University of Southern California Dissertations and Theses

The role of altered folate one-carbon metabolism (FOCM) in the development of colorectal cancer

(USC Thesis Other)

The role of altered folate one-carbon metabolism (FOCM) in the development of colorectal cancer

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content THE ROLE OF ALTERED FOLATE ONE- CARBON METABOLISM (FOCM)
IN THE DEVELOPMENT OF COLORECTAL CANCER
Isaac Asante

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree

DOCTOR OF PHILOSOPHY
CLINICAL AND EXPERIMENTAL THERAPEUTICS

August 2017
i

Dedication
This dissertation is dedicated to my supportive wife, Francisca and my beloved
children, Jesuslina, Isaac Jnr., and Chrysolite. I deeply appreciate your enduring
support during this phase of our shared lives.

ii

Acknowledgement
Surely, it is in Him we live and move and have our being. The
superabundance of your grace and providence has always inspired me to aspire.
I humbly acknowledge how the Ebenezer God nurtured me for the fulfillment of
His glory.
I would like to thank my mentors, Dr. Stan G. Louie and Dr. David V. Conti.
You have been great teachers and scientists that create the perfect environment
for innovation, learning, and development. Your modest approach to solving
complex issues has given me an indispensable tool for life. I am forever grateful
to you for your confidence and the opportunity you gave me. To Dr. Daryl Davies,
thank you for your warmth and guidance. You always brought out the best in me,
and I am very grateful.
A special prayer of gratitude always swells in my heart for Dr. Kathleen
Rodgers. You are a special vessel of honor who is so pure in heart and exude
grace. You have been a blessing to my family and set a pace for me regarding
the true balance of academic and spiritual excellence. You have truly left an
indelible mark that will be told to generations unborn. I cannot leave out Dr. Julio
Camarero for your guidance. Thank you, Dr. Annie Wong-Beringer for your giving
me the opportunity and inspiring me on. I thank Wade Thompson-Harper and
Rosie Soltero for all your support.
Last but not the least, I thank my great family of laboratory mates- past
and present. Thank you, Hua, Eugene, Lindsay, Darryl, Tamar, and Drs. Sachin
Jadhav, Dezheng Dong, Julie Yoo, Radhika Borra and Rong Yang for all the
support and friendship. Each of you means so much to me.

iii

Table of Contents
List of Figures ................................................................................................ vii
List of Tables .................................................................................................. ix
List of Abbreviations ....................................................................................... x
Abstract .......................................................................................................... xii
Chapter 1: Introduction to Colorectal Cancer and Folate One-Carbon
Metabolism ................................................................................................. 1
1.1 Epidemiology of CRC ..................................................................................... 1
1.2 Stages of Colorectal Cancer .......................................................................... 3
1.3 Current Diagnostic Approaches for Colorectal Cancer ............................... 5
1.4 Molecular mechanism of colorectal cancer .................................................. 6
1.4.1 DNA methylation ...................................................................................... 6
1.4.2 Aberrant methylation in CRC .................................................................... 7
1.4.3 Models of CRC development ................................................................... 8
1.5 FOCM and CRC risk ........................................................................................ 9
1.6 Review of previous epidemiological studies on folate............................... 11
1.6.1 The role of folate metabolites in CRC development ................................ 11
1.6.2 The current gap in the role of folate metabolites in CRC development ... 13
1.6.3 Addressing the gap in folates’ role in CRC development ........................ 13
1.7 Using metabolomics to dissect association between folates and CRC
risk ................................................................................................................. 14
1.8 Changes in DNA Methylation in Response to the Diet ............................... 16
1.9 Hypothesis and specific aims ...................................................................... 18
1.10 Outline of the Dissertation ......................................................................... 19
Chapter 1: Bibliography and References ................................................... 21
Chapter 2: Development of Multi-analyte Metabolomics-based LCMS
Assay Platform to Quantify Vitamin B in Plasma Matrix....................... 28
2.1 Introduction ................................................................................................... 28
2.2 Methods ......................................................................................................... 31
2.2.1 Chemicals and Reagents ....................................................................... 31
2.2.2 Biosamples ............................................................................................ 31
2.2.3 Preparation of stock solutions ................................................................ 32
2.2.4 Standards and Quality Control Samples ................................................. 32
2.2.5 Plasma Preparation ................................................................................ 33
2.2.6 Instrumentation and LCMS Conditions ................................................... 33
2.2.7 Method Validation .................................................................................. 35
iv

2.2.8 Method Application ................................................................................. 38
2.3 Results and Discussion ............................................................................... 38
2.3.1 Method optimization ............................................................................... 38
2.3.2 Method validation ................................................................................... 41
2.3.3 Proof of Applicability ............................................................................... 46
2.4 Conclusion .................................................................................................... 47
Chapter 2: Bibliography and References .......................................................... 48
Chapter 3: Identification of Blood-based Biomarkers as Potential Screen
for Colorectal Cancer Using Metabolomics ........................................... 53
3.1. Introduction .................................................................................................. 53
3.2. Materials and Methods ................................................................................ 56
3.2.1. Study samples ...................................................................................... 56
3.2.2. Chemicals and reagents ........................................................................ 57
3.2.3. Sample preparation ............................................................................... 57
3.2.4. LCMS data acquisition for targeted and untargeted scan ...................... 58
3.2.5 Quantitation of plasma methylmalonic acid, homocysteine,
and related metabolites .......................................................................... 59
3.2.6. Data acquisition and processing ............................................................ 61
3.2.7. Statistical analysis ................................................................................. 62
3.3. Results.......................................................................................................... 63
3.3.1. Data reduction and exploratory analyses .............................................. 63
3.3.2. Putative biomarkers from untargeted approach ..................................... 65
3.3.3. Targeted metabolites ............................................................................ 68
3.4 Discussion .................................................................................................... 70
3.5 Conclusions .................................................................................................. 74
Chapter 3: Bibliography and References .......................................................... 76
Chapter 4: Alterations in Folate-dependent One-Carbon Cycle as Colon
Cell Transition from Normal to Cancerous ............................................ 81
4.1 Introduction ................................................................................................... 81
4.2 Materials and Method .......................................................................... 86
4.2.1 Cell lines and culture .............................................................................. 86
4.2.2 Growth rates of colon cell lines............................................................... 87
4.2.3 Extraction and analysis of cellular FOCM metabolites ............................ 88
4.2.4 RNA and DNA isolation .......................................................................... 89
4.2.5 DNA hydrolysis and global methylation measurement by LCMS ............ 89
4.2.6 Quantitative real-time polymerase chain reaction ................................... 90
4.3 Statistical analysis ........................................................................................ 92
4.4 Results........................................................................................................... 93
4.4.1 Growth rates of colon cell lines............................................................... 93
v

4.4.2 Distribution of folate metabolites concentrations in colon cell lines ......... 95
4.4.3 Expression of folate metabolism genes in colon cell lines ...................... 96
4.4.4 Plasma levels of FOCM metabolites in colon cell lines ........................... 97
4.4.5 Global DNA methylation ....................................................................... 100
4.4.6 Expression of FOCM genes across colon cell lines .............................. 101
4.4.7 Correlations between genes, growth rate and CRC status ................... 104
4.5 Discussion .................................................................................................. 105
4.5.1 Folate distribution in colon cells ............................................................ 106
4.5.2 DNA methylation in transformed cells ................................................... 107
4.5.3 Redox homeostasis in transformed colon cell lines .............................. 108
4.5.4 Methionine depletion in transformed cells ............................................ 109
4.5.5 Regulation of proliferation and apoptosis in colon cell lines .................. 110
4.6 Conclusion .................................................................................................. 112
Chapter 4: Bibliography and References ........................................................ 113
Chapter 5: The Effect of Folate Supplementation on colon cell lines at
Different Stages of CRC Development ................................................. 124
5.1 Introduction ................................................................................................. 124
5.2 Materials and Method ................................................................................. 127
5.2.1 Cell lines and culture ............................................................................ 127
5.2.2 Effect of folates on growth rates of colon cell lines ............................... 128
5.2.3 Extraction and analysis of cellular FOCM metabolites .......................... 129
5.2.4 RNA and DNA isolation ........................................................................ 130
5.2.5 DNA hydrolysis and global methylation measurement by LCMS .......... 130
5.2.6 Real time polymerase chain reaction.................................................... 131
5.3 Statistical analysis ...................................................................................... 132
5.4 Results......................................................................................................... 133
5.4.1 Effect of folate treatment on cellular growth.......................................... 133
5.4.2 Effect of folate on colony formation ...................................................... 135
5.4.3 Effect of folate on cellular folate proportions ......................................... 136
5.4.4 Effect of supplementation on cellular FOCM metabolites ..................... 138
5.4.5 Expression of FOCM genes across colon cell lines .............................. 140
5.4.6 Correlations between genes, growth rate and CRC status ................... 141
5.5 Discussion .................................................................................................. 143
5.5.1 Effect of folate supplementation on HCT116 p53+/+ cells .................... 144
5.5.2 Effect of folate supplementation on HCT116 p53-/- cells ...................... 146
5.5.3 Effect of folate supplementation on Caco-2 cells .................................. 146
5.5.4 Effect of folate supplementation on CRL1790 ...................................... 147
5.6 Conclusion .................................................................................................. 148
Chapter 5: Bibliography and References ................................................. 149
Chapter 6: Conclusion and Future Directions .......................................... 155
vi

6.1 Summary ..................................................................................................... 155
6.2 Exploration of CRC biomarkers ................................................................. 157
6.3 The role of FOCM in various colon cell lines ............................................ 159
6.4 Redox homeostasis in transformed colon cell lines ................................ 160
6.5 Methionine depletion in transformed cells ............................................... 162
6.6 Regulation of proliferation and apoptosis in colon cell lines .................. 163
6.7 Folic Acid Supplementation and CRC risk ................................................ 163
6.8 Conclusions and Significance ................................................................... 165
6.9 Future Directions ........................................................................................ 165
Chapter 6: Bibliography and References ................................................. 168
Complete Bibliography and References ................................................... 173

vii

List of Figures
Figure 1.1: Clinical Staging of CRC .............................................................................. 4
Figure 1.2: Comparison of methylation patterns in normal and cancer cells ................. 8
Figure 1.3: The FOCM cycle which drives DNA methylation, synthesis and repair ..... 10
Figure 2.1: The FOCM cycle which drives DNA methylation, synthesis and
repair……………………………………………………………………………………….… 29
Figure 2.2: Comparison of analyte extraction by LLE and SPE……..……………...…..40
Figure 2.3: Stability of FOCM analytes in various stabilizing agents………………...…41
Figure 2.4: Chromatograms of analytes with interfering peaks in the stripped blank
plasma. ....................................................................................................................... 42
Figure 2.5: Representative chromatograms of FOCM analytes ………………………..43
Figure 2.6: Stability of analytes after 18hrs of storage in refrigerated autosampler ..... 46
Figure 3.1: PCA-DA plots of CRC cases and controls…………………………………...65
Figure 3.2: Volcano plot of controls versus CRC cases .............................................. 66
Figure 3.3: Box plots showing the plasma FOCM analytes concentrations in CRC
cases and controls...................................................................................................... 68
Figure 3.4: Box plots showing the plasma concentrations of other FOCM
metabolites in CRC cases and controls……………………………………………………70
Figure 4.1: The FOCM cycle which drives DNA methylation, synthesis and repair. .... 82
Figure 4.2: Growth curve and doubling times of colon cell lines in DMEM . ................ 93
Figure 4.3: The growth rate of cells compared by mean adjusted AUC of
growth curves ............................................................................................................ 94
Figure 4.4: Morphology of colon cell lines. .................................................................. 94
Figure 4.5: Cellular folate distribution in colon cells. ................................................... 96
Figure 4.6: Expression of folate metabolism genes in colon cells ............................... 97
Figure 4.7: Cellular levels of FOCM metabolites are compared in different
colon cells.. ................................................................................................................ 99
viii

Figure 4.8: Expression of DNMT isoforms in colon cells……………………………….101
Figure 4.9: Expression of genes related to transmethylation cycle in colon
cell lines ................................................................................................................... 102
Figure 4.10: Expression of genes related to DNA synthesis in colon cell lines .......... 103
Figure 4.11: Expression of genes related to proliferation and apoptosis in colon
cell lines ................................................................................................................... 103
Figure 4.12: Correlation matrix of FOCM related genes, growth rate and CRC status of
colon cells ................................................................................................................ 105
Figure 5.1: Growth curves of colon cell lines in different treatment media . ............... 134
Figure 5.2: Comparison of the growth rates of colon cell lines in different
treatment media. ...................................................................................................... 135
Figure 5.3: Colony forming assay showing the effect of FA and 5MTHF
supplementation on transitional and transformed colon cells. ................................... 136
Figure 5.4: Cellular folate proportions in CRL1459, APC 10.1, HCT116 p53+/+ and
Caco-2 treated with supplements. ............................................................................ 138
Figure 5.5: Methylation capacity of APC 10.1, Caco-2 and HCT116 p53+/+ grown in
various treatment media. . ........................................................................................ 139
Figure 5.6: Relative global methylation of APC 10.1, Caco-2 and HCT116 p53+/+
grown in various treatment media. . .......................................................................... 139
Figure 5.7: Expression of FOCM genes after treatment with various folate
supplements. . .......................................................................................................... 141
Figure 5.8: Correlation of FOCM related genes in colon cells with FA or 5MTHF
concentration as supplement. . ................................................................................. 142
Figure 5.9: Correlation matrix of FOCM related genes in colon cells
supplemented with different levels of folates. . .......................................................... 143
Figure 6.1: Comparison of folate proportions in clinical and in vitro data. .................. 160
Figure 6.2: Cellular levels of Succinic acid in colon cells .......................................... 167

ix

List of Tables
Table 2.1: Concentration of FOCM components in working solution ........................... 32
Table 2.2: HPLC Pump gradient program for analyte separation ................................ 34
Table 2.3: Analytes, selected MRM transitions, retention time (RT) and
parameter settings ...................................................................................................... 35
Table 2.4: Inter-run accuracy and precision (n=7 for each concentration level) .......... 44
Table 2.5: Recovery rates of analytes at various concentration ranges ...................... 45
Table 2.6: Plasma concentrations (nM) of analytes in healthy volunteers (n=10) ........ 47
Table 3.1: Demographic characteristics of plasma sample donors ............................. 56
Table 3.2: Candidate features whose mean intensities are significantly different in
CRC and controls with respective match in the Human Metabolome Database .......... 67
Table 3.3: Comparison of the plasma FOCM metabolites levels in CRC cases
and controls ................................................................................................................ 69
Table 4.1: Summary of characteristics of colon cell lines ............................................ 87
Table 4.2: Sequences of primers used in RT-PCR in human and mouse species ...... 91
Table 4.3: The effect of p53 mutation on FOCM gene expression in HCT116 .......... 104
Table 5.1: The media treatment conditions of the colon cells .................................... 128
Table 5.2: Comparison of the mean doubling time for the various colon cells in
different treatment media. ......................................................................................... 134

x

List of Abbreviations
4PA 4-Pyridoxic Acid
5,10-MeTHF 5,10-Methylenetetrahydrofolate
5MTHF 5-Methyltetrahydrofolate
APC Adenomatous Polyposis Coli
AUC Area Under the Curve
B2 Riboflavin
B6 Pyridoxine
CBS Cystathionine-β-Synthase
CE Collision Energy
CIMP CpG Island Methylator Phenotype
CIN Chromosomal Instability
CIS Carcinoma In Situ
CRC Colorectal Cancer
CV Coefficient of Variation
CXP Collision Cell Exit Potential
CYSTH Cystathionine
DHF Dihydrofolate
DHFR Dihydrofolate Reductase
DP Declustering Potential
dTMP Deoxythymidine Monophosphate
dUMP Deoxyuridine Monophosphate
EP Entrance Potential
FA Folic Acid
FMN Flavin Mononucleotide
FOBT Fecal Occult Blood Tests
FOCM Folate One-Carbon Metabolism
GSH Glutathione
GWAS Genome-Wide Association Study
HCY Homocysteine
IAP Intracisternal A Particle
IDA Integrated Data Acquisition
LCMS Liquid-Chromatography -Mass Spectrometry
LLE Liquid-Liquid Extraction
LLOQ Lower Limit of Quantification
MAT Methionine Adenosyl Transferase
METH Methionine
MMA Methylmalonic Acid
MMR Mismatch Repair
MRM Multiple Reaction Monitoring
MS Methionine Synthase
xi

MSI Microsatellite Instability
MTHFR Methylenetetrahydrofolate Reductase
MTX Methotrexate
PL Pyridoxal
PM Pyridoxamine
QC Quality Control
RE Recovery
RT Retention Time
SAH S-Adenosyl Homocysteine
SAM S-Adenosyl Methionine
SHMT Serine Hydroxymethyltransferase
SNP Single Nucleotide Polymorphism
SPE Solid-Phase Extraction
TCEP Tris-2-Chloroethyl Phosphate
TGFß Transforming Growth Factor-ß
THF Tetrahydrofolate
TYMS Thymidylate Synthase
USPSTF U.S. Preventive Services Task Force

xii

Abstract
Background: CRC ranks as the third leading cause of cancer-related death in
the industrialized world, responsible for about 50,000 deaths in the United States
annually. Aberrant methylation patterns have been implicated in the
pathogenesis of CRC and thus make it necessary to probe the folate one-carbon
metabolism (FOCM) to explore for any association with CRC risk.
Aims: The main goals of this study were to; 1) Develop a validated
metabolomics-based LCMS assay for the FOCM metabolites; 2) Use these
metabolomic assays as a platform to explore for blood-based biomarkers in
plasma of CRC cases and controls; 3) Explore for differences in the FOCM
metabolites and genes for colon cells at different stages in CRC development
and 4) Evaluate the cellular effect of folic acid (FA) or 5-methyltetrahydrofolate
(5MTHF) supplementation on the FOCM of colon cells at different stages of CRC
development.
Results: The proportions of the folate metabolites were significantly different in
the CRC cases and controls where the accumulation of reduced folate
metabolites was observed in the cases. A similar trend was also seen in the
transformed cells when compared with the transitional and untransformed colon
cells. In addition, there was a significant difference in methionine levels was
observed between the two groups. The transformed cells showed upregulation
in most of the FOCM genes, except for MTR/MTRR, CBS, and MAT where there
were differences based on their microsatellite stability status.
xiii

The effect of FA or 5MTHF supplementation had opposing effects on the
expression of the genes. Although increasing concentrations of FA correlated
with increased growth rate, escalating concentration of 5MTHF did not cause a
significant change in the cellular growth rate. The group with the combination of
2.3 µM 5MTHF and 9 µM FA was observed to slow down the growth rate
observed in the 9 µM FA supplementation group.
Discussion and Conclusions: The accumulation of reduced folate
metabolites may be due to the overexpression of folate metabolism genes and
a downregulation of MS/MTRR genes. Also, there is an overexpression of
proliferation genes as well as their modulation through aberrant methylation
that tend to drive cell proliferation. Supplementation with 9 µM FA was
associated with significant increase in growth rate in untransformed and
transformed colon cell lines, but the addition of 2.3 µM of 5MTHF to the 9 µM
FA (to form the Combo) was associated with reduced growth in the transitional
and transformed cell lines.
1

Chapter 1: Introduction to Colorectal Cancer and Folate One-
Carbon Metabolism
1.1 Epidemiology of CRC
Strategic approaches vary as for how to prevent, manage, or cure various
ailments differ drastically. To prevent a disease, the ability to identify and mitigate
key risks that drive the disease development is critical. No matter which of these
strategies one chooses to utilize, the availability of reliable biomarkers is critical
to guide clinical decision making, initiation of treatment or pharmaceutical
discovery/development.
The development of reliable biomarkers plays a central role in the fight against
colorectal cancer (CRC) disease, which is the third leading cause of cancer-
related death in the industrialized world (Siegel et al., 2016) and major burden
worldwide (Ferlay et al., 2015). In the U.S., it is not only the fourth most
commonly diagnosed cancer but has an annual mortality rate of approximately
50,000 patients (Group, 2016). CRC is cancer thought to evolve from
uncontrolled cellular growth found in the colon or rectum. The annual national
expenditure for CRC treatment is estimated to be $5.5-$6.5 billion out of which
inpatient hospital care accounts for approximately 80% of the total healthcare
expenditure (Seifeldin and Hantsch, 1999). However, early CRC diagnosis and
intervention in the earlier stages have been shown to significantly improved
clinical prognosis compared to CRC diagnosed in later stages (Gupta et al.,
2

2008). Unfortunately, most cases of CRC are rarely diagnosed early, where the
median age of CRC diagnosis and death are 68 years and 73 years respectively
(Group, 2016). Thus it is understandable that most CRC diagnosed in these
patients are often in later stages of the disease, where invasion and metastasis
have already occurred. Also, these tumors have acquired genetic mutations
which make them resistant to chemotherapy such as KRAS. For all of these
reasons, CRC diagnosed in these patients are difficult to treat, with a drastically
reduced five-year survival rate (Burt et al., 2010; O’Connell et al., 2004).
Current screening method uses detection for occult blood in the stool and more
invasive periodic colonoscopy. While Hemoccult detection usually identifies
patients with CRC in more advanced stages, the invasive nature and the cost
associated with colonoscopy have drastically reduced compliance with screening
guideline. These factors contribute to the low screening compliance rates and
thus points to urgent needed for a reliable, sensitive, blood-based CRC detection
approach. The ability to identify pre-cancerous, or CRC in early stages could
potentially save lives. However, this cannot be achieved without the identification
of reliable biomarkers which can inform us of molecular changes that are
occurring in the untransformed epithelial colon cell through the transforming
process to become adenomatous. Before we can develop a rational approach,
we must understand the developmental process of CRC and why the normal
cellular proliferation has changed leading to the formation of adenoma.
3

1.2 Stages of Colorectal Cancer
CRC are staged into one of the five stages: stage 0 (zero) and stages I through
IV (1 through 4). Each stage is further described using the tumor, node, and
metastasis or TNM staging system. TMN is an anatomical or physical-based
scoring system which is widely used to predict CRC prognosis and guide
adjuvant therapy after potentially curative surgery. The T represents the extent
of tumor invasion into the colon, the N is the number of nearby lymph node(s)
where the colon cancer is detected, and the M represents the metastatic status
of the colon cancer. The three characteristics of CRC found in the patient can
be used to assign a clinical staging for the colon cancer, which can be correlated
with clinical course.
T is scored from 1 to 4, where the number represented the extent of tumor found
in the bowel:
 T1 is where the tumor is found only in the inner layer of the bowel
 T2 is where the tumor has invaded into the muscle layer of the bowel
wall
 T3 is where the tumor has invaded into the outer lining of the bowel
wall
 T4 is where the tumor has invaded through the outer lining of the bowel
wall.
4

Also, there are three stages describing whether CRC has been detected in the
adjacent lymph nodes (N).
 N0 is where there is no CRC has been detected in the lymph nodes.
 N1 is where 1 to 3 lymph nodes adjacent to the bowel and was found
to have CRC invasion.
 N2 is where there are cancer cells found in 4 or more nearby lymph
nodes
Thirdly, there are 2 M stages which define the metastases nature of the CRC:
 M0 is where cancer has not spread to other organs
 M1 is where cancer has spread to other parts of the body

Figure 1.1: Clinical Staging of CRC. Adapted from (Winslow, 2005)

5

Disease diagnosis and treatment success depend on tumor stage at detection
(Figure 1.1). This classification begins at the presence of benign polyp (stage 0)
where it infiltrates the mucosa towards serosa with stage progression. Stage 0 is
also referred carcinoma in situ (CIS). Stage 1 CRC in the TNM staging is the
same as T1, N0, M0, or T2, N0, M0. Stage 2 colon cancer has been sub-
classified as 2a and 2b. Stage 2a tumor has migrated into the outer lining of the
bowel wall but has no lymph node involvement or metastases. In contrast, Stage
2b the tumor has penetrated through the outer lining of the outer bowel wall,
similar to Stage 2a where no nodal and metastatic involvement is found. Usually,
at stages 1 and 2, the polyp size becomes big enough to be detected by
colonoscopy. When the tumor progresses to the stages 3 & 4, it spreads into the
circulatory system through the lymph nodes and subsequently to the other
organs. Stage 3 is divided into three stages, where nodal involvement is the key
difference between Stage 2 and 3. No sign of metastasis is noted in Stage 3
staging. Stage 4 is colon cancer that has spread.
1.3 Current Diagnostic Approaches for Colorectal Cancer
There is currently no systemic diagnostic, or predictive biomarker for CRC and
the presence of benign precursor adenomas is identified through invasive
procedures such as colonoscopy. Though with variable risk and benefits, the
U.S. Preventive Services Task Force (USPSTF) recommends CRC screening
using either high-sensitivity fecal occult blood tests (FOBT), sigmoidoscopy, or
colonoscopy in adults, beginning at age 50 years and continuing until age 75
6

years (Pignone and Sox, 2008). High cost, low public education and the poor
patient acceptance of the screening procedures have been cited as some of the
major reasons for low screening rates (Rim et al., 2011; Vernon, 1997). The
development of an affordable, pliable, and minimally invasive test such as blood-
based screening assay that could accurately and precisely detect precursor
lesions would significantly improve screening rates and thereby lead to reduce
CRC-related deaths. The development of reliable biomarkers that can facilitate
early detection, prevention, targeted drug development and improved clinical
outcomes. These biomarkers must be closely linked with the pathways leading
to the development of benign precursor adenomas and ultimately into CRC.
Precursors found in these tumorigenesis pathway(s) will yield the important cues
as to which metabolite(s) or biomarkers to focus on for biomarker development.
1.4 Molecular mechanism of colorectal cancer
1.4.1 DNA methylation
DNA methylation is a key epigenetic alteration that occurs through the addition
of a methyl group onto specific sites found on the DNA that regulate gene
expression. Epigenetic alterations are heritable changes in gene activity and
expression that occur without alteration in DNA sequence (Bird, 2007).
Epigenetic alterations may affect the gene expression patterns through the
interplay of epigenetic substrates (like methyl group), epigenetic regulators (like
methyltransferases) and transcriptions factors. DNA and histone modification
reactions, such as methylation and acetylation, are often used by cells to control
7

the composition, structure, and dynamics of chromatin and thereby regulate gene
expression (Bird, 2002) without affecting the genomic DNA sequence. DNA
methylation is also able to maintain DNA integrity and stability, chromatin
modifications, and reduce the development of mutations.
1.4.2 Aberrant methylation in CRC
The methylated region of the DNA dictates the type of biological function
observed; in cancers, differential DNA methylation occurs at the promoters- CpG
islands or CpG islands shore (2kb from CpG Island). A comparison of the
methylation patterns in normal, and cancer cells reveal differences in their
patterns at CpG islands and globally across the genes (Figure 1.2). Aberrant
DNA methylation in cancer development can manifest as global
hypomethylation, has been associated with chromosomal instability, or
hypermethylation of CpG islands that results in the silence tumor suppressor
genes thus promoting tumorigenesis (Crider et al., 2012). CRC is cancer with
epigenetic alterations and gene mutations in the tumorigenesis pathway(s). The
gene methylation patterns that a cell exhibit during development is established
through the de novo DNA methyltransferases. The DNA methylation process is
highly dependent on two factors 1) methyltransferase and transcription factors
and 2) the level of methyl substrates produced in the folate one-carbon
metabolism (FOCM).

8

Figure 1.2: Comparison of methylation patterns in normal and cancer cells. The orange
circles indicate DNA methylation. Adapted from (Crider et al., 2012)

1.4.3 Models of CRC development
The adenoma-carcinoma sequence model aligns with clinicopathological
changes with genetic alterations along the progression pathway of CRC
development involving genes that regulate cell growth. CRC may develop from
one or a combination of three proposed mechanisms, namely chromosomal
instability (CIN), microsatellite instability (MSI) and CpG island methylator
phenotype (CIMP) (Tariq and Ghias, 2016). The CIMP pathway is known for the
promoter hypermethylation of various genes across the genome, prominent
amongst which are DNA mismatch repair (MMR) genes (e.g. MGMT and MLH1).
This hypermethylation is closely linked to the MSI pathway which involves the
inactivation of genetic alterations in short repeated sequences. The initial step in
CIN adenoma tumorigenesis is associated with loss of adenomatous polyposis
coli (APC), mutations in the GTPase KRAS, followed by loss of chromosome 18q
with SMAD4, which is downstream of transforming growth factor-ß (TGFß).
There are also mutations in TP53 and other relevant genes to drive the
Normal
Cancer
(CRC)
9

tumorigenesis. Most of these mutations and alterations in gene expression may
be mediated by aberrant DNA methylation processes that occur at various points
along the proposed model (Fearon and Vogelstein, 1990). Although the
underlying form of genomic or epigenomic instability determines the types of
mutations that occur in the various types of colon cancer, the selective pressures
that drive the development of the tumors are essentially the same across all
CRCs (Lao, 2011).
1.5 FOCM and CRC risk
Folates are the methyl carriers essential for intracellular transmethylation
reactions including those involved in DNA methylation and DNA synthesis. The
FOCM produces S-adenosyl methionine (SAM), the primary methyl donor used
for DNA methylation. Folates are naturally occurring water-soluble vitamin B and
important nutritional factor that play a major role in the pathogenesis of several
disorders in humans. Studies have shown the role of folates in macrocytic
anemia, cardiovascular disease (Bailey et al., 2003; Boushey et al., 1995;
Collaboration, 2002), thromboembolic processes (Ray, 1998), neural tube and
other congenital defects (MRC Vitamin Study Research Group, 1991; Berry et
al, 1999), adverse pregnancy outcomes (George et al., 2002; Vollset et al.,
2000), neuropsychiatric disorders (Seshadri et al., 2002) and CRC (Benito et al.,
1991; Boyapati et al., 2004; Giovannucci et al., 1998; Keku et al., 2002; Larsson
et al., 2006; Terry et al., 2002). Folates mediate the transfer of methyl group in
10

FOCM to provide the needed substrates for some intracellular methylation
reactions that are critical for gene regulation (Figure 1.3).

Figure 1.3: The FOCM cycle which drives DNA methylation, synthesis and repair.
DHFR, dihydrofolate reductase; 5-MTHF, 5-methyltetrahydrofolate; 5,10-MeTHF,
5,10-methylenetetrahydrofolate; CBS, Cystathionine-β-synthase; MAT, Methionine
adenosyl transferase; MS, methionine synthase; MTHFR,
methylenetetrahydrofolate reductase; SAM, S-adenosylmethionine; SAH, S-
adenosylhomocysteine; SHMT, Serine hydroxymethyltransferase; THF,
tetrahydrofolate; TYMS, thymidylate synthase; dUMP, deoxyuridine
monophosphate; dTMP, deoxythymidine monophosphate.

Folate and its metabolites have been described to maintain genomic stability
through regulating DNA biosynthesis, repair, and methylation (de Vogel et al.,
2011). However, folic acid (FA) is the synthetic form of the vitamin that lacks
coenzyme activity but the natural (dietary) form; 5-methyltetrahydrofolate is
metabolically active. To become an active coenzyme, FA needs to be activated
11

as coenzyme through subsequent reduction to dihydrofolate (DHF) and then to
tetrahydrofolate (THF) (Figure 1.3) by dihydrofolate reductase (DHFR). The
DHFR is a relatively slow enzyme in humans and appears incapable of
completely converting a significant amount of FA to THF (Bailey and Ayling,
2009).
1.6 Review of previous epidemiological studies on folate
1.6.1 The role of folate metabolites in CRC development
As the effect of folate on CRC risks is probed, a parallel interest is a possibility
of reversing epigenetic changes seen in early tumorigenesis (e.g., global
hypomethylation) as rapidly dividing tissue tumors may be constrained by low
folate availability, resulting in global hypomethylation. The reported studies lack
a definitive association between low folate status and global DNA
hypomethylation which also had limited sample size or imprecision of the assays.
There have been several epidemiological studies that have evaluated the
association of folates with the risk of CRC. A large cohort CRC incidence study
followed 88,756 women and found those with higher folate intake (>400ug/d) has
decreased CRC risk than those with intake (<200ug/d) after controlling for
confounders (Giovannucci et al., 1998). Another study which was conducted in
both men and women evaluate the association between total folate, dietary folate
and folic acid intake and CRC risk. Decreased risk of CRC was found to be
significantly associated with increased total folate intake (but not natural folate or
folic acid) (Stevens et al., 2011). These studies were limited by the fact that the
12

dietary folate intake was estimated to be equivalent to the plasma folate
concentration which may not be necessarily true in all cases. The best reflective
measure for folate metabolites to be used for association modeling would have
been the intracellular concentrations which are very challenging to obtain,
especially for clinical samples. However, the extracellular (plasma) concentration
may be a close surrogate marker of the phenotype in the subjects involved and
extrapolate any association. There is usually a homeostatic flux between the
intracellular and extracellular (plasma) levels.
Kim et al. (2011) found high plasma folate among cases (≥ 6.7 ng/ml vs. <4.1
ng/ml) was associated with increased methylation at tumor protein, p73 but not
hMLH1. Pufulete et al. (2005) found decreased cancer risk among patient groups
with higher folate status; CRC and adenoma patients had lower serum folate and
global hypomethylation. The studies that measured serum/plasma folates
showed a better design to overcome the earlier limitation. However, most
epidemiological studies on folates were limited by their sample sizes (most of
them were below 80 per group) which made it difficult to model for several
variables. The development of such multivariate association models requires
large samples sizes which are lacking in most of the reviewed studies. In the
candidate gene pathway-based study, Levine et al. (2010) evaluated the effect
of folates, vitamin B12 and vitamin B6 intake on CRC risks in 1,805 CRC cases
and 2,878 Sibling controls. The study also measured 395 tagSNPs from 13
genes and found an association between DHFR & MS single nucleotide
13

polymorphism (SNP) and decreased the risk of CRC in a folate-fortified
population.
1.6.2 The current gap in the role of folate metabolites in CRC development
The current gap in knowledge is to understand the mechanistic link between
folate levels, methylation capacity, and CRC development. Several studies have
reported contrasting relationship between folates and the CRC risk (Giovannucci
et al., 1995; Giovannucci et al., 1998; Giovannucci et al., 1993; Kim, 2004;
Larsson et al., 2006; Qin et al., 2013; Terry et al., 2002; Ulrich and Potter, 2006).
Among these studies, the ones that measured plasma folates used
microbiological assays to quantify the folates and the other metabolites. The
major limitation of microbiological assays is their inability to distinguish which the
level of specific metabolite(s) found in neither the samples nor the all the major
metabolites required for FOCM.
1.6.3 Addressing the gap in folates’ role in CRC development
Unlike the microbiological assay, liquid-chromatography -mass spectrometry
(LCMS) has the potential advantage of quantifying the individual folate species,
such as the major derivatives, at extracellular and cellular levels, as well as of
relatively minor species, such as unmetabolized folic acid. This advantage is
likely to be a phenomenal platform for phenotyping for more information on
genetic polymorphisms that affect nutritional status and their respective folate
distributions (Shane, 2011). The limitations in the previous studies that might
14

have masked the true relationship that exists between the individual FOCM
metabolites and the CRC risk are addressed in this project. The project proposed
studies that are extensions of the candidate gene pathway-based study and a
genome-wide association study (GWAS). The candidate gene pathway-based
study measured the environmental risks factors associated with the CRC risk
such as nutritional practices, heavy alcohol consumption, cigarette smoking
among others (Haggar and Boushey, 2009). Based on the known gene
polymorphisms in CRC, the studies in this project aimed at exploring the
phenotypic expression of these FOCM-related genes in CRC cases and controls.
The levels of the FOCM metabolites were explored for any association with the
CRC risk, and further developed into blood-based CRC biomarkers.
1.7 Using metabolomics to dissect association between folates and CRC
risk
Exploration for blood-based metabolites as possible biomarkers require sensitive
and specific assays which can quantify the individual intermediates in the FOCM
cycle. Such specific metabolite profiling facilitates the exploration of the possible
association between any FOCM metabolite and CRC risk. In this pursuit,
metabolomics was proposed as the alternative approach to address the
confounding gaps in the previous epidemiological studies and their findings.
Metabolomics is simultaneous profiling of different metabolites’ concentrations/
fluctuations present in a biological sample in response to environmental or
genomic influence. Metabolomics has emerged as a major supporting path of
15

functional genomics, alongside transcriptomics (mRNA profiling) and proteomics
(van der Greef et al., 2003)
At a given point in time, several processes occur in the biological system at the
genomic level through to the metabolites that control the biochemical processes.
A holistic molecular picture of events can be obtained from the aggregate piecing
of information from the “tri-omics”- transcriptomics, proteomics, and
metabolomics. Metabolomics serves as the convergence point of all the “tri-
omics” in addition to the environmental variation like dietary intake and lifestyle
that may affect the biological system. Metabolomics has a unique possibility of
phenotypic and dynamic profiling due to the critical role of metabolites in cellular
controls, communication, and anabolism, and energy transportation. Careful “tri-
omics” profiling and comparison of the biological systems that differ by a disease
condition can be explored to identify the possible pathways that are affected by
the disease. Such profiling approach may lead to the discovery of biomarker(s)
that identifies with the disease condition like CRC. The metabolites of interest
existing in varying concentrations (mostly very low levels) that need the utilization
of platforms that are capable of highly sensitive and specific analysis.
LCMS is a great platform that serves this end by accurately measuring specific
metabolites that make metabolomics a realistic approach. Considering the gaps
in previous studies on the association of folates to CRC risks, it was pertinent to
employ the LCMS-based metabolomics as a unique approach to probe this
association to establish a more reliable association between specific folate
metabolites and CRC risks. The plasma FOCM metabolites were then profiled
16

for cases and controls so that their points of differences explored for pathway
analysis after merging the metabolomics data with their corresponding GWAS
data. Most of these metabolites serve as either substrates or products of key
enzymes of the cycle so the ratio of the product to the substrate can also provide
a powerful analytical tool to compare possible enzymatic functional changes.
1.8 Changes in DNA Methylation in Response to the Diet
The dietary intake contains essential nutrients that modulate critical biochemical
cycles in the biological system. These nutrients may cause epigenetic
modulation that affects gene expression and cell growth in the developmental as
well as later stages of maturation. Altered methylation patterns are the common
epigenetic effect that regulates the several genes, including oncogenes and
proto- oncogenes to affect tumorigenesis. Also, certain environmental exposures
are associated with changes in DNA methylation and alterations in FOCM
metabolites. The amount and the timing of exposure have been found to
influence the effects observed in the biological system.
This observation raises issues of a possible reversal of determined oncogenic
cells by exposing them to environmental conditions that modulate the effects
drive oncogenes by restoring normal methylation patterns. The DNA methylation
capacity is affected by the methyl substrates or the methyltransferase activity
that drive the reaction. When biological systems are exposed to diet rich methyl
substrates, they express a resulting phenotype related to the altered DNA
patterns. In mice, maternal dietary changes in FOCM availability can affect the
17

