)
Ref
−66 (−251−133)
−215 (−401− −9)
Ref
−185 (−357− −25)
40 (−172−226)
Ref
−102 (−280−62)
38 (−182−239)
Ref
50 (−137−235)
−201 (−372− −26)
0.50 0.70 1.0 1.5 2.0
Odds ratio
37
increases the accumulation of me-THF.165 This phenomenon is known as the
folate trap.165
The metabolites in Paper II have only been analyzed in relation to CRC in two
cohorts, EPIC and WHI-OS. Both are large cohorts consisting of
approximately 1300 and 800 case participants, respectively, and their
matched controls.152, 153 In EPIC, plasma concentrations of betaine, choline,
and methionine were associated with a lower CRC risk, although choline
predominantly in women and betaine only in a subgroup with a low folate
status (similar to the folate status in the NSHDS).153 The WHI-OS is an all-
female cohort in which mandatory folic acid fortification of grains was
introduced after sample collection but before CRC diagnosis, 131 as previously
described. Conceivably, this should not be as important regarding the
metabolites studied in Paper II since no direct influence on metabolite
concentrations takes place. In the WHI-OS, plasma concentrations of betaine
were associated with a lower CRC risk, while choline and dimethylglycine were
not associated with CRC risk.152 Plasma concentrations of sarcosine have not
previously been studied in relation to CRC risk.
Taken together, the investigations in the EPIC, WHI-OS, and NSHDS indicate
weak inverse associations between betaine, choline, and methionine and CRC
risk. However, specific conditions that could influence the associations
include a possible interaction with plasma folate for methionine and choline
only being associated with lower CRC risk in women in EPIC but not at all in
the all-female WHI-OS or in the NSHDS.141, 152, 153
Odds ratios and other relative risk estimates can be difficult to interpret and
do not consider the incidence rate of the disease. As an example, each year in
Sweden approximately 100 000 children are born and 20-25 of these have
NTDs, and the yearly CRC incidence is approximately 6000.5, 181 A
hypothetical added exposure with an estimated risk association of OR 2.0
would result in 6000 more diagnosed cases of CRC and 25 more cases of
NTDs. An absolute risk estimate takes the incidence of the outcome in the
studied population into account. In the NSHDS, the cumulative incidence of
CRC from 1987 to 2009 was approximately 830 per 100 000 participants. We
used the incidence rate in each matching group defined by age, cohort, sex,
and sampling year to calculate the average cumulative incidence in our nested
study cohort (approximately 660 per 100 000 participants). The total risk
difference of 200 fewer diagnosed cases of CRC per 100 000 participants for
high plasma concentrations of betaine and methionine versus low plasma
concentrations represents a considerable number of CRC case participants.
38
One of the more interesting findings of the study is the interaction between
low plasma concentrations of folate and high plasma concentrations of
methionine, resulting in a substantial decrease of CRC risk (>60%). As
previously mentioned, folate is active in both the methionine and the folate
cycle, whereas methionine and betaine are only active in the methionine
cycle.84 Therefore, in the NSHDS where low folate status is associated with a
lower CRC risk,132, 205 the association between the methyl donors methionine
and betaine with respect to lower CRC risk suggests that the output of the
folate cycle, and not the methionine cycle, accounts for the increase in CRC
risk. The folate cycle involves the synthesis of nucleotides for DNA,84 essential
for cancer cells,104 and folate deficiency causes misincorporation of uracil into
DNA,103 similar to the commonly used chemotherapy 5-FU.38 The interaction
with methionine also suggests a protective association to CRC risk only when
the methionine cycle is functioning well.
Tumor cells require methionine added to cell media to proliferate, but normal
cells can proliferate with only the precursor homocysteine, a phenomenon
known as the methionine dependency of tumor cells.114 These in vitro results
have been replicated in vivo in animal models fed a low methionine diet.114, 115
Accordingly, methionine restriction has potential as a dietary intervention in
cancer patients.115 However, dietary methionine restriction does not
consistently reduce plasma concentrations of methionine, and degradation of
the surrounding tumor stroma can supply sufficient methionine for the
tumor.115 We found that higher plasma concentrations of methionine were
more strongly associated with lower CRC risk in patient samples collected
closer to diagnosis. This finding indicates that low methionine status did not
inhibit tumor growth and that tumor cell methionine dependency did not
influence the CRC risk in our cohort.
39
Paper III
Main Results
The study contains the same participants as in Paper II and case and control
participant selection is depicted in Figure 8. The multivariable ORs for CRC
according to quartiles of HK:XA, PAr, and PLP are shown in Figure 11. We
found significant associations between a higher CRC risk and the fourth versus
the first quartile of HK:XA and PAr, with ORs of 1.48 (95% CI, 1.08-2.02) and
1.50 (95% CI, 1.10-2.04), respectively. The third versus the first quartile of
plasma concentrations of PLP and sufficient versus deficient vitamin B6 status
was significantly associated with CRC risk, with ORs of 0.60 (95% CI, 0.60-
0.81) and 0.55 (0.37-0.81), respectively. We stratified the full study group into
three groups according to follow-up time between screening and diagnosis.
Higher plasma concentrations of HK:XA and PAr were associated with CRC
risk more strongly in the subgroups with less than 10.5 years of follow-up time,
but no significant association was observed in the subgroup with 10.5 years or
more of follow-up time. Plasma concentrations of PLP were not significantly
influenced by follow-up time.
Figure 11. Odds ratios (ORs) and Risk differences (RDs) of CRC risk for plasma concentrations
of PLP in quartiles and as Vitamin B6 deficiency/sufficiency (cut-off 20 nmol/L), and for quartiles
of PAr and HK:XA. ORs and RDs are adjusted for kidney function, BMI, smoking status, alcohol
intake, dietary fiber intake, recreational and occupational physical activity, and plasma folate,
riboflavin (vitamin B2), and cobalamin (vitamin B12) concentrations. RDs are further adjusted
for the matching variables.
Variable
PLP (quartiles)
1
2
3
4
Vitamin B6 deficient
Vitamin B6 sufficient
PAr (quartiles)
1
2
3
4
HK:XA (quartiles)
1
2
3
4
Cases/controls (n)
190/297
146/296
119/296
153/297
64/78
544/1108
126/152
152/296
155/296
173/296
134/295
148/295
139/295
182/296
Odds ratios (95% CI)
Ref
0.79 (0.58−1.06
0.60 (0.44−0.81)
0.80 (0.59−1.09)
Ref
0.55 (0.37−0.81)
Ref
1.25 (0.92−1.69)
1.29 (0.96−1.74)
1.50 (1.10−2.04)
Ref
1.13 (0.83−1.53)
1.08 (0.79−1.49)
1.48 (1.08−2.02)
Risk differences
cases/100 000
Ref
−139 (−350−84)
−378 (−527−−110)
−115 (−341−120)
Ref
−476 (−859−−17)
Ref
170 (−24−353)
189 (6−356)
273 (58−477)
Ref
64 (−119−242)
37 (−167−233)
211 (−6−412)
0.50 0.70 1.0 1.5 2.0
Odds ratio
40
Interpretation
PLP, the active form of vitamin B6, is the most commonly used marker of
vitamin B6 status and is used to determine vitamin B6 deficiency.164, 211 PLP is
involved in several processes that influence cancer risk and development, such
as angiogenesis, inflammation, cell proliferation, and one-carbon
metabolism. In one-carbon metabolism, PLP is a co-factor for SHMT, which
regulates the influx of labile methyl groups into the folate cycle, and for CBS
and CSE, which regulate the degradation of homocysteine through the
transsulfuration pathway.
In studies on plasma concentrations of PLP in relation to CRC risk, high PLP
concentrations are consistently associated with a lower risk.163, 168 In contrast,
studies on dietary intake of vitamin B6 have found no significant association
with CRC risk.167, 169 The discordant results between dietary intake estimates
and plasma concentration measurements of vitamin B6 could possibly be
explained by confounding as ingestion of high amounts of vitamin B6 may be
associated with a generally healthier lifestyle.163 Another possible explanation
could be the inverse association between systemic inflammation – a risk factor
for CRC 203, 218 – and plasma PLP.163 219
Low plasma concentrations of PLP are associated with several inflammatory
disease conditions, including cancer, IBD, cardiovascular disease, rheumatoid
arthritis, and type II diabetes.220 Furthermore, inflammatory biomarkers are
inversely associated with plasma PLP status.220 In IBD patients, PLP
concentrations are lower in patients with active disease than in patients with
quiescent disease, an observation that could suggest that dietary changes do
not decrease PLP concentrations.221 The association between inflammatory
activity and lower plasma concentrations of PLP indicates that either
inflammation lowers plasma concentrations of PLP or vitamin B6 protects
against inflammation and inflammatory conditions. However, in an
intervention study, dietary restriction of vitamin B6 did not increase the
inflammatory biomarker CRP and in a study of patients with stable angina
pectoris, inverse associations between plasma concentrations of PLP and
inflammatory biomarkers were maintained even after vitamin B6
supplementation.222 In addition, in the population-based NHANES cohort,
vitamin B6 deficiency (PLP <20 nmol/L) was inversely associated with CRP
independent of vitamin B6 intake.223 Taken together, these observations
indicate that inflammatory processes influence PLP concentrations.
Mechanistically, lower PLP concentrations in plasma are a result of the
redistribution of PLP into tissue with inflammation and an increase in PLP
catabolism of PA during inflammation and oxidative stress.220
41
We examined three estimates of vitamin B6 status to account for
inflammatory activity and to allow for a more comprehensive assessment than
relying on PLP alone. PAr is a ratio of the catabolite and the sum of the active
and transport form of vitamin B6. PAr is less influenced than PLP by common
confounders in studies on plasma biomarkers, such as kidney status,
supplementation, and smoking.177 HK:XA is the ratio of plasma
concentrations of the substrate (HK) and product (XA) pair of the vitamin B6
dependent enzyme KAT. HK:XA is negatively associated with plasma
concentrations of PLP and is a sensitive marker of vitamin B6 deficiency.
Compared to HK, it is not as influenced by inflammation, smoking, kidney
status, and BMI.178
We found that vitamin B6 status was associated with CRC risk for all three
markers. PLP deficiency was strongly associated with a higher CRC risk.
Similarly, HK:XA – the inverse functional vitamin B6 status marker – was
associated with a linear increase in CRC risk. PAr, the marker of vitamin B6
flux and catabolism during inflammatory activity, was associated with a
higher CRC risk. The results for PAr resemble those of other inflammatory
markers, such as CRP and neopterin, and to those observed for PAr in relation
to CRC risk in the HUSK cohort. We stratified the full study cohort according
to follow-up time between screening and diagnosis and observed that PAr was
only associated with CRC risk in study participants with a follow-up time
between screening and diagnosis lower than 10.5 years, suggesting a role in
tumor progression rather than initiation. HK:XA is a sensitive marker of
vitamin B6 deficiency, and accordingly replicates the associations observed
for vitamin B6 deficiency defined by PLP. We conclude that inflammation
needs to be considered when studying plasma concentrations of vitamin B6
status, particularly in diseases with an inflammatory aspect.
42
Paper IV
Main Results
The study contains the same participants as in Paper II and III and case and
control participant selection are depicted in Figure 8. We validated the
sensitivity and specificity of the functional B-vitamin markers total
homocysteine (tHcy), and the ratio-based B-vitamin markers hcy:cys, hcy:cre,
and hcy:cys:cre as markers of total B-vitamin status (represented by a
summary score comprising Z-standardized plasma concentrations of betaine,
cobalamin, folate, PLP, and riboflavin). In both case and healthy control
participants, the ratio-based B-vitamin markers outperformed tHcy (Table 1).
The ratio-based B-vitamin markers and total B-vitamin status had similar
associations to CRC risk: approximately a 25% risk decrease associated with
estimated higher B-vitamin status versus low B-vitamin status.
Table 1 Performance of total B-vitamin score in predicting plasma tHcy, Hcy:Cys, Hcy:Cre, and
Hcy:Cys:Cre status in CRC cases and controls.
tHcy Hcy:Cys Hcy:Cre Hcy:Cys:Cre
Cases Controls Cases Controls Cases Controls Cases Controls
Sensitivity, %* 16.7 16.6 26 23.5 23.1 21.5 31.6 27.0
Specificity, %¤ 55.6 55.9 75.1 72.7 76.1 73.7 89.4 83.3
Performance§ 9.3 9.3 19.5 17.1 17.6 15.8 28.2 22.5
Performance ratio 1 (ref) 1 (ref) 2.1 1.8 1.9 1.7 3.0 2.4
* Defined as variance explained by B-vitamin score.
¤ Defined as variance explained by B-vitamin score divided by total variance explained by the model. § Sensitivity multiplied by specificity divided by 100.
Abbreviations: CRC, colorectal cancer; tHcy, total homocysteine; ref, reference category.
Interpretation
Homocysteine is a precursor for both methionine and cysteine synthesis in the
methionine cycle and transsulfuration pathway, respectively.84, 99 Methionine
provides labile methyl groups for SAM synthesis, and creatine synthesis
consumes a major portion of SAM-provided methyl groups in humans.224
Creatine is non-enzymatically degraded to creatinine at a rate of
approximately 2% per day and secreted into the urine.225, 226 The ratio of
homocysteine to creatinine (Hcy:Cre) is thus an estimate of the flux of
homocysteine to creatinine, and disturbances in the involved B-vitamins
result in an increase in Hcy:Cre. This includes folate, riboflavin (co-factor for
MTHFR when transforming m-THF to me-THF), cobalamin (co-factor for MS
when transforming me-THF and homocysteine to methionine), and betaine
(co-factor for BHMT when transforming betaine and homocysteine to
43
methionine). In the transsulfuration pathway, homocysteine provides sulfur
in a series of PLP-dependent enzymatic reactions for cysteine,84, 99 and
consequently PLP deficiency results in increases in the ratio of homocysteine
to cysteine (Hcy:Cys) (Figure 3).
In clinical settings, plasma homocysteine is a marker of impaired B-vitamin
status,157 but is not convincingly associated with CRC risk in prospective case-
control studies.132, 160-162, 205 This result is surprising, considering the evidence
of several one-carbon metabolites, including PLP,163 betaine,153, 214 and
riboflavin,170 being inversely associated with both homocysteine and CRC
risk.157 Furthermore, the MTHFR 677TT polymorphism is associated with a
lower CRC risk and higher plasma concentrations of homocysteine.107 The
similar risk distribution for the ratio-based B-vitamin markers and total B-
vitamin score are consistent with the ratio-based B-vitamin markers ability to
correctly estimate total B-vitamin score, and the previously reported null
association between tHcy and CRC risk is also in line with the lower
performance of tHcy as a marker of total B-vitamin status.
While Hcy:Cys:Cre was observed to have the highest sensitivity and specificity
as a functional marker of total B-vitamin status, both Hcy:Cys and Hcy:Cre
performed markedly better than tHcy. Hcy:Cre is particularly interesting as it
consists of two markers that are already routinely analyzed in a clinical setting
(tHcy and creatinine). Applying a simple algorithm is a cheap and simple way
of drastically increasing sensitivity and specificity when assessing total B-
vitamin status. The ratio-based B-vitamin markers performed equally well in
participants who later developed CRC and in healthy control participants,
further increasing their usefulness. Future research relating the ratio-based
B-vitamin markers to disease risk, particularly in anemia, could potentially
reveal an easy and cost-efficient tool for doctors to increase specificity and
sensitivity in disease diagnosis.
44
Conclusions
Paper I
Low plasma folate concentrations are associated with a lower CRC
risk, whereas plasma concentrations of vitamin B12 and
homocysteine are not associated with CRC risk. The association
between folate and CRC are predominantly observed in case
participants with samples collected closer to diagnosis.
Plasma concentrations of vitamin B12 are associated with lower rectal
cancer risk.
Paper II
Plasma concentrations of betaine and methionine are associated with
a lower CRC risk. Plasma concentrations of choline, dimethylglycine,
and sarcosine are not associated with CRC risk.
The estimated marginal risk difference for high plasma
concentrations versus low plasma concentrations of betaine and
methionine were approximately 200 CRC fewer cases per 100 000
individuals.
In interaction analyses, the lower CRC risk associated with high
plasma concentrations of methionine is modified by folate, and the
combination of low folate and high methionine status is associated
with a substantial decrease in CRC risk.