DNA methylation patterns and phenotype of offspring (Kim et al., 2009). This
effect is well demonstrated in the agouti mouse strain in which the altered
methylation patterns of an inserted intracisternal A particle (IAP) sequence in an
upstream of the agouti gene.
In humans, the association between dietary intake of folates (which are methyl
donors), DNA methylation and subsequently, the risk of CRC is not well
understood. With the food fortification with folic acid in North America, there is a
growing concern for the health consequences due to the altered DNA
methylation levels/patterns. The food fortification program exposes the
population to increased levels of folic acid which may affect the folate
metabolism. However, it is noteworthy that the effect that synthetic folic acid may
render on a biological system may differ from that produced by a dietary folate
that feeds directly into the transmethylation cycle. Because of fortification of the
food supply with folic acid, concerns about folate status have shifted from the
detection of folate deficiency in populations toward possible adverse effects of
high folate intake. Mass spectrometry methods potentially allow the monitoring
of the effects of folic acid accumulation, such as the folate distributions in the
blood (Shane, 2011).
DNA methylation is a genetic regulatory mechanism that serves as a vital
component for establishing and maintaining cell survival, differentiation, and
multipotency. Aberrant DNA methylation is an epigenetic alteration which
activates oncogenes, inactivates tumor suppressors, and drives several
sequential mutations required in CRC development. The microenvironment of
18

the normal epithelium has an influence on the phenotypic expressions. When
this environmental influence interferes with critical cycles in the epithelial cellular
process, it leads to epigenetic changes that eventually may drive other sequential
mutations. The altering of the metabolites that feed into the FOCM of the colonic
epithelia leads to variations in the intracellular levels of FOCM intermediates
which influence the supply of methyl donors and methylation patterns of the cells
thus modulating tumorigenesis. When these alterations in methylation become
aberrant enough, they lead to sequential mutations of the colonic epithelial which
eventually develops into CRC.
1.9 Hypothesis and specific aims
The FOCM modulates the DNA stability and methylation patterns of the normal
and cancer cell system and thus led to our hypothesis that altered circulating
FOCM intermediates are associated with the development of CRC and thus, may
be a good predictive biomarker to identify CRC. The rationale for the proposed
research is that, once some metabolites have been identified as a biomarker for
CRC, it can be used for CRC screening and diagnostic purposes in the clinical
setting when developed into new and innovative bioanalytical assays. The
central hypothesis led to the pursuit of the following specific aims (SA):
SA1: To develop validated metabolomics-based LCMS assays as a platform to
quantify and explore the plasma levels of FOCM intermediates.
SA2: To identify the association between specific FOCM metabolites and
incidence of CRC in humans.
19

SA3: To dissect and verify the role of FOCM intermediates in the tumorigenesis
of CRC using in vitro colon cells at various stages of development.
SA4: To explore the molecular effect of folic acid or 5MTHF supplementation on
the tumorigenesis of CRC using in vitro colon cells at different stages of
development.
1.10 Outline of the Dissertation
This dissertation presents the development metabolomics-based LCMS assays
(Chapter 2) and uses them as platforms to explore the association of FOCM
metabolites in clinical samples (Chapter 3). The role of FOCM in the phenotypic
expression of colon cells at different stages of CRC development will be explored
using in vitro model (Chapter 4). Reviewing the role of FOCM in CRC from
another perspective, the effect of folate supplementation on colon cells at
different stages of CRC development will be assessed using in vitro model
(Chapters 5). The thesis will conclude with a summary of the role of FOCM
explored introducing evidence from the preliminary findings from large
population-based human studies (Chapter 6). The first approach was to validate
the metabolomics-based assays in human plasma, cell and culture media
matrices. Secondly, the assays were used as a platform to explore any
alterations the FOCM metabolites in human CRC and healthy volunteer plasma
samples as well as in colon cells in different stages of cancer development.
Furthermore, the effects of folic acid or 5-methyltetrahydrofolate supplementation
on normal and colon cancer cells were also explored. Lastly, the association
20

between any FOCM metabolite and the risk of developing CRC was assessed
by measuring the metabolite levels in 2,300 CRC cases and their sibling controls.
This model was further enriched by mapping the phenotypic metabolomics data
to the genomic data on the different SNPs that relate to the development of CRC.
21

Chapter 1: Bibliography and References
Bailey, L.B., Rampersaud, G.C., and Kauwell, G.P.A. (2003). Folic acid
supplements and fortification affect the risk for neural tube defects, vascular
disease and cancer: evolving science. J Nutr 133, 1968.
Bailey, S.W., and Ayling, J.E. (2009). 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.
Benito, E., Stiggelbout, A., Bosch, F., Obrador, A., Kaldor, J., Mulet, M., and
Munoz, N. (1991). Nutritional factors in colorectal cancer risk: a case ‐control
study in Majorca. International journal of cancer 49, 161-167.
Bird, A. (2002). DNA methylation patterns and epigenetic memory. Genes &
development 16, 6-21.
Bird, A. (2007). Perceptions of epigenetics. Nature 447, 396-398.
Boushey, C.J., Beresford, S.A., Omenn, G.S., and Motulsky, A.G. (1995). A
quantitative assessment of plasma homocysteine as a risk factor for vascular
disease: probable benefits of increasing folic acid intakes. Jama 274, 1049-
1057.
Boyapati, S.M., Bostick, R.M., McGlynn, K.A., Fina, M.F., Roufail, W.M.,
Geisinger, K.R., Hebert, J.R., Coker, A., and Wargovich, M. (2004). Folate
intake, MTHFR C677T polymorphism, alcohol consumption, and risk for
22

sporadic colorectal adenoma (United States). Cancer Causes & Control 15,
493-501.
Burt, R.W., Barthel, J.S., Dunn, K.B., David, D.S., Drelichman, E., Ford, J.M.,
Giardiello, F.M., Gruber, S.B., Halverson, A.L., and Hamilton, S.R. (2010).
Colorectal cancer screening. Journal of the National Comprehensive Cancer
Network 8, 8-61.
Collaboration, H.S. (2002). Homocysteine and risk of ischemic heart disease
and stroke: a meta-analysis. Jama 288, 2015-2022.
Crider, K.S., Yang, T.P., Berry, R.J., and Bailey, L.B. (2012). Folate and DNA
methylation: a review of molecular mechanisms and the evidence for folate’s
role. Advances in Nutrition: An International Review Journal 3, 21-38.
de Vogel, S., Schneede, J., Ueland, P.M., Vollset, S.E., Meyer, K., Fredriksen,
Å., Midttun, Ø., Bjørge, T., Kampman, E., and Bretthauer, M. (2011).
Biomarkers related to one-carbon metabolism as potential risk factors for distal
colorectal adenomas. Cancer Epidemiology Biomarkers & Prevention 20, 1726-
1735.
Fearon, E.R., and Vogelstein, B. (1990). A genetic model for colorectal
tumorigenesis. Cell 61, 759-767.
Ferlay, J., Soerjomataram, I., Dikshit, R., Eser, S., Mathers, C., Rebelo, M.,
Parkin, D.M., Forman, D., and Bray, F. (2015). Cancer incidence and mortality
23

worldwide: sources, methods and major patterns in GLOBOCAN 2012.
International journal of cancer 136, E359-E386.
George, L., Mills, J.L., Johansson, A.L., Nordmark, A., Olander, B., Granath, F.,
and Cnattingius, S. (2002). Plasma folate levels and risk of spontaneous
abortion. Jama 288, 1867-1873.
Giovannucci, E., Ascherio, A., Rimm, E.B., Stampfer, M.J., Colditz, G.A., and
Willett, W.C. (1995). Intake of carotenoids and retinol in relation to risk of
prostate cancer. Journal of the national cancer institute 87, 1767-1776.
Giovannucci, E., Stampfer, M.J., Colditz, G.A., Hunter, D.J., Fuchs, C., Rosner,
B.A., Speizer, F.E., and Willett, W.C. (1998). Multivitamin use, folate, and colon
cancer in women in the Nurses' Health Study. Annals of Internal Medicine 129,
517-524.
Giovannucci, E., Stampfer, M.J., Colditz, G.A., Rimm, E.B., Trichopoulos, D.,
Rosner, B.A., Speizer, F.E., and Willett, W.C. (1993). Folate, methionine, and
alcohol intake and risk of colorectal adenoma. Journal of the National Cancer
Institute 85, 875-883.
Group, U.C.S.W. (2016). United States cancer statistics: 1999–2013 Incidence
and Mortality Web-based Report. Atlanta: US Department of Health and Human
Services, Centers for Disease Control and Prevention and National Cancer
Institute.
24

Gupta, A.K., Brenner, D.E., and Turgeon, D.K. (2008). Early Detection of Colon
Cancer. Molecular diagnosis & therapy 12, 77-85.
Haggar, F.A., and Boushey, R.P. (2009). Colorectal cancer epidemiology:
incidence, mortality, survival, and risk factors. Clin Colon Rectal Surg 22, 191-
197.
Keku, T., Millikan, R., Worley, K., Winkel, S., Eaton, A., Biscocho, L., Martin, C.,
and Sandler, R. (2002). 5, 10-Methylenetetrahydrofolate reductase codon 677
and 1298 polymorphisms and colon cancer in African Americans and whites.
Cancer Epidemiology Biomarkers & Prevention 11, 1611-1621.
Kim, J.W., Park, H.M., Choi, Y.-K., Chong, S.Y., Oh, D., and Kim, N.K. (2011).
Polymorphisms in genes involved in folate metabolism and plasma DNA
methylation in colorectal cancer patients. Oncol Rep 25, 167-172.
Kim, K.-c., Friso, S., and Choi, S.-W. (2009). DNA methylation, an epigenetic
mechanism connecting folate to healthy embryonic development and aging.
The Journal of nutritional biochemistry 20, 917-926.
Kim, Y.-I. (2004). Folate and DNA methylation: a mechanistic link between
folate deficiency and colorectal cancer? Cancer Epidemiology Biomarkers &
Prevention 13, 511-519.
Lao, V.V., & Grady, W. M. (2011). Epigenetics and Colorectal Cancer. Nature
Reviews Gastroenterology and Hepatology 8, 686- 700.
25

Larsson, S.C., Giovannucci, E., and Wolk, A. (2006). Folate intake, MTHFR
polymorphisms, and risk of esophageal, gastric, and pancreatic cancer: a meta-
analysis. Gastroenterology 131, 1271-1283.
Levine, A.J., Figueiredo, J.C., Lee, W., Conti, D.V., Kennedy, K., Duggan, D.J.,
Poynter, J.N., Campbell, P.T., Newcomb, P., and Martinez, M.E. (2010). A
candidate gene study of folate-associated one carbon metabolism genes and
colorectal cancer risk. Cancer Epidemiology Biomarkers & Prevention 19,
1812-1821.
O’Connell, J.B., Maggard, M.A., and Ko, C.Y. (2004). Colon cancer survival
rates with the new American Joint Committee on Cancer sixth edition staging.
Journal of the National Cancer Institute 96, 1420-1425.
Pignone , M., and Sox, H.C. (2008). Screening for colorectal cancer: US
Preventive Services Task Force recommendation statement. Annals of Internal
Medicine 149.
Pufulete, M., Al-Ghnaniem, R., Rennie, J., Appleby, P., Harris, N., Gout, S.,
Emery, P., and Sanders, T. (2005). Influence of folate status on genomic DNA
methylation in colonic mucosa of subjects without colorectal adenoma or
cancer. Br J Cancer 92, 838-842.
Qin, X., Cui, Y., Shen, L., Sun, N., Zhang, Y., Li, J., Xu, X., Wang, B., Xu, X.,
and Huo, Y. (2013). Folic acid supplementation and cancer risk: A meta ‐
26

analysis of randomized controlled trials. International Journal of Cancer 133,
1033-1041.
Ray, J.G. (1998). Meta-analysis of hyperhomocysteinemia as a risk factor for
venous thromboembolic disease. Arch Intern Med 158, 2101-2106.
Rim, S.H., Joseph, D.A., Steele, C.B., Thompson, T.D., and Seeff, L.C. (2011).
Colorectal cancer screening - United States, 2002, 2004, 2006, and 2008.
MMWR Surveill Summ 60 Suppl, 42-46.
Seifeldin, R., and Hantsch, J.J. (1999). The economic burden associated with
colon cancer in the United States. Clin Ther 21, 1370-1379.
Seshadri, S., Beiser, A., Selhub, J., Jacques, P.F., Rosenberg, I.H., D'Agostino,
R.B., Wilson, P.W.F., and Wolf, P.A. (2002). Plasma homocysteine as a risk
factor for dementia and Alzheimer's disease. N Engl J Med 346, 476-483.
Shane, B. (2011). Folate status assessment history: implications for
measurement of biomarkers in NHANES. The American journal of clinical
nutrition 94, 337S-342S.
Siegel, R. L., Miller, K. D., & Jemal, A. (2016). Cancer statistics, 2016. CA: a
cancer journal for clinicians, 66(1), 7-30.
Tariq, K., and Ghias, K. (2016). Colorectal cancer carcinogenesis: a review of
mechanisms. Cancer biology & medicine 13, 120.
27

Terry, P., Jain, M., Miller, A.B., Howe, G.R., and Rohan, T.E. (2002). Dietary
intake of folic acid and colorectal cancer risk in a cohort of women. International
journal of cancer 97, 864-867.
Ulrich, C.M., and Potter, J.D. (2006). Folate supplementation: too much of a
good thing? Cancer Epidemiology Biomarkers & Prevention 15, 189-193.
van der Greef, J., Davidov, E., Verheij, E., Vogels, J., van der Heijden, R.,
Adourian, A.S., Oresic, M., W., M.E., and Naylor, S. (2003). The Role of
Metabolomics in Systems Biology. In Metabolic Profiling: Its Role in Biomarker
Discovery and Gene Function Analysis, G.G.H.R. Goodacre, ed.
(Boston/Dordrecht/London: Kluwer Academic Publishers ), pp. 171-198.
Vernon, S.W. (1997). Participation in colorectal cancer screening: a review. J
Natl Cancer Inst 89, 1406-1422.
Vollset, S.E., Refsum, H., Irgens, L.M., Emblem, B.M., Tverdal, A., Gjessing,
H.K., Monsen, A.L., and Ueland, P.M. (2000). Plasma total homocysteine,
pregnancy complications, and adverse pregnancy outcomes: the Hordaland
Homocysteine study. Am J Clin Nutr 71, 962-968.
Winslow, T. (2005). Colon Cancer Stages (Cephalic Vein).
28

Chapter 2: Development of Multi-analyte Metabolomics-based
LCMS Assay Platform to Quantify Vitamin B in Plasma Matrix
2.1. Introduction
Vitamin B is a family of water-soluble nutritional factors that play a significant role
in maintaining wellness. In contrast, deficiencies in various isoforms of Vitamin
B have been linked closely to the pathogenesis of several disorders including
macrocytic anemia, cardiovascular disease (Bailey et al., 2003; Boushey et al.,
1995; Collaboration, 2002), thromboembolic processes (Ray, 1998), neural tube
defects (Berry et al., 1999; Group, 1991), adverse pregnancy outcomes (George
et al., 2002; Vollset et al., 2000), neuropsychiatric disorders (Seshadri et al.,
2002) and colorectal cancer (CRC) (Benito et al., 1991; Boyapati et al., 2004;
Giovannucci et al., 1998; Keku et al., 2002; Larsson et al., 2006; Terry et al.,
2002). Vitamin B isoforms include B2, B6, folates (B9), and B12, where compounds
are critical mediators and co-factors mediating biological reactions found in the
folate one-carbon metabolism (FOCM) cycle. The FOCM cycle produces
essential metabolites required for transmethylation reactions such as S-adenosyl
methionine (SAM), the primary methyl donor, which is required for the methylation
of various substrates to biosynthesize DNA, phospholipids, neurotransmitters
and proteins.
Folates are pteroyl glutamates that serve as the primary methyl carriers essential
for intracellular transmethylation reactions such as nucleic acids biosynthesis and
methylation reactions that are critical for gene regulation (Figure 2.1). In addition,
29

folate and its metabolites have been shown to enhance genomic stability through
regulating nucleic acid biosynthesis, repair, and methylation (de Vogel et al.,
2011). Folates are also required to transfer methyl groups that are needed to
synthesize the nucleotides needed to build DNA and modulate gene expressions
through selective DNA site methylation.

Figure 2.1: The FOCM cycle which drives DNA methylation, synthesis and repair.
DHFR, dihydrofolate reductase; 5-MTHF, 5-methyltetrahydrofolate; 5,10-MeTHF,
5,10-methylenetetrahydrofolate; CBS, Cystathionine-β-synthase; MAT, Methionine
adenosyltransferase; MS, methionine synthase; MTHFR,
methylenetetrahydrofolate reductase; SAM, S-adenosylmethionine; SAH, S-
adenosylhomocysteine; SHMT, Serine hydroxymethyltransferase; THF,
tetrahydrofolate; TYMS, thymidylate synthase; dUMP, deoxyuridine
monophosphate; dTMP, deoxythymidine monophosphate.

Dietary folates are readily available source for metabolic reactions when ingested
as either synthetic or supplemented folic acid (FA) or 5-methyltetrahydrofolate
30

(5MTHF). These folates are found in high concentrations in green leafy
vegetables. Metabolically active folate, 5MTHF, serves as the secondary methyl
donor in the FOCM serving as a coenzyme for methionine synthase (MS) to
convert homocysteine to methionine as summarized in Figure 2.1. To become an
active coenzyme folate, FA must be reduced to dihydrofolate (DHF) and then
further metabolized into tetrahydrofolate (THF), a reaction mediated by
dihydrofolate reductase (DHFR).
The levels of vitamin B are important co-factors in the FOCM, where their levels
can impact the rate of enzymatic activities. Previously, studies relating to FOCM
measured plasma folates used microbiological-based assays. Bacteria lacking
specific genes to synthesize various cofactors are employed, where the
exogenous source of these co-factors facilitated survival and growth. As
aforementioned, the limitation of these microbiological growth assays are that
they are unable to distinguish the various folate metabolite(s) that are present in
blood, and thus incapable of identifying the metabolite that is responsible the
biological activity. To more accurately identify which vitamin B metabolite(s) is
responsible for exerting their biological activity(ies), it is necessary to identify and
quantify the specific metabolites accurately and precisely. To address this
information gap, a more specific, accurate, and sensitive assay is required to
quantify the various vitamin B metabolites. This chapter presents a validated
multi-analyte metabolomics-based liquid chromatography- mass spectrometry
(LCMS) method that can adequately quantify the plasma levels of specific vitamin
B metabolites found in the FOCM.
31

2.2. Methods
2.2.1. Chemicals and Reagents
Analytical grades of ascorbic acid, flavin mononucleotide (FMN), folic acid (FA),
pyridoxine (B6), pyridoxal (PL) hydrochloride, pyridoxamine (PM) dihydrochloride,
and 4-pyridoxic acid (4PA) were bought from Sigma (St Louis, MO, USA).
Riboflavin (B2) was bought from Alfa Aesar and 5-methyltetrahydrofolate
(5MTHF), dihydrofolate (DHF) and tetrahydrofolate (THF) were acquired from
Cayman Chemicals (Ann Arbor, MI, USA). Methotrexate (MTX), purchased from
Enzo Life Sciences (Farmingdale, NY, USA), was used as an internal standard
for the assay. The purity of each standard was above 97%, except DHF and THF
which were 90% and 95% respectively. Ultrapure HPLC-grade water, LCMS-
grade methanol, LCMS-grade acetonitrile and formic acid were purchased from
Fisher Scientific (Pittsburg, PA, USA) and were used for sample processing and
mobile phase systems.
2.2.2. Biosamples
Filtered and charcoal-stripped human blank plasma samples obtained from
Bioreclamation Inc (Long Island, NY) were used for assay development and
validation. Samples used to test the assay’s applicability were obtained from
Bioreclamation Inc (Long Island, NY). These samples were collected from
volunteers consented for such analyses. The mean age and standard deviation
of the volunteers were 41 years and 13 respectively. Exactly 50% of the
volunteers were males.
32

2.2.3. Preparation of stock solutions
A stock solution of each standard was prepared at concentrations ranging from
0.34 to 9.6 mg/mL. The stock solutions were stored at -80°C from where they
were retrieved and diluted with 1% ascorbic acid to give an appropriately
stabilized mixture of standards for use.
2.2.4. Standards and Quality Control Samples
The stock solutions of the standards were mixed to form a working solution the
composition of which is defined in Table 2.1. The constituent amount of each
standard in the working solution was estimated at 20 times the upper reference
limit.
Table 2.1: Concentration of FOCM components in working solution
Metabolite Concentration
(ng/mL)
4PA 600
5MTHF 108
B2 460
B6 600
DHF 540
FA 108
FMN 460
PL 600
PM 600
THF 108

The working solution summarized in Table 2.1 was used to prepare calibration
standards and quality control (QC) samples. Working solution was freshly
prepared daily and further diluted using 1% ascorbic acid in water to form
calibration standards and QC’s. Ten calibration standards were prepared at
dilutions of two- to thousand-fold. To each standard 30 ng/ml of methotrexate
33

(MTX) was added as an internal standard. Plasma standards were prepared by
spiking 50 µL of the appropriate standard solution into 50 µL of charcoal-treated
pooled plasma. Suitable factors of working solution were spiked into the blank
plasma to make high, medium and low QC samples by adding the during
calibration curve and sample processing. The nominal concentrations of ten-,
hundred-, and four hundred- fold dilutions of the working solution were used as
high, medium and low QC samples respectively.
2.2.5. Plasma Preparation
To 50 µL of clarified human plasma, 50 µL of 30 ng/ml MTX was added. Ice-cold
precipitation solution (400 µl) composed of 20% 0.2 M ZnSO4 in methanol was
added, vortexed for 30 seconds and stored at -20°C for 30 minutes. The sample
was then centrifuged at 15,000 rpm for 15 min at 4°C, after which 400 µL of the
supernatant was transferred to a clean Eppendorf tube and evaporated to
dryness using a steady stream of dried and HEPA filtered nitrogen gas. The dried
residue was reconstituted using 30 µL of 1% ascorbic acid, where the samples
were centrifuged at 15,000 rpm for a minute at 4°C to remove debris. An aliquot
of 25µL of the clarified supernatant transferred into HPLC vials and 20 µL injected
into LCMS for quantification.
2.2.6. Instrumentation and LCMS Conditions
The analytes were separated and quantified using an LCMS system comprising
of Shimadzu Prominence system linked to an API 4000 LC/MS/MS spectrometer
(Applied Biosystems, Foster City, CA). The data were acquired and processed
34

using Windows 7 platform–based Analyst 1.6.1. Each of the analytes was
quantified using their signature multiple reaction monitoring (MRM) that was
identified after manual tuning of the specific compound.
Analytes were separated using a Kinetex Penta fluorophenyl (30 X 2.1 mm, 2.6
µm) column (Phenomenex Technologies Inc., Torrance, CA). The analytes were
eluted using a gradient mobile phase system consisting of two components.
Component A consisted of 0.1% formic acid in water, while component B is 100%
acetonitrile. The gradient program was set up as shown in Table 2.2.
Table 2.2: HPLC Pump gradient program for analyte separation
Time (Min) Event Parameter
0 Total flow rate 0.2 mL/min
0 Component B concentration 0%
0.40 Component B concentration 10%
0.40 Total flow rate 0.3 mL/min
5.47 Component B concentration 20
6.40 Component B concentration 95
8.80 Component B concentration 95
9.20 Component B concentration 0
13.20 Stop

A flow injection analysis was performed using the transition ions of each analyte
to maximize sensitivity. The analytes were optimized at a source temperature of
500°F, under unit resolution for quadrupoles 1 and 3. The optimal gas pressures
were as follows: collision gas, 10 psi; curtain gas, 40 psi; ion source gas (1), 10
psi; ion source gas (2), 30 psi; and Ion Spray Voltage, 5000 volts (V). The mass
spectrometer operated in the positive MRM mode with a dwell time set at 20 ms.
Analyte-specific settings were determined using Analyst software in the
quantitative optimization mode, as depicted in Table 2.3.
35

Table 2.3: Analytes, selected MRM transitions, retention time (RT) and parameter
settings- collision energy (CE), declustering potential (DP), entrance potential (EP)
and collision cell exit potential (CXP). Each quantifier MRM transition was confirmed
with a respective qualifier transition.
Analyte Q1
Mass,
u
Q3
Mass,
u
RT,
min
DP
(V)
EP
(V)
CE
(eV)
CXP
(V)
Dynamic
range
(nM)
4-pyridoxic acid 184.1 148
2.1
57 8 28.9 12.8 3.3- 1,638
5-methyltetrahydrofolate 460.1 313.2
3.1
79 5 28.4 25.1 0.2- 118
Dihydrofolate 444.3 297.1
3.4
57 7 22.7 8.6 1.0- 514
Flavin mononucleotide 457.1 439.1
3.6
73 9 23.7 13.5 1.0- 504
Folic acid 442.1 295
3.4
91 7 18.1 8.6 0.2- 104
Methotrexate 455.3 308
3.7
95 5 25.1 4.1 NA
Pyridoxal 168 149.7
1.0
57 6 15.3 11.2 3.5- 1,794
Pyridoxamine 169.1 134.1
0.7
57 5 28.9 11.2 3.2- 1,623
Pyridoxine 170.1 151.9
1.1
57 5 17.5 12.2 3.9- 1,933
Riboflavin 376.9 198
3.5
240 5 49.2 16.6 1.2- 611
Tetrahydrofolate 446.1 299.1
3.0
91 4 24.4 8.2 1.0- 234

2.2.7. Method Validation
The method was validated using the FDA guidelines for bioanalytical method
validation (Wallace et al., 2010). The full validation procedure was conducted
since this assay is a metabolomics-based one and the following parameters were
assessed: selectivity and carry-over, cross-analyte interference, limit of
quantification, standard curves, accuracy and precision, recovery and stability.
2.2.7.1. Selectivity and carry-over
Chromatograms for three QC concentrations were compared to those for six
batches of blank plasma before and after spiking. Ion traces of each specific
analyte were checked for potential interferences at the respective retention times.
The acceptance criterion was set at 20% of the lower limit of quantification
(LLOQ) peak area for the analytes and 5% of the peak area for the internal
36

standard. The carry-over effect was evaluated by injecting blank samples after
the high QC sample, and the acceptance criterion was similar to that of selectivity.
2.2.7.2. Cross-analyte interference
Each analyte was injected at the upper limit of quantification and assessed for
interference with other analytes. An interfering peak area of up to 20% of the
mean analyte peak area was deemed significant.
2.2.7.3. Limit of quantification
The concentration that gave intensity at least five-fold higher than the background
in blank plasma was selected as the LLOQ. The limits of quantification were
chosen based on their plasma reference ranges and were set about 3-16 times
beyond the reference limits. Analytes with unavailable reference ranges had that
of their primary precursors being used to guide their limits of quantitation.
2.2.7.4. Standard curves
Calibration curves consisting of eight points were calculated by linear regression,
with single-point calibration measurements instead of multi-point calibration
measurements. This approach was used to increase feasibility (Bjørk et al., 2010;
Peters and Maurer, 2007). The calibration graphs were derived by plotting the
ratio of the analyte peak area to IS versus the standard concentration. Best fit
was selected after exploration of different regression models and weighting
factors.
2.2.7.5. Accuracy and precision
Inter-run accuracy and precision were calculated for the three QC samples, with
measurements on seven experiments done on different days. Intra-run accuracy
37

was not determined due to the duration of freshly prepared standard and sample
preparation as well as run time for the samples. It rather made practical sense to
determine the inter-run accuracy instead. Inter-run accuracy and precision were
evaluated at three concentrations (low, medium and high QCs) by analyzing
replicates at each level, on different batch runs. The inter-run accuracies were
accepted if the calculated values were 15% of the nominal concentration or 20%
at the lower limit of quantification. The precision of samples was acceptable if the
percent coefficient of variation (%CV) was 15% or 20% in the case of the LLOQ.
Accuracy was defined as the percent deviation from the theoretical concentration
by quantifying QC samples with a freshly prepared calibration curve. Precision
was defined as the coefficient of variation (CV) = ([standard deviation/mean of six
measurements] X100%).
2.2.7.6. Recovery rates
Plasma components in the ionization chamber cause batch-specific ion
suppression or enhancement. The respective retention time was visually
assessed for each analyte over the six batches of blank plasma. Three sets of
samples were prepared at low, medium and high concentrations. Chromatograms
were recorded for plasma spiked pre-extraction, plasma spiked post-extraction,
and unspiked standards. In total, nine batches of blank plasma were spiked with
low, medium and high concentrations of the standards in duplicate.
Subsequently, the spiked plasma samples were extracted, as previously
described, and the dry residues were reconstituted and injected into the LCMS
38

for analysis. Recovery (RE) of an analyte was defined as the relative analyte
signal area of samples spiked post-extraction versus pre-extraction.
2.2.7.7. Sample Stability
Storage conditions and periods were chosen to mimic long-term storage of stock
solutions, plasma during freeze-thaw cycles, and plasma extract in the
autosampler awaiting analysis. QC samples were tested for stability when spiked
in plasma and taken through three freeze-thaw cycles as well as keeping the
processed sample for 6, 12, 18 and 24 hours in an autosampler at 4°C. Charcoal-
stripped pooled plasma samples were spiked with a known standard and stored
in -80°C freezer for at least 30 days. Replicates were analyzed on days 3, 7, 15
and 30.
2.2.8. Method Application
The FOCM vitamin B metabolites of ten (10) healthy volunteer plasma samples
(S1-S10) were analyzed using the validated LCMS assay. The results were
subsequently compared with published reference ranges (Fazili et al., 2008; Gori
et al., 2006; Hustad et al., 1999; Lentner, 1981; Midttun et al., 2005).
2.3. Results and Discussion
2.3.1. Method optimization
We report the development and validation of LCMS-based metabolomics assay
that is capable of simultaneously identifying and quantifying the various targeted
metabolites found in the FOCM cycle. This assay offers a significant advantage
over the microbiological approach which was the standard for determining vitamin
39

B metabolites. However, the use of bacteria mutants that are deficient in the
specific enzymes to biosynthesize these co-factors have a limitation because
metabolites need to be pooled. This method is unable to distinguish the specific
metabolite co-factors that may be responsible for various physiological or
biological changes. This LC-MS-based assay requires 50 µl of plasma and is
capable of quantifying ten analytes in the FOCM cycle.
Most observational studies assessing the association between folates and risk of
CRC either used folate intake, or quantified plasma pooled folates as determined
by microbiological assay. As stated earlier, microbiological assays are unable to
distinguish the various folate metabolites like DHF, THF, and 5MTHF. The
inability to differentiate between the concentrations of these FOCM metabolites
may confound metabolite-specific and its related reaction(s) that is responsible
for the phenotypic alteration as seen in normal versus CRC patients.
Analyte extraction is a critical factor in sample processing that ought to maintain
a consistent analyte recovery from the matrix while preserving their stability as
well. Solid-phase extraction (SPE) and liquid-liquid extraction (LLE) were both
evaluated on how they recover the FOCM analytes at low, medium and high QCs.
LLE was selected due to the consistency in extracting the increasing levels of
spiked analytes from plasma (Figure 2.2) and the less processing time required.
Also, a similar assay has successfully used liquid-liquid extraction to extract the
targeted metabolites from plasma(Jagerdeo et al., 2008).
40

Figure 2.2: Comparison of analyte extraction using LLE and SPE for low,
medium and high QCs
To stabilize the metabolites during the extraction process, we had evaluated
some stabilizing reagents including ascorbic acid, tris(2-chloroethyl) phosphate
(TCEP), and the combination (1% ascorbic acid + 1% TCEP). Amongst these,
1% ascorbic acid was found to provide the most consistent measure for the
analytes over the assessed duration of 96 hours (Figure 2.3). An amount 1%
ascorbic acid kept the B6 stable with a relative concentration between 100 and
124%. However, the other stabilizing agents resulted in B6 levels below 50%.
A n a ly te
R e la tiv e e x tra c tio n
B 2 B 6 F A 5 M T H F T H F D H F 4 P A P M P L
0
2 0
4 0
6 0
8 0
1 0 0
1 2 0
1 4 0
L L E h ig h
L L E lo w
L L E m e d iu m
S P E h ig h
S P E lo w
S P E m e d iu m
41

Figure 2.3: Stability of FOCM analytes in various stabilizing agents.
2.3.2. Method validation
2.3.2.1. Selectivity and carry-over
An LC-ESI-MS/MS assay was developed for selective detection as well as for
accurate and precise quantification of vitamin B metabolites. This assay will
establish a metabolomics platform to explore possible targets that are associated
with the FOCM during different disease conditions. The precise quantification of
the various metabolites found in the cycle can reflect the enzymatic defects
associated with a particular disease condition which one can subsequently
investigate at the genomic level. The analytes being endogenous compounds
made it very challenging to define selectivity. However, the LLOQ was set the
standard which was a thousand-fold dilution of the working solution. This
challenge was minimized by the use of filtered and charcoal-stripped human
plasma as blank though was not totally eliminated. Among the analytes who had
S ta b ility o f F A
T im e (h o u rs )
R e la tiv e c o n c e n tra tio n (% )
0 2 4 4 8 7 2 9 6 1 2 0
0
2 0
4 0
6 0
8 0
1 0 0
1 2 0
1 4 0
1 % A s c o rb ic A c id
1 % T C E P
C o m b in a tio n
S ta b ility o f T H F
T im e (h o u rs )
R e la tiv e c o n c e n tra tio n (% )
0 2 4 4 8 7 2 9 6 1 2 0
0
2 0
4 0
6 0
8 0
1 0 0
1 2 0
1 4 0 1 % A s c o rb ic A c id
1 % T C E P
C o m b in a tio n
S ta b ility o f 5 M T H F
T im e (h o u rs )
R e la tiv e c o n c e n tra tio n (% )
0 2 4 4 8 7 2 9 6 1 2 0
0
2 0
4 0
6 0
8 0
1 0 0
1 2 0
1 4 0 1 % A s c o rb ic A c id
1 % T C E P
C o m b in a tio n
S ta b ility o f B
2
T im e (h o u rs )
R e la tiv e c o n c e n tra tio n (% )
0 2 4 4 8 7 2 9 6 1 2 0
0
2 0
4 0
6 0
8 0
1 0 0
1 2 0
1 4 0 1 % A s c o rb ic A c id
1 % T C E P
C o m b in a tio n
S ta b ility o f B
6
T im e (h o u rs )
R e la tiv e c o n c e n tra tio n (% )
0 2 4 4 8 7 2 9 6 1 2 0
0
2 0
4 0
6 0
8 0
1 0 0
1 2 0
1 4 0 1 % A s c o rb ic A c id
1 % T C E P
C o m b in a tio n
42

an interfering peak in the blank plasma were B6, 4PA, and 5MTHF. However,
these peaks were subtracted as background to handle such interference (Figure
2.4). No significant carry-over was observed in blank plasma.