Paper III
Sufficient versus deficient vitamin B6 status as defined by plasma
concentrations of PLP are associated with a lower CRC risk. The
inverse vitamin B6 markers PAr and HK:XA are associated with a
higher CRC risk.
PAr, used to estimate vitamin B6-status in inflammation and
oxidative stress, are predominantly associated with CRC risk in case
participants with samples collected closer to CRC diagnosis.
Paper IV
Total B-vitamin score is associated with a lower CRC risk and the
ratio-based B-vitamin markers have a similar association to CRC risk.
In both case and control participants, the three ratio-based B-vitamin
markers perform better than homocysteine alone at determining
deficient B-vitamin status.
45
Future Considerations
One-carbon metabolism in CRC development has been extensively
investigated. Findings have shown some degree of inconsistency, but balanced
one-carbon metabolism is important for the stability of the genome and
epigenetic regulation and may prevent tumor initiation, whereas an excess of
some components, especially folate, may facilitate the progression of
precancerous lesions.85, 117 CRC develops through distinct pathways resulting
in molecular subtypes differing in clinical characteristics such as response to
treatment.227 These subtypes can be defined by different factors, such as
mutations in the mutually exclusive BRAF and KRAS oncogenes, and/or MSI
and CIMP status.31 Investigating one-carbon metabolism in relation to these
molecularly defined CRC subtypes could give new insights into the
heterogeneous nature of CRC tumorigenesis.
A characteristic of CRC is the variation in outcome depending on tumor
stage,15 but also depending on molecular characteristics such as MSI and
CIMP status, and immune cell infiltration into the tumor.16 Several
components of one-carbon metabolism are related to inflammation, and
relating aspects of one-carbon metabolism to immune cell activity can give
insights not only into tumorigenesis but also patient outcome. A high density
of tumor-infiltrating immune cells is associated with beneficial CRC
prognosis,16 but why the immune system is responding to certain tumors, and
how the immune response is beneficial to patients is not fully understood. A
factor especially important in CRC survival is cancer cachexia - the breakdown
of muscle mass in cancer patients.24 Cachexia is thought to arise from
inflammatory cytokines released by the tumor.21 By combining biobanks at
Umeå University and Västerbotten County Council we can gather information
from pre-diagnostic blood samples, from tumor tissue, and on patient
prognosis. Tumor progression can thus be longitudinally followed and cancer
cachexia can be related to one-carbon metabolism and inflammatory activity
at different stages of disease. This is an interesting future field of research in
both CRC survival and early CRC detection.
46
Acknowledgments
First and foremost, thanks to the participants in the NSHDS cohort. The
strength of our cohort rests in the sheer number of Västerbottningar who
readily gave and continue to give their time, effort, and blood. Thank you.
To my main supervisor Professor Richard Palmqvist. I am very grateful for
having been given the opportunity to do research under your guidance. Thank
you for trusting in my abilities, challenging me, and allowing me room to grow.
To my co-supervisor associate Professor Bethany van Guelpen, for teaching
me scientific thinking and rigor. You are an inspiration to any aspiring
researcher and certainly to me.
To the members of the Colorectal cancer research network, science is a group
effort, and together we can do great things. Thanks to each and every one of
you!
Special thanks to Robin Myte, my closest collaborator, always curious,
creative, hard-working, and enthusiastic. You have a rare combination of
having a deep understanding of statistics and being able to convey your
knowledge to people less naturally gifted in that field. Statistical expertise was
also provided by my co-author associate professor Jenny Häggström, thank
you!
To associate professor Jörn Schneede, who went well beyond the obligations
of a co-author and helped develop my scientific writing and shared his
extensive knowledge. You are my supervisor in all but name.
To Professor Ingegerd Johansson for providing funding for the plasma
analyses and for handling the extensive dietary data collection for the NSHDS.
A mammoth task providing the foundation for the research in this thesis and
for many studies from Umeå University.
To Professor Göran Hallmans, who funded and captained the NSHDS.
Building a biobank from scratch must have seemed like a Sisyphean
undertaking at times. Your perseverance and enthusiasm are unmatched, and
to some degree, contagious.
To my co-authors at Bevital and Bergen University Professor Per Magne
Ueland, Arve Ulvik, and Øivind Midttun. Your expertise has been of utmost
importance in the writing of the articles in this thesis.
47
To the staff at the Department of Biobank Research. Without your support, it
would not be possible to conduct this research.
To the past and present members of the administrative staff at the department
of medical biosciences. Special thanks to Terry Persson who always took her
time to help an easily confused man.
To my good friends at the social corner. For a shared appreciation of good beer
and good company.
To Zülal Keskin, for the beautiful one-carbon metabolism figures. To Işıl
Keskin.
To the Cancer Research Foundation in Northern Sweden and the Swedish
Cancer Society for providing the funding for my research. Special thanks to
the people donating to cancer charities.
Till min släkt och min familj. Margith, Thomas och Kerstin, mina syskon
Johan och Susanna med familjer. Till minnet av Bertil, Gunnar och Irené. Er
för evigt.
48
References
1. Bosman, F.T., World Health Organization. & International Agency for Research on Cancer. WHO classification of tumours of the digestive system, Edn. 4th. (International Agency for Research on Cancer, Lyon; 2010).
2. Ferlay, J. et al. Cancer incidence and mortality worldwide: sources,
methods and major patterns in GLOBOCAN 2012. Int J Cancer 136, E359-386 (2015).
3. Center, M.M., Jemal, A., Smith, R.A. & Ward, E. Worldwide variations in colorectal cancer. CA Cancer J Clin 59, 366-378 (2009).
4. Brenner, H. et al. Progress in colorectal cancer survival in Europe from the late 1980s to the early 21st century: the EUROCARE study. Int J Cancer 131, 1649-1658 (2012).
5. Engholm, G. et al. NORDCAN--a Nordic tool for cancer information,
planning, quality control and research. Acta Oncol 49, 725-736 (2010).
6. Siegel, R.L. et al. Colorectal cancer statistics, 2017. CA Cancer J Clin (2017).
7. Abotchie, P.N., Vernon, S.W. & Du, X.L. Gender differences in colorectal cancer incidence in the United States, 1975-2006. J Womens Health (Larchmt) 21, 393-400 (2012).
8. Cheng, X. et al. Subsite-specific incidence rate and stage of disease in
colorectal cancer by race, gender, and age group in the United States, 1992-1997. Cancer 92, 2547-2554 (2001).
9. Gao, R.N., Neutel, C.I. & Wai, E. Gender differences in colorectal cancer incidence, mortality, hospitalizations and surgical procedures in Canada. J Public Health (Oxf) 30, 194-201 (2008).
10. Murphy, G. et al. Sex disparities in colorectal cancer incidence by anatomic subsite, race and age. Int J Cancer 128, 1668-1675 (2011).
11. Radkiewicz, C., Johansson, A.L.V., Dickman, P.W., Lambe, M. &
Edgren, G. Sex differences in cancer risk and survival: A Swedish cohort study. Eur J Cancer 84, 130-140 (2017).
49
12. Sankaranarayanan, R. et al. Cancer survival in Africa, Asia, and Central America: a population-based study. Lancet Oncol 11, 165-173 (2010).
13. DeSantis, C.E. et al. Cancer treatment and survivorship statistics,
2014. CA Cancer J Clin 64, 252-271 (2014).
14. Brierley, J., Gospodarowicz, M.K. & Wittekind, C. TNM classification of malignant tumours, Edn. Eighth edition. (John Wiley & Sons, Inc., Chichester, West Sussex, UK ; Hoboken, NJ; 2017).
15. O'Connell, J.B., Maggard, M.A. & Ko, C.Y. Colon cancer survival rates with the new American Joint Committee on Cancer sixth edition staging. J Natl Cancer Inst 96, 1420-1425 (2004).
16. Dahlin, A.M. et al. Colorectal cancer prognosis depends on T-cell
infiltration and molecular characteristics of the tumor. Mod Pathol 24, 671-682 (2011).
17. Popat, S., Hubner, R. & Houlston, R.S. Systematic review of microsatellite instability and colorectal cancer prognosis. J Clin Oncol 23, 609-618 (2005).