Figure 2.4: Chromatograms of analytes with interfering peaks in the stripped blank
plasma.

2.3.2.2. Linearity
Plasma calibration standards (n = 8 for analytes) were used to generate standard
curves and fitted using linear regression with a weighting factor of 1/x
2
for B2,
FMN, and DHF but no weighting for the other analytes. The fit was considered
acceptable if the mean calculated values of the calibration standards over the
batches for each value were 15% of the nominal values or 20% at the lower limit
of quantification.
At least six points were used to generate each standard curve. Each batch of
validation and patient samples were run with calibration curves prepared freshly
using the same concentrations (single measurements per level). For routine
analysis, the standard curve and six QC samples (3 concentrations in duplicate)
B
6 4PA
5MTHF
43

were assayed. The standard curve was accepted if at least 4 of the six calculated
concentrations of QC samples were within 20% of the nominal values, with at
least 1 QC sample passing at each concentration. Figure 2.5 shows a
representative chromatogram of the analytes. The calibration curve for each
analyte showed a high linearity with the squared correlation coefficient (r
2
) > 0.99
over the dynamic range (defined in Table 2.1).
Figure 2.5: Representative chromatogram of the analytes at concentration of
five- fold dilution of working solution

M e ta b o lite s o f V ita m in B
T im e , m in
In te n s ity , c p s
0 2 4 6 8 1 0 1 2
0
1 .0  1 0
6
2 .0  1 0
6
3 .0  1 0
6
4 .0  1 0
6
5 .0  1 0
6
5 M T H F
D H F
F A
F M N
M TX
P A
P L
P M
B
6
T H F
B
2
F o la te s
T im e , m in
In te n s ity , c p s
0 1 2 3 4 5
0
2 .0  1 0
5
4 .0  1 0
5
6 .0  1 0
5
8 .0  1 0
5
5 M T H F
D H F
F A
T H F
F la v in s
T im e , m in
In te n s ity , c p s
0 1 2 3 4 5
0
2 .0  1 0
5
4 .0  1 0
5
6 .0  1 0
5
8 .0  1 0
5
F M N
B 2
B
6
M e ta b o lite s
T im e , m in
In te n s ity , c p s
0 1 2 3 4 5
0
1 .0  1 0
6
2 .0  1 0
6
3 .0  1 0
6
4 .0  1 0
6
5 .0  1 0
6
P A
P L
P M
B 6
44

2.3.2.3. Precision and accuracy
Table 2.4 summarizes the inter-run precision (CV %) and accuracy (% deviation
from the nominal concentrations) for six replicates of QCs. All of the replicates
had a CV and accuracy of within 14% and 12.2%, respectively.
Table 2.4: Inter-run accuracy and precision (n=7 for each concentration level)
Analyte
Accuracy (%) Precision (CV%)
Low Medium High Low Medium High
4-Pyridoxic acid 91 109.9 107.4 11% 6% 6%
5-Methyltetrahydrofolate 101 104.3 103.9 10% 11% 7%
Dihydrofolate* 98.4 102.1 95.7 12% 10% 10%
Flavin Mononucleotide 96 98.4 102.3 6% 7% 6%
Folic acid* 105.3 112.2 103.7 1% 4% 8%
Pyridoxal 97.3 98.6 106.7 7% 6% 8%
Pyridoxamine 99.5 103.3 108 13% 7% 3%
Pyridoxine 95.7 105.7 100.6 14% 6% 9%
Riboflavin 101.2 101.9 101.4 10% 8% 8%
Tetrahydrofolate* 104 107.3 105.8 12% 7% 3%
*The accuracy and precision for these analytes are reported for n=5 for each concentration level.

2.3.2.4. Matrix effect and recovery
Recovery rates for the B6 metabolites and the riboflavin were approximately 80%,
however folate analytes were low (ranging from 19% to 80%), with the high QC
groups being the lowest (Table 2.5). The low and medium QCs for PL, PM,
pyridoxine, and riboflavin, had recoveries generally above 100%. The poor
recovery of the folates (especially DHF and THF) may be due to the drying step
in the sample processing during which the very unstable analytes may undergo
chemical conversions. DHF and THF are unstable folate intermediates that may
react readily with any oxygen that might have blown alongside the nitrogen
45

stream during sample drying step. This observation is similar to the finding
reported by Fazili et al. (2008).
Table 2.5: Recovery rates of analytes at various concentration ranges
Analyte Low Medium High
4-Pyridoxic acid 95% 95% 90%
5-Methyltetrahydrofolate 66% 58% 44%
Dihydrofolate 80% 17% 14%
Flavin mononucleotide 36% 37% 40%
Folic acid 54% 50% 25%
Methotrexate 46% 49% 53%
Pyridoxal 113% 98% 82%
Pyridoxamine 158% 131% 85%
Pyridoxine 106% 110% 92%
Riboflavin 148% 101% 88%
Tetrahydrofolate 37% 51% 19%

2.3.2.5. Sample stability
Processed analytes were stable over 18 hours when stored in the refrigerated
autosampler. The accuracy of analytes at the low and high QCs was within 20%
of the initial concentration (Figure 2.6). Analytes were stable for at least one
month when spiked into plasma and stored at -80°C consistent with published
data (Zheng et al., 2015). The linear regression for their respective degradation
curves showed slopes which were not significantly different from zero. Analytes
were thus assumed to be stable at least after one month of storage at -80°C. After
three freeze-thaw cycles (for both low and high QC concentrations), average
remaining levels in extracts ranging between 85-124%.
46

Figure 2.6: Stability of analytes after 18hrs of storage in refrigerated autosampler

2.3.3. Proof of Applicability
The applicability of the assay was tested on the plasma samples derived from ten
healthy volunteers not receiving vitamin supplementation. The individual plasma
concentrations of analytes, the mean plasma analyte concentrations and
standard deviations (SD) are presented in Table 2.6. Most of the plasma analyte
concentrations (greater than 80%) and all the mean (or median) values were
within the defined reference ranges, except PA, B2, and FMN which were below
the specified lower reference limits. The reference limits for B2 and FMN have
been determined with an LCMS assay that used data from 94 patients so the
range may not reflect the population appropriately (Midttun et al., 2005). PA, on
the other hand, is the final metabolite of vitamin B6 with a known high renal
clearance because it is not bound to plasma proteins (Zempleni and Kübler,
1995). The elevated renal clearance of PA may explain the undetectable plasma
PA levels in these volunteers. However, PA has been reported to be elevated in
A n a ly te
A c c u ra c y
4 P A
5 M T H F
B 2
B 6
D H F
F A
F M N
P L
P M
T H F
0
2 0
4 0
6 0
8 0
1 0 0
1 2 0
Q C L O W
Q C H IG H
47

cancers, (Galluzzi et al., 2013) so the assay was tested on some colorectal
cancer samples, and PA was detected in levels between 60- 3,161nM. The
application of the assay to human plasma samples shows its utility for routine
analysis.
Table 2.6: Plasma concentrations (nM) of analytes in healthy volunteers (n=10)

2.4 Conclusion
A sensitive multi-analyte LCMS method was developed employing protein
precipitation extraction to quantify the vitamin B metabolites in the FOCM
accurately thereby providing an alternative to quantifying them by the
microbiological assay. The analytical procedure was validated and successfully
applied on ten plasma samples from healthy volunteers. The assay has shown
enough robustness for use in clinical samples for investigations that relate to
these analytes.
Sample 4PA 5MTHF DHF FMN FA PL PM B6 B2 THF Total
Folates
S1 <3.30 11.63 <1.00 2.04 11.98 6.23 <3.20 <3.9 3.84 1.25
24.86
S2 <3.30 20.83 <1.00 5.38 10.41 107.38 <3.20 <3.9 3.46 1.85
33.09
S3 <3.30 9.07 <1.00 3.83 9.05 99.78 <3.20 <3.9 4.13 1.54
19.65
S4 <3.30 5.10 <1.00 1.89 8.63 107.74 <3.20 <3.9 2.29 0.90
14.63
S5 <3.30 8.83 <1.00 2.92 13.50 85.24 <3.20 <3.9 3.98 1.68
24.01
S6 <3.30 1.94 <1.00 0.55 16.86 86.17 <3.20 <3.9 2.86 1.74
20.54
S7 <3.30 5.05 <1.00 1.35 14.09 165.82 <3.20 <3.9 2.73 1.33
20.46
S8 <3.30 6.53 <1.00 2.93 14.90 13.88 <3.20 <3.9 4.81 2.41
23.85
S9 <3.30 7.14 <1.00 <1.00 15.84 102.86 <3.20 <3.9 3.97 0.62
23.60
S10 <3.30 7.71 <1.00 1.02 11.62 59.97 <3.20 <3.9 3.97 0.86
20.19
Median (nM)
7.43

2.04 12.74 92.98

3.90 1.43 22.07
Mean(nM) 8.38 2.43 12.69 83.51 3.60 1.42
22.49
SD 5.11 1.51 2.82 47.09 0.76 0.54
4.79
Reference
range (nM)
8.7-385 5.9-266 <6 3.3-13.4 <561 2.5-300 <202 <60 4.9-38.4 < 28.9 8.2-642
48

Chapter 2: Bibliography and References
Bailey, L. B., Rampersaud, G. C., and Kauwell, G. P. A. (2003). Folic acid
supplements and fortification affect the risk for neural tube defects, vascular
disease and cancer: evolving science. J Nutr 133, 1968.
Benito, E., Stiggelbout, A., Bosch, F., Obrador, A., Kaldor, J., Mulet, M., and
Munoz, N. (1991). Nutritional factors in colorectal cancer risk: a case ‐control
study in Majorca. International journal of cancer 49, 161-167.
Berry, R. J., Li, Z., Erickson, J. D., Li, S., Moore, C. A., Wang, H., Mulinare, J.,
Zhao, P., Wong, L.-Y. C., and Gindler, J. (1999). Prevention of neural-tube
defects with folic acid in China. New England journal of medicine 341, 1485-
1490.
Bjørk, M. K., Nielsen, M. K., Markussen, L. Ø., Klinke, H. B., and Linnet, K.
(2010). Determination of 19 drugs of abuse and metabolites in whole blood by
high-performance liquid chromatography–tandem mass spectrometry. Analytical
and bioanalytical chemistry 396, 2393-2401.
Boushey, C. J., Beresford, S. A., Omenn, G. S., and Motulsky, A. G. (1995). A
quantitative assessment of plasma homocysteine as a risk factor for vascular
disease: probable benefits of increasing folic acid intakes. Jama 274, 1049-1057.
Boyapati, S. M., Bostick, R. M., McGlynn, K. A., Fina, M. F., Roufail, W. M.,
Geisinger, K. R., Hebert, J. R., Coker, A., and Wargovich, M. (2004). Folate
49

intake, MTHFR C677T polymorphism, alcohol consumption, and risk for sporadic
colorectal adenoma (United States). Cancer Causes & Control 15, 493-501.
Collaboration, H. S. (2002). Homocysteine and risk of ischemic heart disease
and stroke: a meta-analysis. Jama 288, 2015-2022.
de Vogel, S., Schneede, J., Ueland, P. M., Vollset, S. E., Meyer, K., Fredriksen,
Å., Midttun, Ø., Bjørge, T., Kampman, E., and Bretthauer, M. (2011). Biomarkers
related to one-carbon metabolism as potential risk factors for distal colorectal
adenomas. Cancer Epidemiology Biomarkers & Prevention 20, 1726-1735.
Fazili, Z., Pfeiffer, C. M., Zhang, M., Jain, R. B., and Koontz, D. (2008). Influence
of 5, 10-methylenetetrahydrofolate reductase polymorphism on whole-blood
folate concentrations measured by LC-MS/MS, microbiologic assay, and bio-rad
radioassay. Clinical chemistry 54, 197-201.
Galluzzi, L., Vacchelli, E., Michels, J., Garcia, P., Kepp, O., Senovilla, L., Vitale,
I., and Kroemer, G. (2013). Effects of vitamin B6 metabolism on oncogenesis,
tumor progression and therapeutic responses. Oncogene 32, 4995-5004.
George, L., Mills, J. L., Johansson, A. L., Nordmark, A., Olander, B., Granath, F.,
and Cnattingius, S. (2002). Plasma folate levels and risk of spontaneous
abortion. Jama 288, 1867-1873.
Giovannucci, E., Stampfer, M. J., Colditz, G. A., Hunter, D. J., Fuchs, C., Rosner,
B. A., Speizer, F. E., and Willett, W. C. (1998). Multivitamin use, folate, and colon
50

cancer in women in the Nurses' Health Study. Annals of Internal Medicine 129,
517-524.
Gori, A. M., Sofi, F., Corsi, A. M., Gazzini, A., Sestini, I., Lauretani, F., Bandinelli,
S., Gensini, G. F., Ferrucci, L., and Abbate, R. (2006). Predictors of vitamin B6
and folate concentrations in older persons: the InCHIANTI study. Clinical
chemistry 52, 1318-1324.
Group, M. V. S. R. (1991). Prevention of neural tube defects: results of the
Medical Research Council Vitamin Study. The lancet 338, 131-137.
Hustad, S., Ueland, P. M., and Schneede, J. (1999). Quantification of riboflavin,
flavin mononucleotide, and flavin adenine dinucleotide in human plasma by
capillary electrophoresis and laser-induced fluorescence detection. Clinical
chemistry 45, 862-868.
Jagerdeo, E., Montgomery, M. A., LeBeau, M. A., and Sibum, M. (2008). An
automated SPE/LC/MS/MS method for the analysis of cocaine and metabolites
in whole blood. Journal of Chromatography B 874, 15-20.
Keku, T., Millikan, R., Worley, K., Winkel, S., Eaton, A., Biscocho, L., Martin, C.,
and Sandler, R. (2002). 5, 10-Methylenetetrahydrofolate reductase codon 677
and 1298 polymorphisms and colon cancer in African Americans and whites.
Cancer Epidemiology Biomarkers & Prevention 11, 1611-1621.
51

Larsson, S. C., Giovannucci, E., and Wolk, A. (2006). Folate intake, MTHFR
polymorphisms, and risk of esophageal, gastric, and pancreatic cancer: a meta-
analysis. Gastroenterology 131, 1271-1283.
Lentner, C. (1981). Geigy Scientific tables. Vol. 1, Units of measurement, body
fluids, composition of the body, nutrition: Ciba-Geigy Basle).
Midttun, Ø., Hustad, S., Solheim, E., Schneede, J., and Ueland, P. M. (2005).
Multianalyte quantification of vitamin B6 and B2 species in the nanomolar range
in human plasma by liquid chromatography–tandem mass spectrometry. Clinical
Chemistry 51, 1206-1216.
Peters, F. T., and Maurer, H. H. (2007). Systematic comparison of bias and
precision data obtained with multiple-point and one-point calibration in six
validated multianalyte assays for quantification of drugs in human plasma.
Analytical chemistry 79, 4967-4976.
Ray, J. G. (1998). Meta-analysis of hyperhomocysteinemia as a risk factor for
venous thromboembolic disease. Arch Intern Med 158, 2101-2106.
Seshadri, S., Beiser, A., Selhub, J., Jacques, P. F., Rosenberg, I. H., D'Agostino,
R. B., Wilson, P. W. F., and Wolf, P. A. (2002). Plasma homocysteine as a risk
factor for dementia and Alzheimer's disease. N Engl J Med 346, 476-483.
52

Terry, P., Jain, M., Miller, A. B., Howe, G. R., and Rohan, T. E. (2002). Dietary
intake of folic acid and colorectal cancer risk in a cohort of women. International
journal of cancer 97, 864-867.
Vollset, S. E., Refsum, H., Irgens, L. M., Emblem, B. M., Tverdal, A., Gjessing,
H. K., Monsen, A. L., and Ueland, P. M. (2000). Plasma total homocysteine,
pregnancy complications, and adverse pregnancy outcomes: the Hordaland
Homocysteine study. Am J Clin Nutr 71, 962-968.
Wallace, K., Grau, M., Levine, A., Shen, L., Hamdan, R., Chen, X., Gui, J., Haile,
R., Barry, E., Ahnen, D., and al., e. (2010). Association between Folate Levels
and CpG Island Hypermethylation in Normal Colorectal Mucosa Cancer Prev
Res (Phila) 3, 1552-1564.
Zempleni, J., and Kübler, W. (1995). Metabolism of vitamin B 6 by human kidney.
Nutrition Research 15, 187-192.
Zheng, X.-H., Jiang, L.-Y., Zhao, L.-T., Zhang, Q.-Y., and Ding, L. (2015).
Simultaneous quantitation of folic acid and 5-methyltetrahydrofolic acid in
human plasma by HPLC–MS/MS and its application to a pharmacokinetic
study. Journal of Pharmaceutical Analysis 5, 269-275.
53

Chapter 3: Identification of Blood-based Biomarkers as Potential
Screen for Colorectal Cancer Using Metabolomics
3.1. Introduction
CRC is the third leading cause of cancer-related death in the industrialized world
(Siegel et al., 2016). In the United States, CRC is third most commonly diagnosed
cancer, accounting for approximately 50,000 deaths annually. The annual
expenditure for CRC treatment is estimated to be $5.5-$6.5 billion, where
inpatient hospital care accounts for about 80% of the total cost (Seifeldin and
Hantsch, 1999).
CRC evolves as a consequence of dysregulated intestinal cell proliferation. Early
diagnosis of CRC requires highly sensitive and precise detection of cellular
transformational process or while the disease is still in the initial stages exhibiting
only local spread. Immediate effective treatment of localized disease is most
likely to lead to curative modality (Gupta et al., 2008). Unfortunately, most CRC
cases are diagnosed when they are late disease stages with metastasis, making
it harder to achieve complete remission. Furthermore, metastatic disease is
associated with low five-year survival despite aggressive treatment strategies
(Burt et al., 2010; O’Connell et al., 2004). The development of reliable and
predictive biomarkers would be a critical tool to identify individuals with evolving
colorectal cancer (CRC) or presence of early disease.
54

The ability to detect evolving tumorigenesis during transition from anaplasia to
neoplasia may open the door for preemptive interventions. Also, the capacity to
identify early stages of the disease can significantly improve clinical outcomes
after immediate and aggressive therapy. Colonoscopy is currently the choice for
screening and detecting CRC; however, this invasive procedure is not widely
accepted and often avoided. The development of a blood-based molecular
biomarker(s) that can reliably identify the potential of CRC development may
significantly improve screening compliance. To facilitate the development of such
test, an understanding of the aberrant mechanisms underlining the disease can
enable the biomarker development process.
Folates are pteroyl glutamates that serve as the primary methyl carriers essential
for two paths of methyl trafficking in the cell. This includes DNA replication or
repair and formation of S-adenosyl methionine (SAM) that serves as the primary
methyl group donor for the transmethylation reactions (de Vogel et al., 2011).
Specific folate metabolites are also involved in genomic stability processes.
Folate one-carbon metabolism (FOCM) cycle regulates oncogenes and tumor
suppressors involved in cancers. This regulation is accomplished by controlling
the DNA synthesis and methylation. Dysregulation of these critical enzymes
found in the FOCM have been associated with increased risk for CRC
development, thus making this metabolic cycle a good target to probe for in the
development of a CRC biomarker.
55

Epidemiological studies have evaluated the association between folates in
relations with risk for CRC. Most have found an association between increasing
folates intake and decreasing risk of CRC (Giovannucci et al., 1998; Pufulete et
al., 2003; Pufulete et al., 2005; Stevens et al., 2011). Unlike these studies, this
study focuses on the metabolites found in the FOCM to determine whether their
imbalance may be a consequence of cellular transitioning. The most reflective
measure for FOCM metabolites would be the cellular concentrations, but cellular
samples in the affected tissues are difficult to obtain. However, cellular
concentrations may be substituted with the extracellular concentration as a
surrogate. Their levels can be extrapolated to determine the association since
there is usually a homeostatic flux between the two systems.
Metabolomics is a powerful tool useful for exploring metabolites (with molecular
weight <1800 Da) to phenotype the biological system in question (Fiehn, 2002;
Nicholson et al., 1999). It has been extensively used in biomarker discovery to
facilitate disease diagnosis (Madsen et al., 2010) and mechanistic dissection of
disease pathophysiology (Li et al., 2008). Metabolomics approaches can take a
targeted, untargeted or a combination of both. Untargeted metabolomics is
commonly considered to facilitate the measure of all endogenous metabolites in
the biological samples. One advantage of untargeted analysis is that it offers the
opportunity to identify novel targets that may be difficult to identify and
characterize. However, the specificity of the metabolites that are detected is
dependent on the analytical platform. The limitation of this approach is that
analyte of high-abundance are more likely to be detected (Roberts et al., 2012).
56

In contrast, the targeted approach quantifies defined analytes of compatible
chemical characteristics and biochemical annotation. This approach optimizes
the quantification of analytes of interest thereby reducing any interference from
predominant analytes found in the sample. With the emergence of liquid
chromatography mass spectrometry (LCMS) -based metabolomics, it is possible
to profile and even quantify the analytes found in a pathway.
In this study, we used the untargeted metabolomics followed by the targeted
approach to identify biomarkers that may be predictive in determing patients with
CRC and compared with controls. These biomarkers will facilitate early detection,
intervention and decrease CRC-related deaths.
3.2. Materials and Methods
3.2.1. Study samples
Plasma samples from CRC patients (26) were compared with healthy controls
(10) in this study. Samples were collected from participants and stored at -80
o
C
until analysis. Table 3.1 presents the summary of the demographic characteristics
of the patients in this study.
Table 3.1: Demographic characteristics of plasma sample donors
Cases (n=26) Controls (n=10) p-value
Age mean(SD) 64 (13) 41(16) 0.001
Male 50% 50%
Tumor stage
Polyp 1 (4%)
Stage 1 2 (8%)
Stage 2 7 (27%)
Stage 3 6 (23%)
Stage 4 10 (38%)

57

3.2.2. Chemicals and reagents
Analytical grades of flavin mononucleotide (FMN), folic acid (FA), pyridoxine (B6),
pyridoxal (PL) hydrochloride, pyridoxamine (PM) dihydrochloride, 4-pyridoxic
acid (4PA), zinc sulphate, and tris(2-carboxyethyl) phosphine (TCEP) were
purchased from Sigma (St Louis, MO, USA); riboflavin (B2) was purchased from
Alfa Aesar, and 5-methyltetrahydrofolate (5MTHF), dihydrofolate (DHF) and
tetrahydrofolate (THF) were purchased from Cayman Chemicals (Ann Arbor, MI,
USA). Methotrexate (MTX), purchased from Enzo Life Sciences (Farmingdale,
NY, USA), was used as an internal standard for the assay. The purity of each
standard was above 97%, except DHF and THF which were 90% and 95%
respectively. Ultrapure HPLC-grade water, LCMS-grade methanol, LCMS-grade
acetonitrile and formic acid were purchased from Fisher Scientific (Pittsburg, PA,
USA) and were used for sample processing and mobile phase systems. In
addition, homocysteine (HCY), methionine (METH), S-adenosyl methionine
(SAM), S-adenosyl homocysteine (SAH), deuterated SAH, cystathionine
(CYSTH), methylmalonic acid (MMA), and tris(2-carbixyethyl)phosphine (TCEP)
were bought from Sigma (St Louis, MO, USA). The deuterated MMA was bought
from Medical Isotopes (Pelham, NH, USA).
3.2.3. Sample preparation
To human plasma (50 µL), 50 µL of 30ng/ml MTX was added and thoroughly
mixed. The plasma proteins were precipitated with the addition of 400 µL of cold
20% 0.2M ZnSO4 in methanol and kept at -20
o
C for about 30 min. The
sample was then centrifuged at 15,000 rpm for 15 min at 4°C, after which 400µL
58

of the supernatant solution was transferred into a new micro centrifuge tube and
evaporated to dryness using a steady stream of dried and filtered nitrogen gas
at room temperature. The residue reconstituted using 50 μL of 1% ascorbic acid,
where 20 μL was injected onto an LC-MS system linked to a reverse-phase
column (Phenomenex Technologies Inc., Torrance, CA). The analytes were
separated using gradient mobile phase system consisting of two components.
Component A consisted of 0.1% formic acid in water, while component B is 100%
Acetonitrile.
3.2.4. LCMS data acquisition for targeted and untargeted scan
Analytes of the FOCM include Vitamin B2, FMN, B6, 4PA, PL, PM, FA, DHF, THF,
and 5MTHF were targeted during this scan. The samples were analyzed with the
targeted approach using a reverse phase Kinetex Penta fluorophenyl (PFP, 30 X
2.1 mm, 2.6 µm) column (Phenomenex Technologies Inc., Torrance, CA) using
the gradient program in Table 2.2.
The analytes were quantified using an LCMS system consisting of Shimadzu
Prominence HPLC system linked to a Sciex API 4000 LC/MS/MS spectrometer
(Applied Biosystems, Foster City, CA). The data were acquired and processed
using Windows 7 platform–based Analyst 1.6.1. The MS was set in the positive
mode, where settings were as follows: source temperature of 500°F, under unit
resolution for Q1 and Q3. The optimal gas pressures were as follows: collision
gas, 10 psi; curtain gas, 40 psi; ion source gas (1), 10 psi; ion source gas (2), 30
psi; and Ion Spray Voltage, 5000 volts (V).
The untargeted analysis was conducted with similar MS settings, but
59

separation was achieved with a longer reverse phase Kinetex PFP 100A (75 X
3.0 mm, 2.6 µm) column (Phenomenex Technologies Inc., Torrance, CA) at an
extended 33-min gradient. This longer version of the separation was preferred to
facilitate better acquisitions of the features and lessen overlap of features
acquired at a retention time.
3.2.5 Quantitation of plasma methylmalonic acid, homocysteine, and related
metabolites
The plasma levels of MMA, HCY and its related metabolites in the FOCM were
determined using modified procedures as described by Kořínek et al. (2013) and
Fu et al. (2013). The MMA, HCY, METH, SAM, SAH and CYSTH were quantified
using these assays. To 50 µL of plasma, 50 µL of 30ng/ml deuterated SAH
solution, 25 µL of 200 ng/ml deuterated MMA and 25 µL of 0.1M TCEP was added
while working on ice. Ice-cold precipitation solution (350 µL) made up of 20% 0.2
M ZnSO4 in methanol was added, vortexed for 30 seconds and stored at -20°C
for 30 minutes. The sample was then centrifuged at 15,000 rpm for 15 minutes at
4°C and 50 µL of the solution transferred into HPLC vials for the injection of 30
µL unto a reverse phase Shimadzu C18 (50 X 4.6mm, 3 µm) column for
LC/MS/MS analysis. The mobile phase consisted of 0.1% formic acid in water as
Component A and 0.1% formic acid in methanol as Component B running at the
following gradient conditions: starting at 20% of component B, it was maintained
for 1.2min followed by a linear increase to 80% of component B B within 1.3 min,
and then it kept at the same condition for 1.5 min. It was followed by a declined
to 20%B within 0.5 min. The condition of the column was recovered with a 1.5-
60

min run of 20% of component B. The samples were analyzed using an LCMS
system comprising of Shimadzu Prominence HPLC system linked to an API 4000
LC/MS/MS spectrometer (Sciex, Foster City, CA) operating in the positive mode.
MS setting was the following: source temperature of 350°F; collision gas, 10 psi;
curtain gas, 25 psi; ion source gas (1), 40 psi; ion source gas (2), 30 psi; and Ion
Spray Voltage, 5500 V.
About 300 µL of the remaining supernatant was transferred into a clean 1.5 mL
Eppendorf tube and dried under nitrogen gas. The residue was reconstituted into
30uL of 15% methanol in water and transferred into HPLC vials for injection of 20
µL unto Gemini C18 (150 X 4.6 mm, 3 µm) column (Phenomenex Technologies
Inc., Torrance, CA). During analysis, the LCMS/MS operated in the negative
mode with the following settings: source temperature, 350°C; collision gas, 12
psi; curtain gas, 40 psi; ion source gas (1), 50 psi; ion source gas (2), 20 psi; and
Ion Spray Voltage, -3000 V. Chromatographic separation was achieved with an
18.5-min gradient mobile phase system consisting of 0.1% formic acid in water
as Component A and 0.1% formic acid and 10mM ammonium formate in
methanol as Component B. The gradient was as follows: starting at 15% of
component B, it was increased to 95% of component B within 6.2 min, and then
it kept at the same condition for 3.1 min. It was followed by a declined to 15% of
component B within 0.2 min. The condition of the column was recovered with an
8-min run of 15%B.
61

3.2.6. Data acquisition and processing
3.2.6.1. Untargeted Scan
Plasma extracts from samples were analyzed using the LCMS to obtain a Q3
scan with an integrated data acquisition (IDA) criterion which triggered a product
ion scan for the top four hits. The acquired data was loaded into the MarkerView
TM

software (AB Sciex) setting a threshold of 100,000 counts per second and
generating 5,000 features which were present in at least five samples. Peak
finding options were set as follows: subtraction offset, 10 scans; subtraction
multiplication factor, 1.3; noise threshold, 100,000; minimum spectral peak width,
0.4 Da, minimum retention time peak width, two scans and maximum retention
time, 22 mins. Peak alignment options were set as follows: retention time
tolerance, 0.5 min; mass tolerance, 0.4 Da and maximum number of peaks, 5000.
If peaks were found in fewer than five of the samples (17% of all samples), this
feature was automatically discarded using a filter setting of MarkerView
TM
. Using
raw data, peak area integration was performed on each feature normalizing peak
areas with the internal standard.
3.2.6.2. Targeted Scan
Extracted samples were injected into the LCMS to quantify the targeted FOCM
metabolites. The data from the untargeted scan informed the metabolites whose
MRM was targeted. The acquired data was loaded into the MultiQuant 2.1
software setting peak options as follows: Gaussian smooth width, 2 points;
retention time half window, 30 sec and minimum peak width, 3 points. The peak
integration parameters had a 40% noise percentage, 2 min baseline sub-window
62

and a 2-point peak splitting. Using plasma calibration curves, the level of each
metabolite was quantified, normalizing analyte peak areas with that of the internal
standard. The actual concentrations of the analytes were imported into
MarkerView
TM
software for further statistical analysis.
3.2.7. Statistical analysis
All analyses were performed using the MarkerView
TM
and SAS 9.4 (SAS Institute
Inc., Cary, NC) software. The data was log transformed and further normalized
using Pareto normalization settings found in the software. Supervised PCA-DA
was conducted on the samples and Wilcoxon rank sum test was performed on
the acquired data at 5% level of significance (Hoogerbrugge et al., 1983). The
kernel smoothing model was used to calculate the misclassification rate for the
CRC cases and controls based on the assay (Lachenbruch and Mickey, 1968).
The features/ ions that were significantly different in CRC cases and controls and
had more than a 100-fold change in mean intensity for the groups were extracted.
These molecular weights of these features were matched with Human
Metabolome Database (HMDB; http://www.hmdb.ca) for identification of features
of interest. The mass error window for the search was set to 0.2 Da and the
search results manually screened for endogenous metabolites. Features which
did not match any endogenous database entry were not considered for further
investigation.
The plasma concentrations of FOCM analytes in the targeted scan were
compared between the CRC cases and controls using a two-tailed Wilcoxon rank
sum test. Further clinically meaningful ratios of metabolites which may give an
63

index of enzymatic activity were also generated and compared between the
groups. The ratio of product to reactant metabolites was used to estimate the
enzymatic function of some key enzymes in the FOCM. The polyp sample was
added to the CRC cases to facilitate analysis. The analyte concentrations that
were below the lower limit of quantitation (LLOQ) were substituted with the LLOQ
for the analysis. Analyses were done with and without possible outlier sample
analyte concentrations.
3.3. Results
3.3.1. Data reduction and exploratory analyses
Principal Component Analysis (PCA) was used for initial data exploration. This
was accomplished by annotating LCMS peaks in an unsupervised and
supervised fashion. For both targeted and untargeted datasets, there was better
separation in the supervised as compared to unsupervised PCA. In this analysis,
controls were clearly clustered from the cases even in the unsupervised PCA.
Healthy controls (blue) were well separated from the CRC cases (red) as shown
in Figure 3.1. In the targeted, the first two principal components (PC) explained
41.6% of total variance while the first three PC explained 51.1% of total variance
showing how representative the PCA plot is of the original data. When a second
stage discriminant analysis (DA) produced plots with discriminants that explained
100% of the variation, for both targeted and untargeted.
Subsequently, a nonparametric discriminant analysis model was fitted to the PC
of the untargeted (first four PCs) and the targeted (first six PCs) using kernel
64

smoothing to classify each data batch into CRC and controls. The models were
validated by re-substitution and cross-validation approaches (Lachenbruch and
Mickey, 1968). Both batches of data (targeted and untargeted) produced similar
misclassification rates (both total error rate and group error rates). The re-
substitution approach yielded a misclassification error of 0% while that of cross-
validation was 2.8%. Unlike the cross-validation approach, the re-substituted
data are usually part of the model building in the re-substitution approach thereby
decreasing the misclassification error.
Unlike the re-substitution, the leave-one-out cross-validation approach seems
more reliable in evaluating the accuracy of predicting group membership of a
sample by the classifying variables. The procedure is based on repeatedly
withholding one sample at a time, and the complementary training set is used for
the prediction error estimation. The misclassification or prediction error is
calculated by the rate of misclassified samples when predicting for each sample
using the training set. This procedure is repeated, leaving out each patient at a
time until all patients have been classified and then averaging the prediction error
rates over all the possible training sets.
65

Figure 3.1: PCA-DA plots showing the clustering the CRC (red) and controls (blue)
using either (a&c) untargeted scan or (b&d) targeted scan. Panels a&b shows the
loadings of the samples while the right (c&d) shows the loadings of the metabolites.
3.3.2. Putative biomarkers from untargeted approach
We selected 1000 monoisotopic positive ion masses after running a Wilcoxon
rank sum test on the 5000 extracted features. The log (p-values) and the log (fold
change) of the all the extracted features are represented in the volcano plot
(Figure 3.2). Feature selection was conducted to select the best predictors for
CRC screening and possibly diagnoses. If a feature which is greater than 100 Da
showed significant difference between CRC and controls patients assessed by
satisfactory p-value (p-values > 0.05) and a fold change greater than 100, the
feature was chosen for identification through matching in the metabolome
a
b
c
d
66

database annotation. Table 3.2 presents 30 candidate features manually picked
as putative biomarkers after matching with endogenous human metabolome
database. Most of the putative biomarkers (20 out of 30) that differentiated
between CRC patients and healthy controls were involved in FOCM, suggesting
a link to the methylation and nucleotide synthesis processes. This supported the
analytes selection for the targeted approach.