18. Samowitz, W.S. et al. Poor survival associated with the BRAF V600E mutation in microsatellite-stable colon cancers. Cancer Res 65, 6063-6069 (2005).
19. Sinicrope, F.A. et al. Analysis of Molecular Markers by Anatomic
Tumor Site in Stage III Colon Carcinomas from Adjuvant Chemotherapy Trial NCCTG N0147 (Alliance). Clin Cancer Res 21, 5294-5304 (2015).
20. Liao, X. et al. Aspirin use, tumor PIK3CA mutation, and colorectal-cancer survival. N Engl J Med 367, 1596-1606 (2012).
21. Porporato, P.E. Understanding cachexia as a cancer metabolism syndrome. Oncogenesis 5, e200 (2016).
22. Fearon, K.C., Glass, D.J. & Guttridge, D.C. Cancer cachexia:
mediators, signaling, and metabolic pathways. Cell Metab 16, 153-166 (2012).
23. Lee, J., Meyerhardt, J.A., Giovannucci, E. & Jeon, J.Y. Association between body mass index and prognosis of colorectal cancer: a meta-analysis of prospective cohort studies. PLoS One 10, e0120706 (2015).
50
24. Caan, B.J. et al. Explaining the Obesity Paradox: The Association between Body Composition and Colorectal Cancer Survival (C-SCANS Study). Cancer Epidemiol Biomarkers Prev 26, 1008-1015 (2017).
25. Walter, V., Jansen, L., Hoffmeister, M. & Brenner, H. Smoking and
survival of colorectal cancer patients: systematic review and meta-analysis. Ann Oncol 25, 1517-1525 (2014).
26. Fearon, E.R. & Vogelstein, B. A genetic model for colorectal tumorigenesis. Cell 61, 759-767 (1990).
27. Jass, J.R. Classification of colorectal cancer based on correlation of clinical, morphological and molecular features. Histopathology 50, 113-130 (2007).
28. Issa, J.P. CpG island methylator phenotype in cancer. Nat Rev Cancer
4, 988-993 (2004).
29. Toyota, M. et al. CpG island methylator phenotype in colorectal cancer. Proc Natl Acad Sci U S A 96, 8681-8686 (1999).
30. Parsons, M.T., Buchanan, D.D., Thompson, B., Young, J.P. & Spurdle, A.B. Correlation of tumour BRAF mutations and MLH1 methylation with germline mismatch repair (MMR) gene mutation status: a literature review assessing utility of tumour features for MMR variant classification. J Med Genet 49, 151-157 (2012).
31. Guinney, J. et al. The consensus molecular subtypes of colorectal
cancer. Nat Med 21, 1350-1356 (2015).
32. Al-Sukhni, E. et al. Diagnostic accuracy of MRI for assessment of T category, lymph node metastases, and circumferential resection margin involvement in patients with rectal cancer: a systematic review and meta-analysis. Ann Surg Oncol 19, 2212-2223 (2012).
33. Kaur, H. et al. MR imaging for preoperative evaluation of primary rectal cancer: practical considerations. Radiographics 32, 389-409 (2012).
34. van Gijn, W. et al. Preoperative radiotherapy combined with total
mesorectal excision for resectable rectal cancer: 12-year follow-up of the multicentre, randomised controlled TME trial. Lancet Oncol 12, 575-582 (2011).
51
35. Heald, R.J. & Ryall, R.D. Recurrence and survival after total mesorectal excision for rectal cancer. Lancet 1, 1479-1482 (1986).
36. Gill, S. et al. Pooled analysis of fluorouracil-based adjuvant therapy
for stage II and III colon cancer: who benefits and by how much? J Clin Oncol 22, 1797-1806 (2004).
37. Gustavsson, B. et al. A review of the evolution of systemic chemotherapy in the management of colorectal cancer. Clin Colorectal Cancer 14, 1-10 (2015).
38. Longley, D.B., Harkin, D.P. & Johnston, P.G. 5-fluorouracil: mechanisms of action and clinical strategies. Nat Rev Cancer 3, 330-338 (2003).
39. Thirion, P. et al. Modulation of fluorouracil by leucovorin in patients
with advanced colorectal cancer: an updated meta-analysis. J Clin Oncol 22, 3766-3775 (2004).
40. Qiu, M., Hu, J., Yang, D., Cosgrove, D.P. & Xu, R. Pattern of distant metastases in colorectal cancer: a SEER based study. Oncotarget 6, 38658-38666 (2015).
41. Fakih, M.G. Metastatic colorectal cancer: current state and future directions. J Clin Oncol 33, 1809-1824 (2015).
42. Hsu, H.C. et al. Mutations of KRAS/NRAS/BRAF predict cetuximab
resistance in metastatic colorectal cancer patients. Oncotarget 7, 22257-22270 (2016).
43. De Roock, W. et al. Effects of KRAS, BRAF, NRAS, and PIK3CA mutations on the efficacy of cetuximab plus chemotherapy in chemotherapy-refractory metastatic colorectal cancer: a retrospective consortium analysis. Lancet Oncol 11, 753-762 (2010).
44. Lievre, A. et al. KRAS mutations as an independent prognostic factor in patients with advanced colorectal cancer treated with cetuximab. J Clin Oncol 26, 374-379 (2008).
45. Lievre, A. et al. KRAS mutation status is predictive of response to
cetuximab therapy in colorectal cancer. Cancer Res 66, 3992-3995 (2006).
46. Brenner, H., Altenhofen, L., Katalinic, A., Lansdorp-Vogelaar, I. & Hoffmeister, M. Sojourn time of preclinical colorectal cancer by sex
52
and age: estimates from the German national screening colonoscopy database. Am J Epidemiol 174, 1140-1146 (2011).
47. Brenner, H., Altenhofen, L., Stock, C. & Hoffmeister, M. Natural
history of colorectal adenomas: birth cohort analysis among 3.6 million participants of screening colonoscopy. Cancer Epidemiol Biomarkers Prev 22, 1043-1051 (2013).
48. Brenner, H., Stock, C. & Hoffmeister, M. Colorectal cancer screening: the time to act is now. BMC Med 13, 262 (2015).
49. Hewitson, P., Glasziou, P., Watson, E., Towler, B. & Irwig, L. Cochrane systematic review of colorectal cancer screening using the fecal occult blood test (hemoccult): an update. Am J Gastroenterol 103, 1541-1549 (2008).
50. Elmunzer, B.J. et al. Effect of flexible sigmoidoscopy-based screening
on incidence and mortality of colorectal cancer: a systematic review and meta-analysis of randomized controlled trials. PLoS Med 9, e1001352 (2012).
51. Holme, O. et al. Effect of flexible sigmoidoscopy screening on colorectal cancer incidence and mortality: a randomized clinical trial. JAMA 312, 606-615 (2014).
52. Holme, O. et al. Effectiveness of flexible sigmoidoscopy screening in men and women and different age groups: pooled analysis of randomised trials. BMJ 356, i6673 (2017).
53. Brenner, H., Chang-Claude, J., Seiler, C.M., Rickert, A. & Hoffmeister,
M. Protection from colorectal cancer after colonoscopy: a population-based, case-control study. Ann Intern Med 154, 22-30 (2011).
54. Zauber, A.G. et al. Colonoscopic polypectomy and long-term prevention of colorectal-cancer deaths. N Engl J Med 366, 687-696 (2012).
55. Chen, C., Lacke, E., Stock, C., Hoffmeister, M. & Brenner, H. Colonoscopy and sigmoidoscopy use among older adults in different countries: A systematic review. Prev Med 103, 33-42 (2017).
56. Blom, J., Kilpelainen, S., Hultcrantz, R. & Tornberg, S. Five-year
experience of organized colorectal cancer screening in a Swedish population - increased compliance with age, female gender, and subsequent screening round. J Med Screen 21, 144-150 (2014).
53
57. Fritzell, K., Stake Nilsson, K., Jervaeus, A., Hultcrantz, R. & Wengstrom, Y. The importance of people's values and preferences for colorectal cancer screening participation. Eur J Public Health (2017).
58. Schreuders, E.H. et al. Colorectal cancer screening: a global overview
of existing programmes. Gut 64, 1637-1649 (2015).
59. Gillberg, A., Ericsson, E., Granstrom, F. & Olsson, L.I. A population-based audit of the clinical use of faecal occult blood testing in primary care for colorectal cancer. Colorectal Dis 14, e539-546 (2012).
60. Vieira, A.R. et al. Foods and beverages and colorectal cancer risk: a systematic review and meta-analysis of cohort studies, an update of the evidence of the WCRF-AICR Continuous Update Project. Ann Oncol (2017).
61. Jess, T., Rungoe, C. & Peyrin-Biroulet, L. Risk of colorectal cancer in
patients with ulcerative colitis: a meta-analysis of population-based cohort studies. Clin Gastroenterol Hepatol 10, 639-645 (2012).
62. Adami, H.O. et al. The continuing uncertainty about cancer risk in inflammatory bowel disease. Gut 65, 889-893 (2016).
63. Laukoetter, M.G. et al. Intestinal cancer risk in Crohn's disease: a meta-analysis. J Gastrointest Surg 15, 576-583 (2011).
64. Liang, P.S., Chen, T.Y. & Giovannucci, E. Cigarette smoking and
colorectal cancer incidence and mortality: systematic review and meta-analysis. Int J Cancer 124, 2406-2415 (2009).
65. Jiang, Y. et al. Diabetes mellitus and incidence and mortality of colorectal cancer: a systematic review and meta-analysis of cohort studies. Eur J Epidemiol 26, 863-876 (2011).
66. Larsson, S.C. & Wolk, A. Obesity and colon and rectal cancer risk: a meta-analysis of prospective studies. Am J Clin Nutr 86, 556-565 (2007).
67. Ma, Y. et al. Obesity and risk of colorectal cancer: a systematic review
of prospective studies. PLoS One 8, e53916 (2013).
68. Thrift, A.P. et al. Mendelian Randomization Study of Body Mass Index and Colorectal Cancer Risk. Cancer Epidemiol Biomarkers Prev 24, 1024-1031 (2015).
54
69. Gao, C. et al. Mendelian randomization study of adiposity-related traits and risk of breast, ovarian, prostate, lung and colorectal cancer. Int J Epidemiol 45, 896-908 (2016).
70. Jarvis, D. et al. Mendelian randomisation analysis strongly implicates
adiposity with risk of developing colorectal cancer. Br J Cancer 115, 266-272 (2016).
71. Alexander, D.D., Weed, D.L., Miller, P.E. & Mohamed, M.A. Red Meat and Colorectal Cancer: A Quantitative Update on the State of the Epidemiologic Science. J Am Coll Nutr 34, 521-543 (2015).
72. Abid, Z., Cross, A.J. & Sinha, R. Meat, dairy, and cancer. Am J Clin Nutr 100 Suppl 1, 386S-393S (2014).
73. Wolk, A. Potential health hazards of eating red meat. J Intern Med
281, 106-122 (2017).
74. Lin, K.J., Cheung, W.Y., Lai, J.Y. & Giovannucci, E.L. The effect of estrogen vs. combined estrogen-progestogen therapy on the risk of colorectal cancer. Int J Cancer 130, 419-430 (2012).
75. Dou, R. et al. Vitamin D and colorectal cancer: molecular, epidemiological and clinical evidence. Br J Nutr 115, 1643-1660 (2016).
76. Algra, A.M. & Rothwell, P.M. Effects of regular aspirin on long-term
cancer incidence and metastasis: a systematic comparison of evidence from observational studies versus randomised trials. Lancet Oncol 13, 518-527 (2012).
77. Bosetti, C., Rosato, V., Gallus, S., Cuzick, J. & La Vecchia, C. Aspirin and cancer risk: a quantitative review to 2011. Ann Oncol 23, 1403-1415 (2012).
78. Sandler, R.S. et al. A randomized trial of aspirin to prevent colorectal adenomas in patients with previous colorectal cancer. N Engl J Med 348, 883-890 (2003).
79. Lichtenstein, P. et al. Environmental and heritable factors in the
causation of cancer--analyses of cohorts of twins from Sweden, Denmark, and Finland. N Engl J Med 343, 78-85 (2000).
80. Taylor, D.P., Burt, R.W., Williams, M.S., Haug, P.J. & Cannon-Albright, L.A. Population-based family history-specific risks for
55
colorectal cancer: a constellation approach. Gastroenterology 138, 877-885 (2010).
81. Bishehsari, F., Mahdavinia, M., Vacca, M., Malekzadeh, R. & Mariani-
Costantini, R. Epidemiological transition of colorectal cancer in developing countries: environmental factors, molecular pathways, and opportunities for prevention. World J Gastroenterol 20, 6055-6072 (2014).
82. Durko, L. & Malecka-Panas, E. Lifestyle Modifications and Colorectal Cancer. Curr Colorectal Cancer Rep 10, 45-54 (2014).
83. Dunn, J.E. Cancer epidemiology in populations of the United States--with emphasis on Hawaii and California--and Japan. Cancer Res 35, 3240-3245 (1975).
84. Ducker, G.S. & Rabinowitz, J.D. One-Carbon Metabolism in Health
and Disease. Cell Metab 25, 27-42 (2017).
85. Locasale, J.W. Serine, glycine and one-carbon units: cancer metabolism in full circle. Nat Rev Cancer 13, 572-583 (2013).
86. Ueland, P.M. Choline and betaine in health and disease. J Inherit Metab Dis 34, 3-15 (2011).
87. Zempleni, J., Suttie, J.W., Gregory, J.F. & Stover, P.J. Handbook of
vitamins, Edn. Fifth edition. (Taylor & Francis, Boca Raton; 2014).
88. Gao, X., Reid, M.A., Kong, M. & Locasale, J.W. Metabolic interactions with cancer epigenetics. Mol Aspects Med 54, 50-57 (2017).
89. Li, K., Wahlqvist, M.L. & Li, D. Nutrition, One-Carbon Metabolism and Neural Tube Defects: A Review. Nutrients 8 (2016).
90. Patanwala, I. et al. Folic acid handling by the human gut: implications
for food fortification and supplementation. Am J Clin Nutr 100, 593-599 (2014).
91. Wright, A.J., Dainty, J.R. & Finglas, P.M. Folic acid metabolism in human subjects revisited: potential implications for proposed mandatory folic acid fortification in the UK. Br J Nutr 98, 667-675 (2007).
56
92. Bailey, S.W. & Ayling, J.E. The extremely slow and variable activity of dihydrofolate reductase in human liver and its implications for high folic acid intake. Proc Natl Acad Sci U S A 106, 15424-15429 (2009).
93. Plumptre, L. et al. High concentrations of folate and unmetabolized
folic acid in a cohort of pregnant Canadian women and umbilical cord blood. Am J Clin Nutr 102, 848-857 (2015).
94. Pfeiffer, C.M. et al. Unmetabolized folic acid is detected in nearly all serum samples from US children, adolescents, and adults. J Nutr 145, 520-531 (2015).
95. Dalto, D.B. & Matte, J.J. Pyridoxine (Vitamin B(6)) and the Glutathione Peroxidase System; a Link between One-Carbon Metabolism and Antioxidation. Nutrients 9 (2017).
96. Chon, J., Stover, P.J. & Field, M.S. Targeting nuclear thymidylate
biosynthesis. Mol Aspects Med 53, 48-56 (2017).
97. Hustad, S. et al. The methylenetetrahydrofolate reductase 677C-->T polymorphism as a modulator of a B vitamin network with major effects on homocysteine metabolism. Am J Hum Genet 80, 846-855 (2007).
98. Brosnan, J.T. & Brosnan, M.E. The sulfur-containing amino acids: an overview. J Nutr 136, 1636S-1640S (2006).
99. Hannibal, L. & Blom, H.J. Homocysteine and disease: Causal
associations or epiphenomenons? Mol Aspects Med 53, 36-42 (2017).
100. Selhub, J. Homocysteine metabolism. Annu Rev Nutr 19, 217-246 (1999).
101. Farber, S. & Diamond, L.K. Temporary remissions in acute leukemia in children produced by folic acid antagonist, 4-aminopteroyl-glutamic acid. N Engl J Med 238, 787-793 (1948).
102. Hoffbrand, A.V. & Weir, D.G. The history of folic acid. Br J Haematol
113, 579-589 (2001).