Figure 3.2: Volcano plot of controls versus CRC cases

67

Table 3.2: Candidate features whose mean intensities are significantly different in CRC
cases and controls with respective match in the Human Metabolome Database
Mass-to-
charge ratio
[MH
+
]
p-value* Variation of mean
intensities of feature
to control**
Matched metabolites in database
106.0 0.025 ↑ Serine
136.0 0.022 ↑ Homocysteine, Adenine, Methyl cysteine
147.0 0.015 ↑ Adipic Acid, Dimethyl succinic Acid, Glutamine, Acetylcholine
147.1 <0.001 ↓ Lysine
149.0 <0.001 ↑ 3-Methylmalic Acid
161.0 <0.001 ↑ Methacholine
164.2 <0.001 ↑ Dimethyl aminopurine, Pterin
167.0 0.003 ↑ Methyl xanthine
168.0 <0.001 ↑ Homocysteinesulfinic Acid, Pyridoxal
169.0 0.046 ↑ Pyridoxamine
173.2 0.003 ↑ Glycerol-3-Phosphate
194.0 0.001 ↑ Methyl hippuric Acid
223.0 0.003 ↑ Cystathionine
240.0 0.001 ↑ Dihydrobiopterin
242.0 0.004 ↑ Tetrahydrobiopterin
243.0 0.003 ↑ Thymidine
261.0 <0.001 ↑ Glucose-6-Phosphate
268.0 0.003 ↑ Adenosine or Deoxyguanosine
306.0 0.006 ↑ Methionyl-Arginine or Arginyl-Methionine
309.0 0.001 ↑ Deoxyuridine Monophosphate
345.2 0.002 ↑ Thiamine Monophosphate, Difluorodeoxyuridine
Monophosphate
456.2 0.040 ↑ 5,10-Methenyltetrahydrofolate
457.2 0.006 ↑ Flavin Mononucleotide
460.2 0.007 ↑ 5-Methyltetrahydrofolate
468.8 0.005 ↑ Deoxyuridine Triphosphate
538.6 0.029 ↓ Ceramide
442.2 0.017 ↓ Folic Acid
746.2 0.002 ↓ Reduced Nicotinamide Adenine Dinucleotide Phosphate
377.0 <0.001 ↑ Riboflavin
266.0 0.042 ↑ Thiamine
* Student t-test was used to analyze the difference in means of cases and controls
**The arrows ↑ and ↓ indicate increase and decrease of mean feature intensity in the plasma of CRC patients
as compared to healthy controls, respectively

68

3.3.3. Targeted metabolites
FOCM metabolites levels between the CRC cases and controls are summarized
in Table 3.3. Statistical significance was revealed using Wilcoxon rank test
demonstrating patients with CRC had significantly higher plasma concentrations
of THF, 5MTHF, PA and PL compared to healthy controls (Figure 3.3). However,
the plasma concentrations of B2, FA, SAM as well as methylation capacity
(defined by SAM/SAH ratio), FA/THF ratio and FA/5MTHF ratio were significantly
higher in the controls instead. There was a trend in the distribution of folate
metabolites showed proportionally more reduced folates in cases but more folic
acid in controls (Figures 3.4). These results did not differ after censuring the
possible outliers.

Figure 3.3: Box plots showing the plasma concentrations of a) THF b) 5MTHF c)
FA d) B2 e) 4PA and f) PL in CRC cases and controls.

5 -M e th y lte tr a h y d r o fo la te
C R C s ta tu s
A n a ly te C o n c e n tra tio n (n M )
C a s e s C o n tr o ls
0
5 0
1 0 0
1 5 0
p = .0 0 2
R ib o fla v in
C R C s ta tu s
A n a ly te C o n c e n tra tio n (n M )
C a s e s C o n tr o ls
0
1 0
2 0
3 0
4 0
p = .0 0 2
4 -P y rid o x ic A c id
C R C s ta tu s
A n a ly te C o n c e n tra tio n (n M )
C a s e s C o n tr o ls
0
5 0
1 0 0
1 5 0
p = .0 0 3
P y r id o x a l
C R C s ta tu s
A n a ly te C o n c e n tra tio n (n M )
C a s e s C o n tr o ls
0
2 0 0
4 0 0
6 0 0
8 0 0
1 0 0 0 p = .0 2
F o lic A c id
C R C s ta tu s
A n a ly te C o n c e n tra tio n (n M )
C a s e s C o n tr o ls
0
5
1 0
1 5
2 0
p = .0 0 9
T e tr a h y d r o fo la te
C R C s ta tu s
A n a ly te C o n c e n tra tio n (n M )
C a s e s C o n tr o ls
0
5
1 0
1 5
2 0
p < .0 0 0 1
a b
c
d e f
69

Table 3.3: Comparison of the plasma FOCM metabolites levels in CRC cases and
controls using the median and the interquartile range (IQR)
Metabolite/ Ratio
of metabolites
Controls (n=10) CRC Cases (n=26) p-value**
Median IQR Median IQR
5MTHF (nM) 7.4 5.1- 9.7 23.2 15.4- 66.7 0.001*
B2 (nM) 3.9 2.8- 4.0 1.2 1.2- 1.2 0.002*
CYSTH (nM) 303.2 152.4- 1089.7 261.4 111.0- 476.8 0.447
FA (nM) 12.7 10.1- 15.1 9.7 8.4- 11.1 0.009*
FMN (nM) 2.0 1.0- 3.2 1.00 1.0- 8.4 0.575
HCY (uM) 9.6 8.2- 11.6 8.6 6.8- 12.0 0.621
METH (uM) 29.8 21.8- 39.0 24.5 16.3- 32.1 0.230
MMA (nM) 239.80 217.1- 272.7 394.0 268.0- 496.1 <0.001*
PA (nM) 3.3 3.3- 3.3 12.3 3.3- 51.5 0.002
PL (nM) 93.0 48.4- 107.5 152.7 85.9- 294.5 0.020*
SAH (nM) 2.5 1.3- 8.4 7.5 1.3- 16.0 0.126
SAM (nM) 580.9 311.4- 1074.1 88.3 8.4- 486.4 0.011*
THF (nM) 1.4 0.9- 1.8 5.5 3.4- 10.5 <0.001*
Total folates (nM) 23.1 21.1- 25.2 40.1 31.0- 88.4 <0.001*
5MTHF/folates 0.3 0.3-0.4 0.6 0.4- 0.8 0.005*
5MTHF/THF 5.8 3.5- 9.8 5.9 1.8- 11.6 0.832
FA/5MTHF 1.6 1.0- 2.4 0.3 0.1- 0.6 <0.001*
FA/folates 0.6 0.5- 0.6 0.2 0.1- 0.3 <0.001*
FA/THF 9.6 6.1- 11.3 1.5 0.9- 2.5 <0.001*
HCY/CYSTH 30.0 12.1- 66.4 38.7 19.0- 99.9 0.289
HCY/METH 0.4 0.2- 0.4 0.4 0.3- 0.7 0.437
SAM/SAH 286.9 94.1- 538.8 7.9 0.7- 150.1 0.018*
THF/folates 0.06 0.05-0.08 0.12 0.07- 0.26 0.009*
** Wilcoxon rank sum test was used to analyze the significant difference between the cases and
controls
* Significant difference between median of cases and controls (p-value<0.05)

70

Figure 3.4: Box plots showing a) plasma concentration of SAM b) Methylation
capacity c) plasma concentration of MMA d) ratio of FA to THF conversion
[indicative of DHFR activity]; Folate distribution (e) and normalized folate
distribution in CRC cases and controls (f) are shown with stacked column charts.
3.4 Discussion
CRC is the third leading cause of cancer deaths, accounting for about 150,000
deaths annually. Most of these deaths are CRC patients who diagnosed with late
stages of disease. A blood-based biomarker will be a critical tool in identifying
patients who are at risk or in early stages of CRC, where the ability to identify
disease at early stages may reduce CRC-related deaths. This study presents an
approach that has identified putative biomarkers that may be useful to screen for
CRC.
The combined power of metabolomics and LCMS make it feasible to phenotype
CRC patients and distinguish healthy controls using metabolites as biomarker.
The untargeted approach has shown the power to explore the metabolites that
S -A d e n o s y l M e th io n in e
C R C s ta tu s
A n a ly te C o n c e n tra tio n (n M )
C a s e s C o n tr o ls
0
5 0 0
1 0 0 0
1 5 0 0
2 0 0 0
2 5 0 0
p = .0 1 1
M e th y la tio n C a p a c ity (S A M /S A H R a tio )
C R C s ta tu s
R a tio o f c o n c e n tra tio n
C a s e s C o n tr o ls
0
5 0 0
1 0 0 0
1 5 0 0
2 0 0 0
p = .0 1 8
F o lic A c id /T e tra h y d ro fo la te R a tio
C R C s ta tu s
R a tio o f c o n c e n tra tio n
C a s e s C o n tr o ls
0
1 0
2 0
3 0 p < .0 0 0 1
M e th y lm a lo n ic A c id
C R C s ta tu s
C o n c e n tra tio n (n M )
C a s e s C o n tr o ls
0
2 0 0
4 0 0
6 0 0
8 0 0
p = .0 0 1
0%
20%
40%
60%
80%
100%
Controls Cases
Folate Distribution Profile
FA DHF THF MTHF
a
b
c
d
e
f
71

differentiate between CRC and controls. This approach presents another
challenge with the identification of the differentiating metabolites as putative
biomarkers. Although database matching can be used to make an informed
guess on the metabolite’s identity, the approach leaves an uncertainty gap that
must be bridged subsequently with targeted identification approaches. The
targeted metabolomics approach, however, focuses on the main metabolic
pathway(s) that may have an underlying molecular mechanism to explain the
development and progression of the disease. In this study, the differentiating
metabolites for CRC obtained from the untargeted scan guided our focus on the
FOCM pathway where we selected key metabolites for the targeted analysis.
The FOCM pathway is the primary biological cycle that controls the
transmethylation and DNA nucleotide synthesis- both cycles are altered
significantly in CRC (Hanley and Rosenberg, 2015; Kim, 2004). The use of this
approach to probe the FOCM pathway holds a promise for the for the discovery
of individual biomarkers (Kind et al., 2007) and novel drug targets for drug
development. In this pilot study using a limited number of samples, we have
demonstrated that plasma metabolomics analysis has the capability of clustering
CRC cases from controls.
The analysis of the data from the pilot study reveal differences in the plasma
FOCM metabolite levels of CRC cases and controls. The trends in observation
seem similar to the ‘methyl trapping’ phenomenon that occur during a defect,
deficiency or downregulation of a required enzyme or metabolite which prevents
the methyl groups being transferred in the transmethylation reaction process.
72

FOCM is one of the critical homeostatic biochemical processes that modulates
the transmethylation, transsulfuration, and nucleotide synthesis cycles.
Whenever there is an alteration in any of these cyles, the cells respond in a way
that reflects in the phenotypic metabolite concentration in plasma.
The methyl trapping hypothesis has existed over four decades (Herbert and
Zalusky, 1962). Methyl trapping is a feedback mechanism in which low SAM
levels stimulate MTHFR activity, thereby promoting 5MTHF formation to support
the methylation reactions. However, if methionine synthase (MS) activity is low,
due to cofactor deficiency or MS inhibition, this preferential 5MTHF formation is
counterproductive, as 5MTHF becomes pooled metabolically. The 5MTHF can
proceed only into the forward reaction with MS to produce more THF and its
derivatives which are used for nucleotide synthesis. This same reaction
generates METH in situ from HCY, which is toxic when it accumulates. During
oxidative stress conditions like the microenvironment of cancer cells, redox
enzymes like MS are easily inhibited (Zou and Banerjee, 2005) because MS is
locked up in the oxidized state awaiting re-activation by Methionine Synthase
Reductase. During oxidative stress, the need for glutathione to neutralize reactive
oxidative species may also activate the transsulfuration pathway by upregulating
the expression of cystathione beta- synthase (Scherer et al., 2012). The
transsulfuration pathway activation uses up the HCY at the expense of the
transmethylation leading to methyl trapping.
In this study, CRC patients had median plasma MMA concentrations higher than
the upper reference limit of 290 nM suggesting the presence of vitamin B12
73

deficiency (Fu et al., 2013). Higher 5MTHF (p= 0.002) and the presence of vitamin
B12 deficiency in CRC affirm methyl trapping phenomenon in CRC. The trapped
methyl donor consequentially affects the generation of SAM, thereby decreasing
the methylation capacity leading possibly to global hypomethylation. In CRC
samples, the results indicate significantly lower SAM (p= 0.011) and SAM/SAH
ratio (methylation capacity, p= 0.018) but a significantly higher THF plasma levels
(p<0.0001) due to a better FA-THF conversion (p<0.0001). Such a high
conversion is also expected to compensate for the block in THF regeneration
from the 5MTHF-MTR-THF route. Also, the significantly higher plasma levels of
THF in CRC patients may contribute to rapid profileration of CRC. CRC uses THF
to transfer methyl groups into the nucleotide biosynthesis and cell division. The
low SAM levels resulting from impaired methionine regeneration may upregulate
the activity of methylene tetrahydrofolate reductase (MTHFR) resulting in
increased utilization of cofactors like vitamin B2 which is significantly reduced in
CRC cases.
The total folates, as well as the proportions of folate metabolite, also revealed an
interesting trend in the two groups. Total folates and the proportions of reduced
folates (THF and 5MTHF) were significantly higher in CRC than controls. The
reverse trend was observed for the level or proportion of folic acid. Folic acid is
known to be one of the feedback regulatory metabolites of the FOCM inhibiting
DHFR and MTHFR when it accumulates (Jarabak and Bachur, 1971; Matthews
and Daubner, 1982; Morales and Greenberg, 1964; Xia et al., 2009). This
regulatory switch modulates how much of reduced folates join the cycle, but it
74

seems to be less efficient in the CRC cases thereby shuttling more reduced
folates for DNA synthesis or methylation. In the CRC cases, the DNA synthesis
is the preferred pathway due to the block in the methylation of HCY due to the
vitamin B12 deficiency. However, this may provide the needed DNA bases for the
high proliferation of tumor cells but aberrant methylation to drive the disease to
advance stages in the cases.
Converse to expectation is plasma HCY which is not significantly different in CRC
when compared to controls. HCY is expected to be higher in CRC because of its
association with risk of cancer (Chen et al., 1996; Ma et al., 1997). However, the
highly inflammatory environment associated with CRC may have driven excess
HCY through the cysteine-glutathione pathway. Because B6 catalyzes HCY
conversions, the significantly higher metabolites of B6 like PL and PA (p=0.02 and
0.003, respectively) may be resulting from this biochemical conversion of HCY.
In the present study, we have identified sets of plasma metabolites including
5MTHF, THF, FA, B2, PA, PL, SAM, SAH and MMA that are altered in CRC and
thus may be used as biomarkers for CRC screening. The misclassification error
rate of models developed in this pilot study based on these metabolites is 2.8%
which suggest that this finding may be a reliable screening assay as an alternative
to colonoscopy. The convenience and minimal invasion of blood-based assays
makes them highly needful in the population-based CRC screening.
3.5 Conclusions
We have developed a metabolomic approach to distinguish between CRC and
controls. This was accomplished through the identification of nine putative
75

biomarkers that can detect the presence of CRC. With the minimal
misclassification error rates of the grouping models used, the putative biomarkers
can be applied as an alternative screening approach, and this study is promising
for the reduction of CRC-related deaths.

76

Chapter 3: Bibliography and References

Burt, R. W., Barthel, J. S., Dunn, K. B., David, D. S., Drelichman, E., Ford, J. M.,
et al. (2010). Colorectal cancer screening. Journal of the National
Comprehensive Cancer Network, 8(1), 8-61.
Chen, J., Giovannucci, E., Kelsey, K., Rimm, E. B., Stampfer, M. J., Colditz, G.
A., et al. (1996). A methylenetetrahydrofolate reductase polymorphism and the
risk of colorectal cancer. Cancer research, 56(21), 4862-4864.
de Vogel, S., Schneede, J., Ueland, P. M., Vollset, S. E., Meyer, K., Fredriksen,
Å., et al. (2011). Biomarkers related to one-carbon metabolism as potential risk
factors for distal colorectal adenomas. Cancer Epidemiology Biomarkers &
Prevention, 20(8), 1726-1735.
Fiehn, O. (2002). Metabolomics–the link between genotypes and phenotypes.
Plant molecular biology, 48(1-2), 155-171.
Fu, X., Xu, Y.-K., Chan, P., & Pattengale, P. K. (2013). Simple, fast, and
simultaneous detection of plasma total homocysteine, methylmalonic acid,
methionine, and 2-methylcitric acid using liquid chromatography and mass
spectrometry (LC/MS/MS). In JIMD Reports-Volume 10 (pp. 69-78): Springer.
Giovannucci, E., Stampfer, M. J., Colditz, G. A., Hunter, D. J., Fuchs, C., Rosner,
B. A., et al. (1998). Multivitamin use, folate, and colon cancer in women in the
Nurses' Health Study. Annals of Internal Medicine, 129(7), 517-524.
77

Gupta, A. K., Brenner, D. E., & Turgeon, D. K. (2008). Early Detection of Colon
Cancer. Molecular diagnosis & therapy, 12(2), 77-85.
Hanley, M. P., & Rosenberg, D. W. (2015). One-Carbon Metabolism and
Colorectal Cancer: Potential Mechanisms of Chemoprevention. Current
pharmacology reports, 1(3), 197-205.
Herbert, V., & Zalusky, R. (1962). Interrelations of vitamin B12 and folic acid
metabolism: folic acid clearance studies. Journal of Clinical Investigation, 41(6),
1263.
Hoogerbrugge, R., Willig, S. J., & Kistemaker, P. G. (1983). Discriminant analysis
by double stage principal component analysis. Analytical chemistry, 55(11),
1710-1712.
Jarabak, J., & Bachur, N. R. (1971). A soluble dihydrofolate reductase from
human placenta: purification and properties. Archives of biochemistry and
biophysics, 142(2), 417-425.
Kim, Y.-I. (2004). Folate and DNA methylation: a mechanistic link between folate
deficiency and colorectal cancer? Cancer Epidemiology Biomarkers &
Prevention, 13(4), 511-519.
Kind, T., Tolstikov, V., Fiehn, O., & Weiss, R. H. (2007). A comprehensive urinary
metabolomic approach for identifying kidney cancer. Analytical biochemistry,
363(2), 185-195.
78

Kořínek, M., Šístek, V., Mládková, J., Mikeš, P., Jiráček, J., & Selicharová, I.
(2013). Quantification of homocysteine ‐related metabolites and the role of
betaine–homocysteine S ‐methyltransferase in HepG2 cells. Biomedical
Chromatography, 27(1), 111-121.
Lachenbruch, P. A., & Mickey, M. R. (1968). Estimation of error rates in
discriminant analysis. Technometrics, 10(1), 1-11.
Li, M., Wang, B., Zhang, M., Rantalainen, M., Wang, S., Zhou, H., et al. (2008).
Symbiotic gut microbes modulate human metabolic phenotypes. Proceedings of
the National Academy of Sciences, 105(6), 2117-2122.
Ma, J., Stampfer, M. J., Giovannucci, E., Artigas, C., Hunter, D. J., Fuchs, C., et
al. (1997). Methylenetetrahydrofolate reductase polymorphism, dietary
interactions, and risk of colorectal cancer. Cancer research, 57(6), 1098-1102.
Madsen, R., Lundstedt, T., & Trygg, J. (2010). Chemometrics in metabolomics—
a review in human disease diagnosis. Analytica Chimica Acta, 659(1), 23-33.
Matthews, R. G., & Daubner, S. C. (1982). Modulation of
methylenetetrahydrofolate reductase activity by S-adenosylmethionine and by
dihydrofolate and its polyglutamate analogues. Adv Enzyme Regul, 20, 123-131.
Morales, D. R., & Greenberg, D. M. (1964). Purification and properties of
dihydrofolate reductase of sheep liver. Biochimica et Biophysica Acta (BBA)-
Specialized Section on Enzymological Subjects, 85(3), 360-376.
79

Nicholson, J. K., Lindon, J. C., & Holmes, E. (1999). 'Metabonomics':
understanding the metabolic responses of living systems to pathophysiological
stimuli via multivariate statistical analysis of biological NMR spectroscopic data.
Xenobiotica, 29(11), 1181-1189.
O’Connell, J. B., Maggard, M. A., & Ko, C. Y. (2004). Colon cancer survival rates
with the new American Joint Committee on Cancer sixth edition staging. Journal
of the National Cancer Institute, 96(19), 1420-1425.
Pufulete, M., Al-Ghnaniem, R., Leather, A. J., Appleby, P., Gout, S., Terry, C., et
al. (2003). Folate status, genomic DNA hypomethylation, and risk of colorectal
adenoma and cancer: a case control study. Gastroenterology, 124(5), 1240-
1248.
Pufulete, M., Al-Ghnaniem, R., Rennie, J., Appleby, P., Harris, N., Gout, S., et
al. (2005). Influence of folate status on genomic DNA methylation in colonic
mucosa of subjects without colorectal adenoma or cancer. Br J Cancer, 92, 838-
842.
Roberts, L. D., Souza, A. L., Gerszten, R. E., & Clish, C. B. (2012). Targeted
metabolomics. Current protocols in molecular biology, 30.32. 31-30.32. 24.
Scherer, E., Schmitz, F., Vuaden, F., Savio, L., Ferreira, A., Tasca, R., et al.
(2012). Mild hyperhomocysteinemia alters extracellular adenine metabolism in
rat brain. Neuroscience, 223, 28-34.
80

Seifeldin, R., & Hantsch, J. J. (1999). The economic burden associated with
colon cancer in the United States. Clin Ther, 21(8), 1370-1379.
Siegel, R. L., Miller, K. D., & Jemal, A. (2016). Cancer statistics, 2016. CA: a
cancer journal for clinicians, 66(1), 7-30.
Stevens, V. L., McCullough, M. L., Sun, J., Jacobs, E. J., Campbell, P. T., &
Gapstur, S. M. (2011). High levels of folate from supplements and fortification
are not associated with increased risk of colorectal cancer. Gastroenterology,
141(1), 98-105. e101.
Xia, W., Hilgenbrink, A. R., Matteson, E. L., Lockwood, M. B., Cheng, J.-X., &
Low, P. S. (2009). A functional folate receptor is induced during macrophage
activation and can be used to target drugs to activated macrophages. Blood,
113(2), 438-446.
Zou, C.-G., & Banerjee, R. (2005). Homocysteine and redox signaling.
Antioxidants & redox signaling, 7(5-6), 547-559.

81

Chapter 4: Alterations in Folate-dependent One-Carbon Cycle as
Colon Cell Transition from Normal to Cancerous
4.1 Introduction
Folates (vitamin B9) are water-soluble B vitamins that play a critical modulatory
role in the development and prevention of several malignancies. Current thoughts
suggest that there is an inverse association between folate levels and colorectal
cancer (CRC) risk (Arnold et al., 2005). Folate and its metabolites play critical
roles in folate-dependent one carbon cycle (FOCM) to maintain genomic stability
through regulating DNA biosynthesis, repair, and methylation (de Vogel et al.,
2011). There are both preclinical and clinical evidence showing that folate
deficiency is associated with DNA strand breaks, impaired DNA repair machinery,
and increased mutations, and that folate supplementation could correct some of
this folate- deficiency induced defects (Bird, 2002; Burt et al., 2010; Gupta et al.,
2008; O’Connell et al., 2004; Rim et al., 2011; Vernon, 1997).
Folates define the group of pteroglutamates that exist in different forms. Folic acid
(FA) is the synthetic and most stable form of the folates that is given in
supplementation or fortification but requires metabolism to the active form
because it lacks coenzyme activity. In contrast, 5-methyltetrahydrofolate
(5MTHF) which is known as the natural (dietary) form is metabolically active. To
become biologically useful for active coenzyme function, FA must be reduced
82

sequentially to dihydrofolate (DHF) and then to tetrahydrofolate (THF) (Figure
4.1) where the reactions are catalyzed by dihydrofolate reductase (DHFR).

Figure 4.1: The FOCM cycle which drives DNA methylation, synthesis and repair.
DHFR, dihydrofolate reductase; 5-MTHF, 5-methyltetrahydrofolate; 5,10-MeTHF,
5,10-methylenetetrahydrofolate; CBS, Cystathionine-β-synthase; MAT, Methionine
adenosyl transferase; MS, methionine synthase; MTHFR,
methylenetetrahydrofolate reductase; SAM, S-adenosylmethionine; SAH, S-
adenosylhomocysteine; SHMT, Serine hydroxymethyltransferase; THF,
tetrahydrofolate; TYMS, thymidylate synthase; dUMP, deoxyuridine
monophosphate; dTMP, deoxythymidine monophosphate.

The rate of this conversion is modulated by DHFR, which is a relatively slow
enzyme and thus unmetabolized FA accumulates when large quantity is ingested
(Bailey and Ayling, 2009). Folate, in the form of 5MTHF, is the secondary methyl
donor that remethylates homocysteine (HCY) to methionine. Methionine
subsequently gets converted to S-adenosylmethionine (SAM), the primary methyl
83

group donor for a myriad of biological methylations, including that of DNA (Selhub
and Miller, 1992). After the methyl transfer, SAM is converted to S-
adenosylhomocysteine (SAH), a potent inhibitor of most SAM-dependent
methyltransferases (10). Folate deficiency has been proposed to affect DNA
methylation and enhances carcinogenesis, particularly that of the CRC, which
may be mediated through an induction of genomic as well as site- and gene-
specific DNA hypomethylation (Choi and Mason, 2002; Kim, 1999; Lamprecht
and Lipkin, 2003).
The key drivers of the underlying genetic alterations observed in CRC have been
linked to a dysregulation of the FOCM cycle. The FOCM cycle controls the
methylation capacity of any cellular system, thereby affecting the DNA
methylation patterns. The DNA methylation controls gene expression or
repression related to colon tumorigenesis. Although genetic alterations have
been extensively characterized in CRC, there has not been a focused analysis
on the FOCM genes and their corresponding metabolites. The FOCM metabolite
profiling presents a snapshot of the homeostatic flux caused by the downstream
effect of genomic and proteomic alterations that has occurred in the course of
tumorigenesis.
During folate metabolism, the THF is further metabolized into intermediates found
in DNA biosynthesis and repair pathway. This pathway produces the nucleotides
that serve as the building blocks for DNA. DHFR and thymidylate synthase
(TYMS) regulates the translation of proteins through the direct binding of several
84

mRNAs. This can repress the translation of p53 mRNA; a process that might
prevent the activation of this tumor suppressor (Banerjee et al., 2002; Liu et al.,
2002).
Thymidylate deficiency can impair DNA integrity and cellular division. This may
be a consequence of uracil incorporation into DNA leading to unstable DNA,
where editing is required. DNA mismatch repair genes (such as MLH1) are
required to repair for these types DNA errors. An accumulation of these errors
may cause loss of heterozygosity and chromosomal instability. CRC cells exhibit
these features which is called microsatellite stable (MSS). However, when there
is a mutation in the mismatch repair genes, there is an accumulation of repair
errors. MSI is the hallmark of CRC accounting for approximately 15% of all CRC
(Kurzawski et al., 2004). However, there are other forms of CRC that do not have
mutation in their mismatch repair machinery; they are rather a MSS.
5MTHF is the major terminal metabolite that is formed downstream of folate
metabolism. Decreased levels of 5MTHF can also downregulate methionine
biosynthesis, thus affecting protein production and DNA methylation. Decrease
methionine biosynthesis may be a consequence of downregulation of the MTR
gene or inhibition of methionine synthase (MS). MS inhibition may be due to a
feedback inhibition by SAH (Finkelstein, 2007) or lack of reactivation by
methionine synthase reductase (Wilson et al., 1999). Also, high doses of FA
concentrations may saturate DHFR and potentially inhibit the entire folate
metabolism (Bailey and Ayling, 2009), where the overall consequences may alter
85

protein expression (Jhaveri et al., 2001), decreased DNA repair capability and
accumulation of DNA damage (Blount et al., 1997; Duthie and Hawdon, 1998;
Kruman et al., 2002), increased chromosomal aberrations and fragility (Duthie,
1999); events that ultimately reduce growth rate and impair cell division.
The transsulfuration pathway, on the other hand, converts HCY to cysteine, which
can be utilized in various reactions or catabolized under conditions of excess.
Importantly, the transsulfuration pathway provides the limiting reagent, cysteine,
required for the biosynthesis of glutathione (GSH). It is expected that the
enzymes from both transmethylation and transsulfuration pathways compete to
salvage homocysteine in response to the cellular requirements.
High SAM levels can cause a feedback regulation at this junction by activating
cystathionine β-synthase (CBS) (Scherer et al., 2012) while inhibiting
transmethylation by decreasing the activity of methylenetetrahydrofolate
reductase (Singh et al., 2007). Because all the enzymes that compete for HCY
like MS and CBS are also redox enzymes (Chen et al., 1995; Millian and Garrow,
1998; Taoka et al., 1998), it important to understand how they function in the
oxidative stress environment, characteristic of cancer cells.
There is a gap in understanding the fundamental flux differences in FOCM cycle
among colon cells at different stages of tumorigenesis. With the availability of a
specific and sensitive multi-analyte metabolomics-based LCMS assay, the
cellular FOCM metabolites of untransformed, transitional and transformed colon
cells were quantified and compared. The expression of the genes that are
86

associated with the FOCM was also measured and compared. The aim of the
study was to dissect how different the FOCM metabolites and genes are when
compared to colon cells at the various stages of tumorigenesis.
4.2 Materials and Method
4.2.1 Cell lines and culture
Two untransformed human colon cells (CRL1459 and CRL1790) was compared
to three human colon adenocarcinoma cell lines, HCT116 p536+/+, HCT116 p53-
/- and Caco-2 cells. All of these cells were obtained from the American Type
Culture Collection (Manassas, VA). The HCT116 p53-/- and the murine APCmin
cell lines were obtained as a generous gift from Dr. Julio Camarero (University of
Southern California, CA) and Dr. Carla De Giovanni (University of Bologna, Italy)
respectively. The published characteristics of the colon cell lines that are relevant
to the study are summarized in Table 4.1 (Ahmed et al., 2013). The cells were
cultured in standard DMEM medium (Invitrogen, Gaithersburg, MD). Growth
medium was supplemented with 10% fetal bovine serum, 1% non-essential
amino acids and 1% antibiotic-antimycotic. The cells were maintained at 37
o
C in
95% humidity and 5% CO2 and passaged every four days. All cells in treatment
groups were harvested after 10-14 days of growth. The cellular FOCM metabolite
concentrations were determined at the last day of growth as well as their global
methylation levels. The expression of genes related to the folate metabolism,
87

transmethylation pathway, transsulfuration pathway, DNA synthesis & repair, and
proliferation & apoptosis.
Table 4.1: Summary of characteristics of colon cell lines
Cell Line Description CRC status CIN* TP53
a
MLH1
a
KRAS
a
MTHFR
a

CRL1459 Human fibroblast, 2.5m Untransformed ? ? ? ? ?
CRL1790 Fetal epithelial, 21w Untransformed ? ? ? ? ?
APC 10.1 Mice, APC +/- Transitional ? ? ? ? ?
HCT116 p53+/+ Human CRC, MSI Transformed - - + + +
HCT116 p53-/- Human CRC, MSI Transformed - + + + +
Caco-2 Human CRC, MSS Transformed + + - - -
*CIN- Chromosomal Instability
a
Mutation status of gene
4.2.2 Growth rates of colon cell lines
The different cell lines shown in Table 4.1 were treated with the DMEM media.
Approximately 500 cells of each cell line were seeded into each well using growth
media containing about 17% of Alamar blue. The fluorescence reading at each
time point was recorded and normalized using the baseline as a reference. The
growth curves were fitted with logistic curve model while sharing the carrying
capacity of the system across all cell lines. The growth rates of colon cells were
compared using doubling time and mean adjusted area under the curve (AUC).
Mean adjusted AUC has been proven a valid summary statistic for repeated
measures conducted over a period of a tumor growth study (Frison and Pocock,
1992; Pham et al., 1999). In the case of modeling tumor growth, adjusted AUC is
88

the standardized measure that can compare across treatment groups (Qian et
al., 2000).
4.2.3 Extraction and analysis of cellular FOCM metabolites
The validated LCMS assay in Chapter 2 of this dissertation was used to analyze
FOCM metabolites in the media and within the cells using three different
biological replicates. About 800 µL of the media used for growing the cells was
pipetted into an Eppendorf tube and centrifuged at 1500rpm for 10mins to
separate the supernatant from any floating cells. The treated cells were
trypsinized and washed three times with cold PBS. The cell pellets obtained from
the PBS wash was stored at −80°C until analysis.
The extracellular and cellular FOCM metabolites were extracted by taking 50 uL
of the media or cell pellet, adding the required internal standards, reducing agents
and the precipitating solution made of 20% 0.2M Zinc Sulphate in methanol. After
keeping this mixture in the -20°C for 30mins, it was centrifuged at 15,000 rpm for
15mins. About 100 uL of the supernatant was taken for the analysis of HCY and
its metabolites using a published assay (Kořínek et al., 2013) while the remaining
was transferred into a new 1.5 mL micro centrifuge tubes for drying under nitrogen
gas. Just before LCMS analysis, the samples were reconstituted in 30 uL of 1%
ascorbic acid, centrifuged at 5000 rpm for a minute and transferred into HPLC
vials for analysis using the validated assay.
89

4.2.4 RNA and DNA isolation
Genomic DNA was isolated from the cells using TRIzol (Thermo Fisher, Waltham,
MA, USA) per the manufacturer's protocol. The concentration and purity were
evaluated at the absorbance at 230, 260, and 280 nm using a Nanodrop ND-1000
(Thermo Scientific, Waltham, MA, USA). Samples that did not attain the required
purity were further purified by precipitating the genomic material, washing, and
resuspension.
4.2.5 DNA hydrolysis and global methylation measurement by LCMS
The hydrolysis of the DNA and the analysis of the global DNA methylation was
performed after modifying procedures published by Li and Franke (2011). Digest
Mix (enough for 100 samples) was prepared using 250 U Benzonase, 300 mU
phosphodiesterase, 200 U alkaline phosphatase, and 0.5 ml 10× Buffer [100 mM
Tris–HCl, 500 mM NaCl, 100 mM MgCl2 (pH 7.9)], and made up to 5 ml with
water. DNA samples (1 µg in 50-μl water) were digested by adding 50 μL Digest
Mix and incubating at 37°C for six hours. After hydrolysis, 900 𝜇 L of HPLC-grade
water was added to each sample, protecting the sample from light.
Global DNA methylation was quantified using a Shimadzu Prominence HPLC
system linked to an API 4000 LCMS/MS spectrometer (Sciex, Foster City, CA)
operating in the positive mode. Ten μL of DNA digest containing the internal
standard was injected onto a HyPurity C18 column (50 mm × 4.6 mm, 3 μm,
Thermo), using a mobile phase consisting of 0.1% aq. Formic acid (0.1%) in water
90

and 0.1% formic acid in methanol at a flow rate of 4000 μL/min using a gradient
elution over a total run time of 4.6 min.
During analysis, the LCMS/MS operated in the positive mode with the source
temperature at 500 °C, collision gas at 12 psi and curtain gas at 10 psi. The ion
source gas (1) was set at 50 psi, the ion source gas (2) at 50 psi and Ion Spray
Voltage at 55000 V. Data acquisition and analysis were performed using the Sciex’
MultiQuant software.
4.2.6 Quantitative real-time polymerase chain reaction
The extracted RNA was purified and converted to cDNA using standard operating
procedures from the reagent manufacturer. Complementary DNAs (cDNAs) were
generated from 2 μg total RNA by reverse transcriptase (RT) using oligodT
primers and the SuperScript III RNase H-Reverse Transcriptase (Invitrogen)
according to the manufacturer's protocol.
Levels of mRNAs encoding the expression levels of target genes related to the
FOCM were analyzed by quantitative real-time RT-PCR assay carried out in
quadruplicates using an ABI OpenArray Real-Time PCR (Applied Biosystems,
USA). The primers used for the RT-PCR (Table 4.2) were placed in different
exons to minimize the likelihood of amplifying contaminated genomic DNA. The
PCR amplification was performed in reaction mixtures consisting of 5 μL SYBR
Green master mix (Thermo Scientific, USA), 0.2 μL each of forward and reverse
primers (0·2 μM final concentration) and 0.8 μL of cDNA template in a 10-μl final
91

reaction volume. cDNA samples were amplified with a precycling heat activation
at 95 °C for 10 min, followed by 40 cycles of heat denaturation at 95 °C for 15 s
and annealing and extension at 60 °C for 1 min. The mRNA levels of the target
genes in colon cells are expressed as the ratio of the target gene mRNA to β-
actin (internal control) mRNA for each sample. The relative quantification of the
mRNA in each sample was determined using the 2
-ΔΔCt
method (Livak and
Schmittgen, 2001). Target genes and their primers are presented in Table 4.2.
Table 4.2: Sequences of primers used in RT-PCR in human and mouse species
Gene Human Primers Mouse Primers
ACTB
F: GAAAGCGGATAAAAGCAGTACCATC
R: ACAGGATGCAGAAAGAAATCATAACC
F: GAAAGCGGATAAAAGCAGTACCATC
R: ACAGGATGCAGAAAGAAATCATAACC
BCL-2
F: TCGGACTGAGAAACGCAAG
R: CTCGGTCACACTCAGAACTTAC
F: GAGATACGGATTGCACAGGAG
R: CGGAAGATAAAGCGTAACAGTTG
CBS
F: CAAGCAGTTCAAACAGATCCG
R: CACCCCGAACACCATCTG
F: CAGATCCAATCACGAGACCAG
R: CACGAAGTTTAGCAGGTCAATG
c-MYC
F: TTCGGGTAGTGGAAAACCAG
R: AGTAGAAATACGGCTGCACC
F: GCTGTTTGAAGGCTGGATTTC
R: GATGAAATAGGGCTGTACGGAG
CTH
F: CCGTTCTGGAAATCCCACTA
R: CCAAATTCAGATGCCACTTG
F: AACATTTCAAGAATGGGATGG
R: CTTAGCATGCTGCAGAGCAC
DHFR
F: CGCTGTGTCCCAGAACATG
R: GGTCTTCTTACCCATAATCACCAG
F: CATGGTTTGGATAGTCGGAGG
R: GTCACAAAGAGTCTGAGGTGG
DNMT1
F: CCAGAGAACGAGTTGCTAGAC
R: CAGTTTCTGTTTGGGTGTTGG
F: GACCTACTTGAGAGCATCCAG
R: TTCCCTTTCCCTTTGTTCCC
DNMT3A
F: TCTCTTTGATGGAATCGCTACAG
R: GTACATGATCTTCCCCTGGTG