103. Blount, B.C. et al. Folate deficiency causes uracil misincorporation into human DNA and chromosome breakage: implications for cancer and neuronal damage. Proc Natl Acad Sci U S A 94, 3290-3295 (1997).
57
104. Bester, A.C. et al. Nucleotide deficiency promotes genomic instability in early stages of cancer development. Cell 145, 435-446 (2011).
105. Frosst, P. et al. A candidate genetic risk factor for vascular disease: a
common mutation in methylenetetrahydrofolate reductase. Nat Genet 10, 111-113 (1995).
106. Figueiredo, J.C., Levine, A.J., Crott, J.W., Baurley, J. & Haile, R.W. Folate-genetics and colorectal neoplasia: what we know and need to know next. Mol Nutr Food Res 57, 607-627 (2013).
107. Kennedy, D.A. et al. Folate Intake, MTHFR Polymorphisms, and the Risk of Colorectal Cancer: A Systematic Review and Meta-Analysis. J Cancer Epidemiol 2012, 952508 (2012).
108. Hanahan, D. & Weinberg, R.A. Hallmarks of cancer: the next
generation. Cell 144, 646-674 (2011).
109. Mentch, S.J. & Locasale, J.W. One-carbon metabolism and epigenetics: understanding the specificity. Ann N Y Acad Sci 1363, 91-98 (2016).
110. Mentch, S.J. et al. Histone Methylation Dynamics and Gene Regulation Occur through the Sensing of One-Carbon Metabolism. Cell Metab 22, 861-873 (2015).
111. Shyh-Chang, N. et al. Influence of threonine metabolism on S-
adenosylmethionine and histone methylation. Science 339, 222-226 (2013).
112. Anderson, O.S., Sant, K.E. & Dolinoy, D.C. Nutrition and epigenetics: an interplay of dietary methyl donors, one-carbon metabolism and DNA methylation. J Nutr Biochem 23, 853-859 (2012).
113. Paz, M.F. et al. Germ-line variants in methyl-group metabolism genes and susceptibility to DNA methylation in normal tissues and human primary tumors. Cancer Res 62, 4519-4524 (2002).
114. Cellarier, E. et al. Methionine dependency and cancer treatment.
Cancer Treat Rev 29, 489-499 (2003).
115. Vernieri, C. et al. Targeting Cancer Metabolism: Dietary and Pharmacologic Interventions. Cancer Discov 6, 1315-1333 (2016).
58
116. Kim, Y.I. Folate, colorectal carcinogenesis, and DNA methylation: lessons from animal studies. Environ Mol Mutagen 44, 10-25 (2004).
117. Kim, Y.I. Folate: a magic bullet or a double edged sword for colorectal
cancer prevention? Gut 55, 1387-1389 (2006).
118. Atta, C.A. et al. Global Birth Prevalence of Spina Bifida by Folic Acid Fortification Status: A Systematic Review and Meta-Analysis. Am J Public Health 106, e24-34 (2016).
119. Burton, A. Folic acid: time for Europe to mandate fortified flour? Lancet Neurol 15, 1208-1209 (2016).
120. Mason, J.B. et al. A temporal association between folic acid
fortification and an increase in colorectal cancer rates may be illuminating important biological principles: a hypothesis. Cancer Epidemiol Biomarkers Prev 16, 1325-1329 (2007).
121. Lee, J.E. et al. Folate intake and risk of colorectal cancer and adenoma: modification by time. Am J Clin Nutr 93, 817-825 (2011).
122. Kennedy, D.A. et al. Folate intake and the risk of colorectal cancer: a systematic review and meta-analysis. Cancer Epidemiol 35, 2-10 (2011).
123. Gibson, T.M. et al. Pre- and postfortification intake of folate and risk
of colorectal cancer in a large prospective cohort study in the United States. Am J Clin Nutr 94, 1053-1062 (2011).
124. Ebbing, M. et al. Cancer incidence and mortality after treatment with folic acid and vitamin B12. JAMA 302, 2119-2126 (2009).
125. Vollset, S.E. et al. Effects of folic acid supplementation on overall and site-specific cancer incidence during the randomised trials: meta-analyses of data on 50,000 individuals. Lancet 381, 1029-1036 (2013).
126. Glynn, S.A. et al. Colorectal cancer and folate status: a nested case-
control study among male smokers. Cancer Epidemiol Biomarkers Prev 5, 487-494 (1996).
127. Kato, I. et al. Serum folate, homocysteine and colorectal cancer risk in women: a nested case-control study. Br J Cancer 79, 1917-1922 (1999).
59
128. Ma, J. et al. Methylenetetrahydrofolate reductase polymorphism, dietary interactions, and risk of colorectal cancer. Cancer Res 57, 1098-1102 (1997).
129. Otani, T. et al. Plasma folate and risk of colorectal cancer in a nested
case-control study: the Japan Public Health Center-based prospective study. Cancer Causes Control 19, 67-74 (2008).
130. Takata, Y. et al. Plasma folate concentrations and colorectal cancer risk: a case-control study nested within the Shanghai Men's Health Study. Int J Cancer 135, 2191-2198 (2014).
131. Neuhouser, M.L. et al. Red blood cell folate and plasma folate are not associated with risk of incident colorectal cancer in the Women's Health Initiative observational study. Int J Cancer 137, 930-939 (2015).
132. Van Guelpen, B. et al. Low folate levels may protect against colorectal
cancer. Gut 55, 1461-1466 (2006).
133. Lee, J.E. et al. Plasma folate, methylenetetrahydrofolate reductase (MTHFR), and colorectal cancer risk in three large nested case-control studies. Cancer Causes Control 23, 537-545 (2012).
134. Eussen, S.J. et al. Plasma folate, related genetic variants, and colorectal cancer risk in EPIC. Cancer Epidemiol Biomarkers Prev 19, 1328-1340 (2010).
135. de Vogel, S. et al. Biomarkers related to one-carbon metabolism as
potential risk factors for distal colorectal adenomas. Cancer Epidemiol Biomarkers Prev 20, 1726-1735 (2011).
136. Keum, N. & Giovannucci, E.L. Folic acid fortification and colorectal cancer risk. Am J Prev Med 46, S65-72 (2014).
137. Cho, E. et al. Unmetabolized Folic Acid in Prediagnostic Plasma and the Risk of Colorectal Cancer. J Natl Cancer Inst 107, djv260 (2015).
138. Mason, J.B. Unraveling the complex relationship between folate and
cancer risk. Biofactors 37, 253-260 (2011).
139. Mason, J.B. & Tang, S.Y. Folate status and colorectal cancer risk: A 2016 update. Mol Aspects Med 53, 73-79 (2017).
60
140. Troen, A.M. et al. Unmetabolized folic acid in plasma is associated with reduced natural killer cell cytotoxicity among postmenopausal women. J Nutr 136, 189-194 (2006).
141. Myte, R. et al. Untangling the role of one-carbon metabolism in
colorectal cancer risk: a comprehensive Bayesian network analysis. Sci Rep 7, 43434 (2017).
142. Labuschagne, C.F., van den Broek, N.J., Mackay, G.M., Vousden, K.H. & Maddocks, O.D. Serine, but not glycine, supports one-carbon metabolism and proliferation of cancer cells. Cell Rep 7, 1248-1258 (2014).
143. Maddocks, O.D. et al. Serine starvation induces stress and p53-dependent metabolic remodelling in cancer cells. Nature 493, 542-546 (2013).
144. Bauerle, M.R., Schwalm, E.L. & Booker, S.J. Mechanistic diversity of
radical S-adenosylmethionine (SAM)-dependent methylation. J Biol Chem 290, 3995-4002 (2015).
145. Zeisel, S.H., Mar, M.H., Howe, J.C. & Holden, J.M. Concentrations of choline-containing compounds and betaine in common foods. J Nutr 133, 1302-1307 (2003).
146. Mehedint, M.G. & Zeisel, S.H. Choline's role in maintaining liver function: new evidence for epigenetic mechanisms. Curr Opin Clin Nutr Metab Care 16, 339-345 (2013).
147. Resseguie, M. et al. Phosphatidylethanolamine N-methyltransferase
(PEMT) gene expression is induced by estrogen in human and mouse primary hepatocytes. FASEB J 21, 2622-2632 (2007).
148. Cho, E. et al. Dietary choline and betaine and the risk of distal colorectal adenoma in women. J Natl Cancer Inst 99, 1224-1231 (2007).
149. Lee, J.E. et al. Choline and betaine intake and the risk of colorectal cancer in men. Cancer Epidemiol Biomarkers Prev 19, 884-887 (2010).
150. Lu, M.S. et al. Choline and betaine intake and colorectal cancer risk
in Chinese population: a case-control study. PLoS One 10, e0118661 (2015).
61
151. Zhou, Z.Y., Wan, X.Y. & Cao, J.W. Dietary methionine intake and risk of incident colorectal cancer: a meta-analysis of 8 prospective studies involving 431,029 participants. PLoS One 8, e83588 (2013).
152. Bae, S. et al. Plasma choline metabolites and colorectal cancer risk in
the Women's Health Initiative Observational Study. Cancer Res 74, 7442-7452 (2014).
153. Nitter, M. et al. Plasma methionine, choline, betaine, and dimethylglycine in relation to colorectal cancer risk in the European Prospective Investigation into Cancer and Nutrition (EPIC). Ann Oncol 25, 1609-1615 (2014).
154. Ganguly, P. & Alam, S.F. Role of homocysteine in the development of cardiovascular disease. Nutr J 14, 6 (2015).
155. Marti-Carvajal, A.J., Sola, I. & Lathyris, D. Homocysteine-lowering
interventions for preventing cardiovascular events. Cochrane Database Syst Rev 1, CD006612 (2015).
156. Selhub, J. & Miller, J.W. The pathogenesis of homocysteinemia: interruption of the coordinate regulation by S-adenosylmethionine of the remethylation and transsulfuration of homocysteine. Am J Clin Nutr 55, 131-138 (1992).
157. Kumar, T., Sharma, G.S. & Singh, L.R. Homocystinuria: Therapeutic approach. Clin Chim Acta 458, 55-62 (2016).
158. McNulty, H. et al. Riboflavin lowers homocysteine in individuals
homozygous for the MTHFR 677C->T polymorphism. Circulation 113, 74-80 (2006).
159. Ulvik, A. et al. Ratios of One-Carbon Metabolites Are Functional Markers of B-Vitamin Status in a Norwegian Coronary Angiography Screening Cohort. J Nutr (2017).
160. Le Marchand, L. et al. Plasma levels of B vitamins and colorectal cancer risk: the multiethnic cohort study. Cancer Epidemiol Biomarkers Prev 18, 2195-2201 (2009).
161. Weinstein, S.J. et al. One-carbon metabolism biomarkers and risk of
colon and rectal cancers. Cancer Epidemiol Biomarkers Prev 17, 3233-3240 (2008).
62
162. Miller, J.W. et al. Homocysteine, cysteine, and risk of incident colorectal cancer in the Women's Health Initiative observational cohort. Am J Clin Nutr 97, 827-834 (2013).
163. Zhang, X.H., Ma, J., Smith-Warner, S.A., Lee, J.E. & Giovannucci, E.
Vitamin B6 and colorectal cancer: current evidence and future directions. World J Gastroenterol 19, 1005-1010 (2013).
164. Ueland, P.M., Ulvik, A., Rios-Avila, L., Midttun, O. & Gregory, J.F. Direct and Functional Biomarkers of Vitamin B6 Status. Annu Rev Nutr 35, 33-70 (2015).
165. Scott, J.M. & Weir, D.G. The methyl folate trap. A physiological response in man to prevent methyl group deficiency in kwashiorkor (methionine deficiency) and an explanation for folic-acid induced exacerbation of subacute combined degeneration in pernicious anaemia. Lancet 2, 337-340 (1981).
166. Palmer, A.M., Kamynina, E., Field, M.S. & Stover, P.J. Folate rescues
vitamin B12 depletion-induced inhibition of nuclear thymidylate biosynthesis and genome instability. Proc Natl Acad Sci U S A 114, E4095-E4102 (2017).
167. Song, M., Garrett, W.S. & Chan, A.T. Nutrients, foods, and colorectal cancer prevention. Gastroenterology 148, 1244-1260 e1216 (2015).
168. Larsson, S.C., Orsini, N. & Wolk, A. Vitamin B6 and risk of colorectal cancer: a meta-analysis of prospective studies. JAMA 303, 1077-1083 (2010).
169. Larsson, S.C. et al. Folate, vitamin B6, vitamin B12, and methionine
intakes and risk of stroke subtypes in male smokers. Am J Epidemiol 167, 954-961 (2008).
170. Eussen, S.J. et al. Plasma vitamins B2, B6, and B12, and related genetic variants as predictors of colorectal cancer risk. Cancer Epidemiol Biomarkers Prev 19, 2549-2561 (2010).
171. Hustad, S. et al. Riboflavin, flavin mononucleotide, and flavin adenine dinucleotide in human plasma and erythrocytes at baseline and after low-dose riboflavin supplementation. Clin Chem 48, 1571-1577 (2002).
172. van Kapel, J., Spijkers, L.J., Lindemans, J. & Abels, J. Improved
distribution analysis of cobalamins and cobalamin analogues in
63
human plasma in which the use of thiol-blocking agents is a prerequisite. Clin Chim Acta 131, 211-224 (1983).
173. Vashi, P., Edwin, P., Popiel, B., Lammersfeld, C. & Gupta, D.
Methylmalonic Acid and Homocysteine as Indicators of Vitamin B-12 Deficiency in Cancer. PLoS One 11, e0147843 (2016).
174. Dahlin, A.M. et al. Plasma vitamin B12 concentrations and the risk of colorectal cancer: a nested case-referent study. Int J Cancer 122, 2057-2061 (2008).
175. Percudani, R. & Peracchi, A. The B6 database: a tool for the description and classification of vitamin B6-dependent enzymatic activities and of the corresponding protein families. BMC Bioinformatics 10, 273 (2009).
176. Badawy, A.A. Kynurenine Pathway of Tryptophan Metabolism:
Regulatory and Functional Aspects. Int J Tryptophan Res 10, 1178646917691938 (2017).
177. Ulvik, A. et al. Evidence for increased catabolism of vitamin B-6 during systemic inflammation. Am J Clin Nutr 100, 250-255 (2014).
178. Ulvik, A. et al. Substrate product ratios of enzymes in the kynurenine pathway measured in plasma as indicators of functional vitamin B-6 status. Am J Clin Nutr 98, 934-940 (2013).
179. Zuo, H. et al. Plasma Biomarkers of Inflammation, the Kynurenine
Pathway, and Risks of All-Cause, Cancer, and Cardiovascular Disease Mortality: The Hordaland Health Study. Am J Epidemiol 183, 249-258 (2016).
180. Zuo, H. et al. Markers of vitamin B6 status and metabolism as predictors of incident cancer: the Hordaland Health Study. Int J Cancer 136, 2932-2939 (2015).
181. SBU. Benefits and risks of fortifying flour with folic acid to reduce the risk of neural tube defects. Stockholm: Swedish Council on Health Technology Assessment in Health Care (SBU); 2007. SBU report no 183 (in Swedish).
182. Weinehall, L., Hallgren, C.G., Westman, G., Janlert, U. & Wall, S.
Reduction of selection bias in primary prevention of cardiovascular disease through involvement of primary health care. Scand J Prim Health Care 16, 171-176 (1998).
64
183. Morris, M.S., Jacques, P.F., Rosenberg, I.H. & Selhub, J. Folate and vitamin B-12 status in relation to anemia, macrocytosis, and cognitive impairment in older Americans in the age of folic acid fortification. Am J Clin Nutr 85, 193-200 (2007).
184. Selhub, J. & Paul, L. Folic acid fortification: why not vitamin B12 also?
Biofactors 37, 269-271 (2011).
185. Obeid, R. The metabolic burden of methyl donor deficiency with focus on the betaine homocysteine methyltransferase pathway. Nutrients 5, 3481-3495 (2013).
186. Zeisel, S. Choline, Other Methyl-Donors and Epigenetics. Nutrients 9 (2017).
187. Klein, E.A. et al. Vitamin E and the risk of prostate cancer: the
Selenium and Vitamin E Cancer Prevention Trial (SELECT). JAMA 306, 1549-1556 (2011).