F: GGACTTTATGAGGGTACTGGC
R: GATGTCCCTCTTGTCACTAACG
DNMT3B
F: CCCATTCGAGTCCTGTCATTG
R: TTGATATTCCCCTCGTGCTTC
F: GTACCCCATCAGTTGACTTGAG
R: TTGATCTTTCCCCACACGAG
FOLR1
F: AAGAGGACTGTGAGCAATGG
R: GAAATGGAAAGGTTGGCAGG
F: TGAGGACAATTTACACGACCAG
R: TCATAGTTCCGCAGTGGTTC
GSS
F: GTGAGCTATGCCCCATTCAC
R: TGTCAAAGAGACGAGCGGTA
F: CACCATCAAAAAGGACGACT
R: AAGCTGGCAGAGATAGTGTTGA
H19
F: ATGGTGCTACCCAGCTCAAG
R: GTGGCCATGAAGATGGAGTC
F: TGTGGTCAATGTGACAGAAAGA
R: AGGATCCAGAGAGCAGCAGA
IGF2
F: CCTGATTGCTCTACCCACC
R: AACCTGATGGAAACGTCCG
F: TCAGTTTGTCTGTTCGGACC
R: CACTCTTCCACGATGCCAC
92

MAT2A
F: GCACATTCCTTTTCACCTCAG
R: CAGTTTCACAAGCTACTTTGGC
F: AGAAAGTGGTTCGTGAAGCC
R: TTCCTCATTCCGGTCAAGATG
MGMT
F: GCTGAATGCCTATTTCCACC
R: CACTTCTCCGAATTTCACAACC
F: CATGGGATACGGTTGCTCAG
R: TGCGGGTTCACGGAAATAG
MLH1
F: GGCACAGCATCAAACCAAG
R: CAAGCATGGCAAGGTCAAAG
F: GGAAGAGATTAGTGAGCGGTG
R: CTGAGCTTGGTAGTGTTGAGG
MTHFR
F: ATCACTTGCCCCATCGTC
R: CATCGTTGTCTTTGATTGGCTC

F: CATCCTCACCATCAACTCTCAG
R: CCTCCACAGTTTCACGGG
MTR
F: TCTCATCTGGAATAAAGACCCTG
R: TTCACAAGGGCATACTCAAGG
F: CCAACTTATCCTTCTCCTTCCG
R: ATCATACACAGGGAGGTTGC
MTRR
F: AGACCCACCCGACACAGC
R: TAGAAATGCCGGGCTCCA
F: CCACATTGGTTGCTCCATTC
R: AGTTCCTGGACCCACCATTA
RASSF1A
F: ACACCTGACCTTTCTCAAGC
R: TGAAGCCTGTGTAAGAACCG
F: ACACCCGATCTTTCTCAAGC
R: AACAGGACGCACTAGTTTCAG
SHMT1
F: CAACTATGACCAGCTGGAGGA
R: GCTGATGTGAGCCATGTCC
Not done
SHMT2
F: GGTGATTCCCTCGCCTTT
R: TGATTCGGTCCTCAAATGTGT
Not done
TP53
F: GCCATCTACAAGCAGTCACAG
R: TCATCCAAATACTCCACACGC
F: ATGTTCCGGGAGCTGAATG
R: CCCCACTTTCTTGACCATTG
TYMS
F: TGAATCACATCGAGCCACTG
R: TTGGATGCGGATTGTACCC

F: TCCTCTGCTCACAACCAAAG
R: AAGCTGTCCAGAAAGTCTCG

4.3 Statistical analysis
Quantitative data was analyzed using GraphPad Prism software (Graph Pad
Software Inc., La Jolla, CA, USA). Using Analysis of Variance (ANOVA) analysis,
the means + SEM of the data will be compared among the different cell lines.
Post-hoc analysis was conducted on metabolites that show significant differences
between the cell lines. At least three biological replicates were used for each
statistical analysis, and treatments were considered significantly different if
statistical tests produced a p-value ≤0.05. The correlations that exist between
cellular and media FOCM metabolite levels as well as correlations between gene
expressions levels and growth rate were tested at a p-value ≤0.01.
93

4.4 Results
4.4.1 Growth rates of colon cell lines
The various cells cultured in folate sufficient medium (DMEM with 9 µM FA) for
72 hours, grew at widely different rates with the fastest rate in HCT116 p53+/+
followed by HCT116 p53-/-, CRL1459, CRL1790, Caco-2 and APC 10.1, in the
order of decreasing rates of growth. The slowest growth rate was observed in
APC cells (Figures 4.2 & 4.3). The doubling time of HCT116 p53+/+ was
significantly different when it was compared to the doubling times of APC 10.1
(p=0.036) and that of Caco-2 (p=0.0385). However, using the mean adjusted
AUC of fluorescence showed that the growth rates of the cells were significantly
different in each cell pairs (p<0.01) except among CRL1459, CRL1790 and Caco-
2. The morphology of the cells is shown in Figure 4.4.

Figure 4.2: Growth curve and doubling times of colon cell lines cultured in standard
DMEM media. (*) depicts curves whose doubling times are significantly different.
H o u rs o f C u ltu re
R e la tiv e F lu o re s c e n c e to B a s e lin e
0 2 0 4 0 6 0
0
2 0
4 0
6 0
8 0
1 0 0
A P C 1 0 .1
C R L 1 7 9 0
C R L 1 4 5 9
H C T 1 1 6 p 5 3 + /+
H C T 1 1 6 p 5 3 -/-
C A C O -2
*
*
94

Figure 4.3: The growth rate of cells compared by mean adjusted AUC of growth curves.
The growth rates were significantly different in each pair except CRL1459,
CRL1790 and Caco-2

Figure 4.4: Morphology of colon cell lines; CRL1459 (A), CRL1790 (B), APC 10.1 (C),
Caco-2 (D), HCT116 p53+/+ (E) and HCT116 p53-/- (F).

M e a n a d ju s te d A U C
o f flu o re s c e n c e
C R L 1 4 59
C R L 1 7 9 0
A P C 1 0 .1
C A CO -2
H C T 1 1 6 p 5 3 -/-
H C T 1 1 6 p 5 3 + /+
0
1 0
2 0
3 0
4 0
5 0
A B C
D E F
95

4.4.2 Distribution of folate metabolites concentrations in colon cell lines
Cellular folate distributions in the various colon cell lines from non-malignant
(CRL1459 and 1790) to transitional cells (APC 10.1) and finally to malignant cells
(wildtype HCT116, Caco-2) cultured in the DMEM medium were significantly
different between untransformed cells and transitional and transformed cells
(Figure 4.5). Caco-2 showed the highest total cellular folates, where THF was
most abundant. Cellular FA concentrations declined from non-malignant to
malignant cells. In contrast, cellular MTHF levels increased from non-malignant
to malignant cells. There was difference in MTHF levels in the two CRC cell lines
evaluated in this study. Comparing the proportion of folates metabolites
normalized to the total cellular folates, CRL1459 showed high proportions of FA
and DHF (75% and 24%, respectively), whereas APC 10.1, HCT116 and Caco-2
showed a relatively higher proportion of THF and MTHF (approximately 88% with
2% of 5MTHF). The malignant cells had higher proportions of 5MTHF (8% and
67% in Caco-2 and wildtype HCT116, respectively) when compared to CRL1459
and APC 10.1. A measure of the proportion of metabolized folate to
unmetabolized folates showed a sharply increasing trend with CRC progression.
96

Figure 4.5: Cellular folate distribution; both FA & DHF are significantly lower in transition
(APC) and malignant cells when compared to CRL1459. In contrast, the trend was
reversed for the reduced folates (THF & 5MTHF). The colon cells are compared
with their actual cellular levels (A) and the levels of the major folate metabolites
normalized to the total folates (B).

4.4.3 Expressions of folate metabolism genes in colon cell lines
The folate metabolism genes were upregulated in the transitional, fetal and
transformed colon cells (Figure 4.6). The reduced folate carrier gene, RFC was
overexpressed in HCT116 but normally expressed in the other human colon cell
lines. The SHMT1 and SHMT2 genes were highly overexpressed in HCT116
p53+/+ than the expressions in the other colon cells. These findings suggest that
the metabolome difference may be a consequence of altered expressions of
folate metabolism genes. Additionally, the expressions of FOLR1 in malignant
cells suggest they have increased ability to transport folates.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
CRL-1459 Human
APC
CACO-2 HCT116
Folate Distribution in Colon Cells
FA DHF THF 5MTHF
A B
97

Figure 4.6: Expression of folate metabolism genes in colon cells. Genes that regulate
the folate intake to folate reduction to form THF (A) and the genes that regulate
further downstream folate metabolism to form 5MTHF (B).

4.4.4 Plasma levels of FOCM metabolites in colon cell lines
We also evaluated the SAM, SAH and HCY cellular concentrations for CRL1459,
APC 10.1, Caco-2 and wildtype HCT116. In comparisons to CRL1459,
methionine levels were statistically lower in APC 10.1 and both malignant cancer
line. However, the cellular methionine was significantly higher in Caco-2 and
wildtype HCT-116 when compared to APC 10.1 suggesting that the transitional
98

transformation may have the most depletion of methionine, followed by the
transformed stages. The transition to CRC also showed a trend increased cellular
SAM, with reduced SAH and HCY. CYSTH was significantly higher in the Caco-
2 compared to the other cells (Figure 4.7A). The ratio of product to reactant
metabolites may also give an index of enzyme activity. The ratio of THF to FA
(comparable to activity of DHFR) was upregulated in the transformed cells and
that of CYSTH to HCY (comparable to activity of CBS) was upregulated in Caco-
2 cells (Figure 4.7B). The ratios of THF to FA metabolites were consistent with
the expressions of DHFR in the cell lines.
One of the key regulators of the metabolites of biological systems are the
enzymes that catalyze the interconversions. The enzymes catalyze the
metabolism a substrate into a product in the presence co-factors. Based on the
assumption that the cells depict a system in equilibrium, key metabolites ratios
may provide relevant information about the enzymatic function of some key
enzymes. The SAM/SAH ratio is an established index for methylation capacity of
a system. The methylation capacity was comparable in the transformed cells and
their values were higher than that of the other cells. The methylation capacity was
also comparable in the normal and transitional cells. The inverse of the
methylation capacity (SAH/SAM), which is a measure of DNMT function, was
significantly lower in the transformed cells.
99

Figure 4.7: Cellular levels of FOCM metabolites are compared in different colon cells.
The absolute cellular levels (A) are presented alongside the ratio of the metabolites
(B).

However, the THF/FA ratio showed a significantly lower ratio in CRL1459. The
ratio of THF to 5MTHF was significantly highest in the transitional cell but lowest
in HCT116. SAM/METH ratio signifies the activity of Methionine
adenosyltransferase (MAT), which catalyzes the reaction that synthesizes of
100

SAM, the primary methyl donor. The SAM/METH ratio was significantly lower in
the CRL1459 compared to the others, which had seemingly comparable levels.
The CYSTH/HCY clearly showed that CBS function in Caco-2 was significantly
different among the cell lines.
4.4.5 Global DNA methylation
To evaluate whether these changes impacted on DNA methylation, the gene
expression levels of DNMT1, DNMT3A, and DNMT3B, were compared in the
various colon cell lines, to attempt to correlate their expression profiles with the
overall global methylation level in the cell lines. The relative gene expression
levels of the prominent DNMTs and their respective level of global methylation is
depicted in Figure 4.8. Using the CRL1459 as the reference control, we observed
a lower global methylation in Caco-2 but a higher one in the HCT116. The relative
global methylation in APC 10.1 was comparable to the CRL1459 and CRL1790,
a fetal colon epithelial cell line. Except for APC cell line, the untransformed and
the transformed had the trend in their 5MTHF/THF ratios consistent with their
relative global methylation. The DNMTs were overexpressed in all the
transformed cell lines, and there was an overexpression of DNMT3B was
observed in the CRL1790.
101

Figure 4.8: Expression of DNMT isoforms among colon cells (A), their relative
levels of global methylation (B) and 5MTHF/THF ratio (C).

4.4.6 Expression of FOCM genes across colon cell lines
The gene expression levels of FOCM related genes were compared in the various
colon cell lines and correlate their expression profiles with each other as well as
with their related enzyme and metabolite levels in the cell lines. There was an
upregulation of all the FOCM related genes, except CBS which was
downregulated in APC 10.1 (data not shown). The relative gene expression levels
of the FOCM related genes are summarized in Figures 4.9-4.11. Using the
CRL1459 as the control, we observed a downregulation in the TP53, MTRR and
MAT genes in Caco-2 cell line. The expression levels of MTR and MLH1 genes
in Caco-2 cell line and TP53 in HCT116 p53-/- were comparable to that in the
control cell line. There was an upregulation in most of the FOCM related genes
C. 5MTHF/THF Ratio
F o ld c h a n g e
M T H F _ T H F
0 .1
1
1 0
G e n e s
F o ld c h a n g e
D N M T 1 D N M T 3 B D N M T 3 A
0 .1
1
1 0
1 0 0
C R L -1 4 5 9
H C T 1 1 6
C A C O -2
C R L 1 7 9 0
0 .5
2
B. Global Methylation
C e ll
re la tiv e % 5 m C
C R L 1 4 5 9 AP C C AC O -2 H C T 1 1 6
0 .0
0 .5
1 .0
1 .5
A. DNMT Expression
102

in the transformed colon cell lines. There was CBS overexpression in Caco-2 and
CRL1790 but a downregulation in HCT116 p53+/+ and HCT116 p53-/-. The
activation of the transsulfuration pathway channels more HCY through the
CYSTH route, as confirmed by Caco-2 showing significantly higher levels of
CYSTH than the other cells. The overexpression of CBS in Caco-2 was consistent
with the ratio of CYSTH to HCY metabolites. CRL1790 cell line showed an
overexpression in most of the FOCM related genes except for FOLR1, MTR, H19
and TP53 genes; CRL1790 had a downregulation in the H19 gene. The effect of
p53 mutation was also evaluated in the FOCM related gene expression in
HCT116 p53+/+ and HCT116 p53-/- (Table 4.3). There was an upregulation of
DNMTs, MTHFR and IGF2 in HCT116 p53+/+ compared to the HCT116 p53-/-.
Figure 4.9: Expression of genes related to transmethylation cycle in colon cell
lines
103

Figure 4.10: Expression of genes related to DNA synthesis in colon cell lines

Figure 4.11: Expression of genes related to proliferation and apoptosis in colon
cell lines

F o ld c h a n g e
IG F 2 H 1 9
0 .0 0 1
0 .0 1
0 .1
1
1 0
1 0 0
F o ld c h a n g e
B C L -2 C -M Y C
0 .0 0 1
0 .0 1
0 .1
1
1 0
1 0 0
C-Myc
ARF
p53
Apoptosis
Bcl-X Bcl-2
F o ld c h a n g e
T P 5 3 R A S S F 1 A
0 .0 0 1
0 .0 1
0 .1
1
1 0
104

Table 4.3: Differences in FOCM related gene expression for HCT116 p53-/- and
HCT116 p53+/+. The gene expression of CRL1459 was used as reference.
Gene HCT116 p53+/+ HCT116 p53-/-
TP53 20 1
DNMT1 151 27
DNMT 3B 126 26
MTHFR 31 4
IGF2 156 1
RASSF1A 1 6

4.4.7 Correlations between genes, growth rate and CRC status
The CRC status was ranked from 1-5 based on their characteristics in the
following order: CRL1459, CRL1790, Caco-2, HCT116 p53-/- and HCT116
p53+/+. Most of the genes and growth rate had positive correlations with the CRC,
except CBS, RASSF1A and MLH1. Most of the genes also correlated positively
with each other except CBS which had negative correlations with MTR/MTRR,
GSS, DHFR, TYMS and growth rate of the colon cells (Figure 4.12). The growth
rate and the ranked stages in CRC development correlated with most of the
FOCM related genes.
105

Figure 4.12: Correlation matrix of FOCM related genes, growth rate and CRC status of
colon cells. The circles show pairs with significant correlation (p<0.01).

4.5 Discussion
We determined whether there are FOCM metabolic differences that exist
between colon cell lines in different stages of CRC development. The primary
goal was to probe whether expression of FOCM genes and metabolites’ cellular
concentrations are increased or decreased as the colon cells transition to CRC.
We further evaluated if any observed changes genomics or metabolomics reflect
on the cellular phenotypic characteristics. Our approach compared the FOCM
metabolites and their related genes in untransformed, transitional and
transformed colon cells focusing on genes that relate to DNA biosynthesis,
Correlation coefficient, r
106

methylation, folate metabolism and cell proliferation, all of which are relevant to
carcinogenesis.
4.5.1 Folate distribution in colon cells
FOLR1 regulates the folate receptors that transports folic acid and reduced
folates into the cell. An up-regulation of FOLR1 in the transformed cells was
expected because the high FA content of DMEM will be needed by fast
proliferating cells to support the high folate requirement for DNA nucleotides. A
similar trend was seen in other downstream genes in the folate metabolism
pathway, which might have adapted to the high demands with up-regulation in
the genes. The major differences in the folate metabolism among the colon cells
lie in the proportions of the various major metabolites of folates that may be
driving the direction of the folates- into either DNA biosynthesis or DNA
methylation. The significantly higher levels of reduced folates in the transformed
cells corroborate the upregulation observed in the folate metabolism genes
especially MTHFR. Although the heterozygous C677T MTHFR polymorphism
observed in HCT116 is expected to have about 60% activity of the wild-type found
in Caco-2 (Frosst et al., 1995), there was higher levels of terminal folate, 5MTHF
formed in the HCT116 p53+/+. The upregulation of the MTHFR in the HCT116
p53+/+ was about seven times higher than that in the Caco-2. Also, the
comparatively reduced expression of SHMTs in Caco-2 compared to HCT116
may result in the decreased metabolism of THF resulting in its pooling. The
overexpression of MTHFR and SHMTs may explain the significantly higher
107

5MTHF to THF ratio observed in HCT116 p53+/+. The very low levels of FA and
DHF in the transformed cells may lessen the inhibitory controls that may be
available to modulate the enzymatic metabolism of folates downstream (Bailey
and Ayling, 2009).
4.5.2 DNA methylation in transformed cells
Transformed cells exhibit accumulation of reduced folates supporting the
hypothesis of methyl-trapping (Herbert and Zalusky, 1962). A trend in the ratio
5MTHF/THF was detected from the untransformed to the transformed cells.
Except for the transitional cells, this trend was consistent with the global
methylation trends and may be predictive of the global methylation at the cellular
level. The HCT116 p53+/+ had the highest global methylation among the colon
cells studied. This may be a consequence of DNMTs up-regulation and the
availability of substrates to facilitate DNA methylation. This is supported by
inverse relationship of SAH/SAM ratio in non-malignant to transformed cells.
However, Caco-2 cell lines depart from this trend with global methylation far lower
than that of the control, although Caco-2 also has a high methylation capacity.
Probing the problem upstream of the cycle reveals MAT gene down-regulation
which encodes for the enzyme catalyzing the formation of SAM from METH.
There may be a rate limiting constraint in the cycle at this conversion step that
may result in the back pooling of FOCM metabolites. However, the up-regulation
of the CBS gene tend to facilitate the transsulfuration pathway as the needed vent
to metabolize the back-pooled HCY into CYSTH and subsequently through to
108

GSH. It is thus not surprising to observe a back pool of reduced folates in Caco-
2 cells, which have total folate levels at least five-fold higher than the other cell
lines. Consequently, this results in the overall downregulation of the
transmethylation pathway leading to decreased DNA methylation. This
observation is consistent with the observed decreased global methylation in the
Caco-2 cell line.
4.5.3 Redox homeostasis in transformed colon cell lines
Untransformed and transformed cells may also differ in their levels of FOCM
metabolites based on their environment. Whereas untransformed cells may have
a homeostatic balance under normoxic stress condition, transformed cells thrive
under oxidative stress conditions (Mosharov et al., 2000). The FOCM is the
pivotal cycle that responds sensitively to oxidative stress to restore normoxic
conditions. HCY the FOCM metabolite that lies central to this homeostatic
process thereby drives other aspects of the cycle to affect DNA methylation,
synthesis, and proliferation. The fate of HCY may be either as a substrate in the
transmethylation reaction or transsulfuration reaction. During oxidative stress in
transformed cells, MS activity is inhibited, resulting in the diversion of
homocysteine into the transsulfuration reaction to produce the antioxidant GSH,
providing an important adaptive response (Zou and Banerjee, 2005). However,
oxidative inhibition of methionine synthase may trigger resultant epigenetic
changes and an up-regulation of the MTR gene. The epigenetic changes in gene
expression can recruit further adaptive responses to oxidative stress. Decreased
109

methionine synthase activity decreases the level of the methyl donor SAM
favoring methylation inhibition by SAH (Matthews et al., 1998).
During oxidative stress, SAM levels can activate CBS (Scherer et al., 2012) while
depressing transmethylation by decreasing the activity of
methylenetetrahydrofolate reductase (Singh et al., 2007). The up-regulated CBS
can increase levels of GSH in response to increased oxidative stress. Studies
have shown that approximately 50% of cysteine in GSH is derived from HCY
metabolism via the transsulfuration pathway (Beatty and Reed, 1980; Mosharov
et al., 2000). The transformed colon cells respond to their oxidative stress
environment differently to affect the one-carbon cycle genes and metabolites.
Caco-2 up-regulate CBS expression to produce more GSH to neutralize the
reactive oxidative species. This result in lower proportions of the HCY feeding
into the transmethylation cycle to result in lower global methylation in Caco-2
compared to the CRL1459 cell. HCT116 p53+/+, on the other hand, struggle to
activate the transsulfuration pathway due to the downregulated CBS gene.
HCT116 p53+/+ and HCT116 p53-/- cells, however, upregulate the GSS gene to
compensate for the downregulated CBS resulting in a struggle to neutralize the
ROS creating the oxidative stress environment that inhibits the MS further. The
cells may respond to the MS inhibition by up-regulating the MTR and MTRR.
4.5.4 Methionine depletion in transformed cells
The cellular response to environmental changes can also be assessed with
METH can the rate limiting substrate since it is the source of the generated HCY.
110

The standard DMEM medium used to grow these colon cells contain 30mg/L of
METH but has no HCY added. METH is pivotal as an essential amino acid, which
is required for the formation of HCY, SAM (the primary methyl donor), and as the
initiation amino acid for protein synthesis. The rate of METH depletion can be an
indicator for the direction of metabolic flux within the cycle. When there are
decreased cellular levels of SAM, it puts important methylation reactions at risk
thereby initiating mechanisms to maintain supplies of methionine and SAM as a
priority. The folate co-factors are directed through the transmethylation cycle at
the expense of the cycles that produce purines and pyrimidines for DNA
synthesis.
In the case of HCT116, the oxidative stress environment may lead to a vicious
cycle that accumulates 5MTHF because the MTR is inhibited and the inhibitory
feedback loop on the MTHFR is broken. The METH stores are depleted further
to generate HCY and subsequently, GSH. The down-regulated CBS may not
generate enough GSH, and there may be an increase in homocysteine
thiolactone formation (Beatty and Reed, 1980). Caco-2, on the other hand, has a
downregulated MTR/MTRR and MAT genes. However, the prominence placed
on methylation reactions will allow the production of SAM but the biosynthesis of
METH becomes challenging. This may lead to a depletion of the METH stores.
4.5.5 Regulation of proliferation and apoptosis in colon cell lines
One of the genes that were found upregulated in the highly proliferating cells was
IGF2, which is a fetal growth factor known to undergo loss of imprinting (LOI) in
111

CRC. This LOI mechanism differs from what occurs in other cancer types which
involve the repression of IGF2 gene due to methylated long noncoding RNA, H19
(Kelemen, 2006; Popat et al., 2004). In CRC, IGF2 LOI may result from a
hypomethylation of the differentially methylated region of the H19 gene in both
alleles leading to overexpression of IGF2 (Cui et al., 2002). H19 was
overexpressed in HCT116 p53+/+ and HCT116 p53-/- but under expressed in
CRL1790. H19 has been identified as the most significant long non-coding RNA
associated with CRC patient survival. The H19 function is mediated by RB1-E2F1
function and β-catenin activity as essential upstream regulators (Ohtsuka et al.,
2016). The overexpression of the IGF2 leads to the activation of the
phosphatidylinositol 3-kinase- and mitogen-activated protein (MAP) kinase-
dependent signaling pathway which regulates MS activation (Waly et al., 2004).
The consequential effects observed in these cells are cell growth and
proliferation. Tumors overexpressing both IGF2 and p53 have however been
found to be of the advanced stage with poor survival (Zhao and Casson, 2011).
The regulation of cell growth requires p53 which maintains the genomic stability,
cellular senescence (Kaffer et al., 2001; Mason et al., 2007) and induces
apoptosis (Shephard, 2011). Under normoxic conditions, p53 levels are kept at a
very low level unless the cells are activated by signals from DNA damage or other
cellular stresses (Maulik and Maulik, 2010). When there is DNA damage, and
cellular stresses, p53 protein expression level is upregulated, leading to cell cycle
arrest, DNA repair or apoptosis. Tumor protein p53 is thus very critical in the
inhibition of cancer cell division due to the continual oxidative stress environment.
112

However, the co-expression of BCL-2 and the oncogene c-MYC efficiently
antagonizes effects of p53 thereby modulating proliferation and apoptosis,
although BCL-2 on its own can affect antagonize the effect of p53 to induce
apoptosis (Chiarugi and Ruggiero, 1995). The effect of BCL-2 and c-MYC co-
expression may explain the high proliferation of HCT116 p53+/+ cells,
irrespective of the overexpression of TP53 in these cells.
4.6 Conclusion
The FOCM is altered in colon cells as they transition from normal to cancer. The
folate distribution and methionine levels are very different between the
transformed and untransformed colon cells. Also, expression of CBS, MTRR, and
MAT are key in distinguishing between untransformed and transformed colon
cells. It is likely that the dysregulation of the FOCM, specifically these metabolites
and genes may be the driver of the transformation from normal epithelium to
CRC.
113

Chapter 4: Bibliography and References

Ahmed, D., Eide, P., Eilertsen, I., Danielsen, S., Eknaes, M., Hektoen, M., Lind,
G., and Lothe, R. (2013). Epigenetic and genetic features of 24 colon cancer cell
lines. Oncogenesis 2, e71.
Arnold, C. N., Goel, A., Blum, H. E., and Boland, C. R. (2005). Molecular
pathogenesis of colorectal cancer: implications for molecular diagnosis. Cancer
104, 2035-2047.
Bailey, S. W., and Ayling, J. E. (2009). 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.
Banerjee, D., Mayer-Kuckuk, P., Capiaux, G., Budak-Alpdogan, T., Gorlick, R.,
and Bertino, J. R. (2002). Novel aspects of resistance to drugs targeted to
dihydrofolate reductase and thymidylate synthase. Biochimica et Biophysica
Acta (BBA)-Molecular Basis of Disease 1587, 164-173.
Beatty, P. W., and Reed, D. J. (1980). Involvement of the cystathionine pathway
in the biosynthesis of glutathione by isolated rat hepatocytes. Archives of
Biochemistry and Biophysics 204, 80-87.
Bird, A. (2002). DNA methylation patterns and epigenetic memory. Genes &
development 16, 6-21.
114

Blount, B. C., Mack, M. M., Wehr, C. M., MacGregor, J. T., Hiatt, R. A., Wang,
G., Wickramasinghe, S. N., Everson, R. B., and Ames, B. N. (1997). 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.
Burt, R. W., Barthel, J. S., Dunn, K. B., David, D. S., Drelichman, E., Ford, J. M.,
Giardiello, F. M., Gruber, S. B., Halverson, A. L., and Hamilton, S. R. (2010).
Colorectal cancer screening. Journal of the National Comprehensive Cancer
Network 8, 8-61.
Chen, Z., Chakraborty, S., and Banerjee, R. (1995). Demonstration that
mammalian methionine synthases are predominantly cobalamin-loaded. Journal
of Biological Chemistry 270, 19246-19249.
Chiarugi, V., and Ruggiero, M. (1995). Role of three cancer" master genes" p53,
bcl2 and c-myc on the apoptotic process. Tumori 82, 205-209.
Choi, S.-W., and Mason, J. B. (2002). Folate status: effects on pathways of
colorectal carcinogenesis. The Journal of nutrition 132, 2413S-2418S.
Crider, K.S., Yang, T.P., Berry, R.J., and Bailey, L.B. (2012). Folate and DNA
methylation: a review of molecular mechanisms and the evidence for folate’s
role. Advances in Nutrition: An International Review Journal 3, 21-38.
115

Cui, H., Onyango, P., Brandenburg, S., Wu, Y., Hsieh, C.-L., and Feinberg, A. P.
(2002). Loss of imprinting in colorectal cancer linked to hypomethylation of H19
and IGF2. Cancer research 62, 6442-6446.
de Vogel, S., Schneede, J., Ueland, P. M., Vollset, S. E., Meyer, K., Fredriksen,
Å., Midttun, Ø., Bjørge, T., Kampman, E., and Bretthauer, M. (2011). Biomarkers
related to one-carbon metabolism as potential risk factors for distal colorectal
adenomas. Cancer Epidemiology Biomarkers & Prevention 20, 1726-1735.
DeVries, J.W., Rader, J.I., Keagy, P.M., Hudson, C.A., Angyal, G., Arcot, J.,
Castelli, M., Doreanu, N., Hudson, C., Lawrence, P., et al. (2005). Microbiological
assay-trienzyme procedure for total folates in cereals and cereal foods:
collaborative study. J AOAC Int 88, 5-15.
Duthie, S. J. (1999). Folic acid deficiency and cancer: mechanisms of DNA
instability. Br Med Bull 55, 578-592.
Duthie, S. J., and Hawdon, A. (1998). DNA instability (strand breakage, uracil
misincorporation, and defective repair) is increased by folic acid depletion in
human lymphocytes in vitro. FASEB J 12, 1491-1497.
Ferlay, J., Soerjomataram, I., Dikshit, R., Eser, S., Mathers, C., Rebelo, M.,
Parkin, D.M., Forman, D., and Bray, F. (2015). Cancer incidence and mortality
worldwide: sources, methods and major patterns in GLOBOCAN 2012.
International journal of cancer 136, E359-E386.
116

Finkelstein, J. D. (2007). Metabolic regulatory properties of S-
adenosylmethionine and S-adenosylhomocysteine. Clinical Chemical Laboratory
Medicine 45, 1694-1699.
Frison, L., and Pocock, S. J. (1992). Repeated measures in clinical trials:
analysis using mean summary statistics and its implications for design. Statistics
in medicine 11, 1685-1704.
Frosst, P., Blom, H. J., Milos, R., Goyette, P., Sheppard, C. A., Matthews, R. G.,
Boers, G. J., den Heijer, M., Kluijtmans, L. A., and Van den Heuvel, L. (1995). A
candidate genetic risk factor for vascular disease: a common mutation in
methylenetetrahydrofolate reductase.
Fu, X., Xu, Y.-K., Chan, P., and Pattengale, P.K. (2013). Simple, fast, and
simultaneous detection of plasma total homocysteine, methylmalonic acid,
methionine, and 2-methylcitric acid using liquid chromatography and mass
spectrometry (LC/MS/MS). In JIMD Reports-Volume 10 (Springer), pp. 69-78.
Group, M.V.S.R. (1991). Prevention of neural tube defects: results of the Medical
Research Council Vitamin Study. The lancet 338, 131-137.
Group, U.C.S.W. (2016). United States cancer statistics: 1999–2013 Incidence
and Mortality Web-based Report. Atlanta: US Department of Health and Human
Services, Centers for Disease Control and Prevention and National Cancer
Institute.
117

Gupta, A. K., Brenner, D. E., and Turgeon, D. K. (2008). Early Detection of Colon
Cancer. Molecular diagnosis & therapy 12, 77-85.
Haggar, F.A., and Boushey, R.P. (2009). Colorectal cancer epidemiology:
incidence, mortality, survival, and risk factors. Clin Colon Rectal Surg 22, 191-
197.
Herbert, V., and Zalusky, R. (1962). Interrelations of vitamin B12 and folic acid
metabolism: folic acid clearance studies. Journal of Clinical Investigation 41,
1263.
Jhaveri, M. S., Wagner, C., and Trepel, J. B. (2001). Impact of extracellular folate
levels on global gene expression. Mol Pharmacol 60, 1288-1295.
Kaffer, C. R., Grinberg, A., and Pfeifer, K. (2001). Regulatory Mechanisms at the
MouseIgf2/H19 Locus. Molecular and cellular biology 21, 8189-8196.
Kelemen, L. E. (2006). The role of folate receptor α in cancer development,
progression and treatment: cause, consequence or innocent bystander?
International journal of cancer 119, 243-250.
Kim, Y. I. (1999). Folate and carcinogenesis: evidence, mechanisms, and
implications. J Nutr Biochem 10, 66-88.
Kim, Y.I. (2003). Role of folate in colon cancer development and progression. J
Nutr 133, 3731S-3739S.
118

Kořínek, M., Šístek, V., Mládková, J., Mikeš, P., Jiráček, J., and Selicharová, I.
(2013). Quantification of homocysteine ‐related metabolites and the role of
betaine–homocysteine S ‐methyltransferase in HepG2 cells. Biomedical
Chromatography 27, 111-121.
Kruman, I. I., Kumaravel, T. S., Lohani, A., Pedersen, W. A., Cutler, R. G.,
Kruman, Y., Haughey, N., Lee, J., Evans, M., and Mattson, M. P. (2002). Folic
acid deficiency and homocysteine impair DNA repair in hippocampal neurons
and sensitize them to amyloid toxicity in experimental models of Alzheimer's
disease. J Neurosci 22, 1752-1762.
Kurzawski, G., Suchy, J., Debniak, T., Kladny, J., and Lubinski, J. (2004).
Importance of microsatellite instability (MSI) in colorectal cancer: MSI as a
diagnostic tool. Ann Oncol 15, 283-284.
Lamprecht, S. A., and Lipkin, M. (2003). Chemoprevention of colon cancer by
calcium, vitamin D and folate: molecular mechanisms. Nature Reviews Cancer
3, 601-614.
Li, X., and Franke, A. A. (2011). High-throughput and cost-effective global DNA
methylation assay by liquid chromatography-mass spectrometry. Analytica
chimica acta 703, 58-63.
Liu, J., Schmitz, J. C., Lin, X., Tai, N., Yan, W., Farrell, M., Bailly, M., Chen, T.-
m., and Chu, E. (2002). Thymidylate synthase as a translational regulator of
119

cellular gene expression. Biochimica et Biophysica Acta (BBA)-Molecular Basis
of Disease 1587, 174-182.
Livak, K. J., and Schmittgen, T. D. (2001). Analysis of relative gene expression
data using real-time quantitative PCR and the 2− ΔΔCT method. methods 25,
402-408.
Mason, J. B., Dickstein, A., Jacques, P. F., Haggarty, P., Selhub, J., Dallal, G.,
and Rosenberg, I. H. (2007). A temporal association between folic acid
fortification and an increase in colorectal cancer rates may be illuminating
important biological principles: a hypothesis. Cancer Epidemiology and
Prevention Biomarkers 16, 1325-1329.
Matthews, R. G., Sheppard, C., and Goulding, C. (1998).
Methylenetetrahydrofolate reductase and methionine synthase: biochemistry
and molecular biology. European journal of pediatrics 157, S54-S59.
Matthews, R.G., and Daubner, S.C. (1982). Modulation of
methylenetetrahydrofolate reductase activity by S-adenosylmethionine and by
dihydrofolate and its polyglutamate analogues. Adv Enzyme Regul 20, 123-131.
Maulik, N., and Maulik, G. (2010). Nutrition, epigenetic mechanisms, and human
disease: CRC Press).
120