188. Cole, B.F. et al. Folic acid for the prevention of colorectal adenomas: a randomized clinical trial. JAMA 297, 2351-2359 (2007).
189. Norberg, M. et al. Community participation and sustainability--evidence over 25 years in the Vasterbotten Intervention Programme. Glob Health Action 5, 1-9 (2012).
190. Pukkala, E. et al. Nordic biological specimen banks as basis for studies
of cancer causes and control--more than 2 million sample donors, 25 million person years and 100,000 prospective cancers. Acta Oncol 46, 286-307 (2007).
191. Norberg, M., Wall, S., Boman, K. & Weinehall, L. The Vasterbotten Intervention Programme: background, design and implications. Glob Health Action 3 (2010).
192. Van Guelpen, B. et al. Plasma folate and total homocysteine levels are associated with the risk of myocardial infarction, independently of each other and of renal function. J Intern Med 266, 182-195 (2009).
193. Windelberg, A., Arseth, O., Kvalheim, G. & Ueland, P.M. Automated
assay for the determination of methylmalonic acid, total homocysteine, and related amino acids in human serum or plasma by means of methylchloroformate derivatization and gas chromatography-mass spectrometry. Clin Chem 51, 2103-2109 (2005).
65
194. Kelleher, B.P. & Broin, S.D. Microbiological assay for vitamin B12 performed in 96-well microtitre plates. J Clin Pathol 44, 592-595 (1991).
195. Molloy, A.M. & Scott, J.M. Microbiological assay for serum, plasma,
and red cell folate using cryopreserved, microtiter plate method. Methods Enzymol 281, 43-53 (1997).
196. Midttun, O., Hustad, S. & Ueland, P.M. Quantitative profiling of biomarkers related to B-vitamin status, tryptophan metabolism and inflammation in human plasma by liquid chromatography/tandem mass spectrometry. Rapid Commun Mass Spectrom 23, 1371-1379 (2009).
197. Midttun, O., Kvalheim, G. & Ueland, P.M. High-throughput, low-volume, multianalyte quantification of plasma metabolites related to one-carbon metabolism using HPLC-MS/MS. Anal Bioanal Chem 405, 2009-2017 (2013).
198. Meyer, K., Fredriksen, A. & Ueland, P.M. MALDI-TOF MS genotyping
of polymorphisms related to 1-carbon metabolism using common and mass-modified terminators. Clin Chem 55, 139-149 (2009).
199. Cockcroft, D.W. & Gault, M.H. Prediction of creatinine clearance from serum creatinine. Nephron 16, 31-41 (1976).
200. Zuo, H. et al. Interferon-gamma-induced inflammatory markers and the risk of cancer: the Hordaland Health Study. Cancer 120, 3370-3377 (2014).
201. Allin, K.H., Nordestgaard, B.G., Zacho, J., Tybjaerg-Hansen, A. &
Bojesen, S.E. C-reactive protein and the risk of cancer: a mendelian randomization study. J Natl Cancer Inst 102, 202-206 (2010).
202. Toriola, A.T. et al. Biomarkers of inflammation are associated with colorectal cancer risk in women but are not suitable as early detection markers. Int J Cancer 132, 2648-2658 (2013).
203. Brenner, D.R. et al. A review of the application of inflammatory biomarkers in epidemiologic cancer research. Cancer Epidemiol Biomarkers Prev 23, 1729-1751 (2014).
204. Rundle, A.G., Vineis, P. & Ahsan, H. Design options for molecular
epidemiology research within cohort studies. Cancer Epidemiol Biomarkers Prev 14, 1899-1907 (2005).
66
205. Gylling, B. et al. Low folate levels are associated with reduced risk of colorectal cancer in a population with low folate status. Cancer Epidemiol Biomarkers Prev 23, 2136-2144 (2014).
206. Wang, M. et al. Statistical methods for studying disease subtype
heterogeneity. Stat Med 35, 782-800 (2016).
207. Ganna, A. et al. Risk prediction measures for case-cohort and nested case-control designs: an application to cardiovascular disease. Am J Epidemiol 175, 715-724 (2012).
208. Persson, E. & Waernbaum, I. Estimating a marginal causal odds ratio in a case-control design: analyzing the effect of low birth weight on the risk of type 1 diabetes mellitus. Stat Med 32, 2500-2512 (2013).
209. Carpenter, J. & Bithell, J. Bootstrap confidence intervals: when,
which, what? A practical guide for medical statisticians. Stat Med 19, 1141-1164 (2000).
210. Bostom, A.G. et al. Post-methionine load hyperhomocysteinemia in persons with normal fasting total plasma homocysteine: initial results from the NHLBI Family Heart Study. Atherosclerosis 116, 147-151 (1995).
211. Institute of Medicine (US) Standing Committee on the Scientific Evaluation of Dietary Reference Intakes and its Panel on Folate, O.B.V., and Choline. Dietary Reference Intakes for Thiamin, Riboflavin, Niacin, Vitamin B6, Folate, Vitamin B12, Pantothenic Acid, Biotin, and Choline. Washington (DC): National Academies Press (US). (1998).
212. Eussen, S.J. et al. North-south gradients in plasma concentrations of
B-vitamins and other components of one-carbon metabolism in Western Europe: results from the European Prospective Investigation into Cancer and Nutrition (EPIC) Study. Br J Nutr 110, 363-374 (2013).
213. Jacques, P.F., Selhub, J., Bostom, A.G., Wilson, P.W. & Rosenberg, I.H. The effect of folic acid fortification on plasma folate and total homocysteine concentrations. N Engl J Med 340, 1449-1454 (1999).
214. Myte, R. et al. Components of One-carbon Metabolism Other than Folate and Colorectal Cancer Risk. Epidemiology 27, 787-796 (2016).
67
215. Schlesselman, J.J. & Stolley, P.D. Case-control studies : design, conduct, analysis. (Oxford University Press, New York; 1982).
216. Rofail, D., Colligs, A., Abetz, L., Lindemann, M. & Maguire, L. Factors
contributing to the success of folic acid public health campaigns. J Public Health (Oxf) 34, 90-99 (2012).
217. Zauber, A.G. The impact of screening on colorectal cancer mortality and incidence: has it really made a difference? Dig Dis Sci 60, 681-691 (2015).
218. Elinav, E. et al. Inflammation-induced cancer: crosstalk between tumours, immune cells and microorganisms. Nat Rev Cancer 13, 759-771 (2013).
219. Paul, L., Ueland, P.M. & Selhub, J. Mechanistic perspective on the
relationship between pyridoxal 5'-phosphate and inflammation. Nutr Rev 71, 239-244 (2013).
220. Ueland, P.M., McCann, A., Midttun, O. & Ulvik, A. Inflammation, vitamin B6 and related pathways. Mol Aspects Med 53, 10-27 (2017).
221. Saibeni, S. et al. Low vitamin B(6) plasma levels, a risk factor for thrombosis, in inflammatory bowel disease: role of inflammation and correlation with acute phase reactants. Am J Gastroenterol 98, 112-117 (2003).
222. Ulvik, A., Midttun, O., Pedersen, E.R., Nygard, O. & Ueland, P.M.
Association of plasma B-6 vitamers with systemic markers of inflammation before and after pyridoxine treatment in patients with stable angina pectoris. Am J Clin Nutr 95, 1072-1078 (2012).
223. Morris, M.S., Sakakeeny, L., Jacques, P.F., Picciano, M.F. & Selhub, J. Vitamin B-6 intake is inversely related to, and the requirement is affected by, inflammation status. J Nutr 140, 103-110 (2010).
224. Stead, L.M., Brosnan, J.T., Brosnan, M.E., Vance, D.E. & Jacobs, R.L. Is it time to reevaluate methyl balance in humans? Am J Clin Nutr 83, 5-10 (2006).
225. Wyss, M. & Kaddurah-Daouk, R. Creatine and creatinine metabolism.
Physiol Rev 80, 1107-1213 (2000).
226. Mudd, S.H. et al. Methyl balance and transmethylation fluxes in humans. Am J Clin Nutr 85, 19-25 (2007).
68
227. Eklof, V. et al. The prognostic role of KRAS, BRAF, PIK3CA and PTEN in colorectal cancer. Br J Cancer 108, 2153-2163 (2013).