Millian, N. S., and Garrow, T. A. (1998). Human betaine–homocysteine
methyltransferase is a zinc metalloenzyme. Archives of biochemistry and
biophysics 356, 93-98.
Mosharov, E., Cranford, M. R., and Banerjee, R. (2000). The quantitatively
important relationship between homocysteine metabolism and glutathione
synthesis by the transsulfuration pathway and its regulation by redox changes.
Biochemistry 39, 13005-13011.
O’Connell, J. B., Maggard, M. A., and Ko, C. Y. (2004). Colon cancer survival
rates with the new American Joint Committee on Cancer sixth edition staging.
Journal of the National Cancer Institute 96, 1420-1425.
Ohtsuka, M., Ling, H., Ivan, C., Pichler, M., Matsushita, D., Goblirsch, M.,
Stiegelbauer, V., Shigeyasu, K., Zhang, X., and Chen, M. (2016). H19 Noncoding
RNA, an Independent Prognostic Factor, Regulates Essential Rb-E2F and
CDK8-β-Catenin Signaling in Colorectal Cancer. EBioMedicine 13, 113-124.
Pham, B., Cranney, A., Boers, M., Verhoeven, A., Wells, G., and Tugwell, P.
(1999). Validity of area-under-the-curve analysis to summarize effect in
rheumatoid arthritis clinical trials. The Journal of rheumatology 26, 712-716.
Popat, S., Matakidou, A., and Houlston, R. S. (2004). Thymidylate synthase
expression and prognosis in colorectal cancer: a systematic review and meta-
analysis. Journal of Clinical Oncology 22, 529-536.
121

Qian, W., Parmar, M., Sambrook, R., Fayers, P., Girling, D., and Stephens, R.
(2000). Analysis of messy longitudinal data from a randomized clinical trial.
Statistics in medicine 19, 2657-2674.
Rim, S. H., Joseph, D. A., Steele, C. B., Thompson, T. D., and Seeff, L. C. (2011).
Colorectal cancer screening - United States, 2002, 2004, 2006, and 2008.
MMWR Surveill Summ 60 Suppl, 42-46.
Roberts, L.D., Souza, A.L., Gerszten, R.E., and Clish, C.B. (2012). Targeted
metabolomics. Current protocols in molecular biology, 30.32. 31-30.32. 24.
Sauer, J., Mason, J.B., and Choi, S.-W. (2009). Too much folate: a risk factor for
cancer and cardiovascular disease? Curr Opin Clin Nutr Metab Care 12, 30-36.
Scherer, E., Schmitz, F., Vuaden, F., Savio, L., Ferreira, A., Tasca, R., Casali,
E., Bogo, M., Bonan, C., and Wyse, A. (2012). Mild hyperhomocysteinemia alters
extracellular adenine metabolism in rat brain. Neuroscience 223, 28-34.
Selhub, J., and Miller, J. W. (1992). The pathogenesis of homocysteinemia:
interruption of the coordinate regulation by S-adenosylmethionine of the
remethylation and transsulfuration of homocysteine. The American journal of
clinical nutrition 55, 131-138.
Shane, B., Tamura, T., and Stokstad, E.L. (1980). Folate assay: a comparison of
radioassay and microbiological methods. Clin Chim Acta 100, 13-19.
122

Shephard, R. J. (2011). Nutrition, epigenetic mechanisms and human disease.
Applied Physiology, Nutrition, and Metabolism 36, 311-313.
Shephard, R.J. (2011). Nutrition, epigenetic mechanisms and human disease.
Applied Physiology, Nutrition, and Metabolism 36, 311-313.
Singh, S., Madzelan, P., and Banerjee, R. (2007). Properties of an unusual heme
cofactor in PLP-dependent cystathionine β-synthase. Natural product reports 24,
631-639.
Taoka, S., Ohja, S., Shan, X., Kruger, W. D., and Banerjee, R. (1998). Evidence
for heme-mediated redox regulation of human cystathionine β-synthase activity.
Journal of Biological Chemistry 273, 25179-25184.
van der Greef, J., Davidov, E., Verheij, E., Vogels, J., van der Heijden, R.,
Adourian, A.S., Oresic, M., W., M.E., and Naylor, S. (2003). The Role of
Metabolomics in Systems Biology. In Metabolic Profiling: Its Role in Biomarker
Discovery and Gene Function Analysis, G.G.H.R. Goodacre, ed.
(Boston/Dordrecht/London: Kluwer Academic Publishers ), pp. 171-198.
Vernon, S. W. (1997). Participation in colorectal cancer screening: a review. J
Natl Cancer Inst 89, 1406-1422.
Waly, M., Olteanu, H., Banerjee, R., Choi, S., Mason, J., Parker, B., Sukumar,
S., Shim, S., Sharma, A., and Benzecry, J. (2004). Activation of methionine
123

synthase by insulin-like growth factor-1 and dopamine: a target for
neurodevelopmental toxins and thimerosal. Molecular psychiatry 9, 358-370.
Wilson, A., Leclerc, D., Rosenblatt, D., and Gravel, R. (1999). Molecular basis
for methionine synthase reductase deficiency in patients belonging to the cblE
complementation group of disorders in folate/cobalamin metabolism. Human
molecular genetics 8, 2009-2016.
Zhao, R., and Casson, A. G. (2011). Interactions between P53 and IGF2 in
human esophageal adenocarcinoma tissues and cell lines. In, (AACR).
Zou, C.-G., and Banerjee, R. (2005). Homocysteine and redox signaling.
Antioxidants & redox signaling 7, 547-559.
124

Chapter 5: The Effect of Folate Supplementation on Colon cell
lines at Different Stages of CRC Development
5.1 Introduction
Folate supplementation has stimulated an ongoing debate about the association
of folates to the risk of developing colorectal cancer. This public health concern
is warranted due to the report increase in the CRC incidence after the food
supplementation policy in Northern America. However, epidemiological studies
referenced in the ongoing debate have several confounders that may be
challenging to address in clinical studies. There is the need to use in vitro model
models to mimic what happens in the humans to build a substantial evidence to
contribute to the debate.
Folates are water-soluble B vitamins that are required for cell growth and
maintenance of critical metabolic processes. Folate and its metabolites have
been described to maintain genomic stability through regulating DNA
biosynthesis, repair and methylation (de Vogel et al., 2011). Evidence from in
vitro, animal and human studies indicates that folate deficiency is associated with
DNA strand breaks, impaired DNA repair and increased mutations, and that folate
supplementation could correct some of these defects (Bird, 2002; Burt et al.,
2010; Gupta et al., 2008; O’Connell et al., 2004; Rim et al., 2011; Vernon, 1997).
Supplements required for such efforts ought to be stable for consistent dosing
over a period, though they may not necessarily be the active form.
125

Folates may be ingested as folic acid (FA) or reduced folates, prominent amongst
which is 5-methyltetrahydrofolate (5MTHF). FA is the synthetic form of the vitamin
that lacks coenzyme activity but the natural (dietary) form, 5MTHF is
metabolically active. Synthetic FA (pteroylmonoglutamic acid) has a fully oxidized
pteridine ring and is very stable under most conditions (like temperature and pH);
it is the folate form used for supplements and food fortification (Shane, 2010).
During folate metabolism, FA must be reduced sequentially to dihydrofolate
(DHF) and then to tetrahydrofolate (THF) by dihydrofolate reductase (DHFR),
which is relatively slow enzyme in humans and appears incapable of completely
converting large amount of FA to THF (Bailey and Ayling, 2009). Although FA is
water soluble, there is an increased risk of FA accumulation as unmetabolized
FA when there is a continual supplementation beyond the daily requirements.
However, unmetabolized FA has been shown to be associated with reduced
natural killer cells and increased risk of CRC (Cho et al., 2015; Troen et al., 2006).
Folate, in the form of 5-methyltetrahydrofolate, is involved in the transmethylation
pathway in which homocysteine is converted to methionine, which is a precursor
of S-adenosylmethionine (SAM). SAM is the primary methyl group donor for
several biological methylations, including that of DNA (Selhub and Miller, 1992).
After transfer of the methyl group, SAM is converted to S-adenosylhomocysteine
(SAH), a potent inhibitory modulator of most SAM-dependent methyltransferases
(Selhub and Miller, 1992). Studies have shown that folate status and timing can
affect DNA methylation and carcinogenesis (Kim, 1999, 2003; Ulrich and Potter,
2006, 2007). It is proposed that a mechanism by which folate deficiency
126

enhances carcinogenesis, particularly that of the colorectum, might be through
an induction of genomic and site and gene-specific DNA hypomethylation (Choi
and Mason, 2002; Kim, 1999; Lamprecht and Lipkin, 2003).
Theoretically, increased blood levels of FA may interfere with cellular folate
transport and metabolism, or regulatory functions of the coenzyme by blocking
the enzymes and/or carrier proteins of natural folate through competitive inhibition
(Sauer et al., 2009). Thus, high doses of FA may induce effects like those
produced by deficiency of the bioactive forms of folate. High intake of FA may
exert an antagonistic effect towards natural folates due to accumulation of DHF,
which allosterically modulates thymidylate synthase (TS) and MTHFR, leading to
decreased cellular levels of thymidylate and 5MTHF(Matthews and Daubner,
1982; Sauer et al., 2009). DHFR and TS regulates translation through the direct
binding of several mRNAs, thereby repressing the translation of p53 mRNA, a
process that might prevent the activation of this tumor suppressor genes
(Banerjee et al., 2002; Liu et al., 2002). Also, thymidylate deficiency can impair
DNA integrity and cellular division. Decreased levels of 5-MTHF also can
decrease methionine biosynthesis, thereby affecting protein production and DNA
methylation. In addition, high doses of FA may simply saturate DHFR and
potentially inhibit the entire folate metabolism (Bailey and Ayling, 2009). The
consequences of this is altered protein expression (Jhaveri et al., 2001),
decreased DNA repair capability and accumulation of DNA damage (Blount et al.,
1997; Duthie and Hawdon, 1998; Kruman et al., 2002), increased chromosomal
127

aberrations and fragility (Duthie, 1999); events that ultimately reduce growth rate
and impair cell division.
The results of these studies may seem convincing but has failed to translate to
epidemiological studies. The current gap is to understand the specific role that
supplementation of different doses of either form of folate- FA or 5MTHF, will
have on colon cells in different stages of CRC development. With the availability
of a specific and sensitive multi-analyte metabolomics-based LCMS assay, the
dynamics of the FOCM metabolites and genes will provide a useful insight to the
ongoing debate.
5.2 Materials and Method
5.2.1 Cell lines and culture
Two untransformed human colon cells (CRL1459 and CRL1790 and three human
colon adenocarcinoma cell lines, HCT116 p536+/+, HCT116 P53 -/- and Caco-2
cells, were obtained from the American Type Culture Collection (Manassas, VA).
The murine APCmin cell line obtained from Dr. Carla De Giovanni (University of
Bologna, Italy). The published characteristics of the colon cell lines that are
relevant to the study is presented in Table 4.1 (Ahmed et al., 2013). The cells
were cultured and were cultured in Dulbecco's Modified Eagle's Medium (DMEM;
Invitrogen, Gaithersburg, MD) with altered folate concentrations (Table 5.1). The
FA concentrations of 2.3 µM and 9 µM were selected because those are the FA
concentrations in recommended media like Roswell Park Memorial Institute
(RPMI) medium and DMEM respectively. Similar concentrations were used as
128

medium and high dose 5MTHF. A last treatment group consisting of low dose
5MTHF and high FA was also added to test for any interaction between the two
folate metabolites. Growth medium was supplemented with 10% fetal bovine
serum, 1% non-essential amino acids and 1% antibiotic-antimycotic. The cells
were maintained at 37
o
C in 95% humidity and 5% CO2 and passaged every four
days. All cells in treatment groups were harvested after 10-14 days of growth.
The cellular FOCM metabolite concentrations were determined at the last day of
growth as well as their global methylation levels. The expression of genes related
to the folate metabolism, transmethylation pathway, transsulfuration pathway,
DNA synthesis & repair, and proliferation & apoptosis.
Table 5.1: Study design showing the media treatment conditions of the colon cells
DMEM containing
CRL1459 CRL1790 APC 10.1 Caco-2 HCT116 P53 +/+ HCT116 P53 -/-
Folic Acid
0µM √ √ √ √ √ √
2.3µM √ √ √ √ √ √
9µM √ √ √ √ √ √
5MTHF
2.3µM √ √ √ √ √ √
9µM √ √ √ √ √ √
2.3µM + 9µM FA
(Combo)
√ √ √ √ √ √

5.2.2 Effect of folates on growth rates of colon cell lines
The different cell lines shown in Table 4.1 were treated with the DMEM media
conditions in Table 5.1. Approximately 500 cells of each cell line were seeded
129

into the treated media containing about 17% of Alamar blue. The baseline growth,
as well as the growth at various time points, were taken using the fluorescence
reading taken at Ex/Em of 545/590 nm which correlates with the number of cells
within a well. The fluorescence at each time point was calculated using the
baseline as a reference. The growth curves were fitted with logistic curve model
while sharing the carrying capacity of the system across all cell lines. The growth
rates of the different treatment conditions for a colon cell type were compared
using doubling time and mean adjusted area under the curve (AUC). Mean
adjusted AUC has been proven a valid summary statistic for repeated measures
conducted over a period of a tumor growth study (Frison and Pocock, 1992; Pham
et al., 1999). In the case of modeling tumor growth, adjusted AUC is the
standardized measure that can compare across treatment groups (Qian et al.,
2000).
5.2.3 Extraction and analysis of cellular FOCM metabolites
The validated LCMS assay in Chapter 2 of this dissertation was used to analyze
FOCM metabolites in the media and within the cells using three different
biological replicates. About 800 uL of the media used for growing the cells was
pipetted into an Eppendorf tube and centrifuged at 1500rpm for 10mins to
separate the supernatant from any floating cells. The supernatant was separated
for extracellular FOCM metabolite analysis. The treated cells were trypsinized
and washed three times with cold PBS. The cell pellets obtained from the PBS
wash was stored at −80°C until analysis.
130

The extracellular and cellular FOCM metabolites were extracted by taking 50 uL
of the media or cell pellet, adding the required internal standards, reducing agents
and the precipitating solution made of 20% 0.2M Zinc Sulphate in methanol. After
keeping this mixture in the -20°C for 30mins, it was centrifuged at 15,000 rpm for
15mins. About 100 uL of the supernatant was taken for the analysis of HCY and
its metabolites using a published assay (Kořínek et al., 2013) while the remaining
was transferred into a new 1.5 mL micro centrifuge tubes for drying under nitrogen
gas. Just before LCMS analysis, the samples were reconstituted in 30 uL of 1%
ascorbic acid, centrifuged at 5000 rpm for a minute and transferred into HPLC
vials for analysis using the validated assay.
5.2.4 RNA and DNA isolation
Genomic DNA was isolated from folate-treated cells using TRIzol (Thermo Fisher,
Waltham, MA, USA) per the manufacturer's protocol. The concentration and
purity were evaluated at the absorbance at 230, 260, and 280 nm using a
Nanodrop ND-1000 (Thermo Scientific, Waltham, MA, USA). Samples that did
not attain the required purity were further purified by precipitating the genomic
material, washing, and resuspension.
5.2.5 DNA hydrolysis and global methylation measurement by LCMS
The hydrolysis of the DNA and the analysis of the global DNA methylation was
performed after modifying procedures published by Li and Franke (2011). Digest
Mix (enough for 100 samples) was prepared using 250 U Benzonase, 300 mU
131

phosphodiesterase, 200 U alkaline phosphatase, and 0.5 ml 10× Buffer [100 mM
Tris–HCl, 500 mM NaCl, 100 mM MgCl2 (pH 7.9)], and made up to 5 ml with
water. DNA samples (1 µg in 50-μl water) were digested by adding 50 μL of Digest
Mix and incubating at 37°C for six hours. After hydrolysis, 900 𝜇 L of HPLC-grade
water was added to each sample, protecting the sample from light.
Global DNA methylation was quantified using a Shimadzu Prominence HPLC
system linked to an API 4000 LCMS/MS spectrometer (Applied Biosystems,
Foster City, CA) operating in the positive mode. Ten microliters of DNA digest
containing the internal standard was injected onto a HyPurity C18 column (50 mm
× 4.6 mm, 3 μm, Thermo), using a mobile phase consisting of 0.1% aq. Formic
acid in water and 0.1% formic acid in methanol at a flow rate of 4000 μL/min using
a gradient elution over a total run time of 4.6 min.
During analysis, the LCMS/MS operated in the positive mode with the source
temperature at 500 °C, collision gas at 12 psi and curtain gas at 10 psi. The ion
source gas (1) was set at 50 psi, the ion source gas (2) at 50 psi and Ion Spray
Voltage at 55000 V. Data acquisition and analysis were performed using the Sciex’
MultiQuant software.
5.2.6 Real time polymerase chain reaction
The extracted RNA was purified and converted to cDNA using standard operating
procedures from the reagent manufacturer. Complementary DNAs (cDNAs) were
generated from 2 μg total RNA by reverse transcriptase (RT) using oligodT
132

primers and the SuperScript III RNase H-Reverse Transcriptase (Invitrogen) per
the manufacturer's protocol.
Levels of mRNAs encoding the expression levels of target genes related to the
FOCM were analyzed by quantitative real-time RT-PCR assay carried out in
quadruplicates using an ABI OpenArray Real-Time PCR (Applied Biosystems,
USA). The primers used for the RT-PCR (Table 5.3) were placed in different
exons to minimize the likelihood of amplifying contaminated genomic DNA. The
PCR amplification was performed in reaction mixtures consisting of 5 μL SYBR
Green master mix (Thermo Scientific, USA), 0.2 μL each of forward and reverse
primers (0·2 μM final concentration) and 0.8 μL of cDNA template in a 10-μl final
reaction volume. cDNA samples were amplified with a precycling heat activation
at 95 °C for 10 min, followed by 40 cycles of heat denaturation at 95 °C for 15 s
and annealing and extension at 60 °C for 1 min. The mRNA levels of the target
genes in colon cells are expressed as the ratio of the target gene mRNA to β-
ACTIN (internal control) mRNA for each sample. The relative quantification of the
mRNA in each sample was determined using the 2
-ΔΔCt
method (Livak and
Schmittgen, 2001). Target genes and their primers are presented in Table 4.3.
5.3 Statistical analysis
Quantitative data was analyzed using GraphPad Prism software (GraphPad
Software Inc., La Jolla, CA, USA). Using Analysis of Variance (ANOVA) analysis,
the means + SEM of the data will be compared within the treatment groups for
the different cell lines. Post-hoc analysis was conducted on data that show
133

significant differences between the treatment groups within a cell line. At least
three biological replicates were used for each statistical analysis, and treatments
were considered significantly different if statistical tests produced a p-value ≤0.05,
except correlations between the genes which were tested at a p-value ≤0.01.
5.4 Results
5.4.1 Effect of folate treatment on cellular growth
Each of the colon cell type grew in the various treatment media at different rates
expressing significant differences in their doubling time (p-values <0.05, Table
5.2). In all the colon cell lines, the introduction of 2.3 µM 5MTHF to the standard
DMEM with 9 M FA (Combo) resulted in a significant change in doubling time.
Further analysis of the growth rates of the cells (measured by mean adjusted area
under the curve (AUC) of the fluorescence) showed significant differences in all
cell types. However, some media treatment groups for some cell type did not
show significant differences in their mean adjusted AUC. In all the colon cell lines,
except APC 10.1, the growth rates increased with increasing folic acid
concentration (Figures 5.1 & 5.2).
134

Figure 5.1: Growth curves of colon cell lines in different treatment media showing
increasing growth in increasing FA concentration, except APC 10.1.

Table 5.2: Comparison of the mean doubling time (+SEM) in hours for the various colon
cell lines grown in different treatment media.
Colon Cell No Folate 2.3uM FA 9uM FA 2.3uM 5MTHF 9uM 5MTHF Combo ANOVA
CRL1459 17.8+0.5 16.9+0.5 11.5+0.3 18.8+0.6 18.4+0.5 9.9+0.3 <0.0001
CRL1790 20.3+0.7 17.6+0.5 13.0+0.4 18.9+0.6 18.2+0.6 14.9+0.5 <0.0001
APC 10.1 14.8+0.8 12.8+0.7 23.5+1.4 15.1+0.8 14.3+0.8 37.0+3.1 <0.0001
Caco-2 16.0+0.7 15.5+0.7 14.8+0.6 15.2+0.6 15.9+0.7 17.4+0.8 0.013
HCT116 p53-/- 21.4+0.7 18.9+0.6 13.5+0.4 20.0+0.6 19.3+0.6 14.4+0.4 <0.0001
HCT116 p53+/+ 20.3+0.7 17.6+0.5 13.1+0.4 18.9+0.6 18.2+0.6 14.9+0.5 <0.0001

H o u rs o f C u ltu re
R e la tiv e F lu o re s c e n c e to B a s e lin e
0 2 0 4 0 6 0 8 0 1 0 0
0
2
4
6
8
1 0
H o u rs o f C u ltu re
R e la tiv e F lu o re s c e n c e to B a s e lin e
0 2 0 4 0 6 0 8 0
0
1 0
2 0
3 0
4 0
5 0
H o u rs o f C u ltu re
R e la tiv e F lu o re s c e n c e to B a s e lin e
0 2 0 4 0 6 0 8 0
0
1 0
2 0
3 0
4 0
H o u rs o f C u ltu re
R e la tiv e F lu o re s c e n c e to B a s e lin e
0 2 0 4 0 6 0 8 0 1 0 0
0
5 0
1 0 0
1 5 0
H o u rs o f C u ltu re
R e la tiv e F lu o re s c e n c e to B a s e lin e
0 2 0 4 0 6 0 8 0 1 0 0 1 2 0 1 4 0
0
1 0
2 0
3 0
4 0
H o u rs o f C u ltu re
R e la tiv e F lu o re s c e n c e to B a s e lin e
0 2 0 4 0 6 0 8 0 1 0 0
0
2 0
4 0
6 0
8 0
1 0 0
CRL 1459 CRL 1790 APC 10.1
CACO-2 HCT 116 p53+/+ HCT 116 p53-/-
135

Figure 5.2: Comparison of the growth rates (indicated by mean adjusted AUC of
fluorescence) of colon cell lines in different treatment media.

5.4.2 Effect of folate on colony formation
The various cells (except APC 10.1) cultured in folate deficient medium (DMEM)
for 10-14 days, grew despite the relative absence of folic acid in the medium.
Also, the CRL1459 failed to form colonies over the 14-day period for the various
media conditions. The HCT116 p53+/+ and APC 10.1 cells grown rather 5MTHF-
sufficient medium showed an intermediate growth rate but altered morphology in
the cells (data not shown). In the transitional and transformed cell lines, the
presence of 5MTHF resulted in the reduction in the number of colonies
significantly, with the most drastic suppression of colony formation occurring in
C R L 1 4 5 9
M e a n a d ju s te d A U C
o f flu o re s c e n c e
N o F o la te
2 .3 u M F A
9 u M F A
2 .3 u M 5 M T H F
9 u M 5 M T H F
C o m b o
0
5
1 0
1 5
2 0
C R L 1 7 9 0
M e a n a d ju s te d A U C
o f flu o re s c e n c e
N o F o la te
2 .3 u M F A
9 u M F A
2 .3 u M 5 M T H F
9 u M 5 M T H F
C o m b o
0
5
1 0
1 5
2 0
A P C 1 0 .1
M e a n a d ju s te d A U C
o f flu o re s c e n c e
N o F o la te
2 .3 u M F A
9 u M F A
2 .3 u M 5 M T H F
9 u M 5 M T H F
C o m b o
0
2
4
6
C A C O -2
M e a n a d ju s te d A U C
o f flu o re s c e n c e
N o F o la te
2 .3 u M F A
9 u M F A
2 .3 u M 5 M T H F
9 u M 5 M T H F
C o m b o
0
5
1 0
1 5
2 0
H C T 1 1 6 p 5 3 + /+
M e a n a d ju s te d A U C
o f flu o re s c e n c e
N o F o la te
2 .3 u M F A
9 u M F A
2 .3 u M 5 M T H F
9 u M 5 M T H F
C o m b o
0
1 0
2 0
3 0
4 0
5 0
H C T 1 1 6 p 5 3 -/-
M e a n a d ju s te d A U C
o f flu o re s c e n c e
N o F o la te
2 .3 u M F A
9 u M F A
2 .3 u M 5 M T H F
9 u M 5 M T H F
C o m b o
0
1 0
2 0
3 0
4 0
136

HCT116 p53+/+. There was enhanced growth in the HCT116 p53-/- cells when
they were supplemented with 5MTHF (Figure 5.3).

Figure 5.3: Colony forming assay showing the effect of FA and 5MTHF supplementation
on transitional and transformed colon cells.

5.4.3 Effect of folate on cellular folate proportions
Cellular folate proportions were compared in some colon cells cultured in the
different media conditions. There were significant differences in the levels and
proportions of folates across the various media supplementations (Figure 5.4).
The cellular total folate levels increased with increasing concentration of FA or
5MTHF supplement. The gradient of total folate change was highest in HCT116
p53+/+, followed by APC 10.1 and then Caco-2. In HCT116 p53+/+ and APC
10.1, the cellular 5MTHF levels increased with increasing 5MTHF
137

supplementation, that trend was not seen for the FA supplementation. This trend
was consistent with the methyl trapping phenomenon observed in these cells.
Although the 5MTHF was absorbed in a dose dependent manner, they were
trapped at the cellular level within the FOCM.
In contrast, Caco-2 minimal increase in cellular 5MTHF with increasing
supplementation of 5MTHF; even the 9 µM FA group showed the highest level of
5MTHF. This may infer the folate receptors in Caco-2 transport FA more
efficiently than 5MTHF. Also, Caco-2 may have the most efficient conversion of
FA to 5MTHF compared to the other cells. APC 10.1 cells had higher proportions
of FA and DHF accumulated than the other cells. The residual levels of FA and
DHF may be relevant for their role as enzyme modulators in the FOCM. It is also
noteworthy that the proportion of THF was so high in HCT116 p53+/+ and APC
10.1 (even up to 90%) in some treatment groups, diminishing in groups with high
5MTHF accumulation. Also, the folate metabolism genes were upregulated in the
transitional and transformed colon cells (Figure 5.7). CRL1790 also showed an
overexpression of DHFR and MTHFR.
138

Figure 5.4: Cellular folate proportions in CRL1459, APC 10.1, HCT116 p53+/+ and
Caco-2. The absolute cellular levels in ng/mg of cell protein are indicated as (A) and
the cellular proportions indicated as (B). Total folates increased with increasing
concentrations of FA or 5MTHF supplement. The cellular 5MTHF also increased in
a dose response manner.

5.4.4 Effect of folate supplementation on FOCM metabolites in Colon Cell lines
The levels of FOCM metabolites varied across treatment groups and cell types.
The SAM/SAH ratio which indicates the methylation capacity was compared
across the treatment media for each cell type. The 9 µM showed the highest
methylation capacity in CRL1459, APC 10.1 and HCT116 p53+/+. In the case of
Caco-2, the Combo group had the highest methylation capacity. The methylation
capacity generally increased with increasing FA concentration and with an
increase from 2.3 µM 5MTHF to 9 µM 5MTHF (Figure 5.5). The global methylation
0
500
1000
1500
2000
2500
APC 10.1 A
0
1000
2000
3000
4000
5000
6000
CACO-2 A
0
5000
10000
15000
20000
25000
30000
HCT 116 p53+/+ A
0%
20%
40%
60%
80%
100%
APC 10.1 B
0%
20%
40%
60%
80%
100%
CACO-2 B
0%
20%
40%
60%
80%
100%
HCT 116 p53+/+ B
0
1000
2000
3000
4000
5000
Folate level (ng/mg of cell protein)
CRL 1459 A
0%
20%
40%
60%
80%
100%
CRL 1459 B
139

of cells decreased as one moves from 2.3 µM to 9 µM folate (FA or 5MTHF)
supplements. The Combo group in Caco-2 had the lowest global methylation
(Figure 5.6.)

Figure 5.5: Methylation capacity of APC 10.1, Caco-2 and HCT116 p53+/+ grown in
various treatment media. No Folate group for APC 10.1 is missing due to insufficient
sample.

Figure 5.6: Relative global methylation of APC 10.1, Caco-2 and HCT116 p53+/+ grown
in various treatment media. No Folate group for APC 10.1 is missing due to
insufficient sample.

M e d ia C o n d itio n s
S A M /S A H R a tio
N o F o la te
2 .3 u M F A
9 u M F A
2 .3 u M M T H F
9 u M M T H F
C o m b o
1 0
0
1 0
1
1 0
2
1 0
3
1 0
4
M e d ia C o n d itio n s
S A M /S A H R a tio
N o F o la te
2 .3 u M F A
9 u M F A
2 .3 u M M T H F
9 u M M T H F
C o m b o
1 0
0
1 0
1
1 0
2
1 0
3
1 0
4
M e d ia C o n d itio n s
S A M /S A H R a tio
2 .3 u M F A
9 u M F A
2 .3 u M M T H F
9 u M M T H F
C o m b o
0
2 0 0
4 0 0
6 0 0
8 0 0
1 0 0 0
APC 10.1
HCT 116 p53+/+
CACO-2
C A C O -2
M e d ia C o n d itio n s
re la tiv e % 5 m C
N o F o la te
2 .3 u M
9 u M
2 .3 u M 5 M T H F
9 u M 5 M T H F
C o m b o
0 .0
0 .5
1 .0
1 .5
2 .0
H C T 1 1 6 p 5 3 + /+
M e d ia C o n d itio n s
re la tiv e % 5 m C
N o F o la te
2 .3 u M F A
9 u M F A
2 .3 u M 5 M T H F
9 u M 5 M T H F
C o m b o
0 .0
0 .5
1 .0
A P C 1 0 .1
M e d ia C o n d itio n s
re la tiv e % 5 m C
2 .3 u M F A
9 u M F A
2 .3 u M M T H F
9 u M M T H F
C o m b o
0 .0
0 .5
1 .0
1 .5
140

5.4.5 Expression of FOCM genes across colon cell lines
The gene expression levels of FOCM related genes were compared within colon
cell lines and across treatment groups, using the 9 µM FA as reference. There
was a general upregulation in the genes related to folate metabolism,
proliferation, transmethylation and transsulfuration in the HCT116 p53+/+ folate
deficient and 2.3 µM FA groups (which are decreased FA groups). The Combo
group rather had these genes downregulated alongside DNA methylation genes.
It can be inferred that increasing FA concentration caused the decline in
expression of the genes related to folate metabolism, proliferation,
transmethylation and transsulfuration in the HCT116 p53+/+. In the case of
HCT116 p53-/-, all the treatment groups except the Combo had a downregulation
in the folate metabolism, proliferation and transmethylation genes. In the case of
Caco-2, all groups except the Combo had an upregulation in almost all the genes
but CRL1790 had downregulation in FOLR1, MTHFR, BCL-2 and DNMT3A
(Figure 5.7). The Caco-2 Combo group had the least growth rate and may be
attributable to overexpression in the DHFR, CBS, IGF2 and DNMT3B genes.
141

Figure 5.7: Expression of genes of cell lines after treatment with various folate
supplements. Red color with (↑) indicates an upregulation, the blue color with (↓)
indicates a downregulation and (↔) indicates the normal expression. The blank one
had no expression.

5.4.6 Correlations between genes, growth rate and CRC status
The correlation of the genes with increasing FA or 5MTHF supplementation
differed among the cell lines. Most of the trends observed in the FA
supplementation was the reverse of that in 5MTHF supplementation (Figure 5.8).
It can thus be inferred that the effect of FA supplementation is very different from
that of 5MTHF supplementation in untransformed or transformed cells. The genes
FOLR1 MTHFR DHFR MLH1 MGMT TYMS C-MYC BCL-2 RASSF1A TP53 IGF2 H19 MTR MTRR MAT CBS GSS DNMT1 DNMT3A DNMT3B
0 uM ↑ ↑ ↑ ↑ ↔ ↑ ↑ ↑ ↔ ↔ ↔ ↔ ↔ ↑ ↑ ↔ ↑ ↑ ↔ ↔
2.3 uM FA ↔ ↔ ↑ ↔ ↔ ↑ ↔ ↑ ↔ ↔ ↔ ↔ ↔ ↑ ↑ ↑ ↑ ↔ ↔ ↔
9 uM FA ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔
2.3 uM 5MTHF ↓ ↔ ↓ ↔ ↔ ↓ ↔ ↑ ↔ ↑ ↔ ↓ ↔ ↔ ↔ ↔ ↑ ↔ ↔ ↔
9 uM 5MTHF ↓ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↑ ↔ ↔ ↑ ↔ ↔ ↔
Combo ↓ ↓ ↓ ↔ ↔ ↓ ↓ ↓ ↔ ↔ ↓ ↓ ↓ ↔ ↔ ↔ ↔ ↓ ↓ ↔
0 uM ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↔ ↔ ↓ ↓ ↓ ↔ ↓ ↔ ↔ ↓ ↔ ↔
2.3 uM FA ↓ ↓ ↓ ↔ ↔ ↓ ↔ ↔ ↔ ↔ ↓ ↓ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↑
9 uM FA ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔
2.3 uM 5MTHF ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↔ ↔ ↓ ↓ ↓ ↓ ↓ ↓ ↔ ↓ ↔ ↔
9 uM 5MTHF ↓ ↓ ↓ ↓ ↓ ↓ ↓ ↔ ↔ ↑ ↓ ↓ ↔ ↔ ↓ ↔ ↔ ↓ ↔ ↑
Combo ↔ ↔ ↓ ↓ ↔ ↓ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↓ ↔ ↔
0 uM ↑ ↔ ↑ ↑ ↑ ↑ ↑ ↔ ↑ ↑ ↑ ↔ ↑ ↑ ↑ ↑ ↑ ↑ ↔ ↔
2.3 uM FA ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↔ ↑ ↑ ↑ ↑ ↑ ↑ ↔ ↑
9 uM FA ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔
2.3 uM 5MTHF ↑ ↔ ↑ ↑ ↔ ↑ ↓ ↑ ↔ ↑ ↑ ↔ ↑ ↑ ↑ ↑ ↑ ↔ ↑ ↔
9 uM 5MTHF ↑ ↑ ↑ ↑ ↑ ↑ ↑ ↔ ↑ ↑ ↑ ↔ ↑ ↑ ↑ ↑ ↑ ↑ ↔ ↑
Combo ↔ ↔ ↑ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↑ ↔ ↔ ↔ ↔ ↑ ↔ ↑ ↔ ↑
0 uM ↔ ↑ ↓ ↔ ↔ ↓ ↔ ↑ ↓ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↑ ↓
2.3 uM FA ↑ ↑ ↓ ↔ ↔ ↔ ↔ ↑ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↓
9 uM FA ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔
2.3 uM 5MTHF ↔ ↔ ↓ ↔ ↔ ↓ ↔ ↑ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↓
9 uM 5MTHF ↑ ↑ ↓ ↔ ↔ ↔ ↔ ↑ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↑ ↓
Combo ↔ ↑ ↓ ↔ ↔ ↓ ↔ ↑ ↔ ↑ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↑ ↓
0 uM ↓ ↔ ↔ ↔ ↔ ↔ ↓ ↔ ↔ ↔ ↓ ↔ ↔
2.3 uM FA ↓ ↔ ↓ ↔ ↔ ↔ ↓ ↔ ↔ ↔ ↔ ↔ ↔
9 uM FA ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔
2.3 uM 5MTHF
9 uM 5MTHF ↓ ↔ ↔ ↔ ↔ ↓ ↔ ↔ ↔ ↔ ↔ ↔ ↔
Combo ↔ ↔ ↑ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔ ↔
HCT 116 p+13+/+ HCT 116 P+13-/- CACO-2 CRL1790 APC10.1
142

were also evaluated for correlation with each other. HCT116 p53+/+ showed the
highest degree of positive correlation among the genes while HCT116 p53-/-
showed the least positive correlation with a significant negative correlation
between DNMT3B and both DNMT3A and BCL-2 (Figure 5.9).

Figure 5.8: Correlation of FOCM related genes in colon cells with FA or 5MTHF
concentration as supplement. The blue bars show significant positive correlation
and the red, significant negative correlation (p<0.01).

FA CONC 5MTHF CONC FA CONC 5MTHF CONC FA CONC 5MTHF CONC FA CONC 5MTHF CONC
FOLR1 0.9 -0.94 0.94 0.96
DHFR 0.95 0.92 -0.8 0.96 -0.86
MTHFR -0.87 0.96 0.88 -0.92
MTR -0.83 0.95
MTRR -0.88
MAT 0.95
CBS 0.85
TYMS 0.9 -0.86 0.96 -0.89 -0.84
DNMT1 -0.87 0.89 -0.83
DNMT3A -0.94 0.8 0.92 -0.89
DNMT3B 0.97 0.85 -0.84
MLH1 -0.86 0.91 -0.96
TP53 -0.92 -0.84
MGMT 0.93 -0.89
C-MYC
BCL2 -0.87 -0.8
RASSF1A -0.83 0.8 0.93
IGF2 0.94
H19 0.92
GSS -0.85 0.91 0.92 -0.98
HCT 116 p53+/+ CRL 1790 CACO-2 HCT 116 p53-/-
143

Figure 5.9: Correlation matrix of FOCM related genes in colon cells supplemented with
different levels of folates. The circles show pairs with significant correlation
(p<0.01).

5.5 Discussion
We studied the impact of FA or 5MTHF on the FOCM of colon cell lines in different
stages of CRC development. The primary goal was to probe whether expression
of FOCM genes and metabolites’ cellular concentrations are altered when folate
supplements are provided in their environment. An alteration in the FOCM
metabolites and their related genes may affect DNA biosynthesis, methylation,
folate metabolism and cell proliferation, all of which are relevant to
144

carcinogenesis. The various cell lines were grown in media with different
concentrations of FA and 5-MTHF. 5-MTHF can be readily feed into the
transmethylation pathway but folic acid must first be processed through several
enzymatic steps before it can be used as a methyl donor. This implies that 5MTHF
supplementation will circumvent any defects in the upstream of folate metabolism
pathway. We examined the effect of each folate type in each cell type.
5.5.1 Effect of folate supplementation on HCT116 p53+/+ cells
HCT116 p53+/+ cells did not have any noticeable trends in growth rate when
grown in different concentrations of FA or 5-MTHF. However, the growth rate of
the cells was the slowest when grown in the Combo group. The Combo group
was expected to cause the HCT116 p53+/+ cells to grow the fastest because the
folate metabolites in the growth medium were in excess. The cells should have
had more than enough folates for DNA synthesis, methylation, and other key
metabolic processes. However, the excessive folate metabolites in the growth
medium caused a downregulation of the FOLR1, MTHFR, and DHFR genes,
which encode enzymes necessary for metabolizing FA into useful reduced forms
like THF or 5-MTHF. Previous studies have also shown that FOLR1 expression
is inversely dependent on the extracellular folate concentration (Kelemen, 2006).
Such downregulation of folate metabolism genes may be due to a negative
feedback inhibition.
Also, consistent with the decreased growth rate of the cells in Combo group was
the significant downregulation of thymidylate synthase, an enzyme which is
145

necessary for the synthesis of thymidine, which is required for DNA synthesis.
Thymidylate synthase has also been shown to be upregulated in cancers and
used as a druggable target to downregulate to control cancer proliferation (Popat
et al., 2004). The Combo seemed to offer the desired downregulation and
consistently slowed proliferation in the cancer cells. The genes associated with
DNA damage repair, MGMT and MLH1, were also downregulated under excess
folate conditions, which is consistent with the lower growth rate because less DNA
is being synthesized, so there is a lower need for DNA repair. Surprisingly,
several genes involved in cancer progression and cell cycle regulation, BCL-2, c-
MYC, and TP53, showed a significant downregulation as extracellular folate
levels were increased.
One of the key genes that drives cell growth and proliferation is IGF2. IGF2 is
downregulated in the Combo group, which may be the cause for the decrease
growth rate of the cells. MTR, MAT, and GSS, which are key genes involved in
the transsulfuration/transmethylation pathway, also showed strong
downregulation with increasing FA and 5-MTHF levels. A downregulation of these
genes may be bad for HCT116 p53+/+ who depend on these genes to drive
hypermethylation of tumor suppressor genes and generate GSH to buffer reactive
oxidative species. The downregulation of genes in this pathway, along with the
downregulation of DNMT1 and DNMT3B, suggests that global methylation is
being decreased under high folate conditions. Overall, the general trend of gene
expression involving folate metabolism, methylation, and cell cycle regulation in
HCT116 p53+/+ cells decreased when FA and 5-MTHF was increased.
146

5.5.2 Effect of folate supplementation on HCT116 p53-/- cells
HCT116 p53-/- cells had similar growth rates when grown in the different types of
folate media. Unlike the gene expression trend found in HCT116 p53+/+, the
expression of FOLR1 and MTHFR in HCT116 p53-/- cells increased as folate
levels in the growth media were increased. This trend was opposite of what was
seen in HCT116 p53+/+ cells. It appears that the P53 status of the cells is the
reason for this difference in gene expression; however, the mechanism by which
this occurs is unclear. MLH1, MGMT, and TYMS were upregulated in 2.3uM and
9uM FA. There were no clear trends in the oncogenes BCL, c-MYC, and
RASSF1A. Increasing folate concentrations caused an overexpression in IGF2,
but this was not reflected in the growth curve. The reason for this is most likely
due to H19 expression, a noncoding RNA which negatively regulates IGF2 (Kaffer
et al., 2001). H19 expression showed a far greater fold change than IGF2
expression, which would explain the unchanged growth rate.
5.5.3 Effect of folate supplementation on Caco-2 cells
Caco-2 cells had the lowest growth rate in the Combo. The growth rate at other
concentrations of FA and 5-MTHF were relatively similar. Although both HCT116
and Caco-2 are CRC cell lines, the main differences are that Caco-2 has a p53
mutation and that it is microsatellite stable. Unlike the HCT116 p53+/+ and
HCT116 p53-/-, Caco-2 did not show any changes in FOLR1 expression in
response to changing extracellular folate levels. The expression of MTHFR and
DHFR showed no clear trends, but expression was significantly lower in 9uM FA
147

and Combo groups. The genes involved in DNA damage repair and DNA
synthesis also showed decreased expression in the presence of high FA.
Interestingly, most the genes studied were highly downregulated in 9uM FA, but
the downregulation was not as great as in the Combo. This suggests that 2.3uM
5-MTHF used in the Combo is somewhat reversing the gene expression changes
caused by 9uM FA. However, the presence of 5-MTHF alone did not lead to the
upregulation of the genes. Compared to HCT116 WT, Caco-2 cells had far fewer
genes that were affected by changing folate conditions in the growth medium.
5.5.4 Effect of folate supplementation on CRL1790
The gene expression changes under different media folate concentrations were
also examined in the CRL1790 cell line. The CRL1790 cell line is derived from
the human fetal colon. Although these cells are not cancerous, they are still
rapidly growing, which is to be expected in the developing fetus. FOLR1
expression increased with increasing 5-MTHF levels without FA. The expression
of MTHFR showed a decreasing trend with increasing FA concentration. The
expression of DHFR was inversely related to MTHFR expression. This is likely
because DHFR is a negative regulator of MTHFR in the folate metabolism
pathway. TYMS expression showed an increasing trend with increasing
concentration of FA, but the gene was downregulated in the Combo. The
expression of transmethylation/ transsulfuration pathway genes was unaffected
by changing FA or 5-MTHF levels. Interestingly, the gene expressions of
148

DNMT3A and of DNMT3B were inversely related, but there were no clear trends
with increasing folate concentrations.
5.6 Conclusion
The FOCM is altered when colon cells are supplemented with different levels of
FA or 5MTHF supplement. The cellular total folates generally increased with
increasing supplementation concentration but folate proportions differed based
on the cell type. Supplementation with 9 µM FA was associated with significant
increase in growth rate in untransformed and transformed cell lines but slowed
growth in the transitional ones (APC 10.1). The introduction of 2.3 µM of 5MTHF
to the 9 µM FA to make the Combo was associated with reduced growth in the
transitional and transformed cell lines. Also, DNA global methylation was found
to increase with increasing dose of supplement.
149

Chapter 5: Bibliography and References
Ahmed, D., Eide, P., Eilertsen, I., Danielsen, S., Eknaes, M., Hektoen, M., Lind,
G., and Lothe, R. (2013). Epigenetic and genetic features of 24 colon cancer cell
lines. Oncogenesis 2, e71.
Bailey, S.W., and Ayling, J.E. (2009). 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.
Banerjee, D., Mayer-Kuckuk, P., Capiaux, G., Budak-Alpdogan, T., Gorlick, R.,
and Bertino, J.R. (2002). Novel aspects of resistance to drugs targeted to
dihydrofolate reductase and thymidylate synthase. Biochimica et Biophysica
Acta (BBA)-Molecular Basis of Disease 1587, 164-173.
Bird, A. (2002). DNA methylation patterns and epigenetic memory. Genes &
development 16, 6-21.
Blount, B.C., Mack, M.M., Wehr, C.M., MacGregor, J.T., Hiatt, R.A., Wang, G.,
Wickramasinghe, S.N., Everson, R.B., and Ames, B.N. (1997). 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.
Burt, R.W., Barthel, J.S., Dunn, K.B., David, D.S., Drelichman, E., Ford, J.M.,
Giardiello, F.M., Gruber, S.B., Halverson, A.L., and Hamilton, S.R. (2010).
150

Colorectal cancer screening. Journal of the National Comprehensive Cancer
Network 8, 8-61.
Cho, E., Zhang, X., Townsend, M.K., Selhub, J., Paul, L., Rosner, B., Fuchs,
C.S., Willett, W.C., and Giovannucci, E.L. (2015). Unmetabolized Folic Acid in
Prediagnostic Plasma and the Risk for Colorectal Cancer. Journal of the National
Cancer Institute 107, djv260.
Choi, S.-W., and Mason, J.B. (2002). Folate status: effects on pathways of
colorectal carcinogenesis. The Journal of nutrition 132, 2413S-2418S.
de Vogel, S., Schneede, J., Ueland, P.M., Vollset, S.E., Meyer, K., Fredriksen,
Å., Midttun, Ø., Bjørge, T., Kampman, E., and Bretthauer, M. (2011). Biomarkers
related to one-carbon metabolism as potential risk factors for distal colorectal
adenomas. Cancer Epidemiology Biomarkers & Prevention 20, 1726-1735.
Duthie, S.J. (1999). Folic acid deficiency and cancer: mechanisms of DNA
instability. Br Med Bull 55, 578-592.
Duthie, S.J., and Hawdon, A. (1998). DNA instability (strand breakage, uracil
misincorporation, and defective repair) is increased by folic acid depletion in
human lymphocytes in vitro. FASEB J 12, 1491-1497.
Frison, L., and Pocock, S.J. (1992). Repeated measures in clinical trials: analysis
using mean summary statistics and its implications for design. Statistics in
medicine 11, 1685-1704.
151

Gupta, A.K., Brenner, D.E., and Turgeon, D.K. (2008). Early Detection of Colon
Cancer. Molecular diagnosis & therapy 12, 77-85.
Jhaveri, M.S., Wagner, C., and Trepel, J.B. (2001). Impact of extracellular folate
levels on global gene expression. Mol Pharmacol 60, 1288-1295.
Kaffer, C.R., Grinberg, A., and Pfeifer, K. (2001). Regulatory Mechanisms at the
MouseIgf2/H19 Locus. Molecular and cellular biology 21, 8189-8196.
Kelemen, L.E. (2006). The role of folate receptor α in cancer development,
progression and treatment: cause, consequence or innocent bystander?
International journal of cancer 119, 243-250.
Kim, Y.I. (1999). Folate and carcinogenesis: evidence, mechanisms, and
implications. J Nutr Biochem 10, 66-88.
Kim, Y.I. (2003). Role of folate in colon cancer development and progression. J
Nutr 133, 3731S-3739S.
Kořínek, M., Šístek, V., Mládková, J., Mikeš, P., Jiráček, J., and Selicharová, I.
(2013). Quantification of homocysteine ‐related metabolites and the role of
betaine–homocysteine S ‐methyltransferase in HepG2 cells. Biomedical
Chromatography 27, 111-121.
Kruman, I.I., Kumaravel, T.S., Lohani, A., Pedersen, W.A., Cutler, R.G., Kruman,
Y., Haughey, N., Lee, J., Evans, M., and Mattson, M.P. (2002). Folic acid
deficiency and homocysteine impair DNA repair in hippocampal neurons and
152

sensitize them to amyloid toxicity in experimental models of Alzheimer's disease.
J Neurosci 22, 1752-1762.
Lamprecht, S.A., and Lipkin, M. (2003). Chemoprevention of colon cancer by
calcium, vitamin D and folate: molecular mechanisms. Nature Reviews Cancer
3, 601-614.
Li, X., and Franke, A.A. (2011). High-throughput and cost-effective global DNA
methylation assay by liquid chromatography-mass spectrometry. Analytica
chimica acta 703, 58-63.
Liu, J., Schmitz, J.C., Lin, X., Tai, N., Yan, W., Farrell, M., Bailly, M., Chen, T.-
m., and Chu, E. (2002). Thymidylate synthase as a translational regulator of
cellular gene expression. Biochimica et Biophysica Acta (BBA)-Molecular Basis
of Disease 1587, 174-182.
Livak, K.J., and Schmittgen, T.D. (2001). Analysis of relative gene expression
data using real-time quantitative PCR and the 2− ΔΔCT method. methods 25,
402-408.
Matthews, R.G., and Daubner, S.C. (1982). Modulation of
methylenetetrahydrofolate reductase activity by S-adenosylmethionine and by
dihydrofolate and its polyglutamate analogues. Adv Enzyme Regul 20, 123-131.
153

O’Connell, J.B., Maggard, M.A., and Ko, C.Y. (2004). Colon cancer survival rates
with the new American Joint Committee on Cancer sixth edition staging. Journal
of the National Cancer Institute 96, 1420-1425.
Pham, B., Cranney, A., Boers, M., Verhoeven, A., Wells, G., and Tugwell, P.
(1999). Validity of area-under-the-curve analysis to summarize effect in
rheumatoid arthritis clinical trials. The Journal of rheumatology 26, 712-716.
Popat, S., Matakidou, A., and Houlston, R.S. (2004). Thymidylate synthase
expression and prognosis in colorectal cancer: a systematic review and meta-
analysis. Journal of Clinical Oncology 22, 529-536.
Qian, W., Parmar, M., Sambrook, R., Fayers, P., Girling, D., and Stephens, R.
(2000). Analysis of messy longitudinal data from a randomized clinical trial.
Statistics in medicine 19, 2657-2674.
Rim, S.H., Joseph, D.A., Steele, C.B., Thompson, T.D., and Seeff, L.C. (2011).
Colorectal cancer screening - United States, 2002, 2004, 2006, and 2008.
MMWR Surveill Summ 60 Suppl, 42-46.
Sauer, J., Mason, J.B., and Choi, S.-W. (2009). Too much folate: a risk factor for
cancer and cardiovascular disease? Curr Opin Clin Nutr Metab Care 12, 30-36.
Selhub, J., and Miller, J.W. (1992). The pathogenesis of homocysteinemia:
interruption of the coordinate regulation by S-adenosylmethionine of the
154

remethylation and transsulfuration of homocysteine. The American journal of
clinical nutrition 55, 131-138.
Shane, B. (2010). Folate chemistry and metabolism. In Folate in Health and
Disease, L.B. Bailey, ed. (New York, NY, USA: CRC press), pp. 2-3.
Troen, A.M., Mitchell, B., Sorensen, B., Wener, M.H., Johnston, A., Wood, B.,
Selhub, J., McTiernan, A., Yasui, Y., Oral, E., et al. (2006). Unmetabolized folic
acid in plasma is associated with reduced natural killer cell cytotoxicity among
postmenopausal women. J Nutr 136, 189-194.
Ulrich, C.M., and Potter, J.D. (2006). Folate supplementation: too much of a good
thing? Cancer Epidemiology Biomarkers & Prevention 15, 189-193.
Ulrich, C.M., and Potter, J.D. (2007). Folate and cancer—timing is everything.
Jama 297, 2408-2409.
Vernon, S.W. (1997). Participation in colorectal cancer screening: a review. J
Natl Cancer Inst 89, 1406-1422.

155

Chapter 6: Conclusion and Future Directions
6.1 Summary
The development of reliable blood-based biomarkers for colorectal cancer (CRC)
disease is long overdue. CRC ranks as the third leading cause of cancer-related
death in the industrialized world (Siegel et al., 2016) and major burden worldwide
(Ferlay et al., 2015). In the United States, it is not only the fourth most commonly
diagnosed cancer but has an annual mortality rate of approximately 50,000
patients (Group, 2016). CRC results from the accumulation of sequential genetic
changes colorectal cells resulting in their transformation from normal glandular
epithelium into invasive adenocarcinoma.
A comparison of the methylation patterns in normal, and cancer cells reveal
differences in their patterns at CpG islands and globally across the genes. This
finding is supported by previous efforts by others. Aberrant DNA methylation in
cancer development can manifest as global hypomethylation, which correlated
with chromosomal instability. Additionally, hypermethylation of CpG islands results
in the silence tumor suppressor genes thus promoting tumorigenesis (Crider et al.,
2012). CRC is cancer with methylation-dependent epigenetic alterations and
sequential gene mutations in the tumorigenesis pathway(s). The folate one-carbon
metabolism (FOCM) has a critical role in driving DNA methylation patterns
because it is the primary source of methyl groups required for DNA methylation
and synthesis. The goal of this dissertation was to probe the role of the FOCM in
156

the development of CRC. Also, due to the high rate of CRC-related deaths, the
dissertation set as a secondary objective to use the insights into the role of FOCM
in CRC development to explore for blood-based biomarkers for CRC.
This dissertation was staged to contribute meaningfully to the contextual debate
on whether the folate supplementation project adopted by North America in 1996
has led to an increased CRC risk. After a thorough review of most of the
epidemiological studies that had established an association between folates and
CRC risk, it is apparent that the major gap in their analyses was associated with
the microbiological assay used to assess the plasma levels of folates. Limitation
of microbiological growth assays includes inability to distinguish between the
various folate metabolite(s) that are present in blood, thus unable to identify the
specific metabolite responsible for the biological activity (DeVries et al., 2005;
Shane et al., 1980). To more accurately tease out which folate metabolite(s) may
show association with CRC risk, it was necessary to identify and quantify the
specific FOCM and folate metabolites accurately and precisely.
To address the gaps in the microbiological assays as well as develop an effectively
method to quantitate FOCM metabolites for further profiling analysis, we have
proposed a metabolomics approach. Metabolomics presents an economical
approach to simultaneous profile and quantifies different metabolites present in a
biological sample in response to environmental or genomic influence.
Metabolomics has been a major supporting path of functional genomics, alongside
transcriptomics (mRNA profiling) and proteomics (van der Greef et al., 2003).
157

Liquid chromatography-mass spectrometry (LCMS) has been demonstrated to be
a precise platform that can accurately measuring specific metabolites, making
metabolomics a realistic approach. The researcher developed and validated a
metabolomics-based LCMS assays as a platform to quantify and explore the
plasma levels of FOCM intermediates. This assay was economical and sensitive
due to its multi-analyte nature and its requirement of just 50 µL of plasma. The
assay also presented a useful approach to quantify the vitamin B metabolites in
the FOCM accurately, thereby providing an alternative to quantifying them by the
microbiological assay. The chapter 2 of this dissertation presented the validation
of the analytical procedure and its successful application on plasma samples from
healthy volunteers and diseased patients. The assay has shown enough
robustness for use in clinical samples for investigations that relate to these
analytes.
6.2 Exploration of CRC biomarkers
With the stage set for the profiling of FOCM metabolites in CRC versus controls,
we explored CRC blood-based biomarkers using both targeted and untargeted
metabolomics (Roberts et al., 2012). Chapter 3 of this dissertation expands on the
combined power of metabolomics and LCMS which makes it a feasible and novel
approach of exploring for blood-based biomarkers for CRC by profiling the plasma
samples of CRC patients (cases) and normal volunteers (controls). Untargeted
metabolomics measured all endogenous metabolites in the biological samples and
populated the list of metabolites found to be significantly different between the
158

cases and the controls for identification in the human metabolome database. When
this exploratory analysis was performed in the clinical plasma samples from 26
CRC cases and 10 controls, 30 putative metabolites were obtained out of which
20 were linked to the FOCM.
Knowing the key role of FOCM in the aberrant methylation characteristic of CRC
pathogenesis pathway, a targeted metabolomics was used to quantify the FOCM
metabolites using the validated assay developed in Chapter 2. Our aim at this point
was to identify whether specific FOCM metabolites could be used to prediction of
CRC status in humans. The results showed differences in folate proportions
between the cases and controls with an accumulation of reduced folates at the
expense of unmetabolized folic acid in the CRC cases. There was also a sign of
vitamin B12 deficiency in the CRC cases indicated by their median plasma MMA
concentrations being higher than the upper reference limit of 290nM (Fu et al.,
2013). We hypothesize that the oxidative stress conditions existent in the
microenvironment of CRC patients where vitamin B12 deficiency may cause redox
enzymes like MS to be inhibited (Zou and Banerjee, 2005) because MS is in the
oxidized state awaiting re-activation by Methionine Synthase Reductase. This may
explain the plasma pooling of reduced folates like 5-methyltetrahydrofolate
(5MTHF) in the CRC cases.
The study identified several plasma metabolites including 5MTHF, THF, FA, B2,
PA, PL, SAM, SAH and MMA that are altered in CRC and thus may be used as
potential biomarkers for screen for the presence of CRC and even detect the
159

transitioning from polyp to cancer. The group prediction model developed from this
clinical metabolites data yielded a misclassification error rate of 2.8% which may
present a more convenient and reliable screening assay as an alternative to
colonoscopy. The convenience and minimal invasion of blood-based assays
makes them highly needful in the population-based CRC screening.
6.3 The role of FOCM in various colon cell lines
Clinical data has several covariates like folate supplementation and dietary
information that were not adjusted for and thus posed as a great limitation to the
models developed. We decided to corroborate the clinical data by probing whether
the FOCM in colon cells at different stages of CRC development differ significantly.
We selected the colon cells including untransformed, transitional and transformed
along the CRC pathogenesis to address this objective. The FOCM metabolites and
genes expressions were analyzed and compared. There was a general
upregulation in the folate metabolism genes in the transformed cells. This was
consistent with a higher proportion of reduced folates in transformed cells when
compared to non-malignant cells. Lower cellular FA and DHF levels in the
transformed cells can reduce the inhibitory controls required to modulate the
enzymatic metabolism of downstream folates (Bailey and Ayling, 2009).
160

Figure 6.1: Comparison of folate proportions in clinical and in vitro data.

There was consistency in trends of folate proportions observed in both clinical and
in vitro data (Figure 6.1). This suggest that mean folate proportions may be a good
biomarker to classify a sample or a cell as transformed or untransformed. Due to
our uncertainty on how cellular and extracellular metabolite concentrations
compare to each other, we determined in another experiment that there was
generally a high positive correlation between cellular and extracellular metabolite
concentrations.
6.4 Redox homeostasis in transformed colon cell lines
The upregulated DNA methyltransferases in Caco-2 seemed inconsistent with the
lower global methylation (hypomethylation) which is consistent with Caco-2 cells’
chromosomal instability pathway in CRC pathogenesis. To explain, the FOCM
pattern in the transformed cells, one must understand how the various colon cells
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
CRL1459 APC CACO-2 HCT116
FA DHF THF 5MTHF
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Controls Cases
FA DHF THF MTHF
Clinical data In vitro data
161

handle oxidative stress as proposed mechanisms behind their differences in
FOCM.
Untransformed and transformed cells may also differ in their levels of FOCM
metabolites based on their environment. Whereas untransformed cells may have
a homeostatic balance under normoxic stress condition, transformed cells thrive
under oxidative stress conditions (Mosharov et al., 2000). The dissertation shows
how the stressors in the transformed cells environment tend to affect their FOCM-
related metabolites and gene expressions. During oxidative stress in transformed
cells, MS activity is inhibited, resulting in the diversion of homocysteine into the
transsulfuration reaction to produce the antioxidant GSH, providing an important
adaptive response (Scherer et al., 2012; Zou and Banerjee, 2005). Caco-2 which
has a downregulated MTR/MTRR but upregulated CBS may adapt easily by
producing more GSH to buffer the reactive oxidative species in the environment.
However, the story is different for HCT116 cells. The CBS gene is totally
downregulated in HCT116 cells so they may have to continue to upregulate
MTR/MTRR to remove homocysteine, a toxic metabolite which may worsen the
oxidative stress. However, oxidative inhibition of methionine synthase may trigger
resultant epigenetic changes and an up-regulation of the MTR gene. Decreased
methionine synthase activity in turn decreases the level of the methyl donor SAM
favoring methylation inhibition by SAH (Matthews et al., 1998).
162

6.5 Methionine depletion in transformed cells
The cellular response to environmental changes can also be assessed with METH
can the rate limiting substrate since it is the source of the generated HCY. METH
is pivotal as an essential amino acid, which is required for the formation of HCY,
SAM (the primary methyl donor), and as the initiation amino acid for protein
synthesis. The rate of METH depletion can be a great indicator for the direction of
metabolic flux within the cycle. When there are decreased cellular levels of SAM, it
puts important methylation reactions at risk thereby initiating mechanisms to
maintain supplies of methionine and SAM as a priority. The folate co-factors are
directed through the transmethylation cycle at the expense of the cycles that
produce purines and pyrimidines for DNA synthesis.
In the case of HCT116, the oxidative stress environment may lead to a vicious cycle
that accumulates 5MTHF because the MTR is inhibited and the inhibitory feedback
loop on the MTHFR is broken. The METH stores are depleted further to generate
HCY and subsequently, GSH. The down-regulated CBS may not generate enough
GSH, and there may be an increase in homocysteine thiolactone formation (Beatty
and Reed, 1980). Caco-2, on the other hand, has a downregulated MTR/MTRR
and MAT genes. However, the prominence placed on methylation reactions will
allow the production of SAM but the biosynthesis of METH becomes challenging.
This may lead to a depletion of the METH stores.
163

6.6 Regulation of proliferation and apoptosis in colon cell lines
One of the genes that were found upregulated in the highly proliferating cells was
IGF2, which is a fetal growth factor known to undergo loss of imprinting (LOI) in
CRC. This LOI mechanism differs from what occurs in other cancer types which
involves the repression of IGF2 gene due to methylated the noncoding RNA, H19
(Kelemen, 2006; Popat et al., 2004). In CRC, IGF2 LOI may result from a
hypomethylation of the differentially methylated region of the H19 gene in both
alleles leading to overexpression of IGF2 (Cui et al., 2002). H19 was upregulated
in HCT116 p53+/+ and HCT116 p53-/- but downregulated in CRL1790.
The regulation of cell growth requires p53 which maintains the genomic stability,
cellular senescence (Kaffer et al., 2001; Mason et al., 2007) and induces apoptosis
(Shephard, 2011). However, the co-expression of BCL-2 and the oncogene c-MYC
efficiently antagonizes effects of p53 thereby modulating proliferation and
apoptosis, although BCL-2 on its own can affect antagonize the effect of p53 to
induce apoptosis (Chiarugi and Ruggiero, 1995). The effect of BCL-2 and c-MYC
co-expression may explain the high proliferation of HCT116 p53+/+ cells,
irrespective of the overexpression of TP53 in these cells.
6.7 Folic Acid Supplementation and CRC risk
Folic acid supplementation, upon introduction, had great success in reducing the
incidence of neural tube defects [NTDs; (Group, 1991)]. Building on this success,
the US Food and Drug Administration issued a mandate in 1996, requiring
164

fortification of flour with folic acid (140 μg folic acid per 100 g flour) with the aim of
reducing NTDs. Although folic acid is a water-soluble B vitamin, it can easily
accumulate with continual use. It was thus not surprising that FA accumulation
occurred when intakes of FA from fortified foods were more than two times the
level originally predicted. Analysis of CRC incidence pre- and post-
supplementation policy implementation showed a spike which coincided with the
period that FA supplementation was implemented (Mason et al., 2007). More
recent probes into the issue of FA accumulation has revealed the role of
unmetabolized FA in the FA metabolism and how it alters the FOCM through
selective enzyme inhibition (Bailey and Ayling, 2009; Matthews and Daubner,
1982; Sauer et al., 2009). Also, there have been concerns raised on the timing and
the dose of folate supplementation being linked with increased risk of CRC (Kim,
1999, 2003).
In this dissertation, we probed the impact of folates (FA and/or 5MTHF) into growth
media at various concentrations on colon cells at various stages in CRC
development. The FOCM is altered when colon cells are supplemented with
different levels of FA or 5MTHF supplement. The total cellular folates increased
with increasing supplementation concentration, however folate proportions differed
based on the cell type. Although adding 9 µM FA was associated with significant
increased growth rates in both untransformed and transformed cells, the addition
of these folates slowed growth in APC 10.1. The addition of 2.3 µM 5MTHF with 9
µM FA (Combo) was associated with reduced growth in the transitional and
transformed cell lines. In addition, DNA global methylation was found to increase
165

with increasing concentration of folates. The consequence of this finding is still
unclear. Additional interrogation into these findings may lead to some preventive
measures that may reduce the risk of malignant transformation
6.8 Conclusions and Significance
The FOCM has a critical role in the pathogenesis of CRC and it is altered as colon
cells transform from normal to cancer. Whatever genomic changes that underlie
the transformation process are reflected by the phenotypic cellular and
extracellular metabolite levels. These metabolites can be used to phenotype a cell
or person and used to predict whether to classify as normal or cancerous. Also,
the intake of supplements may affect the growth rate of untransformed and
transformed cells. However, in the presence of at least 2.3 µM of dietary folate
(5MTHF) supplement, growth rates of transformed colon cells were slower.
This dissertation can be used to inform the development of blood-based
biomarkers as population screening tools for CRC to reduce CRC-related death. It
can also inform the debate on the food supplementation policy and clinical
management of CRC.
6.9 Future Directions
The project proposed studies that are extensions of the candidate gene pathway-
based study and a genome-wide association study (GWAS). The candidate gene
pathway-based study measured the environmental risks factors associated with
the CRC risk such as nutritional practices, heavy alcohol consumption, cigarette
166

smoking among others (Haggar and Boushey, 2009). The known gene
polymorphisms in subjects can be matched with the FOCM-related metabolites to
build models that can be predictive of CRC cases and controls status. The models
will be indispensable in the battle against CRC-related death through its support
for CRC screening programs and disease management.
There should be a further validation of this work in vitro experiments using 3D colon
organoids which will replicate the characteristics of the cells in the human body.
The results of such 3D experiments with organoids will make the transfers of the
findings to humans more reliable. These organoids can be grown from biopsies
taken from patients and characterized based on the findings of this dissertation for
individualized treatment.
Lastly, there have been recent findings that low-dose aspirin reduces the risk of
CRC. Future studies should employ the metabolomics approach to profile the pro-
inflammatory and anti-inflammatory bioactive lipids and further explore them as
potential biomarkers that can tighten the error margins at predicting or monitoring
CRC progression. A preliminary analysis found increasing levels of cellular
succinic acid, an onco-metabolite with the colon cells at different stages in CRC
development (Figure 6.2). It will be interesting to explore the levels of these organic
acids as biomarkers associated with the risk of CRC.
167

Figure 6.2: Cellular levels of Succinic acid in colon cells at different stages of CRC
development.

.
C o n c e n tra tio n
(n g p e r m g o f c e ll p ro te in )
C R L -1 4 5 9
C R L 1 7 9 0
A P C 1 0 .1
C A C O -2
H C T 1 1 6 p 5 3 + /+
H C T P 5 3 -/-
0
5 0 0 0
1 0 0 0 0
1 5 0 0 0
2 0 0 0 0
2 5 0 0 0
A P C 10 .1
C A C O -2
C R L -1 4 5 9
C R L 1 7 9 0
H C T 1 1 6 p 5 3 + /+
H C T P 5 3 -/-
168

Chapter 6: Bibliography and References
Bailey, S.W., and Ayling, J.E. (2009). 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.
Beatty, P.W., and Reed, D.J. (1980). Involvement of the cystathionine pathway in
the biosynthesis of glutathione by isolated rat hepatocytes. Archives of
Biochemistry and Biophysics 204, 80-87.
Chiarugi, V., and Ruggiero, M. (1995). Role of three cancer" master genes" p53,
bcl2 and c-myc on the apoptotic process. Tumori 82, 205-209.
Crider, K.S., Yang, T.P., Berry, R.J., and Bailey, L.B. (2012). Folate and DNA
methylation: a review of molecular mechanisms and the evidence for folate’s
role. Advances in Nutrition: An International Review Journal 3, 21-38.
Cui, H., Onyango, P., Brandenburg, S., Wu, Y., Hsieh, C.-L., and Feinberg, A.P.
(2002). Loss of imprinting in colorectal cancer linked to hypomethylation of H19
and IGF2. Cancer research 62, 6442-6446.
DeVries, J.W., Rader, J.I., Keagy, P.M., Hudson, C.A., Angyal, G., Arcot, J.,
Castelli, M., Doreanu, N., Hudson, C., Lawrence, P., et al. (2005). Microbiological
assay-trienzyme procedure for total folates in cereals and cereal foods:
collaborative study. J AOAC Int 88, 5-15.
169

Ferlay, J., Soerjomataram, I., Dikshit, R., Eser, S., Mathers, C., Rebelo, M.,
Parkin, D.M., Forman, D., and Bray, F. (2015). Cancer incidence and mortality
worldwide: sources, methods and major patterns in GLOBOCAN 2012.
International journal of cancer 136, E359-E386.
Fu, X., Xu, Y.-K., Chan, P., and Pattengale, P.K. (2013). Simple, fast, and
simultaneous detection of plasma total homocysteine, methylmalonic acid,
methionine, and 2-methylcitric acid using liquid chromatography and mass
spectrometry (LC/MS/MS). In JIMD Reports-Volume 10 (Springer), pp. 69-78.
Group, M.V.S.R. (1991). Prevention of neural tube defects: results of the Medical
Research Council Vitamin Study. The lancet 338, 131-137.
Group, U.C.S.W. (2016). United States cancer statistics: 1999–2013 Incidence
and Mortality Web-based Report. Atlanta: US Department of Health and Human
Services, Centers for Disease Control and Prevention and National Cancer
Institute.
Haggar, F.A., and Boushey, R.P. (2009). Colorectal cancer epidemiology:
incidence, mortality, survival, and risk factors. Clin Colon Rectal Surg 22, 191-
197.
Kaffer, C.R., Grinberg, A., and Pfeifer, K. (2001). Regulatory Mechanisms at the
MouseIgf2/H19 Locus. Molecular and cellular biology 21, 8189-8196.
170

Kelemen, L.E. (2006). The role of folate receptor α in cancer development,
progression and treatment: cause, consequence or innocent bystander?
International journal of cancer 119, 243-250.
Kim, Y.I. (1999). Folate and carcinogenesis: evidence, mechanisms, and
implications. J Nutr Biochem 10, 66-88.
Kim, Y.I. (2003). Role of folate in colon cancer development and progression. J
Nutr 133, 3731S-3739S.
Mason, J.B., Dickstein, A., Jacques, P.F., Haggarty, P., Selhub, J., Dallal, G.,
and Rosenberg, I.H. (2007). A temporal association between folic acid
fortification and an increase in colorectal cancer rates may be illuminating
important biological principles: a hypothesis. Cancer Epidemiology and
Prevention Biomarkers 16, 1325-1329.
Matthews, R.G., and Daubner, S.C. (1982). Modulation of
methylenetetrahydrofolate reductase activity by S-adenosylmethionine and by
dihydrofolate and its polyglutamate analogues. Adv Enzyme Regul 20, 123-131.
Matthews, R.G., Sheppard, C., and Goulding, C. (1998).
Methylenetetrahydrofolate reductase and methionine synthase: biochemistry and
molecular biology. European journal of pediatrics 157, S54-S59.
Mosharov, E., Cranford, M.R., and Banerjee, R. (2000). The quantitatively
important relationship between homocysteine metabolism and glutathione
171

synthesis by the transsulfuration pathway and its regulation by redox changes.
Biochemistry 39, 13005-13011.
Popat, S., Matakidou, A., and Houlston, R.S. (2004). Thymidylate synthase
expression and prognosis in colorectal cancer: a systematic review and meta-
analysis. Journal of Clinical Oncology 22, 529-536.
Roberts, L.D., Souza, A.L., Gerszten, R.E., and Clish, C.B. (2012). Targeted
metabolomics. Current protocols in molecular biology, 30.32. 31-30.32. 24.
Sauer, J., Mason, J.B., and Choi, S.-W. (2009). Too much folate: a risk factor for
cancer and cardiovascular disease? Curr Opin Clin Nutr Metab Care 12, 30-36.
Scherer, E., Schmitz, F., Vuaden, F., Savio, L., Ferreira, A., Tasca, R., Casali, E.,
Bogo, M., Bonan, C., and Wyse, A. (2012). Mild hyperhomocysteinemia alters
extracellular adenine metabolism in rat brain. Neuroscience 223, 28-34.
Shane, B., Tamura, T., and Stokstad, E.L. (1980). Folate assay: a comparison of
radioassay and microbiological methods. Clin Chim Acta 100, 13-19.
Shephard, R.J. (2011). Nutrition, epigenetic mechanisms and human disease.
Applied Physiology, Nutrition, and Metabolism 36, 311-313.
Siegel, R. L., Miller, K. D., & Jemal, A. (2016). Cancer statistics, 2016. CA: a cancer
journal for clinicians, 66(1), 7-30.
172

van der Greef, J., Davidov, E., Verheij, E., Vogels, J., van der Heijden, R.,
Adourian, A.S., Oresic, M., W., M.E., and Naylor, S. (2003). The Role of
Metabolomics in Systems Biology. In Metabolic Profiling: Its Role in Biomarker
Discovery and Gene Function Analysis, G.G.H.R. Goodacre, ed.
(Boston/Dordrecht/London: Kluwer Academic Publishers ), pp. 171-198.
Zou, C.-G., and Banerjee, R. (2005). Homocysteine and redox signaling.
Antioxidants & redox signaling 7, 547-559.

173

Complete Bibliography and References
Ahmed, D., Eide, P., Eilertsen, I., Danielsen, S., Eknaes, M., Hektoen, M., Lind,
G., and Lothe, R. (2013). Epigenetic and genetic features of 24 colon cancer cell
lines. Oncogenesis 2, e71.
Arnold, C.N., Goel, A., Blum, H.E., and Boland, C.R. (2005). Molecular
pathogenesis of colorectal cancer: implications for molecular diagnosis. Cancer
104, 2035-2047.
Bailey, L.B., Rampersaud, G.C., and Kauwell, G.P.A. (2003). Folic acid
supplements and fortification affect the risk for neural tube defects, vascular
disease and cancer: evolving science. J Nutr 133, 1968.
Bailey, S.W., and Ayling, J.E. (2009). 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.
Banerjee, D., Mayer-Kuckuk, P., Capiaux, G., Budak-Alpdogan, T., Gorlick, R., and
Bertino, J.R. (2002). Novel aspects of resistance to drugs targeted to dihydrofolate
reductase and thymidylate synthase. Biochimica et Biophysica Acta (BBA)-
Molecular Basis of Disease 1587, 164-173.
Beatty, P.W., and Reed, D.J. (1980). Involvement of the cystathionine pathway in
the biosynthesis of glutathione by isolated rat hepatocytes. Archives of
Biochemistry and Biophysics 204, 80-87.
174

Benito, E., Stiggelbout, A., Bosch, F., Obrador, A., Kaldor, J., Mulet, M., and
Munoz, N. (1991). Nutritional factors in colorectal cancer risk: a case ‐control study
in Majorca. International journal of cancer 49, 161-167.
Berry, R.J., Li, Z., Erickson, J.D., Li, S., Moore, C.A., Wang, H., Mulinare, J., Zhao,
P., Wong, L.-Y.C., and Gindler, J. (1999). Prevention of neural-tube defects with
folic acid in China. New England journal of medicine 341, 1485-1490.
Bird, A. (2002). DNA methylation patterns and epigenetic memory. Genes &
development 16, 6-21.
Bird, A. (2007). Perceptions of epigenetics. Nature 447, 396-398.
Bjørk, M.K., Nielsen, M.K., Markussen, L.Ø., Klinke, H.B., and Linnet, K. (2010).
Determination of 19 drugs of abuse and metabolites in whole blood by high-
performance liquid chromatography–tandem mass spectrometry. Analytical and
bioanalytical chemistry 396, 2393-2401.
Blount, B.C., Mack, M.M., Wehr, C.M., MacGregor, J.T., Hiatt, R.A., Wang, G.,
Wickramasinghe, S.N., Everson, R.B., and Ames, B.N. (1997). 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.
175

Boushey, C.J., Beresford, S.A., Omenn, G.S., and Motulsky, A.G. (1995). A
quantitative assessment of plasma homocysteine as a risk factor for vascular
disease: probable benefits of increasing folic acid intakes. Jama 274, 1049-1057.
Boyapati, S.M., Bostick, R.M., McGlynn, K.A., Fina, M.F., Roufail, W.M., Geisinger,
K.R., Hebert, J.R., Coker, A., and Wargovich, M. (2004). Folate intake, MTHFR
C677T polymorphism, alcohol consumption, and risk for sporadic colorectal
adenoma (United States). Cancer Causes & Control 15, 493-501.
Burt, R.W., Barthel, J.S., Dunn, K.B., David, D.S., Drelichman, E., Ford, J.M.,
Giardiello, F.M., Gruber, S.B., Halverson, A.L., and Hamilton, S.R. (2010).
Colorectal cancer screening. Journal of the National Comprehensive Cancer
Network 8, 8-61.
Chen, J., Giovannucci, E., Kelsey, K., Rimm, E.B., Stampfer, M.J., Colditz, G.A.,
Spiegelman, D., Willett, W.C., and Hunter, D.J. (1996). A
methylenetetrahydrofolate reductase polymorphism and the risk of colorectal
cancer. Cancer Research 56, 4862-4864.
Chen, Z., Chakraborty, S., and Banerjee, R. (1995). Demonstration that
mammalian methionine synthases are predominantly cobalamin-loaded. Journal
of Biological Chemistry 270, 19246-19249.
Chiarugi, V., and Ruggiero, M. (1995). Role of three cancer" master genes" p53,
bcl2 and c-myc on the apoptotic process. Tumori 82, 205-209.
176

Cho, E., Zhang, X., Townsend, M.K., Selhub, J., Paul, L., Rosner, B., Fuchs, C.S.,
Willett, W.C., and Giovannucci, E.L. (2015). Unmetabolized Folic Acid in
Prediagnostic Plasma and the Risk for Colorectal Cancer. Journal of the National
Cancer Institute 107, djv260.
Choi, S.-W., and Mason, J.B. (2002). Folate status: effects on pathways of
colorectal carcinogenesis. The Journal of nutrition 132, 2413S-2418S.
Collaboration, H.S. (2002). Homocysteine and risk of ischemic heart disease and
stroke: a meta-analysis. Jama 288, 2015-2022.
Crider, K.S., Yang, T.P., Berry, R.J., and Bailey, L.B. (2012). Folate and DNA
methylation: a review of molecular mechanisms and the evidence for folate’s role.
Advances in Nutrition: An International Review Journal 3, 21-38.
Cui, H., Onyango, P., Brandenburg, S., Wu, Y., Hsieh, C.-L., and Feinberg, A.P.
(2002). Loss of imprinting in colorectal cancer linked to hypomethylation of H19
and IGF2. Cancer research 62, 6442-6446.
de Vogel, S., Schneede, J., Ueland, P.M., Vollset, S.E., Meyer, K., Fredriksen, Å.,
Midttun, Ø., Bjørge, T., Kampman, E., and Bretthauer, M. (2011). Biomarkers
related to one-carbon metabolism as potential risk factors for distal colorectal
adenomas. Cancer Epidemiology Biomarkers & Prevention 20, 1726-1735.
DeVries, J.W., Rader, J.I., Keagy, P.M., Hudson, C.A., Angyal, G., Arcot, J.,
Castelli, M., Doreanu, N., Hudson, C., Lawrence, P., et al. (2005). Microbiological
177

assay-trienzyme procedure for total folates in cereals and cereal foods:
collaborative study. J AOAC Int 88, 5-15.
Duthie, S.J. (1999). Folic acid deficiency and cancer: mechanisms of DNA
instability. Br Med Bull 55, 578-592.
Duthie, S.J., and Hawdon, A. (1998). DNA instability (strand breakage, uracil
misincorporation, and defective repair) is increased by folic acid depletion in
human lymphocytes in vitro. FASEB J 12, 1491-1497.
Fazili, Z., Pfeiffer, C.M., Zhang, M., Jain, R.B., and Koontz, D. (2008). Influence of
5, 10-methylenetetrahydrofolate reductase polymorphism on whole-blood folate
concentrations measured by LC-MS/MS, microbiologic assay, and bio-rad
radioassay. Clinical chemistry 54, 197-201.
Fearon, E.R., and Vogelstein, B. (1990). A genetic model for colorectal
tumorigenesis. Cell 61, 759-767.
Ferlay, J., Soerjomataram, I., Dikshit, R., Eser, S., Mathers, C., Rebelo, M., Parkin,
D.M., Forman, D., and Bray, F. (2015). Cancer incidence and mortality worldwide:
sources, methods and major patterns in GLOBOCAN 2012. International journal
of cancer 136, E359-E386.
Fiehn, O. (2002). Metabolomics–the link between genotypes and phenotypes.
Plant molecular biology 48, 155-171.
178

Finkelstein, J.D. (2007). Metabolic regulatory properties of S-adenosylmethionine
and S-adenosylhomocysteine. Clinical Chemical Laboratory Medicine 45, 1694-
1699.
Frison, L., and Pocock, S.J. (1992). Repeated measures in clinical trials: analysis
using mean summary statistics and its implications for design. Statistics in
medicine 11, 1685-1704.
Frosst, P., Blom, H.J., Milos, R., Goyette, P., Sheppard, C.A., Matthews, R.G.,
Boers, G.J., den Heijer, M., Kluijtmans, L.A., and Van den Heuvel, L. (1995). A
candidate genetic risk factor for vascular disease: a common mutation in
methylenetetrahydrofolate reductase.
Fu, X., Xu, Y.-K., Chan, P., and Pattengale, P.K. (2013). Simple, fast, and
simultaneous detection of plasma total homocysteine, methylmalonic acid,
methionine, and 2-methylcitric acid using liquid chromatography and mass
spectrometry (LC/MS/MS). In JIMD Reports-Volume 10 (Springer), pp. 69-78.
Galluzzi, L., Vacchelli, E., Michels, J., Garcia, P., Kepp, O., Senovilla, L., Vitale, I.,
and Kroemer, G. (2013). Effects of vitamin B6 metabolism on oncogenesis, tumor
progression and therapeutic responses. Oncogene 32, 4995-5004.
George, L., Mills, J.L., Johansson, A.L., Nordmark, A., Olander, B., Granath, F.,
and Cnattingius, S. (2002). Plasma folate levels and risk of spontaneous abortion.
Jama 288, 1867-1873.
179

Giovannucci, E., Ascherio, A., Rimm, E.B., Stampfer, M.J., Colditz, G.A., and
Willett, W.C. (1995). Intake of carotenoids and retino in relation to risk of prostate
cancer. Journal of the national cancer institute 87, 1767-1776.
Giovannucci, E., Stampfer, M.J., Colditz, G.A., Hunter, D.J., Fuchs, C., Rosner,
B.A., Speizer, F.E., and Willett, W.C. (1998). Multivitamin use, folate, and colon
cancer in women in the Nurses' Health Study. Annals of Internal Medicine 129,
517-524.
Giovannucci, E., Stampfer, M.J., Colditz, G.A., Rimm, E.B., Trichopoulos, D.,
Rosner, B.A., Speizer, F.E., and Willett, W.C. (1993). Folate, methionine, and
alcohol intake and risk of colorectal adenoma. Journal of the National Cancer
Institute 85, 875-883.
Gori, A.M., Sofi, F., Corsi, A.M., Gazzini, A., Sestini, I., Lauretani, F., Bandinelli,
S., Gensini, G.F., Ferrucci, L., and Abbate, R. (2006). Predictors of vitamin B6 and
folate concentrations in older persons: the InCHIANTI study. Clinical chemistry 52,
1318-1324.
Group, M.V.S.R. (1991). Prevention of neural tube defects: results of the Medical
Research Council Vitamin Study. The lancet 338, 131-137.
Group, U.C.S.W. (2016). United States cancer statistics: 1999–2013 Incidence
and Mortality Web-based Report. Atlanta: US Department of Health and Human
Services, Centers for Disease Control and Prevention and National Cancer
Institute.
180

Gupta, A.K., Brenner, D.E., and Turgeon, D.K. (2008). Early Detection of Colon
Cancer. Molecular diagnosis & therapy 12, 77-85.
Haggar, F.A., and Boushey, R.P. (2009). Colorectal cancer epidemiology:
incidence, mortality, survival, and risk factors. Clin Colon Rectal Surg 22, 191-197.
Hanley, M.P., and Rosenberg, D.W. (2015). One-Carbon Metabolism and
Colorectal Cancer: Potential Mechanisms of Chemoprevention. Current
pharmacology reports 1, 197-205.
Herbert, V., and Zalusky, R. (1962). Interrelations of vitamin B12 and folic acid
metabolism: folic acid clearance studies. Journal of Clinical Investigation 41, 1263.
Hoogerbrugge, R., Willig, S.J., and Kistemaker, P.G. (1983). Discriminant analysis
by double stage principal component analysis. Analytical chemistry 55, 1710-1712.
Hustad, S., Ueland, P.M., and Schneede, J. (1999). Quantification of riboflavin,
flavin mononucleotide, and flavin adenine dinucleotide in human plasma by
capillary electrophoresis and laser-induced fluorescence detection. Clinical
chemistry 45, 862-868.
Jagerdeo, E., Montgomery, M.A., LeBeau, M.A., and Sibum, M. (2008). An
automated SPE/LC/MS/MS method for the analysis of cocaine and metabolites in
whole blood. Journal of Chromatography B 874, 15-20.
181

Jarabak, J., and Bachur, N.R. (1971). A soluble dihydrofolate reductase from
human placenta: purification and properties. Archives of biochemistry and
biophysics 142, 417-425.
Jhaveri, M.S., Wagner, C., and Trepel, J.B. (2001). Impact of extracellular folate
levels on global gene expression. Mol Pharmacol 60, 1288-1295.
Kaffer, C.R., Grinberg, A., and Pfeifer, K. (2001). Regulatory Mechanisms at the
MouseIgf2/H19 Locus. Molecular and cellular biology 21, 8189-8196.
Keku, T., Millikan, R., Worley, K., Winkel, S., Eaton, A., Biscocho, L., Martin, C.,
and Sandler, R. (2002). 5, 10-Methylenetetrahydrofolate reductase codon 677 and
1298 polymorphisms and colon cancer in African Americans and whites. Cancer
Epidemiology Biomarkers & Prevention 11, 1611-1621.
Kelemen, L.E. (2006). The role of folate receptor α in cancer development,
progression and treatment: cause, consequence or innocent bystander?
International journal of cancer 119, 243-250.
Kim, J.W., Park, H.M., Choi, Y.-K., Chong, S.Y., Oh, D., and Kim, N.K. (2011).
Polymorphisms in genes involved in folate metabolism and plasma DNA
methylation in colorectal cancer patients. Oncol Rep 25, 167-172.
Kim, K.-c., Friso, S., and Choi, S.-W. (2009). DNA methylation, an epigenetic
mechanism connecting folate to healthy embryonic development and aging. The
Journal of nutritional biochemistry 20, 917-926.
182

Kim, Y.-I. (2004). Folate and DNA methylation: a mechanistic link between folate
deficiency and colorectal cancer? Cancer Epidemiology Biomarkers & Prevention
13, 511-519.
Kim, Y.I. (1999). Folate and carcinogenesis: evidence, mechanisms, and
implications. J Nutr Biochem 10, 66-88.
Kim, Y.I. (2003). Role of folate in colon cancer development and progression. J
Nutr 133, 3731S-3739S.
Kind, T., Tolstikov, V., Fiehn, O., and Weiss, R.H. (2007). A comprehensive urinary
metabolomic approach for identifying kidney cancer. Analytical biochemistry 363,
185-195.
Kořínek, M., Šístek, V., Mládková, J., Mikeš, P., Jiráček, J., and Selicharová, I.
(2013). Quantification of homocysteine ‐related metabolites and the role of
betaine–homocysteine S ‐methyltransferase in HepG2 cells. Biomedical
Chromatography 27, 111-121.
Kruman, I.I., Kumaravel, T.S., Lohani, A., Pedersen, W.A., Cutler, R.G., Kruman,
Y., Haughey, N., Lee, J., Evans, M., and Mattson, M.P. (2002). Folic acid
deficiency and homocysteine impair DNA repair in hippocampal neurons and
sensitize them to amyloid toxicity in experimental models of Alzheimer's disease.
J Neurosci 22, 1752-1762.
183

Kurzawski, G., Suchy, J., Debniak, T., Kladny, J., and Lubinski, J. (2004).
Importance of microsatellite instability (MSI) in colorectal cancer: MSI as a
diagnostic tool. Ann Oncol 15, 283-284.
Lachenbruch, P.A., and Mickey, M.R. (1968). Estimation of error rates in
discriminant analysis. Technometrics 10, 1-11.
Lamprecht, S.A., and Lipkin, M. (2003). Chemoprevention of colon cancer by
calcium, vitamin D and folate: molecular mechanisms. Nature Reviews Cancer 3,
601-614.
Lao, V.V., & Grady, W. M. (2011). Epigenetics and Colorectal Cancer. Nature
Reviews Gastroenterology and Hepatology 8, 686- 700.
Larsson, S.C., Giovannucci, E., and Wolk, A. (2006). Folate intake, MTHFR
polymorphisms, and risk of esophageal, gastric, and pancreatic cancer: a meta-
analysis. Gastroenterology 131, 1271-1283.
Lentner, C. (1981). Geigy Scientific tables. Vol. 1, Units of measurement, body
fluids, composition of the body, nutrition (Ciba-Geigy Basle).
Levine, A.J., Figueiredo, J.C., Lee, W., Conti, D.V., Kennedy, K., Duggan, D.J.,
Poynter, J.N., Campbell, P.T., Newcomb, P., and Martinez, M.E. (2010). A
candidate gene study of folate-associated one carbon metabolism genes and
colorectal cancer risk. Cancer Epidemiology Biomarkers & Prevention 19, 1812-
1821.
184

Li, M., Wang, B., Zhang, M., Rantalainen, M., Wang, S., Zhou, H., Zhang, Y., Shen,
J., Pang, X., and Zhang, M. (2008). Symbiotic gut microbes modulate human
metabolic phenotypes. Proceedings of the National Academy of Sciences 105,
2117-2122.
Li, X., and Franke, A.A. (2011). High-throughput and cost-effective global DNA
methylation assay by liquid chromatography-mass spectrometry. Analytica
chimica acta 703, 58-63.
Liu, J., Schmitz, J.C., Lin, X., Tai, N., Yan, W., Farrell, M., Bailly, M., Chen, T.-m.,
and Chu, E. (2002). Thymidylate synthase as a translational regulator of cellular
gene expression. Biochimica et Biophysica Acta (BBA)-Molecular Basis of Disease
1587, 174-182.
Livak, K.J., and Schmittgen, T.D. (2001). Analysis of relative gene expression data
using real-time quantitative PCR and the 2− ΔΔCT method. methods 25, 402-408.
Ma, J., Stampfer, M.J., Giovannucci, E., Artigas, C., Hunter, D.J., Fuchs, C.,
Willett, W.C., Selhub, J., Hennekens, C.H., and Rozen, R. (1997).
Methylenetetrahydrofolate reductase polymorphism, dietary interactions, and risk
of colorectal cancer. Cancer research 57, 1098-1102.
Madsen, R., Lundstedt, T., and Trygg, J. (2010). Chemometrics in metabolomics—
a review in human disease diagnosis. Analytica Chimica Acta 659, 23-33.
185

Mason, J.B., Dickstein, A., Jacques, P.F., Haggarty, P., Selhub, J., Dallal, G., and
Rosenberg, I.H. (2007). A temporal association between folic acid fortification and
an increase in colorectal cancer rates may be illuminating important biological
principles: a hypothesis. Cancer Epidemiology and Prevention Biomarkers 16,
1325-1329.
Matthews, R.G., and Daubner, S.C. (1982). Modulation of
methylenetetrahydrofolate reductase activity by S-adenosylmethionine and by
dihydrofolate and its polyglutamate analogues. Adv Enzyme Regul 20, 123-131.
Matthews, R.G., Sheppard, C., and Goulding, C. (1998).
Methylenetetrahydrofolate reductase and methionine synthase: biochemistry and
molecular biology. European journal of pediatrics 157, S54-S59.
Maulik, N., and Maulik, G. (2010). Nutrition, epigenetic mechanisms, and human
disease (CRC Press).
Midttun, Ø., Hustad, S., Solheim, E., Schneede, J., and Ueland, P.M. (2005).
Multianalyte quantification of vitamin B6 and B2 species in the nanomolar range in
human plasma by liquid chromatography–tandem mass spectrometry. Clinical
Chemistry 51, 1206-1216.
Millian, N.S., and Garrow, T.A. (1998). Human betaine–homocysteine
methyltransferase is a zinc metalloenzyme. Archives of biochemistry and
biophysics 356, 93-98.
186

Morales, D.R., and Greenberg, D.M. (1964). Purification and properties of
dihydrofolate reductase of sheep liver. Biochimica et Biophysica Acta (BBA)-
Specialized Section on Enzymological Subjects 85, 360-376.
Mosharov, E., Cranford, M.R., and Banerjee, R. (2000). The quantitatively
important relationship between homocysteine metabolism and glutathione
synthesis by the transsulfuration pathway and its regulation by redox changes.
Biochemistry 39, 13005-13011.
Nicholson, J.K., Lindon, J.C., and Holmes, E. (1999). 'Metabonomics':
understanding the metabolic responses of living systems to pathophysiological
stimuli via multivariate statistical analysis of biological NMR spectroscopic data.
Xenobiotica 29, 1181-1189.
O’Connell, J.B., Maggard, M.A., and Ko, C.Y. (2004). Colon cancer survival rates
with the new American Joint Committee on Cancer sixth edition staging. Journal
of the National Cancer Institute 96, 1420-1425.
Ohtsuka, M., Ling, H., Ivan, C., Pichler, M., Matsushita, D., Goblirsch, M.,
Stiegelbauer, V., Shigeyasu, K., Zhang, X., and Chen, M. (2016). H19 Noncoding
RNA, an Independent Prognostic Factor, Regulates Essential Rb-E2F and CDK8-
β-Catenin Signaling in Colorectal Cancer. EBioMedicine 13, 113-124.
Peters, F.T., and Maurer, H.H. (2007). Systematic comparison of bias and
precision data obtained with multiple-point and one-point calibration in six validated
187

multianalyte assays for quantification of drugs in human plasma. Analytical
chemistry 79, 4967-4976.
Pham, B., Cranney, A., Boers, M., Verhoeven, A., Wells, G., and Tugwell, P.
(1999). Validity of area-under-the-curve analysis to summarize effect in
rheumatoid arthritis clinical trials. The Journal of rheumatology 26, 712-716.
Pignone , M., and Sox, H.C. (2008). Screening for colorectal cancer: US
Preventive Services Task Force recommendation statement. Annals of Internal
Medicine 149.
Popat, S., Matakidou, A., and Houlston, R.S. (2004). Thymidylate synthase
expression and prognosis in colorectal cancer: a systematic review and meta-
analysis. Journal of Clinical Oncology 22, 529-536.
Pufulete, M., Al-Ghnaniem, R., Leather, A.J., Appleby, P., Gout, S., Terry, C.,
Emery, P.W., and Sanders, T.A. (2003). Folate status, genomic DNA
hypomethylation, and risk of colorectal adenoma and cancer: a case control study.
Gastroenterology 124, 1240-1248.
Pufulete, M., Al-Ghnaniem, R., Rennie, J., Appleby, P., Harris, N., Gout, S., Emery,
P., and Sanders, T. (2005). Influence of folate status on genomic DNA methylation
in colonic mucosa of subjects without colorectal adenoma or cancer. Br J Cancer
92, 838-842.
188

Qian, W., Parmar, M., Sambrook, R., Fayers, P., Girling, D., and Stephens, R.
(2000). Analysis of messy longitudinal data from a randomized clinical trial.
Statistics in medicine 19, 2657-2674.
Qin, X., Cui, Y., Shen, L., Sun, N., Zhang, Y., Li, J., Xu, X., Wang, B., Xu, X., and
Huo, Y. (2013). Folic acid supplementation and cancer risk: A meta ‐analysis of
randomized controlled trials. International Journal of Cancer 133, 1033-1041.
Ray, J.G. (1998). Meta-analysis of hyperhomocysteinemia as a risk factor for
venous thromboembolic disease. Arch Intern Med 158, 2101-2106.
Rim, S.H., Joseph, D.A., Steele, C.B., Thompson, T.D., and Seeff, L.C. (2011).
Colorectal cancer screening - United States, 2002, 2004, 2006, and 2008. MMWR
Surveill Summ 60 Suppl, 42-46.
Roberts, L.D., Souza, A.L., Gerszten, R.E., and Clish, C.B. (2012). Targeted
metabolomics. Current protocols in molecular biology, 30.32. 31-30.32. 24.
Sauer, J., Mason, J.B., and Choi, S.-W. (2009). Too much folate: a risk factor for
cancer and cardiovascular disease? Curr Opin Clin Nutr Metab Care 12, 30-36.
Scherer, E., Schmitz, F., Vuaden, F., Savio, L., Ferreira, A., Tasca, R., Casali, E.,
Bogo, M., Bonan, C., and Wyse, A. (2012). Mild hyperhomocysteinemia alters
extracellular adenine metabolism in rat brain. Neuroscience 223, 28-34.
Seifeldin, R., and Hantsch, J.J. (1999). The economic burden associated with
colon cancer in the United States. Clin Ther 21, 1370-1379.
189

Selhub, J., and Miller, J.W. (1992). The pathogenesis of homocysteinemia:
interruption of the coordinate regulation by S-adenosylmethionine of the
remethylation and transsulfuration of homocysteine. The American journal of
clinical nutrition 55, 131-138.
Seshadri, S., Beiser, A., Selhub, J., Jacques, P.F., Rosenberg, I.H., D'Agostino,
R.B., Wilson, P.W.F., and Wolf, P.A. (2002). Plasma homocysteine as a risk factor
for dementia and Alzheimer's disease. N Engl J Med 346, 476-483.
Shane, B. (2010). Folate chemistry and metabolism. In Folate in Health and
Disease, L.B. Bailey, ed. (New York, NY, USA: CRC press), pp. 2-3.
Shane, B. (2011). Folate status assessment history: implications for measurement
of biomarkers in NHANES. The American journal of clinical nutrition 94, 337S-
342S.
Shane, B., Tamura, T., and Stokstad, E.L. (1980). Folate assay: a comparison of
radioassay and microbiological methods. Clin Chim Acta 100, 13-19.
Shephard, R.J. (2011). Nutrition, epigenetic mechanisms and human disease.
Applied Physiology, Nutrition, and Metabolism 36, 311-313.
Siegel, R.L., Miller, K.D., and Jemal, A. (2016). Cancer statistics, 2016. CA: a
cancer journal for clinicians 66, 7-30.
190

Singh, S., Madzelan, P., and Banerjee, R. (2007). Properties of an unusual heme
cofactor in PLP-dependent cystathionine β-synthase. Natural product reports 24,
631-639.
Stevens, V.L., McCullough, M.L., Sun, J., Jacobs, E.J., Campbell, P.T., and
Gapstur, S.M. (2011). High levels of folate from supplements and fortification are
not associated with increased risk of colorectal cancer. Gastroenterology 141, 98-
105. e101.
Taoka, S., Ohja, S., Shan, X., Kruger, W.D., and Banerjee, R. (1998). Evidence
for heme-mediated redox regulation of human cystathionine β-synthase activity.
Journal of Biological Chemistry 273, 25179-25184.
Tariq, K., and Ghias, K. (2016). Colorectal cancer carcinogenesis: a review of
mechanisms. Cancer biology & medicine 13, 120.
Terry, P., Jain, M., Miller, A.B., Howe, G.R., and Rohan, T.E. (2002). Dietary intake
of folic acid and colorectal cancer risk in a cohort of women. International journal
of cancer 97, 864-867.
Troen, A.M., Mitchell, B., Sorensen, B., Wener, M.H., Johnston, A., Wood, B.,
Selhub, J., McTiernan, A., Yasui, Y., Oral, E., et al. (2006). Unmetabolized folic
acid in plasma is associated with reduced natural killer cell cytotoxicity among
postmenopausal women. J Nutr 136, 189-194.
191

Ulrich, C.M., and Potter, J.D. (2006). Folate supplementation: too much of a good
thing? Cancer Epidemiology Biomarkers & Prevention 15, 189-193.
Ulrich, C.M., and Potter, J.D. (2007). Folate and cancer—timing is everything.
Jama 297, 2408-2409.
van der Greef, J., Davidov, E., Verheij, E., Vogels, J., van der Heijden, R.,
Adourian, A.S., Oresic, M., W., M.E., and Naylor, S. (2003). The Role of
Metabolomics in Systems Biology. In Metabolic Profiling: Its Role in Biomarker
Discovery and Gene Function Analysis, G.G.H.R. Goodacre, ed.
(Boston/Dordrecht/London: Kluwer Academic Publishers ), pp. 171-198.
Vernon, S.W. (1997). Participation in colorectal cancer screening: a review. J Natl
Cancer Inst 89, 1406-1422.
Vollset, S.E., Refsum, H., Irgens, L.M., Emblem, B.M., Tverdal, A., Gjessing, H.K.,
Monsen, A.L., and Ueland, P.M. (2000). Plasma total homocysteine, pregnancy
complications, and adverse pregnancy outcomes: the Hordaland Homocysteine
study. Am J Clin Nutr 71, 962-968.
Wallace, K., Grau, M., Levine, A., Shen, L., Hamdan, R., Chen, X., Gui, J., Haile,
R., Barry, E., Ahnen, D., et al. (2010). Association between Folate Levels and CpG
Island Hypermethylation in Normal Colorectal Mucosa Cancer Prev Res (Phila) 3,
1552-1564.
192

Waly, M., Olteanu, H., Banerjee, R., Choi, S., Mason, J., Parker, B., Sukumar, S.,
Shim, S., Sharma, A., and Benzecry, J. (2004). Activation of methionine synthase
by insulin-like growth factor-1 and dopamine: a target for neurodevelopmental
toxins and thimerosal. Molecular psychiatry 9, 358-370.
Wilson, A., Leclerc, D., Rosenblatt, D., and Gravel, R. (1999). Molecular basis for
methionine synthase reductase deficiency in patients belonging to the cblE
complementation group of disorders in folate/cobalamin metabolism. Human
molecular genetics 8, 2009-2016.
Winslow, T. (2005). Colon Cancer Stages (Cephalic Vein).
Xia, W., Hilgenbrink, A.R., Matteson, E.L., Lockwood, M.B., Cheng, J.-X., and Low,
P.S. (2009). A functional folate receptor is induced during macrophage activation
and can be used to target drugs to activated macrophages. Blood 113, 438-446.
Zempleni, J., and Kübler, W. (1995). Metabolism of vitamin B 6 by human kidney.
Nutrition Research 15, 187-192.
Zhao, R., and Casson, A.G. (2011). Interactions between P53 and IGF2 in human
esophageal adenocarcinoma tissues and cell lines (AACR).
Zheng, X.-H., Jiang, L.-Y., Zhao, L.-T., Zhang, Q.-Y., and Ding, L. (2015).
Simultaneous quantitation of folic acid and 5-methyltetrahydrofolic acid in human
plasma by HPLC–MS/MS and its application to a pharmacokinetic study. Journal
of Pharmaceutical Analysis 5, 269-275.
193

Zou, C.-G., and Banerjee, R. (2005). Homocysteine and redox signaling.
Antioxidants & redox signaling 7, 547-559.

Abstract (if available)

Abstract Background: CRC ranks as the third leading cause of cancer-related death in the industrialized world, responsible for about 50,000 deaths in the United States annually. Aberrant methylation patterns have been implicated in the pathogenesis of CRC and thus make it necessary to probe the folate one-carbon metabolism (FOCM) to explore for any association with CRC risk. ❧ Aims: The main goals of this study were to: 1) Develop a validated metabolomics-based LCMS assay for the FOCM metabolites

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Factors that may impact drug disposition and metabolism

PDF

Renin-angiotensin system modulation for the prevention and treatment of metabolic dysfunction

PDF

Laser induced modulation of ocular renin-angiotensin system for treatment of dry age-related macular degeneration

PDF

Alterations in renin-angiotensin system signaling in SARS-CoV2 infection

PDF

Investigation into the role of the intestinal microbiome on ethanol consumption behaviors

PDF

The preclinical evaluation of moxidectin as a platform for drug development for alcohol use disorder

PDF

Exosomal inhibitory activity of clinically approved tertiary amines for the treatment of Kaposi’s sarcoma tumorigenesis

PDF

Effects of glucose modulation on adipocytes maturation and metabolism

PDF

Exploration and expansion of novel therapeutic strategies targeting brain metabolism in neurodegenerative diseases

PDF

Modulation of immune response in Systemic Lupus Erythematosus by synthetic lipoxin analogue, NAP1051

PDF

Development of dihydromyricetin (DHM) as a novel therapy for alcoholic liver disease (ALD) and alcohol use disorder (AUD)

PDF

NAP-2100 and analogues have improved anti-cancer effect in TNBC through SERCA2 depletion induced ER stress

PDF

Modulation of the renin-angiotensin system (RAS) for the treatment of COVID-19

PDF

Investigating sodium butyrate as a potential treatment for alcohol liver disease through the gut-liver axis

PDF

Averting dementia: renin-angiotensin system and angiogenic cells in cognitive decline

PDF

Investigating sodium butyrate effects on liver in a disrupted gut microbiome model of binge-like alcohol consumption in mice

PDF

Small molecule lipoxin analogue for prevention of retinal degeneration

PDF

Inhibition of tumor cell growth by mefloquine via multimechanistic effects involving increased cellular stress, inhibition of autophagy, and impairment of cellular energy metabolism

PDF

Perimenopausal transition increases blood brain permeability: implications for neurodegenerative diseases

PDF

Discovery of novel small molecules targeting cancer cell metabolism

Asset Metadata

Creator Asante, Isaac (author)

Core Title The role of altered folate one-carbon metabolism (FOCM) in the development of colorectal cancer

School School of Pharmacy

Degree Doctor of Philosophy

Degree Program Clinical and Experimental Therapeutics

Publication Date 06/21/2018

Defense Date 03/20/2017

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag colorectal cancer,folate,LCMS,metabolites,OAI-PMH Harvest,one-carbon metabolism,vitamin B

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Louie, Stan (committee chair), Conti, David (committee member), Davies, Daryl (committee member)

Creator Email asante@usc.edu,asante46@yahoo.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-390405

Unique identifier UC11263189

Identifier etd-AsanteIsaa-5436.pdf (filename),usctheses-c40-390405 (legacy record id)

Legacy Identifier etd-AsanteIsaa-5436.pdf

Dmrecord 390405

Document Type Dissertation

Rights Asante, Isaac

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA